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 1.2 or
19 any later version published by the Free Software Foundation; with the
20 Invariant Sections being ``GNU General Public License'' and ``Funding
21 Free Software'', the Front-Cover texts being (a) (see below), and with
22 the Back-Cover Texts being (b) (see below). A copy of the license is
23 included in the gfdl(7) man page.
25 (a) The FSF's Front-Cover Text is:
29 (b) The FSF's Back-Cover Text is:
31 You have freedom to copy and modify this GNU Manual, like GNU
32 software. Copies published by the Free Software Foundation raise
33 funds for GNU development.
35 @c Set file name and title for the man page.
37 @settitle GNU project C and C++ compiler
39 gcc [@option{-c}|@option{-S}|@option{-E}] [@option{-std=}@var{standard}]
40 [@option{-g}] [@option{-pg}] [@option{-O}@var{level}]
41 [@option{-W}@var{warn}@dots{}] [@option{-pedantic}]
42 [@option{-I}@var{dir}@dots{}] [@option{-L}@var{dir}@dots{}]
43 [@option{-D}@var{macro}[=@var{defn}]@dots{}] [@option{-U}@var{macro}]
44 [@option{-f}@var{option}@dots{}] [@option{-m}@var{machine-option}@dots{}]
45 [@option{-o} @var{outfile}] [@@@var{file}] @var{infile}@dots{}
47 Only the most useful options are listed here; see below for the
48 remainder. @samp{g++} accepts mostly the same options as @samp{gcc}.
51 gpl(7), gfdl(7), fsf-funding(7),
52 cpp(1), gcov(1), as(1), ld(1), gdb(1), adb(1), dbx(1), sdb(1)
53 and the Info entries for @file{gcc}, @file{cpp}, @file{as},
54 @file{ld}, @file{binutils} and @file{gdb}.
57 For instructions on reporting bugs, see
61 See the Info entry for @command{gcc}, or
62 @w{@uref{http://gcc.gnu.org/onlinedocs/gcc/Contributors.html}},
63 for contributors to GCC@.
68 @chapter GCC Command Options
69 @cindex GCC command options
70 @cindex command options
71 @cindex options, GCC command
73 @c man begin DESCRIPTION
74 When you invoke GCC, it normally does preprocessing, compilation,
75 assembly and linking. The ``overall options'' allow you to stop this
76 process at an intermediate stage. For example, the @option{-c} option
77 says not to run the linker. Then the output consists of object files
78 output by the assembler.
80 Other options are passed on to one stage of processing. Some options
81 control the preprocessor and others the compiler itself. Yet other
82 options control the assembler and linker; most of these are not
83 documented here, since you rarely need to use any of them.
85 @cindex C compilation options
86 Most of the command line options that you can use with GCC are useful
87 for C programs; when an option is only useful with another language
88 (usually C++), the explanation says so explicitly. If the description
89 for a particular option does not mention a source language, you can use
90 that option with all supported languages.
92 @cindex C++ compilation options
93 @xref{Invoking G++,,Compiling C++ Programs}, for a summary of special
94 options for compiling C++ programs.
96 @cindex grouping options
97 @cindex options, grouping
98 The @command{gcc} program accepts options and file names as operands. Many
99 options have multi-letter names; therefore multiple single-letter options
100 may @emph{not} be grouped: @option{-dv} is very different from @w{@samp{-d
103 @cindex order of options
104 @cindex options, order
105 You can mix options and other arguments. For the most part, the order
106 you use doesn't matter. Order does matter when you use several
107 options of the same kind; for example, if you specify @option{-L} more
108 than once, the directories are searched in the order specified. Also,
109 the placement of the @option{-l} option is significant.
111 Many options have long names starting with @samp{-f} or with
112 @samp{-W}---for example,
113 @option{-fmove-loop-invariants}, @option{-Wformat} and so on. Most of
114 these have both positive and negative forms; the negative form of
115 @option{-ffoo} would be @option{-fno-foo}. This manual documents
116 only one of these two forms, whichever one is not the default.
120 @xref{Option Index}, for an index to GCC's options.
123 * Option Summary:: Brief list of all options, without explanations.
124 * Overall Options:: Controlling the kind of output:
125 an executable, object files, assembler files,
126 or preprocessed source.
127 * Invoking G++:: Compiling C++ programs.
128 * C Dialect Options:: Controlling the variant of C language compiled.
129 * C++ Dialect Options:: Variations on C++.
130 * Objective-C and Objective-C++ Dialect Options:: Variations on Objective-C
132 * Language Independent Options:: Controlling how diagnostics should be
134 * Warning Options:: How picky should the compiler be?
135 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
136 * Optimize Options:: How much optimization?
137 * Preprocessor Options:: Controlling header files and macro definitions.
138 Also, getting dependency information for Make.
139 * Assembler Options:: Passing options to the assembler.
140 * Link Options:: Specifying libraries and so on.
141 * Directory Options:: Where to find header files and libraries.
142 Where to find the compiler executable files.
143 * Spec Files:: How to pass switches to sub-processes.
144 * Target Options:: Running a cross-compiler, or an old version of GCC.
145 * Submodel Options:: Specifying minor hardware or convention variations,
146 such as 68010 vs 68020.
147 * Code Gen Options:: Specifying conventions for function calls, data layout
149 * Environment Variables:: Env vars that affect GCC.
150 * Precompiled Headers:: Compiling a header once, and using it many times.
156 @section Option Summary
158 Here is a summary of all the options, grouped by type. Explanations are
159 in the following sections.
162 @item Overall Options
163 @xref{Overall Options,,Options Controlling the Kind of Output}.
164 @gccoptlist{-c -S -E -o @var{file} -combine -no-canonical-prefixes @gol
165 -pipe -pass-exit-codes @gol
166 -x @var{language} -v -### --help@r{[}=@var{class}@r{[},@dots{}@r{]]} --target-help @gol
167 --version -wrapper@@@var{file} -fplugin=@var{file} -fplugin-arg-@var{name}=@var{arg}}
169 @item C Language Options
170 @xref{C Dialect Options,,Options Controlling C Dialect}.
171 @gccoptlist{-ansi -std=@var{standard} -fgnu89-inline @gol
172 -aux-info @var{filename} @gol
173 -fno-asm -fno-builtin -fno-builtin-@var{function} @gol
174 -fhosted -ffreestanding -fopenmp -fms-extensions @gol
175 -trigraphs -no-integrated-cpp -traditional -traditional-cpp @gol
176 -fallow-single-precision -fcond-mismatch -flax-vector-conversions @gol
177 -fsigned-bitfields -fsigned-char @gol
178 -funsigned-bitfields -funsigned-char}
180 @item C++ Language Options
181 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}.
182 @gccoptlist{-fabi-version=@var{n} -fno-access-control -fcheck-new @gol
183 -fconserve-space -ffriend-injection @gol
184 -fno-elide-constructors @gol
185 -fno-enforce-eh-specs @gol
186 -ffor-scope -fno-for-scope -fno-gnu-keywords @gol
187 -fno-implicit-templates @gol
188 -fno-implicit-inline-templates @gol
189 -fno-implement-inlines -fms-extensions @gol
190 -fno-nonansi-builtins -fno-operator-names @gol
191 -fno-optional-diags -fpermissive @gol
192 -fno-pretty-templates @gol
193 -frepo -fno-rtti -fstats -ftemplate-depth=@var{n} @gol
194 -fno-threadsafe-statics -fuse-cxa-atexit -fno-weak -nostdinc++ @gol
195 -fno-default-inline -fvisibility-inlines-hidden @gol
196 -fvisibility-ms-compat @gol
197 -Wabi -Wconversion-null -Wctor-dtor-privacy @gol
198 -Wnon-virtual-dtor -Wreorder @gol
199 -Weffc++ -Wstrict-null-sentinel @gol
200 -Wno-non-template-friend -Wold-style-cast @gol
201 -Woverloaded-virtual -Wno-pmf-conversions @gol
204 @item Objective-C and Objective-C++ Language Options
205 @xref{Objective-C and Objective-C++ Dialect Options,,Options Controlling
206 Objective-C and Objective-C++ Dialects}.
207 @gccoptlist{-fconstant-string-class=@var{class-name} @gol
208 -fgnu-runtime -fnext-runtime @gol
209 -fno-nil-receivers @gol
210 -fobjc-call-cxx-cdtors @gol
211 -fobjc-direct-dispatch @gol
212 -fobjc-exceptions @gol
214 -freplace-objc-classes @gol
217 -Wassign-intercept @gol
218 -Wno-protocol -Wselector @gol
219 -Wstrict-selector-match @gol
220 -Wundeclared-selector}
222 @item Language Independent Options
223 @xref{Language Independent Options,,Options to Control Diagnostic Messages Formatting}.
224 @gccoptlist{-fmessage-length=@var{n} @gol
225 -fdiagnostics-show-location=@r{[}once@r{|}every-line@r{]} @gol
226 -fdiagnostics-show-option}
228 @item Warning Options
229 @xref{Warning Options,,Options to Request or Suppress Warnings}.
230 @gccoptlist{-fsyntax-only -pedantic -pedantic-errors @gol
231 -w -Wextra -Wall -Waddress -Waggregate-return -Warray-bounds @gol
232 -Wno-attributes -Wno-builtin-macro-redefined @gol
233 -Wc++-compat -Wc++0x-compat -Wcast-align -Wcast-qual @gol
234 -Wchar-subscripts -Wclobbered -Wcomment @gol
235 -Wconversion -Wcoverage-mismatch -Wno-deprecated @gol
236 -Wno-deprecated-declarations -Wdisabled-optimization @gol
237 -Wno-div-by-zero -Wempty-body -Wenum-compare -Wno-endif-labels @gol
238 -Werror -Werror=* @gol
239 -Wfatal-errors -Wfloat-equal -Wformat -Wformat=2 @gol
240 -Wno-format-contains-nul -Wno-format-extra-args -Wformat-nonliteral @gol
241 -Wformat-security -Wformat-y2k @gol
242 -Wframe-larger-than=@var{len} -Wjump-misses-init -Wignored-qualifiers @gol
243 -Wimplicit -Wimplicit-function-declaration -Wimplicit-int @gol
244 -Winit-self -Winline @gol
245 -Wno-int-to-pointer-cast -Wno-invalid-offsetof @gol
246 -Winvalid-pch -Wlarger-than=@var{len} -Wunsafe-loop-optimizations @gol
247 -Wlogical-op -Wlong-long @gol
248 -Wmain -Wmissing-braces -Wmissing-field-initializers @gol
249 -Wmissing-format-attribute -Wmissing-include-dirs @gol
250 -Wmissing-noreturn -Wno-mudflap @gol
251 -Wno-multichar -Wnonnull -Wno-overflow @gol
252 -Woverlength-strings -Wpacked -Wpacked-bitfield-compat -Wpadded @gol
253 -Wparentheses -Wpedantic-ms-format -Wno-pedantic-ms-format @gol
254 -Wpointer-arith -Wno-pointer-to-int-cast @gol
255 -Wredundant-decls @gol
256 -Wreturn-type -Wsequence-point -Wshadow @gol
257 -Wsign-compare -Wsign-conversion -Wstack-protector @gol
258 -Wstrict-aliasing -Wstrict-aliasing=n @gol
259 -Wstrict-overflow -Wstrict-overflow=@var{n} @gol
260 -Wswitch -Wswitch-default -Wswitch-enum -Wsync-nand @gol
261 -Wsystem-headers -Wtrigraphs -Wtype-limits -Wundef -Wuninitialized @gol
262 -Wunknown-pragmas -Wno-pragmas @gol
263 -Wunsuffixed-float-constants -Wunused -Wunused-function @gol
264 -Wunused-label -Wunused-parameter -Wno-unused-result -Wunused-value -Wunused-variable @gol
265 -Wvariadic-macros -Wvla @gol
266 -Wvolatile-register-var -Wwrite-strings}
268 @item C and Objective-C-only Warning Options
269 @gccoptlist{-Wbad-function-cast -Wmissing-declarations @gol
270 -Wmissing-parameter-type -Wmissing-prototypes -Wnested-externs @gol
271 -Wold-style-declaration -Wold-style-definition @gol
272 -Wstrict-prototypes -Wtraditional -Wtraditional-conversion @gol
273 -Wdeclaration-after-statement -Wpointer-sign}
275 @item Debugging Options
276 @xref{Debugging Options,,Options for Debugging Your Program or GCC}.
277 @gccoptlist{-d@var{letters} -dumpspecs -dumpmachine -dumpversion @gol
278 -fdbg-cnt-list -fdbg-cnt=@var{counter-value-list} @gol
279 -fdump-noaddr -fdump-unnumbered -fdump-unnumbered-links @gol
280 -fdump-translation-unit@r{[}-@var{n}@r{]} @gol
281 -fdump-class-hierarchy@r{[}-@var{n}@r{]} @gol
282 -fdump-ipa-all -fdump-ipa-cgraph -fdump-ipa-inline @gol
283 -fdump-statistics @gol
285 -fdump-tree-original@r{[}-@var{n}@r{]} @gol
286 -fdump-tree-optimized@r{[}-@var{n}@r{]} @gol
287 -fdump-tree-cfg -fdump-tree-vcg -fdump-tree-alias @gol
289 -fdump-tree-ssa@r{[}-@var{n}@r{]} -fdump-tree-pre@r{[}-@var{n}@r{]} @gol
290 -fdump-tree-ccp@r{[}-@var{n}@r{]} -fdump-tree-dce@r{[}-@var{n}@r{]} @gol
291 -fdump-tree-gimple@r{[}-raw@r{]} -fdump-tree-mudflap@r{[}-@var{n}@r{]} @gol
292 -fdump-tree-dom@r{[}-@var{n}@r{]} @gol
293 -fdump-tree-dse@r{[}-@var{n}@r{]} @gol
294 -fdump-tree-phiprop@r{[}-@var{n}@r{]} @gol
295 -fdump-tree-phiopt@r{[}-@var{n}@r{]} @gol
296 -fdump-tree-forwprop@r{[}-@var{n}@r{]} @gol
297 -fdump-tree-copyrename@r{[}-@var{n}@r{]} @gol
298 -fdump-tree-nrv -fdump-tree-vect @gol
299 -fdump-tree-sink @gol
300 -fdump-tree-sra@r{[}-@var{n}@r{]} @gol
301 -fdump-tree-forwprop@r{[}-@var{n}@r{]} @gol
302 -fdump-tree-fre@r{[}-@var{n}@r{]} @gol
303 -fdump-tree-vrp@r{[}-@var{n}@r{]} @gol
304 -ftree-vectorizer-verbose=@var{n} @gol
305 -fdump-tree-storeccp@r{[}-@var{n}@r{]} @gol
306 -fdump-final-insns=@var{file} @gol
307 -fcompare-debug@r{[}=@var{opts}@r{]} -fcompare-debug-second @gol
308 -feliminate-dwarf2-dups -feliminate-unused-debug-types @gol
309 -feliminate-unused-debug-symbols -femit-class-debug-always @gol
310 -fenable-icf-debug @gol
311 -fmem-report -fpre-ipa-mem-report -fpost-ipa-mem-report -fprofile-arcs @gol
312 -frandom-seed=@var{string} -fsched-verbose=@var{n} @gol
313 -fsel-sched-verbose -fsel-sched-dump-cfg -fsel-sched-pipelining-verbose @gol
314 -ftest-coverage -ftime-report -fvar-tracking @gol
315 -fvar-tracking-assignments -fvar-tracking-assignments-toggle @gol
316 -g -g@var{level} -gtoggle -gcoff -gdwarf-@var{version} @gol
317 -ggdb -gstabs -gstabs+ -gstrict-dwarf -gno-strict-dwarf @gol
318 -gvms -gxcoff -gxcoff+ @gol
319 -fno-merge-debug-strings -fno-dwarf2-cfi-asm @gol
320 -fdebug-prefix-map=@var{old}=@var{new} @gol
321 -femit-struct-debug-baseonly -femit-struct-debug-reduced @gol
322 -femit-struct-debug-detailed@r{[}=@var{spec-list}@r{]} @gol
323 -p -pg -print-file-name=@var{library} -print-libgcc-file-name @gol
324 -print-multi-directory -print-multi-lib -print-multi-os-directory @gol
325 -print-prog-name=@var{program} -print-search-dirs -Q @gol
326 -print-sysroot -print-sysroot-headers-suffix @gol
327 -save-temps -save-temps=cwd -save-temps=obj -time@r{[}=@var{file}@r{]}}
329 @item Optimization Options
330 @xref{Optimize Options,,Options that Control Optimization}.
332 -falign-functions[=@var{n}] -falign-jumps[=@var{n}] @gol
333 -falign-labels[=@var{n}] -falign-loops[=@var{n}] -fassociative-math @gol
334 -fauto-inc-dec -fbranch-probabilities -fbranch-target-load-optimize @gol
335 -fbranch-target-load-optimize2 -fbtr-bb-exclusive -fcaller-saves @gol
336 -fcheck-data-deps -fconserve-stack -fcprop-registers -fcrossjumping @gol
337 -fcse-follow-jumps -fcse-skip-blocks -fcx-fortran-rules -fcx-limited-range @gol
338 -fdata-sections -fdce -fdce @gol
339 -fdelayed-branch -fdelete-null-pointer-checks -fdse -fdse @gol
340 -fearly-inlining -fipa-sra -fexpensive-optimizations -ffast-math @gol
341 -ffinite-math-only -ffloat-store -fexcess-precision=@var{style} @gol
342 -fforward-propagate -ffunction-sections @gol
343 -fgcse -fgcse-after-reload -fgcse-las -fgcse-lm @gol
344 -fgcse-sm -fif-conversion -fif-conversion2 -findirect-inlining @gol
345 -finline-functions -finline-functions-called-once -finline-limit=@var{n} @gol
346 -finline-small-functions -fipa-cp -fipa-cp-clone -fipa-matrix-reorg -fipa-pta @gol
347 -fipa-pure-const -fipa-reference -fipa-struct-reorg @gol
348 -fipa-type-escape -fira-algorithm=@var{algorithm} @gol
349 -fira-region=@var{region} -fira-coalesce @gol
350 -fira-loop-pressure -fno-ira-share-save-slots @gol
351 -fno-ira-share-spill-slots -fira-verbose=@var{n} @gol
352 -fivopts -fkeep-inline-functions -fkeep-static-consts @gol
353 -floop-block -floop-interchange -floop-strip-mine -fgraphite-identity @gol
354 -floop-parallelize-all -flto -flto-compression-level -flto-report -fltrans @gol
355 -fltrans-output-list -fmerge-all-constants -fmerge-constants -fmodulo-sched @gol
356 -fmodulo-sched-allow-regmoves -fmove-loop-invariants -fmudflap @gol
357 -fmudflapir -fmudflapth -fno-branch-count-reg -fno-default-inline @gol
358 -fno-defer-pop -fno-function-cse -fno-guess-branch-probability @gol
359 -fno-inline -fno-math-errno -fno-peephole -fno-peephole2 @gol
360 -fno-sched-interblock -fno-sched-spec -fno-signed-zeros @gol
361 -fno-toplevel-reorder -fno-trapping-math -fno-zero-initialized-in-bss @gol
362 -fomit-frame-pointer -foptimize-register-move -foptimize-sibling-calls @gol
363 -fpeel-loops -fpredictive-commoning -fprefetch-loop-arrays @gol
364 -fprofile-correction -fprofile-dir=@var{path} -fprofile-generate @gol
365 -fprofile-generate=@var{path} @gol
366 -fprofile-use -fprofile-use=@var{path} -fprofile-values @gol
367 -freciprocal-math -fregmove -frename-registers -freorder-blocks @gol
368 -freorder-blocks-and-partition -freorder-functions @gol
369 -frerun-cse-after-loop -freschedule-modulo-scheduled-loops @gol
370 -frounding-math -fsched2-use-superblocks -fsched-pressure @gol
371 -fsched-spec-load -fsched-spec-load-dangerous @gol
372 -fsched-stalled-insns-dep[=@var{n}] -fsched-stalled-insns[=@var{n}] @gol
373 -fsched-group-heuristic -fsched-critical-path-heuristic @gol
374 -fsched-spec-insn-heuristic -fsched-rank-heuristic @gol
375 -fsched-last-insn-heuristic -fsched-dep-count-heuristic @gol
376 -fschedule-insns -fschedule-insns2 -fsection-anchors @gol
377 -fselective-scheduling -fselective-scheduling2 @gol
378 -fsel-sched-pipelining -fsel-sched-pipelining-outer-loops @gol
379 -fsignaling-nans -fsingle-precision-constant -fsplit-ivs-in-unroller @gol
380 -fsplit-wide-types -fstack-protector -fstack-protector-all @gol
381 -fstrict-aliasing -fstrict-overflow -fthread-jumps -ftracer @gol
382 -ftree-builtin-call-dce -ftree-ccp -ftree-ch -ftree-copy-prop @gol
383 -ftree-copyrename -ftree-dce @gol
384 -ftree-dominator-opts -ftree-dse -ftree-forwprop -ftree-fre -ftree-loop-im @gol
385 -ftree-phiprop -ftree-loop-distribution @gol
386 -ftree-loop-ivcanon -ftree-loop-linear -ftree-loop-optimize @gol
387 -ftree-parallelize-loops=@var{n} -ftree-pre -ftree-pta -ftree-reassoc @gol
388 -ftree-sink -ftree-sra -ftree-switch-conversion @gol
389 -ftree-ter -ftree-vect-loop-version -ftree-vectorize -ftree-vrp @gol
390 -funit-at-a-time -funroll-all-loops -funroll-loops @gol
391 -funsafe-loop-optimizations -funsafe-math-optimizations -funswitch-loops @gol
392 -fvariable-expansion-in-unroller -fvect-cost-model -fvpt -fweb @gol
393 -fwhole-program -fwhopr -fwpa -fuse-linker-plugin @gol
394 --param @var{name}=@var{value}
395 -O -O0 -O1 -O2 -O3 -Os}
397 @item Preprocessor Options
398 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
399 @gccoptlist{-A@var{question}=@var{answer} @gol
400 -A-@var{question}@r{[}=@var{answer}@r{]} @gol
401 -C -dD -dI -dM -dN @gol
402 -D@var{macro}@r{[}=@var{defn}@r{]} -E -H @gol
403 -idirafter @var{dir} @gol
404 -include @var{file} -imacros @var{file} @gol
405 -iprefix @var{file} -iwithprefix @var{dir} @gol
406 -iwithprefixbefore @var{dir} -isystem @var{dir} @gol
407 -imultilib @var{dir} -isysroot @var{dir} @gol
408 -M -MM -MF -MG -MP -MQ -MT -nostdinc @gol
409 -P -fworking-directory -remap @gol
410 -trigraphs -undef -U@var{macro} -Wp,@var{option} @gol
411 -Xpreprocessor @var{option}}
413 @item Assembler Option
414 @xref{Assembler Options,,Passing Options to the Assembler}.
415 @gccoptlist{-Wa,@var{option} -Xassembler @var{option}}
418 @xref{Link Options,,Options for Linking}.
419 @gccoptlist{@var{object-file-name} -l@var{library} @gol
420 -nostartfiles -nodefaultlibs -nostdlib -pie -rdynamic @gol
421 -s -static -static-libgcc -static-libstdc++ -shared @gol
422 -shared-libgcc -symbolic @gol
423 -T @var{script} -Wl,@var{option} -Xlinker @var{option} @gol
426 @item Directory Options
427 @xref{Directory Options,,Options for Directory Search}.
428 @gccoptlist{-B@var{prefix} -I@var{dir} -iquote@var{dir} -L@var{dir}
429 -specs=@var{file} -I- --sysroot=@var{dir}}
432 @c I wrote this xref this way to avoid overfull hbox. -- rms
433 @xref{Target Options}.
434 @gccoptlist{-V @var{version} -b @var{machine}}
436 @item Machine Dependent Options
437 @xref{Submodel Options,,Hardware Models and Configurations}.
438 @c This list is ordered alphanumerically by subsection name.
439 @c Try and put the significant identifier (CPU or system) first,
440 @c so users have a clue at guessing where the ones they want will be.
443 @gccoptlist{-EB -EL @gol
444 -mmangle-cpu -mcpu=@var{cpu} -mtext=@var{text-section} @gol
445 -mdata=@var{data-section} -mrodata=@var{readonly-data-section}}
448 @gccoptlist{-mapcs-frame -mno-apcs-frame @gol
449 -mabi=@var{name} @gol
450 -mapcs-stack-check -mno-apcs-stack-check @gol
451 -mapcs-float -mno-apcs-float @gol
452 -mapcs-reentrant -mno-apcs-reentrant @gol
453 -msched-prolog -mno-sched-prolog @gol
454 -mlittle-endian -mbig-endian -mwords-little-endian @gol
455 -mfloat-abi=@var{name} -msoft-float -mhard-float -mfpe @gol
456 -mfp16-format=@var{name}
457 -mthumb-interwork -mno-thumb-interwork @gol
458 -mcpu=@var{name} -march=@var{name} -mfpu=@var{name} @gol
459 -mstructure-size-boundary=@var{n} @gol
460 -mabort-on-noreturn @gol
461 -mlong-calls -mno-long-calls @gol
462 -msingle-pic-base -mno-single-pic-base @gol
463 -mpic-register=@var{reg} @gol
464 -mnop-fun-dllimport @gol
465 -mcirrus-fix-invalid-insns -mno-cirrus-fix-invalid-insns @gol
466 -mpoke-function-name @gol
468 -mtpcs-frame -mtpcs-leaf-frame @gol
469 -mcaller-super-interworking -mcallee-super-interworking @gol
471 -mword-relocations @gol
472 -mfix-cortex-m3-ldrd}
475 @gccoptlist{-mmcu=@var{mcu} -mno-interrupts @gol
476 -mcall-prologues -mtiny-stack -mint8}
478 @emph{Blackfin Options}
479 @gccoptlist{-mcpu=@var{cpu}@r{[}-@var{sirevision}@r{]} @gol
480 -msim -momit-leaf-frame-pointer -mno-omit-leaf-frame-pointer @gol
481 -mspecld-anomaly -mno-specld-anomaly -mcsync-anomaly -mno-csync-anomaly @gol
482 -mlow-64k -mno-low64k -mstack-check-l1 -mid-shared-library @gol
483 -mno-id-shared-library -mshared-library-id=@var{n} @gol
484 -mleaf-id-shared-library -mno-leaf-id-shared-library @gol
485 -msep-data -mno-sep-data -mlong-calls -mno-long-calls @gol
486 -mfast-fp -minline-plt -mmulticore -mcorea -mcoreb -msdram @gol
490 @gccoptlist{-mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol
491 -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol
492 -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol
493 -mstack-align -mdata-align -mconst-align @gol
494 -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol
495 -melf -maout -melinux -mlinux -sim -sim2 @gol
496 -mmul-bug-workaround -mno-mul-bug-workaround}
499 @gccoptlist{-mmac -mpush-args}
501 @emph{Darwin Options}
502 @gccoptlist{-all_load -allowable_client -arch -arch_errors_fatal @gol
503 -arch_only -bind_at_load -bundle -bundle_loader @gol
504 -client_name -compatibility_version -current_version @gol
506 -dependency-file -dylib_file -dylinker_install_name @gol
507 -dynamic -dynamiclib -exported_symbols_list @gol
508 -filelist -flat_namespace -force_cpusubtype_ALL @gol
509 -force_flat_namespace -headerpad_max_install_names @gol
511 -image_base -init -install_name -keep_private_externs @gol
512 -multi_module -multiply_defined -multiply_defined_unused @gol
513 -noall_load -no_dead_strip_inits_and_terms @gol
514 -nofixprebinding -nomultidefs -noprebind -noseglinkedit @gol
515 -pagezero_size -prebind -prebind_all_twolevel_modules @gol
516 -private_bundle -read_only_relocs -sectalign @gol
517 -sectobjectsymbols -whyload -seg1addr @gol
518 -sectcreate -sectobjectsymbols -sectorder @gol
519 -segaddr -segs_read_only_addr -segs_read_write_addr @gol
520 -seg_addr_table -seg_addr_table_filename -seglinkedit @gol
521 -segprot -segs_read_only_addr -segs_read_write_addr @gol
522 -single_module -static -sub_library -sub_umbrella @gol
523 -twolevel_namespace -umbrella -undefined @gol
524 -unexported_symbols_list -weak_reference_mismatches @gol
525 -whatsloaded -F -gused -gfull -mmacosx-version-min=@var{version} @gol
526 -mkernel -mone-byte-bool}
528 @emph{DEC Alpha Options}
529 @gccoptlist{-mno-fp-regs -msoft-float -malpha-as -mgas @gol
530 -mieee -mieee-with-inexact -mieee-conformant @gol
531 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol
532 -mtrap-precision=@var{mode} -mbuild-constants @gol
533 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} @gol
534 -mbwx -mmax -mfix -mcix @gol
535 -mfloat-vax -mfloat-ieee @gol
536 -mexplicit-relocs -msmall-data -mlarge-data @gol
537 -msmall-text -mlarge-text @gol
538 -mmemory-latency=@var{time}}
540 @emph{DEC Alpha/VMS Options}
541 @gccoptlist{-mvms-return-codes -mdebug-main=@var{prefix} -mmalloc64}
544 @gccoptlist{-msmall-model -mno-lsim}
547 @gccoptlist{-mgpr-32 -mgpr-64 -mfpr-32 -mfpr-64 @gol
548 -mhard-float -msoft-float @gol
549 -malloc-cc -mfixed-cc -mdword -mno-dword @gol
550 -mdouble -mno-double @gol
551 -mmedia -mno-media -mmuladd -mno-muladd @gol
552 -mfdpic -minline-plt -mgprel-ro -multilib-library-pic @gol
553 -mlinked-fp -mlong-calls -malign-labels @gol
554 -mlibrary-pic -macc-4 -macc-8 @gol
555 -mpack -mno-pack -mno-eflags -mcond-move -mno-cond-move @gol
556 -moptimize-membar -mno-optimize-membar @gol
557 -mscc -mno-scc -mcond-exec -mno-cond-exec @gol
558 -mvliw-branch -mno-vliw-branch @gol
559 -mmulti-cond-exec -mno-multi-cond-exec -mnested-cond-exec @gol
560 -mno-nested-cond-exec -mtomcat-stats @gol
564 @emph{GNU/Linux Options}
565 @gccoptlist{-muclibc}
567 @emph{H8/300 Options}
568 @gccoptlist{-mrelax -mh -ms -mn -mint32 -malign-300}
571 @gccoptlist{-march=@var{architecture-type} @gol
572 -mbig-switch -mdisable-fpregs -mdisable-indexing @gol
573 -mfast-indirect-calls -mgas -mgnu-ld -mhp-ld @gol
574 -mfixed-range=@var{register-range} @gol
575 -mjump-in-delay -mlinker-opt -mlong-calls @gol
576 -mlong-load-store -mno-big-switch -mno-disable-fpregs @gol
577 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
578 -mno-jump-in-delay -mno-long-load-store @gol
579 -mno-portable-runtime -mno-soft-float @gol
580 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
581 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
582 -mschedule=@var{cpu-type} -mspace-regs -msio -mwsio @gol
583 -munix=@var{unix-std} -nolibdld -static -threads}
585 @emph{i386 and x86-64 Options}
586 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
587 -mfpmath=@var{unit} @gol
588 -masm=@var{dialect} -mno-fancy-math-387 @gol
589 -mno-fp-ret-in-387 -msoft-float @gol
590 -mno-wide-multiply -mrtd -malign-double @gol
591 -mpreferred-stack-boundary=@var{num}
592 -mincoming-stack-boundary=@var{num}
593 -mcld -mcx16 -msahf -mmovbe -mcrc32 -mrecip @gol
594 -mmmx -msse -msse2 -msse3 -mssse3 -msse4.1 -msse4.2 -msse4 -mavx @gol
595 -maes -mpclmul -mfused-madd @gol
596 -msse4a -m3dnow -mpopcnt -mabm -mfma4 -mxop -mlwp @gol
597 -mthreads -mno-align-stringops -minline-all-stringops @gol
598 -minline-stringops-dynamically -mstringop-strategy=@var{alg} @gol
599 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
600 -m96bit-long-double -mregparm=@var{num} -msseregparm @gol
601 -mveclibabi=@var{type} -mpc32 -mpc64 -mpc80 -mstackrealign @gol
602 -momit-leaf-frame-pointer -mno-red-zone -mno-tls-direct-seg-refs @gol
603 -mcmodel=@var{code-model} -mabi=@var{name} @gol
604 -m32 -m64 -mlarge-data-threshold=@var{num} @gol
608 @gccoptlist{-mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
609 -mvolatile-asm-stop -mregister-names -msdata -mno-sdata @gol
610 -mconstant-gp -mauto-pic -mfused-madd @gol
611 -minline-float-divide-min-latency @gol
612 -minline-float-divide-max-throughput @gol
613 -mno-inline-float-divide @gol
614 -minline-int-divide-min-latency @gol
615 -minline-int-divide-max-throughput @gol
616 -mno-inline-int-divide @gol
617 -minline-sqrt-min-latency -minline-sqrt-max-throughput @gol
618 -mno-inline-sqrt @gol
619 -mdwarf2-asm -mearly-stop-bits @gol
620 -mfixed-range=@var{register-range} -mtls-size=@var{tls-size} @gol
621 -mtune=@var{cpu-type} -milp32 -mlp64 @gol
622 -msched-br-data-spec -msched-ar-data-spec -msched-control-spec @gol
623 -msched-br-in-data-spec -msched-ar-in-data-spec -msched-in-control-spec @gol
624 -msched-spec-ldc -msched-spec-control-ldc @gol
625 -msched-prefer-non-data-spec-insns -msched-prefer-non-control-spec-insns @gol
626 -msched-stop-bits-after-every-cycle -msched-count-spec-in-critical-path @gol
627 -msel-sched-dont-check-control-spec -msched-fp-mem-deps-zero-cost @gol
628 -msched-max-memory-insns-hard-limit -msched-max-memory-insns=@var{max-insns}}
630 @emph{IA-64/VMS Options}
631 @gccoptlist{-mvms-return-codes -mdebug-main=@var{prefix} -mmalloc64}
634 @gccoptlist{-mbarrel-shift-enabled -mdivide-enabled -mmultiply-enabled @gol
635 -msign-extend-enabled -muser-enabled}
637 @emph{M32R/D Options}
638 @gccoptlist{-m32r2 -m32rx -m32r @gol
640 -malign-loops -mno-align-loops @gol
641 -missue-rate=@var{number} @gol
642 -mbranch-cost=@var{number} @gol
643 -mmodel=@var{code-size-model-type} @gol
644 -msdata=@var{sdata-type} @gol
645 -mno-flush-func -mflush-func=@var{name} @gol
646 -mno-flush-trap -mflush-trap=@var{number} @gol
650 @gccoptlist{-mcpu=@var{cpu} -msim -memregs=@var{number}}
652 @emph{M680x0 Options}
653 @gccoptlist{-march=@var{arch} -mcpu=@var{cpu} -mtune=@var{tune}
654 -m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
655 -m68060 -mcpu32 -m5200 -m5206e -m528x -m5307 -m5407 @gol
656 -mcfv4e -mbitfield -mno-bitfield -mc68000 -mc68020 @gol
657 -mnobitfield -mrtd -mno-rtd -mdiv -mno-div -mshort @gol
658 -mno-short -mhard-float -m68881 -msoft-float -mpcrel @gol
659 -malign-int -mstrict-align -msep-data -mno-sep-data @gol
660 -mshared-library-id=n -mid-shared-library -mno-id-shared-library @gol
663 @emph{M68hc1x Options}
664 @gccoptlist{-m6811 -m6812 -m68hc11 -m68hc12 -m68hcs12 @gol
665 -mauto-incdec -minmax -mlong-calls -mshort @gol
666 -msoft-reg-count=@var{count}}
669 @gccoptlist{-mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
670 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
671 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
672 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
673 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
676 @gccoptlist{-mabsdiff -mall-opts -maverage -mbased=@var{n} -mbitops @gol
677 -mc=@var{n} -mclip -mconfig=@var{name} -mcop -mcop32 -mcop64 -mivc2 @gol
678 -mdc -mdiv -meb -mel -mio-volatile -ml -mleadz -mm -mminmax @gol
679 -mmult -mno-opts -mrepeat -ms -msatur -msdram -msim -msimnovec -mtf @gol
683 @gccoptlist{-EL -EB -march=@var{arch} -mtune=@var{arch} @gol
684 -mips1 -mips2 -mips3 -mips4 -mips32 -mips32r2 @gol
685 -mips64 -mips64r2 @gol
686 -mips16 -mno-mips16 -mflip-mips16 @gol
687 -minterlink-mips16 -mno-interlink-mips16 @gol
688 -mabi=@var{abi} -mabicalls -mno-abicalls @gol
689 -mshared -mno-shared -mplt -mno-plt -mxgot -mno-xgot @gol
690 -mgp32 -mgp64 -mfp32 -mfp64 -mhard-float -msoft-float @gol
691 -msingle-float -mdouble-float -mdsp -mno-dsp -mdspr2 -mno-dspr2 @gol
692 -mfpu=@var{fpu-type} @gol
693 -msmartmips -mno-smartmips @gol
694 -mpaired-single -mno-paired-single -mdmx -mno-mdmx @gol
695 -mips3d -mno-mips3d -mmt -mno-mt -mllsc -mno-llsc @gol
696 -mlong64 -mlong32 -msym32 -mno-sym32 @gol
697 -G@var{num} -mlocal-sdata -mno-local-sdata @gol
698 -mextern-sdata -mno-extern-sdata -mgpopt -mno-gopt @gol
699 -membedded-data -mno-embedded-data @gol
700 -muninit-const-in-rodata -mno-uninit-const-in-rodata @gol
701 -mcode-readable=@var{setting} @gol
702 -msplit-addresses -mno-split-addresses @gol
703 -mexplicit-relocs -mno-explicit-relocs @gol
704 -mcheck-zero-division -mno-check-zero-division @gol
705 -mdivide-traps -mdivide-breaks @gol
706 -mmemcpy -mno-memcpy -mlong-calls -mno-long-calls @gol
707 -mmad -mno-mad -mfused-madd -mno-fused-madd -nocpp @gol
708 -mfix-r4000 -mno-fix-r4000 -mfix-r4400 -mno-fix-r4400 @gol
709 -mfix-r10000 -mno-fix-r10000 -mfix-vr4120 -mno-fix-vr4120 @gol
710 -mfix-vr4130 -mno-fix-vr4130 -mfix-sb1 -mno-fix-sb1 @gol
711 -mflush-func=@var{func} -mno-flush-func @gol
712 -mbranch-cost=@var{num} -mbranch-likely -mno-branch-likely @gol
713 -mfp-exceptions -mno-fp-exceptions @gol
714 -mvr4130-align -mno-vr4130-align -msynci -mno-synci @gol
715 -mrelax-pic-calls -mno-relax-pic-calls -mmcount-ra-address}
718 @gccoptlist{-mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
719 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
720 -melf -mbranch-predict -mno-branch-predict -mbase-addresses @gol
721 -mno-base-addresses -msingle-exit -mno-single-exit}
723 @emph{MN10300 Options}
724 @gccoptlist{-mmult-bug -mno-mult-bug @gol
725 -mam33 -mno-am33 @gol
726 -mam33-2 -mno-am33-2 @gol
727 -mreturn-pointer-on-d0 @gol
730 @emph{PDP-11 Options}
731 @gccoptlist{-mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol
732 -mbcopy -mbcopy-builtin -mint32 -mno-int16 @gol
733 -mint16 -mno-int32 -mfloat32 -mno-float64 @gol
734 -mfloat64 -mno-float32 -mabshi -mno-abshi @gol
735 -mbranch-expensive -mbranch-cheap @gol
736 -msplit -mno-split -munix-asm -mdec-asm}
738 @emph{picoChip Options}
739 @gccoptlist{-mae=@var{ae_type} -mvliw-lookahead=@var{N}
740 -msymbol-as-address -mno-inefficient-warnings}
742 @emph{PowerPC Options}
743 See RS/6000 and PowerPC Options.
745 @emph{RS/6000 and PowerPC Options}
746 @gccoptlist{-mcpu=@var{cpu-type} @gol
747 -mtune=@var{cpu-type} @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}
790 @gccoptlist{-m64bit-doubles -m32bit-doubles -fpu -nofpu@gol
792 -mbig-endian-data -mlittle-endian-data @gol
795 -mas100-syntax -mno-as100-syntax@gol
797 -mmax-constant-size=@gol
799 -msave-acc-in-interrupts}
801 @emph{S/390 and zSeries Options}
802 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
803 -mhard-float -msoft-float -mhard-dfp -mno-hard-dfp @gol
804 -mlong-double-64 -mlong-double-128 @gol
805 -mbackchain -mno-backchain -mpacked-stack -mno-packed-stack @gol
806 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
807 -m64 -m31 -mdebug -mno-debug -mesa -mzarch @gol
808 -mtpf-trace -mno-tpf-trace -mfused-madd -mno-fused-madd @gol
809 -mwarn-framesize -mwarn-dynamicstack -mstack-size -mstack-guard}
812 @gccoptlist{-meb -mel @gol
816 -mscore5 -mscore5u -mscore7 -mscore7d}
819 @gccoptlist{-m1 -m2 -m2e @gol
820 -m2a-nofpu -m2a-single-only -m2a-single -m2a @gol
822 -m4-nofpu -m4-single-only -m4-single -m4 @gol
823 -m4a-nofpu -m4a-single-only -m4a-single -m4a -m4al @gol
824 -m5-64media -m5-64media-nofpu @gol
825 -m5-32media -m5-32media-nofpu @gol
826 -m5-compact -m5-compact-nofpu @gol
827 -mb -ml -mdalign -mrelax @gol
828 -mbigtable -mfmovd -mhitachi -mrenesas -mno-renesas -mnomacsave @gol
829 -mieee -mbitops -misize -minline-ic_invalidate -mpadstruct -mspace @gol
830 -mprefergot -musermode -multcost=@var{number} -mdiv=@var{strategy} @gol
831 -mdivsi3_libfunc=@var{name} -mfixed-range=@var{register-range} @gol
832 -madjust-unroll -mindexed-addressing -mgettrcost=@var{number} -mpt-fixed @gol
836 @gccoptlist{-mcpu=@var{cpu-type} @gol
837 -mtune=@var{cpu-type} @gol
838 -mcmodel=@var{code-model} @gol
839 -m32 -m64 -mapp-regs -mno-app-regs @gol
840 -mfaster-structs -mno-faster-structs @gol
841 -mfpu -mno-fpu -mhard-float -msoft-float @gol
842 -mhard-quad-float -msoft-quad-float @gol
843 -mimpure-text -mno-impure-text -mlittle-endian @gol
844 -mstack-bias -mno-stack-bias @gol
845 -munaligned-doubles -mno-unaligned-doubles @gol
846 -mv8plus -mno-v8plus -mvis -mno-vis
847 -threads -pthreads -pthread}
850 @gccoptlist{-mwarn-reloc -merror-reloc @gol
851 -msafe-dma -munsafe-dma @gol
853 -msmall-mem -mlarge-mem -mstdmain @gol
854 -mfixed-range=@var{register-range} @gol
856 -maddress-space-conversion -mno-address-space-conversion @gol
857 -mcache-size=@var{cache-size} @gol
858 -matomic-updates -mno-atomic-updates}
860 @emph{System V Options}
861 @gccoptlist{-Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
864 @gccoptlist{-mlong-calls -mno-long-calls -mep -mno-ep @gol
865 -mprolog-function -mno-prolog-function -mspace @gol
866 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
867 -mapp-regs -mno-app-regs @gol
868 -mdisable-callt -mno-disable-callt @gol
874 @gccoptlist{-mg -mgnu -munix}
876 @emph{VxWorks Options}
877 @gccoptlist{-mrtp -non-static -Bstatic -Bdynamic @gol
878 -Xbind-lazy -Xbind-now}
880 @emph{x86-64 Options}
881 See i386 and x86-64 Options.
883 @emph{i386 and x86-64 Windows Options}
884 @gccoptlist{-mconsole -mcygwin -mno-cygwin -mdll
885 -mnop-fun-dllimport -mthread -municode -mwin32 -mwindows
886 -fno-set-stack-executable}
888 @emph{Xstormy16 Options}
891 @emph{Xtensa Options}
892 @gccoptlist{-mconst16 -mno-const16 @gol
893 -mfused-madd -mno-fused-madd @gol
894 -mserialize-volatile -mno-serialize-volatile @gol
895 -mtext-section-literals -mno-text-section-literals @gol
896 -mtarget-align -mno-target-align @gol
897 -mlongcalls -mno-longcalls}
899 @emph{zSeries Options}
900 See S/390 and zSeries Options.
902 @item Code Generation Options
903 @xref{Code Gen Options,,Options for Code Generation Conventions}.
904 @gccoptlist{-fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
905 -ffixed-@var{reg} -fexceptions @gol
906 -fnon-call-exceptions -funwind-tables @gol
907 -fasynchronous-unwind-tables @gol
908 -finhibit-size-directive -finstrument-functions @gol
909 -finstrument-functions-exclude-function-list=@var{sym},@var{sym},@dots{} @gol
910 -finstrument-functions-exclude-file-list=@var{file},@var{file},@dots{} @gol
911 -fno-common -fno-ident @gol
912 -fpcc-struct-return -fpic -fPIC -fpie -fPIE @gol
913 -fno-jump-tables @gol
914 -frecord-gcc-switches @gol
915 -freg-struct-return -fshort-enums @gol
916 -fshort-double -fshort-wchar @gol
917 -fverbose-asm -fpack-struct[=@var{n}] -fstack-check @gol
918 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
919 -fno-stack-limit @gol
920 -fleading-underscore -ftls-model=@var{model} @gol
921 -ftrapv -fwrapv -fbounds-check @gol
926 * Overall Options:: Controlling the kind of output:
927 an executable, object files, assembler files,
928 or preprocessed source.
929 * C Dialect Options:: Controlling the variant of C language compiled.
930 * C++ Dialect Options:: Variations on C++.
931 * Objective-C and Objective-C++ Dialect Options:: Variations on Objective-C
933 * Language Independent Options:: Controlling how diagnostics should be
935 * Warning Options:: How picky should the compiler be?
936 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
937 * Optimize Options:: How much optimization?
938 * Preprocessor Options:: Controlling header files and macro definitions.
939 Also, getting dependency information for Make.
940 * Assembler Options:: Passing options to the assembler.
941 * Link Options:: Specifying libraries and so on.
942 * Directory Options:: Where to find header files and libraries.
943 Where to find the compiler executable files.
944 * Spec Files:: How to pass switches to sub-processes.
945 * Target Options:: Running a cross-compiler, or an old version of GCC.
948 @node Overall Options
949 @section Options Controlling the Kind of Output
951 Compilation can involve up to four stages: preprocessing, compilation
952 proper, assembly and linking, always in that order. GCC is capable of
953 preprocessing and compiling several files either into several
954 assembler input files, or into one assembler input file; then each
955 assembler input file produces an object file, and linking combines all
956 the object files (those newly compiled, and those specified as input)
957 into an executable file.
959 @cindex file name suffix
960 For any given input file, the file name suffix determines what kind of
965 C source code which must be preprocessed.
968 C source code which should not be preprocessed.
971 C++ source code which should not be preprocessed.
974 Objective-C source code. Note that you must link with the @file{libobjc}
975 library to make an Objective-C program work.
978 Objective-C source code which should not be preprocessed.
982 Objective-C++ source code. Note that you must link with the @file{libobjc}
983 library to make an Objective-C++ program work. Note that @samp{.M} refers
984 to a literal capital M@.
987 Objective-C++ source code which should not be preprocessed.
990 C, C++, Objective-C or Objective-C++ header file to be turned into a
995 @itemx @var{file}.cxx
996 @itemx @var{file}.cpp
997 @itemx @var{file}.CPP
998 @itemx @var{file}.c++
1000 C++ source code which must be preprocessed. Note that in @samp{.cxx},
1001 the last two letters must both be literally @samp{x}. Likewise,
1002 @samp{.C} refers to a literal capital C@.
1006 Objective-C++ source code which must be preprocessed.
1008 @item @var{file}.mii
1009 Objective-C++ source code which should not be preprocessed.
1013 @itemx @var{file}.hp
1014 @itemx @var{file}.hxx
1015 @itemx @var{file}.hpp
1016 @itemx @var{file}.HPP
1017 @itemx @var{file}.h++
1018 @itemx @var{file}.tcc
1019 C++ header file to be turned into a precompiled header.
1022 @itemx @var{file}.for
1023 @itemx @var{file}.ftn
1024 Fixed form Fortran source code which should not be preprocessed.
1027 @itemx @var{file}.FOR
1028 @itemx @var{file}.fpp
1029 @itemx @var{file}.FPP
1030 @itemx @var{file}.FTN
1031 Fixed form Fortran source code which must be preprocessed (with the traditional
1034 @item @var{file}.f90
1035 @itemx @var{file}.f95
1036 @itemx @var{file}.f03
1037 @itemx @var{file}.f08
1038 Free form Fortran source code which should not be preprocessed.
1040 @item @var{file}.F90
1041 @itemx @var{file}.F95
1042 @itemx @var{file}.F03
1043 @itemx @var{file}.F08
1044 Free form Fortran source code which must be preprocessed (with the
1045 traditional preprocessor).
1047 @c FIXME: Descriptions of Java file types.
1053 @item @var{file}.ads
1054 Ada source code file which contains a library unit declaration (a
1055 declaration of a package, subprogram, or generic, or a generic
1056 instantiation), or a library unit renaming declaration (a package,
1057 generic, or subprogram renaming declaration). Such files are also
1060 @item @var{file}.adb
1061 Ada source code file containing a library unit body (a subprogram or
1062 package body). Such files are also called @dfn{bodies}.
1064 @c GCC also knows about some suffixes for languages not yet included:
1075 @itemx @var{file}.sx
1076 Assembler code which must be preprocessed.
1079 An object file to be fed straight into linking.
1080 Any file name with no recognized suffix is treated this way.
1084 You can specify the input language explicitly with the @option{-x} option:
1087 @item -x @var{language}
1088 Specify explicitly the @var{language} for the following input files
1089 (rather than letting the compiler choose a default based on the file
1090 name suffix). This option applies to all following input files until
1091 the next @option{-x} option. Possible values for @var{language} are:
1093 c c-header c-cpp-output
1094 c++ c++-header c++-cpp-output
1095 objective-c objective-c-header objective-c-cpp-output
1096 objective-c++ objective-c++-header objective-c++-cpp-output
1097 assembler assembler-with-cpp
1099 f77 f77-cpp-input f95 f95-cpp-input
1104 Turn off any specification of a language, so that subsequent files are
1105 handled according to their file name suffixes (as they are if @option{-x}
1106 has not been used at all).
1108 @item -pass-exit-codes
1109 @opindex pass-exit-codes
1110 Normally the @command{gcc} program will exit with the code of 1 if any
1111 phase of the compiler returns a non-success return code. If you specify
1112 @option{-pass-exit-codes}, the @command{gcc} program will instead return with
1113 numerically highest error produced by any phase that returned an error
1114 indication. The C, C++, and Fortran frontends return 4, if an internal
1115 compiler error is encountered.
1118 If you only want some of the stages of compilation, you can use
1119 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
1120 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
1121 @command{gcc} is to stop. Note that some combinations (for example,
1122 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
1127 Compile or assemble the source files, but do not link. The linking
1128 stage simply is not done. The ultimate output is in the form of an
1129 object file for each source file.
1131 By default, the object file name for a source file is made by replacing
1132 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
1134 Unrecognized input files, not requiring compilation or assembly, are
1139 Stop after the stage of compilation proper; do not assemble. The output
1140 is in the form of an assembler code file for each non-assembler input
1143 By default, the assembler file name for a source file is made by
1144 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
1146 Input files that don't require compilation are ignored.
1150 Stop after the preprocessing stage; do not run the compiler proper. The
1151 output is in the form of preprocessed source code, which is sent to the
1154 Input files which don't require preprocessing are ignored.
1156 @cindex output file option
1159 Place output in file @var{file}. This applies regardless to whatever
1160 sort of output is being produced, whether it be an executable file,
1161 an object file, an assembler file or preprocessed C code.
1163 If @option{-o} is not specified, the default is to put an executable
1164 file in @file{a.out}, the object file for
1165 @file{@var{source}.@var{suffix}} in @file{@var{source}.o}, its
1166 assembler file in @file{@var{source}.s}, a precompiled header file in
1167 @file{@var{source}.@var{suffix}.gch}, and all preprocessed C source on
1172 Print (on standard error output) the commands executed to run the stages
1173 of compilation. Also print the version number of the compiler driver
1174 program and of the preprocessor and the compiler proper.
1178 Like @option{-v} except the commands are not executed and all command
1179 arguments are quoted. This is useful for shell scripts to capture the
1180 driver-generated command lines.
1184 Use pipes rather than temporary files for communication between the
1185 various stages of compilation. This fails to work on some systems where
1186 the assembler is unable to read from a pipe; but the GNU assembler has
1191 If you are compiling multiple source files, this option tells the driver
1192 to pass all the source files to the compiler at once (for those
1193 languages for which the compiler can handle this). This will allow
1194 intermodule analysis (IMA) to be performed by the compiler. Currently the only
1195 language for which this is supported is C@. If you pass source files for
1196 multiple languages to the driver, using this option, the driver will invoke
1197 the compiler(s) that support IMA once each, passing each compiler all the
1198 source files appropriate for it. For those languages that do not support
1199 IMA this option will be ignored, and the compiler will be invoked once for
1200 each source file in that language. If you use this option in conjunction
1201 with @option{-save-temps}, the compiler will generate multiple
1203 (one for each source file), but only one (combined) @file{.o} or
1208 Print (on the standard output) a description of the command line options
1209 understood by @command{gcc}. If the @option{-v} option is also specified
1210 then @option{--help} will also be passed on to the various processes
1211 invoked by @command{gcc}, so that they can display the command line options
1212 they accept. If the @option{-Wextra} option has also been specified
1213 (prior to the @option{--help} option), then command line options which
1214 have no documentation associated with them will also be displayed.
1217 @opindex target-help
1218 Print (on the standard output) a description of target-specific command
1219 line options for each tool. For some targets extra target-specific
1220 information may also be printed.
1222 @item --help=@{@var{class}@r{|[}^@r{]}@var{qualifier}@}@r{[},@dots{}@r{]}
1223 Print (on the standard output) a description of the command line
1224 options understood by the compiler that fit into all specified classes
1225 and qualifiers. These are the supported classes:
1228 @item @samp{optimizers}
1229 This will display all of the optimization options supported by the
1232 @item @samp{warnings}
1233 This will display all of the options controlling warning messages
1234 produced by the compiler.
1237 This will display target-specific options. Unlike the
1238 @option{--target-help} option however, target-specific options of the
1239 linker and assembler will not be displayed. This is because those
1240 tools do not currently support the extended @option{--help=} syntax.
1243 This will display the values recognized by the @option{--param}
1246 @item @var{language}
1247 This will display the options supported for @var{language}, where
1248 @var{language} is the name of one of the languages supported in this
1252 This will display the options that are common to all languages.
1255 These are the supported qualifiers:
1258 @item @samp{undocumented}
1259 Display only those options which are undocumented.
1262 Display options which take an argument that appears after an equal
1263 sign in the same continuous piece of text, such as:
1264 @samp{--help=target}.
1266 @item @samp{separate}
1267 Display options which take an argument that appears as a separate word
1268 following the original option, such as: @samp{-o output-file}.
1271 Thus for example to display all the undocumented target-specific
1272 switches supported by the compiler the following can be used:
1275 --help=target,undocumented
1278 The sense of a qualifier can be inverted by prefixing it with the
1279 @samp{^} character, so for example to display all binary warning
1280 options (i.e., ones that are either on or off and that do not take an
1281 argument), which have a description the following can be used:
1284 --help=warnings,^joined,^undocumented
1287 The argument to @option{--help=} should not consist solely of inverted
1290 Combining several classes is possible, although this usually
1291 restricts the output by so much that there is nothing to display. One
1292 case where it does work however is when one of the classes is
1293 @var{target}. So for example to display all the target-specific
1294 optimization options the following can be used:
1297 --help=target,optimizers
1300 The @option{--help=} option can be repeated on the command line. Each
1301 successive use will display its requested class of options, skipping
1302 those that have already been displayed.
1304 If the @option{-Q} option appears on the command line before the
1305 @option{--help=} option, then the descriptive text displayed by
1306 @option{--help=} is changed. Instead of describing the displayed
1307 options, an indication is given as to whether the option is enabled,
1308 disabled or set to a specific value (assuming that the compiler
1309 knows this at the point where the @option{--help=} option is used).
1311 Here is a truncated example from the ARM port of @command{gcc}:
1314 % gcc -Q -mabi=2 --help=target -c
1315 The following options are target specific:
1317 -mabort-on-noreturn [disabled]
1321 The output is sensitive to the effects of previous command line
1322 options, so for example it is possible to find out which optimizations
1323 are enabled at @option{-O2} by using:
1326 -Q -O2 --help=optimizers
1329 Alternatively you can discover which binary optimizations are enabled
1330 by @option{-O3} by using:
1333 gcc -c -Q -O3 --help=optimizers > /tmp/O3-opts
1334 gcc -c -Q -O2 --help=optimizers > /tmp/O2-opts
1335 diff /tmp/O2-opts /tmp/O3-opts | grep enabled
1338 @item -no-canonical-prefixes
1339 @opindex no-canonical-prefixes
1340 Do not expand any symbolic links, resolve references to @samp{/../}
1341 or @samp{/./}, or make the path absolute when generating a relative
1346 Display the version number and copyrights of the invoked GCC@.
1350 Invoke all subcommands under a wrapper program. It takes a single
1351 comma separated list as an argument, which will be used to invoke
1355 gcc -c t.c -wrapper gdb,--args
1358 This will invoke all subprograms of gcc under "gdb --args",
1359 thus cc1 invocation will be "gdb --args cc1 ...".
1361 @item -fplugin=@var{name}.so
1362 Load the plugin code in file @var{name}.so, assumed to be a
1363 shared object to be dlopen'd by the compiler. The base name of
1364 the shared object file is used to identify the plugin for the
1365 purposes of argument parsing (See
1366 @option{-fplugin-arg-@var{name}-@var{key}=@var{value}} below).
1367 Each plugin should define the callback functions specified in the
1370 @item -fplugin-arg-@var{name}-@var{key}=@var{value}
1371 Define an argument called @var{key} with a value of @var{value}
1372 for the plugin called @var{name}.
1374 @include @value{srcdir}/../libiberty/at-file.texi
1378 @section Compiling C++ Programs
1380 @cindex suffixes for C++ source
1381 @cindex C++ source file suffixes
1382 C++ source files conventionally use one of the suffixes @samp{.C},
1383 @samp{.cc}, @samp{.cpp}, @samp{.CPP}, @samp{.c++}, @samp{.cp}, or
1384 @samp{.cxx}; C++ header files often use @samp{.hh}, @samp{.hpp},
1385 @samp{.H}, or (for shared template code) @samp{.tcc}; and
1386 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
1387 files with these names and compiles them as C++ programs even if you
1388 call the compiler the same way as for compiling C programs (usually
1389 with the name @command{gcc}).
1393 However, the use of @command{gcc} does not add the C++ library.
1394 @command{g++} is a program that calls GCC and treats @samp{.c},
1395 @samp{.h} and @samp{.i} files as C++ source files instead of C source
1396 files unless @option{-x} is used, and automatically specifies linking
1397 against the C++ library. This program is also useful when
1398 precompiling a C header file with a @samp{.h} extension for use in C++
1399 compilations. On many systems, @command{g++} is also installed with
1400 the name @command{c++}.
1402 @cindex invoking @command{g++}
1403 When you compile C++ programs, you may specify many of the same
1404 command-line options that you use for compiling programs in any
1405 language; or command-line options meaningful for C and related
1406 languages; or options that are meaningful only for C++ programs.
1407 @xref{C Dialect Options,,Options Controlling C Dialect}, for
1408 explanations of options for languages related to C@.
1409 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
1410 explanations of options that are meaningful only for C++ programs.
1412 @node C Dialect Options
1413 @section Options Controlling C Dialect
1414 @cindex dialect options
1415 @cindex language dialect options
1416 @cindex options, dialect
1418 The following options control the dialect of C (or languages derived
1419 from C, such as C++, Objective-C and Objective-C++) that the compiler
1423 @cindex ANSI support
1427 In C mode, this is equivalent to @samp{-std=c90}. In C++ mode, it is
1428 equivalent to @samp{-std=c++98}.
1430 This turns off certain features of GCC that are incompatible with ISO
1431 C90 (when compiling C code), or of standard C++ (when compiling C++ code),
1432 such as the @code{asm} and @code{typeof} keywords, and
1433 predefined macros such as @code{unix} and @code{vax} that identify the
1434 type of system you are using. It also enables the undesirable and
1435 rarely used ISO trigraph feature. For the C compiler,
1436 it disables recognition of C++ style @samp{//} comments as well as
1437 the @code{inline} keyword.
1439 The alternate keywords @code{__asm__}, @code{__extension__},
1440 @code{__inline__} and @code{__typeof__} continue to work despite
1441 @option{-ansi}. You would not want to use them in an ISO C program, of
1442 course, but it is useful to put them in header files that might be included
1443 in compilations done with @option{-ansi}. Alternate predefined macros
1444 such as @code{__unix__} and @code{__vax__} are also available, with or
1445 without @option{-ansi}.
1447 The @option{-ansi} option does not cause non-ISO programs to be
1448 rejected gratuitously. For that, @option{-pedantic} is required in
1449 addition to @option{-ansi}. @xref{Warning Options}.
1451 The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi}
1452 option is used. Some header files may notice this macro and refrain
1453 from declaring certain functions or defining certain macros that the
1454 ISO standard doesn't call for; this is to avoid interfering with any
1455 programs that might use these names for other things.
1457 Functions that would normally be built in but do not have semantics
1458 defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in
1459 functions when @option{-ansi} is used. @xref{Other Builtins,,Other
1460 built-in functions provided by GCC}, for details of the functions
1465 Determine the language standard. @xref{Standards,,Language Standards
1466 Supported by GCC}, for details of these standard versions. This option
1467 is currently only supported when compiling C or C++.
1469 The compiler can accept several base standards, such as @samp{c90} or
1470 @samp{c++98}, and GNU dialects of those standards, such as
1471 @samp{gnu90} or @samp{gnu++98}. By specifying a base standard, the
1472 compiler will accept all programs following that standard and those
1473 using GNU extensions that do not contradict it. For example,
1474 @samp{-std=c90} turns off certain features of GCC that are
1475 incompatible with ISO C90, such as the @code{asm} and @code{typeof}
1476 keywords, but not other GNU extensions that do not have a meaning in
1477 ISO C90, such as omitting the middle term of a @code{?:}
1478 expression. On the other hand, by specifying a GNU dialect of a
1479 standard, all features the compiler support are enabled, even when
1480 those features change the meaning of the base standard and some
1481 strict-conforming programs may be rejected. The particular standard
1482 is used by @option{-pedantic} to identify which features are GNU
1483 extensions given that version of the standard. For example
1484 @samp{-std=gnu90 -pedantic} would warn about C++ style @samp{//}
1485 comments, while @samp{-std=gnu99 -pedantic} would not.
1487 A value for this option must be provided; possible values are
1493 Support all ISO C90 programs (certain GNU extensions that conflict
1494 with ISO C90 are disabled). Same as @option{-ansi} for C code.
1496 @item iso9899:199409
1497 ISO C90 as modified in amendment 1.
1503 ISO C99. Note that this standard is not yet fully supported; see
1504 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The
1505 names @samp{c9x} and @samp{iso9899:199x} are deprecated.
1509 GNU dialect of ISO C90 (including some C99 features). This
1510 is the default for C code.
1514 GNU dialect of ISO C99. When ISO C99 is fully implemented in GCC,
1515 this will become the default. The name @samp{gnu9x} is deprecated.
1518 The 1998 ISO C++ standard plus amendments. Same as @option{-ansi} for
1522 GNU dialect of @option{-std=c++98}. This is the default for
1526 The working draft of the upcoming ISO C++0x standard. This option
1527 enables experimental features that are likely to be included in
1528 C++0x. The working draft is constantly changing, and any feature that is
1529 enabled by this flag may be removed from future versions of GCC if it is
1530 not part of the C++0x standard.
1533 GNU dialect of @option{-std=c++0x}. This option enables
1534 experimental features that may be removed in future versions of GCC.
1537 @item -fgnu89-inline
1538 @opindex fgnu89-inline
1539 The option @option{-fgnu89-inline} tells GCC to use the traditional
1540 GNU semantics for @code{inline} functions when in C99 mode.
1541 @xref{Inline,,An Inline Function is As Fast As a Macro}. This option
1542 is accepted and ignored by GCC versions 4.1.3 up to but not including
1543 4.3. In GCC versions 4.3 and later it changes the behavior of GCC in
1544 C99 mode. Using this option is roughly equivalent to adding the
1545 @code{gnu_inline} function attribute to all inline functions
1546 (@pxref{Function Attributes}).
1548 The option @option{-fno-gnu89-inline} explicitly tells GCC to use the
1549 C99 semantics for @code{inline} when in C99 or gnu99 mode (i.e., it
1550 specifies the default behavior). This option was first supported in
1551 GCC 4.3. This option is not supported in @option{-std=c90} or
1552 @option{-std=gnu90} mode.
1554 The preprocessor macros @code{__GNUC_GNU_INLINE__} and
1555 @code{__GNUC_STDC_INLINE__} may be used to check which semantics are
1556 in effect for @code{inline} functions. @xref{Common Predefined
1557 Macros,,,cpp,The C Preprocessor}.
1559 @item -aux-info @var{filename}
1561 Output to the given filename prototyped declarations for all functions
1562 declared and/or defined in a translation unit, including those in header
1563 files. This option is silently ignored in any language other than C@.
1565 Besides declarations, the file indicates, in comments, the origin of
1566 each declaration (source file and line), whether the declaration was
1567 implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or
1568 @samp{O} for old, respectively, in the first character after the line
1569 number and the colon), and whether it came from a declaration or a
1570 definition (@samp{C} or @samp{F}, respectively, in the following
1571 character). In the case of function definitions, a K&R-style list of
1572 arguments followed by their declarations is also provided, inside
1573 comments, after the declaration.
1577 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
1578 keyword, so that code can use these words as identifiers. You can use
1579 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
1580 instead. @option{-ansi} implies @option{-fno-asm}.
1582 In C++, this switch only affects the @code{typeof} keyword, since
1583 @code{asm} and @code{inline} are standard keywords. You may want to
1584 use the @option{-fno-gnu-keywords} flag instead, which has the same
1585 effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this
1586 switch only affects the @code{asm} and @code{typeof} keywords, since
1587 @code{inline} is a standard keyword in ISO C99.
1590 @itemx -fno-builtin-@var{function}
1591 @opindex fno-builtin
1592 @cindex built-in functions
1593 Don't recognize built-in functions that do not begin with
1594 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
1595 functions provided by GCC}, for details of the functions affected,
1596 including those which are not built-in functions when @option{-ansi} or
1597 @option{-std} options for strict ISO C conformance are used because they
1598 do not have an ISO standard meaning.
1600 GCC normally generates special code to handle certain built-in functions
1601 more efficiently; for instance, calls to @code{alloca} may become single
1602 instructions that adjust the stack directly, and calls to @code{memcpy}
1603 may become inline copy loops. The resulting code is often both smaller
1604 and faster, but since the function calls no longer appear as such, you
1605 cannot set a breakpoint on those calls, nor can you change the behavior
1606 of the functions by linking with a different library. In addition,
1607 when a function is recognized as a built-in function, GCC may use
1608 information about that function to warn about problems with calls to
1609 that function, or to generate more efficient code, even if the
1610 resulting code still contains calls to that function. For example,
1611 warnings are given with @option{-Wformat} for bad calls to
1612 @code{printf}, when @code{printf} is built in, and @code{strlen} is
1613 known not to modify global memory.
1615 With the @option{-fno-builtin-@var{function}} option
1616 only the built-in function @var{function} is
1617 disabled. @var{function} must not begin with @samp{__builtin_}. If a
1618 function is named that is not built-in in this version of GCC, this
1619 option is ignored. There is no corresponding
1620 @option{-fbuiltin-@var{function}} option; if you wish to enable
1621 built-in functions selectively when using @option{-fno-builtin} or
1622 @option{-ffreestanding}, you may define macros such as:
1625 #define abs(n) __builtin_abs ((n))
1626 #define strcpy(d, s) __builtin_strcpy ((d), (s))
1631 @cindex hosted environment
1633 Assert that compilation takes place in a hosted environment. This implies
1634 @option{-fbuiltin}. A hosted environment is one in which the
1635 entire standard library is available, and in which @code{main} has a return
1636 type of @code{int}. Examples are nearly everything except a kernel.
1637 This is equivalent to @option{-fno-freestanding}.
1639 @item -ffreestanding
1640 @opindex ffreestanding
1641 @cindex hosted environment
1643 Assert that compilation takes place in a freestanding environment. This
1644 implies @option{-fno-builtin}. A freestanding environment
1645 is one in which the standard library may not exist, and program startup may
1646 not necessarily be at @code{main}. The most obvious example is an OS kernel.
1647 This is equivalent to @option{-fno-hosted}.
1649 @xref{Standards,,Language Standards Supported by GCC}, for details of
1650 freestanding and hosted environments.
1654 @cindex openmp parallel
1655 Enable handling of OpenMP directives @code{#pragma omp} in C/C++ and
1656 @code{!$omp} in Fortran. When @option{-fopenmp} is specified, the
1657 compiler generates parallel code according to the OpenMP Application
1658 Program Interface v3.0 @w{@uref{http://www.openmp.org/}}. This option
1659 implies @option{-pthread}, and thus is only supported on targets that
1660 have support for @option{-pthread}.
1662 @item -fms-extensions
1663 @opindex fms-extensions
1664 Accept some non-standard constructs used in Microsoft header files.
1666 Some cases of unnamed fields in structures and unions are only
1667 accepted with this option. @xref{Unnamed Fields,,Unnamed struct/union
1668 fields within structs/unions}, for details.
1672 Support ISO C trigraphs. The @option{-ansi} option (and @option{-std}
1673 options for strict ISO C conformance) implies @option{-trigraphs}.
1675 @item -no-integrated-cpp
1676 @opindex no-integrated-cpp
1677 Performs a compilation in two passes: preprocessing and compiling. This
1678 option allows a user supplied "cc1", "cc1plus", or "cc1obj" via the
1679 @option{-B} option. The user supplied compilation step can then add in
1680 an additional preprocessing step after normal preprocessing but before
1681 compiling. The default is to use the integrated cpp (internal cpp)
1683 The semantics of this option will change if "cc1", "cc1plus", and
1684 "cc1obj" are merged.
1686 @cindex traditional C language
1687 @cindex C language, traditional
1689 @itemx -traditional-cpp
1690 @opindex traditional-cpp
1691 @opindex traditional
1692 Formerly, these options caused GCC to attempt to emulate a pre-standard
1693 C compiler. They are now only supported with the @option{-E} switch.
1694 The preprocessor continues to support a pre-standard mode. See the GNU
1695 CPP manual for details.
1697 @item -fcond-mismatch
1698 @opindex fcond-mismatch
1699 Allow conditional expressions with mismatched types in the second and
1700 third arguments. The value of such an expression is void. This option
1701 is not supported for C++.
1703 @item -flax-vector-conversions
1704 @opindex flax-vector-conversions
1705 Allow implicit conversions between vectors with differing numbers of
1706 elements and/or incompatible element types. This option should not be
1709 @item -funsigned-char
1710 @opindex funsigned-char
1711 Let the type @code{char} be unsigned, like @code{unsigned char}.
1713 Each kind of machine has a default for what @code{char} should
1714 be. It is either like @code{unsigned char} by default or like
1715 @code{signed char} by default.
1717 Ideally, a portable program should always use @code{signed char} or
1718 @code{unsigned char} when it depends on the signedness of an object.
1719 But many programs have been written to use plain @code{char} and
1720 expect it to be signed, or expect it to be unsigned, depending on the
1721 machines they were written for. This option, and its inverse, let you
1722 make such a program work with the opposite default.
1724 The type @code{char} is always a distinct type from each of
1725 @code{signed char} or @code{unsigned char}, even though its behavior
1726 is always just like one of those two.
1729 @opindex fsigned-char
1730 Let the type @code{char} be signed, like @code{signed char}.
1732 Note that this is equivalent to @option{-fno-unsigned-char}, which is
1733 the negative form of @option{-funsigned-char}. Likewise, the option
1734 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
1736 @item -fsigned-bitfields
1737 @itemx -funsigned-bitfields
1738 @itemx -fno-signed-bitfields
1739 @itemx -fno-unsigned-bitfields
1740 @opindex fsigned-bitfields
1741 @opindex funsigned-bitfields
1742 @opindex fno-signed-bitfields
1743 @opindex fno-unsigned-bitfields
1744 These options control whether a bit-field is signed or unsigned, when the
1745 declaration does not use either @code{signed} or @code{unsigned}. By
1746 default, such a bit-field is signed, because this is consistent: the
1747 basic integer types such as @code{int} are signed types.
1750 @node C++ Dialect Options
1751 @section Options Controlling C++ Dialect
1753 @cindex compiler options, C++
1754 @cindex C++ options, command line
1755 @cindex options, C++
1756 This section describes the command-line options that are only meaningful
1757 for C++ programs; but you can also use most of the GNU compiler options
1758 regardless of what language your program is in. For example, you
1759 might compile a file @code{firstClass.C} like this:
1762 g++ -g -frepo -O -c firstClass.C
1766 In this example, only @option{-frepo} is an option meant
1767 only for C++ programs; you can use the other options with any
1768 language supported by GCC@.
1770 Here is a list of options that are @emph{only} for compiling C++ programs:
1774 @item -fabi-version=@var{n}
1775 @opindex fabi-version
1776 Use version @var{n} of the C++ ABI@. Version 2 is the version of the
1777 C++ ABI that first appeared in G++ 3.4. Version 1 is the version of
1778 the C++ ABI that first appeared in G++ 3.2. Version 0 will always be
1779 the version that conforms most closely to the C++ ABI specification.
1780 Therefore, the ABI obtained using version 0 will change as ABI bugs
1783 The default is version 2.
1785 Version 3 corrects an error in mangling a constant address as a
1788 Version 4 implements a standard mangling for vector types.
1790 See also @option{-Wabi}.
1792 @item -fno-access-control
1793 @opindex fno-access-control
1794 Turn off all access checking. This switch is mainly useful for working
1795 around bugs in the access control code.
1799 Check that the pointer returned by @code{operator new} is non-null
1800 before attempting to modify the storage allocated. This check is
1801 normally unnecessary because the C++ standard specifies that
1802 @code{operator new} will only return @code{0} if it is declared
1803 @samp{throw()}, in which case the compiler will always check the
1804 return value even without this option. In all other cases, when
1805 @code{operator new} has a non-empty exception specification, memory
1806 exhaustion is signalled by throwing @code{std::bad_alloc}. See also
1807 @samp{new (nothrow)}.
1809 @item -fconserve-space
1810 @opindex fconserve-space
1811 Put uninitialized or runtime-initialized global variables into the
1812 common segment, as C does. This saves space in the executable at the
1813 cost of not diagnosing duplicate definitions. If you compile with this
1814 flag and your program mysteriously crashes after @code{main()} has
1815 completed, you may have an object that is being destroyed twice because
1816 two definitions were merged.
1818 This option is no longer useful on most targets, now that support has
1819 been added for putting variables into BSS without making them common.
1821 @item -fno-deduce-init-list
1822 @opindex fno-deduce-init-list
1823 Disable deduction of a template type parameter as
1824 std::initializer_list from a brace-enclosed initializer list, i.e.
1827 template <class T> auto forward(T t) -> decltype (realfn (t))
1834 forward(@{1,2@}); // call forward<std::initializer_list<int>>
1838 This option is present because this deduction is an extension to the
1839 current specification in the C++0x working draft, and there was
1840 some concern about potential overload resolution problems.
1842 @item -ffriend-injection
1843 @opindex ffriend-injection
1844 Inject friend functions into the enclosing namespace, so that they are
1845 visible outside the scope of the class in which they are declared.
1846 Friend functions were documented to work this way in the old Annotated
1847 C++ Reference Manual, and versions of G++ before 4.1 always worked
1848 that way. However, in ISO C++ a friend function which is not declared
1849 in an enclosing scope can only be found using argument dependent
1850 lookup. This option causes friends to be injected as they were in
1853 This option is for compatibility, and may be removed in a future
1856 @item -fno-elide-constructors
1857 @opindex fno-elide-constructors
1858 The C++ standard allows an implementation to omit creating a temporary
1859 which is only used to initialize another object of the same type.
