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, 2010
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, 2010
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 1.3 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} @gol
168 -fdump-ada-spec@r{[}-slim@r{]}}
170 @item C Language Options
171 @xref{C Dialect Options,,Options Controlling C Dialect}.
172 @gccoptlist{-ansi -std=@var{standard} -fgnu89-inline @gol
173 -aux-info @var{filename} @gol
174 -fno-asm -fno-builtin -fno-builtin-@var{function} @gol
175 -fhosted -ffreestanding -fopenmp -fms-extensions @gol
176 -trigraphs -no-integrated-cpp -traditional -traditional-cpp @gol
177 -fallow-single-precision -fcond-mismatch -flax-vector-conversions @gol
178 -fsigned-bitfields -fsigned-char @gol
179 -funsigned-bitfields -funsigned-char}
181 @item C++ Language Options
182 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}.
183 @gccoptlist{-fabi-version=@var{n} -fno-access-control -fcheck-new @gol
184 -fconserve-space -ffriend-injection @gol
185 -fno-elide-constructors @gol
186 -fno-enforce-eh-specs @gol
187 -ffor-scope -fno-for-scope -fno-gnu-keywords @gol
188 -fno-implicit-templates @gol
189 -fno-implicit-inline-templates @gol
190 -fno-implement-inlines -fms-extensions @gol
191 -fno-nonansi-builtins -fnothrow-opt -fno-operator-names @gol
192 -fno-optional-diags -fpermissive @gol
193 -fno-pretty-templates @gol
194 -frepo -fno-rtti -fstats -ftemplate-depth=@var{n} @gol
195 -fno-threadsafe-statics -fuse-cxa-atexit -fno-weak -nostdinc++ @gol
196 -fno-default-inline -fvisibility-inlines-hidden @gol
197 -fvisibility-ms-compat @gol
198 -Wabi -Wconversion-null -Wctor-dtor-privacy @gol
199 -Wnoexcept -Wnon-virtual-dtor -Wreorder @gol
200 -Weffc++ -Wstrict-null-sentinel @gol
201 -Wno-non-template-friend -Wold-style-cast @gol
202 -Woverloaded-virtual -Wno-pmf-conversions @gol
205 @item Objective-C and Objective-C++ Language Options
206 @xref{Objective-C and Objective-C++ Dialect Options,,Options Controlling
207 Objective-C and Objective-C++ Dialects}.
208 @gccoptlist{-fconstant-string-class=@var{class-name} @gol
209 -fgnu-runtime -fnext-runtime @gol
210 -fno-nil-receivers @gol
211 -fobjc-call-cxx-cdtors @gol
212 -fobjc-direct-dispatch @gol
213 -fobjc-exceptions @gol
215 -freplace-objc-classes @gol
218 -Wassign-intercept @gol
219 -Wno-protocol -Wselector @gol
220 -Wstrict-selector-match @gol
221 -Wundeclared-selector}
223 @item Language Independent Options
224 @xref{Language Independent Options,,Options to Control Diagnostic Messages Formatting}.
225 @gccoptlist{-fmessage-length=@var{n} @gol
226 -fdiagnostics-show-location=@r{[}once@r{|}every-line@r{]} @gol
227 -fdiagnostics-show-option}
229 @item Warning Options
230 @xref{Warning Options,,Options to Request or Suppress Warnings}.
231 @gccoptlist{-fsyntax-only -pedantic -pedantic-errors @gol
232 -w -Wextra -Wall -Waddress -Waggregate-return -Warray-bounds @gol
233 -Wno-attributes -Wno-builtin-macro-redefined @gol
234 -Wc++-compat -Wc++0x-compat -Wcast-align -Wcast-qual @gol
235 -Wchar-subscripts -Wclobbered -Wcomment @gol
236 -Wconversion -Wcoverage-mismatch -Wcpp -Wno-deprecated @gol
237 -Wno-deprecated-declarations -Wdisabled-optimization @gol
238 -Wno-div-by-zero -Wempty-body -Wenum-compare -Wno-endif-labels @gol
239 -Werror -Werror=* @gol
240 -Wfatal-errors -Wfloat-equal -Wformat -Wformat=2 @gol
241 -Wno-format-contains-nul -Wno-format-extra-args -Wformat-nonliteral @gol
242 -Wformat-security -Wformat-y2k @gol
243 -Wframe-larger-than=@var{len} -Wjump-misses-init -Wignored-qualifiers @gol
244 -Wimplicit -Wimplicit-function-declaration -Wimplicit-int @gol
245 -Winit-self -Winline @gol
246 -Wno-int-to-pointer-cast -Wno-invalid-offsetof @gol
247 -Winvalid-pch -Wlarger-than=@var{len} -Wunsafe-loop-optimizations @gol
248 -Wlogical-op -Wlong-long @gol
249 -Wmain -Wmissing-braces -Wmissing-field-initializers @gol
250 -Wmissing-format-attribute -Wmissing-include-dirs @gol
252 -Wno-multichar -Wnonnull -Wno-overflow @gol
253 -Woverlength-strings -Wpacked -Wpacked-bitfield-compat -Wpadded @gol
254 -Wparentheses -Wpedantic-ms-format -Wno-pedantic-ms-format @gol
255 -Wpointer-arith -Wno-pointer-to-int-cast @gol
256 -Wredundant-decls @gol
257 -Wreturn-type -Wsequence-point -Wshadow @gol
258 -Wsign-compare -Wsign-conversion -Wstack-protector @gol
259 -Wstrict-aliasing -Wstrict-aliasing=n @gol
260 -Wstrict-overflow -Wstrict-overflow=@var{n} @gol
261 -Wsuggest-attribute=@r{[}pure@r{|}const@r{|}noreturn@r{]} @gol
262 -Wswitch -Wswitch-default -Wswitch-enum -Wsync-nand @gol
263 -Wsystem-headers -Wtrigraphs -Wtype-limits -Wundef -Wuninitialized @gol
264 -Wunknown-pragmas -Wno-pragmas @gol
265 -Wunsuffixed-float-constants -Wunused -Wunused-function @gol
266 -Wunused-label -Wunused-parameter -Wno-unused-result -Wunused-value -Wunused-variable @gol
267 -Wunused-but-set-parameter -Wunused-but-set-variable -Wvariadic-macros -Wvla @gol
268 -Wvolatile-register-var -Wwrite-strings}
270 @item C and Objective-C-only Warning Options
271 @gccoptlist{-Wbad-function-cast -Wmissing-declarations @gol
272 -Wmissing-parameter-type -Wmissing-prototypes -Wnested-externs @gol
273 -Wold-style-declaration -Wold-style-definition @gol
274 -Wstrict-prototypes -Wtraditional -Wtraditional-conversion @gol
275 -Wdeclaration-after-statement -Wpointer-sign}
277 @item Debugging Options
278 @xref{Debugging Options,,Options for Debugging Your Program or GCC}.
279 @gccoptlist{-d@var{letters} -dumpspecs -dumpmachine -dumpversion @gol
280 -fdbg-cnt-list -fdbg-cnt=@var{counter-value-list} @gol
281 -fdump-noaddr -fdump-unnumbered -fdump-unnumbered-links @gol
282 -fdump-translation-unit@r{[}-@var{n}@r{]} @gol
283 -fdump-class-hierarchy@r{[}-@var{n}@r{]} @gol
284 -fdump-ipa-all -fdump-ipa-cgraph -fdump-ipa-inline @gol
285 -fdump-statistics @gol
287 -fdump-tree-original@r{[}-@var{n}@r{]} @gol
288 -fdump-tree-optimized@r{[}-@var{n}@r{]} @gol
289 -fdump-tree-cfg -fdump-tree-vcg -fdump-tree-alias @gol
291 -fdump-tree-ssa@r{[}-@var{n}@r{]} -fdump-tree-pre@r{[}-@var{n}@r{]} @gol
292 -fdump-tree-ccp@r{[}-@var{n}@r{]} -fdump-tree-dce@r{[}-@var{n}@r{]} @gol
293 -fdump-tree-gimple@r{[}-raw@r{]} -fdump-tree-mudflap@r{[}-@var{n}@r{]} @gol
294 -fdump-tree-dom@r{[}-@var{n}@r{]} @gol
295 -fdump-tree-dse@r{[}-@var{n}@r{]} @gol
296 -fdump-tree-phiprop@r{[}-@var{n}@r{]} @gol
297 -fdump-tree-phiopt@r{[}-@var{n}@r{]} @gol
298 -fdump-tree-forwprop@r{[}-@var{n}@r{]} @gol
299 -fdump-tree-copyrename@r{[}-@var{n}@r{]} @gol
300 -fdump-tree-nrv -fdump-tree-vect @gol
301 -fdump-tree-sink @gol
302 -fdump-tree-sra@r{[}-@var{n}@r{]} @gol
303 -fdump-tree-forwprop@r{[}-@var{n}@r{]} @gol
304 -fdump-tree-fre@r{[}-@var{n}@r{]} @gol
305 -fdump-tree-vrp@r{[}-@var{n}@r{]} @gol
306 -ftree-vectorizer-verbose=@var{n} @gol
307 -fdump-tree-storeccp@r{[}-@var{n}@r{]} @gol
308 -fdump-final-insns=@var{file} @gol
309 -fcompare-debug@r{[}=@var{opts}@r{]} -fcompare-debug-second @gol
310 -feliminate-dwarf2-dups -feliminate-unused-debug-types @gol
311 -feliminate-unused-debug-symbols -femit-class-debug-always @gol
312 -fenable-icf-debug @gol
313 -fmem-report -fpre-ipa-mem-report -fpost-ipa-mem-report -fprofile-arcs @gol
314 -frandom-seed=@var{string} -fsched-verbose=@var{n} @gol
315 -fsel-sched-verbose -fsel-sched-dump-cfg -fsel-sched-pipelining-verbose @gol
316 -ftest-coverage -ftime-report -fvar-tracking @gol
317 -fvar-tracking-assignments -fvar-tracking-assignments-toggle @gol
318 -g -g@var{level} -gtoggle -gcoff -gdwarf-@var{version} @gol
319 -ggdb -gstabs -gstabs+ -gstrict-dwarf -gno-strict-dwarf @gol
320 -gvms -gxcoff -gxcoff+ @gol
321 -fno-merge-debug-strings -fno-dwarf2-cfi-asm @gol
322 -fdebug-prefix-map=@var{old}=@var{new} @gol
323 -femit-struct-debug-baseonly -femit-struct-debug-reduced @gol
324 -femit-struct-debug-detailed@r{[}=@var{spec-list}@r{]} @gol
325 -p -pg -print-file-name=@var{library} -print-libgcc-file-name @gol
326 -print-multi-directory -print-multi-lib -print-multi-os-directory @gol
327 -print-prog-name=@var{program} -print-search-dirs -Q @gol
328 -print-sysroot -print-sysroot-headers-suffix @gol
329 -save-temps -save-temps=cwd -save-temps=obj -time@r{[}=@var{file}@r{]}}
331 @item Optimization Options
332 @xref{Optimize Options,,Options that Control Optimization}.
334 -falign-functions[=@var{n}] -falign-jumps[=@var{n}] @gol
335 -falign-labels[=@var{n}] -falign-loops[=@var{n}] -fassociative-math @gol
336 -fauto-inc-dec -fbranch-probabilities -fbranch-target-load-optimize @gol
337 -fbranch-target-load-optimize2 -fbtr-bb-exclusive -fcaller-saves @gol
338 -fcheck-data-deps -fconserve-stack -fcprop-registers -fcrossjumping @gol
339 -fcse-follow-jumps -fcse-skip-blocks -fcx-fortran-rules -fcx-limited-range @gol
340 -fdata-sections -fdce -fdce @gol
341 -fdelayed-branch -fdelete-null-pointer-checks -fdse -fdse @gol
342 -fearly-inlining -fipa-sra -fexpensive-optimizations -ffast-math @gol
343 -ffinite-math-only -ffloat-store -fexcess-precision=@var{style} @gol
344 -fforward-propagate -ffunction-sections @gol
345 -fgcse -fgcse-after-reload -fgcse-las -fgcse-lm @gol
346 -fgcse-sm -fif-conversion -fif-conversion2 -findirect-inlining @gol
347 -finline-functions -finline-functions-called-once -finline-limit=@var{n} @gol
348 -finline-small-functions -fipa-cp -fipa-cp-clone -fipa-matrix-reorg -fipa-pta @gol
349 -fipa-profile -fipa-pure-const -fipa-reference -fipa-struct-reorg @gol
350 -fira-algorithm=@var{algorithm} @gol
351 -fira-region=@var{region} -fira-coalesce @gol
352 -fira-loop-pressure -fno-ira-share-save-slots @gol
353 -fno-ira-share-spill-slots -fira-verbose=@var{n} @gol
354 -fivopts -fkeep-inline-functions -fkeep-static-consts @gol
355 -floop-block -floop-interchange -floop-strip-mine -fgraphite-identity @gol
356 -floop-parallelize-all -flto -flto-compression-level -flto-report -fltrans @gol
357 -fltrans-output-list -fmerge-all-constants -fmerge-constants -fmodulo-sched @gol
358 -fmodulo-sched-allow-regmoves -fmove-loop-invariants -fmudflap @gol
359 -fmudflapir -fmudflapth -fno-branch-count-reg -fno-default-inline @gol
360 -fno-defer-pop -fno-function-cse -fno-guess-branch-probability @gol
361 -fno-inline -fno-math-errno -fno-peephole -fno-peephole2 @gol
362 -fno-sched-interblock -fno-sched-spec -fno-signed-zeros @gol
363 -fno-toplevel-reorder -fno-trapping-math -fno-zero-initialized-in-bss @gol
364 -fomit-frame-pointer -foptimize-register-move -foptimize-sibling-calls @gol
365 -fpartial-inlining -fpeel-loops -fpredictive-commoning -fprefetch-loop-arrays @gol
366 -fprofile-correction -fprofile-dir=@var{path} -fprofile-generate @gol
367 -fprofile-generate=@var{path} @gol
368 -fprofile-use -fprofile-use=@var{path} -fprofile-values @gol
369 -freciprocal-math -fregmove -frename-registers -freorder-blocks @gol
370 -freorder-blocks-and-partition -freorder-functions @gol
371 -frerun-cse-after-loop -freschedule-modulo-scheduled-loops @gol
372 -frounding-math -fsched2-use-superblocks -fsched-pressure @gol
373 -fsched-spec-load -fsched-spec-load-dangerous @gol
374 -fsched-stalled-insns-dep[=@var{n}] -fsched-stalled-insns[=@var{n}] @gol
375 -fsched-group-heuristic -fsched-critical-path-heuristic @gol
376 -fsched-spec-insn-heuristic -fsched-rank-heuristic @gol
377 -fsched-last-insn-heuristic -fsched-dep-count-heuristic @gol
378 -fschedule-insns -fschedule-insns2 -fsection-anchors @gol
379 -fselective-scheduling -fselective-scheduling2 @gol
380 -fsel-sched-pipelining -fsel-sched-pipelining-outer-loops @gol
381 -fsignaling-nans -fsingle-precision-constant -fsplit-ivs-in-unroller @gol
382 -fsplit-wide-types -fstack-protector -fstack-protector-all @gol
383 -fstrict-aliasing -fstrict-overflow -fthread-jumps -ftracer @gol
384 -ftree-builtin-call-dce -ftree-ccp -ftree-ch -ftree-copy-prop @gol
385 -ftree-copyrename -ftree-dce @gol
386 -ftree-dominator-opts -ftree-dse -ftree-forwprop -ftree-fre -ftree-loop-im @gol
387 -ftree-phiprop -ftree-loop-distribution @gol
388 -ftree-loop-ivcanon -ftree-loop-linear -ftree-loop-optimize @gol
389 -ftree-parallelize-loops=@var{n} -ftree-pre -ftree-pta -ftree-reassoc @gol
390 -ftree-sink -ftree-sra -ftree-switch-conversion @gol
391 -ftree-ter -ftree-vect-loop-version -ftree-vectorize -ftree-vrp @gol
392 -funit-at-a-time -funroll-all-loops -funroll-loops @gol
393 -funsafe-loop-optimizations -funsafe-math-optimizations -funswitch-loops @gol
394 -fvariable-expansion-in-unroller -fvect-cost-model -fvpt -fweb @gol
395 -fwhole-program -fwhopr[=@var{n}] -fwpa -fuse-linker-plugin @gol
396 --param @var{name}=@var{value}
397 -O -O0 -O1 -O2 -O3 -Os -Ofast}
399 @item Preprocessor Options
400 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
401 @gccoptlist{-A@var{question}=@var{answer} @gol
402 -A-@var{question}@r{[}=@var{answer}@r{]} @gol
403 -C -dD -dI -dM -dN @gol
404 -D@var{macro}@r{[}=@var{defn}@r{]} -E -H @gol
405 -idirafter @var{dir} @gol
406 -include @var{file} -imacros @var{file} @gol
407 -iprefix @var{file} -iwithprefix @var{dir} @gol
408 -iwithprefixbefore @var{dir} -isystem @var{dir} @gol
409 -imultilib @var{dir} -isysroot @var{dir} @gol
410 -M -MM -MF -MG -MP -MQ -MT -nostdinc @gol
411 -P -fworking-directory -remap @gol
412 -trigraphs -undef -U@var{macro} -Wp,@var{option} @gol
413 -Xpreprocessor @var{option}}
415 @item Assembler Option
416 @xref{Assembler Options,,Passing Options to the Assembler}.
417 @gccoptlist{-Wa,@var{option} -Xassembler @var{option}}
420 @xref{Link Options,,Options for Linking}.
421 @gccoptlist{@var{object-file-name} -l@var{library} @gol
422 -nostartfiles -nodefaultlibs -nostdlib -pie -rdynamic @gol
423 -s -static -static-libgcc -static-libstdc++ -shared @gol
424 -shared-libgcc -symbolic @gol
425 -T @var{script} -Wl,@var{option} -Xlinker @var{option} @gol
428 @item Directory Options
429 @xref{Directory Options,,Options for Directory Search}.
430 @gccoptlist{-B@var{prefix} -I@var{dir} -iplugindir=@var{dir}}
431 -iquote@var{dir} -L@var{dir} -specs=@var{file} -I-
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} -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{-mglibc -muclibc -mbionic -mandroid @gol
564 -tno-android-cc -tno-android-ld}
566 @emph{H8/300 Options}
567 @gccoptlist{-mrelax -mh -ms -mn -mint32 -malign-300}
570 @gccoptlist{-march=@var{architecture-type} @gol
571 -mbig-switch -mdisable-fpregs -mdisable-indexing @gol
572 -mfast-indirect-calls -mgas -mgnu-ld -mhp-ld @gol
573 -mfixed-range=@var{register-range} @gol
574 -mjump-in-delay -mlinker-opt -mlong-calls @gol
575 -mlong-load-store -mno-big-switch -mno-disable-fpregs @gol
576 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
577 -mno-jump-in-delay -mno-long-load-store @gol
578 -mno-portable-runtime -mno-soft-float @gol
579 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
580 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
581 -mschedule=@var{cpu-type} -mspace-regs -msio -mwsio @gol
582 -munix=@var{unix-std} -nolibdld -static -threads}
584 @emph{i386 and x86-64 Options}
585 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
586 -mfpmath=@var{unit} @gol
587 -masm=@var{dialect} -mno-fancy-math-387 @gol
588 -mno-fp-ret-in-387 -msoft-float @gol
589 -mno-wide-multiply -mrtd -malign-double @gol
590 -mpreferred-stack-boundary=@var{num}
591 -mincoming-stack-boundary=@var{num}
592 -mcld -mcx16 -msahf -mmovbe -mcrc32 -mrecip @gol
593 -mmmx -msse -msse2 -msse3 -mssse3 -msse4.1 -msse4.2 -msse4 -mavx @gol
594 -maes -mpclmul -mfused-madd @gol
595 -msse4a -m3dnow -mpopcnt -mabm -mfma4 -mxop -mlwp @gol
596 -mthreads -mno-align-stringops -minline-all-stringops @gol
597 -minline-stringops-dynamically -mstringop-strategy=@var{alg} @gol
598 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
599 -m96bit-long-double -mregparm=@var{num} -msseregparm @gol
600 -mveclibabi=@var{type} -mpc32 -mpc64 -mpc80 -mstackrealign @gol
601 -momit-leaf-frame-pointer -mno-red-zone -mno-tls-direct-seg-refs @gol
602 -mcmodel=@var{code-model} -mabi=@var{name} @gol
603 -m32 -m64 -mlarge-data-threshold=@var{num} @gol
607 @gccoptlist{-mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
608 -mvolatile-asm-stop -mregister-names -msdata -mno-sdata @gol
609 -mconstant-gp -mauto-pic -mfused-madd @gol
610 -minline-float-divide-min-latency @gol
611 -minline-float-divide-max-throughput @gol
612 -mno-inline-float-divide @gol
613 -minline-int-divide-min-latency @gol
614 -minline-int-divide-max-throughput @gol
615 -mno-inline-int-divide @gol
616 -minline-sqrt-min-latency -minline-sqrt-max-throughput @gol
617 -mno-inline-sqrt @gol
618 -mdwarf2-asm -mearly-stop-bits @gol
619 -mfixed-range=@var{register-range} -mtls-size=@var{tls-size} @gol
620 -mtune=@var{cpu-type} -milp32 -mlp64 @gol
621 -msched-br-data-spec -msched-ar-data-spec -msched-control-spec @gol
622 -msched-br-in-data-spec -msched-ar-in-data-spec -msched-in-control-spec @gol
623 -msched-spec-ldc -msched-spec-control-ldc @gol
624 -msched-prefer-non-data-spec-insns -msched-prefer-non-control-spec-insns @gol
625 -msched-stop-bits-after-every-cycle -msched-count-spec-in-critical-path @gol
626 -msel-sched-dont-check-control-spec -msched-fp-mem-deps-zero-cost @gol
627 -msched-max-memory-insns-hard-limit -msched-max-memory-insns=@var{max-insns}}
629 @emph{IA-64/VMS Options}
630 @gccoptlist{-mvms-return-codes -mdebug-main=@var{prefix} -mmalloc64}
633 @gccoptlist{-mbarrel-shift-enabled -mdivide-enabled -mmultiply-enabled @gol
634 -msign-extend-enabled -muser-enabled}
636 @emph{M32R/D Options}
637 @gccoptlist{-m32r2 -m32rx -m32r @gol
639 -malign-loops -mno-align-loops @gol
640 -missue-rate=@var{number} @gol
641 -mbranch-cost=@var{number} @gol
642 -mmodel=@var{code-size-model-type} @gol
643 -msdata=@var{sdata-type} @gol
644 -mno-flush-func -mflush-func=@var{name} @gol
645 -mno-flush-trap -mflush-trap=@var{number} @gol
649 @gccoptlist{-mcpu=@var{cpu} -msim -memregs=@var{number}}
651 @emph{M680x0 Options}
652 @gccoptlist{-march=@var{arch} -mcpu=@var{cpu} -mtune=@var{tune}
653 -m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
654 -m68060 -mcpu32 -m5200 -m5206e -m528x -m5307 -m5407 @gol
655 -mcfv4e -mbitfield -mno-bitfield -mc68000 -mc68020 @gol
656 -mnobitfield -mrtd -mno-rtd -mdiv -mno-div -mshort @gol
657 -mno-short -mhard-float -m68881 -msoft-float -mpcrel @gol
658 -malign-int -mstrict-align -msep-data -mno-sep-data @gol
659 -mshared-library-id=n -mid-shared-library -mno-id-shared-library @gol
662 @emph{M68hc1x Options}
663 @gccoptlist{-m6811 -m6812 -m68hc11 -m68hc12 -m68hcs12 @gol
664 -mauto-incdec -minmax -mlong-calls -mshort @gol
665 -msoft-reg-count=@var{count}}
668 @gccoptlist{-mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
669 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
670 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
671 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
672 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
675 @gccoptlist{-mabsdiff -mall-opts -maverage -mbased=@var{n} -mbitops @gol
676 -mc=@var{n} -mclip -mconfig=@var{name} -mcop -mcop32 -mcop64 -mivc2 @gol
677 -mdc -mdiv -meb -mel -mio-volatile -ml -mleadz -mm -mminmax @gol
678 -mmult -mno-opts -mrepeat -ms -msatur -msdram -msim -msimnovec -mtf @gol
682 @gccoptlist{-EL -EB -march=@var{arch} -mtune=@var{arch} @gol
683 -mips1 -mips2 -mips3 -mips4 -mips32 -mips32r2 @gol
684 -mips64 -mips64r2 @gol
685 -mips16 -mno-mips16 -mflip-mips16 @gol
686 -minterlink-mips16 -mno-interlink-mips16 @gol
687 -mabi=@var{abi} -mabicalls -mno-abicalls @gol
688 -mshared -mno-shared -mplt -mno-plt -mxgot -mno-xgot @gol
689 -mgp32 -mgp64 -mfp32 -mfp64 -mhard-float -msoft-float @gol
690 -msingle-float -mdouble-float -mdsp -mno-dsp -mdspr2 -mno-dspr2 @gol
691 -mfpu=@var{fpu-type} @gol
692 -msmartmips -mno-smartmips @gol
693 -mpaired-single -mno-paired-single -mdmx -mno-mdmx @gol
694 -mips3d -mno-mips3d -mmt -mno-mt -mllsc -mno-llsc @gol
695 -mlong64 -mlong32 -msym32 -mno-sym32 @gol
696 -G@var{num} -mlocal-sdata -mno-local-sdata @gol
697 -mextern-sdata -mno-extern-sdata -mgpopt -mno-gopt @gol
698 -membedded-data -mno-embedded-data @gol
699 -muninit-const-in-rodata -mno-uninit-const-in-rodata @gol
700 -mcode-readable=@var{setting} @gol
701 -msplit-addresses -mno-split-addresses @gol
702 -mexplicit-relocs -mno-explicit-relocs @gol
703 -mcheck-zero-division -mno-check-zero-division @gol
704 -mdivide-traps -mdivide-breaks @gol
705 -mmemcpy -mno-memcpy -mlong-calls -mno-long-calls @gol
706 -mmad -mno-mad -mfused-madd -mno-fused-madd -nocpp @gol
707 -mfix-r4000 -mno-fix-r4000 -mfix-r4400 -mno-fix-r4400 @gol
708 -mfix-r10000 -mno-fix-r10000 -mfix-vr4120 -mno-fix-vr4120 @gol
709 -mfix-vr4130 -mno-fix-vr4130 -mfix-sb1 -mno-fix-sb1 @gol
710 -mflush-func=@var{func} -mno-flush-func @gol
711 -mbranch-cost=@var{num} -mbranch-likely -mno-branch-likely @gol
712 -mfp-exceptions -mno-fp-exceptions @gol
713 -mvr4130-align -mno-vr4130-align -msynci -mno-synci @gol
714 -mrelax-pic-calls -mno-relax-pic-calls -mmcount-ra-address}
717 @gccoptlist{-mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
718 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
719 -melf -mbranch-predict -mno-branch-predict -mbase-addresses @gol
720 -mno-base-addresses -msingle-exit -mno-single-exit}
722 @emph{MN10300 Options}
723 @gccoptlist{-mmult-bug -mno-mult-bug @gol
724 -mam33 -mno-am33 @gol
725 -mam33-2 -mno-am33-2 @gol
726 -mreturn-pointer-on-d0 @gol
729 @emph{PDP-11 Options}
730 @gccoptlist{-mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol
731 -mbcopy -mbcopy-builtin -mint32 -mno-int16 @gol
732 -mint16 -mno-int32 -mfloat32 -mno-float64 @gol
733 -mfloat64 -mno-float32 -mabshi -mno-abshi @gol
734 -mbranch-expensive -mbranch-cheap @gol
735 -msplit -mno-split -munix-asm -mdec-asm}
737 @emph{picoChip Options}
738 @gccoptlist{-mae=@var{ae_type} -mvliw-lookahead=@var{N}
739 -msymbol-as-address -mno-inefficient-warnings}
741 @emph{PowerPC Options}
742 See RS/6000 and PowerPC Options.
744 @emph{RS/6000 and PowerPC Options}
745 @gccoptlist{-mcpu=@var{cpu-type} @gol
746 -mtune=@var{cpu-type} @gol
747 -mcmodel=@var{code-model} @gol
748 -mpower -mno-power -mpower2 -mno-power2 @gol
749 -mpowerpc -mpowerpc64 -mno-powerpc @gol
750 -maltivec -mno-altivec @gol
751 -mpowerpc-gpopt -mno-powerpc-gpopt @gol
752 -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol
753 -mmfcrf -mno-mfcrf -mpopcntb -mno-popcntb -mpopcntd -mno-popcntd @gol
754 -mfprnd -mno-fprnd @gol
755 -mcmpb -mno-cmpb -mmfpgpr -mno-mfpgpr -mhard-dfp -mno-hard-dfp @gol
756 -mnew-mnemonics -mold-mnemonics @gol
757 -mfull-toc -mminimal-toc -mno-fp-in-toc -mno-sum-in-toc @gol
758 -m64 -m32 -mxl-compat -mno-xl-compat -mpe @gol
759 -malign-power -malign-natural @gol
760 -msoft-float -mhard-float -mmultiple -mno-multiple @gol
761 -msingle-float -mdouble-float -msimple-fpu @gol
762 -mstring -mno-string -mupdate -mno-update @gol
763 -mavoid-indexed-addresses -mno-avoid-indexed-addresses @gol
764 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol
765 -mstrict-align -mno-strict-align -mrelocatable @gol
766 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol
767 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol
768 -mdynamic-no-pic -maltivec -mswdiv @gol
769 -mprioritize-restricted-insns=@var{priority} @gol
770 -msched-costly-dep=@var{dependence_type} @gol
771 -minsert-sched-nops=@var{scheme} @gol
772 -mcall-sysv -mcall-netbsd @gol
773 -maix-struct-return -msvr4-struct-return @gol
774 -mabi=@var{abi-type} -msecure-plt -mbss-plt @gol
775 -misel -mno-isel @gol
776 -misel=yes -misel=no @gol
778 -mspe=yes -mspe=no @gol
780 -mgen-cell-microcode -mwarn-cell-microcode @gol
781 -mvrsave -mno-vrsave @gol
782 -mmulhw -mno-mulhw @gol
783 -mdlmzb -mno-dlmzb @gol
784 -mfloat-gprs=yes -mfloat-gprs=no -mfloat-gprs=single -mfloat-gprs=double @gol
785 -mprototype -mno-prototype @gol
786 -msim -mmvme -mads -myellowknife -memb -msdata @gol
787 -msdata=@var{opt} -mvxworks -G @var{num} -pthread @gol
788 -mrecip -mrecip=@var{opt} -mno-recip -mrecip-precision -mno-recip-precision}
791 @gccoptlist{-m64bit-doubles -m32bit-doubles -fpu -nofpu@gol
793 -mbig-endian-data -mlittle-endian-data @gol
796 -mas100-syntax -mno-as100-syntax@gol
798 -mmax-constant-size=@gol
800 -msave-acc-in-interrupts}
802 @emph{S/390 and zSeries Options}
803 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
804 -mhard-float -msoft-float -mhard-dfp -mno-hard-dfp @gol
805 -mlong-double-64 -mlong-double-128 @gol
806 -mbackchain -mno-backchain -mpacked-stack -mno-packed-stack @gol
807 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
808 -m64 -m31 -mdebug -mno-debug -mesa -mzarch @gol
809 -mtpf-trace -mno-tpf-trace -mfused-madd -mno-fused-madd @gol
810 -mwarn-framesize -mwarn-dynamicstack -mstack-size -mstack-guard}
813 @gccoptlist{-meb -mel @gol
817 -mscore5 -mscore5u -mscore7 -mscore7d}
820 @gccoptlist{-m1 -m2 -m2e @gol
821 -m2a-nofpu -m2a-single-only -m2a-single -m2a @gol
823 -m4-nofpu -m4-single-only -m4-single -m4 @gol
824 -m4a-nofpu -m4a-single-only -m4a-single -m4a -m4al @gol
825 -m5-64media -m5-64media-nofpu @gol
826 -m5-32media -m5-32media-nofpu @gol
827 -m5-compact -m5-compact-nofpu @gol
828 -mb -ml -mdalign -mrelax @gol
829 -mbigtable -mfmovd -mhitachi -mrenesas -mno-renesas -mnomacsave @gol
830 -mieee -mbitops -misize -minline-ic_invalidate -mpadstruct -mspace @gol
831 -mprefergot -musermode -multcost=@var{number} -mdiv=@var{strategy} @gol
832 -mdivsi3_libfunc=@var{name} -mfixed-range=@var{register-range} @gol
833 -madjust-unroll -mindexed-addressing -mgettrcost=@var{number} -mpt-fixed @gol
834 -maccumulate-outgoing-args -minvalid-symbols}
837 @gccoptlist{-mcpu=@var{cpu-type} @gol
838 -mtune=@var{cpu-type} @gol
839 -mcmodel=@var{code-model} @gol
840 -m32 -m64 -mapp-regs -mno-app-regs @gol
841 -mfaster-structs -mno-faster-structs @gol
842 -mfpu -mno-fpu -mhard-float -msoft-float @gol
843 -mhard-quad-float -msoft-quad-float @gol
844 -mimpure-text -mno-impure-text -mlittle-endian @gol
845 -mstack-bias -mno-stack-bias @gol
846 -munaligned-doubles -mno-unaligned-doubles @gol
847 -mv8plus -mno-v8plus -mvis -mno-vis
848 -threads -pthreads -pthread}
851 @gccoptlist{-mwarn-reloc -merror-reloc @gol
852 -msafe-dma -munsafe-dma @gol
854 -msmall-mem -mlarge-mem -mstdmain @gol
855 -mfixed-range=@var{register-range} @gol
857 -maddress-space-conversion -mno-address-space-conversion @gol
858 -mcache-size=@var{cache-size} @gol
859 -matomic-updates -mno-atomic-updates}
861 @emph{System V Options}
862 @gccoptlist{-Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
865 @gccoptlist{-mlong-calls -mno-long-calls -mep -mno-ep @gol
866 -mprolog-function -mno-prolog-function -mspace @gol
867 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
868 -mapp-regs -mno-app-regs @gol
869 -mdisable-callt -mno-disable-callt @gol
875 @gccoptlist{-mg -mgnu -munix}
877 @emph{VxWorks Options}
878 @gccoptlist{-mrtp -non-static -Bstatic -Bdynamic @gol
879 -Xbind-lazy -Xbind-now}
881 @emph{x86-64 Options}
882 See i386 and x86-64 Options.
884 @emph{i386 and x86-64 Windows Options}
885 @gccoptlist{-mconsole -mcygwin -mno-cygwin -mdll
886 -mnop-fun-dllimport -mthread -municode -mwin32 -mwindows
887 -fno-set-stack-executable}
889 @emph{Xstormy16 Options}
892 @emph{Xtensa Options}
893 @gccoptlist{-mconst16 -mno-const16 @gol
894 -mfused-madd -mno-fused-madd @gol
896 -mserialize-volatile -mno-serialize-volatile @gol
897 -mtext-section-literals -mno-text-section-literals @gol
898 -mtarget-align -mno-target-align @gol
899 -mlongcalls -mno-longcalls}
901 @emph{zSeries Options}
902 See S/390 and zSeries Options.
904 @item Code Generation Options
905 @xref{Code Gen Options,,Options for Code Generation Conventions}.
906 @gccoptlist{-fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
907 -ffixed-@var{reg} -fexceptions @gol
908 -fnon-call-exceptions -funwind-tables @gol
909 -fasynchronous-unwind-tables @gol
910 -finhibit-size-directive -finstrument-functions @gol
911 -finstrument-functions-exclude-function-list=@var{sym},@var{sym},@dots{} @gol
912 -finstrument-functions-exclude-file-list=@var{file},@var{file},@dots{} @gol
913 -fno-common -fno-ident @gol
914 -fpcc-struct-return -fpic -fPIC -fpie -fPIE @gol
915 -fno-jump-tables @gol
916 -frecord-gcc-switches @gol
917 -freg-struct-return -fshort-enums @gol
918 -fshort-double -fshort-wchar @gol
919 -fverbose-asm -fpack-struct[=@var{n}] -fstack-check @gol
920 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
921 -fno-stack-limit @gol
922 -fleading-underscore -ftls-model=@var{model} @gol
923 -ftrapv -fwrapv -fbounds-check @gol
928 * Overall Options:: Controlling the kind of output:
929 an executable, object files, assembler files,
930 or preprocessed source.
931 * C Dialect Options:: Controlling the variant of C language compiled.
932 * C++ Dialect Options:: Variations on C++.
933 * Objective-C and Objective-C++ Dialect Options:: Variations on Objective-C
935 * Language Independent Options:: Controlling how diagnostics should be
937 * Warning Options:: How picky should the compiler be?
938 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
939 * Optimize Options:: How much optimization?
940 * Preprocessor Options:: Controlling header files and macro definitions.
941 Also, getting dependency information for Make.
942 * Assembler Options:: Passing options to the assembler.
943 * Link Options:: Specifying libraries and so on.
944 * Directory Options:: Where to find header files and libraries.
945 Where to find the compiler executable files.
946 * Spec Files:: How to pass switches to sub-processes.
947 * Target Options:: Running a cross-compiler, or an old version of GCC.
950 @node Overall Options
951 @section Options Controlling the Kind of Output
953 Compilation can involve up to four stages: preprocessing, compilation
954 proper, assembly and linking, always in that order. GCC is capable of
955 preprocessing and compiling several files either into several
956 assembler input files, or into one assembler input file; then each
957 assembler input file produces an object file, and linking combines all
958 the object files (those newly compiled, and those specified as input)
959 into an executable file.
961 @cindex file name suffix
962 For any given input file, the file name suffix determines what kind of
967 C source code which must be preprocessed.
970 C source code which should not be preprocessed.
973 C++ source code which should not be preprocessed.
976 Objective-C source code. Note that you must link with the @file{libobjc}
977 library to make an Objective-C program work.
980 Objective-C source code which should not be preprocessed.
984 Objective-C++ source code. Note that you must link with the @file{libobjc}
985 library to make an Objective-C++ program work. Note that @samp{.M} refers
986 to a literal capital M@.
989 Objective-C++ source code which should not be preprocessed.
992 C, C++, Objective-C or Objective-C++ header file to be turned into a
993 precompiled header (default), or C, C++ header file to be turned into an
994 Ada spec (via the @option{-fdump-ada-spec} switch).
998 @itemx @var{file}.cxx
999 @itemx @var{file}.cpp
1000 @itemx @var{file}.CPP
1001 @itemx @var{file}.c++
1003 C++ source code which must be preprocessed. Note that in @samp{.cxx},
1004 the last two letters must both be literally @samp{x}. Likewise,
1005 @samp{.C} refers to a literal capital C@.
1009 Objective-C++ source code which must be preprocessed.
1011 @item @var{file}.mii
1012 Objective-C++ source code which should not be preprocessed.
1016 @itemx @var{file}.hp
1017 @itemx @var{file}.hxx
1018 @itemx @var{file}.hpp
1019 @itemx @var{file}.HPP
1020 @itemx @var{file}.h++
1021 @itemx @var{file}.tcc
1022 C++ header file to be turned into a precompiled header or Ada spec.
1025 @itemx @var{file}.for
1026 @itemx @var{file}.ftn
1027 Fixed form Fortran source code which should not be preprocessed.
1030 @itemx @var{file}.FOR
1031 @itemx @var{file}.fpp
1032 @itemx @var{file}.FPP
1033 @itemx @var{file}.FTN
1034 Fixed form Fortran source code which must be preprocessed (with the traditional
1037 @item @var{file}.f90
1038 @itemx @var{file}.f95
1039 @itemx @var{file}.f03
1040 @itemx @var{file}.f08
1041 Free form Fortran source code which should not be preprocessed.
1043 @item @var{file}.F90
1044 @itemx @var{file}.F95
1045 @itemx @var{file}.F03
1046 @itemx @var{file}.F08
1047 Free form Fortran source code which must be preprocessed (with the
1048 traditional preprocessor).
1050 @c FIXME: Descriptions of Java file types.
1056 @item @var{file}.ads
1057 Ada source code file which contains a library unit declaration (a
1058 declaration of a package, subprogram, or generic, or a generic
1059 instantiation), or a library unit renaming declaration (a package,
1060 generic, or subprogram renaming declaration). Such files are also
1063 @item @var{file}.adb
1064 Ada source code file containing a library unit body (a subprogram or
1065 package body). Such files are also called @dfn{bodies}.
1067 @c GCC also knows about some suffixes for languages not yet included:
1078 @itemx @var{file}.sx
1079 Assembler code which must be preprocessed.
1082 An object file to be fed straight into linking.
1083 Any file name with no recognized suffix is treated this way.
1087 You can specify the input language explicitly with the @option{-x} option:
1090 @item -x @var{language}
1091 Specify explicitly the @var{language} for the following input files
1092 (rather than letting the compiler choose a default based on the file
1093 name suffix). This option applies to all following input files until
1094 the next @option{-x} option. Possible values for @var{language} are:
1096 c c-header c-cpp-output
1097 c++ c++-header c++-cpp-output
1098 objective-c objective-c-header objective-c-cpp-output
1099 objective-c++ objective-c++-header objective-c++-cpp-output
1100 assembler assembler-with-cpp
1102 f77 f77-cpp-input f95 f95-cpp-input
1107 Turn off any specification of a language, so that subsequent files are
1108 handled according to their file name suffixes (as they are if @option{-x}
1109 has not been used at all).
1111 @item -pass-exit-codes
1112 @opindex pass-exit-codes
1113 Normally the @command{gcc} program will exit with the code of 1 if any
1114 phase of the compiler returns a non-success return code. If you specify
1115 @option{-pass-exit-codes}, the @command{gcc} program will instead return with
1116 numerically highest error produced by any phase that returned an error
1117 indication. The C, C++, and Fortran frontends return 4, if an internal
1118 compiler error is encountered.
1121 If you only want some of the stages of compilation, you can use
1122 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
1123 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
1124 @command{gcc} is to stop. Note that some combinations (for example,
1125 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
1130 Compile or assemble the source files, but do not link. The linking
1131 stage simply is not done. The ultimate output is in the form of an
1132 object file for each source file.
1134 By default, the object file name for a source file is made by replacing
1135 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
1137 Unrecognized input files, not requiring compilation or assembly, are
1142 Stop after the stage of compilation proper; do not assemble. The output
1143 is in the form of an assembler code file for each non-assembler input
1146 By default, the assembler file name for a source file is made by
1147 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
1149 Input files that don't require compilation are ignored.
1153 Stop after the preprocessing stage; do not run the compiler proper. The
1154 output is in the form of preprocessed source code, which is sent to the
1157 Input files which don't require preprocessing are ignored.
1159 @cindex output file option
1162 Place output in file @var{file}. This applies regardless to whatever
1163 sort of output is being produced, whether it be an executable file,
1164 an object file, an assembler file or preprocessed C code.
1166 If @option{-o} is not specified, the default is to put an executable
1167 file in @file{a.out}, the object file for
1168 @file{@var{source}.@var{suffix}} in @file{@var{source}.o}, its
1169 assembler file in @file{@var{source}.s}, a precompiled header file in
1170 @file{@var{source}.@var{suffix}.gch}, and all preprocessed C source on
1175 Print (on standard error output) the commands executed to run the stages
1176 of compilation. Also print the version number of the compiler driver
1177 program and of the preprocessor and the compiler proper.
1181 Like @option{-v} except the commands are not executed and arguments
1182 are quoted unless they contain only alphanumeric characters or @code{./-_}.
1183 This is useful for shell scripts to capture the driver-generated command lines.
1187 Use pipes rather than temporary files for communication between the
1188 various stages of compilation. This fails to work on some systems where
1189 the assembler is unable to read from a pipe; but the GNU assembler has
1194 If you are compiling multiple source files, this option tells the driver
1195 to pass all the source files to the compiler at once (for those
1196 languages for which the compiler can handle this). This will allow
1197 intermodule analysis (IMA) to be performed by the compiler. Currently the only
1198 language for which this is supported is C@. If you pass source files for
1199 multiple languages to the driver, using this option, the driver will invoke
1200 the compiler(s) that support IMA once each, passing each compiler all the
1201 source files appropriate for it. For those languages that do not support
1202 IMA this option will be ignored, and the compiler will be invoked once for
1203 each source file in that language. If you use this option in conjunction
1204 with @option{-save-temps}, the compiler will generate multiple
1206 (one for each source file), but only one (combined) @file{.o} or
1211 Print (on the standard output) a description of the command line options
1212 understood by @command{gcc}. If the @option{-v} option is also specified
1213 then @option{--help} will also be passed on to the various processes
1214 invoked by @command{gcc}, so that they can display the command line options
1215 they accept. If the @option{-Wextra} option has also been specified
1216 (prior to the @option{--help} option), then command line options which
1217 have no documentation associated with them will also be displayed.
1220 @opindex target-help
1221 Print (on the standard output) a description of target-specific command
1222 line options for each tool. For some targets extra target-specific
1223 information may also be printed.
1225 @item --help=@{@var{class}@r{|[}^@r{]}@var{qualifier}@}@r{[},@dots{}@r{]}
1226 Print (on the standard output) a description of the command line
1227 options understood by the compiler that fit into all specified classes
1228 and qualifiers. These are the supported classes:
1231 @item @samp{optimizers}
1232 This will display all of the optimization options supported by the
1235 @item @samp{warnings}
1236 This will display all of the options controlling warning messages
1237 produced by the compiler.
1240 This will display target-specific options. Unlike the
1241 @option{--target-help} option however, target-specific options of the
1242 linker and assembler will not be displayed. This is because those
1243 tools do not currently support the extended @option{--help=} syntax.
1246 This will display the values recognized by the @option{--param}
1249 @item @var{language}
1250 This will display the options supported for @var{language}, where
1251 @var{language} is the name of one of the languages supported in this
1255 This will display the options that are common to all languages.
1258 These are the supported qualifiers:
1261 @item @samp{undocumented}
1262 Display only those options which are undocumented.
1265 Display options which take an argument that appears after an equal
1266 sign in the same continuous piece of text, such as:
1267 @samp{--help=target}.
1269 @item @samp{separate}
1270 Display options which take an argument that appears as a separate word
1271 following the original option, such as: @samp{-o output-file}.
1274 Thus for example to display all the undocumented target-specific
1275 switches supported by the compiler the following can be used:
1278 --help=target,undocumented
1281 The sense of a qualifier can be inverted by prefixing it with the
1282 @samp{^} character, so for example to display all binary warning
1283 options (i.e., ones that are either on or off and that do not take an
1284 argument), which have a description the following can be used:
1287 --help=warnings,^joined,^undocumented
1290 The argument to @option{--help=} should not consist solely of inverted
1293 Combining several classes is possible, although this usually
1294 restricts the output by so much that there is nothing to display. One
1295 case where it does work however is when one of the classes is
1296 @var{target}. So for example to display all the target-specific
1297 optimization options the following can be used:
1300 --help=target,optimizers
1303 The @option{--help=} option can be repeated on the command line. Each
1304 successive use will display its requested class of options, skipping
1305 those that have already been displayed.
1307 If the @option{-Q} option appears on the command line before the
1308 @option{--help=} option, then the descriptive text displayed by
1309 @option{--help=} is changed. Instead of describing the displayed
1310 options, an indication is given as to whether the option is enabled,
1311 disabled or set to a specific value (assuming that the compiler
1312 knows this at the point where the @option{--help=} option is used).
1314 Here is a truncated example from the ARM port of @command{gcc}:
1317 % gcc -Q -mabi=2 --help=target -c
1318 The following options are target specific:
1320 -mabort-on-noreturn [disabled]
1324 The output is sensitive to the effects of previous command line
1325 options, so for example it is possible to find out which optimizations
1326 are enabled at @option{-O2} by using:
1329 -Q -O2 --help=optimizers
1332 Alternatively you can discover which binary optimizations are enabled
1333 by @option{-O3} by using:
1336 gcc -c -Q -O3 --help=optimizers > /tmp/O3-opts
1337 gcc -c -Q -O2 --help=optimizers > /tmp/O2-opts
1338 diff /tmp/O2-opts /tmp/O3-opts | grep enabled
1341 @item -no-canonical-prefixes
1342 @opindex no-canonical-prefixes
1343 Do not expand any symbolic links, resolve references to @samp{/../}
1344 or @samp{/./}, or make the path absolute when generating a relative
1349 Display the version number and copyrights of the invoked GCC@.
1353 Invoke all subcommands under a wrapper program. It takes a single
1354 comma separated list as an argument, which will be used to invoke
1358 gcc -c t.c -wrapper gdb,--args
1361 This will invoke all subprograms of gcc under "gdb --args",
1362 thus cc1 invocation will be "gdb --args cc1 ...".
1364 @item -fplugin=@var{name}.so
1365 Load the plugin code in file @var{name}.so, assumed to be a
1366 shared object to be dlopen'd by the compiler. The base name of
1367 the shared object file is used to identify the plugin for the
1368 purposes of argument parsing (See
1369 @option{-fplugin-arg-@var{name}-@var{key}=@var{value}} below).
1370 Each plugin should define the callback functions specified in the
1373 @item -fplugin-arg-@var{name}-@var{key}=@var{value}
1374 Define an argument called @var{key} with a value of @var{value}
1375 for the plugin called @var{name}.
1377 @item -fdump-ada-spec@r{[}-slim@r{]}
1378 For C and C++ source and include files, generate corresponding Ada
1379 specs. @xref{Generating Ada Bindings for C and C++ headers,,, gnat_ugn,
1380 GNAT User's Guide}, which provides detailed documentation on this feature.
1382 @include @value{srcdir}/../libiberty/at-file.texi
1386 @section Compiling C++ Programs
1388 @cindex suffixes for C++ source
1389 @cindex C++ source file suffixes
1390 C++ source files conventionally use one of the suffixes @samp{.C},
1391 @samp{.cc}, @samp{.cpp}, @samp{.CPP}, @samp{.c++}, @samp{.cp}, or
1392 @samp{.cxx}; C++ header files often use @samp{.hh}, @samp{.hpp},
1393 @samp{.H}, or (for shared template code) @samp{.tcc}; and
1394 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
1395 files with these names and compiles them as C++ programs even if you
1396 call the compiler the same way as for compiling C programs (usually
1397 with the name @command{gcc}).
1401 However, the use of @command{gcc} does not add the C++ library.
1402 @command{g++} is a program that calls GCC and treats @samp{.c},
1403 @samp{.h} and @samp{.i} files as C++ source files instead of C source
1404 files unless @option{-x} is used, and automatically specifies linking
1405 against the C++ library. This program is also useful when
1406 precompiling a C header file with a @samp{.h} extension for use in C++
1407 compilations. On many systems, @command{g++} is also installed with
1408 the name @command{c++}.
1410 @cindex invoking @command{g++}
1411 When you compile C++ programs, you may specify many of the same
1412 command-line options that you use for compiling programs in any
1413 language; or command-line options meaningful for C and related
1414 languages; or options that are meaningful only for C++ programs.
1415 @xref{C Dialect Options,,Options Controlling C Dialect}, for
1416 explanations of options for languages related to C@.
1417 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
1418 explanations of options that are meaningful only for C++ programs.
1420 @node C Dialect Options
1421 @section Options Controlling C Dialect
1422 @cindex dialect options
1423 @cindex language dialect options
1424 @cindex options, dialect
1426 The following options control the dialect of C (or languages derived
1427 from C, such as C++, Objective-C and Objective-C++) that the compiler
1431 @cindex ANSI support
1435 In C mode, this is equivalent to @samp{-std=c90}. In C++ mode, it is
1436 equivalent to @samp{-std=c++98}.
1438 This turns off certain features of GCC that are incompatible with ISO
1439 C90 (when compiling C code), or of standard C++ (when compiling C++ code),
1440 such as the @code{asm} and @code{typeof} keywords, and
1441 predefined macros such as @code{unix} and @code{vax} that identify the
1442 type of system you are using. It also enables the undesirable and
1443 rarely used ISO trigraph feature. For the C compiler,
1444 it disables recognition of C++ style @samp{//} comments as well as
1445 the @code{inline} keyword.
1447 The alternate keywords @code{__asm__}, @code{__extension__},
1448 @code{__inline__} and @code{__typeof__} continue to work despite
1449 @option{-ansi}. You would not want to use them in an ISO C program, of
1450 course, but it is useful to put them in header files that might be included
1451 in compilations done with @option{-ansi}. Alternate predefined macros
1452 such as @code{__unix__} and @code{__vax__} are also available, with or
1453 without @option{-ansi}.
1455 The @option{-ansi} option does not cause non-ISO programs to be
1456 rejected gratuitously. For that, @option{-pedantic} is required in
1457 addition to @option{-ansi}. @xref{Warning Options}.
1459 The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi}
1460 option is used. Some header files may notice this macro and refrain
1461 from declaring certain functions or defining certain macros that the
1462 ISO standard doesn't call for; this is to avoid interfering with any
1463 programs that might use these names for other things.
1465 Functions that would normally be built in but do not have semantics
1466 defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in
1467 functions when @option{-ansi} is used. @xref{Other Builtins,,Other
1468 built-in functions provided by GCC}, for details of the functions
1473 Determine the language standard. @xref{Standards,,Language Standards
1474 Supported by GCC}, for details of these standard versions. This option
1475 is currently only supported when compiling C or C++.
1477 The compiler can accept several base standards, such as @samp{c90} or
1478 @samp{c++98}, and GNU dialects of those standards, such as
1479 @samp{gnu90} or @samp{gnu++98}. By specifying a base standard, the
1480 compiler will accept all programs following that standard and those
1481 using GNU extensions that do not contradict it. For example,
1482 @samp{-std=c90} turns off certain features of GCC that are
1483 incompatible with ISO C90, such as the @code{asm} and @code{typeof}
1484 keywords, but not other GNU extensions that do not have a meaning in
1485 ISO C90, such as omitting the middle term of a @code{?:}
1486 expression. On the other hand, by specifying a GNU dialect of a
1487 standard, all features the compiler support are enabled, even when
1488 those features change the meaning of the base standard and some
1489 strict-conforming programs may be rejected. The particular standard
1490 is used by @option{-pedantic} to identify which features are GNU
1491 extensions given that version of the standard. For example
1492 @samp{-std=gnu90 -pedantic} would warn about C++ style @samp{//}
1493 comments, while @samp{-std=gnu99 -pedantic} would not.
1495 A value for this option must be provided; possible values are
1501 Support all ISO C90 programs (certain GNU extensions that conflict
1502 with ISO C90 are disabled). Same as @option{-ansi} for C code.
1504 @item iso9899:199409
1505 ISO C90 as modified in amendment 1.
1511 ISO C99. Note that this standard is not yet fully supported; see
1512 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The
1513 names @samp{c9x} and @samp{iso9899:199x} are deprecated.
1516 ISO C1X, the draft of the next revision of the ISO C standard.
1517 Support is limited and experimental and features enabled by this
1518 option may be changed or removed if changed in or removed from the
1523 GNU dialect of ISO C90 (including some C99 features). This
1524 is the default for C code.
1528 GNU dialect of ISO C99. When ISO C99 is fully implemented in GCC,
1529 this will become the default. The name @samp{gnu9x} is deprecated.
1532 GNU dialect of ISO C1X. Support is limited and experimental and
1533 features enabled by this option may be changed or removed if changed
1534 in or removed from the standard draft.
1537 The 1998 ISO C++ standard plus amendments. Same as @option{-ansi} for
1541 GNU dialect of @option{-std=c++98}. This is the default for
1545 The working draft of the upcoming ISO C++0x standard. This option
1546 enables experimental features that are likely to be included in
1547 C++0x. The working draft is constantly changing, and any feature that is
1548 enabled by this flag may be removed from future versions of GCC if it is
1549 not part of the C++0x standard.
1552 GNU dialect of @option{-std=c++0x}. This option enables
1553 experimental features that may be removed in future versions of GCC.
1556 @item -fgnu89-inline
1557 @opindex fgnu89-inline
1558 The option @option{-fgnu89-inline} tells GCC to use the traditional
1559 GNU semantics for @code{inline} functions when in C99 mode.
1560 @xref{Inline,,An Inline Function is As Fast As a Macro}. This option
1561 is accepted and ignored by GCC versions 4.1.3 up to but not including
1562 4.3. In GCC versions 4.3 and later it changes the behavior of GCC in
1563 C99 mode. Using this option is roughly equivalent to adding the
1564 @code{gnu_inline} function attribute to all inline functions
1565 (@pxref{Function Attributes}).
1567 The option @option{-fno-gnu89-inline} explicitly tells GCC to use the
1568 C99 semantics for @code{inline} when in C99 or gnu99 mode (i.e., it
1569 specifies the default behavior). This option was first supported in
1570 GCC 4.3. This option is not supported in @option{-std=c90} or
1571 @option{-std=gnu90} mode.
1573 The preprocessor macros @code{__GNUC_GNU_INLINE__} and
1574 @code{__GNUC_STDC_INLINE__} may be used to check which semantics are
1575 in effect for @code{inline} functions. @xref{Common Predefined
1576 Macros,,,cpp,The C Preprocessor}.
1578 @item -aux-info @var{filename}
1580 Output to the given filename prototyped declarations for all functions
1581 declared and/or defined in a translation unit, including those in header
1582 files. This option is silently ignored in any language other than C@.
1584 Besides declarations, the file indicates, in comments, the origin of
1585 each declaration (source file and line), whether the declaration was
1586 implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or
1587 @samp{O} for old, respectively, in the first character after the line
1588 number and the colon), and whether it came from a declaration or a
1589 definition (@samp{C} or @samp{F}, respectively, in the following
1590 character). In the case of function definitions, a K&R-style list of
1591 arguments followed by their declarations is also provided, inside
1592 comments, after the declaration.
1596 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
1597 keyword, so that code can use these words as identifiers. You can use
1598 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
1599 instead. @option{-ansi} implies @option{-fno-asm}.
1601 In C++, this switch only affects the @code{typeof} keyword, since
1602 @code{asm} and @code{inline} are standard keywords. You may want to
1603 use the @option{-fno-gnu-keywords} flag instead, which has the same
1604 effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this
1605 switch only affects the @code{asm} and @code{typeof} keywords, since
1606 @code{inline} is a standard keyword in ISO C99.
1609 @itemx -fno-builtin-@var{function}
1610 @opindex fno-builtin
1611 @cindex built-in functions
1612 Don't recognize built-in functions that do not begin with
1613 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
1614 functions provided by GCC}, for details of the functions affected,
1615 including those which are not built-in functions when @option{-ansi} or
1616 @option{-std} options for strict ISO C conformance are used because they
1617 do not have an ISO standard meaning.
1619 GCC normally generates special code to handle certain built-in functions
1620 more efficiently; for instance, calls to @code{alloca} may become single
1621 instructions that adjust the stack directly, and calls to @code{memcpy}
1622 may become inline copy loops. The resulting code is often both smaller
1623 and faster, but since the function calls no longer appear as such, you
1624 cannot set a breakpoint on those calls, nor can you change the behavior
1625 of the functions by linking with a different library. In addition,
1626 when a function is recognized as a built-in function, GCC may use
1627 information about that function to warn about problems with calls to
1628 that function, or to generate more efficient code, even if the
1629 resulting code still contains calls to that function. For example,
1630 warnings are given with @option{-Wformat} for bad calls to
1631 @code{printf}, when @code{printf} is built in, and @code{strlen} is
1632 known not to modify global memory.
1634 With the @option{-fno-builtin-@var{function}} option
1635 only the built-in function @var{function} is
1636 disabled. @var{function} must not begin with @samp{__builtin_}. If a
1637 function is named that is not built-in in this version of GCC, this
1638 option is ignored. There is no corresponding
1639 @option{-fbuiltin-@var{function}} option; if you wish to enable
1640 built-in functions selectively when using @option{-fno-builtin} or
1641 @option{-ffreestanding}, you may define macros such as:
1644 #define abs(n) __builtin_abs ((n))
1645 #define strcpy(d, s) __builtin_strcpy ((d), (s))
1650 @cindex hosted environment
1652 Assert that compilation takes place in a hosted environment. This implies
1653 @option{-fbuiltin}. A hosted environment is one in which the
1654 entire standard library is available, and in which @code{main} has a return
1655 type of @code{int}. Examples are nearly everything except a kernel.
1656 This is equivalent to @option{-fno-freestanding}.
1658 @item -ffreestanding
1659 @opindex ffreestanding
1660 @cindex hosted environment
1662 Assert that compilation takes place in a freestanding environment. This
1663 implies @option{-fno-builtin}. A freestanding environment
1664 is one in which the standard library may not exist, and program startup may
1665 not necessarily be at @code{main}. The most obvious example is an OS kernel.
1666 This is equivalent to @option{-fno-hosted}.
1668 @xref{Standards,,Language Standards Supported by GCC}, for details of
1669 freestanding and hosted environments.
1673 @cindex openmp parallel
1674 Enable handling of OpenMP directives @code{#pragma omp} in C/C++ and
1675 @code{!$omp} in Fortran. When @option{-fopenmp} is specified, the
1676 compiler generates parallel code according to the OpenMP Application
1677 Program Interface v3.0 @w{@uref{http://www.openmp.org/}}. This option
1678 implies @option{-pthread}, and thus is only supported on targets that
1679 have support for @option{-pthread}.
1681 @item -fms-extensions
1682 @opindex fms-extensions
1683 Accept some non-standard constructs used in Microsoft header files.
1685 Some cases of unnamed fields in structures and unions are only
1686 accepted with this option. @xref{Unnamed Fields,,Unnamed struct/union
1687 fields within structs/unions}, for details.
1691 Support ISO C trigraphs. The @option{-ansi} option (and @option{-std}
1692 options for strict ISO C conformance) implies @option{-trigraphs}.
1694 @item -no-integrated-cpp
1695 @opindex no-integrated-cpp
1696 Performs a compilation in two passes: preprocessing and compiling. This
1697 option allows a user supplied "cc1", "cc1plus", or "cc1obj" via the
1698 @option{-B} option. The user supplied compilation step can then add in
1699 an additional preprocessing step after normal preprocessing but before
1700 compiling. The default is to use the integrated cpp (internal cpp)
1702 The semantics of this option will change if "cc1", "cc1plus", and
1703 "cc1obj" are merged.
1705 @cindex traditional C language
1706 @cindex C language, traditional
1708 @itemx -traditional-cpp
1709 @opindex traditional-cpp
1710 @opindex traditional
1711 Formerly, these options caused GCC to attempt to emulate a pre-standard
1712 C compiler. They are now only supported with the @option{-E} switch.
1713 The preprocessor continues to support a pre-standard mode. See the GNU
1714 CPP manual for details.
1716 @item -fcond-mismatch
1717 @opindex fcond-mismatch
1718 Allow conditional expressions with mismatched types in the second and
1719 third arguments. The value of such an expression is void. This option
1720 is not supported for C++.
1722 @item -flax-vector-conversions
1723 @opindex flax-vector-conversions
1724 Allow implicit conversions between vectors with differing numbers of
1725 elements and/or incompatible element types. This option should not be
1728 @item -funsigned-char
1729 @opindex funsigned-char
1730 Let the type @code{char} be unsigned, like @code{unsigned char}.
1732 Each kind of machine has a default for what @code{char} should
1733 be. It is either like @code{unsigned char} by default or like
1734 @code{signed char} by default.
1736 Ideally, a portable program should always use @code{signed char} or
1737 @code{unsigned char} when it depends on the signedness of an object.
1738 But many programs have been written to use plain @code{char} and
1739 expect it to be signed, or expect it to be unsigned, depending on the
1740 machines they were written for. This option, and its inverse, let you
1741 make such a program work with the opposite default.
1743 The type @code{char} is always a distinct type from each of
1744 @code{signed char} or @code{unsigned char}, even though its behavior
1745 is always just like one of those two.
1748 @opindex fsigned-char
1749 Let the type @code{char} be signed, like @code{signed char}.
1751 Note that this is equivalent to @option{-fno-unsigned-char}, which is
1752 the negative form of @option{-funsigned-char}. Likewise, the option
1753 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
1755 @item -fsigned-bitfields
1756 @itemx -funsigned-bitfields
1757 @itemx -fno-signed-bitfields
1758 @itemx -fno-unsigned-bitfields
1759 @opindex fsigned-bitfields
1760 @opindex funsigned-bitfields
1761 @opindex fno-signed-bitfields
1762 @opindex fno-unsigned-bitfields
1763 These options control whether a bit-field is signed or unsigned, when the
1764 declaration does not use either @code{signed} or @code{unsigned}. By
1765 default, such a bit-field is signed, because this is consistent: the
1766 basic integer types such as @code{int} are signed types.
1769 @node C++ Dialect Options
1770 @section Options Controlling C++ Dialect
1772 @cindex compiler options, C++
1773 @cindex C++ options, command line
1774 @cindex options, C++
1775 This section describes the command-line options that are only meaningful
1776 for C++ programs; but you can also use most of the GNU compiler options
1777 regardless of what language your program is in. For example, you
1778 might compile a file @code{firstClass.C} like this:
1781 g++ -g -frepo -O -c firstClass.C
1785 In this example, only @option{-frepo} is an option meant
1786 only for C++ programs; you can use the other options with any
1787 language supported by GCC@.
1789 Here is a list of options that are @emph{only} for compiling C++ programs:
1793 @item -fabi-version=@var{n}
1794 @opindex fabi-version
1795 Use version @var{n} of the C++ ABI@. Version 2 is the version of the
1796 C++ ABI that first appeared in G++ 3.4. Version 1 is the version of
1797 the C++ ABI that first appeared in G++ 3.2. Version 0 will always be
1798 the version that conforms most closely to the C++ ABI specification.
1799 Therefore, the ABI obtained using version 0 will change as ABI bugs
1802 The default is version 2.
1804 Version 3 corrects an error in mangling a constant address as a
1807 Version 4 implements a standard mangling for vector types.
1809 See also @option{-Wabi}.
1811 @item -fno-access-control
1812 @opindex fno-access-control
1813 Turn off all access checking. This switch is mainly useful for working
1814 around bugs in the access control code.
1818 Check that the pointer returned by @code{operator new} is non-null
1819 before attempting to modify the storage allocated. This check is
1820 normally unnecessary because the C++ standard specifies that
1821 @code{operator new} will only return @code{0} if it is declared
1822 @samp{throw()}, in which case the compiler will always check the
1823 return value even without this option. In all other cases, when
1824 @code{operator new} has a non-empty exception specification, memory
1825 exhaustion is signalled by throwing @code{std::bad_alloc}. See also
1826 @samp{new (nothrow)}.
1828 @item -fconserve-space
1829 @opindex fconserve-space
1830 Put uninitialized or runtime-initialized global variables into the
1831 common segment, as C does. This saves space in the executable at the
1832 cost of not diagnosing duplicate definitions. If you compile with this
1833 flag and your program mysteriously crashes after @code{main()} has
1834 completed, you may have an object that is being destroyed twice because
1835 two definitions were merged.
1837 This option is no longer useful on most targets, now that support has
1838 been added for putting variables into BSS without making them common.
1840 @item -fno-deduce-init-list
1841 @opindex fno-deduce-init-list
1842 Disable deduction of a template type parameter as
1843 std::initializer_list from a brace-enclosed initializer list, i.e.
1846 template <class T> auto forward(T t) -> decltype (realfn (t))
1853 forward(@{1,2@}); // call forward<std::initializer_list<int>>
1857 This option is present because this deduction is an extension to the
1858 current specification in the C++0x working draft, and there was
1859 some concern about potential overload resolution problems.
1861 @item -ffriend-injection
1862 @opindex ffriend-injection
1863 Inject friend functions into the enclosing namespace, so that they are
1864 visible outside the scope of the class in which they are declared.
1865 Friend functions were documented to work this way in the old Annotated
1866 C++ Reference Manual, and versions of G++ before 4.1 always worked
1867 that way. However, in ISO C++ a friend function which is not declared
1868 in an enclosing scope can only be found using argument dependent
1869 lookup. This option causes friends to be injected as they were in
1872 This option is for compatibility, and may be removed in a future
1875 @item -fno-elide-constructors
1876 @opindex fno-elide-constructors
1877 The C++ standard allows an implementation to omit creating a temporary
1878 which is only used to initialize another object of the same type.
1879 Specifying this option disables that optimization, and forces G++ to
1880 call the copy constructor in all cases.
1882 @item -fno-enforce-eh-specs
1883 @opindex fno-enforce-eh-specs
1884 Don't generate code to check for violation of exception specifications
1885 at runtime. This option violates the C++ standard, but may be useful
1886 for reducing code size in production builds, much like defining
1887 @samp{NDEBUG}. This does not give user code permission to throw
1888 exceptions in violation of the exception specifications; the compiler
1889 will still optimize based on the specifications, so throwing an
1890 unexpected exception will result in undefined behavior.
1893 @itemx -fno-for-scope
1895 @opindex fno-for-scope
1896 If @option{-ffor-scope} is specified, the scope of variables declared in
1897 a @i{for-init-statement} is limited to the @samp{for} loop itself,
1898 as specified by the C++ standard.
1899 If @option{-fno-for-scope} is specified, the scope of variables declared in
1900 a @i{for-init-statement} extends to the end of the enclosing scope,
1901 as was the case in old versions of G++, and other (traditional)
1902 implementations of C++.
1904 The default if neither flag is given to follow the standard,
1905 but to allow and give a warning for old-style code that would
1906 otherwise be invalid, or have different behavior.
1908 @item -fno-gnu-keywords
1909 @opindex fno-gnu-keywords
1910 Do not recognize @code{typeof} as a keyword, so that code can use this
1911 word as an identifier. You can use the keyword @code{__typeof__} instead.
1912 @option{-ansi} implies @option{-fno-gnu-keywords}.
1914 @item -fno-implicit-templates
1915 @opindex fno-implicit-templates
1916 Never emit code for non-inline templates which are instantiated
1917 implicitly (i.e.@: by use); only emit code for explicit instantiations.
1918 @xref{Template Instantiation}, for more information.
1920 @item -fno-implicit-inline-templates
1921 @opindex fno-implicit-inline-templates
1922 Don't emit code for implicit instantiations of inline templates, either.
1923 The default is to handle inlines differently so that compiles with and
1924 without optimization will need the same set of explicit instantiations.
1926 @item -fno-implement-inlines
1927 @opindex fno-implement-inlines
1928 To save space, do not emit out-of-line copies of inline functions
1929 controlled by @samp{#pragma implementation}. This will cause linker
1930 errors if these functions are not inlined everywhere they are called.
1932 @item -fms-extensions
1933 @opindex fms-extensions
1934 Disable pedantic warnings about constructs used in MFC, such as implicit
1935 int and getting a pointer to member function via non-standard syntax.
1937 @item -fno-nonansi-builtins
1938 @opindex fno-nonansi-builtins
1939 Disable built-in declarations of functions that are not mandated by
1940 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
1941 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
1944 @opindex fnothrow-opt
1945 Treat a @code{throw()} exception specification as though it were a
1946 @code{noexcept} specification to reduce or eliminate the text size
1947 overhead relative to a function with no exception specification. If
1948 the function has local variables of types with non-trivial
1949 destructors, the exception specification will actually make the
1950 function smaller because the EH cleanups for those variables can be
1951 optimized away. The semantic effect is that an exception thrown out of
1952 a function with such an exception specification will result in a call
1953 to @code{terminate} rather than @code{unexpected}.
1955 @item -fno-operator-names
1956 @opindex fno-operator-names
1957 Do not treat the operator name keywords @code{and}, @code{bitand},
1958 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
1959 synonyms as keywords.
1961 @item -fno-optional-diags
1962 @opindex fno-optional-diags
1963 Disable diagnostics that the standard says a compiler does not need to
1964 issue. Currently, the only such diagnostic issued by G++ is the one for
1965 a name having multiple meanings within a class.
1968 @opindex fpermissive
1969 Downgrade some diagnostics about nonconformant code from errors to
1970 warnings. Thus, using @option{-fpermissive} will allow some
1971 nonconforming code to compile.
1973 @item -fno-pretty-templates
1974 @opindex fno-pretty-templates
1975 When an error message refers to a specialization of a function
1976 template, the compiler will normally print the signature of the
1977 template followed by the template arguments and any typedefs or
1978 typenames in the signature (e.g. @code{void f(T) [with T = int]}
1979 rather than @code{void f(int)}) so that it's clear which template is
1980 involved. When an error message refers to a specialization of a class
1981 template, the compiler will omit any template arguments which match
1982 the default template arguments for that template. If either of these
1983 behaviors make it harder to understand the error message rather than
1984 easier, using @option{-fno-pretty-templates} will disable them.
1988 Enable automatic template instantiation at link time. This option also
1989 implies @option{-fno-implicit-templates}. @xref{Template
1990 Instantiation}, for more information.
1994 Disable generation of information about every class with virtual
1995 functions for use by the C++ runtime type identification features
1996 (@samp{dynamic_cast} and @samp{typeid}). If you don't use those parts
1997 of the language, you can save some space by using this flag. Note that
1998 exception handling uses the same information, but it will generate it as
1999 needed. The @samp{dynamic_cast} operator can still be used for casts that
2000 do not require runtime type information, i.e.@: casts to @code{void *} or to
2001 unambiguous base classes.
2005 Emit statistics about front-end processing at the end of the compilation.
2006 This information is generally only useful to the G++ development team.
2008 @item -fstrict-enums
2009 @opindex fstrict-enums
2010 Allow the compiler to optimize using the assumption that a value of
2011 enumeration type can only be one of the values of the enumeration (as
2012 defined in the C++ standard; basically, a value which can be
2013 represented in the minimum number of bits needed to represent all the
2014 enumerators). This assumption may not be valid if the program uses a
2015 cast to convert an arbitrary integer value to the enumeration type.
2017 @item -ftemplate-depth=@var{n}
2018 @opindex ftemplate-depth
2019 Set the maximum instantiation depth for template classes to @var{n}.
2020 A limit on the template instantiation depth is needed to detect
2021 endless recursions during template class instantiation. ANSI/ISO C++
2022 conforming programs must not rely on a maximum depth greater than 17
2023 (changed to 1024 in C++0x).
2025 @item -fno-threadsafe-statics
2026 @opindex fno-threadsafe-statics
2027 Do not emit the extra code to use the routines specified in the C++
2028 ABI for thread-safe initialization of local statics. You can use this
2029 option to reduce code size slightly in code that doesn't need to be
2032 @item -fuse-cxa-atexit
2033 @opindex fuse-cxa-atexit
2034 Register destructors for objects with static storage duration with the
2035 @code{__cxa_atexit} function rather than the @code{atexit} function.
2036 This option is required for fully standards-compliant handling of static
2037 destructors, but will only work if your C library supports
2038 @code{__cxa_atexit}.
2040 @item -fno-use-cxa-get-exception-ptr
2041 @opindex fno-use-cxa-get-exception-ptr
2042 Don't use the @code{__cxa_get_exception_ptr} runtime routine. This
2043 will cause @code{std::uncaught_exception} to be incorrect, but is necessary
2044 if the runtime routine is not available.
2046 @item -fvisibility-inlines-hidden
2047 @opindex fvisibility-inlines-hidden
2048 This switch declares that the user does not attempt to compare
2049 pointers to inline methods where the addresses of the two functions
2050 were taken in different shared objects.
2052 The effect of this is that GCC may, effectively, mark inline methods with
2053 @code{__attribute__ ((visibility ("hidden")))} so that they do not
2054 appear in the export table of a DSO and do not require a PLT indirection
2055 when used within the DSO@. Enabling this option can have a dramatic effect
2056 on load and link times of a DSO as it massively reduces the size of the
2057 dynamic export table when the library makes heavy use of templates.
2059 The behavior of this switch is not quite the same as marking the
2060 methods as hidden directly, because it does not affect static variables
2061 local to the function or cause the compiler to deduce that
2062 the function is defined in only one shared object.
2064 You may mark a method as having a visibility explicitly to negate the
2065 effect of the switch for that method. For example, if you do want to
2066 compare pointers to a particular inline method, you might mark it as
2067 having default visibility. Marking the enclosing class with explicit
2068 visibility will have no effect.
2070 Explicitly instantiated inline methods are unaffected by this option
2071 as their linkage might otherwise cross a shared library boundary.
2072 @xref{Template Instantiation}.
2074 @item -fvisibility-ms-compat
2075 @opindex fvisibility-ms-compat
2076 This flag attempts to use visibility settings to make GCC's C++
2077 linkage model compatible with that of Microsoft Visual Studio.
2079 The flag makes these changes to GCC's linkage model:
2083 It sets the default visibility to @code{hidden}, like
2084 @option{-fvisibility=hidden}.
2087 Types, but not their members, are not hidden by default.
2090 The One Definition Rule is relaxed for types without explicit
2091 visibility specifications which are defined in more than one different
2092 shared object: those declarations are permitted if they would have
2093 been permitted when this option was not used.
2096 In new code it is better to use @option{-fvisibility=hidden} and
2097 export those classes which are intended to be externally visible.
2098 Unfortunately it is possible for code to rely, perhaps accidentally,
2099 on the Visual Studio behavior.
2101 Among the consequences of these changes are that static data members
2102 of the same type with the same name but defined in different shared
2103 objects will be different, so changing one will not change the other;
2104 and that pointers to function members defined in different shared
2105 objects may not compare equal. When this flag is given, it is a
2106 violation of the ODR to define types with the same name differently.
2110 Do not use weak symbol support, even if it is provided by the linker.
2111 By default, G++ will use weak symbols if they are available. This
2112 option exists only for testing, and should not be used by end-users;
2113 it will result in inferior code and has no benefits. This option may
2114 be removed in a future release of G++.
2118 Do not search for header files in the standard directories specific to
2119 C++, but do still search the other standard directories. (This option
2120 is used when building the C++ library.)
2123 In addition, these optimization, warning, and code generation options
2124 have meanings only for C++ programs:
2127 @item -fno-default-inline
2128 @opindex fno-default-inline
2129 Do not assume @samp{inline} for functions defined inside a class scope.
2130 @xref{Optimize Options,,Options That Control Optimization}. Note that these
2131 functions will have linkage like inline functions; they just won't be
2134 @item -Wabi @r{(C, Objective-C, C++ and Objective-C++ only)}
2137 Warn when G++ generates code that is probably not compatible with the
2138 vendor-neutral C++ ABI@. Although an effort has been made to warn about
2139 all such cases, there are probably some cases that are not warned about,
2140 even though G++ is generating incompatible code. There may also be
2141 cases where warnings are emitted even though the code that is generated
2144 You should rewrite your code to avoid these warnings if you are
2145 concerned about the fact that code generated by G++ may not be binary
2146 compatible with code generated by other compilers.
2148 The known incompatibilities in @option{-fabi-version=2} (the default) include:
2153 A template with a non-type template parameter of reference type is
2154 mangled incorrectly:
2157 template <int &> struct S @{@};
2161 This is fixed in @option{-fabi-version=3}.
2164 SIMD vector types declared using @code{__attribute ((vector_size))} are
2165 mangled in a non-standard way that does not allow for overloading of
2166 functions taking vectors of different sizes.
2168 The mangling is changed in @option{-fabi-version=4}.
2171 The known incompatibilities in @option{-fabi-version=1} include:
2176 Incorrect handling of tail-padding for bit-fields. G++ may attempt to
2177 pack data into the same byte as a base class. For example:
2180 struct A @{ virtual void f(); int f1 : 1; @};
2181 struct B : public A @{ int f2 : 1; @};
2185 In this case, G++ will place @code{B::f2} into the same byte
2186 as@code{A::f1}; other compilers will not. You can avoid this problem
2187 by explicitly padding @code{A} so that its size is a multiple of the
2188 byte size on your platform; that will cause G++ and other compilers to
2189 layout @code{B} identically.
2192 Incorrect handling of tail-padding for virtual bases. G++ does not use
2193 tail padding when laying out virtual bases. For example:
2196 struct A @{ virtual void f(); char c1; @};
2197 struct B @{ B(); char c2; @};
2198 struct C : public A, public virtual B @{@};
2202 In this case, G++ will not place @code{B} into the tail-padding for
2203 @code{A}; other compilers will. You can avoid this problem by
2204 explicitly padding @code{A} so that its size is a multiple of its
2205 alignment (ignoring virtual base classes); that will cause G++ and other
2206 compilers to layout @code{C} identically.
2209 Incorrect handling of bit-fields with declared widths greater than that
2210 of their underlying types, when the bit-fields appear in a union. For
2214 union U @{ int i : 4096; @};
2218 Assuming that an @code{int} does not have 4096 bits, G++ will make the
2219 union too small by the number of bits in an @code{int}.
2222 Empty classes can be placed at incorrect offsets. For example:
2232 struct C : public B, public A @{@};
2236 G++ will place the @code{A} base class of @code{C} at a nonzero offset;
2237 it should be placed at offset zero. G++ mistakenly believes that the
2238 @code{A} data member of @code{B} is already at offset zero.
2241 Names of template functions whose types involve @code{typename} or
2242 template template parameters can be mangled incorrectly.