1860 Specifying this option disables that optimization, and forces G++ to
1861 call the copy constructor in all cases.
1863 @item -fno-enforce-eh-specs
1864 @opindex fno-enforce-eh-specs
1865 Don't generate code to check for violation of exception specifications
1866 at runtime. This option violates the C++ standard, but may be useful
1867 for reducing code size in production builds, much like defining
1868 @samp{NDEBUG}. This does not give user code permission to throw
1869 exceptions in violation of the exception specifications; the compiler
1870 will still optimize based on the specifications, so throwing an
1871 unexpected exception will result in undefined behavior.
1874 @itemx -fno-for-scope
1876 @opindex fno-for-scope
1877 If @option{-ffor-scope} is specified, the scope of variables declared in
1878 a @i{for-init-statement} is limited to the @samp{for} loop itself,
1879 as specified by the C++ standard.
1880 If @option{-fno-for-scope} is specified, the scope of variables declared in
1881 a @i{for-init-statement} extends to the end of the enclosing scope,
1882 as was the case in old versions of G++, and other (traditional)
1883 implementations of C++.
1885 The default if neither flag is given to follow the standard,
1886 but to allow and give a warning for old-style code that would
1887 otherwise be invalid, or have different behavior.
1889 @item -fno-gnu-keywords
1890 @opindex fno-gnu-keywords
1891 Do not recognize @code{typeof} as a keyword, so that code can use this
1892 word as an identifier. You can use the keyword @code{__typeof__} instead.
1893 @option{-ansi} implies @option{-fno-gnu-keywords}.
1895 @item -fno-implicit-templates
1896 @opindex fno-implicit-templates
1897 Never emit code for non-inline templates which are instantiated
1898 implicitly (i.e.@: by use); only emit code for explicit instantiations.
1899 @xref{Template Instantiation}, for more information.
1901 @item -fno-implicit-inline-templates
1902 @opindex fno-implicit-inline-templates
1903 Don't emit code for implicit instantiations of inline templates, either.
1904 The default is to handle inlines differently so that compiles with and
1905 without optimization will need the same set of explicit instantiations.
1907 @item -fno-implement-inlines
1908 @opindex fno-implement-inlines
1909 To save space, do not emit out-of-line copies of inline functions
1910 controlled by @samp{#pragma implementation}. This will cause linker
1911 errors if these functions are not inlined everywhere they are called.
1913 @item -fms-extensions
1914 @opindex fms-extensions
1915 Disable pedantic warnings about constructs used in MFC, such as implicit
1916 int and getting a pointer to member function via non-standard syntax.
1918 @item -fno-nonansi-builtins
1919 @opindex fno-nonansi-builtins
1920 Disable built-in declarations of functions that are not mandated by
1921 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
1922 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
1924 @item -fno-operator-names
1925 @opindex fno-operator-names
1926 Do not treat the operator name keywords @code{and}, @code{bitand},
1927 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
1928 synonyms as keywords.
1930 @item -fno-optional-diags
1931 @opindex fno-optional-diags
1932 Disable diagnostics that the standard says a compiler does not need to
1933 issue. Currently, the only such diagnostic issued by G++ is the one for
1934 a name having multiple meanings within a class.
1937 @opindex fpermissive
1938 Downgrade some diagnostics about nonconformant code from errors to
1939 warnings. Thus, using @option{-fpermissive} will allow some
1940 nonconforming code to compile.
1942 @item -fno-pretty-templates
1943 @opindex fno-pretty-templates
1944 When an error message refers to a specialization of a function
1945 template, the compiler will normally print the signature of the
1946 template followed by the template arguments and any typedefs or
1947 typenames in the signature (e.g. @code{void f(T) [with T = int]}
1948 rather than @code{void f(int)}) so that it's clear which template is
1949 involved. When an error message refers to a specialization of a class
1950 template, the compiler will omit any template arguments which match
1951 the default template arguments for that template. If either of these
1952 behaviors make it harder to understand the error message rather than
1953 easier, using @option{-fno-pretty-templates} will disable them.
1957 Enable automatic template instantiation at link time. This option also
1958 implies @option{-fno-implicit-templates}. @xref{Template
1959 Instantiation}, for more information.
1963 Disable generation of information about every class with virtual
1964 functions for use by the C++ runtime type identification features
1965 (@samp{dynamic_cast} and @samp{typeid}). If you don't use those parts
1966 of the language, you can save some space by using this flag. Note that
1967 exception handling uses the same information, but it will generate it as
1968 needed. The @samp{dynamic_cast} operator can still be used for casts that
1969 do not require runtime type information, i.e.@: casts to @code{void *} or to
1970 unambiguous base classes.
1974 Emit statistics about front-end processing at the end of the compilation.
1975 This information is generally only useful to the G++ development team.
1977 @item -ftemplate-depth=@var{n}
1978 @opindex ftemplate-depth
1979 Set the maximum instantiation depth for template classes to @var{n}.
1980 A limit on the template instantiation depth is needed to detect
1981 endless recursions during template class instantiation. ANSI/ISO C++
1982 conforming programs must not rely on a maximum depth greater than 17
1983 (changed to 1024 in C++0x).
1985 @item -fno-threadsafe-statics
1986 @opindex fno-threadsafe-statics
1987 Do not emit the extra code to use the routines specified in the C++
1988 ABI for thread-safe initialization of local statics. You can use this
1989 option to reduce code size slightly in code that doesn't need to be
1992 @item -fuse-cxa-atexit
1993 @opindex fuse-cxa-atexit
1994 Register destructors for objects with static storage duration with the
1995 @code{__cxa_atexit} function rather than the @code{atexit} function.
1996 This option is required for fully standards-compliant handling of static
1997 destructors, but will only work if your C library supports
1998 @code{__cxa_atexit}.
2000 @item -fno-use-cxa-get-exception-ptr
2001 @opindex fno-use-cxa-get-exception-ptr
2002 Don't use the @code{__cxa_get_exception_ptr} runtime routine. This
2003 will cause @code{std::uncaught_exception} to be incorrect, but is necessary
2004 if the runtime routine is not available.
2006 @item -fvisibility-inlines-hidden
2007 @opindex fvisibility-inlines-hidden
2008 This switch declares that the user does not attempt to compare
2009 pointers to inline methods where the addresses of the two functions
2010 were taken in different shared objects.
2012 The effect of this is that GCC may, effectively, mark inline methods with
2013 @code{__attribute__ ((visibility ("hidden")))} so that they do not
2014 appear in the export table of a DSO and do not require a PLT indirection
2015 when used within the DSO@. Enabling this option can have a dramatic effect
2016 on load and link times of a DSO as it massively reduces the size of the
2017 dynamic export table when the library makes heavy use of templates.
2019 The behavior of this switch is not quite the same as marking the
2020 methods as hidden directly, because it does not affect static variables
2021 local to the function or cause the compiler to deduce that
2022 the function is defined in only one shared object.
2024 You may mark a method as having a visibility explicitly to negate the
2025 effect of the switch for that method. For example, if you do want to
2026 compare pointers to a particular inline method, you might mark it as
2027 having default visibility. Marking the enclosing class with explicit
2028 visibility will have no effect.
2030 Explicitly instantiated inline methods are unaffected by this option
2031 as their linkage might otherwise cross a shared library boundary.
2032 @xref{Template Instantiation}.
2034 @item -fvisibility-ms-compat
2035 @opindex fvisibility-ms-compat
2036 This flag attempts to use visibility settings to make GCC's C++
2037 linkage model compatible with that of Microsoft Visual Studio.
2039 The flag makes these changes to GCC's linkage model:
2043 It sets the default visibility to @code{hidden}, like
2044 @option{-fvisibility=hidden}.
2047 Types, but not their members, are not hidden by default.
2050 The One Definition Rule is relaxed for types without explicit
2051 visibility specifications which are defined in more than one different
2052 shared object: those declarations are permitted if they would have
2053 been permitted when this option was not used.
2056 In new code it is better to use @option{-fvisibility=hidden} and
2057 export those classes which are intended to be externally visible.
2058 Unfortunately it is possible for code to rely, perhaps accidentally,
2059 on the Visual Studio behavior.
2061 Among the consequences of these changes are that static data members
2062 of the same type with the same name but defined in different shared
2063 objects will be different, so changing one will not change the other;
2064 and that pointers to function members defined in different shared
2065 objects may not compare equal. When this flag is given, it is a
2066 violation of the ODR to define types with the same name differently.
2070 Do not use weak symbol support, even if it is provided by the linker.
2071 By default, G++ will use weak symbols if they are available. This
2072 option exists only for testing, and should not be used by end-users;
2073 it will result in inferior code and has no benefits. This option may
2074 be removed in a future release of G++.
2078 Do not search for header files in the standard directories specific to
2079 C++, but do still search the other standard directories. (This option
2080 is used when building the C++ library.)
2083 In addition, these optimization, warning, and code generation options
2084 have meanings only for C++ programs:
2087 @item -fno-default-inline
2088 @opindex fno-default-inline
2089 Do not assume @samp{inline} for functions defined inside a class scope.
2090 @xref{Optimize Options,,Options That Control Optimization}. Note that these
2091 functions will have linkage like inline functions; they just won't be
2094 @item -Wabi @r{(C, Objective-C, C++ and Objective-C++ only)}
2097 Warn when G++ generates code that is probably not compatible with the
2098 vendor-neutral C++ ABI@. Although an effort has been made to warn about
2099 all such cases, there are probably some cases that are not warned about,
2100 even though G++ is generating incompatible code. There may also be
2101 cases where warnings are emitted even though the code that is generated
2104 You should rewrite your code to avoid these warnings if you are
2105 concerned about the fact that code generated by G++ may not be binary
2106 compatible with code generated by other compilers.
2108 The known incompatibilities in @option{-fabi-version=2} (the default) include:
2113 A template with a non-type template parameter of reference type is
2114 mangled incorrectly:
2117 template <int &> struct S @{@};
2121 This is fixed in @option{-fabi-version=3}.
2124 SIMD vector types declared using @code{__attribute ((vector_size))} are
2125 mangled in a non-standard way that does not allow for overloading of
2126 functions taking vectors of different sizes.
2128 The mangling is changed in @option{-fabi-version=4}.
2131 The known incompatibilities in @option{-fabi-version=1} include:
2136 Incorrect handling of tail-padding for bit-fields. G++ may attempt to
2137 pack data into the same byte as a base class. For example:
2140 struct A @{ virtual void f(); int f1 : 1; @};
2141 struct B : public A @{ int f2 : 1; @};
2145 In this case, G++ will place @code{B::f2} into the same byte
2146 as@code{A::f1}; other compilers will not. You can avoid this problem
2147 by explicitly padding @code{A} so that its size is a multiple of the
2148 byte size on your platform; that will cause G++ and other compilers to
2149 layout @code{B} identically.
2152 Incorrect handling of tail-padding for virtual bases. G++ does not use
2153 tail padding when laying out virtual bases. For example:
2156 struct A @{ virtual void f(); char c1; @};
2157 struct B @{ B(); char c2; @};
2158 struct C : public A, public virtual B @{@};
2162 In this case, G++ will not place @code{B} into the tail-padding for
2163 @code{A}; other compilers will. You can avoid this problem by
2164 explicitly padding @code{A} so that its size is a multiple of its
2165 alignment (ignoring virtual base classes); that will cause G++ and other
2166 compilers to layout @code{C} identically.
2169 Incorrect handling of bit-fields with declared widths greater than that
2170 of their underlying types, when the bit-fields appear in a union. For
2174 union U @{ int i : 4096; @};
2178 Assuming that an @code{int} does not have 4096 bits, G++ will make the
2179 union too small by the number of bits in an @code{int}.
2182 Empty classes can be placed at incorrect offsets. For example:
2192 struct C : public B, public A @{@};
2196 G++ will place the @code{A} base class of @code{C} at a nonzero offset;
2197 it should be placed at offset zero. G++ mistakenly believes that the
2198 @code{A} data member of @code{B} is already at offset zero.
2201 Names of template functions whose types involve @code{typename} or
2202 template template parameters can be mangled incorrectly.
2205 template <typename Q>
2206 void f(typename Q::X) @{@}
2208 template <template <typename> class Q>
2209 void f(typename Q<int>::X) @{@}
2213 Instantiations of these templates may be mangled incorrectly.
2217 It also warns psABI related changes. The known psABI changes at this
2223 For SYSV/x86-64, when passing union with long double, it is changed to
2224 pass in memory as specified in psABI. For example:
2234 @code{union U} will always be passed in memory.
2238 @item -Wctor-dtor-privacy @r{(C++ and Objective-C++ only)}
2239 @opindex Wctor-dtor-privacy
2240 @opindex Wno-ctor-dtor-privacy
2241 Warn when a class seems unusable because all the constructors or
2242 destructors in that class are private, and it has neither friends nor
2243 public static member functions.
2245 @item -Wnon-virtual-dtor @r{(C++ and Objective-C++ only)}
2246 @opindex Wnon-virtual-dtor
2247 @opindex Wno-non-virtual-dtor
2248 Warn when a class has virtual functions and accessible non-virtual
2249 destructor, in which case it would be possible but unsafe to delete
2250 an instance of a derived class through a pointer to the base class.
2251 This warning is also enabled if -Weffc++ is specified.
2253 @item -Wreorder @r{(C++ and Objective-C++ only)}
2255 @opindex Wno-reorder
2256 @cindex reordering, warning
2257 @cindex warning for reordering of member initializers
2258 Warn when the order of member initializers given in the code does not
2259 match the order in which they must be executed. For instance:
2265 A(): j (0), i (1) @{ @}
2269 The compiler will rearrange the member initializers for @samp{i}
2270 and @samp{j} to match the declaration order of the members, emitting
2271 a warning to that effect. This warning is enabled by @option{-Wall}.
2274 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
2277 @item -Weffc++ @r{(C++ and Objective-C++ only)}
2280 Warn about violations of the following style guidelines from Scott Meyers'
2281 @cite{Effective C++} book:
2285 Item 11: Define a copy constructor and an assignment operator for classes
2286 with dynamically allocated memory.
2289 Item 12: Prefer initialization to assignment in constructors.
2292 Item 14: Make destructors virtual in base classes.
2295 Item 15: Have @code{operator=} return a reference to @code{*this}.
2298 Item 23: Don't try to return a reference when you must return an object.
2302 Also warn about violations of the following style guidelines from
2303 Scott Meyers' @cite{More Effective C++} book:
2307 Item 6: Distinguish between prefix and postfix forms of increment and
2308 decrement operators.
2311 Item 7: Never overload @code{&&}, @code{||}, or @code{,}.
2315 When selecting this option, be aware that the standard library
2316 headers do not obey all of these guidelines; use @samp{grep -v}
2317 to filter out those warnings.
2319 @item -Wstrict-null-sentinel @r{(C++ and Objective-C++ only)}
2320 @opindex Wstrict-null-sentinel
2321 @opindex Wno-strict-null-sentinel
2322 Warn also about the use of an uncasted @code{NULL} as sentinel. When
2323 compiling only with GCC this is a valid sentinel, as @code{NULL} is defined
2324 to @code{__null}. Although it is a null pointer constant not a null pointer,
2325 it is guaranteed to be of the same size as a pointer. But this use is
2326 not portable across different compilers.
2328 @item -Wno-non-template-friend @r{(C++ and Objective-C++ only)}
2329 @opindex Wno-non-template-friend
2330 @opindex Wnon-template-friend
2331 Disable warnings when non-templatized friend functions are declared
2332 within a template. Since the advent of explicit template specification
2333 support in G++, if the name of the friend is an unqualified-id (i.e.,
2334 @samp{friend foo(int)}), the C++ language specification demands that the
2335 friend declare or define an ordinary, nontemplate function. (Section
2336 14.5.3). Before G++ implemented explicit specification, unqualified-ids
2337 could be interpreted as a particular specialization of a templatized
2338 function. Because this non-conforming behavior is no longer the default
2339 behavior for G++, @option{-Wnon-template-friend} allows the compiler to
2340 check existing code for potential trouble spots and is on by default.
2341 This new compiler behavior can be turned off with
2342 @option{-Wno-non-template-friend} which keeps the conformant compiler code
2343 but disables the helpful warning.
2345 @item -Wold-style-cast @r{(C++ and Objective-C++ only)}
2346 @opindex Wold-style-cast
2347 @opindex Wno-old-style-cast
2348 Warn if an old-style (C-style) cast to a non-void type is used within
2349 a C++ program. The new-style casts (@samp{dynamic_cast},
2350 @samp{static_cast}, @samp{reinterpret_cast}, and @samp{const_cast}) are
2351 less vulnerable to unintended effects and much easier to search for.
2353 @item -Woverloaded-virtual @r{(C++ and Objective-C++ only)}
2354 @opindex Woverloaded-virtual
2355 @opindex Wno-overloaded-virtual
2356 @cindex overloaded virtual fn, warning
2357 @cindex warning for overloaded virtual fn
2358 Warn when a function declaration hides virtual functions from a
2359 base class. For example, in:
2366 struct B: public A @{
2371 the @code{A} class version of @code{f} is hidden in @code{B}, and code
2379 will fail to compile.
2381 @item -Wno-pmf-conversions @r{(C++ and Objective-C++ only)}
2382 @opindex Wno-pmf-conversions
2383 @opindex Wpmf-conversions
2384 Disable the diagnostic for converting a bound pointer to member function
2387 @item -Wsign-promo @r{(C++ and Objective-C++ only)}
2388 @opindex Wsign-promo
2389 @opindex Wno-sign-promo
2390 Warn when overload resolution chooses a promotion from unsigned or
2391 enumerated type to a signed type, over a conversion to an unsigned type of
2392 the same size. Previous versions of G++ would try to preserve
2393 unsignedness, but the standard mandates the current behavior.
2398 A& operator = (int);
2408 In this example, G++ will synthesize a default @samp{A& operator =
2409 (const A&);}, while cfront will use the user-defined @samp{operator =}.
2412 @node Objective-C and Objective-C++ Dialect Options
2413 @section Options Controlling Objective-C and Objective-C++ Dialects
2415 @cindex compiler options, Objective-C and Objective-C++
2416 @cindex Objective-C and Objective-C++ options, command line
2417 @cindex options, Objective-C and Objective-C++
2418 (NOTE: This manual does not describe the Objective-C and Objective-C++
2419 languages themselves. See @xref{Standards,,Language Standards
2420 Supported by GCC}, for references.)
2422 This section describes the command-line options that are only meaningful
2423 for Objective-C and Objective-C++ programs, but you can also use most of
2424 the language-independent GNU compiler options.
2425 For example, you might compile a file @code{some_class.m} like this:
2428 gcc -g -fgnu-runtime -O -c some_class.m
2432 In this example, @option{-fgnu-runtime} is an option meant only for
2433 Objective-C and Objective-C++ programs; you can use the other options with
2434 any language supported by GCC@.
2436 Note that since Objective-C is an extension of the C language, Objective-C
2437 compilations may also use options specific to the C front-end (e.g.,
2438 @option{-Wtraditional}). Similarly, Objective-C++ compilations may use
2439 C++-specific options (e.g., @option{-Wabi}).
2441 Here is a list of options that are @emph{only} for compiling Objective-C
2442 and Objective-C++ programs:
2445 @item -fconstant-string-class=@var{class-name}
2446 @opindex fconstant-string-class
2447 Use @var{class-name} as the name of the class to instantiate for each
2448 literal string specified with the syntax @code{@@"@dots{}"}. The default
2449 class name is @code{NXConstantString} if the GNU runtime is being used, and
2450 @code{NSConstantString} if the NeXT runtime is being used (see below). The
2451 @option{-fconstant-cfstrings} option, if also present, will override the
2452 @option{-fconstant-string-class} setting and cause @code{@@"@dots{}"} literals
2453 to be laid out as constant CoreFoundation strings.
2456 @opindex fgnu-runtime
2457 Generate object code compatible with the standard GNU Objective-C
2458 runtime. This is the default for most types of systems.
2460 @item -fnext-runtime
2461 @opindex fnext-runtime
2462 Generate output compatible with the NeXT runtime. This is the default
2463 for NeXT-based systems, including Darwin and Mac OS X@. The macro
2464 @code{__NEXT_RUNTIME__} is predefined if (and only if) this option is
2467 @item -fno-nil-receivers
2468 @opindex fno-nil-receivers
2469 Assume that all Objective-C message dispatches (e.g.,
2470 @code{[receiver message:arg]}) in this translation unit ensure that the receiver
2471 is not @code{nil}. This allows for more efficient entry points in the runtime
2472 to be used. Currently, this option is only available in conjunction with
2473 the NeXT runtime on Mac OS X 10.3 and later.
2475 @item -fobjc-call-cxx-cdtors
2476 @opindex fobjc-call-cxx-cdtors
2477 For each Objective-C class, check if any of its instance variables is a
2478 C++ object with a non-trivial default constructor. If so, synthesize a
2479 special @code{- (id) .cxx_construct} instance method that will run
2480 non-trivial default constructors on any such instance variables, in order,
2481 and then return @code{self}. Similarly, check if any instance variable
2482 is a C++ object with a non-trivial destructor, and if so, synthesize a
2483 special @code{- (void) .cxx_destruct} method that will run
2484 all such default destructors, in reverse order.
2486 The @code{- (id) .cxx_construct} and/or @code{- (void) .cxx_destruct} methods
2487 thusly generated will only operate on instance variables declared in the
2488 current Objective-C class, and not those inherited from superclasses. It
2489 is the responsibility of the Objective-C runtime to invoke all such methods
2490 in an object's inheritance hierarchy. The @code{- (id) .cxx_construct} methods
2491 will be invoked by the runtime immediately after a new object
2492 instance is allocated; the @code{- (void) .cxx_destruct} methods will
2493 be invoked immediately before the runtime deallocates an object instance.
2495 As of this writing, only the NeXT runtime on Mac OS X 10.4 and later has
2496 support for invoking the @code{- (id) .cxx_construct} and
2497 @code{- (void) .cxx_destruct} methods.
2499 @item -fobjc-direct-dispatch
2500 @opindex fobjc-direct-dispatch
2501 Allow fast jumps to the message dispatcher. On Darwin this is
2502 accomplished via the comm page.
2504 @item -fobjc-exceptions
2505 @opindex fobjc-exceptions
2506 Enable syntactic support for structured exception handling in Objective-C,
2507 similar to what is offered by C++ and Java. This option is
2508 unavailable in conjunction with the NeXT runtime on Mac OS X 10.2 and
2517 @@catch (AnObjCClass *exc) @{
2524 @@catch (AnotherClass *exc) @{
2527 @@catch (id allOthers) @{
2537 The @code{@@throw} statement may appear anywhere in an Objective-C or
2538 Objective-C++ program; when used inside of a @code{@@catch} block, the
2539 @code{@@throw} may appear without an argument (as shown above), in which case
2540 the object caught by the @code{@@catch} will be rethrown.
2542 Note that only (pointers to) Objective-C objects may be thrown and
2543 caught using this scheme. When an object is thrown, it will be caught
2544 by the nearest @code{@@catch} clause capable of handling objects of that type,
2545 analogously to how @code{catch} blocks work in C++ and Java. A
2546 @code{@@catch(id @dots{})} clause (as shown above) may also be provided to catch
2547 any and all Objective-C exceptions not caught by previous @code{@@catch}
2550 The @code{@@finally} clause, if present, will be executed upon exit from the
2551 immediately preceding @code{@@try @dots{} @@catch} section. This will happen
2552 regardless of whether any exceptions are thrown, caught or rethrown
2553 inside the @code{@@try @dots{} @@catch} section, analogously to the behavior
2554 of the @code{finally} clause in Java.
2556 There are several caveats to using the new exception mechanism:
2560 Although currently designed to be binary compatible with @code{NS_HANDLER}-style
2561 idioms provided by the @code{NSException} class, the new
2562 exceptions can only be used on Mac OS X 10.3 (Panther) and later
2563 systems, due to additional functionality needed in the (NeXT) Objective-C
2567 As mentioned above, the new exceptions do not support handling
2568 types other than Objective-C objects. Furthermore, when used from
2569 Objective-C++, the Objective-C exception model does not interoperate with C++
2570 exceptions at this time. This means you cannot @code{@@throw} an exception
2571 from Objective-C and @code{catch} it in C++, or vice versa
2572 (i.e., @code{throw @dots{} @@catch}).
2575 The @option{-fobjc-exceptions} switch also enables the use of synchronization
2576 blocks for thread-safe execution:
2579 @@synchronized (ObjCClass *guard) @{
2584 Upon entering the @code{@@synchronized} block, a thread of execution shall
2585 first check whether a lock has been placed on the corresponding @code{guard}
2586 object by another thread. If it has, the current thread shall wait until
2587 the other thread relinquishes its lock. Once @code{guard} becomes available,
2588 the current thread will place its own lock on it, execute the code contained in
2589 the @code{@@synchronized} block, and finally relinquish the lock (thereby
2590 making @code{guard} available to other threads).
2592 Unlike Java, Objective-C does not allow for entire methods to be marked
2593 @code{@@synchronized}. Note that throwing exceptions out of
2594 @code{@@synchronized} blocks is allowed, and will cause the guarding object
2595 to be unlocked properly.
2599 Enable garbage collection (GC) in Objective-C and Objective-C++ programs.
2601 @item -freplace-objc-classes
2602 @opindex freplace-objc-classes
2603 Emit a special marker instructing @command{ld(1)} not to statically link in
2604 the resulting object file, and allow @command{dyld(1)} to load it in at
2605 run time instead. This is used in conjunction with the Fix-and-Continue
2606 debugging mode, where the object file in question may be recompiled and
2607 dynamically reloaded in the course of program execution, without the need
2608 to restart the program itself. Currently, Fix-and-Continue functionality
2609 is only available in conjunction with the NeXT runtime on Mac OS X 10.3
2614 When compiling for the NeXT runtime, the compiler ordinarily replaces calls
2615 to @code{objc_getClass("@dots{}")} (when the name of the class is known at
2616 compile time) with static class references that get initialized at load time,
2617 which improves run-time performance. Specifying the @option{-fzero-link} flag
2618 suppresses this behavior and causes calls to @code{objc_getClass("@dots{}")}
2619 to be retained. This is useful in Zero-Link debugging mode, since it allows
2620 for individual class implementations to be modified during program execution.
2624 Dump interface declarations for all classes seen in the source file to a
2625 file named @file{@var{sourcename}.decl}.
2627 @item -Wassign-intercept @r{(Objective-C and Objective-C++ only)}
2628 @opindex Wassign-intercept
2629 @opindex Wno-assign-intercept
2630 Warn whenever an Objective-C assignment is being intercepted by the
2633 @item -Wno-protocol @r{(Objective-C and Objective-C++ only)}
2634 @opindex Wno-protocol
2636 If a class is declared to implement a protocol, a warning is issued for
2637 every method in the protocol that is not implemented by the class. The
2638 default behavior is to issue a warning for every method not explicitly
2639 implemented in the class, even if a method implementation is inherited
2640 from the superclass. If you use the @option{-Wno-protocol} option, then
2641 methods inherited from the superclass are considered to be implemented,
2642 and no warning is issued for them.
2644 @item -Wselector @r{(Objective-C and Objective-C++ only)}
2646 @opindex Wno-selector
2647 Warn if multiple methods of different types for the same selector are
2648 found during compilation. The check is performed on the list of methods
2649 in the final stage of compilation. Additionally, a check is performed
2650 for each selector appearing in a @code{@@selector(@dots{})}
2651 expression, and a corresponding method for that selector has been found
2652 during compilation. Because these checks scan the method table only at
2653 the end of compilation, these warnings are not produced if the final
2654 stage of compilation is not reached, for example because an error is
2655 found during compilation, or because the @option{-fsyntax-only} option is
2658 @item -Wstrict-selector-match @r{(Objective-C and Objective-C++ only)}
2659 @opindex Wstrict-selector-match
2660 @opindex Wno-strict-selector-match
2661 Warn if multiple methods with differing argument and/or return types are
2662 found for a given selector when attempting to send a message using this
2663 selector to a receiver of type @code{id} or @code{Class}. When this flag
2664 is off (which is the default behavior), the compiler will omit such warnings
2665 if any differences found are confined to types which share the same size
2668 @item -Wundeclared-selector @r{(Objective-C and Objective-C++ only)}
2669 @opindex Wundeclared-selector
2670 @opindex Wno-undeclared-selector
2671 Warn if a @code{@@selector(@dots{})} expression referring to an
2672 undeclared selector is found. A selector is considered undeclared if no
2673 method with that name has been declared before the
2674 @code{@@selector(@dots{})} expression, either explicitly in an
2675 @code{@@interface} or @code{@@protocol} declaration, or implicitly in
2676 an @code{@@implementation} section. This option always performs its
2677 checks as soon as a @code{@@selector(@dots{})} expression is found,
2678 while @option{-Wselector} only performs its checks in the final stage of
2679 compilation. This also enforces the coding style convention
2680 that methods and selectors must be declared before being used.
2682 @item -print-objc-runtime-info
2683 @opindex print-objc-runtime-info
2684 Generate C header describing the largest structure that is passed by
2689 @node Language Independent Options
2690 @section Options to Control Diagnostic Messages Formatting
2691 @cindex options to control diagnostics formatting
2692 @cindex diagnostic messages
2693 @cindex message formatting
2695 Traditionally, diagnostic messages have been formatted irrespective of
2696 the output device's aspect (e.g.@: its width, @dots{}). The options described
2697 below can be used to control the diagnostic messages formatting
2698 algorithm, e.g.@: how many characters per line, how often source location
2699 information should be reported. Right now, only the C++ front end can
2700 honor these options. However it is expected, in the near future, that
2701 the remaining front ends would be able to digest them correctly.
2704 @item -fmessage-length=@var{n}
2705 @opindex fmessage-length
2706 Try to format error messages so that they fit on lines of about @var{n}
2707 characters. The default is 72 characters for @command{g++} and 0 for the rest of
2708 the front ends supported by GCC@. If @var{n} is zero, then no
2709 line-wrapping will be done; each error message will appear on a single
2712 @opindex fdiagnostics-show-location
2713 @item -fdiagnostics-show-location=once
2714 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
2715 reporter to emit @emph{once} source location information; that is, in
2716 case the message is too long to fit on a single physical line and has to
2717 be wrapped, the source location won't be emitted (as prefix) again,
2718 over and over, in subsequent continuation lines. This is the default
2721 @item -fdiagnostics-show-location=every-line
2722 Only meaningful in line-wrapping mode. Instructs the diagnostic
2723 messages reporter to emit the same source location information (as
2724 prefix) for physical lines that result from the process of breaking
2725 a message which is too long to fit on a single line.
2727 @item -fdiagnostics-show-option
2728 @opindex fdiagnostics-show-option
2729 This option instructs the diagnostic machinery to add text to each
2730 diagnostic emitted, which indicates which command line option directly
2731 controls that diagnostic, when such an option is known to the
2732 diagnostic machinery.
2734 @item -Wcoverage-mismatch
2735 @opindex Wcoverage-mismatch
2736 Warn if feedback profiles do not match when using the
2737 @option{-fprofile-use} option.
2738 If a source file was changed between @option{-fprofile-gen} and
2739 @option{-fprofile-use}, the files with the profile feedback can fail
2740 to match the source file and GCC can not use the profile feedback
2741 information. By default, GCC emits an error message in this case.
2742 The option @option{-Wcoverage-mismatch} emits a warning instead of an
2743 error. GCC does not use appropriate feedback profiles, so using this
2744 option can result in poorly optimized code. This option is useful
2745 only in the case of very minor changes such as bug fixes to an
2750 @node Warning Options
2751 @section Options to Request or Suppress Warnings
2752 @cindex options to control warnings
2753 @cindex warning messages
2754 @cindex messages, warning
2755 @cindex suppressing warnings
2757 Warnings are diagnostic messages that report constructions which
2758 are not inherently erroneous but which are risky or suggest there
2759 may have been an error.
2761 The following language-independent options do not enable specific
2762 warnings but control the kinds of diagnostics produced by GCC.
2765 @cindex syntax checking
2767 @opindex fsyntax-only
2768 Check the code for syntax errors, but don't do anything beyond that.
2772 Inhibit all warning messages.
2777 Make all warnings into errors.
2782 Make the specified warning into an error. The specifier for a warning
2783 is appended, for example @option{-Werror=switch} turns the warnings
2784 controlled by @option{-Wswitch} into errors. This switch takes a
2785 negative form, to be used to negate @option{-Werror} for specific
2786 warnings, for example @option{-Wno-error=switch} makes
2787 @option{-Wswitch} warnings not be errors, even when @option{-Werror}
2788 is in effect. You can use the @option{-fdiagnostics-show-option}
2789 option to have each controllable warning amended with the option which
2790 controls it, to determine what to use with this option.
2792 Note that specifying @option{-Werror=}@var{foo} automatically implies
2793 @option{-W}@var{foo}. However, @option{-Wno-error=}@var{foo} does not
2796 @item -Wfatal-errors
2797 @opindex Wfatal-errors
2798 @opindex Wno-fatal-errors
2799 This option causes the compiler to abort compilation on the first error
2800 occurred rather than trying to keep going and printing further error
2805 You can request many specific warnings with options beginning
2806 @samp{-W}, for example @option{-Wimplicit} to request warnings on
2807 implicit declarations. Each of these specific warning options also
2808 has a negative form beginning @samp{-Wno-} to turn off warnings; for
2809 example, @option{-Wno-implicit}. This manual lists only one of the
2810 two forms, whichever is not the default. For further,
2811 language-specific options also refer to @ref{C++ Dialect Options} and
2812 @ref{Objective-C and Objective-C++ Dialect Options}.
2817 Issue all the warnings demanded by strict ISO C and ISO C++;
2818 reject all programs that use forbidden extensions, and some other
2819 programs that do not follow ISO C and ISO C++. For ISO C, follows the
2820 version of the ISO C standard specified by any @option{-std} option used.
2822 Valid ISO C and ISO C++ programs should compile properly with or without
2823 this option (though a rare few will require @option{-ansi} or a
2824 @option{-std} option specifying the required version of ISO C)@. However,
2825 without this option, certain GNU extensions and traditional C and C++
2826 features are supported as well. With this option, they are rejected.
2828 @option{-pedantic} does not cause warning messages for use of the
2829 alternate keywords whose names begin and end with @samp{__}. Pedantic
2830 warnings are also disabled in the expression that follows
2831 @code{__extension__}. However, only system header files should use
2832 these escape routes; application programs should avoid them.
2833 @xref{Alternate Keywords}.
2835 Some users try to use @option{-pedantic} to check programs for strict ISO
2836 C conformance. They soon find that it does not do quite what they want:
2837 it finds some non-ISO practices, but not all---only those for which
2838 ISO C @emph{requires} a diagnostic, and some others for which
2839 diagnostics have been added.
2841 A feature to report any failure to conform to ISO C might be useful in
2842 some instances, but would require considerable additional work and would
2843 be quite different from @option{-pedantic}. We don't have plans to
2844 support such a feature in the near future.
2846 Where the standard specified with @option{-std} represents a GNU
2847 extended dialect of C, such as @samp{gnu90} or @samp{gnu99}, there is a
2848 corresponding @dfn{base standard}, the version of ISO C on which the GNU
2849 extended dialect is based. Warnings from @option{-pedantic} are given
2850 where they are required by the base standard. (It would not make sense
2851 for such warnings to be given only for features not in the specified GNU
2852 C dialect, since by definition the GNU dialects of C include all
2853 features the compiler supports with the given option, and there would be
2854 nothing to warn about.)
2856 @item -pedantic-errors
2857 @opindex pedantic-errors
2858 Like @option{-pedantic}, except that errors are produced rather than
2864 This enables all the warnings about constructions that some users
2865 consider questionable, and that are easy to avoid (or modify to
2866 prevent the warning), even in conjunction with macros. This also
2867 enables some language-specific warnings described in @ref{C++ Dialect
2868 Options} and @ref{Objective-C and Objective-C++ Dialect Options}.
2870 @option{-Wall} turns on the following warning flags:
2872 @gccoptlist{-Waddress @gol
2873 -Warray-bounds @r{(only with} @option{-O2}@r{)} @gol
2875 -Wchar-subscripts @gol
2876 -Wenum-compare @r{(in C/Objc; this is on by default in C++)} @gol
2878 -Wimplicit-function-declaration @gol
2881 -Wmain @r{(only for C/ObjC and unless} @option{-ffreestanding}@r{)} @gol
2882 -Wmissing-braces @gol
2888 -Wsequence-point @gol
2889 -Wsign-compare @r{(only in C++)} @gol
2890 -Wstrict-aliasing @gol
2891 -Wstrict-overflow=1 @gol
2894 -Wuninitialized @gol
2895 -Wunknown-pragmas @gol
2896 -Wunused-function @gol
2899 -Wunused-variable @gol
2900 -Wvolatile-register-var @gol
2903 Note that some warning flags are not implied by @option{-Wall}. Some of
2904 them warn about constructions that users generally do not consider
2905 questionable, but which occasionally you might wish to check for;
2906 others warn about constructions that are necessary or hard to avoid in
2907 some cases, and there is no simple way to modify the code to suppress
2908 the warning. Some of them are enabled by @option{-Wextra} but many of
2909 them must be enabled individually.
2915 This enables some extra warning flags that are not enabled by
2916 @option{-Wall}. (This option used to be called @option{-W}. The older
2917 name is still supported, but the newer name is more descriptive.)
2919 @gccoptlist{-Wclobbered @gol
2921 -Wignored-qualifiers @gol
2922 -Wmissing-field-initializers @gol
2923 -Wmissing-parameter-type @r{(C only)} @gol
2924 -Wold-style-declaration @r{(C only)} @gol
2925 -Woverride-init @gol
2928 -Wuninitialized @gol
2929 -Wunused-parameter @r{(only with} @option{-Wunused} @r{or} @option{-Wall}@r{)} @gol
2932 The option @option{-Wextra} also prints warning messages for the
2938 A pointer is compared against integer zero with @samp{<}, @samp{<=},
2939 @samp{>}, or @samp{>=}.
2942 (C++ only) An enumerator and a non-enumerator both appear in a
2943 conditional expression.
2946 (C++ only) Ambiguous virtual bases.
2949 (C++ only) Subscripting an array which has been declared @samp{register}.
2952 (C++ only) Taking the address of a variable which has been declared
2956 (C++ only) A base class is not initialized in a derived class' copy
2961 @item -Wchar-subscripts
2962 @opindex Wchar-subscripts
2963 @opindex Wno-char-subscripts
2964 Warn if an array subscript has type @code{char}. This is a common cause
2965 of error, as programmers often forget that this type is signed on some
2967 This warning is enabled by @option{-Wall}.
2971 @opindex Wno-comment
2972 Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*}
2973 comment, or whenever a Backslash-Newline appears in a @samp{//} comment.
2974 This warning is enabled by @option{-Wall}.
2979 @opindex ffreestanding
2980 @opindex fno-builtin
2981 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
2982 the arguments supplied have types appropriate to the format string
2983 specified, and that the conversions specified in the format string make
2984 sense. This includes standard functions, and others specified by format
2985 attributes (@pxref{Function Attributes}), in the @code{printf},
2986 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
2987 not in the C standard) families (or other target-specific families).
2988 Which functions are checked without format attributes having been
2989 specified depends on the standard version selected, and such checks of
2990 functions without the attribute specified are disabled by
2991 @option{-ffreestanding} or @option{-fno-builtin}.
2993 The formats are checked against the format features supported by GNU
2994 libc version 2.2. These include all ISO C90 and C99 features, as well
2995 as features from the Single Unix Specification and some BSD and GNU
2996 extensions. Other library implementations may not support all these
2997 features; GCC does not support warning about features that go beyond a
2998 particular library's limitations. However, if @option{-pedantic} is used
2999 with @option{-Wformat}, warnings will be given about format features not
3000 in the selected standard version (but not for @code{strfmon} formats,
3001 since those are not in any version of the C standard). @xref{C Dialect
3002 Options,,Options Controlling C Dialect}.
3004 Since @option{-Wformat} also checks for null format arguments for
3005 several functions, @option{-Wformat} also implies @option{-Wnonnull}.
3007 @option{-Wformat} is included in @option{-Wall}. For more control over some
3008 aspects of format checking, the options @option{-Wformat-y2k},
3009 @option{-Wno-format-extra-args}, @option{-Wno-format-zero-length},
3010 @option{-Wformat-nonliteral}, @option{-Wformat-security}, and
3011 @option{-Wformat=2} are available, but are not included in @option{-Wall}.
3014 @opindex Wformat-y2k
3015 @opindex Wno-format-y2k
3016 If @option{-Wformat} is specified, also warn about @code{strftime}
3017 formats which may yield only a two-digit year.
3019 @item -Wno-format-contains-nul
3020 @opindex Wno-format-contains-nul
3021 @opindex Wformat-contains-nul
3022 If @option{-Wformat} is specified, do not warn about format strings that
3025 @item -Wno-format-extra-args
3026 @opindex Wno-format-extra-args
3027 @opindex Wformat-extra-args
3028 If @option{-Wformat} is specified, do not warn about excess arguments to a
3029 @code{printf} or @code{scanf} format function. The C standard specifies
3030 that such arguments are ignored.
3032 Where the unused arguments lie between used arguments that are
3033 specified with @samp{$} operand number specifications, normally
3034 warnings are still given, since the implementation could not know what
3035 type to pass to @code{va_arg} to skip the unused arguments. However,
3036 in the case of @code{scanf} formats, this option will suppress the
3037 warning if the unused arguments are all pointers, since the Single
3038 Unix Specification says that such unused arguments are allowed.
3040 @item -Wno-format-zero-length @r{(C and Objective-C only)}
3041 @opindex Wno-format-zero-length
3042 @opindex Wformat-zero-length
3043 If @option{-Wformat} is specified, do not warn about zero-length formats.
3044 The C standard specifies that zero-length formats are allowed.
3046 @item -Wformat-nonliteral
3047 @opindex Wformat-nonliteral
3048 @opindex Wno-format-nonliteral
3049 If @option{-Wformat} is specified, also warn if the format string is not a
3050 string literal and so cannot be checked, unless the format function
3051 takes its format arguments as a @code{va_list}.
3053 @item -Wformat-security
3054 @opindex Wformat-security
3055 @opindex Wno-format-security
3056 If @option{-Wformat} is specified, also warn about uses of format
3057 functions that represent possible security problems. At present, this
3058 warns about calls to @code{printf} and @code{scanf} functions where the
3059 format string is not a string literal and there are no format arguments,
3060 as in @code{printf (foo);}. This may be a security hole if the format
3061 string came from untrusted input and contains @samp{%n}. (This is
3062 currently a subset of what @option{-Wformat-nonliteral} warns about, but
3063 in future warnings may be added to @option{-Wformat-security} that are not
3064 included in @option{-Wformat-nonliteral}.)
3068 @opindex Wno-format=2
3069 Enable @option{-Wformat} plus format checks not included in
3070 @option{-Wformat}. Currently equivalent to @samp{-Wformat
3071 -Wformat-nonliteral -Wformat-security -Wformat-y2k}.
3073 @item -Wnonnull @r{(C and Objective-C only)}
3075 @opindex Wno-nonnull
3076 Warn about passing a null pointer for arguments marked as
3077 requiring a non-null value by the @code{nonnull} function attribute.
3079 @option{-Wnonnull} is included in @option{-Wall} and @option{-Wformat}. It
3080 can be disabled with the @option{-Wno-nonnull} option.
3082 @item -Winit-self @r{(C, C++, Objective-C and Objective-C++ only)}
3084 @opindex Wno-init-self
3085 Warn about uninitialized variables which are initialized with themselves.
3086 Note this option can only be used with the @option{-Wuninitialized} option.
3088 For example, GCC will warn about @code{i} being uninitialized in the
3089 following snippet only when @option{-Winit-self} has been specified:
3100 @item -Wimplicit-int @r{(C and Objective-C only)}
3101 @opindex Wimplicit-int
3102 @opindex Wno-implicit-int
3103 Warn when a declaration does not specify a type.
3104 This warning is enabled by @option{-Wall}.
3106 @item -Wimplicit-function-declaration @r{(C and Objective-C only)}
3107 @opindex Wimplicit-function-declaration
3108 @opindex Wno-implicit-function-declaration
3109 Give a warning whenever a function is used before being declared. In
3110 C99 mode (@option{-std=c99} or @option{-std=gnu99}), this warning is
3111 enabled by default and it is made into an error by
3112 @option{-pedantic-errors}. This warning is also enabled by
3117 @opindex Wno-implicit
3118 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
3119 This warning is enabled by @option{-Wall}.
3121 @item -Wignored-qualifiers @r{(C and C++ only)}
3122 @opindex Wignored-qualifiers
3123 @opindex Wno-ignored-qualifiers
3124 Warn if the return type of a function has a type qualifier
3125 such as @code{const}. For ISO C such a type qualifier has no effect,
3126 since the value returned by a function is not an lvalue.
3127 For C++, the warning is only emitted for scalar types or @code{void}.
3128 ISO C prohibits qualified @code{void} return types on function
3129 definitions, so such return types always receive a warning
3130 even without this option.
3132 This warning is also enabled by @option{-Wextra}.
3137 Warn if the type of @samp{main} is suspicious. @samp{main} should be
3138 a function with external linkage, returning int, taking either zero
3139 arguments, two, or three arguments of appropriate types. This warning
3140 is enabled by default in C++ and is enabled by either @option{-Wall}
3141 or @option{-pedantic}.
3143 @item -Wmissing-braces
3144 @opindex Wmissing-braces
3145 @opindex Wno-missing-braces
3146 Warn if an aggregate or union initializer is not fully bracketed. In
3147 the following example, the initializer for @samp{a} is not fully
3148 bracketed, but that for @samp{b} is fully bracketed.
3151 int a[2][2] = @{ 0, 1, 2, 3 @};
3152 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
3155 This warning is enabled by @option{-Wall}.
3157 @item -Wmissing-include-dirs @r{(C, C++, Objective-C and Objective-C++ only)}
3158 @opindex Wmissing-include-dirs
3159 @opindex Wno-missing-include-dirs
3160 Warn if a user-supplied include directory does not exist.
3163 @opindex Wparentheses
3164 @opindex Wno-parentheses
3165 Warn if parentheses are omitted in certain contexts, such
3166 as when there is an assignment in a context where a truth value
3167 is expected, or when operators are nested whose precedence people
3168 often get confused about.
3170 Also warn if a comparison like @samp{x<=y<=z} appears; this is
3171 equivalent to @samp{(x<=y ? 1 : 0) <= z}, which is a different
3172 interpretation from that of ordinary mathematical notation.
3174 Also warn about constructions where there may be confusion to which
3175 @code{if} statement an @code{else} branch belongs. Here is an example of
3190 In C/C++, every @code{else} branch belongs to the innermost possible
3191 @code{if} statement, which in this example is @code{if (b)}. This is
3192 often not what the programmer expected, as illustrated in the above
3193 example by indentation the programmer chose. When there is the
3194 potential for this confusion, GCC will issue a warning when this flag
3195 is specified. To eliminate the warning, add explicit braces around
3196 the innermost @code{if} statement so there is no way the @code{else}
3197 could belong to the enclosing @code{if}. The resulting code would
3214 This warning is enabled by @option{-Wall}.
3216 @item -Wsequence-point
3217 @opindex Wsequence-point
3218 @opindex Wno-sequence-point
3219 Warn about code that may have undefined semantics because of violations
3220 of sequence point rules in the C and C++ standards.
3222 The C and C++ standards defines the order in which expressions in a C/C++
3223 program are evaluated in terms of @dfn{sequence points}, which represent
3224 a partial ordering between the execution of parts of the program: those
3225 executed before the sequence point, and those executed after it. These
3226 occur after the evaluation of a full expression (one which is not part
3227 of a larger expression), after the evaluation of the first operand of a
3228 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
3229 function is called (but after the evaluation of its arguments and the
3230 expression denoting the called function), and in certain other places.
3231 Other than as expressed by the sequence point rules, the order of
3232 evaluation of subexpressions of an expression is not specified. All
3233 these rules describe only a partial order rather than a total order,
3234 since, for example, if two functions are called within one expression
3235 with no sequence point between them, the order in which the functions
3236 are called is not specified. However, the standards committee have
3237 ruled that function calls do not overlap.
3239 It is not specified when between sequence points modifications to the
3240 values of objects take effect. Programs whose behavior depends on this
3241 have undefined behavior; the C and C++ standards specify that ``Between
3242 the previous and next sequence point an object shall have its stored
3243 value modified at most once by the evaluation of an expression.
3244 Furthermore, the prior value shall be read only to determine the value
3245 to be stored.''. If a program breaks these rules, the results on any
3246 particular implementation are entirely unpredictable.
3248 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
3249 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
3250 diagnosed by this option, and it may give an occasional false positive
3251 result, but in general it has been found fairly effective at detecting
3252 this sort of problem in programs.
3254 The standard is worded confusingly, therefore there is some debate
3255 over the precise meaning of the sequence point rules in subtle cases.
3256 Links to discussions of the problem, including proposed formal
3257 definitions, may be found on the GCC readings page, at
3258 @w{@uref{http://gcc.gnu.org/readings.html}}.
3260 This warning is enabled by @option{-Wall} for C and C++.
3263 @opindex Wreturn-type
3264 @opindex Wno-return-type
3265 Warn whenever a function is defined with a return-type that defaults
3266 to @code{int}. Also warn about any @code{return} statement with no
3267 return-value in a function whose return-type is not @code{void}
3268 (falling off the end of the function body is considered returning
3269 without a value), and about a @code{return} statement with an
3270 expression in a function whose return-type is @code{void}.
3272 For C++, a function without return type always produces a diagnostic
3273 message, even when @option{-Wno-return-type} is specified. The only
3274 exceptions are @samp{main} and functions defined in system headers.
3276 This warning is enabled by @option{-Wall}.
3281 Warn whenever a @code{switch} statement has an index of enumerated type
3282 and lacks a @code{case} for one or more of the named codes of that
3283 enumeration. (The presence of a @code{default} label prevents this
3284 warning.) @code{case} labels outside the enumeration range also
3285 provoke warnings when this option is used (even if there is a
3286 @code{default} label).
3287 This warning is enabled by @option{-Wall}.
3289 @item -Wswitch-default
3290 @opindex Wswitch-default
3291 @opindex Wno-switch-default
3292 Warn whenever a @code{switch} statement does not have a @code{default}
3296 @opindex Wswitch-enum
3297 @opindex Wno-switch-enum
3298 Warn whenever a @code{switch} statement has an index of enumerated type
3299 and lacks a @code{case} for one or more of the named codes of that
3300 enumeration. @code{case} labels outside the enumeration range also
3301 provoke warnings when this option is used. The only difference
3302 between @option{-Wswitch} and this option is that this option gives a
3303 warning about an omitted enumeration code even if there is a
3304 @code{default} label.
3306 @item -Wsync-nand @r{(C and C++ only)}
3308 @opindex Wno-sync-nand
3309 Warn when @code{__sync_fetch_and_nand} and @code{__sync_nand_and_fetch}
3310 built-in functions are used. These functions changed semantics in GCC 4.4.
3314 @opindex Wno-trigraphs
3315 Warn if any trigraphs are encountered that might change the meaning of
3316 the program (trigraphs within comments are not warned about).
3317 This warning is enabled by @option{-Wall}.
3319 @item -Wunused-function
3320 @opindex Wunused-function
3321 @opindex Wno-unused-function
3322 Warn whenever a static function is declared but not defined or a
3323 non-inline static function is unused.
3324 This warning is enabled by @option{-Wall}.
3326 @item -Wunused-label
3327 @opindex Wunused-label
3328 @opindex Wno-unused-label
3329 Warn whenever a label is declared but not used.
3330 This warning is enabled by @option{-Wall}.
3332 To suppress this warning use the @samp{unused} attribute
3333 (@pxref{Variable Attributes}).
3335 @item -Wunused-parameter
3336 @opindex Wunused-parameter
3337 @opindex Wno-unused-parameter
3338 Warn whenever a function parameter is unused aside from its declaration.
3340 To suppress this warning use the @samp{unused} attribute
3341 (@pxref{Variable Attributes}).
3343 @item -Wno-unused-result
3344 @opindex Wunused-result
3345 @opindex Wno-unused-result
3346 Do not warn if a caller of a function marked with attribute
3347 @code{warn_unused_result} (@pxref{Variable Attributes}) does not use
3348 its return value. The default is @option{-Wunused-result}.
3350 @item -Wunused-variable
3351 @opindex Wunused-variable
3352 @opindex Wno-unused-variable
3353 Warn whenever a local variable or non-constant static variable is unused
3354 aside from its declaration.
3355 This warning is enabled by @option{-Wall}.
3357 To suppress this warning use the @samp{unused} attribute
3358 (@pxref{Variable Attributes}).
3360 @item -Wunused-value
3361 @opindex Wunused-value
3362 @opindex Wno-unused-value
3363 Warn whenever a statement computes a result that is explicitly not
3364 used. To suppress this warning cast the unused expression to
3365 @samp{void}. This includes an expression-statement or the left-hand
3366 side of a comma expression that contains no side effects. For example,
3367 an expression such as @samp{x[i,j]} will cause a warning, while
3368 @samp{x[(void)i,j]} will not.
3370 This warning is enabled by @option{-Wall}.
3375 All the above @option{-Wunused} options combined.
3377 In order to get a warning about an unused function parameter, you must
3378 either specify @samp{-Wextra -Wunused} (note that @samp{-Wall} implies
3379 @samp{-Wunused}), or separately specify @option{-Wunused-parameter}.
3381 @item -Wuninitialized
3382 @opindex Wuninitialized
3383 @opindex Wno-uninitialized
3384 Warn if an automatic variable is used without first being initialized
3385 or if a variable may be clobbered by a @code{setjmp} call. In C++,
3386 warn if a non-static reference or non-static @samp{const} member
3387 appears in a class without constructors.
3389 If you want to warn about code which uses the uninitialized value of the
3390 variable in its own initializer, use the @option{-Winit-self} option.
3392 These warnings occur for individual uninitialized or clobbered
3393 elements of structure, union or array variables as well as for
3394 variables which are uninitialized or clobbered as a whole. They do
3395 not occur for variables or elements declared @code{volatile}. Because
3396 these warnings depend on optimization, the exact variables or elements
3397 for which there are warnings will depend on the precise optimization
3398 options and version of GCC used.
3400 Note that there may be no warning about a variable that is used only
3401 to compute a value that itself is never used, because such
3402 computations may be deleted by data flow analysis before the warnings
3405 These warnings are made optional because GCC is not smart
3406 enough to see all the reasons why the code might be correct
3407 despite appearing to have an error. Here is one example of how
3428 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
3429 always initialized, but GCC doesn't know this. Here is
3430 another common case:
3435 if (change_y) save_y = y, y = new_y;
3437 if (change_y) y = save_y;
3442 This has no bug because @code{save_y} is used only if it is set.
3444 @cindex @code{longjmp} warnings
3445 This option also warns when a non-volatile automatic variable might be
3446 changed by a call to @code{longjmp}. These warnings as well are possible
3447 only in optimizing compilation.
3449 The compiler sees only the calls to @code{setjmp}. It cannot know
3450 where @code{longjmp} will be called; in fact, a signal handler could
3451 call it at any point in the code. As a result, you may get a warning
3452 even when there is in fact no problem because @code{longjmp} cannot
3453 in fact be called at the place which would cause a problem.
3455 Some spurious warnings can be avoided if you declare all the functions
3456 you use that never return as @code{noreturn}. @xref{Function
3459 This warning is enabled by @option{-Wall} or @option{-Wextra}.
3461 @item -Wunknown-pragmas
3462 @opindex Wunknown-pragmas
3463 @opindex Wno-unknown-pragmas
3464 @cindex warning for unknown pragmas
3465 @cindex unknown pragmas, warning
3466 @cindex pragmas, warning of unknown
3467 Warn when a #pragma directive is encountered which is not understood by
3468 GCC@. If this command line option is used, warnings will even be issued
3469 for unknown pragmas in system header files. This is not the case if
3470 the warnings were only enabled by the @option{-Wall} command line option.
3473 @opindex Wno-pragmas
3475 Do not warn about misuses of pragmas, such as incorrect parameters,
3476 invalid syntax, or conflicts between pragmas. See also
3477 @samp{-Wunknown-pragmas}.
3479 @item -Wstrict-aliasing
3480 @opindex Wstrict-aliasing
3481 @opindex Wno-strict-aliasing
3482 This option is only active when @option{-fstrict-aliasing} is active.
3483 It warns about code which might break the strict aliasing rules that the
3484 compiler is using for optimization. The warning does not catch all
3485 cases, but does attempt to catch the more common pitfalls. It is
3486 included in @option{-Wall}.
3487 It is equivalent to @option{-Wstrict-aliasing=3}
3489 @item -Wstrict-aliasing=n
3490 @opindex Wstrict-aliasing=n
3491 @opindex Wno-strict-aliasing=n
3492 This option is only active when @option{-fstrict-aliasing} is active.
3493 It warns about code which might break the strict aliasing rules that the
3494 compiler is using for optimization.
3495 Higher levels correspond to higher accuracy (fewer false positives).
3496 Higher levels also correspond to more effort, similar to the way -O works.
3497 @option{-Wstrict-aliasing} is equivalent to @option{-Wstrict-aliasing=n},
3500 Level 1: Most aggressive, quick, least accurate.
3501 Possibly useful when higher levels
3502 do not warn but -fstrict-aliasing still breaks the code, as it has very few
3503 false negatives. However, it has many false positives.
3504 Warns for all pointer conversions between possibly incompatible types,
3505 even if never dereferenced. Runs in the frontend only.
3507 Level 2: Aggressive, quick, not too precise.
3508 May still have many false positives (not as many as level 1 though),
3509 and few false negatives (but possibly more than level 1).
3510 Unlike level 1, it only warns when an address is taken. Warns about
3511 incomplete types. Runs in the frontend only.
3513 Level 3 (default for @option{-Wstrict-aliasing}):
3514 Should have very few false positives and few false
3515 negatives. Slightly slower than levels 1 or 2 when optimization is enabled.
3516 Takes care of the common pun+dereference pattern in the frontend:
3517 @code{*(int*)&some_float}.
3518 If optimization is enabled, it also runs in the backend, where it deals
3519 with multiple statement cases using flow-sensitive points-to information.
3520 Only warns when the converted pointer is dereferenced.
3521 Does not warn about incomplete types.
3523 @item -Wstrict-overflow
3524 @itemx -Wstrict-overflow=@var{n}
3525 @opindex Wstrict-overflow
3526 @opindex Wno-strict-overflow
3527 This option is only active when @option{-fstrict-overflow} is active.
3528 It warns about cases where the compiler optimizes based on the
3529 assumption that signed overflow does not occur. Note that it does not
3530 warn about all cases where the code might overflow: it only warns
3531 about cases where the compiler implements some optimization. Thus
3532 this warning depends on the optimization level.
3534 An optimization which assumes that signed overflow does not occur is
3535 perfectly safe if the values of the variables involved are such that
3536 overflow never does, in fact, occur. Therefore this warning can
3537 easily give a false positive: a warning about code which is not
3538 actually a problem. To help focus on important issues, several
3539 warning levels are defined. No warnings are issued for the use of
3540 undefined signed overflow when estimating how many iterations a loop
3541 will require, in particular when determining whether a loop will be
3545 @item -Wstrict-overflow=1
3546 Warn about cases which are both questionable and easy to avoid. For
3547 example: @code{x + 1 > x}; with @option{-fstrict-overflow}, the
3548 compiler will simplify this to @code{1}. This level of
3549 @option{-Wstrict-overflow} is enabled by @option{-Wall}; higher levels
3550 are not, and must be explicitly requested.
3552 @item -Wstrict-overflow=2
3553 Also warn about other cases where a comparison is simplified to a
3554 constant. For example: @code{abs (x) >= 0}. This can only be
3555 simplified when @option{-fstrict-overflow} is in effect, because
3556 @code{abs (INT_MIN)} overflows to @code{INT_MIN}, which is less than
3557 zero. @option{-Wstrict-overflow} (with no level) is the same as
3558 @option{-Wstrict-overflow=2}.
3560 @item -Wstrict-overflow=3
3561 Also warn about other cases where a comparison is simplified. For
3562 example: @code{x + 1 > 1} will be simplified to @code{x > 0}.
3564 @item -Wstrict-overflow=4
3565 Also warn about other simplifications not covered by the above cases.
3566 For example: @code{(x * 10) / 5} will be simplified to @code{x * 2}.
3568 @item -Wstrict-overflow=5
3569 Also warn about cases where the compiler reduces the magnitude of a
3570 constant involved in a comparison. For example: @code{x + 2 > y} will
3571 be simplified to @code{x + 1 >= y}. This is reported only at the
3572 highest warning level because this simplification applies to many
3573 comparisons, so this warning level will give a very large number of
3577 @item -Warray-bounds
3578 @opindex Wno-array-bounds
3579 @opindex Warray-bounds
3580 This option is only active when @option{-ftree-vrp} is active
3581 (default for -O2 and above). It warns about subscripts to arrays
3582 that are always out of bounds. This warning is enabled by @option{-Wall}.
3584 @item -Wno-div-by-zero
3585 @opindex Wno-div-by-zero
3586 @opindex Wdiv-by-zero
3587 Do not warn about compile-time integer division by zero. Floating point
3588 division by zero is not warned about, as it can be a legitimate way of
3589 obtaining infinities and NaNs.
3591 @item -Wsystem-headers
3592 @opindex Wsystem-headers
3593 @opindex Wno-system-headers
3594 @cindex warnings from system headers
3595 @cindex system headers, warnings from
3596 Print warning messages for constructs found in system header files.
3597 Warnings from system headers are normally suppressed, on the assumption
3598 that they usually do not indicate real problems and would only make the
3599 compiler output harder to read. Using this command line option tells
3600 GCC to emit warnings from system headers as if they occurred in user
3601 code. However, note that using @option{-Wall} in conjunction with this
3602 option will @emph{not} warn about unknown pragmas in system
3603 headers---for that, @option{-Wunknown-pragmas} must also be used.
3606 @opindex Wfloat-equal
3607 @opindex Wno-float-equal
3608 Warn if floating point values are used in equality comparisons.
3610 The idea behind this is that sometimes it is convenient (for the
3611 programmer) to consider floating-point values as approximations to
3612 infinitely precise real numbers. If you are doing this, then you need
3613 to compute (by analyzing the code, or in some other way) the maximum or
3614 likely maximum error that the computation introduces, and allow for it
3615 when performing comparisons (and when producing output, but that's a
3616 different problem). In particular, instead of testing for equality, you
3617 would check to see whether the two values have ranges that overlap; and
3618 this is done with the relational operators, so equality comparisons are
3621 @item -Wtraditional @r{(C and Objective-C only)}
3622 @opindex Wtraditional
3623 @opindex Wno-traditional
3624 Warn about certain constructs that behave differently in traditional and
3625 ISO C@. Also warn about ISO C constructs that have no traditional C
3626 equivalent, and/or problematic constructs which should be avoided.
3630 Macro parameters that appear within string literals in the macro body.
3631 In traditional C macro replacement takes place within string literals,
3632 but does not in ISO C@.
3635 In traditional C, some preprocessor directives did not exist.
3636 Traditional preprocessors would only consider a line to be a directive
3637 if the @samp{#} appeared in column 1 on the line. Therefore
3638 @option{-Wtraditional} warns about directives that traditional C
3639 understands but would ignore because the @samp{#} does not appear as the
3640 first character on the line. It also suggests you hide directives like
3641 @samp{#pragma} not understood by traditional C by indenting them. Some
3642 traditional implementations would not recognize @samp{#elif}, so it
3643 suggests avoiding it altogether.
3646 A function-like macro that appears without arguments.
3649 The unary plus operator.
3652 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating point
3653 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
3654 constants.) Note, these suffixes appear in macros defined in the system
3655 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
3656 Use of these macros in user code might normally lead to spurious
3657 warnings, however GCC's integrated preprocessor has enough context to
3658 avoid warning in these cases.
3661 A function declared external in one block and then used after the end of
3665 A @code{switch} statement has an operand of type @code{long}.
3668 A non-@code{static} function declaration follows a @code{static} one.
3669 This construct is not accepted by some traditional C compilers.
3672 The ISO type of an integer constant has a different width or
3673 signedness from its traditional type. This warning is only issued if
3674 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
3675 typically represent bit patterns, are not warned about.
3678 Usage of ISO string concatenation is detected.
3681 Initialization of automatic aggregates.
3684 Identifier conflicts with labels. Traditional C lacks a separate
3685 namespace for labels.
3688 Initialization of unions. If the initializer is zero, the warning is
3689 omitted. This is done under the assumption that the zero initializer in
3690 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
3691 initializer warnings and relies on default initialization to zero in the
3695 Conversions by prototypes between fixed/floating point values and vice
3696 versa. The absence of these prototypes when compiling with traditional
3697 C would cause serious problems. This is a subset of the possible
3698 conversion warnings, for the full set use @option{-Wtraditional-conversion}.
3701 Use of ISO C style function definitions. This warning intentionally is
3702 @emph{not} issued for prototype declarations or variadic functions
3703 because these ISO C features will appear in your code when using
3704 libiberty's traditional C compatibility macros, @code{PARAMS} and
3705 @code{VPARAMS}. This warning is also bypassed for nested functions
3706 because that feature is already a GCC extension and thus not relevant to
3707 traditional C compatibility.
3710 @item -Wtraditional-conversion @r{(C and Objective-C only)}
3711 @opindex Wtraditional-conversion
3712 @opindex Wno-traditional-conversion
3713 Warn if a prototype causes a type conversion that is different from what
3714 would happen to the same argument in the absence of a prototype. This
3715 includes conversions of fixed point to floating and vice versa, and
3716 conversions changing the width or signedness of a fixed point argument
3717 except when the same as the default promotion.
3719 @item -Wdeclaration-after-statement @r{(C and Objective-C only)}
3720 @opindex Wdeclaration-after-statement
3721 @opindex Wno-declaration-after-statement
3722 Warn when a declaration is found after a statement in a block. This
3723 construct, known from C++, was introduced with ISO C99 and is by default
3724 allowed in GCC@. It is not supported by ISO C90 and was not supported by
3725 GCC versions before GCC 3.0. @xref{Mixed Declarations}.
3730 Warn if an undefined identifier is evaluated in an @samp{#if} directive.
3732 @item -Wno-endif-labels
3733 @opindex Wno-endif-labels
3734 @opindex Wendif-labels
3735 Do not warn whenever an @samp{#else} or an @samp{#endif} are followed by text.
3740 Warn whenever a local variable shadows another local variable, parameter or
3741 global variable or whenever a built-in function is shadowed.
3743 @item -Wlarger-than=@var{len}
3744 @opindex Wlarger-than=@var{len}
3745 @opindex Wlarger-than-@var{len}
3746 Warn whenever an object of larger than @var{len} bytes is defined.
3748 @item -Wframe-larger-than=@var{len}
3749 @opindex Wframe-larger-than
3750 Warn if the size of a function frame is larger than @var{len} bytes.
3751 The computation done to determine the stack frame size is approximate
3752 and not conservative.
3753 The actual requirements may be somewhat greater than @var{len}
3754 even if you do not get a warning. In addition, any space allocated
3755 via @code{alloca}, variable-length arrays, or related constructs
3756 is not included by the compiler when determining
3757 whether or not to issue a warning.
3759 @item -Wunsafe-loop-optimizations
3760 @opindex Wunsafe-loop-optimizations
3761 @opindex Wno-unsafe-loop-optimizations
3762 Warn if the loop cannot be optimized because the compiler could not
3763 assume anything on the bounds of the loop indices. With
3764 @option{-funsafe-loop-optimizations} warn if the compiler made
3767 @item -Wno-pedantic-ms-format @r{(MinGW targets only)}
3768 @opindex Wno-pedantic-ms-format
3769 @opindex Wpedantic-ms-format
3770 Disables the warnings about non-ISO @code{printf} / @code{scanf} format
3771 width specifiers @code{I32}, @code{I64}, and @code{I} used on Windows targets
3772 depending on the MS runtime, when you are using the options @option{-Wformat}
3773 and @option{-pedantic} without gnu-extensions.
3775 @item -Wpointer-arith
3776 @opindex Wpointer-arith
3777 @opindex Wno-pointer-arith
3778 Warn about anything that depends on the ``size of'' a function type or
3779 of @code{void}. GNU C assigns these types a size of 1, for
3780 convenience in calculations with @code{void *} pointers and pointers
3781 to functions. In C++, warn also when an arithmetic operation involves
3782 @code{NULL}. This warning is also enabled by @option{-pedantic}.
3785 @opindex Wtype-limits
3786 @opindex Wno-type-limits
3787 Warn if a comparison is always true or always false due to the limited
3788 range of the data type, but do not warn for constant expressions. For
3789 example, warn if an unsigned variable is compared against zero with
3790 @samp{<} or @samp{>=}. This warning is also enabled by
3793 @item -Wbad-function-cast @r{(C and Objective-C only)}
3794 @opindex Wbad-function-cast
3795 @opindex Wno-bad-function-cast
3796 Warn whenever a function call is cast to a non-matching type.
3797 For example, warn if @code{int malloc()} is cast to @code{anything *}.
3799 @item -Wc++-compat @r{(C and Objective-C only)}
3800 Warn about ISO C constructs that are outside of the common subset of
3801 ISO C and ISO C++, e.g.@: request for implicit conversion from
3802 @code{void *} to a pointer to non-@code{void} type.
3804 @item -Wc++0x-compat @r{(C++ and Objective-C++ only)}
3805 Warn about C++ constructs whose meaning differs between ISO C++ 1998 and
3806 ISO C++ 200x, e.g., identifiers in ISO C++ 1998 that will become keywords
3807 in ISO C++ 200x. This warning is enabled by @option{-Wall}.
3811 @opindex Wno-cast-qual
3812 Warn whenever a pointer is cast so as to remove a type qualifier from
3813 the target type. For example, warn if a @code{const char *} is cast
3814 to an ordinary @code{char *}.
3816 Also warn when making a cast which introduces a type qualifier in an
3817 unsafe way. For example, casting @code{char **} to @code{const char **}
3818 is unsafe, as in this example:
3821 /* p is char ** value. */
3822 const char **q = (const char **) p;
3823 /* Assignment of readonly string to const char * is OK. */
3825 /* Now char** pointer points to read-only memory. */
3830 @opindex Wcast-align
3831 @opindex Wno-cast-align
3832 Warn whenever a pointer is cast such that the required alignment of the
3833 target is increased. For example, warn if a @code{char *} is cast to
3834 an @code{int *} on machines where integers can only be accessed at
3835 two- or four-byte boundaries.
3837 @item -Wwrite-strings
3838 @opindex Wwrite-strings
3839 @opindex Wno-write-strings
3840 When compiling C, give string constants the type @code{const
3841 char[@var{length}]} so that copying the address of one into a
3842 non-@code{const} @code{char *} pointer will get a warning. These
3843 warnings will help you find at compile time code that can try to write
3844 into a string constant, but only if you have been very careful about
3845 using @code{const} in declarations and prototypes. Otherwise, it will
3846 just be a nuisance. This is why we did not make @option{-Wall} request
3849 When compiling C++, warn about the deprecated conversion from string
3850 literals to @code{char *}. This warning is enabled by default for C++
3855 @opindex Wno-clobbered
3856 Warn for variables that might be changed by @samp{longjmp} or
3857 @samp{vfork}. This warning is also enabled by @option{-Wextra}.
3860 @opindex Wconversion
3861 @opindex Wno-conversion
3862 Warn for implicit conversions that may alter a value. This includes
3863 conversions between real and integer, like @code{abs (x)} when
3864 @code{x} is @code{double}; conversions between signed and unsigned,
3865 like @code{unsigned ui = -1}; and conversions to smaller types, like
3866 @code{sqrtf (M_PI)}. Do not warn for explicit casts like @code{abs
3867 ((int) x)} and @code{ui = (unsigned) -1}, or if the value is not
3868 changed by the conversion like in @code{abs (2.0)}. Warnings about
3869 conversions between signed and unsigned integers can be disabled by
3870 using @option{-Wno-sign-conversion}.
3872 For C++, also warn for confusing overload resolution for user-defined
3873 conversions; and conversions that will never use a type conversion
3874 operator: conversions to @code{void}, the same type, a base class or a
3875 reference to them. Warnings about conversions between signed and
3876 unsigned integers are disabled by default in C++ unless
3877 @option{-Wsign-conversion} is explicitly enabled.
3879 @item -Wno-conversion-null @r{(C++ and Objective-C++ only)}
3880 @opindex Wconversion-null
3881 @opindex Wno-conversion-null
3882 Do not warn for conversions between @code{NULL} and non-pointer
3883 types. @option{-Wconversion-null} is enabled by default.
3886 @opindex Wempty-body
3887 @opindex Wno-empty-body
3888 Warn if an empty body occurs in an @samp{if}, @samp{else} or @samp{do
3889 while} statement. This warning is also enabled by @option{-Wextra}.
3891 @item -Wenum-compare
3892 @opindex Wenum-compare
3893 @opindex Wno-enum-compare
3894 Warn about a comparison between values of different enum types. In C++
3895 this warning is enabled by default. In C this warning is enabled by
3898 @item -Wjump-misses-init @r{(C, Objective-C only)}
3899 @opindex Wjump-misses-init
3900 @opindex Wno-jump-misses-init
3901 Warn if a @code{goto} statement or a @code{switch} statement jumps
3902 forward across the initialization of a variable, or jumps backward to a
3903 label after the variable has been initialized. This only warns about
3904 variables which are initialized when they are declared. This warning is
3905 only supported for C and Objective C; in C++ this sort of branch is an
3908 @option{-Wjump-misses-init} is included in @option{-Wc++-compat}. It
3909 can be disabled with the @option{-Wno-jump-misses-init} option.
3911 @item -Wsign-compare
3912 @opindex Wsign-compare
3913 @opindex Wno-sign-compare
3914 @cindex warning for comparison of signed and unsigned values
3915 @cindex comparison of signed and unsigned values, warning
3916 @cindex signed and unsigned values, comparison warning
3917 Warn when a comparison between signed and unsigned values could produce
3918 an incorrect result when the signed value is converted to unsigned.
3919 This warning is also enabled by @option{-Wextra}; to get the other warnings
3920 of @option{-Wextra} without this warning, use @samp{-Wextra -Wno-sign-compare}.
3922 @item -Wsign-conversion
3923 @opindex Wsign-conversion
3924 @opindex Wno-sign-conversion
3925 Warn for implicit conversions that may change the sign of an integer
3926 value, like assigning a signed integer expression to an unsigned
3927 integer variable. An explicit cast silences the warning. In C, this
3928 option is enabled also by @option{-Wconversion}.
3932 @opindex Wno-address
3933 Warn about suspicious uses of memory addresses. These include using
3934 the address of a function in a conditional expression, such as
3935 @code{void func(void); if (func)}, and comparisons against the memory
3936 address of a string literal, such as @code{if (x == "abc")}. Such
3937 uses typically indicate a programmer error: the address of a function
3938 always evaluates to true, so their use in a conditional usually
3939 indicate that the programmer forgot the parentheses in a function
3940 call; and comparisons against string literals result in unspecified
3941 behavior and are not portable in C, so they usually indicate that the
3942 programmer intended to use @code{strcmp}. This warning is enabled by
3946 @opindex Wlogical-op
3947 @opindex Wno-logical-op
3948 Warn about suspicious uses of logical operators in expressions.
3949 This includes using logical operators in contexts where a
3950 bit-wise operator is likely to be expected.
3952 @item -Waggregate-return
3953 @opindex Waggregate-return
3954 @opindex Wno-aggregate-return
3955 Warn if any functions that return structures or unions are defined or
3956 called. (In languages where you can return an array, this also elicits
3959 @item -Wno-attributes
3960 @opindex Wno-attributes
3961 @opindex Wattributes
3962 Do not warn if an unexpected @code{__attribute__} is used, such as
3963 unrecognized attributes, function attributes applied to variables,
3964 etc. This will not stop errors for incorrect use of supported
3967 @item -Wno-builtin-macro-redefined
3968 @opindex Wno-builtin-macro-redefined
3969 @opindex Wbuiltin-macro-redefined
3970 Do not warn if certain built-in macros are redefined. This suppresses
3971 warnings for redefinition of @code{__TIMESTAMP__}, @code{__TIME__},
3972 @code{__DATE__}, @code{__FILE__}, and @code{__BASE_FILE__}.
3974 @item -Wstrict-prototypes @r{(C and Objective-C only)}
3975 @opindex Wstrict-prototypes
3976 @opindex Wno-strict-prototypes
3977 Warn if a function is declared or defined without specifying the
3978 argument types. (An old-style function definition is permitted without
3979 a warning if preceded by a declaration which specifies the argument
3982 @item -Wold-style-declaration @r{(C and Objective-C only)}
3983 @opindex Wold-style-declaration
3984 @opindex Wno-old-style-declaration
3985 Warn for obsolescent usages, according to the C Standard, in a
3986 declaration. For example, warn if storage-class specifiers like
3987 @code{static} are not the first things in a declaration. This warning
3988 is also enabled by @option{-Wextra}.
3990 @item -Wold-style-definition @r{(C and Objective-C only)}
3991 @opindex Wold-style-definition
3992 @opindex Wno-old-style-definition
3993 Warn if an old-style function definition is used. A warning is given
3994 even if there is a previous prototype.
3996 @item -Wmissing-parameter-type @r{(C and Objective-C only)}
3997 @opindex Wmissing-parameter-type
3998 @opindex Wno-missing-parameter-type
3999 A function parameter is declared without a type specifier in K&R-style
4006 This warning is also enabled by @option{-Wextra}.
4008 @item -Wmissing-prototypes @r{(C and Objective-C only)}
4009 @opindex Wmissing-prototypes
4010 @opindex Wno-missing-prototypes
4011 Warn if a global function is defined without a previous prototype
4012 declaration. This warning is issued even if the definition itself
4013 provides a prototype. The aim is to detect global functions that fail
4014 to be declared in header files.
4016 @item -Wmissing-declarations
4017 @opindex Wmissing-declarations
4018 @opindex Wno-missing-declarations
4019 Warn if a global function is defined without a previous declaration.
4020 Do so even if the definition itself provides a prototype.
4021 Use this option to detect global functions that are not declared in
4022 header files. In C++, no warnings are issued for function templates,
4023 or for inline functions, or for functions in anonymous namespaces.
4025 @item -Wmissing-field-initializers
4026 @opindex Wmissing-field-initializers
4027 @opindex Wno-missing-field-initializers
4031 Warn if a structure's initializer has some fields missing. For
4032 example, the following code would cause such a warning, because
4033 @code{x.h} is implicitly zero:
4036 struct s @{ int f, g, h; @};
4037 struct s x = @{ 3, 4 @};
4040 This option does not warn about designated initializers, so the following
4041 modification would not trigger a warning:
4044 struct s @{ int f, g, h; @};
4045 struct s x = @{ .f = 3, .g = 4 @};
4048 This warning is included in @option{-Wextra}. To get other @option{-Wextra}
4049 warnings without this one, use @samp{-Wextra -Wno-missing-field-initializers}.
4051 @item -Wmissing-noreturn
4052 @opindex Wmissing-noreturn
4053 @opindex Wno-missing-noreturn
4054 Warn about functions which might be candidates for attribute @code{noreturn}.
4055 Note these are only possible candidates, not absolute ones. Care should
4056 be taken to manually verify functions actually do not ever return before
4057 adding the @code{noreturn} attribute, otherwise subtle code generation
4058 bugs could be introduced. You will not get a warning for @code{main} in
4059 hosted C environments.
4061 @item -Wmissing-format-attribute
4062 @opindex Wmissing-format-attribute
4063 @opindex Wno-missing-format-attribute
4066 Warn about function pointers which might be candidates for @code{format}
4067 attributes. Note these are only possible candidates, not absolute ones.
4068 GCC will guess that function pointers with @code{format} attributes that
4069 are used in assignment, initialization, parameter passing or return
4070 statements should have a corresponding @code{format} attribute in the
4071 resulting type. I.e.@: the left-hand side of the assignment or
4072 initialization, the type of the parameter variable, or the return type
4073 of the containing function respectively should also have a @code{format}
4074 attribute to avoid the warning.
4076 GCC will also warn about function definitions which might be
4077 candidates for @code{format} attributes. Again, these are only
4078 possible candidates. GCC will guess that @code{format} attributes
4079 might be appropriate for any function that calls a function like
4080 @code{vprintf} or @code{vscanf}, but this might not always be the
4081 case, and some functions for which @code{format} attributes are
4082 appropriate may not be detected.
4084 @item -Wno-multichar
4085 @opindex Wno-multichar
4087 Do not warn if a multicharacter constant (@samp{'FOOF'}) is used.
4088 Usually they indicate a typo in the user's code, as they have
4089 implementation-defined values, and should not be used in portable code.
4091 @item -Wnormalized=<none|id|nfc|nfkc>
4092 @opindex Wnormalized=
4095 @cindex character set, input normalization
4096 In ISO C and ISO C++, two identifiers are different if they are
4097 different sequences of characters. However, sometimes when characters
4098 outside the basic ASCII character set are used, you can have two
4099 different character sequences that look the same. To avoid confusion,
4100 the ISO 10646 standard sets out some @dfn{normalization rules} which
4101 when applied ensure that two sequences that look the same are turned into
4102 the same sequence. GCC can warn you if you are using identifiers which
4103 have not been normalized; this option controls that warning.
4105 There are four levels of warning that GCC supports. The default is
4106 @option{-Wnormalized=nfc}, which warns about any identifier which is
4107 not in the ISO 10646 ``C'' normalized form, @dfn{NFC}. NFC is the
4108 recommended form for most uses.
4110 Unfortunately, there are some characters which ISO C and ISO C++ allow
4111 in identifiers that when turned into NFC aren't allowable as
4112 identifiers. That is, there's no way to use these symbols in portable
4113 ISO C or C++ and have all your identifiers in NFC@.
4114 @option{-Wnormalized=id} suppresses the warning for these characters.
4115 It is hoped that future versions of the standards involved will correct
4116 this, which is why this option is not the default.
4118 You can switch the warning off for all characters by writing
4119 @option{-Wnormalized=none}. You would only want to do this if you
4120 were using some other normalization scheme (like ``D''), because
4121 otherwise you can easily create bugs that are literally impossible to see.
4123 Some characters in ISO 10646 have distinct meanings but look identical
4124 in some fonts or display methodologies, especially once formatting has
4125 been applied. For instance @code{\u207F}, ``SUPERSCRIPT LATIN SMALL
4126 LETTER N'', will display just like a regular @code{n} which has been
4127 placed in a superscript. ISO 10646 defines the @dfn{NFKC}
4128 normalization scheme to convert all these into a standard form as
4129 well, and GCC will warn if your code is not in NFKC if you use
4130 @option{-Wnormalized=nfkc}. This warning is comparable to warning
4131 about every identifier that contains the letter O because it might be
4132 confused with the digit 0, and so is not the default, but may be
4133 useful as a local coding convention if the programming environment is
4134 unable to be fixed to display these characters distinctly.
4136 @item -Wno-deprecated
4137 @opindex Wno-deprecated
4138 @opindex Wdeprecated
4139 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
4141 @item -Wno-deprecated-declarations
4142 @opindex Wno-deprecated-declarations
4143 @opindex Wdeprecated-declarations
4144 Do not warn about uses of functions (@pxref{Function Attributes}),
4145 variables (@pxref{Variable Attributes}), and types (@pxref{Type
4146 Attributes}) marked as deprecated by using the @code{deprecated}
4150 @opindex Wno-overflow
4152 Do not warn about compile-time overflow in constant expressions.
4154 @item -Woverride-init @r{(C and Objective-C only)}
4155 @opindex Woverride-init
4156 @opindex Wno-override-init
4160 Warn if an initialized field without side effects is overridden when
4161 using designated initializers (@pxref{Designated Inits, , Designated
4164 This warning is included in @option{-Wextra}. To get other
4165 @option{-Wextra} warnings without this one, use @samp{-Wextra
4166 -Wno-override-init}.
4171 Warn if a structure is given the packed attribute, but the packed
4172 attribute has no effect on the layout or size of the structure.
4173 Such structures may be mis-aligned for little benefit. For
4174 instance, in this code, the variable @code{f.x} in @code{struct bar}
4175 will be misaligned even though @code{struct bar} does not itself
4176 have the packed attribute:
4183 @} __attribute__((packed));
4191 @item -Wpacked-bitfield-compat
4192 @opindex Wpacked-bitfield-compat
4193 @opindex Wno-packed-bitfield-compat
4194 The 4.1, 4.2 and 4.3 series of GCC ignore the @code{packed} attribute
4195 on bit-fields of type @code{char}. This has been fixed in GCC 4.4 but
4196 the change can lead to differences in the structure layout. GCC
4197 informs you when the offset of such a field has changed in GCC 4.4.
4198 For example there is no longer a 4-bit padding between field @code{a}
4199 and @code{b} in this structure:
4206 @} __attribute__ ((packed));
4209 This warning is enabled by default. Use
4210 @option{-Wno-packed-bitfield-compat} to disable this warning.
4215 Warn if padding is included in a structure, either to align an element
4216 of the structure or to align the whole structure. Sometimes when this
4217 happens it is possible to rearrange the fields of the structure to
4218 reduce the padding and so make the structure smaller.
4220 @item -Wredundant-decls
4221 @opindex Wredundant-decls
4222 @opindex Wno-redundant-decls
4223 Warn if anything is declared more than once in the same scope, even in
4224 cases where multiple declaration is valid and changes nothing.
4226 @item -Wnested-externs @r{(C and Objective-C only)}
4227 @opindex Wnested-externs
4228 @opindex Wno-nested-externs
4229 Warn if an @code{extern} declaration is encountered within a function.
4234 Warn if a function can not be inlined and it was declared as inline.
4235 Even with this option, the compiler will not warn about failures to
4236 inline functions declared in system headers.
4238 The compiler uses a variety of heuristics to determine whether or not
4239 to inline a function. For example, the compiler takes into account
4240 the size of the function being inlined and the amount of inlining
4241 that has already been done in the current function. Therefore,
4242 seemingly insignificant changes in the source program can cause the
4243 warnings produced by @option{-Winline} to appear or disappear.
4245 @item -Wno-invalid-offsetof @r{(C++ and Objective-C++ only)}
4246 @opindex Wno-invalid-offsetof
4247 @opindex Winvalid-offsetof
4248 Suppress warnings from applying the @samp{offsetof} macro to a non-POD
4249 type. According to the 1998 ISO C++ standard, applying @samp{offsetof}
4250 to a non-POD type is undefined. In existing C++ implementations,
4251 however, @samp{offsetof} typically gives meaningful results even when
4252 applied to certain kinds of non-POD types. (Such as a simple
4253 @samp{struct} that fails to be a POD type only by virtue of having a
4254 constructor.) This flag is for users who are aware that they are
4255 writing nonportable code and who have deliberately chosen to ignore the
4258 The restrictions on @samp{offsetof} may be relaxed in a future version
4259 of the C++ standard.
4261 @item -Wno-int-to-pointer-cast @r{(C and Objective-C only)}
4262 @opindex Wno-int-to-pointer-cast
4263 @opindex Wint-to-pointer-cast
4264 Suppress warnings from casts to pointer type of an integer of a
4267 @item -Wno-pointer-to-int-cast @r{(C and Objective-C only)}
4268 @opindex Wno-pointer-to-int-cast
4269 @opindex Wpointer-to-int-cast
4270 Suppress warnings from casts from a pointer to an integer type of a
4274 @opindex Winvalid-pch
4275 @opindex Wno-invalid-pch
4276 Warn if a precompiled header (@pxref{Precompiled Headers}) is found in
4277 the search path but can't be used.
4281 @opindex Wno-long-long
4282 Warn if @samp{long long} type is used. This is enabled by either
4283 @option{-pedantic} or @option{-Wtraditional} in ISO C90 and C++98
4284 modes. To inhibit the warning messages, use @option{-Wno-long-long}.
4286 @item -Wvariadic-macros
4287 @opindex Wvariadic-macros
4288 @opindex Wno-variadic-macros
4289 Warn if variadic macros are used in pedantic ISO C90 mode, or the GNU
4290 alternate syntax when in pedantic ISO C99 mode. This is default.
4291 To inhibit the warning messages, use @option{-Wno-variadic-macros}.
4296 Warn if variable length array is used in the code.
4297 @option{-Wno-vla} will prevent the @option{-pedantic} warning of
4298 the variable length array.
4300 @item -Wvolatile-register-var
4301 @opindex Wvolatile-register-var
4302 @opindex Wno-volatile-register-var
4303 Warn if a register variable is declared volatile. The volatile
4304 modifier does not inhibit all optimizations that may eliminate reads
4305 and/or writes to register variables. This warning is enabled by
4308 @item -Wdisabled-optimization
4309 @opindex Wdisabled-optimization
4310 @opindex Wno-disabled-optimization
4311 Warn if a requested optimization pass is disabled. This warning does
4312 not generally indicate that there is anything wrong with your code; it
4313 merely indicates that GCC's optimizers were unable to handle the code
4314 effectively. Often, the problem is that your code is too big or too
4315 complex; GCC will refuse to optimize programs when the optimization
4316 itself is likely to take inordinate amounts of time.
4318 @item -Wpointer-sign @r{(C and Objective-C only)}
4319 @opindex Wpointer-sign
4320 @opindex Wno-pointer-sign
4321 Warn for pointer argument passing or assignment with different signedness.
4322 This option is only supported for C and Objective-C@. It is implied by
4323 @option{-Wall} and by @option{-pedantic}, which can be disabled with
4324 @option{-Wno-pointer-sign}.
4326 @item -Wstack-protector
4327 @opindex Wstack-protector
4328 @opindex Wno-stack-protector
4329 This option is only active when @option{-fstack-protector} is active. It
4330 warns about functions that will not be protected against stack smashing.
4333 @opindex Wno-mudflap
4334 Suppress warnings about constructs that cannot be instrumented by
4337 @item -Woverlength-strings
4338 @opindex Woverlength-strings
4339 @opindex Wno-overlength-strings
4340 Warn about string constants which are longer than the ``minimum
4341 maximum'' length specified in the C standard. Modern compilers
4342 generally allow string constants which are much longer than the
4343 standard's minimum limit, but very portable programs should avoid
4344 using longer strings.
4346 The limit applies @emph{after} string constant concatenation, and does
4347 not count the trailing NUL@. In C90, the limit was 509 characters; in
4348 C99, it was raised to 4095. C++98 does not specify a normative
4349 minimum maximum, so we do not diagnose overlength strings in C++@.
4351 This option is implied by @option{-pedantic}, and can be disabled with
4352 @option{-Wno-overlength-strings}.
4354 @item -Wunsuffixed-float-constants @r{(C and Objective-C only)}
4355 @opindex Wunsuffixed-float-constants
4357 GCC will issue a warning for any floating constant that does not have
4358 a suffix. When used together with @option{-Wsystem-headers} it will
4359 warn about such constants in system header files. This can be useful
4360 when preparing code to use with the @code{FLOAT_CONST_DECIMAL64} pragma
4361 from the decimal floating-point extension to C99.
4364 @node Debugging Options
4365 @section Options for Debugging Your Program or GCC
4366 @cindex options, debugging
4367 @cindex debugging information options
4369 GCC has various special options that are used for debugging
4370 either your program or GCC:
4375 Produce debugging information in the operating system's native format
4376 (stabs, COFF, XCOFF, or DWARF 2)@. GDB can work with this debugging
4379 On most systems that use stabs format, @option{-g} enables use of extra
4380 debugging information that only GDB can use; this extra information
4381 makes debugging work better in GDB but will probably make other debuggers
4383 refuse to read the program. If you want to control for certain whether
4384 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
4385 @option{-gxcoff+}, @option{-gxcoff}, or @option{-gvms} (see below).
4387 GCC allows you to use @option{-g} with
4388 @option{-O}. The shortcuts taken by optimized code may occasionally
4389 produce surprising results: some variables you declared may not exist
4390 at all; flow of control may briefly move where you did not expect it;
4391 some statements may not be executed because they compute constant
4392 results or their values were already at hand; some statements may
4393 execute in different places because they were moved out of loops.
4395 Nevertheless it proves possible to debug optimized output. This makes
4396 it reasonable to use the optimizer for programs that might have bugs.
4398 The following options are useful when GCC is generated with the
4399 capability for more than one debugging format.
4403 Produce debugging information for use by GDB@. This means to use the
4404 most expressive format available (DWARF 2, stabs, or the native format
4405 if neither of those are supported), including GDB extensions if at all
4410 Produce debugging information in stabs format (if that is supported),
4411 without GDB extensions. This is the format used by DBX on most BSD
4412 systems. On MIPS, Alpha and System V Release 4 systems this option
4413 produces stabs debugging output which is not understood by DBX or SDB@.
4414 On System V Release 4 systems this option requires the GNU assembler.
4416 @item -feliminate-unused-debug-symbols
4417 @opindex feliminate-unused-debug-symbols
4418 Produce debugging information in stabs format (if that is supported),
4419 for only symbols that are actually used.
4421 @item -femit-class-debug-always
4422 Instead of emitting debugging information for a C++ class in only one
4423 object file, emit it in all object files using the class. This option
4424 should be used only with debuggers that are unable to handle the way GCC
4425 normally emits debugging information for classes because using this
4426 option will increase the size of debugging information by as much as a
4431 Produce debugging information in stabs format (if that is supported),
4432 using GNU extensions understood only by the GNU debugger (GDB)@. The
4433 use of these extensions is likely to make other debuggers crash or
4434 refuse to read the program.
4438 Produce debugging information in COFF format (if that is supported).
4439 This is the format used by SDB on most System V systems prior to
4444 Produce debugging information in XCOFF format (if that is supported).
4445 This is the format used by the DBX debugger on IBM RS/6000 systems.
4449 Produce debugging information in XCOFF format (if that is supported),
4450 using GNU extensions understood only by the GNU debugger (GDB)@. The
4451 use of these extensions is likely to make other debuggers crash or
4452 refuse to read the program, and may cause assemblers other than the GNU
4453 assembler (GAS) to fail with an error.
4455 @item -gdwarf-@var{version}
4456 @opindex gdwarf-@var{version}
4457 Produce debugging information in DWARF format (if that is
4458 supported). This is the format used by DBX on IRIX 6. The value
4459 of @var{version} may be either 2, 3 or 4; the default version is 2.
4461 Note that with DWARF version 2 some ports require, and will always
4462 use, some non-conflicting DWARF 3 extensions in the unwind tables.
4464 Version 4 may require GDB 7.0 and @option{-fvar-tracking-assignments}
4465 for maximum benefit.
4467 @item -gstrict-dwarf
4468 @opindex gstrict-dwarf
4469 Disallow using extensions of later DWARF standard version than selected
4470 with @option{-gdwarf-@var{version}}. On most targets using non-conflicting
4471 DWARF extensions from later standard versions is allowed.
4473 @item -gno-strict-dwarf
4474 @opindex gno-strict-dwarf
4475 Allow using extensions of later DWARF standard version than selected with
4476 @option{-gdwarf-@var{version}}.
4480 Produce debugging information in VMS debug format (if that is
4481 supported). This is the format used by DEBUG on VMS systems.
4484 @itemx -ggdb@var{level}
4485 @itemx -gstabs@var{level}
4486 @itemx -gcoff@var{level}
4487 @itemx -gxcoff@var{level}
4488 @itemx -gvms@var{level}
4489 Request debugging information and also use @var{level} to specify how
4490 much information. The default level is 2.
4492 Level 0 produces no debug information at all. Thus, @option{-g0} negates
4495 Level 1 produces minimal information, enough for making backtraces in
4496 parts of the program that you don't plan to debug. This includes
4497 descriptions of functions and external variables, but no information
4498 about local variables and no line numbers.
4500 Level 3 includes extra information, such as all the macro definitions
4501 present in the program. Some debuggers support macro expansion when
4502 you use @option{-g3}.
4504 @option{-gdwarf-2} does not accept a concatenated debug level, because
4505 GCC used to support an option @option{-gdwarf} that meant to generate
4506 debug information in version 1 of the DWARF format (which is very
4507 different from version 2), and it would have been too confusing. That
4508 debug format is long obsolete, but the option cannot be changed now.
4509 Instead use an additional @option{-g@var{level}} option to change the
4510 debug level for DWARF.
4514 Turn off generation of debug info, if leaving out this option would have
4515 generated it, or turn it on at level 2 otherwise. The position of this
4516 argument in the command line does not matter, it takes effect after all
4517 other options are processed, and it does so only once, no matter how
4518 many times it is given. This is mainly intended to be used with
4519 @option{-fcompare-debug}.
4521 @item -fdump-final-insns@r{[}=@var{file}@r{]}
4522 @opindex fdump-final-insns
4523 Dump the final internal representation (RTL) to @var{file}. If the
4524 optional argument is omitted (or if @var{file} is @code{.}), the name
4525 of the dump file will be determined by appending @code{.gkd} to the
4526 compilation output file name.
4528 @item -fcompare-debug@r{[}=@var{opts}@r{]}
4529 @opindex fcompare-debug
4530 @opindex fno-compare-debug
4531 If no error occurs during compilation, run the compiler a second time,
4532 adding @var{opts} and @option{-fcompare-debug-second} to the arguments
4533 passed to the second compilation. Dump the final internal
4534 representation in both compilations, and print an error if they differ.
4536 If the equal sign is omitted, the default @option{-gtoggle} is used.
4538 The environment variable @env{GCC_COMPARE_DEBUG}, if defined, non-empty
4539 and nonzero, implicitly enables @option{-fcompare-debug}. If
4540 @env{GCC_COMPARE_DEBUG} is defined to a string starting with a dash,
4541 then it is used for @var{opts}, otherwise the default @option{-gtoggle}
4544 @option{-fcompare-debug=}, with the equal sign but without @var{opts},
4545 is equivalent to @option{-fno-compare-debug}, which disables the dumping
4546 of the final representation and the second compilation, preventing even
4547 @env{GCC_COMPARE_DEBUG} from taking effect.
4549 To verify full coverage during @option{-fcompare-debug} testing, set
4550 @env{GCC_COMPARE_DEBUG} to say @samp{-fcompare-debug-not-overridden},
4551 which GCC will reject as an invalid option in any actual compilation
4552 (rather than preprocessing, assembly or linking). To get just a
4553 warning, setting @env{GCC_COMPARE_DEBUG} to @samp{-w%n-fcompare-debug
4554 not overridden} will do.
4556 @item -fcompare-debug-second
4557 @opindex fcompare-debug-second
4558 This option is implicitly passed to the compiler for the second
4559 compilation requested by @option{-fcompare-debug}, along with options to
4560 silence warnings, and omitting other options that would cause
4561 side-effect compiler outputs to files or to the standard output. Dump
4562 files and preserved temporary files are renamed so as to contain the
4563 @code{.gk} additional extension during the second compilation, to avoid
4564 overwriting those generated by the first.
4566 When this option is passed to the compiler driver, it causes the
4567 @emph{first} compilation to be skipped, which makes it useful for little
4568 other than debugging the compiler proper.
4570 @item -feliminate-dwarf2-dups
4571 @opindex feliminate-dwarf2-dups
4572 Compress DWARF2 debugging information by eliminating duplicated
4573 information about each symbol. This option only makes sense when
4574 generating DWARF2 debugging information with @option{-gdwarf-2}.
4576 @item -femit-struct-debug-baseonly
4577 Emit debug information for struct-like types
4578 only when the base name of the compilation source file
4579 matches the base name of file in which the struct was defined.
4581 This option substantially reduces the size of debugging information,
4582 but at significant potential loss in type information to the debugger.
4583 See @option{-femit-struct-debug-reduced} for a less aggressive option.
4584 See @option{-femit-struct-debug-detailed} for more detailed control.
4586 This option works only with DWARF 2.
4588 @item -femit-struct-debug-reduced
4589 Emit debug information for struct-like types
4590 only when the base name of the compilation source file
4591 matches the base name of file in which the type was defined,
4592 unless the struct is a template or defined in a system header.
4594 This option significantly reduces the size of debugging information,
4595 with some potential loss in type information to the debugger.
4596 See @option{-femit-struct-debug-baseonly} for a more aggressive option.
4597 See @option{-femit-struct-debug-detailed} for more detailed control.
4599 This option works only with DWARF 2.
4601 @item -femit-struct-debug-detailed@r{[}=@var{spec-list}@r{]}
4602 Specify the struct-like types
4603 for which the compiler will generate debug information.
4604 The intent is to reduce duplicate struct debug information
4605 between different object files within the same program.
4607 This option is a detailed version of
4608 @option{-femit-struct-debug-reduced} and @option{-femit-struct-debug-baseonly},
4609 which will serve for most needs.
4611 A specification has the syntax
4612 [@samp{dir:}|@samp{ind:}][@samp{ord:}|@samp{gen:}](@samp{any}|@samp{sys}|@samp{base}|@samp{none})
4614 The optional first word limits the specification to
4615 structs that are used directly (@samp{dir:}) or used indirectly (@samp{ind:}).
4616 A struct type is used directly when it is the type of a variable, member.
4617 Indirect uses arise through pointers to structs.
4618 That is, when use of an incomplete struct would be legal, the use is indirect.
4620 @samp{struct one direct; struct two * indirect;}.
4622 The optional second word limits the specification to
4623 ordinary structs (@samp{ord:}) or generic structs (@samp{gen:}).
4624 Generic structs are a bit complicated to explain.
4625 For C++, these are non-explicit specializations of template classes,
4626 or non-template classes within the above.
4627 Other programming languages have generics,
4628 but @samp{-femit-struct-debug-detailed} does not yet implement them.
4630 The third word specifies the source files for those
4631 structs for which the compiler will emit debug information.
4632 The values @samp{none} and @samp{any} have the normal meaning.
4633 The value @samp{base} means that
4634 the base of name of the file in which the type declaration appears
4635 must match the base of the name of the main compilation file.
4636 In practice, this means that
4637 types declared in @file{foo.c} and @file{foo.h} will have debug information,
4638 but types declared in other header will not.
4639 The value @samp{sys} means those types satisfying @samp{base}
4640 or declared in system or compiler headers.
4642 You may need to experiment to determine the best settings for your application.
4644 The default is @samp{-femit-struct-debug-detailed=all}.
4646 This option works only with DWARF 2.
4648 @item -fenable-icf-debug
4649 @opindex fenable-icf-debug
4650 Generate additional debug information to support identical code folding (ICF).
4651 This option only works with DWARF version 2 or higher.
4653 @item -fno-merge-debug-strings
4654 @opindex fmerge-debug-strings
4655 @opindex fno-merge-debug-strings
4656 Direct the linker to not merge together strings in the debugging
4657 information which are identical in different object files. Merging is
4658 not supported by all assemblers or linkers. Merging decreases the size
4659 of the debug information in the output file at the cost of increasing
4660 link processing time. Merging is enabled by default.
4662 @item -fdebug-prefix-map=@var{old}=@var{new}
4663 @opindex fdebug-prefix-map
4664 When compiling files in directory @file{@var{old}}, record debugging
4665 information describing them as in @file{@var{new}} instead.
4667 @item -fno-dwarf2-cfi-asm
4668 @opindex fdwarf2-cfi-asm
4669 @opindex fno-dwarf2-cfi-asm
4670 Emit DWARF 2 unwind info as compiler generated @code{.eh_frame} section
4671 instead of using GAS @code{.cfi_*} directives.
4673 @cindex @command{prof}
4676 Generate extra code to write profile information suitable for the
4677 analysis program @command{prof}. You must use this option when compiling
4678 the source files you want data about, and you must also use it when
4681 @cindex @command{gprof}
4684 Generate extra code to write profile information suitable for the
4685 analysis program @command{gprof}. You must use this option when compiling
4686 the source files you want data about, and you must also use it when
4691 Makes the compiler print out each function name as it is compiled, and
4692 print some statistics about each pass when it finishes.
4695 @opindex ftime-report
4696 Makes the compiler print some statistics about the time consumed by each
4697 pass when it finishes.
4700 @opindex fmem-report
4701 Makes the compiler print some statistics about permanent memory
4702 allocation when it finishes.
4704 @item -fpre-ipa-mem-report
4705 @opindex fpre-ipa-mem-report
4706 @item -fpost-ipa-mem-report
4707 @opindex fpost-ipa-mem-report
4708 Makes the compiler print some statistics about permanent memory
4709 allocation before or after interprocedural optimization.
4711 @item -fprofile-arcs
4712 @opindex fprofile-arcs
4713 Add code so that program flow @dfn{arcs} are instrumented. During
4714 execution the program records how many times each branch and call is
4715 executed and how many times it is taken or returns. When the compiled
4716 program exits it saves this data to a file called
4717 @file{@var{auxname}.gcda} for each source file. The data may be used for
4718 profile-directed optimizations (@option{-fbranch-probabilities}), or for
4719 test coverage analysis (@option{-ftest-coverage}). Each object file's
4720 @var{auxname} is generated from the name of the output file, if
4721 explicitly specified and it is not the final executable, otherwise it is
4722 the basename of the source file. In both cases any suffix is removed
4723 (e.g.@: @file{foo.gcda} for input file @file{dir/foo.c}, or
4724 @file{dir/foo.gcda} for output file specified as @option{-o dir/foo.o}).
4725 @xref{Cross-profiling}.
4727 @cindex @command{gcov}
4731 This option is used to compile and link code instrumented for coverage
4732 analysis. The option is a synonym for @option{-fprofile-arcs}
4733 @option{-ftest-coverage} (when compiling) and @option{-lgcov} (when
4734 linking). See the documentation for those options for more details.
4739 Compile the source files with @option{-fprofile-arcs} plus optimization
4740 and code generation options. For test coverage analysis, use the
4741 additional @option{-ftest-coverage} option. You do not need to profile
4742 every source file in a program.
4745 Link your object files with @option{-lgcov} or @option{-fprofile-arcs}
4746 (the latter implies the former).
4749 Run the program on a representative workload to generate the arc profile
4750 information. This may be repeated any number of times. You can run
4751 concurrent instances of your program, and provided that the file system
4752 supports locking, the data files will be correctly updated. Also
4753 @code{fork} calls are detected and correctly handled (double counting
4757 For profile-directed optimizations, compile the source files again with
4758 the same optimization and code generation options plus
4759 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
4760 Control Optimization}).
4763 For test coverage analysis, use @command{gcov} to produce human readable
4764 information from the @file{.gcno} and @file{.gcda} files. Refer to the
4765 @command{gcov} documentation for further information.
4769 With @option{-fprofile-arcs}, for each function of your program GCC
4770 creates a program flow graph, then finds a spanning tree for the graph.
4771 Only arcs that are not on the spanning tree have to be instrumented: the
4772 compiler adds code to count the number of times that these arcs are
4773 executed. When an arc is the only exit or only entrance to a block, the
4774 instrumentation code can be added to the block; otherwise, a new basic
4775 block must be created to hold the instrumentation code.
4778 @item -ftest-coverage
4779 @opindex ftest-coverage
4780 Produce a notes file that the @command{gcov} code-coverage utility
4781 (@pxref{Gcov,, @command{gcov}---a Test Coverage Program}) can use to
4782 show program coverage. Each source file's note file is called
4783 @file{@var{auxname}.gcno}. Refer to the @option{-fprofile-arcs} option
4784 above for a description of @var{auxname} and instructions on how to
4785 generate test coverage data. Coverage data will match the source files
4786 more closely, if you do not optimize.
4788 @item -fdbg-cnt-list
4789 @opindex fdbg-cnt-list
4790 Print the name and the counter upperbound for all debug counters.
4792 @item -fdbg-cnt=@var{counter-value-list}
4794 Set the internal debug counter upperbound. @var{counter-value-list}
4795 is a comma-separated list of @var{name}:@var{value} pairs
4796 which sets the upperbound of each debug counter @var{name} to @var{value}.
4797 All debug counters have the initial upperbound of @var{UINT_MAX},
4798 thus dbg_cnt() returns true always unless the upperbound is set by this option.
4799 e.g. With -fdbg-cnt=dce:10,tail_call:0
4800 dbg_cnt(dce) will return true only for first 10 invocations
4801 and dbg_cnt(tail_call) will return false always.
4803 @item -d@var{letters}
4804 @itemx -fdump-rtl-@var{pass}
4806 Says to make debugging dumps during compilation at times specified by
4807 @var{letters}. This is used for debugging the RTL-based passes of the
4808 compiler. The file names for most of the dumps are made by appending
4809 a pass number and a word to the @var{dumpname}, and the files are
4810 created in the directory of the output file. @var{dumpname} is
4811 generated from the name of the output file, if explicitly specified
4812 and it is not an executable, otherwise it is the basename of the
4813 source file. These switches may have different effects when
4814 @option{-E} is used for preprocessing.
4816 Debug dumps can be enabled with a @option{-fdump-rtl} switch or some
4817 @option{-d} option @var{letters}. Here are the possible
4818 letters for use in @var{pass} and @var{letters}, and their meanings:
4822 @item -fdump-rtl-alignments
4823 @opindex fdump-rtl-alignments
4824 Dump after branch alignments have been computed.
4826 @item -fdump-rtl-asmcons
4827 @opindex fdump-rtl-asmcons
4828 Dump after fixing rtl statements that have unsatisfied in/out constraints.
4830 @item -fdump-rtl-auto_inc_dec
4831 @opindex fdump-rtl-auto_inc_dec
4832 Dump after auto-inc-dec discovery. This pass is only run on
4833 architectures that have auto inc or auto dec instructions.
4835 @item -fdump-rtl-barriers
4836 @opindex fdump-rtl-barriers
4837 Dump after cleaning up the barrier instructions.
4839 @item -fdump-rtl-bbpart
4840 @opindex fdump-rtl-bbpart
4841 Dump after partitioning hot and cold basic blocks.
4843 @item -fdump-rtl-bbro
4844 @opindex fdump-rtl-bbro
4845 Dump after block reordering.
4847 @item -fdump-rtl-btl1
4848 @itemx -fdump-rtl-btl2
4849 @opindex fdump-rtl-btl2
4850 @opindex fdump-rtl-btl2
4851 @option{-fdump-rtl-btl1} and @option{-fdump-rtl-btl2} enable dumping
4852 after the two branch
4853 target load optimization passes.
4855 @item -fdump-rtl-bypass
4856 @opindex fdump-rtl-bypass
4857 Dump after jump bypassing and control flow optimizations.
4859 @item -fdump-rtl-combine
4860 @opindex fdump-rtl-combine
4861 Dump after the RTL instruction combination pass.
4863 @item -fdump-rtl-compgotos
4864 @opindex fdump-rtl-compgotos
4865 Dump after duplicating the computed gotos.
4867 @item -fdump-rtl-ce1
4868 @itemx -fdump-rtl-ce2
4869 @itemx -fdump-rtl-ce3
4870 @opindex fdump-rtl-ce1
4871 @opindex fdump-rtl-ce2
4872 @opindex fdump-rtl-ce3
4873 @option{-fdump-rtl-ce1}, @option{-fdump-rtl-ce2}, and
4874 @option{-fdump-rtl-ce3} enable dumping after the three
4875 if conversion passes.
4877 @itemx -fdump-rtl-cprop_hardreg
4878 @opindex fdump-rtl-cprop_hardreg
4879 Dump after hard register copy propagation.
4881 @itemx -fdump-rtl-csa
4882 @opindex fdump-rtl-csa
4883 Dump after combining stack adjustments.
4885 @item -fdump-rtl-cse1
4886 @itemx -fdump-rtl-cse2
4887 @opindex fdump-rtl-cse1
4888 @opindex fdump-rtl-cse2
4889 @option{-fdump-rtl-cse1} and @option{-fdump-rtl-cse2} enable dumping after
4890 the two common sub-expression elimination passes.
4892 @itemx -fdump-rtl-dce
4893 @opindex fdump-rtl-dce
4894 Dump after the standalone dead code elimination passes.
4896 @itemx -fdump-rtl-dbr
4897 @opindex fdump-rtl-dbr
4898 Dump after delayed branch scheduling.
4900 @item -fdump-rtl-dce1
4901 @itemx -fdump-rtl-dce2
4902 @opindex fdump-rtl-dce1
4903 @opindex fdump-rtl-dce2
4904 @option{-fdump-rtl-dce1} and @option{-fdump-rtl-dce2} enable dumping after
4905 the two dead store elimination passes.
4908 @opindex fdump-rtl-eh
4909 Dump after finalization of EH handling code.
4911 @item -fdump-rtl-eh_ranges
4912 @opindex fdump-rtl-eh_ranges
4913 Dump after conversion of EH handling range regions.
4915 @item -fdump-rtl-expand
4916 @opindex fdump-rtl-expand
4917 Dump after RTL generation.
4919 @item -fdump-rtl-fwprop1
4920 @itemx -fdump-rtl-fwprop2
4921 @opindex fdump-rtl-fwprop1
4922 @opindex fdump-rtl-fwprop2
4923 @option{-fdump-rtl-fwprop1} and @option{-fdump-rtl-fwprop2} enable
4924 dumping after the two forward propagation passes.
4926 @item -fdump-rtl-gcse1
4927 @itemx -fdump-rtl-gcse2
4928 @opindex fdump-rtl-gcse1
4929 @opindex fdump-rtl-gcse2
4930 @option{-fdump-rtl-gcse1} and @option{-fdump-rtl-gcse2} enable dumping
4931 after global common subexpression elimination.
4933 @item -fdump-rtl-init-regs
4934 @opindex fdump-rtl-init-regs
4935 Dump after the initialization of the registers.
4937 @item -fdump-rtl-initvals
4938 @opindex fdump-rtl-initvals
4939 Dump after the computation of the initial value sets.
4941 @itemx -fdump-rtl-into_cfglayout
4942 @opindex fdump-rtl-into_cfglayout
4943 Dump after converting to cfglayout mode.
4945 @item -fdump-rtl-ira
4946 @opindex fdump-rtl-ira
4947 Dump after iterated register allocation.
4949 @item -fdump-rtl-jump
4950 @opindex fdump-rtl-jump
4951 Dump after the second jump optimization.
4953 @item -fdump-rtl-loop2
4954 @opindex fdump-rtl-loop2
4955 @option{-fdump-rtl-loop2} enables dumping after the rtl
4956 loop optimization passes.
4958 @item -fdump-rtl-mach
4959 @opindex fdump-rtl-mach
4960 Dump after performing the machine dependent reorganization pass, if that
4963 @item -fdump-rtl-mode_sw
4964 @opindex fdump-rtl-mode_sw
4965 Dump after removing redundant mode switches.
4967 @item -fdump-rtl-rnreg
4968 @opindex fdump-rtl-rnreg
4969 Dump after register renumbering.
4971 @itemx -fdump-rtl-outof_cfglayout
4972 @opindex fdump-rtl-outof_cfglayout
4973 Dump after converting from cfglayout mode.
4975 @item -fdump-rtl-peephole2
4976 @opindex fdump-rtl-peephole2
4977 Dump after the peephole pass.
4979 @item -fdump-rtl-postreload
4980 @opindex fdump-rtl-postreload
4981 Dump after post-reload optimizations.
4983 @itemx -fdump-rtl-pro_and_epilogue
4984 @opindex fdump-rtl-pro_and_epilogue
4985 Dump after generating the function pro and epilogues.
4987 @item -fdump-rtl-regmove
4988 @opindex fdump-rtl-regmove
4989 Dump after the register move pass.
4991 @item -fdump-rtl-sched1
4992 @itemx -fdump-rtl-sched2
4993 @opindex fdump-rtl-sched1
4994 @opindex fdump-rtl-sched2
4995 @option{-fdump-rtl-sched1} and @option{-fdump-rtl-sched2} enable dumping
4996 after the basic block scheduling passes.
4998 @item -fdump-rtl-see
4999 @opindex fdump-rtl-see
5000 Dump after sign extension elimination.
5002 @item -fdump-rtl-seqabstr
5003 @opindex fdump-rtl-seqabstr
5004 Dump after common sequence discovery.
5006 @item -fdump-rtl-shorten
5007 @opindex fdump-rtl-shorten
5008 Dump after shortening branches.
5010 @item -fdump-rtl-sibling
5011 @opindex fdump-rtl-sibling
5012 Dump after sibling call optimizations.
5014 @item -fdump-rtl-split1
5015 @itemx -fdump-rtl-split2
5016 @itemx -fdump-rtl-split3
5017 @itemx -fdump-rtl-split4
5018 @itemx -fdump-rtl-split5
5019 @opindex fdump-rtl-split1
5020 @opindex fdump-rtl-split2
5021 @opindex fdump-rtl-split3
5022 @opindex fdump-rtl-split4
5023 @opindex fdump-rtl-split5
5024 @option{-fdump-rtl-split1}, @option{-fdump-rtl-split2},
5025 @option{-fdump-rtl-split3}, @option{-fdump-rtl-split4} and
5026 @option{-fdump-rtl-split5} enable dumping after five rounds of
5027 instruction splitting.
5029 @item -fdump-rtl-sms
5030 @opindex fdump-rtl-sms
5031 Dump after modulo scheduling. This pass is only run on some
5034 @item -fdump-rtl-stack
5035 @opindex fdump-rtl-stack
5036 Dump after conversion from GCC's "flat register file" registers to the
5037 x87's stack-like registers. This pass is only run on x86 variants.
5039 @item -fdump-rtl-subreg1
5040 @itemx -fdump-rtl-subreg2
5041 @opindex fdump-rtl-subreg1
5042 @opindex fdump-rtl-subreg2
5043 @option{-fdump-rtl-subreg1} and @option{-fdump-rtl-subreg2} enable dumping after
5044 the two subreg expansion passes.
5046 @item -fdump-rtl-unshare
5047 @opindex fdump-rtl-unshare
5048 Dump after all rtl has been unshared.
5050 @item -fdump-rtl-vartrack
5051 @opindex fdump-rtl-vartrack
5052 Dump after variable tracking.
5054 @item -fdump-rtl-vregs
5055 @opindex fdump-rtl-vregs
5056 Dump after converting virtual registers to hard registers.
5058 @item -fdump-rtl-web
5059 @opindex fdump-rtl-web
5060 Dump after live range splitting.
5062 @item -fdump-rtl-regclass
5063 @itemx -fdump-rtl-subregs_of_mode_init
5064 @itemx -fdump-rtl-subregs_of_mode_finish
5065 @itemx -fdump-rtl-dfinit
5066 @itemx -fdump-rtl-dfinish
5067 @opindex fdump-rtl-regclass
5068 @opindex fdump-rtl-subregs_of_mode_init
5069 @opindex fdump-rtl-subregs_of_mode_finish
5070 @opindex fdump-rtl-dfinit
5071 @opindex fdump-rtl-dfinish
5072 These dumps are defined but always produce empty files.
5074 @item -fdump-rtl-all
5075 @opindex fdump-rtl-all
5076 Produce all the dumps listed above.
5080 Annotate the assembler output with miscellaneous debugging information.
5084 Dump all macro definitions, at the end of preprocessing, in addition to
5089 Produce a core dump whenever an error occurs.
5093 Print statistics on memory usage, at the end of the run, to
5098 Annotate the assembler output with a comment indicating which
5099 pattern and alternative was used. The length of each instruction is
5104 Dump the RTL in the assembler output as a comment before each instruction.
5105 Also turns on @option{-dp} annotation.
5109 For each of the other indicated dump files (@option{-fdump-rtl-@var{pass}}),
5110 dump a representation of the control flow graph suitable for viewing with VCG
5111 to @file{@var{file}.@var{pass}.vcg}.
5115 Just generate RTL for a function instead of compiling it. Usually used
5116 with @option{-fdump-rtl-expand}.
5120 Dump debugging information during parsing, to standard error.
5124 @opindex fdump-noaddr
5125 When doing debugging dumps, suppress address output. This makes it more
5126 feasible to use diff on debugging dumps for compiler invocations with
5127 different compiler binaries and/or different
5128 text / bss / data / heap / stack / dso start locations.
5130 @item -fdump-unnumbered
5131 @opindex fdump-unnumbered
5132 When doing debugging dumps, suppress instruction numbers and address output.
5133 This makes it more feasible to use diff on debugging dumps for compiler
5134 invocations with different options, in particular with and without
5137 @item -fdump-unnumbered-links
5138 @opindex fdump-unnumbered-links
5139 When doing debugging dumps (see @option{-d} option above), suppress
5140 instruction numbers for the links to the previous and next instructions
5143 @item -fdump-translation-unit @r{(C++ only)}
5144 @itemx -fdump-translation-unit-@var{options} @r{(C++ only)}
5145 @opindex fdump-translation-unit
5146 Dump a representation of the tree structure for the entire translation
5147 unit to a file. The file name is made by appending @file{.tu} to the
5148 source file name, and the file is created in the same directory as the
5149 output file. If the @samp{-@var{options}} form is used, @var{options}
5150 controls the details of the dump as described for the
5151 @option{-fdump-tree} options.
5153 @item -fdump-class-hierarchy @r{(C++ only)}
5154 @itemx -fdump-class-hierarchy-@var{options} @r{(C++ only)}
5155 @opindex fdump-class-hierarchy
5156 Dump a representation of each class's hierarchy and virtual function
5157 table layout to a file. The file name is made by appending
5158 @file{.class} to the source file name, and the file is created in the
5159 same directory as the output file. If the @samp{-@var{options}} form
5160 is used, @var{options} controls the details of the dump as described
5161 for the @option{-fdump-tree} options.
5163 @item -fdump-ipa-@var{switch}
5165 Control the dumping at various stages of inter-procedural analysis
5166 language tree to a file. The file name is generated by appending a
5167 switch specific suffix to the source file name, and the file is created
5168 in the same directory as the output file. The following dumps are
5173 Enables all inter-procedural analysis dumps.
5176 Dumps information about call-graph optimization, unused function removal,
5177 and inlining decisions.
5180 Dump after function inlining.
5184 @item -fdump-statistics-@var{option}
5185 @opindex fdump-statistics
5186 Enable and control dumping of pass statistics in a separate file. The
5187 file name is generated by appending a suffix ending in
5188 @samp{.statistics} to the source file name, and the file is created in
5189 the same directory as the output file. If the @samp{-@var{option}}
5190 form is used, @samp{-stats} will cause counters to be summed over the
5191 whole compilation unit while @samp{-details} will dump every event as
5192 the passes generate them. The default with no option is to sum
5193 counters for each function compiled.
5195 @item -fdump-tree-@var{switch}
5196 @itemx -fdump-tree-@var{switch}-@var{options}
5198 Control the dumping at various stages of processing the intermediate
5199 language tree to a file. The file name is generated by appending a
5200 switch specific suffix to the source file name, and the file is
5201 created in the same directory as the output file. If the
5202 @samp{-@var{options}} form is used, @var{options} is a list of
5203 @samp{-} separated options that control the details of the dump. Not
5204 all options are applicable to all dumps, those which are not
5205 meaningful will be ignored. The following options are available
5209 Print the address of each node. Usually this is not meaningful as it
5210 changes according to the environment and source file. Its primary use
5211 is for tying up a dump file with a debug environment.
5213 If @code{DECL_ASSEMBLER_NAME} has been set for a given decl, use that
5214 in the dump instead of @code{DECL_NAME}. Its primary use is ease of
5215 use working backward from mangled names in the assembly file.
5217 Inhibit dumping of members of a scope or body of a function merely
5218 because that scope has been reached. Only dump such items when they
5219 are directly reachable by some other path. When dumping pretty-printed
5220 trees, this option inhibits dumping the bodies of control structures.
5222 Print a raw representation of the tree. By default, trees are
5223 pretty-printed into a C-like representation.
5225 Enable more detailed dumps (not honored by every dump option).
5227 Enable dumping various statistics about the pass (not honored by every dump
5230 Enable showing basic block boundaries (disabled in raw dumps).
5232 Enable showing virtual operands for every statement.
5234 Enable showing line numbers for statements.
5236 Enable showing the unique ID (@code{DECL_UID}) for each variable.
5238 Enable showing the tree dump for each statement.
5240 Enable showing the EH region number holding each statement.
5242 Turn on all options, except @option{raw}, @option{slim}, @option{verbose}
5243 and @option{lineno}.
5246 The following tree dumps are possible:
5250 @opindex fdump-tree-original
5251 Dump before any tree based optimization, to @file{@var{file}.original}.
5254 @opindex fdump-tree-optimized
5255 Dump after all tree based optimization, to @file{@var{file}.optimized}.
5258 @opindex fdump-tree-gimple
5259 Dump each function before and after the gimplification pass to a file. The
5260 file name is made by appending @file{.gimple} to the source file name.
5263 @opindex fdump-tree-cfg
5264 Dump the control flow graph of each function to a file. The file name is
5265 made by appending @file{.cfg} to the source file name.
5268 @opindex fdump-tree-vcg
5269 Dump the control flow graph of each function to a file in VCG format. The
5270 file name is made by appending @file{.vcg} to the source file name. Note
5271 that if the file contains more than one function, the generated file cannot
5272 be used directly by VCG@. You will need to cut and paste each function's
5273 graph into its own separate file first.
5276 @opindex fdump-tree-ch
5277 Dump each function after copying loop headers. The file name is made by
5278 appending @file{.ch} to the source file name.
5281 @opindex fdump-tree-ssa
5282 Dump SSA related information to a file. The file name is made by appending
5283 @file{.ssa} to the source file name.
5286 @opindex fdump-tree-alias
5287 Dump aliasing information for each function. The file name is made by
5288 appending @file{.alias} to the source file name.
5291 @opindex fdump-tree-ccp
5292 Dump each function after CCP@. The file name is made by appending
5293 @file{.ccp} to the source file name.
5296 @opindex fdump-tree-storeccp
5297 Dump each function after STORE-CCP@. The file name is made by appending
5298 @file{.storeccp} to the source file name.
5301 @opindex fdump-tree-pre
5302 Dump trees after partial redundancy elimination. The file name is made
5303 by appending @file{.pre} to the source file name.
5306 @opindex fdump-tree-fre
5307 Dump trees after full redundancy elimination. The file name is made
5308 by appending @file{.fre} to the source file name.
5311 @opindex fdump-tree-copyprop
5312 Dump trees after copy propagation. The file name is made
5313 by appending @file{.copyprop} to the source file name.
5315 @item store_copyprop
5316 @opindex fdump-tree-store_copyprop
5317 Dump trees after store copy-propagation. The file name is made
5318 by appending @file{.store_copyprop} to the source file name.
5321 @opindex fdump-tree-dce
5322 Dump each function after dead code elimination. The file name is made by
5323 appending @file{.dce} to the source file name.
5326 @opindex fdump-tree-mudflap
5327 Dump each function after adding mudflap instrumentation. The file name is
5328 made by appending @file{.mudflap} to the source file name.
5331 @opindex fdump-tree-sra
5332 Dump each function after performing scalar replacement of aggregates. The
5333 file name is made by appending @file{.sra} to the source file name.
5336 @opindex fdump-tree-sink
5337 Dump each function after performing code sinking. The file name is made
5338 by appending @file{.sink} to the source file name.
5341 @opindex fdump-tree-dom
5342 Dump each function after applying dominator tree optimizations. The file
5343 name is made by appending @file{.dom} to the source file name.
5346 @opindex fdump-tree-dse
5347 Dump each function after applying dead store elimination. The file
5348 name is made by appending @file{.dse} to the source file name.
5351 @opindex fdump-tree-phiopt
5352 Dump each function after optimizing PHI nodes into straightline code. The file
5353 name is made by appending @file{.phiopt} to the source file name.
5356 @opindex fdump-tree-forwprop
5357 Dump each function after forward propagating single use variables. The file
5358 name is made by appending @file{.forwprop} to the source file name.
5361 @opindex fdump-tree-copyrename
5362 Dump each function after applying the copy rename optimization. The file
5363 name is made by appending @file{.copyrename} to the source file name.
5366 @opindex fdump-tree-nrv
5367 Dump each function after applying the named return value optimization on
5368 generic trees. The file name is made by appending @file{.nrv} to the source
5372 @opindex fdump-tree-vect
5373 Dump each function after applying vectorization of loops. The file name is
5374 made by appending @file{.vect} to the source file name.
5377 @opindex fdump-tree-slp
5378 Dump each function after applying vectorization of basic blocks. The file name
5379 is made by appending @file{.slp} to the source file name.
5382 @opindex fdump-tree-vrp
5383 Dump each function after Value Range Propagation (VRP). The file name
5384 is made by appending @file{.vrp} to the source file name.
5387 @opindex fdump-tree-all
5388 Enable all the available tree dumps with the flags provided in this option.
5391 @item -ftree-vectorizer-verbose=@var{n}
5392 @opindex ftree-vectorizer-verbose
5393 This option controls the amount of debugging output the vectorizer prints.
5394 This information is written to standard error, unless
5395 @option{-fdump-tree-all} or @option{-fdump-tree-vect} is specified,
5396 in which case it is output to the usual dump listing file, @file{.vect}.
5397 For @var{n}=0 no diagnostic information is reported.
5398 If @var{n}=1 the vectorizer reports each loop that got vectorized,
5399 and the total number of loops that got vectorized.
5400 If @var{n}=2 the vectorizer also reports non-vectorized loops that passed
5401 the first analysis phase (vect_analyze_loop_form) - i.e.@: countable,
5402 inner-most, single-bb, single-entry/exit loops. This is the same verbosity
5403 level that @option{-fdump-tree-vect-stats} uses.
5404 Higher verbosity levels mean either more information dumped for each
5405 reported loop, or same amount of information reported for more loops:
5406 if @var{n}=3, vectorizer cost model information is reported.
5407 If @var{n}=4, alignment related information is added to the reports.
5408 If @var{n}=5, data-references related information (e.g.@: memory dependences,
5409 memory access-patterns) is added to the reports.
5410 If @var{n}=6, the vectorizer reports also non-vectorized inner-most loops
5411 that did not pass the first analysis phase (i.e., may not be countable, or
5412 may have complicated control-flow).
5413 If @var{n}=7, the vectorizer reports also non-vectorized nested loops.
5414 If @var{n}=8, SLP related information is added to the reports.
5415 For @var{n}=9, all the information the vectorizer generates during its
5416 analysis and transformation is reported. This is the same verbosity level
5417 that @option{-fdump-tree-vect-details} uses.
5419 @item -frandom-seed=@var{string}
5420 @opindex frandom-seed
5421 This option provides a seed that GCC uses when it would otherwise use
5422 random numbers. It is used to generate certain symbol names
5423 that have to be different in every compiled file. It is also used to
5424 place unique stamps in coverage data files and the object files that
5425 produce them. You can use the @option{-frandom-seed} option to produce
5426 reproducibly identical object files.
5428 The @var{string} should be different for every file you compile.
5430 @item -fsched-verbose=@var{n}
5431 @opindex fsched-verbose
5432 On targets that use instruction scheduling, this option controls the
5433 amount of debugging output the scheduler prints. This information is
5434 written to standard error, unless @option{-fdump-rtl-sched1} or
5435 @option{-fdump-rtl-sched2} is specified, in which case it is output
5436 to the usual dump listing file, @file{.sched1} or @file{.sched2}
5437 respectively. However for @var{n} greater than nine, the output is
5438 always printed to standard error.
5440 For @var{n} greater than zero, @option{-fsched-verbose} outputs the
5441 same information as @option{-fdump-rtl-sched1} and @option{-fdump-rtl-sched2}.
5442 For @var{n} greater than one, it also output basic block probabilities,
5443 detailed ready list information and unit/insn info. For @var{n} greater
5444 than two, it includes RTL at abort point, control-flow and regions info.
5445 And for @var{n} over four, @option{-fsched-verbose} also includes
5449 @itemx -save-temps=cwd
5451 Store the usual ``temporary'' intermediate files permanently; place them
5452 in the current directory and name them based on the source file. Thus,
5453 compiling @file{foo.c} with @samp{-c -save-temps} would produce files
5454 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
5455 preprocessed @file{foo.i} output file even though the compiler now
5456 normally uses an integrated preprocessor.
5458 When used in combination with the @option{-x} command line option,
5459 @option{-save-temps} is sensible enough to avoid over writing an
5460 input source file with the same extension as an intermediate file.
5461 The corresponding intermediate file may be obtained by renaming the
5462 source file before using @option{-save-temps}.
5464 If you invoke GCC in parallel, compiling several different source
5465 files that share a common base name in different subdirectories or the
5466 same source file compiled for multiple output destinations, it is
5467 likely that the different parallel compilers will interfere with each
5468 other, and overwrite the temporary files. For instance:
5471 gcc -save-temps -o outdir1/foo.o indir1/foo.c&
5472 gcc -save-temps -o outdir2/foo.o indir2/foo.c&
5475 may result in @file{foo.i} and @file{foo.o} being written to
5476 simultaneously by both compilers.
5478 @item -save-temps=obj
5479 @opindex save-temps=obj
5480 Store the usual ``temporary'' intermediate files permanently. If the
5481 @option{-o} option is used, the temporary files are based on the
5482 object file. If the @option{-o} option is not used, the
5483 @option{-save-temps=obj} switch behaves like @option{-save-temps}.
5488 gcc -save-temps=obj -c foo.c
5489 gcc -save-temps=obj -c bar.c -o dir/xbar.o
5490 gcc -save-temps=obj foobar.c -o dir2/yfoobar
5493 would create @file{foo.i}, @file{foo.s}, @file{dir/xbar.i},
5494 @file{dir/xbar.s}, @file{dir2/yfoobar.i}, @file{dir2/yfoobar.s}, and
5495 @file{dir2/yfoobar.o}.
5497 @item -time@r{[}=@var{file}@r{]}
5499 Report the CPU time taken by each subprocess in the compilation
5500 sequence. For C source files, this is the compiler proper and assembler
5501 (plus the linker if linking is done).
5503 Without the specification of an output file, the output looks like this:
5510 The first number on each line is the ``user time'', that is time spent
5511 executing the program itself. The second number is ``system time'',
5512 time spent executing operating system routines on behalf of the program.
5513 Both numbers are in seconds.
5515 With the specification of an output file, the output is appended to the
5516 named file, and it looks like this:
5519 0.12 0.01 cc1 @var{options}
5520 0.00 0.01 as @var{options}
5523 The ``user time'' and the ``system time'' are moved before the program
5524 name, and the options passed to the program are displayed, so that one
5525 can later tell what file was being compiled, and with which options.
5527 @item -fvar-tracking
5528 @opindex fvar-tracking
5529 Run variable tracking pass. It computes where variables are stored at each
5530 position in code. Better debugging information is then generated
5531 (if the debugging information format supports this information).
5533 It is enabled by default when compiling with optimization (@option{-Os},
5534 @option{-O}, @option{-O2}, @dots{}), debugging information (@option{-g}) and
5535 the debug info format supports it.
5537 @item -fvar-tracking-assignments
5538 @opindex fvar-tracking-assignments
5539 @opindex fno-var-tracking-assignments
5540 Annotate assignments to user variables early in the compilation and
5541 attempt to carry the annotations over throughout the compilation all the
5542 way to the end, in an attempt to improve debug information while
5543 optimizing. Use of @option{-gdwarf-4} is recommended along with it.
5545 It can be enabled even if var-tracking is disabled, in which case
5546 annotations will be created and maintained, but discarded at the end.
5548 @item -fvar-tracking-assignments-toggle
5549 @opindex fvar-tracking-assignments-toggle
5550 @opindex fno-var-tracking-assignments-toggle
5551 Toggle @option{-fvar-tracking-assignments}, in the same way that
5552 @option{-gtoggle} toggles @option{-g}.
5554 @item -print-file-name=@var{library}
5555 @opindex print-file-name
5556 Print the full absolute name of the library file @var{library} that
5557 would be used when linking---and don't do anything else. With this
5558 option, GCC does not compile or link anything; it just prints the
5561 @item -print-multi-directory
5562 @opindex print-multi-directory
5563 Print the directory name corresponding to the multilib selected by any
5564 other switches present in the command line. This directory is supposed
5565 to exist in @env{GCC_EXEC_PREFIX}.
5567 @item -print-multi-lib
5568 @opindex print-multi-lib
5569 Print the mapping from multilib directory names to compiler switches
5570 that enable them. The directory name is separated from the switches by
5571 @samp{;}, and each switch starts with an @samp{@@} instead of the
5572 @samp{-}, without spaces between multiple switches. This is supposed to
5573 ease shell-processing.
5575 @item -print-multi-os-directory
5576 @opindex print-multi-os-directory
5577 Print the path to OS libraries for the selected
5578 multilib, relative to some @file{lib} subdirectory. If OS libraries are
5579 present in the @file{lib} subdirectory and no multilibs are used, this is
5580 usually just @file{.}, if OS libraries are present in @file{lib@var{suffix}}
5581 sibling directories this prints e.g.@: @file{../lib64}, @file{../lib} or
5582 @file{../lib32}, or if OS libraries are present in @file{lib/@var{subdir}}
5583 subdirectories it prints e.g.@: @file{amd64}, @file{sparcv9} or @file{ev6}.
5585 @item -print-prog-name=@var{program}
5586 @opindex print-prog-name
5587 Like @option{-print-file-name}, but searches for a program such as @samp{cpp}.
5589 @item -print-libgcc-file-name
5590 @opindex print-libgcc-file-name
5591 Same as @option{-print-file-name=libgcc.a}.
5593 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
5594 but you do want to link with @file{libgcc.a}. You can do
5597 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
5600 @item -print-search-dirs
5601 @opindex print-search-dirs
5602 Print the name of the configured installation directory and a list of
5603 program and library directories @command{gcc} will search---and don't do anything else.
5605 This is useful when @command{gcc} prints the error message
5606 @samp{installation problem, cannot exec cpp0: No such file or directory}.
5607 To resolve this you either need to put @file{cpp0} and the other compiler
5608 components where @command{gcc} expects to find them, or you can set the environment
5609 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
5610 Don't forget the trailing @samp{/}.
5611 @xref{Environment Variables}.
5613 @item -print-sysroot
5614 @opindex print-sysroot
5615 Print the target sysroot directory that will be used during
5616 compilation. This is the target sysroot specified either at configure
5617 time or using the @option{--sysroot} option, possibly with an extra
5618 suffix that depends on compilation options. If no target sysroot is
5619 specified, the option prints nothing.
5621 @item -print-sysroot-headers-suffix
5622 @opindex print-sysroot-headers-suffix
5623 Print the suffix added to the target sysroot when searching for
5624 headers, or give an error if the compiler is not configured with such
5625 a suffix---and don't do anything else.
5628 @opindex dumpmachine
5629 Print the compiler's target machine (for example,
5630 @samp{i686-pc-linux-gnu})---and don't do anything else.
5633 @opindex dumpversion
5634 Print the compiler version (for example, @samp{3.0})---and don't do
5639 Print the compiler's built-in specs---and don't do anything else. (This
5640 is used when GCC itself is being built.) @xref{Spec Files}.
5642 @item -feliminate-unused-debug-types
5643 @opindex feliminate-unused-debug-types
5644 Normally, when producing DWARF2 output, GCC will emit debugging
5645 information for all types declared in a compilation
5646 unit, regardless of whether or not they are actually used
5647 in that compilation unit. Sometimes this is useful, such as
5648 if, in the debugger, you want to cast a value to a type that is
5649 not actually used in your program (but is declared). More often,
5650 however, this results in a significant amount of wasted space.
5651 With this option, GCC will avoid producing debug symbol output
5652 for types that are nowhere used in the source file being compiled.
5655 @node Optimize Options
5656 @section Options That Control Optimization
5657 @cindex optimize options
5658 @cindex options, optimization
5660 These options control various sorts of optimizations.
5662 Without any optimization option, the compiler's goal is to reduce the
5663 cost of compilation and to make debugging produce the expected
5664 results. Statements are independent: if you stop the program with a
5665 breakpoint between statements, you can then assign a new value to any
5666 variable or change the program counter to any other statement in the
5667 function and get exactly the results you would expect from the source
5670 Turning on optimization flags makes the compiler attempt to improve
5671 the performance and/or code size at the expense of compilation time
5672 and possibly the ability to debug the program.
5674 The compiler performs optimization based on the knowledge it has of the
5675 program. Compiling multiple files at once to a single output file mode allows
5676 the compiler to use information gained from all of the files when compiling
5679 Not all optimizations are controlled directly by a flag. Only
5680 optimizations that have a flag are listed in this section.
5682 Most optimizations are only enabled if an @option{-O} level is set on
5683 the command line. Otherwise they are disabled, even if individual
5684 optimization flags are specified.
5686 Depending on the target and how GCC was configured, a slightly different
5687 set of optimizations may be enabled at each @option{-O} level than
5688 those listed here. You can invoke GCC with @samp{-Q --help=optimizers}
5689 to find out the exact set of optimizations that are enabled at each level.
5690 @xref{Overall Options}, for examples.
5697 Optimize. Optimizing compilation takes somewhat more time, and a lot
5698 more memory for a large function.
5700 With @option{-O}, the compiler tries to reduce code size and execution
5701 time, without performing any optimizations that take a great deal of
5704 @option{-O} turns on the following optimization flags:
5707 -fcprop-registers @gol
5710 -fdelayed-branch @gol
5712 -fguess-branch-probability @gol
5713 -fif-conversion2 @gol
5714 -fif-conversion @gol
5715 -fipa-pure-const @gol
5716 -fipa-reference @gol
5718 -fsplit-wide-types @gol
5719 -ftree-builtin-call-dce @gol
5722 -ftree-copyrename @gol
5724 -ftree-dominator-opts @gol
5726 -ftree-forwprop @gol
5734 @option{-O} also turns on @option{-fomit-frame-pointer} on machines
5735 where doing so does not interfere with debugging.
5739 Optimize even more. GCC performs nearly all supported optimizations
5740 that do not involve a space-speed tradeoff.
5741 As compared to @option{-O}, this option increases both compilation time
5742 and the performance of the generated code.
5744 @option{-O2} turns on all optimization flags specified by @option{-O}. It
5745 also turns on the following optimization flags:
5746 @gccoptlist{-fthread-jumps @gol
5747 -falign-functions -falign-jumps @gol
5748 -falign-loops -falign-labels @gol
5751 -fcse-follow-jumps -fcse-skip-blocks @gol
5752 -fdelete-null-pointer-checks @gol
5753 -fexpensive-optimizations @gol
5754 -fgcse -fgcse-lm @gol
5755 -finline-small-functions @gol
5756 -findirect-inlining @gol
5758 -foptimize-sibling-calls @gol
5761 -freorder-blocks -freorder-functions @gol
5762 -frerun-cse-after-loop @gol
5763 -fsched-interblock -fsched-spec @gol
5764 -fschedule-insns -fschedule-insns2 @gol
5765 -fstrict-aliasing -fstrict-overflow @gol
5766 -ftree-switch-conversion @gol
5770 Please note the warning under @option{-fgcse} about
5771 invoking @option{-O2} on programs that use computed gotos.
5775 Optimize yet more. @option{-O3} turns on all optimizations specified
5776 by @option{-O2} and also turns on the @option{-finline-functions},
5777 @option{-funswitch-loops}, @option{-fpredictive-commoning},
5778 @option{-fgcse-after-reload} and @option{-ftree-vectorize} options.
5782 Reduce compilation time and make debugging produce the expected
5783 results. This is the default.
5787 Optimize for size. @option{-Os} enables all @option{-O2} optimizations that
5788 do not typically increase code size. It also performs further
5789 optimizations designed to reduce code size.
5791 @option{-Os} disables the following optimization flags:
5792 @gccoptlist{-falign-functions -falign-jumps -falign-loops @gol
5793 -falign-labels -freorder-blocks -freorder-blocks-and-partition @gol
5794 -fprefetch-loop-arrays -ftree-vect-loop-version}
5796 If you use multiple @option{-O} options, with or without level numbers,
5797 the last such option is the one that is effective.
5800 Options of the form @option{-f@var{flag}} specify machine-independent
5801 flags. Most flags have both positive and negative forms; the negative
5802 form of @option{-ffoo} would be @option{-fno-foo}. In the table
5803 below, only one of the forms is listed---the one you typically will
5804 use. You can figure out the other form by either removing @samp{no-}
5807 The following options control specific optimizations. They are either
5808 activated by @option{-O} options or are related to ones that are. You
5809 can use the following flags in the rare cases when ``fine-tuning'' of
5810 optimizations to be performed is desired.
5813 @item -fno-default-inline
5814 @opindex fno-default-inline
5815 Do not make member functions inline by default merely because they are
5816 defined inside the class scope (C++ only). Otherwise, when you specify
5817 @w{@option{-O}}, member functions defined inside class scope are compiled
5818 inline by default; i.e., you don't need to add @samp{inline} in front of
5819 the member function name.
5821 @item -fno-defer-pop
5822 @opindex fno-defer-pop
5823 Always pop the arguments to each function call as soon as that function
5824 returns. For machines which must pop arguments after a function call,
5825 the compiler normally lets arguments accumulate on the stack for several
5826 function calls and pops them all at once.
5828 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5830 @item -fforward-propagate
5831 @opindex fforward-propagate
5832 Perform a forward propagation pass on RTL@. The pass tries to combine two
5833 instructions and checks if the result can be simplified. If loop unrolling
5834 is active, two passes are performed and the second is scheduled after
5837 This option is enabled by default at optimization levels @option{-O},
5838 @option{-O2}, @option{-O3}, @option{-Os}.
5840 @item -fomit-frame-pointer
5841 @opindex fomit-frame-pointer
5842 Don't keep the frame pointer in a register for functions that
5843 don't need one. This avoids the instructions to save, set up and
5844 restore frame pointers; it also makes an extra register available
5845 in many functions. @strong{It also makes debugging impossible on
5848 On some machines, such as the VAX, this flag has no effect, because
5849 the standard calling sequence automatically handles the frame pointer
5850 and nothing is saved by pretending it doesn't exist. The
5851 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
5852 whether a target machine supports this flag. @xref{Registers,,Register
5853 Usage, gccint, GNU Compiler Collection (GCC) Internals}.
5855 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5857 @item -foptimize-sibling-calls
5858 @opindex foptimize-sibling-calls
5859 Optimize sibling and tail recursive calls.
5861 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
5865 Don't pay attention to the @code{inline} keyword. Normally this option
5866 is used to keep the compiler from expanding any functions inline.
5867 Note that if you are not optimizing, no functions can be expanded inline.
5869 @item -finline-small-functions
5870 @opindex finline-small-functions
5871 Integrate functions into their callers when their body is smaller than expected
5872 function call code (so overall size of program gets smaller). The compiler
5873 heuristically decides which functions are simple enough to be worth integrating
5876 Enabled at level @option{-O2}.
5878 @item -findirect-inlining
5879 @opindex findirect-inlining
5880 Inline also indirect calls that are discovered to be known at compile
5881 time thanks to previous inlining. This option has any effect only
5882 when inlining itself is turned on by the @option{-finline-functions}
5883 or @option{-finline-small-functions} options.
5885 Enabled at level @option{-O2}.
5887 @item -finline-functions
5888 @opindex finline-functions
5889 Integrate all simple functions into their callers. The compiler
5890 heuristically decides which functions are simple enough to be worth
5891 integrating in this way.
5893 If all calls to a given function are integrated, and the function is
5894 declared @code{static}, then the function is normally not output as
5895 assembler code in its own right.
5897 Enabled at level @option{-O3}.
5899 @item -finline-functions-called-once
5900 @opindex finline-functions-called-once
5901 Consider all @code{static} functions called once for inlining into their
5902 caller even if they are not marked @code{inline}. If a call to a given
5903 function is integrated, then the function is not output as assembler code
5906 Enabled at levels @option{-O1}, @option{-O2}, @option{-O3} and @option{-Os}.
5908 @item -fearly-inlining
5909 @opindex fearly-inlining
5910 Inline functions marked by @code{always_inline} and functions whose body seems
5911 smaller than the function call overhead early before doing
5912 @option{-fprofile-generate} instrumentation and real inlining pass. Doing so
5913 makes profiling significantly cheaper and usually inlining faster on programs
5914 having large chains of nested wrapper functions.
5920 Perform interprocedural scalar replacement of aggregates, removal of
5921 unused parameters and replacement of parameters passed by reference
5922 by parameters passed by value.
5924 Enabled at levels @option{-O2}, @option{-O3} and @option{-Os}.
5926 @item -finline-limit=@var{n}
5927 @opindex finline-limit
5928 By default, GCC limits the size of functions that can be inlined. This flag
5929 allows coarse control of this limit. @var{n} is the size of functions that
5930 can be inlined in number of pseudo instructions.
5932 Inlining is actually controlled by a number of parameters, which may be
5933 specified individually by using @option{--param @var{name}=@var{value}}.
5934 The @option{-finline-limit=@var{n}} option sets some of these parameters
5938 @item max-inline-insns-single
5939 is set to @var{n}/2.
5940 @item max-inline-insns-auto
5941 is set to @var{n}/2.
5944 See below for a documentation of the individual
5945 parameters controlling inlining and for the defaults of these parameters.
5947 @emph{Note:} there may be no value to @option{-finline-limit} that results
5948 in default behavior.
5950 @emph{Note:} pseudo instruction represents, in this particular context, an
5951 abstract measurement of function's size. In no way does it represent a count
5952 of assembly instructions and as such its exact meaning might change from one
5953 release to an another.
5955 @item -fkeep-inline-functions
5956 @opindex fkeep-inline-functions
5957 In C, emit @code{static} functions that are declared @code{inline}
5958 into the object file, even if the function has been inlined into all
5959 of its callers. This switch does not affect functions using the
5960 @code{extern inline} extension in GNU C90@. In C++, emit any and all
5961 inline functions into the object file.
5963 @item -fkeep-static-consts
5964 @opindex fkeep-static-consts
5965 Emit variables declared @code{static const} when optimization isn't turned
5966 on, even if the variables aren't referenced.
5968 GCC enables this option by default. If you want to force the compiler to
5969 check if the variable was referenced, regardless of whether or not
5970 optimization is turned on, use the @option{-fno-keep-static-consts} option.
5972 @item -fmerge-constants
5973 @opindex fmerge-constants
5974 Attempt to merge identical constants (string constants and floating point
5975 constants) across compilation units.
5977 This option is the default for optimized compilation if the assembler and
5978 linker support it. Use @option{-fno-merge-constants} to inhibit this
5981 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
5983 @item -fmerge-all-constants
5984 @opindex fmerge-all-constants
5985 Attempt to merge identical constants and identical variables.
5987 This option implies @option{-fmerge-constants}. In addition to
5988 @option{-fmerge-constants} this considers e.g.@: even constant initialized
5989 arrays or initialized constant variables with integral or floating point
5990 types. Languages like C or C++ require each variable, including multiple
5991 instances of the same variable in recursive calls, to have distinct locations,
5992 so using this option will result in non-conforming
5995 @item -fmodulo-sched
5996 @opindex fmodulo-sched
5997 Perform swing modulo scheduling immediately before the first scheduling
5998 pass. This pass looks at innermost loops and reorders their
5999 instructions by overlapping different iterations.
6001 @item -fmodulo-sched-allow-regmoves
6002 @opindex fmodulo-sched-allow-regmoves
6003 Perform more aggressive SMS based modulo scheduling with register moves
6004 allowed. By setting this flag certain anti-dependences edges will be
6005 deleted which will trigger the generation of reg-moves based on the
6006 life-range analysis. This option is effective only with
6007 @option{-fmodulo-sched} enabled.
6009 @item -fno-branch-count-reg
6010 @opindex fno-branch-count-reg
6011 Do not use ``decrement and branch'' instructions on a count register,
6012 but instead generate a sequence of instructions that decrement a
6013 register, compare it against zero, then branch based upon the result.
6014 This option is only meaningful on architectures that support such
6015 instructions, which include x86, PowerPC, IA-64 and S/390.
6017 The default is @option{-fbranch-count-reg}.
6019 @item -fno-function-cse
6020 @opindex fno-function-cse
6021 Do not put function addresses in registers; make each instruction that
6022 calls a constant function contain the function's address explicitly.
6024 This option results in less efficient code, but some strange hacks
6025 that alter the assembler output may be confused by the optimizations
6026 performed when this option is not used.
6028 The default is @option{-ffunction-cse}
6030 @item -fno-zero-initialized-in-bss
6031 @opindex fno-zero-initialized-in-bss
6032 If the target supports a BSS section, GCC by default puts variables that
6033 are initialized to zero into BSS@. This can save space in the resulting
6036 This option turns off this behavior because some programs explicitly
6037 rely on variables going to the data section. E.g., so that the
6038 resulting executable can find the beginning of that section and/or make
6039 assumptions based on that.
6041 The default is @option{-fzero-initialized-in-bss}.
6043 @item -fmudflap -fmudflapth -fmudflapir
6047 @cindex bounds checking
6049 For front-ends that support it (C and C++), instrument all risky
6050 pointer/array dereferencing operations, some standard library
6051 string/heap functions, and some other associated constructs with
6052 range/validity tests. Modules so instrumented should be immune to
6053 buffer overflows, invalid heap use, and some other classes of C/C++
6054 programming errors. The instrumentation relies on a separate runtime
6055 library (@file{libmudflap}), which will be linked into a program if
6056 @option{-fmudflap} is given at link time. Run-time behavior of the
6057 instrumented program is controlled by the @env{MUDFLAP_OPTIONS}
6058 environment variable. See @code{env MUDFLAP_OPTIONS=-help a.out}
6061 Use @option{-fmudflapth} instead of @option{-fmudflap} to compile and to
6062 link if your program is multi-threaded. Use @option{-fmudflapir}, in
6063 addition to @option{-fmudflap} or @option{-fmudflapth}, if
6064 instrumentation should ignore pointer reads. This produces less
6065 instrumentation (and therefore faster execution) and still provides
6066 some protection against outright memory corrupting writes, but allows
6067 erroneously read data to propagate within a program.