2245 template <typename Q>
2246 void f(typename Q::X) @{@}
2248 template <template <typename> class Q>
2249 void f(typename Q<int>::X) @{@}
2253 Instantiations of these templates may be mangled incorrectly.
2257 It also warns psABI related changes. The known psABI changes at this
2263 For SYSV/x86-64, when passing union with long double, it is changed to
2264 pass in memory as specified in psABI. For example:
2274 @code{union U} will always be passed in memory.
2278 @item -Wctor-dtor-privacy @r{(C++ and Objective-C++ only)}
2279 @opindex Wctor-dtor-privacy
2280 @opindex Wno-ctor-dtor-privacy
2281 Warn when a class seems unusable because all the constructors or
2282 destructors in that class are private, and it has neither friends nor
2283 public static member functions.
2285 @item -Wnoexcept @r{(C++ and Objective-C++ only)}
2287 @opindex Wno-noexcept
2288 Warn when a noexcept-expression evaluates to false because of a call
2289 to a function that does not have a non-throwing exception
2290 specification (i.e. @samp{throw()} or @samp{noexcept}) but is known by
2291 the compiler to never throw an exception.
2293 @item -Wnon-virtual-dtor @r{(C++ and Objective-C++ only)}
2294 @opindex Wnon-virtual-dtor
2295 @opindex Wno-non-virtual-dtor
2296 Warn when a class has virtual functions and accessible non-virtual
2297 destructor, in which case it would be possible but unsafe to delete
2298 an instance of a derived class through a pointer to the base class.
2299 This warning is also enabled if -Weffc++ is specified.
2301 @item -Wreorder @r{(C++ and Objective-C++ only)}
2303 @opindex Wno-reorder
2304 @cindex reordering, warning
2305 @cindex warning for reordering of member initializers
2306 Warn when the order of member initializers given in the code does not
2307 match the order in which they must be executed. For instance:
2313 A(): j (0), i (1) @{ @}
2317 The compiler will rearrange the member initializers for @samp{i}
2318 and @samp{j} to match the declaration order of the members, emitting
2319 a warning to that effect. This warning is enabled by @option{-Wall}.
2322 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
2325 @item -Weffc++ @r{(C++ and Objective-C++ only)}
2328 Warn about violations of the following style guidelines from Scott Meyers'
2329 @cite{Effective C++} book:
2333 Item 11: Define a copy constructor and an assignment operator for classes
2334 with dynamically allocated memory.
2337 Item 12: Prefer initialization to assignment in constructors.
2340 Item 14: Make destructors virtual in base classes.
2343 Item 15: Have @code{operator=} return a reference to @code{*this}.
2346 Item 23: Don't try to return a reference when you must return an object.
2350 Also warn about violations of the following style guidelines from
2351 Scott Meyers' @cite{More Effective C++} book:
2355 Item 6: Distinguish between prefix and postfix forms of increment and
2356 decrement operators.
2359 Item 7: Never overload @code{&&}, @code{||}, or @code{,}.
2363 When selecting this option, be aware that the standard library
2364 headers do not obey all of these guidelines; use @samp{grep -v}
2365 to filter out those warnings.
2367 @item -Wstrict-null-sentinel @r{(C++ and Objective-C++ only)}
2368 @opindex Wstrict-null-sentinel
2369 @opindex Wno-strict-null-sentinel
2370 Warn also about the use of an uncasted @code{NULL} as sentinel. When
2371 compiling only with GCC this is a valid sentinel, as @code{NULL} is defined
2372 to @code{__null}. Although it is a null pointer constant not a null pointer,
2373 it is guaranteed to be of the same size as a pointer. But this use is
2374 not portable across different compilers.
2376 @item -Wno-non-template-friend @r{(C++ and Objective-C++ only)}
2377 @opindex Wno-non-template-friend
2378 @opindex Wnon-template-friend
2379 Disable warnings when non-templatized friend functions are declared
2380 within a template. Since the advent of explicit template specification
2381 support in G++, if the name of the friend is an unqualified-id (i.e.,
2382 @samp{friend foo(int)}), the C++ language specification demands that the
2383 friend declare or define an ordinary, nontemplate function. (Section
2384 14.5.3). Before G++ implemented explicit specification, unqualified-ids
2385 could be interpreted as a particular specialization of a templatized
2386 function. Because this non-conforming behavior is no longer the default
2387 behavior for G++, @option{-Wnon-template-friend} allows the compiler to
2388 check existing code for potential trouble spots and is on by default.
2389 This new compiler behavior can be turned off with
2390 @option{-Wno-non-template-friend} which keeps the conformant compiler code
2391 but disables the helpful warning.
2393 @item -Wold-style-cast @r{(C++ and Objective-C++ only)}
2394 @opindex Wold-style-cast
2395 @opindex Wno-old-style-cast
2396 Warn if an old-style (C-style) cast to a non-void type is used within
2397 a C++ program. The new-style casts (@samp{dynamic_cast},
2398 @samp{static_cast}, @samp{reinterpret_cast}, and @samp{const_cast}) are
2399 less vulnerable to unintended effects and much easier to search for.
2401 @item -Woverloaded-virtual @r{(C++ and Objective-C++ only)}
2402 @opindex Woverloaded-virtual
2403 @opindex Wno-overloaded-virtual
2404 @cindex overloaded virtual fn, warning
2405 @cindex warning for overloaded virtual fn
2406 Warn when a function declaration hides virtual functions from a
2407 base class. For example, in:
2414 struct B: public A @{
2419 the @code{A} class version of @code{f} is hidden in @code{B}, and code
2427 will fail to compile.
2429 @item -Wno-pmf-conversions @r{(C++ and Objective-C++ only)}
2430 @opindex Wno-pmf-conversions
2431 @opindex Wpmf-conversions
2432 Disable the diagnostic for converting a bound pointer to member function
2435 @item -Wsign-promo @r{(C++ and Objective-C++ only)}
2436 @opindex Wsign-promo
2437 @opindex Wno-sign-promo
2438 Warn when overload resolution chooses a promotion from unsigned or
2439 enumerated type to a signed type, over a conversion to an unsigned type of
2440 the same size. Previous versions of G++ would try to preserve
2441 unsignedness, but the standard mandates the current behavior.
2446 A& operator = (int);
2456 In this example, G++ will synthesize a default @samp{A& operator =
2457 (const A&);}, while cfront will use the user-defined @samp{operator =}.
2460 @node Objective-C and Objective-C++ Dialect Options
2461 @section Options Controlling Objective-C and Objective-C++ Dialects
2463 @cindex compiler options, Objective-C and Objective-C++
2464 @cindex Objective-C and Objective-C++ options, command line
2465 @cindex options, Objective-C and Objective-C++
2466 (NOTE: This manual does not describe the Objective-C and Objective-C++
2467 languages themselves. See @xref{Standards,,Language Standards
2468 Supported by GCC}, for references.)
2470 This section describes the command-line options that are only meaningful
2471 for Objective-C and Objective-C++ programs, but you can also use most of
2472 the language-independent GNU compiler options.
2473 For example, you might compile a file @code{some_class.m} like this:
2476 gcc -g -fgnu-runtime -O -c some_class.m
2480 In this example, @option{-fgnu-runtime} is an option meant only for
2481 Objective-C and Objective-C++ programs; you can use the other options with
2482 any language supported by GCC@.
2484 Note that since Objective-C is an extension of the C language, Objective-C
2485 compilations may also use options specific to the C front-end (e.g.,
2486 @option{-Wtraditional}). Similarly, Objective-C++ compilations may use
2487 C++-specific options (e.g., @option{-Wabi}).
2489 Here is a list of options that are @emph{only} for compiling Objective-C
2490 and Objective-C++ programs:
2493 @item -fconstant-string-class=@var{class-name}
2494 @opindex fconstant-string-class
2495 Use @var{class-name} as the name of the class to instantiate for each
2496 literal string specified with the syntax @code{@@"@dots{}"}. The default
2497 class name is @code{NXConstantString} if the GNU runtime is being used, and
2498 @code{NSConstantString} if the NeXT runtime is being used (see below). The
2499 @option{-fconstant-cfstrings} option, if also present, will override the
2500 @option{-fconstant-string-class} setting and cause @code{@@"@dots{}"} literals
2501 to be laid out as constant CoreFoundation strings.
2504 @opindex fgnu-runtime
2505 Generate object code compatible with the standard GNU Objective-C
2506 runtime. This is the default for most types of systems.
2508 @item -fnext-runtime
2509 @opindex fnext-runtime
2510 Generate output compatible with the NeXT runtime. This is the default
2511 for NeXT-based systems, including Darwin and Mac OS X@. The macro
2512 @code{__NEXT_RUNTIME__} is predefined if (and only if) this option is
2515 @item -fno-nil-receivers
2516 @opindex fno-nil-receivers
2517 Assume that all Objective-C message dispatches (e.g.,
2518 @code{[receiver message:arg]}) in this translation unit ensure that the receiver
2519 is not @code{nil}. This allows for more efficient entry points in the runtime
2520 to be used. Currently, this option is only available in conjunction with
2521 the NeXT runtime on Mac OS X 10.3 and later.
2523 @item -fobjc-call-cxx-cdtors
2524 @opindex fobjc-call-cxx-cdtors
2525 For each Objective-C class, check if any of its instance variables is a
2526 C++ object with a non-trivial default constructor. If so, synthesize a
2527 special @code{- (id) .cxx_construct} instance method that will run
2528 non-trivial default constructors on any such instance variables, in order,
2529 and then return @code{self}. Similarly, check if any instance variable
2530 is a C++ object with a non-trivial destructor, and if so, synthesize a
2531 special @code{- (void) .cxx_destruct} method that will run
2532 all such default destructors, in reverse order.
2534 The @code{- (id) .cxx_construct} and/or @code{- (void) .cxx_destruct} methods
2535 thusly generated will only operate on instance variables declared in the
2536 current Objective-C class, and not those inherited from superclasses. It
2537 is the responsibility of the Objective-C runtime to invoke all such methods
2538 in an object's inheritance hierarchy. The @code{- (id) .cxx_construct} methods
2539 will be invoked by the runtime immediately after a new object
2540 instance is allocated; the @code{- (void) .cxx_destruct} methods will
2541 be invoked immediately before the runtime deallocates an object instance.
2543 As of this writing, only the NeXT runtime on Mac OS X 10.4 and later has
2544 support for invoking the @code{- (id) .cxx_construct} and
2545 @code{- (void) .cxx_destruct} methods.
2547 @item -fobjc-direct-dispatch
2548 @opindex fobjc-direct-dispatch
2549 Allow fast jumps to the message dispatcher. On Darwin this is
2550 accomplished via the comm page.
2552 @item -fobjc-exceptions
2553 @opindex fobjc-exceptions
2554 Enable syntactic support for structured exception handling in Objective-C,
2555 similar to what is offered by C++ and Java. This option is
2556 unavailable in conjunction with the NeXT runtime on Mac OS X 10.2 and
2565 @@catch (AnObjCClass *exc) @{
2572 @@catch (AnotherClass *exc) @{
2575 @@catch (id allOthers) @{
2585 The @code{@@throw} statement may appear anywhere in an Objective-C or
2586 Objective-C++ program; when used inside of a @code{@@catch} block, the
2587 @code{@@throw} may appear without an argument (as shown above), in which case
2588 the object caught by the @code{@@catch} will be rethrown.
2590 Note that only (pointers to) Objective-C objects may be thrown and
2591 caught using this scheme. When an object is thrown, it will be caught
2592 by the nearest @code{@@catch} clause capable of handling objects of that type,
2593 analogously to how @code{catch} blocks work in C++ and Java. A
2594 @code{@@catch(id @dots{})} clause (as shown above) may also be provided to catch
2595 any and all Objective-C exceptions not caught by previous @code{@@catch}
2598 The @code{@@finally} clause, if present, will be executed upon exit from the
2599 immediately preceding @code{@@try @dots{} @@catch} section. This will happen
2600 regardless of whether any exceptions are thrown, caught or rethrown
2601 inside the @code{@@try @dots{} @@catch} section, analogously to the behavior
2602 of the @code{finally} clause in Java.
2604 There are several caveats to using the new exception mechanism:
2608 Although currently designed to be binary compatible with @code{NS_HANDLER}-style
2609 idioms provided by the @code{NSException} class, the new
2610 exceptions can only be used on Mac OS X 10.3 (Panther) and later
2611 systems, due to additional functionality needed in the (NeXT) Objective-C
2615 As mentioned above, the new exceptions do not support handling
2616 types other than Objective-C objects. Furthermore, when used from
2617 Objective-C++, the Objective-C exception model does not interoperate with C++
2618 exceptions at this time. This means you cannot @code{@@throw} an exception
2619 from Objective-C and @code{catch} it in C++, or vice versa
2620 (i.e., @code{throw @dots{} @@catch}).
2623 The @option{-fobjc-exceptions} switch also enables the use of synchronization
2624 blocks for thread-safe execution:
2627 @@synchronized (ObjCClass *guard) @{
2632 Upon entering the @code{@@synchronized} block, a thread of execution shall
2633 first check whether a lock has been placed on the corresponding @code{guard}
2634 object by another thread. If it has, the current thread shall wait until
2635 the other thread relinquishes its lock. Once @code{guard} becomes available,
2636 the current thread will place its own lock on it, execute the code contained in
2637 the @code{@@synchronized} block, and finally relinquish the lock (thereby
2638 making @code{guard} available to other threads).
2640 Unlike Java, Objective-C does not allow for entire methods to be marked
2641 @code{@@synchronized}. Note that throwing exceptions out of
2642 @code{@@synchronized} blocks is allowed, and will cause the guarding object
2643 to be unlocked properly.
2647 Enable garbage collection (GC) in Objective-C and Objective-C++ programs.
2649 @item -freplace-objc-classes
2650 @opindex freplace-objc-classes
2651 Emit a special marker instructing @command{ld(1)} not to statically link in
2652 the resulting object file, and allow @command{dyld(1)} to load it in at
2653 run time instead. This is used in conjunction with the Fix-and-Continue
2654 debugging mode, where the object file in question may be recompiled and
2655 dynamically reloaded in the course of program execution, without the need
2656 to restart the program itself. Currently, Fix-and-Continue functionality
2657 is only available in conjunction with the NeXT runtime on Mac OS X 10.3
2662 When compiling for the NeXT runtime, the compiler ordinarily replaces calls
2663 to @code{objc_getClass("@dots{}")} (when the name of the class is known at
2664 compile time) with static class references that get initialized at load time,
2665 which improves run-time performance. Specifying the @option{-fzero-link} flag
2666 suppresses this behavior and causes calls to @code{objc_getClass("@dots{}")}
2667 to be retained. This is useful in Zero-Link debugging mode, since it allows
2668 for individual class implementations to be modified during program execution.
2672 Dump interface declarations for all classes seen in the source file to a
2673 file named @file{@var{sourcename}.decl}.
2675 @item -Wassign-intercept @r{(Objective-C and Objective-C++ only)}
2676 @opindex Wassign-intercept
2677 @opindex Wno-assign-intercept
2678 Warn whenever an Objective-C assignment is being intercepted by the
2681 @item -Wno-protocol @r{(Objective-C and Objective-C++ only)}
2682 @opindex Wno-protocol
2684 If a class is declared to implement a protocol, a warning is issued for
2685 every method in the protocol that is not implemented by the class. The
2686 default behavior is to issue a warning for every method not explicitly
2687 implemented in the class, even if a method implementation is inherited
2688 from the superclass. If you use the @option{-Wno-protocol} option, then
2689 methods inherited from the superclass are considered to be implemented,
2690 and no warning is issued for them.
2692 @item -Wselector @r{(Objective-C and Objective-C++ only)}
2694 @opindex Wno-selector
2695 Warn if multiple methods of different types for the same selector are
2696 found during compilation. The check is performed on the list of methods
2697 in the final stage of compilation. Additionally, a check is performed
2698 for each selector appearing in a @code{@@selector(@dots{})}
2699 expression, and a corresponding method for that selector has been found
2700 during compilation. Because these checks scan the method table only at
2701 the end of compilation, these warnings are not produced if the final
2702 stage of compilation is not reached, for example because an error is
2703 found during compilation, or because the @option{-fsyntax-only} option is
2706 @item -Wstrict-selector-match @r{(Objective-C and Objective-C++ only)}
2707 @opindex Wstrict-selector-match
2708 @opindex Wno-strict-selector-match
2709 Warn if multiple methods with differing argument and/or return types are
2710 found for a given selector when attempting to send a message using this
2711 selector to a receiver of type @code{id} or @code{Class}. When this flag
2712 is off (which is the default behavior), the compiler will omit such warnings
2713 if any differences found are confined to types which share the same size
2716 @item -Wundeclared-selector @r{(Objective-C and Objective-C++ only)}
2717 @opindex Wundeclared-selector
2718 @opindex Wno-undeclared-selector
2719 Warn if a @code{@@selector(@dots{})} expression referring to an
2720 undeclared selector is found. A selector is considered undeclared if no
2721 method with that name has been declared before the
2722 @code{@@selector(@dots{})} expression, either explicitly in an
2723 @code{@@interface} or @code{@@protocol} declaration, or implicitly in
2724 an @code{@@implementation} section. This option always performs its
2725 checks as soon as a @code{@@selector(@dots{})} expression is found,
2726 while @option{-Wselector} only performs its checks in the final stage of
2727 compilation. This also enforces the coding style convention
2728 that methods and selectors must be declared before being used.
2730 @item -print-objc-runtime-info
2731 @opindex print-objc-runtime-info
2732 Generate C header describing the largest structure that is passed by
2737 @node Language Independent Options
2738 @section Options to Control Diagnostic Messages Formatting
2739 @cindex options to control diagnostics formatting
2740 @cindex diagnostic messages
2741 @cindex message formatting
2743 Traditionally, diagnostic messages have been formatted irrespective of
2744 the output device's aspect (e.g.@: its width, @dots{}). The options described
2745 below can be used to control the diagnostic messages formatting
2746 algorithm, e.g.@: how many characters per line, how often source location
2747 information should be reported. Right now, only the C++ front end can
2748 honor these options. However it is expected, in the near future, that
2749 the remaining front ends would be able to digest them correctly.
2752 @item -fmessage-length=@var{n}
2753 @opindex fmessage-length
2754 Try to format error messages so that they fit on lines of about @var{n}
2755 characters. The default is 72 characters for @command{g++} and 0 for the rest of
2756 the front ends supported by GCC@. If @var{n} is zero, then no
2757 line-wrapping will be done; each error message will appear on a single
2760 @opindex fdiagnostics-show-location
2761 @item -fdiagnostics-show-location=once
2762 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
2763 reporter to emit @emph{once} source location information; that is, in
2764 case the message is too long to fit on a single physical line and has to
2765 be wrapped, the source location won't be emitted (as prefix) again,
2766 over and over, in subsequent continuation lines. This is the default
2769 @item -fdiagnostics-show-location=every-line
2770 Only meaningful in line-wrapping mode. Instructs the diagnostic
2771 messages reporter to emit the same source location information (as
2772 prefix) for physical lines that result from the process of breaking
2773 a message which is too long to fit on a single line.
2775 @item -fdiagnostics-show-option
2776 @opindex fdiagnostics-show-option
2777 This option instructs the diagnostic machinery to add text to each
2778 diagnostic emitted, which indicates which command line option directly
2779 controls that diagnostic, when such an option is known to the
2780 diagnostic machinery.
2782 @item -Wcoverage-mismatch
2783 @opindex Wcoverage-mismatch
2784 Warn if feedback profiles do not match when using the
2785 @option{-fprofile-use} option.
2786 If a source file was changed between @option{-fprofile-gen} and
2787 @option{-fprofile-use}, the files with the profile feedback can fail
2788 to match the source file and GCC can not use the profile feedback
2789 information. By default, this warning is enabled and is treated as an
2790 error. @option{-Wno-coverage-mismatch} can be used to disable the
2791 warning or @option{-Wno-error=coverage-mismatch} can be used to
2792 disable the error. Disable the error for this warning can result in
2793 poorly optimized code, so disabling the error is useful only in the
2794 case of very minor changes such as bug fixes to an existing code-base.
2795 Completely disabling the warning is not recommended.
2799 @node Warning Options
2800 @section Options to Request or Suppress Warnings
2801 @cindex options to control warnings
2802 @cindex warning messages
2803 @cindex messages, warning
2804 @cindex suppressing warnings
2806 Warnings are diagnostic messages that report constructions which
2807 are not inherently erroneous but which are risky or suggest there
2808 may have been an error.
2810 The following language-independent options do not enable specific
2811 warnings but control the kinds of diagnostics produced by GCC.
2814 @cindex syntax checking
2816 @opindex fsyntax-only
2817 Check the code for syntax errors, but don't do anything beyond that.
2821 Inhibit all warning messages.
2826 Make all warnings into errors.
2831 Make the specified warning into an error. The specifier for a warning
2832 is appended, for example @option{-Werror=switch} turns the warnings
2833 controlled by @option{-Wswitch} into errors. This switch takes a
2834 negative form, to be used to negate @option{-Werror} for specific
2835 warnings, for example @option{-Wno-error=switch} makes
2836 @option{-Wswitch} warnings not be errors, even when @option{-Werror}
2837 is in effect. You can use the @option{-fdiagnostics-show-option}
2838 option to have each controllable warning amended with the option which
2839 controls it, to determine what to use with this option.
2841 Note that specifying @option{-Werror=}@var{foo} automatically implies
2842 @option{-W}@var{foo}. However, @option{-Wno-error=}@var{foo} does not
2845 @item -Wfatal-errors
2846 @opindex Wfatal-errors
2847 @opindex Wno-fatal-errors
2848 This option causes the compiler to abort compilation on the first error
2849 occurred rather than trying to keep going and printing further error
2854 You can request many specific warnings with options beginning
2855 @samp{-W}, for example @option{-Wimplicit} to request warnings on
2856 implicit declarations. Each of these specific warning options also
2857 has a negative form beginning @samp{-Wno-} to turn off warnings; for
2858 example, @option{-Wno-implicit}. This manual lists only one of the
2859 two forms, whichever is not the default. For further,
2860 language-specific options also refer to @ref{C++ Dialect Options} and
2861 @ref{Objective-C and Objective-C++ Dialect Options}.
2863 When an unrecognized warning label is requested (e.g.,
2864 @option{-Wunknown-warning}), GCC will emit a diagnostic stating
2865 that the option is not recognized. However, if the @samp{-Wno-} form
2866 is used, the behavior is slightly different: No diagnostic will be
2867 produced for @option{-Wno-unknown-warning} unless other diagnostics
2868 are being produced. This allows the use of new @option{-Wno-} options
2869 with old compilers, but if something goes wrong, the compiler will
2870 warn that an unrecognized option was used.
2875 Issue all the warnings demanded by strict ISO C and ISO C++;
2876 reject all programs that use forbidden extensions, and some other
2877 programs that do not follow ISO C and ISO C++. For ISO C, follows the
2878 version of the ISO C standard specified by any @option{-std} option used.
2880 Valid ISO C and ISO C++ programs should compile properly with or without
2881 this option (though a rare few will require @option{-ansi} or a
2882 @option{-std} option specifying the required version of ISO C)@. However,
2883 without this option, certain GNU extensions and traditional C and C++
2884 features are supported as well. With this option, they are rejected.
2886 @option{-pedantic} does not cause warning messages for use of the
2887 alternate keywords whose names begin and end with @samp{__}. Pedantic
2888 warnings are also disabled in the expression that follows
2889 @code{__extension__}. However, only system header files should use
2890 these escape routes; application programs should avoid them.
2891 @xref{Alternate Keywords}.
2893 Some users try to use @option{-pedantic} to check programs for strict ISO
2894 C conformance. They soon find that it does not do quite what they want:
2895 it finds some non-ISO practices, but not all---only those for which
2896 ISO C @emph{requires} a diagnostic, and some others for which
2897 diagnostics have been added.
2899 A feature to report any failure to conform to ISO C might be useful in
2900 some instances, but would require considerable additional work and would
2901 be quite different from @option{-pedantic}. We don't have plans to
2902 support such a feature in the near future.
2904 Where the standard specified with @option{-std} represents a GNU
2905 extended dialect of C, such as @samp{gnu90} or @samp{gnu99}, there is a
2906 corresponding @dfn{base standard}, the version of ISO C on which the GNU
2907 extended dialect is based. Warnings from @option{-pedantic} are given
2908 where they are required by the base standard. (It would not make sense
2909 for such warnings to be given only for features not in the specified GNU
2910 C dialect, since by definition the GNU dialects of C include all
2911 features the compiler supports with the given option, and there would be
2912 nothing to warn about.)
2914 @item -pedantic-errors
2915 @opindex pedantic-errors
2916 Like @option{-pedantic}, except that errors are produced rather than
2922 This enables all the warnings about constructions that some users
2923 consider questionable, and that are easy to avoid (or modify to
2924 prevent the warning), even in conjunction with macros. This also
2925 enables some language-specific warnings described in @ref{C++ Dialect
2926 Options} and @ref{Objective-C and Objective-C++ Dialect Options}.
2928 @option{-Wall} turns on the following warning flags:
2930 @gccoptlist{-Waddress @gol
2931 -Warray-bounds @r{(only with} @option{-O2}@r{)} @gol
2933 -Wchar-subscripts @gol
2934 -Wenum-compare @r{(in C/Objc; this is on by default in C++)} @gol
2935 -Wimplicit-int @r{(C and Objective-C only)} @gol
2936 -Wimplicit-function-declaration @r{(C and Objective-C only)} @gol
2939 -Wmain @r{(only for C/ObjC and unless} @option{-ffreestanding}@r{)} @gol
2940 -Wmissing-braces @gol
2946 -Wsequence-point @gol
2947 -Wsign-compare @r{(only in C++)} @gol
2948 -Wstrict-aliasing @gol
2949 -Wstrict-overflow=1 @gol
2952 -Wuninitialized @gol
2953 -Wunknown-pragmas @gol
2954 -Wunused-function @gol
2957 -Wunused-variable @gol
2958 -Wvolatile-register-var @gol
2961 Note that some warning flags are not implied by @option{-Wall}. Some of
2962 them warn about constructions that users generally do not consider
2963 questionable, but which occasionally you might wish to check for;
2964 others warn about constructions that are necessary or hard to avoid in
2965 some cases, and there is no simple way to modify the code to suppress
2966 the warning. Some of them are enabled by @option{-Wextra} but many of
2967 them must be enabled individually.
2973 This enables some extra warning flags that are not enabled by
2974 @option{-Wall}. (This option used to be called @option{-W}. The older
2975 name is still supported, but the newer name is more descriptive.)
2977 @gccoptlist{-Wclobbered @gol
2979 -Wignored-qualifiers @gol
2980 -Wmissing-field-initializers @gol
2981 -Wmissing-parameter-type @r{(C only)} @gol
2982 -Wold-style-declaration @r{(C only)} @gol
2983 -Woverride-init @gol
2986 -Wuninitialized @gol
2987 -Wunused-parameter @r{(only with} @option{-Wunused} @r{or} @option{-Wall}@r{)} @gol
2988 -Wunused-but-set-parameter @r{(only with} @option{-Wunused} @r{or} @option{-Wall}@r{)} @gol
2991 The option @option{-Wextra} also prints warning messages for the
2997 A pointer is compared against integer zero with @samp{<}, @samp{<=},
2998 @samp{>}, or @samp{>=}.
3001 (C++ only) An enumerator and a non-enumerator both appear in a
3002 conditional expression.
3005 (C++ only) Ambiguous virtual bases.
3008 (C++ only) Subscripting an array which has been declared @samp{register}.
3011 (C++ only) Taking the address of a variable which has been declared
3015 (C++ only) A base class is not initialized in a derived class' copy
3020 @item -Wchar-subscripts
3021 @opindex Wchar-subscripts
3022 @opindex Wno-char-subscripts
3023 Warn if an array subscript has type @code{char}. This is a common cause
3024 of error, as programmers often forget that this type is signed on some
3026 This warning is enabled by @option{-Wall}.
3030 @opindex Wno-comment
3031 Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*}
3032 comment, or whenever a Backslash-Newline appears in a @samp{//} comment.
3033 This warning is enabled by @option{-Wall}.
3036 @r{(C, Objective-C, C++, Objective-C++ and Fortran only)}
3038 Suppress warning messages emitted by @code{#warning} directives.
3043 @opindex ffreestanding
3044 @opindex fno-builtin
3045 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
3046 the arguments supplied have types appropriate to the format string
3047 specified, and that the conversions specified in the format string make
3048 sense. This includes standard functions, and others specified by format
3049 attributes (@pxref{Function Attributes}), in the @code{printf},
3050 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
3051 not in the C standard) families (or other target-specific families).
3052 Which functions are checked without format attributes having been
3053 specified depends on the standard version selected, and such checks of
3054 functions without the attribute specified are disabled by
3055 @option{-ffreestanding} or @option{-fno-builtin}.
3057 The formats are checked against the format features supported by GNU
3058 libc version 2.2. These include all ISO C90 and C99 features, as well
3059 as features from the Single Unix Specification and some BSD and GNU
3060 extensions. Other library implementations may not support all these
3061 features; GCC does not support warning about features that go beyond a
3062 particular library's limitations. However, if @option{-pedantic} is used
3063 with @option{-Wformat}, warnings will be given about format features not
3064 in the selected standard version (but not for @code{strfmon} formats,
3065 since those are not in any version of the C standard). @xref{C Dialect
3066 Options,,Options Controlling C Dialect}.
3068 Since @option{-Wformat} also checks for null format arguments for
3069 several functions, @option{-Wformat} also implies @option{-Wnonnull}.
3071 @option{-Wformat} is included in @option{-Wall}. For more control over some
3072 aspects of format checking, the options @option{-Wformat-y2k},
3073 @option{-Wno-format-extra-args}, @option{-Wno-format-zero-length},
3074 @option{-Wformat-nonliteral}, @option{-Wformat-security}, and
3075 @option{-Wformat=2} are available, but are not included in @option{-Wall}.
3078 @opindex Wformat-y2k
3079 @opindex Wno-format-y2k
3080 If @option{-Wformat} is specified, also warn about @code{strftime}
3081 formats which may yield only a two-digit year.
3083 @item -Wno-format-contains-nul
3084 @opindex Wno-format-contains-nul
3085 @opindex Wformat-contains-nul
3086 If @option{-Wformat} is specified, do not warn about format strings that
3089 @item -Wno-format-extra-args
3090 @opindex Wno-format-extra-args
3091 @opindex Wformat-extra-args
3092 If @option{-Wformat} is specified, do not warn about excess arguments to a
3093 @code{printf} or @code{scanf} format function. The C standard specifies
3094 that such arguments are ignored.
3096 Where the unused arguments lie between used arguments that are
3097 specified with @samp{$} operand number specifications, normally
3098 warnings are still given, since the implementation could not know what
3099 type to pass to @code{va_arg} to skip the unused arguments. However,
3100 in the case of @code{scanf} formats, this option will suppress the
3101 warning if the unused arguments are all pointers, since the Single
3102 Unix Specification says that such unused arguments are allowed.
3104 @item -Wno-format-zero-length @r{(C and Objective-C only)}
3105 @opindex Wno-format-zero-length
3106 @opindex Wformat-zero-length
3107 If @option{-Wformat} is specified, do not warn about zero-length formats.
3108 The C standard specifies that zero-length formats are allowed.
3110 @item -Wformat-nonliteral
3111 @opindex Wformat-nonliteral
3112 @opindex Wno-format-nonliteral
3113 If @option{-Wformat} is specified, also warn if the format string is not a
3114 string literal and so cannot be checked, unless the format function
3115 takes its format arguments as a @code{va_list}.
3117 @item -Wformat-security
3118 @opindex Wformat-security
3119 @opindex Wno-format-security
3120 If @option{-Wformat} is specified, also warn about uses of format
3121 functions that represent possible security problems. At present, this
3122 warns about calls to @code{printf} and @code{scanf} functions where the
3123 format string is not a string literal and there are no format arguments,
3124 as in @code{printf (foo);}. This may be a security hole if the format
3125 string came from untrusted input and contains @samp{%n}. (This is
3126 currently a subset of what @option{-Wformat-nonliteral} warns about, but
3127 in future warnings may be added to @option{-Wformat-security} that are not
3128 included in @option{-Wformat-nonliteral}.)
3132 @opindex Wno-format=2
3133 Enable @option{-Wformat} plus format checks not included in
3134 @option{-Wformat}. Currently equivalent to @samp{-Wformat
3135 -Wformat-nonliteral -Wformat-security -Wformat-y2k}.
3137 @item -Wnonnull @r{(C and Objective-C only)}
3139 @opindex Wno-nonnull
3140 Warn about passing a null pointer for arguments marked as
3141 requiring a non-null value by the @code{nonnull} function attribute.
3143 @option{-Wnonnull} is included in @option{-Wall} and @option{-Wformat}. It
3144 can be disabled with the @option{-Wno-nonnull} option.
3146 @item -Winit-self @r{(C, C++, Objective-C and Objective-C++ only)}
3148 @opindex Wno-init-self
3149 Warn about uninitialized variables which are initialized with themselves.
3150 Note this option can only be used with the @option{-Wuninitialized} option.
3152 For example, GCC will warn about @code{i} being uninitialized in the
3153 following snippet only when @option{-Winit-self} has been specified:
3164 @item -Wimplicit-int @r{(C and Objective-C only)}
3165 @opindex Wimplicit-int
3166 @opindex Wno-implicit-int
3167 Warn when a declaration does not specify a type.
3168 This warning is enabled by @option{-Wall}.
3170 @item -Wimplicit-function-declaration @r{(C and Objective-C only)}
3171 @opindex Wimplicit-function-declaration
3172 @opindex Wno-implicit-function-declaration
3173 Give a warning whenever a function is used before being declared. In
3174 C99 mode (@option{-std=c99} or @option{-std=gnu99}), this warning is
3175 enabled by default and it is made into an error by
3176 @option{-pedantic-errors}. This warning is also enabled by
3179 @item -Wimplicit @r{(C and Objective-C only)}
3181 @opindex Wno-implicit
3182 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
3183 This warning is enabled by @option{-Wall}.
3185 @item -Wignored-qualifiers @r{(C and C++ only)}
3186 @opindex Wignored-qualifiers
3187 @opindex Wno-ignored-qualifiers
3188 Warn if the return type of a function has a type qualifier
3189 such as @code{const}. For ISO C such a type qualifier has no effect,
3190 since the value returned by a function is not an lvalue.
3191 For C++, the warning is only emitted for scalar types or @code{void}.
3192 ISO C prohibits qualified @code{void} return types on function
3193 definitions, so such return types always receive a warning
3194 even without this option.
3196 This warning is also enabled by @option{-Wextra}.
3201 Warn if the type of @samp{main} is suspicious. @samp{main} should be
3202 a function with external linkage, returning int, taking either zero
3203 arguments, two, or three arguments of appropriate types. This warning
3204 is enabled by default in C++ and is enabled by either @option{-Wall}
3205 or @option{-pedantic}.
3207 @item -Wmissing-braces
3208 @opindex Wmissing-braces
3209 @opindex Wno-missing-braces
3210 Warn if an aggregate or union initializer is not fully bracketed. In
3211 the following example, the initializer for @samp{a} is not fully
3212 bracketed, but that for @samp{b} is fully bracketed.
3215 int a[2][2] = @{ 0, 1, 2, 3 @};
3216 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
3219 This warning is enabled by @option{-Wall}.
3221 @item -Wmissing-include-dirs @r{(C, C++, Objective-C and Objective-C++ only)}
3222 @opindex Wmissing-include-dirs
3223 @opindex Wno-missing-include-dirs
3224 Warn if a user-supplied include directory does not exist.
3227 @opindex Wparentheses
3228 @opindex Wno-parentheses
3229 Warn if parentheses are omitted in certain contexts, such
3230 as when there is an assignment in a context where a truth value
3231 is expected, or when operators are nested whose precedence people
3232 often get confused about.
3234 Also warn if a comparison like @samp{x<=y<=z} appears; this is
3235 equivalent to @samp{(x<=y ? 1 : 0) <= z}, which is a different
3236 interpretation from that of ordinary mathematical notation.
3238 Also warn about constructions where there may be confusion to which
3239 @code{if} statement an @code{else} branch belongs. Here is an example of
3254 In C/C++, every @code{else} branch belongs to the innermost possible
3255 @code{if} statement, which in this example is @code{if (b)}. This is
3256 often not what the programmer expected, as illustrated in the above
3257 example by indentation the programmer chose. When there is the
3258 potential for this confusion, GCC will issue a warning when this flag
3259 is specified. To eliminate the warning, add explicit braces around
3260 the innermost @code{if} statement so there is no way the @code{else}
3261 could belong to the enclosing @code{if}. The resulting code would
3278 Also warn for dangerous uses of the
3279 ?: with omitted middle operand GNU extension. When the condition
3280 in the ?: operator is a boolean expression the omitted value will
3281 be always 1. Often the user expects it to be a value computed
3282 inside the conditional expression instead.
3284 This warning is enabled by @option{-Wall}.
3286 @item -Wsequence-point
3287 @opindex Wsequence-point
3288 @opindex Wno-sequence-point
3289 Warn about code that may have undefined semantics because of violations
3290 of sequence point rules in the C and C++ standards.
3292 The C and C++ standards defines the order in which expressions in a C/C++
3293 program are evaluated in terms of @dfn{sequence points}, which represent
3294 a partial ordering between the execution of parts of the program: those
3295 executed before the sequence point, and those executed after it. These
3296 occur after the evaluation of a full expression (one which is not part
3297 of a larger expression), after the evaluation of the first operand of a
3298 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
3299 function is called (but after the evaluation of its arguments and the
3300 expression denoting the called function), and in certain other places.
3301 Other than as expressed by the sequence point rules, the order of
3302 evaluation of subexpressions of an expression is not specified. All
3303 these rules describe only a partial order rather than a total order,
3304 since, for example, if two functions are called within one expression
3305 with no sequence point between them, the order in which the functions
3306 are called is not specified. However, the standards committee have
3307 ruled that function calls do not overlap.
3309 It is not specified when between sequence points modifications to the
3310 values of objects take effect. Programs whose behavior depends on this
3311 have undefined behavior; the C and C++ standards specify that ``Between
3312 the previous and next sequence point an object shall have its stored
3313 value modified at most once by the evaluation of an expression.
3314 Furthermore, the prior value shall be read only to determine the value
3315 to be stored.''. If a program breaks these rules, the results on any
3316 particular implementation are entirely unpredictable.
3318 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
3319 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
3320 diagnosed by this option, and it may give an occasional false positive
3321 result, but in general it has been found fairly effective at detecting
3322 this sort of problem in programs.
3324 The standard is worded confusingly, therefore there is some debate
3325 over the precise meaning of the sequence point rules in subtle cases.
3326 Links to discussions of the problem, including proposed formal
3327 definitions, may be found on the GCC readings page, at
3328 @w{@uref{http://gcc.gnu.org/readings.html}}.
3330 This warning is enabled by @option{-Wall} for C and C++.
3333 @opindex Wreturn-type
3334 @opindex Wno-return-type
3335 Warn whenever a function is defined with a return-type that defaults
3336 to @code{int}. Also warn about any @code{return} statement with no
3337 return-value in a function whose return-type is not @code{void}
3338 (falling off the end of the function body is considered returning
3339 without a value), and about a @code{return} statement with an
3340 expression in a function whose return-type is @code{void}.
3342 For C++, a function without return type always produces a diagnostic
3343 message, even when @option{-Wno-return-type} is specified. The only
3344 exceptions are @samp{main} and functions defined in system headers.
3346 This warning is enabled by @option{-Wall}.
3351 Warn whenever a @code{switch} statement has an index of enumerated type
3352 and lacks a @code{case} for one or more of the named codes of that
3353 enumeration. (The presence of a @code{default} label prevents this
3354 warning.) @code{case} labels outside the enumeration range also
3355 provoke warnings when this option is used (even if there is a
3356 @code{default} label).
3357 This warning is enabled by @option{-Wall}.
3359 @item -Wswitch-default
3360 @opindex Wswitch-default
3361 @opindex Wno-switch-default
3362 Warn whenever a @code{switch} statement does not have a @code{default}
3366 @opindex Wswitch-enum
3367 @opindex Wno-switch-enum
3368 Warn whenever a @code{switch} statement has an index of enumerated type
3369 and lacks a @code{case} for one or more of the named codes of that
3370 enumeration. @code{case} labels outside the enumeration range also
3371 provoke warnings when this option is used. The only difference
3372 between @option{-Wswitch} and this option is that this option gives a
3373 warning about an omitted enumeration code even if there is a
3374 @code{default} label.
3376 @item -Wsync-nand @r{(C and C++ only)}
3378 @opindex Wno-sync-nand
3379 Warn when @code{__sync_fetch_and_nand} and @code{__sync_nand_and_fetch}
3380 built-in functions are used. These functions changed semantics in GCC 4.4.
3384 @opindex Wno-trigraphs
3385 Warn if any trigraphs are encountered that might change the meaning of
3386 the program (trigraphs within comments are not warned about).
3387 This warning is enabled by @option{-Wall}.
3389 @item -Wunused-but-set-parameter
3390 @opindex Wunused-but-set-parameter
3391 @opindex Wno-unused-but-set-parameter
3392 Warn whenever a function parameter is assigned to, but otherwise unused
3393 (aside from its declaration).
3395 To suppress this warning use the @samp{unused} attribute
3396 (@pxref{Variable Attributes}).
3398 This warning is also enabled by @option{-Wunused} together with
3401 @item -Wunused-but-set-variable
3402 @opindex Wunused-but-set-variable
3403 @opindex Wno-unused-but-set-variable
3404 Warn whenever a local variable is assigned to, but otherwise unused
3405 (aside from its declaration).
3406 This warning is enabled by @option{-Wall}.
3408 To suppress this warning use the @samp{unused} attribute
3409 (@pxref{Variable Attributes}).
3411 This warning is also enabled by @option{-Wunused}, which is enabled
3414 @item -Wunused-function
3415 @opindex Wunused-function
3416 @opindex Wno-unused-function
3417 Warn whenever a static function is declared but not defined or a
3418 non-inline static function is unused.
3419 This warning is enabled by @option{-Wall}.
3421 @item -Wunused-label
3422 @opindex Wunused-label
3423 @opindex Wno-unused-label
3424 Warn whenever a label is declared but not used.
3425 This warning is enabled by @option{-Wall}.
3427 To suppress this warning use the @samp{unused} attribute
3428 (@pxref{Variable Attributes}).
3430 @item -Wunused-parameter
3431 @opindex Wunused-parameter
3432 @opindex Wno-unused-parameter
3433 Warn whenever a function parameter is unused aside from its declaration.
3435 To suppress this warning use the @samp{unused} attribute
3436 (@pxref{Variable Attributes}).
3438 @item -Wno-unused-result
3439 @opindex Wunused-result
3440 @opindex Wno-unused-result
3441 Do not warn if a caller of a function marked with attribute
3442 @code{warn_unused_result} (@pxref{Variable Attributes}) does not use
3443 its return value. The default is @option{-Wunused-result}.
3445 @item -Wunused-variable
3446 @opindex Wunused-variable
3447 @opindex Wno-unused-variable
3448 Warn whenever a local variable or non-constant static variable is unused
3449 aside from its declaration.
3450 This warning is enabled by @option{-Wall}.
3452 To suppress this warning use the @samp{unused} attribute
3453 (@pxref{Variable Attributes}).
3455 @item -Wunused-value
3456 @opindex Wunused-value
3457 @opindex Wno-unused-value
3458 Warn whenever a statement computes a result that is explicitly not
3459 used. To suppress this warning cast the unused expression to
3460 @samp{void}. This includes an expression-statement or the left-hand
3461 side of a comma expression that contains no side effects. For example,
3462 an expression such as @samp{x[i,j]} will cause a warning, while
3463 @samp{x[(void)i,j]} will not.
3465 This warning is enabled by @option{-Wall}.
3470 All the above @option{-Wunused} options combined.
3472 In order to get a warning about an unused function parameter, you must
3473 either specify @samp{-Wextra -Wunused} (note that @samp{-Wall} implies
3474 @samp{-Wunused}), or separately specify @option{-Wunused-parameter}.
3476 @item -Wuninitialized
3477 @opindex Wuninitialized
3478 @opindex Wno-uninitialized
3479 Warn if an automatic variable is used without first being initialized
3480 or if a variable may be clobbered by a @code{setjmp} call. In C++,
3481 warn if a non-static reference or non-static @samp{const} member
3482 appears in a class without constructors.
3484 If you want to warn about code which uses the uninitialized value of the
3485 variable in its own initializer, use the @option{-Winit-self} option.
3487 These warnings occur for individual uninitialized or clobbered
3488 elements of structure, union or array variables as well as for
3489 variables which are uninitialized or clobbered as a whole. They do
3490 not occur for variables or elements declared @code{volatile}. Because
3491 these warnings depend on optimization, the exact variables or elements
3492 for which there are warnings will depend on the precise optimization
3493 options and version of GCC used.
3495 Note that there may be no warning about a variable that is used only
3496 to compute a value that itself is never used, because such
3497 computations may be deleted by data flow analysis before the warnings
3500 These warnings are made optional because GCC is not smart
3501 enough to see all the reasons why the code might be correct
3502 despite appearing to have an error. Here is one example of how
3523 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
3524 always initialized, but GCC doesn't know this. Here is
3525 another common case:
3530 if (change_y) save_y = y, y = new_y;
3532 if (change_y) y = save_y;
3537 This has no bug because @code{save_y} is used only if it is set.
3539 @cindex @code{longjmp} warnings
3540 This option also warns when a non-volatile automatic variable might be
3541 changed by a call to @code{longjmp}. These warnings as well are possible
3542 only in optimizing compilation.
3544 The compiler sees only the calls to @code{setjmp}. It cannot know
3545 where @code{longjmp} will be called; in fact, a signal handler could
3546 call it at any point in the code. As a result, you may get a warning
3547 even when there is in fact no problem because @code{longjmp} cannot
3548 in fact be called at the place which would cause a problem.
3550 Some spurious warnings can be avoided if you declare all the functions
3551 you use that never return as @code{noreturn}. @xref{Function
3554 This warning is enabled by @option{-Wall} or @option{-Wextra}.
3556 @item -Wunknown-pragmas
3557 @opindex Wunknown-pragmas
3558 @opindex Wno-unknown-pragmas
3559 @cindex warning for unknown pragmas
3560 @cindex unknown pragmas, warning
3561 @cindex pragmas, warning of unknown
3562 Warn when a #pragma directive is encountered which is not understood by
3563 GCC@. If this command line option is used, warnings will even be issued
3564 for unknown pragmas in system header files. This is not the case if
3565 the warnings were only enabled by the @option{-Wall} command line option.
3568 @opindex Wno-pragmas
3570 Do not warn about misuses of pragmas, such as incorrect parameters,
3571 invalid syntax, or conflicts between pragmas. See also
3572 @samp{-Wunknown-pragmas}.
3574 @item -Wstrict-aliasing
3575 @opindex Wstrict-aliasing
3576 @opindex Wno-strict-aliasing
3577 This option is only active when @option{-fstrict-aliasing} is active.
3578 It warns about code which might break the strict aliasing rules that the
3579 compiler is using for optimization. The warning does not catch all
3580 cases, but does attempt to catch the more common pitfalls. It is
3581 included in @option{-Wall}.
3582 It is equivalent to @option{-Wstrict-aliasing=3}
3584 @item -Wstrict-aliasing=n
3585 @opindex Wstrict-aliasing=n
3586 @opindex Wno-strict-aliasing=n
3587 This option is only active when @option{-fstrict-aliasing} is active.
3588 It warns about code which might break the strict aliasing rules that the
3589 compiler is using for optimization.
3590 Higher levels correspond to higher accuracy (fewer false positives).
3591 Higher levels also correspond to more effort, similar to the way -O works.
3592 @option{-Wstrict-aliasing} is equivalent to @option{-Wstrict-aliasing=n},
3595 Level 1: Most aggressive, quick, least accurate.
3596 Possibly useful when higher levels
3597 do not warn but -fstrict-aliasing still breaks the code, as it has very few
3598 false negatives. However, it has many false positives.
3599 Warns for all pointer conversions between possibly incompatible types,
3600 even if never dereferenced. Runs in the frontend only.
3602 Level 2: Aggressive, quick, not too precise.
3603 May still have many false positives (not as many as level 1 though),
3604 and few false negatives (but possibly more than level 1).
3605 Unlike level 1, it only warns when an address is taken. Warns about
3606 incomplete types. Runs in the frontend only.
3608 Level 3 (default for @option{-Wstrict-aliasing}):
3609 Should have very few false positives and few false
3610 negatives. Slightly slower than levels 1 or 2 when optimization is enabled.
3611 Takes care of the common pun+dereference pattern in the frontend:
3612 @code{*(int*)&some_float}.
3613 If optimization is enabled, it also runs in the backend, where it deals
3614 with multiple statement cases using flow-sensitive points-to information.
3615 Only warns when the converted pointer is dereferenced.
3616 Does not warn about incomplete types.
3618 @item -Wstrict-overflow
3619 @itemx -Wstrict-overflow=@var{n}
3620 @opindex Wstrict-overflow
3621 @opindex Wno-strict-overflow
3622 This option is only active when @option{-fstrict-overflow} is active.
3623 It warns about cases where the compiler optimizes based on the
3624 assumption that signed overflow does not occur. Note that it does not
3625 warn about all cases where the code might overflow: it only warns
3626 about cases where the compiler implements some optimization. Thus
3627 this warning depends on the optimization level.
3629 An optimization which assumes that signed overflow does not occur is
3630 perfectly safe if the values of the variables involved are such that
3631 overflow never does, in fact, occur. Therefore this warning can
3632 easily give a false positive: a warning about code which is not
3633 actually a problem. To help focus on important issues, several
3634 warning levels are defined. No warnings are issued for the use of
3635 undefined signed overflow when estimating how many iterations a loop
3636 will require, in particular when determining whether a loop will be
3640 @item -Wstrict-overflow=1
3641 Warn about cases which are both questionable and easy to avoid. For
3642 example: @code{x + 1 > x}; with @option{-fstrict-overflow}, the
3643 compiler will simplify this to @code{1}. This level of
3644 @option{-Wstrict-overflow} is enabled by @option{-Wall}; higher levels
3645 are not, and must be explicitly requested.
3647 @item -Wstrict-overflow=2
3648 Also warn about other cases where a comparison is simplified to a
3649 constant. For example: @code{abs (x) >= 0}. This can only be
3650 simplified when @option{-fstrict-overflow} is in effect, because
3651 @code{abs (INT_MIN)} overflows to @code{INT_MIN}, which is less than
3652 zero. @option{-Wstrict-overflow} (with no level) is the same as
3653 @option{-Wstrict-overflow=2}.
3655 @item -Wstrict-overflow=3
3656 Also warn about other cases where a comparison is simplified. For
3657 example: @code{x + 1 > 1} will be simplified to @code{x > 0}.
3659 @item -Wstrict-overflow=4
3660 Also warn about other simplifications not covered by the above cases.
3661 For example: @code{(x * 10) / 5} will be simplified to @code{x * 2}.
3663 @item -Wstrict-overflow=5
3664 Also warn about cases where the compiler reduces the magnitude of a
3665 constant involved in a comparison. For example: @code{x + 2 > y} will
3666 be simplified to @code{x + 1 >= y}. This is reported only at the
3667 highest warning level because this simplification applies to many
3668 comparisons, so this warning level will give a very large number of
3672 @item -Wsuggest-attribute=@r{[}pure@r{|}const@r{|}noreturn@r{]}
3673 @opindex Wsuggest-attribute=
3674 @opindex Wno-suggest-attribute=
3675 Warn for cases where adding an attribute may be beneficial. The
3676 attributes currently supported are listed below.
3679 @item -Wsuggest-attribute=pure
3680 @itemx -Wsuggest-attribute=const
3681 @itemx -Wsuggest-attribute=noreturn
3682 @opindex Wsuggest-attribute=pure
3683 @opindex Wno-suggest-attribute=pure
3684 @opindex Wsuggest-attribute=const
3685 @opindex Wno-suggest-attribute=const
3686 @opindex Wsuggest-attribute=noreturn
3687 @opindex Wno-suggest-attribute=noreturn
3689 Warn about functions which might be candidates for attributes
3690 @code{pure}, @code{const} or @code{noreturn}. The compiler only warns for
3691 functions visible in other compilation units or (in the case of @code{pure} and
3692 @code{const}) if it cannot prove that the function returns normally. A function
3693 returns normally if it doesn't contain an infinite loop nor returns abnormally
3694 by throwing, calling @code{abort()} or trapping. This analysis requires option
3695 @option{-fipa-pure-const}, which is enabled by default at @option{-O} and
3696 higher. Higher optimization levels improve the accuracy of the analysis.
3699 @item -Warray-bounds
3700 @opindex Wno-array-bounds
3701 @opindex Warray-bounds
3702 This option is only active when @option{-ftree-vrp} is active
3703 (default for @option{-O2} and above). It warns about subscripts to arrays
3704 that are always out of bounds. This warning is enabled by @option{-Wall}.
3706 @item -Wno-div-by-zero
3707 @opindex Wno-div-by-zero
3708 @opindex Wdiv-by-zero
3709 Do not warn about compile-time integer division by zero. Floating point
3710 division by zero is not warned about, as it can be a legitimate way of
3711 obtaining infinities and NaNs.
3713 @item -Wsystem-headers
3714 @opindex Wsystem-headers
3715 @opindex Wno-system-headers
3716 @cindex warnings from system headers
3717 @cindex system headers, warnings from
3718 Print warning messages for constructs found in system header files.
3719 Warnings from system headers are normally suppressed, on the assumption
3720 that they usually do not indicate real problems and would only make the
3721 compiler output harder to read. Using this command line option tells
3722 GCC to emit warnings from system headers as if they occurred in user
3723 code. However, note that using @option{-Wall} in conjunction with this
3724 option will @emph{not} warn about unknown pragmas in system
3725 headers---for that, @option{-Wunknown-pragmas} must also be used.
3728 @opindex Wfloat-equal
3729 @opindex Wno-float-equal
3730 Warn if floating point values are used in equality comparisons.
3732 The idea behind this is that sometimes it is convenient (for the
3733 programmer) to consider floating-point values as approximations to
3734 infinitely precise real numbers. If you are doing this, then you need
3735 to compute (by analyzing the code, or in some other way) the maximum or
3736 likely maximum error that the computation introduces, and allow for it
3737 when performing comparisons (and when producing output, but that's a
3738 different problem). In particular, instead of testing for equality, you
3739 would check to see whether the two values have ranges that overlap; and
3740 this is done with the relational operators, so equality comparisons are
3743 @item -Wtraditional @r{(C and Objective-C only)}
3744 @opindex Wtraditional
3745 @opindex Wno-traditional
3746 Warn about certain constructs that behave differently in traditional and
3747 ISO C@. Also warn about ISO C constructs that have no traditional C
3748 equivalent, and/or problematic constructs which should be avoided.
3752 Macro parameters that appear within string literals in the macro body.
3753 In traditional C macro replacement takes place within string literals,
3754 but does not in ISO C@.
3757 In traditional C, some preprocessor directives did not exist.
3758 Traditional preprocessors would only consider a line to be a directive
3759 if the @samp{#} appeared in column 1 on the line. Therefore
3760 @option{-Wtraditional} warns about directives that traditional C
3761 understands but would ignore because the @samp{#} does not appear as the
3762 first character on the line. It also suggests you hide directives like
3763 @samp{#pragma} not understood by traditional C by indenting them. Some
3764 traditional implementations would not recognize @samp{#elif}, so it
3765 suggests avoiding it altogether.
3768 A function-like macro that appears without arguments.
3771 The unary plus operator.
3774 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating point
3775 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
3776 constants.) Note, these suffixes appear in macros defined in the system
3777 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
3778 Use of these macros in user code might normally lead to spurious
3779 warnings, however GCC's integrated preprocessor has enough context to
3780 avoid warning in these cases.
3783 A function declared external in one block and then used after the end of
3787 A @code{switch} statement has an operand of type @code{long}.
3790 A non-@code{static} function declaration follows a @code{static} one.
3791 This construct is not accepted by some traditional C compilers.
3794 The ISO type of an integer constant has a different width or
3795 signedness from its traditional type. This warning is only issued if
3796 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
3797 typically represent bit patterns, are not warned about.
3800 Usage of ISO string concatenation is detected.
3803 Initialization of automatic aggregates.
3806 Identifier conflicts with labels. Traditional C lacks a separate
3807 namespace for labels.
3810 Initialization of unions. If the initializer is zero, the warning is
3811 omitted. This is done under the assumption that the zero initializer in
3812 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
3813 initializer warnings and relies on default initialization to zero in the
3817 Conversions by prototypes between fixed/floating point values and vice
3818 versa. The absence of these prototypes when compiling with traditional
3819 C would cause serious problems. This is a subset of the possible
3820 conversion warnings, for the full set use @option{-Wtraditional-conversion}.
3823 Use of ISO C style function definitions. This warning intentionally is
3824 @emph{not} issued for prototype declarations or variadic functions
3825 because these ISO C features will appear in your code when using
3826 libiberty's traditional C compatibility macros, @code{PARAMS} and
3827 @code{VPARAMS}. This warning is also bypassed for nested functions
3828 because that feature is already a GCC extension and thus not relevant to
3829 traditional C compatibility.
3832 @item -Wtraditional-conversion @r{(C and Objective-C only)}
3833 @opindex Wtraditional-conversion
3834 @opindex Wno-traditional-conversion
3835 Warn if a prototype causes a type conversion that is different from what
3836 would happen to the same argument in the absence of a prototype. This
3837 includes conversions of fixed point to floating and vice versa, and
3838 conversions changing the width or signedness of a fixed point argument
3839 except when the same as the default promotion.
3841 @item -Wdeclaration-after-statement @r{(C and Objective-C only)}
3842 @opindex Wdeclaration-after-statement
3843 @opindex Wno-declaration-after-statement
3844 Warn when a declaration is found after a statement in a block. This
3845 construct, known from C++, was introduced with ISO C99 and is by default
3846 allowed in GCC@. It is not supported by ISO C90 and was not supported by
3847 GCC versions before GCC 3.0. @xref{Mixed Declarations}.
3852 Warn if an undefined identifier is evaluated in an @samp{#if} directive.
3854 @item -Wno-endif-labels
3855 @opindex Wno-endif-labels
3856 @opindex Wendif-labels
3857 Do not warn whenever an @samp{#else} or an @samp{#endif} are followed by text.
3862 Warn whenever a local variable shadows another local variable, parameter or
3863 global variable or whenever a built-in function is shadowed.
3865 @item -Wlarger-than=@var{len}
3866 @opindex Wlarger-than=@var{len}
3867 @opindex Wlarger-than-@var{len}
3868 Warn whenever an object of larger than @var{len} bytes is defined.
3870 @item -Wframe-larger-than=@var{len}
3871 @opindex Wframe-larger-than
3872 Warn if the size of a function frame is larger than @var{len} bytes.
3873 The computation done to determine the stack frame size is approximate
3874 and not conservative.
3875 The actual requirements may be somewhat greater than @var{len}
3876 even if you do not get a warning. In addition, any space allocated
3877 via @code{alloca}, variable-length arrays, or related constructs
3878 is not included by the compiler when determining
3879 whether or not to issue a warning.
3881 @item -Wunsafe-loop-optimizations
3882 @opindex Wunsafe-loop-optimizations
3883 @opindex Wno-unsafe-loop-optimizations
3884 Warn if the loop cannot be optimized because the compiler could not
3885 assume anything on the bounds of the loop indices. With
3886 @option{-funsafe-loop-optimizations} warn if the compiler made
3889 @item -Wno-pedantic-ms-format @r{(MinGW targets only)}
3890 @opindex Wno-pedantic-ms-format
3891 @opindex Wpedantic-ms-format
3892 Disables the warnings about non-ISO @code{printf} / @code{scanf} format
3893 width specifiers @code{I32}, @code{I64}, and @code{I} used on Windows targets
3894 depending on the MS runtime, when you are using the options @option{-Wformat}
3895 and @option{-pedantic} without gnu-extensions.
3897 @item -Wpointer-arith
3898 @opindex Wpointer-arith
3899 @opindex Wno-pointer-arith
3900 Warn about anything that depends on the ``size of'' a function type or
3901 of @code{void}. GNU C assigns these types a size of 1, for
3902 convenience in calculations with @code{void *} pointers and pointers
3903 to functions. In C++, warn also when an arithmetic operation involves
3904 @code{NULL}. This warning is also enabled by @option{-pedantic}.
3907 @opindex Wtype-limits
3908 @opindex Wno-type-limits
3909 Warn if a comparison is always true or always false due to the limited
3910 range of the data type, but do not warn for constant expressions. For
3911 example, warn if an unsigned variable is compared against zero with
3912 @samp{<} or @samp{>=}. This warning is also enabled by
3915 @item -Wbad-function-cast @r{(C and Objective-C only)}
3916 @opindex Wbad-function-cast
3917 @opindex Wno-bad-function-cast
3918 Warn whenever a function call is cast to a non-matching type.
3919 For example, warn if @code{int malloc()} is cast to @code{anything *}.
3921 @item -Wc++-compat @r{(C and Objective-C only)}
3922 Warn about ISO C constructs that are outside of the common subset of
3923 ISO C and ISO C++, e.g.@: request for implicit conversion from
3924 @code{void *} to a pointer to non-@code{void} type.
3926 @item -Wc++0x-compat @r{(C++ and Objective-C++ only)}
3927 Warn about C++ constructs whose meaning differs between ISO C++ 1998 and
3928 ISO C++ 200x, e.g., identifiers in ISO C++ 1998 that will become keywords
3929 in ISO C++ 200x. This warning is enabled by @option{-Wall}.
3933 @opindex Wno-cast-qual
3934 Warn whenever a pointer is cast so as to remove a type qualifier from
3935 the target type. For example, warn if a @code{const char *} is cast
3936 to an ordinary @code{char *}.
3938 Also warn when making a cast which introduces a type qualifier in an
3939 unsafe way. For example, casting @code{char **} to @code{const char **}
3940 is unsafe, as in this example:
3943 /* p is char ** value. */
3944 const char **q = (const char **) p;
3945 /* Assignment of readonly string to const char * is OK. */
3947 /* Now char** pointer points to read-only memory. */
3952 @opindex Wcast-align
3953 @opindex Wno-cast-align
3954 Warn whenever a pointer is cast such that the required alignment of the
3955 target is increased. For example, warn if a @code{char *} is cast to
3956 an @code{int *} on machines where integers can only be accessed at
3957 two- or four-byte boundaries.
3959 @item -Wwrite-strings
3960 @opindex Wwrite-strings
3961 @opindex Wno-write-strings
3962 When compiling C, give string constants the type @code{const
3963 char[@var{length}]} so that copying the address of one into a
3964 non-@code{const} @code{char *} pointer will get a warning. These
3965 warnings will help you find at compile time code that can try to write
3966 into a string constant, but only if you have been very careful about
3967 using @code{const} in declarations and prototypes. Otherwise, it will
3968 just be a nuisance. This is why we did not make @option{-Wall} request
3971 When compiling C++, warn about the deprecated conversion from string
3972 literals to @code{char *}. This warning is enabled by default for C++
3977 @opindex Wno-clobbered
3978 Warn for variables that might be changed by @samp{longjmp} or
3979 @samp{vfork}. This warning is also enabled by @option{-Wextra}.
3982 @opindex Wconversion
3983 @opindex Wno-conversion
3984 Warn for implicit conversions that may alter a value. This includes
3985 conversions between real and integer, like @code{abs (x)} when
3986 @code{x} is @code{double}; conversions between signed and unsigned,
3987 like @code{unsigned ui = -1}; and conversions to smaller types, like
3988 @code{sqrtf (M_PI)}. Do not warn for explicit casts like @code{abs
3989 ((int) x)} and @code{ui = (unsigned) -1}, or if the value is not
3990 changed by the conversion like in @code{abs (2.0)}. Warnings about
3991 conversions between signed and unsigned integers can be disabled by
3992 using @option{-Wno-sign-conversion}.
3994 For C++, also warn for confusing overload resolution for user-defined
3995 conversions; and conversions that will never use a type conversion
3996 operator: conversions to @code{void}, the same type, a base class or a
3997 reference to them. Warnings about conversions between signed and
3998 unsigned integers are disabled by default in C++ unless
3999 @option{-Wsign-conversion} is explicitly enabled.
4001 @item -Wno-conversion-null @r{(C++ and Objective-C++ only)}
4002 @opindex Wconversion-null
4003 @opindex Wno-conversion-null
4004 Do not warn for conversions between @code{NULL} and non-pointer
4005 types. @option{-Wconversion-null} is enabled by default.
4008 @opindex Wempty-body
4009 @opindex Wno-empty-body
4010 Warn if an empty body occurs in an @samp{if}, @samp{else} or @samp{do
4011 while} statement. This warning is also enabled by @option{-Wextra}.
4013 @item -Wenum-compare
4014 @opindex Wenum-compare
4015 @opindex Wno-enum-compare
4016 Warn about a comparison between values of different enum types. In C++
4017 this warning is enabled by default. In C this warning is enabled by
4020 @item -Wjump-misses-init @r{(C, Objective-C only)}
4021 @opindex Wjump-misses-init
4022 @opindex Wno-jump-misses-init
4023 Warn if a @code{goto} statement or a @code{switch} statement jumps
4024 forward across the initialization of a variable, or jumps backward to a
4025 label after the variable has been initialized. This only warns about
4026 variables which are initialized when they are declared. This warning is
4027 only supported for C and Objective C; in C++ this sort of branch is an
4030 @option{-Wjump-misses-init} is included in @option{-Wc++-compat}. It
4031 can be disabled with the @option{-Wno-jump-misses-init} option.
4033 @item -Wsign-compare
4034 @opindex Wsign-compare
4035 @opindex Wno-sign-compare
4036 @cindex warning for comparison of signed and unsigned values
4037 @cindex comparison of signed and unsigned values, warning
4038 @cindex signed and unsigned values, comparison warning
4039 Warn when a comparison between signed and unsigned values could produce
4040 an incorrect result when the signed value is converted to unsigned.
4041 This warning is also enabled by @option{-Wextra}; to get the other warnings
4042 of @option{-Wextra} without this warning, use @samp{-Wextra -Wno-sign-compare}.
4044 @item -Wsign-conversion
4045 @opindex Wsign-conversion
4046 @opindex Wno-sign-conversion
4047 Warn for implicit conversions that may change the sign of an integer
4048 value, like assigning a signed integer expression to an unsigned
4049 integer variable. An explicit cast silences the warning. In C, this
4050 option is enabled also by @option{-Wconversion}.
4054 @opindex Wno-address
4055 Warn about suspicious uses of memory addresses. These include using
4056 the address of a function in a conditional expression, such as
4057 @code{void func(void); if (func)}, and comparisons against the memory
4058 address of a string literal, such as @code{if (x == "abc")}. Such
4059 uses typically indicate a programmer error: the address of a function
4060 always evaluates to true, so their use in a conditional usually
4061 indicate that the programmer forgot the parentheses in a function
4062 call; and comparisons against string literals result in unspecified
4063 behavior and are not portable in C, so they usually indicate that the
4064 programmer intended to use @code{strcmp}. This warning is enabled by
4068 @opindex Wlogical-op
4069 @opindex Wno-logical-op
4070 Warn about suspicious uses of logical operators in expressions.
4071 This includes using logical operators in contexts where a
4072 bit-wise operator is likely to be expected.
4074 @item -Waggregate-return
4075 @opindex Waggregate-return
4076 @opindex Wno-aggregate-return
4077 Warn if any functions that return structures or unions are defined or
4078 called. (In languages where you can return an array, this also elicits
4081 @item -Wno-attributes
4082 @opindex Wno-attributes
4083 @opindex Wattributes
4084 Do not warn if an unexpected @code{__attribute__} is used, such as
4085 unrecognized attributes, function attributes applied to variables,
4086 etc. This will not stop errors for incorrect use of supported
4089 @item -Wno-builtin-macro-redefined
4090 @opindex Wno-builtin-macro-redefined
4091 @opindex Wbuiltin-macro-redefined
4092 Do not warn if certain built-in macros are redefined. This suppresses
4093 warnings for redefinition of @code{__TIMESTAMP__}, @code{__TIME__},
4094 @code{__DATE__}, @code{__FILE__}, and @code{__BASE_FILE__}.
4096 @item -Wstrict-prototypes @r{(C and Objective-C only)}
4097 @opindex Wstrict-prototypes
4098 @opindex Wno-strict-prototypes
4099 Warn if a function is declared or defined without specifying the
4100 argument types. (An old-style function definition is permitted without
4101 a warning if preceded by a declaration which specifies the argument
4104 @item -Wold-style-declaration @r{(C and Objective-C only)}
4105 @opindex Wold-style-declaration
4106 @opindex Wno-old-style-declaration
4107 Warn for obsolescent usages, according to the C Standard, in a
4108 declaration. For example, warn if storage-class specifiers like
4109 @code{static} are not the first things in a declaration. This warning
4110 is also enabled by @option{-Wextra}.
4112 @item -Wold-style-definition @r{(C and Objective-C only)}
4113 @opindex Wold-style-definition
4114 @opindex Wno-old-style-definition
4115 Warn if an old-style function definition is used. A warning is given
4116 even if there is a previous prototype.
4118 @item -Wmissing-parameter-type @r{(C and Objective-C only)}
4119 @opindex Wmissing-parameter-type
4120 @opindex Wno-missing-parameter-type
4121 A function parameter is declared without a type specifier in K&R-style
4128 This warning is also enabled by @option{-Wextra}.
4130 @item -Wmissing-prototypes @r{(C and Objective-C only)}
4131 @opindex Wmissing-prototypes
4132 @opindex Wno-missing-prototypes
4133 Warn if a global function is defined without a previous prototype
4134 declaration. This warning is issued even if the definition itself
4135 provides a prototype. The aim is to detect global functions that fail
4136 to be declared in header files.
4138 @item -Wmissing-declarations
4139 @opindex Wmissing-declarations
4140 @opindex Wno-missing-declarations
4141 Warn if a global function is defined without a previous declaration.
4142 Do so even if the definition itself provides a prototype.
4143 Use this option to detect global functions that are not declared in
4144 header files. In C++, no warnings are issued for function templates,
4145 or for inline functions, or for functions in anonymous namespaces.
4147 @item -Wmissing-field-initializers
4148 @opindex Wmissing-field-initializers
4149 @opindex Wno-missing-field-initializers
4153 Warn if a structure's initializer has some fields missing. For
4154 example, the following code would cause such a warning, because
4155 @code{x.h} is implicitly zero:
4158 struct s @{ int f, g, h; @};
4159 struct s x = @{ 3, 4 @};
4162 This option does not warn about designated initializers, so the following
4163 modification would not trigger a warning:
4166 struct s @{ int f, g, h; @};
4167 struct s x = @{ .f = 3, .g = 4 @};
4170 This warning is included in @option{-Wextra}. To get other @option{-Wextra}
4171 warnings without this one, use @samp{-Wextra -Wno-missing-field-initializers}.
4173 @item -Wmissing-format-attribute
4174 @opindex Wmissing-format-attribute
4175 @opindex Wno-missing-format-attribute
4178 Warn about function pointers which might be candidates for @code{format}
4179 attributes. Note these are only possible candidates, not absolute ones.
4180 GCC will guess that function pointers with @code{format} attributes that
4181 are used in assignment, initialization, parameter passing or return
4182 statements should have a corresponding @code{format} attribute in the
4183 resulting type. I.e.@: the left-hand side of the assignment or
4184 initialization, the type of the parameter variable, or the return type
4185 of the containing function respectively should also have a @code{format}
4186 attribute to avoid the warning.
4188 GCC will also warn about function definitions which might be
4189 candidates for @code{format} attributes. Again, these are only
4190 possible candidates. GCC will guess that @code{format} attributes
4191 might be appropriate for any function that calls a function like
4192 @code{vprintf} or @code{vscanf}, but this might not always be the
4193 case, and some functions for which @code{format} attributes are
4194 appropriate may not be detected.
4196 @item -Wno-multichar
4197 @opindex Wno-multichar
4199 Do not warn if a multicharacter constant (@samp{'FOOF'}) is used.
4200 Usually they indicate a typo in the user's code, as they have
4201 implementation-defined values, and should not be used in portable code.
4203 @item -Wnormalized=<none|id|nfc|nfkc>
4204 @opindex Wnormalized=
4207 @cindex character set, input normalization
4208 In ISO C and ISO C++, two identifiers are different if they are
4209 different sequences of characters. However, sometimes when characters
4210 outside the basic ASCII character set are used, you can have two
4211 different character sequences that look the same. To avoid confusion,
4212 the ISO 10646 standard sets out some @dfn{normalization rules} which
4213 when applied ensure that two sequences that look the same are turned into
4214 the same sequence. GCC can warn you if you are using identifiers which
4215 have not been normalized; this option controls that warning.
4217 There are four levels of warning that GCC supports. The default is
4218 @option{-Wnormalized=nfc}, which warns about any identifier which is
4219 not in the ISO 10646 ``C'' normalized form, @dfn{NFC}. NFC is the
4220 recommended form for most uses.
4222 Unfortunately, there are some characters which ISO C and ISO C++ allow
4223 in identifiers that when turned into NFC aren't allowable as
4224 identifiers. That is, there's no way to use these symbols in portable
4225 ISO C or C++ and have all your identifiers in NFC@.
4226 @option{-Wnormalized=id} suppresses the warning for these characters.
4227 It is hoped that future versions of the standards involved will correct
4228 this, which is why this option is not the default.
4230 You can switch the warning off for all characters by writing
4231 @option{-Wnormalized=none}. You would only want to do this if you
4232 were using some other normalization scheme (like ``D''), because
4233 otherwise you can easily create bugs that are literally impossible to see.
4235 Some characters in ISO 10646 have distinct meanings but look identical
4236 in some fonts or display methodologies, especially once formatting has
4237 been applied. For instance @code{\u207F}, ``SUPERSCRIPT LATIN SMALL
4238 LETTER N'', will display just like a regular @code{n} which has been
4239 placed in a superscript. ISO 10646 defines the @dfn{NFKC}
4240 normalization scheme to convert all these into a standard form as
4241 well, and GCC will warn if your code is not in NFKC if you use
4242 @option{-Wnormalized=nfkc}. This warning is comparable to warning
4243 about every identifier that contains the letter O because it might be
4244 confused with the digit 0, and so is not the default, but may be
4245 useful as a local coding convention if the programming environment is
4246 unable to be fixed to display these characters distinctly.
4248 @item -Wno-deprecated
4249 @opindex Wno-deprecated
4250 @opindex Wdeprecated
4251 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
4253 @item -Wno-deprecated-declarations
4254 @opindex Wno-deprecated-declarations
4255 @opindex Wdeprecated-declarations
4256 Do not warn about uses of functions (@pxref{Function Attributes}),
4257 variables (@pxref{Variable Attributes}), and types (@pxref{Type
4258 Attributes}) marked as deprecated by using the @code{deprecated}
4262 @opindex Wno-overflow
4264 Do not warn about compile-time overflow in constant expressions.
4266 @item -Woverride-init @r{(C and Objective-C only)}
4267 @opindex Woverride-init
4268 @opindex Wno-override-init
4272 Warn if an initialized field without side effects is overridden when
4273 using designated initializers (@pxref{Designated Inits, , Designated
4276 This warning is included in @option{-Wextra}. To get other
4277 @option{-Wextra} warnings without this one, use @samp{-Wextra
4278 -Wno-override-init}.
4283 Warn if a structure is given the packed attribute, but the packed
4284 attribute has no effect on the layout or size of the structure.
4285 Such structures may be mis-aligned for little benefit. For
4286 instance, in this code, the variable @code{f.x} in @code{struct bar}
4287 will be misaligned even though @code{struct bar} does not itself
4288 have the packed attribute:
4295 @} __attribute__((packed));
4303 @item -Wpacked-bitfield-compat
4304 @opindex Wpacked-bitfield-compat
4305 @opindex Wno-packed-bitfield-compat
4306 The 4.1, 4.2 and 4.3 series of GCC ignore the @code{packed} attribute
4307 on bit-fields of type @code{char}. This has been fixed in GCC 4.4 but
4308 the change can lead to differences in the structure layout. GCC
4309 informs you when the offset of such a field has changed in GCC 4.4.
4310 For example there is no longer a 4-bit padding between field @code{a}
4311 and @code{b} in this structure:
4318 @} __attribute__ ((packed));
4321 This warning is enabled by default. Use
4322 @option{-Wno-packed-bitfield-compat} to disable this warning.
4327 Warn if padding is included in a structure, either to align an element
4328 of the structure or to align the whole structure. Sometimes when this
4329 happens it is possible to rearrange the fields of the structure to
4330 reduce the padding and so make the structure smaller.
4332 @item -Wredundant-decls
4333 @opindex Wredundant-decls
4334 @opindex Wno-redundant-decls
4335 Warn if anything is declared more than once in the same scope, even in
4336 cases where multiple declaration is valid and changes nothing.
4338 @item -Wnested-externs @r{(C and Objective-C only)}
4339 @opindex Wnested-externs
4340 @opindex Wno-nested-externs
4341 Warn if an @code{extern} declaration is encountered within a function.
4346 Warn if a function can not be inlined and it was declared as inline.
4347 Even with this option, the compiler will not warn about failures to
4348 inline functions declared in system headers.
4350 The compiler uses a variety of heuristics to determine whether or not
4351 to inline a function. For example, the compiler takes into account
4352 the size of the function being inlined and the amount of inlining
4353 that has already been done in the current function. Therefore,
4354 seemingly insignificant changes in the source program can cause the
4355 warnings produced by @option{-Winline} to appear or disappear.
4357 @item -Wno-invalid-offsetof @r{(C++ and Objective-C++ only)}
4358 @opindex Wno-invalid-offsetof
4359 @opindex Winvalid-offsetof
4360 Suppress warnings from applying the @samp{offsetof} macro to a non-POD
4361 type. According to the 1998 ISO C++ standard, applying @samp{offsetof}
4362 to a non-POD type is undefined. In existing C++ implementations,
4363 however, @samp{offsetof} typically gives meaningful results even when
4364 applied to certain kinds of non-POD types. (Such as a simple
4365 @samp{struct} that fails to be a POD type only by virtue of having a
4366 constructor.) This flag is for users who are aware that they are
4367 writing nonportable code and who have deliberately chosen to ignore the
4370 The restrictions on @samp{offsetof} may be relaxed in a future version
4371 of the C++ standard.
4373 @item -Wno-int-to-pointer-cast
4374 @opindex Wno-int-to-pointer-cast
4375 @opindex Wint-to-pointer-cast
4376 Suppress warnings from casts to pointer type of an integer of a
4377 different size. In C++, casting to a pointer type of smaller size is
4378 an error. @option{Wint-to-pointer-cast} is enabled by default.
4381 @item -Wno-pointer-to-int-cast @r{(C and Objective-C only)}
4382 @opindex Wno-pointer-to-int-cast
4383 @opindex Wpointer-to-int-cast
4384 Suppress warnings from casts from a pointer to an integer type of a
4388 @opindex Winvalid-pch
4389 @opindex Wno-invalid-pch
4390 Warn if a precompiled header (@pxref{Precompiled Headers}) is found in
4391 the search path but can't be used.
4395 @opindex Wno-long-long
4396 Warn if @samp{long long} type is used. This is enabled by either
4397 @option{-pedantic} or @option{-Wtraditional} in ISO C90 and C++98
4398 modes. To inhibit the warning messages, use @option{-Wno-long-long}.
4400 @item -Wvariadic-macros
4401 @opindex Wvariadic-macros
4402 @opindex Wno-variadic-macros
4403 Warn if variadic macros are used in pedantic ISO C90 mode, or the GNU
4404 alternate syntax when in pedantic ISO C99 mode. This is default.
4405 To inhibit the warning messages, use @option{-Wno-variadic-macros}.
4410 Warn if variable length array is used in the code.
4411 @option{-Wno-vla} will prevent the @option{-pedantic} warning of
4412 the variable length array.
4414 @item -Wvolatile-register-var
4415 @opindex Wvolatile-register-var
4416 @opindex Wno-volatile-register-var
4417 Warn if a register variable is declared volatile. The volatile
4418 modifier does not inhibit all optimizations that may eliminate reads
4419 and/or writes to register variables. This warning is enabled by
4422 @item -Wdisabled-optimization
4423 @opindex Wdisabled-optimization
4424 @opindex Wno-disabled-optimization
4425 Warn if a requested optimization pass is disabled. This warning does
4426 not generally indicate that there is anything wrong with your code; it
4427 merely indicates that GCC's optimizers were unable to handle the code
4428 effectively. Often, the problem is that your code is too big or too
4429 complex; GCC will refuse to optimize programs when the optimization
4430 itself is likely to take inordinate amounts of time.
4432 @item -Wpointer-sign @r{(C and Objective-C only)}
4433 @opindex Wpointer-sign
4434 @opindex Wno-pointer-sign
4435 Warn for pointer argument passing or assignment with different signedness.
4436 This option is only supported for C and Objective-C@. It is implied by
4437 @option{-Wall} and by @option{-pedantic}, which can be disabled with
4438 @option{-Wno-pointer-sign}.
4440 @item -Wstack-protector
4441 @opindex Wstack-protector
4442 @opindex Wno-stack-protector
4443 This option is only active when @option{-fstack-protector} is active. It
4444 warns about functions that will not be protected against stack smashing.
4447 @opindex Wno-mudflap
4448 Suppress warnings about constructs that cannot be instrumented by
4451 @item -Woverlength-strings
4452 @opindex Woverlength-strings
4453 @opindex Wno-overlength-strings
4454 Warn about string constants which are longer than the ``minimum
4455 maximum'' length specified in the C standard. Modern compilers
4456 generally allow string constants which are much longer than the
4457 standard's minimum limit, but very portable programs should avoid
4458 using longer strings.
4460 The limit applies @emph{after} string constant concatenation, and does
4461 not count the trailing NUL@. In C90, the limit was 509 characters; in
4462 C99, it was raised to 4095. C++98 does not specify a normative
4463 minimum maximum, so we do not diagnose overlength strings in C++@.
4465 This option is implied by @option{-pedantic}, and can be disabled with
4466 @option{-Wno-overlength-strings}.
4468 @item -Wunsuffixed-float-constants @r{(C and Objective-C only)}
4469 @opindex Wunsuffixed-float-constants
4471 GCC will issue a warning for any floating constant that does not have
4472 a suffix. When used together with @option{-Wsystem-headers} it will
4473 warn about such constants in system header files. This can be useful
4474 when preparing code to use with the @code{FLOAT_CONST_DECIMAL64} pragma
4475 from the decimal floating-point extension to C99.
4478 @node Debugging Options
4479 @section Options for Debugging Your Program or GCC
4480 @cindex options, debugging
4481 @cindex debugging information options
4483 GCC has various special options that are used for debugging
4484 either your program or GCC:
4489 Produce debugging information in the operating system's native format
4490 (stabs, COFF, XCOFF, or DWARF 2)@. GDB can work with this debugging
4493 On most systems that use stabs format, @option{-g} enables use of extra
4494 debugging information that only GDB can use; this extra information
4495 makes debugging work better in GDB but will probably make other debuggers
4497 refuse to read the program. If you want to control for certain whether
4498 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
4499 @option{-gxcoff+}, @option{-gxcoff}, or @option{-gvms} (see below).
4501 GCC allows you to use @option{-g} with
4502 @option{-O}. The shortcuts taken by optimized code may occasionally
4503 produce surprising results: some variables you declared may not exist
4504 at all; flow of control may briefly move where you did not expect it;
4505 some statements may not be executed because they compute constant
4506 results or their values were already at hand; some statements may
4507 execute in different places because they were moved out of loops.
4509 Nevertheless it proves possible to debug optimized output. This makes
4510 it reasonable to use the optimizer for programs that might have bugs.
4512 The following options are useful when GCC is generated with the
4513 capability for more than one debugging format.
4517 Produce debugging information for use by GDB@. This means to use the
4518 most expressive format available (DWARF 2, stabs, or the native format
4519 if neither of those are supported), including GDB extensions if at all
4524 Produce debugging information in stabs format (if that is supported),
4525 without GDB extensions. This is the format used by DBX on most BSD
4526 systems. On MIPS, Alpha and System V Release 4 systems this option
4527 produces stabs debugging output which is not understood by DBX or SDB@.
4528 On System V Release 4 systems this option requires the GNU assembler.
4530 @item -feliminate-unused-debug-symbols
4531 @opindex feliminate-unused-debug-symbols
4532 Produce debugging information in stabs format (if that is supported),
4533 for only symbols that are actually used.
4535 @item -femit-class-debug-always
4536 Instead of emitting debugging information for a C++ class in only one
4537 object file, emit it in all object files using the class. This option
4538 should be used only with debuggers that are unable to handle the way GCC
4539 normally emits debugging information for classes because using this
4540 option will increase the size of debugging information by as much as a
4545 Produce debugging information in stabs format (if that is supported),
4546 using GNU extensions understood only by the GNU debugger (GDB)@. The
4547 use of these extensions is likely to make other debuggers crash or
4548 refuse to read the program.
4552 Produce debugging information in COFF format (if that is supported).
4553 This is the format used by SDB on most System V systems prior to
4558 Produce debugging information in XCOFF format (if that is supported).
4559 This is the format used by the DBX debugger on IBM RS/6000 systems.
4563 Produce debugging information in XCOFF format (if that is supported),
4564 using GNU extensions understood only by the GNU debugger (GDB)@. The
4565 use of these extensions is likely to make other debuggers crash or
4566 refuse to read the program, and may cause assemblers other than the GNU
4567 assembler (GAS) to fail with an error.
4569 @item -gdwarf-@var{version}
4570 @opindex gdwarf-@var{version}
4571 Produce debugging information in DWARF format (if that is
4572 supported). This is the format used by DBX on IRIX 6. The value
4573 of @var{version} may be either 2, 3 or 4; the default version is 2.
4575 Note that with DWARF version 2 some ports require, and will always
4576 use, some non-conflicting DWARF 3 extensions in the unwind tables.
4578 Version 4 may require GDB 7.0 and @option{-fvar-tracking-assignments}
4579 for maximum benefit.
4581 @item -gstrict-dwarf
4582 @opindex gstrict-dwarf
4583 Disallow using extensions of later DWARF standard version than selected
4584 with @option{-gdwarf-@var{version}}. On most targets using non-conflicting
4585 DWARF extensions from later standard versions is allowed.
4587 @item -gno-strict-dwarf
4588 @opindex gno-strict-dwarf
4589 Allow using extensions of later DWARF standard version than selected with
4590 @option{-gdwarf-@var{version}}.
4594 Produce debugging information in VMS debug format (if that is
4595 supported). This is the format used by DEBUG on VMS systems.
4598 @itemx -ggdb@var{level}
4599 @itemx -gstabs@var{level}
4600 @itemx -gcoff@var{level}
4601 @itemx -gxcoff@var{level}
4602 @itemx -gvms@var{level}
4603 Request debugging information and also use @var{level} to specify how
4604 much information. The default level is 2.
4606 Level 0 produces no debug information at all. Thus, @option{-g0} negates
4609 Level 1 produces minimal information, enough for making backtraces in
4610 parts of the program that you don't plan to debug. This includes
4611 descriptions of functions and external variables, but no information
4612 about local variables and no line numbers.
4614 Level 3 includes extra information, such as all the macro definitions
4615 present in the program. Some debuggers support macro expansion when
4616 you use @option{-g3}.
4618 @option{-gdwarf-2} does not accept a concatenated debug level, because
4619 GCC used to support an option @option{-gdwarf} that meant to generate
4620 debug information in version 1 of the DWARF format (which is very
4621 different from version 2), and it would have been too confusing. That
4622 debug format is long obsolete, but the option cannot be changed now.
4623 Instead use an additional @option{-g@var{level}} option to change the
4624 debug level for DWARF.
4628 Turn off generation of debug info, if leaving out this option would have
4629 generated it, or turn it on at level 2 otherwise. The position of this
4630 argument in the command line does not matter, it takes effect after all
4631 other options are processed, and it does so only once, no matter how
4632 many times it is given. This is mainly intended to be used with
4633 @option{-fcompare-debug}.
4635 @item -fdump-final-insns@r{[}=@var{file}@r{]}
4636 @opindex fdump-final-insns
4637 Dump the final internal representation (RTL) to @var{file}. If the
4638 optional argument is omitted (or if @var{file} is @code{.}), the name
4639 of the dump file will be determined by appending @code{.gkd} to the
4640 compilation output file name.
4642 @item -fcompare-debug@r{[}=@var{opts}@r{]}
4643 @opindex fcompare-debug
4644 @opindex fno-compare-debug
4645 If no error occurs during compilation, run the compiler a second time,
4646 adding @var{opts} and @option{-fcompare-debug-second} to the arguments
4647 passed to the second compilation. Dump the final internal
4648 representation in both compilations, and print an error if they differ.
4650 If the equal sign is omitted, the default @option{-gtoggle} is used.
4652 The environment variable @env{GCC_COMPARE_DEBUG}, if defined, non-empty
4653 and nonzero, implicitly enables @option{-fcompare-debug}. If
4654 @env{GCC_COMPARE_DEBUG} is defined to a string starting with a dash,
4655 then it is used for @var{opts}, otherwise the default @option{-gtoggle}
4658 @option{-fcompare-debug=}, with the equal sign but without @var{opts},
4659 is equivalent to @option{-fno-compare-debug}, which disables the dumping
4660 of the final representation and the second compilation, preventing even
4661 @env{GCC_COMPARE_DEBUG} from taking effect.
4663 To verify full coverage during @option{-fcompare-debug} testing, set
4664 @env{GCC_COMPARE_DEBUG} to say @samp{-fcompare-debug-not-overridden},
4665 which GCC will reject as an invalid option in any actual compilation
4666 (rather than preprocessing, assembly or linking). To get just a
4667 warning, setting @env{GCC_COMPARE_DEBUG} to @samp{-w%n-fcompare-debug
4668 not overridden} will do.
4670 @item -fcompare-debug-second
4671 @opindex fcompare-debug-second
4672 This option is implicitly passed to the compiler for the second
4673 compilation requested by @option{-fcompare-debug}, along with options to
4674 silence warnings, and omitting other options that would cause
4675 side-effect compiler outputs to files or to the standard output. Dump
4676 files and preserved temporary files are renamed so as to contain the
4677 @code{.gk} additional extension during the second compilation, to avoid
4678 overwriting those generated by the first.
4680 When this option is passed to the compiler driver, it causes the
4681 @emph{first} compilation to be skipped, which makes it useful for little
4682 other than debugging the compiler proper.
4684 @item -feliminate-dwarf2-dups
4685 @opindex feliminate-dwarf2-dups
4686 Compress DWARF2 debugging information by eliminating duplicated
4687 information about each symbol. This option only makes sense when
4688 generating DWARF2 debugging information with @option{-gdwarf-2}.
4690 @item -femit-struct-debug-baseonly
4691 Emit debug information for struct-like types
4692 only when the base name of the compilation source file
4693 matches the base name of file in which the struct was defined.
4695 This option substantially reduces the size of debugging information,
4696 but at significant potential loss in type information to the debugger.
4697 See @option{-femit-struct-debug-reduced} for a less aggressive option.
4698 See @option{-femit-struct-debug-detailed} for more detailed control.
4700 This option works only with DWARF 2.
4702 @item -femit-struct-debug-reduced
4703 Emit debug information for struct-like types
4704 only when the base name of the compilation source file
4705 matches the base name of file in which the type was defined,
4706 unless the struct is a template or defined in a system header.
4708 This option significantly reduces the size of debugging information,
4709 with some potential loss in type information to the debugger.
4710 See @option{-femit-struct-debug-baseonly} for a more aggressive option.
4711 See @option{-femit-struct-debug-detailed} for more detailed control.
4713 This option works only with DWARF 2.
4715 @item -femit-struct-debug-detailed@r{[}=@var{spec-list}@r{]}
4716 Specify the struct-like types
4717 for which the compiler will generate debug information.
4718 The intent is to reduce duplicate struct debug information
4719 between different object files within the same program.
4721 This option is a detailed version of
4722 @option{-femit-struct-debug-reduced} and @option{-femit-struct-debug-baseonly},
4723 which will serve for most needs.
4725 A specification has the syntax
4726 [@samp{dir:}|@samp{ind:}][@samp{ord:}|@samp{gen:}](@samp{any}|@samp{sys}|@samp{base}|@samp{none})
4728 The optional first word limits the specification to
4729 structs that are used directly (@samp{dir:}) or used indirectly (@samp{ind:}).
4730 A struct type is used directly when it is the type of a variable, member.
4731 Indirect uses arise through pointers to structs.
4732 That is, when use of an incomplete struct would be legal, the use is indirect.
4734 @samp{struct one direct; struct two * indirect;}.
4736 The optional second word limits the specification to
4737 ordinary structs (@samp{ord:}) or generic structs (@samp{gen:}).
4738 Generic structs are a bit complicated to explain.
4739 For C++, these are non-explicit specializations of template classes,
4740 or non-template classes within the above.
4741 Other programming languages have generics,
4742 but @samp{-femit-struct-debug-detailed} does not yet implement them.
4744 The third word specifies the source files for those
4745 structs for which the compiler will emit debug information.
4746 The values @samp{none} and @samp{any} have the normal meaning.
4747 The value @samp{base} means that
4748 the base of name of the file in which the type declaration appears
4749 must match the base of the name of the main compilation file.
4750 In practice, this means that
4751 types declared in @file{foo.c} and @file{foo.h} will have debug information,
4752 but types declared in other header will not.
4753 The value @samp{sys} means those types satisfying @samp{base}
4754 or declared in system or compiler headers.
4756 You may need to experiment to determine the best settings for your application.
4758 The default is @samp{-femit-struct-debug-detailed=all}.
4760 This option works only with DWARF 2.
4762 @item -fenable-icf-debug
4763 @opindex fenable-icf-debug
4764 Generate additional debug information to support identical code folding (ICF).
4765 This option only works with DWARF version 2 or higher.
4767 @item -fno-merge-debug-strings
4768 @opindex fmerge-debug-strings
4769 @opindex fno-merge-debug-strings
4770 Direct the linker to not merge together strings in the debugging
4771 information which are identical in different object files. Merging is
4772 not supported by all assemblers or linkers. Merging decreases the size
4773 of the debug information in the output file at the cost of increasing
4774 link processing time. Merging is enabled by default.
4776 @item -fdebug-prefix-map=@var{old}=@var{new}
4777 @opindex fdebug-prefix-map
4778 When compiling files in directory @file{@var{old}}, record debugging
4779 information describing them as in @file{@var{new}} instead.
4781 @item -fno-dwarf2-cfi-asm
4782 @opindex fdwarf2-cfi-asm
4783 @opindex fno-dwarf2-cfi-asm
4784 Emit DWARF 2 unwind info as compiler generated @code{.eh_frame} section
4785 instead of using GAS @code{.cfi_*} directives.
4787 @cindex @command{prof}
4790 Generate extra code to write profile information suitable for the
4791 analysis program @command{prof}. You must use this option when compiling
4792 the source files you want data about, and you must also use it when
4795 @cindex @command{gprof}
4798 Generate extra code to write profile information suitable for the
4799 analysis program @command{gprof}. You must use this option when compiling
4800 the source files you want data about, and you must also use it when
4805 Makes the compiler print out each function name as it is compiled, and
4806 print some statistics about each pass when it finishes.
4809 @opindex ftime-report
4810 Makes the compiler print some statistics about the time consumed by each
4811 pass when it finishes.
4814 @opindex fmem-report
4815 Makes the compiler print some statistics about permanent memory
4816 allocation when it finishes.
4818 @item -fpre-ipa-mem-report
4819 @opindex fpre-ipa-mem-report
4820 @item -fpost-ipa-mem-report
4821 @opindex fpost-ipa-mem-report
4822 Makes the compiler print some statistics about permanent memory
4823 allocation before or after interprocedural optimization.
4825 @item -fprofile-arcs
4826 @opindex fprofile-arcs
4827 Add code so that program flow @dfn{arcs} are instrumented. During
4828 execution the program records how many times each branch and call is
4829 executed and how many times it is taken or returns. When the compiled
4830 program exits it saves this data to a file called
4831 @file{@var{auxname}.gcda} for each source file. The data may be used for
4832 profile-directed optimizations (@option{-fbranch-probabilities}), or for
4833 test coverage analysis (@option{-ftest-coverage}). Each object file's
4834 @var{auxname} is generated from the name of the output file, if
4835 explicitly specified and it is not the final executable, otherwise it is
4836 the basename of the source file. In both cases any suffix is removed
4837 (e.g.@: @file{foo.gcda} for input file @file{dir/foo.c}, or
4838 @file{dir/foo.gcda} for output file specified as @option{-o dir/foo.o}).
4839 @xref{Cross-profiling}.
4841 @cindex @command{gcov}
4845 This option is used to compile and link code instrumented for coverage
4846 analysis. The option is a synonym for @option{-fprofile-arcs}
4847 @option{-ftest-coverage} (when compiling) and @option{-lgcov} (when
4848 linking). See the documentation for those options for more details.
4853 Compile the source files with @option{-fprofile-arcs} plus optimization
4854 and code generation options. For test coverage analysis, use the
4855 additional @option{-ftest-coverage} option. You do not need to profile
4856 every source file in a program.
4859 Link your object files with @option{-lgcov} or @option{-fprofile-arcs}
4860 (the latter implies the former).
4863 Run the program on a representative workload to generate the arc profile
4864 information. This may be repeated any number of times. You can run
4865 concurrent instances of your program, and provided that the file system
4866 supports locking, the data files will be correctly updated. Also
4867 @code{fork} calls are detected and correctly handled (double counting
4871 For profile-directed optimizations, compile the source files again with
4872 the same optimization and code generation options plus
4873 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
4874 Control Optimization}).
4877 For test coverage analysis, use @command{gcov} to produce human readable
4878 information from the @file{.gcno} and @file{.gcda} files. Refer to the
4879 @command{gcov} documentation for further information.
4883 With @option{-fprofile-arcs}, for each function of your program GCC
4884 creates a program flow graph, then finds a spanning tree for the graph.
4885 Only arcs that are not on the spanning tree have to be instrumented: the
4886 compiler adds code to count the number of times that these arcs are
4887 executed. When an arc is the only exit or only entrance to a block, the
4888 instrumentation code can be added to the block; otherwise, a new basic
4889 block must be created to hold the instrumentation code.
4892 @item -ftest-coverage
4893 @opindex ftest-coverage
4894 Produce a notes file that the @command{gcov} code-coverage utility
4895 (@pxref{Gcov,, @command{gcov}---a Test Coverage Program}) can use to
4896 show program coverage. Each source file's note file is called
4897 @file{@var{auxname}.gcno}. Refer to the @option{-fprofile-arcs} option
4898 above for a description of @var{auxname} and instructions on how to
4899 generate test coverage data. Coverage data will match the source files
4900 more closely, if you do not optimize.
4902 @item -fdbg-cnt-list
4903 @opindex fdbg-cnt-list
4904 Print the name and the counter upperbound for all debug counters.
4906 @item -fdbg-cnt=@var{counter-value-list}
4908 Set the internal debug counter upperbound. @var{counter-value-list}
4909 is a comma-separated list of @var{name}:@var{value} pairs
4910 which sets the upperbound of each debug counter @var{name} to @var{value}.
4911 All debug counters have the initial upperbound of @var{UINT_MAX},
4912 thus dbg_cnt() returns true always unless the upperbound is set by this option.
4913 e.g. With -fdbg-cnt=dce:10,tail_call:0
4914 dbg_cnt(dce) will return true only for first 10 invocations
4915 and dbg_cnt(tail_call) will return false always.
4917 @item -d@var{letters}
4918 @itemx -fdump-rtl-@var{pass}
4920 Says to make debugging dumps during compilation at times specified by
4921 @var{letters}. This is used for debugging the RTL-based passes of the
4922 compiler. The file names for most of the dumps are made by appending
4923 a pass number and a word to the @var{dumpname}, and the files are
4924 created in the directory of the output file. @var{dumpname} is
4925 generated from the name of the output file, if explicitly specified
4926 and it is not an executable, otherwise it is the basename of the
4927 source file. These switches may have different effects when
4928 @option{-E} is used for preprocessing.
4930 Debug dumps can be enabled with a @option{-fdump-rtl} switch or some
4931 @option{-d} option @var{letters}. Here are the possible
4932 letters for use in @var{pass} and @var{letters}, and their meanings:
4936 @item -fdump-rtl-alignments
4937 @opindex fdump-rtl-alignments
4938 Dump after branch alignments have been computed.
4940 @item -fdump-rtl-asmcons
4941 @opindex fdump-rtl-asmcons
4942 Dump after fixing rtl statements that have unsatisfied in/out constraints.
4944 @item -fdump-rtl-auto_inc_dec
4945 @opindex fdump-rtl-auto_inc_dec
4946 Dump after auto-inc-dec discovery. This pass is only run on
4947 architectures that have auto inc or auto dec instructions.
4949 @item -fdump-rtl-barriers
4950 @opindex fdump-rtl-barriers
4951 Dump after cleaning up the barrier instructions.
4953 @item -fdump-rtl-bbpart
4954 @opindex fdump-rtl-bbpart
4955 Dump after partitioning hot and cold basic blocks.
4957 @item -fdump-rtl-bbro
4958 @opindex fdump-rtl-bbro
4959 Dump after block reordering.
4961 @item -fdump-rtl-btl1
4962 @itemx -fdump-rtl-btl2
4963 @opindex fdump-rtl-btl2
4964 @opindex fdump-rtl-btl2
4965 @option{-fdump-rtl-btl1} and @option{-fdump-rtl-btl2} enable dumping
4966 after the two branch
4967 target load optimization passes.
4969 @item -fdump-rtl-bypass
4970 @opindex fdump-rtl-bypass
4971 Dump after jump bypassing and control flow optimizations.
4973 @item -fdump-rtl-combine
4974 @opindex fdump-rtl-combine
4975 Dump after the RTL instruction combination pass.
4977 @item -fdump-rtl-compgotos
4978 @opindex fdump-rtl-compgotos
4979 Dump after duplicating the computed gotos.
4981 @item -fdump-rtl-ce1
4982 @itemx -fdump-rtl-ce2
4983 @itemx -fdump-rtl-ce3
4984 @opindex fdump-rtl-ce1
4985 @opindex fdump-rtl-ce2
4986 @opindex fdump-rtl-ce3
4987 @option{-fdump-rtl-ce1}, @option{-fdump-rtl-ce2}, and
4988 @option{-fdump-rtl-ce3} enable dumping after the three
4989 if conversion passes.
4991 @itemx -fdump-rtl-cprop_hardreg
4992 @opindex fdump-rtl-cprop_hardreg
4993 Dump after hard register copy propagation.
4995 @itemx -fdump-rtl-csa
4996 @opindex fdump-rtl-csa
4997 Dump after combining stack adjustments.
4999 @item -fdump-rtl-cse1
5000 @itemx -fdump-rtl-cse2
5001 @opindex fdump-rtl-cse1
5002 @opindex fdump-rtl-cse2
5003 @option{-fdump-rtl-cse1} and @option{-fdump-rtl-cse2} enable dumping after
5004 the two common sub-expression elimination passes.
5006 @itemx -fdump-rtl-dce
5007 @opindex fdump-rtl-dce
5008 Dump after the standalone dead code elimination passes.
5010 @itemx -fdump-rtl-dbr
5011 @opindex fdump-rtl-dbr
5012 Dump after delayed branch scheduling.
5014 @item -fdump-rtl-dce1
5015 @itemx -fdump-rtl-dce2
5016 @opindex fdump-rtl-dce1
5017 @opindex fdump-rtl-dce2
5018 @option{-fdump-rtl-dce1} and @option{-fdump-rtl-dce2} enable dumping after
5019 the two dead store elimination passes.
5022 @opindex fdump-rtl-eh
5023 Dump after finalization of EH handling code.
5025 @item -fdump-rtl-eh_ranges
5026 @opindex fdump-rtl-eh_ranges
5027 Dump after conversion of EH handling range regions.
5029 @item -fdump-rtl-expand
5030 @opindex fdump-rtl-expand
5031 Dump after RTL generation.
5033 @item -fdump-rtl-fwprop1
5034 @itemx -fdump-rtl-fwprop2
5035 @opindex fdump-rtl-fwprop1
5036 @opindex fdump-rtl-fwprop2
5037 @option{-fdump-rtl-fwprop1} and @option{-fdump-rtl-fwprop2} enable
5038 dumping after the two forward propagation passes.
5040 @item -fdump-rtl-gcse1
5041 @itemx -fdump-rtl-gcse2
5042 @opindex fdump-rtl-gcse1
5043 @opindex fdump-rtl-gcse2
5044 @option{-fdump-rtl-gcse1} and @option{-fdump-rtl-gcse2} enable dumping
5045 after global common subexpression elimination.
5047 @item -fdump-rtl-init-regs
5048 @opindex fdump-rtl-init-regs
5049 Dump after the initialization of the registers.
5051 @item -fdump-rtl-initvals
5052 @opindex fdump-rtl-initvals
5053 Dump after the computation of the initial value sets.
5055 @itemx -fdump-rtl-into_cfglayout
5056 @opindex fdump-rtl-into_cfglayout
5057 Dump after converting to cfglayout mode.
5059 @item -fdump-rtl-ira
5060 @opindex fdump-rtl-ira
5061 Dump after iterated register allocation.
5063 @item -fdump-rtl-jump
5064 @opindex fdump-rtl-jump
5065 Dump after the second jump optimization.
5067 @item -fdump-rtl-loop2
5068 @opindex fdump-rtl-loop2
5069 @option{-fdump-rtl-loop2} enables dumping after the rtl
5070 loop optimization passes.
5072 @item -fdump-rtl-mach
5073 @opindex fdump-rtl-mach
5074 Dump after performing the machine dependent reorganization pass, if that
5077 @item -fdump-rtl-mode_sw
5078 @opindex fdump-rtl-mode_sw
5079 Dump after removing redundant mode switches.
5081 @item -fdump-rtl-rnreg
5082 @opindex fdump-rtl-rnreg
5083 Dump after register renumbering.
5085 @itemx -fdump-rtl-outof_cfglayout
5086 @opindex fdump-rtl-outof_cfglayout
5087 Dump after converting from cfglayout mode.
5089 @item -fdump-rtl-peephole2
5090 @opindex fdump-rtl-peephole2
5091 Dump after the peephole pass.
5093 @item -fdump-rtl-postreload
5094 @opindex fdump-rtl-postreload
5095 Dump after post-reload optimizations.
5097 @itemx -fdump-rtl-pro_and_epilogue
5098 @opindex fdump-rtl-pro_and_epilogue
5099 Dump after generating the function pro and epilogues.
5101 @item -fdump-rtl-regmove
5102 @opindex fdump-rtl-regmove
5103 Dump after the register move pass.
5105 @item -fdump-rtl-sched1
5106 @itemx -fdump-rtl-sched2
5107 @opindex fdump-rtl-sched1
5108 @opindex fdump-rtl-sched2
5109 @option{-fdump-rtl-sched1} and @option{-fdump-rtl-sched2} enable dumping
5110 after the basic block scheduling passes.
5112 @item -fdump-rtl-see
5113 @opindex fdump-rtl-see
5114 Dump after sign extension elimination.
5116 @item -fdump-rtl-seqabstr
5117 @opindex fdump-rtl-seqabstr
5118 Dump after common sequence discovery.
5120 @item -fdump-rtl-shorten
5121 @opindex fdump-rtl-shorten
5122 Dump after shortening branches.
5124 @item -fdump-rtl-sibling
5125 @opindex fdump-rtl-sibling
5126 Dump after sibling call optimizations.
5128 @item -fdump-rtl-split1
5129 @itemx -fdump-rtl-split2
5130 @itemx -fdump-rtl-split3
5131 @itemx -fdump-rtl-split4
5132 @itemx -fdump-rtl-split5
5133 @opindex fdump-rtl-split1
5134 @opindex fdump-rtl-split2
5135 @opindex fdump-rtl-split3
5136 @opindex fdump-rtl-split4
5137 @opindex fdump-rtl-split5
5138 @option{-fdump-rtl-split1}, @option{-fdump-rtl-split2},
5139 @option{-fdump-rtl-split3}, @option{-fdump-rtl-split4} and
5140 @option{-fdump-rtl-split5} enable dumping after five rounds of
5141 instruction splitting.
5143 @item -fdump-rtl-sms
5144 @opindex fdump-rtl-sms
5145 Dump after modulo scheduling. This pass is only run on some
5148 @item -fdump-rtl-stack
5149 @opindex fdump-rtl-stack
5150 Dump after conversion from GCC's "flat register file" registers to the
5151 x87's stack-like registers. This pass is only run on x86 variants.
5153 @item -fdump-rtl-subreg1
5154 @itemx -fdump-rtl-subreg2
5155 @opindex fdump-rtl-subreg1
5156 @opindex fdump-rtl-subreg2
5157 @option{-fdump-rtl-subreg1} and @option{-fdump-rtl-subreg2} enable dumping after
5158 the two subreg expansion passes.
5160 @item -fdump-rtl-unshare
5161 @opindex fdump-rtl-unshare
5162 Dump after all rtl has been unshared.
5164 @item -fdump-rtl-vartrack
5165 @opindex fdump-rtl-vartrack
5166 Dump after variable tracking.
5168 @item -fdump-rtl-vregs
5169 @opindex fdump-rtl-vregs
5170 Dump after converting virtual registers to hard registers.
5172 @item -fdump-rtl-web
5173 @opindex fdump-rtl-web
5174 Dump after live range splitting.
5176 @item -fdump-rtl-regclass
5177 @itemx -fdump-rtl-subregs_of_mode_init
5178 @itemx -fdump-rtl-subregs_of_mode_finish
5179 @itemx -fdump-rtl-dfinit
5180 @itemx -fdump-rtl-dfinish
5181 @opindex fdump-rtl-regclass
5182 @opindex fdump-rtl-subregs_of_mode_init
5183 @opindex fdump-rtl-subregs_of_mode_finish
5184 @opindex fdump-rtl-dfinit
5185 @opindex fdump-rtl-dfinish
5186 These dumps are defined but always produce empty files.
5188 @item -fdump-rtl-all
5189 @opindex fdump-rtl-all
5190 Produce all the dumps listed above.
5194 Annotate the assembler output with miscellaneous debugging information.
5198 Dump all macro definitions, at the end of preprocessing, in addition to
5203 Produce a core dump whenever an error occurs.
5207 Print statistics on memory usage, at the end of the run, to
5212 Annotate the assembler output with a comment indicating which
5213 pattern and alternative was used. The length of each instruction is
5218 Dump the RTL in the assembler output as a comment before each instruction.
5219 Also turns on @option{-dp} annotation.
5223 For each of the other indicated dump files (@option{-fdump-rtl-@var{pass}}),
5224 dump a representation of the control flow graph suitable for viewing with VCG
5225 to @file{@var{file}.@var{pass}.vcg}.
5229 Just generate RTL for a function instead of compiling it. Usually used
5230 with @option{-fdump-rtl-expand}.
5234 Dump debugging information during parsing, to standard error.
5238 @opindex fdump-noaddr
5239 When doing debugging dumps, suppress address output. This makes it more
5240 feasible to use diff on debugging dumps for compiler invocations with
5241 different compiler binaries and/or different
5242 text / bss / data / heap / stack / dso start locations.
5244 @item -fdump-unnumbered
5245 @opindex fdump-unnumbered
5246 When doing debugging dumps, suppress instruction numbers and address output.
5247 This makes it more feasible to use diff on debugging dumps for compiler
5248 invocations with different options, in particular with and without
5251 @item -fdump-unnumbered-links
5252 @opindex fdump-unnumbered-links
5253 When doing debugging dumps (see @option{-d} option above), suppress
5254 instruction numbers for the links to the previous and next instructions
5257 @item -fdump-translation-unit @r{(C++ only)}
5258 @itemx -fdump-translation-unit-@var{options} @r{(C++ only)}
5259 @opindex fdump-translation-unit
5260 Dump a representation of the tree structure for the entire translation
5261 unit to a file. The file name is made by appending @file{.tu} to the
5262 source file name, and the file is created in the same directory as the
5263 output file. If the @samp{-@var{options}} form is used, @var{options}
5264 controls the details of the dump as described for the
5265 @option{-fdump-tree} options.
5267 @item -fdump-class-hierarchy @r{(C++ only)}
5268 @itemx -fdump-class-hierarchy-@var{options} @r{(C++ only)}
5269 @opindex fdump-class-hierarchy
5270 Dump a representation of each class's hierarchy and virtual function
5271 table layout to a file. The file name is made by appending
5272 @file{.class} to the source file name, and the file is created in the
5273 same directory as the output file. If the @samp{-@var{options}} form
5274 is used, @var{options} controls the details of the dump as described
5275 for the @option{-fdump-tree} options.
5277 @item -fdump-ipa-@var{switch}
5279 Control the dumping at various stages of inter-procedural analysis
5280 language tree to a file. The file name is generated by appending a
5281 switch specific suffix to the source file name, and the file is created
5282 in the same directory as the output file. The following dumps are
5287 Enables all inter-procedural analysis dumps.
5290 Dumps information about call-graph optimization, unused function removal,
5291 and inlining decisions.
5294 Dump after function inlining.
5298 @item -fdump-statistics-@var{option}
5299 @opindex fdump-statistics
5300 Enable and control dumping of pass statistics in a separate file. The
5301 file name is generated by appending a suffix ending in
5302 @samp{.statistics} to the source file name, and the file is created in
5303 the same directory as the output file. If the @samp{-@var{option}}
5304 form is used, @samp{-stats} will cause counters to be summed over the
5305 whole compilation unit while @samp{-details} will dump every event as
5306 the passes generate them. The default with no option is to sum
5307 counters for each function compiled.
5309 @item -fdump-tree-@var{switch}
5310 @itemx -fdump-tree-@var{switch}-@var{options}
5312 Control the dumping at various stages of processing the intermediate
5313 language tree to a file. The file name is generated by appending a
5314 switch specific suffix to the source file name, and the file is
5315 created in the same directory as the output file. If the
5316 @samp{-@var{options}} form is used, @var{options} is a list of
5317 @samp{-} separated options that control the details of the dump. Not
5318 all options are applicable to all dumps, those which are not
5319 meaningful will be ignored. The following options are available
5323 Print the address of each node. Usually this is not meaningful as it
5324 changes according to the environment and source file. Its primary use
5325 is for tying up a dump file with a debug environment.
5327 If @code{DECL_ASSEMBLER_NAME} has been set for a given decl, use that
5328 in the dump instead of @code{DECL_NAME}. Its primary use is ease of
5329 use working backward from mangled names in the assembly file.
5331 Inhibit dumping of members of a scope or body of a function merely
5332 because that scope has been reached. Only dump such items when they
5333 are directly reachable by some other path. When dumping pretty-printed
5334 trees, this option inhibits dumping the bodies of control structures.
5336 Print a raw representation of the tree. By default, trees are
5337 pretty-printed into a C-like representation.
5339 Enable more detailed dumps (not honored by every dump option).
5341 Enable dumping various statistics about the pass (not honored by every dump
5344 Enable showing basic block boundaries (disabled in raw dumps).
5346 Enable showing virtual operands for every statement.
5348 Enable showing line numbers for statements.
5350 Enable showing the unique ID (@code{DECL_UID}) for each variable.
5352 Enable showing the tree dump for each statement.
5354 Enable showing the EH region number holding each statement.
5356 Turn on all options, except @option{raw}, @option{slim}, @option{verbose}
5357 and @option{lineno}.
5360 The following tree dumps are possible:
5364 @opindex fdump-tree-original
5365 Dump before any tree based optimization, to @file{@var{file}.original}.
5368 @opindex fdump-tree-optimized
5369 Dump after all tree based optimization, to @file{@var{file}.optimized}.
5372 @opindex fdump-tree-gimple
5373 Dump each function before and after the gimplification pass to a file. The
5374 file name is made by appending @file{.gimple} to the source file name.
5377 @opindex fdump-tree-cfg
5378 Dump the control flow graph of each function to a file. The file name is
5379 made by appending @file{.cfg} to the source file name.
5382 @opindex fdump-tree-vcg
5383 Dump the control flow graph of each function to a file in VCG format. The
5384 file name is made by appending @file{.vcg} to the source file name. Note
5385 that if the file contains more than one function, the generated file cannot
5386 be used directly by VCG@. You will need to cut and paste each function's
5387 graph into its own separate file first.
5390 @opindex fdump-tree-ch
5391 Dump each function after copying loop headers. The file name is made by
5392 appending @file{.ch} to the source file name.
5395 @opindex fdump-tree-ssa
5396 Dump SSA related information to a file. The file name is made by appending
5397 @file{.ssa} to the source file name.
5400 @opindex fdump-tree-alias
5401 Dump aliasing information for each function. The file name is made by
5402 appending @file{.alias} to the source file name.
5405 @opindex fdump-tree-ccp
5406 Dump each function after CCP@. The file name is made by appending
5407 @file{.ccp} to the source file name.
5410 @opindex fdump-tree-storeccp
5411 Dump each function after STORE-CCP@. The file name is made by appending
5412 @file{.storeccp} to the source file name.
5415 @opindex fdump-tree-pre
5416 Dump trees after partial redundancy elimination. The file name is made
5417 by appending @file{.pre} to the source file name.
5420 @opindex fdump-tree-fre
5421 Dump trees after full redundancy elimination. The file name is made
5422 by appending @file{.fre} to the source file name.
5425 @opindex fdump-tree-copyprop
5426 Dump trees after copy propagation. The file name is made
5427 by appending @file{.copyprop} to the source file name.
5429 @item store_copyprop
5430 @opindex fdump-tree-store_copyprop
5431 Dump trees after store copy-propagation. The file name is made
5432 by appending @file{.store_copyprop} to the source file name.
5435 @opindex fdump-tree-dce
5436 Dump each function after dead code elimination. The file name is made by
5437 appending @file{.dce} to the source file name.
5440 @opindex fdump-tree-mudflap
5441 Dump each function after adding mudflap instrumentation. The file name is
5442 made by appending @file{.mudflap} to the source file name.
5445 @opindex fdump-tree-sra
5446 Dump each function after performing scalar replacement of aggregates. The
5447 file name is made by appending @file{.sra} to the source file name.
5450 @opindex fdump-tree-sink
5451 Dump each function after performing code sinking. The file name is made
5452 by appending @file{.sink} to the source file name.
5455 @opindex fdump-tree-dom
5456 Dump each function after applying dominator tree optimizations. The file
5457 name is made by appending @file{.dom} to the source file name.
5460 @opindex fdump-tree-dse
5461 Dump each function after applying dead store elimination. The file
5462 name is made by appending @file{.dse} to the source file name.
5465 @opindex fdump-tree-phiopt
5466 Dump each function after optimizing PHI nodes into straightline code. The file
5467 name is made by appending @file{.phiopt} to the source file name.
5470 @opindex fdump-tree-forwprop
5471 Dump each function after forward propagating single use variables. The file
5472 name is made by appending @file{.forwprop} to the source file name.
5475 @opindex fdump-tree-copyrename
5476 Dump each function after applying the copy rename optimization. The file
5477 name is made by appending @file{.copyrename} to the source file name.
5480 @opindex fdump-tree-nrv
5481 Dump each function after applying the named return value optimization on
5482 generic trees. The file name is made by appending @file{.nrv} to the source
5486 @opindex fdump-tree-vect
5487 Dump each function after applying vectorization of loops. The file name is
5488 made by appending @file{.vect} to the source file name.
5491 @opindex fdump-tree-slp
5492 Dump each function after applying vectorization of basic blocks. The file name
5493 is made by appending @file{.slp} to the source file name.
5496 @opindex fdump-tree-vrp
5497 Dump each function after Value Range Propagation (VRP). The file name
5498 is made by appending @file{.vrp} to the source file name.
5501 @opindex fdump-tree-all
5502 Enable all the available tree dumps with the flags provided in this option.
5505 @item -ftree-vectorizer-verbose=@var{n}
5506 @opindex ftree-vectorizer-verbose
5507 This option controls the amount of debugging output the vectorizer prints.
5508 This information is written to standard error, unless
5509 @option{-fdump-tree-all} or @option{-fdump-tree-vect} is specified,
5510 in which case it is output to the usual dump listing file, @file{.vect}.
5511 For @var{n}=0 no diagnostic information is reported.
5512 If @var{n}=1 the vectorizer reports each loop that got vectorized,
5513 and the total number of loops that got vectorized.
5514 If @var{n}=2 the vectorizer also reports non-vectorized loops that passed
5515 the first analysis phase (vect_analyze_loop_form) - i.e.@: countable,
5516 inner-most, single-bb, single-entry/exit loops. This is the same verbosity
5517 level that @option{-fdump-tree-vect-stats} uses.
5518 Higher verbosity levels mean either more information dumped for each
5519 reported loop, or same amount of information reported for more loops:
5520 if @var{n}=3, vectorizer cost model information is reported.
5521 If @var{n}=4, alignment related information is added to the reports.
5522 If @var{n}=5, data-references related information (e.g.@: memory dependences,
5523 memory access-patterns) is added to the reports.
5524 If @var{n}=6, the vectorizer reports also non-vectorized inner-most loops
5525 that did not pass the first analysis phase (i.e., may not be countable, or
5526 may have complicated control-flow).
5527 If @var{n}=7, the vectorizer reports also non-vectorized nested loops.
5528 If @var{n}=8, SLP related information is added to the reports.
5529 For @var{n}=9, all the information the vectorizer generates during its
5530 analysis and transformation is reported. This is the same verbosity level
5531 that @option{-fdump-tree-vect-details} uses.
5533 @item -frandom-seed=@var{string}
5534 @opindex frandom-seed
5535 This option provides a seed that GCC uses when it would otherwise use
5536 random numbers. It is used to generate certain symbol names
5537 that have to be different in every compiled file. It is also used to
5538 place unique stamps in coverage data files and the object files that
5539 produce them. You can use the @option{-frandom-seed} option to produce
5540 reproducibly identical object files.
5542 The @var{string} should be different for every file you compile.
5544 @item -fsched-verbose=@var{n}
5545 @opindex fsched-verbose
5546 On targets that use instruction scheduling, this option controls the
5547 amount of debugging output the scheduler prints. This information is
5548 written to standard error, unless @option{-fdump-rtl-sched1} or
5549 @option{-fdump-rtl-sched2} is specified, in which case it is output
5550 to the usual dump listing file, @file{.sched1} or @file{.sched2}
5551 respectively. However for @var{n} greater than nine, the output is
5552 always printed to standard error.
5554 For @var{n} greater than zero, @option{-fsched-verbose} outputs the
5555 same information as @option{-fdump-rtl-sched1} and @option{-fdump-rtl-sched2}.
5556 For @var{n} greater than one, it also output basic block probabilities,
5557 detailed ready list information and unit/insn info. For @var{n} greater
5558 than two, it includes RTL at abort point, control-flow and regions info.
5559 And for @var{n} over four, @option{-fsched-verbose} also includes
5563 @itemx -save-temps=cwd
5565 Store the usual ``temporary'' intermediate files permanently; place them
5566 in the current directory and name them based on the source file. Thus,
5567 compiling @file{foo.c} with @samp{-c -save-temps} would produce files
5568 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
5569 preprocessed @file{foo.i} output file even though the compiler now
5570 normally uses an integrated preprocessor.
5572 When used in combination with the @option{-x} command line option,
5573 @option{-save-temps} is sensible enough to avoid over writing an
5574 input source file with the same extension as an intermediate file.
5575 The corresponding intermediate file may be obtained by renaming the
5576 source file before using @option{-save-temps}.
5578 If you invoke GCC in parallel, compiling several different source
5579 files that share a common base name in different subdirectories or the
5580 same source file compiled for multiple output destinations, it is
5581 likely that the different parallel compilers will interfere with each
5582 other, and overwrite the temporary files. For instance:
5585 gcc -save-temps -o outdir1/foo.o indir1/foo.c&
5586 gcc -save-temps -o outdir2/foo.o indir2/foo.c&
5589 may result in @file{foo.i} and @file{foo.o} being written to
5590 simultaneously by both compilers.
5592 @item -save-temps=obj
5593 @opindex save-temps=obj
5594 Store the usual ``temporary'' intermediate files permanently. If the
5595 @option{-o} option is used, the temporary files are based on the
5596 object file. If the @option{-o} option is not used, the
5597 @option{-save-temps=obj} switch behaves like @option{-save-temps}.
5602 gcc -save-temps=obj -c foo.c
5603 gcc -save-temps=obj -c bar.c -o dir/xbar.o
5604 gcc -save-temps=obj foobar.c -o dir2/yfoobar
5607 would create @file{foo.i}, @file{foo.s}, @file{dir/xbar.i},
5608 @file{dir/xbar.s}, @file{dir2/yfoobar.i}, @file{dir2/yfoobar.s}, and
5609 @file{dir2/yfoobar.o}.
5611 @item -time@r{[}=@var{file}@r{]}
5613 Report the CPU time taken by each subprocess in the compilation
5614 sequence. For C source files, this is the compiler proper and assembler
5615 (plus the linker if linking is done).
5617 Without the specification of an output file, the output looks like this:
5624 The first number on each line is the ``user time'', that is time spent
5625 executing the program itself. The second number is ``system time'',
5626 time spent executing operating system routines on behalf of the program.
5627 Both numbers are in seconds.
5629 With the specification of an output file, the output is appended to the
5630 named file, and it looks like this:
5633 0.12 0.01 cc1 @var{options}
5634 0.00 0.01 as @var{options}
5637 The ``user time'' and the ``system time'' are moved before the program
5638 name, and the options passed to the program are displayed, so that one
5639 can later tell what file was being compiled, and with which options.
5641 @item -fvar-tracking
5642 @opindex fvar-tracking
5643 Run variable tracking pass. It computes where variables are stored at each
5644 position in code. Better debugging information is then generated
5645 (if the debugging information format supports this information).
5647 It is enabled by default when compiling with optimization (@option{-Os},
5648 @option{-O}, @option{-O2}, @dots{}), debugging information (@option{-g}) and
5649 the debug info format supports it.
5651 @item -fvar-tracking-assignments
5652 @opindex fvar-tracking-assignments
5653 @opindex fno-var-tracking-assignments
5654 Annotate assignments to user variables early in the compilation and
5655 attempt to carry the annotations over throughout the compilation all the
5656 way to the end, in an attempt to improve debug information while
5657 optimizing. Use of @option{-gdwarf-4} is recommended along with it.
5659 It can be enabled even if var-tracking is disabled, in which case
5660 annotations will be created and maintained, but discarded at the end.
5662 @item -fvar-tracking-assignments-toggle
5663 @opindex fvar-tracking-assignments-toggle
5664 @opindex fno-var-tracking-assignments-toggle
5665 Toggle @option{-fvar-tracking-assignments}, in the same way that
5666 @option{-gtoggle} toggles @option{-g}.
5668 @item -print-file-name=@var{library}
5669 @opindex print-file-name
5670 Print the full absolute name of the library file @var{library} that
5671 would be used when linking---and don't do anything else. With this
5672 option, GCC does not compile or link anything; it just prints the
5675 @item -print-multi-directory
5676 @opindex print-multi-directory
5677 Print the directory name corresponding to the multilib selected by any
5678 other switches present in the command line. This directory is supposed
5679 to exist in @env{GCC_EXEC_PREFIX}.
5681 @item -print-multi-lib
5682 @opindex print-multi-lib
5683 Print the mapping from multilib directory names to compiler switches
5684 that enable them. The directory name is separated from the switches by
5685 @samp{;}, and each switch starts with an @samp{@@} instead of the
5686 @samp{-}, without spaces between multiple switches. This is supposed to
5687 ease shell-processing.
5689 @item -print-multi-os-directory
5690 @opindex print-multi-os-directory
5691 Print the path to OS libraries for the selected
5692 multilib, relative to some @file{lib} subdirectory. If OS libraries are
5693 present in the @file{lib} subdirectory and no multilibs are used, this is
5694 usually just @file{.}, if OS libraries are present in @file{lib@var{suffix}}
5695 sibling directories this prints e.g.@: @file{../lib64}, @file{../lib} or
5696 @file{../lib32}, or if OS libraries are present in @file{lib/@var{subdir}}
5697 subdirectories it prints e.g.@: @file{amd64}, @file{sparcv9} or @file{ev6}.
5699 @item -print-prog-name=@var{program}
5700 @opindex print-prog-name
5701 Like @option{-print-file-name}, but searches for a program such as @samp{cpp}.
5703 @item -print-libgcc-file-name
5704 @opindex print-libgcc-file-name
5705 Same as @option{-print-file-name=libgcc.a}.
5707 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
5708 but you do want to link with @file{libgcc.a}. You can do
5711 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
5714 @item -print-search-dirs
5715 @opindex print-search-dirs
5716 Print the name of the configured installation directory and a list of
5717 program and library directories @command{gcc} will search---and don't do anything else.
5719 This is useful when @command{gcc} prints the error message
5720 @samp{installation problem, cannot exec cpp0: No such file or directory}.
5721 To resolve this you either need to put @file{cpp0} and the other compiler
5722 components where @command{gcc} expects to find them, or you can set the environment
5723 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
5724 Don't forget the trailing @samp{/}.
5725 @xref{Environment Variables}.
5727 @item -print-sysroot
5728 @opindex print-sysroot
5729 Print the target sysroot directory that will be used during
5730 compilation. This is the target sysroot specified either at configure
5731 time or using the @option{--sysroot} option, possibly with an extra
5732 suffix that depends on compilation options. If no target sysroot is
5733 specified, the option prints nothing.
5735 @item -print-sysroot-headers-suffix
5736 @opindex print-sysroot-headers-suffix
5737 Print the suffix added to the target sysroot when searching for
5738 headers, or give an error if the compiler is not configured with such
5739 a suffix---and don't do anything else.
5742 @opindex dumpmachine
5743 Print the compiler's target machine (for example,
5744 @samp{i686-pc-linux-gnu})---and don't do anything else.
5747 @opindex dumpversion
5748 Print the compiler version (for example, @samp{3.0})---and don't do
5753 Print the compiler's built-in specs---and don't do anything else. (This
5754 is used when GCC itself is being built.) @xref{Spec Files}.
5756 @item -feliminate-unused-debug-types
5757 @opindex feliminate-unused-debug-types
5758 Normally, when producing DWARF2 output, GCC will emit debugging
5759 information for all types declared in a compilation
5760 unit, regardless of whether or not they are actually used
5761 in that compilation unit. Sometimes this is useful, such as
5762 if, in the debugger, you want to cast a value to a type that is
5763 not actually used in your program (but is declared). More often,
5764 however, this results in a significant amount of wasted space.
5765 With this option, GCC will avoid producing debug symbol output
5766 for types that are nowhere used in the source file being compiled.
5769 @node Optimize Options
5770 @section Options That Control Optimization
5771 @cindex optimize options
5772 @cindex options, optimization
5774 These options control various sorts of optimizations.
5776 Without any optimization option, the compiler's goal is to reduce the
5777 cost of compilation and to make debugging produce the expected
5778 results. Statements are independent: if you stop the program with a
5779 breakpoint between statements, you can then assign a new value to any
5780 variable or change the program counter to any other statement in the
5781 function and get exactly the results you would expect from the source
5784 Turning on optimization flags makes the compiler attempt to improve
5785 the performance and/or code size at the expense of compilation time
5786 and possibly the ability to debug the program.
5788 The compiler performs optimization based on the knowledge it has of the
5789 program. Compiling multiple files at once to a single output file mode allows
5790 the compiler to use information gained from all of the files when compiling
5793 Not all optimizations are controlled directly by a flag. Only
5794 optimizations that have a flag are listed in this section.
5796 Most optimizations are only enabled if an @option{-O} level is set on
5797 the command line. Otherwise they are disabled, even if individual
5798 optimization flags are specified.
5800 Depending on the target and how GCC was configured, a slightly different
5801 set of optimizations may be enabled at each @option{-O} level than
5802 those listed here. You can invoke GCC with @samp{-Q --help=optimizers}
5803 to find out the exact set of optimizations that are enabled at each level.
5804 @xref{Overall Options}, for examples.
5811 Optimize. Optimizing compilation takes somewhat more time, and a lot
5812 more memory for a large function.
5814 With @option{-O}, the compiler tries to reduce code size and execution
5815 time, without performing any optimizations that take a great deal of
5818 @option{-O} turns on the following optimization flags:
5821 -fcprop-registers @gol
5824 -fdelayed-branch @gol
5826 -fguess-branch-probability @gol
5827 -fif-conversion2 @gol
5828 -fif-conversion @gol
5829 -fipa-pure-const @gol
5831 -fipa-reference @gol
5833 -fsplit-wide-types @gol
5834 -ftree-builtin-call-dce @gol
5837 -ftree-copyrename @gol
5839 -ftree-dominator-opts @gol
5841 -ftree-forwprop @gol
5849 @option{-O} also turns on @option{-fomit-frame-pointer} on machines
5850 where doing so does not interfere with debugging.
5854 Optimize even more. GCC performs nearly all supported optimizations
5855 that do not involve a space-speed tradeoff.
5856 As compared to @option{-O}, this option increases both compilation time
5857 and the performance of the generated code.
5859 @option{-O2} turns on all optimization flags specified by @option{-O}. It
5860 also turns on the following optimization flags:
5861 @gccoptlist{-fthread-jumps @gol
5862 -falign-functions -falign-jumps @gol
5863 -falign-loops -falign-labels @gol
5866 -fcse-follow-jumps -fcse-skip-blocks @gol
5867 -fdelete-null-pointer-checks @gol
5868 -fexpensive-optimizations @gol
5869 -fgcse -fgcse-lm @gol
5870 -finline-small-functions @gol
5871 -findirect-inlining @gol
5873 -foptimize-sibling-calls @gol
5874 -fpartial-inlining @gol
5877 -freorder-blocks -freorder-functions @gol
5878 -frerun-cse-after-loop @gol
5879 -fsched-interblock -fsched-spec @gol
5880 -fschedule-insns -fschedule-insns2 @gol
5881 -fstrict-aliasing -fstrict-overflow @gol
5882 -ftree-switch-conversion @gol
5886 Please note the warning under @option{-fgcse} about
5887 invoking @option{-O2} on programs that use computed gotos.
5891 Optimize yet more. @option{-O3} turns on all optimizations specified
5892 by @option{-O2} and also turns on the @option{-finline-functions},
5893 @option{-funswitch-loops}, @option{-fpredictive-commoning},
5894 @option{-fgcse-after-reload} and @option{-ftree-vectorize} options.
5898 Reduce compilation time and make debugging produce the expected
5899 results. This is the default.
5903 Optimize for size. @option{-Os} enables all @option{-O2} optimizations that
5904 do not typically increase code size. It also performs further
5905 optimizations designed to reduce code size.
5907 @option{-Os} disables the following optimization flags:
5908 @gccoptlist{-falign-functions -falign-jumps -falign-loops @gol
5909 -falign-labels -freorder-blocks -freorder-blocks-and-partition @gol
5910 -fprefetch-loop-arrays -ftree-vect-loop-version}
5914 Disregard strict standards compliance. @option{-Ofast} enables all
5915 @option{-O3} optimizations. It also enables optimizations that are not
5916 valid for all standard compliant programs.
5917 It turns on @option{-ffast-math}.
5919 If you use multiple @option{-O} options, with or without level numbers,
5920 the last such option is the one that is effective.
5923 Options of the form @option{-f@var{flag}} specify machine-independent
5924 flags. Most flags have both positive and negative forms; the negative
5925 form of @option{-ffoo} would be @option{-fno-foo}. In the table
5926 below, only one of the forms is listed---the one you typically will
5927 use. You can figure out the other form by either removing @samp{no-}
5930 The following options control specific optimizations. They are either
5931 activated by @option{-O} options or are related to ones that are. You
5932 can use the following flags in the rare cases when ``fine-tuning'' of
5933 optimizations to be performed is desired.
5936 @item -fno-default-inline
5937 @opindex fno-default-inline
5938 Do not make member functions inline by default merely because they are
5939 defined inside the class scope (C++ only). Otherwise, when you specify
5940 @w{@option{-O}}, member functions defined inside class scope are compiled
5941 inline by default; i.e., you don't need to add @samp{inline} in front of
5942 the member function name.
5944 @item -fno-defer-pop
5945 @opindex fno-defer-pop
5946 Always pop the arguments to each function call as soon as that function
5947 returns. For machines which must pop arguments after a function call,
5948 the compiler normally lets arguments accumulate on the stack for several
5949 function calls and pops them all at once.
5951 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5953 @item -fforward-propagate
5954 @opindex fforward-propagate
5955 Perform a forward propagation pass on RTL@. The pass tries to combine two
5956 instructions and checks if the result can be simplified. If loop unrolling
5957 is active, two passes are performed and the second is scheduled after
5960 This option is enabled by default at optimization levels @option{-O},
5961 @option{-O2}, @option{-O3}, @option{-Os}.
5963 @item -fomit-frame-pointer
5964 @opindex fomit-frame-pointer
5965 Don't keep the frame pointer in a register for functions that
5966 don't need one. This avoids the instructions to save, set up and
5967 restore frame pointers; it also makes an extra register available
5968 in many functions. @strong{It also makes debugging impossible on
5971 On some machines, such as the VAX, this flag has no effect, because
5972 the standard calling sequence automatically handles the frame pointer
5973 and nothing is saved by pretending it doesn't exist. The
5974 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
5975 whether a target machine supports this flag. @xref{Registers,,Register
5976 Usage, gccint, GNU Compiler Collection (GCC) Internals}.
5978 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5980 @item -foptimize-sibling-calls
5981 @opindex foptimize-sibling-calls
5982 Optimize sibling and tail recursive calls.
5984 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5988 Don't pay attention to the @code{inline} keyword. Normally this option
5989 is used to keep the compiler from expanding any functions inline.
5990 Note that if you are not optimizing, no functions can be expanded inline.
5992 @item -finline-small-functions
5993 @opindex finline-small-functions
5994 Integrate functions into their callers when their body is smaller than expected
5995 function call code (so overall size of program gets smaller). The compiler
5996 heuristically decides which functions are simple enough to be worth integrating
5999 Enabled at level @option{-O2}.
6001 @item -findirect-inlining
6002 @opindex findirect-inlining
6003 Inline also indirect calls that are discovered to be known at compile
6004 time thanks to previous inlining. This option has any effect only
6005 when inlining itself is turned on by the @option{-finline-functions}
6006 or @option{-finline-small-functions} options.
6008 Enabled at level @option{-O2}.
6010 @item -finline-functions
6011 @opindex finline-functions
6012 Integrate all simple functions into their callers. The compiler
6013 heuristically decides which functions are simple enough to be worth
6014 integrating in this way.
6016 If all calls to a given function are integrated, and the function is
6017 declared @code{static}, then the function is normally not output as
6018 assembler code in its own right.
6020 Enabled at level @option{-O3}.
6022 @item -finline-functions-called-once
6023 @opindex finline-functions-called-once
6024 Consider all @code{static} functions called once for inlining into their
6025 caller even if they are not marked @code{inline}. If a call to a given
6026 function is integrated, then the function is not output as assembler code
6029 Enabled at levels @option{-O1}, @option{-O2}, @option{-O3} and @option{-Os}.
6031 @item -fearly-inlining
6032 @opindex fearly-inlining
6033 Inline functions marked by @code{always_inline} and functions whose body seems
6034 smaller than the function call overhead early before doing
6035 @option{-fprofile-generate} instrumentation and real inlining pass. Doing so
6036 makes profiling significantly cheaper and usually inlining faster on programs
6037 having large chains of nested wrapper functions.
6043 Perform interprocedural scalar replacement of aggregates, removal of
6044 unused parameters and replacement of parameters passed by reference
6045 by parameters passed by value.
6047 Enabled at levels @option{-O2}, @option{-O3} and @option{-Os}.
6049 @item -finline-limit=@var{n}
6050 @opindex finline-limit
6051 By default, GCC limits the size of functions that can be inlined. This flag
6052 allows coarse control of this limit. @var{n} is the size of functions that
6053 can be inlined in number of pseudo instructions.
6055 Inlining is actually controlled by a number of parameters, which may be
6056 specified individually by using @option{--param @var{name}=@var{value}}.
6057 The @option{-finline-limit=@var{n}} option sets some of these parameters
6061 @item max-inline-insns-single
6062 is set to @var{n}/2.
6063 @item max-inline-insns-auto
6064 is set to @var{n}/2.
6067 See below for a documentation of the individual
6068 parameters controlling inlining and for the defaults of these parameters.
6070 @emph{Note:} there may be no value to @option{-finline-limit} that results
6071 in default behavior.
6073 @emph{Note:} pseudo instruction represents, in this particular context, an
6074 abstract measurement of function's size. In no way does it represent a count
6075 of assembly instructions and as such its exact meaning might change from one
6076 release to an another.
6078 @item -fkeep-inline-functions
6079 @opindex fkeep-inline-functions
6080 In C, emit @code{static} functions that are declared @code{inline}
6081 into the object file, even if the function has been inlined into all
6082 of its callers. This switch does not affect functions using the
6083 @code{extern inline} extension in GNU C90@. In C++, emit any and all
6084 inline functions into the object file.
6086 @item -fkeep-static-consts
6087 @opindex fkeep-static-consts
6088 Emit variables declared @code{static const} when optimization isn't turned
6089 on, even if the variables aren't referenced.
6091 GCC enables this option by default. If you want to force the compiler to
6092 check if the variable was referenced, regardless of whether or not
6093 optimization is turned on, use the @option{-fno-keep-static-consts} option.
6095 @item -fmerge-constants
6096 @opindex fmerge-constants
6097 Attempt to merge identical constants (string constants and floating point
6098 constants) across compilation units.
6100 This option is the default for optimized compilation if the assembler and
6101 linker support it. Use @option{-fno-merge-constants} to inhibit this
6104 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
6106 @item -fmerge-all-constants
6107 @opindex fmerge-all-constants
6108 Attempt to merge identical constants and identical variables.
6110 This option implies @option{-fmerge-constants}. In addition to
6111 @option{-fmerge-constants} this considers e.g.@: even constant initialized
6112 arrays or initialized constant variables with integral or floating point
6113 types. Languages like C or C++ require each variable, including multiple
6114 instances of the same variable in recursive calls, to have distinct locations,
6115 so using this option will result in non-conforming
6118 @item -fmodulo-sched
6119 @opindex fmodulo-sched
6120 Perform swing modulo scheduling immediately before the first scheduling
6121 pass. This pass looks at innermost loops and reorders their
6122 instructions by overlapping different iterations.
6124 @item -fmodulo-sched-allow-regmoves
6125 @opindex fmodulo-sched-allow-regmoves
6126 Perform more aggressive SMS based modulo scheduling with register moves
6127 allowed. By setting this flag certain anti-dependences edges will be
6128 deleted which will trigger the generation of reg-moves based on the
6129 life-range analysis. This option is effective only with
6130 @option{-fmodulo-sched} enabled.
6132 @item -fno-branch-count-reg
6133 @opindex fno-branch-count-reg
6134 Do not use ``decrement and branch'' instructions on a count register,
6135 but instead generate a sequence of instructions that decrement a
6136 register, compare it against zero, then branch based upon the result.
6137 This option is only meaningful on architectures that support such
6138 instructions, which include x86, PowerPC, IA-64 and S/390.
6140 The default is @option{-fbranch-count-reg}.
6142 @item -fno-function-cse
6143 @opindex fno-function-cse
6144 Do not put function addresses in registers; make each instruction that
6145 calls a constant function contain the function's address explicitly.
6147 This option results in less efficient code, but some strange hacks
6148 that alter the assembler output may be confused by the optimizations
6149 performed when this option is not used.
6151 The default is @option{-ffunction-cse}
6153 @item -fno-zero-initialized-in-bss
6154 @opindex fno-zero-initialized-in-bss
6155 If the target supports a BSS section, GCC by default puts variables that
6156 are initialized to zero into BSS@. This can save space in the resulting
6159 This option turns off this behavior because some programs explicitly
6160 rely on variables going to the data section. E.g., so that the
6161 resulting executable can find the beginning of that section and/or make
6162 assumptions based on that.
6164 The default is @option{-fzero-initialized-in-bss}.
6166 @item -fmudflap -fmudflapth -fmudflapir
6170 @cindex bounds checking
6172 For front-ends that support it (C and C++), instrument all risky
6173 pointer/array dereferencing operations, some standard library
6174 string/heap functions, and some other associated constructs with
6175 range/validity tests. Modules so instrumented should be immune to
6176 buffer overflows, invalid heap use, and some other classes of C/C++
6177 programming errors. The instrumentation relies on a separate runtime
6178 library (@file{libmudflap}), which will be linked into a program if
6179 @option{-fmudflap} is given at link time. Run-time behavior of the
6180 instrumented program is controlled by the @env{MUDFLAP_OPTIONS}
6181 environment variable. See @code{env MUDFLAP_OPTIONS=-help a.out}
6184 Use @option{-fmudflapth} instead of @option{-fmudflap} to compile and to
6185 link if your program is multi-threaded. Use @option{-fmudflapir}, in
6186 addition to @option{-fmudflap} or @option{-fmudflapth}, if
6187 instrumentation should ignore pointer reads. This produces less
6188 instrumentation (and therefore faster execution) and still provides
6189 some protection against outright memory corrupting writes, but allows
6190 erroneously read data to propagate within a program.
6192 @item -fthread-jumps
6193 @opindex fthread-jumps
6194 Perform optimizations where we check to see if a jump branches to a
6195 location where another comparison subsumed by the first is found. If
6196 so, the first branch is redirected to either the destination of the
6197 second branch or a point immediately following it, depending on whether
6198 the condition is known to be true or false.
6200 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6202 @item -fsplit-wide-types
6203 @opindex fsplit-wide-types
6204 When using a type that occupies multiple registers, such as @code{long
6205 long} on a 32-bit system, split the registers apart and allocate them
6206 independently. This normally generates better code for those types,
6207 but may make debugging more difficult.
6209 Enabled at levels @option{-O}, @option{-O2}, @option{-O3},
6212 @item -fcse-follow-jumps
6213 @opindex fcse-follow-jumps
6214 In common subexpression elimination (CSE), scan through jump instructions
6215 when the target of the jump is not reached by any other path. For
6216 example, when CSE encounters an @code{if} statement with an
6217 @code{else} clause, CSE will follow the jump when the condition
6220 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6222 @item -fcse-skip-blocks
6223 @opindex fcse-skip-blocks
6224 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
6225 follow jumps which conditionally skip over blocks. When CSE
6226 encounters a simple @code{if} statement with no else clause,
6227 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
6228 body of the @code{if}.
6230 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6232 @item -frerun-cse-after-loop
6233 @opindex frerun-cse-after-loop
6234 Re-run common subexpression elimination after loop optimizations has been
6237 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6241 Perform a global common subexpression elimination pass.
6242 This pass also performs global constant and copy propagation.
6244 @emph{Note:} When compiling a program using computed gotos, a GCC
6245 extension, you may get better runtime performance if you disable
6246 the global common subexpression elimination pass by adding
6247 @option{-fno-gcse} to the command line.
6249 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6253 When @option{-fgcse-lm} is enabled, global common subexpression elimination will
6254 attempt to move loads which are only killed by stores into themselves. This
6255 allows a loop containing a load/store sequence to be changed to a load outside
6256 the loop, and a copy/store within the loop.
6258 Enabled by default when gcse is enabled.
6262 When @option{-fgcse-sm} is enabled, a store motion pass is run after
6263 global common subexpression elimination. This pass will attempt to move
6264 stores out of loops. When used in conjunction with @option{-fgcse-lm},
6265 loops containing a load/store sequence can be changed to a load before
6266 the loop and a store after the loop.
6268 Not enabled at any optimization level.
6272 When @option{-fgcse-las} is enabled, the global common subexpression
6273 elimination pass eliminates redundant loads that come after stores to the
6274 same memory location (both partial and full redundancies).
6276 Not enabled at any optimization level.
6278 @item -fgcse-after-reload
6279 @opindex fgcse-after-reload
6280 When @option{-fgcse-after-reload} is enabled, a redundant load elimination
6281 pass is performed after reload. The purpose of this pass is to cleanup
6284 @item -funsafe-loop-optimizations
6285 @opindex funsafe-loop-optimizations
6286 If given, the loop optimizer will assume that loop indices do not
6287 overflow, and that the loops with nontrivial exit condition are not
6288 infinite. This enables a wider range of loop optimizations even if
6289 the loop optimizer itself cannot prove that these assumptions are valid.
6290 Using @option{-Wunsafe-loop-optimizations}, the compiler will warn you
6291 if it finds this kind of loop.
6293 @item -fcrossjumping
6294 @opindex fcrossjumping
6295 Perform cross-jumping transformation. This transformation unifies equivalent code and save code size. The
6296 resulting code may or may not perform better than without cross-jumping.
6298 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6300 @item -fauto-inc-dec
6301 @opindex fauto-inc-dec
6302 Combine increments or decrements of addresses with memory accesses.
6303 This pass is always skipped on architectures that do not have
6304 instructions to support this. Enabled by default at @option{-O} and
6305 higher on architectures that support this.
6309 Perform dead code elimination (DCE) on RTL@.
6310 Enabled by default at @option{-O} and higher.
6314 Perform dead store elimination (DSE) on RTL@.
6315 Enabled by default at @option{-O} and higher.
6317 @item -fif-conversion
6318 @opindex fif-conversion
6319 Attempt to transform conditional jumps into branch-less equivalents. This
6320 include use of conditional moves, min, max, set flags and abs instructions, and
6321 some tricks doable by standard arithmetics. The use of conditional execution
6322 on chips where it is available is controlled by @code{if-conversion2}.
6324 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
6326 @item -fif-conversion2
6327 @opindex fif-conversion2
6328 Use conditional execution (where available) to transform conditional jumps into
6329 branch-less equivalents.
6331 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
6333 @item -fdelete-null-pointer-checks
6334 @opindex fdelete-null-pointer-checks
6335 Assume that programs cannot safely dereference null pointers, and that
6336 no code or data element resides there. This enables simple constant
6337 folding optimizations at all optimization levels. In addition, other
6338 optimization passes in GCC use this flag to control global dataflow
6339 analyses that eliminate useless checks for null pointers; these assume
6340 that if a pointer is checked after it has already been dereferenced,
6343 Note however that in some environments this assumption is not true.
6344 Use @option{-fno-delete-null-pointer-checks} to disable this optimization
6345 for programs which depend on that behavior.
6347 Some targets, especially embedded ones, disable this option at all levels.
6348 Otherwise it is enabled at all levels: @option{-O0}, @option{-O1},
6349 @option{-O2}, @option{-O3}, @option{-Os}. Passes that use the information
6350 are enabled independently at different optimization levels.
6352 @item -fexpensive-optimizations
6353 @opindex fexpensive-optimizations
6354 Perform a number of minor optimizations that are relatively expensive.
6356 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6358 @item -foptimize-register-move
6360 @opindex foptimize-register-move
6362 Attempt to reassign register numbers in move instructions and as
6363 operands of other simple instructions in order to maximize the amount of
6364 register tying. This is especially helpful on machines with two-operand
6367 Note @option{-fregmove} and @option{-foptimize-register-move} are the same
6370 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6372 @item -fira-algorithm=@var{algorithm}
6373 Use specified coloring algorithm for the integrated register
6374 allocator. The @var{algorithm} argument should be @code{priority} or
6375 @code{CB}. The first algorithm specifies Chow's priority coloring,
6376 the second one specifies Chaitin-Briggs coloring. The second
6377 algorithm can be unimplemented for some architectures. If it is
6378 implemented, it is the default because Chaitin-Briggs coloring as a
6379 rule generates a better code.
6381 @item -fira-region=@var{region}
6382 Use specified regions for the integrated register allocator. The
6383 @var{region} argument should be one of @code{all}, @code{mixed}, or
6384 @code{one}. The first value means using all loops as register
6385 allocation regions, the second value which is the default means using
6386 all loops except for loops with small register pressure as the
6387 regions, and third one means using all function as a single region.
6388 The first value can give best result for machines with small size and
6389 irregular register set, the third one results in faster and generates
6390 decent code and the smallest size code, and the default value usually
6391 give the best results in most cases and for most architectures.
6393 @item -fira-coalesce
6394 @opindex fira-coalesce
6395 Do optimistic register coalescing. This option might be profitable for
6396 architectures with big regular register files.
6398 @item -fira-loop-pressure
6399 @opindex fira-loop-pressure
6400 Use IRA to evaluate register pressure in loops for decision to move
6401 loop invariants. Usage of this option usually results in generation
6402 of faster and smaller code on machines with big register files (>= 32
6403 registers) but it can slow compiler down.
6405 This option is enabled at level @option{-O3} for some targets.
6407 @item -fno-ira-share-save-slots
6408 @opindex fno-ira-share-save-slots
6409 Switch off sharing stack slots used for saving call used hard
6410 registers living through a call. Each hard register will get a
6411 separate stack slot and as a result function stack frame will be
6414 @item -fno-ira-share-spill-slots
6415 @opindex fno-ira-share-spill-slots
6416 Switch off sharing stack slots allocated for pseudo-registers. Each
6417 pseudo-register which did not get a hard register will get a separate
6418 stack slot and as a result function stack frame will be bigger.
6420 @item -fira-verbose=@var{n}
6421 @opindex fira-verbose
6422 Set up how verbose dump file for the integrated register allocator
6423 will be. Default value is 5. If the value is greater or equal to 10,
6424 the dump file will be stderr as if the value were @var{n} minus 10.
6426 @item -fdelayed-branch
6427 @opindex fdelayed-branch
6428 If supported for the target machine, attempt to reorder instructions
6429 to exploit instruction slots available after delayed branch
6432 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
6434 @item -fschedule-insns
6435 @opindex fschedule-insns
6436 If supported for the target machine, attempt to reorder instructions to
6437 eliminate execution stalls due to required data being unavailable. This
6438 helps machines that have slow floating point or memory load instructions
6439 by allowing other instructions to be issued until the result of the load
6440 or floating point instruction is required.
6442 Enabled at levels @option{-O2}, @option{-O3}.
6444 @item -fschedule-insns2
6445 @opindex fschedule-insns2
6446 Similar to @option{-fschedule-insns}, but requests an additional pass of
6447 instruction scheduling after register allocation has been done. This is
6448 especially useful on machines with a relatively small number of
6449 registers and where memory load instructions take more than one cycle.
6451 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6453 @item -fno-sched-interblock
6454 @opindex fno-sched-interblock
6455 Don't schedule instructions across basic blocks. This is normally
6456 enabled by default when scheduling before register allocation, i.e.@:
6457 with @option{-fschedule-insns} or at @option{-O2} or higher.
6459 @item -fno-sched-spec
6460 @opindex fno-sched-spec
6461 Don't allow speculative motion of non-load instructions. This is normally
6462 enabled by default when scheduling before register allocation, i.e.@:
6463 with @option{-fschedule-insns} or at @option{-O2} or higher.
6465 @item -fsched-pressure
6466 @opindex fsched-pressure
6467 Enable register pressure sensitive insn scheduling before the register
6468 allocation. This only makes sense when scheduling before register
6469 allocation is enabled, i.e.@: with @option{-fschedule-insns} or at
6470 @option{-O2} or higher. Usage of this option can improve the
6471 generated code and decrease its size by preventing register pressure
6472 increase above the number of available hard registers and as a
6473 consequence register spills in the register allocation.
6475 @item -fsched-spec-load
6476 @opindex fsched-spec-load
6477 Allow speculative motion of some load instructions. This only makes
6478 sense when scheduling before register allocation, i.e.@: with
6479 @option{-fschedule-insns} or at @option{-O2} or higher.
6481 @item -fsched-spec-load-dangerous
6482 @opindex fsched-spec-load-dangerous
6483 Allow speculative motion of more load instructions. This only makes
6484 sense when scheduling before register allocation, i.e.@: with
6485 @option{-fschedule-insns} or at @option{-O2} or higher.
6487 @item -fsched-stalled-insns
6488 @itemx -fsched-stalled-insns=@var{n}
6489 @opindex fsched-stalled-insns
6490 Define how many insns (if any) can be moved prematurely from the queue
6491 of stalled insns into the ready list, during the second scheduling pass.
6492 @option{-fno-sched-stalled-insns} means that no insns will be moved
6493 prematurely, @option{-fsched-stalled-insns=0} means there is no limit
6494 on how many queued insns can be moved prematurely.
6495 @option{-fsched-stalled-insns} without a value is equivalent to
6496 @option{-fsched-stalled-insns=1}.
6498 @item -fsched-stalled-insns-dep
6499 @itemx -fsched-stalled-insns-dep=@var{n}
6500 @opindex fsched-stalled-insns-dep
6501 Define how many insn groups (cycles) will be examined for a dependency
6502 on a stalled insn that is candidate for premature removal from the queue
6503 of stalled insns. This has an effect only during the second scheduling pass,
6504 and only if @option{-fsched-stalled-insns} is used.
6505 @option{-fno-sched-stalled-insns-dep} is equivalent to
6506 @option{-fsched-stalled-insns-dep=0}.
6507 @option{-fsched-stalled-insns-dep} without a value is equivalent to
6508 @option{-fsched-stalled-insns-dep=1}.
6510 @item -fsched2-use-superblocks
6511 @opindex fsched2-use-superblocks
6512 When scheduling after register allocation, do use superblock scheduling
6513 algorithm. Superblock scheduling allows motion across basic block boundaries
6514 resulting on faster schedules. This option is experimental, as not all machine
6515 descriptions used by GCC model the CPU closely enough to avoid unreliable
6516 results from the algorithm.
6518 This only makes sense when scheduling after register allocation, i.e.@: with
6519 @option{-fschedule-insns2} or at @option{-O2} or higher.
6521 @item -fsched-group-heuristic
6522 @opindex fsched-group-heuristic
6523 Enable the group heuristic in the scheduler. This heuristic favors
6524 the instruction that belongs to a schedule group. This is enabled
6525 by default when scheduling is enabled, i.e.@: with @option{-fschedule-insns}
6526 or @option{-fschedule-insns2} or at @option{-O2} or higher.
6528 @item -fsched-critical-path-heuristic
6529 @opindex fsched-critical-path-heuristic
6530 Enable the critical-path heuristic in the scheduler. This heuristic favors
6531 instructions on the critical path. This is enabled by default when
6532 scheduling is enabled, i.e.@: with @option{-fschedule-insns}
6533 or @option{-fschedule-insns2} or at @option{-O2} or higher.
6535 @item -fsched-spec-insn-heuristic
6536 @opindex fsched-spec-insn-heuristic
6537 Enable the speculative instruction heuristic in the scheduler. This
6538 heuristic favors speculative instructions with greater dependency weakness.
6539 This is enabled by default when scheduling is enabled, i.e.@:
6540 with @option{-fschedule-insns} or @option{-fschedule-insns2}
6541 or at @option{-O2} or higher.
6543 @item -fsched-rank-heuristic
6544 @opindex fsched-rank-heuristic
6545 Enable the rank heuristic in the scheduler. This heuristic favors
6546 the instruction belonging to a basic block with greater size or frequency.
6547 This is enabled by default when scheduling is enabled, i.e.@:
6548 with @option{-fschedule-insns} or @option{-fschedule-insns2} or
6549 at @option{-O2} or higher.
6551 @item -fsched-last-insn-heuristic
6552 @opindex fsched-last-insn-heuristic
6553 Enable the last-instruction heuristic in the scheduler. This heuristic
6554 favors the instruction that is less dependent on the last instruction
6555 scheduled. This is enabled by default when scheduling is enabled,
6556 i.e.@: with @option{-fschedule-insns} or @option{-fschedule-insns2} or
6557 at @option{-O2} or higher.
6559 @item -fsched-dep-count-heuristic
6560 @opindex fsched-dep-count-heuristic
6561 Enable the dependent-count heuristic in the scheduler. This heuristic
6562 favors the instruction that has more instructions depending on it.
6563 This is enabled by default when scheduling is enabled, i.e.@:
6564 with @option{-fschedule-insns} or @option{-fschedule-insns2} or
6565 at @option{-O2} or higher.
6567 @item -freschedule-modulo-scheduled-loops
6568 @opindex freschedule-modulo-scheduled-loops
6569 The modulo scheduling comes before the traditional scheduling, if a loop
6570 was modulo scheduled we may want to prevent the later scheduling passes
6571 from changing its schedule, we use this option to control that.
6573 @item -fselective-scheduling
6574 @opindex fselective-scheduling
6575 Schedule instructions using selective scheduling algorithm. Selective
6576 scheduling runs instead of the first scheduler pass.
6578 @item -fselective-scheduling2
6579 @opindex fselective-scheduling2
6580 Schedule instructions using selective scheduling algorithm. Selective
6581 scheduling runs instead of the second scheduler pass.
6583 @item -fsel-sched-pipelining
6584 @opindex fsel-sched-pipelining
6585 Enable software pipelining of innermost loops during selective scheduling.
6586 This option has no effect until one of @option{-fselective-scheduling} or
6587 @option{-fselective-scheduling2} is turned on.
6589 @item -fsel-sched-pipelining-outer-loops
6590 @opindex fsel-sched-pipelining-outer-loops
6591 When pipelining loops during selective scheduling, also pipeline outer loops.
6592 This option has no effect until @option{-fsel-sched-pipelining} is turned on.
6594 @item -fcaller-saves
6595 @opindex fcaller-saves
6596 Enable values to be allocated in registers that will be clobbered by
6597 function calls, by emitting extra instructions to save and restore the
6598 registers around such calls. Such allocation is done only when it
6599 seems to result in better code than would otherwise be produced.
6601 This option is always enabled by default on certain machines, usually
6602 those which have no call-preserved registers to use instead.