6069 @item -fthread-jumps
6070 @opindex fthread-jumps
6071 Perform optimizations where we check to see if a jump branches to a
6072 location where another comparison subsumed by the first is found. If
6073 so, the first branch is redirected to either the destination of the
6074 second branch or a point immediately following it, depending on whether
6075 the condition is known to be true or false.
6077 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6079 @item -fsplit-wide-types
6080 @opindex fsplit-wide-types
6081 When using a type that occupies multiple registers, such as @code{long
6082 long} on a 32-bit system, split the registers apart and allocate them
6083 independently. This normally generates better code for those types,
6084 but may make debugging more difficult.
6086 Enabled at levels @option{-O}, @option{-O2}, @option{-O3},
6089 @item -fcse-follow-jumps
6090 @opindex fcse-follow-jumps
6091 In common subexpression elimination (CSE), scan through jump instructions
6092 when the target of the jump is not reached by any other path. For
6093 example, when CSE encounters an @code{if} statement with an
6094 @code{else} clause, CSE will follow the jump when the condition
6097 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6099 @item -fcse-skip-blocks
6100 @opindex fcse-skip-blocks
6101 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
6102 follow jumps which conditionally skip over blocks. When CSE
6103 encounters a simple @code{if} statement with no else clause,
6104 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
6105 body of the @code{if}.
6107 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6109 @item -frerun-cse-after-loop
6110 @opindex frerun-cse-after-loop
6111 Re-run common subexpression elimination after loop optimizations has been
6114 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6118 Perform a global common subexpression elimination pass.
6119 This pass also performs global constant and copy propagation.
6121 @emph{Note:} When compiling a program using computed gotos, a GCC
6122 extension, you may get better runtime performance if you disable
6123 the global common subexpression elimination pass by adding
6124 @option{-fno-gcse} to the command line.
6126 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6130 When @option{-fgcse-lm} is enabled, global common subexpression elimination will
6131 attempt to move loads which are only killed by stores into themselves. This
6132 allows a loop containing a load/store sequence to be changed to a load outside
6133 the loop, and a copy/store within the loop.
6135 Enabled by default when gcse is enabled.
6139 When @option{-fgcse-sm} is enabled, a store motion pass is run after
6140 global common subexpression elimination. This pass will attempt to move
6141 stores out of loops. When used in conjunction with @option{-fgcse-lm},
6142 loops containing a load/store sequence can be changed to a load before
6143 the loop and a store after the loop.
6145 Not enabled at any optimization level.
6149 When @option{-fgcse-las} is enabled, the global common subexpression
6150 elimination pass eliminates redundant loads that come after stores to the
6151 same memory location (both partial and full redundancies).
6153 Not enabled at any optimization level.
6155 @item -fgcse-after-reload
6156 @opindex fgcse-after-reload
6157 When @option{-fgcse-after-reload} is enabled, a redundant load elimination
6158 pass is performed after reload. The purpose of this pass is to cleanup
6161 @item -funsafe-loop-optimizations
6162 @opindex funsafe-loop-optimizations
6163 If given, the loop optimizer will assume that loop indices do not
6164 overflow, and that the loops with nontrivial exit condition are not
6165 infinite. This enables a wider range of loop optimizations even if
6166 the loop optimizer itself cannot prove that these assumptions are valid.
6167 Using @option{-Wunsafe-loop-optimizations}, the compiler will warn you
6168 if it finds this kind of loop.
6170 @item -fcrossjumping
6171 @opindex fcrossjumping
6172 Perform cross-jumping transformation. This transformation unifies equivalent code and save code size. The
6173 resulting code may or may not perform better than without cross-jumping.
6175 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6177 @item -fauto-inc-dec
6178 @opindex fauto-inc-dec
6179 Combine increments or decrements of addresses with memory accesses.
6180 This pass is always skipped on architectures that do not have
6181 instructions to support this. Enabled by default at @option{-O} and
6182 higher on architectures that support this.
6186 Perform dead code elimination (DCE) on RTL@.
6187 Enabled by default at @option{-O} and higher.
6191 Perform dead store elimination (DSE) on RTL@.
6192 Enabled by default at @option{-O} and higher.
6194 @item -fif-conversion
6195 @opindex fif-conversion
6196 Attempt to transform conditional jumps into branch-less equivalents. This
6197 include use of conditional moves, min, max, set flags and abs instructions, and
6198 some tricks doable by standard arithmetics. The use of conditional execution
6199 on chips where it is available is controlled by @code{if-conversion2}.
6201 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
6203 @item -fif-conversion2
6204 @opindex fif-conversion2
6205 Use conditional execution (where available) to transform conditional jumps into
6206 branch-less equivalents.
6208 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
6210 @item -fdelete-null-pointer-checks
6211 @opindex fdelete-null-pointer-checks
6212 Assume that programs cannot safely dereference null pointers, and that
6213 no code or data element resides there. This enables simple constant
6214 folding optimizations at all optimization levels. In addition, other
6215 optimization passes in GCC use this flag to control global dataflow
6216 analyses that eliminate useless checks for null pointers; these assume
6217 that if a pointer is checked after it has already been dereferenced,
6220 Note however that in some environments this assumption is not true.
6221 Use @option{-fno-delete-null-pointer-checks} to disable this optimization
6222 for programs which depend on that behavior.
6224 Some targets, especially embedded ones, disable this option at all levels.
6225 Otherwise it is enabled at all levels: @option{-O0}, @option{-O1},
6226 @option{-O2}, @option{-O3}, @option{-Os}. Passes that use the information
6227 are enabled independently at different optimization levels.
6229 @item -fexpensive-optimizations
6230 @opindex fexpensive-optimizations
6231 Perform a number of minor optimizations that are relatively expensive.
6233 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6235 @item -foptimize-register-move
6237 @opindex foptimize-register-move
6239 Attempt to reassign register numbers in move instructions and as
6240 operands of other simple instructions in order to maximize the amount of
6241 register tying. This is especially helpful on machines with two-operand
6244 Note @option{-fregmove} and @option{-foptimize-register-move} are the same
6247 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6249 @item -fira-algorithm=@var{algorithm}
6250 Use specified coloring algorithm for the integrated register
6251 allocator. The @var{algorithm} argument should be @code{priority} or
6252 @code{CB}. The first algorithm specifies Chow's priority coloring,
6253 the second one specifies Chaitin-Briggs coloring. The second
6254 algorithm can be unimplemented for some architectures. If it is
6255 implemented, it is the default because Chaitin-Briggs coloring as a
6256 rule generates a better code.
6258 @item -fira-region=@var{region}
6259 Use specified regions for the integrated register allocator. The
6260 @var{region} argument should be one of @code{all}, @code{mixed}, or
6261 @code{one}. The first value means using all loops as register
6262 allocation regions, the second value which is the default means using
6263 all loops except for loops with small register pressure as the
6264 regions, and third one means using all function as a single region.
6265 The first value can give best result for machines with small size and
6266 irregular register set, the third one results in faster and generates
6267 decent code and the smallest size code, and the default value usually
6268 give the best results in most cases and for most architectures.
6270 @item -fira-coalesce
6271 @opindex fira-coalesce
6272 Do optimistic register coalescing. This option might be profitable for
6273 architectures with big regular register files.
6275 @item -fira-loop-pressure
6276 @opindex fira-loop-pressure
6277 Use IRA to evaluate register pressure in loops for decision to move
6278 loop invariants. Usage of this option usually results in generation
6279 of faster and smaller code on machines with big register files (>= 32
6280 registers) but it can slow compiler down.
6282 This option is enabled at level @option{-O3} for some targets.
6284 @item -fno-ira-share-save-slots
6285 @opindex fno-ira-share-save-slots
6286 Switch off sharing stack slots used for saving call used hard
6287 registers living through a call. Each hard register will get a
6288 separate stack slot and as a result function stack frame will be
6291 @item -fno-ira-share-spill-slots
6292 @opindex fno-ira-share-spill-slots
6293 Switch off sharing stack slots allocated for pseudo-registers. Each
6294 pseudo-register which did not get a hard register will get a separate
6295 stack slot and as a result function stack frame will be bigger.
6297 @item -fira-verbose=@var{n}
6298 @opindex fira-verbose
6299 Set up how verbose dump file for the integrated register allocator
6300 will be. Default value is 5. If the value is greater or equal to 10,
6301 the dump file will be stderr as if the value were @var{n} minus 10.
6303 @item -fdelayed-branch
6304 @opindex fdelayed-branch
6305 If supported for the target machine, attempt to reorder instructions
6306 to exploit instruction slots available after delayed branch
6309 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
6311 @item -fschedule-insns
6312 @opindex fschedule-insns
6313 If supported for the target machine, attempt to reorder instructions to
6314 eliminate execution stalls due to required data being unavailable. This
6315 helps machines that have slow floating point or memory load instructions
6316 by allowing other instructions to be issued until the result of the load
6317 or floating point instruction is required.
6319 Enabled at levels @option{-O2}, @option{-O3}.
6321 @item -fschedule-insns2
6322 @opindex fschedule-insns2
6323 Similar to @option{-fschedule-insns}, but requests an additional pass of
6324 instruction scheduling after register allocation has been done. This is
6325 especially useful on machines with a relatively small number of
6326 registers and where memory load instructions take more than one cycle.
6328 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6330 @item -fno-sched-interblock
6331 @opindex fno-sched-interblock
6332 Don't schedule instructions across basic blocks. This is normally
6333 enabled by default when scheduling before register allocation, i.e.@:
6334 with @option{-fschedule-insns} or at @option{-O2} or higher.
6336 @item -fno-sched-spec
6337 @opindex fno-sched-spec
6338 Don't allow speculative motion of non-load instructions. This is normally
6339 enabled by default when scheduling before register allocation, i.e.@:
6340 with @option{-fschedule-insns} or at @option{-O2} or higher.
6342 @item -fsched-pressure
6343 @opindex fsched-pressure
6344 Enable register pressure sensitive insn scheduling before the register
6345 allocation. This only makes sense when scheduling before register
6346 allocation is enabled, i.e.@: with @option{-fschedule-insns} or at
6347 @option{-O2} or higher. Usage of this option can improve the
6348 generated code and decrease its size by preventing register pressure
6349 increase above the number of available hard registers and as a
6350 consequence register spills in the register allocation.
6352 @item -fsched-spec-load
6353 @opindex fsched-spec-load
6354 Allow speculative motion of some load instructions. This only makes
6355 sense when scheduling before register allocation, i.e.@: with
6356 @option{-fschedule-insns} or at @option{-O2} or higher.
6358 @item -fsched-spec-load-dangerous
6359 @opindex fsched-spec-load-dangerous
6360 Allow speculative motion of more load instructions. This only makes
6361 sense when scheduling before register allocation, i.e.@: with
6362 @option{-fschedule-insns} or at @option{-O2} or higher.
6364 @item -fsched-stalled-insns
6365 @itemx -fsched-stalled-insns=@var{n}
6366 @opindex fsched-stalled-insns
6367 Define how many insns (if any) can be moved prematurely from the queue
6368 of stalled insns into the ready list, during the second scheduling pass.
6369 @option{-fno-sched-stalled-insns} means that no insns will be moved
6370 prematurely, @option{-fsched-stalled-insns=0} means there is no limit
6371 on how many queued insns can be moved prematurely.
6372 @option{-fsched-stalled-insns} without a value is equivalent to
6373 @option{-fsched-stalled-insns=1}.
6375 @item -fsched-stalled-insns-dep
6376 @itemx -fsched-stalled-insns-dep=@var{n}
6377 @opindex fsched-stalled-insns-dep
6378 Define how many insn groups (cycles) will be examined for a dependency
6379 on a stalled insn that is candidate for premature removal from the queue
6380 of stalled insns. This has an effect only during the second scheduling pass,
6381 and only if @option{-fsched-stalled-insns} is used.
6382 @option{-fno-sched-stalled-insns-dep} is equivalent to
6383 @option{-fsched-stalled-insns-dep=0}.
6384 @option{-fsched-stalled-insns-dep} without a value is equivalent to
6385 @option{-fsched-stalled-insns-dep=1}.
6387 @item -fsched2-use-superblocks
6388 @opindex fsched2-use-superblocks
6389 When scheduling after register allocation, do use superblock scheduling
6390 algorithm. Superblock scheduling allows motion across basic block boundaries
6391 resulting on faster schedules. This option is experimental, as not all machine
6392 descriptions used by GCC model the CPU closely enough to avoid unreliable
6393 results from the algorithm.
6395 This only makes sense when scheduling after register allocation, i.e.@: with
6396 @option{-fschedule-insns2} or at @option{-O2} or higher.
6398 @item -fsched-group-heuristic
6399 @opindex fsched-group-heuristic
6400 Enable the group heuristic in the scheduler. This heuristic favors
6401 the instruction that belongs to a schedule group. This is enabled
6402 by default when scheduling is enabled, i.e.@: with @option{-fschedule-insns}
6403 or @option{-fschedule-insns2} or at @option{-O2} or higher.
6405 @item -fsched-critical-path-heuristic
6406 @opindex fsched-critical-path-heuristic
6407 Enable the critical-path heuristic in the scheduler. This heuristic favors
6408 instructions on the critical path. This is enabled by default when
6409 scheduling is enabled, i.e.@: with @option{-fschedule-insns}
6410 or @option{-fschedule-insns2} or at @option{-O2} or higher.
6412 @item -fsched-spec-insn-heuristic
6413 @opindex fsched-spec-insn-heuristic
6414 Enable the speculative instruction heuristic in the scheduler. This
6415 heuristic favors speculative instructions with greater dependency weakness.
6416 This is enabled by default when scheduling is enabled, i.e.@:
6417 with @option{-fschedule-insns} or @option{-fschedule-insns2}
6418 or at @option{-O2} or higher.
6420 @item -fsched-rank-heuristic
6421 @opindex fsched-rank-heuristic
6422 Enable the rank heuristic in the scheduler. This heuristic favors
6423 the instruction belonging to a basic block with greater size or frequency.
6424 This is enabled by default when scheduling is enabled, i.e.@:
6425 with @option{-fschedule-insns} or @option{-fschedule-insns2} or
6426 at @option{-O2} or higher.
6428 @item -fsched-last-insn-heuristic
6429 @opindex fsched-last-insn-heuristic
6430 Enable the last-instruction heuristic in the scheduler. This heuristic
6431 favors the instruction that is less dependent on the last instruction
6432 scheduled. This is enabled by default when scheduling is enabled,
6433 i.e.@: with @option{-fschedule-insns} or @option{-fschedule-insns2} or
6434 at @option{-O2} or higher.
6436 @item -fsched-dep-count-heuristic
6437 @opindex fsched-dep-count-heuristic
6438 Enable the dependent-count heuristic in the scheduler. This heuristic
6439 favors the instruction that has more instructions depending on it.
6440 This is enabled by default when scheduling is enabled, i.e.@:
6441 with @option{-fschedule-insns} or @option{-fschedule-insns2} or
6442 at @option{-O2} or higher.
6444 @item -freschedule-modulo-scheduled-loops
6445 @opindex freschedule-modulo-scheduled-loops
6446 The modulo scheduling comes before the traditional scheduling, if a loop
6447 was modulo scheduled we may want to prevent the later scheduling passes
6448 from changing its schedule, we use this option to control that.
6450 @item -fselective-scheduling
6451 @opindex fselective-scheduling
6452 Schedule instructions using selective scheduling algorithm. Selective
6453 scheduling runs instead of the first scheduler pass.
6455 @item -fselective-scheduling2
6456 @opindex fselective-scheduling2
6457 Schedule instructions using selective scheduling algorithm. Selective
6458 scheduling runs instead of the second scheduler pass.
6460 @item -fsel-sched-pipelining
6461 @opindex fsel-sched-pipelining
6462 Enable software pipelining of innermost loops during selective scheduling.
6463 This option has no effect until one of @option{-fselective-scheduling} or
6464 @option{-fselective-scheduling2} is turned on.
6466 @item -fsel-sched-pipelining-outer-loops
6467 @opindex fsel-sched-pipelining-outer-loops
6468 When pipelining loops during selective scheduling, also pipeline outer loops.
6469 This option has no effect until @option{-fsel-sched-pipelining} is turned on.
6471 @item -fcaller-saves
6472 @opindex fcaller-saves
6473 Enable values to be allocated in registers that will be clobbered by
6474 function calls, by emitting extra instructions to save and restore the
6475 registers around such calls. Such allocation is done only when it
6476 seems to result in better code than would otherwise be produced.
6478 This option is always enabled by default on certain machines, usually
6479 those which have no call-preserved registers to use instead.
6481 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6483 @item -fconserve-stack
6484 @opindex fconserve-stack
6485 Attempt to minimize stack usage. The compiler will attempt to use less
6486 stack space, even if that makes the program slower. This option
6487 implies setting the @option{large-stack-frame} parameter to 100
6488 and the @option{large-stack-frame-growth} parameter to 400.
6490 @item -ftree-reassoc
6491 @opindex ftree-reassoc
6492 Perform reassociation on trees. This flag is enabled by default
6493 at @option{-O} and higher.
6497 Perform partial redundancy elimination (PRE) on trees. This flag is
6498 enabled by default at @option{-O2} and @option{-O3}.
6500 @item -ftree-forwprop
6501 @opindex ftree-forwprop
6502 Perform forward propagation on trees. This flag is enabled by default
6503 at @option{-O} and higher.
6507 Perform full redundancy elimination (FRE) on trees. The difference
6508 between FRE and PRE is that FRE only considers expressions
6509 that are computed on all paths leading to the redundant computation.
6510 This analysis is faster than PRE, though it exposes fewer redundancies.
6511 This flag is enabled by default at @option{-O} and higher.
6513 @item -ftree-phiprop
6514 @opindex ftree-phiprop
6515 Perform hoisting of loads from conditional pointers on trees. This
6516 pass is enabled by default at @option{-O} and higher.
6518 @item -ftree-copy-prop
6519 @opindex ftree-copy-prop
6520 Perform copy propagation on trees. This pass eliminates unnecessary
6521 copy operations. This flag is enabled by default at @option{-O} and
6524 @item -fipa-pure-const
6525 @opindex fipa-pure-const
6526 Discover which functions are pure or constant.
6527 Enabled by default at @option{-O} and higher.
6529 @item -fipa-reference
6530 @opindex fipa-reference
6531 Discover which static variables do not escape cannot escape the
6533 Enabled by default at @option{-O} and higher.
6535 @item -fipa-struct-reorg
6536 @opindex fipa-struct-reorg
6537 Perform structure reorganization optimization, that change C-like structures
6538 layout in order to better utilize spatial locality. This transformation is
6539 affective for programs containing arrays of structures. Available in two
6540 compilation modes: profile-based (enabled with @option{-fprofile-generate})
6541 or static (which uses built-in heuristics). Require @option{-fipa-type-escape}
6542 to provide the safety of this transformation. It works only in whole program
6543 mode, so it requires @option{-fwhole-program} and @option{-combine} to be
6544 enabled. Structures considered @samp{cold} by this transformation are not
6545 affected (see @option{--param struct-reorg-cold-struct-ratio=@var{value}}).
6547 With this flag, the program debug info reflects a new structure layout.
6551 Perform interprocedural pointer analysis. This option is experimental
6552 and does not affect generated code.
6556 Perform interprocedural constant propagation.
6557 This optimization analyzes the program to determine when values passed
6558 to functions are constants and then optimizes accordingly.
6559 This optimization can substantially increase performance
6560 if the application has constants passed to functions.
6561 This flag is enabled by default at @option{-O2}, @option{-Os} and @option{-O3}.
6563 @item -fipa-cp-clone
6564 @opindex fipa-cp-clone
6565 Perform function cloning to make interprocedural constant propagation stronger.
6566 When enabled, interprocedural constant propagation will perform function cloning
6567 when externally visible function can be called with constant arguments.
6568 Because this optimization can create multiple copies of functions,
6569 it may significantly increase code size
6570 (see @option{--param ipcp-unit-growth=@var{value}}).
6571 This flag is enabled by default at @option{-O3}.
6573 @item -fipa-matrix-reorg
6574 @opindex fipa-matrix-reorg
6575 Perform matrix flattening and transposing.
6576 Matrix flattening tries to replace an @math{m}-dimensional matrix
6577 with its equivalent @math{n}-dimensional matrix, where @math{n < m}.
6578 This reduces the level of indirection needed for accessing the elements
6579 of the matrix. The second optimization is matrix transposing that
6580 attempts to change the order of the matrix's dimensions in order to
6581 improve cache locality.
6582 Both optimizations need the @option{-fwhole-program} flag.
6583 Transposing is enabled only if profiling information is available.
6587 Perform forward store motion on trees. This flag is
6588 enabled by default at @option{-O} and higher.
6592 Perform sparse conditional constant propagation (CCP) on trees. This
6593 pass only operates on local scalar variables and is enabled by default
6594 at @option{-O} and higher.
6596 @item -ftree-switch-conversion
6597 Perform conversion of simple initializations in a switch to
6598 initializations from a scalar array. This flag is enabled by default
6599 at @option{-O2} and higher.
6603 Perform dead code elimination (DCE) on trees. This flag is enabled by
6604 default at @option{-O} and higher.
6606 @item -ftree-builtin-call-dce
6607 @opindex ftree-builtin-call-dce
6608 Perform conditional dead code elimination (DCE) for calls to builtin functions
6609 that may set @code{errno} but are otherwise side-effect free. This flag is
6610 enabled by default at @option{-O2} and higher if @option{-Os} is not also
6613 @item -ftree-dominator-opts
6614 @opindex ftree-dominator-opts
6615 Perform a variety of simple scalar cleanups (constant/copy
6616 propagation, redundancy elimination, range propagation and expression
6617 simplification) based on a dominator tree traversal. This also
6618 performs jump threading (to reduce jumps to jumps). This flag is
6619 enabled by default at @option{-O} and higher.
6623 Perform dead store elimination (DSE) on trees. A dead store is a store into
6624 a memory location which will later be overwritten by another store without
6625 any intervening loads. In this case the earlier store can be deleted. This
6626 flag is enabled by default at @option{-O} and higher.
6630 Perform loop header copying on trees. This is beneficial since it increases
6631 effectiveness of code motion optimizations. It also saves one jump. This flag
6632 is enabled by default at @option{-O} and higher. It is not enabled
6633 for @option{-Os}, since it usually increases code size.
6635 @item -ftree-loop-optimize
6636 @opindex ftree-loop-optimize
6637 Perform loop optimizations on trees. This flag is enabled by default
6638 at @option{-O} and higher.
6640 @item -ftree-loop-linear
6641 @opindex ftree-loop-linear
6642 Perform linear loop transformations on tree. This flag can improve cache
6643 performance and allow further loop optimizations to take place.
6645 @item -floop-interchange
6646 Perform loop interchange transformations on loops. Interchanging two
6647 nested loops switches the inner and outer loops. For example, given a
6652 A(J, I) = A(J, I) * C
6656 loop interchange will transform the loop as if the user had written:
6660 A(J, I) = A(J, I) * C
6664 which can be beneficial when @code{N} is larger than the caches,
6665 because in Fortran, the elements of an array are stored in memory
6666 contiguously by column, and the original loop iterates over rows,
6667 potentially creating at each access a cache miss. This optimization
6668 applies to all the languages supported by GCC and is not limited to
6669 Fortran. To use this code transformation, GCC has to be configured
6670 with @option{--with-ppl} and @option{--with-cloog} to enable the
6671 Graphite loop transformation infrastructure.
6673 @item -floop-strip-mine
6674 Perform loop strip mining transformations on loops. Strip mining
6675 splits a loop into two nested loops. The outer loop has strides
6676 equal to the strip size and the inner loop has strides of the
6677 original loop within a strip. The strip length can be changed
6678 using the @option{loop-block-tile-size} parameter. For example,
6685 loop strip mining will transform the loop as if the user had written:
6688 DO I = II, min (II + 50, N)
6693 This optimization applies to all the languages supported by GCC and is
6694 not limited to Fortran. To use this code transformation, GCC has to
6695 be configured with @option{--with-ppl} and @option{--with-cloog} to
6696 enable the Graphite loop transformation infrastructure.
6699 Perform loop blocking transformations on loops. Blocking strip mines
6700 each loop in the loop nest such that the memory accesses of the
6701 element loops fit inside caches. The strip length can be changed
6702 using the @option{loop-block-tile-size} parameter. For example, given
6707 A(J, I) = B(I) + C(J)
6711 loop blocking will transform the loop as if the user had written:
6715 DO I = II, min (II + 50, N)
6716 DO J = JJ, min (JJ + 50, M)
6717 A(J, I) = B(I) + C(J)
6723 which can be beneficial when @code{M} is larger than the caches,
6724 because the innermost loop will iterate over a smaller amount of data
6725 that can be kept in the caches. This optimization applies to all the
6726 languages supported by GCC and is not limited to Fortran. To use this
6727 code transformation, GCC has to be configured with @option{--with-ppl}
6728 and @option{--with-cloog} to enable the Graphite loop transformation
6731 @item -fgraphite-identity
6732 @opindex fgraphite-identity
6733 Enable the identity transformation for graphite. For every SCoP we generate
6734 the polyhedral representation and transform it back to gimple. Using
6735 @option{-fgraphite-identity} we can check the costs or benefits of the
6736 GIMPLE -> GRAPHITE -> GIMPLE transformation. Some minimal optimizations
6737 are also performed by the code generator CLooG, like index splitting and
6738 dead code elimination in loops.
6740 @item -floop-parallelize-all
6741 Use the Graphite data dependence analysis to identify loops that can
6742 be parallelized. Parallelize all the loops that can be analyzed to
6743 not contain loop carried dependences without checking that it is
6744 profitable to parallelize the loops.
6746 @item -fcheck-data-deps
6747 @opindex fcheck-data-deps
6748 Compare the results of several data dependence analyzers. This option
6749 is used for debugging the data dependence analyzers.
6751 @item -ftree-loop-distribution
6752 Perform loop distribution. This flag can improve cache performance on
6753 big loop bodies and allow further loop optimizations, like
6754 parallelization or vectorization, to take place. For example, the loop
6771 @item -ftree-loop-im
6772 @opindex ftree-loop-im
6773 Perform loop invariant motion on trees. This pass moves only invariants that
6774 would be hard to handle at RTL level (function calls, operations that expand to
6775 nontrivial sequences of insns). With @option{-funswitch-loops} it also moves
6776 operands of conditions that are invariant out of the loop, so that we can use
6777 just trivial invariantness analysis in loop unswitching. The pass also includes
6780 @item -ftree-loop-ivcanon
6781 @opindex ftree-loop-ivcanon
6782 Create a canonical counter for number of iterations in the loop for that
6783 determining number of iterations requires complicated analysis. Later
6784 optimizations then may determine the number easily. Useful especially
6785 in connection with unrolling.
6789 Perform induction variable optimizations (strength reduction, induction
6790 variable merging and induction variable elimination) on trees.
6792 @item -ftree-parallelize-loops=n
6793 @opindex ftree-parallelize-loops
6794 Parallelize loops, i.e., split their iteration space to run in n threads.
6795 This is only possible for loops whose iterations are independent
6796 and can be arbitrarily reordered. The optimization is only
6797 profitable on multiprocessor machines, for loops that are CPU-intensive,
6798 rather than constrained e.g.@: by memory bandwidth. This option
6799 implies @option{-pthread}, and thus is only supported on targets
6800 that have support for @option{-pthread}.
6804 Perform function-local points-to analysis on trees. This flag is
6805 enabled by default at @option{-O} and higher.
6809 Perform scalar replacement of aggregates. This pass replaces structure
6810 references with scalars to prevent committing structures to memory too
6811 early. This flag is enabled by default at @option{-O} and higher.
6813 @item -ftree-copyrename
6814 @opindex ftree-copyrename
6815 Perform copy renaming on trees. This pass attempts to rename compiler
6816 temporaries to other variables at copy locations, usually resulting in
6817 variable names which more closely resemble the original variables. This flag
6818 is enabled by default at @option{-O} and higher.
6822 Perform temporary expression replacement during the SSA->normal phase. Single
6823 use/single def temporaries are replaced at their use location with their
6824 defining expression. This results in non-GIMPLE code, but gives the expanders
6825 much more complex trees to work on resulting in better RTL generation. This is
6826 enabled by default at @option{-O} and higher.
6828 @item -ftree-vectorize
6829 @opindex ftree-vectorize
6830 Perform loop vectorization on trees. This flag is enabled by default at
6833 @item -ftree-slp-vectorize
6834 @opindex ftree-slp-vectorize
6835 Perform basic block vectorization on trees. This flag is enabled by default at
6836 @option{-O3} and when @option{-ftree-vectorize} is enabled.
6838 @item -ftree-vect-loop-version
6839 @opindex ftree-vect-loop-version
6840 Perform loop versioning when doing loop vectorization on trees. When a loop
6841 appears to be vectorizable except that data alignment or data dependence cannot
6842 be determined at compile time then vectorized and non-vectorized versions of
6843 the loop are generated along with runtime checks for alignment or dependence
6844 to control which version is executed. This option is enabled by default
6845 except at level @option{-Os} where it is disabled.
6847 @item -fvect-cost-model
6848 @opindex fvect-cost-model
6849 Enable cost model for vectorization.
6853 Perform Value Range Propagation on trees. This is similar to the
6854 constant propagation pass, but instead of values, ranges of values are
6855 propagated. This allows the optimizers to remove unnecessary range
6856 checks like array bound checks and null pointer checks. This is
6857 enabled by default at @option{-O2} and higher. Null pointer check
6858 elimination is only done if @option{-fdelete-null-pointer-checks} is
6863 Perform tail duplication to enlarge superblock size. This transformation
6864 simplifies the control flow of the function allowing other optimizations to do
6867 @item -funroll-loops
6868 @opindex funroll-loops
6869 Unroll loops whose number of iterations can be determined at compile
6870 time or upon entry to the loop. @option{-funroll-loops} implies
6871 @option{-frerun-cse-after-loop}. This option makes code larger,
6872 and may or may not make it run faster.
6874 @item -funroll-all-loops
6875 @opindex funroll-all-loops
6876 Unroll all loops, even if their number of iterations is uncertain when
6877 the loop is entered. This usually makes programs run more slowly.
6878 @option{-funroll-all-loops} implies the same options as
6879 @option{-funroll-loops},
6881 @item -fsplit-ivs-in-unroller
6882 @opindex fsplit-ivs-in-unroller
6883 Enables expressing of values of induction variables in later iterations
6884 of the unrolled loop using the value in the first iteration. This breaks
6885 long dependency chains, thus improving efficiency of the scheduling passes.
6887 Combination of @option{-fweb} and CSE is often sufficient to obtain the
6888 same effect. However in cases the loop body is more complicated than
6889 a single basic block, this is not reliable. It also does not work at all
6890 on some of the architectures due to restrictions in the CSE pass.
6892 This optimization is enabled by default.
6894 @item -fvariable-expansion-in-unroller
6895 @opindex fvariable-expansion-in-unroller
6896 With this option, the compiler will create multiple copies of some
6897 local variables when unrolling a loop which can result in superior code.
6899 @item -fpredictive-commoning
6900 @opindex fpredictive-commoning
6901 Perform predictive commoning optimization, i.e., reusing computations
6902 (especially memory loads and stores) performed in previous
6903 iterations of loops.
6905 This option is enabled at level @option{-O3}.
6907 @item -fprefetch-loop-arrays
6908 @opindex fprefetch-loop-arrays
6909 If supported by the target machine, generate instructions to prefetch
6910 memory to improve the performance of loops that access large arrays.
6912 This option may generate better or worse code; results are highly
6913 dependent on the structure of loops within the source code.
6915 Disabled at level @option{-Os}.
6918 @itemx -fno-peephole2
6919 @opindex fno-peephole
6920 @opindex fno-peephole2
6921 Disable any machine-specific peephole optimizations. The difference
6922 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
6923 are implemented in the compiler; some targets use one, some use the
6924 other, a few use both.
6926 @option{-fpeephole} is enabled by default.
6927 @option{-fpeephole2} enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6929 @item -fno-guess-branch-probability
6930 @opindex fno-guess-branch-probability
6931 Do not guess branch probabilities using heuristics.
6933 GCC will use heuristics to guess branch probabilities if they are
6934 not provided by profiling feedback (@option{-fprofile-arcs}). These
6935 heuristics are based on the control flow graph. If some branch probabilities
6936 are specified by @samp{__builtin_expect}, then the heuristics will be
6937 used to guess branch probabilities for the rest of the control flow graph,
6938 taking the @samp{__builtin_expect} info into account. The interactions
6939 between the heuristics and @samp{__builtin_expect} can be complex, and in
6940 some cases, it may be useful to disable the heuristics so that the effects
6941 of @samp{__builtin_expect} are easier to understand.
6943 The default is @option{-fguess-branch-probability} at levels
6944 @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
6946 @item -freorder-blocks
6947 @opindex freorder-blocks
6948 Reorder basic blocks in the compiled function in order to reduce number of
6949 taken branches and improve code locality.
6951 Enabled at levels @option{-O2}, @option{-O3}.
6953 @item -freorder-blocks-and-partition
6954 @opindex freorder-blocks-and-partition
6955 In addition to reordering basic blocks in the compiled function, in order
6956 to reduce number of taken branches, partitions hot and cold basic blocks
6957 into separate sections of the assembly and .o files, to improve
6958 paging and cache locality performance.
6960 This optimization is automatically turned off in the presence of
6961 exception handling, for linkonce sections, for functions with a user-defined
6962 section attribute and on any architecture that does not support named
6965 @item -freorder-functions
6966 @opindex freorder-functions
6967 Reorder functions in the object file in order to
6968 improve code locality. This is implemented by using special
6969 subsections @code{.text.hot} for most frequently executed functions and
6970 @code{.text.unlikely} for unlikely executed functions. Reordering is done by
6971 the linker so object file format must support named sections and linker must
6972 place them in a reasonable way.
6974 Also profile feedback must be available in to make this option effective. See
6975 @option{-fprofile-arcs} for details.
6977 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
6979 @item -fstrict-aliasing
6980 @opindex fstrict-aliasing
6981 Allow the compiler to assume the strictest aliasing rules applicable to
6982 the language being compiled. For C (and C++), this activates
6983 optimizations based on the type of expressions. In particular, an
6984 object of one type is assumed never to reside at the same address as an
6985 object of a different type, unless the types are almost the same. For
6986 example, an @code{unsigned int} can alias an @code{int}, but not a
6987 @code{void*} or a @code{double}. A character type may alias any other
6990 @anchor{Type-punning}Pay special attention to code like this:
7003 The practice of reading from a different union member than the one most
7004 recently written to (called ``type-punning'') is common. Even with
7005 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
7006 is accessed through the union type. So, the code above will work as
7007 expected. @xref{Structures unions enumerations and bit-fields
7008 implementation}. However, this code might not:
7019 Similarly, access by taking the address, casting the resulting pointer
7020 and dereferencing the result has undefined behavior, even if the cast
7021 uses a union type, e.g.:
7025 return ((union a_union *) &d)->i;
7029 The @option{-fstrict-aliasing} option is enabled at levels
7030 @option{-O2}, @option{-O3}, @option{-Os}.
7032 @item -fstrict-overflow
7033 @opindex fstrict-overflow
7034 Allow the compiler to assume strict signed overflow rules, depending
7035 on the language being compiled. For C (and C++) this means that
7036 overflow when doing arithmetic with signed numbers is undefined, which
7037 means that the compiler may assume that it will not happen. This
7038 permits various optimizations. For example, the compiler will assume
7039 that an expression like @code{i + 10 > i} will always be true for
7040 signed @code{i}. This assumption is only valid if signed overflow is
7041 undefined, as the expression is false if @code{i + 10} overflows when
7042 using twos complement arithmetic. When this option is in effect any
7043 attempt to determine whether an operation on signed numbers will
7044 overflow must be written carefully to not actually involve overflow.
7046 This option also allows the compiler to assume strict pointer
7047 semantics: given a pointer to an object, if adding an offset to that
7048 pointer does not produce a pointer to the same object, the addition is
7049 undefined. This permits the compiler to conclude that @code{p + u >
7050 p} is always true for a pointer @code{p} and unsigned integer
7051 @code{u}. This assumption is only valid because pointer wraparound is
7052 undefined, as the expression is false if @code{p + u} overflows using
7053 twos complement arithmetic.
7055 See also the @option{-fwrapv} option. Using @option{-fwrapv} means
7056 that integer signed overflow is fully defined: it wraps. When
7057 @option{-fwrapv} is used, there is no difference between
7058 @option{-fstrict-overflow} and @option{-fno-strict-overflow} for
7059 integers. With @option{-fwrapv} certain types of overflow are
7060 permitted. For example, if the compiler gets an overflow when doing
7061 arithmetic on constants, the overflowed value can still be used with
7062 @option{-fwrapv}, but not otherwise.
7064 The @option{-fstrict-overflow} option is enabled at levels
7065 @option{-O2}, @option{-O3}, @option{-Os}.
7067 @item -falign-functions
7068 @itemx -falign-functions=@var{n}
7069 @opindex falign-functions
7070 Align the start of functions to the next power-of-two greater than
7071 @var{n}, skipping up to @var{n} bytes. For instance,
7072 @option{-falign-functions=32} aligns functions to the next 32-byte
7073 boundary, but @option{-falign-functions=24} would align to the next
7074 32-byte boundary only if this can be done by skipping 23 bytes or less.
7076 @option{-fno-align-functions} and @option{-falign-functions=1} are
7077 equivalent and mean that functions will not be aligned.
7079 Some assemblers only support this flag when @var{n} is a power of two;
7080 in that case, it is rounded up.
7082 If @var{n} is not specified or is zero, use a machine-dependent default.
7084 Enabled at levels @option{-O2}, @option{-O3}.
7086 @item -falign-labels
7087 @itemx -falign-labels=@var{n}
7088 @opindex falign-labels
7089 Align all branch targets to a power-of-two boundary, skipping up to
7090 @var{n} bytes like @option{-falign-functions}. This option can easily
7091 make code slower, because it must insert dummy operations for when the
7092 branch target is reached in the usual flow of the code.
7094 @option{-fno-align-labels} and @option{-falign-labels=1} are
7095 equivalent and mean that labels will not be aligned.
7097 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
7098 are greater than this value, then their values are used instead.
7100 If @var{n} is not specified or is zero, use a machine-dependent default
7101 which is very likely to be @samp{1}, meaning no alignment.
7103 Enabled at levels @option{-O2}, @option{-O3}.
7106 @itemx -falign-loops=@var{n}
7107 @opindex falign-loops
7108 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
7109 like @option{-falign-functions}. The hope is that the loop will be
7110 executed many times, which will make up for any execution of the dummy
7113 @option{-fno-align-loops} and @option{-falign-loops=1} are
7114 equivalent and mean that loops will not be aligned.
7116 If @var{n} is not specified or is zero, use a machine-dependent default.
7118 Enabled at levels @option{-O2}, @option{-O3}.
7121 @itemx -falign-jumps=@var{n}
7122 @opindex falign-jumps
7123 Align branch targets to a power-of-two boundary, for branch targets
7124 where the targets can only be reached by jumping, skipping up to @var{n}
7125 bytes like @option{-falign-functions}. In this case, no dummy operations
7128 @option{-fno-align-jumps} and @option{-falign-jumps=1} are
7129 equivalent and mean that loops will not be aligned.
7131 If @var{n} is not specified or is zero, use a machine-dependent default.
7133 Enabled at levels @option{-O2}, @option{-O3}.
7135 @item -funit-at-a-time
7136 @opindex funit-at-a-time
7137 This option is left for compatibility reasons. @option{-funit-at-a-time}
7138 has no effect, while @option{-fno-unit-at-a-time} implies
7139 @option{-fno-toplevel-reorder} and @option{-fno-section-anchors}.
7143 @item -fno-toplevel-reorder
7144 @opindex fno-toplevel-reorder
7145 Do not reorder top-level functions, variables, and @code{asm}
7146 statements. Output them in the same order that they appear in the
7147 input file. When this option is used, unreferenced static variables
7148 will not be removed. This option is intended to support existing code
7149 which relies on a particular ordering. For new code, it is better to
7152 Enabled at level @option{-O0}. When disabled explicitly, it also imply
7153 @option{-fno-section-anchors} that is otherwise enabled at @option{-O0} on some
7158 Constructs webs as commonly used for register allocation purposes and assign
7159 each web individual pseudo register. This allows the register allocation pass
7160 to operate on pseudos directly, but also strengthens several other optimization
7161 passes, such as CSE, loop optimizer and trivial dead code remover. It can,
7162 however, make debugging impossible, since variables will no longer stay in a
7165 Enabled by default with @option{-funroll-loops}.
7167 @item -fwhole-program
7168 @opindex fwhole-program
7169 Assume that the current compilation unit represents the whole program being
7170 compiled. All public functions and variables with the exception of @code{main}
7171 and those merged by attribute @code{externally_visible} become static functions
7172 and in effect are optimized more aggressively by interprocedural optimizers.
7173 While this option is equivalent to proper use of the @code{static} keyword for
7174 programs consisting of a single file, in combination with option
7175 @option{-combine}, @option{-flto} or @option{-fwhopr} this flag can be used to
7176 compile many smaller scale programs since the functions and variables become
7177 local for the whole combined compilation unit, not for the single source file
7180 This option implies @option{-fwhole-file} for Fortran programs.
7184 This option runs the standard link-time optimizer. When invoked
7185 with source code, it generates GIMPLE (one of GCC's internal
7186 representations) and writes it to special ELF sections in the object
7187 file. When the object files are linked together, all the function
7188 bodies are read from these ELF sections and instantiated as if they
7189 had been part of the same translation unit.
7191 To use the link-timer optimizer, @option{-flto} needs to be specified at
7192 compile time and during the final link. For example,
7195 gcc -c -O2 -flto foo.c
7196 gcc -c -O2 -flto bar.c
7197 gcc -o myprog -flto -O2 foo.o bar.o
7200 The first two invocations to GCC will save a bytecode representation
7201 of GIMPLE into special ELF sections inside @file{foo.o} and
7202 @file{bar.o}. The final invocation will read the GIMPLE bytecode from
7203 @file{foo.o} and @file{bar.o}, merge the two files into a single
7204 internal image, and compile the result as usual. Since both
7205 @file{foo.o} and @file{bar.o} are merged into a single image, this
7206 causes all the inter-procedural analyses and optimizations in GCC to
7207 work across the two files as if they were a single one. This means,
7208 for example, that the inliner will be able to inline functions in
7209 @file{bar.o} into functions in @file{foo.o} and vice-versa.
7211 Another (simpler) way to enable link-time optimization is,
7214 gcc -o myprog -flto -O2 foo.c bar.c
7217 The above will generate bytecode for @file{foo.c} and @file{bar.c},
7218 merge them together into a single GIMPLE representation and optimize
7219 them as usual to produce @file{myprog}.
7221 The only important thing to keep in mind is that to enable link-time
7222 optimizations the @option{-flto} flag needs to be passed to both the
7223 compile and the link commands.
7225 Note that when a file is compiled with @option{-flto}, the generated
7226 object file will be larger than a regular object file because it will
7227 contain GIMPLE bytecodes and the usual final code. This means that
7228 object files with LTO information can be linked as a normal object
7229 file. So, in the previous example, if the final link is done with
7232 gcc -o myprog foo.o bar.o
7235 The only difference will be that no inter-procedural optimizations
7236 will be applied to produce @file{myprog}. The two object files
7237 @file{foo.o} and @file{bar.o} will be simply sent to the regular
7240 Additionally, the optimization flags used to compile individual files
7241 are not necessarily related to those used at link-time. For instance,
7244 gcc -c -O0 -flto foo.c
7245 gcc -c -O0 -flto bar.c
7246 gcc -o myprog -flto -O3 foo.o bar.o
7249 This will produce individual object files with unoptimized assembler
7250 code, but the resulting binary @file{myprog} will be optimized at
7251 @option{-O3}. Now, if the final binary is generated without
7252 @option{-flto}, then @file{myprog} will not be optimized.
7254 When producing the final binary with @option{-flto}, GCC will only
7255 apply link-time optimizations to those files that contain bytecode.
7256 Therefore, you can mix and match object files and libraries with
7257 GIMPLE bytecodes and final object code. GCC will automatically select
7258 which files to optimize in LTO mode and which files to link without
7261 There are some code generation flags that GCC will preserve when
7262 generating bytecodes, as they need to be used during the final link
7263 stage. Currently, the following options are saved into the GIMPLE
7264 bytecode files: @option{-fPIC}, @option{-fcommon} and all the
7265 @option{-m} target flags.
7267 At link time, these options are read-in and reapplied. Note that the
7268 current implementation makes no attempt at recognizing conflicting
7269 values for these options. If two or more files have a conflicting
7270 value (e.g., one file is compiled with @option{-fPIC} and another
7271 isn't), the compiler will simply use the last value read from the
7272 bytecode files. It is recommended, then, that all the files
7273 participating in the same link be compiled with the same options.
7275 Another feature of LTO is that it is possible to apply interprocedural
7276 optimizations on files written in different languages. This requires
7277 some support in the language front end. Currently, the C, C++ and
7278 Fortran front ends are capable of emitting GIMPLE bytecodes, so
7279 something like this should work
7284 gfortran -c -flto baz.f90
7285 g++ -o myprog -flto -O3 foo.o bar.o baz.o -lgfortran
7288 Notice that the final link is done with @command{g++} to get the C++
7289 runtime libraries and @option{-lgfortran} is added to get the Fortran
7290 runtime libraries. In general, when mixing languages in LTO mode, you
7291 should use the same link command used when mixing languages in a
7292 regular (non-LTO) compilation. This means that if your build process
7293 was mixing languages before, all you need to add is @option{-flto} to
7294 all the compile and link commands.
7296 If LTO encounters objects with C linkage declared with incompatible
7297 types in separate translation units to be linked together (undefined
7298 behavior according to ISO C99 6.2.7), a non-fatal diagnostic may be
7299 issued. The behavior is still undefined at runtime.
7301 If object files containing GIMPLE bytecode are stored in a library
7302 archive, say @file{libfoo.a}, it is possible to extract and use them
7303 in an LTO link if you are using @command{gold} as the linker (which,
7304 in turn requires GCC to be configured with @option{--enable-gold}).
7305 To enable this feature, use the flag @option{-fuse-linker-plugin} at
7309 gcc -o myprog -O2 -flto -fuse-linker-plugin a.o b.o -lfoo
7312 With the linker plugin enabled, @command{gold} will extract the needed
7313 GIMPLE files from @file{libfoo.a} and pass them on to the running GCC
7314 to make them part of the aggregated GIMPLE image to be optimized.
7316 If you are not using @command{gold} and/or do not specify
7317 @option{-fuse-linker-plugin} then the objects inside @file{libfoo.a}
7318 will be extracted and linked as usual, but they will not participate
7319 in the LTO optimization process.
7321 Link time optimizations do not require the presence of the whole
7322 program to operate. If the program does not require any symbols to
7323 be exported, it is possible to combine @option{-flto} and
7324 @option{-fwhopr} with @option{-fwhole-program} to allow the
7325 interprocedural optimizers to use more aggressive assumptions which
7326 may lead to improved optimization opportunities.
7328 Regarding portability: the current implementation of LTO makes no
7329 attempt at generating bytecode that can be ported between different
7330 types of hosts. The bytecode files are versioned and there is a
7331 strict version check, so bytecode files generated in one version of
7332 GCC will not work with an older/newer version of GCC.
7334 Link time optimization does not play well with generating debugging
7335 information. Combining @option{-flto} or @option{-fwhopr} with
7336 @option{-g} is experimental.
7338 This option is disabled by default.
7342 This option is identical in functionality to @option{-flto} but it
7343 differs in how the final link stage is executed. Instead of loading
7344 all the function bodies in memory, the callgraph is analyzed and
7345 optimization decisions are made (whole program analysis or WPA). Once
7346 optimization decisions are made, the callgraph is partitioned and the
7347 different sections are compiled separately (local transformations or
7348 LTRANS)@. This process allows optimizations on very large programs
7349 that otherwise would not fit in memory. This option enables
7350 @option{-fwpa} and @option{-fltrans} automatically.
7352 Disabled by default.
7354 This option is experimental.
7358 This is an internal option used by GCC when compiling with
7359 @option{-fwhopr}. You should never need to use it.
7361 This option runs the link-time optimizer in the whole-program-analysis
7362 (WPA) mode, which reads in summary information from all inputs and
7363 performs a whole-program analysis based on summary information only.
7364 It generates object files for subsequent runs of the link-time
7365 optimizer where individual object files are optimized using both
7366 summary information from the WPA mode and the actual function bodies.
7367 It then drives the LTRANS phase.
7369 Disabled by default.
7373 This is an internal option used by GCC when compiling with
7374 @option{-fwhopr}. You should never need to use it.
7376 This option runs the link-time optimizer in the local-transformation (LTRANS)
7377 mode, which reads in output from a previous run of the LTO in WPA mode.
7378 In the LTRANS mode, LTO optimizes an object and produces the final assembly.
7380 Disabled by default.
7382 @item -fltrans-output-list=@var{file}
7383 @opindex fltrans-output-list
7384 This is an internal option used by GCC when compiling with
7385 @option{-fwhopr}. You should never need to use it.
7387 This option specifies a file to which the names of LTRANS output files are
7388 written. This option is only meaningful in conjunction with @option{-fwpa}.
7390 Disabled by default.
7392 @item -flto-compression-level=@var{n}
7393 This option specifies the level of compression used for intermediate
7394 language written to LTO object files, and is only meaningful in
7395 conjunction with LTO mode (@option{-fwhopr}, @option{-flto}). Valid
7396 values are 0 (no compression) to 9 (maximum compression). Values
7397 outside this range are clamped to either 0 or 9. If the option is not
7398 given, a default balanced compression setting is used.
7401 Prints a report with internal details on the workings of the link-time
7402 optimizer. The contents of this report vary from version to version,
7403 it is meant to be useful to GCC developers when processing object
7404 files in LTO mode (via @option{-fwhopr} or @option{-flto}).
7406 Disabled by default.
7408 @item -fuse-linker-plugin
7409 Enables the extraction of objects with GIMPLE bytecode information
7410 from library archives. This option relies on features available only
7411 in @command{gold}, so to use this you must configure GCC with
7412 @option{--enable-gold}. See @option{-flto} for a description on the
7413 effect of this flag and how to use it.
7415 Disabled by default.
7417 @item -fcprop-registers
7418 @opindex fcprop-registers
7419 After register allocation and post-register allocation instruction splitting,
7420 we perform a copy-propagation pass to try to reduce scheduling dependencies
7421 and occasionally eliminate the copy.
7423 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
7425 @item -fprofile-correction
7426 @opindex fprofile-correction
7427 Profiles collected using an instrumented binary for multi-threaded programs may
7428 be inconsistent due to missed counter updates. When this option is specified,
7429 GCC will use heuristics to correct or smooth out such inconsistencies. By
7430 default, GCC will emit an error message when an inconsistent profile is detected.
7432 @item -fprofile-dir=@var{path}
7433 @opindex fprofile-dir
7435 Set the directory to search the profile data files in to @var{path}.
7436 This option affects only the profile data generated by
7437 @option{-fprofile-generate}, @option{-ftest-coverage}, @option{-fprofile-arcs}
7438 and used by @option{-fprofile-use} and @option{-fbranch-probabilities}
7439 and its related options.
7440 By default, GCC will use the current directory as @var{path}
7441 thus the profile data file will appear in the same directory as the object file.
7443 @item -fprofile-generate
7444 @itemx -fprofile-generate=@var{path}
7445 @opindex fprofile-generate
7447 Enable options usually used for instrumenting application to produce
7448 profile useful for later recompilation with profile feedback based
7449 optimization. You must use @option{-fprofile-generate} both when
7450 compiling and when linking your program.
7452 The following options are enabled: @code{-fprofile-arcs}, @code{-fprofile-values}, @code{-fvpt}.
7454 If @var{path} is specified, GCC will look at the @var{path} to find
7455 the profile feedback data files. See @option{-fprofile-dir}.
7458 @itemx -fprofile-use=@var{path}
7459 @opindex fprofile-use
7460 Enable profile feedback directed optimizations, and optimizations
7461 generally profitable only with profile feedback available.
7463 The following options are enabled: @code{-fbranch-probabilities}, @code{-fvpt},
7464 @code{-funroll-loops}, @code{-fpeel-loops}, @code{-ftracer}
7466 By default, GCC emits an error message if the feedback profiles do not
7467 match the source code. This error can be turned into a warning by using
7468 @option{-Wcoverage-mismatch}. Note this may result in poorly optimized
7471 If @var{path} is specified, GCC will look at the @var{path} to find
7472 the profile feedback data files. See @option{-fprofile-dir}.
7475 The following options control compiler behavior regarding floating
7476 point arithmetic. These options trade off between speed and
7477 correctness. All must be specifically enabled.
7481 @opindex ffloat-store
7482 Do not store floating point variables in registers, and inhibit other
7483 options that might change whether a floating point value is taken from a
7486 @cindex floating point precision
7487 This option prevents undesirable excess precision on machines such as
7488 the 68000 where the floating registers (of the 68881) keep more
7489 precision than a @code{double} is supposed to have. Similarly for the
7490 x86 architecture. For most programs, the excess precision does only
7491 good, but a few programs rely on the precise definition of IEEE floating
7492 point. Use @option{-ffloat-store} for such programs, after modifying
7493 them to store all pertinent intermediate computations into variables.
7495 @item -fexcess-precision=@var{style}
7496 @opindex fexcess-precision
7497 This option allows further control over excess precision on machines
7498 where floating-point registers have more precision than the IEEE
7499 @code{float} and @code{double} types and the processor does not
7500 support operations rounding to those types. By default,
7501 @option{-fexcess-precision=fast} is in effect; this means that
7502 operations are carried out in the precision of the registers and that
7503 it is unpredictable when rounding to the types specified in the source
7504 code takes place. When compiling C, if
7505 @option{-fexcess-precision=standard} is specified then excess
7506 precision will follow the rules specified in ISO C99; in particular,
7507 both casts and assignments cause values to be rounded to their
7508 semantic types (whereas @option{-ffloat-store} only affects
7509 assignments). This option is enabled by default for C if a strict
7510 conformance option such as @option{-std=c99} is used.
7513 @option{-fexcess-precision=standard} is not implemented for languages
7514 other than C, and has no effect if
7515 @option{-funsafe-math-optimizations} or @option{-ffast-math} is
7516 specified. On the x86, it also has no effect if @option{-mfpmath=sse}
7517 or @option{-mfpmath=sse+387} is specified; in the former case, IEEE
7518 semantics apply without excess precision, and in the latter, rounding
7523 Sets @option{-fno-math-errno}, @option{-funsafe-math-optimizations},
7524 @option{-ffinite-math-only}, @option{-fno-rounding-math},
7525 @option{-fno-signaling-nans} and @option{-fcx-limited-range}.
7527 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
7529 This option is not turned on by any @option{-O} option since
7530 it can result in incorrect output for programs which depend on
7531 an exact implementation of IEEE or ISO rules/specifications for
7532 math functions. It may, however, yield faster code for programs
7533 that do not require the guarantees of these specifications.
7535 @item -fno-math-errno
7536 @opindex fno-math-errno
7537 Do not set ERRNO after calling math functions that are executed
7538 with a single instruction, e.g., sqrt. A program that relies on
7539 IEEE exceptions for math error handling may want to use this flag
7540 for speed while maintaining IEEE arithmetic compatibility.
7542 This option is not turned on by any @option{-O} option since
7543 it can result in incorrect output for programs which depend on
7544 an exact implementation of IEEE or ISO rules/specifications for
7545 math functions. It may, however, yield faster code for programs
7546 that do not require the guarantees of these specifications.
7548 The default is @option{-fmath-errno}.
7550 On Darwin systems, the math library never sets @code{errno}. There is
7551 therefore no reason for the compiler to consider the possibility that
7552 it might, and @option{-fno-math-errno} is the default.
7554 @item -funsafe-math-optimizations
7555 @opindex funsafe-math-optimizations
7557 Allow optimizations for floating-point arithmetic that (a) assume
7558 that arguments and results are valid and (b) may violate IEEE or
7559 ANSI standards. When used at link-time, it may include libraries
7560 or startup files that change the default FPU control word or other
7561 similar optimizations.
7563 This option is not turned on by any @option{-O} option since
7564 it can result in incorrect output for programs which depend on
7565 an exact implementation of IEEE or ISO rules/specifications for
7566 math functions. It may, however, yield faster code for programs
7567 that do not require the guarantees of these specifications.
7568 Enables @option{-fno-signed-zeros}, @option{-fno-trapping-math},
7569 @option{-fassociative-math} and @option{-freciprocal-math}.
7571 The default is @option{-fno-unsafe-math-optimizations}.
7573 @item -fassociative-math
7574 @opindex fassociative-math
7576 Allow re-association of operands in series of floating-point operations.
7577 This violates the ISO C and C++ language standard by possibly changing
7578 computation result. NOTE: re-ordering may change the sign of zero as
7579 well as ignore NaNs and inhibit or create underflow or overflow (and
7580 thus cannot be used on a code which relies on rounding behavior like
7581 @code{(x + 2**52) - 2**52)}. May also reorder floating-point comparisons
7582 and thus may not be used when ordered comparisons are required.
7583 This option requires that both @option{-fno-signed-zeros} and
7584 @option{-fno-trapping-math} be in effect. Moreover, it doesn't make
7585 much sense with @option{-frounding-math}. For Fortran the option
7586 is automatically enabled when both @option{-fno-signed-zeros} and
7587 @option{-fno-trapping-math} are in effect.
7589 The default is @option{-fno-associative-math}.
7591 @item -freciprocal-math
7592 @opindex freciprocal-math
7594 Allow the reciprocal of a value to be used instead of dividing by
7595 the value if this enables optimizations. For example @code{x / y}
7596 can be replaced with @code{x * (1/y)} which is useful if @code{(1/y)}
7597 is subject to common subexpression elimination. Note that this loses
7598 precision and increases the number of flops operating on the value.
7600 The default is @option{-fno-reciprocal-math}.
7602 @item -ffinite-math-only
7603 @opindex ffinite-math-only
7604 Allow optimizations for floating-point arithmetic that assume
7605 that arguments and results are not NaNs or +-Infs.
7607 This option is not turned on by any @option{-O} option since
7608 it can result in incorrect output for programs which depend on
7609 an exact implementation of IEEE or ISO rules/specifications for
7610 math functions. It may, however, yield faster code for programs
7611 that do not require the guarantees of these specifications.
7613 The default is @option{-fno-finite-math-only}.
7615 @item -fno-signed-zeros
7616 @opindex fno-signed-zeros
7617 Allow optimizations for floating point arithmetic that ignore the
7618 signedness of zero. IEEE arithmetic specifies the behavior of
7619 distinct +0.0 and @minus{}0.0 values, which then prohibits simplification
7620 of expressions such as x+0.0 or 0.0*x (even with @option{-ffinite-math-only}).
7621 This option implies that the sign of a zero result isn't significant.
7623 The default is @option{-fsigned-zeros}.
7625 @item -fno-trapping-math
7626 @opindex fno-trapping-math
7627 Compile code assuming that floating-point operations cannot generate
7628 user-visible traps. These traps include division by zero, overflow,
7629 underflow, inexact result and invalid operation. This option requires
7630 that @option{-fno-signaling-nans} be in effect. Setting this option may
7631 allow faster code if one relies on ``non-stop'' IEEE arithmetic, for example.
7633 This option should never be turned on by any @option{-O} option since
7634 it can result in incorrect output for programs which depend on
7635 an exact implementation of IEEE or ISO rules/specifications for
7638 The default is @option{-ftrapping-math}.
7640 @item -frounding-math
7641 @opindex frounding-math
7642 Disable transformations and optimizations that assume default floating
7643 point rounding behavior. This is round-to-zero for all floating point
7644 to integer conversions, and round-to-nearest for all other arithmetic
7645 truncations. This option should be specified for programs that change
7646 the FP rounding mode dynamically, or that may be executed with a
7647 non-default rounding mode. This option disables constant folding of
7648 floating point expressions at compile-time (which may be affected by
7649 rounding mode) and arithmetic transformations that are unsafe in the
7650 presence of sign-dependent rounding modes.
7652 The default is @option{-fno-rounding-math}.
7654 This option is experimental and does not currently guarantee to
7655 disable all GCC optimizations that are affected by rounding mode.
7656 Future versions of GCC may provide finer control of this setting
7657 using C99's @code{FENV_ACCESS} pragma. This command line option
7658 will be used to specify the default state for @code{FENV_ACCESS}.
7660 @item -fsignaling-nans
7661 @opindex fsignaling-nans
7662 Compile code assuming that IEEE signaling NaNs may generate user-visible
7663 traps during floating-point operations. Setting this option disables
7664 optimizations that may change the number of exceptions visible with
7665 signaling NaNs. This option implies @option{-ftrapping-math}.
7667 This option causes the preprocessor macro @code{__SUPPORT_SNAN__} to
7670 The default is @option{-fno-signaling-nans}.
7672 This option is experimental and does not currently guarantee to
7673 disable all GCC optimizations that affect signaling NaN behavior.
7675 @item -fsingle-precision-constant
7676 @opindex fsingle-precision-constant
7677 Treat floating point constant as single precision constant instead of
7678 implicitly converting it to double precision constant.
7680 @item -fcx-limited-range
7681 @opindex fcx-limited-range
7682 When enabled, this option states that a range reduction step is not
7683 needed when performing complex division. Also, there is no checking
7684 whether the result of a complex multiplication or division is @code{NaN
7685 + I*NaN}, with an attempt to rescue the situation in that case. The
7686 default is @option{-fno-cx-limited-range}, but is enabled by
7687 @option{-ffast-math}.
7689 This option controls the default setting of the ISO C99
7690 @code{CX_LIMITED_RANGE} pragma. Nevertheless, the option applies to
7693 @item -fcx-fortran-rules
7694 @opindex fcx-fortran-rules
7695 Complex multiplication and division follow Fortran rules. Range
7696 reduction is done as part of complex division, but there is no checking
7697 whether the result of a complex multiplication or division is @code{NaN
7698 + I*NaN}, with an attempt to rescue the situation in that case.
7700 The default is @option{-fno-cx-fortran-rules}.
7704 The following options control optimizations that may improve
7705 performance, but are not enabled by any @option{-O} options. This
7706 section includes experimental options that may produce broken code.
7709 @item -fbranch-probabilities
7710 @opindex fbranch-probabilities
7711 After running a program compiled with @option{-fprofile-arcs}
7712 (@pxref{Debugging Options,, Options for Debugging Your Program or
7713 @command{gcc}}), you can compile it a second time using
7714 @option{-fbranch-probabilities}, to improve optimizations based on
7715 the number of times each branch was taken. When the program
7716 compiled with @option{-fprofile-arcs} exits it saves arc execution
7717 counts to a file called @file{@var{sourcename}.gcda} for each source
7718 file. The information in this data file is very dependent on the
7719 structure of the generated code, so you must use the same source code
7720 and the same optimization options for both compilations.
7722 With @option{-fbranch-probabilities}, GCC puts a
7723 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
7724 These can be used to improve optimization. Currently, they are only
7725 used in one place: in @file{reorg.c}, instead of guessing which path a
7726 branch is mostly to take, the @samp{REG_BR_PROB} values are used to
7727 exactly determine which path is taken more often.
7729 @item -fprofile-values
7730 @opindex fprofile-values
7731 If combined with @option{-fprofile-arcs}, it adds code so that some
7732 data about values of expressions in the program is gathered.
7734 With @option{-fbranch-probabilities}, it reads back the data gathered
7735 from profiling values of expressions and adds @samp{REG_VALUE_PROFILE}
7736 notes to instructions for their later usage in optimizations.
7738 Enabled with @option{-fprofile-generate} and @option{-fprofile-use}.
7742 If combined with @option{-fprofile-arcs}, it instructs the compiler to add
7743 a code to gather information about values of expressions.
7745 With @option{-fbranch-probabilities}, it reads back the data gathered
7746 and actually performs the optimizations based on them.
7747 Currently the optimizations include specialization of division operation
7748 using the knowledge about the value of the denominator.
7750 @item -frename-registers
7751 @opindex frename-registers
7752 Attempt to avoid false dependencies in scheduled code by making use
7753 of registers left over after register allocation. This optimization
7754 will most benefit processors with lots of registers. Depending on the
7755 debug information format adopted by the target, however, it can
7756 make debugging impossible, since variables will no longer stay in
7757 a ``home register''.
7759 Enabled by default with @option{-funroll-loops} and @option{-fpeel-loops}.
7763 Perform tail duplication to enlarge superblock size. This transformation
7764 simplifies the control flow of the function allowing other optimizations to do
7767 Enabled with @option{-fprofile-use}.
7769 @item -funroll-loops
7770 @opindex funroll-loops
7771 Unroll loops whose number of iterations can be determined at compile time or
7772 upon entry to the loop. @option{-funroll-loops} implies
7773 @option{-frerun-cse-after-loop}, @option{-fweb} and @option{-frename-registers}.
7774 It also turns on complete loop peeling (i.e.@: complete removal of loops with
7775 small constant number of iterations). This option makes code larger, and may
7776 or may not make it run faster.
7778 Enabled with @option{-fprofile-use}.
7780 @item -funroll-all-loops
7781 @opindex funroll-all-loops
7782 Unroll all loops, even if their number of iterations is uncertain when
7783 the loop is entered. This usually makes programs run more slowly.
7784 @option{-funroll-all-loops} implies the same options as
7785 @option{-funroll-loops}.
7788 @opindex fpeel-loops
7789 Peels the loops for that there is enough information that they do not
7790 roll much (from profile feedback). It also turns on complete loop peeling
7791 (i.e.@: complete removal of loops with small constant number of iterations).
7793 Enabled with @option{-fprofile-use}.
7795 @item -fmove-loop-invariants
7796 @opindex fmove-loop-invariants
7797 Enables the loop invariant motion pass in the RTL loop optimizer. Enabled
7798 at level @option{-O1}
7800 @item -funswitch-loops
7801 @opindex funswitch-loops
7802 Move branches with loop invariant conditions out of the loop, with duplicates
7803 of the loop on both branches (modified according to result of the condition).
7805 @item -ffunction-sections
7806 @itemx -fdata-sections
7807 @opindex ffunction-sections
7808 @opindex fdata-sections
7809 Place each function or data item into its own section in the output
7810 file if the target supports arbitrary sections. The name of the
7811 function or the name of the data item determines the section's name
7814 Use these options on systems where the linker can perform optimizations
7815 to improve locality of reference in the instruction space. Most systems
7816 using the ELF object format and SPARC processors running Solaris 2 have
7817 linkers with such optimizations. AIX may have these optimizations in
7820 Only use these options when there are significant benefits from doing
7821 so. When you specify these options, the assembler and linker will
7822 create larger object and executable files and will also be slower.
7823 You will not be able to use @code{gprof} on all systems if you
7824 specify this option and you may have problems with debugging if
7825 you specify both this option and @option{-g}.
7827 @item -fbranch-target-load-optimize
7828 @opindex fbranch-target-load-optimize
7829 Perform branch target register load optimization before prologue / epilogue
7831 The use of target registers can typically be exposed only during reload,
7832 thus hoisting loads out of loops and doing inter-block scheduling needs
7833 a separate optimization pass.
7835 @item -fbranch-target-load-optimize2
7836 @opindex fbranch-target-load-optimize2
7837 Perform branch target register load optimization after prologue / epilogue
7840 @item -fbtr-bb-exclusive
7841 @opindex fbtr-bb-exclusive
7842 When performing branch target register load optimization, don't reuse
7843 branch target registers in within any basic block.
7845 @item -fstack-protector
7846 @opindex fstack-protector
7847 Emit extra code to check for buffer overflows, such as stack smashing
7848 attacks. This is done by adding a guard variable to functions with
7849 vulnerable objects. This includes functions that call alloca, and
7850 functions with buffers larger than 8 bytes. The guards are initialized
7851 when a function is entered and then checked when the function exits.
7852 If a guard check fails, an error message is printed and the program exits.
7854 @item -fstack-protector-all
7855 @opindex fstack-protector-all
7856 Like @option{-fstack-protector} except that all functions are protected.
7858 @item -fsection-anchors
7859 @opindex fsection-anchors
7860 Try to reduce the number of symbolic address calculations by using
7861 shared ``anchor'' symbols to address nearby objects. This transformation
7862 can help to reduce the number of GOT entries and GOT accesses on some
7865 For example, the implementation of the following function @code{foo}:
7869 int foo (void) @{ return a + b + c; @}
7872 would usually calculate the addresses of all three variables, but if you
7873 compile it with @option{-fsection-anchors}, it will access the variables
7874 from a common anchor point instead. The effect is similar to the
7875 following pseudocode (which isn't valid C):
7880 register int *xr = &x;
7881 return xr[&a - &x] + xr[&b - &x] + xr[&c - &x];
7885 Not all targets support this option.
7887 @item --param @var{name}=@var{value}
7889 In some places, GCC uses various constants to control the amount of
7890 optimization that is done. For example, GCC will not inline functions
7891 that contain more that a certain number of instructions. You can
7892 control some of these constants on the command-line using the
7893 @option{--param} option.
7895 The names of specific parameters, and the meaning of the values, are
7896 tied to the internals of the compiler, and are subject to change
7897 without notice in future releases.
7899 In each case, the @var{value} is an integer. The allowable choices for
7900 @var{name} are given in the following table:
7903 @item struct-reorg-cold-struct-ratio
7904 The threshold ratio (as a percentage) between a structure frequency
7905 and the frequency of the hottest structure in the program. This parameter
7906 is used by struct-reorg optimization enabled by @option{-fipa-struct-reorg}.
7907 We say that if the ratio of a structure frequency, calculated by profiling,
7908 to the hottest structure frequency in the program is less than this
7909 parameter, then structure reorganization is not applied to this structure.
7912 @item predictable-branch-outcome
7913 When branch is predicted to be taken with probability lower than this threshold
7914 (in percent), then it is considered well predictable. The default is 10.
7916 @item max-crossjump-edges
7917 The maximum number of incoming edges to consider for crossjumping.
7918 The algorithm used by @option{-fcrossjumping} is @math{O(N^2)} in
7919 the number of edges incoming to each block. Increasing values mean
7920 more aggressive optimization, making the compile time increase with
7921 probably small improvement in executable size.
7923 @item min-crossjump-insns
7924 The minimum number of instructions which must be matched at the end
7925 of two blocks before crossjumping will be performed on them. This
7926 value is ignored in the case where all instructions in the block being
7927 crossjumped from are matched. The default value is 5.
7929 @item max-grow-copy-bb-insns
7930 The maximum code size expansion factor when copying basic blocks
7931 instead of jumping. The expansion is relative to a jump instruction.
7932 The default value is 8.
7934 @item max-goto-duplication-insns
7935 The maximum number of instructions to duplicate to a block that jumps
7936 to a computed goto. To avoid @math{O(N^2)} behavior in a number of
7937 passes, GCC factors computed gotos early in the compilation process,
7938 and unfactors them as late as possible. Only computed jumps at the
7939 end of a basic blocks with no more than max-goto-duplication-insns are
7940 unfactored. The default value is 8.
7942 @item max-delay-slot-insn-search
7943 The maximum number of instructions to consider when looking for an
7944 instruction to fill a delay slot. If more than this arbitrary number of
7945 instructions is searched, the time savings from filling the delay slot
7946 will be minimal so stop searching. Increasing values mean more
7947 aggressive optimization, making the compile time increase with probably
7948 small improvement in executable run time.
7950 @item max-delay-slot-live-search
7951 When trying to fill delay slots, the maximum number of instructions to
7952 consider when searching for a block with valid live register
7953 information. Increasing this arbitrarily chosen value means more
7954 aggressive optimization, increasing the compile time. This parameter
7955 should be removed when the delay slot code is rewritten to maintain the
7958 @item max-gcse-memory
7959 The approximate maximum amount of memory that will be allocated in
7960 order to perform the global common subexpression elimination
7961 optimization. If more memory than specified is required, the
7962 optimization will not be done.
7964 @item max-pending-list-length
7965 The maximum number of pending dependencies scheduling will allow
7966 before flushing the current state and starting over. Large functions
7967 with few branches or calls can create excessively large lists which
7968 needlessly consume memory and resources.
7970 @item max-inline-insns-single
7971 Several parameters control the tree inliner used in gcc.
7972 This number sets the maximum number of instructions (counted in GCC's
7973 internal representation) in a single function that the tree inliner
7974 will consider for inlining. This only affects functions declared
7975 inline and methods implemented in a class declaration (C++).
7976 The default value is 300.
7978 @item max-inline-insns-auto
7979 When you use @option{-finline-functions} (included in @option{-O3}),
7980 a lot of functions that would otherwise not be considered for inlining
7981 by the compiler will be investigated. To those functions, a different
7982 (more restrictive) limit compared to functions declared inline can
7984 The default value is 50.
7986 @item large-function-insns
7987 The limit specifying really large functions. For functions larger than this
7988 limit after inlining, inlining is constrained by
7989 @option{--param large-function-growth}. This parameter is useful primarily
7990 to avoid extreme compilation time caused by non-linear algorithms used by the
7992 The default value is 2700.
7994 @item large-function-growth
7995 Specifies maximal growth of large function caused by inlining in percents.
7996 The default value is 100 which limits large function growth to 2.0 times
7999 @item large-unit-insns
8000 The limit specifying large translation unit. Growth caused by inlining of
8001 units larger than this limit is limited by @option{--param inline-unit-growth}.
8002 For small units this might be too tight (consider unit consisting of function A
8003 that is inline and B that just calls A three time. If B is small relative to
8004 A, the growth of unit is 300\% and yet such inlining is very sane. For very
8005 large units consisting of small inlineable functions however the overall unit
8006 growth limit is needed to avoid exponential explosion of code size. Thus for
8007 smaller units, the size is increased to @option{--param large-unit-insns}
8008 before applying @option{--param inline-unit-growth}. The default is 10000
8010 @item inline-unit-growth
8011 Specifies maximal overall growth of the compilation unit caused by inlining.
8012 The default value is 30 which limits unit growth to 1.3 times the original
8015 @item ipcp-unit-growth
8016 Specifies maximal overall growth of the compilation unit caused by
8017 interprocedural constant propagation. The default value is 10 which limits
8018 unit growth to 1.1 times the original size.
8020 @item large-stack-frame
8021 The limit specifying large stack frames. While inlining the algorithm is trying
8022 to not grow past this limit too much. Default value is 256 bytes.
8024 @item large-stack-frame-growth
8025 Specifies maximal growth of large stack frames caused by inlining in percents.
8026 The default value is 1000 which limits large stack frame growth to 11 times
8029 @item max-inline-insns-recursive
8030 @itemx max-inline-insns-recursive-auto
8031 Specifies maximum number of instructions out-of-line copy of self recursive inline
8032 function can grow into by performing recursive inlining.
8034 For functions declared inline @option{--param max-inline-insns-recursive} is
8035 taken into account. For function not declared inline, recursive inlining
8036 happens only when @option{-finline-functions} (included in @option{-O3}) is
8037 enabled and @option{--param max-inline-insns-recursive-auto} is used. The
8038 default value is 450.
8040 @item max-inline-recursive-depth
8041 @itemx max-inline-recursive-depth-auto
8042 Specifies maximum recursion depth used by the recursive inlining.
8044 For functions declared inline @option{--param max-inline-recursive-depth} is
8045 taken into account. For function not declared inline, recursive inlining
8046 happens only when @option{-finline-functions} (included in @option{-O3}) is
8047 enabled and @option{--param max-inline-recursive-depth-auto} is used. The
8050 @item min-inline-recursive-probability
8051 Recursive inlining is profitable only for function having deep recursion
8052 in average and can hurt for function having little recursion depth by
8053 increasing the prologue size or complexity of function body to other
8056 When profile feedback is available (see @option{-fprofile-generate}) the actual
8057 recursion depth can be guessed from probability that function will recurse via
8058 given call expression. This parameter limits inlining only to call expression
8059 whose probability exceeds given threshold (in percents). The default value is
8062 @item early-inlining-insns
8063 Specify growth that early inliner can make. In effect it increases amount of
8064 inlining for code having large abstraction penalty. The default value is 8.
8066 @item max-early-inliner-iterations
8067 @itemx max-early-inliner-iterations
8068 Limit of iterations of early inliner. This basically bounds number of nested
8069 indirect calls early inliner can resolve. Deeper chains are still handled by
8072 @item min-vect-loop-bound
8073 The minimum number of iterations under which a loop will not get vectorized
8074 when @option{-ftree-vectorize} is used. The number of iterations after
8075 vectorization needs to be greater than the value specified by this option
8076 to allow vectorization. The default value is 0.
8078 @item max-unrolled-insns
8079 The maximum number of instructions that a loop should have if that loop
8080 is unrolled, and if the loop is unrolled, it determines how many times
8081 the loop code is unrolled.