6604 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6606 @item -fconserve-stack
6607 @opindex fconserve-stack
6608 Attempt to minimize stack usage. The compiler will attempt to use less
6609 stack space, even if that makes the program slower. This option
6610 implies setting the @option{large-stack-frame} parameter to 100
6611 and the @option{large-stack-frame-growth} parameter to 400.
6613 @item -ftree-reassoc
6614 @opindex ftree-reassoc
6615 Perform reassociation on trees. This flag is enabled by default
6616 at @option{-O} and higher.
6620 Perform partial redundancy elimination (PRE) on trees. This flag is
6621 enabled by default at @option{-O2} and @option{-O3}.
6623 @item -ftree-forwprop
6624 @opindex ftree-forwprop
6625 Perform forward propagation on trees. This flag is enabled by default
6626 at @option{-O} and higher.
6630 Perform full redundancy elimination (FRE) on trees. The difference
6631 between FRE and PRE is that FRE only considers expressions
6632 that are computed on all paths leading to the redundant computation.
6633 This analysis is faster than PRE, though it exposes fewer redundancies.
6634 This flag is enabled by default at @option{-O} and higher.
6636 @item -ftree-phiprop
6637 @opindex ftree-phiprop
6638 Perform hoisting of loads from conditional pointers on trees. This
6639 pass is enabled by default at @option{-O} and higher.
6641 @item -ftree-copy-prop
6642 @opindex ftree-copy-prop
6643 Perform copy propagation on trees. This pass eliminates unnecessary
6644 copy operations. This flag is enabled by default at @option{-O} and
6647 @item -fipa-pure-const
6648 @opindex fipa-pure-const
6649 Discover which functions are pure or constant.
6650 Enabled by default at @option{-O} and higher.
6652 @item -fipa-reference
6653 @opindex fipa-reference
6654 Discover which static variables do not escape cannot escape the
6656 Enabled by default at @option{-O} and higher.
6658 @item -fipa-struct-reorg
6659 @opindex fipa-struct-reorg
6660 Perform structure reorganization optimization, that change C-like structures
6661 layout in order to better utilize spatial locality. This transformation is
6662 affective for programs containing arrays of structures. Available in two
6663 compilation modes: profile-based (enabled with @option{-fprofile-generate})
6664 or static (which uses built-in heuristics). It works only in whole program
6665 mode, so it requires @option{-fwhole-program} and @option{-combine} to be
6666 enabled. Structures considered @samp{cold} by this transformation are not
6667 affected (see @option{--param struct-reorg-cold-struct-ratio=@var{value}}).
6669 With this flag, the program debug info reflects a new structure layout.
6673 Perform interprocedural pointer analysis and interprocedural modification
6674 and reference analysis. This option can cause excessive memory and
6675 compile-time usage on large compilation units. It is not enabled by
6676 default at any optimization level.
6679 @opindex fipa-profile
6680 Perform interprocedural profile propagation. The functions called only from
6681 cold functions are marked as cold. Also functions executed once (such as
6682 @code{cold}, @code{noreturn}, static constructors or destructors) are identified. Cold
6683 functions and loop less parts of functions executed once are then optimized for
6685 Enabled by default at @option{-O} and higher.
6689 Perform interprocedural constant propagation.
6690 This optimization analyzes the program to determine when values passed
6691 to functions are constants and then optimizes accordingly.
6692 This optimization can substantially increase performance
6693 if the application has constants passed to functions.
6694 This flag is enabled by default at @option{-O2}, @option{-Os} and @option{-O3}.
6696 @item -fipa-cp-clone
6697 @opindex fipa-cp-clone
6698 Perform function cloning to make interprocedural constant propagation stronger.
6699 When enabled, interprocedural constant propagation will perform function cloning
6700 when externally visible function can be called with constant arguments.
6701 Because this optimization can create multiple copies of functions,
6702 it may significantly increase code size
6703 (see @option{--param ipcp-unit-growth=@var{value}}).
6704 This flag is enabled by default at @option{-O3}.
6706 @item -fipa-matrix-reorg
6707 @opindex fipa-matrix-reorg
6708 Perform matrix flattening and transposing.
6709 Matrix flattening tries to replace an @math{m}-dimensional matrix
6710 with its equivalent @math{n}-dimensional matrix, where @math{n < m}.
6711 This reduces the level of indirection needed for accessing the elements
6712 of the matrix. The second optimization is matrix transposing that
6713 attempts to change the order of the matrix's dimensions in order to
6714 improve cache locality.
6715 Both optimizations need the @option{-fwhole-program} flag.
6716 Transposing is enabled only if profiling information is available.
6720 Perform forward store motion on trees. This flag is
6721 enabled by default at @option{-O} and higher.
6725 Perform sparse conditional constant propagation (CCP) on trees. This
6726 pass only operates on local scalar variables and is enabled by default
6727 at @option{-O} and higher.
6729 @item -ftree-switch-conversion
6730 Perform conversion of simple initializations in a switch to
6731 initializations from a scalar array. This flag is enabled by default
6732 at @option{-O2} and higher.
6736 Perform dead code elimination (DCE) on trees. This flag is enabled by
6737 default at @option{-O} and higher.
6739 @item -ftree-builtin-call-dce
6740 @opindex ftree-builtin-call-dce
6741 Perform conditional dead code elimination (DCE) for calls to builtin functions
6742 that may set @code{errno} but are otherwise side-effect free. This flag is
6743 enabled by default at @option{-O2} and higher if @option{-Os} is not also
6746 @item -ftree-dominator-opts
6747 @opindex ftree-dominator-opts
6748 Perform a variety of simple scalar cleanups (constant/copy
6749 propagation, redundancy elimination, range propagation and expression
6750 simplification) based on a dominator tree traversal. This also
6751 performs jump threading (to reduce jumps to jumps). This flag is
6752 enabled by default at @option{-O} and higher.
6756 Perform dead store elimination (DSE) on trees. A dead store is a store into
6757 a memory location which will later be overwritten by another store without
6758 any intervening loads. In this case the earlier store can be deleted. This
6759 flag is enabled by default at @option{-O} and higher.
6763 Perform loop header copying on trees. This is beneficial since it increases
6764 effectiveness of code motion optimizations. It also saves one jump. This flag
6765 is enabled by default at @option{-O} and higher. It is not enabled
6766 for @option{-Os}, since it usually increases code size.
6768 @item -ftree-loop-optimize
6769 @opindex ftree-loop-optimize
6770 Perform loop optimizations on trees. This flag is enabled by default
6771 at @option{-O} and higher.
6773 @item -ftree-loop-linear
6774 @opindex ftree-loop-linear
6775 Perform linear loop transformations on tree. This flag can improve cache
6776 performance and allow further loop optimizations to take place.
6778 @item -floop-interchange
6779 Perform loop interchange transformations on loops. Interchanging two
6780 nested loops switches the inner and outer loops. For example, given a
6785 A(J, I) = A(J, I) * C
6789 loop interchange will transform the loop as if the user had written:
6793 A(J, I) = A(J, I) * C
6797 which can be beneficial when @code{N} is larger than the caches,
6798 because in Fortran, the elements of an array are stored in memory
6799 contiguously by column, and the original loop iterates over rows,
6800 potentially creating at each access a cache miss. This optimization
6801 applies to all the languages supported by GCC and is not limited to
6802 Fortran. To use this code transformation, GCC has to be configured
6803 with @option{--with-ppl} and @option{--with-cloog} to enable the
6804 Graphite loop transformation infrastructure.
6806 @item -floop-strip-mine
6807 Perform loop strip mining transformations on loops. Strip mining
6808 splits a loop into two nested loops. The outer loop has strides
6809 equal to the strip size and the inner loop has strides of the
6810 original loop within a strip. The strip length can be changed
6811 using the @option{loop-block-tile-size} parameter. For example,
6818 loop strip mining will transform the loop as if the user had written:
6821 DO I = II, min (II + 50, N)
6826 This optimization applies to all the languages supported by GCC and is
6827 not limited to Fortran. To use this code transformation, GCC has to
6828 be configured with @option{--with-ppl} and @option{--with-cloog} to
6829 enable the Graphite loop transformation infrastructure.
6832 Perform loop blocking transformations on loops. Blocking strip mines
6833 each loop in the loop nest such that the memory accesses of the
6834 element loops fit inside caches. The strip length can be changed
6835 using the @option{loop-block-tile-size} parameter. For example, given
6840 A(J, I) = B(I) + C(J)
6844 loop blocking will transform the loop as if the user had written:
6848 DO I = II, min (II + 50, N)
6849 DO J = JJ, min (JJ + 50, M)
6850 A(J, I) = B(I) + C(J)
6856 which can be beneficial when @code{M} is larger than the caches,
6857 because the innermost loop will iterate over a smaller amount of data
6858 that can be kept in the caches. This optimization applies to all the
6859 languages supported by GCC and is not limited to Fortran. To use this
6860 code transformation, GCC has to be configured with @option{--with-ppl}
6861 and @option{--with-cloog} to enable the Graphite loop transformation
6864 @item -fgraphite-identity
6865 @opindex fgraphite-identity
6866 Enable the identity transformation for graphite. For every SCoP we generate
6867 the polyhedral representation and transform it back to gimple. Using
6868 @option{-fgraphite-identity} we can check the costs or benefits of the
6869 GIMPLE -> GRAPHITE -> GIMPLE transformation. Some minimal optimizations
6870 are also performed by the code generator CLooG, like index splitting and
6871 dead code elimination in loops.
6873 @item -floop-parallelize-all
6874 Use the Graphite data dependence analysis to identify loops that can
6875 be parallelized. Parallelize all the loops that can be analyzed to
6876 not contain loop carried dependences without checking that it is
6877 profitable to parallelize the loops.
6879 @item -fcheck-data-deps
6880 @opindex fcheck-data-deps
6881 Compare the results of several data dependence analyzers. This option
6882 is used for debugging the data dependence analyzers.
6884 @item -ftree-loop-distribution
6885 Perform loop distribution. This flag can improve cache performance on
6886 big loop bodies and allow further loop optimizations, like
6887 parallelization or vectorization, to take place. For example, the loop
6904 @item -ftree-loop-im
6905 @opindex ftree-loop-im
6906 Perform loop invariant motion on trees. This pass moves only invariants that
6907 would be hard to handle at RTL level (function calls, operations that expand to
6908 nontrivial sequences of insns). With @option{-funswitch-loops} it also moves
6909 operands of conditions that are invariant out of the loop, so that we can use
6910 just trivial invariantness analysis in loop unswitching. The pass also includes
6913 @item -ftree-loop-ivcanon
6914 @opindex ftree-loop-ivcanon
6915 Create a canonical counter for number of iterations in the loop for that
6916 determining number of iterations requires complicated analysis. Later
6917 optimizations then may determine the number easily. Useful especially
6918 in connection with unrolling.
6922 Perform induction variable optimizations (strength reduction, induction
6923 variable merging and induction variable elimination) on trees.
6925 @item -ftree-parallelize-loops=n
6926 @opindex ftree-parallelize-loops
6927 Parallelize loops, i.e., split their iteration space to run in n threads.
6928 This is only possible for loops whose iterations are independent
6929 and can be arbitrarily reordered. The optimization is only
6930 profitable on multiprocessor machines, for loops that are CPU-intensive,
6931 rather than constrained e.g.@: by memory bandwidth. This option
6932 implies @option{-pthread}, and thus is only supported on targets
6933 that have support for @option{-pthread}.
6937 Perform function-local points-to analysis on trees. This flag is
6938 enabled by default at @option{-O} and higher.
6942 Perform scalar replacement of aggregates. This pass replaces structure
6943 references with scalars to prevent committing structures to memory too
6944 early. This flag is enabled by default at @option{-O} and higher.
6946 @item -ftree-copyrename
6947 @opindex ftree-copyrename
6948 Perform copy renaming on trees. This pass attempts to rename compiler
6949 temporaries to other variables at copy locations, usually resulting in
6950 variable names which more closely resemble the original variables. This flag
6951 is enabled by default at @option{-O} and higher.
6955 Perform temporary expression replacement during the SSA->normal phase. Single
6956 use/single def temporaries are replaced at their use location with their
6957 defining expression. This results in non-GIMPLE code, but gives the expanders
6958 much more complex trees to work on resulting in better RTL generation. This is
6959 enabled by default at @option{-O} and higher.
6961 @item -ftree-vectorize
6962 @opindex ftree-vectorize
6963 Perform loop vectorization on trees. This flag is enabled by default at
6966 @item -ftree-slp-vectorize
6967 @opindex ftree-slp-vectorize
6968 Perform basic block vectorization on trees. This flag is enabled by default at
6969 @option{-O3} and when @option{-ftree-vectorize} is enabled.
6971 @item -ftree-vect-loop-version
6972 @opindex ftree-vect-loop-version
6973 Perform loop versioning when doing loop vectorization on trees. When a loop
6974 appears to be vectorizable except that data alignment or data dependence cannot
6975 be determined at compile time then vectorized and non-vectorized versions of
6976 the loop are generated along with runtime checks for alignment or dependence
6977 to control which version is executed. This option is enabled by default
6978 except at level @option{-Os} where it is disabled.
6980 @item -fvect-cost-model
6981 @opindex fvect-cost-model
6982 Enable cost model for vectorization.
6986 Perform Value Range Propagation on trees. This is similar to the
6987 constant propagation pass, but instead of values, ranges of values are
6988 propagated. This allows the optimizers to remove unnecessary range
6989 checks like array bound checks and null pointer checks. This is
6990 enabled by default at @option{-O2} and higher. Null pointer check
6991 elimination is only done if @option{-fdelete-null-pointer-checks} is
6996 Perform tail duplication to enlarge superblock size. This transformation
6997 simplifies the control flow of the function allowing other optimizations to do
7000 @item -funroll-loops
7001 @opindex funroll-loops
7002 Unroll loops whose number of iterations can be determined at compile
7003 time or upon entry to the loop. @option{-funroll-loops} implies
7004 @option{-frerun-cse-after-loop}. This option makes code larger,
7005 and may or may not make it run faster.
7007 @item -funroll-all-loops
7008 @opindex funroll-all-loops
7009 Unroll all loops, even if their number of iterations is uncertain when
7010 the loop is entered. This usually makes programs run more slowly.
7011 @option{-funroll-all-loops} implies the same options as
7012 @option{-funroll-loops},
7014 @item -fsplit-ivs-in-unroller
7015 @opindex fsplit-ivs-in-unroller
7016 Enables expressing of values of induction variables in later iterations
7017 of the unrolled loop using the value in the first iteration. This breaks
7018 long dependency chains, thus improving efficiency of the scheduling passes.
7020 Combination of @option{-fweb} and CSE is often sufficient to obtain the
7021 same effect. However in cases the loop body is more complicated than
7022 a single basic block, this is not reliable. It also does not work at all
7023 on some of the architectures due to restrictions in the CSE pass.
7025 This optimization is enabled by default.
7027 @item -fvariable-expansion-in-unroller
7028 @opindex fvariable-expansion-in-unroller
7029 With this option, the compiler will create multiple copies of some
7030 local variables when unrolling a loop which can result in superior code.
7032 @item -fpartial-inlining
7033 @opindex fpartial-inlining
7034 Inline parts of functions. This option has any effect only
7035 when inlining itself is turned on by the @option{-finline-functions}
7036 or @option{-finline-small-functions} options.
7038 Enabled at level @option{-O2}.
7040 @item -fpredictive-commoning
7041 @opindex fpredictive-commoning
7042 Perform predictive commoning optimization, i.e., reusing computations
7043 (especially memory loads and stores) performed in previous
7044 iterations of loops.
7046 This option is enabled at level @option{-O3}.
7048 @item -fprefetch-loop-arrays
7049 @opindex fprefetch-loop-arrays
7050 If supported by the target machine, generate instructions to prefetch
7051 memory to improve the performance of loops that access large arrays.
7053 This option may generate better or worse code; results are highly
7054 dependent on the structure of loops within the source code.
7056 Disabled at level @option{-Os}.
7059 @itemx -fno-peephole2
7060 @opindex fno-peephole
7061 @opindex fno-peephole2
7062 Disable any machine-specific peephole optimizations. The difference
7063 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
7064 are implemented in the compiler; some targets use one, some use the
7065 other, a few use both.
7067 @option{-fpeephole} is enabled by default.
7068 @option{-fpeephole2} enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
7070 @item -fno-guess-branch-probability
7071 @opindex fno-guess-branch-probability
7072 Do not guess branch probabilities using heuristics.
7074 GCC will use heuristics to guess branch probabilities if they are
7075 not provided by profiling feedback (@option{-fprofile-arcs}). These
7076 heuristics are based on the control flow graph. If some branch probabilities
7077 are specified by @samp{__builtin_expect}, then the heuristics will be
7078 used to guess branch probabilities for the rest of the control flow graph,
7079 taking the @samp{__builtin_expect} info into account. The interactions
7080 between the heuristics and @samp{__builtin_expect} can be complex, and in
7081 some cases, it may be useful to disable the heuristics so that the effects
7082 of @samp{__builtin_expect} are easier to understand.
7084 The default is @option{-fguess-branch-probability} at levels
7085 @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
7087 @item -freorder-blocks
7088 @opindex freorder-blocks
7089 Reorder basic blocks in the compiled function in order to reduce number of
7090 taken branches and improve code locality.
7092 Enabled at levels @option{-O2}, @option{-O3}.
7094 @item -freorder-blocks-and-partition
7095 @opindex freorder-blocks-and-partition
7096 In addition to reordering basic blocks in the compiled function, in order
7097 to reduce number of taken branches, partitions hot and cold basic blocks
7098 into separate sections of the assembly and .o files, to improve
7099 paging and cache locality performance.
7101 This optimization is automatically turned off in the presence of
7102 exception handling, for linkonce sections, for functions with a user-defined
7103 section attribute and on any architecture that does not support named
7106 @item -freorder-functions
7107 @opindex freorder-functions
7108 Reorder functions in the object file in order to
7109 improve code locality. This is implemented by using special
7110 subsections @code{.text.hot} for most frequently executed functions and
7111 @code{.text.unlikely} for unlikely executed functions. Reordering is done by
7112 the linker so object file format must support named sections and linker must
7113 place them in a reasonable way.
7115 Also profile feedback must be available in to make this option effective. See
7116 @option{-fprofile-arcs} for details.
7118 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
7120 @item -fstrict-aliasing
7121 @opindex fstrict-aliasing
7122 Allow the compiler to assume the strictest aliasing rules applicable to
7123 the language being compiled. For C (and C++), this activates
7124 optimizations based on the type of expressions. In particular, an
7125 object of one type is assumed never to reside at the same address as an
7126 object of a different type, unless the types are almost the same. For
7127 example, an @code{unsigned int} can alias an @code{int}, but not a
7128 @code{void*} or a @code{double}. A character type may alias any other
7131 @anchor{Type-punning}Pay special attention to code like this:
7144 The practice of reading from a different union member than the one most
7145 recently written to (called ``type-punning'') is common. Even with
7146 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
7147 is accessed through the union type. So, the code above will work as
7148 expected. @xref{Structures unions enumerations and bit-fields
7149 implementation}. However, this code might not:
7160 Similarly, access by taking the address, casting the resulting pointer
7161 and dereferencing the result has undefined behavior, even if the cast
7162 uses a union type, e.g.:
7166 return ((union a_union *) &d)->i;
7170 The @option{-fstrict-aliasing} option is enabled at levels
7171 @option{-O2}, @option{-O3}, @option{-Os}.
7173 @item -fstrict-overflow
7174 @opindex fstrict-overflow
7175 Allow the compiler to assume strict signed overflow rules, depending
7176 on the language being compiled. For C (and C++) this means that
7177 overflow when doing arithmetic with signed numbers is undefined, which
7178 means that the compiler may assume that it will not happen. This
7179 permits various optimizations. For example, the compiler will assume
7180 that an expression like @code{i + 10 > i} will always be true for
7181 signed @code{i}. This assumption is only valid if signed overflow is
7182 undefined, as the expression is false if @code{i + 10} overflows when
7183 using twos complement arithmetic. When this option is in effect any
7184 attempt to determine whether an operation on signed numbers will
7185 overflow must be written carefully to not actually involve overflow.
7187 This option also allows the compiler to assume strict pointer
7188 semantics: given a pointer to an object, if adding an offset to that
7189 pointer does not produce a pointer to the same object, the addition is
7190 undefined. This permits the compiler to conclude that @code{p + u >
7191 p} is always true for a pointer @code{p} and unsigned integer
7192 @code{u}. This assumption is only valid because pointer wraparound is
7193 undefined, as the expression is false if @code{p + u} overflows using
7194 twos complement arithmetic.
7196 See also the @option{-fwrapv} option. Using @option{-fwrapv} means
7197 that integer signed overflow is fully defined: it wraps. When
7198 @option{-fwrapv} is used, there is no difference between
7199 @option{-fstrict-overflow} and @option{-fno-strict-overflow} for
7200 integers. With @option{-fwrapv} certain types of overflow are
7201 permitted. For example, if the compiler gets an overflow when doing
7202 arithmetic on constants, the overflowed value can still be used with
7203 @option{-fwrapv}, but not otherwise.
7205 The @option{-fstrict-overflow} option is enabled at levels
7206 @option{-O2}, @option{-O3}, @option{-Os}.
7208 @item -falign-functions
7209 @itemx -falign-functions=@var{n}
7210 @opindex falign-functions
7211 Align the start of functions to the next power-of-two greater than
7212 @var{n}, skipping up to @var{n} bytes. For instance,
7213 @option{-falign-functions=32} aligns functions to the next 32-byte
7214 boundary, but @option{-falign-functions=24} would align to the next
7215 32-byte boundary only if this can be done by skipping 23 bytes or less.
7217 @option{-fno-align-functions} and @option{-falign-functions=1} are
7218 equivalent and mean that functions will not be aligned.
7220 Some assemblers only support this flag when @var{n} is a power of two;
7221 in that case, it is rounded up.
7223 If @var{n} is not specified or is zero, use a machine-dependent default.
7225 Enabled at levels @option{-O2}, @option{-O3}.
7227 @item -falign-labels
7228 @itemx -falign-labels=@var{n}
7229 @opindex falign-labels
7230 Align all branch targets to a power-of-two boundary, skipping up to
7231 @var{n} bytes like @option{-falign-functions}. This option can easily
7232 make code slower, because it must insert dummy operations for when the
7233 branch target is reached in the usual flow of the code.
7235 @option{-fno-align-labels} and @option{-falign-labels=1} are
7236 equivalent and mean that labels will not be aligned.
7238 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
7239 are greater than this value, then their values are used instead.
7241 If @var{n} is not specified or is zero, use a machine-dependent default
7242 which is very likely to be @samp{1}, meaning no alignment.
7244 Enabled at levels @option{-O2}, @option{-O3}.
7247 @itemx -falign-loops=@var{n}
7248 @opindex falign-loops
7249 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
7250 like @option{-falign-functions}. The hope is that the loop will be
7251 executed many times, which will make up for any execution of the dummy
7254 @option{-fno-align-loops} and @option{-falign-loops=1} are
7255 equivalent and mean that loops will not be aligned.
7257 If @var{n} is not specified or is zero, use a machine-dependent default.
7259 Enabled at levels @option{-O2}, @option{-O3}.
7262 @itemx -falign-jumps=@var{n}
7263 @opindex falign-jumps
7264 Align branch targets to a power-of-two boundary, for branch targets
7265 where the targets can only be reached by jumping, skipping up to @var{n}
7266 bytes like @option{-falign-functions}. In this case, no dummy operations
7269 @option{-fno-align-jumps} and @option{-falign-jumps=1} are
7270 equivalent and mean that loops will not be aligned.
7272 If @var{n} is not specified or is zero, use a machine-dependent default.
7274 Enabled at levels @option{-O2}, @option{-O3}.
7276 @item -funit-at-a-time
7277 @opindex funit-at-a-time
7278 This option is left for compatibility reasons. @option{-funit-at-a-time}
7279 has no effect, while @option{-fno-unit-at-a-time} implies
7280 @option{-fno-toplevel-reorder} and @option{-fno-section-anchors}.
7284 @item -fno-toplevel-reorder
7285 @opindex fno-toplevel-reorder
7286 Do not reorder top-level functions, variables, and @code{asm}
7287 statements. Output them in the same order that they appear in the
7288 input file. When this option is used, unreferenced static variables
7289 will not be removed. This option is intended to support existing code
7290 which relies on a particular ordering. For new code, it is better to
7293 Enabled at level @option{-O0}. When disabled explicitly, it also imply
7294 @option{-fno-section-anchors} that is otherwise enabled at @option{-O0} on some
7299 Constructs webs as commonly used for register allocation purposes and assign
7300 each web individual pseudo register. This allows the register allocation pass
7301 to operate on pseudos directly, but also strengthens several other optimization
7302 passes, such as CSE, loop optimizer and trivial dead code remover. It can,
7303 however, make debugging impossible, since variables will no longer stay in a
7306 Enabled by default with @option{-funroll-loops}.
7308 @item -fwhole-program
7309 @opindex fwhole-program
7310 Assume that the current compilation unit represents the whole program being
7311 compiled. All public functions and variables with the exception of @code{main}
7312 and those merged by attribute @code{externally_visible} become static functions
7313 and in effect are optimized more aggressively by interprocedural optimizers. If @command{gold} is used as the linker plugin, @code{externally_visible} attributes are automatically added to functions (not variable yet due to a current @command{gold} issue) that are accessed outside of LTO objects according to resolution file produced by @command{gold}. For other linkers that cannot generate resolution file, explicit @code{externally_visible} attributes are still necessary.
7314 While this option is equivalent to proper use of the @code{static} keyword for
7315 programs consisting of a single file, in combination with option
7316 @option{-combine}, @option{-flto} or @option{-fwhopr} this flag can be used to
7317 compile many smaller scale programs since the functions and variables become
7318 local for the whole combined compilation unit, not for the single source file
7321 This option implies @option{-fwhole-file} for Fortran programs.
7325 This option runs the standard link-time optimizer. When invoked
7326 with source code, it generates GIMPLE (one of GCC's internal
7327 representations) and writes it to special ELF sections in the object
7328 file. When the object files are linked together, all the function
7329 bodies are read from these ELF sections and instantiated as if they
7330 had been part of the same translation unit.
7332 To use the link-timer optimizer, @option{-flto} needs to be specified at
7333 compile time and during the final link. For example,
7336 gcc -c -O2 -flto foo.c
7337 gcc -c -O2 -flto bar.c
7338 gcc -o myprog -flto -O2 foo.o bar.o
7341 The first two invocations to GCC will save a bytecode representation
7342 of GIMPLE into special ELF sections inside @file{foo.o} and
7343 @file{bar.o}. The final invocation will read the GIMPLE bytecode from
7344 @file{foo.o} and @file{bar.o}, merge the two files into a single
7345 internal image, and compile the result as usual. Since both
7346 @file{foo.o} and @file{bar.o} are merged into a single image, this
7347 causes all the inter-procedural analyses and optimizations in GCC to
7348 work across the two files as if they were a single one. This means,
7349 for example, that the inliner will be able to inline functions in
7350 @file{bar.o} into functions in @file{foo.o} and vice-versa.
7352 Another (simpler) way to enable link-time optimization is,
7355 gcc -o myprog -flto -O2 foo.c bar.c
7358 The above will generate bytecode for @file{foo.c} and @file{bar.c},
7359 merge them together into a single GIMPLE representation and optimize
7360 them as usual to produce @file{myprog}.
7362 The only important thing to keep in mind is that to enable link-time
7363 optimizations the @option{-flto} flag needs to be passed to both the
7364 compile and the link commands.
7366 Note that when a file is compiled with @option{-flto}, the generated
7367 object file will be larger than a regular object file because it will
7368 contain GIMPLE bytecodes and the usual final code. This means that
7369 object files with LTO information can be linked as a normal object
7370 file. So, in the previous example, if the final link is done with
7373 gcc -o myprog foo.o bar.o
7376 The only difference will be that no inter-procedural optimizations
7377 will be applied to produce @file{myprog}. The two object files
7378 @file{foo.o} and @file{bar.o} will be simply sent to the regular
7381 Additionally, the optimization flags used to compile individual files
7382 are not necessarily related to those used at link-time. For instance,
7385 gcc -c -O0 -flto foo.c
7386 gcc -c -O0 -flto bar.c
7387 gcc -o myprog -flto -O3 foo.o bar.o
7390 This will produce individual object files with unoptimized assembler
7391 code, but the resulting binary @file{myprog} will be optimized at
7392 @option{-O3}. Now, if the final binary is generated without
7393 @option{-flto}, then @file{myprog} will not be optimized.
7395 When producing the final binary with @option{-flto}, GCC will only
7396 apply link-time optimizations to those files that contain bytecode.
7397 Therefore, you can mix and match object files and libraries with
7398 GIMPLE bytecodes and final object code. GCC will automatically select
7399 which files to optimize in LTO mode and which files to link without
7402 There are some code generation flags that GCC will preserve when
7403 generating bytecodes, as they need to be used during the final link
7404 stage. Currently, the following options are saved into the GIMPLE
7405 bytecode files: @option{-fPIC}, @option{-fcommon} and all the
7406 @option{-m} target flags.
7408 At link time, these options are read-in and reapplied. Note that the
7409 current implementation makes no attempt at recognizing conflicting
7410 values for these options. If two or more files have a conflicting
7411 value (e.g., one file is compiled with @option{-fPIC} and another
7412 isn't), the compiler will simply use the last value read from the
7413 bytecode files. It is recommended, then, that all the files
7414 participating in the same link be compiled with the same options.
7416 Another feature of LTO is that it is possible to apply interprocedural
7417 optimizations on files written in different languages. This requires
7418 some support in the language front end. Currently, the C, C++ and
7419 Fortran front ends are capable of emitting GIMPLE bytecodes, so
7420 something like this should work
7425 gfortran -c -flto baz.f90
7426 g++ -o myprog -flto -O3 foo.o bar.o baz.o -lgfortran
7429 Notice that the final link is done with @command{g++} to get the C++
7430 runtime libraries and @option{-lgfortran} is added to get the Fortran
7431 runtime libraries. In general, when mixing languages in LTO mode, you
7432 should use the same link command used when mixing languages in a
7433 regular (non-LTO) compilation. This means that if your build process
7434 was mixing languages before, all you need to add is @option{-flto} to
7435 all the compile and link commands.
7437 If LTO encounters objects with C linkage declared with incompatible
7438 types in separate translation units to be linked together (undefined
7439 behavior according to ISO C99 6.2.7), a non-fatal diagnostic may be
7440 issued. The behavior is still undefined at runtime.
7442 If object files containing GIMPLE bytecode are stored in a library
7443 archive, say @file{libfoo.a}, it is possible to extract and use them
7444 in an LTO link if you are using @command{gold} as the linker (which,
7445 in turn requires GCC to be configured with @option{--enable-gold}).
7446 To enable this feature, use the flag @option{-fuse-linker-plugin} at
7450 gcc -o myprog -O2 -flto -fuse-linker-plugin a.o b.o -lfoo
7453 With the linker plugin enabled, @command{gold} will extract the needed
7454 GIMPLE files from @file{libfoo.a} and pass them on to the running GCC
7455 to make them part of the aggregated GIMPLE image to be optimized.
7457 If you are not using @command{gold} and/or do not specify
7458 @option{-fuse-linker-plugin} then the objects inside @file{libfoo.a}
7459 will be extracted and linked as usual, but they will not participate
7460 in the LTO optimization process.
7462 Link time optimizations do not require the presence of the whole
7463 program to operate. If the program does not require any symbols to
7464 be exported, it is possible to combine @option{-flto} and
7465 @option{-fwhopr} with @option{-fwhole-program} to allow the
7466 interprocedural optimizers to use more aggressive assumptions which
7467 may lead to improved optimization opportunities.
7469 Regarding portability: the current implementation of LTO makes no
7470 attempt at generating bytecode that can be ported between different
7471 types of hosts. The bytecode files are versioned and there is a
7472 strict version check, so bytecode files generated in one version of
7473 GCC will not work with an older/newer version of GCC.
7475 Link time optimization does not play well with generating debugging
7476 information. Combining @option{-flto} or @option{-fwhopr} with
7477 @option{-g} is experimental.
7479 This option is disabled by default.
7481 @item -fwhopr[=@var{n}]
7483 This option is identical in functionality to @option{-flto} but it
7484 differs in how the final link stage is executed. Instead of loading
7485 all the function bodies in memory, the callgraph is analyzed and
7486 optimization decisions are made (whole program analysis or WPA). Once
7487 optimization decisions are made, the callgraph is partitioned and the
7488 different sections are compiled separately (local transformations or
7489 LTRANS)@. This process allows optimizations on very large programs
7490 that otherwise would not fit in memory. This option enables
7491 @option{-fwpa} and @option{-fltrans} automatically.
7493 If you specify the optional @var{n} the link stage is executed in
7494 parallel using @var{n} parallel jobs by utilizing an installed
7495 @command{make} program. The environment variable @env{MAKE} may be
7496 used to override the program used.
7498 Disabled by default.
7502 This is an internal option used by GCC when compiling with
7503 @option{-fwhopr}. You should never need to use it.
7505 This option runs the link-time optimizer in the whole-program-analysis
7506 (WPA) mode, which reads in summary information from all inputs and
7507 performs a whole-program analysis based on summary information only.
7508 It generates object files for subsequent runs of the link-time
7509 optimizer where individual object files are optimized using both
7510 summary information from the WPA mode and the actual function bodies.
7511 It then drives the LTRANS phase.
7513 Disabled by default.
7517 This is an internal option used by GCC when compiling with
7518 @option{-fwhopr}. You should never need to use it.
7520 This option runs the link-time optimizer in the local-transformation (LTRANS)
7521 mode, which reads in output from a previous run of the LTO in WPA mode.
7522 In the LTRANS mode, LTO optimizes an object and produces the final assembly.
7524 Disabled by default.
7526 @item -fltrans-output-list=@var{file}
7527 @opindex fltrans-output-list
7528 This is an internal option used by GCC when compiling with
7529 @option{-fwhopr}. You should never need to use it.
7531 This option specifies a file to which the names of LTRANS output files are
7532 written. This option is only meaningful in conjunction with @option{-fwpa}.
7534 Disabled by default.
7536 @item -flto-compression-level=@var{n}
7537 This option specifies the level of compression used for intermediate
7538 language written to LTO object files, and is only meaningful in
7539 conjunction with LTO mode (@option{-fwhopr}, @option{-flto}). Valid
7540 values are 0 (no compression) to 9 (maximum compression). Values
7541 outside this range are clamped to either 0 or 9. If the option is not
7542 given, a default balanced compression setting is used.
7545 Prints a report with internal details on the workings of the link-time
7546 optimizer. The contents of this report vary from version to version,
7547 it is meant to be useful to GCC developers when processing object
7548 files in LTO mode (via @option{-fwhopr} or @option{-flto}).
7550 Disabled by default.
7552 @item -fuse-linker-plugin
7553 Enables the extraction of objects with GIMPLE bytecode information
7554 from library archives. This option relies on features available only
7555 in @command{gold}, so to use this you must configure GCC with
7556 @option{--enable-gold}. See @option{-flto} for a description on the
7557 effect of this flag and how to use it.
7559 Disabled by default.
7561 @item -fcprop-registers
7562 @opindex fcprop-registers
7563 After register allocation and post-register allocation instruction splitting,
7564 we perform a copy-propagation pass to try to reduce scheduling dependencies
7565 and occasionally eliminate the copy.
7567 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
7569 @item -fprofile-correction
7570 @opindex fprofile-correction
7571 Profiles collected using an instrumented binary for multi-threaded programs may
7572 be inconsistent due to missed counter updates. When this option is specified,
7573 GCC will use heuristics to correct or smooth out such inconsistencies. By
7574 default, GCC will emit an error message when an inconsistent profile is detected.
7576 @item -fprofile-dir=@var{path}
7577 @opindex fprofile-dir
7579 Set the directory to search the profile data files in to @var{path}.
7580 This option affects only the profile data generated by
7581 @option{-fprofile-generate}, @option{-ftest-coverage}, @option{-fprofile-arcs}
7582 and used by @option{-fprofile-use} and @option{-fbranch-probabilities}
7583 and its related options.
7584 By default, GCC will use the current directory as @var{path}
7585 thus the profile data file will appear in the same directory as the object file.
7587 @item -fprofile-generate
7588 @itemx -fprofile-generate=@var{path}
7589 @opindex fprofile-generate
7591 Enable options usually used for instrumenting application to produce
7592 profile useful for later recompilation with profile feedback based
7593 optimization. You must use @option{-fprofile-generate} both when
7594 compiling and when linking your program.
7596 The following options are enabled: @code{-fprofile-arcs}, @code{-fprofile-values}, @code{-fvpt}.
7598 If @var{path} is specified, GCC will look at the @var{path} to find
7599 the profile feedback data files. See @option{-fprofile-dir}.
7602 @itemx -fprofile-use=@var{path}
7603 @opindex fprofile-use
7604 Enable profile feedback directed optimizations, and optimizations
7605 generally profitable only with profile feedback available.
7607 The following options are enabled: @code{-fbranch-probabilities}, @code{-fvpt},
7608 @code{-funroll-loops}, @code{-fpeel-loops}, @code{-ftracer}
7610 By default, GCC emits an error message if the feedback profiles do not
7611 match the source code. This error can be turned into a warning by using
7612 @option{-Wcoverage-mismatch}. Note this may result in poorly optimized
7615 If @var{path} is specified, GCC will look at the @var{path} to find
7616 the profile feedback data files. See @option{-fprofile-dir}.
7619 The following options control compiler behavior regarding floating
7620 point arithmetic. These options trade off between speed and
7621 correctness. All must be specifically enabled.
7625 @opindex ffloat-store
7626 Do not store floating point variables in registers, and inhibit other
7627 options that might change whether a floating point value is taken from a
7630 @cindex floating point precision
7631 This option prevents undesirable excess precision on machines such as
7632 the 68000 where the floating registers (of the 68881) keep more
7633 precision than a @code{double} is supposed to have. Similarly for the
7634 x86 architecture. For most programs, the excess precision does only
7635 good, but a few programs rely on the precise definition of IEEE floating
7636 point. Use @option{-ffloat-store} for such programs, after modifying
7637 them to store all pertinent intermediate computations into variables.
7639 @item -fexcess-precision=@var{style}
7640 @opindex fexcess-precision
7641 This option allows further control over excess precision on machines
7642 where floating-point registers have more precision than the IEEE
7643 @code{float} and @code{double} types and the processor does not
7644 support operations rounding to those types. By default,
7645 @option{-fexcess-precision=fast} is in effect; this means that
7646 operations are carried out in the precision of the registers and that
7647 it is unpredictable when rounding to the types specified in the source
7648 code takes place. When compiling C, if
7649 @option{-fexcess-precision=standard} is specified then excess
7650 precision will follow the rules specified in ISO C99; in particular,
7651 both casts and assignments cause values to be rounded to their
7652 semantic types (whereas @option{-ffloat-store} only affects
7653 assignments). This option is enabled by default for C if a strict
7654 conformance option such as @option{-std=c99} is used.
7657 @option{-fexcess-precision=standard} is not implemented for languages
7658 other than C, and has no effect if
7659 @option{-funsafe-math-optimizations} or @option{-ffast-math} is
7660 specified. On the x86, it also has no effect if @option{-mfpmath=sse}
7661 or @option{-mfpmath=sse+387} is specified; in the former case, IEEE
7662 semantics apply without excess precision, and in the latter, rounding
7667 Sets @option{-fno-math-errno}, @option{-funsafe-math-optimizations},
7668 @option{-ffinite-math-only}, @option{-fno-rounding-math},
7669 @option{-fno-signaling-nans} and @option{-fcx-limited-range}.
7671 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
7673 This option is not turned on by any @option{-O} option since
7674 it can result in incorrect output for programs which depend on
7675 an exact implementation of IEEE or ISO rules/specifications for
7676 math functions. It may, however, yield faster code for programs
7677 that do not require the guarantees of these specifications.
7679 @item -fno-math-errno
7680 @opindex fno-math-errno
7681 Do not set ERRNO after calling math functions that are executed
7682 with a single instruction, e.g., sqrt. A program that relies on
7683 IEEE exceptions for math error handling may want to use this flag
7684 for speed while maintaining IEEE arithmetic compatibility.
7686 This option is not turned on by any @option{-O} option since
7687 it can result in incorrect output for programs which depend on
7688 an exact implementation of IEEE or ISO rules/specifications for
7689 math functions. It may, however, yield faster code for programs
7690 that do not require the guarantees of these specifications.
7692 The default is @option{-fmath-errno}.
7694 On Darwin systems, the math library never sets @code{errno}. There is
7695 therefore no reason for the compiler to consider the possibility that
7696 it might, and @option{-fno-math-errno} is the default.
7698 @item -funsafe-math-optimizations
7699 @opindex funsafe-math-optimizations
7701 Allow optimizations for floating-point arithmetic that (a) assume
7702 that arguments and results are valid and (b) may violate IEEE or
7703 ANSI standards. When used at link-time, it may include libraries
7704 or startup files that change the default FPU control word or other
7705 similar optimizations.
7707 This option is not turned on by any @option{-O} option since
7708 it can result in incorrect output for programs which depend on
7709 an exact implementation of IEEE or ISO rules/specifications for
7710 math functions. It may, however, yield faster code for programs
7711 that do not require the guarantees of these specifications.
7712 Enables @option{-fno-signed-zeros}, @option{-fno-trapping-math},
7713 @option{-fassociative-math} and @option{-freciprocal-math}.
7715 The default is @option{-fno-unsafe-math-optimizations}.
7717 @item -fassociative-math
7718 @opindex fassociative-math
7720 Allow re-association of operands in series of floating-point operations.
7721 This violates the ISO C and C++ language standard by possibly changing
7722 computation result. NOTE: re-ordering may change the sign of zero as
7723 well as ignore NaNs and inhibit or create underflow or overflow (and
7724 thus cannot be used on a code which relies on rounding behavior like
7725 @code{(x + 2**52) - 2**52)}. May also reorder floating-point comparisons
7726 and thus may not be used when ordered comparisons are required.
7727 This option requires that both @option{-fno-signed-zeros} and
7728 @option{-fno-trapping-math} be in effect. Moreover, it doesn't make
7729 much sense with @option{-frounding-math}. For Fortran the option
7730 is automatically enabled when both @option{-fno-signed-zeros} and
7731 @option{-fno-trapping-math} are in effect.
7733 The default is @option{-fno-associative-math}.
7735 @item -freciprocal-math
7736 @opindex freciprocal-math
7738 Allow the reciprocal of a value to be used instead of dividing by
7739 the value if this enables optimizations. For example @code{x / y}
7740 can be replaced with @code{x * (1/y)} which is useful if @code{(1/y)}
7741 is subject to common subexpression elimination. Note that this loses
7742 precision and increases the number of flops operating on the value.
7744 The default is @option{-fno-reciprocal-math}.
7746 @item -ffinite-math-only
7747 @opindex ffinite-math-only
7748 Allow optimizations for floating-point arithmetic that assume
7749 that arguments and results are not NaNs or +-Infs.
7751 This option is not turned on by any @option{-O} option since
7752 it can result in incorrect output for programs which depend on
7753 an exact implementation of IEEE or ISO rules/specifications for
7754 math functions. It may, however, yield faster code for programs
7755 that do not require the guarantees of these specifications.
7757 The default is @option{-fno-finite-math-only}.
7759 @item -fno-signed-zeros
7760 @opindex fno-signed-zeros
7761 Allow optimizations for floating point arithmetic that ignore the
7762 signedness of zero. IEEE arithmetic specifies the behavior of
7763 distinct +0.0 and @minus{}0.0 values, which then prohibits simplification
7764 of expressions such as x+0.0 or 0.0*x (even with @option{-ffinite-math-only}).
7765 This option implies that the sign of a zero result isn't significant.
7767 The default is @option{-fsigned-zeros}.
7769 @item -fno-trapping-math
7770 @opindex fno-trapping-math
7771 Compile code assuming that floating-point operations cannot generate
7772 user-visible traps. These traps include division by zero, overflow,
7773 underflow, inexact result and invalid operation. This option requires
7774 that @option{-fno-signaling-nans} be in effect. Setting this option may
7775 allow faster code if one relies on ``non-stop'' IEEE arithmetic, for example.
7777 This option should never be turned on by any @option{-O} option since
7778 it can result in incorrect output for programs which depend on
7779 an exact implementation of IEEE or ISO rules/specifications for
7782 The default is @option{-ftrapping-math}.
7784 @item -frounding-math
7785 @opindex frounding-math
7786 Disable transformations and optimizations that assume default floating
7787 point rounding behavior. This is round-to-zero for all floating point
7788 to integer conversions, and round-to-nearest for all other arithmetic
7789 truncations. This option should be specified for programs that change
7790 the FP rounding mode dynamically, or that may be executed with a
7791 non-default rounding mode. This option disables constant folding of
7792 floating point expressions at compile-time (which may be affected by
7793 rounding mode) and arithmetic transformations that are unsafe in the
7794 presence of sign-dependent rounding modes.
7796 The default is @option{-fno-rounding-math}.
7798 This option is experimental and does not currently guarantee to
7799 disable all GCC optimizations that are affected by rounding mode.
7800 Future versions of GCC may provide finer control of this setting
7801 using C99's @code{FENV_ACCESS} pragma. This command line option
7802 will be used to specify the default state for @code{FENV_ACCESS}.
7804 @item -fsignaling-nans
7805 @opindex fsignaling-nans
7806 Compile code assuming that IEEE signaling NaNs may generate user-visible
7807 traps during floating-point operations. Setting this option disables
7808 optimizations that may change the number of exceptions visible with
7809 signaling NaNs. This option implies @option{-ftrapping-math}.
7811 This option causes the preprocessor macro @code{__SUPPORT_SNAN__} to
7814 The default is @option{-fno-signaling-nans}.
7816 This option is experimental and does not currently guarantee to
7817 disable all GCC optimizations that affect signaling NaN behavior.
7819 @item -fsingle-precision-constant
7820 @opindex fsingle-precision-constant
7821 Treat floating point constant as single precision constant instead of
7822 implicitly converting it to double precision constant.
7824 @item -fcx-limited-range
7825 @opindex fcx-limited-range
7826 When enabled, this option states that a range reduction step is not
7827 needed when performing complex division. Also, there is no checking
7828 whether the result of a complex multiplication or division is @code{NaN
7829 + I*NaN}, with an attempt to rescue the situation in that case. The
7830 default is @option{-fno-cx-limited-range}, but is enabled by
7831 @option{-ffast-math}.
7833 This option controls the default setting of the ISO C99
7834 @code{CX_LIMITED_RANGE} pragma. Nevertheless, the option applies to
7837 @item -fcx-fortran-rules
7838 @opindex fcx-fortran-rules
7839 Complex multiplication and division follow Fortran rules. Range
7840 reduction is done as part of complex division, but there is no checking
7841 whether the result of a complex multiplication or division is @code{NaN
7842 + I*NaN}, with an attempt to rescue the situation in that case.
7844 The default is @option{-fno-cx-fortran-rules}.
7848 The following options control optimizations that may improve
7849 performance, but are not enabled by any @option{-O} options. This
7850 section includes experimental options that may produce broken code.
7853 @item -fbranch-probabilities
7854 @opindex fbranch-probabilities
7855 After running a program compiled with @option{-fprofile-arcs}
7856 (@pxref{Debugging Options,, Options for Debugging Your Program or
7857 @command{gcc}}), you can compile it a second time using
7858 @option{-fbranch-probabilities}, to improve optimizations based on
7859 the number of times each branch was taken. When the program
7860 compiled with @option{-fprofile-arcs} exits it saves arc execution
7861 counts to a file called @file{@var{sourcename}.gcda} for each source
7862 file. The information in this data file is very dependent on the
7863 structure of the generated code, so you must use the same source code
7864 and the same optimization options for both compilations.
7866 With @option{-fbranch-probabilities}, GCC puts a
7867 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
7868 These can be used to improve optimization. Currently, they are only
7869 used in one place: in @file{reorg.c}, instead of guessing which path a
7870 branch is mostly to take, the @samp{REG_BR_PROB} values are used to
7871 exactly determine which path is taken more often.
7873 @item -fprofile-values
7874 @opindex fprofile-values
7875 If combined with @option{-fprofile-arcs}, it adds code so that some
7876 data about values of expressions in the program is gathered.
7878 With @option{-fbranch-probabilities}, it reads back the data gathered
7879 from profiling values of expressions and adds @samp{REG_VALUE_PROFILE}
7880 notes to instructions for their later usage in optimizations.
7882 Enabled with @option{-fprofile-generate} and @option{-fprofile-use}.
7886 If combined with @option{-fprofile-arcs}, it instructs the compiler to add
7887 a code to gather information about values of expressions.
7889 With @option{-fbranch-probabilities}, it reads back the data gathered
7890 and actually performs the optimizations based on them.
7891 Currently the optimizations include specialization of division operation
7892 using the knowledge about the value of the denominator.
7894 @item -frename-registers
7895 @opindex frename-registers
7896 Attempt to avoid false dependencies in scheduled code by making use
7897 of registers left over after register allocation. This optimization
7898 will most benefit processors with lots of registers. Depending on the
7899 debug information format adopted by the target, however, it can
7900 make debugging impossible, since variables will no longer stay in
7901 a ``home register''.
7903 Enabled by default with @option{-funroll-loops} and @option{-fpeel-loops}.
7907 Perform tail duplication to enlarge superblock size. This transformation
7908 simplifies the control flow of the function allowing other optimizations to do
7911 Enabled with @option{-fprofile-use}.
7913 @item -funroll-loops
7914 @opindex funroll-loops
7915 Unroll loops whose number of iterations can be determined at compile time or
7916 upon entry to the loop. @option{-funroll-loops} implies
7917 @option{-frerun-cse-after-loop}, @option{-fweb} and @option{-frename-registers}.
7918 It also turns on complete loop peeling (i.e.@: complete removal of loops with
7919 small constant number of iterations). This option makes code larger, and may
7920 or may not make it run faster.
7922 Enabled with @option{-fprofile-use}.
7924 @item -funroll-all-loops
7925 @opindex funroll-all-loops
7926 Unroll all loops, even if their number of iterations is uncertain when
7927 the loop is entered. This usually makes programs run more slowly.
7928 @option{-funroll-all-loops} implies the same options as
7929 @option{-funroll-loops}.
7932 @opindex fpeel-loops
7933 Peels the loops for that there is enough information that they do not
7934 roll much (from profile feedback). It also turns on complete loop peeling
7935 (i.e.@: complete removal of loops with small constant number of iterations).
7937 Enabled with @option{-fprofile-use}.
7939 @item -fmove-loop-invariants
7940 @opindex fmove-loop-invariants
7941 Enables the loop invariant motion pass in the RTL loop optimizer. Enabled
7942 at level @option{-O1}
7944 @item -funswitch-loops
7945 @opindex funswitch-loops
7946 Move branches with loop invariant conditions out of the loop, with duplicates
7947 of the loop on both branches (modified according to result of the condition).
7949 @item -ffunction-sections
7950 @itemx -fdata-sections
7951 @opindex ffunction-sections
7952 @opindex fdata-sections
7953 Place each function or data item into its own section in the output
7954 file if the target supports arbitrary sections. The name of the
7955 function or the name of the data item determines the section's name
7958 Use these options on systems where the linker can perform optimizations
7959 to improve locality of reference in the instruction space. Most systems
7960 using the ELF object format and SPARC processors running Solaris 2 have
7961 linkers with such optimizations. AIX may have these optimizations in
7964 Only use these options when there are significant benefits from doing
7965 so. When you specify these options, the assembler and linker will
7966 create larger object and executable files and will also be slower.
7967 You will not be able to use @code{gprof} on all systems if you
7968 specify this option and you may have problems with debugging if
7969 you specify both this option and @option{-g}.
7971 @item -fbranch-target-load-optimize
7972 @opindex fbranch-target-load-optimize
7973 Perform branch target register load optimization before prologue / epilogue
7975 The use of target registers can typically be exposed only during reload,
7976 thus hoisting loads out of loops and doing inter-block scheduling needs
7977 a separate optimization pass.
7979 @item -fbranch-target-load-optimize2
7980 @opindex fbranch-target-load-optimize2
7981 Perform branch target register load optimization after prologue / epilogue
7984 @item -fbtr-bb-exclusive
7985 @opindex fbtr-bb-exclusive
7986 When performing branch target register load optimization, don't reuse
7987 branch target registers in within any basic block.
7989 @item -fstack-protector
7990 @opindex fstack-protector
7991 Emit extra code to check for buffer overflows, such as stack smashing
7992 attacks. This is done by adding a guard variable to functions with
7993 vulnerable objects. This includes functions that call alloca, and
7994 functions with buffers larger than 8 bytes. The guards are initialized
7995 when a function is entered and then checked when the function exits.
7996 If a guard check fails, an error message is printed and the program exits.
7998 @item -fstack-protector-all
7999 @opindex fstack-protector-all
8000 Like @option{-fstack-protector} except that all functions are protected.
8002 @item -fsection-anchors
8003 @opindex fsection-anchors
8004 Try to reduce the number of symbolic address calculations by using
8005 shared ``anchor'' symbols to address nearby objects. This transformation
8006 can help to reduce the number of GOT entries and GOT accesses on some
8009 For example, the implementation of the following function @code{foo}:
8013 int foo (void) @{ return a + b + c; @}
8016 would usually calculate the addresses of all three variables, but if you
8017 compile it with @option{-fsection-anchors}, it will access the variables
8018 from a common anchor point instead. The effect is similar to the
8019 following pseudocode (which isn't valid C):
8024 register int *xr = &x;
8025 return xr[&a - &x] + xr[&b - &x] + xr[&c - &x];
8029 Not all targets support this option.
8031 @item --param @var{name}=@var{value}
8033 In some places, GCC uses various constants to control the amount of
8034 optimization that is done. For example, GCC will not inline functions
8035 that contain more that a certain number of instructions. You can
8036 control some of these constants on the command-line using the
8037 @option{--param} option.
8039 The names of specific parameters, and the meaning of the values, are
8040 tied to the internals of the compiler, and are subject to change
8041 without notice in future releases.
8043 In each case, the @var{value} is an integer. The allowable choices for
8044 @var{name} are given in the following table:
8047 @item struct-reorg-cold-struct-ratio
8048 The threshold ratio (as a percentage) between a structure frequency
8049 and the frequency of the hottest structure in the program. This parameter
8050 is used by struct-reorg optimization enabled by @option{-fipa-struct-reorg}.
8051 We say that if the ratio of a structure frequency, calculated by profiling,
8052 to the hottest structure frequency in the program is less than this
8053 parameter, then structure reorganization is not applied to this structure.
8056 @item predictable-branch-outcome
8057 When branch is predicted to be taken with probability lower than this threshold
8058 (in percent), then it is considered well predictable. The default is 10.
8060 @item max-crossjump-edges
8061 The maximum number of incoming edges to consider for crossjumping.
8062 The algorithm used by @option{-fcrossjumping} is @math{O(N^2)} in
8063 the number of edges incoming to each block. Increasing values mean
8064 more aggressive optimization, making the compile time increase with
8065 probably small improvement in executable size.
8067 @item min-crossjump-insns
8068 The minimum number of instructions which must be matched at the end
8069 of two blocks before crossjumping will be performed on them. This
8070 value is ignored in the case where all instructions in the block being
8071 crossjumped from are matched. The default value is 5.
8073 @item max-grow-copy-bb-insns
8074 The maximum code size expansion factor when copying basic blocks
8075 instead of jumping. The expansion is relative to a jump instruction.
8076 The default value is 8.
8078 @item max-goto-duplication-insns
8079 The maximum number of instructions to duplicate to a block that jumps
8080 to a computed goto. To avoid @math{O(N^2)} behavior in a number of
8081 passes, GCC factors computed gotos early in the compilation process,
8082 and unfactors them as late as possible. Only computed jumps at the
8083 end of a basic blocks with no more than max-goto-duplication-insns are
8084 unfactored. The default value is 8.
8086 @item max-delay-slot-insn-search
8087 The maximum number of instructions to consider when looking for an
8088 instruction to fill a delay slot. If more than this arbitrary number of
8089 instructions is searched, the time savings from filling the delay slot
8090 will be minimal so stop searching. Increasing values mean more
8091 aggressive optimization, making the compile time increase with probably
8092 small improvement in executable run time.
8094 @item max-delay-slot-live-search
8095 When trying to fill delay slots, the maximum number of instructions to
8096 consider when searching for a block with valid live register
8097 information. Increasing this arbitrarily chosen value means more
8098 aggressive optimization, increasing the compile time. This parameter
8099 should be removed when the delay slot code is rewritten to maintain the
8102 @item max-gcse-memory
8103 The approximate maximum amount of memory that will be allocated in
8104 order to perform the global common subexpression elimination
8105 optimization. If more memory than specified is required, the
8106 optimization will not be done.
8108 @item max-pending-list-length
8109 The maximum number of pending dependencies scheduling will allow
8110 before flushing the current state and starting over. Large functions
8111 with few branches or calls can create excessively large lists which
8112 needlessly consume memory and resources.
8114 @item max-inline-insns-single
8115 Several parameters control the tree inliner used in gcc.
8116 This number sets the maximum number of instructions (counted in GCC's
8117 internal representation) in a single function that the tree inliner
8118 will consider for inlining. This only affects functions declared
8119 inline and methods implemented in a class declaration (C++).
8120 The default value is 300.
8122 @item max-inline-insns-auto
8123 When you use @option{-finline-functions} (included in @option{-O3}),
8124 a lot of functions that would otherwise not be considered for inlining
8125 by the compiler will be investigated. To those functions, a different
8126 (more restrictive) limit compared to functions declared inline can
8128 The default value is 40.
8130 @item large-function-insns
8131 The limit specifying really large functions. For functions larger than this
8132 limit after inlining, inlining is constrained by
8133 @option{--param large-function-growth}. This parameter is useful primarily
8134 to avoid extreme compilation time caused by non-linear algorithms used by the
8136 The default value is 2700.
8138 @item large-function-growth
8139 Specifies maximal growth of large function caused by inlining in percents.
8140 The default value is 100 which limits large function growth to 2.0 times
8143 @item large-unit-insns
8144 The limit specifying large translation unit. Growth caused by inlining of
8145 units larger than this limit is limited by @option{--param inline-unit-growth}.
8146 For small units this might be too tight (consider unit consisting of function A
8147 that is inline and B that just calls A three time. If B is small relative to
8148 A, the growth of unit is 300\% and yet such inlining is very sane. For very
8149 large units consisting of small inlineable functions however the overall unit
8150 growth limit is needed to avoid exponential explosion of code size. Thus for
8151 smaller units, the size is increased to @option{--param large-unit-insns}
8152 before applying @option{--param inline-unit-growth}. The default is 10000
8154 @item inline-unit-growth
8155 Specifies maximal overall growth of the compilation unit caused by inlining.
8156 The default value is 30 which limits unit growth to 1.3 times the original
8159 @item ipcp-unit-growth
8160 Specifies maximal overall growth of the compilation unit caused by
8161 interprocedural constant propagation. The default value is 10 which limits
8162 unit growth to 1.1 times the original size.
8164 @item large-stack-frame
8165 The limit specifying large stack frames. While inlining the algorithm is trying
8166 to not grow past this limit too much. Default value is 256 bytes.
8168 @item large-stack-frame-growth
8169 Specifies maximal growth of large stack frames caused by inlining in percents.
8170 The default value is 1000 which limits large stack frame growth to 11 times
8173 @item max-inline-insns-recursive
8174 @itemx max-inline-insns-recursive-auto
8175 Specifies maximum number of instructions out-of-line copy of self recursive inline
8176 function can grow into by performing recursive inlining.
8178 For functions declared inline @option{--param max-inline-insns-recursive} is
8179 taken into account. For function not declared inline, recursive inlining
8180 happens only when @option{-finline-functions} (included in @option{-O3}) is
8181 enabled and @option{--param max-inline-insns-recursive-auto} is used. The
8182 default value is 450.
8184 @item max-inline-recursive-depth
8185 @itemx max-inline-recursive-depth-auto
8186 Specifies maximum recursion depth used by the recursive inlining.
8188 For functions declared inline @option{--param max-inline-recursive-depth} is
8189 taken into account. For function not declared inline, recursive inlining
8190 happens only when @option{-finline-functions} (included in @option{-O3}) is
8191 enabled and @option{--param max-inline-recursive-depth-auto} is used. The
8194 @item min-inline-recursive-probability
8195 Recursive inlining is profitable only for function having deep recursion
8196 in average and can hurt for function having little recursion depth by
8197 increasing the prologue size or complexity of function body to other
8200 When profile feedback is available (see @option{-fprofile-generate}) the actual
8201 recursion depth can be guessed from probability that function will recurse via
8202 given call expression. This parameter limits inlining only to call expression
8203 whose probability exceeds given threshold (in percents). The default value is
8206 @item early-inlining-insns
8207 Specify growth that early inliner can make. In effect it increases amount of
8208 inlining for code having large abstraction penalty. The default value is 8.
8210 @item max-early-inliner-iterations
8211 @itemx max-early-inliner-iterations
8212 Limit of iterations of early inliner. This basically bounds number of nested
8213 indirect calls early inliner can resolve. Deeper chains are still handled by
8216 @item min-vect-loop-bound
8217 The minimum number of iterations under which a loop will not get vectorized
8218 when @option{-ftree-vectorize} is used. The number of iterations after
8219 vectorization needs to be greater than the value specified by this option
8220 to allow vectorization. The default value is 0.
8222 @item max-unrolled-insns
8223 The maximum number of instructions that a loop should have if that loop
8224 is unrolled, and if the loop is unrolled, it determines how many times
8225 the loop code is unrolled.
8227 @item max-average-unrolled-insns
8228 The maximum number of instructions biased by probabilities of their execution
8229 that a loop should have if that loop is unrolled, and if the loop is unrolled,
8230 it determines how many times the loop code is unrolled.
8232 @item max-unroll-times
8233 The maximum number of unrollings of a single loop.
8235 @item max-peeled-insns
8236 The maximum number of instructions that a loop should have if that loop
8237 is peeled, and if the loop is peeled, it determines how many times
8238 the loop code is peeled.
8240 @item max-peel-times
8241 The maximum number of peelings of a single loop.
8243 @item max-completely-peeled-insns
8244 The maximum number of insns of a completely peeled loop.
8246 @item max-completely-peel-times
8247 The maximum number of iterations of a loop to be suitable for complete peeling.
8249 @item max-completely-peel-loop-nest-depth
8250 The maximum depth of a loop nest suitable for complete peeling.
8252 @item max-unswitch-insns
8253 The maximum number of insns of an unswitched loop.
8255 @item max-unswitch-level
8256 The maximum number of branches unswitched in a single loop.
8259 The minimum cost of an expensive expression in the loop invariant motion.
8261 @item iv-consider-all-candidates-bound
8262 Bound on number of candidates for induction variables below that
8263 all candidates are considered for each use in induction variable
8264 optimizations. Only the most relevant candidates are considered
8265 if there are more candidates, to avoid quadratic time complexity.
8267 @item iv-max-considered-uses
8268 The induction variable optimizations give up on loops that contain more
8269 induction variable uses.
8271 @item iv-always-prune-cand-set-bound
8272 If number of candidates in the set is smaller than this value,
8273 we always try to remove unnecessary ivs from the set during its
8274 optimization when a new iv is added to the set.
8276 @item scev-max-expr-size
8277 Bound on size of expressions used in the scalar evolutions analyzer.
8278 Large expressions slow the analyzer.
8280 @item omega-max-vars
8281 The maximum number of variables in an Omega constraint system.
8282 The default value is 128.
8284 @item omega-max-geqs
8285 The maximum number of inequalities in an Omega constraint system.
8286 The default value is 256.
8289 The maximum number of equalities in an Omega constraint system.
8290 The default value is 128.
8292 @item omega-max-wild-cards
8293 The maximum number of wildcard variables that the Omega solver will
8294 be able to insert. The default value is 18.
8296 @item omega-hash-table-size
8297 The size of the hash table in the Omega solver. The default value is
8300 @item omega-max-keys
8301 The maximal number of keys used by the Omega solver. The default
8304 @item omega-eliminate-redundant-constraints
8305 When set to 1, use expensive methods to eliminate all redundant
8306 constraints. The default value is 0.
8308 @item vect-max-version-for-alignment-checks
8309 The maximum number of runtime checks that can be performed when
8310 doing loop versioning for alignment in the vectorizer. See option
8311 ftree-vect-loop-version for more information.
8313 @item vect-max-version-for-alias-checks
8314 The maximum number of runtime checks that can be performed when
8315 doing loop versioning for alias in the vectorizer. See option
8316 ftree-vect-loop-version for more information.
8318 @item max-iterations-to-track
8320 The maximum number of iterations of a loop the brute force algorithm
8321 for analysis of # of iterations of the loop tries to evaluate.
8323 @item hot-bb-count-fraction
8324 Select fraction of the maximal count of repetitions of basic block in program
8325 given basic block needs to have to be considered hot.
8327 @item hot-bb-frequency-fraction
8328 Select fraction of the maximal frequency of executions of basic block in
8329 function given basic block needs to have to be considered hot
8331 @item max-predicted-iterations
8332 The maximum number of loop iterations we predict statically. This is useful
8333 in cases where function contain single loop with known bound and other loop
8334 with unknown. We predict the known number of iterations correctly, while
8335 the unknown number of iterations average to roughly 10. This means that the
8336 loop without bounds would appear artificially cold relative to the other one.
8338 @item align-threshold
8340 Select fraction of the maximal frequency of executions of basic block in
8341 function given basic block will get aligned.
8343 @item align-loop-iterations
8345 A loop expected to iterate at lest the selected number of iterations will get
8348 @item tracer-dynamic-coverage
8349 @itemx tracer-dynamic-coverage-feedback
8351 This value is used to limit superblock formation once the given percentage of
8352 executed instructions is covered. This limits unnecessary code size
8355 The @option{tracer-dynamic-coverage-feedback} is used only when profile
8356 feedback is available. The real profiles (as opposed to statically estimated
8357 ones) are much less balanced allowing the threshold to be larger value.
8359 @item tracer-max-code-growth
8360 Stop tail duplication once code growth has reached given percentage. This is
8361 rather hokey argument, as most of the duplicates will be eliminated later in
8362 cross jumping, so it may be set to much higher values than is the desired code
8365 @item tracer-min-branch-ratio
8367 Stop reverse growth when the reverse probability of best edge is less than this
8368 threshold (in percent).
8370 @item tracer-min-branch-ratio
8371 @itemx tracer-min-branch-ratio-feedback
8373 Stop forward growth if the best edge do have probability lower than this
8376 Similarly to @option{tracer-dynamic-coverage} two values are present, one for
8377 compilation for profile feedback and one for compilation without. The value
8378 for compilation with profile feedback needs to be more conservative (higher) in
8379 order to make tracer effective.
8381 @item max-cse-path-length
8383 Maximum number of basic blocks on path that cse considers. The default is 10.
8386 The maximum instructions CSE process before flushing. The default is 1000.
8388 @item ggc-min-expand
8390 GCC uses a garbage collector to manage its own memory allocation. This
8391 parameter specifies the minimum percentage by which the garbage
8392 collector's heap should be allowed to expand between collections.
8393 Tuning this may improve compilation speed; it has no effect on code
8396 The default is 30% + 70% * (RAM/1GB) with an upper bound of 100% when
8397 RAM >= 1GB@. If @code{getrlimit} is available, the notion of "RAM" is
8398 the smallest of actual RAM and @code{RLIMIT_DATA} or @code{RLIMIT_AS}. If
8399 GCC is not able to calculate RAM on a particular platform, the lower
8400 bound of 30% is used. Setting this parameter and
8401 @option{ggc-min-heapsize} to zero causes a full collection to occur at
8402 every opportunity. This is extremely slow, but can be useful for
8405 @item ggc-min-heapsize
8407 Minimum size of the garbage collector's heap before it begins bothering
8408 to collect garbage. The first collection occurs after the heap expands
8409 by @option{ggc-min-expand}% beyond @option{ggc-min-heapsize}. Again,
8410 tuning this may improve compilation speed, and has no effect on code
8413 The default is the smaller of RAM/8, RLIMIT_RSS, or a limit which
8414 tries to ensure that RLIMIT_DATA or RLIMIT_AS are not exceeded, but
8415 with a lower bound of 4096 (four megabytes) and an upper bound of
8416 131072 (128 megabytes). If GCC is not able to calculate RAM on a
8417 particular platform, the lower bound is used. Setting this parameter
8418 very large effectively disables garbage collection. Setting this
8419 parameter and @option{ggc-min-expand} to zero causes a full collection
8420 to occur at every opportunity.
8422 @item max-reload-search-insns
8423 The maximum number of instruction reload should look backward for equivalent
8424 register. Increasing values mean more aggressive optimization, making the
8425 compile time increase with probably slightly better performance. The default
8428 @item max-cselib-memory-locations
8429 The maximum number of memory locations cselib should take into account.
8430 Increasing values mean more aggressive optimization, making the compile time
8431 increase with probably slightly better performance. The default value is 500.
8433 @item reorder-blocks-duplicate
8434 @itemx reorder-blocks-duplicate-feedback
8436 Used by basic block reordering pass to decide whether to use unconditional
8437 branch or duplicate the code on its destination. Code is duplicated when its
8438 estimated size is smaller than this value multiplied by the estimated size of
8439 unconditional jump in the hot spots of the program.
8441 The @option{reorder-block-duplicate-feedback} is used only when profile
8442 feedback is available and may be set to higher values than
8443 @option{reorder-block-duplicate} since information about the hot spots is more
8446 @item max-sched-ready-insns
8447 The maximum number of instructions ready to be issued the scheduler should
8448 consider at any given time during the first scheduling pass. Increasing
8449 values mean more thorough searches, making the compilation time increase
8450 with probably little benefit. The default value is 100.
8452 @item max-sched-region-blocks
8453 The maximum number of blocks in a region to be considered for
8454 interblock scheduling. The default value is 10.
8456 @item max-pipeline-region-blocks
8457 The maximum number of blocks in a region to be considered for
8458 pipelining in the selective scheduler. The default value is 15.
8460 @item max-sched-region-insns
8461 The maximum number of insns in a region to be considered for
8462 interblock scheduling. The default value is 100.
8464 @item max-pipeline-region-insns
8465 The maximum number of insns in a region to be considered for
8466 pipelining in the selective scheduler. The default value is 200.
8469 The minimum probability (in percents) of reaching a source block
8470 for interblock speculative scheduling. The default value is 40.
8472 @item max-sched-extend-regions-iters
8473 The maximum number of iterations through CFG to extend regions.
8474 0 - disable region extension,
8475 N - do at most N iterations.
8476 The default value is 0.
8478 @item max-sched-insn-conflict-delay
8479 The maximum conflict delay for an insn to be considered for speculative motion.
8480 The default value is 3.
8482 @item sched-spec-prob-cutoff
8483 The minimal probability of speculation success (in percents), so that
8484 speculative insn will be scheduled.
8485 The default value is 40.
8487 @item sched-mem-true-dep-cost
8488 Minimal distance (in CPU cycles) between store and load targeting same
8489 memory locations. The default value is 1.
8491 @item selsched-max-lookahead
8492 The maximum size of the lookahead window of selective scheduling. It is a
8493 depth of search for available instructions.
8494 The default value is 50.
8496 @item selsched-max-sched-times
8497 The maximum number of times that an instruction will be scheduled during
8498 selective scheduling. This is the limit on the number of iterations
8499 through which the instruction may be pipelined. The default value is 2.
8501 @item selsched-max-insns-to-rename
8502 The maximum number of best instructions in the ready list that are considered
8503 for renaming in the selective scheduler. The default value is 2.
8505 @item max-last-value-rtl
8506 The maximum size measured as number of RTLs that can be recorded in an expression
8507 in combiner for a pseudo register as last known value of that register. The default
8510 @item integer-share-limit
8511 Small integer constants can use a shared data structure, reducing the
8512 compiler's memory usage and increasing its speed. This sets the maximum
8513 value of a shared integer constant. The default value is 256.
8515 @item min-virtual-mappings
8516 Specifies the minimum number of virtual mappings in the incremental
8517 SSA updater that should be registered to trigger the virtual mappings
8518 heuristic defined by virtual-mappings-ratio. The default value is
8521 @item virtual-mappings-ratio
8522 If the number of virtual mappings is virtual-mappings-ratio bigger
8523 than the number of virtual symbols to be updated, then the incremental
8524 SSA updater switches to a full update for those symbols. The default
8527 @item ssp-buffer-size
8528 The minimum size of buffers (i.e.@: arrays) that will receive stack smashing
8529 protection when @option{-fstack-protection} is used.
8531 @item max-jump-thread-duplication-stmts
8532 Maximum number of statements allowed in a block that needs to be
8533 duplicated when threading jumps.
8535 @item max-fields-for-field-sensitive
8536 Maximum number of fields in a structure we will treat in
8537 a field sensitive manner during pointer analysis. The default is zero
8538 for -O0, and -O1 and 100 for -Os, -O2, and -O3.
8540 @item prefetch-latency
8541 Estimate on average number of instructions that are executed before
8542 prefetch finishes. The distance we prefetch ahead is proportional
8543 to this constant. Increasing this number may also lead to less
8544 streams being prefetched (see @option{simultaneous-prefetches}).
8546 @item simultaneous-prefetches
8547 Maximum number of prefetches that can run at the same time.
8549 @item l1-cache-line-size
8550 The size of cache line in L1 cache, in bytes.
8553 The size of L1 cache, in kilobytes.
8556 The size of L2 cache, in kilobytes.
8558 @item min-insn-to-prefetch-ratio
8559 The minimum ratio between the number of instructions and the
8560 number of prefetches to enable prefetching in a loop.
8562 @item prefetch-min-insn-to-mem-ratio
8563 The minimum ratio between the number of instructions and the
8564 number of memory references to enable prefetching in a loop.
8566 @item use-canonical-types
8567 Whether the compiler should use the ``canonical'' type system. By
8568 default, this should always be 1, which uses a more efficient internal
8569 mechanism for comparing types in C++ and Objective-C++. However, if
8570 bugs in the canonical type system are causing compilation failures,
8571 set this value to 0 to disable canonical types.
8573 @item switch-conversion-max-branch-ratio
8574 Switch initialization conversion will refuse to create arrays that are
8575 bigger than @option{switch-conversion-max-branch-ratio} times the number of
8576 branches in the switch.
8578 @item max-partial-antic-length
8579 Maximum length of the partial antic set computed during the tree
8580 partial redundancy elimination optimization (@option{-ftree-pre}) when
8581 optimizing at @option{-O3} and above. For some sorts of source code
8582 the enhanced partial redundancy elimination optimization can run away,
8583 consuming all of the memory available on the host machine. This
8584 parameter sets a limit on the length of the sets that are computed,
8585 which prevents the runaway behavior. Setting a value of 0 for
8586 this parameter will allow an unlimited set length.
8588 @item sccvn-max-scc-size
8589 Maximum size of a strongly connected component (SCC) during SCCVN
8590 processing. If this limit is hit, SCCVN processing for the whole
8591 function will not be done and optimizations depending on it will
8592 be disabled. The default maximum SCC size is 10000.
8594 @item ira-max-loops-num
8595 IRA uses a regional register allocation by default. If a function
8596 contains loops more than number given by the parameter, only at most
8597 given number of the most frequently executed loops will form regions
8598 for the regional register allocation. The default value of the
8601 @item ira-max-conflict-table-size
8602 Although IRA uses a sophisticated algorithm of compression conflict
8603 table, the table can be still big for huge functions. If the conflict
8604 table for a function could be more than size in MB given by the
8605 parameter, the conflict table is not built and faster, simpler, and
8606 lower quality register allocation algorithm will be used. The
8607 algorithm do not use pseudo-register conflicts. The default value of
8608 the parameter is 2000.
8610 @item ira-loop-reserved-regs
8611 IRA can be used to evaluate more accurate register pressure in loops
8612 for decision to move loop invariants (see @option{-O3}). The number
8613 of available registers reserved for some other purposes is described
8614 by this parameter. The default value of the parameter is 2 which is
8615 minimal number of registers needed for execution of typical
8616 instruction. This value is the best found from numerous experiments.
8618 @item loop-invariant-max-bbs-in-loop
8619 Loop invariant motion can be very expensive, both in compile time and
8620 in amount of needed compile time memory, with very large loops. Loops
8621 with more basic blocks than this parameter won't have loop invariant
8622 motion optimization performed on them. The default value of the
8623 parameter is 1000 for -O1 and 10000 for -O2 and above.
8625 @item max-vartrack-size
8626 Sets a maximum number of hash table slots to use during variable
8627 tracking dataflow analysis of any function. If this limit is exceeded
8628 with variable tracking at assignments enabled, analysis for that
8629 function is retried without it, after removing all debug insns from
8630 the function. If the limit is exceeded even without debug insns, var
8631 tracking analysis is completely disabled for the function. Setting
8632 the parameter to zero makes it unlimited.
8634 @item min-nondebug-insn-uid
8635 Use uids starting at this parameter for nondebug insns. The range below
8636 the parameter is reserved exclusively for debug insns created by
8637 @option{-fvar-tracking-assignments}, but debug insns may get
8638 (non-overlapping) uids above it if the reserved range is exhausted.
8640 @item ipa-sra-ptr-growth-factor
8641 IPA-SRA will replace a pointer to an aggregate with one or more new
8642 parameters only when their cumulative size is less or equal to
8643 @option{ipa-sra-ptr-growth-factor} times the size of the original
8646 @item graphite-max-nb-scop-params
8647 To avoid exponential effects in the Graphite loop transforms, the
8648 number of parameters in a Static Control Part (SCoP) is bounded. The
8649 default value is 10 parameters. A variable whose value is unknown at
8650 compile time and defined outside a SCoP is a parameter of the SCoP.
8652 @item graphite-max-bbs-per-function
8653 To avoid exponential effects in the detection of SCoPs, the size of
8654 the functions analyzed by Graphite is bounded. The default value is
8657 @item loop-block-tile-size
8658 Loop blocking or strip mining transforms, enabled with
8659 @option{-floop-block} or @option{-floop-strip-mine}, strip mine each
8660 loop in the loop nest by a given number of iterations. The strip
8661 length can be changed using the @option{loop-block-tile-size}
8662 parameter. The default value is 51 iterations.
8667 @node Preprocessor Options
8668 @section Options Controlling the Preprocessor
8669 @cindex preprocessor options
8670 @cindex options, preprocessor
8672 These options control the C preprocessor, which is run on each C source
8673 file before actual compilation.
8675 If you use the @option{-E} option, nothing is done except preprocessing.
8676 Some of these options make sense only together with @option{-E} because
8677 they cause the preprocessor output to be unsuitable for actual
8681 @item -Wp,@var{option}
8683 You can use @option{-Wp,@var{option}} to bypass the compiler driver
8684 and pass @var{option} directly through to the preprocessor. If
8685 @var{option} contains commas, it is split into multiple options at the
8686 commas. However, many options are modified, translated or interpreted
8687 by the compiler driver before being passed to the preprocessor, and
8688 @option{-Wp} forcibly bypasses this phase. The preprocessor's direct
8689 interface is undocumented and subject to change, so whenever possible
8690 you should avoid using @option{-Wp} and let the driver handle the
8693 @item -Xpreprocessor @var{option}
8694 @opindex Xpreprocessor
8695 Pass @var{option} as an option to the preprocessor. You can use this to
8696 supply system-specific preprocessor options which GCC does not know how to
8699 If you want to pass an option that takes an argument, you must use
8700 @option{-Xpreprocessor} twice, once for the option and once for the argument.
8703 @include cppopts.texi
8705 @node Assembler Options
8706 @section Passing Options to the Assembler
8708 @c prevent bad page break with this line
8709 You can pass options to the assembler.
8712 @item -Wa,@var{option}
8714 Pass @var{option} as an option to the assembler. If @var{option}
8715 contains commas, it is split into multiple options at the commas.
8717 @item -Xassembler @var{option}
8719 Pass @var{option} as an option to the assembler. You can use this to
8720 supply system-specific assembler options which GCC does not know how to
8723 If you want to pass an option that takes an argument, you must use
8724 @option{-Xassembler} twice, once for the option and once for the argument.
8729 @section Options for Linking
8730 @cindex link options
8731 @cindex options, linking
8733 These options come into play when the compiler links object files into
8734 an executable output file. They are meaningless if the compiler is
8735 not doing a link step.
8739 @item @var{object-file-name}
8740 A file name that does not end in a special recognized suffix is
8741 considered to name an object file or library. (Object files are
8742 distinguished from libraries by the linker according to the file
8743 contents.) If linking is done, these object files are used as input
8752 If any of these options is used, then the linker is not run, and
8753 object file names should not be used as arguments. @xref{Overall
8757 @item -l@var{library}
8758 @itemx -l @var{library}
8760 Search the library named @var{library} when linking. (The second
8761 alternative with the library as a separate argument is only for
8762 POSIX compliance and is not recommended.)
8764 It makes a difference where in the command you write this option; the
8765 linker searches and processes libraries and object files in the order they
8766 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
8767 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
8768 to functions in @samp{z}, those functions may not be loaded.