8083 @item max-average-unrolled-insns
8084 The maximum number of instructions biased by probabilities of their execution
8085 that a loop should have if that loop is unrolled, and if the loop is unrolled,
8086 it determines how many times the loop code is unrolled.
8088 @item max-unroll-times
8089 The maximum number of unrollings of a single loop.
8091 @item max-peeled-insns
8092 The maximum number of instructions that a loop should have if that loop
8093 is peeled, and if the loop is peeled, it determines how many times
8094 the loop code is peeled.
8096 @item max-peel-times
8097 The maximum number of peelings of a single loop.
8099 @item max-completely-peeled-insns
8100 The maximum number of insns of a completely peeled loop.
8102 @item max-completely-peel-times
8103 The maximum number of iterations of a loop to be suitable for complete peeling.
8105 @item max-completely-peel-loop-nest-depth
8106 The maximum depth of a loop nest suitable for complete peeling.
8108 @item max-unswitch-insns
8109 The maximum number of insns of an unswitched loop.
8111 @item max-unswitch-level
8112 The maximum number of branches unswitched in a single loop.
8115 The minimum cost of an expensive expression in the loop invariant motion.
8117 @item iv-consider-all-candidates-bound
8118 Bound on number of candidates for induction variables below that
8119 all candidates are considered for each use in induction variable
8120 optimizations. Only the most relevant candidates are considered
8121 if there are more candidates, to avoid quadratic time complexity.
8123 @item iv-max-considered-uses
8124 The induction variable optimizations give up on loops that contain more
8125 induction variable uses.
8127 @item iv-always-prune-cand-set-bound
8128 If number of candidates in the set is smaller than this value,
8129 we always try to remove unnecessary ivs from the set during its
8130 optimization when a new iv is added to the set.
8132 @item scev-max-expr-size
8133 Bound on size of expressions used in the scalar evolutions analyzer.
8134 Large expressions slow the analyzer.
8136 @item omega-max-vars
8137 The maximum number of variables in an Omega constraint system.
8138 The default value is 128.
8140 @item omega-max-geqs
8141 The maximum number of inequalities in an Omega constraint system.
8142 The default value is 256.
8145 The maximum number of equalities in an Omega constraint system.
8146 The default value is 128.
8148 @item omega-max-wild-cards
8149 The maximum number of wildcard variables that the Omega solver will
8150 be able to insert. The default value is 18.
8152 @item omega-hash-table-size
8153 The size of the hash table in the Omega solver. The default value is
8156 @item omega-max-keys
8157 The maximal number of keys used by the Omega solver. The default
8160 @item omega-eliminate-redundant-constraints
8161 When set to 1, use expensive methods to eliminate all redundant
8162 constraints. The default value is 0.
8164 @item vect-max-version-for-alignment-checks
8165 The maximum number of runtime checks that can be performed when
8166 doing loop versioning for alignment in the vectorizer. See option
8167 ftree-vect-loop-version for more information.
8169 @item vect-max-version-for-alias-checks
8170 The maximum number of runtime checks that can be performed when
8171 doing loop versioning for alias in the vectorizer. See option
8172 ftree-vect-loop-version for more information.
8174 @item max-iterations-to-track
8176 The maximum number of iterations of a loop the brute force algorithm
8177 for analysis of # of iterations of the loop tries to evaluate.
8179 @item hot-bb-count-fraction
8180 Select fraction of the maximal count of repetitions of basic block in program
8181 given basic block needs to have to be considered hot.
8183 @item hot-bb-frequency-fraction
8184 Select fraction of the maximal frequency of executions of basic block in
8185 function given basic block needs to have to be considered hot
8187 @item max-predicted-iterations
8188 The maximum number of loop iterations we predict statically. This is useful
8189 in cases where function contain single loop with known bound and other loop
8190 with unknown. We predict the known number of iterations correctly, while
8191 the unknown number of iterations average to roughly 10. This means that the
8192 loop without bounds would appear artificially cold relative to the other one.
8194 @item align-threshold
8196 Select fraction of the maximal frequency of executions of basic block in
8197 function given basic block will get aligned.
8199 @item align-loop-iterations
8201 A loop expected to iterate at lest the selected number of iterations will get
8204 @item tracer-dynamic-coverage
8205 @itemx tracer-dynamic-coverage-feedback
8207 This value is used to limit superblock formation once the given percentage of
8208 executed instructions is covered. This limits unnecessary code size
8211 The @option{tracer-dynamic-coverage-feedback} is used only when profile
8212 feedback is available. The real profiles (as opposed to statically estimated
8213 ones) are much less balanced allowing the threshold to be larger value.
8215 @item tracer-max-code-growth
8216 Stop tail duplication once code growth has reached given percentage. This is
8217 rather hokey argument, as most of the duplicates will be eliminated later in
8218 cross jumping, so it may be set to much higher values than is the desired code
8221 @item tracer-min-branch-ratio
8223 Stop reverse growth when the reverse probability of best edge is less than this
8224 threshold (in percent).
8226 @item tracer-min-branch-ratio
8227 @itemx tracer-min-branch-ratio-feedback
8229 Stop forward growth if the best edge do have probability lower than this
8232 Similarly to @option{tracer-dynamic-coverage} two values are present, one for
8233 compilation for profile feedback and one for compilation without. The value
8234 for compilation with profile feedback needs to be more conservative (higher) in
8235 order to make tracer effective.
8237 @item max-cse-path-length
8239 Maximum number of basic blocks on path that cse considers. The default is 10.
8242 The maximum instructions CSE process before flushing. The default is 1000.
8244 @item ggc-min-expand
8246 GCC uses a garbage collector to manage its own memory allocation. This
8247 parameter specifies the minimum percentage by which the garbage
8248 collector's heap should be allowed to expand between collections.
8249 Tuning this may improve compilation speed; it has no effect on code
8252 The default is 30% + 70% * (RAM/1GB) with an upper bound of 100% when
8253 RAM >= 1GB@. If @code{getrlimit} is available, the notion of "RAM" is
8254 the smallest of actual RAM and @code{RLIMIT_DATA} or @code{RLIMIT_AS}. If
8255 GCC is not able to calculate RAM on a particular platform, the lower
8256 bound of 30% is used. Setting this parameter and
8257 @option{ggc-min-heapsize} to zero causes a full collection to occur at
8258 every opportunity. This is extremely slow, but can be useful for
8261 @item ggc-min-heapsize
8263 Minimum size of the garbage collector's heap before it begins bothering
8264 to collect garbage. The first collection occurs after the heap expands
8265 by @option{ggc-min-expand}% beyond @option{ggc-min-heapsize}. Again,
8266 tuning this may improve compilation speed, and has no effect on code
8269 The default is the smaller of RAM/8, RLIMIT_RSS, or a limit which
8270 tries to ensure that RLIMIT_DATA or RLIMIT_AS are not exceeded, but
8271 with a lower bound of 4096 (four megabytes) and an upper bound of
8272 131072 (128 megabytes). If GCC is not able to calculate RAM on a
8273 particular platform, the lower bound is used. Setting this parameter
8274 very large effectively disables garbage collection. Setting this
8275 parameter and @option{ggc-min-expand} to zero causes a full collection
8276 to occur at every opportunity.
8278 @item max-reload-search-insns
8279 The maximum number of instruction reload should look backward for equivalent
8280 register. Increasing values mean more aggressive optimization, making the
8281 compile time increase with probably slightly better performance. The default
8284 @item max-cselib-memory-locations
8285 The maximum number of memory locations cselib should take into account.
8286 Increasing values mean more aggressive optimization, making the compile time
8287 increase with probably slightly better performance. The default value is 500.
8289 @item reorder-blocks-duplicate
8290 @itemx reorder-blocks-duplicate-feedback
8292 Used by basic block reordering pass to decide whether to use unconditional
8293 branch or duplicate the code on its destination. Code is duplicated when its
8294 estimated size is smaller than this value multiplied by the estimated size of
8295 unconditional jump in the hot spots of the program.
8297 The @option{reorder-block-duplicate-feedback} is used only when profile
8298 feedback is available and may be set to higher values than
8299 @option{reorder-block-duplicate} since information about the hot spots is more
8302 @item max-sched-ready-insns
8303 The maximum number of instructions ready to be issued the scheduler should
8304 consider at any given time during the first scheduling pass. Increasing
8305 values mean more thorough searches, making the compilation time increase
8306 with probably little benefit. The default value is 100.
8308 @item max-sched-region-blocks
8309 The maximum number of blocks in a region to be considered for
8310 interblock scheduling. The default value is 10.
8312 @item max-pipeline-region-blocks
8313 The maximum number of blocks in a region to be considered for
8314 pipelining in the selective scheduler. The default value is 15.
8316 @item max-sched-region-insns
8317 The maximum number of insns in a region to be considered for
8318 interblock scheduling. The default value is 100.
8320 @item max-pipeline-region-insns
8321 The maximum number of insns in a region to be considered for
8322 pipelining in the selective scheduler. The default value is 200.
8325 The minimum probability (in percents) of reaching a source block
8326 for interblock speculative scheduling. The default value is 40.
8328 @item max-sched-extend-regions-iters
8329 The maximum number of iterations through CFG to extend regions.
8330 0 - disable region extension,
8331 N - do at most N iterations.
8332 The default value is 0.
8334 @item max-sched-insn-conflict-delay
8335 The maximum conflict delay for an insn to be considered for speculative motion.
8336 The default value is 3.
8338 @item sched-spec-prob-cutoff
8339 The minimal probability of speculation success (in percents), so that
8340 speculative insn will be scheduled.
8341 The default value is 40.
8343 @item sched-mem-true-dep-cost
8344 Minimal distance (in CPU cycles) between store and load targeting same
8345 memory locations. The default value is 1.
8347 @item selsched-max-lookahead
8348 The maximum size of the lookahead window of selective scheduling. It is a
8349 depth of search for available instructions.
8350 The default value is 50.
8352 @item selsched-max-sched-times
8353 The maximum number of times that an instruction will be scheduled during
8354 selective scheduling. This is the limit on the number of iterations
8355 through which the instruction may be pipelined. The default value is 2.
8357 @item selsched-max-insns-to-rename
8358 The maximum number of best instructions in the ready list that are considered
8359 for renaming in the selective scheduler. The default value is 2.
8361 @item max-last-value-rtl
8362 The maximum size measured as number of RTLs that can be recorded in an expression
8363 in combiner for a pseudo register as last known value of that register. The default
8366 @item integer-share-limit
8367 Small integer constants can use a shared data structure, reducing the
8368 compiler's memory usage and increasing its speed. This sets the maximum
8369 value of a shared integer constant. The default value is 256.
8371 @item min-virtual-mappings
8372 Specifies the minimum number of virtual mappings in the incremental
8373 SSA updater that should be registered to trigger the virtual mappings
8374 heuristic defined by virtual-mappings-ratio. The default value is
8377 @item virtual-mappings-ratio
8378 If the number of virtual mappings is virtual-mappings-ratio bigger
8379 than the number of virtual symbols to be updated, then the incremental
8380 SSA updater switches to a full update for those symbols. The default
8383 @item ssp-buffer-size
8384 The minimum size of buffers (i.e.@: arrays) that will receive stack smashing
8385 protection when @option{-fstack-protection} is used.
8387 @item max-jump-thread-duplication-stmts
8388 Maximum number of statements allowed in a block that needs to be
8389 duplicated when threading jumps.
8391 @item max-fields-for-field-sensitive
8392 Maximum number of fields in a structure we will treat in
8393 a field sensitive manner during pointer analysis. The default is zero
8394 for -O0, and -O1 and 100 for -Os, -O2, and -O3.
8396 @item prefetch-latency
8397 Estimate on average number of instructions that are executed before
8398 prefetch finishes. The distance we prefetch ahead is proportional
8399 to this constant. Increasing this number may also lead to less
8400 streams being prefetched (see @option{simultaneous-prefetches}).
8402 @item simultaneous-prefetches
8403 Maximum number of prefetches that can run at the same time.
8405 @item l1-cache-line-size
8406 The size of cache line in L1 cache, in bytes.
8409 The size of L1 cache, in kilobytes.
8412 The size of L2 cache, in kilobytes.
8414 @item min-insn-to-prefetch-ratio
8415 The minimum ratio between the number of instructions and the
8416 number of prefetches to enable prefetching in a loop with an
8419 @item prefetch-min-insn-to-mem-ratio
8420 The minimum ratio between the number of instructions and the
8421 number of memory references to enable prefetching in a loop.
8423 @item use-canonical-types
8424 Whether the compiler should use the ``canonical'' type system. By
8425 default, this should always be 1, which uses a more efficient internal
8426 mechanism for comparing types in C++ and Objective-C++. However, if
8427 bugs in the canonical type system are causing compilation failures,
8428 set this value to 0 to disable canonical types.
8430 @item switch-conversion-max-branch-ratio
8431 Switch initialization conversion will refuse to create arrays that are
8432 bigger than @option{switch-conversion-max-branch-ratio} times the number of
8433 branches in the switch.
8435 @item max-partial-antic-length
8436 Maximum length of the partial antic set computed during the tree
8437 partial redundancy elimination optimization (@option{-ftree-pre}) when
8438 optimizing at @option{-O3} and above. For some sorts of source code
8439 the enhanced partial redundancy elimination optimization can run away,
8440 consuming all of the memory available on the host machine. This
8441 parameter sets a limit on the length of the sets that are computed,
8442 which prevents the runaway behavior. Setting a value of 0 for
8443 this parameter will allow an unlimited set length.
8445 @item sccvn-max-scc-size
8446 Maximum size of a strongly connected component (SCC) during SCCVN
8447 processing. If this limit is hit, SCCVN processing for the whole
8448 function will not be done and optimizations depending on it will
8449 be disabled. The default maximum SCC size is 10000.
8451 @item ira-max-loops-num
8452 IRA uses a regional register allocation by default. If a function
8453 contains loops more than number given by the parameter, only at most
8454 given number of the most frequently executed loops will form regions
8455 for the regional register allocation. The default value of the
8458 @item ira-max-conflict-table-size
8459 Although IRA uses a sophisticated algorithm of compression conflict
8460 table, the table can be still big for huge functions. If the conflict
8461 table for a function could be more than size in MB given by the
8462 parameter, the conflict table is not built and faster, simpler, and
8463 lower quality register allocation algorithm will be used. The
8464 algorithm do not use pseudo-register conflicts. The default value of
8465 the parameter is 2000.
8467 @item ira-loop-reserved-regs
8468 IRA can be used to evaluate more accurate register pressure in loops
8469 for decision to move loop invariants (see @option{-O3}). The number
8470 of available registers reserved for some other purposes is described
8471 by this parameter. The default value of the parameter is 2 which is
8472 minimal number of registers needed for execution of typical
8473 instruction. This value is the best found from numerous experiments.
8475 @item loop-invariant-max-bbs-in-loop
8476 Loop invariant motion can be very expensive, both in compile time and
8477 in amount of needed compile time memory, with very large loops. Loops
8478 with more basic blocks than this parameter won't have loop invariant
8479 motion optimization performed on them. The default value of the
8480 parameter is 1000 for -O1 and 10000 for -O2 and above.
8482 @item max-vartrack-size
8483 Sets a maximum number of hash table slots to use during variable
8484 tracking dataflow analysis of any function. If this limit is exceeded
8485 with variable tracking at assignments enabled, analysis for that
8486 function is retried without it, after removing all debug insns from
8487 the function. If the limit is exceeded even without debug insns, var
8488 tracking analysis is completely disabled for the function. Setting
8489 the parameter to zero makes it unlimited.
8491 @item min-nondebug-insn-uid
8492 Use uids starting at this parameter for nondebug insns. The range below
8493 the parameter is reserved exclusively for debug insns created by
8494 @option{-fvar-tracking-assignments}, but debug insns may get
8495 (non-overlapping) uids above it if the reserved range is exhausted.
8497 @item ipa-sra-ptr-growth-factor
8498 IPA-SRA will replace a pointer to an aggregate with one or more new
8499 parameters only when their cumulative size is less or equal to
8500 @option{ipa-sra-ptr-growth-factor} times the size of the original
8503 @item graphite-max-nb-scop-params
8504 To avoid exponential effects in the Graphite loop transforms, the
8505 number of parameters in a Static Control Part (SCoP) is bounded. The
8506 default value is 10 parameters. A variable whose value is unknown at
8507 compile time and defined outside a SCoP is a parameter of the SCoP.
8509 @item graphite-max-bbs-per-function
8510 To avoid exponential effects in the detection of SCoPs, the size of
8511 the functions analyzed by Graphite is bounded. The default value is
8514 @item loop-block-tile-size
8515 Loop blocking or strip mining transforms, enabled with
8516 @option{-floop-block} or @option{-floop-strip-mine}, strip mine each
8517 loop in the loop nest by a given number of iterations. The strip
8518 length can be changed using the @option{loop-block-tile-size}
8519 parameter. The default value is 51 iterations.
8524 @node Preprocessor Options
8525 @section Options Controlling the Preprocessor
8526 @cindex preprocessor options
8527 @cindex options, preprocessor
8529 These options control the C preprocessor, which is run on each C source
8530 file before actual compilation.
8532 If you use the @option{-E} option, nothing is done except preprocessing.
8533 Some of these options make sense only together with @option{-E} because
8534 they cause the preprocessor output to be unsuitable for actual
8538 @item -Wp,@var{option}
8540 You can use @option{-Wp,@var{option}} to bypass the compiler driver
8541 and pass @var{option} directly through to the preprocessor. If
8542 @var{option} contains commas, it is split into multiple options at the
8543 commas. However, many options are modified, translated or interpreted
8544 by the compiler driver before being passed to the preprocessor, and
8545 @option{-Wp} forcibly bypasses this phase. The preprocessor's direct
8546 interface is undocumented and subject to change, so whenever possible
8547 you should avoid using @option{-Wp} and let the driver handle the
8550 @item -Xpreprocessor @var{option}
8551 @opindex Xpreprocessor
8552 Pass @var{option} as an option to the preprocessor. You can use this to
8553 supply system-specific preprocessor options which GCC does not know how to
8556 If you want to pass an option that takes an argument, you must use
8557 @option{-Xpreprocessor} twice, once for the option and once for the argument.
8560 @include cppopts.texi
8562 @node Assembler Options
8563 @section Passing Options to the Assembler
8565 @c prevent bad page break with this line
8566 You can pass options to the assembler.
8569 @item -Wa,@var{option}
8571 Pass @var{option} as an option to the assembler. If @var{option}
8572 contains commas, it is split into multiple options at the commas.
8574 @item -Xassembler @var{option}
8576 Pass @var{option} as an option to the assembler. You can use this to
8577 supply system-specific assembler options which GCC does not know how to
8580 If you want to pass an option that takes an argument, you must use
8581 @option{-Xassembler} twice, once for the option and once for the argument.
8586 @section Options for Linking
8587 @cindex link options
8588 @cindex options, linking
8590 These options come into play when the compiler links object files into
8591 an executable output file. They are meaningless if the compiler is
8592 not doing a link step.
8596 @item @var{object-file-name}
8597 A file name that does not end in a special recognized suffix is
8598 considered to name an object file or library. (Object files are
8599 distinguished from libraries by the linker according to the file
8600 contents.) If linking is done, these object files are used as input
8609 If any of these options is used, then the linker is not run, and
8610 object file names should not be used as arguments. @xref{Overall
8614 @item -l@var{library}
8615 @itemx -l @var{library}
8617 Search the library named @var{library} when linking. (The second
8618 alternative with the library as a separate argument is only for
8619 POSIX compliance and is not recommended.)
8621 It makes a difference where in the command you write this option; the
8622 linker searches and processes libraries and object files in the order they
8623 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
8624 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
8625 to functions in @samp{z}, those functions may not be loaded.
8627 The linker searches a standard list of directories for the library,
8628 which is actually a file named @file{lib@var{library}.a}. The linker
8629 then uses this file as if it had been specified precisely by name.
8631 The directories searched include several standard system directories
8632 plus any that you specify with @option{-L}.
8634 Normally the files found this way are library files---archive files
8635 whose members are object files. The linker handles an archive file by
8636 scanning through it for members which define symbols that have so far
8637 been referenced but not defined. But if the file that is found is an
8638 ordinary object file, it is linked in the usual fashion. The only
8639 difference between using an @option{-l} option and specifying a file name
8640 is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
8641 and searches several directories.
8645 You need this special case of the @option{-l} option in order to
8646 link an Objective-C or Objective-C++ program.
8649 @opindex nostartfiles
8650 Do not use the standard system startup files when linking.
8651 The standard system libraries are used normally, unless @option{-nostdlib}
8652 or @option{-nodefaultlibs} is used.
8654 @item -nodefaultlibs
8655 @opindex nodefaultlibs
8656 Do not use the standard system libraries when linking.
8657 Only the libraries you specify will be passed to the linker, options
8658 specifying linkage of the system libraries, such as @code{-static-libgcc}
8659 or @code{-shared-libgcc}, will be ignored.
8660 The standard startup files are used normally, unless @option{-nostartfiles}
8661 is used. The compiler may generate calls to @code{memcmp},
8662 @code{memset}, @code{memcpy} and @code{memmove}.
8663 These entries are usually resolved by entries in
8664 libc. These entry points should be supplied through some other
8665 mechanism when this option is specified.
8669 Do not use the standard system startup files or libraries when linking.
8670 No startup files and only the libraries you specify will be passed to
8671 the linker, options specifying linkage of the system libraries, such as
8672 @code{-static-libgcc} or @code{-shared-libgcc}, will be ignored.
8673 The compiler may generate calls to @code{memcmp}, @code{memset},
8674 @code{memcpy} and @code{memmove}.
8675 These entries are usually resolved by entries in
8676 libc. These entry points should be supplied through some other
8677 mechanism when this option is specified.
8679 @cindex @option{-lgcc}, use with @option{-nostdlib}
8680 @cindex @option{-nostdlib} and unresolved references
8681 @cindex unresolved references and @option{-nostdlib}
8682 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
8683 @cindex @option{-nodefaultlibs} and unresolved references
8684 @cindex unresolved references and @option{-nodefaultlibs}
8685 One of the standard libraries bypassed by @option{-nostdlib} and
8686 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
8687 that GCC uses to overcome shortcomings of particular machines, or special
8688 needs for some languages.
8689 (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler
8690 Collection (GCC) Internals},
8691 for more discussion of @file{libgcc.a}.)
8692 In most cases, you need @file{libgcc.a} even when you want to avoid
8693 other standard libraries. In other words, when you specify @option{-nostdlib}
8694 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
8695 This ensures that you have no unresolved references to internal GCC
8696 library subroutines. (For example, @samp{__main}, used to ensure C++
8697 constructors will be called; @pxref{Collect2,,@code{collect2}, gccint,
8698 GNU Compiler Collection (GCC) Internals}.)
8702 Produce a position independent executable on targets which support it.
8703 For predictable results, you must also specify the same set of options
8704 that were used to generate code (@option{-fpie}, @option{-fPIE},
8705 or model suboptions) when you specify this option.
8709 Pass the flag @option{-export-dynamic} to the ELF linker, on targets
8710 that support it. This instructs the linker to add all symbols, not
8711 only used ones, to the dynamic symbol table. This option is needed
8712 for some uses of @code{dlopen} or to allow obtaining backtraces
8713 from within a program.
8717 Remove all symbol table and relocation information from the executable.
8721 On systems that support dynamic linking, this prevents linking with the shared
8722 libraries. On other systems, this option has no effect.
8726 Produce a shared object which can then be linked with other objects to
8727 form an executable. Not all systems support this option. For predictable
8728 results, you must also specify the same set of options that were used to
8729 generate code (@option{-fpic}, @option{-fPIC}, or model suboptions)
8730 when you specify this option.@footnote{On some systems, @samp{gcc -shared}
8731 needs to build supplementary stub code for constructors to work. On
8732 multi-libbed systems, @samp{gcc -shared} must select the correct support
8733 libraries to link against. Failing to supply the correct flags may lead
8734 to subtle defects. Supplying them in cases where they are not necessary
8737 @item -shared-libgcc
8738 @itemx -static-libgcc
8739 @opindex shared-libgcc
8740 @opindex static-libgcc
8741 On systems that provide @file{libgcc} as a shared library, these options
8742 force the use of either the shared or static version respectively.
8743 If no shared version of @file{libgcc} was built when the compiler was
8744 configured, these options have no effect.
8746 There are several situations in which an application should use the
8747 shared @file{libgcc} instead of the static version. The most common
8748 of these is when the application wishes to throw and catch exceptions
8749 across different shared libraries. In that case, each of the libraries
8750 as well as the application itself should use the shared @file{libgcc}.
8752 Therefore, the G++ and GCJ drivers automatically add
8753 @option{-shared-libgcc} whenever you build a shared library or a main
8754 executable, because C++ and Java programs typically use exceptions, so
8755 this is the right thing to do.
8757 If, instead, you use the GCC driver to create shared libraries, you may
8758 find that they will not always be linked with the shared @file{libgcc}.
8759 If GCC finds, at its configuration time, that you have a non-GNU linker
8760 or a GNU linker that does not support option @option{--eh-frame-hdr},
8761 it will link the shared version of @file{libgcc} into shared libraries
8762 by default. Otherwise, it will take advantage of the linker and optimize
8763 away the linking with the shared version of @file{libgcc}, linking with
8764 the static version of libgcc by default. This allows exceptions to
8765 propagate through such shared libraries, without incurring relocation
8766 costs at library load time.
8768 However, if a library or main executable is supposed to throw or catch
8769 exceptions, you must link it using the G++ or GCJ driver, as appropriate
8770 for the languages used in the program, or using the option
8771 @option{-shared-libgcc}, such that it is linked with the shared
8774 @item -static-libstdc++
8775 When the @command{g++} program is used to link a C++ program, it will
8776 normally automatically link against @option{libstdc++}. If
8777 @file{libstdc++} is available as a shared library, and the
8778 @option{-static} option is not used, then this will link against the
8779 shared version of @file{libstdc++}. That is normally fine. However, it
8780 is sometimes useful to freeze the version of @file{libstdc++} used by
8781 the program without going all the way to a fully static link. The
8782 @option{-static-libstdc++} option directs the @command{g++} driver to
8783 link @file{libstdc++} statically, without necessarily linking other
8784 libraries statically.
8788 Bind references to global symbols when building a shared object. Warn
8789 about any unresolved references (unless overridden by the link editor
8790 option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support
8793 @item -T @var{script}
8795 @cindex linker script
8796 Use @var{script} as the linker script. This option is supported by most
8797 systems using the GNU linker. On some targets, such as bare-board
8798 targets without an operating system, the @option{-T} option may be required
8799 when linking to avoid references to undefined symbols.
8801 @item -Xlinker @var{option}
8803 Pass @var{option} as an option to the linker. You can use this to
8804 supply system-specific linker options which GCC does not know how to
8807 If you want to pass an option that takes a separate argument, you must use
8808 @option{-Xlinker} twice, once for the option and once for the argument.
8809 For example, to pass @option{-assert definitions}, you must write
8810 @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write
8811 @option{-Xlinker "-assert definitions"}, because this passes the entire
8812 string as a single argument, which is not what the linker expects.
8814 When using the GNU linker, it is usually more convenient to pass
8815 arguments to linker options using the @option{@var{option}=@var{value}}
8816 syntax than as separate arguments. For example, you can specify
8817 @samp{-Xlinker -Map=output.map} rather than
8818 @samp{-Xlinker -Map -Xlinker output.map}. Other linkers may not support
8819 this syntax for command-line options.
8821 @item -Wl,@var{option}
8823 Pass @var{option} as an option to the linker. If @var{option} contains
8824 commas, it is split into multiple options at the commas. You can use this
8825 syntax to pass an argument to the option.
8826 For example, @samp{-Wl,-Map,output.map} passes @samp{-Map output.map} to the
8827 linker. When using the GNU linker, you can also get the same effect with
8828 @samp{-Wl,-Map=output.map}.
8830 @item -u @var{symbol}
8832 Pretend the symbol @var{symbol} is undefined, to force linking of
8833 library modules to define it. You can use @option{-u} multiple times with
8834 different symbols to force loading of additional library modules.
8837 @node Directory Options
8838 @section Options for Directory Search
8839 @cindex directory options
8840 @cindex options, directory search
8843 These options specify directories to search for header files, for
8844 libraries and for parts of the compiler:
8849 Add the directory @var{dir} to the head of the list of directories to be
8850 searched for header files. This can be used to override a system header
8851 file, substituting your own version, since these directories are
8852 searched before the system header file directories. However, you should
8853 not use this option to add directories that contain vendor-supplied
8854 system header files (use @option{-isystem} for that). If you use more than
8855 one @option{-I} option, the directories are scanned in left-to-right
8856 order; the standard system directories come after.
8858 If a standard system include directory, or a directory specified with
8859 @option{-isystem}, is also specified with @option{-I}, the @option{-I}
8860 option will be ignored. The directory will still be searched but as a
8861 system directory at its normal position in the system include chain.
8862 This is to ensure that GCC's procedure to fix buggy system headers and
8863 the ordering for the include_next directive are not inadvertently changed.
8864 If you really need to change the search order for system directories,
8865 use the @option{-nostdinc} and/or @option{-isystem} options.
8867 @item -iquote@var{dir}
8869 Add the directory @var{dir} to the head of the list of directories to
8870 be searched for header files only for the case of @samp{#include
8871 "@var{file}"}; they are not searched for @samp{#include <@var{file}>},
8872 otherwise just like @option{-I}.
8876 Add directory @var{dir} to the list of directories to be searched
8879 @item -B@var{prefix}
8881 This option specifies where to find the executables, libraries,
8882 include files, and data files of the compiler itself.
8884 The compiler driver program runs one or more of the subprograms
8885 @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries
8886 @var{prefix} as a prefix for each program it tries to run, both with and
8887 without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}).
8889 For each subprogram to be run, the compiler driver first tries the
8890 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
8891 was not specified, the driver tries two standard prefixes, which are
8892 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc/}. If neither of
8893 those results in a file name that is found, the unmodified program
8894 name is searched for using the directories specified in your
8895 @env{PATH} environment variable.
8897 The compiler will check to see if the path provided by the @option{-B}
8898 refers to a directory, and if necessary it will add a directory
8899 separator character at the end of the path.
8901 @option{-B} prefixes that effectively specify directory names also apply
8902 to libraries in the linker, because the compiler translates these
8903 options into @option{-L} options for the linker. They also apply to
8904 includes files in the preprocessor, because the compiler translates these
8905 options into @option{-isystem} options for the preprocessor. In this case,
8906 the compiler appends @samp{include} to the prefix.
8908 The run-time support file @file{libgcc.a} can also be searched for using
8909 the @option{-B} prefix, if needed. If it is not found there, the two
8910 standard prefixes above are tried, and that is all. The file is left
8911 out of the link if it is not found by those means.
8913 Another way to specify a prefix much like the @option{-B} prefix is to use
8914 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
8917 As a special kludge, if the path provided by @option{-B} is
8918 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
8919 9, then it will be replaced by @file{[dir/]include}. This is to help
8920 with boot-strapping the compiler.
8922 @item -specs=@var{file}
8924 Process @var{file} after the compiler reads in the standard @file{specs}
8925 file, in order to override the defaults that the @file{gcc} driver
8926 program uses when determining what switches to pass to @file{cc1},
8927 @file{cc1plus}, @file{as}, @file{ld}, etc. More than one
8928 @option{-specs=@var{file}} can be specified on the command line, and they
8929 are processed in order, from left to right.
8931 @item --sysroot=@var{dir}
8933 Use @var{dir} as the logical root directory for headers and libraries.
8934 For example, if the compiler would normally search for headers in
8935 @file{/usr/include} and libraries in @file{/usr/lib}, it will instead
8936 search @file{@var{dir}/usr/include} and @file{@var{dir}/usr/lib}.
8938 If you use both this option and the @option{-isysroot} option, then
8939 the @option{--sysroot} option will apply to libraries, but the
8940 @option{-isysroot} option will apply to header files.
8942 The GNU linker (beginning with version 2.16) has the necessary support
8943 for this option. If your linker does not support this option, the
8944 header file aspect of @option{--sysroot} will still work, but the
8945 library aspect will not.
8949 This option has been deprecated. Please use @option{-iquote} instead for
8950 @option{-I} directories before the @option{-I-} and remove the @option{-I-}.
8951 Any directories you specify with @option{-I} options before the @option{-I-}
8952 option are searched only for the case of @samp{#include "@var{file}"};
8953 they are not searched for @samp{#include <@var{file}>}.
8955 If additional directories are specified with @option{-I} options after
8956 the @option{-I-}, these directories are searched for all @samp{#include}
8957 directives. (Ordinarily @emph{all} @option{-I} directories are used
8960 In addition, the @option{-I-} option inhibits the use of the current
8961 directory (where the current input file came from) as the first search
8962 directory for @samp{#include "@var{file}"}. There is no way to
8963 override this effect of @option{-I-}. With @option{-I.} you can specify
8964 searching the directory which was current when the compiler was
8965 invoked. That is not exactly the same as what the preprocessor does
8966 by default, but it is often satisfactory.
8968 @option{-I-} does not inhibit the use of the standard system directories
8969 for header files. Thus, @option{-I-} and @option{-nostdinc} are
8976 @section Specifying subprocesses and the switches to pass to them
8979 @command{gcc} is a driver program. It performs its job by invoking a
8980 sequence of other programs to do the work of compiling, assembling and
8981 linking. GCC interprets its command-line parameters and uses these to
8982 deduce which programs it should invoke, and which command-line options
8983 it ought to place on their command lines. This behavior is controlled
8984 by @dfn{spec strings}. In most cases there is one spec string for each
8985 program that GCC can invoke, but a few programs have multiple spec
8986 strings to control their behavior. The spec strings built into GCC can
8987 be overridden by using the @option{-specs=} command-line switch to specify
8990 @dfn{Spec files} are plaintext files that are used to construct spec
8991 strings. They consist of a sequence of directives separated by blank
8992 lines. The type of directive is determined by the first non-whitespace
8993 character on the line and it can be one of the following:
8996 @item %@var{command}
8997 Issues a @var{command} to the spec file processor. The commands that can
9001 @item %include <@var{file}>
9003 Search for @var{file} and insert its text at the current point in the
9006 @item %include_noerr <@var{file}>
9007 @cindex %include_noerr
9008 Just like @samp{%include}, but do not generate an error message if the include
9009 file cannot be found.
9011 @item %rename @var{old_name} @var{new_name}
9013 Rename the spec string @var{old_name} to @var{new_name}.
9017 @item *[@var{spec_name}]:
9018 This tells the compiler to create, override or delete the named spec
9019 string. All lines after this directive up to the next directive or
9020 blank line are considered to be the text for the spec string. If this
9021 results in an empty string then the spec will be deleted. (Or, if the
9022 spec did not exist, then nothing will happened.) Otherwise, if the spec
9023 does not currently exist a new spec will be created. If the spec does
9024 exist then its contents will be overridden by the text of this
9025 directive, unless the first character of that text is the @samp{+}
9026 character, in which case the text will be appended to the spec.
9028 @item [@var{suffix}]:
9029 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
9030 and up to the next directive or blank line are considered to make up the
9031 spec string for the indicated suffix. When the compiler encounters an
9032 input file with the named suffix, it will processes the spec string in
9033 order to work out how to compile that file. For example:
9040 This says that any input file whose name ends in @samp{.ZZ} should be
9041 passed to the program @samp{z-compile}, which should be invoked with the
9042 command-line switch @option{-input} and with the result of performing the
9043 @samp{%i} substitution. (See below.)
9045 As an alternative to providing a spec string, the text that follows a
9046 suffix directive can be one of the following:
9049 @item @@@var{language}
9050 This says that the suffix is an alias for a known @var{language}. This is
9051 similar to using the @option{-x} command-line switch to GCC to specify a
9052 language explicitly. For example:
9059 Says that .ZZ files are, in fact, C++ source files.
9062 This causes an error messages saying:
9065 @var{name} compiler not installed on this system.
9069 GCC already has an extensive list of suffixes built into it.
9070 This directive will add an entry to the end of the list of suffixes, but
9071 since the list is searched from the end backwards, it is effectively
9072 possible to override earlier entries using this technique.
9076 GCC has the following spec strings built into it. Spec files can
9077 override these strings or create their own. Note that individual
9078 targets can also add their own spec strings to this list.
9081 asm Options to pass to the assembler
9082 asm_final Options to pass to the assembler post-processor
9083 cpp Options to pass to the C preprocessor
9084 cc1 Options to pass to the C compiler
9085 cc1plus Options to pass to the C++ compiler
9086 endfile Object files to include at the end of the link
9087 link Options to pass to the linker
9088 lib Libraries to include on the command line to the linker
9089 libgcc Decides which GCC support library to pass to the linker
9090 linker Sets the name of the linker
9091 predefines Defines to be passed to the C preprocessor
9092 signed_char Defines to pass to CPP to say whether @code{char} is signed
9094 startfile Object files to include at the start of the link
9097 Here is a small example of a spec file:
9103 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
9106 This example renames the spec called @samp{lib} to @samp{old_lib} and
9107 then overrides the previous definition of @samp{lib} with a new one.
9108 The new definition adds in some extra command-line options before
9109 including the text of the old definition.
9111 @dfn{Spec strings} are a list of command-line options to be passed to their
9112 corresponding program. In addition, the spec strings can contain
9113 @samp{%}-prefixed sequences to substitute variable text or to
9114 conditionally insert text into the command line. Using these constructs
9115 it is possible to generate quite complex command lines.
9117 Here is a table of all defined @samp{%}-sequences for spec
9118 strings. Note that spaces are not generated automatically around the
9119 results of expanding these sequences. Therefore you can concatenate them
9120 together or combine them with constant text in a single argument.
9124 Substitute one @samp{%} into the program name or argument.
9127 Substitute the name of the input file being processed.
9130 Substitute the basename of the input file being processed.
9131 This is the substring up to (and not including) the last period
9132 and not including the directory.
9135 This is the same as @samp{%b}, but include the file suffix (text after
9139 Marks the argument containing or following the @samp{%d} as a
9140 temporary file name, so that that file will be deleted if GCC exits
9141 successfully. Unlike @samp{%g}, this contributes no text to the
9144 @item %g@var{suffix}
9145 Substitute a file name that has suffix @var{suffix} and is chosen
9146 once per compilation, and mark the argument in the same way as
9147 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
9148 name is now chosen in a way that is hard to predict even when previously
9149 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
9150 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
9151 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
9152 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
9153 was simply substituted with a file name chosen once per compilation,
9154 without regard to any appended suffix (which was therefore treated
9155 just like ordinary text), making such attacks more likely to succeed.
9157 @item %u@var{suffix}
9158 Like @samp{%g}, but generates a new temporary file name even if
9159 @samp{%u@var{suffix}} was already seen.
9161 @item %U@var{suffix}
9162 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
9163 new one if there is no such last file name. In the absence of any
9164 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
9165 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
9166 would involve the generation of two distinct file names, one
9167 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
9168 simply substituted with a file name chosen for the previous @samp{%u},
9169 without regard to any appended suffix.
9171 @item %j@var{suffix}
9172 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
9173 writable, and if save-temps is off; otherwise, substitute the name
9174 of a temporary file, just like @samp{%u}. This temporary file is not
9175 meant for communication between processes, but rather as a junk
9178 @item %|@var{suffix}
9179 @itemx %m@var{suffix}
9180 Like @samp{%g}, except if @option{-pipe} is in effect. In that case
9181 @samp{%|} substitutes a single dash and @samp{%m} substitutes nothing at
9182 all. These are the two most common ways to instruct a program that it
9183 should read from standard input or write to standard output. If you
9184 need something more elaborate you can use an @samp{%@{pipe:@code{X}@}}
9185 construct: see for example @file{f/lang-specs.h}.
9187 @item %.@var{SUFFIX}
9188 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
9189 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
9190 terminated by the next space or %.
9193 Marks the argument containing or following the @samp{%w} as the
9194 designated output file of this compilation. This puts the argument
9195 into the sequence of arguments that @samp{%o} will substitute later.
9198 Substitutes the names of all the output files, with spaces
9199 automatically placed around them. You should write spaces
9200 around the @samp{%o} as well or the results are undefined.
9201 @samp{%o} is for use in the specs for running the linker.
9202 Input files whose names have no recognized suffix are not compiled
9203 at all, but they are included among the output files, so they will
9207 Substitutes the suffix for object files. Note that this is
9208 handled specially when it immediately follows @samp{%g, %u, or %U},
9209 because of the need for those to form complete file names. The
9210 handling is such that @samp{%O} is treated exactly as if it had already
9211 been substituted, except that @samp{%g, %u, and %U} do not currently
9212 support additional @var{suffix} characters following @samp{%O} as they would
9213 following, for example, @samp{.o}.
9216 Substitutes the standard macro predefinitions for the
9217 current target machine. Use this when running @code{cpp}.
9220 Like @samp{%p}, but puts @samp{__} before and after the name of each
9221 predefined macro, except for macros that start with @samp{__} or with
9222 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
9226 Substitute any of @option{-iprefix} (made from @env{GCC_EXEC_PREFIX}),
9227 @option{-isysroot} (made from @env{TARGET_SYSTEM_ROOT}),
9228 @option{-isystem} (made from @env{COMPILER_PATH} and @option{-B} options)
9229 and @option{-imultilib} as necessary.
9232 Current argument is the name of a library or startup file of some sort.
9233 Search for that file in a standard list of directories and substitute
9234 the full name found. The current working directory is included in the
9235 list of directories scanned.
9238 Current argument is the name of a linker script. Search for that file
9239 in the current list of directories to scan for libraries. If the file
9240 is located insert a @option{--script} option into the command line
9241 followed by the full path name found. If the file is not found then
9242 generate an error message. Note: the current working directory is not
9246 Print @var{str} as an error message. @var{str} is terminated by a newline.
9247 Use this when inconsistent options are detected.
9250 Substitute the contents of spec string @var{name} at this point.
9253 Like @samp{%(@dots{})} but put @samp{__} around @option{-D} arguments.
9255 @item %x@{@var{option}@}
9256 Accumulate an option for @samp{%X}.
9259 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
9263 Output the accumulated assembler options specified by @option{-Wa}.
9266 Output the accumulated preprocessor options specified by @option{-Wp}.
9269 Process the @code{asm} spec. This is used to compute the
9270 switches to be passed to the assembler.
9273 Process the @code{asm_final} spec. This is a spec string for
9274 passing switches to an assembler post-processor, if such a program is
9278 Process the @code{link} spec. This is the spec for computing the
9279 command line passed to the linker. Typically it will make use of the
9280 @samp{%L %G %S %D and %E} sequences.
9283 Dump out a @option{-L} option for each directory that GCC believes might
9284 contain startup files. If the target supports multilibs then the
9285 current multilib directory will be prepended to each of these paths.
9288 Process the @code{lib} spec. This is a spec string for deciding which
9289 libraries should be included on the command line to the linker.
9292 Process the @code{libgcc} spec. This is a spec string for deciding
9293 which GCC support library should be included on the command line to the linker.
9296 Process the @code{startfile} spec. This is a spec for deciding which
9297 object files should be the first ones passed to the linker. Typically
9298 this might be a file named @file{crt0.o}.
9301 Process the @code{endfile} spec. This is a spec string that specifies
9302 the last object files that will be passed to the linker.
9305 Process the @code{cpp} spec. This is used to construct the arguments
9306 to be passed to the C preprocessor.
9309 Process the @code{cc1} spec. This is used to construct the options to be
9310 passed to the actual C compiler (@samp{cc1}).
9313 Process the @code{cc1plus} spec. This is used to construct the options to be
9314 passed to the actual C++ compiler (@samp{cc1plus}).
9317 Substitute the variable part of a matched option. See below.
9318 Note that each comma in the substituted string is replaced by
9322 Remove all occurrences of @code{-S} from the command line. Note---this
9323 command is position dependent. @samp{%} commands in the spec string
9324 before this one will see @code{-S}, @samp{%} commands in the spec string
9325 after this one will not.
9327 @item %:@var{function}(@var{args})
9328 Call the named function @var{function}, passing it @var{args}.
9329 @var{args} is first processed as a nested spec string, then split
9330 into an argument vector in the usual fashion. The function returns
9331 a string which is processed as if it had appeared literally as part
9332 of the current spec.
9334 The following built-in spec functions are provided:
9338 The @code{getenv} spec function takes two arguments: an environment
9339 variable name and a string. If the environment variable is not
9340 defined, a fatal error is issued. Otherwise, the return value is the
9341 value of the environment variable concatenated with the string. For
9342 example, if @env{TOPDIR} is defined as @file{/path/to/top}, then:
9345 %:getenv(TOPDIR /include)
9348 expands to @file{/path/to/top/include}.
9350 @item @code{if-exists}
9351 The @code{if-exists} spec function takes one argument, an absolute
9352 pathname to a file. If the file exists, @code{if-exists} returns the
9353 pathname. Here is a small example of its usage:
9357 crt0%O%s %:if-exists(crti%O%s) crtbegin%O%s
9360 @item @code{if-exists-else}
9361 The @code{if-exists-else} spec function is similar to the @code{if-exists}
9362 spec function, except that it takes two arguments. The first argument is
9363 an absolute pathname to a file. If the file exists, @code{if-exists-else}
9364 returns the pathname. If it does not exist, it returns the second argument.
9365 This way, @code{if-exists-else} can be used to select one file or another,
9366 based on the existence of the first. Here is a small example of its usage:
9370 crt0%O%s %:if-exists(crti%O%s) \
9371 %:if-exists-else(crtbeginT%O%s crtbegin%O%s)
9374 @item @code{replace-outfile}
9375 The @code{replace-outfile} spec function takes two arguments. It looks for the
9376 first argument in the outfiles array and replaces it with the second argument. Here
9377 is a small example of its usage:
9380 %@{fgnu-runtime:%:replace-outfile(-lobjc -lobjc-gnu)@}
9383 @item @code{print-asm-header}
9384 The @code{print-asm-header} function takes no arguments and simply
9385 prints a banner like:
9391 Use "-Wa,OPTION" to pass "OPTION" to the assembler.
9394 It is used to separate compiler options from assembler options
9395 in the @option{--target-help} output.
9399 Substitutes the @code{-S} switch, if that switch was given to GCC@.
9400 If that switch was not specified, this substitutes nothing. Note that
9401 the leading dash is omitted when specifying this option, and it is
9402 automatically inserted if the substitution is performed. Thus the spec
9403 string @samp{%@{foo@}} would match the command-line option @option{-foo}
9404 and would output the command line option @option{-foo}.
9406 @item %W@{@code{S}@}
9407 Like %@{@code{S}@} but mark last argument supplied within as a file to be
9410 @item %@{@code{S}*@}
9411 Substitutes all the switches specified to GCC whose names start
9412 with @code{-S}, but which also take an argument. This is used for
9413 switches like @option{-o}, @option{-D}, @option{-I}, etc.
9414 GCC considers @option{-o foo} as being
9415 one switch whose names starts with @samp{o}. %@{o*@} would substitute this
9416 text, including the space. Thus two arguments would be generated.
9418 @item %@{@code{S}*&@code{T}*@}
9419 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
9420 (the order of @code{S} and @code{T} in the spec is not significant).
9421 There can be any number of ampersand-separated variables; for each the
9422 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
9424 @item %@{@code{S}:@code{X}@}
9425 Substitutes @code{X}, if the @samp{-S} switch was given to GCC@.
9427 @item %@{!@code{S}:@code{X}@}
9428 Substitutes @code{X}, if the @samp{-S} switch was @emph{not} given to GCC@.
9430 @item %@{@code{S}*:@code{X}@}
9431 Substitutes @code{X} if one or more switches whose names start with
9432 @code{-S} are specified to GCC@. Normally @code{X} is substituted only
9433 once, no matter how many such switches appeared. However, if @code{%*}
9434 appears somewhere in @code{X}, then @code{X} will be substituted once
9435 for each matching switch, with the @code{%*} replaced by the part of
9436 that switch that matched the @code{*}.
9438 @item %@{.@code{S}:@code{X}@}
9439 Substitutes @code{X}, if processing a file with suffix @code{S}.
9441 @item %@{!.@code{S}:@code{X}@}
9442 Substitutes @code{X}, if @emph{not} processing a file with suffix @code{S}.
9444 @item %@{,@code{S}:@code{X}@}
9445 Substitutes @code{X}, if processing a file for language @code{S}.
9447 @item %@{!,@code{S}:@code{X}@}
9448 Substitutes @code{X}, if not processing a file for language @code{S}.
9450 @item %@{@code{S}|@code{P}:@code{X}@}
9451 Substitutes @code{X} if either @code{-S} or @code{-P} was given to
9452 GCC@. This may be combined with @samp{!}, @samp{.}, @samp{,}, and
9453 @code{*} sequences as well, although they have a stronger binding than
9454 the @samp{|}. If @code{%*} appears in @code{X}, all of the
9455 alternatives must be starred, and only the first matching alternative
9458 For example, a spec string like this:
9461 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
9464 will output the following command-line options from the following input
9465 command-line options:
9470 -d fred.c -foo -baz -boggle
9471 -d jim.d -bar -baz -boggle
9474 @item %@{S:X; T:Y; :D@}
9476 If @code{S} was given to GCC, substitutes @code{X}; else if @code{T} was
9477 given to GCC, substitutes @code{Y}; else substitutes @code{D}. There can
9478 be as many clauses as you need. This may be combined with @code{.},
9479 @code{,}, @code{!}, @code{|}, and @code{*} as needed.
9484 The conditional text @code{X} in a %@{@code{S}:@code{X}@} or similar
9485 construct may contain other nested @samp{%} constructs or spaces, or
9486 even newlines. They are processed as usual, as described above.
9487 Trailing white space in @code{X} is ignored. White space may also
9488 appear anywhere on the left side of the colon in these constructs,
9489 except between @code{.} or @code{*} and the corresponding word.
9491 The @option{-O}, @option{-f}, @option{-m}, and @option{-W} switches are
9492 handled specifically in these constructs. If another value of
9493 @option{-O} or the negated form of a @option{-f}, @option{-m}, or
9494 @option{-W} switch is found later in the command line, the earlier
9495 switch value is ignored, except with @{@code{S}*@} where @code{S} is
9496 just one letter, which passes all matching options.
9498 The character @samp{|} at the beginning of the predicate text is used to
9499 indicate that a command should be piped to the following command, but
9500 only if @option{-pipe} is specified.
9502 It is built into GCC which switches take arguments and which do not.
9503 (You might think it would be useful to generalize this to allow each
9504 compiler's spec to say which switches take arguments. But this cannot
9505 be done in a consistent fashion. GCC cannot even decide which input
9506 files have been specified without knowing which switches take arguments,
9507 and it must know which input files to compile in order to tell which
9510 GCC also knows implicitly that arguments starting in @option{-l} are to be
9511 treated as compiler output files, and passed to the linker in their
9512 proper position among the other output files.
9514 @c man begin OPTIONS
9516 @node Target Options
9517 @section Specifying Target Machine and Compiler Version
9518 @cindex target options
9519 @cindex cross compiling
9520 @cindex specifying machine version
9521 @cindex specifying compiler version and target machine
9522 @cindex compiler version, specifying
9523 @cindex target machine, specifying
9525 The usual way to run GCC is to run the executable called @file{gcc}, or
9526 @file{<machine>-gcc} when cross-compiling, or
9527 @file{<machine>-gcc-<version>} to run a version other than the one that
9528 was installed last. Sometimes this is inconvenient, so GCC provides
9529 options that will switch to another cross-compiler or version.
9532 @item -b @var{machine}
9534 The argument @var{machine} specifies the target machine for compilation.
9536 The value to use for @var{machine} is the same as was specified as the
9537 machine type when configuring GCC as a cross-compiler. For
9538 example, if a cross-compiler was configured with @samp{configure
9539 arm-elf}, meaning to compile for an arm processor with elf binaries,
9540 then you would specify @option{-b arm-elf} to run that cross compiler.
9541 Because there are other options beginning with @option{-b}, the
9542 configuration must contain a hyphen, or @option{-b} alone should be one
9543 argument followed by the configuration in the next argument.
9545 @item -V @var{version}
9547 The argument @var{version} specifies which version of GCC to run.
9548 This is useful when multiple versions are installed. For example,
9549 @var{version} might be @samp{4.0}, meaning to run GCC version 4.0.
9552 The @option{-V} and @option{-b} options work by running the
9553 @file{<machine>-gcc-<version>} executable, so there's no real reason to
9554 use them if you can just run that directly.
9556 @node Submodel Options
9557 @section Hardware Models and Configurations
9558 @cindex submodel options
9559 @cindex specifying hardware config
9560 @cindex hardware models and configurations, specifying
9561 @cindex machine dependent options
9563 Earlier we discussed the standard option @option{-b} which chooses among
9564 different installed compilers for completely different target
9565 machines, such as VAX vs.@: 68000 vs.@: 80386.
9567 In addition, each of these target machine types can have its own
9568 special options, starting with @samp{-m}, to choose among various
9569 hardware models or configurations---for example, 68010 vs 68020,
9570 floating coprocessor or none. A single installed version of the
9571 compiler can compile for any model or configuration, according to the
9574 Some configurations of the compiler also support additional special
9575 options, usually for compatibility with other compilers on the same
9578 @c This list is ordered alphanumerically by subsection name.
9579 @c It should be the same order and spelling as these options are listed
9580 @c in Machine Dependent Options
9586 * Blackfin Options::
9590 * DEC Alpha Options::
9591 * DEC Alpha/VMS Options::
9594 * GNU/Linux Options::
9597 * i386 and x86-64 Options::
9598 * i386 and x86-64 Windows Options::
9600 * IA-64/VMS Options::
9612 * picoChip Options::
9614 * RS/6000 and PowerPC Options::
9616 * S/390 and zSeries Options::
9621 * System V Options::
9626 * Xstormy16 Options::
9632 @subsection ARC Options
9635 These options are defined for ARC implementations:
9640 Compile code for little endian mode. This is the default.
9644 Compile code for big endian mode.
9647 @opindex mmangle-cpu
9648 Prepend the name of the cpu to all public symbol names.
9649 In multiple-processor systems, there are many ARC variants with different
9650 instruction and register set characteristics. This flag prevents code
9651 compiled for one cpu to be linked with code compiled for another.
9652 No facility exists for handling variants that are ``almost identical''.
9653 This is an all or nothing option.
9655 @item -mcpu=@var{cpu}
9657 Compile code for ARC variant @var{cpu}.
9658 Which variants are supported depend on the configuration.
9659 All variants support @option{-mcpu=base}, this is the default.
9661 @item -mtext=@var{text-section}
9662 @itemx -mdata=@var{data-section}
9663 @itemx -mrodata=@var{readonly-data-section}
9667 Put functions, data, and readonly data in @var{text-section},
9668 @var{data-section}, and @var{readonly-data-section} respectively
9669 by default. This can be overridden with the @code{section} attribute.
9670 @xref{Variable Attributes}.
9672 @item -mfix-cortex-m3-ldrd
9673 @opindex mfix-cortex-m3-ldrd
9674 Some Cortex-M3 cores can cause data corruption when @code{ldrd} instructions
9675 with overlapping destination and base registers are used. This option avoids
9676 generating these instructions. This option is enabled by default when
9677 @option{-mcpu=cortex-m3} is specified.
9682 @subsection ARM Options
9685 These @samp{-m} options are defined for Advanced RISC Machines (ARM)
9689 @item -mabi=@var{name}
9691 Generate code for the specified ABI@. Permissible values are: @samp{apcs-gnu},
9692 @samp{atpcs}, @samp{aapcs}, @samp{aapcs-linux} and @samp{iwmmxt}.
9695 @opindex mapcs-frame
9696 Generate a stack frame that is compliant with the ARM Procedure Call
9697 Standard for all functions, even if this is not strictly necessary for
9698 correct execution of the code. Specifying @option{-fomit-frame-pointer}
9699 with this option will cause the stack frames not to be generated for
9700 leaf functions. The default is @option{-mno-apcs-frame}.
9704 This is a synonym for @option{-mapcs-frame}.
9707 @c not currently implemented
9708 @item -mapcs-stack-check
9709 @opindex mapcs-stack-check
9710 Generate code to check the amount of stack space available upon entry to
9711 every function (that actually uses some stack space). If there is
9712 insufficient space available then either the function
9713 @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be
9714 called, depending upon the amount of stack space required. The run time
9715 system is required to provide these functions. The default is
9716 @option{-mno-apcs-stack-check}, since this produces smaller code.
9718 @c not currently implemented
9720 @opindex mapcs-float
9721 Pass floating point arguments using the float point registers. This is
9722 one of the variants of the APCS@. This option is recommended if the
9723 target hardware has a floating point unit or if a lot of floating point
9724 arithmetic is going to be performed by the code. The default is
9725 @option{-mno-apcs-float}, since integer only code is slightly increased in
9726 size if @option{-mapcs-float} is used.
9728 @c not currently implemented
9729 @item -mapcs-reentrant
9730 @opindex mapcs-reentrant
9731 Generate reentrant, position independent code. The default is
9732 @option{-mno-apcs-reentrant}.
9735 @item -mthumb-interwork
9736 @opindex mthumb-interwork
9737 Generate code which supports calling between the ARM and Thumb
9738 instruction sets. Without this option the two instruction sets cannot
9739 be reliably used inside one program. The default is
9740 @option{-mno-thumb-interwork}, since slightly larger code is generated
9741 when @option{-mthumb-interwork} is specified.
9743 @item -mno-sched-prolog
9744 @opindex mno-sched-prolog
9745 Prevent the reordering of instructions in the function prolog, or the
9746 merging of those instruction with the instructions in the function's
9747 body. This means that all functions will start with a recognizable set
9748 of instructions (or in fact one of a choice from a small set of
9749 different function prologues), and this information can be used to
9750 locate the start if functions inside an executable piece of code. The
9751 default is @option{-msched-prolog}.
9753 @item -mfloat-abi=@var{name}
9755 Specifies which floating-point ABI to use. Permissible values
9756 are: @samp{soft}, @samp{softfp} and @samp{hard}.
9758 Specifying @samp{soft} causes GCC to generate output containing
9759 library calls for floating-point operations.
9760 @samp{softfp} allows the generation of code using hardware floating-point
9761 instructions, but still uses the soft-float calling conventions.
9762 @samp{hard} allows generation of floating-point instructions
9763 and uses FPU-specific calling conventions.
9765 The default depends on the specific target configuration. Note that
9766 the hard-float and soft-float ABIs are not link-compatible; you must
9767 compile your entire program with the same ABI, and link with a
9768 compatible set of libraries.
9771 @opindex mhard-float
9772 Equivalent to @option{-mfloat-abi=hard}.
9775 @opindex msoft-float
9776 Equivalent to @option{-mfloat-abi=soft}.
9778 @item -mlittle-endian
9779 @opindex mlittle-endian
9780 Generate code for a processor running in little-endian mode. This is
9781 the default for all standard configurations.
9784 @opindex mbig-endian
9785 Generate code for a processor running in big-endian mode; the default is
9786 to compile code for a little-endian processor.
9788 @item -mwords-little-endian
9789 @opindex mwords-little-endian
9790 This option only applies when generating code for big-endian processors.
9791 Generate code for a little-endian word order but a big-endian byte
9792 order. That is, a byte order of the form @samp{32107654}. Note: this
9793 option should only be used if you require compatibility with code for
9794 big-endian ARM processors generated by versions of the compiler prior to
9797 @item -mcpu=@var{name}
9799 This specifies the name of the target ARM processor. GCC uses this name
9800 to determine what kind of instructions it can emit when generating
9801 assembly code. Permissible names are: @samp{arm2}, @samp{arm250},
9802 @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610},
9803 @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm},
9804 @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700},
9805 @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100},
9807 @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm7tdmi-s},
9808 @samp{arm710t}, @samp{arm720t}, @samp{arm740t},
9809 @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100},
9810 @samp{strongarm1110},
9811 @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920},
9812 @samp{arm920t}, @samp{arm922t}, @samp{arm946e-s}, @samp{arm966e-s},
9813 @samp{arm968e-s}, @samp{arm926ej-s}, @samp{arm940t}, @samp{arm9tdmi},
9814 @samp{arm10tdmi}, @samp{arm1020t}, @samp{arm1026ej-s},
9815 @samp{arm10e}, @samp{arm1020e}, @samp{arm1022e},
9816 @samp{arm1136j-s}, @samp{arm1136jf-s}, @samp{mpcore}, @samp{mpcorenovfp},
9817 @samp{arm1156t2-s}, @samp{arm1156t2f-s}, @samp{arm1176jz-s}, @samp{arm1176jzf-s},
9818 @samp{cortex-a5}, @samp{cortex-a8}, @samp{cortex-a9},
9819 @samp{cortex-r4}, @samp{cortex-r4f}, @samp{cortex-m3},
9822 @samp{xscale}, @samp{iwmmxt}, @samp{iwmmxt2}, @samp{ep9312}.
9824 @item -mtune=@var{name}
9826 This option is very similar to the @option{-mcpu=} option, except that
9827 instead of specifying the actual target processor type, and hence
9828 restricting which instructions can be used, it specifies that GCC should
9829 tune the performance of the code as if the target were of the type
9830 specified in this option, but still choosing the instructions that it
9831 will generate based on the cpu specified by a @option{-mcpu=} option.
9832 For some ARM implementations better performance can be obtained by using
9835 @item -march=@var{name}
9837 This specifies the name of the target ARM architecture. GCC uses this
9838 name to determine what kind of instructions it can emit when generating
9839 assembly code. This option can be used in conjunction with or instead
9840 of the @option{-mcpu=} option. Permissible names are: @samp{armv2},
9841 @samp{armv2a}, @samp{armv3}, @samp{armv3m}, @samp{armv4}, @samp{armv4t},
9842 @samp{armv5}, @samp{armv5t}, @samp{armv5e}, @samp{armv5te},
9843 @samp{armv6}, @samp{armv6j},
9844 @samp{armv6t2}, @samp{armv6z}, @samp{armv6zk}, @samp{armv6-m},
9845 @samp{armv7}, @samp{armv7-a}, @samp{armv7-r}, @samp{armv7-m},
9846 @samp{iwmmxt}, @samp{iwmmxt2}, @samp{ep9312}.
9848 @item -mfpu=@var{name}
9849 @itemx -mfpe=@var{number}
9850 @itemx -mfp=@var{number}
9854 This specifies what floating point hardware (or hardware emulation) is
9855 available on the target. Permissible names are: @samp{fpa}, @samp{fpe2},
9856 @samp{fpe3}, @samp{maverick}, @samp{vfp}, @samp{vfpv3}, @samp{vfpv3-fp16},
9857 @samp{vfpv3-d16}, @samp{vfpv3-d16-fp16}, @samp{vfpv3xd}, @samp{vfpv3xd-fp16},
9858 @samp{neon}, @samp{neon-fp16}, @samp{vfpv4}, @samp{vfpv4-d16},
9859 @samp{fpv4-sp-d16} and @samp{neon-vfpv4}.
9860 @option{-mfp} and @option{-mfpe} are synonyms for
9861 @option{-mfpu}=@samp{fpe}@var{number}, for compatibility with older versions
9864 If @option{-msoft-float} is specified this specifies the format of
9865 floating point values.
9867 @item -mfp16-format=@var{name}
9868 @opindex mfp16-format
9869 Specify the format of the @code{__fp16} half-precision floating-point type.
9870 Permissible names are @samp{none}, @samp{ieee}, and @samp{alternative};
9871 the default is @samp{none}, in which case the @code{__fp16} type is not
9872 defined. @xref{Half-Precision}, for more information.
9874 @item -mstructure-size-boundary=@var{n}
9875 @opindex mstructure-size-boundary
9876 The size of all structures and unions will be rounded up to a multiple
9877 of the number of bits set by this option. Permissible values are 8, 32
9878 and 64. The default value varies for different toolchains. For the COFF
9879 targeted toolchain the default value is 8. A value of 64 is only allowed
9880 if the underlying ABI supports it.
9882 Specifying the larger number can produce faster, more efficient code, but
9883 can also increase the size of the program. Different values are potentially
9884 incompatible. Code compiled with one value cannot necessarily expect to
9885 work with code or libraries compiled with another value, if they exchange
9886 information using structures or unions.
9888 @item -mabort-on-noreturn
9889 @opindex mabort-on-noreturn
9890 Generate a call to the function @code{abort} at the end of a
9891 @code{noreturn} function. It will be executed if the function tries to
9895 @itemx -mno-long-calls
9896 @opindex mlong-calls
9897 @opindex mno-long-calls
9898 Tells the compiler to perform function calls by first loading the
9899 address of the function into a register and then performing a subroutine
9900 call on this register. This switch is needed if the target function
9901 will lie outside of the 64 megabyte addressing range of the offset based
9902 version of subroutine call instruction.
9904 Even if this switch is enabled, not all function calls will be turned
9905 into long calls. The heuristic is that static functions, functions
9906 which have the @samp{short-call} attribute, functions that are inside
9907 the scope of a @samp{#pragma no_long_calls} directive and functions whose
9908 definitions have already been compiled within the current compilation
9909 unit, will not be turned into long calls. The exception to this rule is
9910 that weak function definitions, functions with the @samp{long-call}
9911 attribute or the @samp{section} attribute, and functions that are within
9912 the scope of a @samp{#pragma long_calls} directive, will always be
9913 turned into long calls.
9915 This feature is not enabled by default. Specifying
9916 @option{-mno-long-calls} will restore the default behavior, as will
9917 placing the function calls within the scope of a @samp{#pragma
9918 long_calls_off} directive. Note these switches have no effect on how
9919 the compiler generates code to handle function calls via function
9922 @item -msingle-pic-base
9923 @opindex msingle-pic-base
9924 Treat the register used for PIC addressing as read-only, rather than
9925 loading it in the prologue for each function. The run-time system is
9926 responsible for initializing this register with an appropriate value
9927 before execution begins.
9929 @item -mpic-register=@var{reg}
9930 @opindex mpic-register
9931 Specify the register to be used for PIC addressing. The default is R10
9932 unless stack-checking is enabled, when R9 is used.
9934 @item -mcirrus-fix-invalid-insns
9935 @opindex mcirrus-fix-invalid-insns
9936 @opindex mno-cirrus-fix-invalid-insns
9937 Insert NOPs into the instruction stream to in order to work around
9938 problems with invalid Maverick instruction combinations. This option
9939 is only valid if the @option{-mcpu=ep9312} option has been used to
9940 enable generation of instructions for the Cirrus Maverick floating
9941 point co-processor. This option is not enabled by default, since the
9942 problem is only present in older Maverick implementations. The default
9943 can be re-enabled by use of the @option{-mno-cirrus-fix-invalid-insns}
9946 @item -mpoke-function-name
9947 @opindex mpoke-function-name
9948 Write the name of each function into the text section, directly
9949 preceding the function prologue. The generated code is similar to this:
9953 .ascii "arm_poke_function_name", 0
9956 .word 0xff000000 + (t1 - t0)
9957 arm_poke_function_name
9959 stmfd sp!, @{fp, ip, lr, pc@}
9963 When performing a stack backtrace, code can inspect the value of
9964 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
9965 location @code{pc - 12} and the top 8 bits are set, then we know that
9966 there is a function name embedded immediately preceding this location
9967 and has length @code{((pc[-3]) & 0xff000000)}.
9971 Generate code for the Thumb instruction set. The default is to
9972 use the 32-bit ARM instruction set.
9973 This option automatically enables either 16-bit Thumb-1 or
9974 mixed 16/32-bit Thumb-2 instructions based on the @option{-mcpu=@var{name}}
9975 and @option{-march=@var{name}} options. This option is not passed to the
9976 assembler. If you want to force assembler files to be interpreted as Thumb code,
9977 either add a @samp{.thumb} directive to the source or pass the @option{-mthumb}
9978 option directly to the assembler by prefixing it with @option{-Wa}.
9981 @opindex mtpcs-frame
9982 Generate a stack frame that is compliant with the Thumb Procedure Call
9983 Standard for all non-leaf functions. (A leaf function is one that does
9984 not call any other functions.) The default is @option{-mno-tpcs-frame}.
9986 @item -mtpcs-leaf-frame
9987 @opindex mtpcs-leaf-frame
9988 Generate a stack frame that is compliant with the Thumb Procedure Call
9989 Standard for all leaf functions. (A leaf function is one that does
9990 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
9992 @item -mcallee-super-interworking
9993 @opindex mcallee-super-interworking
9994 Gives all externally visible functions in the file being compiled an ARM
9995 instruction set header which switches to Thumb mode before executing the
9996 rest of the function. This allows these functions to be called from
9997 non-interworking code. This option is not valid in AAPCS configurations
9998 because interworking is enabled by default.
10000 @item -mcaller-super-interworking
10001 @opindex mcaller-super-interworking
10002 Allows calls via function pointers (including virtual functions) to
10003 execute correctly regardless of whether the target code has been
10004 compiled for interworking or not. There is a small overhead in the cost
10005 of executing a function pointer if this option is enabled. This option
10006 is not valid in AAPCS configurations because interworking is enabled
10009 @item -mtp=@var{name}
10011 Specify the access model for the thread local storage pointer. The valid
10012 models are @option{soft}, which generates calls to @code{__aeabi_read_tp},
10013 @option{cp15}, which fetches the thread pointer from @code{cp15} directly
10014 (supported in the arm6k architecture), and @option{auto}, which uses the
10015 best available method for the selected processor. The default setting is
10018 @item -mword-relocations
10019 @opindex mword-relocations
10020 Only generate absolute relocations on word sized values (i.e. R_ARM_ABS32).
10021 This is enabled by default on targets (uClinux, SymbianOS) where the runtime
10022 loader imposes this restriction, and when @option{-fpic} or @option{-fPIC}
10028 @subsection AVR Options
10029 @cindex AVR Options
10031 These options are defined for AVR implementations:
10034 @item -mmcu=@var{mcu}
10036 Specify ATMEL AVR instruction set or MCU type.
10038 Instruction set avr1 is for the minimal AVR core, not supported by the C
10039 compiler, only for assembler programs (MCU types: at90s1200, attiny10,
10040 attiny11, attiny12, attiny15, attiny28).
10042 Instruction set avr2 (default) is for the classic AVR core with up to
10043 8K program memory space (MCU types: at90s2313, at90s2323, attiny22,
10044 at90s2333, at90s2343, at90s4414, at90s4433, at90s4434, at90s8515,
10045 at90c8534, at90s8535).
10047 Instruction set avr3 is for the classic AVR core with up to 128K program
10048 memory space (MCU types: atmega103, atmega603, at43usb320, at76c711).
10050 Instruction set avr4 is for the enhanced AVR core with up to 8K program
10051 memory space (MCU types: atmega8, atmega83, atmega85).
10053 Instruction set avr5 is for the enhanced AVR core with up to 128K program
10054 memory space (MCU types: atmega16, atmega161, atmega163, atmega32, atmega323,
10055 atmega64, atmega128, at43usb355, at94k).
10057 @item -mno-interrupts
10058 @opindex mno-interrupts
10059 Generated code is not compatible with hardware interrupts.
10060 Code size will be smaller.
10062 @item -mcall-prologues
10063 @opindex mcall-prologues
10064 Functions prologues/epilogues expanded as call to appropriate
10065 subroutines. Code size will be smaller.
10068 @opindex mtiny-stack
10069 Change only the low 8 bits of the stack pointer.
10073 Assume int to be 8 bit integer. This affects the sizes of all types: A
10074 char will be 1 byte, an int will be 1 byte, a long will be 2 bytes
10075 and long long will be 4 bytes. Please note that this option does not
10076 comply to the C standards, but it will provide you with smaller code
10080 @node Blackfin Options
10081 @subsection Blackfin Options
10082 @cindex Blackfin Options
10085 @item -mcpu=@var{cpu}@r{[}-@var{sirevision}@r{]}
10087 Specifies the name of the target Blackfin processor. Currently, @var{cpu}
10088 can be one of @samp{bf512}, @samp{bf514}, @samp{bf516}, @samp{bf518},
10089 @samp{bf522}, @samp{bf523}, @samp{bf524}, @samp{bf525}, @samp{bf526},
10090 @samp{bf527}, @samp{bf531}, @samp{bf532}, @samp{bf533},
10091 @samp{bf534}, @samp{bf536}, @samp{bf537}, @samp{bf538}, @samp{bf539},
10092 @samp{bf542}, @samp{bf544}, @samp{bf547}, @samp{bf548}, @samp{bf549},
10093 @samp{bf542m}, @samp{bf544m}, @samp{bf547m}, @samp{bf548m}, @samp{bf549m},
10095 The optional @var{sirevision} specifies the silicon revision of the target
10096 Blackfin processor. Any workarounds available for the targeted silicon revision
10097 will be enabled. If @var{sirevision} is @samp{none}, no workarounds are enabled.
10098 If @var{sirevision} is @samp{any}, all workarounds for the targeted processor
10099 will be enabled. The @code{__SILICON_REVISION__} macro is defined to two
10100 hexadecimal digits representing the major and minor numbers in the silicon
10101 revision. If @var{sirevision} is @samp{none}, the @code{__SILICON_REVISION__}
10102 is not defined. If @var{sirevision} is @samp{any}, the
10103 @code{__SILICON_REVISION__} is defined to be @code{0xffff}.
10104 If this optional @var{sirevision} is not used, GCC assumes the latest known
10105 silicon revision of the targeted Blackfin processor.
10107 Support for @samp{bf561} is incomplete. For @samp{bf561},
10108 Only the processor macro is defined.
10109 Without this option, @samp{bf532} is used as the processor by default.
10110 The corresponding predefined processor macros for @var{cpu} is to
10111 be defined. And for @samp{bfin-elf} toolchain, this causes the hardware BSP
10112 provided by libgloss to be linked in if @option{-msim} is not given.
10116 Specifies that the program will be run on the simulator. This causes
10117 the simulator BSP provided by libgloss to be linked in. This option
10118 has effect only for @samp{bfin-elf} toolchain.
10119 Certain other options, such as @option{-mid-shared-library} and
10120 @option{-mfdpic}, imply @option{-msim}.
10122 @item -momit-leaf-frame-pointer
10123 @opindex momit-leaf-frame-pointer
10124 Don't keep the frame pointer in a register for leaf functions. This
10125 avoids the instructions to save, set up and restore frame pointers and
10126 makes an extra register available in leaf functions. The option
10127 @option{-fomit-frame-pointer} removes the frame pointer for all functions
10128 which might make debugging harder.
10130 @item -mspecld-anomaly
10131 @opindex mspecld-anomaly
10132 When enabled, the compiler will ensure that the generated code does not
10133 contain speculative loads after jump instructions. If this option is used,
10134 @code{__WORKAROUND_SPECULATIVE_LOADS} is defined.
10136 @item -mno-specld-anomaly
10137 @opindex mno-specld-anomaly
10138 Don't generate extra code to prevent speculative loads from occurring.
10140 @item -mcsync-anomaly
10141 @opindex mcsync-anomaly
10142 When enabled, the compiler will ensure that the generated code does not
10143 contain CSYNC or SSYNC instructions too soon after conditional branches.
10144 If this option is used, @code{__WORKAROUND_SPECULATIVE_SYNCS} is defined.
10146 @item -mno-csync-anomaly
10147 @opindex mno-csync-anomaly
10148 Don't generate extra code to prevent CSYNC or SSYNC instructions from
10149 occurring too soon after a conditional branch.
10153 When enabled, the compiler is free to take advantage of the knowledge that
10154 the entire program fits into the low 64k of memory.
10157 @opindex mno-low-64k
10158 Assume that the program is arbitrarily large. This is the default.
10160 @item -mstack-check-l1
10161 @opindex mstack-check-l1
10162 Do stack checking using information placed into L1 scratchpad memory by the
10165 @item -mid-shared-library
10166 @opindex mid-shared-library
10167 Generate code that supports shared libraries via the library ID method.
10168 This allows for execute in place and shared libraries in an environment
10169 without virtual memory management. This option implies @option{-fPIC}.
10170 With a @samp{bfin-elf} target, this option implies @option{-msim}.
10172 @item -mno-id-shared-library
10173 @opindex mno-id-shared-library
10174 Generate code that doesn't assume ID based shared libraries are being used.
10175 This is the default.
10177 @item -mleaf-id-shared-library
10178 @opindex mleaf-id-shared-library
10179 Generate code that supports shared libraries via the library ID method,
10180 but assumes that this library or executable won't link against any other
10181 ID shared libraries. That allows the compiler to use faster code for jumps
10184 @item -mno-leaf-id-shared-library
10185 @opindex mno-leaf-id-shared-library
10186 Do not assume that the code being compiled won't link against any ID shared
10187 libraries. Slower code will be generated for jump and call insns.
10189 @item -mshared-library-id=n
10190 @opindex mshared-library-id
10191 Specified the identification number of the ID based shared library being
10192 compiled. Specifying a value of 0 will generate more compact code, specifying
10193 other values will force the allocation of that number to the current
10194 library but is no more space or time efficient than omitting this option.