8770 The linker searches a standard list of directories for the library,
8771 which is actually a file named @file{lib@var{library}.a}. The linker
8772 then uses this file as if it had been specified precisely by name.
8774 The directories searched include several standard system directories
8775 plus any that you specify with @option{-L}.
8777 Normally the files found this way are library files---archive files
8778 whose members are object files. The linker handles an archive file by
8779 scanning through it for members which define symbols that have so far
8780 been referenced but not defined. But if the file that is found is an
8781 ordinary object file, it is linked in the usual fashion. The only
8782 difference between using an @option{-l} option and specifying a file name
8783 is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
8784 and searches several directories.
8788 You need this special case of the @option{-l} option in order to
8789 link an Objective-C or Objective-C++ program.
8792 @opindex nostartfiles
8793 Do not use the standard system startup files when linking.
8794 The standard system libraries are used normally, unless @option{-nostdlib}
8795 or @option{-nodefaultlibs} is used.
8797 @item -nodefaultlibs
8798 @opindex nodefaultlibs
8799 Do not use the standard system libraries when linking.
8800 Only the libraries you specify will be passed to the linker, options
8801 specifying linkage of the system libraries, such as @code{-static-libgcc}
8802 or @code{-shared-libgcc}, will be ignored.
8803 The standard startup files are used normally, unless @option{-nostartfiles}
8804 is used. The compiler may generate calls to @code{memcmp},
8805 @code{memset}, @code{memcpy} and @code{memmove}.
8806 These entries are usually resolved by entries in
8807 libc. These entry points should be supplied through some other
8808 mechanism when this option is specified.
8812 Do not use the standard system startup files or libraries when linking.
8813 No startup files and only the libraries you specify will be passed to
8814 the linker, options specifying linkage of the system libraries, such as
8815 @code{-static-libgcc} or @code{-shared-libgcc}, will be ignored.
8816 The compiler may generate calls to @code{memcmp}, @code{memset},
8817 @code{memcpy} and @code{memmove}.
8818 These entries are usually resolved by entries in
8819 libc. These entry points should be supplied through some other
8820 mechanism when this option is specified.
8822 @cindex @option{-lgcc}, use with @option{-nostdlib}
8823 @cindex @option{-nostdlib} and unresolved references
8824 @cindex unresolved references and @option{-nostdlib}
8825 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
8826 @cindex @option{-nodefaultlibs} and unresolved references
8827 @cindex unresolved references and @option{-nodefaultlibs}
8828 One of the standard libraries bypassed by @option{-nostdlib} and
8829 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
8830 that GCC uses to overcome shortcomings of particular machines, or special
8831 needs for some languages.
8832 (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler
8833 Collection (GCC) Internals},
8834 for more discussion of @file{libgcc.a}.)
8835 In most cases, you need @file{libgcc.a} even when you want to avoid
8836 other standard libraries. In other words, when you specify @option{-nostdlib}
8837 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
8838 This ensures that you have no unresolved references to internal GCC
8839 library subroutines. (For example, @samp{__main}, used to ensure C++
8840 constructors will be called; @pxref{Collect2,,@code{collect2}, gccint,
8841 GNU Compiler Collection (GCC) Internals}.)
8845 Produce a position independent executable on targets which support it.
8846 For predictable results, you must also specify the same set of options
8847 that were used to generate code (@option{-fpie}, @option{-fPIE},
8848 or model suboptions) when you specify this option.
8852 Pass the flag @option{-export-dynamic} to the ELF linker, on targets
8853 that support it. This instructs the linker to add all symbols, not
8854 only used ones, to the dynamic symbol table. This option is needed
8855 for some uses of @code{dlopen} or to allow obtaining backtraces
8856 from within a program.
8860 Remove all symbol table and relocation information from the executable.
8864 On systems that support dynamic linking, this prevents linking with the shared
8865 libraries. On other systems, this option has no effect.
8869 Produce a shared object which can then be linked with other objects to
8870 form an executable. Not all systems support this option. For predictable
8871 results, you must also specify the same set of options that were used to
8872 generate code (@option{-fpic}, @option{-fPIC}, or model suboptions)
8873 when you specify this option.@footnote{On some systems, @samp{gcc -shared}
8874 needs to build supplementary stub code for constructors to work. On
8875 multi-libbed systems, @samp{gcc -shared} must select the correct support
8876 libraries to link against. Failing to supply the correct flags may lead
8877 to subtle defects. Supplying them in cases where they are not necessary
8880 @item -shared-libgcc
8881 @itemx -static-libgcc
8882 @opindex shared-libgcc
8883 @opindex static-libgcc
8884 On systems that provide @file{libgcc} as a shared library, these options
8885 force the use of either the shared or static version respectively.
8886 If no shared version of @file{libgcc} was built when the compiler was
8887 configured, these options have no effect.
8889 There are several situations in which an application should use the
8890 shared @file{libgcc} instead of the static version. The most common
8891 of these is when the application wishes to throw and catch exceptions
8892 across different shared libraries. In that case, each of the libraries
8893 as well as the application itself should use the shared @file{libgcc}.
8895 Therefore, the G++ and GCJ drivers automatically add
8896 @option{-shared-libgcc} whenever you build a shared library or a main
8897 executable, because C++ and Java programs typically use exceptions, so
8898 this is the right thing to do.
8900 If, instead, you use the GCC driver to create shared libraries, you may
8901 find that they will not always be linked with the shared @file{libgcc}.
8902 If GCC finds, at its configuration time, that you have a non-GNU linker
8903 or a GNU linker that does not support option @option{--eh-frame-hdr},
8904 it will link the shared version of @file{libgcc} into shared libraries
8905 by default. Otherwise, it will take advantage of the linker and optimize
8906 away the linking with the shared version of @file{libgcc}, linking with
8907 the static version of libgcc by default. This allows exceptions to
8908 propagate through such shared libraries, without incurring relocation
8909 costs at library load time.
8911 However, if a library or main executable is supposed to throw or catch
8912 exceptions, you must link it using the G++ or GCJ driver, as appropriate
8913 for the languages used in the program, or using the option
8914 @option{-shared-libgcc}, such that it is linked with the shared
8917 @item -static-libstdc++
8918 When the @command{g++} program is used to link a C++ program, it will
8919 normally automatically link against @option{libstdc++}. If
8920 @file{libstdc++} is available as a shared library, and the
8921 @option{-static} option is not used, then this will link against the
8922 shared version of @file{libstdc++}. That is normally fine. However, it
8923 is sometimes useful to freeze the version of @file{libstdc++} used by
8924 the program without going all the way to a fully static link. The
8925 @option{-static-libstdc++} option directs the @command{g++} driver to
8926 link @file{libstdc++} statically, without necessarily linking other
8927 libraries statically.
8931 Bind references to global symbols when building a shared object. Warn
8932 about any unresolved references (unless overridden by the link editor
8933 option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support
8936 @item -T @var{script}
8938 @cindex linker script
8939 Use @var{script} as the linker script. This option is supported by most
8940 systems using the GNU linker. On some targets, such as bare-board
8941 targets without an operating system, the @option{-T} option may be required
8942 when linking to avoid references to undefined symbols.
8944 @item -Xlinker @var{option}
8946 Pass @var{option} as an option to the linker. You can use this to
8947 supply system-specific linker options which GCC does not know how to
8950 If you want to pass an option that takes a separate argument, you must use
8951 @option{-Xlinker} twice, once for the option and once for the argument.
8952 For example, to pass @option{-assert definitions}, you must write
8953 @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write
8954 @option{-Xlinker "-assert definitions"}, because this passes the entire
8955 string as a single argument, which is not what the linker expects.
8957 When using the GNU linker, it is usually more convenient to pass
8958 arguments to linker options using the @option{@var{option}=@var{value}}
8959 syntax than as separate arguments. For example, you can specify
8960 @samp{-Xlinker -Map=output.map} rather than
8961 @samp{-Xlinker -Map -Xlinker output.map}. Other linkers may not support
8962 this syntax for command-line options.
8964 @item -Wl,@var{option}
8966 Pass @var{option} as an option to the linker. If @var{option} contains
8967 commas, it is split into multiple options at the commas. You can use this
8968 syntax to pass an argument to the option.
8969 For example, @samp{-Wl,-Map,output.map} passes @samp{-Map output.map} to the
8970 linker. When using the GNU linker, you can also get the same effect with
8971 @samp{-Wl,-Map=output.map}.
8973 @item -u @var{symbol}
8975 Pretend the symbol @var{symbol} is undefined, to force linking of
8976 library modules to define it. You can use @option{-u} multiple times with
8977 different symbols to force loading of additional library modules.
8980 @node Directory Options
8981 @section Options for Directory Search
8982 @cindex directory options
8983 @cindex options, directory search
8986 These options specify directories to search for header files, for
8987 libraries and for parts of the compiler:
8992 Add the directory @var{dir} to the head of the list of directories to be
8993 searched for header files. This can be used to override a system header
8994 file, substituting your own version, since these directories are
8995 searched before the system header file directories. However, you should
8996 not use this option to add directories that contain vendor-supplied
8997 system header files (use @option{-isystem} for that). If you use more than
8998 one @option{-I} option, the directories are scanned in left-to-right
8999 order; the standard system directories come after.
9001 If a standard system include directory, or a directory specified with
9002 @option{-isystem}, is also specified with @option{-I}, the @option{-I}
9003 option will be ignored. The directory will still be searched but as a
9004 system directory at its normal position in the system include chain.
9005 This is to ensure that GCC's procedure to fix buggy system headers and
9006 the ordering for the include_next directive are not inadvertently changed.
9007 If you really need to change the search order for system directories,
9008 use the @option{-nostdinc} and/or @option{-isystem} options.
9010 @item -iplugindir=@var{dir}
9011 Set the directory to search for plugins which are passed
9012 by @option{-fplugin=@var{name}} instead of
9013 @option{-fplugin=@var{path}/@var{name}.so}. This option is not meant
9014 to be used by the user, but only passed by the driver.
9016 @item -iquote@var{dir}
9018 Add the directory @var{dir} to the head of the list of directories to
9019 be searched for header files only for the case of @samp{#include
9020 "@var{file}"}; they are not searched for @samp{#include <@var{file}>},
9021 otherwise just like @option{-I}.
9025 Add directory @var{dir} to the list of directories to be searched
9028 @item -B@var{prefix}
9030 This option specifies where to find the executables, libraries,
9031 include files, and data files of the compiler itself.
9033 The compiler driver program runs one or more of the subprograms
9034 @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries
9035 @var{prefix} as a prefix for each program it tries to run, both with and
9036 without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}).
9038 For each subprogram to be run, the compiler driver first tries the
9039 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
9040 was not specified, the driver tries two standard prefixes, which are
9041 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc/}. If neither of
9042 those results in a file name that is found, the unmodified program
9043 name is searched for using the directories specified in your
9044 @env{PATH} environment variable.
9046 The compiler will check to see if the path provided by the @option{-B}
9047 refers to a directory, and if necessary it will add a directory
9048 separator character at the end of the path.
9050 @option{-B} prefixes that effectively specify directory names also apply
9051 to libraries in the linker, because the compiler translates these
9052 options into @option{-L} options for the linker. They also apply to
9053 includes files in the preprocessor, because the compiler translates these
9054 options into @option{-isystem} options for the preprocessor. In this case,
9055 the compiler appends @samp{include} to the prefix.
9057 The run-time support file @file{libgcc.a} can also be searched for using
9058 the @option{-B} prefix, if needed. If it is not found there, the two
9059 standard prefixes above are tried, and that is all. The file is left
9060 out of the link if it is not found by those means.
9062 Another way to specify a prefix much like the @option{-B} prefix is to use
9063 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
9066 As a special kludge, if the path provided by @option{-B} is
9067 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
9068 9, then it will be replaced by @file{[dir/]include}. This is to help
9069 with boot-strapping the compiler.
9071 @item -specs=@var{file}
9073 Process @var{file} after the compiler reads in the standard @file{specs}
9074 file, in order to override the defaults that the @file{gcc} driver
9075 program uses when determining what switches to pass to @file{cc1},
9076 @file{cc1plus}, @file{as}, @file{ld}, etc. More than one
9077 @option{-specs=@var{file}} can be specified on the command line, and they
9078 are processed in order, from left to right.
9080 @item --sysroot=@var{dir}
9082 Use @var{dir} as the logical root directory for headers and libraries.
9083 For example, if the compiler would normally search for headers in
9084 @file{/usr/include} and libraries in @file{/usr/lib}, it will instead
9085 search @file{@var{dir}/usr/include} and @file{@var{dir}/usr/lib}.
9087 If you use both this option and the @option{-isysroot} option, then
9088 the @option{--sysroot} option will apply to libraries, but the
9089 @option{-isysroot} option will apply to header files.
9091 The GNU linker (beginning with version 2.16) has the necessary support
9092 for this option. If your linker does not support this option, the
9093 header file aspect of @option{--sysroot} will still work, but the
9094 library aspect will not.
9098 This option has been deprecated. Please use @option{-iquote} instead for
9099 @option{-I} directories before the @option{-I-} and remove the @option{-I-}.
9100 Any directories you specify with @option{-I} options before the @option{-I-}
9101 option are searched only for the case of @samp{#include "@var{file}"};
9102 they are not searched for @samp{#include <@var{file}>}.
9104 If additional directories are specified with @option{-I} options after
9105 the @option{-I-}, these directories are searched for all @samp{#include}
9106 directives. (Ordinarily @emph{all} @option{-I} directories are used
9109 In addition, the @option{-I-} option inhibits the use of the current
9110 directory (where the current input file came from) as the first search
9111 directory for @samp{#include "@var{file}"}. There is no way to
9112 override this effect of @option{-I-}. With @option{-I.} you can specify
9113 searching the directory which was current when the compiler was
9114 invoked. That is not exactly the same as what the preprocessor does
9115 by default, but it is often satisfactory.
9117 @option{-I-} does not inhibit the use of the standard system directories
9118 for header files. Thus, @option{-I-} and @option{-nostdinc} are
9125 @section Specifying subprocesses and the switches to pass to them
9128 @command{gcc} is a driver program. It performs its job by invoking a
9129 sequence of other programs to do the work of compiling, assembling and
9130 linking. GCC interprets its command-line parameters and uses these to
9131 deduce which programs it should invoke, and which command-line options
9132 it ought to place on their command lines. This behavior is controlled
9133 by @dfn{spec strings}. In most cases there is one spec string for each
9134 program that GCC can invoke, but a few programs have multiple spec
9135 strings to control their behavior. The spec strings built into GCC can
9136 be overridden by using the @option{-specs=} command-line switch to specify
9139 @dfn{Spec files} are plaintext files that are used to construct spec
9140 strings. They consist of a sequence of directives separated by blank
9141 lines. The type of directive is determined by the first non-whitespace
9142 character on the line and it can be one of the following:
9145 @item %@var{command}
9146 Issues a @var{command} to the spec file processor. The commands that can
9150 @item %include <@var{file}>
9152 Search for @var{file} and insert its text at the current point in the
9155 @item %include_noerr <@var{file}>
9156 @cindex %include_noerr
9157 Just like @samp{%include}, but do not generate an error message if the include
9158 file cannot be found.
9160 @item %rename @var{old_name} @var{new_name}
9162 Rename the spec string @var{old_name} to @var{new_name}.
9166 @item *[@var{spec_name}]:
9167 This tells the compiler to create, override or delete the named spec
9168 string. All lines after this directive up to the next directive or
9169 blank line are considered to be the text for the spec string. If this
9170 results in an empty string then the spec will be deleted. (Or, if the
9171 spec did not exist, then nothing will happened.) Otherwise, if the spec
9172 does not currently exist a new spec will be created. If the spec does
9173 exist then its contents will be overridden by the text of this
9174 directive, unless the first character of that text is the @samp{+}
9175 character, in which case the text will be appended to the spec.
9177 @item [@var{suffix}]:
9178 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
9179 and up to the next directive or blank line are considered to make up the
9180 spec string for the indicated suffix. When the compiler encounters an
9181 input file with the named suffix, it will processes the spec string in
9182 order to work out how to compile that file. For example:
9189 This says that any input file whose name ends in @samp{.ZZ} should be
9190 passed to the program @samp{z-compile}, which should be invoked with the
9191 command-line switch @option{-input} and with the result of performing the
9192 @samp{%i} substitution. (See below.)
9194 As an alternative to providing a spec string, the text that follows a
9195 suffix directive can be one of the following:
9198 @item @@@var{language}
9199 This says that the suffix is an alias for a known @var{language}. This is
9200 similar to using the @option{-x} command-line switch to GCC to specify a
9201 language explicitly. For example:
9208 Says that .ZZ files are, in fact, C++ source files.
9211 This causes an error messages saying:
9214 @var{name} compiler not installed on this system.
9218 GCC already has an extensive list of suffixes built into it.
9219 This directive will add an entry to the end of the list of suffixes, but
9220 since the list is searched from the end backwards, it is effectively
9221 possible to override earlier entries using this technique.
9225 GCC has the following spec strings built into it. Spec files can
9226 override these strings or create their own. Note that individual
9227 targets can also add their own spec strings to this list.
9230 asm Options to pass to the assembler
9231 asm_final Options to pass to the assembler post-processor
9232 cpp Options to pass to the C preprocessor
9233 cc1 Options to pass to the C compiler
9234 cc1plus Options to pass to the C++ compiler
9235 endfile Object files to include at the end of the link
9236 link Options to pass to the linker
9237 lib Libraries to include on the command line to the linker
9238 libgcc Decides which GCC support library to pass to the linker
9239 linker Sets the name of the linker
9240 predefines Defines to be passed to the C preprocessor
9241 signed_char Defines to pass to CPP to say whether @code{char} is signed
9243 startfile Object files to include at the start of the link
9246 Here is a small example of a spec file:
9252 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
9255 This example renames the spec called @samp{lib} to @samp{old_lib} and
9256 then overrides the previous definition of @samp{lib} with a new one.
9257 The new definition adds in some extra command-line options before
9258 including the text of the old definition.
9260 @dfn{Spec strings} are a list of command-line options to be passed to their
9261 corresponding program. In addition, the spec strings can contain
9262 @samp{%}-prefixed sequences to substitute variable text or to
9263 conditionally insert text into the command line. Using these constructs
9264 it is possible to generate quite complex command lines.
9266 Here is a table of all defined @samp{%}-sequences for spec
9267 strings. Note that spaces are not generated automatically around the
9268 results of expanding these sequences. Therefore you can concatenate them
9269 together or combine them with constant text in a single argument.
9273 Substitute one @samp{%} into the program name or argument.
9276 Substitute the name of the input file being processed.
9279 Substitute the basename of the input file being processed.
9280 This is the substring up to (and not including) the last period
9281 and not including the directory.
9284 This is the same as @samp{%b}, but include the file suffix (text after
9288 Marks the argument containing or following the @samp{%d} as a
9289 temporary file name, so that that file will be deleted if GCC exits
9290 successfully. Unlike @samp{%g}, this contributes no text to the
9293 @item %g@var{suffix}
9294 Substitute a file name that has suffix @var{suffix} and is chosen
9295 once per compilation, and mark the argument in the same way as
9296 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
9297 name is now chosen in a way that is hard to predict even when previously
9298 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
9299 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
9300 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
9301 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
9302 was simply substituted with a file name chosen once per compilation,
9303 without regard to any appended suffix (which was therefore treated
9304 just like ordinary text), making such attacks more likely to succeed.
9306 @item %u@var{suffix}
9307 Like @samp{%g}, but generates a new temporary file name even if
9308 @samp{%u@var{suffix}} was already seen.
9310 @item %U@var{suffix}
9311 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
9312 new one if there is no such last file name. In the absence of any
9313 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
9314 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
9315 would involve the generation of two distinct file names, one
9316 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
9317 simply substituted with a file name chosen for the previous @samp{%u},
9318 without regard to any appended suffix.
9320 @item %j@var{suffix}
9321 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
9322 writable, and if save-temps is off; otherwise, substitute the name
9323 of a temporary file, just like @samp{%u}. This temporary file is not
9324 meant for communication between processes, but rather as a junk
9327 @item %|@var{suffix}
9328 @itemx %m@var{suffix}
9329 Like @samp{%g}, except if @option{-pipe} is in effect. In that case
9330 @samp{%|} substitutes a single dash and @samp{%m} substitutes nothing at
9331 all. These are the two most common ways to instruct a program that it
9332 should read from standard input or write to standard output. If you
9333 need something more elaborate you can use an @samp{%@{pipe:@code{X}@}}
9334 construct: see for example @file{f/lang-specs.h}.
9336 @item %.@var{SUFFIX}
9337 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
9338 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
9339 terminated by the next space or %.
9342 Marks the argument containing or following the @samp{%w} as the
9343 designated output file of this compilation. This puts the argument
9344 into the sequence of arguments that @samp{%o} will substitute later.
9347 Substitutes the names of all the output files, with spaces
9348 automatically placed around them. You should write spaces
9349 around the @samp{%o} as well or the results are undefined.
9350 @samp{%o} is for use in the specs for running the linker.
9351 Input files whose names have no recognized suffix are not compiled
9352 at all, but they are included among the output files, so they will
9356 Substitutes the suffix for object files. Note that this is
9357 handled specially when it immediately follows @samp{%g, %u, or %U},
9358 because of the need for those to form complete file names. The
9359 handling is such that @samp{%O} is treated exactly as if it had already
9360 been substituted, except that @samp{%g, %u, and %U} do not currently
9361 support additional @var{suffix} characters following @samp{%O} as they would
9362 following, for example, @samp{.o}.
9365 Substitutes the standard macro predefinitions for the
9366 current target machine. Use this when running @code{cpp}.
9369 Like @samp{%p}, but puts @samp{__} before and after the name of each
9370 predefined macro, except for macros that start with @samp{__} or with
9371 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
9375 Substitute any of @option{-iprefix} (made from @env{GCC_EXEC_PREFIX}),
9376 @option{-isysroot} (made from @env{TARGET_SYSTEM_ROOT}),
9377 @option{-isystem} (made from @env{COMPILER_PATH} and @option{-B} options)
9378 and @option{-imultilib} as necessary.
9381 Current argument is the name of a library or startup file of some sort.
9382 Search for that file in a standard list of directories and substitute
9383 the full name found. The current working directory is included in the
9384 list of directories scanned.
9387 Current argument is the name of a linker script. Search for that file
9388 in the current list of directories to scan for libraries. If the file
9389 is located insert a @option{--script} option into the command line
9390 followed by the full path name found. If the file is not found then
9391 generate an error message. Note: the current working directory is not
9395 Print @var{str} as an error message. @var{str} is terminated by a newline.
9396 Use this when inconsistent options are detected.
9399 Substitute the contents of spec string @var{name} at this point.
9402 Like @samp{%(@dots{})} but put @samp{__} around @option{-D} arguments.
9404 @item %x@{@var{option}@}
9405 Accumulate an option for @samp{%X}.
9408 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
9412 Output the accumulated assembler options specified by @option{-Wa}.
9415 Output the accumulated preprocessor options specified by @option{-Wp}.
9418 Process the @code{asm} spec. This is used to compute the
9419 switches to be passed to the assembler.
9422 Process the @code{asm_final} spec. This is a spec string for
9423 passing switches to an assembler post-processor, if such a program is
9427 Process the @code{link} spec. This is the spec for computing the
9428 command line passed to the linker. Typically it will make use of the
9429 @samp{%L %G %S %D and %E} sequences.
9432 Dump out a @option{-L} option for each directory that GCC believes might
9433 contain startup files. If the target supports multilibs then the
9434 current multilib directory will be prepended to each of these paths.
9437 Process the @code{lib} spec. This is a spec string for deciding which
9438 libraries should be included on the command line to the linker.
9441 Process the @code{libgcc} spec. This is a spec string for deciding
9442 which GCC support library should be included on the command line to the linker.
9445 Process the @code{startfile} spec. This is a spec for deciding which
9446 object files should be the first ones passed to the linker. Typically
9447 this might be a file named @file{crt0.o}.
9450 Process the @code{endfile} spec. This is a spec string that specifies
9451 the last object files that will be passed to the linker.
9454 Process the @code{cpp} spec. This is used to construct the arguments
9455 to be passed to the C preprocessor.
9458 Process the @code{cc1} spec. This is used to construct the options to be
9459 passed to the actual C compiler (@samp{cc1}).
9462 Process the @code{cc1plus} spec. This is used to construct the options to be
9463 passed to the actual C++ compiler (@samp{cc1plus}).
9466 Substitute the variable part of a matched option. See below.
9467 Note that each comma in the substituted string is replaced by
9471 Remove all occurrences of @code{-S} from the command line. Note---this
9472 command is position dependent. @samp{%} commands in the spec string
9473 before this one will see @code{-S}, @samp{%} commands in the spec string
9474 after this one will not.
9476 @item %:@var{function}(@var{args})
9477 Call the named function @var{function}, passing it @var{args}.
9478 @var{args} is first processed as a nested spec string, then split
9479 into an argument vector in the usual fashion. The function returns
9480 a string which is processed as if it had appeared literally as part
9481 of the current spec.
9483 The following built-in spec functions are provided:
9487 The @code{getenv} spec function takes two arguments: an environment
9488 variable name and a string. If the environment variable is not
9489 defined, a fatal error is issued. Otherwise, the return value is the
9490 value of the environment variable concatenated with the string. For
9491 example, if @env{TOPDIR} is defined as @file{/path/to/top}, then:
9494 %:getenv(TOPDIR /include)
9497 expands to @file{/path/to/top/include}.
9499 @item @code{if-exists}
9500 The @code{if-exists} spec function takes one argument, an absolute
9501 pathname to a file. If the file exists, @code{if-exists} returns the
9502 pathname. Here is a small example of its usage:
9506 crt0%O%s %:if-exists(crti%O%s) crtbegin%O%s
9509 @item @code{if-exists-else}
9510 The @code{if-exists-else} spec function is similar to the @code{if-exists}
9511 spec function, except that it takes two arguments. The first argument is
9512 an absolute pathname to a file. If the file exists, @code{if-exists-else}
9513 returns the pathname. If it does not exist, it returns the second argument.
9514 This way, @code{if-exists-else} can be used to select one file or another,
9515 based on the existence of the first. Here is a small example of its usage:
9519 crt0%O%s %:if-exists(crti%O%s) \
9520 %:if-exists-else(crtbeginT%O%s crtbegin%O%s)
9523 @item @code{replace-outfile}
9524 The @code{replace-outfile} spec function takes two arguments. It looks for the
9525 first argument in the outfiles array and replaces it with the second argument. Here
9526 is a small example of its usage:
9529 %@{fgnu-runtime:%:replace-outfile(-lobjc -lobjc-gnu)@}
9532 @item @code{print-asm-header}
9533 The @code{print-asm-header} function takes no arguments and simply
9534 prints a banner like:
9540 Use "-Wa,OPTION" to pass "OPTION" to the assembler.
9543 It is used to separate compiler options from assembler options
9544 in the @option{--target-help} output.
9548 Substitutes the @code{-S} switch, if that switch was given to GCC@.
9549 If that switch was not specified, this substitutes nothing. Note that
9550 the leading dash is omitted when specifying this option, and it is
9551 automatically inserted if the substitution is performed. Thus the spec
9552 string @samp{%@{foo@}} would match the command-line option @option{-foo}
9553 and would output the command line option @option{-foo}.
9555 @item %W@{@code{S}@}
9556 Like %@{@code{S}@} but mark last argument supplied within as a file to be
9559 @item %@{@code{S}*@}
9560 Substitutes all the switches specified to GCC whose names start
9561 with @code{-S}, but which also take an argument. This is used for
9562 switches like @option{-o}, @option{-D}, @option{-I}, etc.
9563 GCC considers @option{-o foo} as being
9564 one switch whose names starts with @samp{o}. %@{o*@} would substitute this
9565 text, including the space. Thus two arguments would be generated.
9567 @item %@{@code{S}*&@code{T}*@}
9568 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
9569 (the order of @code{S} and @code{T} in the spec is not significant).
9570 There can be any number of ampersand-separated variables; for each the
9571 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
9573 @item %@{@code{S}:@code{X}@}
9574 Substitutes @code{X}, if the @samp{-S} switch was given to GCC@.
9576 @item %@{!@code{S}:@code{X}@}
9577 Substitutes @code{X}, if the @samp{-S} switch was @emph{not} given to GCC@.
9579 @item %@{@code{S}*:@code{X}@}
9580 Substitutes @code{X} if one or more switches whose names start with
9581 @code{-S} are specified to GCC@. Normally @code{X} is substituted only
9582 once, no matter how many such switches appeared. However, if @code{%*}
9583 appears somewhere in @code{X}, then @code{X} will be substituted once
9584 for each matching switch, with the @code{%*} replaced by the part of
9585 that switch that matched the @code{*}.
9587 @item %@{.@code{S}:@code{X}@}
9588 Substitutes @code{X}, if processing a file with suffix @code{S}.
9590 @item %@{!.@code{S}:@code{X}@}
9591 Substitutes @code{X}, if @emph{not} processing a file with suffix @code{S}.
9593 @item %@{,@code{S}:@code{X}@}
9594 Substitutes @code{X}, if processing a file for language @code{S}.
9596 @item %@{!,@code{S}:@code{X}@}
9597 Substitutes @code{X}, if not processing a file for language @code{S}.
9599 @item %@{@code{S}|@code{P}:@code{X}@}
9600 Substitutes @code{X} if either @code{-S} or @code{-P} was given to
9601 GCC@. This may be combined with @samp{!}, @samp{.}, @samp{,}, and
9602 @code{*} sequences as well, although they have a stronger binding than
9603 the @samp{|}. If @code{%*} appears in @code{X}, all of the
9604 alternatives must be starred, and only the first matching alternative
9607 For example, a spec string like this:
9610 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
9613 will output the following command-line options from the following input
9614 command-line options:
9619 -d fred.c -foo -baz -boggle
9620 -d jim.d -bar -baz -boggle
9623 @item %@{S:X; T:Y; :D@}
9625 If @code{S} was given to GCC, substitutes @code{X}; else if @code{T} was
9626 given to GCC, substitutes @code{Y}; else substitutes @code{D}. There can
9627 be as many clauses as you need. This may be combined with @code{.},
9628 @code{,}, @code{!}, @code{|}, and @code{*} as needed.
9633 The conditional text @code{X} in a %@{@code{S}:@code{X}@} or similar
9634 construct may contain other nested @samp{%} constructs or spaces, or
9635 even newlines. They are processed as usual, as described above.
9636 Trailing white space in @code{X} is ignored. White space may also
9637 appear anywhere on the left side of the colon in these constructs,
9638 except between @code{.} or @code{*} and the corresponding word.
9640 The @option{-O}, @option{-f}, @option{-m}, and @option{-W} switches are
9641 handled specifically in these constructs. If another value of
9642 @option{-O} or the negated form of a @option{-f}, @option{-m}, or
9643 @option{-W} switch is found later in the command line, the earlier
9644 switch value is ignored, except with @{@code{S}*@} where @code{S} is
9645 just one letter, which passes all matching options.
9647 The character @samp{|} at the beginning of the predicate text is used to
9648 indicate that a command should be piped to the following command, but
9649 only if @option{-pipe} is specified.
9651 It is built into GCC which switches take arguments and which do not.
9652 (You might think it would be useful to generalize this to allow each
9653 compiler's spec to say which switches take arguments. But this cannot
9654 be done in a consistent fashion. GCC cannot even decide which input
9655 files have been specified without knowing which switches take arguments,
9656 and it must know which input files to compile in order to tell which
9659 GCC also knows implicitly that arguments starting in @option{-l} are to be
9660 treated as compiler output files, and passed to the linker in their
9661 proper position among the other output files.
9663 @c man begin OPTIONS
9665 @node Target Options
9666 @section Specifying Target Machine and Compiler Version
9667 @cindex target options
9668 @cindex cross compiling
9669 @cindex specifying machine version
9670 @cindex specifying compiler version and target machine
9671 @cindex compiler version, specifying
9672 @cindex target machine, specifying
9674 The usual way to run GCC is to run the executable called @file{gcc}, or
9675 @file{<machine>-gcc} when cross-compiling, or
9676 @file{<machine>-gcc-<version>} to run a version other than the one that
9679 @node Submodel Options
9680 @section Hardware Models and Configurations
9681 @cindex submodel options
9682 @cindex specifying hardware config
9683 @cindex hardware models and configurations, specifying
9684 @cindex machine dependent options
9686 Each target machine types can have its own
9687 special options, starting with @samp{-m}, to choose among various
9688 hardware models or configurations---for example, 68010 vs 68020,
9689 floating coprocessor or none. A single installed version of the
9690 compiler can compile for any model or configuration, according to the
9693 Some configurations of the compiler also support additional special
9694 options, usually for compatibility with other compilers on the same
9697 @c This list is ordered alphanumerically by subsection name.
9698 @c It should be the same order and spelling as these options are listed
9699 @c in Machine Dependent Options
9705 * Blackfin Options::
9709 * DEC Alpha Options::
9710 * DEC Alpha/VMS Options::
9713 * GNU/Linux Options::
9716 * i386 and x86-64 Options::
9717 * i386 and x86-64 Windows Options::
9719 * IA-64/VMS Options::
9731 * picoChip Options::
9733 * RS/6000 and PowerPC Options::
9735 * S/390 and zSeries Options::
9740 * System V Options::
9745 * Xstormy16 Options::
9751 @subsection ARC Options
9754 These options are defined for ARC implementations:
9759 Compile code for little endian mode. This is the default.
9763 Compile code for big endian mode.
9766 @opindex mmangle-cpu
9767 Prepend the name of the cpu to all public symbol names.
9768 In multiple-processor systems, there are many ARC variants with different
9769 instruction and register set characteristics. This flag prevents code
9770 compiled for one cpu to be linked with code compiled for another.
9771 No facility exists for handling variants that are ``almost identical''.
9772 This is an all or nothing option.
9774 @item -mcpu=@var{cpu}
9776 Compile code for ARC variant @var{cpu}.
9777 Which variants are supported depend on the configuration.
9778 All variants support @option{-mcpu=base}, this is the default.
9780 @item -mtext=@var{text-section}
9781 @itemx -mdata=@var{data-section}
9782 @itemx -mrodata=@var{readonly-data-section}
9786 Put functions, data, and readonly data in @var{text-section},
9787 @var{data-section}, and @var{readonly-data-section} respectively
9788 by default. This can be overridden with the @code{section} attribute.
9789 @xref{Variable Attributes}.
9794 @subsection ARM Options
9797 These @samp{-m} options are defined for Advanced RISC Machines (ARM)
9801 @item -mabi=@var{name}
9803 Generate code for the specified ABI@. Permissible values are: @samp{apcs-gnu},
9804 @samp{atpcs}, @samp{aapcs}, @samp{aapcs-linux} and @samp{iwmmxt}.
9807 @opindex mapcs-frame
9808 Generate a stack frame that is compliant with the ARM Procedure Call
9809 Standard for all functions, even if this is not strictly necessary for
9810 correct execution of the code. Specifying @option{-fomit-frame-pointer}
9811 with this option will cause the stack frames not to be generated for
9812 leaf functions. The default is @option{-mno-apcs-frame}.
9816 This is a synonym for @option{-mapcs-frame}.
9819 @c not currently implemented
9820 @item -mapcs-stack-check
9821 @opindex mapcs-stack-check
9822 Generate code to check the amount of stack space available upon entry to
9823 every function (that actually uses some stack space). If there is
9824 insufficient space available then either the function
9825 @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be
9826 called, depending upon the amount of stack space required. The run time
9827 system is required to provide these functions. The default is
9828 @option{-mno-apcs-stack-check}, since this produces smaller code.
9830 @c not currently implemented
9832 @opindex mapcs-float
9833 Pass floating point arguments using the float point registers. This is
9834 one of the variants of the APCS@. This option is recommended if the
9835 target hardware has a floating point unit or if a lot of floating point
9836 arithmetic is going to be performed by the code. The default is
9837 @option{-mno-apcs-float}, since integer only code is slightly increased in
9838 size if @option{-mapcs-float} is used.
9840 @c not currently implemented
9841 @item -mapcs-reentrant
9842 @opindex mapcs-reentrant
9843 Generate reentrant, position independent code. The default is
9844 @option{-mno-apcs-reentrant}.
9847 @item -mthumb-interwork
9848 @opindex mthumb-interwork
9849 Generate code which supports calling between the ARM and Thumb
9850 instruction sets. Without this option the two instruction sets cannot
9851 be reliably used inside one program. The default is
9852 @option{-mno-thumb-interwork}, since slightly larger code is generated
9853 when @option{-mthumb-interwork} is specified.
9855 @item -mno-sched-prolog
9856 @opindex mno-sched-prolog
9857 Prevent the reordering of instructions in the function prolog, or the
9858 merging of those instruction with the instructions in the function's
9859 body. This means that all functions will start with a recognizable set
9860 of instructions (or in fact one of a choice from a small set of
9861 different function prologues), and this information can be used to
9862 locate the start if functions inside an executable piece of code. The
9863 default is @option{-msched-prolog}.
9865 @item -mfloat-abi=@var{name}
9867 Specifies which floating-point ABI to use. Permissible values
9868 are: @samp{soft}, @samp{softfp} and @samp{hard}.
9870 Specifying @samp{soft} causes GCC to generate output containing
9871 library calls for floating-point operations.
9872 @samp{softfp} allows the generation of code using hardware floating-point
9873 instructions, but still uses the soft-float calling conventions.
9874 @samp{hard} allows generation of floating-point instructions
9875 and uses FPU-specific calling conventions.
9877 The default depends on the specific target configuration. Note that
9878 the hard-float and soft-float ABIs are not link-compatible; you must
9879 compile your entire program with the same ABI, and link with a
9880 compatible set of libraries.
9883 @opindex mhard-float
9884 Equivalent to @option{-mfloat-abi=hard}.
9887 @opindex msoft-float
9888 Equivalent to @option{-mfloat-abi=soft}.
9890 @item -mlittle-endian
9891 @opindex mlittle-endian
9892 Generate code for a processor running in little-endian mode. This is
9893 the default for all standard configurations.
9896 @opindex mbig-endian
9897 Generate code for a processor running in big-endian mode; the default is
9898 to compile code for a little-endian processor.
9900 @item -mwords-little-endian
9901 @opindex mwords-little-endian
9902 This option only applies when generating code for big-endian processors.
9903 Generate code for a little-endian word order but a big-endian byte
9904 order. That is, a byte order of the form @samp{32107654}. Note: this
9905 option should only be used if you require compatibility with code for
9906 big-endian ARM processors generated by versions of the compiler prior to
9909 @item -mcpu=@var{name}
9911 This specifies the name of the target ARM processor. GCC uses this name
9912 to determine what kind of instructions it can emit when generating
9913 assembly code. Permissible names are: @samp{arm2}, @samp{arm250},
9914 @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610},
9915 @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm},
9916 @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700},
9917 @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100},
9919 @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm7tdmi-s},
9920 @samp{arm710t}, @samp{arm720t}, @samp{arm740t},
9921 @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100},
9922 @samp{strongarm1110},
9923 @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920},
9924 @samp{arm920t}, @samp{arm922t}, @samp{arm946e-s}, @samp{arm966e-s},
9925 @samp{arm968e-s}, @samp{arm926ej-s}, @samp{arm940t}, @samp{arm9tdmi},
9926 @samp{arm10tdmi}, @samp{arm1020t}, @samp{arm1026ej-s},
9927 @samp{arm10e}, @samp{arm1020e}, @samp{arm1022e},
9928 @samp{arm1136j-s}, @samp{arm1136jf-s}, @samp{mpcore}, @samp{mpcorenovfp},
9929 @samp{arm1156t2-s}, @samp{arm1156t2f-s}, @samp{arm1176jz-s}, @samp{arm1176jzf-s},
9930 @samp{cortex-a5}, @samp{cortex-a8}, @samp{cortex-a9},
9931 @samp{cortex-r4}, @samp{cortex-r4f}, @samp{cortex-m4}, @samp{cortex-m3},
9934 @samp{xscale}, @samp{iwmmxt}, @samp{iwmmxt2}, @samp{ep9312}.
9936 @item -mtune=@var{name}
9938 This option is very similar to the @option{-mcpu=} option, except that
9939 instead of specifying the actual target processor type, and hence
9940 restricting which instructions can be used, it specifies that GCC should
9941 tune the performance of the code as if the target were of the type
9942 specified in this option, but still choosing the instructions that it
9943 will generate based on the cpu specified by a @option{-mcpu=} option.
9944 For some ARM implementations better performance can be obtained by using
9947 @item -march=@var{name}
9949 This specifies the name of the target ARM architecture. GCC uses this
9950 name to determine what kind of instructions it can emit when generating
9951 assembly code. This option can be used in conjunction with or instead
9952 of the @option{-mcpu=} option. Permissible names are: @samp{armv2},
9953 @samp{armv2a}, @samp{armv3}, @samp{armv3m}, @samp{armv4}, @samp{armv4t},
9954 @samp{armv5}, @samp{armv5t}, @samp{armv5e}, @samp{armv5te},
9955 @samp{armv6}, @samp{armv6j},
9956 @samp{armv6t2}, @samp{armv6z}, @samp{armv6zk}, @samp{armv6-m},
9957 @samp{armv7}, @samp{armv7-a}, @samp{armv7-r}, @samp{armv7-m},
9958 @samp{iwmmxt}, @samp{iwmmxt2}, @samp{ep9312}.
9960 @item -mfpu=@var{name}
9961 @itemx -mfpe=@var{number}
9962 @itemx -mfp=@var{number}
9966 This specifies what floating point hardware (or hardware emulation) is
9967 available on the target. Permissible names are: @samp{fpa}, @samp{fpe2},
9968 @samp{fpe3}, @samp{maverick}, @samp{vfp}, @samp{vfpv3}, @samp{vfpv3-fp16},
9969 @samp{vfpv3-d16}, @samp{vfpv3-d16-fp16}, @samp{vfpv3xd}, @samp{vfpv3xd-fp16},
9970 @samp{neon}, @samp{neon-fp16}, @samp{vfpv4}, @samp{vfpv4-d16},
9971 @samp{fpv4-sp-d16} and @samp{neon-vfpv4}.
9972 @option{-mfp} and @option{-mfpe} are synonyms for
9973 @option{-mfpu}=@samp{fpe}@var{number}, for compatibility with older versions
9976 If @option{-msoft-float} is specified this specifies the format of
9977 floating point values.
9979 If the selected floating-point hardware includes the NEON extension
9980 (e.g. @option{-mfpu}=@samp{neon}), note that floating-point
9981 operations will not be used by GCC's auto-vectorization pass unless
9982 @option{-funsafe-math-optimizations} is also specified. This is
9983 because NEON hardware does not fully implement the IEEE 754 standard for
9984 floating-point arithmetic (in particular denormal values are treated as
9985 zero), so the use of NEON instructions may lead to a loss of precision.
9987 @item -mfp16-format=@var{name}
9988 @opindex mfp16-format
9989 Specify the format of the @code{__fp16} half-precision floating-point type.
9990 Permissible names are @samp{none}, @samp{ieee}, and @samp{alternative};
9991 the default is @samp{none}, in which case the @code{__fp16} type is not
9992 defined. @xref{Half-Precision}, for more information.
9994 @item -mstructure-size-boundary=@var{n}
9995 @opindex mstructure-size-boundary
9996 The size of all structures and unions will be rounded up to a multiple
9997 of the number of bits set by this option. Permissible values are 8, 32
9998 and 64. The default value varies for different toolchains. For the COFF
9999 targeted toolchain the default value is 8. A value of 64 is only allowed
10000 if the underlying ABI supports it.
10002 Specifying the larger number can produce faster, more efficient code, but
10003 can also increase the size of the program. Different values are potentially
10004 incompatible. Code compiled with one value cannot necessarily expect to
10005 work with code or libraries compiled with another value, if they exchange
10006 information using structures or unions.
10008 @item -mabort-on-noreturn
10009 @opindex mabort-on-noreturn
10010 Generate a call to the function @code{abort} at the end of a
10011 @code{noreturn} function. It will be executed if the function tries to
10015 @itemx -mno-long-calls
10016 @opindex mlong-calls
10017 @opindex mno-long-calls
10018 Tells the compiler to perform function calls by first loading the
10019 address of the function into a register and then performing a subroutine
10020 call on this register. This switch is needed if the target function
10021 will lie outside of the 64 megabyte addressing range of the offset based
10022 version of subroutine call instruction.
10024 Even if this switch is enabled, not all function calls will be turned
10025 into long calls. The heuristic is that static functions, functions
10026 which have the @samp{short-call} attribute, functions that are inside
10027 the scope of a @samp{#pragma no_long_calls} directive and functions whose
10028 definitions have already been compiled within the current compilation
10029 unit, will not be turned into long calls. The exception to this rule is
10030 that weak function definitions, functions with the @samp{long-call}
10031 attribute or the @samp{section} attribute, and functions that are within
10032 the scope of a @samp{#pragma long_calls} directive, will always be
10033 turned into long calls.
10035 This feature is not enabled by default. Specifying
10036 @option{-mno-long-calls} will restore the default behavior, as will
10037 placing the function calls within the scope of a @samp{#pragma
10038 long_calls_off} directive. Note these switches have no effect on how
10039 the compiler generates code to handle function calls via function
10042 @item -msingle-pic-base
10043 @opindex msingle-pic-base
10044 Treat the register used for PIC addressing as read-only, rather than
10045 loading it in the prologue for each function. The run-time system is
10046 responsible for initializing this register with an appropriate value
10047 before execution begins.
10049 @item -mpic-register=@var{reg}
10050 @opindex mpic-register
10051 Specify the register to be used for PIC addressing. The default is R10
10052 unless stack-checking is enabled, when R9 is used.
10054 @item -mcirrus-fix-invalid-insns
10055 @opindex mcirrus-fix-invalid-insns
10056 @opindex mno-cirrus-fix-invalid-insns
10057 Insert NOPs into the instruction stream to in order to work around
10058 problems with invalid Maverick instruction combinations. This option
10059 is only valid if the @option{-mcpu=ep9312} option has been used to
10060 enable generation of instructions for the Cirrus Maverick floating
10061 point co-processor. This option is not enabled by default, since the
10062 problem is only present in older Maverick implementations. The default
10063 can be re-enabled by use of the @option{-mno-cirrus-fix-invalid-insns}
10066 @item -mpoke-function-name
10067 @opindex mpoke-function-name
10068 Write the name of each function into the text section, directly
10069 preceding the function prologue. The generated code is similar to this:
10073 .ascii "arm_poke_function_name", 0
10076 .word 0xff000000 + (t1 - t0)
10077 arm_poke_function_name
10079 stmfd sp!, @{fp, ip, lr, pc@}
10083 When performing a stack backtrace, code can inspect the value of
10084 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
10085 location @code{pc - 12} and the top 8 bits are set, then we know that
10086 there is a function name embedded immediately preceding this location
10087 and has length @code{((pc[-3]) & 0xff000000)}.
10091 Generate code for the Thumb instruction set. The default is to
10092 use the 32-bit ARM instruction set.
10093 This option automatically enables either 16-bit Thumb-1 or
10094 mixed 16/32-bit Thumb-2 instructions based on the @option{-mcpu=@var{name}}
10095 and @option{-march=@var{name}} options. This option is not passed to the
10096 assembler. If you want to force assembler files to be interpreted as Thumb code,
10097 either add a @samp{.thumb} directive to the source or pass the @option{-mthumb}
10098 option directly to the assembler by prefixing it with @option{-Wa}.
10101 @opindex mtpcs-frame
10102 Generate a stack frame that is compliant with the Thumb Procedure Call
10103 Standard for all non-leaf functions. (A leaf function is one that does
10104 not call any other functions.) The default is @option{-mno-tpcs-frame}.
10106 @item -mtpcs-leaf-frame
10107 @opindex mtpcs-leaf-frame
10108 Generate a stack frame that is compliant with the Thumb Procedure Call
10109 Standard for all leaf functions. (A leaf function is one that does
10110 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
10112 @item -mcallee-super-interworking
10113 @opindex mcallee-super-interworking
10114 Gives all externally visible functions in the file being compiled an ARM
10115 instruction set header which switches to Thumb mode before executing the
10116 rest of the function. This allows these functions to be called from
10117 non-interworking code. This option is not valid in AAPCS configurations
10118 because interworking is enabled by default.
10120 @item -mcaller-super-interworking
10121 @opindex mcaller-super-interworking
10122 Allows calls via function pointers (including virtual functions) to
10123 execute correctly regardless of whether the target code has been
10124 compiled for interworking or not. There is a small overhead in the cost
10125 of executing a function pointer if this option is enabled. This option
10126 is not valid in AAPCS configurations because interworking is enabled
10129 @item -mtp=@var{name}
10131 Specify the access model for the thread local storage pointer. The valid
10132 models are @option{soft}, which generates calls to @code{__aeabi_read_tp},
10133 @option{cp15}, which fetches the thread pointer from @code{cp15} directly
10134 (supported in the arm6k architecture), and @option{auto}, which uses the
10135 best available method for the selected processor. The default setting is
10138 @item -mword-relocations
10139 @opindex mword-relocations
10140 Only generate absolute relocations on word sized values (i.e. R_ARM_ABS32).
10141 This is enabled by default on targets (uClinux, SymbianOS) where the runtime
10142 loader imposes this restriction, and when @option{-fpic} or @option{-fPIC}
10145 @item -mfix-cortex-m3-ldrd
10146 @opindex mfix-cortex-m3-ldrd
10147 Some Cortex-M3 cores can cause data corruption when @code{ldrd} instructions
10148 with overlapping destination and base registers are used. This option avoids
10149 generating these instructions. This option is enabled by default when
10150 @option{-mcpu=cortex-m3} is specified.
10155 @subsection AVR Options
10156 @cindex AVR Options
10158 These options are defined for AVR implementations:
10161 @item -mmcu=@var{mcu}
10163 Specify ATMEL AVR instruction set or MCU type.
10165 Instruction set avr1 is for the minimal AVR core, not supported by the C
10166 compiler, only for assembler programs (MCU types: at90s1200, attiny10,
10167 attiny11, attiny12, attiny15, attiny28).
10169 Instruction set avr2 (default) is for the classic AVR core with up to
10170 8K program memory space (MCU types: at90s2313, at90s2323, attiny22,
10171 at90s2333, at90s2343, at90s4414, at90s4433, at90s4434, at90s8515,
10172 at90c8534, at90s8535).
10174 Instruction set avr3 is for the classic AVR core with up to 128K program
10175 memory space (MCU types: atmega103, atmega603, at43usb320, at76c711).
10177 Instruction set avr4 is for the enhanced AVR core with up to 8K program
10178 memory space (MCU types: atmega8, atmega83, atmega85).
10180 Instruction set avr5 is for the enhanced AVR core with up to 128K program
10181 memory space (MCU types: atmega16, atmega161, atmega163, atmega32, atmega323,
10182 atmega64, atmega128, at43usb355, at94k).
10184 @item -mno-interrupts
10185 @opindex mno-interrupts
10186 Generated code is not compatible with hardware interrupts.
10187 Code size will be smaller.
10189 @item -mcall-prologues
10190 @opindex mcall-prologues
10191 Functions prologues/epilogues expanded as call to appropriate
10192 subroutines. Code size will be smaller.
10195 @opindex mtiny-stack
10196 Change only the low 8 bits of the stack pointer.
10200 Assume int to be 8 bit integer. This affects the sizes of all types: A
10201 char will be 1 byte, an int will be 1 byte, a long will be 2 bytes
10202 and long long will be 4 bytes. Please note that this option does not
10203 comply to the C standards, but it will provide you with smaller code
10207 @node Blackfin Options
10208 @subsection Blackfin Options
10209 @cindex Blackfin Options
10212 @item -mcpu=@var{cpu}@r{[}-@var{sirevision}@r{]}
10214 Specifies the name of the target Blackfin processor. Currently, @var{cpu}
10215 can be one of @samp{bf512}, @samp{bf514}, @samp{bf516}, @samp{bf518},
10216 @samp{bf522}, @samp{bf523}, @samp{bf524}, @samp{bf525}, @samp{bf526},
10217 @samp{bf527}, @samp{bf531}, @samp{bf532}, @samp{bf533},
10218 @samp{bf534}, @samp{bf536}, @samp{bf537}, @samp{bf538}, @samp{bf539},
10219 @samp{bf542}, @samp{bf544}, @samp{bf547}, @samp{bf548}, @samp{bf549},
10220 @samp{bf542m}, @samp{bf544m}, @samp{bf547m}, @samp{bf548m}, @samp{bf549m},
10222 The optional @var{sirevision} specifies the silicon revision of the target
10223 Blackfin processor. Any workarounds available for the targeted silicon revision
10224 will be enabled. If @var{sirevision} is @samp{none}, no workarounds are enabled.
10225 If @var{sirevision} is @samp{any}, all workarounds for the targeted processor
10226 will be enabled. The @code{__SILICON_REVISION__} macro is defined to two
10227 hexadecimal digits representing the major and minor numbers in the silicon
10228 revision. If @var{sirevision} is @samp{none}, the @code{__SILICON_REVISION__}
10229 is not defined. If @var{sirevision} is @samp{any}, the
10230 @code{__SILICON_REVISION__} is defined to be @code{0xffff}.
10231 If this optional @var{sirevision} is not used, GCC assumes the latest known
10232 silicon revision of the targeted Blackfin processor.
10234 Support for @samp{bf561} is incomplete. For @samp{bf561},
10235 Only the processor macro is defined.
10236 Without this option, @samp{bf532} is used as the processor by default.
10237 The corresponding predefined processor macros for @var{cpu} is to
10238 be defined. And for @samp{bfin-elf} toolchain, this causes the hardware BSP
10239 provided by libgloss to be linked in if @option{-msim} is not given.
10243 Specifies that the program will be run on the simulator. This causes
10244 the simulator BSP provided by libgloss to be linked in. This option
10245 has effect only for @samp{bfin-elf} toolchain.
10246 Certain other options, such as @option{-mid-shared-library} and
10247 @option{-mfdpic}, imply @option{-msim}.
10249 @item -momit-leaf-frame-pointer
10250 @opindex momit-leaf-frame-pointer
10251 Don't keep the frame pointer in a register for leaf functions. This
10252 avoids the instructions to save, set up and restore frame pointers and
10253 makes an extra register available in leaf functions. The option
10254 @option{-fomit-frame-pointer} removes the frame pointer for all functions
10255 which might make debugging harder.
10257 @item -mspecld-anomaly
10258 @opindex mspecld-anomaly
10259 When enabled, the compiler will ensure that the generated code does not
10260 contain speculative loads after jump instructions. If this option is used,
10261 @code{__WORKAROUND_SPECULATIVE_LOADS} is defined.
10263 @item -mno-specld-anomaly
10264 @opindex mno-specld-anomaly
10265 Don't generate extra code to prevent speculative loads from occurring.
10267 @item -mcsync-anomaly
10268 @opindex mcsync-anomaly
10269 When enabled, the compiler will ensure that the generated code does not
10270 contain CSYNC or SSYNC instructions too soon after conditional branches.
10271 If this option is used, @code{__WORKAROUND_SPECULATIVE_SYNCS} is defined.
10273 @item -mno-csync-anomaly
10274 @opindex mno-csync-anomaly
10275 Don't generate extra code to prevent CSYNC or SSYNC instructions from
10276 occurring too soon after a conditional branch.
10280 When enabled, the compiler is free to take advantage of the knowledge that
10281 the entire program fits into the low 64k of memory.
10284 @opindex mno-low-64k
10285 Assume that the program is arbitrarily large. This is the default.
10287 @item -mstack-check-l1
10288 @opindex mstack-check-l1
10289 Do stack checking using information placed into L1 scratchpad memory by the
10292 @item -mid-shared-library
10293 @opindex mid-shared-library
10294 Generate code that supports shared libraries via the library ID method.
10295 This allows for execute in place and shared libraries in an environment
10296 without virtual memory management. This option implies @option{-fPIC}.
10297 With a @samp{bfin-elf} target, this option implies @option{-msim}.
10299 @item -mno-id-shared-library
10300 @opindex mno-id-shared-library
10301 Generate code that doesn't assume ID based shared libraries are being used.
10302 This is the default.
10304 @item -mleaf-id-shared-library
10305 @opindex mleaf-id-shared-library
10306 Generate code that supports shared libraries via the library ID method,
10307 but assumes that this library or executable won't link against any other
10308 ID shared libraries. That allows the compiler to use faster code for jumps
10311 @item -mno-leaf-id-shared-library
10312 @opindex mno-leaf-id-shared-library
10313 Do not assume that the code being compiled won't link against any ID shared
10314 libraries. Slower code will be generated for jump and call insns.
10316 @item -mshared-library-id=n
10317 @opindex mshared-library-id
10318 Specified the identification number of the ID based shared library being
10319 compiled. Specifying a value of 0 will generate more compact code, specifying
10320 other values will force the allocation of that number to the current
10321 library but is no more space or time efficient than omitting this option.
10325 Generate code that allows the data segment to be located in a different
10326 area of memory from the text segment. This allows for execute in place in
10327 an environment without virtual memory management by eliminating relocations
10328 against the text section.
10330 @item -mno-sep-data
10331 @opindex mno-sep-data
10332 Generate code that assumes that the data segment follows the text segment.
10333 This is the default.
10336 @itemx -mno-long-calls
10337 @opindex mlong-calls
10338 @opindex mno-long-calls
10339 Tells the compiler to perform function calls by first loading the
10340 address of the function into a register and then performing a subroutine
10341 call on this register. This switch is needed if the target function
10342 will lie outside of the 24 bit addressing range of the offset based
10343 version of subroutine call instruction.
10345 This feature is not enabled by default. Specifying
10346 @option{-mno-long-calls} will restore the default behavior. Note these
10347 switches have no effect on how the compiler generates code to handle
10348 function calls via function pointers.
10352 Link with the fast floating-point library. This library relaxes some of
10353 the IEEE floating-point standard's rules for checking inputs against
10354 Not-a-Number (NAN), in the interest of performance.
10357 @opindex minline-plt
10358 Enable inlining of PLT entries in function calls to functions that are
10359 not known to bind locally. It has no effect without @option{-mfdpic}.
10362 @opindex mmulticore
10363 Build standalone application for multicore Blackfin processor. Proper
10364 start files and link scripts will be used to support multicore.
10365 This option defines @code{__BFIN_MULTICORE}. It can only be used with
10366 @option{-mcpu=bf561@r{[}-@var{sirevision}@r{]}}. It can be used with
10367 @option{-mcorea} or @option{-mcoreb}. If it's used without
10368 @option{-mcorea} or @option{-mcoreb}, single application/dual core
10369 programming model is used. In this model, the main function of Core B
10370 should be named as coreb_main. If it's used with @option{-mcorea} or
10371 @option{-mcoreb}, one application per core programming model is used.
10372 If this option is not used, single core application programming
10377 Build standalone application for Core A of BF561 when using
10378 one application per core programming model. Proper start files
10379 and link scripts will be used to support Core A. This option
10380 defines @code{__BFIN_COREA}. It must be used with @option{-mmulticore}.
10384 Build standalone application for Core B of BF561 when using
10385 one application per core programming model. Proper start files
10386 and link scripts will be used to support Core B. This option
10387 defines @code{__BFIN_COREB}. When this option is used, coreb_main
10388 should be used instead of main. It must be used with
10389 @option{-mmulticore}.
10393 Build standalone application for SDRAM. Proper start files and
10394 link scripts will be used to put the application into SDRAM.
10395 Loader should initialize SDRAM before loading the application
10396 into SDRAM. This option defines @code{__BFIN_SDRAM}.
10400 Assume that ICPLBs are enabled at runtime. This has an effect on certain
10401 anomaly workarounds. For Linux targets, the default is to assume ICPLBs
10402 are enabled; for standalone applications the default is off.
10406 @subsection CRIS Options
10407 @cindex CRIS Options
10409 These options are defined specifically for the CRIS ports.
10412 @item -march=@var{architecture-type}
10413 @itemx -mcpu=@var{architecture-type}
10416 Generate code for the specified architecture. The choices for
10417 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
10418 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX@.
10419 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
10422 @item -mtune=@var{architecture-type}
10424 Tune to @var{architecture-type} everything applicable about the generated
10425 code, except for the ABI and the set of available instructions. The
10426 choices for @var{architecture-type} are the same as for
10427 @option{-march=@var{architecture-type}}.
10429 @item -mmax-stack-frame=@var{n}
10430 @opindex mmax-stack-frame
10431 Warn when the stack frame of a function exceeds @var{n} bytes.
10437 The options @option{-metrax4} and @option{-metrax100} are synonyms for
10438 @option{-march=v3} and @option{-march=v8} respectively.
10440 @item -mmul-bug-workaround
10441 @itemx -mno-mul-bug-workaround
10442 @opindex mmul-bug-workaround
10443 @opindex mno-mul-bug-workaround
10444 Work around a bug in the @code{muls} and @code{mulu} instructions for CPU
10445 models where it applies. This option is active by default.
10449 Enable CRIS-specific verbose debug-related information in the assembly
10450 code. This option also has the effect to turn off the @samp{#NO_APP}
10451 formatted-code indicator to the assembler at the beginning of the
10456 Do not use condition-code results from previous instruction; always emit
10457 compare and test instructions before use of condition codes.
10459 @item -mno-side-effects
10460 @opindex mno-side-effects
10461 Do not emit instructions with side-effects in addressing modes other than
10464 @item -mstack-align
10465 @itemx -mno-stack-align
10466 @itemx -mdata-align
10467 @itemx -mno-data-align
10468 @itemx -mconst-align
10469 @itemx -mno-const-align
10470 @opindex mstack-align
10471 @opindex mno-stack-align
10472 @opindex mdata-align
10473 @opindex mno-data-align
10474 @opindex mconst-align
10475 @opindex mno-const-align
10476 These options (no-options) arranges (eliminate arrangements) for the
10477 stack-frame, individual data and constants to be aligned for the maximum
10478 single data access size for the chosen CPU model. The default is to
10479 arrange for 32-bit alignment. ABI details such as structure layout are
10480 not affected by these options.
10488 Similar to the stack- data- and const-align options above, these options
10489 arrange for stack-frame, writable data and constants to all be 32-bit,
10490 16-bit or 8-bit aligned. The default is 32-bit alignment.
10492 @item -mno-prologue-epilogue
10493 @itemx -mprologue-epilogue
10494 @opindex mno-prologue-epilogue
10495 @opindex mprologue-epilogue
10496 With @option{-mno-prologue-epilogue}, the normal function prologue and
10497 epilogue that sets up the stack-frame are omitted and no return
10498 instructions or return sequences are generated in the code. Use this
10499 option only together with visual inspection of the compiled code: no
10500 warnings or errors are generated when call-saved registers must be saved,
10501 or storage for local variable needs to be allocated.
10505 @opindex mno-gotplt
10507 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
10508 instruction sequences that load addresses for functions from the PLT part
10509 of the GOT rather than (traditional on other architectures) calls to the
10510 PLT@. The default is @option{-mgotplt}.
10514 Legacy no-op option only recognized with the cris-axis-elf and
10515 cris-axis-linux-gnu targets.
10519 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
10523 This option, recognized for the cris-axis-elf arranges
10524 to link with input-output functions from a simulator library. Code,
10525 initialized data and zero-initialized data are allocated consecutively.
10529 Like @option{-sim}, but pass linker options to locate initialized data at
10530 0x40000000 and zero-initialized data at 0x80000000.
10534 @subsection CRX Options
10535 @cindex CRX Options
10537 These options are defined specifically for the CRX ports.
10543 Enable the use of multiply-accumulate instructions. Disabled by default.
10546 @opindex mpush-args
10547 Push instructions will be used to pass outgoing arguments when functions
10548 are called. Enabled by default.
10551 @node Darwin Options
10552 @subsection Darwin Options
10553 @cindex Darwin options
10555 These options are defined for all architectures running the Darwin operating
10558 FSF GCC on Darwin does not create ``fat'' object files; it will create
10559 an object file for the single architecture that it was built to
10560 target. Apple's GCC on Darwin does create ``fat'' files if multiple
10561 @option{-arch} options are used; it does so by running the compiler or
10562 linker multiple times and joining the results together with
10565 The subtype of the file created (like @samp{ppc7400} or @samp{ppc970} or
10566 @samp{i686}) is determined by the flags that specify the ISA
10567 that GCC is targetting, like @option{-mcpu} or @option{-march}. The
10568 @option{-force_cpusubtype_ALL} option can be used to override this.
10570 The Darwin tools vary in their behavior when presented with an ISA
10571 mismatch. The assembler, @file{as}, will only permit instructions to
10572 be used that are valid for the subtype of the file it is generating,
10573 so you cannot put 64-bit instructions in a @samp{ppc750} object file.
10574 The linker for shared libraries, @file{/usr/bin/libtool}, will fail
10575 and print an error if asked to create a shared library with a less
10576 restrictive subtype than its input files (for instance, trying to put
10577 a @samp{ppc970} object file in a @samp{ppc7400} library). The linker
10578 for executables, @file{ld}, will quietly give the executable the most
10579 restrictive subtype of any of its input files.
10584 Add the framework directory @var{dir} to the head of the list of
10585 directories to be searched for header files. These directories are
10586 interleaved with those specified by @option{-I} options and are
10587 scanned in a left-to-right order.
10589 A framework directory is a directory with frameworks in it. A
10590 framework is a directory with a @samp{"Headers"} and/or
10591 @samp{"PrivateHeaders"} directory contained directly in it that ends
10592 in @samp{".framework"}. The name of a framework is the name of this
10593 directory excluding the @samp{".framework"}. Headers associated with
10594 the framework are found in one of those two directories, with
10595 @samp{"Headers"} being searched first. A subframework is a framework
10596 directory that is in a framework's @samp{"Frameworks"} directory.
10597 Includes of subframework headers can only appear in a header of a
10598 framework that contains the subframework, or in a sibling subframework
10599 header. Two subframeworks are siblings if they occur in the same
10600 framework. A subframework should not have the same name as a
10601 framework, a warning will be issued if this is violated. Currently a
10602 subframework cannot have subframeworks, in the future, the mechanism
10603 may be extended to support this. The standard frameworks can be found
10604 in @samp{"/System/Library/Frameworks"} and
10605 @samp{"/Library/Frameworks"}. An example include looks like
10606 @code{#include <Framework/header.h>}, where @samp{Framework} denotes
10607 the name of the framework and header.h is found in the
10608 @samp{"PrivateHeaders"} or @samp{"Headers"} directory.
10610 @item -iframework@var{dir}
10611 @opindex iframework
10612 Like @option{-F} except the directory is a treated as a system
10613 directory. The main difference between this @option{-iframework} and
10614 @option{-F} is that with @option{-iframework} the compiler does not
10615 warn about constructs contained within header files found via
10616 @var{dir}. This option is valid only for the C family of languages.
10620 Emit debugging information for symbols that are used. For STABS
10621 debugging format, this enables @option{-feliminate-unused-debug-symbols}.
10622 This is by default ON@.
10626 Emit debugging information for all symbols and types.
10628 @item -mmacosx-version-min=@var{version}
10629 The earliest version of MacOS X that this executable will run on
10630 is @var{version}. Typical values of @var{version} include @code{10.1},
10631 @code{10.2}, and @code{10.3.9}.
10633 If the compiler was built to use the system's headers by default,
10634 then the default for this option is the system version on which the
10635 compiler is running, otherwise the default is to make choices which
10636 are compatible with as many systems and code bases as possible.
10640 Enable kernel development mode. The @option{-mkernel} option sets
10641 @option{-static}, @option{-fno-common}, @option{-fno-cxa-atexit},
10642 @option{-fno-exceptions}, @option{-fno-non-call-exceptions},
10643 @option{-fapple-kext}, @option{-fno-weak} and @option{-fno-rtti} where
10644 applicable. This mode also sets @option{-mno-altivec},
10645 @option{-msoft-float}, @option{-fno-builtin} and
10646 @option{-mlong-branch} for PowerPC targets.
10648 @item -mone-byte-bool
10649 @opindex mone-byte-bool
10650 Override the defaults for @samp{bool} so that @samp{sizeof(bool)==1}.
10651 By default @samp{sizeof(bool)} is @samp{4} when compiling for
10652 Darwin/PowerPC and @samp{1} when compiling for Darwin/x86, so this
10653 option has no effect on x86.
10655 @strong{Warning:} The @option{-mone-byte-bool} switch causes GCC
10656 to generate code that is not binary compatible with code generated
10657 without that switch. Using this switch may require recompiling all
10658 other modules in a program, including system libraries. Use this
10659 switch to conform to a non-default data model.
10661 @item -mfix-and-continue
10662 @itemx -ffix-and-continue
10663 @itemx -findirect-data
10664 @opindex mfix-and-continue
10665 @opindex ffix-and-continue
10666 @opindex findirect-data
10667 Generate code suitable for fast turn around development. Needed to
10668 enable gdb to dynamically load @code{.o} files into already running
10669 programs. @option{-findirect-data} and @option{-ffix-and-continue}
10670 are provided for backwards compatibility.
10674 Loads all members of static archive libraries.
10675 See man ld(1) for more information.
10677 @item -arch_errors_fatal
10678 @opindex arch_errors_fatal
10679 Cause the errors having to do with files that have the wrong architecture
10682 @item -bind_at_load
10683 @opindex bind_at_load
10684 Causes the output file to be marked such that the dynamic linker will
10685 bind all undefined references when the file is loaded or launched.
10689 Produce a Mach-o bundle format file.
10690 See man ld(1) for more information.
10692 @item -bundle_loader @var{executable}
10693 @opindex bundle_loader
10694 This option specifies the @var{executable} that will be loading the build
10695 output file being linked. See man ld(1) for more information.
10698 @opindex dynamiclib
10699 When passed this option, GCC will produce a dynamic library instead of
10700 an executable when linking, using the Darwin @file{libtool} command.
10702 @item -force_cpusubtype_ALL
10703 @opindex force_cpusubtype_ALL
10704 This causes GCC's output file to have the @var{ALL} subtype, instead of
10705 one controlled by the @option{-mcpu} or @option{-march} option.
10707 @item -allowable_client @var{client_name}
10708 @itemx -client_name
10709 @itemx -compatibility_version
10710 @itemx -current_version
10712 @itemx -dependency-file
10714 @itemx -dylinker_install_name
10716 @itemx -exported_symbols_list
10718 @itemx -flat_namespace
10719 @itemx -force_flat_namespace
10720 @itemx -headerpad_max_install_names
10723 @itemx -install_name
10724 @itemx -keep_private_externs
10725 @itemx -multi_module
10726 @itemx -multiply_defined
10727 @itemx -multiply_defined_unused
10729 @itemx -no_dead_strip_inits_and_terms
10730 @itemx -nofixprebinding
10731 @itemx -nomultidefs
10733 @itemx -noseglinkedit
10734 @itemx -pagezero_size
10736 @itemx -prebind_all_twolevel_modules
10737 @itemx -private_bundle
10738 @itemx -read_only_relocs
10740 @itemx -sectobjectsymbols
10744 @itemx -sectobjectsymbols
10747 @itemx -segs_read_only_addr
10748 @itemx -segs_read_write_addr
10749 @itemx -seg_addr_table
10750 @itemx -seg_addr_table_filename
10751 @itemx -seglinkedit
10753 @itemx -segs_read_only_addr
10754 @itemx -segs_read_write_addr
10755 @itemx -single_module
10757 @itemx -sub_library
10758 @itemx -sub_umbrella
10759 @itemx -twolevel_namespace
10762 @itemx -unexported_symbols_list
10763 @itemx -weak_reference_mismatches
10764 @itemx -whatsloaded
10765 @opindex allowable_client
10766 @opindex client_name
10767 @opindex compatibility_version
10768 @opindex current_version
10769 @opindex dead_strip
10770 @opindex dependency-file
10771 @opindex dylib_file
10772 @opindex dylinker_install_name
10774 @opindex exported_symbols_list
10776 @opindex flat_namespace
10777 @opindex force_flat_namespace
10778 @opindex headerpad_max_install_names
10779 @opindex image_base
10781 @opindex install_name
10782 @opindex keep_private_externs
10783 @opindex multi_module
10784 @opindex multiply_defined
10785 @opindex multiply_defined_unused
10786 @opindex noall_load
10787 @opindex no_dead_strip_inits_and_terms
10788 @opindex nofixprebinding
10789 @opindex nomultidefs
10791 @opindex noseglinkedit
10792 @opindex pagezero_size
10794 @opindex prebind_all_twolevel_modules
10795 @opindex private_bundle
10796 @opindex read_only_relocs
10798 @opindex sectobjectsymbols
10801 @opindex sectcreate
10802 @opindex sectobjectsymbols
10805 @opindex segs_read_only_addr
10806 @opindex segs_read_write_addr
10807 @opindex seg_addr_table
10808 @opindex seg_addr_table_filename
10809 @opindex seglinkedit
10811 @opindex segs_read_only_addr
10812 @opindex segs_read_write_addr
10813 @opindex single_module
10815 @opindex sub_library
10816 @opindex sub_umbrella
10817 @opindex twolevel_namespace
10820 @opindex unexported_symbols_list
10821 @opindex weak_reference_mismatches
10822 @opindex whatsloaded
10823 These options are passed to the Darwin linker. The Darwin linker man page
10824 describes them in detail.
10827 @node DEC Alpha Options
10828 @subsection DEC Alpha Options
10830 These @samp{-m} options are defined for the DEC Alpha implementations:
10833 @item -mno-soft-float
10834 @itemx -msoft-float
10835 @opindex mno-soft-float
10836 @opindex msoft-float
10837 Use (do not use) the hardware floating-point instructions for
10838 floating-point operations. When @option{-msoft-float} is specified,
10839 functions in @file{libgcc.a} will be used to perform floating-point
10840 operations. Unless they are replaced by routines that emulate the
10841 floating-point operations, or compiled in such a way as to call such
10842 emulations routines, these routines will issue floating-point
10843 operations. If you are compiling for an Alpha without floating-point
10844 operations, you must ensure that the library is built so as not to call
10847 Note that Alpha implementations without floating-point operations are
10848 required to have floating-point registers.
10851 @itemx -mno-fp-regs
10853 @opindex mno-fp-regs
10854 Generate code that uses (does not use) the floating-point register set.
10855 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
10856 register set is not used, floating point operands are passed in integer
10857 registers as if they were integers and floating-point results are passed
10858 in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence,
10859 so any function with a floating-point argument or return value called by code
10860 compiled with @option{-mno-fp-regs} must also be compiled with that
10863 A typical use of this option is building a kernel that does not use,
10864 and hence need not save and restore, any floating-point registers.
10868 The Alpha architecture implements floating-point hardware optimized for
10869 maximum performance. It is mostly compliant with the IEEE floating
10870 point standard. However, for full compliance, software assistance is
10871 required. This option generates code fully IEEE compliant code
10872 @emph{except} that the @var{inexact-flag} is not maintained (see below).
10873 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
10874 defined during compilation. The resulting code is less efficient but is
10875 able to correctly support denormalized numbers and exceptional IEEE
10876 values such as not-a-number and plus/minus infinity. Other Alpha
10877 compilers call this option @option{-ieee_with_no_inexact}.
10879 @item -mieee-with-inexact
10880 @opindex mieee-with-inexact
10881 This is like @option{-mieee} except the generated code also maintains
10882 the IEEE @var{inexact-flag}. Turning on this option causes the
10883 generated code to implement fully-compliant IEEE math. In addition to
10884 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
10885 macro. On some Alpha implementations the resulting code may execute
10886 significantly slower than the code generated by default. Since there is
10887 very little code that depends on the @var{inexact-flag}, you should
10888 normally not specify this option. Other Alpha compilers call this
10889 option @option{-ieee_with_inexact}.
10891 @item -mfp-trap-mode=@var{trap-mode}
10892 @opindex mfp-trap-mode
10893 This option controls what floating-point related traps are enabled.
10894 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
10895 The trap mode can be set to one of four values:
10899 This is the default (normal) setting. The only traps that are enabled
10900 are the ones that cannot be disabled in software (e.g., division by zero
10904 In addition to the traps enabled by @samp{n}, underflow traps are enabled
10908 Like @samp{u}, but the instructions are marked to be safe for software
10909 completion (see Alpha architecture manual for details).
10912 Like @samp{su}, but inexact traps are enabled as well.
10915 @item -mfp-rounding-mode=@var{rounding-mode}
10916 @opindex mfp-rounding-mode
10917 Selects the IEEE rounding mode. Other Alpha compilers call this option
10918 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
10923 Normal IEEE rounding mode. Floating point numbers are rounded towards
10924 the nearest machine number or towards the even machine number in case
10928 Round towards minus infinity.
10931 Chopped rounding mode. Floating point numbers are rounded towards zero.
10934 Dynamic rounding mode. A field in the floating point control register
10935 (@var{fpcr}, see Alpha architecture reference manual) controls the
10936 rounding mode in effect. The C library initializes this register for
10937 rounding towards plus infinity. Thus, unless your program modifies the
10938 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
10941 @item -mtrap-precision=@var{trap-precision}
10942 @opindex mtrap-precision
10943 In the Alpha architecture, floating point traps are imprecise. This
10944 means without software assistance it is impossible to recover from a
10945 floating trap and program execution normally needs to be terminated.
10946 GCC can generate code that can assist operating system trap handlers
10947 in determining the exact location that caused a floating point trap.
10948 Depending on the requirements of an application, different levels of
10949 precisions can be selected:
10953 Program precision. This option is the default and means a trap handler
10954 can only identify which program caused a floating point exception.
10957 Function precision. The trap handler can determine the function that
10958 caused a floating point exception.
10961 Instruction precision. The trap handler can determine the exact
10962 instruction that caused a floating point exception.
10965 Other Alpha compilers provide the equivalent options called
10966 @option{-scope_safe} and @option{-resumption_safe}.
10968 @item -mieee-conformant
10969 @opindex mieee-conformant
10970 This option marks the generated code as IEEE conformant. You must not
10971 use this option unless you also specify @option{-mtrap-precision=i} and either
10972 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
10973 is to emit the line @samp{.eflag 48} in the function prologue of the
10974 generated assembly file. Under DEC Unix, this has the effect that
10975 IEEE-conformant math library routines will be linked in.
10977 @item -mbuild-constants
10978 @opindex mbuild-constants
10979 Normally GCC examines a 32- or 64-bit integer constant to
10980 see if it can construct it from smaller constants in two or three
10981 instructions. If it cannot, it will output the constant as a literal and
10982 generate code to load it from the data segment at runtime.
10984 Use this option to require GCC to construct @emph{all} integer constants
10985 using code, even if it takes more instructions (the maximum is six).
10987 You would typically use this option to build a shared library dynamic
10988 loader. Itself a shared library, it must relocate itself in memory
10989 before it can find the variables and constants in its own data segment.
10995 Select whether to generate code to be assembled by the vendor-supplied
10996 assembler (@option{-malpha-as}) or by the GNU assembler @option{-mgas}.
11014 Indicate whether GCC should generate code to use the optional BWX,
11015 CIX, FIX and MAX instruction sets. The default is to use the instruction
11016 sets supported by the CPU type specified via @option{-mcpu=} option or that
11017 of the CPU on which GCC was built if none was specified.
11020 @itemx -mfloat-ieee
11021 @opindex mfloat-vax
11022 @opindex mfloat-ieee
11023 Generate code that uses (does not use) VAX F and G floating point
11024 arithmetic instead of IEEE single and double precision.
11026 @item -mexplicit-relocs
11027 @itemx -mno-explicit-relocs
11028 @opindex mexplicit-relocs
11029 @opindex mno-explicit-relocs
11030 Older Alpha assemblers provided no way to generate symbol relocations
11031 except via assembler macros. Use of these macros does not allow
11032 optimal instruction scheduling. GNU binutils as of version 2.12
11033 supports a new syntax that allows the compiler to explicitly mark
11034 which relocations should apply to which instructions. This option
11035 is mostly useful for debugging, as GCC detects the capabilities of
11036 the assembler when it is built and sets the default accordingly.
11039 @itemx -mlarge-data
11040 @opindex msmall-data
11041 @opindex mlarge-data
11042 When @option{-mexplicit-relocs} is in effect, static data is
11043 accessed via @dfn{gp-relative} relocations. When @option{-msmall-data}
11044 is used, objects 8 bytes long or smaller are placed in a @dfn{small data area}
11045 (the @code{.sdata} and @code{.sbss} sections) and are accessed via
11046 16-bit relocations off of the @code{$gp} register. This limits the
11047 size of the small data area to 64KB, but allows the variables to be
11048 directly accessed via a single instruction.
11050 The default is @option{-mlarge-data}. With this option the data area
11051 is limited to just below 2GB@. Programs that require more than 2GB of
11052 data must use @code{malloc} or @code{mmap} to allocate the data in the
11053 heap instead of in the program's data segment.
11055 When generating code for shared libraries, @option{-fpic} implies
11056 @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}.
11059 @itemx -mlarge-text
11060 @opindex msmall-text
11061 @opindex mlarge-text
11062 When @option{-msmall-text} is used, the compiler assumes that the
11063 code of the entire program (or shared library) fits in 4MB, and is
11064 thus reachable with a branch instruction. When @option{-msmall-data}
11065 is used, the compiler can assume that all local symbols share the
11066 same @code{$gp} value, and thus reduce the number of instructions
11067 required for a function call from 4 to 1.
11069 The default is @option{-mlarge-text}.
11071 @item -mcpu=@var{cpu_type}
11073 Set the instruction set and instruction scheduling parameters for
11074 machine type @var{cpu_type}. You can specify either the @samp{EV}
11075 style name or the corresponding chip number. GCC supports scheduling
11076 parameters for the EV4, EV5 and EV6 family of processors and will
11077 choose the default values for the instruction set from the processor
11078 you specify. If you do not specify a processor type, GCC will default
11079 to the processor on which the compiler was built.
11081 Supported values for @var{cpu_type} are
11087 Schedules as an EV4 and has no instruction set extensions.
11091 Schedules as an EV5 and has no instruction set extensions.
11095 Schedules as an EV5 and supports the BWX extension.
11100 Schedules as an EV5 and supports the BWX and MAX extensions.
11104 Schedules as an EV6 and supports the BWX, FIX, and MAX extensions.
11108 Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions.
11111 Native Linux/GNU toolchains also support the value @samp{native},
11112 which selects the best architecture option for the host processor.
11113 @option{-mcpu=native} has no effect if GCC does not recognize
11116 @item -mtune=@var{cpu_type}
11118 Set only the instruction scheduling parameters for machine type
11119 @var{cpu_type}. The instruction set is not changed.
11121 Native Linux/GNU toolchains also support the value @samp{native},
11122 which selects the best architecture option for the host processor.
11123 @option{-mtune=native} has no effect if GCC does not recognize
11126 @item -mmemory-latency=@var{time}
11127 @opindex mmemory-latency
11128 Sets the latency the scheduler should assume for typical memory
11129 references as seen by the application. This number is highly
11130 dependent on the memory access patterns used by the application
11131 and the size of the external cache on the machine.
11133 Valid options for @var{time} are
11137 A decimal number representing clock cycles.
11143 The compiler contains estimates of the number of clock cycles for
11144 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
11145 (also called Dcache, Scache, and Bcache), as well as to main memory.
11146 Note that L3 is only valid for EV5.
11151 @node DEC Alpha/VMS Options
11152 @subsection DEC Alpha/VMS Options
11154 These @samp{-m} options are defined for the DEC Alpha/VMS implementations:
11157 @item -mvms-return-codes
11158 @opindex mvms-return-codes
11159 Return VMS condition codes from main. The default is to return POSIX
11160 style condition (e.g.@: error) codes.
11162 @item -mdebug-main=@var{prefix}
11163 @opindex mdebug-main=@var{prefix}
11164 Flag the first routine whose name starts with @var{prefix} as the main
11165 routine for the debugger.
11169 Default to 64bit memory allocation routines.
11173 @subsection FR30 Options
11174 @cindex FR30 Options
11176 These options are defined specifically for the FR30 port.
11180 @item -msmall-model
11181 @opindex msmall-model
11182 Use the small address space model. This can produce smaller code, but
11183 it does assume that all symbolic values and addresses will fit into a
11188 Assume that run-time support has been provided and so there is no need
11189 to include the simulator library (@file{libsim.a}) on the linker
11195 @subsection FRV Options
11196 @cindex FRV Options
11202 Only use the first 32 general purpose registers.
11207 Use all 64 general purpose registers.
11212 Use only the first 32 floating point registers.
11217 Use all 64 floating point registers
11220 @opindex mhard-float
11222 Use hardware instructions for floating point operations.
11225 @opindex msoft-float
11227 Use library routines for floating point operations.
11232 Dynamically allocate condition code registers.
11237 Do not try to dynamically allocate condition code registers, only
11238 use @code{icc0} and @code{fcc0}.
11243 Change ABI to use double word insns.
11248 Do not use double word instructions.
11253 Use floating point double instructions.
11256 @opindex mno-double
11258 Do not use floating point double instructions.
11263 Use media instructions.
11268 Do not use media instructions.
11273 Use multiply and add/subtract instructions.
11276 @opindex mno-muladd
11278 Do not use multiply and add/subtract instructions.
11283 Select the FDPIC ABI, that uses function descriptors to represent
11284 pointers to functions. Without any PIC/PIE-related options, it
11285 implies @option{-fPIE}. With @option{-fpic} or @option{-fpie}, it
11286 assumes GOT entries and small data are within a 12-bit range from the
11287 GOT base address; with @option{-fPIC} or @option{-fPIE}, GOT offsets
11288 are computed with 32 bits.
11289 With a @samp{bfin-elf} target, this option implies @option{-msim}.
11292 @opindex minline-plt
11294 Enable inlining of PLT entries in function calls to functions that are
11295 not known to bind locally. It has no effect without @option{-mfdpic}.
11296 It's enabled by default if optimizing for speed and compiling for
11297 shared libraries (i.e., @option{-fPIC} or @option{-fpic}), or when an
11298 optimization option such as @option{-O3} or above is present in the
11304 Assume a large TLS segment when generating thread-local code.
11309 Do not assume a large TLS segment when generating thread-local code.
11314 Enable the use of @code{GPREL} relocations in the FDPIC ABI for data
11315 that is known to be in read-only sections. It's enabled by default,
11316 except for @option{-fpic} or @option{-fpie}: even though it may help
11317 make the global offset table smaller, it trades 1 instruction for 4.
11318 With @option{-fPIC} or @option{-fPIE}, it trades 3 instructions for 4,
11319 one of which may be shared by multiple symbols, and it avoids the need
11320 for a GOT entry for the referenced symbol, so it's more likely to be a
11321 win. If it is not, @option{-mno-gprel-ro} can be used to disable it.
11323 @item -multilib-library-pic
11324 @opindex multilib-library-pic
11326 Link with the (library, not FD) pic libraries. It's implied by
11327 @option{-mlibrary-pic}, as well as by @option{-fPIC} and
11328 @option{-fpic} without @option{-mfdpic}. You should never have to use
11332 @opindex mlinked-fp
11334 Follow the EABI requirement of always creating a frame pointer whenever
11335 a stack frame is allocated. This option is enabled by default and can
11336 be disabled with @option{-mno-linked-fp}.
11339 @opindex mlong-calls
11341 Use indirect addressing to call functions outside the current
11342 compilation unit. This allows the functions to be placed anywhere
11343 within the 32-bit address space.
11345 @item -malign-labels
11346 @opindex malign-labels
11348 Try to align labels to an 8-byte boundary by inserting nops into the
11349 previous packet. This option only has an effect when VLIW packing
11350 is enabled. It doesn't create new packets; it merely adds nops to
11353 @item -mlibrary-pic
11354 @opindex mlibrary-pic
11356 Generate position-independent EABI code.
11361 Use only the first four media accumulator registers.
11366 Use all eight media accumulator registers.
11371 Pack VLIW instructions.
11376 Do not pack VLIW instructions.
11379 @opindex mno-eflags
11381 Do not mark ABI switches in e_flags.
11384 @opindex mcond-move
11386 Enable the use of conditional-move instructions (default).
11388 This switch is mainly for debugging the compiler and will likely be removed
11389 in a future version.
11391 @item -mno-cond-move
11392 @opindex mno-cond-move
11394 Disable the use of conditional-move instructions.
11396 This switch is mainly for debugging the compiler and will likely be removed
11397 in a future version.
11402 Enable the use of conditional set instructions (default).
11404 This switch is mainly for debugging the compiler and will likely be removed
11405 in a future version.
11410 Disable the use of conditional set instructions.
11412 This switch is mainly for debugging the compiler and will likely be removed
11413 in a future version.
11416 @opindex mcond-exec
11418 Enable the use of conditional execution (default).
11420 This switch is mainly for debugging the compiler and will likely be removed
11421 in a future version.
11423 @item -mno-cond-exec
11424 @opindex mno-cond-exec
11426 Disable the use of conditional execution.
11428 This switch is mainly for debugging the compiler and will likely be removed
11429 in a future version.
11431 @item -mvliw-branch
11432 @opindex mvliw-branch
11434 Run a pass to pack branches into VLIW instructions (default).
11436 This switch is mainly for debugging the compiler and will likely be removed
11437 in a future version.
11439 @item -mno-vliw-branch
11440 @opindex mno-vliw-branch
11442 Do not run a pass to pack branches into VLIW instructions.
11444 This switch is mainly for debugging the compiler and will likely be removed
11445 in a future version.
11447 @item -mmulti-cond-exec
11448 @opindex mmulti-cond-exec
11450 Enable optimization of @code{&&} and @code{||} in conditional execution
11453 This switch is mainly for debugging the compiler and will likely be removed
11454 in a future version.
11456 @item -mno-multi-cond-exec
11457 @opindex mno-multi-cond-exec
11459 Disable optimization of @code{&&} and @code{||} in conditional execution.
11461 This switch is mainly for debugging the compiler and will likely be removed
11462 in a future version.
11464 @item -mnested-cond-exec
11465 @opindex mnested-cond-exec
11467 Enable nested conditional execution optimizations (default).
11469 This switch is mainly for debugging the compiler and will likely be removed
11470 in a future version.
11472 @item -mno-nested-cond-exec
11473 @opindex mno-nested-cond-exec
11475 Disable nested conditional execution optimizations.
11477 This switch is mainly for debugging the compiler and will likely be removed
11478 in a future version.
11480 @item -moptimize-membar
11481 @opindex moptimize-membar
11483 This switch removes redundant @code{membar} instructions from the
11484 compiler generated code. It is enabled by default.
11486 @item -mno-optimize-membar
11487 @opindex mno-optimize-membar
11489 This switch disables the automatic removal of redundant @code{membar}
11490 instructions from the generated code.
11492 @item -mtomcat-stats
11493 @opindex mtomcat-stats
11495 Cause gas to print out tomcat statistics.
11497 @item -mcpu=@var{cpu}
11500 Select the processor type for which to generate code. Possible values are
11501 @samp{frv}, @samp{fr550}, @samp{tomcat}, @samp{fr500}, @samp{fr450},
11502 @samp{fr405}, @samp{fr400}, @samp{fr300} and @samp{simple}.
11506 @node GNU/Linux Options
11507 @subsection GNU/Linux Options
11509 These @samp{-m} options are defined for GNU/Linux targets:
11514 Use the GNU C library. This is the default except
11515 on @samp{*-*-linux-*uclibc*} and @samp{*-*-linux-*android*} targets.
11519 Use uClibc C library. This is the default on
11520 @samp{*-*-linux-*uclibc*} targets.
11524 Use Bionic C library. This is the default on
11525 @samp{*-*-linux-*android*} targets.
11529 Compile code compatible with Android platform. This is the default on
11530 @samp{*-*-linux-*android*} targets.
11532 When compiling, this option enables @option{-mbionic}, @option{-fPIC},
11533 @option{-fno-exceptions} and @option{-fno-rtti} by default. When linking,
11534 this option makes the GCC driver pass Android-specific options to the linker.
11535 Finally, this option causes the preprocessor macro @code{__ANDROID__}
11538 @item -tno-android-cc
11539 @opindex tno-android-cc
11540 Disable compilation effects of @option{-mandroid}, i.e., do not enable
11541 @option{-mbionic}, @option{-fPIC}, @option{-fno-exceptions} and
11542 @option{-fno-rtti} by default.
11544 @item -tno-android-ld
11545 @opindex tno-android-ld
11546 Disable linking effects of @option{-mandroid}, i.e., pass standard Linux
11547 linking options to the linker.
11551 @node H8/300 Options
11552 @subsection H8/300 Options
11554 These @samp{-m} options are defined for the H8/300 implementations:
11559 Shorten some address references at link time, when possible; uses the
11560 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
11561 ld, Using ld}, for a fuller description.
11565 Generate code for the H8/300H@.
11569 Generate code for the H8S@.
11573 Generate code for the H8S and H8/300H in the normal mode. This switch
11574 must be used either with @option{-mh} or @option{-ms}.
11578 Generate code for the H8S/2600. This switch must be used with @option{-ms}.
11582 Make @code{int} data 32 bits by default.
11585 @opindex malign-300
11586 On the H8/300H and H8S, use the same alignment rules as for the H8/300.
11587 The default for the H8/300H and H8S is to align longs and floats on 4
11589 @option{-malign-300} causes them to be aligned on 2 byte boundaries.
11590 This option has no effect on the H8/300.
11594 @subsection HPPA Options
11595 @cindex HPPA Options
11597 These @samp{-m} options are defined for the HPPA family of computers:
11600 @item -march=@var{architecture-type}
11602 Generate code for the specified architecture. The choices for
11603 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
11604 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
11605 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
11606 architecture option for your machine. Code compiled for lower numbered
11607 architectures will run on higher numbered architectures, but not the
11610 @item -mpa-risc-1-0
11611 @itemx -mpa-risc-1-1
11612 @itemx -mpa-risc-2-0
11613 @opindex mpa-risc-1-0
11614 @opindex mpa-risc-1-1
11615 @opindex mpa-risc-2-0
11616 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
11619 @opindex mbig-switch
11620 Generate code suitable for big switch tables. Use this option only if
11621 the assembler/linker complain about out of range branches within a switch
11624 @item -mjump-in-delay
11625 @opindex mjump-in-delay
11626 Fill delay slots of function calls with unconditional jump instructions
11627 by modifying the return pointer for the function call to be the target
11628 of the conditional jump.
11630 @item -mdisable-fpregs
11631 @opindex mdisable-fpregs
11632 Prevent floating point registers from being used in any manner. This is
11633 necessary for compiling kernels which perform lazy context switching of
11634 floating point registers. If you use this option and attempt to perform
11635 floating point operations, the compiler will abort.
11637 @item -mdisable-indexing
11638 @opindex mdisable-indexing
11639 Prevent the compiler from using indexing address modes. This avoids some
11640 rather obscure problems when compiling MIG generated code under MACH@.
11642 @item -mno-space-regs
11643 @opindex mno-space-regs
11644 Generate code that assumes the target has no space registers. This allows
11645 GCC to generate faster indirect calls and use unscaled index address modes.
11647 Such code is suitable for level 0 PA systems and kernels.
11649 @item -mfast-indirect-calls
11650 @opindex mfast-indirect-calls
11651 Generate code that assumes calls never cross space boundaries. This
11652 allows GCC to emit code which performs faster indirect calls.
11654 This option will not work in the presence of shared libraries or nested
11657 @item -mfixed-range=@var{register-range}
11658 @opindex mfixed-range
11659 Generate code treating the given register range as fixed registers.
11660 A fixed register is one that the register allocator can not use. This is
11661 useful when compiling kernel code. A register range is specified as
11662 two registers separated by a dash. Multiple register ranges can be
11663 specified separated by a comma.
11665 @item -mlong-load-store
11666 @opindex mlong-load-store
11667 Generate 3-instruction load and store sequences as sometimes required by
11668 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
11671 @item -mportable-runtime
11672 @opindex mportable-runtime
11673 Use the portable calling conventions proposed by HP for ELF systems.
11677 Enable the use of assembler directives only GAS understands.
11679 @item -mschedule=@var{cpu-type}
11681 Schedule code according to the constraints for the machine type
11682 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
11683 @samp{7100}, @samp{7100LC}, @samp{7200}, @samp{7300} and @samp{8000}. Refer
11684 to @file{/usr/lib/sched.models} on an HP-UX system to determine the
11685 proper scheduling option for your machine. The default scheduling is
11689 @opindex mlinker-opt
11690 Enable the optimization pass in the HP-UX linker. Note this makes symbolic
11691 debugging impossible. It also triggers a bug in the HP-UX 8 and HP-UX 9
11692 linkers in which they give bogus error messages when linking some programs.
11695 @opindex msoft-float
11696 Generate output containing library calls for floating point.
11697 @strong{Warning:} the requisite libraries are not available for all HPPA
11698 targets. Normally the facilities of the machine's usual C compiler are
11699 used, but this cannot be done directly in cross-compilation. You must make
11700 your own arrangements to provide suitable library functions for
11703 @option{-msoft-float} changes the calling convention in the output file;
11704 therefore, it is only useful if you compile @emph{all} of a program with
11705 this option. In particular, you need to compile @file{libgcc.a}, the
11706 library that comes with GCC, with @option{-msoft-float} in order for
11711 Generate the predefine, @code{_SIO}, for server IO@. The default is
11712 @option{-mwsio}. This generates the predefines, @code{__hp9000s700},
11713 @code{__hp9000s700__} and @code{_WSIO}, for workstation IO@. These
11714 options are available under HP-UX and HI-UX@.
11718 Use GNU ld specific options. This passes @option{-shared} to ld when
11719 building a shared library. It is the default when GCC is configured,
11720 explicitly or implicitly, with the GNU linker. This option does not
11721 have any affect on which ld is called, it only changes what parameters
11722 are passed to that ld. The ld that is called is determined by the
11723 @option{--with-ld} configure option, GCC's program search path, and
11724 finally by the user's @env{PATH}. The linker used by GCC can be printed
11725 using @samp{which `gcc -print-prog-name=ld`}. This option is only available
11726 on the 64 bit HP-UX GCC, i.e.@: configured with @samp{hppa*64*-*-hpux*}.
11730 Use HP ld specific options. This passes @option{-b} to ld when building
11731 a shared library and passes @option{+Accept TypeMismatch} to ld on all
11732 links. It is the default when GCC is configured, explicitly or
11733 implicitly, with the HP linker. This option does not have any affect on
11734 which ld is called, it only changes what parameters are passed to that
11735 ld. The ld that is called is determined by the @option{--with-ld}
11736 configure option, GCC's program search path, and finally by the user's
11737 @env{PATH}. The linker used by GCC can be printed using @samp{which
11738 `gcc -print-prog-name=ld`}. This option is only available on the 64 bit
11739 HP-UX GCC, i.e.@: configured with @samp{hppa*64*-*-hpux*}.
11742 @opindex mno-long-calls
11743 Generate code that uses long call sequences. This ensures that a call
11744 is always able to reach linker generated stubs. The default is to generate
11745 long calls only when the distance from the call site to the beginning
11746 of the function or translation unit, as the case may be, exceeds a
11747 predefined limit set by the branch type being used. The limits for
11748 normal calls are 7,600,000 and 240,000 bytes, respectively for the
11749 PA 2.0 and PA 1.X architectures. Sibcalls are always limited at
11752 Distances are measured from the beginning of functions when using the
11753 @option{-ffunction-sections} option, or when using the @option{-mgas}
11754 and @option{-mno-portable-runtime} options together under HP-UX with
11757 It is normally not desirable to use this option as it will degrade
11758 performance. However, it may be useful in large applications,
11759 particularly when partial linking is used to build the application.
11761 The types of long calls used depends on the capabilities of the
11762 assembler and linker, and the type of code being generated. The
11763 impact on systems that support long absolute calls, and long pic
11764 symbol-difference or pc-relative calls should be relatively small.
11765 However, an indirect call is used on 32-bit ELF systems in pic code
11766 and it is quite long.
11768 @item -munix=@var{unix-std}
11770 Generate compiler predefines and select a startfile for the specified
11771 UNIX standard. The choices for @var{unix-std} are @samp{93}, @samp{95}
11772 and @samp{98}. @samp{93} is supported on all HP-UX versions. @samp{95}
11773 is available on HP-UX 10.10 and later. @samp{98} is available on HP-UX
11774 11.11 and later. The default values are @samp{93} for HP-UX 10.00,
11775 @samp{95} for HP-UX 10.10 though to 11.00, and @samp{98} for HP-UX 11.11
11778 @option{-munix=93} provides the same predefines as GCC 3.3 and 3.4.
11779 @option{-munix=95} provides additional predefines for @code{XOPEN_UNIX}
11780 and @code{_XOPEN_SOURCE_EXTENDED}, and the startfile @file{unix95.o}.
11781 @option{-munix=98} provides additional predefines for @code{_XOPEN_UNIX},
11782 @code{_XOPEN_SOURCE_EXTENDED}, @code{_INCLUDE__STDC_A1_SOURCE} and
11783 @code{_INCLUDE_XOPEN_SOURCE_500}, and the startfile @file{unix98.o}.
11785 It is @emph{important} to note that this option changes the interfaces
11786 for various library routines. It also affects the operational behavior
11787 of the C library. Thus, @emph{extreme} care is needed in using this
11790 Library code that is intended to operate with more than one UNIX
11791 standard must test, set and restore the variable @var{__xpg4_extended_mask}
11792 as appropriate. Most GNU software doesn't provide this capability.
11796 Suppress the generation of link options to search libdld.sl when the
11797 @option{-static} option is specified on HP-UX 10 and later.
11801 The HP-UX implementation of setlocale in libc has a dependency on
11802 libdld.sl. There isn't an archive version of libdld.sl. Thus,
11803 when the @option{-static} option is specified, special link options
11804 are needed to resolve this dependency.
11806 On HP-UX 10 and later, the GCC driver adds the necessary options to
11807 link with libdld.sl when the @option{-static} option is specified.
11808 This causes the resulting binary to be dynamic. On the 64-bit port,
11809 the linkers generate dynamic binaries by default in any case. The
11810 @option{-nolibdld} option can be used to prevent the GCC driver from
11811 adding these link options.
11815 Add support for multithreading with the @dfn{dce thread} library
11816 under HP-UX@. This option sets flags for both the preprocessor and
11820 @node i386 and x86-64 Options
11821 @subsection Intel 386 and AMD x86-64 Options
11822 @cindex i386 Options
11823 @cindex x86-64 Options
11824 @cindex Intel 386 Options
11825 @cindex AMD x86-64 Options
11827 These @samp{-m} options are defined for the i386 and x86-64 family of
11831 @item -mtune=@var{cpu-type}
11833 Tune to @var{cpu-type} everything applicable about the generated code, except
11834 for the ABI and the set of available instructions. The choices for
11835 @var{cpu-type} are:
11838 Produce code optimized for the most common IA32/AMD64/EM64T processors.
11839 If you know the CPU on which your code will run, then you should use
11840 the corresponding @option{-mtune} option instead of
11841 @option{-mtune=generic}. But, if you do not know exactly what CPU users
11842 of your application will have, then you should use this option.
11844 As new processors are deployed in the marketplace, the behavior of this
11845 option will change. Therefore, if you upgrade to a newer version of
11846 GCC, the code generated option will change to reflect the processors
11847 that were most common when that version of GCC was released.
11849 There is no @option{-march=generic} option because @option{-march}
11850 indicates the instruction set the compiler can use, and there is no
11851 generic instruction set applicable to all processors. In contrast,
11852 @option{-mtune} indicates the processor (or, in this case, collection of
11853 processors) for which the code is optimized.
11855 This selects the CPU to tune for at compilation time by determining
11856 the processor type of the compiling machine. Using @option{-mtune=native}
11857 will produce code optimized for the local machine under the constraints
11858 of the selected instruction set. Using @option{-march=native} will
11859 enable all instruction subsets supported by the local machine (hence
11860 the result might not run on different machines).
11862 Original Intel's i386 CPU@.
11864 Intel's i486 CPU@. (No scheduling is implemented for this chip.)
11865 @item i586, pentium
11866 Intel Pentium CPU with no MMX support.
11868 Intel PentiumMMX CPU based on Pentium core with MMX instruction set support.
11870 Intel PentiumPro CPU@.
11872 Same as @code{generic}, but when used as @code{march} option, PentiumPro
11873 instruction set will be used, so the code will run on all i686 family chips.
11875 Intel Pentium2 CPU based on PentiumPro core with MMX instruction set support.
11876 @item pentium3, pentium3m
11877 Intel Pentium3 CPU based on PentiumPro core with MMX and SSE instruction set
11880 Low power version of Intel Pentium3 CPU with MMX, SSE and SSE2 instruction set
11881 support. Used by Centrino notebooks.
11882 @item pentium4, pentium4m
11883 Intel Pentium4 CPU with MMX, SSE and SSE2 instruction set support.
11885 Improved version of Intel Pentium4 CPU with MMX, SSE, SSE2 and SSE3 instruction
11888 Improved version of Intel Pentium4 CPU with 64-bit extensions, MMX, SSE,
11889 SSE2 and SSE3 instruction set support.
11891 Intel Core2 CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3 and SSSE3
11892 instruction set support.
11894 Intel Atom CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3 and SSSE3
11895 instruction set support.
11897 AMD K6 CPU with MMX instruction set support.
11899 Improved versions of AMD K6 CPU with MMX and 3DNow!@: instruction set support.
11900 @item athlon, athlon-tbird
11901 AMD Athlon CPU with MMX, 3dNOW!, enhanced 3DNow!@: and SSE prefetch instructions
11903 @item athlon-4, athlon-xp, athlon-mp
11904 Improved AMD Athlon CPU with MMX, 3DNow!, enhanced 3DNow!@: and full SSE
11905 instruction set support.
11906 @item k8, opteron, athlon64, athlon-fx
11907 AMD K8 core based CPUs with x86-64 instruction set support. (This supersets
11908 MMX, SSE, SSE2, 3DNow!, enhanced 3DNow!@: and 64-bit instruction set extensions.)
11909 @item k8-sse3, opteron-sse3, athlon64-sse3
11910 Improved versions of k8, opteron and athlon64 with SSE3 instruction set support.
11911 @item amdfam10, barcelona
11912 AMD Family 10h core based CPUs with x86-64 instruction set support. (This
11913 supersets MMX, SSE, SSE2, SSE3, SSE4A, 3DNow!, enhanced 3DNow!, ABM and 64-bit
11914 instruction set extensions.)
11916 IDT Winchip C6 CPU, dealt in same way as i486 with additional MMX instruction
11919 IDT Winchip2 CPU, dealt in same way as i486 with additional MMX and 3DNow!@:
11920 instruction set support.
11922 Via C3 CPU with MMX and 3DNow!@: instruction set support. (No scheduling is
11923 implemented for this chip.)
11925 Via C3-2 CPU with MMX and SSE instruction set support. (No scheduling is
11926 implemented for this chip.)
11928 Embedded AMD CPU with MMX and 3DNow!@: instruction set support.
11931 While picking a specific @var{cpu-type} will schedule things appropriately
11932 for that particular chip, the compiler will not generate any code that
11933 does not run on the i386 without the @option{-march=@var{cpu-type}} option
11936 @item -march=@var{cpu-type}
11938 Generate instructions for the machine type @var{cpu-type}. The choices
11939 for @var{cpu-type} are the same as for @option{-mtune}. Moreover,
11940 specifying @option{-march=@var{cpu-type}} implies @option{-mtune=@var{cpu-type}}.
11942 @item -mcpu=@var{cpu-type}
11944 A deprecated synonym for @option{-mtune}.
11946 @item -mfpmath=@var{unit}
11948 Generate floating point arithmetics for selected unit @var{unit}. The choices
11949 for @var{unit} are:
11953 Use the standard 387 floating point coprocessor present majority of chips and
11954 emulated otherwise. Code compiled with this option will run almost everywhere.
11955 The temporary results are computed in 80bit precision instead of precision
11956 specified by the type resulting in slightly different results compared to most
11957 of other chips. See @option{-ffloat-store} for more detailed description.
11959 This is the default choice for i386 compiler.
11962 Use scalar floating point instructions present in the SSE instruction set.
11963 This instruction set is supported by Pentium3 and newer chips, in the AMD line
11964 by Athlon-4, Athlon-xp and Athlon-mp chips. The earlier version of SSE
11965 instruction set supports only single precision arithmetics, thus the double and
11966 extended precision arithmetics is still done using 387. Later version, present
11967 only in Pentium4 and the future AMD x86-64 chips supports double precision
11970 For the i386 compiler, you need to use @option{-march=@var{cpu-type}}, @option{-msse}
11971 or @option{-msse2} switches to enable SSE extensions and make this option
11972 effective. For the x86-64 compiler, these extensions are enabled by default.
11974 The resulting code should be considerably faster in the majority of cases and avoid
11975 the numerical instability problems of 387 code, but may break some existing
11976 code that expects temporaries to be 80bit.
11978 This is the default choice for the x86-64 compiler.
11983 Attempt to utilize both instruction sets at once. This effectively double the
11984 amount of available registers and on chips with separate execution units for
11985 387 and SSE the execution resources too. Use this option with care, as it is
11986 still experimental, because the GCC register allocator does not model separate
11987 functional units well resulting in instable performance.
11990 @item -masm=@var{dialect}
11991 @opindex masm=@var{dialect}
11992 Output asm instructions using selected @var{dialect}. Supported
11993 choices are @samp{intel} or @samp{att} (the default one). Darwin does
11994 not support @samp{intel}.
11997 @itemx -mno-ieee-fp
11999 @opindex mno-ieee-fp
12000 Control whether or not the compiler uses IEEE floating point
12001 comparisons. These handle correctly the case where the result of a
12002 comparison is unordered.
12005 @opindex msoft-float
12006 Generate output containing library calls for floating point.
12007 @strong{Warning:} the requisite libraries are not part of GCC@.
12008 Normally the facilities of the machine's usual C compiler are used, but
12009 this can't be done directly in cross-compilation. You must make your
12010 own arrangements to provide suitable library functions for
12013 On machines where a function returns floating point results in the 80387
12014 register stack, some floating point opcodes may be emitted even if
12015 @option{-msoft-float} is used.
12017 @item -mno-fp-ret-in-387
12018 @opindex mno-fp-ret-in-387
12019 Do not use the FPU registers for return values of functions.
12021 The usual calling convention has functions return values of types
12022 @code{float} and @code{double} in an FPU register, even if there
12023 is no FPU@. The idea is that the operating system should emulate
12026 The option @option{-mno-fp-ret-in-387} causes such values to be returned
12027 in ordinary CPU registers instead.
12029 @item -mno-fancy-math-387
12030 @opindex mno-fancy-math-387
12031 Some 387 emulators do not support the @code{sin}, @code{cos} and
12032 @code{sqrt} instructions for the 387. Specify this option to avoid
12033 generating those instructions. This option is the default on FreeBSD,
12034 OpenBSD and NetBSD@. This option is overridden when @option{-march}
12035 indicates that the target cpu will always have an FPU and so the
12036 instruction will not need emulation. As of revision 2.6.1, these
12037 instructions are not generated unless you also use the
12038 @option{-funsafe-math-optimizations} switch.
12040 @item -malign-double
12041 @itemx -mno-align-double
12042 @opindex malign-double
12043 @opindex mno-align-double
12044 Control whether GCC aligns @code{double}, @code{long double}, and
12045 @code{long long} variables on a two word boundary or a one word
12046 boundary. Aligning @code{double} variables on a two word boundary will
12047 produce code that runs somewhat faster on a @samp{Pentium} at the
12048 expense of more memory.
12050 On x86-64, @option{-malign-double} is enabled by default.
12052 @strong{Warning:} if you use the @option{-malign-double} switch,
12053 structures containing the above types will be aligned differently than
12054 the published application binary interface specifications for the 386
12055 and will not be binary compatible with structures in code compiled
12056 without that switch.
12058 @item -m96bit-long-double
12059 @itemx -m128bit-long-double
12060 @opindex m96bit-long-double
12061 @opindex m128bit-long-double
12062 These switches control the size of @code{long double} type. The i386
12063 application binary interface specifies the size to be 96 bits,
12064 so @option{-m96bit-long-double} is the default in 32 bit mode.
12066 Modern architectures (Pentium and newer) would prefer @code{long double}
12067 to be aligned to an 8 or 16 byte boundary. In arrays or structures
12068 conforming to the ABI, this would not be possible. So specifying a
12069 @option{-m128bit-long-double} will align @code{long double}
12070 to a 16 byte boundary by padding the @code{long double} with an additional
12073 In the x86-64 compiler, @option{-m128bit-long-double} is the default choice as
12074 its ABI specifies that @code{long double} is to be aligned on 16 byte boundary.
12076 Notice that neither of these options enable any extra precision over the x87
12077 standard of 80 bits for a @code{long double}.
12079 @strong{Warning:} if you override the default value for your target ABI, the
12080 structures and arrays containing @code{long double} variables will change
12081 their size as well as function calling convention for function taking
12082 @code{long double} will be modified. Hence they will not be binary
12083 compatible with arrays or structures in code compiled without that switch.
12085 @item -mlarge-data-threshold=@var{number}
12086 @opindex mlarge-data-threshold=@var{number}
12087 When @option{-mcmodel=medium} is specified, the data greater than
12088 @var{threshold} are placed in large data section. This value must be the
12089 same across all object linked into the binary and defaults to 65535.
12093 Use a different function-calling convention, in which functions that
12094 take a fixed number of arguments return with the @code{ret} @var{num}
12095 instruction, which pops their arguments while returning. This saves one
12096 instruction in the caller since there is no need to pop the arguments
12099 You can specify that an individual function is called with this calling
12100 sequence with the function attribute @samp{stdcall}. You can also
12101 override the @option{-mrtd} option by using the function attribute
12102 @samp{cdecl}. @xref{Function Attributes}.
12104 @strong{Warning:} this calling convention is incompatible with the one
12105 normally used on Unix, so you cannot use it if you need to call
12106 libraries compiled with the Unix compiler.
12108 Also, you must provide function prototypes for all functions that
12109 take variable numbers of arguments (including @code{printf});
12110 otherwise incorrect code will be generated for calls to those
12113 In addition, seriously incorrect code will result if you call a
12114 function with too many arguments. (Normally, extra arguments are
12115 harmlessly ignored.)
12117 @item -mregparm=@var{num}
12119 Control how many registers are used to pass integer arguments. By
12120 default, no registers are used to pass arguments, and at most 3
12121 registers can be used. You can control this behavior for a specific
12122 function by using the function attribute @samp{regparm}.
12123 @xref{Function Attributes}.
12125 @strong{Warning:} if you use this switch, and
12126 @var{num} is nonzero, then you must build all modules with the same
12127 value, including any libraries. This includes the system libraries and
12131 @opindex msseregparm
12132 Use SSE register passing conventions for float and double arguments
12133 and return values. You can control this behavior for a specific
12134 function by using the function attribute @samp{sseregparm}.
12135 @xref{Function Attributes}.
12137 @strong{Warning:} if you use this switch then you must build all
12138 modules with the same value, including any libraries. This includes
12139 the system libraries and startup modules.
12148 Set 80387 floating-point precision to 32, 64 or 80 bits. When @option{-mpc32}
12149 is specified, the significands of results of floating-point operations are
12150 rounded to 24 bits (single precision); @option{-mpc64} rounds the
12151 significands of results of floating-point operations to 53 bits (double
12152 precision) and @option{-mpc80} rounds the significands of results of
12153 floating-point operations to 64 bits (extended double precision), which is
12154 the default. When this option is used, floating-point operations in higher
12155 precisions are not available to the programmer without setting the FPU
12156 control word explicitly.
12158 Setting the rounding of floating-point operations to less than the default
12159 80 bits can speed some programs by 2% or more. Note that some mathematical
12160 libraries assume that extended precision (80 bit) floating-point operations
12161 are enabled by default; routines in such libraries could suffer significant
12162 loss of accuracy, typically through so-called "catastrophic cancellation",
12163 when this option is used to set the precision to less than extended precision.
12165 @item -mstackrealign
12166 @opindex mstackrealign
12167 Realign the stack at entry. On the Intel x86, the @option{-mstackrealign}
12168 option will generate an alternate prologue and epilogue that realigns the
12169 runtime stack if necessary. This supports mixing legacy codes that keep
12170 a 4-byte aligned stack with modern codes that keep a 16-byte stack for
12171 SSE compatibility. See also the attribute @code{force_align_arg_pointer},
12172 applicable to individual functions.
12174 @item -mpreferred-stack-boundary=@var{num}
12175 @opindex mpreferred-stack-boundary
12176 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
12177 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
12178 the default is 4 (16 bytes or 128 bits).
12180 @item -mincoming-stack-boundary=@var{num}
12181 @opindex mincoming-stack-boundary
12182 Assume the incoming stack is aligned to a 2 raised to @var{num} byte
12183 boundary. If @option{-mincoming-stack-boundary} is not specified,
12184 the one specified by @option{-mpreferred-stack-boundary} will be used.
12186 On Pentium and PentiumPro, @code{double} and @code{long double} values
12187 should be aligned to an 8 byte boundary (see @option{-malign-double}) or
12188 suffer significant run time performance penalties. On Pentium III, the
12189 Streaming SIMD Extension (SSE) data type @code{__m128} may not work
12190 properly if it is not 16 byte aligned.
12192 To ensure proper alignment of this values on the stack, the stack boundary
12193 must be as aligned as that required by any value stored on the stack.
12194 Further, every function must be generated such that it keeps the stack
12195 aligned. Thus calling a function compiled with a higher preferred
12196 stack boundary from a function compiled with a lower preferred stack
12197 boundary will most likely misalign the stack. It is recommended that
12198 libraries that use callbacks always use the default setting.
12200 This extra alignment does consume extra stack space, and generally
12201 increases code size. Code that is sensitive to stack space usage, such
12202 as embedded systems and operating system kernels, may want to reduce the
12203 preferred alignment to @option{-mpreferred-stack-boundary=2}.
12247 These switches enable or disable the use of instructions in the MMX,
12248 SSE, SSE2, SSE3, SSSE3, SSE4.1, AVX, AES, PCLMUL, SSE4A, FMA4, XOP,
12249 LWP, ABM or 3DNow!@: extended instruction sets.
12250 These extensions are also available as built-in functions: see
12251 @ref{X86 Built-in Functions}, for details of the functions enabled and
12252 disabled by these switches.
12254 To have SSE/SSE2 instructions generated automatically from floating-point
12255 code (as opposed to 387 instructions), see @option{-mfpmath=sse}.
12257 GCC depresses SSEx instructions when @option{-mavx} is used. Instead, it
12258 generates new AVX instructions or AVX equivalence for all SSEx instructions
12261 These options will enable GCC to use these extended instructions in
12262 generated code, even without @option{-mfpmath=sse}. Applications which
12263 perform runtime CPU detection must compile separate files for each
12264 supported architecture, using the appropriate flags. In particular,
12265 the file containing the CPU detection code should be compiled without
12269 @itemx -mno-fused-madd
12270 @opindex mfused-madd
12271 @opindex mno-fused-madd
12272 Do (don't) generate code that uses the fused multiply/add or multiply/subtract
12273 instructions. The default is to use these instructions.
12277 This option instructs GCC to emit a @code{cld} instruction in the prologue
12278 of functions that use string instructions. String instructions depend on
12279 the DF flag to select between autoincrement or autodecrement mode. While the
12280 ABI specifies the DF flag to be cleared on function entry, some operating
12281 systems violate this specification by not clearing the DF flag in their
12282 exception dispatchers. The exception handler can be invoked with the DF flag
12283 set which leads to wrong direction mode, when string instructions are used.
12284 This option can be enabled by default on 32-bit x86 targets by configuring
12285 GCC with the @option{--enable-cld} configure option. Generation of @code{cld}
12286 instructions can be suppressed with the @option{-mno-cld} compiler option
12291 This option will enable GCC to use CMPXCHG16B instruction in generated code.
12292 CMPXCHG16B allows for atomic operations on 128-bit double quadword (or oword)
12293 data types. This is useful for high resolution counters that could be updated
12294 by multiple processors (or cores). This instruction is generated as part of
12295 atomic built-in functions: see @ref{Atomic Builtins} for details.
12299 This option will enable GCC to use SAHF instruction in generated 64-bit code.
12300 Early Intel CPUs with Intel 64 lacked LAHF and SAHF instructions supported
12301 by AMD64 until introduction of Pentium 4 G1 step in December 2005. LAHF and
12302 SAHF are load and store instructions, respectively, for certain status flags.
12303 In 64-bit mode, SAHF instruction is used to optimize @code{fmod}, @code{drem}
12304 or @code{remainder} built-in functions: see @ref{Other Builtins} for details.
12308 This option will enable GCC to use movbe instruction to implement
12309 @code{__builtin_bswap32} and @code{__builtin_bswap64}.
12313 This option will enable built-in functions, @code{__builtin_ia32_crc32qi},
12314 @code{__builtin_ia32_crc32hi}. @code{__builtin_ia32_crc32si} and
12315 @code{__builtin_ia32_crc32di} to generate the crc32 machine instruction.
12319 This option will enable GCC to use RCPSS and RSQRTSS instructions (and their
12320 vectorized variants RCPPS and RSQRTPS) with an additional Newton-Raphson step
12321 to increase precision instead of DIVSS and SQRTSS (and their vectorized
12322 variants) for single precision floating point arguments. These instructions
12323 are generated only when @option{-funsafe-math-optimizations} is enabled
12324 together with @option{-finite-math-only} and @option{-fno-trapping-math}.
12325 Note that while the throughput of the sequence is higher than the throughput
12326 of the non-reciprocal instruction, the precision of the sequence can be
12327 decreased by up to 2 ulp (i.e. the inverse of 1.0 equals 0.99999994).
12329 Note that GCC implements 1.0f/sqrtf(x) in terms of RSQRTSS (or RSQRTPS)
12330 already with @option{-ffast-math} (or the above option combination), and
12331 doesn't need @option{-mrecip}.
12333 @item -mveclibabi=@var{type}
12334 @opindex mveclibabi
12335 Specifies the ABI type to use for vectorizing intrinsics using an
12336 external library. Supported types are @code{svml} for the Intel short
12337 vector math library and @code{acml} for the AMD math core library style
12338 of interfacing. GCC will currently emit calls to @code{vmldExp2},
12339 @code{vmldLn2}, @code{vmldLog102}, @code{vmldLog102}, @code{vmldPow2},
12340 @code{vmldTanh2}, @code{vmldTan2}, @code{vmldAtan2}, @code{vmldAtanh2},
12341 @code{vmldCbrt2}, @code{vmldSinh2}, @code{vmldSin2}, @code{vmldAsinh2},
12342 @code{vmldAsin2}, @code{vmldCosh2}, @code{vmldCos2}, @code{vmldAcosh2},
12343 @code{vmldAcos2}, @code{vmlsExp4}, @code{vmlsLn4}, @code{vmlsLog104},
12344 @code{vmlsLog104}, @code{vmlsPow4}, @code{vmlsTanh4}, @code{vmlsTan4},
12345 @code{vmlsAtan4}, @code{vmlsAtanh4}, @code{vmlsCbrt4}, @code{vmlsSinh4},
12346 @code{vmlsSin4}, @code{vmlsAsinh4}, @code{vmlsAsin4}, @code{vmlsCosh4},
12347 @code{vmlsCos4}, @code{vmlsAcosh4} and @code{vmlsAcos4} for corresponding
12348 function type when @option{-mveclibabi=svml} is used and @code{__vrd2_sin},
12349 @code{__vrd2_cos}, @code{__vrd2_exp}, @code{__vrd2_log}, @code{__vrd2_log2},
12350 @code{__vrd2_log10}, @code{__vrs4_sinf}, @code{__vrs4_cosf},
12351 @code{__vrs4_expf}, @code{__vrs4_logf}, @code{__vrs4_log2f},
12352 @code{__vrs4_log10f} and @code{__vrs4_powf} for corresponding function type
12353 when @option{-mveclibabi=acml} is used. Both @option{-ftree-vectorize} and
12354 @option{-funsafe-math-optimizations} have to be enabled. A SVML or ACML ABI
12355 compatible library will have to be specified at link time.
12357 @item -mabi=@var{name}
12359 Generate code for the specified calling convention. Permissible values
12360 are: @samp{sysv} for the ABI used on GNU/Linux and other systems and
12361 @samp{ms} for the Microsoft ABI. The default is to use the Microsoft
12362 ABI when targeting Windows. On all other systems, the default is the
12363 SYSV ABI. You can control this behavior for a specific function by
12364 using the function attribute @samp{ms_abi}/@samp{sysv_abi}.
12365 @xref{Function Attributes}.
12368 @itemx -mno-push-args
12369 @opindex mpush-args
12370 @opindex mno-push-args
12371 Use PUSH operations to store outgoing parameters. This method is shorter
12372 and usually equally fast as method using SUB/MOV operations and is enabled
12373 by default. In some cases disabling it may improve performance because of
12374 improved scheduling and reduced dependencies.
12376 @item -maccumulate-outgoing-args
12377 @opindex maccumulate-outgoing-args
12378 If enabled, the maximum amount of space required for outgoing arguments will be
12379 computed in the function prologue. This is faster on most modern CPUs
12380 because of reduced dependencies, improved scheduling and reduced stack usage
12381 when preferred stack boundary is not equal to 2. The drawback is a notable
12382 increase in code size. This switch implies @option{-mno-push-args}.
12386 Support thread-safe exception handling on @samp{Mingw32}. Code that relies
12387 on thread-safe exception handling must compile and link all code with the
12388 @option{-mthreads} option. When compiling, @option{-mthreads} defines
12389 @option{-D_MT}; when linking, it links in a special thread helper library
12390 @option{-lmingwthrd} which cleans up per thread exception handling data.
12392 @item -mno-align-stringops
12393 @opindex mno-align-stringops
12394 Do not align destination of inlined string operations. This switch reduces
12395 code size and improves performance in case the destination is already aligned,
12396 but GCC doesn't know about it.
12398 @item -minline-all-stringops
12399 @opindex minline-all-stringops
12400 By default GCC inlines string operations only when destination is known to be
12401 aligned at least to 4 byte boundary. This enables more inlining, increase code
12402 size, but may improve performance of code that depends on fast memcpy, strlen
12403 and memset for short lengths.
12405 @item -minline-stringops-dynamically
12406 @opindex minline-stringops-dynamically
12407 For string operation of unknown size, inline runtime checks so for small
12408 blocks inline code is used, while for large blocks library call is used.
12410 @item -mstringop-strategy=@var{alg}
12411 @opindex mstringop-strategy=@var{alg}
12412 Overwrite internal decision heuristic about particular algorithm to inline
12413 string operation with. The allowed values are @code{rep_byte},
12414 @code{rep_4byte}, @code{rep_8byte} for expanding using i386 @code{rep} prefix
12415 of specified size, @code{byte_loop}, @code{loop}, @code{unrolled_loop} for
12416 expanding inline loop, @code{libcall} for always expanding library call.
12418 @item -momit-leaf-frame-pointer
12419 @opindex momit-leaf-frame-pointer
12420 Don't keep the frame pointer in a register for leaf functions. This
12421 avoids the instructions to save, set up and restore frame pointers and
12422 makes an extra register available in leaf functions. The option
12423 @option{-fomit-frame-pointer} removes the frame pointer for all functions
12424 which might make debugging harder.
12426 @item -mtls-direct-seg-refs
12427 @itemx -mno-tls-direct-seg-refs
12428 @opindex mtls-direct-seg-refs
12429 Controls whether TLS variables may be accessed with offsets from the
12430 TLS segment register (@code{%gs} for 32-bit, @code{%fs} for 64-bit),
12431 or whether the thread base pointer must be added. Whether or not this
12432 is legal depends on the operating system, and whether it maps the
12433 segment to cover the entire TLS area.
12435 For systems that use GNU libc, the default is on.
12438 @itemx -mno-sse2avx
12440 Specify that the assembler should encode SSE instructions with VEX
12441 prefix. The option @option{-mavx} turns this on by default.
12444 These @samp{-m} switches are supported in addition to the above
12445 on AMD x86-64 processors in 64-bit environments.
12452 Generate code for a 32-bit or 64-bit environment.
12453 The 32-bit environment sets int, long and pointer to 32 bits and
12454 generates code that runs on any i386 system.
12455 The 64-bit environment sets int to 32 bits and long and pointer
12456 to 64 bits and generates code for AMD's x86-64 architecture. For
12457 darwin only the -m64 option turns off the @option{-fno-pic} and
12458 @option{-mdynamic-no-pic} options.
12460 @item -mno-red-zone
12461 @opindex mno-red-zone
12462 Do not use a so called red zone for x86-64 code. The red zone is mandated
12463 by the x86-64 ABI, it is a 128-byte area beyond the location of the
12464 stack pointer that will not be modified by signal or interrupt handlers
12465 and therefore can be used for temporary data without adjusting the stack
12466 pointer. The flag @option{-mno-red-zone} disables this red zone.
12468 @item -mcmodel=small
12469 @opindex mcmodel=small
12470 Generate code for the small code model: the program and its symbols must
12471 be linked in the lower 2 GB of the address space. Pointers are 64 bits.
12472 Programs can be statically or dynamically linked. This is the default
12475 @item -mcmodel=kernel
12476 @opindex mcmodel=kernel
12477 Generate code for the kernel code model. The kernel runs in the
12478 negative 2 GB of the address space.
12479 This model has to be used for Linux kernel code.
12481 @item -mcmodel=medium
12482 @opindex mcmodel=medium
12483 Generate code for the medium model: The program is linked in the lower 2
12484 GB of the address space. Small symbols are also placed there. Symbols
12485 with sizes larger than @option{-mlarge-data-threshold} are put into
12486 large data or bss sections and can be located above 2GB. Programs can
12487 be statically or dynamically linked.
12489 @item -mcmodel=large
12490 @opindex mcmodel=large
12491 Generate code for the large model: This model makes no assumptions
12492 about addresses and sizes of sections.
12495 @node IA-64 Options
12496 @subsection IA-64 Options
12497 @cindex IA-64 Options
12499 These are the @samp{-m} options defined for the Intel IA-64 architecture.
12503 @opindex mbig-endian
12504 Generate code for a big endian target. This is the default for HP-UX@.
12506 @item -mlittle-endian
12507 @opindex mlittle-endian
12508 Generate code for a little endian target. This is the default for AIX5
12514 @opindex mno-gnu-as
12515 Generate (or don't) code for the GNU assembler. This is the default.
12516 @c Also, this is the default if the configure option @option{--with-gnu-as}
12522 @opindex mno-gnu-ld
12523 Generate (or don't) code for the GNU linker. This is the default.
12524 @c Also, this is the default if the configure option @option{--with-gnu-ld}
12529 Generate code that does not use a global pointer register. The result
12530 is not position independent code, and violates the IA-64 ABI@.
12532 @item -mvolatile-asm-stop
12533 @itemx -mno-volatile-asm-stop
12534 @opindex mvolatile-asm-stop
12535 @opindex mno-volatile-asm-stop
12536 Generate (or don't) a stop bit immediately before and after volatile asm
12539 @item -mregister-names
12540 @itemx -mno-register-names
12541 @opindex mregister-names
12542 @opindex mno-register-names
12543 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
12544 the stacked registers. This may make assembler output more readable.
12550 Disable (or enable) optimizations that use the small data section. This may
12551 be useful for working around optimizer bugs.
12553 @item -mconstant-gp
12554 @opindex mconstant-gp
12555 Generate code that uses a single constant global pointer value. This is
12556 useful when compiling kernel code.
12560 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
12561 This is useful when compiling firmware code.
12563 @item -minline-float-divide-min-latency
12564 @opindex minline-float-divide-min-latency
12565 Generate code for inline divides of floating point values
12566 using the minimum latency algorithm.
12568 @item -minline-float-divide-max-throughput
12569 @opindex minline-float-divide-max-throughput
12570 Generate code for inline divides of floating point values
12571 using the maximum throughput algorithm.
12573 @item -mno-inline-float-divide
12574 @opindex mno-inline-float-divide
12575 Do not generate inline code for divides of floating point values.
12577 @item -minline-int-divide-min-latency
12578 @opindex minline-int-divide-min-latency
12579 Generate code for inline divides of integer values
12580 using the minimum latency algorithm.
12582 @item -minline-int-divide-max-throughput
12583 @opindex minline-int-divide-max-throughput
12584 Generate code for inline divides of integer values
12585 using the maximum throughput algorithm.
12587 @item -mno-inline-int-divide
12588 @opindex mno-inline-int-divide
12589 Do not generate inline code for divides of integer values.
12591 @item -minline-sqrt-min-latency
12592 @opindex minline-sqrt-min-latency
12593 Generate code for inline square roots
12594 using the minimum latency algorithm.
12596 @item -minline-sqrt-max-throughput
12597 @opindex minline-sqrt-max-throughput
12598 Generate code for inline square roots
12599 using the maximum throughput algorithm.
12601 @item -mno-inline-sqrt
12602 @opindex mno-inline-sqrt
12603 Do not generate inline code for sqrt.
12606 @itemx -mno-fused-madd
12607 @opindex mfused-madd
12608 @opindex mno-fused-madd
12609 Do (don't) generate code that uses the fused multiply/add or multiply/subtract
12610 instructions. The default is to use these instructions.
12612 @item -mno-dwarf2-asm
12613 @itemx -mdwarf2-asm
12614 @opindex mno-dwarf2-asm
12615 @opindex mdwarf2-asm
12616 Don't (or do) generate assembler code for the DWARF2 line number debugging
12617 info. This may be useful when not using the GNU assembler.
12619 @item -mearly-stop-bits
12620 @itemx -mno-early-stop-bits
12621 @opindex mearly-stop-bits
12622 @opindex mno-early-stop-bits
12623 Allow stop bits to be placed earlier than immediately preceding the
12624 instruction that triggered the stop bit. This can improve instruction
12625 scheduling, but does not always do so.
12627 @item -mfixed-range=@var{register-range}
12628 @opindex mfixed-range
12629 Generate code treating the given register range as fixed registers.
12630 A fixed register is one that the register allocator can not use. This is
12631 useful when compiling kernel code. A register range is specified as
12632 two registers separated by a dash. Multiple register ranges can be
12633 specified separated by a comma.
12635 @item -mtls-size=@var{tls-size}
12637 Specify bit size of immediate TLS offsets. Valid values are 14, 22, and
12640 @item -mtune=@var{cpu-type}
12642 Tune the instruction scheduling for a particular CPU, Valid values are
12643 itanium, itanium1, merced, itanium2, and mckinley.
12649 Generate code for a 32-bit or 64-bit environment.
12650 The 32-bit environment sets int, long and pointer to 32 bits.
12651 The 64-bit environment sets int to 32 bits and long and pointer
12652 to 64 bits. These are HP-UX specific flags.
12654 @item -mno-sched-br-data-spec
12655 @itemx -msched-br-data-spec
12656 @opindex mno-sched-br-data-spec
12657 @opindex msched-br-data-spec
12658 (Dis/En)able data speculative scheduling before reload.
12659 This will result in generation of the ld.a instructions and
12660 the corresponding check instructions (ld.c / chk.a).
12661 The default is 'disable'.
12663 @item -msched-ar-data-spec
12664 @itemx -mno-sched-ar-data-spec
12665 @opindex msched-ar-data-spec
12666 @opindex mno-sched-ar-data-spec
12667 (En/Dis)able data speculative scheduling after reload.
12668 This will result in generation of the ld.a instructions and
12669 the corresponding check instructions (ld.c / chk.a).
12670 The default is 'enable'.
12672 @item -mno-sched-control-spec
12673 @itemx -msched-control-spec
12674 @opindex mno-sched-control-spec
12675 @opindex msched-control-spec
12676 (Dis/En)able control speculative scheduling. This feature is
12677 available only during region scheduling (i.e.@: before reload).
12678 This will result in generation of the ld.s instructions and
12679 the corresponding check instructions chk.s .
12680 The default is 'disable'.
12682 @item -msched-br-in-data-spec
12683 @itemx -mno-sched-br-in-data-spec
12684 @opindex msched-br-in-data-spec
12685 @opindex mno-sched-br-in-data-spec
12686 (En/Dis)able speculative scheduling of the instructions that
12687 are dependent on the data speculative loads before reload.
12688 This is effective only with @option{-msched-br-data-spec} enabled.
12689 The default is 'enable'.
12691 @item -msched-ar-in-data-spec
12692 @itemx -mno-sched-ar-in-data-spec
12693 @opindex msched-ar-in-data-spec
12694 @opindex mno-sched-ar-in-data-spec
12695 (En/Dis)able speculative scheduling of the instructions that
12696 are dependent on the data speculative loads after reload.
12697 This is effective only with @option{-msched-ar-data-spec} enabled.
12698 The default is 'enable'.
12700 @item -msched-in-control-spec
12701 @itemx -mno-sched-in-control-spec
12702 @opindex msched-in-control-spec
12703 @opindex mno-sched-in-control-spec
12704 (En/Dis)able speculative scheduling of the instructions that
12705 are dependent on the control speculative loads.
12706 This is effective only with @option{-msched-control-spec} enabled.
12707 The default is 'enable'.
12709 @item -mno-sched-prefer-non-data-spec-insns
12710 @itemx -msched-prefer-non-data-spec-insns
12711 @opindex mno-sched-prefer-non-data-spec-insns
12712 @opindex msched-prefer-non-data-spec-insns
12713 If enabled, data speculative instructions will be chosen for schedule
12714 only if there are no other choices at the moment. This will make
12715 the use of the data speculation much more conservative.
12716 The default is 'disable'.
12718 @item -mno-sched-prefer-non-control-spec-insns
12719 @itemx -msched-prefer-non-control-spec-insns
12720 @opindex mno-sched-prefer-non-control-spec-insns
12721 @opindex msched-prefer-non-control-spec-insns
12722 If enabled, control speculative instructions will be chosen for schedule
12723 only if there are no other choices at the moment. This will make
12724 the use of the control speculation much more conservative.
12725 The default is 'disable'.
12727 @item -mno-sched-count-spec-in-critical-path
12728 @itemx -msched-count-spec-in-critical-path
12729 @opindex mno-sched-count-spec-in-critical-path
12730 @opindex msched-count-spec-in-critical-path
12731 If enabled, speculative dependencies will be considered during
12732 computation of the instructions priorities. This will make the use of the
12733 speculation a bit more conservative.
12734 The default is 'disable'.
12736 @item -msched-spec-ldc
12737 @opindex msched-spec-ldc
12738 Use a simple data speculation check. This option is on by default.
12740 @item -msched-control-spec-ldc
12741 @opindex msched-spec-ldc
12742 Use a simple check for control speculation. This option is on by default.
12744 @item -msched-stop-bits-after-every-cycle
12745 @opindex msched-stop-bits-after-every-cycle
12746 Place a stop bit after every cycle when scheduling. This option is on
12749 @item -msched-fp-mem-deps-zero-cost
12750 @opindex msched-fp-mem-deps-zero-cost
12751 Assume that floating-point stores and loads are not likely to cause a conflict
12752 when placed into the same instruction group. This option is disabled by
12755 @item -msel-sched-dont-check-control-spec
12756 @opindex msel-sched-dont-check-control-spec
12757 Generate checks for control speculation in selective scheduling.
12758 This flag is disabled by default.
12760 @item -msched-max-memory-insns=@var{max-insns}
12761 @opindex msched-max-memory-insns
12762 Limit on the number of memory insns per instruction group, giving lower
12763 priority to subsequent memory insns attempting to schedule in the same
12764 instruction group. Frequently useful to prevent cache bank conflicts.
12765 The default value is 1.
12767 @item -msched-max-memory-insns-hard-limit
12768 @opindex msched-max-memory-insns-hard-limit
12769 Disallow more than `msched-max-memory-insns' in instruction group.
12770 Otherwise, limit is `soft' meaning that we would prefer non-memory operations
12771 when limit is reached but may still schedule memory operations.
12775 @node IA-64/VMS Options
12776 @subsection IA-64/VMS Options
12778 These @samp{-m} options are defined for the IA-64/VMS implementations:
12781 @item -mvms-return-codes
12782 @opindex mvms-return-codes
12783 Return VMS condition codes from main. The default is to return POSIX
12784 style condition (e.g.@ error) codes.
12786 @item -mdebug-main=@var{prefix}
12787 @opindex mdebug-main=@var{prefix}
12788 Flag the first routine whose name starts with @var{prefix} as the main
12789 routine for the debugger.
12793 Default to 64bit memory allocation routines.
12797 @subsection LM32 Options
12798 @cindex LM32 options
12800 These @option{-m} options are defined for the Lattice Mico32 architecture:
12803 @item -mbarrel-shift-enabled
12804 @opindex mbarrel-shift-enabled
12805 Enable barrel-shift instructions.
12807 @item -mdivide-enabled
12808 @opindex mdivide-enabled
12809 Enable divide and modulus instructions.
12811 @item -mmultiply-enabled
12812 @opindex multiply-enabled
12813 Enable multiply instructions.
12815 @item -msign-extend-enabled
12816 @opindex msign-extend-enabled
12817 Enable sign extend instructions.
12819 @item -muser-enabled
12820 @opindex muser-enabled
12821 Enable user-defined instructions.
12826 @subsection M32C Options
12827 @cindex M32C options
12830 @item -mcpu=@var{name}
12832 Select the CPU for which code is generated. @var{name} may be one of
12833 @samp{r8c} for the R8C/Tiny series, @samp{m16c} for the M16C (up to
12834 /60) series, @samp{m32cm} for the M16C/80 series, or @samp{m32c} for
12835 the M32C/80 series.
12839 Specifies that the program will be run on the simulator. This causes
12840 an alternate runtime library to be linked in which supports, for
12841 example, file I/O@. You must not use this option when generating
12842 programs that will run on real hardware; you must provide your own
12843 runtime library for whatever I/O functions are needed.
12845 @item -memregs=@var{number}
12847 Specifies the number of memory-based pseudo-registers GCC will use
12848 during code generation. These pseudo-registers will be used like real
12849 registers, so there is a tradeoff between GCC's ability to fit the
12850 code into available registers, and the performance penalty of using
12851 memory instead of registers. Note that all modules in a program must
12852 be compiled with the same value for this option. Because of that, you
12853 must not use this option with the default runtime libraries gcc
12858 @node M32R/D Options
12859 @subsection M32R/D Options
12860 @cindex M32R/D options
12862 These @option{-m} options are defined for Renesas M32R/D architectures:
12867 Generate code for the M32R/2@.
12871 Generate code for the M32R/X@.
12875 Generate code for the M32R@. This is the default.
12877 @item -mmodel=small
12878 @opindex mmodel=small
12879 Assume all objects live in the lower 16MB of memory (so that their addresses
12880 can be loaded with the @code{ld24} instruction), and assume all subroutines
12881 are reachable with the @code{bl} instruction.
12882 This is the default.
12884 The addressability of a particular object can be set with the
12885 @code{model} attribute.
12887 @item -mmodel=medium
12888 @opindex mmodel=medium
12889 Assume objects may be anywhere in the 32-bit address space (the compiler
12890 will generate @code{seth/add3} instructions to load their addresses), and
12891 assume all subroutines are reachable with the @code{bl} instruction.
12893 @item -mmodel=large
12894 @opindex mmodel=large
12895 Assume objects may be anywhere in the 32-bit address space (the compiler
12896 will generate @code{seth/add3} instructions to load their addresses), and
12897 assume subroutines may not be reachable with the @code{bl} instruction
12898 (the compiler will generate the much slower @code{seth/add3/jl}
12899 instruction sequence).
12902 @opindex msdata=none
12903 Disable use of the small data area. Variables will be put into
12904 one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the
12905 @code{section} attribute has been specified).
12906 This is the default.
12908 The small data area consists of sections @samp{.sdata} and @samp{.sbss}.
12909 Objects may be explicitly put in the small data area with the
12910 @code{section} attribute using one of these sections.
12912 @item -msdata=sdata
12913 @opindex msdata=sdata
12914 Put small global and static data in the small data area, but do not
12915 generate special code to reference them.
12918 @opindex msdata=use
12919 Put small global and static data in the small data area, and generate
12920 special instructions to reference them.
12924 @cindex smaller data references
12925 Put global and static objects less than or equal to @var{num} bytes
12926 into the small data or bss sections instead of the normal data or bss
12927 sections. The default value of @var{num} is 8.
12928 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
12929 for this option to have any effect.
12931 All modules should be compiled with the same @option{-G @var{num}} value.
12932 Compiling with different values of @var{num} may or may not work; if it
12933 doesn't the linker will give an error message---incorrect code will not be
12938 Makes the M32R specific code in the compiler display some statistics
12939 that might help in debugging programs.
12941 @item -malign-loops
12942 @opindex malign-loops
12943 Align all loops to a 32-byte boundary.
12945 @item -mno-align-loops
12946 @opindex mno-align-loops
12947 Do not enforce a 32-byte alignment for loops. This is the default.
12949 @item -missue-rate=@var{number}
12950 @opindex missue-rate=@var{number}
12951 Issue @var{number} instructions per cycle. @var{number} can only be 1
12954 @item -mbranch-cost=@var{number}
12955 @opindex mbranch-cost=@var{number}
12956 @var{number} can only be 1 or 2. If it is 1 then branches will be
12957 preferred over conditional code, if it is 2, then the opposite will
12960 @item -mflush-trap=@var{number}
12961 @opindex mflush-trap=@var{number}
12962 Specifies the trap number to use to flush the cache. The default is
12963 12. Valid numbers are between 0 and 15 inclusive.
12965 @item -mno-flush-trap
12966 @opindex mno-flush-trap
12967 Specifies that the cache cannot be flushed by using a trap.
12969 @item -mflush-func=@var{name}
12970 @opindex mflush-func=@var{name}
12971 Specifies the name of the operating system function to call to flush
12972 the cache. The default is @emph{_flush_cache}, but a function call
12973 will only be used if a trap is not available.
12975 @item -mno-flush-func
12976 @opindex mno-flush-func
12977 Indicates that there is no OS function for flushing the cache.
12981 @node M680x0 Options
12982 @subsection M680x0 Options
12983 @cindex M680x0 options
12985 These are the @samp{-m} options defined for M680x0 and ColdFire processors.
12986 The default settings depend on which architecture was selected when
12987 the compiler was configured; the defaults for the most common choices
12991 @item -march=@var{arch}
12993 Generate code for a specific M680x0 or ColdFire instruction set
12994 architecture. Permissible values of @var{arch} for M680x0
12995 architectures are: @samp{68000}, @samp{68010}, @samp{68020},
12996 @samp{68030}, @samp{68040}, @samp{68060} and @samp{cpu32}. ColdFire
12997 architectures are selected according to Freescale's ISA classification
12998 and the permissible values are: @samp{isaa}, @samp{isaaplus},
12999 @samp{isab} and @samp{isac}.
13001 gcc defines a macro @samp{__mcf@var{arch}__} whenever it is generating
13002 code for a ColdFire target. The @var{arch} in this macro is one of the
13003 @option{-march} arguments given above.
13005 When used together, @option{-march} and @option{-mtune} select code
13006 that runs on a family of similar processors but that is optimized
13007 for a particular microarchitecture.
13009 @item -mcpu=@var{cpu}
13011 Generate code for a specific M680x0 or ColdFire processor.
13012 The M680x0 @var{cpu}s are: @samp{68000}, @samp{68010}, @samp{68020},
13013 @samp{68030}, @samp{68040}, @samp{68060}, @samp{68302}, @samp{68332}
13014 and @samp{cpu32}. The ColdFire @var{cpu}s are given by the table
13015 below, which also classifies the CPUs into families:
13017 @multitable @columnfractions 0.20 0.80
13018 @item @strong{Family} @tab @strong{@samp{-mcpu} arguments}
13019 @item @samp{51} @tab @samp{51} @samp{51ac} @samp{51cn} @samp{51em} @samp{51qe}
13020 @item @samp{5206} @tab @samp{5202} @samp{5204} @samp{5206}
13021 @item @samp{5206e} @tab @samp{5206e}
13022 @item @samp{5208} @tab @samp{5207} @samp{5208}
13023 @item @samp{5211a} @tab @samp{5210a} @samp{5211a}
13024 @item @samp{5213} @tab @samp{5211} @samp{5212} @samp{5213}
13025 @item @samp{5216} @tab @samp{5214} @samp{5216}
13026 @item @samp{52235} @tab @samp{52230} @samp{52231} @samp{52232} @samp{52233} @samp{52234} @samp{52235}
13027 @item @samp{5225} @tab @samp{5224} @samp{5225}
13028 @item @samp{52259} @tab @samp{52252} @samp{52254} @samp{52255} @samp{52256} @samp{52258} @samp{52259}
13029 @item @samp{5235} @tab @samp{5232} @samp{5233} @samp{5234} @samp{5235} @samp{523x}
13030 @item @samp{5249} @tab @samp{5249}
13031 @item @samp{5250} @tab @samp{5250}
13032 @item @samp{5271} @tab @samp{5270} @samp{5271}
13033 @item @samp{5272} @tab @samp{5272}
13034 @item @samp{5275} @tab @samp{5274} @samp{5275}
13035 @item @samp{5282} @tab @samp{5280} @samp{5281} @samp{5282} @samp{528x}
13036 @item @samp{53017} @tab @samp{53011} @samp{53012} @samp{53013} @samp{53014} @samp{53015} @samp{53016} @samp{53017}
13037 @item @samp{5307} @tab @samp{5307}
13038 @item @samp{5329} @tab @samp{5327} @samp{5328} @samp{5329} @samp{532x}
13039 @item @samp{5373} @tab @samp{5372} @samp{5373} @samp{537x}
13040 @item @samp{5407} @tab @samp{5407}
13041 @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}
13044 @option{-mcpu=@var{cpu}} overrides @option{-march=@var{arch}} if
13045 @var{arch} is compatible with @var{cpu}. Other combinations of
13046 @option{-mcpu} and @option{-march} are rejected.
13048 gcc defines the macro @samp{__mcf_cpu_@var{cpu}} when ColdFire target
13049 @var{cpu} is selected. It also defines @samp{__mcf_family_@var{family}},
13050 where the value of @var{family} is given by the table above.
13052 @item -mtune=@var{tune}
13054 Tune the code for a particular microarchitecture, within the
13055 constraints set by @option{-march} and @option{-mcpu}.
13056 The M680x0 microarchitectures are: @samp{68000}, @samp{68010},
13057 @samp{68020}, @samp{68030}, @samp{68040}, @samp{68060}
13058 and @samp{cpu32}. The ColdFire microarchitectures
13059 are: @samp{cfv1}, @samp{cfv2}, @samp{cfv3}, @samp{cfv4} and @samp{cfv4e}.
13061 You can also use @option{-mtune=68020-40} for code that needs
13062 to run relatively well on 68020, 68030 and 68040 targets.
13063 @option{-mtune=68020-60} is similar but includes 68060 targets
13064 as well. These two options select the same tuning decisions as
13065 @option{-m68020-40} and @option{-m68020-60} respectively.
13067 gcc defines the macros @samp{__mc@var{arch}} and @samp{__mc@var{arch}__}
13068 when tuning for 680x0 architecture @var{arch}. It also defines
13069 @samp{mc@var{arch}} unless either @option{-ansi} or a non-GNU @option{-std}
13070 option is used. If gcc is tuning for a range of architectures,
13071 as selected by @option{-mtune=68020-40} or @option{-mtune=68020-60},
13072 it defines the macros for every architecture in the range.
13074 gcc also defines the macro @samp{__m@var{uarch}__} when tuning for
13075 ColdFire microarchitecture @var{uarch}, where @var{uarch} is one
13076 of the arguments given above.
13082 Generate output for a 68000. This is the default
13083 when the compiler is configured for 68000-based systems.
13084 It is equivalent to @option{-march=68000}.
13086 Use this option for microcontrollers with a 68000 or EC000 core,
13087 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
13091 Generate output for a 68010. This is the default
13092 when the compiler is configured for 68010-based systems.
13093 It is equivalent to @option{-march=68010}.
13099 Generate output for a 68020. This is the default
13100 when the compiler is configured for 68020-based systems.
13101 It is equivalent to @option{-march=68020}.
13105 Generate output for a 68030. This is the default when the compiler is
13106 configured for 68030-based systems. It is equivalent to
13107 @option{-march=68030}.
13111 Generate output for a 68040. This is the default when the compiler is
13112 configured for 68040-based systems. It is equivalent to
13113 @option{-march=68040}.
13115 This option inhibits the use of 68881/68882 instructions that have to be
13116 emulated by software on the 68040. Use this option if your 68040 does not
13117 have code to emulate those instructions.
13121 Generate output for a 68060. This is the default when the compiler is
13122 configured for 68060-based systems. It is equivalent to
13123 @option{-march=68060}.
13125 This option inhibits the use of 68020 and 68881/68882 instructions that
13126 have to be emulated by software on the 68060. Use this option if your 68060
13127 does not have code to emulate those instructions.
13131 Generate output for a CPU32. This is the default
13132 when the compiler is configured for CPU32-based systems.
13133 It is equivalent to @option{-march=cpu32}.
13135 Use this option for microcontrollers with a
13136 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
13137 68336, 68340, 68341, 68349 and 68360.
13141 Generate output for a 520X ColdFire CPU@. This is the default
13142 when the compiler is configured for 520X-based systems.
13143 It is equivalent to @option{-mcpu=5206}, and is now deprecated
13144 in favor of that option.
13146 Use this option for microcontroller with a 5200 core, including
13147 the MCF5202, MCF5203, MCF5204 and MCF5206.
13151 Generate output for a 5206e ColdFire CPU@. The option is now
13152 deprecated in favor of the equivalent @option{-mcpu=5206e}.
13156 Generate output for a member of the ColdFire 528X family.
13157 The option is now deprecated in favor of the equivalent
13158 @option{-mcpu=528x}.
13162 Generate output for a ColdFire 5307 CPU@. The option is now deprecated
13163 in favor of the equivalent @option{-mcpu=5307}.
13167 Generate output for a ColdFire 5407 CPU@. The option is now deprecated
13168 in favor of the equivalent @option{-mcpu=5407}.
13172 Generate output for a ColdFire V4e family CPU (e.g.@: 547x/548x).
13173 This includes use of hardware floating point instructions.
13174 The option is equivalent to @option{-mcpu=547x}, and is now
13175 deprecated in favor of that option.
13179 Generate output for a 68040, without using any of the new instructions.
13180 This results in code which can run relatively efficiently on either a
13181 68020/68881 or a 68030 or a 68040. The generated code does use the
13182 68881 instructions that are emulated on the 68040.
13184 The option is equivalent to @option{-march=68020} @option{-mtune=68020-40}.
13188 Generate output for a 68060, without using any of the new instructions.
13189 This results in code which can run relatively efficiently on either a
13190 68020/68881 or a 68030 or a 68040. The generated code does use the
13191 68881 instructions that are emulated on the 68060.
13193 The option is equivalent to @option{-march=68020} @option{-mtune=68020-60}.
13197 @opindex mhard-float
13199 Generate floating-point instructions. This is the default for 68020
13200 and above, and for ColdFire devices that have an FPU@. It defines the
13201 macro @samp{__HAVE_68881__} on M680x0 targets and @samp{__mcffpu__}
13202 on ColdFire targets.
13205 @opindex msoft-float
13206 Do not generate floating-point instructions; use library calls instead.
13207 This is the default for 68000, 68010, and 68832 targets. It is also
13208 the default for ColdFire devices that have no FPU.
13214 Generate (do not generate) ColdFire hardware divide and remainder
13215 instructions. If @option{-march} is used without @option{-mcpu},
13216 the default is ``on'' for ColdFire architectures and ``off'' for M680x0
13217 architectures. Otherwise, the default is taken from the target CPU
13218 (either the default CPU, or the one specified by @option{-mcpu}). For
13219 example, the default is ``off'' for @option{-mcpu=5206} and ``on'' for
13220 @option{-mcpu=5206e}.
13222 gcc defines the macro @samp{__mcfhwdiv__} when this option is enabled.
13226 Consider type @code{int} to be 16 bits wide, like @code{short int}.
13227 Additionally, parameters passed on the stack are also aligned to a
13228 16-bit boundary even on targets whose API mandates promotion to 32-bit.
13232 Do not consider type @code{int} to be 16 bits wide. This is the default.
13235 @itemx -mno-bitfield
13236 @opindex mnobitfield
13237 @opindex mno-bitfield
13238 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
13239 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
13243 Do use the bit-field instructions. The @option{-m68020} option implies
13244 @option{-mbitfield}. This is the default if you use a configuration
13245 designed for a 68020.
13249 Use a different function-calling convention, in which functions
13250 that take a fixed number of arguments return with the @code{rtd}
13251 instruction, which pops their arguments while returning. This
13252 saves one instruction in the caller since there is no need to pop
13253 the arguments there.
13255 This calling convention is incompatible with the one normally
13256 used on Unix, so you cannot use it if you need to call libraries
13257 compiled with the Unix compiler.
13259 Also, you must provide function prototypes for all functions that
13260 take variable numbers of arguments (including @code{printf});
13261 otherwise incorrect code will be generated for calls to those
13264 In addition, seriously incorrect code will result if you call a
13265 function with too many arguments. (Normally, extra arguments are
13266 harmlessly ignored.)
13268 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
13269 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
13273 Do not use the calling conventions selected by @option{-mrtd}.
13274 This is the default.
13277 @itemx -mno-align-int
13278 @opindex malign-int
13279 @opindex mno-align-int
13280 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
13281 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
13282 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
13283 Aligning variables on 32-bit boundaries produces code that runs somewhat
13284 faster on processors with 32-bit busses at the expense of more memory.
13286 @strong{Warning:} if you use the @option{-malign-int} switch, GCC will
13287 align structures containing the above types differently than
13288 most published application binary interface specifications for the m68k.
13292 Use the pc-relative addressing mode of the 68000 directly, instead of
13293 using a global offset table. At present, this option implies @option{-fpic},
13294 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
13295 not presently supported with @option{-mpcrel}, though this could be supported for
13296 68020 and higher processors.
13298 @item -mno-strict-align
13299 @itemx -mstrict-align
13300 @opindex mno-strict-align
13301 @opindex mstrict-align
13302 Do not (do) assume that unaligned memory references will be handled by
13306 Generate code that allows the data segment to be located in a different
13307 area of memory from the text segment. This allows for execute in place in
13308 an environment without virtual memory management. This option implies
13311 @item -mno-sep-data
13312 Generate code that assumes that the data segment follows the text segment.
13313 This is the default.
13315 @item -mid-shared-library
13316 Generate code that supports shared libraries via the library ID method.
13317 This allows for execute in place and shared libraries in an environment
13318 without virtual memory management. This option implies @option{-fPIC}.
13320 @item -mno-id-shared-library
13321 Generate code that doesn't assume ID based shared libraries are being used.
13322 This is the default.
13324 @item -mshared-library-id=n
13325 Specified the identification number of the ID based shared library being
13326 compiled. Specifying a value of 0 will generate more compact code, specifying
13327 other values will force the allocation of that number to the current
13328 library but is no more space or time efficient than omitting this option.
13334 When generating position-independent code for ColdFire, generate code
13335 that works if the GOT has more than 8192 entries. This code is
13336 larger and slower than code generated without this option. On M680x0
13337 processors, this option is not needed; @option{-fPIC} suffices.
13339 GCC normally uses a single instruction to load values from the GOT@.
13340 While this is relatively efficient, it only works if the GOT
13341 is smaller than about 64k. Anything larger causes the linker
13342 to report an error such as:
13344 @cindex relocation truncated to fit (ColdFire)
13346 relocation truncated to fit: R_68K_GOT16O foobar
13349 If this happens, you should recompile your code with @option{-mxgot}.
13350 It should then work with very large GOTs. However, code generated with
13351 @option{-mxgot} is less efficient, since it takes 4 instructions to fetch
13352 the value of a global symbol.
13354 Note that some linkers, including newer versions of the GNU linker,
13355 can create multiple GOTs and sort GOT entries. If you have such a linker,
13356 you should only need to use @option{-mxgot} when compiling a single
13357 object file that accesses more than 8192 GOT entries. Very few do.
13359 These options have no effect unless GCC is generating
13360 position-independent code.
13364 @node M68hc1x Options
13365 @subsection M68hc1x Options
13366 @cindex M68hc1x options
13368 These are the @samp{-m} options defined for the 68hc11 and 68hc12
13369 microcontrollers. The default values for these options depends on
13370 which style of microcontroller was selected when the compiler was configured;
13371 the defaults for the most common choices are given below.
13378 Generate output for a 68HC11. This is the default
13379 when the compiler is configured for 68HC11-based systems.
13385 Generate output for a 68HC12. This is the default
13386 when the compiler is configured for 68HC12-based systems.
13392 Generate output for a 68HCS12.
13394 @item -mauto-incdec
13395 @opindex mauto-incdec
13396 Enable the use of 68HC12 pre and post auto-increment and auto-decrement
13403 Enable the use of 68HC12 min and max instructions.
13406 @itemx -mno-long-calls
13407 @opindex mlong-calls
13408 @opindex mno-long-calls
13409 Treat all calls as being far away (near). If calls are assumed to be
13410 far away, the compiler will use the @code{call} instruction to
13411 call a function and the @code{rtc} instruction for returning.
13415 Consider type @code{int} to be 16 bits wide, like @code{short int}.
13417 @item -msoft-reg-count=@var{count}
13418 @opindex msoft-reg-count
13419 Specify the number of pseudo-soft registers which are used for the
13420 code generation. The maximum number is 32. Using more pseudo-soft
13421 register may or may not result in better code depending on the program.
13422 The default is 4 for 68HC11 and 2 for 68HC12.
13426 @node MCore Options
13427 @subsection MCore Options
13428 @cindex MCore options
13430 These are the @samp{-m} options defined for the Motorola M*Core
13436 @itemx -mno-hardlit
13438 @opindex mno-hardlit
13439 Inline constants into the code stream if it can be done in two
13440 instructions or less.