10198 Generate code that allows the data segment to be located in a different
10199 area of memory from the text segment. This allows for execute in place in
10200 an environment without virtual memory management by eliminating relocations
10201 against the text section.
10203 @item -mno-sep-data
10204 @opindex mno-sep-data
10205 Generate code that assumes that the data segment follows the text segment.
10206 This is the default.
10209 @itemx -mno-long-calls
10210 @opindex mlong-calls
10211 @opindex mno-long-calls
10212 Tells the compiler to perform function calls by first loading the
10213 address of the function into a register and then performing a subroutine
10214 call on this register. This switch is needed if the target function
10215 will lie outside of the 24 bit addressing range of the offset based
10216 version of subroutine call instruction.
10218 This feature is not enabled by default. Specifying
10219 @option{-mno-long-calls} will restore the default behavior. Note these
10220 switches have no effect on how the compiler generates code to handle
10221 function calls via function pointers.
10225 Link with the fast floating-point library. This library relaxes some of
10226 the IEEE floating-point standard's rules for checking inputs against
10227 Not-a-Number (NAN), in the interest of performance.
10230 @opindex minline-plt
10231 Enable inlining of PLT entries in function calls to functions that are
10232 not known to bind locally. It has no effect without @option{-mfdpic}.
10235 @opindex mmulticore
10236 Build standalone application for multicore Blackfin processor. Proper
10237 start files and link scripts will be used to support multicore.
10238 This option defines @code{__BFIN_MULTICORE}. It can only be used with
10239 @option{-mcpu=bf561@r{[}-@var{sirevision}@r{]}}. It can be used with
10240 @option{-mcorea} or @option{-mcoreb}. If it's used without
10241 @option{-mcorea} or @option{-mcoreb}, single application/dual core
10242 programming model is used. In this model, the main function of Core B
10243 should be named as coreb_main. If it's used with @option{-mcorea} or
10244 @option{-mcoreb}, one application per core programming model is used.
10245 If this option is not used, single core application programming
10250 Build standalone application for Core A of BF561 when using
10251 one application per core programming model. Proper start files
10252 and link scripts will be used to support Core A. This option
10253 defines @code{__BFIN_COREA}. It must be used with @option{-mmulticore}.
10257 Build standalone application for Core B of BF561 when using
10258 one application per core programming model. Proper start files
10259 and link scripts will be used to support Core B. This option
10260 defines @code{__BFIN_COREB}. When this option is used, coreb_main
10261 should be used instead of main. It must be used with
10262 @option{-mmulticore}.
10266 Build standalone application for SDRAM. Proper start files and
10267 link scripts will be used to put the application into SDRAM.
10268 Loader should initialize SDRAM before loading the application
10269 into SDRAM. This option defines @code{__BFIN_SDRAM}.
10273 Assume that ICPLBs are enabled at runtime. This has an effect on certain
10274 anomaly workarounds. For Linux targets, the default is to assume ICPLBs
10275 are enabled; for standalone applications the default is off.
10279 @subsection CRIS Options
10280 @cindex CRIS Options
10282 These options are defined specifically for the CRIS ports.
10285 @item -march=@var{architecture-type}
10286 @itemx -mcpu=@var{architecture-type}
10289 Generate code for the specified architecture. The choices for
10290 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
10291 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX@.
10292 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
10295 @item -mtune=@var{architecture-type}
10297 Tune to @var{architecture-type} everything applicable about the generated
10298 code, except for the ABI and the set of available instructions. The
10299 choices for @var{architecture-type} are the same as for
10300 @option{-march=@var{architecture-type}}.
10302 @item -mmax-stack-frame=@var{n}
10303 @opindex mmax-stack-frame
10304 Warn when the stack frame of a function exceeds @var{n} bytes.
10310 The options @option{-metrax4} and @option{-metrax100} are synonyms for
10311 @option{-march=v3} and @option{-march=v8} respectively.
10313 @item -mmul-bug-workaround
10314 @itemx -mno-mul-bug-workaround
10315 @opindex mmul-bug-workaround
10316 @opindex mno-mul-bug-workaround
10317 Work around a bug in the @code{muls} and @code{mulu} instructions for CPU
10318 models where it applies. This option is active by default.
10322 Enable CRIS-specific verbose debug-related information in the assembly
10323 code. This option also has the effect to turn off the @samp{#NO_APP}
10324 formatted-code indicator to the assembler at the beginning of the
10329 Do not use condition-code results from previous instruction; always emit
10330 compare and test instructions before use of condition codes.
10332 @item -mno-side-effects
10333 @opindex mno-side-effects
10334 Do not emit instructions with side-effects in addressing modes other than
10337 @item -mstack-align
10338 @itemx -mno-stack-align
10339 @itemx -mdata-align
10340 @itemx -mno-data-align
10341 @itemx -mconst-align
10342 @itemx -mno-const-align
10343 @opindex mstack-align
10344 @opindex mno-stack-align
10345 @opindex mdata-align
10346 @opindex mno-data-align
10347 @opindex mconst-align
10348 @opindex mno-const-align
10349 These options (no-options) arranges (eliminate arrangements) for the
10350 stack-frame, individual data and constants to be aligned for the maximum
10351 single data access size for the chosen CPU model. The default is to
10352 arrange for 32-bit alignment. ABI details such as structure layout are
10353 not affected by these options.
10361 Similar to the stack- data- and const-align options above, these options
10362 arrange for stack-frame, writable data and constants to all be 32-bit,
10363 16-bit or 8-bit aligned. The default is 32-bit alignment.
10365 @item -mno-prologue-epilogue
10366 @itemx -mprologue-epilogue
10367 @opindex mno-prologue-epilogue
10368 @opindex mprologue-epilogue
10369 With @option{-mno-prologue-epilogue}, the normal function prologue and
10370 epilogue that sets up the stack-frame are omitted and no return
10371 instructions or return sequences are generated in the code. Use this
10372 option only together with visual inspection of the compiled code: no
10373 warnings or errors are generated when call-saved registers must be saved,
10374 or storage for local variable needs to be allocated.
10378 @opindex mno-gotplt
10380 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
10381 instruction sequences that load addresses for functions from the PLT part
10382 of the GOT rather than (traditional on other architectures) calls to the
10383 PLT@. The default is @option{-mgotplt}.
10387 Legacy no-op option only recognized with the cris-axis-elf and
10388 cris-axis-linux-gnu targets.
10392 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
10396 This option, recognized for the cris-axis-elf arranges
10397 to link with input-output functions from a simulator library. Code,
10398 initialized data and zero-initialized data are allocated consecutively.
10402 Like @option{-sim}, but pass linker options to locate initialized data at
10403 0x40000000 and zero-initialized data at 0x80000000.
10407 @subsection CRX Options
10408 @cindex CRX Options
10410 These options are defined specifically for the CRX ports.
10416 Enable the use of multiply-accumulate instructions. Disabled by default.
10419 @opindex mpush-args
10420 Push instructions will be used to pass outgoing arguments when functions
10421 are called. Enabled by default.
10424 @node Darwin Options
10425 @subsection Darwin Options
10426 @cindex Darwin options
10428 These options are defined for all architectures running the Darwin operating
10431 FSF GCC on Darwin does not create ``fat'' object files; it will create
10432 an object file for the single architecture that it was built to
10433 target. Apple's GCC on Darwin does create ``fat'' files if multiple
10434 @option{-arch} options are used; it does so by running the compiler or
10435 linker multiple times and joining the results together with
10438 The subtype of the file created (like @samp{ppc7400} or @samp{ppc970} or
10439 @samp{i686}) is determined by the flags that specify the ISA
10440 that GCC is targetting, like @option{-mcpu} or @option{-march}. The
10441 @option{-force_cpusubtype_ALL} option can be used to override this.
10443 The Darwin tools vary in their behavior when presented with an ISA
10444 mismatch. The assembler, @file{as}, will only permit instructions to
10445 be used that are valid for the subtype of the file it is generating,
10446 so you cannot put 64-bit instructions in a @samp{ppc750} object file.
10447 The linker for shared libraries, @file{/usr/bin/libtool}, will fail
10448 and print an error if asked to create a shared library with a less
10449 restrictive subtype than its input files (for instance, trying to put
10450 a @samp{ppc970} object file in a @samp{ppc7400} library). The linker
10451 for executables, @file{ld}, will quietly give the executable the most
10452 restrictive subtype of any of its input files.
10457 Add the framework directory @var{dir} to the head of the list of
10458 directories to be searched for header files. These directories are
10459 interleaved with those specified by @option{-I} options and are
10460 scanned in a left-to-right order.
10462 A framework directory is a directory with frameworks in it. A
10463 framework is a directory with a @samp{"Headers"} and/or
10464 @samp{"PrivateHeaders"} directory contained directly in it that ends
10465 in @samp{".framework"}. The name of a framework is the name of this
10466 directory excluding the @samp{".framework"}. Headers associated with
10467 the framework are found in one of those two directories, with
10468 @samp{"Headers"} being searched first. A subframework is a framework
10469 directory that is in a framework's @samp{"Frameworks"} directory.
10470 Includes of subframework headers can only appear in a header of a
10471 framework that contains the subframework, or in a sibling subframework
10472 header. Two subframeworks are siblings if they occur in the same
10473 framework. A subframework should not have the same name as a
10474 framework, a warning will be issued if this is violated. Currently a
10475 subframework cannot have subframeworks, in the future, the mechanism
10476 may be extended to support this. The standard frameworks can be found
10477 in @samp{"/System/Library/Frameworks"} and
10478 @samp{"/Library/Frameworks"}. An example include looks like
10479 @code{#include <Framework/header.h>}, where @samp{Framework} denotes
10480 the name of the framework and header.h is found in the
10481 @samp{"PrivateHeaders"} or @samp{"Headers"} directory.
10483 @item -iframework@var{dir}
10484 @opindex iframework
10485 Like @option{-F} except the directory is a treated as a system
10486 directory. The main difference between this @option{-iframework} and
10487 @option{-F} is that with @option{-iframework} the compiler does not
10488 warn about constructs contained within header files found via
10489 @var{dir}. This option is valid only for the C family of languages.
10493 Emit debugging information for symbols that are used. For STABS
10494 debugging format, this enables @option{-feliminate-unused-debug-symbols}.
10495 This is by default ON@.
10499 Emit debugging information for all symbols and types.
10501 @item -mmacosx-version-min=@var{version}
10502 The earliest version of MacOS X that this executable will run on
10503 is @var{version}. Typical values of @var{version} include @code{10.1},
10504 @code{10.2}, and @code{10.3.9}.
10506 If the compiler was built to use the system's headers by default,
10507 then the default for this option is the system version on which the
10508 compiler is running, otherwise the default is to make choices which
10509 are compatible with as many systems and code bases as possible.
10513 Enable kernel development mode. The @option{-mkernel} option sets
10514 @option{-static}, @option{-fno-common}, @option{-fno-cxa-atexit},
10515 @option{-fno-exceptions}, @option{-fno-non-call-exceptions},
10516 @option{-fapple-kext}, @option{-fno-weak} and @option{-fno-rtti} where
10517 applicable. This mode also sets @option{-mno-altivec},
10518 @option{-msoft-float}, @option{-fno-builtin} and
10519 @option{-mlong-branch} for PowerPC targets.
10521 @item -mone-byte-bool
10522 @opindex mone-byte-bool
10523 Override the defaults for @samp{bool} so that @samp{sizeof(bool)==1}.
10524 By default @samp{sizeof(bool)} is @samp{4} when compiling for
10525 Darwin/PowerPC and @samp{1} when compiling for Darwin/x86, so this
10526 option has no effect on x86.
10528 @strong{Warning:} The @option{-mone-byte-bool} switch causes GCC
10529 to generate code that is not binary compatible with code generated
10530 without that switch. Using this switch may require recompiling all
10531 other modules in a program, including system libraries. Use this
10532 switch to conform to a non-default data model.
10534 @item -mfix-and-continue
10535 @itemx -ffix-and-continue
10536 @itemx -findirect-data
10537 @opindex mfix-and-continue
10538 @opindex ffix-and-continue
10539 @opindex findirect-data
10540 Generate code suitable for fast turn around development. Needed to
10541 enable gdb to dynamically load @code{.o} files into already running
10542 programs. @option{-findirect-data} and @option{-ffix-and-continue}
10543 are provided for backwards compatibility.
10547 Loads all members of static archive libraries.
10548 See man ld(1) for more information.
10550 @item -arch_errors_fatal
10551 @opindex arch_errors_fatal
10552 Cause the errors having to do with files that have the wrong architecture
10555 @item -bind_at_load
10556 @opindex bind_at_load
10557 Causes the output file to be marked such that the dynamic linker will
10558 bind all undefined references when the file is loaded or launched.
10562 Produce a Mach-o bundle format file.
10563 See man ld(1) for more information.
10565 @item -bundle_loader @var{executable}
10566 @opindex bundle_loader
10567 This option specifies the @var{executable} that will be loading the build
10568 output file being linked. See man ld(1) for more information.
10571 @opindex dynamiclib
10572 When passed this option, GCC will produce a dynamic library instead of
10573 an executable when linking, using the Darwin @file{libtool} command.
10575 @item -force_cpusubtype_ALL
10576 @opindex force_cpusubtype_ALL
10577 This causes GCC's output file to have the @var{ALL} subtype, instead of
10578 one controlled by the @option{-mcpu} or @option{-march} option.
10580 @item -allowable_client @var{client_name}
10581 @itemx -client_name
10582 @itemx -compatibility_version
10583 @itemx -current_version
10585 @itemx -dependency-file
10587 @itemx -dylinker_install_name
10589 @itemx -exported_symbols_list
10591 @itemx -flat_namespace
10592 @itemx -force_flat_namespace
10593 @itemx -headerpad_max_install_names
10596 @itemx -install_name
10597 @itemx -keep_private_externs
10598 @itemx -multi_module
10599 @itemx -multiply_defined
10600 @itemx -multiply_defined_unused
10602 @itemx -no_dead_strip_inits_and_terms
10603 @itemx -nofixprebinding
10604 @itemx -nomultidefs
10606 @itemx -noseglinkedit
10607 @itemx -pagezero_size
10609 @itemx -prebind_all_twolevel_modules
10610 @itemx -private_bundle
10611 @itemx -read_only_relocs
10613 @itemx -sectobjectsymbols
10617 @itemx -sectobjectsymbols
10620 @itemx -segs_read_only_addr
10621 @itemx -segs_read_write_addr
10622 @itemx -seg_addr_table
10623 @itemx -seg_addr_table_filename
10624 @itemx -seglinkedit
10626 @itemx -segs_read_only_addr
10627 @itemx -segs_read_write_addr
10628 @itemx -single_module
10630 @itemx -sub_library
10631 @itemx -sub_umbrella
10632 @itemx -twolevel_namespace
10635 @itemx -unexported_symbols_list
10636 @itemx -weak_reference_mismatches
10637 @itemx -whatsloaded
10638 @opindex allowable_client
10639 @opindex client_name
10640 @opindex compatibility_version
10641 @opindex current_version
10642 @opindex dead_strip
10643 @opindex dependency-file
10644 @opindex dylib_file
10645 @opindex dylinker_install_name
10647 @opindex exported_symbols_list
10649 @opindex flat_namespace
10650 @opindex force_flat_namespace
10651 @opindex headerpad_max_install_names
10652 @opindex image_base
10654 @opindex install_name
10655 @opindex keep_private_externs
10656 @opindex multi_module
10657 @opindex multiply_defined
10658 @opindex multiply_defined_unused
10659 @opindex noall_load
10660 @opindex no_dead_strip_inits_and_terms
10661 @opindex nofixprebinding
10662 @opindex nomultidefs
10664 @opindex noseglinkedit
10665 @opindex pagezero_size
10667 @opindex prebind_all_twolevel_modules
10668 @opindex private_bundle
10669 @opindex read_only_relocs
10671 @opindex sectobjectsymbols
10674 @opindex sectcreate
10675 @opindex sectobjectsymbols
10678 @opindex segs_read_only_addr
10679 @opindex segs_read_write_addr
10680 @opindex seg_addr_table
10681 @opindex seg_addr_table_filename
10682 @opindex seglinkedit
10684 @opindex segs_read_only_addr
10685 @opindex segs_read_write_addr
10686 @opindex single_module
10688 @opindex sub_library
10689 @opindex sub_umbrella
10690 @opindex twolevel_namespace
10693 @opindex unexported_symbols_list
10694 @opindex weak_reference_mismatches
10695 @opindex whatsloaded
10696 These options are passed to the Darwin linker. The Darwin linker man page
10697 describes them in detail.
10700 @node DEC Alpha Options
10701 @subsection DEC Alpha Options
10703 These @samp{-m} options are defined for the DEC Alpha implementations:
10706 @item -mno-soft-float
10707 @itemx -msoft-float
10708 @opindex mno-soft-float
10709 @opindex msoft-float
10710 Use (do not use) the hardware floating-point instructions for
10711 floating-point operations. When @option{-msoft-float} is specified,
10712 functions in @file{libgcc.a} will be used to perform floating-point
10713 operations. Unless they are replaced by routines that emulate the
10714 floating-point operations, or compiled in such a way as to call such
10715 emulations routines, these routines will issue floating-point
10716 operations. If you are compiling for an Alpha without floating-point
10717 operations, you must ensure that the library is built so as not to call
10720 Note that Alpha implementations without floating-point operations are
10721 required to have floating-point registers.
10724 @itemx -mno-fp-regs
10726 @opindex mno-fp-regs
10727 Generate code that uses (does not use) the floating-point register set.
10728 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
10729 register set is not used, floating point operands are passed in integer
10730 registers as if they were integers and floating-point results are passed
10731 in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence,
10732 so any function with a floating-point argument or return value called by code
10733 compiled with @option{-mno-fp-regs} must also be compiled with that
10736 A typical use of this option is building a kernel that does not use,
10737 and hence need not save and restore, any floating-point registers.
10741 The Alpha architecture implements floating-point hardware optimized for
10742 maximum performance. It is mostly compliant with the IEEE floating
10743 point standard. However, for full compliance, software assistance is
10744 required. This option generates code fully IEEE compliant code
10745 @emph{except} that the @var{inexact-flag} is not maintained (see below).
10746 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
10747 defined during compilation. The resulting code is less efficient but is
10748 able to correctly support denormalized numbers and exceptional IEEE
10749 values such as not-a-number and plus/minus infinity. Other Alpha
10750 compilers call this option @option{-ieee_with_no_inexact}.
10752 @item -mieee-with-inexact
10753 @opindex mieee-with-inexact
10754 This is like @option{-mieee} except the generated code also maintains
10755 the IEEE @var{inexact-flag}. Turning on this option causes the
10756 generated code to implement fully-compliant IEEE math. In addition to
10757 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
10758 macro. On some Alpha implementations the resulting code may execute
10759 significantly slower than the code generated by default. Since there is
10760 very little code that depends on the @var{inexact-flag}, you should
10761 normally not specify this option. Other Alpha compilers call this
10762 option @option{-ieee_with_inexact}.
10764 @item -mfp-trap-mode=@var{trap-mode}
10765 @opindex mfp-trap-mode
10766 This option controls what floating-point related traps are enabled.
10767 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
10768 The trap mode can be set to one of four values:
10772 This is the default (normal) setting. The only traps that are enabled
10773 are the ones that cannot be disabled in software (e.g., division by zero
10777 In addition to the traps enabled by @samp{n}, underflow traps are enabled
10781 Like @samp{u}, but the instructions are marked to be safe for software
10782 completion (see Alpha architecture manual for details).
10785 Like @samp{su}, but inexact traps are enabled as well.
10788 @item -mfp-rounding-mode=@var{rounding-mode}
10789 @opindex mfp-rounding-mode
10790 Selects the IEEE rounding mode. Other Alpha compilers call this option
10791 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
10796 Normal IEEE rounding mode. Floating point numbers are rounded towards
10797 the nearest machine number or towards the even machine number in case
10801 Round towards minus infinity.
10804 Chopped rounding mode. Floating point numbers are rounded towards zero.
10807 Dynamic rounding mode. A field in the floating point control register
10808 (@var{fpcr}, see Alpha architecture reference manual) controls the
10809 rounding mode in effect. The C library initializes this register for
10810 rounding towards plus infinity. Thus, unless your program modifies the
10811 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
10814 @item -mtrap-precision=@var{trap-precision}
10815 @opindex mtrap-precision
10816 In the Alpha architecture, floating point traps are imprecise. This
10817 means without software assistance it is impossible to recover from a
10818 floating trap and program execution normally needs to be terminated.
10819 GCC can generate code that can assist operating system trap handlers
10820 in determining the exact location that caused a floating point trap.
10821 Depending on the requirements of an application, different levels of
10822 precisions can be selected:
10826 Program precision. This option is the default and means a trap handler
10827 can only identify which program caused a floating point exception.
10830 Function precision. The trap handler can determine the function that
10831 caused a floating point exception.
10834 Instruction precision. The trap handler can determine the exact
10835 instruction that caused a floating point exception.
10838 Other Alpha compilers provide the equivalent options called
10839 @option{-scope_safe} and @option{-resumption_safe}.
10841 @item -mieee-conformant
10842 @opindex mieee-conformant
10843 This option marks the generated code as IEEE conformant. You must not
10844 use this option unless you also specify @option{-mtrap-precision=i} and either
10845 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
10846 is to emit the line @samp{.eflag 48} in the function prologue of the
10847 generated assembly file. Under DEC Unix, this has the effect that
10848 IEEE-conformant math library routines will be linked in.
10850 @item -mbuild-constants
10851 @opindex mbuild-constants
10852 Normally GCC examines a 32- or 64-bit integer constant to
10853 see if it can construct it from smaller constants in two or three
10854 instructions. If it cannot, it will output the constant as a literal and
10855 generate code to load it from the data segment at runtime.
10857 Use this option to require GCC to construct @emph{all} integer constants
10858 using code, even if it takes more instructions (the maximum is six).
10860 You would typically use this option to build a shared library dynamic
10861 loader. Itself a shared library, it must relocate itself in memory
10862 before it can find the variables and constants in its own data segment.
10868 Select whether to generate code to be assembled by the vendor-supplied
10869 assembler (@option{-malpha-as}) or by the GNU assembler @option{-mgas}.
10887 Indicate whether GCC should generate code to use the optional BWX,
10888 CIX, FIX and MAX instruction sets. The default is to use the instruction
10889 sets supported by the CPU type specified via @option{-mcpu=} option or that
10890 of the CPU on which GCC was built if none was specified.
10893 @itemx -mfloat-ieee
10894 @opindex mfloat-vax
10895 @opindex mfloat-ieee
10896 Generate code that uses (does not use) VAX F and G floating point
10897 arithmetic instead of IEEE single and double precision.
10899 @item -mexplicit-relocs
10900 @itemx -mno-explicit-relocs
10901 @opindex mexplicit-relocs
10902 @opindex mno-explicit-relocs
10903 Older Alpha assemblers provided no way to generate symbol relocations
10904 except via assembler macros. Use of these macros does not allow
10905 optimal instruction scheduling. GNU binutils as of version 2.12
10906 supports a new syntax that allows the compiler to explicitly mark
10907 which relocations should apply to which instructions. This option
10908 is mostly useful for debugging, as GCC detects the capabilities of
10909 the assembler when it is built and sets the default accordingly.
10912 @itemx -mlarge-data
10913 @opindex msmall-data
10914 @opindex mlarge-data
10915 When @option{-mexplicit-relocs} is in effect, static data is
10916 accessed via @dfn{gp-relative} relocations. When @option{-msmall-data}
10917 is used, objects 8 bytes long or smaller are placed in a @dfn{small data area}
10918 (the @code{.sdata} and @code{.sbss} sections) and are accessed via
10919 16-bit relocations off of the @code{$gp} register. This limits the
10920 size of the small data area to 64KB, but allows the variables to be
10921 directly accessed via a single instruction.
10923 The default is @option{-mlarge-data}. With this option the data area
10924 is limited to just below 2GB@. Programs that require more than 2GB of
10925 data must use @code{malloc} or @code{mmap} to allocate the data in the
10926 heap instead of in the program's data segment.
10928 When generating code for shared libraries, @option{-fpic} implies
10929 @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}.
10932 @itemx -mlarge-text
10933 @opindex msmall-text
10934 @opindex mlarge-text
10935 When @option{-msmall-text} is used, the compiler assumes that the
10936 code of the entire program (or shared library) fits in 4MB, and is
10937 thus reachable with a branch instruction. When @option{-msmall-data}
10938 is used, the compiler can assume that all local symbols share the
10939 same @code{$gp} value, and thus reduce the number of instructions
10940 required for a function call from 4 to 1.
10942 The default is @option{-mlarge-text}.
10944 @item -mcpu=@var{cpu_type}
10946 Set the instruction set and instruction scheduling parameters for
10947 machine type @var{cpu_type}. You can specify either the @samp{EV}
10948 style name or the corresponding chip number. GCC supports scheduling
10949 parameters for the EV4, EV5 and EV6 family of processors and will
10950 choose the default values for the instruction set from the processor
10951 you specify. If you do not specify a processor type, GCC will default
10952 to the processor on which the compiler was built.
10954 Supported values for @var{cpu_type} are
10960 Schedules as an EV4 and has no instruction set extensions.
10964 Schedules as an EV5 and has no instruction set extensions.
10968 Schedules as an EV5 and supports the BWX extension.
10973 Schedules as an EV5 and supports the BWX and MAX extensions.
10977 Schedules as an EV6 and supports the BWX, FIX, and MAX extensions.
10981 Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions.
10984 Native Linux/GNU toolchains also support the value @samp{native},
10985 which selects the best architecture option for the host processor.
10986 @option{-mcpu=native} has no effect if GCC does not recognize
10989 @item -mtune=@var{cpu_type}
10991 Set only the instruction scheduling parameters for machine type
10992 @var{cpu_type}. The instruction set is not changed.
10994 Native Linux/GNU toolchains also support the value @samp{native},
10995 which selects the best architecture option for the host processor.
10996 @option{-mtune=native} has no effect if GCC does not recognize
10999 @item -mmemory-latency=@var{time}
11000 @opindex mmemory-latency
11001 Sets the latency the scheduler should assume for typical memory
11002 references as seen by the application. This number is highly
11003 dependent on the memory access patterns used by the application
11004 and the size of the external cache on the machine.
11006 Valid options for @var{time} are
11010 A decimal number representing clock cycles.
11016 The compiler contains estimates of the number of clock cycles for
11017 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
11018 (also called Dcache, Scache, and Bcache), as well as to main memory.
11019 Note that L3 is only valid for EV5.
11024 @node DEC Alpha/VMS Options
11025 @subsection DEC Alpha/VMS Options
11027 These @samp{-m} options are defined for the DEC Alpha/VMS implementations:
11030 @item -mvms-return-codes
11031 @opindex mvms-return-codes
11032 Return VMS condition codes from main. The default is to return POSIX
11033 style condition (e.g.@: error) codes.
11035 @item -mdebug-main=@var{prefix}
11036 @opindex mdebug-main=@var{prefix}
11037 Flag the first routine whose name starts with @var{prefix} as the main
11038 routine for the debugger.
11042 Default to 64bit memory allocation routines.
11046 @subsection FR30 Options
11047 @cindex FR30 Options
11049 These options are defined specifically for the FR30 port.
11053 @item -msmall-model
11054 @opindex msmall-model
11055 Use the small address space model. This can produce smaller code, but
11056 it does assume that all symbolic values and addresses will fit into a
11061 Assume that run-time support has been provided and so there is no need
11062 to include the simulator library (@file{libsim.a}) on the linker
11068 @subsection FRV Options
11069 @cindex FRV Options
11075 Only use the first 32 general purpose registers.
11080 Use all 64 general purpose registers.
11085 Use only the first 32 floating point registers.
11090 Use all 64 floating point registers
11093 @opindex mhard-float
11095 Use hardware instructions for floating point operations.
11098 @opindex msoft-float
11100 Use library routines for floating point operations.
11105 Dynamically allocate condition code registers.
11110 Do not try to dynamically allocate condition code registers, only
11111 use @code{icc0} and @code{fcc0}.
11116 Change ABI to use double word insns.
11121 Do not use double word instructions.
11126 Use floating point double instructions.
11129 @opindex mno-double
11131 Do not use floating point double instructions.
11136 Use media instructions.
11141 Do not use media instructions.
11146 Use multiply and add/subtract instructions.
11149 @opindex mno-muladd
11151 Do not use multiply and add/subtract instructions.
11156 Select the FDPIC ABI, that uses function descriptors to represent
11157 pointers to functions. Without any PIC/PIE-related options, it
11158 implies @option{-fPIE}. With @option{-fpic} or @option{-fpie}, it
11159 assumes GOT entries and small data are within a 12-bit range from the
11160 GOT base address; with @option{-fPIC} or @option{-fPIE}, GOT offsets
11161 are computed with 32 bits.
11162 With a @samp{bfin-elf} target, this option implies @option{-msim}.
11165 @opindex minline-plt
11167 Enable inlining of PLT entries in function calls to functions that are
11168 not known to bind locally. It has no effect without @option{-mfdpic}.
11169 It's enabled by default if optimizing for speed and compiling for
11170 shared libraries (i.e., @option{-fPIC} or @option{-fpic}), or when an
11171 optimization option such as @option{-O3} or above is present in the
11177 Assume a large TLS segment when generating thread-local code.
11182 Do not assume a large TLS segment when generating thread-local code.
11187 Enable the use of @code{GPREL} relocations in the FDPIC ABI for data
11188 that is known to be in read-only sections. It's enabled by default,
11189 except for @option{-fpic} or @option{-fpie}: even though it may help
11190 make the global offset table smaller, it trades 1 instruction for 4.
11191 With @option{-fPIC} or @option{-fPIE}, it trades 3 instructions for 4,
11192 one of which may be shared by multiple symbols, and it avoids the need
11193 for a GOT entry for the referenced symbol, so it's more likely to be a
11194 win. If it is not, @option{-mno-gprel-ro} can be used to disable it.
11196 @item -multilib-library-pic
11197 @opindex multilib-library-pic
11199 Link with the (library, not FD) pic libraries. It's implied by
11200 @option{-mlibrary-pic}, as well as by @option{-fPIC} and
11201 @option{-fpic} without @option{-mfdpic}. You should never have to use
11205 @opindex mlinked-fp
11207 Follow the EABI requirement of always creating a frame pointer whenever
11208 a stack frame is allocated. This option is enabled by default and can
11209 be disabled with @option{-mno-linked-fp}.
11212 @opindex mlong-calls
11214 Use indirect addressing to call functions outside the current
11215 compilation unit. This allows the functions to be placed anywhere
11216 within the 32-bit address space.
11218 @item -malign-labels
11219 @opindex malign-labels
11221 Try to align labels to an 8-byte boundary by inserting nops into the
11222 previous packet. This option only has an effect when VLIW packing
11223 is enabled. It doesn't create new packets; it merely adds nops to
11226 @item -mlibrary-pic
11227 @opindex mlibrary-pic
11229 Generate position-independent EABI code.
11234 Use only the first four media accumulator registers.
11239 Use all eight media accumulator registers.
11244 Pack VLIW instructions.
11249 Do not pack VLIW instructions.
11252 @opindex mno-eflags
11254 Do not mark ABI switches in e_flags.
11257 @opindex mcond-move
11259 Enable the use of conditional-move instructions (default).
11261 This switch is mainly for debugging the compiler and will likely be removed
11262 in a future version.
11264 @item -mno-cond-move
11265 @opindex mno-cond-move
11267 Disable the use of conditional-move instructions.
11269 This switch is mainly for debugging the compiler and will likely be removed
11270 in a future version.
11275 Enable the use of conditional set instructions (default).
11277 This switch is mainly for debugging the compiler and will likely be removed
11278 in a future version.
11283 Disable the use of conditional set instructions.
11285 This switch is mainly for debugging the compiler and will likely be removed
11286 in a future version.
11289 @opindex mcond-exec
11291 Enable the use of conditional execution (default).
11293 This switch is mainly for debugging the compiler and will likely be removed
11294 in a future version.
11296 @item -mno-cond-exec
11297 @opindex mno-cond-exec
11299 Disable the use of conditional execution.
11301 This switch is mainly for debugging the compiler and will likely be removed
11302 in a future version.
11304 @item -mvliw-branch
11305 @opindex mvliw-branch
11307 Run a pass to pack branches into VLIW instructions (default).
11309 This switch is mainly for debugging the compiler and will likely be removed
11310 in a future version.
11312 @item -mno-vliw-branch
11313 @opindex mno-vliw-branch
11315 Do not run a pass to pack branches into VLIW instructions.
11317 This switch is mainly for debugging the compiler and will likely be removed
11318 in a future version.
11320 @item -mmulti-cond-exec
11321 @opindex mmulti-cond-exec
11323 Enable optimization of @code{&&} and @code{||} in conditional execution
11326 This switch is mainly for debugging the compiler and will likely be removed
11327 in a future version.
11329 @item -mno-multi-cond-exec
11330 @opindex mno-multi-cond-exec
11332 Disable optimization of @code{&&} and @code{||} in conditional execution.
11334 This switch is mainly for debugging the compiler and will likely be removed
11335 in a future version.
11337 @item -mnested-cond-exec
11338 @opindex mnested-cond-exec
11340 Enable nested conditional execution optimizations (default).
11342 This switch is mainly for debugging the compiler and will likely be removed
11343 in a future version.
11345 @item -mno-nested-cond-exec
11346 @opindex mno-nested-cond-exec
11348 Disable nested conditional execution optimizations.
11350 This switch is mainly for debugging the compiler and will likely be removed
11351 in a future version.
11353 @item -moptimize-membar
11354 @opindex moptimize-membar
11356 This switch removes redundant @code{membar} instructions from the
11357 compiler generated code. It is enabled by default.
11359 @item -mno-optimize-membar
11360 @opindex mno-optimize-membar
11362 This switch disables the automatic removal of redundant @code{membar}
11363 instructions from the generated code.
11365 @item -mtomcat-stats
11366 @opindex mtomcat-stats
11368 Cause gas to print out tomcat statistics.
11370 @item -mcpu=@var{cpu}
11373 Select the processor type for which to generate code. Possible values are
11374 @samp{frv}, @samp{fr550}, @samp{tomcat}, @samp{fr500}, @samp{fr450},
11375 @samp{fr405}, @samp{fr400}, @samp{fr300} and @samp{simple}.
11379 @node GNU/Linux Options
11380 @subsection GNU/Linux Options
11382 These @samp{-m} options are defined for GNU/Linux targets:
11387 Use the GNU C library instead of uClibc. This is the default except
11388 on @samp{*-*-linux-*uclibc*} targets.
11392 Use uClibc instead of the GNU C library. This is the default on
11393 @samp{*-*-linux-*uclibc*} targets.
11396 @node H8/300 Options
11397 @subsection H8/300 Options
11399 These @samp{-m} options are defined for the H8/300 implementations:
11404 Shorten some address references at link time, when possible; uses the
11405 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
11406 ld, Using ld}, for a fuller description.
11410 Generate code for the H8/300H@.
11414 Generate code for the H8S@.
11418 Generate code for the H8S and H8/300H in the normal mode. This switch
11419 must be used either with @option{-mh} or @option{-ms}.
11423 Generate code for the H8S/2600. This switch must be used with @option{-ms}.
11427 Make @code{int} data 32 bits by default.
11430 @opindex malign-300
11431 On the H8/300H and H8S, use the same alignment rules as for the H8/300.
11432 The default for the H8/300H and H8S is to align longs and floats on 4
11434 @option{-malign-300} causes them to be aligned on 2 byte boundaries.
11435 This option has no effect on the H8/300.
11439 @subsection HPPA Options
11440 @cindex HPPA Options
11442 These @samp{-m} options are defined for the HPPA family of computers:
11445 @item -march=@var{architecture-type}
11447 Generate code for the specified architecture. The choices for
11448 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
11449 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
11450 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
11451 architecture option for your machine. Code compiled for lower numbered
11452 architectures will run on higher numbered architectures, but not the
11455 @item -mpa-risc-1-0
11456 @itemx -mpa-risc-1-1
11457 @itemx -mpa-risc-2-0
11458 @opindex mpa-risc-1-0
11459 @opindex mpa-risc-1-1
11460 @opindex mpa-risc-2-0
11461 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
11464 @opindex mbig-switch
11465 Generate code suitable for big switch tables. Use this option only if
11466 the assembler/linker complain about out of range branches within a switch
11469 @item -mjump-in-delay
11470 @opindex mjump-in-delay
11471 Fill delay slots of function calls with unconditional jump instructions
11472 by modifying the return pointer for the function call to be the target
11473 of the conditional jump.
11475 @item -mdisable-fpregs
11476 @opindex mdisable-fpregs
11477 Prevent floating point registers from being used in any manner. This is
11478 necessary for compiling kernels which perform lazy context switching of
11479 floating point registers. If you use this option and attempt to perform
11480 floating point operations, the compiler will abort.
11482 @item -mdisable-indexing
11483 @opindex mdisable-indexing
11484 Prevent the compiler from using indexing address modes. This avoids some
11485 rather obscure problems when compiling MIG generated code under MACH@.
11487 @item -mno-space-regs
11488 @opindex mno-space-regs
11489 Generate code that assumes the target has no space registers. This allows
11490 GCC to generate faster indirect calls and use unscaled index address modes.
11492 Such code is suitable for level 0 PA systems and kernels.
11494 @item -mfast-indirect-calls
11495 @opindex mfast-indirect-calls
11496 Generate code that assumes calls never cross space boundaries. This
11497 allows GCC to emit code which performs faster indirect calls.
11499 This option will not work in the presence of shared libraries or nested
11502 @item -mfixed-range=@var{register-range}
11503 @opindex mfixed-range
11504 Generate code treating the given register range as fixed registers.
11505 A fixed register is one that the register allocator can not use. This is
11506 useful when compiling kernel code. A register range is specified as
11507 two registers separated by a dash. Multiple register ranges can be
11508 specified separated by a comma.
11510 @item -mlong-load-store
11511 @opindex mlong-load-store
11512 Generate 3-instruction load and store sequences as sometimes required by
11513 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
11516 @item -mportable-runtime
11517 @opindex mportable-runtime
11518 Use the portable calling conventions proposed by HP for ELF systems.
11522 Enable the use of assembler directives only GAS understands.
11524 @item -mschedule=@var{cpu-type}
11526 Schedule code according to the constraints for the machine type
11527 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
11528 @samp{7100}, @samp{7100LC}, @samp{7200}, @samp{7300} and @samp{8000}. Refer
11529 to @file{/usr/lib/sched.models} on an HP-UX system to determine the
11530 proper scheduling option for your machine. The default scheduling is
11534 @opindex mlinker-opt
11535 Enable the optimization pass in the HP-UX linker. Note this makes symbolic
11536 debugging impossible. It also triggers a bug in the HP-UX 8 and HP-UX 9
11537 linkers in which they give bogus error messages when linking some programs.
11540 @opindex msoft-float
11541 Generate output containing library calls for floating point.
11542 @strong{Warning:} the requisite libraries are not available for all HPPA
11543 targets. Normally the facilities of the machine's usual C compiler are
11544 used, but this cannot be done directly in cross-compilation. You must make
11545 your own arrangements to provide suitable library functions for
11548 @option{-msoft-float} changes the calling convention in the output file;
11549 therefore, it is only useful if you compile @emph{all} of a program with
11550 this option. In particular, you need to compile @file{libgcc.a}, the
11551 library that comes with GCC, with @option{-msoft-float} in order for
11556 Generate the predefine, @code{_SIO}, for server IO@. The default is
11557 @option{-mwsio}. This generates the predefines, @code{__hp9000s700},
11558 @code{__hp9000s700__} and @code{_WSIO}, for workstation IO@. These
11559 options are available under HP-UX and HI-UX@.
11563 Use GNU ld specific options. This passes @option{-shared} to ld when
11564 building a shared library. It is the default when GCC is configured,
11565 explicitly or implicitly, with the GNU linker. This option does not
11566 have any affect on which ld is called, it only changes what parameters
11567 are passed to that ld. The ld that is called is determined by the
11568 @option{--with-ld} configure option, GCC's program search path, and
11569 finally by the user's @env{PATH}. The linker used by GCC can be printed
11570 using @samp{which `gcc -print-prog-name=ld`}. This option is only available
11571 on the 64 bit HP-UX GCC, i.e.@: configured with @samp{hppa*64*-*-hpux*}.
11575 Use HP ld specific options. This passes @option{-b} to ld when building
11576 a shared library and passes @option{+Accept TypeMismatch} to ld on all
11577 links. It is the default when GCC is configured, explicitly or
11578 implicitly, with the HP linker. This option does not have any affect on
11579 which ld is called, it only changes what parameters are passed to that
11580 ld. The ld that is called is determined by the @option{--with-ld}
11581 configure option, GCC's program search path, and finally by the user's
11582 @env{PATH}. The linker used by GCC can be printed using @samp{which
11583 `gcc -print-prog-name=ld`}. This option is only available on the 64 bit
11584 HP-UX GCC, i.e.@: configured with @samp{hppa*64*-*-hpux*}.
11587 @opindex mno-long-calls
11588 Generate code that uses long call sequences. This ensures that a call
11589 is always able to reach linker generated stubs. The default is to generate
11590 long calls only when the distance from the call site to the beginning
11591 of the function or translation unit, as the case may be, exceeds a
11592 predefined limit set by the branch type being used. The limits for
11593 normal calls are 7,600,000 and 240,000 bytes, respectively for the
11594 PA 2.0 and PA 1.X architectures. Sibcalls are always limited at
11597 Distances are measured from the beginning of functions when using the
11598 @option{-ffunction-sections} option, or when using the @option{-mgas}
11599 and @option{-mno-portable-runtime} options together under HP-UX with
11602 It is normally not desirable to use this option as it will degrade
11603 performance. However, it may be useful in large applications,
11604 particularly when partial linking is used to build the application.
11606 The types of long calls used depends on the capabilities of the
11607 assembler and linker, and the type of code being generated. The
11608 impact on systems that support long absolute calls, and long pic
11609 symbol-difference or pc-relative calls should be relatively small.
11610 However, an indirect call is used on 32-bit ELF systems in pic code
11611 and it is quite long.
11613 @item -munix=@var{unix-std}
11615 Generate compiler predefines and select a startfile for the specified
11616 UNIX standard. The choices for @var{unix-std} are @samp{93}, @samp{95}
11617 and @samp{98}. @samp{93} is supported on all HP-UX versions. @samp{95}
11618 is available on HP-UX 10.10 and later. @samp{98} is available on HP-UX
11619 11.11 and later. The default values are @samp{93} for HP-UX 10.00,
11620 @samp{95} for HP-UX 10.10 though to 11.00, and @samp{98} for HP-UX 11.11
11623 @option{-munix=93} provides the same predefines as GCC 3.3 and 3.4.
11624 @option{-munix=95} provides additional predefines for @code{XOPEN_UNIX}
11625 and @code{_XOPEN_SOURCE_EXTENDED}, and the startfile @file{unix95.o}.
11626 @option{-munix=98} provides additional predefines for @code{_XOPEN_UNIX},
11627 @code{_XOPEN_SOURCE_EXTENDED}, @code{_INCLUDE__STDC_A1_SOURCE} and
11628 @code{_INCLUDE_XOPEN_SOURCE_500}, and the startfile @file{unix98.o}.
11630 It is @emph{important} to note that this option changes the interfaces
11631 for various library routines. It also affects the operational behavior
11632 of the C library. Thus, @emph{extreme} care is needed in using this
11635 Library code that is intended to operate with more than one UNIX
11636 standard must test, set and restore the variable @var{__xpg4_extended_mask}
11637 as appropriate. Most GNU software doesn't provide this capability.
11641 Suppress the generation of link options to search libdld.sl when the
11642 @option{-static} option is specified on HP-UX 10 and later.
11646 The HP-UX implementation of setlocale in libc has a dependency on
11647 libdld.sl. There isn't an archive version of libdld.sl. Thus,
11648 when the @option{-static} option is specified, special link options
11649 are needed to resolve this dependency.
11651 On HP-UX 10 and later, the GCC driver adds the necessary options to
11652 link with libdld.sl when the @option{-static} option is specified.
11653 This causes the resulting binary to be dynamic. On the 64-bit port,
11654 the linkers generate dynamic binaries by default in any case. The
11655 @option{-nolibdld} option can be used to prevent the GCC driver from
11656 adding these link options.
11660 Add support for multithreading with the @dfn{dce thread} library
11661 under HP-UX@. This option sets flags for both the preprocessor and
11665 @node i386 and x86-64 Options
11666 @subsection Intel 386 and AMD x86-64 Options
11667 @cindex i386 Options
11668 @cindex x86-64 Options
11669 @cindex Intel 386 Options
11670 @cindex AMD x86-64 Options
11672 These @samp{-m} options are defined for the i386 and x86-64 family of
11676 @item -mtune=@var{cpu-type}
11678 Tune to @var{cpu-type} everything applicable about the generated code, except
11679 for the ABI and the set of available instructions. The choices for
11680 @var{cpu-type} are:
11683 Produce code optimized for the most common IA32/AMD64/EM64T processors.
11684 If you know the CPU on which your code will run, then you should use
11685 the corresponding @option{-mtune} option instead of
11686 @option{-mtune=generic}. But, if you do not know exactly what CPU users
11687 of your application will have, then you should use this option.
11689 As new processors are deployed in the marketplace, the behavior of this
11690 option will change. Therefore, if you upgrade to a newer version of
11691 GCC, the code generated option will change to reflect the processors
11692 that were most common when that version of GCC was released.
11694 There is no @option{-march=generic} option because @option{-march}
11695 indicates the instruction set the compiler can use, and there is no
11696 generic instruction set applicable to all processors. In contrast,
11697 @option{-mtune} indicates the processor (or, in this case, collection of
11698 processors) for which the code is optimized.
11700 This selects the CPU to tune for at compilation time by determining
11701 the processor type of the compiling machine. Using @option{-mtune=native}
11702 will produce code optimized for the local machine under the constraints
11703 of the selected instruction set. Using @option{-march=native} will
11704 enable all instruction subsets supported by the local machine (hence
11705 the result might not run on different machines).
11707 Original Intel's i386 CPU@.
11709 Intel's i486 CPU@. (No scheduling is implemented for this chip.)
11710 @item i586, pentium
11711 Intel Pentium CPU with no MMX support.
11713 Intel PentiumMMX CPU based on Pentium core with MMX instruction set support.
11715 Intel PentiumPro CPU@.
11717 Same as @code{generic}, but when used as @code{march} option, PentiumPro
11718 instruction set will be used, so the code will run on all i686 family chips.
11720 Intel Pentium2 CPU based on PentiumPro core with MMX instruction set support.
11721 @item pentium3, pentium3m
11722 Intel Pentium3 CPU based on PentiumPro core with MMX and SSE instruction set
11725 Low power version of Intel Pentium3 CPU with MMX, SSE and SSE2 instruction set
11726 support. Used by Centrino notebooks.
11727 @item pentium4, pentium4m
11728 Intel Pentium4 CPU with MMX, SSE and SSE2 instruction set support.
11730 Improved version of Intel Pentium4 CPU with MMX, SSE, SSE2 and SSE3 instruction
11733 Improved version of Intel Pentium4 CPU with 64-bit extensions, MMX, SSE,
11734 SSE2 and SSE3 instruction set support.
11736 Intel Core2 CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3 and SSSE3
11737 instruction set support.
11739 Intel Atom CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3 and SSSE3
11740 instruction set support.
11742 AMD K6 CPU with MMX instruction set support.
11744 Improved versions of AMD K6 CPU with MMX and 3DNow!@: instruction set support.
11745 @item athlon, athlon-tbird
11746 AMD Athlon CPU with MMX, 3dNOW!, enhanced 3DNow!@: and SSE prefetch instructions
11748 @item athlon-4, athlon-xp, athlon-mp
11749 Improved AMD Athlon CPU with MMX, 3DNow!, enhanced 3DNow!@: and full SSE
11750 instruction set support.
11751 @item k8, opteron, athlon64, athlon-fx
11752 AMD K8 core based CPUs with x86-64 instruction set support. (This supersets
11753 MMX, SSE, SSE2, 3DNow!, enhanced 3DNow!@: and 64-bit instruction set extensions.)
11754 @item k8-sse3, opteron-sse3, athlon64-sse3
11755 Improved versions of k8, opteron and athlon64 with SSE3 instruction set support.
11756 @item amdfam10, barcelona
11757 AMD Family 10h core based CPUs with x86-64 instruction set support. (This
11758 supersets MMX, SSE, SSE2, SSE3, SSE4A, 3DNow!, enhanced 3DNow!, ABM and 64-bit
11759 instruction set extensions.)
11761 IDT Winchip C6 CPU, dealt in same way as i486 with additional MMX instruction
11764 IDT Winchip2 CPU, dealt in same way as i486 with additional MMX and 3DNow!@:
11765 instruction set support.
11767 Via C3 CPU with MMX and 3DNow!@: instruction set support. (No scheduling is
11768 implemented for this chip.)
11770 Via C3-2 CPU with MMX and SSE instruction set support. (No scheduling is
11771 implemented for this chip.)
11773 Embedded AMD CPU with MMX and 3DNow!@: instruction set support.
11776 While picking a specific @var{cpu-type} will schedule things appropriately
11777 for that particular chip, the compiler will not generate any code that
11778 does not run on the i386 without the @option{-march=@var{cpu-type}} option
11781 @item -march=@var{cpu-type}
11783 Generate instructions for the machine type @var{cpu-type}. The choices
11784 for @var{cpu-type} are the same as for @option{-mtune}. Moreover,
11785 specifying @option{-march=@var{cpu-type}} implies @option{-mtune=@var{cpu-type}}.
11787 @item -mcpu=@var{cpu-type}
11789 A deprecated synonym for @option{-mtune}.
11791 @item -mfpmath=@var{unit}
11793 Generate floating point arithmetics for selected unit @var{unit}. The choices
11794 for @var{unit} are:
11798 Use the standard 387 floating point coprocessor present majority of chips and
11799 emulated otherwise. Code compiled with this option will run almost everywhere.
11800 The temporary results are computed in 80bit precision instead of precision
11801 specified by the type resulting in slightly different results compared to most
11802 of other chips. See @option{-ffloat-store} for more detailed description.
11804 This is the default choice for i386 compiler.
11807 Use scalar floating point instructions present in the SSE instruction set.
11808 This instruction set is supported by Pentium3 and newer chips, in the AMD line
11809 by Athlon-4, Athlon-xp and Athlon-mp chips. The earlier version of SSE
11810 instruction set supports only single precision arithmetics, thus the double and
11811 extended precision arithmetics is still done using 387. Later version, present
11812 only in Pentium4 and the future AMD x86-64 chips supports double precision
11815 For the i386 compiler, you need to use @option{-march=@var{cpu-type}}, @option{-msse}
11816 or @option{-msse2} switches to enable SSE extensions and make this option
11817 effective. For the x86-64 compiler, these extensions are enabled by default.
11819 The resulting code should be considerably faster in the majority of cases and avoid
11820 the numerical instability problems of 387 code, but may break some existing
11821 code that expects temporaries to be 80bit.
11823 This is the default choice for the x86-64 compiler.
11828 Attempt to utilize both instruction sets at once. This effectively double the
11829 amount of available registers and on chips with separate execution units for
11830 387 and SSE the execution resources too. Use this option with care, as it is
11831 still experimental, because the GCC register allocator does not model separate
11832 functional units well resulting in instable performance.
11835 @item -masm=@var{dialect}
11836 @opindex masm=@var{dialect}
11837 Output asm instructions using selected @var{dialect}. Supported
11838 choices are @samp{intel} or @samp{att} (the default one). Darwin does
11839 not support @samp{intel}.
11842 @itemx -mno-ieee-fp
11844 @opindex mno-ieee-fp
11845 Control whether or not the compiler uses IEEE floating point
11846 comparisons. These handle correctly the case where the result of a
11847 comparison is unordered.
11850 @opindex msoft-float
11851 Generate output containing library calls for floating point.
11852 @strong{Warning:} the requisite libraries are not part of GCC@.
11853 Normally the facilities of the machine's usual C compiler are used, but
11854 this can't be done directly in cross-compilation. You must make your
11855 own arrangements to provide suitable library functions for
11858 On machines where a function returns floating point results in the 80387
11859 register stack, some floating point opcodes may be emitted even if
11860 @option{-msoft-float} is used.
11862 @item -mno-fp-ret-in-387
11863 @opindex mno-fp-ret-in-387
11864 Do not use the FPU registers for return values of functions.
11866 The usual calling convention has functions return values of types
11867 @code{float} and @code{double} in an FPU register, even if there
11868 is no FPU@. The idea is that the operating system should emulate
11871 The option @option{-mno-fp-ret-in-387} causes such values to be returned
11872 in ordinary CPU registers instead.
11874 @item -mno-fancy-math-387
11875 @opindex mno-fancy-math-387
11876 Some 387 emulators do not support the @code{sin}, @code{cos} and
11877 @code{sqrt} instructions for the 387. Specify this option to avoid
11878 generating those instructions. This option is the default on FreeBSD,
11879 OpenBSD and NetBSD@. This option is overridden when @option{-march}
11880 indicates that the target cpu will always have an FPU and so the
11881 instruction will not need emulation. As of revision 2.6.1, these
11882 instructions are not generated unless you also use the
11883 @option{-funsafe-math-optimizations} switch.
11885 @item -malign-double
11886 @itemx -mno-align-double
11887 @opindex malign-double
11888 @opindex mno-align-double
11889 Control whether GCC aligns @code{double}, @code{long double}, and
11890 @code{long long} variables on a two word boundary or a one word
11891 boundary. Aligning @code{double} variables on a two word boundary will
11892 produce code that runs somewhat faster on a @samp{Pentium} at the
11893 expense of more memory.
11895 On x86-64, @option{-malign-double} is enabled by default.
11897 @strong{Warning:} if you use the @option{-malign-double} switch,
11898 structures containing the above types will be aligned differently than
11899 the published application binary interface specifications for the 386
11900 and will not be binary compatible with structures in code compiled
11901 without that switch.
11903 @item -m96bit-long-double
11904 @itemx -m128bit-long-double
11905 @opindex m96bit-long-double
11906 @opindex m128bit-long-double
11907 These switches control the size of @code{long double} type. The i386
11908 application binary interface specifies the size to be 96 bits,
11909 so @option{-m96bit-long-double} is the default in 32 bit mode.
11911 Modern architectures (Pentium and newer) would prefer @code{long double}
11912 to be aligned to an 8 or 16 byte boundary. In arrays or structures
11913 conforming to the ABI, this would not be possible. So specifying a
11914 @option{-m128bit-long-double} will align @code{long double}
11915 to a 16 byte boundary by padding the @code{long double} with an additional
11918 In the x86-64 compiler, @option{-m128bit-long-double} is the default choice as
11919 its ABI specifies that @code{long double} is to be aligned on 16 byte boundary.
11921 Notice that neither of these options enable any extra precision over the x87
11922 standard of 80 bits for a @code{long double}.
11924 @strong{Warning:} if you override the default value for your target ABI, the
11925 structures and arrays containing @code{long double} variables will change
11926 their size as well as function calling convention for function taking
11927 @code{long double} will be modified. Hence they will not be binary
11928 compatible with arrays or structures in code compiled without that switch.
11930 @item -mlarge-data-threshold=@var{number}
11931 @opindex mlarge-data-threshold=@var{number}
11932 When @option{-mcmodel=medium} is specified, the data greater than
11933 @var{threshold} are placed in large data section. This value must be the
11934 same across all object linked into the binary and defaults to 65535.
11938 Use a different function-calling convention, in which functions that
11939 take a fixed number of arguments return with the @code{ret} @var{num}
11940 instruction, which pops their arguments while returning. This saves one
11941 instruction in the caller since there is no need to pop the arguments
11944 You can specify that an individual function is called with this calling
11945 sequence with the function attribute @samp{stdcall}. You can also
11946 override the @option{-mrtd} option by using the function attribute
11947 @samp{cdecl}. @xref{Function Attributes}.
11949 @strong{Warning:} this calling convention is incompatible with the one
11950 normally used on Unix, so you cannot use it if you need to call
11951 libraries compiled with the Unix compiler.
11953 Also, you must provide function prototypes for all functions that
11954 take variable numbers of arguments (including @code{printf});
11955 otherwise incorrect code will be generated for calls to those
11958 In addition, seriously incorrect code will result if you call a
11959 function with too many arguments. (Normally, extra arguments are
11960 harmlessly ignored.)
11962 @item -mregparm=@var{num}
11964 Control how many registers are used to pass integer arguments. By
11965 default, no registers are used to pass arguments, and at most 3
11966 registers can be used. You can control this behavior for a specific
11967 function by using the function attribute @samp{regparm}.
11968 @xref{Function Attributes}.
11970 @strong{Warning:} if you use this switch, and
11971 @var{num} is nonzero, then you must build all modules with the same
11972 value, including any libraries. This includes the system libraries and
11976 @opindex msseregparm
11977 Use SSE register passing conventions for float and double arguments
11978 and return values. You can control this behavior for a specific
11979 function by using the function attribute @samp{sseregparm}.
11980 @xref{Function Attributes}.
11982 @strong{Warning:} if you use this switch then you must build all
11983 modules with the same value, including any libraries. This includes
11984 the system libraries and startup modules.
11993 Set 80387 floating-point precision to 32, 64 or 80 bits. When @option{-mpc32}
11994 is specified, the significands of results of floating-point operations are
11995 rounded to 24 bits (single precision); @option{-mpc64} rounds the
11996 significands of results of floating-point operations to 53 bits (double
11997 precision) and @option{-mpc80} rounds the significands of results of
11998 floating-point operations to 64 bits (extended double precision), which is
11999 the default. When this option is used, floating-point operations in higher
12000 precisions are not available to the programmer without setting the FPU
12001 control word explicitly.
12003 Setting the rounding of floating-point operations to less than the default
12004 80 bits can speed some programs by 2% or more. Note that some mathematical
12005 libraries assume that extended precision (80 bit) floating-point operations
12006 are enabled by default; routines in such libraries could suffer significant
12007 loss of accuracy, typically through so-called "catastrophic cancellation",
12008 when this option is used to set the precision to less than extended precision.
12010 @item -mstackrealign
12011 @opindex mstackrealign
12012 Realign the stack at entry. On the Intel x86, the @option{-mstackrealign}
12013 option will generate an alternate prologue and epilogue that realigns the
12014 runtime stack if necessary. This supports mixing legacy codes that keep
12015 a 4-byte aligned stack with modern codes that keep a 16-byte stack for
12016 SSE compatibility. See also the attribute @code{force_align_arg_pointer},
12017 applicable to individual functions.
12019 @item -mpreferred-stack-boundary=@var{num}
12020 @opindex mpreferred-stack-boundary
12021 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
12022 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
12023 the default is 4 (16 bytes or 128 bits).
12025 @item -mincoming-stack-boundary=@var{num}
12026 @opindex mincoming-stack-boundary
12027 Assume the incoming stack is aligned to a 2 raised to @var{num} byte
12028 boundary. If @option{-mincoming-stack-boundary} is not specified,
12029 the one specified by @option{-mpreferred-stack-boundary} will be used.
12031 On Pentium and PentiumPro, @code{double} and @code{long double} values
12032 should be aligned to an 8 byte boundary (see @option{-malign-double}) or
12033 suffer significant run time performance penalties. On Pentium III, the
12034 Streaming SIMD Extension (SSE) data type @code{__m128} may not work
12035 properly if it is not 16 byte aligned.
12037 To ensure proper alignment of this values on the stack, the stack boundary
12038 must be as aligned as that required by any value stored on the stack.
12039 Further, every function must be generated such that it keeps the stack
12040 aligned. Thus calling a function compiled with a higher preferred
12041 stack boundary from a function compiled with a lower preferred stack
12042 boundary will most likely misalign the stack. It is recommended that
12043 libraries that use callbacks always use the default setting.
12045 This extra alignment does consume extra stack space, and generally
12046 increases code size. Code that is sensitive to stack space usage, such
12047 as embedded systems and operating system kernels, may want to reduce the
12048 preferred alignment to @option{-mpreferred-stack-boundary=2}.
12092 These switches enable or disable the use of instructions in the MMX,
12093 SSE, SSE2, SSE3, SSSE3, SSE4.1, AVX, AES, PCLMUL, SSE4A, FMA4, XOP,
12094 LWP, ABM or 3DNow!@: extended instruction sets.
12095 These extensions are also available as built-in functions: see
12096 @ref{X86 Built-in Functions}, for details of the functions enabled and
12097 disabled by these switches.
12099 To have SSE/SSE2 instructions generated automatically from floating-point
12100 code (as opposed to 387 instructions), see @option{-mfpmath=sse}.
12102 GCC depresses SSEx instructions when @option{-mavx} is used. Instead, it
12103 generates new AVX instructions or AVX equivalence for all SSEx instructions
12106 These options will enable GCC to use these extended instructions in
12107 generated code, even without @option{-mfpmath=sse}. Applications which
12108 perform runtime CPU detection must compile separate files for each
12109 supported architecture, using the appropriate flags. In particular,
12110 the file containing the CPU detection code should be compiled without
12114 @itemx -mno-fused-madd
12115 @opindex mfused-madd
12116 @opindex mno-fused-madd
12117 Do (don't) generate code that uses the fused multiply/add or multiply/subtract
12118 instructions. The default is to use these instructions.
12122 This option instructs GCC to emit a @code{cld} instruction in the prologue
12123 of functions that use string instructions. String instructions depend on
12124 the DF flag to select between autoincrement or autodecrement mode. While the
12125 ABI specifies the DF flag to be cleared on function entry, some operating
12126 systems violate this specification by not clearing the DF flag in their
12127 exception dispatchers. The exception handler can be invoked with the DF flag
12128 set which leads to wrong direction mode, when string instructions are used.
12129 This option can be enabled by default on 32-bit x86 targets by configuring
12130 GCC with the @option{--enable-cld} configure option. Generation of @code{cld}
12131 instructions can be suppressed with the @option{-mno-cld} compiler option
12136 This option will enable GCC to use CMPXCHG16B instruction in generated code.
12137 CMPXCHG16B allows for atomic operations on 128-bit double quadword (or oword)
12138 data types. This is useful for high resolution counters that could be updated
12139 by multiple processors (or cores). This instruction is generated as part of
12140 atomic built-in functions: see @ref{Atomic Builtins} for details.
12144 This option will enable GCC to use SAHF instruction in generated 64-bit code.
12145 Early Intel CPUs with Intel 64 lacked LAHF and SAHF instructions supported
12146 by AMD64 until introduction of Pentium 4 G1 step in December 2005. LAHF and
12147 SAHF are load and store instructions, respectively, for certain status flags.
12148 In 64-bit mode, SAHF instruction is used to optimize @code{fmod}, @code{drem}
12149 or @code{remainder} built-in functions: see @ref{Other Builtins} for details.
12153 This option will enable GCC to use movbe instruction to implement
12154 @code{__builtin_bswap32} and @code{__builtin_bswap64}.
12158 This option will enable built-in functions, @code{__builtin_ia32_crc32qi},
12159 @code{__builtin_ia32_crc32hi}. @code{__builtin_ia32_crc32si} and
12160 @code{__builtin_ia32_crc32di} to generate the crc32 machine instruction.
12164 This option will enable GCC to use RCPSS and RSQRTSS instructions (and their
12165 vectorized variants RCPPS and RSQRTPS) with an additional Newton-Raphson step
12166 to increase precision instead of DIVSS and SQRTSS (and their vectorized
12167 variants) for single precision floating point arguments. These instructions
12168 are generated only when @option{-funsafe-math-optimizations} is enabled
12169 together with @option{-finite-math-only} and @option{-fno-trapping-math}.
12170 Note that while the throughput of the sequence is higher than the throughput
12171 of the non-reciprocal instruction, the precision of the sequence can be
12172 decreased by up to 2 ulp (i.e. the inverse of 1.0 equals 0.99999994).
12174 Note that GCC implements 1.0f/sqrtf(x) in terms of RSQRTSS (or RSQRTPS)
12175 already with @option{-ffast-math} (or the above option combination), and
12176 doesn't need @option{-mrecip}.
12178 @item -mveclibabi=@var{type}
12179 @opindex mveclibabi
12180 Specifies the ABI type to use for vectorizing intrinsics using an
12181 external library. Supported types are @code{svml} for the Intel short
12182 vector math library and @code{acml} for the AMD math core library style
12183 of interfacing. GCC will currently emit calls to @code{vmldExp2},
12184 @code{vmldLn2}, @code{vmldLog102}, @code{vmldLog102}, @code{vmldPow2},
12185 @code{vmldTanh2}, @code{vmldTan2}, @code{vmldAtan2}, @code{vmldAtanh2},
12186 @code{vmldCbrt2}, @code{vmldSinh2}, @code{vmldSin2}, @code{vmldAsinh2},
12187 @code{vmldAsin2}, @code{vmldCosh2}, @code{vmldCos2}, @code{vmldAcosh2},
12188 @code{vmldAcos2}, @code{vmlsExp4}, @code{vmlsLn4}, @code{vmlsLog104},
12189 @code{vmlsLog104}, @code{vmlsPow4}, @code{vmlsTanh4}, @code{vmlsTan4},
12190 @code{vmlsAtan4}, @code{vmlsAtanh4}, @code{vmlsCbrt4}, @code{vmlsSinh4},
12191 @code{vmlsSin4}, @code{vmlsAsinh4}, @code{vmlsAsin4}, @code{vmlsCosh4},
12192 @code{vmlsCos4}, @code{vmlsAcosh4} and @code{vmlsAcos4} for corresponding
12193 function type when @option{-mveclibabi=svml} is used and @code{__vrd2_sin},
12194 @code{__vrd2_cos}, @code{__vrd2_exp}, @code{__vrd2_log}, @code{__vrd2_log2},
12195 @code{__vrd2_log10}, @code{__vrs4_sinf}, @code{__vrs4_cosf},
12196 @code{__vrs4_expf}, @code{__vrs4_logf}, @code{__vrs4_log2f},
12197 @code{__vrs4_log10f} and @code{__vrs4_powf} for corresponding function type
12198 when @option{-mveclibabi=acml} is used. Both @option{-ftree-vectorize} and
12199 @option{-funsafe-math-optimizations} have to be enabled. A SVML or ACML ABI
12200 compatible library will have to be specified at link time.
12202 @item -mabi=@var{name}
12204 Generate code for the specified calling convention. Permissible values
12205 are: @samp{sysv} for the ABI used on GNU/Linux and other systems and
12206 @samp{ms} for the Microsoft ABI. The default is to use the Microsoft
12207 ABI when targeting Windows. On all other systems, the default is the
12208 SYSV ABI. You can control this behavior for a specific function by
12209 using the function attribute @samp{ms_abi}/@samp{sysv_abi}.
12210 @xref{Function Attributes}.
12213 @itemx -mno-push-args
12214 @opindex mpush-args
12215 @opindex mno-push-args
12216 Use PUSH operations to store outgoing parameters. This method is shorter
12217 and usually equally fast as method using SUB/MOV operations and is enabled
12218 by default. In some cases disabling it may improve performance because of
12219 improved scheduling and reduced dependencies.
12221 @item -maccumulate-outgoing-args
12222 @opindex maccumulate-outgoing-args
12223 If enabled, the maximum amount of space required for outgoing arguments will be
12224 computed in the function prologue. This is faster on most modern CPUs
12225 because of reduced dependencies, improved scheduling and reduced stack usage
12226 when preferred stack boundary is not equal to 2. The drawback is a notable
12227 increase in code size. This switch implies @option{-mno-push-args}.
12231 Support thread-safe exception handling on @samp{Mingw32}. Code that relies
12232 on thread-safe exception handling must compile and link all code with the
12233 @option{-mthreads} option. When compiling, @option{-mthreads} defines
12234 @option{-D_MT}; when linking, it links in a special thread helper library
12235 @option{-lmingwthrd} which cleans up per thread exception handling data.
12237 @item -mno-align-stringops
12238 @opindex mno-align-stringops
12239 Do not align destination of inlined string operations. This switch reduces
12240 code size and improves performance in case the destination is already aligned,
12241 but GCC doesn't know about it.
12243 @item -minline-all-stringops
12244 @opindex minline-all-stringops
12245 By default GCC inlines string operations only when destination is known to be
12246 aligned at least to 4 byte boundary. This enables more inlining, increase code
12247 size, but may improve performance of code that depends on fast memcpy, strlen
12248 and memset for short lengths.
12250 @item -minline-stringops-dynamically
12251 @opindex minline-stringops-dynamically
12252 For string operation of unknown size, inline runtime checks so for small
12253 blocks inline code is used, while for large blocks library call is used.
12255 @item -mstringop-strategy=@var{alg}
12256 @opindex mstringop-strategy=@var{alg}
12257 Overwrite internal decision heuristic about particular algorithm to inline
12258 string operation with. The allowed values are @code{rep_byte},
12259 @code{rep_4byte}, @code{rep_8byte} for expanding using i386 @code{rep} prefix
12260 of specified size, @code{byte_loop}, @code{loop}, @code{unrolled_loop} for
12261 expanding inline loop, @code{libcall} for always expanding library call.
12263 @item -momit-leaf-frame-pointer
12264 @opindex momit-leaf-frame-pointer
12265 Don't keep the frame pointer in a register for leaf functions. This
12266 avoids the instructions to save, set up and restore frame pointers and
12267 makes an extra register available in leaf functions. The option
12268 @option{-fomit-frame-pointer} removes the frame pointer for all functions
12269 which might make debugging harder.
12271 @item -mtls-direct-seg-refs
12272 @itemx -mno-tls-direct-seg-refs
12273 @opindex mtls-direct-seg-refs
12274 Controls whether TLS variables may be accessed with offsets from the
12275 TLS segment register (@code{%gs} for 32-bit, @code{%fs} for 64-bit),
12276 or whether the thread base pointer must be added. Whether or not this
12277 is legal depends on the operating system, and whether it maps the
12278 segment to cover the entire TLS area.
12280 For systems that use GNU libc, the default is on.
12283 @itemx -mno-sse2avx
12285 Specify that the assembler should encode SSE instructions with VEX
12286 prefix. The option @option{-mavx} turns this on by default.
12289 These @samp{-m} switches are supported in addition to the above
12290 on AMD x86-64 processors in 64-bit environments.
12297 Generate code for a 32-bit or 64-bit environment.
12298 The 32-bit environment sets int, long and pointer to 32 bits and
12299 generates code that runs on any i386 system.
12300 The 64-bit environment sets int to 32 bits and long and pointer
12301 to 64 bits and generates code for AMD's x86-64 architecture. For
12302 darwin only the -m64 option turns off the @option{-fno-pic} and
12303 @option{-mdynamic-no-pic} options.
12305 @item -mno-red-zone
12306 @opindex mno-red-zone
12307 Do not use a so called red zone for x86-64 code. The red zone is mandated
12308 by the x86-64 ABI, it is a 128-byte area beyond the location of the
12309 stack pointer that will not be modified by signal or interrupt handlers
12310 and therefore can be used for temporary data without adjusting the stack
12311 pointer. The flag @option{-mno-red-zone} disables this red zone.
12313 @item -mcmodel=small
12314 @opindex mcmodel=small
12315 Generate code for the small code model: the program and its symbols must
12316 be linked in the lower 2 GB of the address space. Pointers are 64 bits.
12317 Programs can be statically or dynamically linked. This is the default
12320 @item -mcmodel=kernel
12321 @opindex mcmodel=kernel
12322 Generate code for the kernel code model. The kernel runs in the
12323 negative 2 GB of the address space.
12324 This model has to be used for Linux kernel code.
12326 @item -mcmodel=medium
12327 @opindex mcmodel=medium
12328 Generate code for the medium model: The program is linked in the lower 2
12329 GB of the address space. Small symbols are also placed there. Symbols
12330 with sizes larger than @option{-mlarge-data-threshold} are put into
12331 large data or bss sections and can be located above 2GB. Programs can
12332 be statically or dynamically linked.
12334 @item -mcmodel=large
12335 @opindex mcmodel=large
12336 Generate code for the large model: This model makes no assumptions
12337 about addresses and sizes of sections.
12340 @node IA-64 Options
12341 @subsection IA-64 Options
12342 @cindex IA-64 Options
12344 These are the @samp{-m} options defined for the Intel IA-64 architecture.
12348 @opindex mbig-endian
12349 Generate code for a big endian target. This is the default for HP-UX@.
12351 @item -mlittle-endian
12352 @opindex mlittle-endian
12353 Generate code for a little endian target. This is the default for AIX5
12359 @opindex mno-gnu-as
12360 Generate (or don't) code for the GNU assembler. This is the default.
12361 @c Also, this is the default if the configure option @option{--with-gnu-as}
12367 @opindex mno-gnu-ld
12368 Generate (or don't) code for the GNU linker. This is the default.
12369 @c Also, this is the default if the configure option @option{--with-gnu-ld}
12374 Generate code that does not use a global pointer register. The result
12375 is not position independent code, and violates the IA-64 ABI@.
12377 @item -mvolatile-asm-stop
12378 @itemx -mno-volatile-asm-stop
12379 @opindex mvolatile-asm-stop
12380 @opindex mno-volatile-asm-stop
12381 Generate (or don't) a stop bit immediately before and after volatile asm
12384 @item -mregister-names
12385 @itemx -mno-register-names
12386 @opindex mregister-names
12387 @opindex mno-register-names
12388 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
12389 the stacked registers. This may make assembler output more readable.
12395 Disable (or enable) optimizations that use the small data section. This may
12396 be useful for working around optimizer bugs.
12398 @item -mconstant-gp
12399 @opindex mconstant-gp
12400 Generate code that uses a single constant global pointer value. This is
12401 useful when compiling kernel code.
12405 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
12406 This is useful when compiling firmware code.
12408 @item -minline-float-divide-min-latency
12409 @opindex minline-float-divide-min-latency
12410 Generate code for inline divides of floating point values
12411 using the minimum latency algorithm.
12413 @item -minline-float-divide-max-throughput
12414 @opindex minline-float-divide-max-throughput
12415 Generate code for inline divides of floating point values
12416 using the maximum throughput algorithm.
12418 @item -mno-inline-float-divide
12419 @opindex mno-inline-float-divide
12420 Do not generate inline code for divides of floating point values.
12422 @item -minline-int-divide-min-latency
12423 @opindex minline-int-divide-min-latency
12424 Generate code for inline divides of integer values
12425 using the minimum latency algorithm.
12427 @item -minline-int-divide-max-throughput
12428 @opindex minline-int-divide-max-throughput
12429 Generate code for inline divides of integer values
12430 using the maximum throughput algorithm.
12432 @item -mno-inline-int-divide
12433 @opindex mno-inline-int-divide
12434 Do not generate inline code for divides of integer values.
12436 @item -minline-sqrt-min-latency
12437 @opindex minline-sqrt-min-latency
12438 Generate code for inline square roots
12439 using the minimum latency algorithm.
12441 @item -minline-sqrt-max-throughput
12442 @opindex minline-sqrt-max-throughput
12443 Generate code for inline square roots
12444 using the maximum throughput algorithm.
12446 @item -mno-inline-sqrt
12447 @opindex mno-inline-sqrt
12448 Do not generate inline code for sqrt.
12451 @itemx -mno-fused-madd
12452 @opindex mfused-madd
12453 @opindex mno-fused-madd
12454 Do (don't) generate code that uses the fused multiply/add or multiply/subtract
12455 instructions. The default is to use these instructions.
12457 @item -mno-dwarf2-asm
12458 @itemx -mdwarf2-asm
12459 @opindex mno-dwarf2-asm
12460 @opindex mdwarf2-asm
12461 Don't (or do) generate assembler code for the DWARF2 line number debugging
12462 info. This may be useful when not using the GNU assembler.
12464 @item -mearly-stop-bits
12465 @itemx -mno-early-stop-bits
12466 @opindex mearly-stop-bits
12467 @opindex mno-early-stop-bits
12468 Allow stop bits to be placed earlier than immediately preceding the
12469 instruction that triggered the stop bit. This can improve instruction
12470 scheduling, but does not always do so.
12472 @item -mfixed-range=@var{register-range}
12473 @opindex mfixed-range
12474 Generate code treating the given register range as fixed registers.
12475 A fixed register is one that the register allocator can not use. This is
12476 useful when compiling kernel code. A register range is specified as
12477 two registers separated by a dash. Multiple register ranges can be
12478 specified separated by a comma.
12480 @item -mtls-size=@var{tls-size}
12482 Specify bit size of immediate TLS offsets. Valid values are 14, 22, and
12485 @item -mtune=@var{cpu-type}
12487 Tune the instruction scheduling for a particular CPU, Valid values are
12488 itanium, itanium1, merced, itanium2, and mckinley.