13446 Use the divide instruction. (Enabled by default).
13448 @item -mrelax-immediate
13449 @itemx -mno-relax-immediate
13450 @opindex mrelax-immediate
13451 @opindex mno-relax-immediate
13452 Allow arbitrary sized immediates in bit operations.
13454 @item -mwide-bitfields
13455 @itemx -mno-wide-bitfields
13456 @opindex mwide-bitfields
13457 @opindex mno-wide-bitfields
13458 Always treat bit-fields as int-sized.
13460 @item -m4byte-functions
13461 @itemx -mno-4byte-functions
13462 @opindex m4byte-functions
13463 @opindex mno-4byte-functions
13464 Force all functions to be aligned to a four byte boundary.
13466 @item -mcallgraph-data
13467 @itemx -mno-callgraph-data
13468 @opindex mcallgraph-data
13469 @opindex mno-callgraph-data
13470 Emit callgraph information.
13473 @itemx -mno-slow-bytes
13474 @opindex mslow-bytes
13475 @opindex mno-slow-bytes
13476 Prefer word access when reading byte quantities.
13478 @item -mlittle-endian
13479 @itemx -mbig-endian
13480 @opindex mlittle-endian
13481 @opindex mbig-endian
13482 Generate code for a little endian target.
13488 Generate code for the 210 processor.
13492 Assume that run-time support has been provided and so omit the
13493 simulator library (@file{libsim.a)} from the linker command line.
13495 @item -mstack-increment=@var{size}
13496 @opindex mstack-increment
13497 Set the maximum amount for a single stack increment operation. Large
13498 values can increase the speed of programs which contain functions
13499 that need a large amount of stack space, but they can also trigger a
13500 segmentation fault if the stack is extended too much. The default
13506 @subsection MeP Options
13507 @cindex MeP options
13513 Enables the @code{abs} instruction, which is the absolute difference
13514 between two registers.
13518 Enables all the optional instructions - average, multiply, divide, bit
13519 operations, leading zero, absolute difference, min/max, clip, and
13525 Enables the @code{ave} instruction, which computes the average of two
13528 @item -mbased=@var{n}
13530 Variables of size @var{n} bytes or smaller will be placed in the
13531 @code{.based} section by default. Based variables use the @code{$tp}
13532 register as a base register, and there is a 128 byte limit to the
13533 @code{.based} section.
13537 Enables the bit operation instructions - bit test (@code{btstm}), set
13538 (@code{bsetm}), clear (@code{bclrm}), invert (@code{bnotm}), and
13539 test-and-set (@code{tas}).
13541 @item -mc=@var{name}
13543 Selects which section constant data will be placed in. @var{name} may
13544 be @code{tiny}, @code{near}, or @code{far}.
13548 Enables the @code{clip} instruction. Note that @code{-mclip} is not
13549 useful unless you also provide @code{-mminmax}.
13551 @item -mconfig=@var{name}
13553 Selects one of the build-in core configurations. Each MeP chip has
13554 one or more modules in it; each module has a core CPU and a variety of
13555 coprocessors, optional instructions, and peripherals. The
13556 @code{MeP-Integrator} tool, not part of GCC, provides these
13557 configurations through this option; using this option is the same as
13558 using all the corresponding command line options. The default
13559 configuration is @code{default}.
13563 Enables the coprocessor instructions. By default, this is a 32-bit
13564 coprocessor. Note that the coprocessor is normally enabled via the
13565 @code{-mconfig=} option.
13569 Enables the 32-bit coprocessor's instructions.
13573 Enables the 64-bit coprocessor's instructions.
13577 Enables IVC2 scheduling. IVC2 is a 64-bit VLIW coprocessor.
13581 Causes constant variables to be placed in the @code{.near} section.
13585 Enables the @code{div} and @code{divu} instructions.
13589 Generate big-endian code.
13593 Generate little-endian code.
13595 @item -mio-volatile
13596 @opindex mio-volatile
13597 Tells the compiler that any variable marked with the @code{io}
13598 attribute is to be considered volatile.
13602 Causes variables to be assigned to the @code{.far} section by default.
13606 Enables the @code{leadz} (leading zero) instruction.
13610 Causes variables to be assigned to the @code{.near} section by default.
13614 Enables the @code{min} and @code{max} instructions.
13618 Enables the multiplication and multiply-accumulate instructions.
13622 Disables all the optional instructions enabled by @code{-mall-opts}.
13626 Enables the @code{repeat} and @code{erepeat} instructions, used for
13627 low-overhead looping.
13631 Causes all variables to default to the @code{.tiny} section. Note
13632 that there is a 65536 byte limit to this section. Accesses to these
13633 variables use the @code{%gp} base register.
13637 Enables the saturation instructions. Note that the compiler does not
13638 currently generate these itself, but this option is included for
13639 compatibility with other tools, like @code{as}.
13643 Link the SDRAM-based runtime instead of the default ROM-based runtime.
13647 Link the simulator runtime libraries.
13651 Link the simulator runtime libraries, excluding built-in support
13652 for reset and exception vectors and tables.
13656 Causes all functions to default to the @code{.far} section. Without
13657 this option, functions default to the @code{.near} section.
13659 @item -mtiny=@var{n}
13661 Variables that are @var{n} bytes or smaller will be allocated to the
13662 @code{.tiny} section. These variables use the @code{$gp} base
13663 register. The default for this option is 4, but note that there's a
13664 65536 byte limit to the @code{.tiny} section.
13669 @subsection MIPS Options
13670 @cindex MIPS options
13676 Generate big-endian code.
13680 Generate little-endian code. This is the default for @samp{mips*el-*-*}
13683 @item -march=@var{arch}
13685 Generate code that will run on @var{arch}, which can be the name of a
13686 generic MIPS ISA, or the name of a particular processor.
13688 @samp{mips1}, @samp{mips2}, @samp{mips3}, @samp{mips4},
13689 @samp{mips32}, @samp{mips32r2}, @samp{mips64} and @samp{mips64r2}.
13690 The processor names are:
13691 @samp{4kc}, @samp{4km}, @samp{4kp}, @samp{4ksc},
13692 @samp{4kec}, @samp{4kem}, @samp{4kep}, @samp{4ksd},
13693 @samp{5kc}, @samp{5kf},
13695 @samp{24kc}, @samp{24kf2_1}, @samp{24kf1_1},
13696 @samp{24kec}, @samp{24kef2_1}, @samp{24kef1_1},
13697 @samp{34kc}, @samp{34kf2_1}, @samp{34kf1_1},
13698 @samp{74kc}, @samp{74kf2_1}, @samp{74kf1_1}, @samp{74kf3_2},
13699 @samp{1004kc}, @samp{1004kf2_1}, @samp{1004kf1_1},
13700 @samp{loongson2e}, @samp{loongson2f},
13704 @samp{r2000}, @samp{r3000}, @samp{r3900}, @samp{r4000}, @samp{r4400},
13705 @samp{r4600}, @samp{r4650}, @samp{r6000}, @samp{r8000},
13706 @samp{rm7000}, @samp{rm9000},
13707 @samp{r10000}, @samp{r12000}, @samp{r14000}, @samp{r16000},
13710 @samp{vr4100}, @samp{vr4111}, @samp{vr4120}, @samp{vr4130}, @samp{vr4300},
13711 @samp{vr5000}, @samp{vr5400}, @samp{vr5500}
13713 The special value @samp{from-abi} selects the
13714 most compatible architecture for the selected ABI (that is,
13715 @samp{mips1} for 32-bit ABIs and @samp{mips3} for 64-bit ABIs)@.
13717 Native Linux/GNU toolchains also support the value @samp{native},
13718 which selects the best architecture option for the host processor.
13719 @option{-march=native} has no effect if GCC does not recognize
13722 In processor names, a final @samp{000} can be abbreviated as @samp{k}
13723 (for example, @samp{-march=r2k}). Prefixes are optional, and
13724 @samp{vr} may be written @samp{r}.
13726 Names of the form @samp{@var{n}f2_1} refer to processors with
13727 FPUs clocked at half the rate of the core, names of the form
13728 @samp{@var{n}f1_1} refer to processors with FPUs clocked at the same
13729 rate as the core, and names of the form @samp{@var{n}f3_2} refer to
13730 processors with FPUs clocked a ratio of 3:2 with respect to the core.
13731 For compatibility reasons, @samp{@var{n}f} is accepted as a synonym
13732 for @samp{@var{n}f2_1} while @samp{@var{n}x} and @samp{@var{b}fx} are
13733 accepted as synonyms for @samp{@var{n}f1_1}.
13735 GCC defines two macros based on the value of this option. The first
13736 is @samp{_MIPS_ARCH}, which gives the name of target architecture, as
13737 a string. The second has the form @samp{_MIPS_ARCH_@var{foo}},
13738 where @var{foo} is the capitalized value of @samp{_MIPS_ARCH}@.
13739 For example, @samp{-march=r2000} will set @samp{_MIPS_ARCH}
13740 to @samp{"r2000"} and define the macro @samp{_MIPS_ARCH_R2000}.
13742 Note that the @samp{_MIPS_ARCH} macro uses the processor names given
13743 above. In other words, it will have the full prefix and will not
13744 abbreviate @samp{000} as @samp{k}. In the case of @samp{from-abi},
13745 the macro names the resolved architecture (either @samp{"mips1"} or
13746 @samp{"mips3"}). It names the default architecture when no
13747 @option{-march} option is given.
13749 @item -mtune=@var{arch}
13751 Optimize for @var{arch}. Among other things, this option controls
13752 the way instructions are scheduled, and the perceived cost of arithmetic
13753 operations. The list of @var{arch} values is the same as for
13756 When this option is not used, GCC will optimize for the processor
13757 specified by @option{-march}. By using @option{-march} and
13758 @option{-mtune} together, it is possible to generate code that will
13759 run on a family of processors, but optimize the code for one
13760 particular member of that family.
13762 @samp{-mtune} defines the macros @samp{_MIPS_TUNE} and
13763 @samp{_MIPS_TUNE_@var{foo}}, which work in the same way as the
13764 @samp{-march} ones described above.
13768 Equivalent to @samp{-march=mips1}.
13772 Equivalent to @samp{-march=mips2}.
13776 Equivalent to @samp{-march=mips3}.
13780 Equivalent to @samp{-march=mips4}.
13784 Equivalent to @samp{-march=mips32}.
13788 Equivalent to @samp{-march=mips32r2}.
13792 Equivalent to @samp{-march=mips64}.
13796 Equivalent to @samp{-march=mips64r2}.
13801 @opindex mno-mips16
13802 Generate (do not generate) MIPS16 code. If GCC is targetting a
13803 MIPS32 or MIPS64 architecture, it will make use of the MIPS16e ASE@.
13805 MIPS16 code generation can also be controlled on a per-function basis
13806 by means of @code{mips16} and @code{nomips16} attributes.
13807 @xref{Function Attributes}, for more information.
13809 @item -mflip-mips16
13810 @opindex mflip-mips16
13811 Generate MIPS16 code on alternating functions. This option is provided
13812 for regression testing of mixed MIPS16/non-MIPS16 code generation, and is
13813 not intended for ordinary use in compiling user code.
13815 @item -minterlink-mips16
13816 @itemx -mno-interlink-mips16
13817 @opindex minterlink-mips16
13818 @opindex mno-interlink-mips16
13819 Require (do not require) that non-MIPS16 code be link-compatible with
13822 For example, non-MIPS16 code cannot jump directly to MIPS16 code;
13823 it must either use a call or an indirect jump. @option{-minterlink-mips16}
13824 therefore disables direct jumps unless GCC knows that the target of the
13825 jump is not MIPS16.
13837 Generate code for the given ABI@.
13839 Note that the EABI has a 32-bit and a 64-bit variant. GCC normally
13840 generates 64-bit code when you select a 64-bit architecture, but you
13841 can use @option{-mgp32} to get 32-bit code instead.
13843 For information about the O64 ABI, see
13844 @w{@uref{http://gcc.gnu.org/projects/mipso64-abi.html}}.
13846 GCC supports a variant of the o32 ABI in which floating-point registers
13847 are 64 rather than 32 bits wide. You can select this combination with
13848 @option{-mabi=32} @option{-mfp64}. This ABI relies on the @samp{mthc1}
13849 and @samp{mfhc1} instructions and is therefore only supported for
13850 MIPS32R2 processors.
13852 The register assignments for arguments and return values remain the
13853 same, but each scalar value is passed in a single 64-bit register
13854 rather than a pair of 32-bit registers. For example, scalar
13855 floating-point values are returned in @samp{$f0} only, not a
13856 @samp{$f0}/@samp{$f1} pair. The set of call-saved registers also
13857 remains the same, but all 64 bits are saved.
13860 @itemx -mno-abicalls
13862 @opindex mno-abicalls
13863 Generate (do not generate) code that is suitable for SVR4-style
13864 dynamic objects. @option{-mabicalls} is the default for SVR4-based
13869 Generate (do not generate) code that is fully position-independent,
13870 and that can therefore be linked into shared libraries. This option
13871 only affects @option{-mabicalls}.
13873 All @option{-mabicalls} code has traditionally been position-independent,
13874 regardless of options like @option{-fPIC} and @option{-fpic}. However,
13875 as an extension, the GNU toolchain allows executables to use absolute
13876 accesses for locally-binding symbols. It can also use shorter GP
13877 initialization sequences and generate direct calls to locally-defined
13878 functions. This mode is selected by @option{-mno-shared}.
13880 @option{-mno-shared} depends on binutils 2.16 or higher and generates
13881 objects that can only be linked by the GNU linker. However, the option
13882 does not affect the ABI of the final executable; it only affects the ABI
13883 of relocatable objects. Using @option{-mno-shared} will generally make
13884 executables both smaller and quicker.
13886 @option{-mshared} is the default.
13892 Assume (do not assume) that the static and dynamic linkers
13893 support PLTs and copy relocations. This option only affects
13894 @samp{-mno-shared -mabicalls}. For the n64 ABI, this option
13895 has no effect without @samp{-msym32}.
13897 You can make @option{-mplt} the default by configuring
13898 GCC with @option{--with-mips-plt}. The default is
13899 @option{-mno-plt} otherwise.
13905 Lift (do not lift) the usual restrictions on the size of the global
13908 GCC normally uses a single instruction to load values from the GOT@.
13909 While this is relatively efficient, it will only work if the GOT
13910 is smaller than about 64k. Anything larger will cause the linker
13911 to report an error such as:
13913 @cindex relocation truncated to fit (MIPS)
13915 relocation truncated to fit: R_MIPS_GOT16 foobar
13918 If this happens, you should recompile your code with @option{-mxgot}.
13919 It should then work with very large GOTs, although it will also be
13920 less efficient, since it will take three instructions to fetch the
13921 value of a global symbol.
13923 Note that some linkers can create multiple GOTs. If you have such a
13924 linker, you should only need to use @option{-mxgot} when a single object
13925 file accesses more than 64k's worth of GOT entries. Very few do.
13927 These options have no effect unless GCC is generating position
13932 Assume that general-purpose registers are 32 bits wide.
13936 Assume that general-purpose registers are 64 bits wide.
13940 Assume that floating-point registers are 32 bits wide.
13944 Assume that floating-point registers are 64 bits wide.
13947 @opindex mhard-float
13948 Use floating-point coprocessor instructions.
13951 @opindex msoft-float
13952 Do not use floating-point coprocessor instructions. Implement
13953 floating-point calculations using library calls instead.
13955 @item -msingle-float
13956 @opindex msingle-float
13957 Assume that the floating-point coprocessor only supports single-precision
13960 @item -mdouble-float
13961 @opindex mdouble-float
13962 Assume that the floating-point coprocessor supports double-precision
13963 operations. This is the default.
13969 Use (do not use) @samp{ll}, @samp{sc}, and @samp{sync} instructions to
13970 implement atomic memory built-in functions. When neither option is
13971 specified, GCC will use the instructions if the target architecture
13974 @option{-mllsc} is useful if the runtime environment can emulate the
13975 instructions and @option{-mno-llsc} can be useful when compiling for
13976 nonstandard ISAs. You can make either option the default by
13977 configuring GCC with @option{--with-llsc} and @option{--without-llsc}
13978 respectively. @option{--with-llsc} is the default for some
13979 configurations; see the installation documentation for details.
13985 Use (do not use) revision 1 of the MIPS DSP ASE@.
13986 @xref{MIPS DSP Built-in Functions}. This option defines the
13987 preprocessor macro @samp{__mips_dsp}. It also defines
13988 @samp{__mips_dsp_rev} to 1.
13994 Use (do not use) revision 2 of the MIPS DSP ASE@.
13995 @xref{MIPS DSP Built-in Functions}. This option defines the
13996 preprocessor macros @samp{__mips_dsp} and @samp{__mips_dspr2}.
13997 It also defines @samp{__mips_dsp_rev} to 2.
14000 @itemx -mno-smartmips
14001 @opindex msmartmips
14002 @opindex mno-smartmips
14003 Use (do not use) the MIPS SmartMIPS ASE.
14005 @item -mpaired-single
14006 @itemx -mno-paired-single
14007 @opindex mpaired-single
14008 @opindex mno-paired-single
14009 Use (do not use) paired-single floating-point instructions.
14010 @xref{MIPS Paired-Single Support}. This option requires
14011 hardware floating-point support to be enabled.
14017 Use (do not use) MIPS Digital Media Extension instructions.
14018 This option can only be used when generating 64-bit code and requires
14019 hardware floating-point support to be enabled.
14024 @opindex mno-mips3d
14025 Use (do not use) the MIPS-3D ASE@. @xref{MIPS-3D Built-in Functions}.
14026 The option @option{-mips3d} implies @option{-mpaired-single}.
14032 Use (do not use) MT Multithreading instructions.
14036 Force @code{long} types to be 64 bits wide. See @option{-mlong32} for
14037 an explanation of the default and the way that the pointer size is
14042 Force @code{long}, @code{int}, and pointer types to be 32 bits wide.
14044 The default size of @code{int}s, @code{long}s and pointers depends on
14045 the ABI@. All the supported ABIs use 32-bit @code{int}s. The n64 ABI
14046 uses 64-bit @code{long}s, as does the 64-bit EABI; the others use
14047 32-bit @code{long}s. Pointers are the same size as @code{long}s,
14048 or the same size as integer registers, whichever is smaller.
14054 Assume (do not assume) that all symbols have 32-bit values, regardless
14055 of the selected ABI@. This option is useful in combination with
14056 @option{-mabi=64} and @option{-mno-abicalls} because it allows GCC
14057 to generate shorter and faster references to symbolic addresses.
14061 Put definitions of externally-visible data in a small data section
14062 if that data is no bigger than @var{num} bytes. GCC can then access
14063 the data more efficiently; see @option{-mgpopt} for details.
14065 The default @option{-G} option depends on the configuration.
14067 @item -mlocal-sdata
14068 @itemx -mno-local-sdata
14069 @opindex mlocal-sdata
14070 @opindex mno-local-sdata
14071 Extend (do not extend) the @option{-G} behavior to local data too,
14072 such as to static variables in C@. @option{-mlocal-sdata} is the
14073 default for all configurations.
14075 If the linker complains that an application is using too much small data,
14076 you might want to try rebuilding the less performance-critical parts with
14077 @option{-mno-local-sdata}. You might also want to build large
14078 libraries with @option{-mno-local-sdata}, so that the libraries leave
14079 more room for the main program.
14081 @item -mextern-sdata
14082 @itemx -mno-extern-sdata
14083 @opindex mextern-sdata
14084 @opindex mno-extern-sdata
14085 Assume (do not assume) that externally-defined data will be in
14086 a small data section if that data is within the @option{-G} limit.
14087 @option{-mextern-sdata} is the default for all configurations.
14089 If you compile a module @var{Mod} with @option{-mextern-sdata} @option{-G
14090 @var{num}} @option{-mgpopt}, and @var{Mod} references a variable @var{Var}
14091 that is no bigger than @var{num} bytes, you must make sure that @var{Var}
14092 is placed in a small data section. If @var{Var} is defined by another
14093 module, you must either compile that module with a high-enough
14094 @option{-G} setting or attach a @code{section} attribute to @var{Var}'s
14095 definition. If @var{Var} is common, you must link the application
14096 with a high-enough @option{-G} setting.
14098 The easiest way of satisfying these restrictions is to compile
14099 and link every module with the same @option{-G} option. However,
14100 you may wish to build a library that supports several different
14101 small data limits. You can do this by compiling the library with
14102 the highest supported @option{-G} setting and additionally using
14103 @option{-mno-extern-sdata} to stop the library from making assumptions
14104 about externally-defined data.
14110 Use (do not use) GP-relative accesses for symbols that are known to be
14111 in a small data section; see @option{-G}, @option{-mlocal-sdata} and
14112 @option{-mextern-sdata}. @option{-mgpopt} is the default for all
14115 @option{-mno-gpopt} is useful for cases where the @code{$gp} register
14116 might not hold the value of @code{_gp}. For example, if the code is
14117 part of a library that might be used in a boot monitor, programs that
14118 call boot monitor routines will pass an unknown value in @code{$gp}.
14119 (In such situations, the boot monitor itself would usually be compiled
14120 with @option{-G0}.)
14122 @option{-mno-gpopt} implies @option{-mno-local-sdata} and
14123 @option{-mno-extern-sdata}.
14125 @item -membedded-data
14126 @itemx -mno-embedded-data
14127 @opindex membedded-data
14128 @opindex mno-embedded-data
14129 Allocate variables to the read-only data section first if possible, then
14130 next in the small data section if possible, otherwise in data. This gives
14131 slightly slower code than the default, but reduces the amount of RAM required
14132 when executing, and thus may be preferred for some embedded systems.
14134 @item -muninit-const-in-rodata
14135 @itemx -mno-uninit-const-in-rodata
14136 @opindex muninit-const-in-rodata
14137 @opindex mno-uninit-const-in-rodata
14138 Put uninitialized @code{const} variables in the read-only data section.
14139 This option is only meaningful in conjunction with @option{-membedded-data}.
14141 @item -mcode-readable=@var{setting}
14142 @opindex mcode-readable
14143 Specify whether GCC may generate code that reads from executable sections.
14144 There are three possible settings:
14147 @item -mcode-readable=yes
14148 Instructions may freely access executable sections. This is the
14151 @item -mcode-readable=pcrel
14152 MIPS16 PC-relative load instructions can access executable sections,
14153 but other instructions must not do so. This option is useful on 4KSc
14154 and 4KSd processors when the code TLBs have the Read Inhibit bit set.
14155 It is also useful on processors that can be configured to have a dual
14156 instruction/data SRAM interface and that, like the M4K, automatically
14157 redirect PC-relative loads to the instruction RAM.
14159 @item -mcode-readable=no
14160 Instructions must not access executable sections. This option can be
14161 useful on targets that are configured to have a dual instruction/data
14162 SRAM interface but that (unlike the M4K) do not automatically redirect
14163 PC-relative loads to the instruction RAM.
14166 @item -msplit-addresses
14167 @itemx -mno-split-addresses
14168 @opindex msplit-addresses
14169 @opindex mno-split-addresses
14170 Enable (disable) use of the @code{%hi()} and @code{%lo()} assembler
14171 relocation operators. This option has been superseded by
14172 @option{-mexplicit-relocs} but is retained for backwards compatibility.
14174 @item -mexplicit-relocs
14175 @itemx -mno-explicit-relocs
14176 @opindex mexplicit-relocs
14177 @opindex mno-explicit-relocs
14178 Use (do not use) assembler relocation operators when dealing with symbolic
14179 addresses. The alternative, selected by @option{-mno-explicit-relocs},
14180 is to use assembler macros instead.
14182 @option{-mexplicit-relocs} is the default if GCC was configured
14183 to use an assembler that supports relocation operators.
14185 @item -mcheck-zero-division
14186 @itemx -mno-check-zero-division
14187 @opindex mcheck-zero-division
14188 @opindex mno-check-zero-division
14189 Trap (do not trap) on integer division by zero.
14191 The default is @option{-mcheck-zero-division}.
14193 @item -mdivide-traps
14194 @itemx -mdivide-breaks
14195 @opindex mdivide-traps
14196 @opindex mdivide-breaks
14197 MIPS systems check for division by zero by generating either a
14198 conditional trap or a break instruction. Using traps results in
14199 smaller code, but is only supported on MIPS II and later. Also, some
14200 versions of the Linux kernel have a bug that prevents trap from
14201 generating the proper signal (@code{SIGFPE}). Use @option{-mdivide-traps} to
14202 allow conditional traps on architectures that support them and
14203 @option{-mdivide-breaks} to force the use of breaks.
14205 The default is usually @option{-mdivide-traps}, but this can be
14206 overridden at configure time using @option{--with-divide=breaks}.
14207 Divide-by-zero checks can be completely disabled using
14208 @option{-mno-check-zero-division}.
14213 @opindex mno-memcpy
14214 Force (do not force) the use of @code{memcpy()} for non-trivial block
14215 moves. The default is @option{-mno-memcpy}, which allows GCC to inline
14216 most constant-sized copies.
14219 @itemx -mno-long-calls
14220 @opindex mlong-calls
14221 @opindex mno-long-calls
14222 Disable (do not disable) use of the @code{jal} instruction. Calling
14223 functions using @code{jal} is more efficient but requires the caller
14224 and callee to be in the same 256 megabyte segment.
14226 This option has no effect on abicalls code. The default is
14227 @option{-mno-long-calls}.
14233 Enable (disable) use of the @code{mad}, @code{madu} and @code{mul}
14234 instructions, as provided by the R4650 ISA@.
14237 @itemx -mno-fused-madd
14238 @opindex mfused-madd
14239 @opindex mno-fused-madd
14240 Enable (disable) use of the floating point multiply-accumulate
14241 instructions, when they are available. The default is
14242 @option{-mfused-madd}.
14244 When multiply-accumulate instructions are used, the intermediate
14245 product is calculated to infinite precision and is not subject to
14246 the FCSR Flush to Zero bit. This may be undesirable in some
14251 Tell the MIPS assembler to not run its preprocessor over user
14252 assembler files (with a @samp{.s} suffix) when assembling them.
14255 @itemx -mno-fix-r4000
14256 @opindex mfix-r4000
14257 @opindex mno-fix-r4000
14258 Work around certain R4000 CPU errata:
14261 A double-word or a variable shift may give an incorrect result if executed
14262 immediately after starting an integer division.
14264 A double-word or a variable shift may give an incorrect result if executed
14265 while an integer multiplication is in progress.
14267 An integer division may give an incorrect result if started in a delay slot
14268 of a taken branch or a jump.
14272 @itemx -mno-fix-r4400
14273 @opindex mfix-r4400
14274 @opindex mno-fix-r4400
14275 Work around certain R4400 CPU errata:
14278 A double-word or a variable shift may give an incorrect result if executed
14279 immediately after starting an integer division.
14283 @itemx -mno-fix-r10000
14284 @opindex mfix-r10000
14285 @opindex mno-fix-r10000
14286 Work around certain R10000 errata:
14289 @code{ll}/@code{sc} sequences may not behave atomically on revisions
14290 prior to 3.0. They may deadlock on revisions 2.6 and earlier.
14293 This option can only be used if the target architecture supports
14294 branch-likely instructions. @option{-mfix-r10000} is the default when
14295 @option{-march=r10000} is used; @option{-mno-fix-r10000} is the default
14299 @itemx -mno-fix-vr4120
14300 @opindex mfix-vr4120
14301 Work around certain VR4120 errata:
14304 @code{dmultu} does not always produce the correct result.
14306 @code{div} and @code{ddiv} do not always produce the correct result if one
14307 of the operands is negative.
14309 The workarounds for the division errata rely on special functions in
14310 @file{libgcc.a}. At present, these functions are only provided by
14311 the @code{mips64vr*-elf} configurations.
14313 Other VR4120 errata require a nop to be inserted between certain pairs of
14314 instructions. These errata are handled by the assembler, not by GCC itself.
14317 @opindex mfix-vr4130
14318 Work around the VR4130 @code{mflo}/@code{mfhi} errata. The
14319 workarounds are implemented by the assembler rather than by GCC,
14320 although GCC will avoid using @code{mflo} and @code{mfhi} if the
14321 VR4130 @code{macc}, @code{macchi}, @code{dmacc} and @code{dmacchi}
14322 instructions are available instead.
14325 @itemx -mno-fix-sb1
14327 Work around certain SB-1 CPU core errata.
14328 (This flag currently works around the SB-1 revision 2
14329 ``F1'' and ``F2'' floating point errata.)
14331 @item -mr10k-cache-barrier=@var{setting}
14332 @opindex mr10k-cache-barrier
14333 Specify whether GCC should insert cache barriers to avoid the
14334 side-effects of speculation on R10K processors.
14336 In common with many processors, the R10K tries to predict the outcome
14337 of a conditional branch and speculatively executes instructions from
14338 the ``taken'' branch. It later aborts these instructions if the
14339 predicted outcome was wrong. However, on the R10K, even aborted
14340 instructions can have side effects.
14342 This problem only affects kernel stores and, depending on the system,
14343 kernel loads. As an example, a speculatively-executed store may load
14344 the target memory into cache and mark the cache line as dirty, even if
14345 the store itself is later aborted. If a DMA operation writes to the
14346 same area of memory before the ``dirty'' line is flushed, the cached
14347 data will overwrite the DMA-ed data. See the R10K processor manual
14348 for a full description, including other potential problems.
14350 One workaround is to insert cache barrier instructions before every memory
14351 access that might be speculatively executed and that might have side
14352 effects even if aborted. @option{-mr10k-cache-barrier=@var{setting}}
14353 controls GCC's implementation of this workaround. It assumes that
14354 aborted accesses to any byte in the following regions will not have
14359 the memory occupied by the current function's stack frame;
14362 the memory occupied by an incoming stack argument;
14365 the memory occupied by an object with a link-time-constant address.
14368 It is the kernel's responsibility to ensure that speculative
14369 accesses to these regions are indeed safe.
14371 If the input program contains a function declaration such as:
14377 then the implementation of @code{foo} must allow @code{j foo} and
14378 @code{jal foo} to be executed speculatively. GCC honors this
14379 restriction for functions it compiles itself. It expects non-GCC
14380 functions (such as hand-written assembly code) to do the same.
14382 The option has three forms:
14385 @item -mr10k-cache-barrier=load-store
14386 Insert a cache barrier before a load or store that might be
14387 speculatively executed and that might have side effects even
14390 @item -mr10k-cache-barrier=store
14391 Insert a cache barrier before a store that might be speculatively
14392 executed and that might have side effects even if aborted.
14394 @item -mr10k-cache-barrier=none
14395 Disable the insertion of cache barriers. This is the default setting.
14398 @item -mflush-func=@var{func}
14399 @itemx -mno-flush-func
14400 @opindex mflush-func
14401 Specifies the function to call to flush the I and D caches, or to not
14402 call any such function. If called, the function must take the same
14403 arguments as the common @code{_flush_func()}, that is, the address of the
14404 memory range for which the cache is being flushed, the size of the
14405 memory range, and the number 3 (to flush both caches). The default
14406 depends on the target GCC was configured for, but commonly is either
14407 @samp{_flush_func} or @samp{__cpu_flush}.
14409 @item mbranch-cost=@var{num}
14410 @opindex mbranch-cost
14411 Set the cost of branches to roughly @var{num} ``simple'' instructions.
14412 This cost is only a heuristic and is not guaranteed to produce
14413 consistent results across releases. A zero cost redundantly selects
14414 the default, which is based on the @option{-mtune} setting.
14416 @item -mbranch-likely
14417 @itemx -mno-branch-likely
14418 @opindex mbranch-likely
14419 @opindex mno-branch-likely
14420 Enable or disable use of Branch Likely instructions, regardless of the
14421 default for the selected architecture. By default, Branch Likely
14422 instructions may be generated if they are supported by the selected
14423 architecture. An exception is for the MIPS32 and MIPS64 architectures
14424 and processors which implement those architectures; for those, Branch
14425 Likely instructions will not be generated by default because the MIPS32
14426 and MIPS64 architectures specifically deprecate their use.
14428 @item -mfp-exceptions
14429 @itemx -mno-fp-exceptions
14430 @opindex mfp-exceptions
14431 Specifies whether FP exceptions are enabled. This affects how we schedule
14432 FP instructions for some processors. The default is that FP exceptions are
14435 For instance, on the SB-1, if FP exceptions are disabled, and we are emitting
14436 64-bit code, then we can use both FP pipes. Otherwise, we can only use one
14439 @item -mvr4130-align
14440 @itemx -mno-vr4130-align
14441 @opindex mvr4130-align
14442 The VR4130 pipeline is two-way superscalar, but can only issue two
14443 instructions together if the first one is 8-byte aligned. When this
14444 option is enabled, GCC will align pairs of instructions that it
14445 thinks should execute in parallel.
14447 This option only has an effect when optimizing for the VR4130.
14448 It normally makes code faster, but at the expense of making it bigger.
14449 It is enabled by default at optimization level @option{-O3}.
14454 Enable (disable) generation of @code{synci} instructions on
14455 architectures that support it. The @code{synci} instructions (if
14456 enabled) will be generated when @code{__builtin___clear_cache()} is
14459 This option defaults to @code{-mno-synci}, but the default can be
14460 overridden by configuring with @code{--with-synci}.
14462 When compiling code for single processor systems, it is generally safe
14463 to use @code{synci}. However, on many multi-core (SMP) systems, it
14464 will not invalidate the instruction caches on all cores and may lead
14465 to undefined behavior.
14467 @item -mrelax-pic-calls
14468 @itemx -mno-relax-pic-calls
14469 @opindex mrelax-pic-calls
14470 Try to turn PIC calls that are normally dispatched via register
14471 @code{$25} into direct calls. This is only possible if the linker can
14472 resolve the destination at link-time and if the destination is within
14473 range for a direct call.
14475 @option{-mrelax-pic-calls} is the default if GCC was configured to use
14476 an assembler and a linker that supports the @code{.reloc} assembly
14477 directive and @code{-mexplicit-relocs} is in effect. With
14478 @code{-mno-explicit-relocs}, this optimization can be performed by the
14479 assembler and the linker alone without help from the compiler.
14481 @item -mmcount-ra-address
14482 @itemx -mno-mcount-ra-address
14483 @opindex mmcount-ra-address
14484 @opindex mno-mcount-ra-address
14485 Emit (do not emit) code that allows @code{_mcount} to modify the
14486 calling function's return address. When enabled, this option extends
14487 the usual @code{_mcount} interface with a new @var{ra-address}
14488 parameter, which has type @code{intptr_t *} and is passed in register
14489 @code{$12}. @code{_mcount} can then modify the return address by
14490 doing both of the following:
14493 Returning the new address in register @code{$31}.
14495 Storing the new address in @code{*@var{ra-address}},
14496 if @var{ra-address} is nonnull.
14499 The default is @option{-mno-mcount-ra-address}.
14504 @subsection MMIX Options
14505 @cindex MMIX Options
14507 These options are defined for the MMIX:
14511 @itemx -mno-libfuncs
14513 @opindex mno-libfuncs
14514 Specify that intrinsic library functions are being compiled, passing all
14515 values in registers, no matter the size.
14518 @itemx -mno-epsilon
14520 @opindex mno-epsilon
14521 Generate floating-point comparison instructions that compare with respect
14522 to the @code{rE} epsilon register.
14524 @item -mabi=mmixware
14526 @opindex mabi=mmixware
14528 Generate code that passes function parameters and return values that (in
14529 the called function) are seen as registers @code{$0} and up, as opposed to
14530 the GNU ABI which uses global registers @code{$231} and up.
14532 @item -mzero-extend
14533 @itemx -mno-zero-extend
14534 @opindex mzero-extend
14535 @opindex mno-zero-extend
14536 When reading data from memory in sizes shorter than 64 bits, use (do not
14537 use) zero-extending load instructions by default, rather than
14538 sign-extending ones.
14541 @itemx -mno-knuthdiv
14543 @opindex mno-knuthdiv
14544 Make the result of a division yielding a remainder have the same sign as
14545 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
14546 remainder follows the sign of the dividend. Both methods are
14547 arithmetically valid, the latter being almost exclusively used.
14549 @item -mtoplevel-symbols
14550 @itemx -mno-toplevel-symbols
14551 @opindex mtoplevel-symbols
14552 @opindex mno-toplevel-symbols
14553 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
14554 code can be used with the @code{PREFIX} assembly directive.
14558 Generate an executable in the ELF format, rather than the default
14559 @samp{mmo} format used by the @command{mmix} simulator.
14561 @item -mbranch-predict
14562 @itemx -mno-branch-predict
14563 @opindex mbranch-predict
14564 @opindex mno-branch-predict
14565 Use (do not use) the probable-branch instructions, when static branch
14566 prediction indicates a probable branch.
14568 @item -mbase-addresses
14569 @itemx -mno-base-addresses
14570 @opindex mbase-addresses
14571 @opindex mno-base-addresses
14572 Generate (do not generate) code that uses @emph{base addresses}. Using a
14573 base address automatically generates a request (handled by the assembler
14574 and the linker) for a constant to be set up in a global register. The
14575 register is used for one or more base address requests within the range 0
14576 to 255 from the value held in the register. The generally leads to short
14577 and fast code, but the number of different data items that can be
14578 addressed is limited. This means that a program that uses lots of static
14579 data may require @option{-mno-base-addresses}.
14581 @item -msingle-exit
14582 @itemx -mno-single-exit
14583 @opindex msingle-exit
14584 @opindex mno-single-exit
14585 Force (do not force) generated code to have a single exit point in each
14589 @node MN10300 Options
14590 @subsection MN10300 Options
14591 @cindex MN10300 options
14593 These @option{-m} options are defined for Matsushita MN10300 architectures:
14598 Generate code to avoid bugs in the multiply instructions for the MN10300
14599 processors. This is the default.
14601 @item -mno-mult-bug
14602 @opindex mno-mult-bug
14603 Do not generate code to avoid bugs in the multiply instructions for the
14604 MN10300 processors.
14608 Generate code which uses features specific to the AM33 processor.
14612 Do not generate code which uses features specific to the AM33 processor. This
14615 @item -mreturn-pointer-on-d0
14616 @opindex mreturn-pointer-on-d0
14617 When generating a function which returns a pointer, return the pointer
14618 in both @code{a0} and @code{d0}. Otherwise, the pointer is returned
14619 only in a0, and attempts to call such functions without a prototype
14620 would result in errors. Note that this option is on by default; use
14621 @option{-mno-return-pointer-on-d0} to disable it.
14625 Do not link in the C run-time initialization object file.
14629 Indicate to the linker that it should perform a relaxation optimization pass
14630 to shorten branches, calls and absolute memory addresses. This option only
14631 has an effect when used on the command line for the final link step.
14633 This option makes symbolic debugging impossible.
14636 @node PDP-11 Options
14637 @subsection PDP-11 Options
14638 @cindex PDP-11 Options
14640 These options are defined for the PDP-11:
14645 Use hardware FPP floating point. This is the default. (FIS floating
14646 point on the PDP-11/40 is not supported.)
14649 @opindex msoft-float
14650 Do not use hardware floating point.
14654 Return floating-point results in ac0 (fr0 in Unix assembler syntax).
14658 Return floating-point results in memory. This is the default.
14662 Generate code for a PDP-11/40.
14666 Generate code for a PDP-11/45. This is the default.
14670 Generate code for a PDP-11/10.
14672 @item -mbcopy-builtin
14673 @opindex mbcopy-builtin
14674 Use inline @code{movmemhi} patterns for copying memory. This is the
14679 Do not use inline @code{movmemhi} patterns for copying memory.
14685 Use 16-bit @code{int}. This is the default.
14691 Use 32-bit @code{int}.
14694 @itemx -mno-float32
14696 @opindex mno-float32
14697 Use 64-bit @code{float}. This is the default.
14700 @itemx -mno-float64
14702 @opindex mno-float64
14703 Use 32-bit @code{float}.
14707 Use @code{abshi2} pattern. This is the default.
14711 Do not use @code{abshi2} pattern.
14713 @item -mbranch-expensive
14714 @opindex mbranch-expensive
14715 Pretend that branches are expensive. This is for experimenting with
14716 code generation only.
14718 @item -mbranch-cheap
14719 @opindex mbranch-cheap
14720 Do not pretend that branches are expensive. This is the default.
14724 Generate code for a system with split I&D@.
14728 Generate code for a system without split I&D@. This is the default.
14732 Use Unix assembler syntax. This is the default when configured for
14733 @samp{pdp11-*-bsd}.
14737 Use DEC assembler syntax. This is the default when configured for any
14738 PDP-11 target other than @samp{pdp11-*-bsd}.
14741 @node picoChip Options
14742 @subsection picoChip Options
14743 @cindex picoChip options
14745 These @samp{-m} options are defined for picoChip implementations:
14749 @item -mae=@var{ae_type}
14751 Set the instruction set, register set, and instruction scheduling
14752 parameters for array element type @var{ae_type}. Supported values
14753 for @var{ae_type} are @samp{ANY}, @samp{MUL}, and @samp{MAC}.
14755 @option{-mae=ANY} selects a completely generic AE type. Code
14756 generated with this option will run on any of the other AE types. The
14757 code will not be as efficient as it would be if compiled for a specific
14758 AE type, and some types of operation (e.g., multiplication) will not
14759 work properly on all types of AE.
14761 @option{-mae=MUL} selects a MUL AE type. This is the most useful AE type
14762 for compiled code, and is the default.
14764 @option{-mae=MAC} selects a DSP-style MAC AE. Code compiled with this
14765 option may suffer from poor performance of byte (char) manipulation,
14766 since the DSP AE does not provide hardware support for byte load/stores.
14768 @item -msymbol-as-address
14769 Enable the compiler to directly use a symbol name as an address in a
14770 load/store instruction, without first loading it into a
14771 register. Typically, the use of this option will generate larger
14772 programs, which run faster than when the option isn't used. However, the
14773 results vary from program to program, so it is left as a user option,
14774 rather than being permanently enabled.
14776 @item -mno-inefficient-warnings
14777 Disables warnings about the generation of inefficient code. These
14778 warnings can be generated, for example, when compiling code which
14779 performs byte-level memory operations on the MAC AE type. The MAC AE has
14780 no hardware support for byte-level memory operations, so all byte
14781 load/stores must be synthesized from word load/store operations. This is
14782 inefficient and a warning will be generated indicating to the programmer
14783 that they should rewrite the code to avoid byte operations, or to target
14784 an AE type which has the necessary hardware support. This option enables
14785 the warning to be turned off.
14789 @node PowerPC Options
14790 @subsection PowerPC Options
14791 @cindex PowerPC options
14793 These are listed under @xref{RS/6000 and PowerPC Options}.
14795 @node RS/6000 and PowerPC Options
14796 @subsection IBM RS/6000 and PowerPC Options
14797 @cindex RS/6000 and PowerPC Options
14798 @cindex IBM RS/6000 and PowerPC Options
14800 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
14807 @itemx -mno-powerpc
14808 @itemx -mpowerpc-gpopt
14809 @itemx -mno-powerpc-gpopt
14810 @itemx -mpowerpc-gfxopt
14811 @itemx -mno-powerpc-gfxopt
14813 @itemx -mno-powerpc64
14817 @itemx -mno-popcntb
14819 @itemx -mno-popcntd
14827 @itemx -mno-hard-dfp
14831 @opindex mno-power2
14833 @opindex mno-powerpc
14834 @opindex mpowerpc-gpopt
14835 @opindex mno-powerpc-gpopt
14836 @opindex mpowerpc-gfxopt
14837 @opindex mno-powerpc-gfxopt
14838 @opindex mpowerpc64
14839 @opindex mno-powerpc64
14843 @opindex mno-popcntb
14845 @opindex mno-popcntd
14851 @opindex mno-mfpgpr
14853 @opindex mno-hard-dfp
14854 GCC supports two related instruction set architectures for the
14855 RS/6000 and PowerPC@. The @dfn{POWER} instruction set are those
14856 instructions supported by the @samp{rios} chip set used in the original
14857 RS/6000 systems and the @dfn{PowerPC} instruction set is the
14858 architecture of the Freescale MPC5xx, MPC6xx, MPC8xx microprocessors, and
14859 the IBM 4xx, 6xx, and follow-on microprocessors.
14861 Neither architecture is a subset of the other. However there is a
14862 large common subset of instructions supported by both. An MQ
14863 register is included in processors supporting the POWER architecture.
14865 You use these options to specify which instructions are available on the
14866 processor you are using. The default value of these options is
14867 determined when configuring GCC@. Specifying the
14868 @option{-mcpu=@var{cpu_type}} overrides the specification of these
14869 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
14870 rather than the options listed above.
14872 The @option{-mpower} option allows GCC to generate instructions that
14873 are found only in the POWER architecture and to use the MQ register.
14874 Specifying @option{-mpower2} implies @option{-power} and also allows GCC
14875 to generate instructions that are present in the POWER2 architecture but
14876 not the original POWER architecture.
14878 The @option{-mpowerpc} option allows GCC to generate instructions that
14879 are found only in the 32-bit subset of the PowerPC architecture.
14880 Specifying @option{-mpowerpc-gpopt} implies @option{-mpowerpc} and also allows
14881 GCC to use the optional PowerPC architecture instructions in the
14882 General Purpose group, including floating-point square root. Specifying
14883 @option{-mpowerpc-gfxopt} implies @option{-mpowerpc} and also allows GCC to
14884 use the optional PowerPC architecture instructions in the Graphics
14885 group, including floating-point select.
14887 The @option{-mmfcrf} option allows GCC to generate the move from
14888 condition register field instruction implemented on the POWER4
14889 processor and other processors that support the PowerPC V2.01
14891 The @option{-mpopcntb} option allows GCC to generate the popcount and
14892 double precision FP reciprocal estimate instruction implemented on the
14893 POWER5 processor and other processors that support the PowerPC V2.02
14895 The @option{-mpopcntd} option allows GCC to generate the popcount
14896 instruction implemented on the POWER7 processor and other processors
14897 that support the PowerPC V2.06 architecture.
14898 The @option{-mfprnd} option allows GCC to generate the FP round to
14899 integer instructions implemented on the POWER5+ processor and other
14900 processors that support the PowerPC V2.03 architecture.
14901 The @option{-mcmpb} option allows GCC to generate the compare bytes
14902 instruction implemented on the POWER6 processor and other processors
14903 that support the PowerPC V2.05 architecture.
14904 The @option{-mmfpgpr} option allows GCC to generate the FP move to/from
14905 general purpose register instructions implemented on the POWER6X
14906 processor and other processors that support the extended PowerPC V2.05
14908 The @option{-mhard-dfp} option allows GCC to generate the decimal floating
14909 point instructions implemented on some POWER processors.
14911 The @option{-mpowerpc64} option allows GCC to generate the additional
14912 64-bit instructions that are found in the full PowerPC64 architecture
14913 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
14914 @option{-mno-powerpc64}.
14916 If you specify both @option{-mno-power} and @option{-mno-powerpc}, GCC
14917 will use only the instructions in the common subset of both
14918 architectures plus some special AIX common-mode calls, and will not use
14919 the MQ register. Specifying both @option{-mpower} and @option{-mpowerpc}
14920 permits GCC to use any instruction from either architecture and to
14921 allow use of the MQ register; specify this for the Motorola MPC601.
14923 @item -mnew-mnemonics
14924 @itemx -mold-mnemonics
14925 @opindex mnew-mnemonics
14926 @opindex mold-mnemonics
14927 Select which mnemonics to use in the generated assembler code. With
14928 @option{-mnew-mnemonics}, GCC uses the assembler mnemonics defined for
14929 the PowerPC architecture. With @option{-mold-mnemonics} it uses the
14930 assembler mnemonics defined for the POWER architecture. Instructions
14931 defined in only one architecture have only one mnemonic; GCC uses that
14932 mnemonic irrespective of which of these options is specified.
14934 GCC defaults to the mnemonics appropriate for the architecture in
14935 use. Specifying @option{-mcpu=@var{cpu_type}} sometimes overrides the
14936 value of these option. Unless you are building a cross-compiler, you
14937 should normally not specify either @option{-mnew-mnemonics} or
14938 @option{-mold-mnemonics}, but should instead accept the default.
14940 @item -mcpu=@var{cpu_type}
14942 Set architecture type, register usage, choice of mnemonics, and
14943 instruction scheduling parameters for machine type @var{cpu_type}.
14944 Supported values for @var{cpu_type} are @samp{401}, @samp{403},
14945 @samp{405}, @samp{405fp}, @samp{440}, @samp{440fp}, @samp{464}, @samp{464fp},
14946 @samp{476}, @samp{476fp}, @samp{505}, @samp{601}, @samp{602}, @samp{603},
14947 @samp{603e}, @samp{604}, @samp{604e}, @samp{620}, @samp{630}, @samp{740},
14948 @samp{7400}, @samp{7450}, @samp{750}, @samp{801}, @samp{821}, @samp{823},
14949 @samp{860}, @samp{970}, @samp{8540}, @samp{a2}, @samp{e300c2},
14950 @samp{e300c3}, @samp{e500mc}, @samp{e500mc64}, @samp{ec603e}, @samp{G3},
14951 @samp{G4}, @samp{G5}, @samp{titan}, @samp{power}, @samp{power2}, @samp{power3},
14952 @samp{power4}, @samp{power5}, @samp{power5+}, @samp{power6}, @samp{power6x},
14953 @samp{power7}, @samp{common}, @samp{powerpc}, @samp{powerpc64}, @samp{rios},
14954 @samp{rios1}, @samp{rios2}, @samp{rsc}, and @samp{rs64}.
14956 @option{-mcpu=common} selects a completely generic processor. Code
14957 generated under this option will run on any POWER or PowerPC processor.
14958 GCC will use only the instructions in the common subset of both
14959 architectures, and will not use the MQ register. GCC assumes a generic
14960 processor model for scheduling purposes.
14962 @option{-mcpu=power}, @option{-mcpu=power2}, @option{-mcpu=powerpc}, and
14963 @option{-mcpu=powerpc64} specify generic POWER, POWER2, pure 32-bit
14964 PowerPC (i.e., not MPC601), and 64-bit PowerPC architecture machine
14965 types, with an appropriate, generic processor model assumed for
14966 scheduling purposes.
14968 The other options specify a specific processor. Code generated under
14969 those options will run best on that processor, and may not run at all on
14972 The @option{-mcpu} options automatically enable or disable the
14975 @gccoptlist{-maltivec -mfprnd -mhard-float -mmfcrf -mmultiple @gol
14976 -mnew-mnemonics -mpopcntb -mpopcntd -mpower -mpower2 -mpowerpc64 @gol
14977 -mpowerpc-gpopt -mpowerpc-gfxopt -msingle-float -mdouble-float @gol
14978 -msimple-fpu -mstring -mmulhw -mdlmzb -mmfpgpr -mvsx}
14980 The particular options set for any particular CPU will vary between
14981 compiler versions, depending on what setting seems to produce optimal
14982 code for that CPU; it doesn't necessarily reflect the actual hardware's
14983 capabilities. If you wish to set an individual option to a particular
14984 value, you may specify it after the @option{-mcpu} option, like
14985 @samp{-mcpu=970 -mno-altivec}.
14987 On AIX, the @option{-maltivec} and @option{-mpowerpc64} options are
14988 not enabled or disabled by the @option{-mcpu} option at present because
14989 AIX does not have full support for these options. You may still
14990 enable or disable them individually if you're sure it'll work in your
14993 @item -mtune=@var{cpu_type}
14995 Set the instruction scheduling parameters for machine type
14996 @var{cpu_type}, but do not set the architecture type, register usage, or
14997 choice of mnemonics, as @option{-mcpu=@var{cpu_type}} would. The same
14998 values for @var{cpu_type} are used for @option{-mtune} as for
14999 @option{-mcpu}. If both are specified, the code generated will use the
15000 architecture, registers, and mnemonics set by @option{-mcpu}, but the
15001 scheduling parameters set by @option{-mtune}.
15003 @item -mcmodel=small
15004 @opindex mcmodel=small
15005 Generate PowerPC64 code for the small model: The TOC is limited to
15008 @item -mcmodel=large
15009 @opindex mcmodel=large
15010 Generate PowerPC64 code for the large model: The TOC may be up to 4G
15011 in size. Other data and code is only limited by the 64-bit address
15015 @itemx -mno-altivec
15017 @opindex mno-altivec
15018 Generate code that uses (does not use) AltiVec instructions, and also
15019 enable the use of built-in functions that allow more direct access to
15020 the AltiVec instruction set. You may also need to set
15021 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
15027 @opindex mno-vrsave
15028 Generate VRSAVE instructions when generating AltiVec code.
15030 @item -mgen-cell-microcode
15031 @opindex mgen-cell-microcode
15032 Generate Cell microcode instructions
15034 @item -mwarn-cell-microcode
15035 @opindex mwarn-cell-microcode
15036 Warning when a Cell microcode instruction is going to emitted. An example
15037 of a Cell microcode instruction is a variable shift.
15040 @opindex msecure-plt
15041 Generate code that allows ld and ld.so to build executables and shared
15042 libraries with non-exec .plt and .got sections. This is a PowerPC
15043 32-bit SYSV ABI option.
15047 Generate code that uses a BSS .plt section that ld.so fills in, and
15048 requires .plt and .got sections that are both writable and executable.
15049 This is a PowerPC 32-bit SYSV ABI option.
15055 This switch enables or disables the generation of ISEL instructions.
15057 @item -misel=@var{yes/no}
15058 This switch has been deprecated. Use @option{-misel} and
15059 @option{-mno-isel} instead.
15065 This switch enables or disables the generation of SPE simd
15071 @opindex mno-paired
15072 This switch enables or disables the generation of PAIRED simd
15075 @item -mspe=@var{yes/no}
15076 This option has been deprecated. Use @option{-mspe} and
15077 @option{-mno-spe} instead.
15083 Generate code that uses (does not use) vector/scalar (VSX)
15084 instructions, and also enable the use of built-in functions that allow
15085 more direct access to the VSX instruction set.
15087 @item -mfloat-gprs=@var{yes/single/double/no}
15088 @itemx -mfloat-gprs
15089 @opindex mfloat-gprs
15090 This switch enables or disables the generation of floating point
15091 operations on the general purpose registers for architectures that
15094 The argument @var{yes} or @var{single} enables the use of
15095 single-precision floating point operations.
15097 The argument @var{double} enables the use of single and
15098 double-precision floating point operations.
15100 The argument @var{no} disables floating point operations on the
15101 general purpose registers.
15103 This option is currently only available on the MPC854x.
15109 Generate code for 32-bit or 64-bit environments of Darwin and SVR4
15110 targets (including GNU/Linux). The 32-bit environment sets int, long
15111 and pointer to 32 bits and generates code that runs on any PowerPC
15112 variant. The 64-bit environment sets int to 32 bits and long and
15113 pointer to 64 bits, and generates code for PowerPC64, as for
15114 @option{-mpowerpc64}.
15117 @itemx -mno-fp-in-toc
15118 @itemx -mno-sum-in-toc
15119 @itemx -mminimal-toc
15121 @opindex mno-fp-in-toc
15122 @opindex mno-sum-in-toc
15123 @opindex mminimal-toc
15124 Modify generation of the TOC (Table Of Contents), which is created for
15125 every executable file. The @option{-mfull-toc} option is selected by
15126 default. In that case, GCC will allocate at least one TOC entry for
15127 each unique non-automatic variable reference in your program. GCC
15128 will also place floating-point constants in the TOC@. However, only
15129 16,384 entries are available in the TOC@.
15131 If you receive a linker error message that saying you have overflowed
15132 the available TOC space, you can reduce the amount of TOC space used
15133 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
15134 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
15135 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
15136 generate code to calculate the sum of an address and a constant at
15137 run-time instead of putting that sum into the TOC@. You may specify one
15138 or both of these options. Each causes GCC to produce very slightly
15139 slower and larger code at the expense of conserving TOC space.
15141 If you still run out of space in the TOC even when you specify both of
15142 these options, specify @option{-mminimal-toc} instead. This option causes
15143 GCC to make only one TOC entry for every file. When you specify this
15144 option, GCC will produce code that is slower and larger but which
15145 uses extremely little TOC space. You may wish to use this option
15146 only on files that contain less frequently executed code.
15152 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
15153 @code{long} type, and the infrastructure needed to support them.
15154 Specifying @option{-maix64} implies @option{-mpowerpc64} and
15155 @option{-mpowerpc}, while @option{-maix32} disables the 64-bit ABI and
15156 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
15159 @itemx -mno-xl-compat
15160 @opindex mxl-compat
15161 @opindex mno-xl-compat
15162 Produce code that conforms more closely to IBM XL compiler semantics
15163 when using AIX-compatible ABI@. Pass floating-point arguments to
15164 prototyped functions beyond the register save area (RSA) on the stack
15165 in addition to argument FPRs. Do not assume that most significant
15166 double in 128-bit long double value is properly rounded when comparing
15167 values and converting to double. Use XL symbol names for long double
15170 The AIX calling convention was extended but not initially documented to
15171 handle an obscure K&R C case of calling a function that takes the
15172 address of its arguments with fewer arguments than declared. IBM XL
15173 compilers access floating point arguments which do not fit in the
15174 RSA from the stack when a subroutine is compiled without
15175 optimization. Because always storing floating-point arguments on the
15176 stack is inefficient and rarely needed, this option is not enabled by
15177 default and only is necessary when calling subroutines compiled by IBM
15178 XL compilers without optimization.
15182 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
15183 application written to use message passing with special startup code to
15184 enable the application to run. The system must have PE installed in the
15185 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
15186 must be overridden with the @option{-specs=} option to specify the
15187 appropriate directory location. The Parallel Environment does not
15188 support threads, so the @option{-mpe} option and the @option{-pthread}
15189 option are incompatible.
15191 @item -malign-natural
15192 @itemx -malign-power
15193 @opindex malign-natural
15194 @opindex malign-power
15195 On AIX, 32-bit Darwin, and 64-bit PowerPC GNU/Linux, the option
15196 @option{-malign-natural} overrides the ABI-defined alignment of larger
15197 types, such as floating-point doubles, on their natural size-based boundary.
15198 The option @option{-malign-power} instructs GCC to follow the ABI-specified
15199 alignment rules. GCC defaults to the standard alignment defined in the ABI@.
15201 On 64-bit Darwin, natural alignment is the default, and @option{-malign-power}
15205 @itemx -mhard-float
15206 @opindex msoft-float
15207 @opindex mhard-float
15208 Generate code that does not use (uses) the floating-point register set.
15209 Software floating point emulation is provided if you use the
15210 @option{-msoft-float} option, and pass the option to GCC when linking.
15212 @item -msingle-float
15213 @itemx -mdouble-float
15214 @opindex msingle-float
15215 @opindex mdouble-float
15216 Generate code for single or double-precision floating point operations.
15217 @option{-mdouble-float} implies @option{-msingle-float}.
15220 @opindex msimple-fpu
15221 Do not generate sqrt and div instructions for hardware floating point unit.
15225 Specify type of floating point unit. Valid values are @var{sp_lite}
15226 (equivalent to -msingle-float -msimple-fpu), @var{dp_lite} (equivalent
15227 to -mdouble-float -msimple-fpu), @var{sp_full} (equivalent to -msingle-float),
15228 and @var{dp_full} (equivalent to -mdouble-float).
15231 @opindex mxilinx-fpu
15232 Perform optimizations for floating point unit on Xilinx PPC 405/440.
15235 @itemx -mno-multiple
15237 @opindex mno-multiple
15238 Generate code that uses (does not use) the load multiple word
15239 instructions and the store multiple word instructions. These
15240 instructions are generated by default on POWER systems, and not
15241 generated on PowerPC systems. Do not use @option{-mmultiple} on little
15242 endian PowerPC systems, since those instructions do not work when the
15243 processor is in little endian mode. The exceptions are PPC740 and
15244 PPC750 which permit the instructions usage in little endian mode.
15249 @opindex mno-string
15250 Generate code that uses (does not use) the load string instructions
15251 and the store string word instructions to save multiple registers and
15252 do small block moves. These instructions are generated by default on
15253 POWER systems, and not generated on PowerPC systems. Do not use
15254 @option{-mstring} on little endian PowerPC systems, since those
15255 instructions do not work when the processor is in little endian mode.
15256 The exceptions are PPC740 and PPC750 which permit the instructions
15257 usage in little endian mode.
15262 @opindex mno-update
15263 Generate code that uses (does not use) the load or store instructions
15264 that update the base register to the address of the calculated memory
15265 location. These instructions are generated by default. If you use
15266 @option{-mno-update}, there is a small window between the time that the
15267 stack pointer is updated and the address of the previous frame is
15268 stored, which means code that walks the stack frame across interrupts or
15269 signals may get corrupted data.
15271 @item -mavoid-indexed-addresses
15272 @itemx -mno-avoid-indexed-addresses
15273 @opindex mavoid-indexed-addresses
15274 @opindex mno-avoid-indexed-addresses
15275 Generate code that tries to avoid (not avoid) the use of indexed load
15276 or store instructions. These instructions can incur a performance
15277 penalty on Power6 processors in certain situations, such as when
15278 stepping through large arrays that cross a 16M boundary. This option
15279 is enabled by default when targetting Power6 and disabled otherwise.
15282 @itemx -mno-fused-madd
15283 @opindex mfused-madd
15284 @opindex mno-fused-madd
15285 Generate code that uses (does not use) the floating point multiply and
15286 accumulate instructions. These instructions are generated by default if
15287 hardware floating is used.
15293 Generate code that uses (does not use) the half-word multiply and
15294 multiply-accumulate instructions on the IBM 405, 440, 464 and 476 processors.
15295 These instructions are generated by default when targetting those
15302 Generate code that uses (does not use) the string-search @samp{dlmzb}
15303 instruction on the IBM 405, 440, 464 and 476 processors. This instruction is
15304 generated by default when targetting those processors.
15306 @item -mno-bit-align
15308 @opindex mno-bit-align
15309 @opindex mbit-align
15310 On System V.4 and embedded PowerPC systems do not (do) force structures
15311 and unions that contain bit-fields to be aligned to the base type of the
15314 For example, by default a structure containing nothing but 8
15315 @code{unsigned} bit-fields of length 1 would be aligned to a 4 byte
15316 boundary and have a size of 4 bytes. By using @option{-mno-bit-align},
15317 the structure would be aligned to a 1 byte boundary and be one byte in
15320 @item -mno-strict-align
15321 @itemx -mstrict-align
15322 @opindex mno-strict-align
15323 @opindex mstrict-align
15324 On System V.4 and embedded PowerPC systems do not (do) assume that
15325 unaligned memory references will be handled by the system.
15327 @item -mrelocatable
15328 @itemx -mno-relocatable
15329 @opindex mrelocatable
15330 @opindex mno-relocatable
15331 On embedded PowerPC systems generate code that allows (does not allow)
15332 the program to be relocated to a different address at runtime. If you
15333 use @option{-mrelocatable} on any module, all objects linked together must
15334 be compiled with @option{-mrelocatable} or @option{-mrelocatable-lib}.
15336 @item -mrelocatable-lib
15337 @itemx -mno-relocatable-lib
15338 @opindex mrelocatable-lib
15339 @opindex mno-relocatable-lib
15340 On embedded PowerPC systems generate code that allows (does not allow)
15341 the program to be relocated to a different address at runtime. Modules
15342 compiled with @option{-mrelocatable-lib} can be linked with either modules
15343 compiled without @option{-mrelocatable} and @option{-mrelocatable-lib} or
15344 with modules compiled with the @option{-mrelocatable} options.
15350 On System V.4 and embedded PowerPC systems do not (do) assume that
15351 register 2 contains a pointer to a global area pointing to the addresses
15352 used in the program.
15355 @itemx -mlittle-endian
15357 @opindex mlittle-endian
15358 On System V.4 and embedded PowerPC systems compile code for the
15359 processor in little endian mode. The @option{-mlittle-endian} option is
15360 the same as @option{-mlittle}.
15363 @itemx -mbig-endian
15365 @opindex mbig-endian
15366 On System V.4 and embedded PowerPC systems compile code for the
15367 processor in big endian mode. The @option{-mbig-endian} option is
15368 the same as @option{-mbig}.
15370 @item -mdynamic-no-pic
15371 @opindex mdynamic-no-pic
15372 On Darwin and Mac OS X systems, compile code so that it is not
15373 relocatable, but that its external references are relocatable. The
15374 resulting code is suitable for applications, but not shared
15377 @item -mprioritize-restricted-insns=@var{priority}
15378 @opindex mprioritize-restricted-insns
15379 This option controls the priority that is assigned to
15380 dispatch-slot restricted instructions during the second scheduling
15381 pass. The argument @var{priority} takes the value @var{0/1/2} to assign
15382 @var{no/highest/second-highest} priority to dispatch slot restricted
15385 @item -msched-costly-dep=@var{dependence_type}
15386 @opindex msched-costly-dep
15387 This option controls which dependences are considered costly
15388 by the target during instruction scheduling. The argument
15389 @var{dependence_type} takes one of the following values:
15390 @var{no}: no dependence is costly,
15391 @var{all}: all dependences are costly,
15392 @var{true_store_to_load}: a true dependence from store to load is costly,
15393 @var{store_to_load}: any dependence from store to load is costly,
15394 @var{number}: any dependence which latency >= @var{number} is costly.
15396 @item -minsert-sched-nops=@var{scheme}
15397 @opindex minsert-sched-nops
15398 This option controls which nop insertion scheme will be used during
15399 the second scheduling pass. The argument @var{scheme} takes one of the
15401 @var{no}: Don't insert nops.
15402 @var{pad}: Pad with nops any dispatch group which has vacant issue slots,
15403 according to the scheduler's grouping.
15404 @var{regroup_exact}: Insert nops to force costly dependent insns into
15405 separate groups. Insert exactly as many nops as needed to force an insn
15406 to a new group, according to the estimated processor grouping.
15407 @var{number}: Insert nops to force costly dependent insns into
15408 separate groups. Insert @var{number} nops to force an insn to a new group.
15411 @opindex mcall-sysv
15412 On System V.4 and embedded PowerPC systems compile code using calling
15413 conventions that adheres to the March 1995 draft of the System V
15414 Application Binary Interface, PowerPC processor supplement. This is the
15415 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
15417 @item -mcall-sysv-eabi
15419 @opindex mcall-sysv-eabi
15420 @opindex mcall-eabi
15421 Specify both @option{-mcall-sysv} and @option{-meabi} options.
15423 @item -mcall-sysv-noeabi
15424 @opindex mcall-sysv-noeabi
15425 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
15427 @item -mcall-aixdesc
15429 On System V.4 and embedded PowerPC systems compile code for the AIX
15433 @opindex mcall-linux
15434 On System V.4 and embedded PowerPC systems compile code for the
15435 Linux-based GNU system.
15439 On System V.4 and embedded PowerPC systems compile code for the
15440 Hurd-based GNU system.
15442 @item -mcall-freebsd
15443 @opindex mcall-freebsd
15444 On System V.4 and embedded PowerPC systems compile code for the
15445 FreeBSD operating system.
15447 @item -mcall-netbsd
15448 @opindex mcall-netbsd
15449 On System V.4 and embedded PowerPC systems compile code for the
15450 NetBSD operating system.
15452 @item -mcall-openbsd
15453 @opindex mcall-netbsd
15454 On System V.4 and embedded PowerPC systems compile code for the
15455 OpenBSD operating system.
15457 @item -maix-struct-return
15458 @opindex maix-struct-return
15459 Return all structures in memory (as specified by the AIX ABI)@.
15461 @item -msvr4-struct-return
15462 @opindex msvr4-struct-return
15463 Return structures smaller than 8 bytes in registers (as specified by the
15466 @item -mabi=@var{abi-type}
15468 Extend the current ABI with a particular extension, or remove such extension.
15469 Valid values are @var{altivec}, @var{no-altivec}, @var{spe},
15470 @var{no-spe}, @var{ibmlongdouble}, @var{ieeelongdouble}@.
15474 Extend the current ABI with SPE ABI extensions. This does not change
15475 the default ABI, instead it adds the SPE ABI extensions to the current
15479 @opindex mabi=no-spe
15480 Disable Booke SPE ABI extensions for the current ABI@.
15482 @item -mabi=ibmlongdouble
15483 @opindex mabi=ibmlongdouble
15484 Change the current ABI to use IBM extended precision long double.
15485 This is a PowerPC 32-bit SYSV ABI option.
15487 @item -mabi=ieeelongdouble
15488 @opindex mabi=ieeelongdouble
15489 Change the current ABI to use IEEE extended precision long double.
15490 This is a PowerPC 32-bit Linux ABI option.
15493 @itemx -mno-prototype
15494 @opindex mprototype
15495 @opindex mno-prototype
15496 On System V.4 and embedded PowerPC systems assume that all calls to
15497 variable argument functions are properly prototyped. Otherwise, the
15498 compiler must insert an instruction before every non prototyped call to
15499 set or clear bit 6 of the condition code register (@var{CR}) to
15500 indicate whether floating point values were passed in the floating point
15501 registers in case the function takes a variable arguments. With
15502 @option{-mprototype}, only calls to prototyped variable argument functions
15503 will set or clear the bit.
15507 On embedded PowerPC systems, assume that the startup module is called
15508 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
15509 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}
15514 On embedded PowerPC systems, assume that the startup module is called
15515 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
15520 On embedded PowerPC systems, assume that the startup module is called
15521 @file{crt0.o} and the standard C libraries are @file{libads.a} and
15524 @item -myellowknife
15525 @opindex myellowknife
15526 On embedded PowerPC systems, assume that the startup module is called
15527 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
15532 On System V.4 and embedded PowerPC systems, specify that you are
15533 compiling for a VxWorks system.
15537 On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags
15538 header to indicate that @samp{eabi} extended relocations are used.
15544 On System V.4 and embedded PowerPC systems do (do not) adhere to the
15545 Embedded Applications Binary Interface (eabi) which is a set of
15546 modifications to the System V.4 specifications. Selecting @option{-meabi}
15547 means that the stack is aligned to an 8 byte boundary, a function
15548 @code{__eabi} is called to from @code{main} to set up the eabi
15549 environment, and the @option{-msdata} option can use both @code{r2} and
15550 @code{r13} to point to two separate small data areas. Selecting
15551 @option{-mno-eabi} means that the stack is aligned to a 16 byte boundary,
15552 do not call an initialization function from @code{main}, and the
15553 @option{-msdata} option will only use @code{r13} to point to a single
15554 small data area. The @option{-meabi} option is on by default if you
15555 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
15558 @opindex msdata=eabi
15559 On System V.4 and embedded PowerPC systems, put small initialized
15560 @code{const} global and static data in the @samp{.sdata2} section, which
15561 is pointed to by register @code{r2}. Put small initialized
15562 non-@code{const} global and static data in the @samp{.sdata} section,
15563 which is pointed to by register @code{r13}. Put small uninitialized
15564 global and static data in the @samp{.sbss} section, which is adjacent to
15565 the @samp{.sdata} section. The @option{-msdata=eabi} option is
15566 incompatible with the @option{-mrelocatable} option. The
15567 @option{-msdata=eabi} option also sets the @option{-memb} option.
15570 @opindex msdata=sysv
15571 On System V.4 and embedded PowerPC systems, put small global and static
15572 data in the @samp{.sdata} section, which is pointed to by register
15573 @code{r13}. Put small uninitialized global and static data in the
15574 @samp{.sbss} section, which is adjacent to the @samp{.sdata} section.
15575 The @option{-msdata=sysv} option is incompatible with the
15576 @option{-mrelocatable} option.
15578 @item -msdata=default
15580 @opindex msdata=default
15582 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
15583 compile code the same as @option{-msdata=eabi}, otherwise compile code the
15584 same as @option{-msdata=sysv}.
15587 @opindex msdata=data
15588 On System V.4 and embedded PowerPC systems, put small global
15589 data in the @samp{.sdata} section. Put small uninitialized global
15590 data in the @samp{.sbss} section. Do not use register @code{r13}
15591 to address small data however. This is the default behavior unless
15592 other @option{-msdata} options are used.
15596 @opindex msdata=none
15598 On embedded PowerPC systems, put all initialized global and static data
15599 in the @samp{.data} section, and all uninitialized data in the
15600 @samp{.bss} section.
15604 @cindex smaller data references (PowerPC)
15605 @cindex .sdata/.sdata2 references (PowerPC)
15606 On embedded PowerPC systems, put global and static items less than or
15607 equal to @var{num} bytes into the small data or bss sections instead of
15608 the normal data or bss section. By default, @var{num} is 8. The
15609 @option{-G @var{num}} switch is also passed to the linker.
15610 All modules should be compiled with the same @option{-G @var{num}} value.
15613 @itemx -mno-regnames
15615 @opindex mno-regnames
15616 On System V.4 and embedded PowerPC systems do (do not) emit register
15617 names in the assembly language output using symbolic forms.
15620 @itemx -mno-longcall
15622 @opindex mno-longcall
15623 By default assume that all calls are far away so that a longer more
15624 expensive calling sequence is required. This is required for calls
15625 further than 32 megabytes (33,554,432 bytes) from the current location.
15626 A short call will be generated if the compiler knows
15627 the call cannot be that far away. This setting can be overridden by
15628 the @code{shortcall} function attribute, or by @code{#pragma
15631 Some linkers are capable of detecting out-of-range calls and generating
15632 glue code on the fly. On these systems, long calls are unnecessary and
15633 generate slower code. As of this writing, the AIX linker can do this,
15634 as can the GNU linker for PowerPC/64. It is planned to add this feature
15635 to the GNU linker for 32-bit PowerPC systems as well.
15637 On Darwin/PPC systems, @code{#pragma longcall} will generate ``jbsr
15638 callee, L42'', plus a ``branch island'' (glue code). The two target
15639 addresses represent the callee and the ``branch island''. The
15640 Darwin/PPC linker will prefer the first address and generate a ``bl
15641 callee'' if the PPC ``bl'' instruction will reach the callee directly;
15642 otherwise, the linker will generate ``bl L42'' to call the ``branch
15643 island''. The ``branch island'' is appended to the body of the
15644 calling function; it computes the full 32-bit address of the callee
15647 On Mach-O (Darwin) systems, this option directs the compiler emit to
15648 the glue for every direct call, and the Darwin linker decides whether
15649 to use or discard it.
15651 In the future, we may cause GCC to ignore all longcall specifications
15652 when the linker is known to generate glue.
15654 @item -mtls-markers
15655 @itemx -mno-tls-markers
15656 @opindex mtls-markers
15657 @opindex mno-tls-markers
15658 Mark (do not mark) calls to @code{__tls_get_addr} with a relocation
15659 specifying the function argument. The relocation allows ld to
15660 reliably associate function call with argument setup instructions for
15661 TLS optimization, which in turn allows gcc to better schedule the
15666 Adds support for multithreading with the @dfn{pthreads} library.
15667 This option sets flags for both the preprocessor and linker.
15672 This option will enable GCC to use the reciprocal estimate and
15673 reciprocal square root estimate instructions with additional
15674 Newton-Raphson steps to increase precision instead of doing a divide or
15675 square root and divide for floating point arguments. You should use
15676 the @option{-ffast-math} option when using @option{-mrecip} (or at
15677 least @option{-funsafe-math-optimizations},
15678 @option{-finite-math-only}, @option{-freciprocal-math} and
15679 @option{-fno-trapping-math}). Note that while the throughput of the
15680 sequence is generally higher than the throughput of the non-reciprocal
15681 instruction, the precision of the sequence can be decreased by up to 2
15682 ulp (i.e. the inverse of 1.0 equals 0.99999994) for reciprocal square
15685 @item -mrecip=@var{opt}
15686 @opindex mrecip=opt
15687 This option allows to control which reciprocal estimate instructions
15688 may be used. @var{opt} is a comma separated list of options, that may
15689 be preceeded by a @code{!} to invert the option:
15690 @code{all}: enable all estimate instructions,
15691 @code{default}: enable the default instructions, equvalent to @option{-mrecip},
15692 @code{none}: disable all estimate instructions, equivalent to @option{-mno-recip};
15693 @code{div}: enable the reciprocal approximation instructions for both single and double precision;
15694 @code{divf}: enable the single precision reciprocal approximation instructions;
15695 @code{divd}: enable the double precision reciprocal approximation instructions;
15696 @code{rsqrt}: enable the reciprocal square root approximation instructions for both single and double precision;
15697 @code{rsqrtf}: enable the single precision reciprocal square root approximation instructions;
15698 @code{rsqrtd}: enable the double precision reciprocal square root approximation instructions;
15700 So for example, @option{-mrecip=all,!rsqrtd} would enable the
15701 all of the reciprocal estimate instructions, except for the
15702 @code{FRSQRTE}, @code{XSRSQRTEDP}, and @code{XVRSQRTEDP} instructions
15703 which handle the double precision reciprocal square root calculations.
15705 @item -mrecip-precision
15706 @itemx -mno-recip-precision
15707 @opindex mrecip-precision
15708 Assume (do not assume) that the reciprocal estimate instructions
15709 provide higher precision estimates than is mandated by the powerpc
15710 ABI. Selecting @option{-mcpu=power6} or @option{-mcpu=power7}
15711 automatically selects @option{-mrecip-precision}. The double
15712 precision square root estimate instructions are not generated by
15713 default on low precision machines, since they do not provide an
15714 estimate that converges after three steps.
15718 @subsection RX Options
15721 These command line options are defined for RX targets:
15724 @item -m64bit-doubles
15725 @itemx -m32bit-doubles
15726 @opindex m64bit-doubles
15727 @opindex m32bit-doubles
15728 Make the @code{double} data type be 64-bits (@option{-m64bit-doubles})
15729 or 32-bits (@option{-m32bit-doubles}) in size. The default is
15730 @option{-m32bit-doubles}. @emph{Note} RX floating point hardware only
15731 works on 32-bit values, which is why the default is
15732 @option{-m32bit-doubles}.
15738 Enables (@option{-fpu}) or disables (@option{-nofpu}) the use of RX
15739 floating point hardware. The default is enabled for the @var{RX600}
15740 series and disabled for the @var{RX200} series.
15742 Floating point instructions will only be generated for 32-bit floating
15743 point values however, so if the @option{-m64bit-doubles} option is in
15744 use then the FPU hardware will not be used for doubles.
15746 @emph{Note} If the @option{-fpu} option is enabled then
15747 @option{-funsafe-math-optimizations} is also enabled automatically.
15748 This is because the RX FPU instructions are themselves unsafe.
15750 @item -mcpu=@var{name}
15751 @itemx -patch=@var{name}
15754 Selects the type of RX CPU to be targeted. Currently three types are
15755 supported, the generic @var{RX600} and @var{RX200} series hardware and
15756 the specific @var{RX610} cpu. The default is @var{RX600}.
15758 The only difference between @var{RX600} and @var{RX610} is that the
15759 @var{RX610} does not support the @code{MVTIPL} instruction.
15761 The @var{RX200} series does not have a hardware floating point unit
15762 and so @option{-nofpu} is enabled by default when this type is
15765 @item -mbig-endian-data
15766 @itemx -mlittle-endian-data
15767 @opindex mbig-endian-data
15768 @opindex mlittle-endian-data
15769 Store data (but not code) in the big-endian format. The default is
15770 @option{-mlittle-endian-data}, ie to store data in the little endian
15773 @item -msmall-data-limit=@var{N}
15774 @opindex msmall-data-limit
15775 Specifies the maximum size in bytes of global and static variables
15776 which can be placed into the small data area. Using the small data
15777 area can lead to smaller and faster code, but the size of area is
15778 limited and it is up to the programmer to ensure that the area does
15779 not overflow. Also when the small data area is used one of the RX's
15780 registers (@code{r13}) is reserved for use pointing to this area, so
15781 it is no longer available for use by the compiler. This could result
15782 in slower and/or larger code if variables which once could have been
15783 held in @code{r13} are now pushed onto the stack.
15785 Note, common variables (variables which have not been initialised) and
15786 constants are not placed into the small data area as they are assigned
15787 to other sections in the output executable.
15789 The default value is zero, which disables this feature. Note, this
15790 feature is not enabled by default with higher optimization levels
15791 (@option{-O2} etc) because of the potentially detrimental effects of
15792 reserving register @code{r13}. It is up to the programmer to
15793 experiment and discover whether this feature is of benefit to their
15800 Use the simulator runtime. The default is to use the libgloss board
15803 @item -mas100-syntax
15804 @itemx -mno-as100-syntax
15805 @opindex mas100-syntax
15806 @opindex mno-as100-syntax
15807 When generating assembler output use a syntax that is compatible with
15808 Renesas's AS100 assembler. This syntax can also be handled by the GAS
15809 assembler but it has some restrictions so generating it is not the
15812 @item -mmax-constant-size=@var{N}
15813 @opindex mmax-constant-size
15814 Specifies the maximum size, in bytes, of a constant that can be used as
15815 an operand in a RX instruction. Although the RX instruction set does
15816 allow constants of up to 4 bytes in length to be used in instructions,
15817 a longer value equates to a longer instruction. Thus in some
15818 circumstances it can be beneficial to restrict the size of constants
15819 that are used in instructions. Constants that are too big are instead
15820 placed into a constant pool and referenced via register indirection.