12494 Generate code for a 32-bit or 64-bit environment.
12495 The 32-bit environment sets int, long and pointer to 32 bits.
12496 The 64-bit environment sets int to 32 bits and long and pointer
12497 to 64 bits. These are HP-UX specific flags.
12499 @item -mno-sched-br-data-spec
12500 @itemx -msched-br-data-spec
12501 @opindex mno-sched-br-data-spec
12502 @opindex msched-br-data-spec
12503 (Dis/En)able data speculative scheduling before reload.
12504 This will result in generation of the ld.a instructions and
12505 the corresponding check instructions (ld.c / chk.a).
12506 The default is 'disable'.
12508 @item -msched-ar-data-spec
12509 @itemx -mno-sched-ar-data-spec
12510 @opindex msched-ar-data-spec
12511 @opindex mno-sched-ar-data-spec
12512 (En/Dis)able data speculative scheduling after reload.
12513 This will result in generation of the ld.a instructions and
12514 the corresponding check instructions (ld.c / chk.a).
12515 The default is 'enable'.
12517 @item -mno-sched-control-spec
12518 @itemx -msched-control-spec
12519 @opindex mno-sched-control-spec
12520 @opindex msched-control-spec
12521 (Dis/En)able control speculative scheduling. This feature is
12522 available only during region scheduling (i.e.@: before reload).
12523 This will result in generation of the ld.s instructions and
12524 the corresponding check instructions chk.s .
12525 The default is 'disable'.
12527 @item -msched-br-in-data-spec
12528 @itemx -mno-sched-br-in-data-spec
12529 @opindex msched-br-in-data-spec
12530 @opindex mno-sched-br-in-data-spec
12531 (En/Dis)able speculative scheduling of the instructions that
12532 are dependent on the data speculative loads before reload.
12533 This is effective only with @option{-msched-br-data-spec} enabled.
12534 The default is 'enable'.
12536 @item -msched-ar-in-data-spec
12537 @itemx -mno-sched-ar-in-data-spec
12538 @opindex msched-ar-in-data-spec
12539 @opindex mno-sched-ar-in-data-spec
12540 (En/Dis)able speculative scheduling of the instructions that
12541 are dependent on the data speculative loads after reload.
12542 This is effective only with @option{-msched-ar-data-spec} enabled.
12543 The default is 'enable'.
12545 @item -msched-in-control-spec
12546 @itemx -mno-sched-in-control-spec
12547 @opindex msched-in-control-spec
12548 @opindex mno-sched-in-control-spec
12549 (En/Dis)able speculative scheduling of the instructions that
12550 are dependent on the control speculative loads.
12551 This is effective only with @option{-msched-control-spec} enabled.
12552 The default is 'enable'.
12554 @item -mno-sched-prefer-non-data-spec-insns
12555 @itemx -msched-prefer-non-data-spec-insns
12556 @opindex mno-sched-prefer-non-data-spec-insns
12557 @opindex msched-prefer-non-data-spec-insns
12558 If enabled, data speculative instructions will be chosen for schedule
12559 only if there are no other choices at the moment. This will make
12560 the use of the data speculation much more conservative.
12561 The default is 'disable'.
12563 @item -mno-sched-prefer-non-control-spec-insns
12564 @itemx -msched-prefer-non-control-spec-insns
12565 @opindex mno-sched-prefer-non-control-spec-insns
12566 @opindex msched-prefer-non-control-spec-insns
12567 If enabled, control speculative instructions will be chosen for schedule
12568 only if there are no other choices at the moment. This will make
12569 the use of the control speculation much more conservative.
12570 The default is 'disable'.
12572 @item -mno-sched-count-spec-in-critical-path
12573 @itemx -msched-count-spec-in-critical-path
12574 @opindex mno-sched-count-spec-in-critical-path
12575 @opindex msched-count-spec-in-critical-path
12576 If enabled, speculative dependencies will be considered during
12577 computation of the instructions priorities. This will make the use of the
12578 speculation a bit more conservative.
12579 The default is 'disable'.
12581 @item -msched-spec-ldc
12582 @opindex msched-spec-ldc
12583 Use a simple data speculation check. This option is on by default.
12585 @item -msched-control-spec-ldc
12586 @opindex msched-spec-ldc
12587 Use a simple check for control speculation. This option is on by default.
12589 @item -msched-stop-bits-after-every-cycle
12590 @opindex msched-stop-bits-after-every-cycle
12591 Place a stop bit after every cycle when scheduling. This option is on
12594 @item -msched-fp-mem-deps-zero-cost
12595 @opindex msched-fp-mem-deps-zero-cost
12596 Assume that floating-point stores and loads are not likely to cause a conflict
12597 when placed into the same instruction group. This option is disabled by
12600 @item -msel-sched-dont-check-control-spec
12601 @opindex msel-sched-dont-check-control-spec
12602 Generate checks for control speculation in selective scheduling.
12603 This flag is disabled by default.
12605 @item -msched-max-memory-insns=@var{max-insns}
12606 @opindex msched-max-memory-insns
12607 Limit on the number of memory insns per instruction group, giving lower
12608 priority to subsequent memory insns attempting to schedule in the same
12609 instruction group. Frequently useful to prevent cache bank conflicts.
12610 The default value is 1.
12612 @item -msched-max-memory-insns-hard-limit
12613 @opindex msched-max-memory-insns-hard-limit
12614 Disallow more than `msched-max-memory-insns' in instruction group.
12615 Otherwise, limit is `soft' meaning that we would prefer non-memory operations
12616 when limit is reached but may still schedule memory operations.
12620 @node IA-64/VMS Options
12621 @subsection IA-64/VMS Options
12623 These @samp{-m} options are defined for the IA-64/VMS implementations:
12626 @item -mvms-return-codes
12627 @opindex mvms-return-codes
12628 Return VMS condition codes from main. The default is to return POSIX
12629 style condition (e.g.@ error) codes.
12631 @item -mdebug-main=@var{prefix}
12632 @opindex mdebug-main=@var{prefix}
12633 Flag the first routine whose name starts with @var{prefix} as the main
12634 routine for the debugger.
12638 Default to 64bit memory allocation routines.
12642 @subsection LM32 Options
12643 @cindex LM32 options
12645 These @option{-m} options are defined for the Lattice Mico32 architecture:
12648 @item -mbarrel-shift-enabled
12649 @opindex mbarrel-shift-enabled
12650 Enable barrel-shift instructions.
12652 @item -mdivide-enabled
12653 @opindex mdivide-enabled
12654 Enable divide and modulus instructions.
12656 @item -mmultiply-enabled
12657 @opindex multiply-enabled
12658 Enable multiply instructions.
12660 @item -msign-extend-enabled
12661 @opindex msign-extend-enabled
12662 Enable sign extend instructions.
12664 @item -muser-enabled
12665 @opindex muser-enabled
12666 Enable user-defined instructions.
12671 @subsection M32C Options
12672 @cindex M32C options
12675 @item -mcpu=@var{name}
12677 Select the CPU for which code is generated. @var{name} may be one of
12678 @samp{r8c} for the R8C/Tiny series, @samp{m16c} for the M16C (up to
12679 /60) series, @samp{m32cm} for the M16C/80 series, or @samp{m32c} for
12680 the M32C/80 series.
12684 Specifies that the program will be run on the simulator. This causes
12685 an alternate runtime library to be linked in which supports, for
12686 example, file I/O@. You must not use this option when generating
12687 programs that will run on real hardware; you must provide your own
12688 runtime library for whatever I/O functions are needed.
12690 @item -memregs=@var{number}
12692 Specifies the number of memory-based pseudo-registers GCC will use
12693 during code generation. These pseudo-registers will be used like real
12694 registers, so there is a tradeoff between GCC's ability to fit the
12695 code into available registers, and the performance penalty of using
12696 memory instead of registers. Note that all modules in a program must
12697 be compiled with the same value for this option. Because of that, you
12698 must not use this option with the default runtime libraries gcc
12703 @node M32R/D Options
12704 @subsection M32R/D Options
12705 @cindex M32R/D options
12707 These @option{-m} options are defined for Renesas M32R/D architectures:
12712 Generate code for the M32R/2@.
12716 Generate code for the M32R/X@.
12720 Generate code for the M32R@. This is the default.
12722 @item -mmodel=small
12723 @opindex mmodel=small
12724 Assume all objects live in the lower 16MB of memory (so that their addresses
12725 can be loaded with the @code{ld24} instruction), and assume all subroutines
12726 are reachable with the @code{bl} instruction.
12727 This is the default.
12729 The addressability of a particular object can be set with the
12730 @code{model} attribute.
12732 @item -mmodel=medium
12733 @opindex mmodel=medium
12734 Assume objects may be anywhere in the 32-bit address space (the compiler
12735 will generate @code{seth/add3} instructions to load their addresses), and
12736 assume all subroutines are reachable with the @code{bl} instruction.
12738 @item -mmodel=large
12739 @opindex mmodel=large
12740 Assume objects may be anywhere in the 32-bit address space (the compiler
12741 will generate @code{seth/add3} instructions to load their addresses), and
12742 assume subroutines may not be reachable with the @code{bl} instruction
12743 (the compiler will generate the much slower @code{seth/add3/jl}
12744 instruction sequence).
12747 @opindex msdata=none
12748 Disable use of the small data area. Variables will be put into
12749 one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the
12750 @code{section} attribute has been specified).
12751 This is the default.
12753 The small data area consists of sections @samp{.sdata} and @samp{.sbss}.
12754 Objects may be explicitly put in the small data area with the
12755 @code{section} attribute using one of these sections.
12757 @item -msdata=sdata
12758 @opindex msdata=sdata
12759 Put small global and static data in the small data area, but do not
12760 generate special code to reference them.
12763 @opindex msdata=use
12764 Put small global and static data in the small data area, and generate
12765 special instructions to reference them.
12769 @cindex smaller data references
12770 Put global and static objects less than or equal to @var{num} bytes
12771 into the small data or bss sections instead of the normal data or bss
12772 sections. The default value of @var{num} is 8.
12773 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
12774 for this option to have any effect.
12776 All modules should be compiled with the same @option{-G @var{num}} value.
12777 Compiling with different values of @var{num} may or may not work; if it
12778 doesn't the linker will give an error message---incorrect code will not be
12783 Makes the M32R specific code in the compiler display some statistics
12784 that might help in debugging programs.
12786 @item -malign-loops
12787 @opindex malign-loops
12788 Align all loops to a 32-byte boundary.
12790 @item -mno-align-loops
12791 @opindex mno-align-loops
12792 Do not enforce a 32-byte alignment for loops. This is the default.
12794 @item -missue-rate=@var{number}
12795 @opindex missue-rate=@var{number}
12796 Issue @var{number} instructions per cycle. @var{number} can only be 1
12799 @item -mbranch-cost=@var{number}
12800 @opindex mbranch-cost=@var{number}
12801 @var{number} can only be 1 or 2. If it is 1 then branches will be
12802 preferred over conditional code, if it is 2, then the opposite will
12805 @item -mflush-trap=@var{number}
12806 @opindex mflush-trap=@var{number}
12807 Specifies the trap number to use to flush the cache. The default is
12808 12. Valid numbers are between 0 and 15 inclusive.
12810 @item -mno-flush-trap
12811 @opindex mno-flush-trap
12812 Specifies that the cache cannot be flushed by using a trap.
12814 @item -mflush-func=@var{name}
12815 @opindex mflush-func=@var{name}
12816 Specifies the name of the operating system function to call to flush
12817 the cache. The default is @emph{_flush_cache}, but a function call
12818 will only be used if a trap is not available.
12820 @item -mno-flush-func
12821 @opindex mno-flush-func
12822 Indicates that there is no OS function for flushing the cache.
12826 @node M680x0 Options
12827 @subsection M680x0 Options
12828 @cindex M680x0 options
12830 These are the @samp{-m} options defined for M680x0 and ColdFire processors.
12831 The default settings depend on which architecture was selected when
12832 the compiler was configured; the defaults for the most common choices
12836 @item -march=@var{arch}
12838 Generate code for a specific M680x0 or ColdFire instruction set
12839 architecture. Permissible values of @var{arch} for M680x0
12840 architectures are: @samp{68000}, @samp{68010}, @samp{68020},
12841 @samp{68030}, @samp{68040}, @samp{68060} and @samp{cpu32}. ColdFire
12842 architectures are selected according to Freescale's ISA classification
12843 and the permissible values are: @samp{isaa}, @samp{isaaplus},
12844 @samp{isab} and @samp{isac}.
12846 gcc defines a macro @samp{__mcf@var{arch}__} whenever it is generating
12847 code for a ColdFire target. The @var{arch} in this macro is one of the
12848 @option{-march} arguments given above.
12850 When used together, @option{-march} and @option{-mtune} select code
12851 that runs on a family of similar processors but that is optimized
12852 for a particular microarchitecture.
12854 @item -mcpu=@var{cpu}
12856 Generate code for a specific M680x0 or ColdFire processor.
12857 The M680x0 @var{cpu}s are: @samp{68000}, @samp{68010}, @samp{68020},
12858 @samp{68030}, @samp{68040}, @samp{68060}, @samp{68302}, @samp{68332}
12859 and @samp{cpu32}. The ColdFire @var{cpu}s are given by the table
12860 below, which also classifies the CPUs into families:
12862 @multitable @columnfractions 0.20 0.80
12863 @item @strong{Family} @tab @strong{@samp{-mcpu} arguments}
12864 @item @samp{51} @tab @samp{51} @samp{51ac} @samp{51cn} @samp{51em} @samp{51qe}
12865 @item @samp{5206} @tab @samp{5202} @samp{5204} @samp{5206}
12866 @item @samp{5206e} @tab @samp{5206e}
12867 @item @samp{5208} @tab @samp{5207} @samp{5208}
12868 @item @samp{5211a} @tab @samp{5210a} @samp{5211a}
12869 @item @samp{5213} @tab @samp{5211} @samp{5212} @samp{5213}
12870 @item @samp{5216} @tab @samp{5214} @samp{5216}
12871 @item @samp{52235} @tab @samp{52230} @samp{52231} @samp{52232} @samp{52233} @samp{52234} @samp{52235}
12872 @item @samp{5225} @tab @samp{5224} @samp{5225}
12873 @item @samp{52259} @tab @samp{52252} @samp{52254} @samp{52255} @samp{52256} @samp{52258} @samp{52259}
12874 @item @samp{5235} @tab @samp{5232} @samp{5233} @samp{5234} @samp{5235} @samp{523x}
12875 @item @samp{5249} @tab @samp{5249}
12876 @item @samp{5250} @tab @samp{5250}
12877 @item @samp{5271} @tab @samp{5270} @samp{5271}
12878 @item @samp{5272} @tab @samp{5272}
12879 @item @samp{5275} @tab @samp{5274} @samp{5275}
12880 @item @samp{5282} @tab @samp{5280} @samp{5281} @samp{5282} @samp{528x}
12881 @item @samp{53017} @tab @samp{53011} @samp{53012} @samp{53013} @samp{53014} @samp{53015} @samp{53016} @samp{53017}
12882 @item @samp{5307} @tab @samp{5307}
12883 @item @samp{5329} @tab @samp{5327} @samp{5328} @samp{5329} @samp{532x}
12884 @item @samp{5373} @tab @samp{5372} @samp{5373} @samp{537x}
12885 @item @samp{5407} @tab @samp{5407}
12886 @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}
12889 @option{-mcpu=@var{cpu}} overrides @option{-march=@var{arch}} if
12890 @var{arch} is compatible with @var{cpu}. Other combinations of
12891 @option{-mcpu} and @option{-march} are rejected.
12893 gcc defines the macro @samp{__mcf_cpu_@var{cpu}} when ColdFire target
12894 @var{cpu} is selected. It also defines @samp{__mcf_family_@var{family}},
12895 where the value of @var{family} is given by the table above.
12897 @item -mtune=@var{tune}
12899 Tune the code for a particular microarchitecture, within the
12900 constraints set by @option{-march} and @option{-mcpu}.
12901 The M680x0 microarchitectures are: @samp{68000}, @samp{68010},
12902 @samp{68020}, @samp{68030}, @samp{68040}, @samp{68060}
12903 and @samp{cpu32}. The ColdFire microarchitectures
12904 are: @samp{cfv1}, @samp{cfv2}, @samp{cfv3}, @samp{cfv4} and @samp{cfv4e}.
12906 You can also use @option{-mtune=68020-40} for code that needs
12907 to run relatively well on 68020, 68030 and 68040 targets.
12908 @option{-mtune=68020-60} is similar but includes 68060 targets
12909 as well. These two options select the same tuning decisions as
12910 @option{-m68020-40} and @option{-m68020-60} respectively.
12912 gcc defines the macros @samp{__mc@var{arch}} and @samp{__mc@var{arch}__}
12913 when tuning for 680x0 architecture @var{arch}. It also defines
12914 @samp{mc@var{arch}} unless either @option{-ansi} or a non-GNU @option{-std}
12915 option is used. If gcc is tuning for a range of architectures,
12916 as selected by @option{-mtune=68020-40} or @option{-mtune=68020-60},
12917 it defines the macros for every architecture in the range.
12919 gcc also defines the macro @samp{__m@var{uarch}__} when tuning for
12920 ColdFire microarchitecture @var{uarch}, where @var{uarch} is one
12921 of the arguments given above.
12927 Generate output for a 68000. This is the default
12928 when the compiler is configured for 68000-based systems.
12929 It is equivalent to @option{-march=68000}.
12931 Use this option for microcontrollers with a 68000 or EC000 core,
12932 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
12936 Generate output for a 68010. This is the default
12937 when the compiler is configured for 68010-based systems.
12938 It is equivalent to @option{-march=68010}.
12944 Generate output for a 68020. This is the default
12945 when the compiler is configured for 68020-based systems.
12946 It is equivalent to @option{-march=68020}.
12950 Generate output for a 68030. This is the default when the compiler is
12951 configured for 68030-based systems. It is equivalent to
12952 @option{-march=68030}.
12956 Generate output for a 68040. This is the default when the compiler is
12957 configured for 68040-based systems. It is equivalent to
12958 @option{-march=68040}.
12960 This option inhibits the use of 68881/68882 instructions that have to be
12961 emulated by software on the 68040. Use this option if your 68040 does not
12962 have code to emulate those instructions.
12966 Generate output for a 68060. This is the default when the compiler is
12967 configured for 68060-based systems. It is equivalent to
12968 @option{-march=68060}.
12970 This option inhibits the use of 68020 and 68881/68882 instructions that
12971 have to be emulated by software on the 68060. Use this option if your 68060
12972 does not have code to emulate those instructions.
12976 Generate output for a CPU32. This is the default
12977 when the compiler is configured for CPU32-based systems.
12978 It is equivalent to @option{-march=cpu32}.
12980 Use this option for microcontrollers with a
12981 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
12982 68336, 68340, 68341, 68349 and 68360.
12986 Generate output for a 520X ColdFire CPU@. This is the default
12987 when the compiler is configured for 520X-based systems.
12988 It is equivalent to @option{-mcpu=5206}, and is now deprecated
12989 in favor of that option.
12991 Use this option for microcontroller with a 5200 core, including
12992 the MCF5202, MCF5203, MCF5204 and MCF5206.
12996 Generate output for a 5206e ColdFire CPU@. The option is now
12997 deprecated in favor of the equivalent @option{-mcpu=5206e}.
13001 Generate output for a member of the ColdFire 528X family.
13002 The option is now deprecated in favor of the equivalent
13003 @option{-mcpu=528x}.
13007 Generate output for a ColdFire 5307 CPU@. The option is now deprecated
13008 in favor of the equivalent @option{-mcpu=5307}.
13012 Generate output for a ColdFire 5407 CPU@. The option is now deprecated
13013 in favor of the equivalent @option{-mcpu=5407}.
13017 Generate output for a ColdFire V4e family CPU (e.g.@: 547x/548x).
13018 This includes use of hardware floating point instructions.
13019 The option is equivalent to @option{-mcpu=547x}, and is now
13020 deprecated in favor of that option.
13024 Generate output for a 68040, without using any of the new instructions.
13025 This results in code which can run relatively efficiently on either a
13026 68020/68881 or a 68030 or a 68040. The generated code does use the
13027 68881 instructions that are emulated on the 68040.
13029 The option is equivalent to @option{-march=68020} @option{-mtune=68020-40}.
13033 Generate output for a 68060, without using any of the new instructions.
13034 This results in code which can run relatively efficiently on either a
13035 68020/68881 or a 68030 or a 68040. The generated code does use the
13036 68881 instructions that are emulated on the 68060.
13038 The option is equivalent to @option{-march=68020} @option{-mtune=68020-60}.
13042 @opindex mhard-float
13044 Generate floating-point instructions. This is the default for 68020
13045 and above, and for ColdFire devices that have an FPU@. It defines the
13046 macro @samp{__HAVE_68881__} on M680x0 targets and @samp{__mcffpu__}
13047 on ColdFire targets.
13050 @opindex msoft-float
13051 Do not generate floating-point instructions; use library calls instead.
13052 This is the default for 68000, 68010, and 68832 targets. It is also
13053 the default for ColdFire devices that have no FPU.
13059 Generate (do not generate) ColdFire hardware divide and remainder
13060 instructions. If @option{-march} is used without @option{-mcpu},
13061 the default is ``on'' for ColdFire architectures and ``off'' for M680x0
13062 architectures. Otherwise, the default is taken from the target CPU
13063 (either the default CPU, or the one specified by @option{-mcpu}). For
13064 example, the default is ``off'' for @option{-mcpu=5206} and ``on'' for
13065 @option{-mcpu=5206e}.
13067 gcc defines the macro @samp{__mcfhwdiv__} when this option is enabled.
13071 Consider type @code{int} to be 16 bits wide, like @code{short int}.
13072 Additionally, parameters passed on the stack are also aligned to a
13073 16-bit boundary even on targets whose API mandates promotion to 32-bit.
13077 Do not consider type @code{int} to be 16 bits wide. This is the default.
13080 @itemx -mno-bitfield
13081 @opindex mnobitfield
13082 @opindex mno-bitfield
13083 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
13084 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
13088 Do use the bit-field instructions. The @option{-m68020} option implies
13089 @option{-mbitfield}. This is the default if you use a configuration
13090 designed for a 68020.
13094 Use a different function-calling convention, in which functions
13095 that take a fixed number of arguments return with the @code{rtd}
13096 instruction, which pops their arguments while returning. This
13097 saves one instruction in the caller since there is no need to pop
13098 the arguments there.
13100 This calling convention is incompatible with the one normally
13101 used on Unix, so you cannot use it if you need to call libraries
13102 compiled with the Unix compiler.
13104 Also, you must provide function prototypes for all functions that
13105 take variable numbers of arguments (including @code{printf});
13106 otherwise incorrect code will be generated for calls to those
13109 In addition, seriously incorrect code will result if you call a
13110 function with too many arguments. (Normally, extra arguments are
13111 harmlessly ignored.)
13113 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
13114 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
13118 Do not use the calling conventions selected by @option{-mrtd}.
13119 This is the default.
13122 @itemx -mno-align-int
13123 @opindex malign-int
13124 @opindex mno-align-int
13125 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
13126 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
13127 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
13128 Aligning variables on 32-bit boundaries produces code that runs somewhat
13129 faster on processors with 32-bit busses at the expense of more memory.
13131 @strong{Warning:} if you use the @option{-malign-int} switch, GCC will
13132 align structures containing the above types differently than
13133 most published application binary interface specifications for the m68k.
13137 Use the pc-relative addressing mode of the 68000 directly, instead of
13138 using a global offset table. At present, this option implies @option{-fpic},
13139 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
13140 not presently supported with @option{-mpcrel}, though this could be supported for
13141 68020 and higher processors.
13143 @item -mno-strict-align
13144 @itemx -mstrict-align
13145 @opindex mno-strict-align
13146 @opindex mstrict-align
13147 Do not (do) assume that unaligned memory references will be handled by
13151 Generate code that allows the data segment to be located in a different
13152 area of memory from the text segment. This allows for execute in place in
13153 an environment without virtual memory management. This option implies
13156 @item -mno-sep-data
13157 Generate code that assumes that the data segment follows the text segment.
13158 This is the default.
13160 @item -mid-shared-library
13161 Generate code that supports shared libraries via the library ID method.
13162 This allows for execute in place and shared libraries in an environment
13163 without virtual memory management. This option implies @option{-fPIC}.
13165 @item -mno-id-shared-library
13166 Generate code that doesn't assume ID based shared libraries are being used.
13167 This is the default.
13169 @item -mshared-library-id=n
13170 Specified the identification number of the ID based shared library being
13171 compiled. Specifying a value of 0 will generate more compact code, specifying
13172 other values will force the allocation of that number to the current
13173 library but is no more space or time efficient than omitting this option.
13179 When generating position-independent code for ColdFire, generate code
13180 that works if the GOT has more than 8192 entries. This code is
13181 larger and slower than code generated without this option. On M680x0
13182 processors, this option is not needed; @option{-fPIC} suffices.
13184 GCC normally uses a single instruction to load values from the GOT@.
13185 While this is relatively efficient, it only works if the GOT
13186 is smaller than about 64k. Anything larger causes the linker
13187 to report an error such as:
13189 @cindex relocation truncated to fit (ColdFire)
13191 relocation truncated to fit: R_68K_GOT16O foobar
13194 If this happens, you should recompile your code with @option{-mxgot}.
13195 It should then work with very large GOTs. However, code generated with
13196 @option{-mxgot} is less efficient, since it takes 4 instructions to fetch
13197 the value of a global symbol.
13199 Note that some linkers, including newer versions of the GNU linker,
13200 can create multiple GOTs and sort GOT entries. If you have such a linker,
13201 you should only need to use @option{-mxgot} when compiling a single
13202 object file that accesses more than 8192 GOT entries. Very few do.
13204 These options have no effect unless GCC is generating
13205 position-independent code.
13209 @node M68hc1x Options
13210 @subsection M68hc1x Options
13211 @cindex M68hc1x options
13213 These are the @samp{-m} options defined for the 68hc11 and 68hc12
13214 microcontrollers. The default values for these options depends on
13215 which style of microcontroller was selected when the compiler was configured;
13216 the defaults for the most common choices are given below.
13223 Generate output for a 68HC11. This is the default
13224 when the compiler is configured for 68HC11-based systems.
13230 Generate output for a 68HC12. This is the default
13231 when the compiler is configured for 68HC12-based systems.
13237 Generate output for a 68HCS12.
13239 @item -mauto-incdec
13240 @opindex mauto-incdec
13241 Enable the use of 68HC12 pre and post auto-increment and auto-decrement
13248 Enable the use of 68HC12 min and max instructions.
13251 @itemx -mno-long-calls
13252 @opindex mlong-calls
13253 @opindex mno-long-calls
13254 Treat all calls as being far away (near). If calls are assumed to be
13255 far away, the compiler will use the @code{call} instruction to
13256 call a function and the @code{rtc} instruction for returning.
13260 Consider type @code{int} to be 16 bits wide, like @code{short int}.
13262 @item -msoft-reg-count=@var{count}
13263 @opindex msoft-reg-count
13264 Specify the number of pseudo-soft registers which are used for the
13265 code generation. The maximum number is 32. Using more pseudo-soft
13266 register may or may not result in better code depending on the program.
13267 The default is 4 for 68HC11 and 2 for 68HC12.
13271 @node MCore Options
13272 @subsection MCore Options
13273 @cindex MCore options
13275 These are the @samp{-m} options defined for the Motorola M*Core
13281 @itemx -mno-hardlit
13283 @opindex mno-hardlit
13284 Inline constants into the code stream if it can be done in two
13285 instructions or less.
13291 Use the divide instruction. (Enabled by default).
13293 @item -mrelax-immediate
13294 @itemx -mno-relax-immediate
13295 @opindex mrelax-immediate
13296 @opindex mno-relax-immediate
13297 Allow arbitrary sized immediates in bit operations.
13299 @item -mwide-bitfields
13300 @itemx -mno-wide-bitfields
13301 @opindex mwide-bitfields
13302 @opindex mno-wide-bitfields
13303 Always treat bit-fields as int-sized.
13305 @item -m4byte-functions
13306 @itemx -mno-4byte-functions
13307 @opindex m4byte-functions
13308 @opindex mno-4byte-functions
13309 Force all functions to be aligned to a four byte boundary.
13311 @item -mcallgraph-data
13312 @itemx -mno-callgraph-data
13313 @opindex mcallgraph-data
13314 @opindex mno-callgraph-data
13315 Emit callgraph information.
13318 @itemx -mno-slow-bytes
13319 @opindex mslow-bytes
13320 @opindex mno-slow-bytes
13321 Prefer word access when reading byte quantities.
13323 @item -mlittle-endian
13324 @itemx -mbig-endian
13325 @opindex mlittle-endian
13326 @opindex mbig-endian
13327 Generate code for a little endian target.
13333 Generate code for the 210 processor.
13337 Assume that run-time support has been provided and so omit the
13338 simulator library (@file{libsim.a)} from the linker command line.
13340 @item -mstack-increment=@var{size}
13341 @opindex mstack-increment
13342 Set the maximum amount for a single stack increment operation. Large
13343 values can increase the speed of programs which contain functions
13344 that need a large amount of stack space, but they can also trigger a
13345 segmentation fault if the stack is extended too much. The default
13351 @subsection MeP Options
13352 @cindex MeP options
13358 Enables the @code{abs} instruction, which is the absolute difference
13359 between two registers.
13363 Enables all the optional instructions - average, multiply, divide, bit
13364 operations, leading zero, absolute difference, min/max, clip, and
13370 Enables the @code{ave} instruction, which computes the average of two
13373 @item -mbased=@var{n}
13375 Variables of size @var{n} bytes or smaller will be placed in the
13376 @code{.based} section by default. Based variables use the @code{$tp}
13377 register as a base register, and there is a 128 byte limit to the
13378 @code{.based} section.
13382 Enables the bit operation instructions - bit test (@code{btstm}), set
13383 (@code{bsetm}), clear (@code{bclrm}), invert (@code{bnotm}), and
13384 test-and-set (@code{tas}).
13386 @item -mc=@var{name}
13388 Selects which section constant data will be placed in. @var{name} may
13389 be @code{tiny}, @code{near}, or @code{far}.
13393 Enables the @code{clip} instruction. Note that @code{-mclip} is not
13394 useful unless you also provide @code{-mminmax}.
13396 @item -mconfig=@var{name}
13398 Selects one of the build-in core configurations. Each MeP chip has
13399 one or more modules in it; each module has a core CPU and a variety of
13400 coprocessors, optional instructions, and peripherals. The
13401 @code{MeP-Integrator} tool, not part of GCC, provides these
13402 configurations through this option; using this option is the same as
13403 using all the corresponding command line options. The default
13404 configuration is @code{default}.
13408 Enables the coprocessor instructions. By default, this is a 32-bit
13409 coprocessor. Note that the coprocessor is normally enabled via the
13410 @code{-mconfig=} option.
13414 Enables the 32-bit coprocessor's instructions.
13418 Enables the 64-bit coprocessor's instructions.
13422 Enables IVC2 scheduling. IVC2 is a 64-bit VLIW coprocessor.
13426 Causes constant variables to be placed in the @code{.near} section.
13430 Enables the @code{div} and @code{divu} instructions.
13434 Generate big-endian code.
13438 Generate little-endian code.
13440 @item -mio-volatile
13441 @opindex mio-volatile
13442 Tells the compiler that any variable marked with the @code{io}
13443 attribute is to be considered volatile.
13447 Causes variables to be assigned to the @code{.far} section by default.
13451 Enables the @code{leadz} (leading zero) instruction.
13455 Causes variables to be assigned to the @code{.near} section by default.
13459 Enables the @code{min} and @code{max} instructions.
13463 Enables the multiplication and multiply-accumulate instructions.
13467 Disables all the optional instructions enabled by @code{-mall-opts}.
13471 Enables the @code{repeat} and @code{erepeat} instructions, used for
13472 low-overhead looping.
13476 Causes all variables to default to the @code{.tiny} section. Note
13477 that there is a 65536 byte limit to this section. Accesses to these
13478 variables use the @code{%gp} base register.
13482 Enables the saturation instructions. Note that the compiler does not
13483 currently generate these itself, but this option is included for
13484 compatibility with other tools, like @code{as}.
13488 Link the SDRAM-based runtime instead of the default ROM-based runtime.
13492 Link the simulator runtime libraries.
13496 Link the simulator runtime libraries, excluding built-in support
13497 for reset and exception vectors and tables.
13501 Causes all functions to default to the @code{.far} section. Without
13502 this option, functions default to the @code{.near} section.
13504 @item -mtiny=@var{n}
13506 Variables that are @var{n} bytes or smaller will be allocated to the
13507 @code{.tiny} section. These variables use the @code{$gp} base
13508 register. The default for this option is 4, but note that there's a
13509 65536 byte limit to the @code{.tiny} section.
13514 @subsection MIPS Options
13515 @cindex MIPS options
13521 Generate big-endian code.
13525 Generate little-endian code. This is the default for @samp{mips*el-*-*}
13528 @item -march=@var{arch}
13530 Generate code that will run on @var{arch}, which can be the name of a
13531 generic MIPS ISA, or the name of a particular processor.
13533 @samp{mips1}, @samp{mips2}, @samp{mips3}, @samp{mips4},
13534 @samp{mips32}, @samp{mips32r2}, @samp{mips64} and @samp{mips64r2}.
13535 The processor names are:
13536 @samp{4kc}, @samp{4km}, @samp{4kp}, @samp{4ksc},
13537 @samp{4kec}, @samp{4kem}, @samp{4kep}, @samp{4ksd},
13538 @samp{5kc}, @samp{5kf},
13540 @samp{24kc}, @samp{24kf2_1}, @samp{24kf1_1},
13541 @samp{24kec}, @samp{24kef2_1}, @samp{24kef1_1},
13542 @samp{34kc}, @samp{34kf2_1}, @samp{34kf1_1},
13543 @samp{74kc}, @samp{74kf2_1}, @samp{74kf1_1}, @samp{74kf3_2},
13544 @samp{1004kc}, @samp{1004kf2_1}, @samp{1004kf1_1},
13545 @samp{loongson2e}, @samp{loongson2f},
13549 @samp{r2000}, @samp{r3000}, @samp{r3900}, @samp{r4000}, @samp{r4400},
13550 @samp{r4600}, @samp{r4650}, @samp{r6000}, @samp{r8000},
13551 @samp{rm7000}, @samp{rm9000},
13552 @samp{r10000}, @samp{r12000}, @samp{r14000}, @samp{r16000},
13555 @samp{vr4100}, @samp{vr4111}, @samp{vr4120}, @samp{vr4130}, @samp{vr4300},
13556 @samp{vr5000}, @samp{vr5400}, @samp{vr5500}
13558 The special value @samp{from-abi} selects the
13559 most compatible architecture for the selected ABI (that is,
13560 @samp{mips1} for 32-bit ABIs and @samp{mips3} for 64-bit ABIs)@.
13562 Native Linux/GNU toolchains also support the value @samp{native},
13563 which selects the best architecture option for the host processor.
13564 @option{-march=native} has no effect if GCC does not recognize
13567 In processor names, a final @samp{000} can be abbreviated as @samp{k}
13568 (for example, @samp{-march=r2k}). Prefixes are optional, and
13569 @samp{vr} may be written @samp{r}.
13571 Names of the form @samp{@var{n}f2_1} refer to processors with
13572 FPUs clocked at half the rate of the core, names of the form
13573 @samp{@var{n}f1_1} refer to processors with FPUs clocked at the same
13574 rate as the core, and names of the form @samp{@var{n}f3_2} refer to
13575 processors with FPUs clocked a ratio of 3:2 with respect to the core.
13576 For compatibility reasons, @samp{@var{n}f} is accepted as a synonym
13577 for @samp{@var{n}f2_1} while @samp{@var{n}x} and @samp{@var{b}fx} are
13578 accepted as synonyms for @samp{@var{n}f1_1}.
13580 GCC defines two macros based on the value of this option. The first
13581 is @samp{_MIPS_ARCH}, which gives the name of target architecture, as
13582 a string. The second has the form @samp{_MIPS_ARCH_@var{foo}},
13583 where @var{foo} is the capitalized value of @samp{_MIPS_ARCH}@.
13584 For example, @samp{-march=r2000} will set @samp{_MIPS_ARCH}
13585 to @samp{"r2000"} and define the macro @samp{_MIPS_ARCH_R2000}.
13587 Note that the @samp{_MIPS_ARCH} macro uses the processor names given
13588 above. In other words, it will have the full prefix and will not
13589 abbreviate @samp{000} as @samp{k}. In the case of @samp{from-abi},
13590 the macro names the resolved architecture (either @samp{"mips1"} or
13591 @samp{"mips3"}). It names the default architecture when no
13592 @option{-march} option is given.
13594 @item -mtune=@var{arch}
13596 Optimize for @var{arch}. Among other things, this option controls
13597 the way instructions are scheduled, and the perceived cost of arithmetic
13598 operations. The list of @var{arch} values is the same as for
13601 When this option is not used, GCC will optimize for the processor
13602 specified by @option{-march}. By using @option{-march} and
13603 @option{-mtune} together, it is possible to generate code that will
13604 run on a family of processors, but optimize the code for one
13605 particular member of that family.
13607 @samp{-mtune} defines the macros @samp{_MIPS_TUNE} and
13608 @samp{_MIPS_TUNE_@var{foo}}, which work in the same way as the
13609 @samp{-march} ones described above.
13613 Equivalent to @samp{-march=mips1}.
13617 Equivalent to @samp{-march=mips2}.
13621 Equivalent to @samp{-march=mips3}.
13625 Equivalent to @samp{-march=mips4}.
13629 Equivalent to @samp{-march=mips32}.
13633 Equivalent to @samp{-march=mips32r2}.
13637 Equivalent to @samp{-march=mips64}.
13641 Equivalent to @samp{-march=mips64r2}.
13646 @opindex mno-mips16
13647 Generate (do not generate) MIPS16 code. If GCC is targetting a
13648 MIPS32 or MIPS64 architecture, it will make use of the MIPS16e ASE@.
13650 MIPS16 code generation can also be controlled on a per-function basis
13651 by means of @code{mips16} and @code{nomips16} attributes.
13652 @xref{Function Attributes}, for more information.
13654 @item -mflip-mips16
13655 @opindex mflip-mips16
13656 Generate MIPS16 code on alternating functions. This option is provided
13657 for regression testing of mixed MIPS16/non-MIPS16 code generation, and is
13658 not intended for ordinary use in compiling user code.
13660 @item -minterlink-mips16
13661 @itemx -mno-interlink-mips16
13662 @opindex minterlink-mips16
13663 @opindex mno-interlink-mips16
13664 Require (do not require) that non-MIPS16 code be link-compatible with
13667 For example, non-MIPS16 code cannot jump directly to MIPS16 code;
13668 it must either use a call or an indirect jump. @option{-minterlink-mips16}
13669 therefore disables direct jumps unless GCC knows that the target of the
13670 jump is not MIPS16.
13682 Generate code for the given ABI@.
13684 Note that the EABI has a 32-bit and a 64-bit variant. GCC normally
13685 generates 64-bit code when you select a 64-bit architecture, but you
13686 can use @option{-mgp32} to get 32-bit code instead.
13688 For information about the O64 ABI, see
13689 @w{@uref{http://gcc.gnu.org/projects/mipso64-abi.html}}.
13691 GCC supports a variant of the o32 ABI in which floating-point registers
13692 are 64 rather than 32 bits wide. You can select this combination with
13693 @option{-mabi=32} @option{-mfp64}. This ABI relies on the @samp{mthc1}
13694 and @samp{mfhc1} instructions and is therefore only supported for
13695 MIPS32R2 processors.
13697 The register assignments for arguments and return values remain the
13698 same, but each scalar value is passed in a single 64-bit register
13699 rather than a pair of 32-bit registers. For example, scalar
13700 floating-point values are returned in @samp{$f0} only, not a
13701 @samp{$f0}/@samp{$f1} pair. The set of call-saved registers also
13702 remains the same, but all 64 bits are saved.
13705 @itemx -mno-abicalls
13707 @opindex mno-abicalls
13708 Generate (do not generate) code that is suitable for SVR4-style
13709 dynamic objects. @option{-mabicalls} is the default for SVR4-based
13714 Generate (do not generate) code that is fully position-independent,
13715 and that can therefore be linked into shared libraries. This option
13716 only affects @option{-mabicalls}.
13718 All @option{-mabicalls} code has traditionally been position-independent,
13719 regardless of options like @option{-fPIC} and @option{-fpic}. However,
13720 as an extension, the GNU toolchain allows executables to use absolute
13721 accesses for locally-binding symbols. It can also use shorter GP
13722 initialization sequences and generate direct calls to locally-defined
13723 functions. This mode is selected by @option{-mno-shared}.
13725 @option{-mno-shared} depends on binutils 2.16 or higher and generates
13726 objects that can only be linked by the GNU linker. However, the option
13727 does not affect the ABI of the final executable; it only affects the ABI
13728 of relocatable objects. Using @option{-mno-shared} will generally make
13729 executables both smaller and quicker.
13731 @option{-mshared} is the default.
13737 Assume (do not assume) that the static and dynamic linkers
13738 support PLTs and copy relocations. This option only affects
13739 @samp{-mno-shared -mabicalls}. For the n64 ABI, this option
13740 has no effect without @samp{-msym32}.
13742 You can make @option{-mplt} the default by configuring
13743 GCC with @option{--with-mips-plt}. The default is
13744 @option{-mno-plt} otherwise.
13750 Lift (do not lift) the usual restrictions on the size of the global
13753 GCC normally uses a single instruction to load values from the GOT@.
13754 While this is relatively efficient, it will only work if the GOT
13755 is smaller than about 64k. Anything larger will cause the linker
13756 to report an error such as:
13758 @cindex relocation truncated to fit (MIPS)
13760 relocation truncated to fit: R_MIPS_GOT16 foobar
13763 If this happens, you should recompile your code with @option{-mxgot}.
13764 It should then work with very large GOTs, although it will also be
13765 less efficient, since it will take three instructions to fetch the
13766 value of a global symbol.
13768 Note that some linkers can create multiple GOTs. If you have such a
13769 linker, you should only need to use @option{-mxgot} when a single object
13770 file accesses more than 64k's worth of GOT entries. Very few do.
13772 These options have no effect unless GCC is generating position
13777 Assume that general-purpose registers are 32 bits wide.
13781 Assume that general-purpose registers are 64 bits wide.
13785 Assume that floating-point registers are 32 bits wide.
13789 Assume that floating-point registers are 64 bits wide.
13792 @opindex mhard-float
13793 Use floating-point coprocessor instructions.
13796 @opindex msoft-float
13797 Do not use floating-point coprocessor instructions. Implement
13798 floating-point calculations using library calls instead.
13800 @item -msingle-float
13801 @opindex msingle-float
13802 Assume that the floating-point coprocessor only supports single-precision
13805 @item -mdouble-float
13806 @opindex mdouble-float
13807 Assume that the floating-point coprocessor supports double-precision
13808 operations. This is the default.
13814 Use (do not use) @samp{ll}, @samp{sc}, and @samp{sync} instructions to
13815 implement atomic memory built-in functions. When neither option is
13816 specified, GCC will use the instructions if the target architecture
13819 @option{-mllsc} is useful if the runtime environment can emulate the
13820 instructions and @option{-mno-llsc} can be useful when compiling for
13821 nonstandard ISAs. You can make either option the default by
13822 configuring GCC with @option{--with-llsc} and @option{--without-llsc}
13823 respectively. @option{--with-llsc} is the default for some
13824 configurations; see the installation documentation for details.
13830 Use (do not use) revision 1 of the MIPS DSP ASE@.
13831 @xref{MIPS DSP Built-in Functions}. This option defines the
13832 preprocessor macro @samp{__mips_dsp}. It also defines
13833 @samp{__mips_dsp_rev} to 1.
13839 Use (do not use) revision 2 of the MIPS DSP ASE@.
13840 @xref{MIPS DSP Built-in Functions}. This option defines the
13841 preprocessor macros @samp{__mips_dsp} and @samp{__mips_dspr2}.
13842 It also defines @samp{__mips_dsp_rev} to 2.
13845 @itemx -mno-smartmips
13846 @opindex msmartmips
13847 @opindex mno-smartmips
13848 Use (do not use) the MIPS SmartMIPS ASE.
13850 @item -mpaired-single
13851 @itemx -mno-paired-single
13852 @opindex mpaired-single
13853 @opindex mno-paired-single
13854 Use (do not use) paired-single floating-point instructions.
13855 @xref{MIPS Paired-Single Support}. This option requires
13856 hardware floating-point support to be enabled.
13862 Use (do not use) MIPS Digital Media Extension instructions.
13863 This option can only be used when generating 64-bit code and requires
13864 hardware floating-point support to be enabled.
13869 @opindex mno-mips3d
13870 Use (do not use) the MIPS-3D ASE@. @xref{MIPS-3D Built-in Functions}.
13871 The option @option{-mips3d} implies @option{-mpaired-single}.
13877 Use (do not use) MT Multithreading instructions.
13881 Force @code{long} types to be 64 bits wide. See @option{-mlong32} for
13882 an explanation of the default and the way that the pointer size is
13887 Force @code{long}, @code{int}, and pointer types to be 32 bits wide.
13889 The default size of @code{int}s, @code{long}s and pointers depends on
13890 the ABI@. All the supported ABIs use 32-bit @code{int}s. The n64 ABI
13891 uses 64-bit @code{long}s, as does the 64-bit EABI; the others use
13892 32-bit @code{long}s. Pointers are the same size as @code{long}s,
13893 or the same size as integer registers, whichever is smaller.
13899 Assume (do not assume) that all symbols have 32-bit values, regardless
13900 of the selected ABI@. This option is useful in combination with
13901 @option{-mabi=64} and @option{-mno-abicalls} because it allows GCC
13902 to generate shorter and faster references to symbolic addresses.
13906 Put definitions of externally-visible data in a small data section
13907 if that data is no bigger than @var{num} bytes. GCC can then access
13908 the data more efficiently; see @option{-mgpopt} for details.
13910 The default @option{-G} option depends on the configuration.
13912 @item -mlocal-sdata
13913 @itemx -mno-local-sdata
13914 @opindex mlocal-sdata
13915 @opindex mno-local-sdata
13916 Extend (do not extend) the @option{-G} behavior to local data too,
13917 such as to static variables in C@. @option{-mlocal-sdata} is the
13918 default for all configurations.
13920 If the linker complains that an application is using too much small data,
13921 you might want to try rebuilding the less performance-critical parts with
13922 @option{-mno-local-sdata}. You might also want to build large
13923 libraries with @option{-mno-local-sdata}, so that the libraries leave
13924 more room for the main program.
13926 @item -mextern-sdata
13927 @itemx -mno-extern-sdata
13928 @opindex mextern-sdata
13929 @opindex mno-extern-sdata
13930 Assume (do not assume) that externally-defined data will be in
13931 a small data section if that data is within the @option{-G} limit.
13932 @option{-mextern-sdata} is the default for all configurations.
13934 If you compile a module @var{Mod} with @option{-mextern-sdata} @option{-G
13935 @var{num}} @option{-mgpopt}, and @var{Mod} references a variable @var{Var}
13936 that is no bigger than @var{num} bytes, you must make sure that @var{Var}
13937 is placed in a small data section. If @var{Var} is defined by another
13938 module, you must either compile that module with a high-enough
13939 @option{-G} setting or attach a @code{section} attribute to @var{Var}'s
13940 definition. If @var{Var} is common, you must link the application
13941 with a high-enough @option{-G} setting.
13943 The easiest way of satisfying these restrictions is to compile
13944 and link every module with the same @option{-G} option. However,
13945 you may wish to build a library that supports several different
13946 small data limits. You can do this by compiling the library with
13947 the highest supported @option{-G} setting and additionally using
13948 @option{-mno-extern-sdata} to stop the library from making assumptions
13949 about externally-defined data.
13955 Use (do not use) GP-relative accesses for symbols that are known to be
13956 in a small data section; see @option{-G}, @option{-mlocal-sdata} and
13957 @option{-mextern-sdata}. @option{-mgpopt} is the default for all
13960 @option{-mno-gpopt} is useful for cases where the @code{$gp} register
13961 might not hold the value of @code{_gp}. For example, if the code is
13962 part of a library that might be used in a boot monitor, programs that
13963 call boot monitor routines will pass an unknown value in @code{$gp}.
13964 (In such situations, the boot monitor itself would usually be compiled
13965 with @option{-G0}.)
13967 @option{-mno-gpopt} implies @option{-mno-local-sdata} and
13968 @option{-mno-extern-sdata}.
13970 @item -membedded-data
13971 @itemx -mno-embedded-data
13972 @opindex membedded-data
13973 @opindex mno-embedded-data
13974 Allocate variables to the read-only data section first if possible, then
13975 next in the small data section if possible, otherwise in data. This gives
13976 slightly slower code than the default, but reduces the amount of RAM required
13977 when executing, and thus may be preferred for some embedded systems.
13979 @item -muninit-const-in-rodata
13980 @itemx -mno-uninit-const-in-rodata
13981 @opindex muninit-const-in-rodata
13982 @opindex mno-uninit-const-in-rodata
13983 Put uninitialized @code{const} variables in the read-only data section.
13984 This option is only meaningful in conjunction with @option{-membedded-data}.
13986 @item -mcode-readable=@var{setting}
13987 @opindex mcode-readable
13988 Specify whether GCC may generate code that reads from executable sections.
13989 There are three possible settings:
13992 @item -mcode-readable=yes
13993 Instructions may freely access executable sections. This is the
13996 @item -mcode-readable=pcrel
13997 MIPS16 PC-relative load instructions can access executable sections,
13998 but other instructions must not do so. This option is useful on 4KSc
13999 and 4KSd processors when the code TLBs have the Read Inhibit bit set.
14000 It is also useful on processors that can be configured to have a dual
14001 instruction/data SRAM interface and that, like the M4K, automatically
14002 redirect PC-relative loads to the instruction RAM.
14004 @item -mcode-readable=no
14005 Instructions must not access executable sections. This option can be
14006 useful on targets that are configured to have a dual instruction/data
14007 SRAM interface but that (unlike the M4K) do not automatically redirect
14008 PC-relative loads to the instruction RAM.
14011 @item -msplit-addresses
14012 @itemx -mno-split-addresses
14013 @opindex msplit-addresses
14014 @opindex mno-split-addresses
14015 Enable (disable) use of the @code{%hi()} and @code{%lo()} assembler
14016 relocation operators. This option has been superseded by
14017 @option{-mexplicit-relocs} but is retained for backwards compatibility.
14019 @item -mexplicit-relocs
14020 @itemx -mno-explicit-relocs
14021 @opindex mexplicit-relocs
14022 @opindex mno-explicit-relocs
14023 Use (do not use) assembler relocation operators when dealing with symbolic
14024 addresses. The alternative, selected by @option{-mno-explicit-relocs},
14025 is to use assembler macros instead.
14027 @option{-mexplicit-relocs} is the default if GCC was configured
14028 to use an assembler that supports relocation operators.
14030 @item -mcheck-zero-division
14031 @itemx -mno-check-zero-division
14032 @opindex mcheck-zero-division
14033 @opindex mno-check-zero-division
14034 Trap (do not trap) on integer division by zero.
14036 The default is @option{-mcheck-zero-division}.
14038 @item -mdivide-traps
14039 @itemx -mdivide-breaks
14040 @opindex mdivide-traps
14041 @opindex mdivide-breaks
14042 MIPS systems check for division by zero by generating either a
14043 conditional trap or a break instruction. Using traps results in
14044 smaller code, but is only supported on MIPS II and later. Also, some
14045 versions of the Linux kernel have a bug that prevents trap from
14046 generating the proper signal (@code{SIGFPE}). Use @option{-mdivide-traps} to
14047 allow conditional traps on architectures that support them and
14048 @option{-mdivide-breaks} to force the use of breaks.
14050 The default is usually @option{-mdivide-traps}, but this can be
14051 overridden at configure time using @option{--with-divide=breaks}.
14052 Divide-by-zero checks can be completely disabled using
14053 @option{-mno-check-zero-division}.
14058 @opindex mno-memcpy
14059 Force (do not force) the use of @code{memcpy()} for non-trivial block
14060 moves. The default is @option{-mno-memcpy}, which allows GCC to inline
14061 most constant-sized copies.
14064 @itemx -mno-long-calls
14065 @opindex mlong-calls
14066 @opindex mno-long-calls
14067 Disable (do not disable) use of the @code{jal} instruction. Calling
14068 functions using @code{jal} is more efficient but requires the caller
14069 and callee to be in the same 256 megabyte segment.
14071 This option has no effect on abicalls code. The default is
14072 @option{-mno-long-calls}.
14078 Enable (disable) use of the @code{mad}, @code{madu} and @code{mul}
14079 instructions, as provided by the R4650 ISA@.
14082 @itemx -mno-fused-madd
14083 @opindex mfused-madd
14084 @opindex mno-fused-madd
14085 Enable (disable) use of the floating point multiply-accumulate
14086 instructions, when they are available. The default is
14087 @option{-mfused-madd}.
14089 When multiply-accumulate instructions are used, the intermediate
14090 product is calculated to infinite precision and is not subject to
14091 the FCSR Flush to Zero bit. This may be undesirable in some
14096 Tell the MIPS assembler to not run its preprocessor over user
14097 assembler files (with a @samp{.s} suffix) when assembling them.
14100 @itemx -mno-fix-r4000
14101 @opindex mfix-r4000
14102 @opindex mno-fix-r4000
14103 Work around certain R4000 CPU errata:
14106 A double-word or a variable shift may give an incorrect result if executed
14107 immediately after starting an integer division.
14109 A double-word or a variable shift may give an incorrect result if executed
14110 while an integer multiplication is in progress.
14112 An integer division may give an incorrect result if started in a delay slot
14113 of a taken branch or a jump.
14117 @itemx -mno-fix-r4400
14118 @opindex mfix-r4400
14119 @opindex mno-fix-r4400
14120 Work around certain R4400 CPU errata:
14123 A double-word or a variable shift may give an incorrect result if executed
14124 immediately after starting an integer division.
14128 @itemx -mno-fix-r10000
14129 @opindex mfix-r10000
14130 @opindex mno-fix-r10000
14131 Work around certain R10000 errata:
14134 @code{ll}/@code{sc} sequences may not behave atomically on revisions
14135 prior to 3.0. They may deadlock on revisions 2.6 and earlier.
14138 This option can only be used if the target architecture supports
14139 branch-likely instructions. @option{-mfix-r10000} is the default when
14140 @option{-march=r10000} is used; @option{-mno-fix-r10000} is the default
14144 @itemx -mno-fix-vr4120
14145 @opindex mfix-vr4120
14146 Work around certain VR4120 errata:
14149 @code{dmultu} does not always produce the correct result.
14151 @code{div} and @code{ddiv} do not always produce the correct result if one
14152 of the operands is negative.
14154 The workarounds for the division errata rely on special functions in
14155 @file{libgcc.a}. At present, these functions are only provided by
14156 the @code{mips64vr*-elf} configurations.
14158 Other VR4120 errata require a nop to be inserted between certain pairs of
14159 instructions. These errata are handled by the assembler, not by GCC itself.
14162 @opindex mfix-vr4130
14163 Work around the VR4130 @code{mflo}/@code{mfhi} errata. The
14164 workarounds are implemented by the assembler rather than by GCC,
14165 although GCC will avoid using @code{mflo} and @code{mfhi} if the
14166 VR4130 @code{macc}, @code{macchi}, @code{dmacc} and @code{dmacchi}
14167 instructions are available instead.
14170 @itemx -mno-fix-sb1
14172 Work around certain SB-1 CPU core errata.
14173 (This flag currently works around the SB-1 revision 2
14174 ``F1'' and ``F2'' floating point errata.)
14176 @item -mr10k-cache-barrier=@var{setting}
14177 @opindex mr10k-cache-barrier
14178 Specify whether GCC should insert cache barriers to avoid the
14179 side-effects of speculation on R10K processors.
14181 In common with many processors, the R10K tries to predict the outcome
14182 of a conditional branch and speculatively executes instructions from
14183 the ``taken'' branch. It later aborts these instructions if the
14184 predicted outcome was wrong. However, on the R10K, even aborted
14185 instructions can have side effects.
14187 This problem only affects kernel stores and, depending on the system,
14188 kernel loads. As an example, a speculatively-executed store may load
14189 the target memory into cache and mark the cache line as dirty, even if
14190 the store itself is later aborted. If a DMA operation writes to the
14191 same area of memory before the ``dirty'' line is flushed, the cached
14192 data will overwrite the DMA-ed data. See the R10K processor manual
14193 for a full description, including other potential problems.
14195 One workaround is to insert cache barrier instructions before every memory
14196 access that might be speculatively executed and that might have side
14197 effects even if aborted. @option{-mr10k-cache-barrier=@var{setting}}
14198 controls GCC's implementation of this workaround. It assumes that
14199 aborted accesses to any byte in the following regions will not have
14204 the memory occupied by the current function's stack frame;
14207 the memory occupied by an incoming stack argument;
14210 the memory occupied by an object with a link-time-constant address.
14213 It is the kernel's responsibility to ensure that speculative
14214 accesses to these regions are indeed safe.
14216 If the input program contains a function declaration such as:
14222 then the implementation of @code{foo} must allow @code{j foo} and
14223 @code{jal foo} to be executed speculatively. GCC honors this
14224 restriction for functions it compiles itself. It expects non-GCC
14225 functions (such as hand-written assembly code) to do the same.
14227 The option has three forms:
14230 @item -mr10k-cache-barrier=load-store
14231 Insert a cache barrier before a load or store that might be
14232 speculatively executed and that might have side effects even
14235 @item -mr10k-cache-barrier=store
14236 Insert a cache barrier before a store that might be speculatively
14237 executed and that might have side effects even if aborted.
14239 @item -mr10k-cache-barrier=none
14240 Disable the insertion of cache barriers. This is the default setting.
14243 @item -mflush-func=@var{func}
14244 @itemx -mno-flush-func
14245 @opindex mflush-func
14246 Specifies the function to call to flush the I and D caches, or to not
14247 call any such function. If called, the function must take the same
14248 arguments as the common @code{_flush_func()}, that is, the address of the
14249 memory range for which the cache is being flushed, the size of the
14250 memory range, and the number 3 (to flush both caches). The default
14251 depends on the target GCC was configured for, but commonly is either
14252 @samp{_flush_func} or @samp{__cpu_flush}.
14254 @item mbranch-cost=@var{num}
14255 @opindex mbranch-cost
14256 Set the cost of branches to roughly @var{num} ``simple'' instructions.
14257 This cost is only a heuristic and is not guaranteed to produce
14258 consistent results across releases. A zero cost redundantly selects
14259 the default, which is based on the @option{-mtune} setting.
14261 @item -mbranch-likely
14262 @itemx -mno-branch-likely
14263 @opindex mbranch-likely
14264 @opindex mno-branch-likely
14265 Enable or disable use of Branch Likely instructions, regardless of the
14266 default for the selected architecture. By default, Branch Likely
14267 instructions may be generated if they are supported by the selected
14268 architecture. An exception is for the MIPS32 and MIPS64 architectures
14269 and processors which implement those architectures; for those, Branch
14270 Likely instructions will not be generated by default because the MIPS32
14271 and MIPS64 architectures specifically deprecate their use.
14273 @item -mfp-exceptions
14274 @itemx -mno-fp-exceptions
14275 @opindex mfp-exceptions
14276 Specifies whether FP exceptions are enabled. This affects how we schedule
14277 FP instructions for some processors. The default is that FP exceptions are
14280 For instance, on the SB-1, if FP exceptions are disabled, and we are emitting
14281 64-bit code, then we can use both FP pipes. Otherwise, we can only use one
14284 @item -mvr4130-align
14285 @itemx -mno-vr4130-align
14286 @opindex mvr4130-align
14287 The VR4130 pipeline is two-way superscalar, but can only issue two
14288 instructions together if the first one is 8-byte aligned. When this
14289 option is enabled, GCC will align pairs of instructions that it
14290 thinks should execute in parallel.
14292 This option only has an effect when optimizing for the VR4130.
14293 It normally makes code faster, but at the expense of making it bigger.
14294 It is enabled by default at optimization level @option{-O3}.
14299 Enable (disable) generation of @code{synci} instructions on
14300 architectures that support it. The @code{synci} instructions (if
14301 enabled) will be generated when @code{__builtin___clear_cache()} is
14304 This option defaults to @code{-mno-synci}, but the default can be
14305 overridden by configuring with @code{--with-synci}.
14307 When compiling code for single processor systems, it is generally safe
14308 to use @code{synci}. However, on many multi-core (SMP) systems, it
14309 will not invalidate the instruction caches on all cores and may lead
14310 to undefined behavior.
14312 @item -mrelax-pic-calls
14313 @itemx -mno-relax-pic-calls
14314 @opindex mrelax-pic-calls
14315 Try to turn PIC calls that are normally dispatched via register
14316 @code{$25} into direct calls. This is only possible if the linker can
14317 resolve the destination at link-time and if the destination is within
14318 range for a direct call.
14320 @option{-mrelax-pic-calls} is the default if GCC was configured to use
14321 an assembler and a linker that supports the @code{.reloc} assembly
14322 directive and @code{-mexplicit-relocs} is in effect. With
14323 @code{-mno-explicit-relocs}, this optimization can be performed by the
14324 assembler and the linker alone without help from the compiler.
14326 @item -mmcount-ra-address
14327 @itemx -mno-mcount-ra-address
14328 @opindex mmcount-ra-address
14329 @opindex mno-mcount-ra-address
14330 Emit (do not emit) code that allows @code{_mcount} to modify the
14331 calling function's return address. When enabled, this option extends
14332 the usual @code{_mcount} interface with a new @var{ra-address}
14333 parameter, which has type @code{intptr_t *} and is passed in register
14334 @code{$12}. @code{_mcount} can then modify the return address by
14335 doing both of the following:
14338 Returning the new address in register @code{$31}.
14340 Storing the new address in @code{*@var{ra-address}},
14341 if @var{ra-address} is nonnull.
14344 The default is @option{-mno-mcount-ra-address}.
14349 @subsection MMIX Options
14350 @cindex MMIX Options
14352 These options are defined for the MMIX:
14356 @itemx -mno-libfuncs
14358 @opindex mno-libfuncs
14359 Specify that intrinsic library functions are being compiled, passing all
14360 values in registers, no matter the size.
14363 @itemx -mno-epsilon
14365 @opindex mno-epsilon
14366 Generate floating-point comparison instructions that compare with respect
14367 to the @code{rE} epsilon register.
14369 @item -mabi=mmixware
14371 @opindex mabi=mmixware
14373 Generate code that passes function parameters and return values that (in
14374 the called function) are seen as registers @code{$0} and up, as opposed to
14375 the GNU ABI which uses global registers @code{$231} and up.
14377 @item -mzero-extend
14378 @itemx -mno-zero-extend
14379 @opindex mzero-extend
14380 @opindex mno-zero-extend
14381 When reading data from memory in sizes shorter than 64 bits, use (do not
14382 use) zero-extending load instructions by default, rather than
14383 sign-extending ones.
14386 @itemx -mno-knuthdiv
14388 @opindex mno-knuthdiv
14389 Make the result of a division yielding a remainder have the same sign as
14390 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
14391 remainder follows the sign of the dividend. Both methods are
14392 arithmetically valid, the latter being almost exclusively used.
14394 @item -mtoplevel-symbols
14395 @itemx -mno-toplevel-symbols
14396 @opindex mtoplevel-symbols
14397 @opindex mno-toplevel-symbols
14398 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
14399 code can be used with the @code{PREFIX} assembly directive.
14403 Generate an executable in the ELF format, rather than the default
14404 @samp{mmo} format used by the @command{mmix} simulator.
14406 @item -mbranch-predict
14407 @itemx -mno-branch-predict
14408 @opindex mbranch-predict
14409 @opindex mno-branch-predict
14410 Use (do not use) the probable-branch instructions, when static branch
14411 prediction indicates a probable branch.
14413 @item -mbase-addresses
14414 @itemx -mno-base-addresses
14415 @opindex mbase-addresses
14416 @opindex mno-base-addresses
14417 Generate (do not generate) code that uses @emph{base addresses}. Using a
14418 base address automatically generates a request (handled by the assembler
14419 and the linker) for a constant to be set up in a global register. The
14420 register is used for one or more base address requests within the range 0
14421 to 255 from the value held in the register. The generally leads to short
14422 and fast code, but the number of different data items that can be
14423 addressed is limited. This means that a program that uses lots of static
14424 data may require @option{-mno-base-addresses}.
14426 @item -msingle-exit
14427 @itemx -mno-single-exit
14428 @opindex msingle-exit
14429 @opindex mno-single-exit
14430 Force (do not force) generated code to have a single exit point in each
14434 @node MN10300 Options
14435 @subsection MN10300 Options
14436 @cindex MN10300 options
14438 These @option{-m} options are defined for Matsushita MN10300 architectures:
14443 Generate code to avoid bugs in the multiply instructions for the MN10300
14444 processors. This is the default.
14446 @item -mno-mult-bug
14447 @opindex mno-mult-bug
14448 Do not generate code to avoid bugs in the multiply instructions for the
14449 MN10300 processors.
14453 Generate code which uses features specific to the AM33 processor.
14457 Do not generate code which uses features specific to the AM33 processor. This
14460 @item -mreturn-pointer-on-d0
14461 @opindex mreturn-pointer-on-d0
14462 When generating a function which returns a pointer, return the pointer
14463 in both @code{a0} and @code{d0}. Otherwise, the pointer is returned
14464 only in a0, and attempts to call such functions without a prototype
14465 would result in errors. Note that this option is on by default; use
14466 @option{-mno-return-pointer-on-d0} to disable it.
14470 Do not link in the C run-time initialization object file.
14474 Indicate to the linker that it should perform a relaxation optimization pass
14475 to shorten branches, calls and absolute memory addresses. This option only
14476 has an effect when used on the command line for the final link step.
14478 This option makes symbolic debugging impossible.
14481 @node PDP-11 Options
14482 @subsection PDP-11 Options
14483 @cindex PDP-11 Options
14485 These options are defined for the PDP-11:
14490 Use hardware FPP floating point. This is the default. (FIS floating
14491 point on the PDP-11/40 is not supported.)
14494 @opindex msoft-float
14495 Do not use hardware floating point.
14499 Return floating-point results in ac0 (fr0 in Unix assembler syntax).
14503 Return floating-point results in memory. This is the default.
14507 Generate code for a PDP-11/40.
14511 Generate code for a PDP-11/45. This is the default.
14515 Generate code for a PDP-11/10.
14517 @item -mbcopy-builtin
14518 @opindex mbcopy-builtin
14519 Use inline @code{movmemhi} patterns for copying memory. This is the
14524 Do not use inline @code{movmemhi} patterns for copying memory.
14530 Use 16-bit @code{int}. This is the default.
14536 Use 32-bit @code{int}.
14539 @itemx -mno-float32
14541 @opindex mno-float32
14542 Use 64-bit @code{float}. This is the default.
14545 @itemx -mno-float64
14547 @opindex mno-float64
14548 Use 32-bit @code{float}.
14552 Use @code{abshi2} pattern. This is the default.
14556 Do not use @code{abshi2} pattern.
14558 @item -mbranch-expensive
14559 @opindex mbranch-expensive
14560 Pretend that branches are expensive. This is for experimenting with
14561 code generation only.
14563 @item -mbranch-cheap
14564 @opindex mbranch-cheap
14565 Do not pretend that branches are expensive. This is the default.
14569 Generate code for a system with split I&D@.
14573 Generate code for a system without split I&D@. This is the default.
14577 Use Unix assembler syntax. This is the default when configured for
14578 @samp{pdp11-*-bsd}.
14582 Use DEC assembler syntax. This is the default when configured for any
14583 PDP-11 target other than @samp{pdp11-*-bsd}.
14586 @node picoChip Options
14587 @subsection picoChip Options
14588 @cindex picoChip options
14590 These @samp{-m} options are defined for picoChip implementations:
14594 @item -mae=@var{ae_type}
14596 Set the instruction set, register set, and instruction scheduling
14597 parameters for array element type @var{ae_type}. Supported values
14598 for @var{ae_type} are @samp{ANY}, @samp{MUL}, and @samp{MAC}.
14600 @option{-mae=ANY} selects a completely generic AE type. Code
14601 generated with this option will run on any of the other AE types. The
14602 code will not be as efficient as it would be if compiled for a specific
14603 AE type, and some types of operation (e.g., multiplication) will not
14604 work properly on all types of AE.
14606 @option{-mae=MUL} selects a MUL AE type. This is the most useful AE type
14607 for compiled code, and is the default.
14609 @option{-mae=MAC} selects a DSP-style MAC AE. Code compiled with this
14610 option may suffer from poor performance of byte (char) manipulation,
14611 since the DSP AE does not provide hardware support for byte load/stores.
14613 @item -msymbol-as-address
14614 Enable the compiler to directly use a symbol name as an address in a
14615 load/store instruction, without first loading it into a
14616 register. Typically, the use of this option will generate larger
14617 programs, which run faster than when the option isn't used. However, the
14618 results vary from program to program, so it is left as a user option,
14619 rather than being permanently enabled.
14621 @item -mno-inefficient-warnings
14622 Disables warnings about the generation of inefficient code. These
14623 warnings can be generated, for example, when compiling code which
14624 performs byte-level memory operations on the MAC AE type. The MAC AE has
14625 no hardware support for byte-level memory operations, so all byte
14626 load/stores must be synthesized from word load/store operations. This is
14627 inefficient and a warning will be generated indicating to the programmer
14628 that they should rewrite the code to avoid byte operations, or to target
14629 an AE type which has the necessary hardware support. This option enables
14630 the warning to be turned off.
14634 @node PowerPC Options
14635 @subsection PowerPC Options
14636 @cindex PowerPC options
14638 These are listed under @xref{RS/6000 and PowerPC Options}.
14640 @node RS/6000 and PowerPC Options
14641 @subsection IBM RS/6000 and PowerPC Options
14642 @cindex RS/6000 and PowerPC Options
14643 @cindex IBM RS/6000 and PowerPC Options
14645 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
14652 @itemx -mno-powerpc
14653 @itemx -mpowerpc-gpopt
14654 @itemx -mno-powerpc-gpopt
14655 @itemx -mpowerpc-gfxopt
14656 @itemx -mno-powerpc-gfxopt
14658 @itemx -mno-powerpc64
14662 @itemx -mno-popcntb
14664 @itemx -mno-popcntd
14672 @itemx -mno-hard-dfp
14676 @opindex mno-power2
14678 @opindex mno-powerpc
14679 @opindex mpowerpc-gpopt
14680 @opindex mno-powerpc-gpopt
14681 @opindex mpowerpc-gfxopt
14682 @opindex mno-powerpc-gfxopt
14683 @opindex mpowerpc64
14684 @opindex mno-powerpc64
14688 @opindex mno-popcntb
14690 @opindex mno-popcntd
14696 @opindex mno-mfpgpr
14698 @opindex mno-hard-dfp
14699 GCC supports two related instruction set architectures for the
14700 RS/6000 and PowerPC@. The @dfn{POWER} instruction set are those
14701 instructions supported by the @samp{rios} chip set used in the original
14702 RS/6000 systems and the @dfn{PowerPC} instruction set is the
14703 architecture of the Freescale MPC5xx, MPC6xx, MPC8xx microprocessors, and
14704 the IBM 4xx, 6xx, and follow-on microprocessors.
14706 Neither architecture is a subset of the other. However there is a
14707 large common subset of instructions supported by both. An MQ
14708 register is included in processors supporting the POWER architecture.
14710 You use these options to specify which instructions are available on the
14711 processor you are using. The default value of these options is
14712 determined when configuring GCC@. Specifying the
14713 @option{-mcpu=@var{cpu_type}} overrides the specification of these
14714 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
14715 rather than the options listed above.
14717 The @option{-mpower} option allows GCC to generate instructions that
14718 are found only in the POWER architecture and to use the MQ register.
14719 Specifying @option{-mpower2} implies @option{-power} and also allows GCC
14720 to generate instructions that are present in the POWER2 architecture but
14721 not the original POWER architecture.
14723 The @option{-mpowerpc} option allows GCC to generate instructions that
14724 are found only in the 32-bit subset of the PowerPC architecture.
14725 Specifying @option{-mpowerpc-gpopt} implies @option{-mpowerpc} and also allows
14726 GCC to use the optional PowerPC architecture instructions in the
14727 General Purpose group, including floating-point square root. Specifying
14728 @option{-mpowerpc-gfxopt} implies @option{-mpowerpc} and also allows GCC to
14729 use the optional PowerPC architecture instructions in the Graphics
14730 group, including floating-point select.
14732 The @option{-mmfcrf} option allows GCC to generate the move from
14733 condition register field instruction implemented on the POWER4
14734 processor and other processors that support the PowerPC V2.01
14736 The @option{-mpopcntb} option allows GCC to generate the popcount and
14737 double precision FP reciprocal estimate instruction implemented on the
14738 POWER5 processor and other processors that support the PowerPC V2.02
14740 The @option{-mpopcntd} option allows GCC to generate the popcount
14741 instruction implemented on the POWER7 processor and other processors
14742 that support the PowerPC V2.06 architecture.
14743 The @option{-mfprnd} option allows GCC to generate the FP round to
14744 integer instructions implemented on the POWER5+ processor and other
14745 processors that support the PowerPC V2.03 architecture.
14746 The @option{-mcmpb} option allows GCC to generate the compare bytes
14747 instruction implemented on the POWER6 processor and other processors
14748 that support the PowerPC V2.05 architecture.
14749 The @option{-mmfpgpr} option allows GCC to generate the FP move to/from
14750 general purpose register instructions implemented on the POWER6X
14751 processor and other processors that support the extended PowerPC V2.05
14753 The @option{-mhard-dfp} option allows GCC to generate the decimal floating
14754 point instructions implemented on some POWER processors.
14756 The @option{-mpowerpc64} option allows GCC to generate the additional
14757 64-bit instructions that are found in the full PowerPC64 architecture
14758 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
14759 @option{-mno-powerpc64}.
14761 If you specify both @option{-mno-power} and @option{-mno-powerpc}, GCC
14762 will use only the instructions in the common subset of both
14763 architectures plus some special AIX common-mode calls, and will not use
14764 the MQ register. Specifying both @option{-mpower} and @option{-mpowerpc}
14765 permits GCC to use any instruction from either architecture and to
14766 allow use of the MQ register; specify this for the Motorola MPC601.
14768 @item -mnew-mnemonics
14769 @itemx -mold-mnemonics
14770 @opindex mnew-mnemonics
14771 @opindex mold-mnemonics
14772 Select which mnemonics to use in the generated assembler code. With
14773 @option{-mnew-mnemonics}, GCC uses the assembler mnemonics defined for
14774 the PowerPC architecture. With @option{-mold-mnemonics} it uses the
14775 assembler mnemonics defined for the POWER architecture. Instructions
14776 defined in only one architecture have only one mnemonic; GCC uses that
14777 mnemonic irrespective of which of these options is specified.
14779 GCC defaults to the mnemonics appropriate for the architecture in
14780 use. Specifying @option{-mcpu=@var{cpu_type}} sometimes overrides the
14781 value of these option. Unless you are building a cross-compiler, you
14782 should normally not specify either @option{-mnew-mnemonics} or
14783 @option{-mold-mnemonics}, but should instead accept the default.
14785 @item -mcpu=@var{cpu_type}
14787 Set architecture type, register usage, choice of mnemonics, and
14788 instruction scheduling parameters for machine type @var{cpu_type}.
14789 Supported values for @var{cpu_type} are @samp{401}, @samp{403},
14790 @samp{405}, @samp{405fp}, @samp{440}, @samp{440fp}, @samp{464}, @samp{464fp},
14791 @samp{476}, @samp{476fp}, @samp{505}, @samp{601}, @samp{602}, @samp{603},
14792 @samp{603e}, @samp{604}, @samp{604e}, @samp{620}, @samp{630}, @samp{740},
14793 @samp{7400}, @samp{7450}, @samp{750}, @samp{801}, @samp{821}, @samp{823},
14794 @samp{860}, @samp{970}, @samp{8540}, @samp{a2}, @samp{e300c2},
14795 @samp{e300c3}, @samp{e500mc}, @samp{e500mc64}, @samp{ec603e}, @samp{G3},
14796 @samp{G4}, @samp{G5}, @samp{power}, @samp{power2}, @samp{power3},
14797 @samp{power4}, @samp{power5}, @samp{power5+}, @samp{power6}, @samp{power6x},
14798 @samp{power7}, @samp{common}, @samp{powerpc}, @samp{powerpc64}, @samp{rios},
14799 @samp{rios1}, @samp{rios2}, @samp{rsc}, and @samp{rs64}.
14801 @option{-mcpu=common} selects a completely generic processor. Code
14802 generated under this option will run on any POWER or PowerPC processor.