15822 The value @var{N} can be between 0 and 4. A value of 0 (the default)
15823 or 4 means that constants of any size are allowed.
15827 Enable linker relaxation. Linker relaxation is a process whereby the
15828 linker will attempt to reduce the size of a program by finding shorter
15829 versions of various instructions. Disabled by default.
15831 @item -mint-register=@var{N}
15832 @opindex mint-register
15833 Specify the number of registers to reserve for fast interrupt handler
15834 functions. The value @var{N} can be between 0 and 4. A value of 1
15835 means that register @code{r13} will be reserved for the exclusive use
15836 of fast interrupt handlers. A value of 2 reserves @code{r13} and
15837 @code{r12}. A value of 3 reserves @code{r13}, @code{r12} and
15838 @code{r11}, and a value of 4 reserves @code{r13} through @code{r10}.
15839 A value of 0, the default, does not reserve any registers.
15841 @item -msave-acc-in-interrupts
15842 @opindex msave-acc-in-interrupts
15843 Specifies that interrupt handler functions should preserve the
15844 accumulator register. This is only necessary if normal code might use
15845 the accumulator register, for example because it performs 64-bit
15846 multiplications. The default is to ignore the accumulator as this
15847 makes the interrupt handlers faster.
15851 @emph{Note:} The generic GCC command line @option{-ffixed-@var{reg}}
15852 has special significance to the RX port when used with the
15853 @code{interrupt} function attribute. This attribute indicates a
15854 function intended to process fast interrupts. GCC will will ensure
15855 that it only uses the registers @code{r10}, @code{r11}, @code{r12}
15856 and/or @code{r13} and only provided that the normal use of the
15857 corresponding registers have been restricted via the
15858 @option{-ffixed-@var{reg}} or @option{-mint-register} command line
15861 @node S/390 and zSeries Options
15862 @subsection S/390 and zSeries Options
15863 @cindex S/390 and zSeries Options
15865 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
15869 @itemx -msoft-float
15870 @opindex mhard-float
15871 @opindex msoft-float
15872 Use (do not use) the hardware floating-point instructions and registers
15873 for floating-point operations. When @option{-msoft-float} is specified,
15874 functions in @file{libgcc.a} will be used to perform floating-point
15875 operations. When @option{-mhard-float} is specified, the compiler
15876 generates IEEE floating-point instructions. This is the default.
15879 @itemx -mno-hard-dfp
15881 @opindex mno-hard-dfp
15882 Use (do not use) the hardware decimal-floating-point instructions for
15883 decimal-floating-point operations. When @option{-mno-hard-dfp} is
15884 specified, functions in @file{libgcc.a} will be used to perform
15885 decimal-floating-point operations. When @option{-mhard-dfp} is
15886 specified, the compiler generates decimal-floating-point hardware
15887 instructions. This is the default for @option{-march=z9-ec} or higher.
15889 @item -mlong-double-64
15890 @itemx -mlong-double-128
15891 @opindex mlong-double-64
15892 @opindex mlong-double-128
15893 These switches control the size of @code{long double} type. A size
15894 of 64bit makes the @code{long double} type equivalent to the @code{double}
15895 type. This is the default.
15898 @itemx -mno-backchain
15899 @opindex mbackchain
15900 @opindex mno-backchain
15901 Store (do not store) the address of the caller's frame as backchain pointer
15902 into the callee's stack frame.
15903 A backchain may be needed to allow debugging using tools that do not understand
15904 DWARF-2 call frame information.
15905 When @option{-mno-packed-stack} is in effect, the backchain pointer is stored
15906 at the bottom of the stack frame; when @option{-mpacked-stack} is in effect,
15907 the backchain is placed into the topmost word of the 96/160 byte register
15910 In general, code compiled with @option{-mbackchain} is call-compatible with
15911 code compiled with @option{-mmo-backchain}; however, use of the backchain
15912 for debugging purposes usually requires that the whole binary is built with
15913 @option{-mbackchain}. Note that the combination of @option{-mbackchain},
15914 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
15915 to build a linux kernel use @option{-msoft-float}.
15917 The default is to not maintain the backchain.
15919 @item -mpacked-stack
15920 @itemx -mno-packed-stack
15921 @opindex mpacked-stack
15922 @opindex mno-packed-stack
15923 Use (do not use) the packed stack layout. When @option{-mno-packed-stack} is
15924 specified, the compiler uses the all fields of the 96/160 byte register save
15925 area only for their default purpose; unused fields still take up stack space.
15926 When @option{-mpacked-stack} is specified, register save slots are densely
15927 packed at the top of the register save area; unused space is reused for other
15928 purposes, allowing for more efficient use of the available stack space.
15929 However, when @option{-mbackchain} is also in effect, the topmost word of
15930 the save area is always used to store the backchain, and the return address
15931 register is always saved two words below the backchain.
15933 As long as the stack frame backchain is not used, code generated with
15934 @option{-mpacked-stack} is call-compatible with code generated with
15935 @option{-mno-packed-stack}. Note that some non-FSF releases of GCC 2.95 for
15936 S/390 or zSeries generated code that uses the stack frame backchain at run
15937 time, not just for debugging purposes. Such code is not call-compatible
15938 with code compiled with @option{-mpacked-stack}. Also, note that the
15939 combination of @option{-mbackchain},
15940 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
15941 to build a linux kernel use @option{-msoft-float}.
15943 The default is to not use the packed stack layout.
15946 @itemx -mno-small-exec
15947 @opindex msmall-exec
15948 @opindex mno-small-exec
15949 Generate (or do not generate) code using the @code{bras} instruction
15950 to do subroutine calls.
15951 This only works reliably if the total executable size does not
15952 exceed 64k. The default is to use the @code{basr} instruction instead,
15953 which does not have this limitation.
15959 When @option{-m31} is specified, generate code compliant to the
15960 GNU/Linux for S/390 ABI@. When @option{-m64} is specified, generate
15961 code compliant to the GNU/Linux for zSeries ABI@. This allows GCC in
15962 particular to generate 64-bit instructions. For the @samp{s390}
15963 targets, the default is @option{-m31}, while the @samp{s390x}
15964 targets default to @option{-m64}.
15970 When @option{-mzarch} is specified, generate code using the
15971 instructions available on z/Architecture.
15972 When @option{-mesa} is specified, generate code using the
15973 instructions available on ESA/390. Note that @option{-mesa} is
15974 not possible with @option{-m64}.
15975 When generating code compliant to the GNU/Linux for S/390 ABI,
15976 the default is @option{-mesa}. When generating code compliant
15977 to the GNU/Linux for zSeries ABI, the default is @option{-mzarch}.
15983 Generate (or do not generate) code using the @code{mvcle} instruction
15984 to perform block moves. When @option{-mno-mvcle} is specified,
15985 use a @code{mvc} loop instead. This is the default unless optimizing for
15992 Print (or do not print) additional debug information when compiling.
15993 The default is to not print debug information.
15995 @item -march=@var{cpu-type}
15997 Generate code that will run on @var{cpu-type}, which is the name of a system
15998 representing a certain processor type. Possible values for
15999 @var{cpu-type} are @samp{g5}, @samp{g6}, @samp{z900}, @samp{z990},
16000 @samp{z9-109}, @samp{z9-ec} and @samp{z10}.
16001 When generating code using the instructions available on z/Architecture,
16002 the default is @option{-march=z900}. Otherwise, the default is
16003 @option{-march=g5}.
16005 @item -mtune=@var{cpu-type}
16007 Tune to @var{cpu-type} everything applicable about the generated code,
16008 except for the ABI and the set of available instructions.
16009 The list of @var{cpu-type} values is the same as for @option{-march}.
16010 The default is the value used for @option{-march}.
16013 @itemx -mno-tpf-trace
16014 @opindex mtpf-trace
16015 @opindex mno-tpf-trace
16016 Generate code that adds (does not add) in TPF OS specific branches to trace
16017 routines in the operating system. This option is off by default, even
16018 when compiling for the TPF OS@.
16021 @itemx -mno-fused-madd
16022 @opindex mfused-madd
16023 @opindex mno-fused-madd
16024 Generate code that uses (does not use) the floating point multiply and
16025 accumulate instructions. These instructions are generated by default if
16026 hardware floating point is used.
16028 @item -mwarn-framesize=@var{framesize}
16029 @opindex mwarn-framesize
16030 Emit a warning if the current function exceeds the given frame size. Because
16031 this is a compile time check it doesn't need to be a real problem when the program
16032 runs. It is intended to identify functions which most probably cause
16033 a stack overflow. It is useful to be used in an environment with limited stack
16034 size e.g.@: the linux kernel.
16036 @item -mwarn-dynamicstack
16037 @opindex mwarn-dynamicstack
16038 Emit a warning if the function calls alloca or uses dynamically
16039 sized arrays. This is generally a bad idea with a limited stack size.
16041 @item -mstack-guard=@var{stack-guard}
16042 @itemx -mstack-size=@var{stack-size}
16043 @opindex mstack-guard
16044 @opindex mstack-size
16045 If these options are provided the s390 back end emits additional instructions in
16046 the function prologue which trigger a trap if the stack size is @var{stack-guard}
16047 bytes above the @var{stack-size} (remember that the stack on s390 grows downward).
16048 If the @var{stack-guard} option is omitted the smallest power of 2 larger than
16049 the frame size of the compiled function is chosen.
16050 These options are intended to be used to help debugging stack overflow problems.
16051 The additionally emitted code causes only little overhead and hence can also be
16052 used in production like systems without greater performance degradation. The given
16053 values have to be exact powers of 2 and @var{stack-size} has to be greater than
16054 @var{stack-guard} without exceeding 64k.
16055 In order to be efficient the extra code makes the assumption that the stack starts
16056 at an address aligned to the value given by @var{stack-size}.
16057 The @var{stack-guard} option can only be used in conjunction with @var{stack-size}.
16060 @node Score Options
16061 @subsection Score Options
16062 @cindex Score Options
16064 These options are defined for Score implementations:
16069 Compile code for big endian mode. This is the default.
16073 Compile code for little endian mode.
16077 Disable generate bcnz instruction.
16081 Enable generate unaligned load and store instruction.
16085 Enable the use of multiply-accumulate instructions. Disabled by default.
16089 Specify the SCORE5 as the target architecture.
16093 Specify the SCORE5U of the target architecture.
16097 Specify the SCORE7 as the target architecture. This is the default.
16101 Specify the SCORE7D as the target architecture.
16105 @subsection SH Options
16107 These @samp{-m} options are defined for the SH implementations:
16112 Generate code for the SH1.
16116 Generate code for the SH2.
16119 Generate code for the SH2e.
16123 Generate code for the SH2a without FPU, or for a SH2a-FPU in such a way
16124 that the floating-point unit is not used.
16126 @item -m2a-single-only
16127 @opindex m2a-single-only
16128 Generate code for the SH2a-FPU, in such a way that no double-precision
16129 floating point operations are used.
16132 @opindex m2a-single
16133 Generate code for the SH2a-FPU assuming the floating-point unit is in
16134 single-precision mode by default.
16138 Generate code for the SH2a-FPU assuming the floating-point unit is in
16139 double-precision mode by default.
16143 Generate code for the SH3.
16147 Generate code for the SH3e.
16151 Generate code for the SH4 without a floating-point unit.
16153 @item -m4-single-only
16154 @opindex m4-single-only
16155 Generate code for the SH4 with a floating-point unit that only
16156 supports single-precision arithmetic.
16160 Generate code for the SH4 assuming the floating-point unit is in
16161 single-precision mode by default.
16165 Generate code for the SH4.
16169 Generate code for the SH4al-dsp, or for a SH4a in such a way that the
16170 floating-point unit is not used.
16172 @item -m4a-single-only
16173 @opindex m4a-single-only
16174 Generate code for the SH4a, in such a way that no double-precision
16175 floating point operations are used.
16178 @opindex m4a-single
16179 Generate code for the SH4a assuming the floating-point unit is in
16180 single-precision mode by default.
16184 Generate code for the SH4a.
16188 Same as @option{-m4a-nofpu}, except that it implicitly passes
16189 @option{-dsp} to the assembler. GCC doesn't generate any DSP
16190 instructions at the moment.
16194 Compile code for the processor in big endian mode.
16198 Compile code for the processor in little endian mode.
16202 Align doubles at 64-bit boundaries. Note that this changes the calling
16203 conventions, and thus some functions from the standard C library will
16204 not work unless you recompile it first with @option{-mdalign}.
16208 Shorten some address references at link time, when possible; uses the
16209 linker option @option{-relax}.
16213 Use 32-bit offsets in @code{switch} tables. The default is to use
16218 Enable the use of bit manipulation instructions on SH2A.
16222 Enable the use of the instruction @code{fmovd}. Check @option{-mdalign} for
16223 alignment constraints.
16227 Comply with the calling conventions defined by Renesas.
16231 Comply with the calling conventions defined by Renesas.
16235 Comply with the calling conventions defined for GCC before the Renesas
16236 conventions were available. This option is the default for all
16237 targets of the SH toolchain except for @samp{sh-symbianelf}.
16240 @opindex mnomacsave
16241 Mark the @code{MAC} register as call-clobbered, even if
16242 @option{-mhitachi} is given.
16246 Increase IEEE-compliance of floating-point code.
16247 At the moment, this is equivalent to @option{-fno-finite-math-only}.
16248 When generating 16 bit SH opcodes, getting IEEE-conforming results for
16249 comparisons of NANs / infinities incurs extra overhead in every
16250 floating point comparison, therefore the default is set to
16251 @option{-ffinite-math-only}.
16253 @item -minline-ic_invalidate
16254 @opindex minline-ic_invalidate
16255 Inline code to invalidate instruction cache entries after setting up
16256 nested function trampolines.
16257 This option has no effect if -musermode is in effect and the selected
16258 code generation option (e.g. -m4) does not allow the use of the icbi
16260 If the selected code generation option does not allow the use of the icbi
16261 instruction, and -musermode is not in effect, the inlined code will
16262 manipulate the instruction cache address array directly with an associative
16263 write. This not only requires privileged mode, but it will also
16264 fail if the cache line had been mapped via the TLB and has become unmapped.
16268 Dump instruction size and location in the assembly code.
16271 @opindex mpadstruct
16272 This option is deprecated. It pads structures to multiple of 4 bytes,
16273 which is incompatible with the SH ABI@.
16277 Optimize for space instead of speed. Implied by @option{-Os}.
16280 @opindex mprefergot
16281 When generating position-independent code, emit function calls using
16282 the Global Offset Table instead of the Procedure Linkage Table.
16286 Don't generate privileged mode only code; implies -mno-inline-ic_invalidate
16287 if the inlined code would not work in user mode.
16288 This is the default when the target is @code{sh-*-linux*}.
16290 @item -multcost=@var{number}
16291 @opindex multcost=@var{number}
16292 Set the cost to assume for a multiply insn.
16294 @item -mdiv=@var{strategy}
16295 @opindex mdiv=@var{strategy}
16296 Set the division strategy to use for SHmedia code. @var{strategy} must be
16297 one of: call, call2, fp, inv, inv:minlat, inv20u, inv20l, inv:call,
16298 inv:call2, inv:fp .
16299 "fp" performs the operation in floating point. This has a very high latency,
16300 but needs only a few instructions, so it might be a good choice if
16301 your code has enough easily exploitable ILP to allow the compiler to
16302 schedule the floating point instructions together with other instructions.
16303 Division by zero causes a floating point exception.
16304 "inv" uses integer operations to calculate the inverse of the divisor,
16305 and then multiplies the dividend with the inverse. This strategy allows
16306 cse and hoisting of the inverse calculation. Division by zero calculates
16307 an unspecified result, but does not trap.
16308 "inv:minlat" is a variant of "inv" where if no cse / hoisting opportunities
16309 have been found, or if the entire operation has been hoisted to the same
16310 place, the last stages of the inverse calculation are intertwined with the
16311 final multiply to reduce the overall latency, at the expense of using a few
16312 more instructions, and thus offering fewer scheduling opportunities with
16314 "call" calls a library function that usually implements the inv:minlat
16316 This gives high code density for m5-*media-nofpu compilations.
16317 "call2" uses a different entry point of the same library function, where it
16318 assumes that a pointer to a lookup table has already been set up, which
16319 exposes the pointer load to cse / code hoisting optimizations.
16320 "inv:call", "inv:call2" and "inv:fp" all use the "inv" algorithm for initial
16321 code generation, but if the code stays unoptimized, revert to the "call",
16322 "call2", or "fp" strategies, respectively. Note that the
16323 potentially-trapping side effect of division by zero is carried by a
16324 separate instruction, so it is possible that all the integer instructions
16325 are hoisted out, but the marker for the side effect stays where it is.
16326 A recombination to fp operations or a call is not possible in that case.
16327 "inv20u" and "inv20l" are variants of the "inv:minlat" strategy. In the case
16328 that the inverse calculation was nor separated from the multiply, they speed
16329 up division where the dividend fits into 20 bits (plus sign where applicable),
16330 by inserting a test to skip a number of operations in this case; this test
16331 slows down the case of larger dividends. inv20u assumes the case of a such
16332 a small dividend to be unlikely, and inv20l assumes it to be likely.
16334 @item -maccumulate-outgoing-args
16335 @opindex maccumulate-outgoing-args
16336 Reserve space once for outgoing arguments in the function prologue rather
16337 than around each call. Generally beneficial for performance and size. Also
16338 needed for unwinding to avoid changing the stack frame around conditional code.
16340 @item -mdivsi3_libfunc=@var{name}
16341 @opindex mdivsi3_libfunc=@var{name}
16342 Set the name of the library function used for 32 bit signed division to
16343 @var{name}. This only affect the name used in the call and inv:call
16344 division strategies, and the compiler will still expect the same
16345 sets of input/output/clobbered registers as if this option was not present.
16347 @item -mfixed-range=@var{register-range}
16348 @opindex mfixed-range
16349 Generate code treating the given register range as fixed registers.
16350 A fixed register is one that the register allocator can not use. This is
16351 useful when compiling kernel code. A register range is specified as
16352 two registers separated by a dash. Multiple register ranges can be
16353 specified separated by a comma.
16355 @item -madjust-unroll
16356 @opindex madjust-unroll
16357 Throttle unrolling to avoid thrashing target registers.
16358 This option only has an effect if the gcc code base supports the
16359 TARGET_ADJUST_UNROLL_MAX target hook.
16361 @item -mindexed-addressing
16362 @opindex mindexed-addressing
16363 Enable the use of the indexed addressing mode for SHmedia32/SHcompact.
16364 This is only safe if the hardware and/or OS implement 32 bit wrap-around
16365 semantics for the indexed addressing mode. The architecture allows the
16366 implementation of processors with 64 bit MMU, which the OS could use to
16367 get 32 bit addressing, but since no current hardware implementation supports
16368 this or any other way to make the indexed addressing mode safe to use in
16369 the 32 bit ABI, the default is -mno-indexed-addressing.
16371 @item -mgettrcost=@var{number}
16372 @opindex mgettrcost=@var{number}
16373 Set the cost assumed for the gettr instruction to @var{number}.
16374 The default is 2 if @option{-mpt-fixed} is in effect, 100 otherwise.
16378 Assume pt* instructions won't trap. This will generally generate better
16379 scheduled code, but is unsafe on current hardware. The current architecture
16380 definition says that ptabs and ptrel trap when the target anded with 3 is 3.
16381 This has the unintentional effect of making it unsafe to schedule ptabs /
16382 ptrel before a branch, or hoist it out of a loop. For example,
16383 __do_global_ctors, a part of libgcc that runs constructors at program
16384 startup, calls functions in a list which is delimited by @minus{}1. With the
16385 -mpt-fixed option, the ptabs will be done before testing against @minus{}1.
16386 That means that all the constructors will be run a bit quicker, but when
16387 the loop comes to the end of the list, the program crashes because ptabs
16388 loads @minus{}1 into a target register. Since this option is unsafe for any
16389 hardware implementing the current architecture specification, the default
16390 is -mno-pt-fixed. Unless the user specifies a specific cost with
16391 @option{-mgettrcost}, -mno-pt-fixed also implies @option{-mgettrcost=100};
16392 this deters register allocation using target registers for storing
16395 @item -minvalid-symbols
16396 @opindex minvalid-symbols
16397 Assume symbols might be invalid. Ordinary function symbols generated by
16398 the compiler will always be valid to load with movi/shori/ptabs or
16399 movi/shori/ptrel, but with assembler and/or linker tricks it is possible
16400 to generate symbols that will cause ptabs / ptrel to trap.
16401 This option is only meaningful when @option{-mno-pt-fixed} is in effect.
16402 It will then prevent cross-basic-block cse, hoisting and most scheduling
16403 of symbol loads. The default is @option{-mno-invalid-symbols}.
16406 @node SPARC Options
16407 @subsection SPARC Options
16408 @cindex SPARC options
16410 These @samp{-m} options are supported on the SPARC:
16413 @item -mno-app-regs
16415 @opindex mno-app-regs
16417 Specify @option{-mapp-regs} to generate output using the global registers
16418 2 through 4, which the SPARC SVR4 ABI reserves for applications. This
16421 To be fully SVR4 ABI compliant at the cost of some performance loss,
16422 specify @option{-mno-app-regs}. You should compile libraries and system
16423 software with this option.
16426 @itemx -mhard-float
16428 @opindex mhard-float
16429 Generate output containing floating point instructions. This is the
16433 @itemx -msoft-float
16435 @opindex msoft-float
16436 Generate output containing library calls for floating point.
16437 @strong{Warning:} the requisite libraries are not available for all SPARC
16438 targets. Normally the facilities of the machine's usual C compiler are
16439 used, but this cannot be done directly in cross-compilation. You must make
16440 your own arrangements to provide suitable library functions for
16441 cross-compilation. The embedded targets @samp{sparc-*-aout} and
16442 @samp{sparclite-*-*} do provide software floating point support.
16444 @option{-msoft-float} changes the calling convention in the output file;
16445 therefore, it is only useful if you compile @emph{all} of a program with
16446 this option. In particular, you need to compile @file{libgcc.a}, the
16447 library that comes with GCC, with @option{-msoft-float} in order for
16450 @item -mhard-quad-float
16451 @opindex mhard-quad-float
16452 Generate output containing quad-word (long double) floating point
16455 @item -msoft-quad-float
16456 @opindex msoft-quad-float
16457 Generate output containing library calls for quad-word (long double)
16458 floating point instructions. The functions called are those specified
16459 in the SPARC ABI@. This is the default.
16461 As of this writing, there are no SPARC implementations that have hardware
16462 support for the quad-word floating point instructions. They all invoke
16463 a trap handler for one of these instructions, and then the trap handler
16464 emulates the effect of the instruction. Because of the trap handler overhead,
16465 this is much slower than calling the ABI library routines. Thus the
16466 @option{-msoft-quad-float} option is the default.
16468 @item -mno-unaligned-doubles
16469 @itemx -munaligned-doubles
16470 @opindex mno-unaligned-doubles
16471 @opindex munaligned-doubles
16472 Assume that doubles have 8 byte alignment. This is the default.
16474 With @option{-munaligned-doubles}, GCC assumes that doubles have 8 byte
16475 alignment only if they are contained in another type, or if they have an
16476 absolute address. Otherwise, it assumes they have 4 byte alignment.
16477 Specifying this option avoids some rare compatibility problems with code
16478 generated by other compilers. It is not the default because it results
16479 in a performance loss, especially for floating point code.
16481 @item -mno-faster-structs
16482 @itemx -mfaster-structs
16483 @opindex mno-faster-structs
16484 @opindex mfaster-structs
16485 With @option{-mfaster-structs}, the compiler assumes that structures
16486 should have 8 byte alignment. This enables the use of pairs of
16487 @code{ldd} and @code{std} instructions for copies in structure
16488 assignment, in place of twice as many @code{ld} and @code{st} pairs.
16489 However, the use of this changed alignment directly violates the SPARC
16490 ABI@. Thus, it's intended only for use on targets where the developer
16491 acknowledges that their resulting code will not be directly in line with
16492 the rules of the ABI@.
16494 @item -mimpure-text
16495 @opindex mimpure-text
16496 @option{-mimpure-text}, used in addition to @option{-shared}, tells
16497 the compiler to not pass @option{-z text} to the linker when linking a
16498 shared object. Using this option, you can link position-dependent
16499 code into a shared object.
16501 @option{-mimpure-text} suppresses the ``relocations remain against
16502 allocatable but non-writable sections'' linker error message.
16503 However, the necessary relocations will trigger copy-on-write, and the
16504 shared object is not actually shared across processes. Instead of
16505 using @option{-mimpure-text}, you should compile all source code with
16506 @option{-fpic} or @option{-fPIC}.
16508 This option is only available on SunOS and Solaris.
16510 @item -mcpu=@var{cpu_type}
16512 Set the instruction set, register set, and instruction scheduling parameters
16513 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
16514 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite},
16515 @samp{f930}, @samp{f934}, @samp{hypersparc}, @samp{sparclite86x},
16516 @samp{sparclet}, @samp{tsc701}, @samp{v9}, @samp{ultrasparc},
16517 @samp{ultrasparc3}, @samp{niagara} and @samp{niagara2}.
16519 Default instruction scheduling parameters are used for values that select
16520 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
16521 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
16523 Here is a list of each supported architecture and their supported
16528 v8: supersparc, hypersparc
16529 sparclite: f930, f934, sparclite86x
16531 v9: ultrasparc, ultrasparc3, niagara, niagara2
16534 By default (unless configured otherwise), GCC generates code for the V7
16535 variant of the SPARC architecture. With @option{-mcpu=cypress}, the compiler
16536 additionally optimizes it for the Cypress CY7C602 chip, as used in the
16537 SPARCStation/SPARCServer 3xx series. This is also appropriate for the older
16538 SPARCStation 1, 2, IPX etc.
16540 With @option{-mcpu=v8}, GCC generates code for the V8 variant of the SPARC
16541 architecture. The only difference from V7 code is that the compiler emits
16542 the integer multiply and integer divide instructions which exist in SPARC-V8
16543 but not in SPARC-V7. With @option{-mcpu=supersparc}, the compiler additionally
16544 optimizes it for the SuperSPARC chip, as used in the SPARCStation 10, 1000 and
16547 With @option{-mcpu=sparclite}, GCC generates code for the SPARClite variant of
16548 the SPARC architecture. This adds the integer multiply, integer divide step
16549 and scan (@code{ffs}) instructions which exist in SPARClite but not in SPARC-V7.
16550 With @option{-mcpu=f930}, the compiler additionally optimizes it for the
16551 Fujitsu MB86930 chip, which is the original SPARClite, with no FPU@. With
16552 @option{-mcpu=f934}, the compiler additionally optimizes it for the Fujitsu
16553 MB86934 chip, which is the more recent SPARClite with FPU@.
16555 With @option{-mcpu=sparclet}, GCC generates code for the SPARClet variant of
16556 the SPARC architecture. This adds the integer multiply, multiply/accumulate,
16557 integer divide step and scan (@code{ffs}) instructions which exist in SPARClet
16558 but not in SPARC-V7. With @option{-mcpu=tsc701}, the compiler additionally
16559 optimizes it for the TEMIC SPARClet chip.
16561 With @option{-mcpu=v9}, GCC generates code for the V9 variant of the SPARC
16562 architecture. This adds 64-bit integer and floating-point move instructions,
16563 3 additional floating-point condition code registers and conditional move
16564 instructions. With @option{-mcpu=ultrasparc}, the compiler additionally
16565 optimizes it for the Sun UltraSPARC I/II/IIi chips. With
16566 @option{-mcpu=ultrasparc3}, the compiler additionally optimizes it for the
16567 Sun UltraSPARC III/III+/IIIi/IIIi+/IV/IV+ chips. With
16568 @option{-mcpu=niagara}, the compiler additionally optimizes it for
16569 Sun UltraSPARC T1 chips. With @option{-mcpu=niagara2}, the compiler
16570 additionally optimizes it for Sun UltraSPARC T2 chips.
16572 @item -mtune=@var{cpu_type}
16574 Set the instruction scheduling parameters for machine type
16575 @var{cpu_type}, but do not set the instruction set or register set that the
16576 option @option{-mcpu=@var{cpu_type}} would.
16578 The same values for @option{-mcpu=@var{cpu_type}} can be used for
16579 @option{-mtune=@var{cpu_type}}, but the only useful values are those
16580 that select a particular cpu implementation. Those are @samp{cypress},
16581 @samp{supersparc}, @samp{hypersparc}, @samp{f930}, @samp{f934},
16582 @samp{sparclite86x}, @samp{tsc701}, @samp{ultrasparc},
16583 @samp{ultrasparc3}, @samp{niagara}, and @samp{niagara2}.
16588 @opindex mno-v8plus
16589 With @option{-mv8plus}, GCC generates code for the SPARC-V8+ ABI@. The
16590 difference from the V8 ABI is that the global and out registers are
16591 considered 64-bit wide. This is enabled by default on Solaris in 32-bit
16592 mode for all SPARC-V9 processors.
16598 With @option{-mvis}, GCC generates code that takes advantage of the UltraSPARC
16599 Visual Instruction Set extensions. The default is @option{-mno-vis}.
16602 These @samp{-m} options are supported in addition to the above
16603 on SPARC-V9 processors in 64-bit environments:
16606 @item -mlittle-endian
16607 @opindex mlittle-endian
16608 Generate code for a processor running in little-endian mode. It is only
16609 available for a few configurations and most notably not on Solaris and Linux.
16615 Generate code for a 32-bit or 64-bit environment.
16616 The 32-bit environment sets int, long and pointer to 32 bits.
16617 The 64-bit environment sets int to 32 bits and long and pointer
16620 @item -mcmodel=medlow
16621 @opindex mcmodel=medlow
16622 Generate code for the Medium/Low code model: 64-bit addresses, programs
16623 must be linked in the low 32 bits of memory. Programs can be statically
16624 or dynamically linked.
16626 @item -mcmodel=medmid
16627 @opindex mcmodel=medmid
16628 Generate code for the Medium/Middle code model: 64-bit addresses, programs
16629 must be linked in the low 44 bits of memory, the text and data segments must
16630 be less than 2GB in size and the data segment must be located within 2GB of
16633 @item -mcmodel=medany
16634 @opindex mcmodel=medany
16635 Generate code for the Medium/Anywhere code model: 64-bit addresses, programs
16636 may be linked anywhere in memory, the text and data segments must be less
16637 than 2GB in size and the data segment must be located within 2GB of the
16640 @item -mcmodel=embmedany
16641 @opindex mcmodel=embmedany
16642 Generate code for the Medium/Anywhere code model for embedded systems:
16643 64-bit addresses, the text and data segments must be less than 2GB in
16644 size, both starting anywhere in memory (determined at link time). The
16645 global register %g4 points to the base of the data segment. Programs
16646 are statically linked and PIC is not supported.
16649 @itemx -mno-stack-bias
16650 @opindex mstack-bias
16651 @opindex mno-stack-bias
16652 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
16653 frame pointer if present, are offset by @minus{}2047 which must be added back
16654 when making stack frame references. This is the default in 64-bit mode.
16655 Otherwise, assume no such offset is present.
16658 These switches are supported in addition to the above on Solaris:
16663 Add support for multithreading using the Solaris threads library. This
16664 option sets flags for both the preprocessor and linker. This option does
16665 not affect the thread safety of object code produced by the compiler or
16666 that of libraries supplied with it.
16670 Add support for multithreading using the POSIX threads library. This
16671 option sets flags for both the preprocessor and linker. This option does
16672 not affect the thread safety of object code produced by the compiler or
16673 that of libraries supplied with it.
16677 This is a synonym for @option{-pthreads}.
16681 @subsection SPU Options
16682 @cindex SPU options
16684 These @samp{-m} options are supported on the SPU:
16688 @itemx -merror-reloc
16689 @opindex mwarn-reloc
16690 @opindex merror-reloc
16692 The loader for SPU does not handle dynamic relocations. By default, GCC
16693 will give an error when it generates code that requires a dynamic
16694 relocation. @option{-mno-error-reloc} disables the error,
16695 @option{-mwarn-reloc} will generate a warning instead.
16698 @itemx -munsafe-dma
16700 @opindex munsafe-dma
16702 Instructions which initiate or test completion of DMA must not be
16703 reordered with respect to loads and stores of the memory which is being
16704 accessed. Users typically address this problem using the volatile
16705 keyword, but that can lead to inefficient code in places where the
16706 memory is known to not change. Rather than mark the memory as volatile
16707 we treat the DMA instructions as potentially effecting all memory. With
16708 @option{-munsafe-dma} users must use the volatile keyword to protect
16711 @item -mbranch-hints
16712 @opindex mbranch-hints
16714 By default, GCC will generate a branch hint instruction to avoid
16715 pipeline stalls for always taken or probably taken branches. A hint
16716 will not be generated closer than 8 instructions away from its branch.
16717 There is little reason to disable them, except for debugging purposes,
16718 or to make an object a little bit smaller.
16722 @opindex msmall-mem
16723 @opindex mlarge-mem
16725 By default, GCC generates code assuming that addresses are never larger
16726 than 18 bits. With @option{-mlarge-mem} code is generated that assumes
16727 a full 32 bit address.
16732 By default, GCC links against startup code that assumes the SPU-style
16733 main function interface (which has an unconventional parameter list).
16734 With @option{-mstdmain}, GCC will link your program against startup
16735 code that assumes a C99-style interface to @code{main}, including a
16736 local copy of @code{argv} strings.
16738 @item -mfixed-range=@var{register-range}
16739 @opindex mfixed-range
16740 Generate code treating the given register range as fixed registers.
16741 A fixed register is one that the register allocator can not use. This is
16742 useful when compiling kernel code. A register range is specified as
16743 two registers separated by a dash. Multiple register ranges can be
16744 specified separated by a comma.
16750 Compile code assuming that pointers to the PPU address space accessed
16751 via the @code{__ea} named address space qualifier are either 32 or 64
16752 bits wide. The default is 32 bits. As this is an ABI changing option,
16753 all object code in an executable must be compiled with the same setting.
16755 @item -maddress-space-conversion
16756 @itemx -mno-address-space-conversion
16757 @opindex maddress-space-conversion
16758 @opindex mno-address-space-conversion
16759 Allow/disallow treating the @code{__ea} address space as superset
16760 of the generic address space. This enables explicit type casts
16761 between @code{__ea} and generic pointer as well as implicit
16762 conversions of generic pointers to @code{__ea} pointers. The
16763 default is to allow address space pointer conversions.
16765 @item -mcache-size=@var{cache-size}
16766 @opindex mcache-size
16767 This option controls the version of libgcc that the compiler links to an
16768 executable and selects a software-managed cache for accessing variables
16769 in the @code{__ea} address space with a particular cache size. Possible
16770 options for @var{cache-size} are @samp{8}, @samp{16}, @samp{32}, @samp{64}
16771 and @samp{128}. The default cache size is 64KB.
16773 @item -matomic-updates
16774 @itemx -mno-atomic-updates
16775 @opindex matomic-updates
16776 @opindex mno-atomic-updates
16777 This option controls the version of libgcc that the compiler links to an
16778 executable and selects whether atomic updates to the software-managed
16779 cache of PPU-side variables are used. If you use atomic updates, changes
16780 to a PPU variable from SPU code using the @code{__ea} named address space
16781 qualifier will not interfere with changes to other PPU variables residing
16782 in the same cache line from PPU code. If you do not use atomic updates,
16783 such interference may occur; however, writing back cache lines will be
16784 more efficient. The default behavior is to use atomic updates.
16787 @itemx -mdual-nops=@var{n}
16788 @opindex mdual-nops
16789 By default, GCC will insert nops to increase dual issue when it expects
16790 it to increase performance. @var{n} can be a value from 0 to 10. A
16791 smaller @var{n} will insert fewer nops. 10 is the default, 0 is the
16792 same as @option{-mno-dual-nops}. Disabled with @option{-Os}.
16794 @item -mhint-max-nops=@var{n}
16795 @opindex mhint-max-nops
16796 Maximum number of nops to insert for a branch hint. A branch hint must
16797 be at least 8 instructions away from the branch it is effecting. GCC
16798 will insert up to @var{n} nops to enforce this, otherwise it will not
16799 generate the branch hint.
16801 @item -mhint-max-distance=@var{n}
16802 @opindex mhint-max-distance
16803 The encoding of the branch hint instruction limits the hint to be within
16804 256 instructions of the branch it is effecting. By default, GCC makes
16805 sure it is within 125.
16808 @opindex msafe-hints
16809 Work around a hardware bug which causes the SPU to stall indefinitely.
16810 By default, GCC will insert the @code{hbrp} instruction to make sure
16811 this stall won't happen.
16815 @node System V Options
16816 @subsection Options for System V
16818 These additional options are available on System V Release 4 for
16819 compatibility with other compilers on those systems:
16824 Create a shared object.
16825 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
16829 Identify the versions of each tool used by the compiler, in a
16830 @code{.ident} assembler directive in the output.
16834 Refrain from adding @code{.ident} directives to the output file (this is
16837 @item -YP,@var{dirs}
16839 Search the directories @var{dirs}, and no others, for libraries
16840 specified with @option{-l}.
16842 @item -Ym,@var{dir}
16844 Look in the directory @var{dir} to find the M4 preprocessor.
16845 The assembler uses this option.
16846 @c This is supposed to go with a -Yd for predefined M4 macro files, but
16847 @c the generic assembler that comes with Solaris takes just -Ym.
16851 @subsection V850 Options
16852 @cindex V850 Options
16854 These @samp{-m} options are defined for V850 implementations:
16858 @itemx -mno-long-calls
16859 @opindex mlong-calls
16860 @opindex mno-long-calls
16861 Treat all calls as being far away (near). If calls are assumed to be
16862 far away, the compiler will always load the functions address up into a
16863 register, and call indirect through the pointer.
16869 Do not optimize (do optimize) basic blocks that use the same index
16870 pointer 4 or more times to copy pointer into the @code{ep} register, and
16871 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
16872 option is on by default if you optimize.
16874 @item -mno-prolog-function
16875 @itemx -mprolog-function
16876 @opindex mno-prolog-function
16877 @opindex mprolog-function
16878 Do not use (do use) external functions to save and restore registers
16879 at the prologue and epilogue of a function. The external functions
16880 are slower, but use less code space if more than one function saves
16881 the same number of registers. The @option{-mprolog-function} option
16882 is on by default if you optimize.
16886 Try to make the code as small as possible. At present, this just turns
16887 on the @option{-mep} and @option{-mprolog-function} options.
16889 @item -mtda=@var{n}
16891 Put static or global variables whose size is @var{n} bytes or less into
16892 the tiny data area that register @code{ep} points to. The tiny data
16893 area can hold up to 256 bytes in total (128 bytes for byte references).
16895 @item -msda=@var{n}
16897 Put static or global variables whose size is @var{n} bytes or less into
16898 the small data area that register @code{gp} points to. The small data
16899 area can hold up to 64 kilobytes.
16901 @item -mzda=@var{n}
16903 Put static or global variables whose size is @var{n} bytes or less into
16904 the first 32 kilobytes of memory.
16908 Specify that the target processor is the V850.
16911 @opindex mbig-switch
16912 Generate code suitable for big switch tables. Use this option only if
16913 the assembler/linker complain about out of range branches within a switch
16918 This option will cause r2 and r5 to be used in the code generated by
16919 the compiler. This setting is the default.
16921 @item -mno-app-regs
16922 @opindex mno-app-regs
16923 This option will cause r2 and r5 to be treated as fixed registers.
16927 Specify that the target processor is the V850E1. The preprocessor
16928 constants @samp{__v850e1__} and @samp{__v850e__} will be defined if
16929 this option is used.
16933 Specify that the target processor is the V850E@. The preprocessor
16934 constant @samp{__v850e__} will be defined if this option is used.
16936 If neither @option{-mv850} nor @option{-mv850e} nor @option{-mv850e1}
16937 are defined then a default target processor will be chosen and the
16938 relevant @samp{__v850*__} preprocessor constant will be defined.
16940 The preprocessor constants @samp{__v850} and @samp{__v851__} are always
16941 defined, regardless of which processor variant is the target.
16943 @item -mdisable-callt
16944 @opindex mdisable-callt
16945 This option will suppress generation of the CALLT instruction for the
16946 v850e and v850e1 flavors of the v850 architecture. The default is
16947 @option{-mno-disable-callt} which allows the CALLT instruction to be used.
16952 @subsection VAX Options
16953 @cindex VAX options
16955 These @samp{-m} options are defined for the VAX:
16960 Do not output certain jump instructions (@code{aobleq} and so on)
16961 that the Unix assembler for the VAX cannot handle across long
16966 Do output those jump instructions, on the assumption that you
16967 will assemble with the GNU assembler.
16971 Output code for g-format floating point numbers instead of d-format.
16974 @node VxWorks Options
16975 @subsection VxWorks Options
16976 @cindex VxWorks Options
16978 The options in this section are defined for all VxWorks targets.
16979 Options specific to the target hardware are listed with the other
16980 options for that target.
16985 GCC can generate code for both VxWorks kernels and real time processes
16986 (RTPs). This option switches from the former to the latter. It also
16987 defines the preprocessor macro @code{__RTP__}.
16990 @opindex non-static
16991 Link an RTP executable against shared libraries rather than static
16992 libraries. The options @option{-static} and @option{-shared} can
16993 also be used for RTPs (@pxref{Link Options}); @option{-static}
17000 These options are passed down to the linker. They are defined for
17001 compatibility with Diab.
17004 @opindex Xbind-lazy
17005 Enable lazy binding of function calls. This option is equivalent to
17006 @option{-Wl,-z,now} and is defined for compatibility with Diab.
17010 Disable lazy binding of function calls. This option is the default and
17011 is defined for compatibility with Diab.
17014 @node x86-64 Options
17015 @subsection x86-64 Options
17016 @cindex x86-64 options
17018 These are listed under @xref{i386 and x86-64 Options}.
17020 @node i386 and x86-64 Windows Options
17021 @subsection i386 and x86-64 Windows Options
17022 @cindex i386 and x86-64 Windows Options
17024 These additional options are available for Windows targets:
17029 This option is available for Cygwin and MinGW targets. It
17030 specifies that a console application is to be generated, by
17031 instructing the linker to set the PE header subsystem type
17032 required for console applications.
17033 This is the default behavior for Cygwin and MinGW targets.
17037 This option is available for Cygwin targets. It specifies that
17038 the Cygwin internal interface is to be used for predefined
17039 preprocessor macros, C runtime libraries and related linker
17040 paths and options. For Cygwin targets this is the default behavior.
17041 This option is deprecated and will be removed in a future release.
17044 @opindex mno-cygwin
17045 This option is available for Cygwin targets. It specifies that
17046 the MinGW internal interface is to be used instead of Cygwin's, by
17047 setting MinGW-related predefined macros and linker paths and default
17049 This option is deprecated and will be removed in a future release.
17053 This option is available for Cygwin and MinGW targets. It
17054 specifies that a DLL - a dynamic link library - is to be
17055 generated, enabling the selection of the required runtime
17056 startup object and entry point.
17058 @item -mnop-fun-dllimport
17059 @opindex mnop-fun-dllimport
17060 This option is available for Cygwin and MinGW targets. It
17061 specifies that the dllimport attribute should be ignored.
17065 This option is available for MinGW targets. It specifies
17066 that MinGW-specific thread support is to be used.
17070 This option is available for mingw-w64 targets. It specifies
17071 that the UNICODE macro is getting pre-defined and that the
17072 unicode capable runtime startup code is chosen.
17076 This option is available for Cygwin and MinGW targets. It
17077 specifies that the typical Windows pre-defined macros are to
17078 be set in the pre-processor, but does not influence the choice
17079 of runtime library/startup code.
17083 This option is available for Cygwin and MinGW targets. It
17084 specifies that a GUI application is to be generated by
17085 instructing the linker to set the PE header subsystem type
17088 @item -fno-set-stack-executable
17089 @opindex fno-set-stack-executable
17090 This option is available for MinGW targets. It specifies that
17091 the executable flag for stack used by nested functions isn't
17092 set. This is necessary for binaries running in kernel mode of
17093 Windows, as there the user32 API, which is used to set executable
17094 privileges, isn't available.
17096 @item -mpe-aligned-commons
17097 @opindex mpe-aligned-commons
17098 This option is available for Cygwin and MinGW targets. It
17099 specifies that the GNU extension to the PE file format that
17100 permits the correct alignment of COMMON variables should be
17101 used when generating code. It will be enabled by default if
17102 GCC detects that the target assembler found during configuration
17103 supports the feature.
17106 See also under @ref{i386 and x86-64 Options} for standard options.
17108 @node Xstormy16 Options
17109 @subsection Xstormy16 Options
17110 @cindex Xstormy16 Options
17112 These options are defined for Xstormy16:
17117 Choose startup files and linker script suitable for the simulator.
17120 @node Xtensa Options
17121 @subsection Xtensa Options
17122 @cindex Xtensa Options
17124 These options are supported for Xtensa targets:
17128 @itemx -mno-const16
17130 @opindex mno-const16
17131 Enable or disable use of @code{CONST16} instructions for loading
17132 constant values. The @code{CONST16} instruction is currently not a
17133 standard option from Tensilica. When enabled, @code{CONST16}
17134 instructions are always used in place of the standard @code{L32R}
17135 instructions. The use of @code{CONST16} is enabled by default only if
17136 the @code{L32R} instruction is not available.
17139 @itemx -mno-fused-madd
17140 @opindex mfused-madd
17141 @opindex mno-fused-madd
17142 Enable or disable use of fused multiply/add and multiply/subtract
17143 instructions in the floating-point option. This has no effect if the
17144 floating-point option is not also enabled. Disabling fused multiply/add
17145 and multiply/subtract instructions forces the compiler to use separate
17146 instructions for the multiply and add/subtract operations. This may be
17147 desirable in some cases where strict IEEE 754-compliant results are
17148 required: the fused multiply add/subtract instructions do not round the
17149 intermediate result, thereby producing results with @emph{more} bits of
17150 precision than specified by the IEEE standard. Disabling fused multiply
17151 add/subtract instructions also ensures that the program output is not
17152 sensitive to the compiler's ability to combine multiply and add/subtract
17155 @item -mserialize-volatile
17156 @itemx -mno-serialize-volatile
17157 @opindex mserialize-volatile
17158 @opindex mno-serialize-volatile
17159 When this option is enabled, GCC inserts @code{MEMW} instructions before
17160 @code{volatile} memory references to guarantee sequential consistency.
17161 The default is @option{-mserialize-volatile}. Use
17162 @option{-mno-serialize-volatile} to omit the @code{MEMW} instructions.
17164 @item -mforce-no-pic
17165 @opindex mforce-no-pic
17166 For targets, like GNU/Linux, where all user-mode Xtensa code must be
17167 position-independent code (PIC), this option disables PIC for compiling
17170 @item -mtext-section-literals
17171 @itemx -mno-text-section-literals
17172 @opindex mtext-section-literals
17173 @opindex mno-text-section-literals
17174 Control the treatment of literal pools. The default is
17175 @option{-mno-text-section-literals}, which places literals in a separate
17176 section in the output file. This allows the literal pool to be placed
17177 in a data RAM/ROM, and it also allows the linker to combine literal
17178 pools from separate object files to remove redundant literals and
17179 improve code size. With @option{-mtext-section-literals}, the literals
17180 are interspersed in the text section in order to keep them as close as
17181 possible to their references. This may be necessary for large assembly
17184 @item -mtarget-align
17185 @itemx -mno-target-align
17186 @opindex mtarget-align
17187 @opindex mno-target-align
17188 When this option is enabled, GCC instructs the assembler to
17189 automatically align instructions to reduce branch penalties at the
17190 expense of some code density. The assembler attempts to widen density
17191 instructions to align branch targets and the instructions following call
17192 instructions. If there are not enough preceding safe density
17193 instructions to align a target, no widening will be performed. The
17194 default is @option{-mtarget-align}. These options do not affect the
17195 treatment of auto-aligned instructions like @code{LOOP}, which the
17196 assembler will always align, either by widening density instructions or
17197 by inserting no-op instructions.
17200 @itemx -mno-longcalls
17201 @opindex mlongcalls
17202 @opindex mno-longcalls
17203 When this option is enabled, GCC instructs the assembler to translate
17204 direct calls to indirect calls unless it can determine that the target
17205 of a direct call is in the range allowed by the call instruction. This
17206 translation typically occurs for calls to functions in other source
17207 files. Specifically, the assembler translates a direct @code{CALL}
17208 instruction into an @code{L32R} followed by a @code{CALLX} instruction.
17209 The default is @option{-mno-longcalls}. This option should be used in
17210 programs where the call target can potentially be out of range. This
17211 option is implemented in the assembler, not the compiler, so the
17212 assembly code generated by GCC will still show direct call
17213 instructions---look at the disassembled object code to see the actual
17214 instructions. Note that the assembler will use an indirect call for
17215 every cross-file call, not just those that really will be out of range.
17218 @node zSeries Options
17219 @subsection zSeries Options
17220 @cindex zSeries options
17222 These are listed under @xref{S/390 and zSeries Options}.
17224 @node Code Gen Options
17225 @section Options for Code Generation Conventions
17226 @cindex code generation conventions
17227 @cindex options, code generation
17228 @cindex run-time options
17230 These machine-independent options control the interface conventions
17231 used in code generation.
17233 Most of them have both positive and negative forms; the negative form
17234 of @option{-ffoo} would be @option{-fno-foo}. In the table below, only
17235 one of the forms is listed---the one which is not the default. You
17236 can figure out the other form by either removing @samp{no-} or adding
17240 @item -fbounds-check
17241 @opindex fbounds-check
17242 For front-ends that support it, generate additional code to check that
17243 indices used to access arrays are within the declared range. This is
17244 currently only supported by the Java and Fortran front-ends, where
17245 this option defaults to true and false respectively.
17249 This option generates traps for signed overflow on addition, subtraction,
17250 multiplication operations.
17254 This option instructs the compiler to assume that signed arithmetic
17255 overflow of addition, subtraction and multiplication wraps around
17256 using twos-complement representation. This flag enables some optimizations
17257 and disables others. This option is enabled by default for the Java
17258 front-end, as required by the Java language specification.
17261 @opindex fexceptions
17262 Enable exception handling. Generates extra code needed to propagate
17263 exceptions. For some targets, this implies GCC will generate frame
17264 unwind information for all functions, which can produce significant data
17265 size overhead, although it does not affect execution. If you do not
17266 specify this option, GCC will enable it by default for languages like
17267 C++ which normally require exception handling, and disable it for
17268 languages like C that do not normally require it. However, you may need
17269 to enable this option when compiling C code that needs to interoperate
17270 properly with exception handlers written in C++. You may also wish to
17271 disable this option if you are compiling older C++ programs that don't
17272 use exception handling.
17274 @item -fnon-call-exceptions
17275 @opindex fnon-call-exceptions
17276 Generate code that allows trapping instructions to throw exceptions.
17277 Note that this requires platform-specific runtime support that does
17278 not exist everywhere. Moreover, it only allows @emph{trapping}
17279 instructions to throw exceptions, i.e.@: memory references or floating
17280 point instructions. It does not allow exceptions to be thrown from
17281 arbitrary signal handlers such as @code{SIGALRM}.
17283 @item -funwind-tables
17284 @opindex funwind-tables
17285 Similar to @option{-fexceptions}, except that it will just generate any needed
17286 static data, but will not affect the generated code in any other way.
17287 You will normally not enable this option; instead, a language processor
17288 that needs this handling would enable it on your behalf.
17290 @item -fasynchronous-unwind-tables
17291 @opindex fasynchronous-unwind-tables
17292 Generate unwind table in dwarf2 format, if supported by target machine. The
17293 table is exact at each instruction boundary, so it can be used for stack
17294 unwinding from asynchronous events (such as debugger or garbage collector).
17296 @item -fpcc-struct-return
17297 @opindex fpcc-struct-return
17298 Return ``short'' @code{struct} and @code{union} values in memory like
17299 longer ones, rather than in registers. This convention is less
17300 efficient, but it has the advantage of allowing intercallability between
17301 GCC-compiled files and files compiled with other compilers, particularly
17302 the Portable C Compiler (pcc).
17304 The precise convention for returning structures in memory depends
17305 on the target configuration macros.
17307 Short structures and unions are those whose size and alignment match
17308 that of some integer type.
17310 @strong{Warning:} code compiled with the @option{-fpcc-struct-return}
17311 switch is not binary compatible with code compiled with the
17312 @option{-freg-struct-return} switch.
17313 Use it to conform to a non-default application binary interface.
17315 @item -freg-struct-return
17316 @opindex freg-struct-return
17317 Return @code{struct} and @code{union} values in registers when possible.
17318 This is more efficient for small structures than
17319 @option{-fpcc-struct-return}.
17321 If you specify neither @option{-fpcc-struct-return} nor
17322 @option{-freg-struct-return}, GCC defaults to whichever convention is
17323 standard for the target. If there is no standard convention, GCC
17324 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
17325 the principal compiler. In those cases, we can choose the standard, and
17326 we chose the more efficient register return alternative.
17328 @strong{Warning:} code compiled with the @option{-freg-struct-return}
17329 switch is not binary compatible with code compiled with the
17330 @option{-fpcc-struct-return} switch.
17331 Use it to conform to a non-default application binary interface.
17333 @item -fshort-enums
17334 @opindex fshort-enums
17335 Allocate to an @code{enum} type only as many bytes as it needs for the
17336 declared range of possible values. Specifically, the @code{enum} type
17337 will be equivalent to the smallest integer type which has enough room.
17339 @strong{Warning:} the @option{-fshort-enums} switch causes GCC to generate
17340 code that is not binary compatible with code generated without that switch.
17341 Use it to conform to a non-default application binary interface.
17343 @item -fshort-double
17344 @opindex fshort-double
17345 Use the same size for @code{double} as for @code{float}.
17347 @strong{Warning:} the @option{-fshort-double} switch causes GCC to generate
17348 code that is not binary compatible with code generated without that switch.
17349 Use it to conform to a non-default application binary interface.
17351 @item -fshort-wchar
17352 @opindex fshort-wchar
17353 Override the underlying type for @samp{wchar_t} to be @samp{short
17354 unsigned int} instead of the default for the target. This option is
17355 useful for building programs to run under WINE@.
17357 @strong{Warning:} the @option{-fshort-wchar} switch causes GCC to generate
17358 code that is not binary compatible with code generated without that switch.
17359 Use it to conform to a non-default application binary interface.
17362 @opindex fno-common
17363 In C code, controls the placement of uninitialized global variables.
17364 Unix C compilers have traditionally permitted multiple definitions of
17365 such variables in different compilation units by placing the variables
17367 This is the behavior specified by @option{-fcommon}, and is the default
17368 for GCC on most targets.
17369 On the other hand, this behavior is not required by ISO C, and on some
17370 targets may carry a speed or code size penalty on variable references.
17371 The @option{-fno-common} option specifies that the compiler should place
17372 uninitialized global variables in the data section of the object file,
17373 rather than generating them as common blocks.
17374 This has the effect that if the same variable is declared
17375 (without @code{extern}) in two different compilations,
17376 you will get a multiple-definition error when you link them.
17377 In this case, you must compile with @option{-fcommon} instead.
17378 Compiling with @option{-fno-common} is useful on targets for which
17379 it provides better performance, or if you wish to verify that the
17380 program will work on other systems which always treat uninitialized
17381 variable declarations this way.
17385 Ignore the @samp{#ident} directive.
17387 @item -finhibit-size-directive
17388 @opindex finhibit-size-directive
17389 Don't output a @code{.size} assembler directive, or anything else that
17390 would cause trouble if the function is split in the middle, and the
17391 two halves are placed at locations far apart in memory. This option is
17392 used when compiling @file{crtstuff.c}; you should not need to use it
17395 @item -fverbose-asm
17396 @opindex fverbose-asm
17397 Put extra commentary information in the generated assembly code to
17398 make it more readable. This option is generally only of use to those
17399 who actually need to read the generated assembly code (perhaps while
17400 debugging the compiler itself).
17402 @option{-fno-verbose-asm}, the default, causes the
17403 extra information to be omitted and is useful when comparing two assembler
17406 @item -frecord-gcc-switches
17407 @opindex frecord-gcc-switches
17408 This switch causes the command line that was used to invoke the
17409 compiler to be recorded into the object file that is being created.
17410 This switch is only implemented on some targets and the exact format
17411 of the recording is target and binary file format dependent, but it
17412 usually takes the form of a section containing ASCII text. This
17413 switch is related to the @option{-fverbose-asm} switch, but that
17414 switch only records information in the assembler output file as
17415 comments, so it never reaches the object file.
17419 @cindex global offset table
17421 Generate position-independent code (PIC) suitable for use in a shared
17422 library, if supported for the target machine. Such code accesses all
17423 constant addresses through a global offset table (GOT)@. The dynamic
17424 loader resolves the GOT entries when the program starts (the dynamic
17425 loader is not part of GCC; it is part of the operating system). If
17426 the GOT size for the linked executable exceeds a machine-specific
17427 maximum size, you get an error message from the linker indicating that
17428 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
17429 instead. (These maximums are 8k on the SPARC and 32k
17430 on the m68k and RS/6000. The 386 has no such limit.)
17432 Position-independent code requires special support, and therefore works
17433 only on certain machines. For the 386, GCC supports PIC for System V
17434 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
17435 position-independent.
17437 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
17442 If supported for the target machine, emit position-independent code,
17443 suitable for dynamic linking and avoiding any limit on the size of the
17444 global offset table. This option makes a difference on the m68k,
17445 PowerPC and SPARC@.
17447 Position-independent code requires special support, and therefore works
17448 only on certain machines.
17450 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
17457 These options are similar to @option{-fpic} and @option{-fPIC}, but
17458 generated position independent code can be only linked into executables.
17459 Usually these options are used when @option{-pie} GCC option will be
17460 used during linking.
17462 @option{-fpie} and @option{-fPIE} both define the macros
17463 @code{__pie__} and @code{__PIE__}. The macros have the value 1
17464 for @option{-fpie} and 2 for @option{-fPIE}.
17466 @item -fno-jump-tables
17467 @opindex fno-jump-tables
17468 Do not use jump tables for switch statements even where it would be
17469 more efficient than other code generation strategies. This option is
17470 of use in conjunction with @option{-fpic} or @option{-fPIC} for
17471 building code which forms part of a dynamic linker and cannot
17472 reference the address of a jump table. On some targets, jump tables
17473 do not require a GOT and this option is not needed.
17475 @item -ffixed-@var{reg}
17477 Treat the register named @var{reg} as a fixed register; generated code
17478 should never refer to it (except perhaps as a stack pointer, frame
17479 pointer or in some other fixed role).
17481 @var{reg} must be the name of a register. The register names accepted
17482 are machine-specific and are defined in the @code{REGISTER_NAMES}
17483 macro in the machine description macro file.
17485 This flag does not have a negative form, because it specifies a
17488 @item -fcall-used-@var{reg}
17489 @opindex fcall-used
17490 Treat the register named @var{reg} as an allocable register that is
17491 clobbered by function calls. It may be allocated for temporaries or
17492 variables that do not live across a call. Functions compiled this way
17493 will not save and restore the register @var{reg}.
17495 It is an error to used this flag with the frame pointer or stack pointer.
17496 Use of this flag for other registers that have fixed pervasive roles in
17497 the machine's execution model will produce disastrous results.
17499 This flag does not have a negative form, because it specifies a
17502 @item -fcall-saved-@var{reg}
17503 @opindex fcall-saved
17504 Treat the register named @var{reg} as an allocable register saved by
17505 functions. It may be allocated even for temporaries or variables that
17506 live across a call. Functions compiled this way will save and restore
17507 the register @var{reg} if they use it.
17509 It is an error to used this flag with the frame pointer or stack pointer.
17510 Use of this flag for other registers that have fixed pervasive roles in
17511 the machine's execution model will produce disastrous results.
17513 A different sort of disaster will result from the use of this flag for
17514 a register in which function values may be returned.
17516 This flag does not have a negative form, because it specifies a
17519 @item -fpack-struct[=@var{n}]
17520 @opindex fpack-struct
17521 Without a value specified, pack all structure members together without
17522 holes. When a value is specified (which must be a small power of two), pack
17523 structure members according to this value, representing the maximum
17524 alignment (that is, objects with default alignment requirements larger than
17525 this will be output potentially unaligned at the next fitting location.
17527 @strong{Warning:} the @option{-fpack-struct} switch causes GCC to generate
17528 code that is not binary compatible with code generated without that switch.
17529 Additionally, it makes the code suboptimal.
17530 Use it to conform to a non-default application binary interface.
17532 @item -finstrument-functions
17533 @opindex finstrument-functions
17534 Generate instrumentation calls for entry and exit to functions. Just
17535 after function entry and just before function exit, the following
17536 profiling functions will be called with the address of the current
17537 function and its call site. (On some platforms,
17538 @code{__builtin_return_address} does not work beyond the current
17539 function, so the call site information may not be available to the
17540 profiling functions otherwise.)
17543 void __cyg_profile_func_enter (void *this_fn,
17545 void __cyg_profile_func_exit (void *this_fn,
17549 The first argument is the address of the start of the current function,
17550 which may be looked up exactly in the symbol table.
17552 This instrumentation is also done for functions expanded inline in other
17553 functions. The profiling calls will indicate where, conceptually, the
17554 inline function is entered and exited. This means that addressable
17555 versions of such functions must be available. If all your uses of a
17556 function are expanded inline, this may mean an additional expansion of
17557 code size. If you use @samp{extern inline} in your C code, an
17558 addressable version of such functions must be provided. (This is
17559 normally the case anyways, but if you get lucky and the optimizer always
17560 expands the functions inline, you might have gotten away without
17561 providing static copies.)
17563 A function may be given the attribute @code{no_instrument_function}, in
17564 which case this instrumentation will not be done. This can be used, for
17565 example, for the profiling functions listed above, high-priority
17566 interrupt routines, and any functions from which the profiling functions
17567 cannot safely be called (perhaps signal handlers, if the profiling
17568 routines generate output or allocate memory).
17570 @item -finstrument-functions-exclude-file-list=@var{file},@var{file},@dots{}
17571 @opindex finstrument-functions-exclude-file-list
17573 Set the list of functions that are excluded from instrumentation (see
17574 the description of @code{-finstrument-functions}). If the file that
17575 contains a function definition matches with one of @var{file}, then
17576 that function is not instrumented. The match is done on substrings:
17577 if the @var{file} parameter is a substring of the file name, it is
17578 considered to be a match.
17581 @code{-finstrument-functions-exclude-file-list=/bits/stl,include/sys}
17582 will exclude any inline function defined in files whose pathnames
17583 contain @code{/bits/stl} or @code{include/sys}.
17585 If, for some reason, you want to include letter @code{','} in one of
17586 @var{sym}, write @code{'\,'}. For example,
17587 @code{-finstrument-functions-exclude-file-list='\,\,tmp'}
17588 (note the single quote surrounding the option).
17590 @item -finstrument-functions-exclude-function-list=@var{sym},@var{sym},@dots{}
17591 @opindex finstrument-functions-exclude-function-list
17593 This is similar to @code{-finstrument-functions-exclude-file-list},
17594 but this option sets the list of function names to be excluded from
17595 instrumentation. The function name to be matched is its user-visible
17596 name, such as @code{vector<int> blah(const vector<int> &)}, not the
17597 internal mangled name (e.g., @code{_Z4blahRSt6vectorIiSaIiEE}). The
17598 match is done on substrings: if the @var{sym} parameter is a substring
17599 of the function name, it is considered to be a match. For C99 and C++
17600 extended identifiers, the function name must be given in UTF-8, not
17601 using universal character names.
17603 @item -fstack-check
17604 @opindex fstack-check
17605 Generate code to verify that you do not go beyond the boundary of the
17606 stack. You should specify this flag if you are running in an
17607 environment with multiple threads, but only rarely need to specify it in
17608 a single-threaded environment since stack overflow is automatically
17609 detected on nearly all systems if there is only one stack.
17611 Note that this switch does not actually cause checking to be done; the
17612 operating system or the language runtime must do that. The switch causes
17613 generation of code to ensure that they see the stack being extended.
17615 You can additionally specify a string parameter: @code{no} means no
17616 checking, @code{generic} means force the use of old-style checking,
17617 @code{specific} means use the best checking method and is equivalent
17618 to bare @option{-fstack-check}.
17620 Old-style checking is a generic mechanism that requires no specific
17621 target support in the compiler but comes with the following drawbacks:
17625 Modified allocation strategy for large objects: they will always be
17626 allocated dynamically if their size exceeds a fixed threshold.
17629 Fixed limit on the size of the static frame of functions: when it is
17630 topped by a particular function, stack checking is not reliable and
17631 a warning is issued by the compiler.
17634 Inefficiency: because of both the modified allocation strategy and the
17635 generic implementation, the performances of the code are hampered.
17638 Note that old-style stack checking is also the fallback method for
17639 @code{specific} if no target support has been added in the compiler.
17641 @item -fstack-limit-register=@var{reg}
17642 @itemx -fstack-limit-symbol=@var{sym}
17643 @itemx -fno-stack-limit
17644 @opindex fstack-limit-register
17645 @opindex fstack-limit-symbol
17646 @opindex fno-stack-limit
17647 Generate code to ensure that the stack does not grow beyond a certain value,
17648 either the value of a register or the address of a symbol. If the stack
17649 would grow beyond the value, a signal is raised. For most targets,
17650 the signal is raised before the stack overruns the boundary, so
17651 it is possible to catch the signal without taking special precautions.
17653 For instance, if the stack starts at absolute address @samp{0x80000000}
17654 and grows downwards, you can use the flags
17655 @option{-fstack-limit-symbol=__stack_limit} and
17656 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
17657 of 128KB@. Note that this may only work with the GNU linker.
17659 @item -fleading-underscore
17660 @opindex fleading-underscore
17661 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
17662 change the way C symbols are represented in the object file. One use
17663 is to help link with legacy assembly code.
17665 @strong{Warning:} the @option{-fleading-underscore} switch causes GCC to
17666 generate code that is not binary compatible with code generated without that
17667 switch. Use it to conform to a non-default application binary interface.
17668 Not all targets provide complete support for this switch.
17670 @item -ftls-model=@var{model}
17671 @opindex ftls-model
17672 Alter the thread-local storage model to be used (@pxref{Thread-Local}).
17673 The @var{model} argument should be one of @code{global-dynamic},
17674 @code{local-dynamic}, @code{initial-exec} or @code{local-exec}.
17676 The default without @option{-fpic} is @code{initial-exec}; with
17677 @option{-fpic} the default is @code{global-dynamic}.
17679 @item -fvisibility=@var{default|internal|hidden|protected}
17680 @opindex fvisibility
17681 Set the default ELF image symbol visibility to the specified option---all
17682 symbols will be marked with this unless overridden within the code.
17683 Using this feature can very substantially improve linking and
17684 load times of shared object libraries, produce more optimized
17685 code, provide near-perfect API export and prevent symbol clashes.
17686 It is @strong{strongly} recommended that you use this in any shared objects
17689 Despite the nomenclature, @code{default} always means public ie;
17690 available to be linked against from outside the shared object.
17691 @code{protected} and @code{internal} are pretty useless in real-world
17692 usage so the only other commonly used option will be @code{hidden}.
17693 The default if @option{-fvisibility} isn't specified is
17694 @code{default}, i.e., make every
17695 symbol public---this causes the same behavior as previous versions of
17698 A good explanation of the benefits offered by ensuring ELF
17699 symbols have the correct visibility is given by ``How To Write
17700 Shared Libraries'' by Ulrich Drepper (which can be found at
17701 @w{@uref{http://people.redhat.com/~drepper/}})---however a superior
17702 solution made possible by this option to marking things hidden when
17703 the default is public is to make the default hidden and mark things
17704 public. This is the norm with DLL's on Windows and with @option{-fvisibility=hidden}
17705 and @code{__attribute__ ((visibility("default")))} instead of
17706 @code{__declspec(dllexport)} you get almost identical semantics with
17707 identical syntax. This is a great boon to those working with
17708 cross-platform projects.
17710 For those adding visibility support to existing code, you may find
17711 @samp{#pragma GCC visibility} of use. This works by you enclosing
17712 the declarations you wish to set visibility for with (for example)
17713 @samp{#pragma GCC visibility push(hidden)} and
17714 @samp{#pragma GCC visibility pop}.
17715 Bear in mind that symbol visibility should be viewed @strong{as
17716 part of the API interface contract} and thus all new code should
17717 always specify visibility when it is not the default ie; declarations
17718 only for use within the local DSO should @strong{always} be marked explicitly
17719 as hidden as so to avoid PLT indirection overheads---making this
17720 abundantly clear also aids readability and self-documentation of the code.
17721 Note that due to ISO C++ specification requirements, operator new and
17722 operator delete must always be of default visibility.
17724 Be aware that headers from outside your project, in particular system
17725 headers and headers from any other library you use, may not be
17726 expecting to be compiled with visibility other than the default. You
17727 may need to explicitly say @samp{#pragma GCC visibility push(default)}
17728 before including any such headers.
17730 @samp{extern} declarations are not affected by @samp{-fvisibility}, so
17731 a lot of code can be recompiled with @samp{-fvisibility=hidden} with
17732 no modifications. However, this means that calls to @samp{extern}
17733 functions with no explicit visibility will use the PLT, so it is more
17734 effective to use @samp{__attribute ((visibility))} and/or
17735 @samp{#pragma GCC visibility} to tell the compiler which @samp{extern}
17736 declarations should be treated as hidden.
17738 Note that @samp{-fvisibility} does affect C++ vague linkage
17739 entities. This means that, for instance, an exception class that will
17740 be thrown between DSOs must be explicitly marked with default
17741 visibility so that the @samp{type_info} nodes will be unified between
17744 An overview of these techniques, their benefits and how to use them
17745 is at @w{@uref{http://gcc.gnu.org/wiki/Visibility}}.
17747 @item -fstrict-volatile-bitfields
17748 This option should be used if accesses to volatile bitfields (or other
17749 structure fields, although the compiler usually honors those types
17750 anyway) should use a single access in a mode of the same size as the
17751 container's type, aligned to a natural alignment if possible. For
17752 example, targets with memory-mapped peripheral registers might require
17753 all such accesses to be 16 bits wide; with this flag the user could
17754 declare all peripheral bitfields as ``unsigned short'' (assuming short
17755 is 16 bits on these targets) to force GCC to use 16 bit accesses
17756 instead of, perhaps, a more efficient 32 bit access.
17758 If this option is disabled, the compiler will use the most efficient
17759 instruction. In the previous example, that might be a 32-bit load
17760 instruction, even though that will access bytes that do not contain
17761 any portion of the bitfield, or memory-mapped registers unrelated to
17762 the one being updated.
17764 If the target requires strict alignment, and honoring the container
17765 type would require violating this alignment, a warning is issued.
17766 However, the access happens as the user requested, under the
17767 assumption that the user knows something about the target hardware
17768 that GCC is unaware of.
17770 The default value of this option is determined by the application binary
17771 interface for the target processor.
17777 @node Environment Variables
17778 @section Environment Variables Affecting GCC
17779 @cindex environment variables
17781 @c man begin ENVIRONMENT
17782 This section describes several environment variables that affect how GCC
17783 operates. Some of them work by specifying directories or prefixes to use
17784 when searching for various kinds of files. Some are used to specify other
17785 aspects of the compilation environment.
17787 Note that you can also specify places to search using options such as
17788 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
17789 take precedence over places specified using environment variables, which
17790 in turn take precedence over those specified by the configuration of GCC@.
17791 @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint,
17792 GNU Compiler Collection (GCC) Internals}.
17797 @c @itemx LC_COLLATE
17799 @c @itemx LC_MONETARY
17800 @c @itemx LC_NUMERIC
17805 @c @findex LC_COLLATE
17806 @findex LC_MESSAGES
17807 @c @findex LC_MONETARY
17808 @c @findex LC_NUMERIC
17812 These environment variables control the way that GCC uses
17813 localization information that allow GCC to work with different
17814 national conventions. GCC inspects the locale categories
17815 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
17816 so. These locale categories can be set to any value supported by your
17817 installation. A typical value is @samp{en_GB.UTF-8} for English in the United
17818 Kingdom encoded in UTF-8.
17820 The @env{LC_CTYPE} environment variable specifies character
17821 classification. GCC uses it to determine the character boundaries in
17822 a string; this is needed for some multibyte encodings that contain quote
17823 and escape characters that would otherwise be interpreted as a string
17826 The @env{LC_MESSAGES} environment variable specifies the language to
17827 use in diagnostic messages.
17829 If the @env{LC_ALL} environment variable is set, it overrides the value
17830 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
17831 and @env{LC_MESSAGES} default to the value of the @env{LANG}
17832 environment variable. If none of these variables are set, GCC
17833 defaults to traditional C English behavior.
17837 If @env{TMPDIR} is set, it specifies the directory to use for temporary
17838 files. GCC uses temporary files to hold the output of one stage of
17839 compilation which is to be used as input to the next stage: for example,
17840 the output of the preprocessor, which is the input to the compiler
17843 @item GCC_EXEC_PREFIX
17844 @findex GCC_EXEC_PREFIX
17845 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
17846 names of the subprograms executed by the compiler. No slash is added
17847 when this prefix is combined with the name of a subprogram, but you can
17848 specify a prefix that ends with a slash if you wish.
17850 If @env{GCC_EXEC_PREFIX} is not set, GCC will attempt to figure out
17851 an appropriate prefix to use based on the pathname it was invoked with.
17853 If GCC cannot find the subprogram using the specified prefix, it
17854 tries looking in the usual places for the subprogram.
17856 The default value of @env{GCC_EXEC_PREFIX} is
17857 @file{@var{prefix}/lib/gcc/} where @var{prefix} is the prefix to
17858 the installed compiler. In many cases @var{prefix} is the value
17859 of @code{prefix} when you ran the @file{configure} script.
17861 Other prefixes specified with @option{-B} take precedence over this prefix.
17863 This prefix is also used for finding files such as @file{crt0.o} that are
17866 In addition, the prefix is used in an unusual way in finding the
17867 directories to search for header files. For each of the standard
17868 directories whose name normally begins with @samp{/usr/local/lib/gcc}
17869 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
17870 replacing that beginning with the specified prefix to produce an
17871 alternate directory name. Thus, with @option{-Bfoo/}, GCC will search
17872 @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}.
17873 These alternate directories are searched first; the standard directories
17874 come next. If a standard directory begins with the configured
17875 @var{prefix} then the value of @var{prefix} is replaced by
17876 @env{GCC_EXEC_PREFIX} when looking for header files.
17878 @item COMPILER_PATH
17879 @findex COMPILER_PATH
17880 The value of @env{COMPILER_PATH} is a colon-separated list of
17881 directories, much like @env{PATH}. GCC tries the directories thus
17882 specified when searching for subprograms, if it can't find the
17883 subprograms using @env{GCC_EXEC_PREFIX}.
17886 @findex LIBRARY_PATH
17887 The value of @env{LIBRARY_PATH} is a colon-separated list of
17888 directories, much like @env{PATH}. When configured as a native compiler,
17889 GCC tries the directories thus specified when searching for special
17890 linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking
17891 using GCC also uses these directories when searching for ordinary
17892 libraries for the @option{-l} option (but directories specified with
17893 @option{-L} come first).
17897 @cindex locale definition
17898 This variable is used to pass locale information to the compiler. One way in
17899 which this information is used is to determine the character set to be used
17900 when character literals, string literals and comments are parsed in C and C++.
17901 When the compiler is configured to allow multibyte characters,
17902 the following values for @env{LANG} are recognized:
17906 Recognize JIS characters.
17908 Recognize SJIS characters.
17910 Recognize EUCJP characters.
17913 If @env{LANG} is not defined, or if it has some other value, then the
17914 compiler will use mblen and mbtowc as defined by the default locale to
17915 recognize and translate multibyte characters.
17919 Some additional environments variables affect the behavior of the
17922 @include cppenv.texi
17926 @node Precompiled Headers
17927 @section Using Precompiled Headers
17928 @cindex precompiled headers
17929 @cindex speed of compilation
17931 Often large projects have many header files that are included in every
17932 source file. The time the compiler takes to process these header files
17933 over and over again can account for nearly all of the time required to
17934 build the project. To make builds faster, GCC allows users to
17935 `precompile' a header file; then, if builds can use the precompiled
17936 header file they will be much faster.
17938 To create a precompiled header file, simply compile it as you would any
17939 other file, if necessary using the @option{-x} option to make the driver
17940 treat it as a C or C++ header file. You will probably want to use a
17941 tool like @command{make} to keep the precompiled header up-to-date when
17942 the headers it contains change.
17944 A precompiled header file will be searched for when @code{#include} is
17945 seen in the compilation. As it searches for the included file
17946 (@pxref{Search Path,,Search Path,cpp,The C Preprocessor}) the
17947 compiler looks for a precompiled header in each directory just before it
17948 looks for the include file in that directory. The name searched for is
17949 the name specified in the @code{#include} with @samp{.gch} appended. If
17950 the precompiled header file can't be used, it is ignored.
17952 For instance, if you have @code{#include "all.h"}, and you have
17953 @file{all.h.gch} in the same directory as @file{all.h}, then the
17954 precompiled header file will be used if possible, and the original
17955 header will be used otherwise.
17957 Alternatively, you might decide to put the precompiled header file in a
17958 directory and use @option{-I} to ensure that directory is searched
17959 before (or instead of) the directory containing the original header.
17960 Then, if you want to check that the precompiled header file is always
17961 used, you can put a file of the same name as the original header in this
17962 directory containing an @code{#error} command.
17964 This also works with @option{-include}. So yet another way to use
17965 precompiled headers, good for projects not designed with precompiled
17966 header files in mind, is to simply take most of the header files used by
17967 a project, include them from another header file, precompile that header
17968 file, and @option{-include} the precompiled header. If the header files
17969 have guards against multiple inclusion, they will be skipped because
17970 they've already been included (in the precompiled header).
17972 If you need to precompile the same header file for different
17973 languages, targets, or compiler options, you can instead make a
17974 @emph{directory} named like @file{all.h.gch}, and put each precompiled
17975 header in the directory, perhaps using @option{-o}. It doesn't matter
17976 what you call the files in the directory, every precompiled header in
17977 the directory will be considered. The first precompiled header
17978 encountered in the directory that is valid for this compilation will
17979 be used; they're searched in no particular order.
17981 There are many other possibilities, limited only by your imagination,
17982 good sense, and the constraints of your build system.
17984 A precompiled header file can be used only when these conditions apply:
17988 Only one precompiled header can be used in a particular compilation.
17991 A precompiled header can't be used once the first C token is seen. You
17992 can have preprocessor directives before a precompiled header; you can
17993 even include a precompiled header from inside another header, so long as
17994 there are no C tokens before the @code{#include}.
17997 The precompiled header file must be produced for the same language as
17998 the current compilation. You can't use a C precompiled header for a C++
18002 The precompiled header file must have been produced by the same compiler
18003 binary as the current compilation is using.
18006 Any macros defined before the precompiled header is included must
18007 either be defined in the same way as when the precompiled header was
18008 generated, or must not affect the precompiled header, which usually
18009 means that they don't appear in the precompiled header at all.
18011 The @option{-D} option is one way to define a macro before a
18012 precompiled header is included; using a @code{#define} can also do it.
18013 There are also some options that define macros implicitly, like
18014 @option{-O} and @option{-Wdeprecated}; the same rule applies to macros
18017 @item If debugging information is output when using the precompiled
18018 header, using @option{-g} or similar, the same kind of debugging information
18019 must have been output when building the precompiled header. However,
18020 a precompiled header built using @option{-g} can be used in a compilation
18021 when no debugging information is being output.
18023 @item The same @option{-m} options must generally be used when building
18024 and using the precompiled header. @xref{Submodel Options},
18025 for any cases where this rule is relaxed.
18027 @item Each of the following options must be the same when building and using
18028 the precompiled header:
18030 @gccoptlist{-fexceptions}
18033 Some other command-line options starting with @option{-f},
18034 @option{-p}, or @option{-O} must be defined in the same way as when
18035 the precompiled header was generated. At present, it's not clear
18036 which options are safe to change and which are not; the safest choice
18037 is to use exactly the same options when generating and using the
18038 precompiled header. The following are known to be safe:
18040 @gccoptlist{-fmessage-length= -fpreprocessed -fsched-interblock @gol
18041 -fsched-spec -fsched-spec-load -fsched-spec-load-dangerous @gol
18042 -fsched-verbose=<number> -fschedule-insns -fvisibility= @gol
18047 For all of these except the last, the compiler will automatically
18048 ignore the precompiled header if the conditions aren't met. If you
18049 find an option combination that doesn't work and doesn't cause the
18050 precompiled header to be ignored, please consider filing a bug report,
18053 If you do use differing options when generating and using the
18054 precompiled header, the actual behavior will be a mixture of the
18055 behavior for the options. For instance, if you use @option{-g} to
18056 generate the precompiled header but not when using it, you may or may
18057 not get debugging information for routines in the precompiled header.