14803 GCC will use only the instructions in the common subset of both
14804 architectures, and will not use the MQ register. GCC assumes a generic
14805 processor model for scheduling purposes.
14807 @option{-mcpu=power}, @option{-mcpu=power2}, @option{-mcpu=powerpc}, and
14808 @option{-mcpu=powerpc64} specify generic POWER, POWER2, pure 32-bit
14809 PowerPC (i.e., not MPC601), and 64-bit PowerPC architecture machine
14810 types, with an appropriate, generic processor model assumed for
14811 scheduling purposes.
14813 The other options specify a specific processor. Code generated under
14814 those options will run best on that processor, and may not run at all on
14817 The @option{-mcpu} options automatically enable or disable the
14820 @gccoptlist{-maltivec -mfprnd -mhard-float -mmfcrf -mmultiple @gol
14821 -mnew-mnemonics -mpopcntb -mpopcntd -mpower -mpower2 -mpowerpc64 @gol
14822 -mpowerpc-gpopt -mpowerpc-gfxopt -msingle-float -mdouble-float @gol
14823 -msimple-fpu -mstring -mmulhw -mdlmzb -mmfpgpr -mvsx}
14825 The particular options set for any particular CPU will vary between
14826 compiler versions, depending on what setting seems to produce optimal
14827 code for that CPU; it doesn't necessarily reflect the actual hardware's
14828 capabilities. If you wish to set an individual option to a particular
14829 value, you may specify it after the @option{-mcpu} option, like
14830 @samp{-mcpu=970 -mno-altivec}.
14832 On AIX, the @option{-maltivec} and @option{-mpowerpc64} options are
14833 not enabled or disabled by the @option{-mcpu} option at present because
14834 AIX does not have full support for these options. You may still
14835 enable or disable them individually if you're sure it'll work in your
14838 @item -mtune=@var{cpu_type}
14840 Set the instruction scheduling parameters for machine type
14841 @var{cpu_type}, but do not set the architecture type, register usage, or
14842 choice of mnemonics, as @option{-mcpu=@var{cpu_type}} would. The same
14843 values for @var{cpu_type} are used for @option{-mtune} as for
14844 @option{-mcpu}. If both are specified, the code generated will use the
14845 architecture, registers, and mnemonics set by @option{-mcpu}, but the
14846 scheduling parameters set by @option{-mtune}.
14852 Generate code to compute division as reciprocal estimate and iterative
14853 refinement, creating opportunities for increased throughput. This
14854 feature requires: optional PowerPC Graphics instruction set for single
14855 precision and FRE instruction for double precision, assuming divides
14856 cannot generate user-visible traps, and the domain values not include
14857 Infinities, denormals or zero denominator.
14860 @itemx -mno-altivec
14862 @opindex mno-altivec
14863 Generate code that uses (does not use) AltiVec instructions, and also
14864 enable the use of built-in functions that allow more direct access to
14865 the AltiVec instruction set. You may also need to set
14866 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
14872 @opindex mno-vrsave
14873 Generate VRSAVE instructions when generating AltiVec code.
14875 @item -mgen-cell-microcode
14876 @opindex mgen-cell-microcode
14877 Generate Cell microcode instructions
14879 @item -mwarn-cell-microcode
14880 @opindex mwarn-cell-microcode
14881 Warning when a Cell microcode instruction is going to emitted. An example
14882 of a Cell microcode instruction is a variable shift.
14885 @opindex msecure-plt
14886 Generate code that allows ld and ld.so to build executables and shared
14887 libraries with non-exec .plt and .got sections. This is a PowerPC
14888 32-bit SYSV ABI option.
14892 Generate code that uses a BSS .plt section that ld.so fills in, and
14893 requires .plt and .got sections that are both writable and executable.
14894 This is a PowerPC 32-bit SYSV ABI option.
14900 This switch enables or disables the generation of ISEL instructions.
14902 @item -misel=@var{yes/no}
14903 This switch has been deprecated. Use @option{-misel} and
14904 @option{-mno-isel} instead.
14910 This switch enables or disables the generation of SPE simd
14916 @opindex mno-paired
14917 This switch enables or disables the generation of PAIRED simd
14920 @item -mspe=@var{yes/no}
14921 This option has been deprecated. Use @option{-mspe} and
14922 @option{-mno-spe} instead.
14928 Generate code that uses (does not use) vector/scalar (VSX)
14929 instructions, and also enable the use of built-in functions that allow
14930 more direct access to the VSX instruction set.
14932 @item -mfloat-gprs=@var{yes/single/double/no}
14933 @itemx -mfloat-gprs
14934 @opindex mfloat-gprs
14935 This switch enables or disables the generation of floating point
14936 operations on the general purpose registers for architectures that
14939 The argument @var{yes} or @var{single} enables the use of
14940 single-precision floating point operations.
14942 The argument @var{double} enables the use of single and
14943 double-precision floating point operations.
14945 The argument @var{no} disables floating point operations on the
14946 general purpose registers.
14948 This option is currently only available on the MPC854x.
14954 Generate code for 32-bit or 64-bit environments of Darwin and SVR4
14955 targets (including GNU/Linux). The 32-bit environment sets int, long
14956 and pointer to 32 bits and generates code that runs on any PowerPC
14957 variant. The 64-bit environment sets int to 32 bits and long and
14958 pointer to 64 bits, and generates code for PowerPC64, as for
14959 @option{-mpowerpc64}.
14962 @itemx -mno-fp-in-toc
14963 @itemx -mno-sum-in-toc
14964 @itemx -mminimal-toc
14966 @opindex mno-fp-in-toc
14967 @opindex mno-sum-in-toc
14968 @opindex mminimal-toc
14969 Modify generation of the TOC (Table Of Contents), which is created for
14970 every executable file. The @option{-mfull-toc} option is selected by
14971 default. In that case, GCC will allocate at least one TOC entry for
14972 each unique non-automatic variable reference in your program. GCC
14973 will also place floating-point constants in the TOC@. However, only
14974 16,384 entries are available in the TOC@.
14976 If you receive a linker error message that saying you have overflowed
14977 the available TOC space, you can reduce the amount of TOC space used
14978 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
14979 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
14980 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
14981 generate code to calculate the sum of an address and a constant at
14982 run-time instead of putting that sum into the TOC@. You may specify one
14983 or both of these options. Each causes GCC to produce very slightly
14984 slower and larger code at the expense of conserving TOC space.
14986 If you still run out of space in the TOC even when you specify both of
14987 these options, specify @option{-mminimal-toc} instead. This option causes
14988 GCC to make only one TOC entry for every file. When you specify this
14989 option, GCC will produce code that is slower and larger but which
14990 uses extremely little TOC space. You may wish to use this option
14991 only on files that contain less frequently executed code.
14997 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
14998 @code{long} type, and the infrastructure needed to support them.
14999 Specifying @option{-maix64} implies @option{-mpowerpc64} and
15000 @option{-mpowerpc}, while @option{-maix32} disables the 64-bit ABI and
15001 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
15004 @itemx -mno-xl-compat
15005 @opindex mxl-compat
15006 @opindex mno-xl-compat
15007 Produce code that conforms more closely to IBM XL compiler semantics
15008 when using AIX-compatible ABI@. Pass floating-point arguments to
15009 prototyped functions beyond the register save area (RSA) on the stack
15010 in addition to argument FPRs. Do not assume that most significant
15011 double in 128-bit long double value is properly rounded when comparing
15012 values and converting to double. Use XL symbol names for long double
15015 The AIX calling convention was extended but not initially documented to
15016 handle an obscure K&R C case of calling a function that takes the
15017 address of its arguments with fewer arguments than declared. IBM XL
15018 compilers access floating point arguments which do not fit in the
15019 RSA from the stack when a subroutine is compiled without
15020 optimization. Because always storing floating-point arguments on the
15021 stack is inefficient and rarely needed, this option is not enabled by
15022 default and only is necessary when calling subroutines compiled by IBM
15023 XL compilers without optimization.
15027 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
15028 application written to use message passing with special startup code to
15029 enable the application to run. The system must have PE installed in the
15030 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
15031 must be overridden with the @option{-specs=} option to specify the
15032 appropriate directory location. The Parallel Environment does not
15033 support threads, so the @option{-mpe} option and the @option{-pthread}
15034 option are incompatible.
15036 @item -malign-natural
15037 @itemx -malign-power
15038 @opindex malign-natural
15039 @opindex malign-power
15040 On AIX, 32-bit Darwin, and 64-bit PowerPC GNU/Linux, the option
15041 @option{-malign-natural} overrides the ABI-defined alignment of larger
15042 types, such as floating-point doubles, on their natural size-based boundary.
15043 The option @option{-malign-power} instructs GCC to follow the ABI-specified
15044 alignment rules. GCC defaults to the standard alignment defined in the ABI@.
15046 On 64-bit Darwin, natural alignment is the default, and @option{-malign-power}
15050 @itemx -mhard-float
15051 @opindex msoft-float
15052 @opindex mhard-float
15053 Generate code that does not use (uses) the floating-point register set.
15054 Software floating point emulation is provided if you use the
15055 @option{-msoft-float} option, and pass the option to GCC when linking.
15057 @item -msingle-float
15058 @itemx -mdouble-float
15059 @opindex msingle-float
15060 @opindex mdouble-float
15061 Generate code for single or double-precision floating point operations.
15062 @option{-mdouble-float} implies @option{-msingle-float}.
15065 @opindex msimple-fpu
15066 Do not generate sqrt and div instructions for hardware floating point unit.
15070 Specify type of floating point unit. Valid values are @var{sp_lite}
15071 (equivalent to -msingle-float -msimple-fpu), @var{dp_lite} (equivalent
15072 to -mdouble-float -msimple-fpu), @var{sp_full} (equivalent to -msingle-float),
15073 and @var{dp_full} (equivalent to -mdouble-float).
15076 @opindex mxilinx-fpu
15077 Perform optimizations for floating point unit on Xilinx PPC 405/440.
15080 @itemx -mno-multiple
15082 @opindex mno-multiple
15083 Generate code that uses (does not use) the load multiple word
15084 instructions and the store multiple word instructions. These
15085 instructions are generated by default on POWER systems, and not
15086 generated on PowerPC systems. Do not use @option{-mmultiple} on little
15087 endian PowerPC systems, since those instructions do not work when the
15088 processor is in little endian mode. The exceptions are PPC740 and
15089 PPC750 which permit the instructions usage in little endian mode.
15094 @opindex mno-string
15095 Generate code that uses (does not use) the load string instructions
15096 and the store string word instructions to save multiple registers and
15097 do small block moves. These instructions are generated by default on
15098 POWER systems, and not generated on PowerPC systems. Do not use
15099 @option{-mstring} on little endian PowerPC systems, since those
15100 instructions do not work when the processor is in little endian mode.
15101 The exceptions are PPC740 and PPC750 which permit the instructions
15102 usage in little endian mode.
15107 @opindex mno-update
15108 Generate code that uses (does not use) the load or store instructions
15109 that update the base register to the address of the calculated memory
15110 location. These instructions are generated by default. If you use
15111 @option{-mno-update}, there is a small window between the time that the
15112 stack pointer is updated and the address of the previous frame is
15113 stored, which means code that walks the stack frame across interrupts or
15114 signals may get corrupted data.
15116 @item -mavoid-indexed-addresses
15117 @itemx -mno-avoid-indexed-addresses
15118 @opindex mavoid-indexed-addresses
15119 @opindex mno-avoid-indexed-addresses
15120 Generate code that tries to avoid (not avoid) the use of indexed load
15121 or store instructions. These instructions can incur a performance
15122 penalty on Power6 processors in certain situations, such as when
15123 stepping through large arrays that cross a 16M boundary. This option
15124 is enabled by default when targetting Power6 and disabled otherwise.
15127 @itemx -mno-fused-madd
15128 @opindex mfused-madd
15129 @opindex mno-fused-madd
15130 Generate code that uses (does not use) the floating point multiply and
15131 accumulate instructions. These instructions are generated by default if
15132 hardware floating is used.
15138 Generate code that uses (does not use) the half-word multiply and
15139 multiply-accumulate instructions on the IBM 405, 440, 464 and 476 processors.
15140 These instructions are generated by default when targetting those
15147 Generate code that uses (does not use) the string-search @samp{dlmzb}
15148 instruction on the IBM 405, 440, 464 and 476 processors. This instruction is
15149 generated by default when targetting those processors.
15151 @item -mno-bit-align
15153 @opindex mno-bit-align
15154 @opindex mbit-align
15155 On System V.4 and embedded PowerPC systems do not (do) force structures
15156 and unions that contain bit-fields to be aligned to the base type of the
15159 For example, by default a structure containing nothing but 8
15160 @code{unsigned} bit-fields of length 1 would be aligned to a 4 byte
15161 boundary and have a size of 4 bytes. By using @option{-mno-bit-align},
15162 the structure would be aligned to a 1 byte boundary and be one byte in
15165 @item -mno-strict-align
15166 @itemx -mstrict-align
15167 @opindex mno-strict-align
15168 @opindex mstrict-align
15169 On System V.4 and embedded PowerPC systems do not (do) assume that
15170 unaligned memory references will be handled by the system.
15172 @item -mrelocatable
15173 @itemx -mno-relocatable
15174 @opindex mrelocatable
15175 @opindex mno-relocatable
15176 On embedded PowerPC systems generate code that allows (does not allow)
15177 the program to be relocated to a different address at runtime. If you
15178 use @option{-mrelocatable} on any module, all objects linked together must
15179 be compiled with @option{-mrelocatable} or @option{-mrelocatable-lib}.
15181 @item -mrelocatable-lib
15182 @itemx -mno-relocatable-lib
15183 @opindex mrelocatable-lib
15184 @opindex mno-relocatable-lib
15185 On embedded PowerPC systems generate code that allows (does not allow)
15186 the program to be relocated to a different address at runtime. Modules
15187 compiled with @option{-mrelocatable-lib} can be linked with either modules
15188 compiled without @option{-mrelocatable} and @option{-mrelocatable-lib} or
15189 with modules compiled with the @option{-mrelocatable} options.
15195 On System V.4 and embedded PowerPC systems do not (do) assume that
15196 register 2 contains a pointer to a global area pointing to the addresses
15197 used in the program.
15200 @itemx -mlittle-endian
15202 @opindex mlittle-endian
15203 On System V.4 and embedded PowerPC systems compile code for the
15204 processor in little endian mode. The @option{-mlittle-endian} option is
15205 the same as @option{-mlittle}.
15208 @itemx -mbig-endian
15210 @opindex mbig-endian
15211 On System V.4 and embedded PowerPC systems compile code for the
15212 processor in big endian mode. The @option{-mbig-endian} option is
15213 the same as @option{-mbig}.
15215 @item -mdynamic-no-pic
15216 @opindex mdynamic-no-pic
15217 On Darwin and Mac OS X systems, compile code so that it is not
15218 relocatable, but that its external references are relocatable. The
15219 resulting code is suitable for applications, but not shared
15222 @item -mprioritize-restricted-insns=@var{priority}
15223 @opindex mprioritize-restricted-insns
15224 This option controls the priority that is assigned to
15225 dispatch-slot restricted instructions during the second scheduling
15226 pass. The argument @var{priority} takes the value @var{0/1/2} to assign
15227 @var{no/highest/second-highest} priority to dispatch slot restricted
15230 @item -msched-costly-dep=@var{dependence_type}
15231 @opindex msched-costly-dep
15232 This option controls which dependences are considered costly
15233 by the target during instruction scheduling. The argument
15234 @var{dependence_type} takes one of the following values:
15235 @var{no}: no dependence is costly,
15236 @var{all}: all dependences are costly,
15237 @var{true_store_to_load}: a true dependence from store to load is costly,
15238 @var{store_to_load}: any dependence from store to load is costly,
15239 @var{number}: any dependence which latency >= @var{number} is costly.
15241 @item -minsert-sched-nops=@var{scheme}
15242 @opindex minsert-sched-nops
15243 This option controls which nop insertion scheme will be used during
15244 the second scheduling pass. The argument @var{scheme} takes one of the
15246 @var{no}: Don't insert nops.
15247 @var{pad}: Pad with nops any dispatch group which has vacant issue slots,
15248 according to the scheduler's grouping.
15249 @var{regroup_exact}: Insert nops to force costly dependent insns into
15250 separate groups. Insert exactly as many nops as needed to force an insn
15251 to a new group, according to the estimated processor grouping.
15252 @var{number}: Insert nops to force costly dependent insns into
15253 separate groups. Insert @var{number} nops to force an insn to a new group.
15256 @opindex mcall-sysv
15257 On System V.4 and embedded PowerPC systems compile code using calling
15258 conventions that adheres to the March 1995 draft of the System V
15259 Application Binary Interface, PowerPC processor supplement. This is the
15260 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
15262 @item -mcall-sysv-eabi
15264 @opindex mcall-sysv-eabi
15265 @opindex mcall-eabi
15266 Specify both @option{-mcall-sysv} and @option{-meabi} options.
15268 @item -mcall-sysv-noeabi
15269 @opindex mcall-sysv-noeabi
15270 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
15272 @item -mcall-aixdesc
15274 On System V.4 and embedded PowerPC systems compile code for the AIX
15278 @opindex mcall-linux
15279 On System V.4 and embedded PowerPC systems compile code for the
15280 Linux-based GNU system.
15284 On System V.4 and embedded PowerPC systems compile code for the
15285 Hurd-based GNU system.
15287 @item -mcall-freebsd
15288 @opindex mcall-freebsd
15289 On System V.4 and embedded PowerPC systems compile code for the
15290 FreeBSD operating system.
15292 @item -mcall-netbsd
15293 @opindex mcall-netbsd
15294 On System V.4 and embedded PowerPC systems compile code for the
15295 NetBSD operating system.
15297 @item -mcall-openbsd
15298 @opindex mcall-netbsd
15299 On System V.4 and embedded PowerPC systems compile code for the
15300 OpenBSD operating system.
15302 @item -maix-struct-return
15303 @opindex maix-struct-return
15304 Return all structures in memory (as specified by the AIX ABI)@.
15306 @item -msvr4-struct-return
15307 @opindex msvr4-struct-return
15308 Return structures smaller than 8 bytes in registers (as specified by the
15311 @item -mabi=@var{abi-type}
15313 Extend the current ABI with a particular extension, or remove such extension.
15314 Valid values are @var{altivec}, @var{no-altivec}, @var{spe},
15315 @var{no-spe}, @var{ibmlongdouble}, @var{ieeelongdouble}@.
15319 Extend the current ABI with SPE ABI extensions. This does not change
15320 the default ABI, instead it adds the SPE ABI extensions to the current
15324 @opindex mabi=no-spe
15325 Disable Booke SPE ABI extensions for the current ABI@.
15327 @item -mabi=ibmlongdouble
15328 @opindex mabi=ibmlongdouble
15329 Change the current ABI to use IBM extended precision long double.
15330 This is a PowerPC 32-bit SYSV ABI option.
15332 @item -mabi=ieeelongdouble
15333 @opindex mabi=ieeelongdouble
15334 Change the current ABI to use IEEE extended precision long double.
15335 This is a PowerPC 32-bit Linux ABI option.
15338 @itemx -mno-prototype
15339 @opindex mprototype
15340 @opindex mno-prototype
15341 On System V.4 and embedded PowerPC systems assume that all calls to
15342 variable argument functions are properly prototyped. Otherwise, the
15343 compiler must insert an instruction before every non prototyped call to
15344 set or clear bit 6 of the condition code register (@var{CR}) to
15345 indicate whether floating point values were passed in the floating point
15346 registers in case the function takes a variable arguments. With
15347 @option{-mprototype}, only calls to prototyped variable argument functions
15348 will set or clear the bit.
15352 On embedded PowerPC systems, assume that the startup module is called
15353 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
15354 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}
15359 On embedded PowerPC systems, assume that the startup module is called
15360 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
15365 On embedded PowerPC systems, assume that the startup module is called
15366 @file{crt0.o} and the standard C libraries are @file{libads.a} and
15369 @item -myellowknife
15370 @opindex myellowknife
15371 On embedded PowerPC systems, assume that the startup module is called
15372 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
15377 On System V.4 and embedded PowerPC systems, specify that you are
15378 compiling for a VxWorks system.
15382 On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags
15383 header to indicate that @samp{eabi} extended relocations are used.
15389 On System V.4 and embedded PowerPC systems do (do not) adhere to the
15390 Embedded Applications Binary Interface (eabi) which is a set of
15391 modifications to the System V.4 specifications. Selecting @option{-meabi}
15392 means that the stack is aligned to an 8 byte boundary, a function
15393 @code{__eabi} is called to from @code{main} to set up the eabi
15394 environment, and the @option{-msdata} option can use both @code{r2} and
15395 @code{r13} to point to two separate small data areas. Selecting
15396 @option{-mno-eabi} means that the stack is aligned to a 16 byte boundary,
15397 do not call an initialization function from @code{main}, and the
15398 @option{-msdata} option will only use @code{r13} to point to a single
15399 small data area. The @option{-meabi} option is on by default if you
15400 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
15403 @opindex msdata=eabi
15404 On System V.4 and embedded PowerPC systems, put small initialized
15405 @code{const} global and static data in the @samp{.sdata2} section, which
15406 is pointed to by register @code{r2}. Put small initialized
15407 non-@code{const} global and static data in the @samp{.sdata} section,
15408 which is pointed to by register @code{r13}. Put small uninitialized
15409 global and static data in the @samp{.sbss} section, which is adjacent to
15410 the @samp{.sdata} section. The @option{-msdata=eabi} option is
15411 incompatible with the @option{-mrelocatable} option. The
15412 @option{-msdata=eabi} option also sets the @option{-memb} option.
15415 @opindex msdata=sysv
15416 On System V.4 and embedded PowerPC systems, put small global and static
15417 data in the @samp{.sdata} section, which is pointed to by register
15418 @code{r13}. Put small uninitialized global and static data in the
15419 @samp{.sbss} section, which is adjacent to the @samp{.sdata} section.
15420 The @option{-msdata=sysv} option is incompatible with the
15421 @option{-mrelocatable} option.
15423 @item -msdata=default
15425 @opindex msdata=default
15427 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
15428 compile code the same as @option{-msdata=eabi}, otherwise compile code the
15429 same as @option{-msdata=sysv}.
15432 @opindex msdata=data
15433 On System V.4 and embedded PowerPC systems, put small global
15434 data in the @samp{.sdata} section. Put small uninitialized global
15435 data in the @samp{.sbss} section. Do not use register @code{r13}
15436 to address small data however. This is the default behavior unless
15437 other @option{-msdata} options are used.
15441 @opindex msdata=none
15443 On embedded PowerPC systems, put all initialized global and static data
15444 in the @samp{.data} section, and all uninitialized data in the
15445 @samp{.bss} section.
15449 @cindex smaller data references (PowerPC)
15450 @cindex .sdata/.sdata2 references (PowerPC)
15451 On embedded PowerPC systems, put global and static items less than or
15452 equal to @var{num} bytes into the small data or bss sections instead of
15453 the normal data or bss section. By default, @var{num} is 8. The
15454 @option{-G @var{num}} switch is also passed to the linker.
15455 All modules should be compiled with the same @option{-G @var{num}} value.
15458 @itemx -mno-regnames
15460 @opindex mno-regnames
15461 On System V.4 and embedded PowerPC systems do (do not) emit register
15462 names in the assembly language output using symbolic forms.
15465 @itemx -mno-longcall
15467 @opindex mno-longcall
15468 By default assume that all calls are far away so that a longer more
15469 expensive calling sequence is required. This is required for calls
15470 further than 32 megabytes (33,554,432 bytes) from the current location.
15471 A short call will be generated if the compiler knows
15472 the call cannot be that far away. This setting can be overridden by
15473 the @code{shortcall} function attribute, or by @code{#pragma
15476 Some linkers are capable of detecting out-of-range calls and generating
15477 glue code on the fly. On these systems, long calls are unnecessary and
15478 generate slower code. As of this writing, the AIX linker can do this,
15479 as can the GNU linker for PowerPC/64. It is planned to add this feature
15480 to the GNU linker for 32-bit PowerPC systems as well.
15482 On Darwin/PPC systems, @code{#pragma longcall} will generate ``jbsr
15483 callee, L42'', plus a ``branch island'' (glue code). The two target
15484 addresses represent the callee and the ``branch island''. The
15485 Darwin/PPC linker will prefer the first address and generate a ``bl
15486 callee'' if the PPC ``bl'' instruction will reach the callee directly;
15487 otherwise, the linker will generate ``bl L42'' to call the ``branch
15488 island''. The ``branch island'' is appended to the body of the
15489 calling function; it computes the full 32-bit address of the callee
15492 On Mach-O (Darwin) systems, this option directs the compiler emit to
15493 the glue for every direct call, and the Darwin linker decides whether
15494 to use or discard it.
15496 In the future, we may cause GCC to ignore all longcall specifications
15497 when the linker is known to generate glue.
15499 @item -mtls-markers
15500 @itemx -mno-tls-markers
15501 @opindex mtls-markers
15502 @opindex mno-tls-markers
15503 Mark (do not mark) calls to @code{__tls_get_addr} with a relocation
15504 specifying the function argument. The relocation allows ld to
15505 reliably associate function call with argument setup instructions for
15506 TLS optimization, which in turn allows gcc to better schedule the
15511 Adds support for multithreading with the @dfn{pthreads} library.
15512 This option sets flags for both the preprocessor and linker.
15517 @subsection RX Options
15520 These command line options are defined for RX targets:
15523 @item -m64bit-doubles
15524 @itemx -m32bit-doubles
15525 @opindex m64bit-doubles
15526 @opindex m32bit-doubles
15527 Make the @code{double} data type be 64-bits (@option{-m64bit-doubles})
15528 or 32-bits (@option{-m32bit-doubles}) in size. The default is
15529 @option{-m32bit-doubles}. @emph{Note} RX floating point hardware only
15530 works on 32-bit values, which is why the default is
15531 @option{-m32bit-doubles}.
15537 Enables (@option{-fpu}) or disables (@option{-nofpu}) the use of RX
15538 floating point hardware. The default is enabled for the @var{RX600}
15539 series and disabled for the @var{RX200} series.
15541 Floating point instructions will only be generated for 32-bit floating
15542 point values however, so if the @option{-m64bit-doubles} option is in
15543 use then the FPU hardware will not be used for doubles.
15545 @emph{Note} If the @option{-fpu} option is enabled then
15546 @option{-funsafe-math-optimizations} is also enabled automatically.
15547 This is because the RX FPU instructions are themselves unsafe.
15549 @item -mcpu=@var{name}
15550 @itemx -patch=@var{name}
15553 Selects the type of RX CPU to be targeted. Currently three types are
15554 supported, the generic @var{RX600} and @var{RX200} series hardware and
15555 the specific @var{RX610} cpu. The default is @var{RX600}.
15557 The only difference between @var{RX600} and @var{RX610} is that the
15558 @var{RX610} does not support the @code{MVTIPL} instruction.
15560 The @var{RX200} series does not have a hardware floating point unit
15561 and so @option{-nofpu} is enabled by default when this type is
15564 @item -mbig-endian-data
15565 @itemx -mlittle-endian-data
15566 @opindex mbig-endian-data
15567 @opindex mlittle-endian-data
15568 Store data (but not code) in the big-endian format. The default is
15569 @option{-mlittle-endian-data}, ie to store data in the little endian
15572 @item -msmall-data-limit=@var{N}
15573 @opindex msmall-data-limit
15574 Specifies the maximum size in bytes of global and static variables
15575 which can be placed into the small data area. Using the small data
15576 area can lead to smaller and faster code, but the size of area is
15577 limited and it is up to the programmer to ensure that the area does
15578 not overflow. Also when the small data area is used one of the RX's
15579 registers (@code{r13}) is reserved for use pointing to this area, so
15580 it is no longer available for use by the compiler. This could result
15581 in slower and/or larger code if variables which once could have been
15582 held in @code{r13} are now pushed onto the stack.
15584 Note, common variables (variables which have not been initialised) and
15585 constants are not placed into the small data area as they are assigned
15586 to other sections in the output executable.
15588 The default value is zero, which disables this feature. Note, this
15589 feature is not enabled by default with higher optimization levels
15590 (@option{-O2} etc) because of the potentially detrimental effects of
15591 reserving register @code{r13}. It is up to the programmer to
15592 experiment and discover whether this feature is of benefit to their
15599 Use the simulator runtime. The default is to use the libgloss board
15602 @item -mas100-syntax
15603 @itemx -mno-as100-syntax
15604 @opindex mas100-syntax
15605 @opindex mno-as100-syntax
15606 When generating assembler output use a syntax that is compatible with
15607 Renesas's AS100 assembler. This syntax can also be handled by the GAS
15608 assembler but it has some restrictions so generating it is not the
15611 @item -mmax-constant-size=@var{N}
15612 @opindex mmax-constant-size
15613 Specifies the maximum size, in bytes, of a constant that can be used as
15614 an operand in a RX instruction. Although the RX instruction set does
15615 allow constants of up to 4 bytes in length to be used in instructions,
15616 a longer value equates to a longer instruction. Thus in some
15617 circumstances it can be beneficial to restrict the size of constants
15618 that are used in instructions. Constants that are too big are instead
15619 placed into a constant pool and referenced via register indirection.
15621 The value @var{N} can be between 0 and 4. A value of 0 (the default)
15622 or 4 means that constants of any size are allowed.
15626 Enable linker relaxation. Linker relaxation is a process whereby the
15627 linker will attempt to reduce the size of a program by finding shorter
15628 versions of various instructions. Disabled by default.
15630 @item -mint-register=@var{N}
15631 @opindex mint-register
15632 Specify the number of registers to reserve for fast interrupt handler
15633 functions. The value @var{N} can be between 0 and 4. A value of 1
15634 means that register @code{r13} will be reserved for the exclusive use
15635 of fast interrupt handlers. A value of 2 reserves @code{r13} and
15636 @code{r12}. A value of 3 reserves @code{r13}, @code{r12} and
15637 @code{r11}, and a value of 4 reserves @code{r13} through @code{r10}.
15638 A value of 0, the default, does not reserve any registers.
15640 @item -msave-acc-in-interrupts
15641 @opindex msave-acc-in-interrupts
15642 Specifies that interrupt handler functions should preserve the
15643 accumulator register. This is only necessary if normal code might use
15644 the accumulator register, for example because it performs 64-bit
15645 multiplications. The default is to ignore the accumulator as this
15646 makes the interrupt handlers faster.
15650 @emph{Note:} The generic GCC command line @option{-ffixed-@var{reg}}
15651 has special significance to the RX port when used with the
15652 @code{interrupt} function attribute. This attribute indicates a
15653 function intended to process fast interrupts. GCC will will ensure
15654 that it only uses the registers @code{r10}, @code{r11}, @code{r12}
15655 and/or @code{r13} and only provided that the normal use of the
15656 corresponding registers have been restricted via the
15657 @option{-ffixed-@var{reg}} or @option{-mint-register} command line
15660 @node S/390 and zSeries Options
15661 @subsection S/390 and zSeries Options
15662 @cindex S/390 and zSeries Options
15664 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
15668 @itemx -msoft-float
15669 @opindex mhard-float
15670 @opindex msoft-float
15671 Use (do not use) the hardware floating-point instructions and registers
15672 for floating-point operations. When @option{-msoft-float} is specified,
15673 functions in @file{libgcc.a} will be used to perform floating-point
15674 operations. When @option{-mhard-float} is specified, the compiler
15675 generates IEEE floating-point instructions. This is the default.
15678 @itemx -mno-hard-dfp
15680 @opindex mno-hard-dfp
15681 Use (do not use) the hardware decimal-floating-point instructions for
15682 decimal-floating-point operations. When @option{-mno-hard-dfp} is
15683 specified, functions in @file{libgcc.a} will be used to perform
15684 decimal-floating-point operations. When @option{-mhard-dfp} is
15685 specified, the compiler generates decimal-floating-point hardware
15686 instructions. This is the default for @option{-march=z9-ec} or higher.
15688 @item -mlong-double-64
15689 @itemx -mlong-double-128
15690 @opindex mlong-double-64
15691 @opindex mlong-double-128
15692 These switches control the size of @code{long double} type. A size
15693 of 64bit makes the @code{long double} type equivalent to the @code{double}
15694 type. This is the default.
15697 @itemx -mno-backchain
15698 @opindex mbackchain
15699 @opindex mno-backchain
15700 Store (do not store) the address of the caller's frame as backchain pointer
15701 into the callee's stack frame.
15702 A backchain may be needed to allow debugging using tools that do not understand
15703 DWARF-2 call frame information.
15704 When @option{-mno-packed-stack} is in effect, the backchain pointer is stored
15705 at the bottom of the stack frame; when @option{-mpacked-stack} is in effect,
15706 the backchain is placed into the topmost word of the 96/160 byte register
15709 In general, code compiled with @option{-mbackchain} is call-compatible with
15710 code compiled with @option{-mmo-backchain}; however, use of the backchain
15711 for debugging purposes usually requires that the whole binary is built with
15712 @option{-mbackchain}. Note that the combination of @option{-mbackchain},
15713 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
15714 to build a linux kernel use @option{-msoft-float}.
15716 The default is to not maintain the backchain.
15718 @item -mpacked-stack
15719 @itemx -mno-packed-stack
15720 @opindex mpacked-stack
15721 @opindex mno-packed-stack
15722 Use (do not use) the packed stack layout. When @option{-mno-packed-stack} is
15723 specified, the compiler uses the all fields of the 96/160 byte register save
15724 area only for their default purpose; unused fields still take up stack space.
15725 When @option{-mpacked-stack} is specified, register save slots are densely
15726 packed at the top of the register save area; unused space is reused for other
15727 purposes, allowing for more efficient use of the available stack space.
15728 However, when @option{-mbackchain} is also in effect, the topmost word of
15729 the save area is always used to store the backchain, and the return address
15730 register is always saved two words below the backchain.
15732 As long as the stack frame backchain is not used, code generated with
15733 @option{-mpacked-stack} is call-compatible with code generated with
15734 @option{-mno-packed-stack}. Note that some non-FSF releases of GCC 2.95 for
15735 S/390 or zSeries generated code that uses the stack frame backchain at run
15736 time, not just for debugging purposes. Such code is not call-compatible
15737 with code compiled with @option{-mpacked-stack}. Also, note that the
15738 combination of @option{-mbackchain},
15739 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
15740 to build a linux kernel use @option{-msoft-float}.
15742 The default is to not use the packed stack layout.
15745 @itemx -mno-small-exec
15746 @opindex msmall-exec
15747 @opindex mno-small-exec
15748 Generate (or do not generate) code using the @code{bras} instruction
15749 to do subroutine calls.
15750 This only works reliably if the total executable size does not
15751 exceed 64k. The default is to use the @code{basr} instruction instead,
15752 which does not have this limitation.
15758 When @option{-m31} is specified, generate code compliant to the
15759 GNU/Linux for S/390 ABI@. When @option{-m64} is specified, generate
15760 code compliant to the GNU/Linux for zSeries ABI@. This allows GCC in
15761 particular to generate 64-bit instructions. For the @samp{s390}
15762 targets, the default is @option{-m31}, while the @samp{s390x}
15763 targets default to @option{-m64}.
15769 When @option{-mzarch} is specified, generate code using the
15770 instructions available on z/Architecture.
15771 When @option{-mesa} is specified, generate code using the
15772 instructions available on ESA/390. Note that @option{-mesa} is
15773 not possible with @option{-m64}.
15774 When generating code compliant to the GNU/Linux for S/390 ABI,
15775 the default is @option{-mesa}. When generating code compliant
15776 to the GNU/Linux for zSeries ABI, the default is @option{-mzarch}.
15782 Generate (or do not generate) code using the @code{mvcle} instruction
15783 to perform block moves. When @option{-mno-mvcle} is specified,
15784 use a @code{mvc} loop instead. This is the default unless optimizing for
15791 Print (or do not print) additional debug information when compiling.
15792 The default is to not print debug information.
15794 @item -march=@var{cpu-type}
15796 Generate code that will run on @var{cpu-type}, which is the name of a system
15797 representing a certain processor type. Possible values for
15798 @var{cpu-type} are @samp{g5}, @samp{g6}, @samp{z900}, @samp{z990},
15799 @samp{z9-109}, @samp{z9-ec} and @samp{z10}.
15800 When generating code using the instructions available on z/Architecture,
15801 the default is @option{-march=z900}. Otherwise, the default is
15802 @option{-march=g5}.
15804 @item -mtune=@var{cpu-type}
15806 Tune to @var{cpu-type} everything applicable about the generated code,
15807 except for the ABI and the set of available instructions.
15808 The list of @var{cpu-type} values is the same as for @option{-march}.
15809 The default is the value used for @option{-march}.
15812 @itemx -mno-tpf-trace
15813 @opindex mtpf-trace
15814 @opindex mno-tpf-trace
15815 Generate code that adds (does not add) in TPF OS specific branches to trace
15816 routines in the operating system. This option is off by default, even
15817 when compiling for the TPF OS@.
15820 @itemx -mno-fused-madd
15821 @opindex mfused-madd
15822 @opindex mno-fused-madd
15823 Generate code that uses (does not use) the floating point multiply and
15824 accumulate instructions. These instructions are generated by default if
15825 hardware floating point is used.
15827 @item -mwarn-framesize=@var{framesize}
15828 @opindex mwarn-framesize
15829 Emit a warning if the current function exceeds the given frame size. Because
15830 this is a compile time check it doesn't need to be a real problem when the program
15831 runs. It is intended to identify functions which most probably cause
15832 a stack overflow. It is useful to be used in an environment with limited stack
15833 size e.g.@: the linux kernel.
15835 @item -mwarn-dynamicstack
15836 @opindex mwarn-dynamicstack
15837 Emit a warning if the function calls alloca or uses dynamically
15838 sized arrays. This is generally a bad idea with a limited stack size.
15840 @item -mstack-guard=@var{stack-guard}
15841 @itemx -mstack-size=@var{stack-size}
15842 @opindex mstack-guard
15843 @opindex mstack-size
15844 If these options are provided the s390 back end emits additional instructions in
15845 the function prologue which trigger a trap if the stack size is @var{stack-guard}
15846 bytes above the @var{stack-size} (remember that the stack on s390 grows downward).
15847 If the @var{stack-guard} option is omitted the smallest power of 2 larger than
15848 the frame size of the compiled function is chosen.
15849 These options are intended to be used to help debugging stack overflow problems.
15850 The additionally emitted code causes only little overhead and hence can also be
15851 used in production like systems without greater performance degradation. The given
15852 values have to be exact powers of 2 and @var{stack-size} has to be greater than
15853 @var{stack-guard} without exceeding 64k.
15854 In order to be efficient the extra code makes the assumption that the stack starts
15855 at an address aligned to the value given by @var{stack-size}.
15856 The @var{stack-guard} option can only be used in conjunction with @var{stack-size}.
15859 @node Score Options
15860 @subsection Score Options
15861 @cindex Score Options
15863 These options are defined for Score implementations:
15868 Compile code for big endian mode. This is the default.
15872 Compile code for little endian mode.
15876 Disable generate bcnz instruction.
15880 Enable generate unaligned load and store instruction.
15884 Enable the use of multiply-accumulate instructions. Disabled by default.
15888 Specify the SCORE5 as the target architecture.
15892 Specify the SCORE5U of the target architecture.
15896 Specify the SCORE7 as the target architecture. This is the default.
15900 Specify the SCORE7D as the target architecture.
15904 @subsection SH Options
15906 These @samp{-m} options are defined for the SH implementations:
15911 Generate code for the SH1.
15915 Generate code for the SH2.
15918 Generate code for the SH2e.
15922 Generate code for the SH2a without FPU, or for a SH2a-FPU in such a way
15923 that the floating-point unit is not used.
15925 @item -m2a-single-only
15926 @opindex m2a-single-only
15927 Generate code for the SH2a-FPU, in such a way that no double-precision
15928 floating point operations are used.
15931 @opindex m2a-single
15932 Generate code for the SH2a-FPU assuming the floating-point unit is in
15933 single-precision mode by default.
15937 Generate code for the SH2a-FPU assuming the floating-point unit is in
15938 double-precision mode by default.
15942 Generate code for the SH3.
15946 Generate code for the SH3e.
15950 Generate code for the SH4 without a floating-point unit.
15952 @item -m4-single-only
15953 @opindex m4-single-only
15954 Generate code for the SH4 with a floating-point unit that only
15955 supports single-precision arithmetic.
15959 Generate code for the SH4 assuming the floating-point unit is in
15960 single-precision mode by default.
15964 Generate code for the SH4.
15968 Generate code for the SH4al-dsp, or for a SH4a in such a way that the
15969 floating-point unit is not used.
15971 @item -m4a-single-only
15972 @opindex m4a-single-only
15973 Generate code for the SH4a, in such a way that no double-precision
15974 floating point operations are used.
15977 @opindex m4a-single
15978 Generate code for the SH4a assuming the floating-point unit is in
15979 single-precision mode by default.
15983 Generate code for the SH4a.
15987 Same as @option{-m4a-nofpu}, except that it implicitly passes
15988 @option{-dsp} to the assembler. GCC doesn't generate any DSP
15989 instructions at the moment.
15993 Compile code for the processor in big endian mode.
15997 Compile code for the processor in little endian mode.
16001 Align doubles at 64-bit boundaries. Note that this changes the calling
16002 conventions, and thus some functions from the standard C library will
16003 not work unless you recompile it first with @option{-mdalign}.
16007 Shorten some address references at link time, when possible; uses the
16008 linker option @option{-relax}.
16012 Use 32-bit offsets in @code{switch} tables. The default is to use
16017 Enable the use of bit manipulation instructions on SH2A.
16021 Enable the use of the instruction @code{fmovd}. Check @option{-mdalign} for
16022 alignment constraints.
16026 Comply with the calling conventions defined by Renesas.
16030 Comply with the calling conventions defined by Renesas.
16034 Comply with the calling conventions defined for GCC before the Renesas
16035 conventions were available. This option is the default for all
16036 targets of the SH toolchain except for @samp{sh-symbianelf}.
16039 @opindex mnomacsave
16040 Mark the @code{MAC} register as call-clobbered, even if
16041 @option{-mhitachi} is given.
16045 Increase IEEE-compliance of floating-point code.
16046 At the moment, this is equivalent to @option{-fno-finite-math-only}.
16047 When generating 16 bit SH opcodes, getting IEEE-conforming results for
16048 comparisons of NANs / infinities incurs extra overhead in every
16049 floating point comparison, therefore the default is set to
16050 @option{-ffinite-math-only}.
16052 @item -minline-ic_invalidate
16053 @opindex minline-ic_invalidate
16054 Inline code to invalidate instruction cache entries after setting up
16055 nested function trampolines.
16056 This option has no effect if -musermode is in effect and the selected
16057 code generation option (e.g. -m4) does not allow the use of the icbi
16059 If the selected code generation option does not allow the use of the icbi
16060 instruction, and -musermode is not in effect, the inlined code will
16061 manipulate the instruction cache address array directly with an associative
16062 write. This not only requires privileged mode, but it will also
16063 fail if the cache line had been mapped via the TLB and has become unmapped.
16067 Dump instruction size and location in the assembly code.
16070 @opindex mpadstruct
16071 This option is deprecated. It pads structures to multiple of 4 bytes,
16072 which is incompatible with the SH ABI@.
16076 Optimize for space instead of speed. Implied by @option{-Os}.
16079 @opindex mprefergot
16080 When generating position-independent code, emit function calls using
16081 the Global Offset Table instead of the Procedure Linkage Table.
16085 Don't generate privileged mode only code; implies -mno-inline-ic_invalidate
16086 if the inlined code would not work in user mode.
16087 This is the default when the target is @code{sh-*-linux*}.
16089 @item -multcost=@var{number}
16090 @opindex multcost=@var{number}
16091 Set the cost to assume for a multiply insn.
16093 @item -mdiv=@var{strategy}
16094 @opindex mdiv=@var{strategy}
16095 Set the division strategy to use for SHmedia code. @var{strategy} must be
16096 one of: call, call2, fp, inv, inv:minlat, inv20u, inv20l, inv:call,
16097 inv:call2, inv:fp .
16098 "fp" performs the operation in floating point. This has a very high latency,
16099 but needs only a few instructions, so it might be a good choice if
16100 your code has enough easily exploitable ILP to allow the compiler to
16101 schedule the floating point instructions together with other instructions.
16102 Division by zero causes a floating point exception.
16103 "inv" uses integer operations to calculate the inverse of the divisor,
16104 and then multiplies the dividend with the inverse. This strategy allows
16105 cse and hoisting of the inverse calculation. Division by zero calculates
16106 an unspecified result, but does not trap.
16107 "inv:minlat" is a variant of "inv" where if no cse / hoisting opportunities
16108 have been found, or if the entire operation has been hoisted to the same
16109 place, the last stages of the inverse calculation are intertwined with the
16110 final multiply to reduce the overall latency, at the expense of using a few
16111 more instructions, and thus offering fewer scheduling opportunities with
16113 "call" calls a library function that usually implements the inv:minlat
16115 This gives high code density for m5-*media-nofpu compilations.
16116 "call2" uses a different entry point of the same library function, where it
16117 assumes that a pointer to a lookup table has already been set up, which
16118 exposes the pointer load to cse / code hoisting optimizations.
16119 "inv:call", "inv:call2" and "inv:fp" all use the "inv" algorithm for initial
16120 code generation, but if the code stays unoptimized, revert to the "call",
16121 "call2", or "fp" strategies, respectively. Note that the
16122 potentially-trapping side effect of division by zero is carried by a
16123 separate instruction, so it is possible that all the integer instructions
16124 are hoisted out, but the marker for the side effect stays where it is.
16125 A recombination to fp operations or a call is not possible in that case.
16126 "inv20u" and "inv20l" are variants of the "inv:minlat" strategy. In the case
16127 that the inverse calculation was nor separated from the multiply, they speed
16128 up division where the dividend fits into 20 bits (plus sign where applicable),
16129 by inserting a test to skip a number of operations in this case; this test
16130 slows down the case of larger dividends. inv20u assumes the case of a such
16131 a small dividend to be unlikely, and inv20l assumes it to be likely.
16133 @item -mdivsi3_libfunc=@var{name}
16134 @opindex mdivsi3_libfunc=@var{name}
16135 Set the name of the library function used for 32 bit signed division to
16136 @var{name}. This only affect the name used in the call and inv:call
16137 division strategies, and the compiler will still expect the same
16138 sets of input/output/clobbered registers as if this option was not present.
16140 @item -mfixed-range=@var{register-range}
16141 @opindex mfixed-range
16142 Generate code treating the given register range as fixed registers.
16143 A fixed register is one that the register allocator can not use. This is
16144 useful when compiling kernel code. A register range is specified as
16145 two registers separated by a dash. Multiple register ranges can be
16146 specified separated by a comma.
16148 @item -madjust-unroll
16149 @opindex madjust-unroll
16150 Throttle unrolling to avoid thrashing target registers.
16151 This option only has an effect if the gcc code base supports the
16152 TARGET_ADJUST_UNROLL_MAX target hook.
16154 @item -mindexed-addressing
16155 @opindex mindexed-addressing
16156 Enable the use of the indexed addressing mode for SHmedia32/SHcompact.
16157 This is only safe if the hardware and/or OS implement 32 bit wrap-around
16158 semantics for the indexed addressing mode. The architecture allows the
16159 implementation of processors with 64 bit MMU, which the OS could use to
16160 get 32 bit addressing, but since no current hardware implementation supports
16161 this or any other way to make the indexed addressing mode safe to use in
16162 the 32 bit ABI, the default is -mno-indexed-addressing.
16164 @item -mgettrcost=@var{number}
16165 @opindex mgettrcost=@var{number}
16166 Set the cost assumed for the gettr instruction to @var{number}.
16167 The default is 2 if @option{-mpt-fixed} is in effect, 100 otherwise.
16171 Assume pt* instructions won't trap. This will generally generate better
16172 scheduled code, but is unsafe on current hardware. The current architecture
16173 definition says that ptabs and ptrel trap when the target anded with 3 is 3.
16174 This has the unintentional effect of making it unsafe to schedule ptabs /
16175 ptrel before a branch, or hoist it out of a loop. For example,
16176 __do_global_ctors, a part of libgcc that runs constructors at program
16177 startup, calls functions in a list which is delimited by @minus{}1. With the
16178 -mpt-fixed option, the ptabs will be done before testing against @minus{}1.
16179 That means that all the constructors will be run a bit quicker, but when
16180 the loop comes to the end of the list, the program crashes because ptabs
16181 loads @minus{}1 into a target register. Since this option is unsafe for any
16182 hardware implementing the current architecture specification, the default
16183 is -mno-pt-fixed. Unless the user specifies a specific cost with
16184 @option{-mgettrcost}, -mno-pt-fixed also implies @option{-mgettrcost=100};
16185 this deters register allocation using target registers for storing
16188 @item -minvalid-symbols
16189 @opindex minvalid-symbols
16190 Assume symbols might be invalid. Ordinary function symbols generated by
16191 the compiler will always be valid to load with movi/shori/ptabs or
16192 movi/shori/ptrel, but with assembler and/or linker tricks it is possible
16193 to generate symbols that will cause ptabs / ptrel to trap.
16194 This option is only meaningful when @option{-mno-pt-fixed} is in effect.
16195 It will then prevent cross-basic-block cse, hoisting and most scheduling
16196 of symbol loads. The default is @option{-mno-invalid-symbols}.
16199 @node SPARC Options
16200 @subsection SPARC Options
16201 @cindex SPARC options
16203 These @samp{-m} options are supported on the SPARC:
16206 @item -mno-app-regs
16208 @opindex mno-app-regs
16210 Specify @option{-mapp-regs} to generate output using the global registers
16211 2 through 4, which the SPARC SVR4 ABI reserves for applications. This
16214 To be fully SVR4 ABI compliant at the cost of some performance loss,
16215 specify @option{-mno-app-regs}. You should compile libraries and system
16216 software with this option.
16219 @itemx -mhard-float
16221 @opindex mhard-float
16222 Generate output containing floating point instructions. This is the
16226 @itemx -msoft-float
16228 @opindex msoft-float
16229 Generate output containing library calls for floating point.
16230 @strong{Warning:} the requisite libraries are not available for all SPARC
16231 targets. Normally the facilities of the machine's usual C compiler are
16232 used, but this cannot be done directly in cross-compilation. You must make
16233 your own arrangements to provide suitable library functions for
16234 cross-compilation. The embedded targets @samp{sparc-*-aout} and
16235 @samp{sparclite-*-*} do provide software floating point support.
16237 @option{-msoft-float} changes the calling convention in the output file;
16238 therefore, it is only useful if you compile @emph{all} of a program with
16239 this option. In particular, you need to compile @file{libgcc.a}, the
16240 library that comes with GCC, with @option{-msoft-float} in order for
16243 @item -mhard-quad-float
16244 @opindex mhard-quad-float
16245 Generate output containing quad-word (long double) floating point
16248 @item -msoft-quad-float
16249 @opindex msoft-quad-float
16250 Generate output containing library calls for quad-word (long double)
16251 floating point instructions. The functions called are those specified
16252 in the SPARC ABI@. This is the default.
16254 As of this writing, there are no SPARC implementations that have hardware
16255 support for the quad-word floating point instructions. They all invoke
16256 a trap handler for one of these instructions, and then the trap handler
16257 emulates the effect of the instruction. Because of the trap handler overhead,
16258 this is much slower than calling the ABI library routines. Thus the
16259 @option{-msoft-quad-float} option is the default.
16261 @item -mno-unaligned-doubles
16262 @itemx -munaligned-doubles
16263 @opindex mno-unaligned-doubles
16264 @opindex munaligned-doubles
16265 Assume that doubles have 8 byte alignment. This is the default.
16267 With @option{-munaligned-doubles}, GCC assumes that doubles have 8 byte
16268 alignment only if they are contained in another type, or if they have an
16269 absolute address. Otherwise, it assumes they have 4 byte alignment.
16270 Specifying this option avoids some rare compatibility problems with code
16271 generated by other compilers. It is not the default because it results
16272 in a performance loss, especially for floating point code.
16274 @item -mno-faster-structs
16275 @itemx -mfaster-structs
16276 @opindex mno-faster-structs
16277 @opindex mfaster-structs
16278 With @option{-mfaster-structs}, the compiler assumes that structures
16279 should have 8 byte alignment. This enables the use of pairs of
16280 @code{ldd} and @code{std} instructions for copies in structure
16281 assignment, in place of twice as many @code{ld} and @code{st} pairs.
16282 However, the use of this changed alignment directly violates the SPARC
16283 ABI@. Thus, it's intended only for use on targets where the developer
16284 acknowledges that their resulting code will not be directly in line with
16285 the rules of the ABI@.
16287 @item -mimpure-text
16288 @opindex mimpure-text
16289 @option{-mimpure-text}, used in addition to @option{-shared}, tells
16290 the compiler to not pass @option{-z text} to the linker when linking a
16291 shared object. Using this option, you can link position-dependent
16292 code into a shared object.
16294 @option{-mimpure-text} suppresses the ``relocations remain against
16295 allocatable but non-writable sections'' linker error message.
16296 However, the necessary relocations will trigger copy-on-write, and the
16297 shared object is not actually shared across processes. Instead of
16298 using @option{-mimpure-text}, you should compile all source code with
16299 @option{-fpic} or @option{-fPIC}.
16301 This option is only available on SunOS and Solaris.
16303 @item -mcpu=@var{cpu_type}
16305 Set the instruction set, register set, and instruction scheduling parameters
16306 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
16307 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite},
16308 @samp{f930}, @samp{f934}, @samp{hypersparc}, @samp{sparclite86x},
16309 @samp{sparclet}, @samp{tsc701}, @samp{v9}, @samp{ultrasparc},
16310 @samp{ultrasparc3}, @samp{niagara} and @samp{niagara2}.
16312 Default instruction scheduling parameters are used for values that select
16313 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
16314 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
16316 Here is a list of each supported architecture and their supported
16321 v8: supersparc, hypersparc
16322 sparclite: f930, f934, sparclite86x
16324 v9: ultrasparc, ultrasparc3, niagara, niagara2
16327 By default (unless configured otherwise), GCC generates code for the V7
16328 variant of the SPARC architecture. With @option{-mcpu=cypress}, the compiler
16329 additionally optimizes it for the Cypress CY7C602 chip, as used in the
16330 SPARCStation/SPARCServer 3xx series. This is also appropriate for the older
16331 SPARCStation 1, 2, IPX etc.
16333 With @option{-mcpu=v8}, GCC generates code for the V8 variant of the SPARC
16334 architecture. The only difference from V7 code is that the compiler emits
16335 the integer multiply and integer divide instructions which exist in SPARC-V8
16336 but not in SPARC-V7. With @option{-mcpu=supersparc}, the compiler additionally
16337 optimizes it for the SuperSPARC chip, as used in the SPARCStation 10, 1000 and
16340 With @option{-mcpu=sparclite}, GCC generates code for the SPARClite variant of
16341 the SPARC architecture. This adds the integer multiply, integer divide step
16342 and scan (@code{ffs}) instructions which exist in SPARClite but not in SPARC-V7.
16343 With @option{-mcpu=f930}, the compiler additionally optimizes it for the
16344 Fujitsu MB86930 chip, which is the original SPARClite, with no FPU@. With
16345 @option{-mcpu=f934}, the compiler additionally optimizes it for the Fujitsu
16346 MB86934 chip, which is the more recent SPARClite with FPU@.
16348 With @option{-mcpu=sparclet}, GCC generates code for the SPARClet variant of
16349 the SPARC architecture. This adds the integer multiply, multiply/accumulate,
16350 integer divide step and scan (@code{ffs}) instructions which exist in SPARClet
16351 but not in SPARC-V7. With @option{-mcpu=tsc701}, the compiler additionally
16352 optimizes it for the TEMIC SPARClet chip.
16354 With @option{-mcpu=v9}, GCC generates code for the V9 variant of the SPARC
16355 architecture. This adds 64-bit integer and floating-point move instructions,
16356 3 additional floating-point condition code registers and conditional move
16357 instructions. With @option{-mcpu=ultrasparc}, the compiler additionally
16358 optimizes it for the Sun UltraSPARC I/II/IIi chips. With
16359 @option{-mcpu=ultrasparc3}, the compiler additionally optimizes it for the
16360 Sun UltraSPARC III/III+/IIIi/IIIi+/IV/IV+ chips. With
16361 @option{-mcpu=niagara}, the compiler additionally optimizes it for
16362 Sun UltraSPARC T1 chips. With @option{-mcpu=niagara2}, the compiler
16363 additionally optimizes it for Sun UltraSPARC T2 chips.
16365 @item -mtune=@var{cpu_type}
16367 Set the instruction scheduling parameters for machine type
16368 @var{cpu_type}, but do not set the instruction set or register set that the
16369 option @option{-mcpu=@var{cpu_type}} would.
16371 The same values for @option{-mcpu=@var{cpu_type}} can be used for
16372 @option{-mtune=@var{cpu_type}}, but the only useful values are those
16373 that select a particular cpu implementation. Those are @samp{cypress},
16374 @samp{supersparc}, @samp{hypersparc}, @samp{f930}, @samp{f934},
16375 @samp{sparclite86x}, @samp{tsc701}, @samp{ultrasparc},
16376 @samp{ultrasparc3}, @samp{niagara}, and @samp{niagara2}.
16381 @opindex mno-v8plus
16382 With @option{-mv8plus}, GCC generates code for the SPARC-V8+ ABI@. The
16383 difference from the V8 ABI is that the global and out registers are
16384 considered 64-bit wide. This is enabled by default on Solaris in 32-bit
16385 mode for all SPARC-V9 processors.
16391 With @option{-mvis}, GCC generates code that takes advantage of the UltraSPARC
16392 Visual Instruction Set extensions. The default is @option{-mno-vis}.
16395 These @samp{-m} options are supported in addition to the above
16396 on SPARC-V9 processors in 64-bit environments:
16399 @item -mlittle-endian
16400 @opindex mlittle-endian
16401 Generate code for a processor running in little-endian mode. It is only
16402 available for a few configurations and most notably not on Solaris and Linux.
16408 Generate code for a 32-bit or 64-bit environment.
16409 The 32-bit environment sets int, long and pointer to 32 bits.
16410 The 64-bit environment sets int to 32 bits and long and pointer
16413 @item -mcmodel=medlow
16414 @opindex mcmodel=medlow
16415 Generate code for the Medium/Low code model: 64-bit addresses, programs
16416 must be linked in the low 32 bits of memory. Programs can be statically
16417 or dynamically linked.
16419 @item -mcmodel=medmid
16420 @opindex mcmodel=medmid
16421 Generate code for the Medium/Middle code model: 64-bit addresses, programs
16422 must be linked in the low 44 bits of memory, the text and data segments must
16423 be less than 2GB in size and the data segment must be located within 2GB of
16426 @item -mcmodel=medany
16427 @opindex mcmodel=medany
16428 Generate code for the Medium/Anywhere code model: 64-bit addresses, programs
16429 may be linked anywhere in memory, the text and data segments must be less
16430 than 2GB in size and the data segment must be located within 2GB of the
16433 @item -mcmodel=embmedany
16434 @opindex mcmodel=embmedany
16435 Generate code for the Medium/Anywhere code model for embedded systems:
16436 64-bit addresses, the text and data segments must be less than 2GB in
16437 size, both starting anywhere in memory (determined at link time). The
16438 global register %g4 points to the base of the data segment. Programs
16439 are statically linked and PIC is not supported.
16442 @itemx -mno-stack-bias
16443 @opindex mstack-bias
16444 @opindex mno-stack-bias
16445 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
16446 frame pointer if present, are offset by @minus{}2047 which must be added back
16447 when making stack frame references. This is the default in 64-bit mode.
16448 Otherwise, assume no such offset is present.
16451 These switches are supported in addition to the above on Solaris:
16456 Add support for multithreading using the Solaris threads library. This
16457 option sets flags for both the preprocessor and linker. This option does
16458 not affect the thread safety of object code produced by the compiler or
16459 that of libraries supplied with it.
16463 Add support for multithreading using the POSIX threads library. This
16464 option sets flags for both the preprocessor and linker. This option does
16465 not affect the thread safety of object code produced by the compiler or
16466 that of libraries supplied with it.
16470 This is a synonym for @option{-pthreads}.
16474 @subsection SPU Options
16475 @cindex SPU options
16477 These @samp{-m} options are supported on the SPU:
16481 @itemx -merror-reloc
16482 @opindex mwarn-reloc
16483 @opindex merror-reloc
16485 The loader for SPU does not handle dynamic relocations. By default, GCC
16486 will give an error when it generates code that requires a dynamic
16487 relocation. @option{-mno-error-reloc} disables the error,
16488 @option{-mwarn-reloc} will generate a warning instead.
16491 @itemx -munsafe-dma
16493 @opindex munsafe-dma
16495 Instructions which initiate or test completion of DMA must not be
16496 reordered with respect to loads and stores of the memory which is being
16497 accessed. Users typically address this problem using the volatile
16498 keyword, but that can lead to inefficient code in places where the
16499 memory is known to not change. Rather than mark the memory as volatile
16500 we treat the DMA instructions as potentially effecting all memory. With
16501 @option{-munsafe-dma} users must use the volatile keyword to protect
16504 @item -mbranch-hints
16505 @opindex mbranch-hints
16507 By default, GCC will generate a branch hint instruction to avoid
16508 pipeline stalls for always taken or probably taken branches. A hint
16509 will not be generated closer than 8 instructions away from its branch.
16510 There is little reason to disable them, except for debugging purposes,
16511 or to make an object a little bit smaller.
16515 @opindex msmall-mem
16516 @opindex mlarge-mem
16518 By default, GCC generates code assuming that addresses are never larger
16519 than 18 bits. With @option{-mlarge-mem} code is generated that assumes
16520 a full 32 bit address.
16525 By default, GCC links against startup code that assumes the SPU-style
16526 main function interface (which has an unconventional parameter list).
16527 With @option{-mstdmain}, GCC will link your program against startup
16528 code that assumes a C99-style interface to @code{main}, including a
16529 local copy of @code{argv} strings.
16531 @item -mfixed-range=@var{register-range}
16532 @opindex mfixed-range
16533 Generate code treating the given register range as fixed registers.
16534 A fixed register is one that the register allocator can not use. This is
16535 useful when compiling kernel code. A register range is specified as
16536 two registers separated by a dash. Multiple register ranges can be
16537 specified separated by a comma.
16543 Compile code assuming that pointers to the PPU address space accessed
16544 via the @code{__ea} named address space qualifier are either 32 or 64
16545 bits wide. The default is 32 bits. As this is an ABI changing option,
16546 all object code in an executable must be compiled with the same setting.
16548 @item -maddress-space-conversion
16549 @itemx -mno-address-space-conversion
16550 @opindex maddress-space-conversion
16551 @opindex mno-address-space-conversion
16552 Allow/disallow treating the @code{__ea} address space as superset
16553 of the generic address space. This enables explicit type casts
16554 between @code{__ea} and generic pointer as well as implicit
16555 conversions of generic pointers to @code{__ea} pointers. The
16556 default is to allow address space pointer conversions.
16558 @item -mcache-size=@var{cache-size}
16559 @opindex mcache-size
16560 This option controls the version of libgcc that the compiler links to an
16561 executable and selects a software-managed cache for accessing variables
16562 in the @code{__ea} address space with a particular cache size. Possible
16563 options for @var{cache-size} are @samp{8}, @samp{16}, @samp{32}, @samp{64}
16564 and @samp{128}. The default cache size is 64KB.
16566 @item -matomic-updates
16567 @itemx -mno-atomic-updates
16568 @opindex matomic-updates
16569 @opindex mno-atomic-updates
16570 This option controls the version of libgcc that the compiler links to an
16571 executable and selects whether atomic updates to the software-managed
16572 cache of PPU-side variables are used. If you use atomic updates, changes
16573 to a PPU variable from SPU code using the @code{__ea} named address space
16574 qualifier will not interfere with changes to other PPU variables residing
16575 in the same cache line from PPU code. If you do not use atomic updates,
16576 such interference may occur; however, writing back cache lines will be
16577 more efficient. The default behavior is to use atomic updates.
16580 @itemx -mdual-nops=@var{n}
16581 @opindex mdual-nops
16582 By default, GCC will insert nops to increase dual issue when it expects
16583 it to increase performance. @var{n} can be a value from 0 to 10. A
16584 smaller @var{n} will insert fewer nops. 10 is the default, 0 is the
16585 same as @option{-mno-dual-nops}. Disabled with @option{-Os}.
16587 @item -mhint-max-nops=@var{n}
16588 @opindex mhint-max-nops
16589 Maximum number of nops to insert for a branch hint. A branch hint must
16590 be at least 8 instructions away from the branch it is effecting. GCC
16591 will insert up to @var{n} nops to enforce this, otherwise it will not
16592 generate the branch hint.
16594 @item -mhint-max-distance=@var{n}
16595 @opindex mhint-max-distance
16596 The encoding of the branch hint instruction limits the hint to be within
16597 256 instructions of the branch it is effecting. By default, GCC makes
16598 sure it is within 125.
16601 @opindex msafe-hints
16602 Work around a hardware bug which causes the SPU to stall indefinitely.
16603 By default, GCC will insert the @code{hbrp} instruction to make sure
16604 this stall won't happen.
16608 @node System V Options
16609 @subsection Options for System V
16611 These additional options are available on System V Release 4 for
16612 compatibility with other compilers on those systems:
16617 Create a shared object.
16618 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
16622 Identify the versions of each tool used by the compiler, in a
16623 @code{.ident} assembler directive in the output.
16627 Refrain from adding @code{.ident} directives to the output file (this is
16630 @item -YP,@var{dirs}
16632 Search the directories @var{dirs}, and no others, for libraries
16633 specified with @option{-l}.
16635 @item -Ym,@var{dir}
16637 Look in the directory @var{dir} to find the M4 preprocessor.
16638 The assembler uses this option.
16639 @c This is supposed to go with a -Yd for predefined M4 macro files, but
16640 @c the generic assembler that comes with Solaris takes just -Ym.
16644 @subsection V850 Options
16645 @cindex V850 Options
16647 These @samp{-m} options are defined for V850 implementations:
16651 @itemx -mno-long-calls
16652 @opindex mlong-calls
16653 @opindex mno-long-calls
16654 Treat all calls as being far away (near). If calls are assumed to be
16655 far away, the compiler will always load the functions address up into a
16656 register, and call indirect through the pointer.
16662 Do not optimize (do optimize) basic blocks that use the same index
16663 pointer 4 or more times to copy pointer into the @code{ep} register, and
16664 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
16665 option is on by default if you optimize.
16667 @item -mno-prolog-function
16668 @itemx -mprolog-function
16669 @opindex mno-prolog-function
16670 @opindex mprolog-function
16671 Do not use (do use) external functions to save and restore registers
16672 at the prologue and epilogue of a function. The external functions
16673 are slower, but use less code space if more than one function saves
16674 the same number of registers. The @option{-mprolog-function} option
16675 is on by default if you optimize.
16679 Try to make the code as small as possible. At present, this just turns
16680 on the @option{-mep} and @option{-mprolog-function} options.
16682 @item -mtda=@var{n}
16684 Put static or global variables whose size is @var{n} bytes or less into
16685 the tiny data area that register @code{ep} points to. The tiny data
16686 area can hold up to 256 bytes in total (128 bytes for byte references).
16688 @item -msda=@var{n}
16690 Put static or global variables whose size is @var{n} bytes or less into
16691 the small data area that register @code{gp} points to. The small data
16692 area can hold up to 64 kilobytes.
16694 @item -mzda=@var{n}
16696 Put static or global variables whose size is @var{n} bytes or less into
16697 the first 32 kilobytes of memory.
16701 Specify that the target processor is the V850.
16704 @opindex mbig-switch
16705 Generate code suitable for big switch tables. Use this option only if
16706 the assembler/linker complain about out of range branches within a switch
16711 This option will cause r2 and r5 to be used in the code generated by
16712 the compiler. This setting is the default.
16714 @item -mno-app-regs
16715 @opindex mno-app-regs
16716 This option will cause r2 and r5 to be treated as fixed registers.
16720 Specify that the target processor is the V850E1. The preprocessor
16721 constants @samp{__v850e1__} and @samp{__v850e__} will be defined if
16722 this option is used.
16726 Specify that the target processor is the V850E@. The preprocessor
16727 constant @samp{__v850e__} will be defined if this option is used.
16729 If neither @option{-mv850} nor @option{-mv850e} nor @option{-mv850e1}
16730 are defined then a default target processor will be chosen and the
16731 relevant @samp{__v850*__} preprocessor constant will be defined.
16733 The preprocessor constants @samp{__v850} and @samp{__v851__} are always
16734 defined, regardless of which processor variant is the target.
16736 @item -mdisable-callt
16737 @opindex mdisable-callt
16738 This option will suppress generation of the CALLT instruction for the
16739 v850e and v850e1 flavors of the v850 architecture. The default is
16740 @option{-mno-disable-callt} which allows the CALLT instruction to be used.
16745 @subsection VAX Options
16746 @cindex VAX options
16748 These @samp{-m} options are defined for the VAX:
16753 Do not output certain jump instructions (@code{aobleq} and so on)
16754 that the Unix assembler for the VAX cannot handle across long
16759 Do output those jump instructions, on the assumption that you
16760 will assemble with the GNU assembler.
16764 Output code for g-format floating point numbers instead of d-format.
16767 @node VxWorks Options
16768 @subsection VxWorks Options
16769 @cindex VxWorks Options
16771 The options in this section are defined for all VxWorks targets.
16772 Options specific to the target hardware are listed with the other
16773 options for that target.
16778 GCC can generate code for both VxWorks kernels and real time processes
16779 (RTPs). This option switches from the former to the latter. It also
16780 defines the preprocessor macro @code{__RTP__}.
16783 @opindex non-static
16784 Link an RTP executable against shared libraries rather than static
16785 libraries. The options @option{-static} and @option{-shared} can
16786 also be used for RTPs (@pxref{Link Options}); @option{-static}
16793 These options are passed down to the linker. They are defined for
16794 compatibility with Diab.
16797 @opindex Xbind-lazy
16798 Enable lazy binding of function calls. This option is equivalent to
16799 @option{-Wl,-z,now} and is defined for compatibility with Diab.
16803 Disable lazy binding of function calls. This option is the default and
16804 is defined for compatibility with Diab.
16807 @node x86-64 Options
16808 @subsection x86-64 Options
16809 @cindex x86-64 options
16811 These are listed under @xref{i386 and x86-64 Options}.
16813 @node i386 and x86-64 Windows Options
16814 @subsection i386 and x86-64 Windows Options
16815 @cindex i386 and x86-64 Windows Options
16817 These additional options are available for Windows targets:
16822 This option is available for Cygwin and MinGW targets. It
16823 specifies that a console application is to be generated, by
16824 instructing the linker to set the PE header subsystem type
16825 required for console applications.
16826 This is the default behavior for Cygwin and MinGW targets.
16830 This option is available for Cygwin targets. It specifies that
16831 the Cygwin internal interface is to be used for predefined
16832 preprocessor macros, C runtime libraries and related linker
16833 paths and options. For Cygwin targets this is the default behavior.
16834 This option is deprecated and will be removed in a future release.
16837 @opindex mno-cygwin
16838 This option is available for Cygwin targets. It specifies that
16839 the MinGW internal interface is to be used instead of Cygwin's, by
16840 setting MinGW-related predefined macros and linker paths and default
16842 This option is deprecated and will be removed in a future release.
16846 This option is available for Cygwin and MinGW targets. It
16847 specifies that a DLL - a dynamic link library - is to be
16848 generated, enabling the selection of the required runtime
16849 startup object and entry point.
16851 @item -mnop-fun-dllimport
16852 @opindex mnop-fun-dllimport
16853 This option is available for Cygwin and MinGW targets. It
16854 specifies that the dllimport attribute should be ignored.
16858 This option is available for MinGW targets. It specifies
16859 that MinGW-specific thread support is to be used.
16863 This option is available for mingw-w64 targets. It specifies
16864 that the UNICODE macro is getting pre-defined and that the
16865 unicode capable runtime startup code is chosen.
16869 This option is available for Cygwin and MinGW targets. It
16870 specifies that the typical Windows pre-defined macros are to
16871 be set in the pre-processor, but does not influence the choice
16872 of runtime library/startup code.
16876 This option is available for Cygwin and MinGW targets. It
16877 specifies that a GUI application is to be generated by
16878 instructing the linker to set the PE header subsystem type
16881 @item -fno-set-stack-executable
16882 @opindex fno-set-stack-executable
16883 This option is available for MinGW targets. It specifies that
16884 the executable flag for stack used by nested functions isn't
16885 set. This is necessary for binaries running in kernel mode of
16886 Windows, as there the user32 API, which is used to set executable
16887 privileges, isn't available.
16889 @item -mpe-aligned-commons
16890 @opindex mpe-aligned-commons
16891 This option is available for Cygwin and MinGW targets. It
16892 specifies that the GNU extension to the PE file format that
16893 permits the correct alignment of COMMON variables should be
16894 used when generating code. It will be enabled by default if
16895 GCC detects that the target assembler found during configuration
16896 supports the feature.
16899 See also under @ref{i386 and x86-64 Options} for standard options.
16901 @node Xstormy16 Options
16902 @subsection Xstormy16 Options
16903 @cindex Xstormy16 Options
16905 These options are defined for Xstormy16:
16910 Choose startup files and linker script suitable for the simulator.
16913 @node Xtensa Options
16914 @subsection Xtensa Options
16915 @cindex Xtensa Options
16917 These options are supported for Xtensa targets:
16921 @itemx -mno-const16
16923 @opindex mno-const16
16924 Enable or disable use of @code{CONST16} instructions for loading
16925 constant values. The @code{CONST16} instruction is currently not a
16926 standard option from Tensilica. When enabled, @code{CONST16}
16927 instructions are always used in place of the standard @code{L32R}
16928 instructions. The use of @code{CONST16} is enabled by default only if
16929 the @code{L32R} instruction is not available.
16932 @itemx -mno-fused-madd
16933 @opindex mfused-madd
16934 @opindex mno-fused-madd
16935 Enable or disable use of fused multiply/add and multiply/subtract
16936 instructions in the floating-point option. This has no effect if the
16937 floating-point option is not also enabled. Disabling fused multiply/add
16938 and multiply/subtract instructions forces the compiler to use separate
16939 instructions for the multiply and add/subtract operations. This may be
16940 desirable in some cases where strict IEEE 754-compliant results are
16941 required: the fused multiply add/subtract instructions do not round the
16942 intermediate result, thereby producing results with @emph{more} bits of
16943 precision than specified by the IEEE standard. Disabling fused multiply
16944 add/subtract instructions also ensures that the program output is not
16945 sensitive to the compiler's ability to combine multiply and add/subtract
16948 @item -mserialize-volatile
16949 @itemx -mno-serialize-volatile
16950 @opindex mserialize-volatile
16951 @opindex mno-serialize-volatile
16952 When this option is enabled, GCC inserts @code{MEMW} instructions before
16953 @code{volatile} memory references to guarantee sequential consistency.
16954 The default is @option{-mserialize-volatile}. Use
16955 @option{-mno-serialize-volatile} to omit the @code{MEMW} instructions.
16957 @item -mtext-section-literals
16958 @itemx -mno-text-section-literals
16959 @opindex mtext-section-literals
16960 @opindex mno-text-section-literals
16961 Control the treatment of literal pools. The default is
16962 @option{-mno-text-section-literals}, which places literals in a separate
16963 section in the output file. This allows the literal pool to be placed
16964 in a data RAM/ROM, and it also allows the linker to combine literal
16965 pools from separate object files to remove redundant literals and
16966 improve code size. With @option{-mtext-section-literals}, the literals
16967 are interspersed in the text section in order to keep them as close as
16968 possible to their references. This may be necessary for large assembly
16971 @item -mtarget-align
16972 @itemx -mno-target-align
16973 @opindex mtarget-align
16974 @opindex mno-target-align
16975 When this option is enabled, GCC instructs the assembler to
16976 automatically align instructions to reduce branch penalties at the
16977 expense of some code density. The assembler attempts to widen density
16978 instructions to align branch targets and the instructions following call
16979 instructions. If there are not enough preceding safe density
16980 instructions to align a target, no widening will be performed. The
16981 default is @option{-mtarget-align}. These options do not affect the
16982 treatment of auto-aligned instructions like @code{LOOP}, which the
16983 assembler will always align, either by widening density instructions or
16984 by inserting no-op instructions.
16987 @itemx -mno-longcalls
16988 @opindex mlongcalls
16989 @opindex mno-longcalls
16990 When this option is enabled, GCC instructs the assembler to translate
16991 direct calls to indirect calls unless it can determine that the target
16992 of a direct call is in the range allowed by the call instruction. This
16993 translation typically occurs for calls to functions in other source
16994 files. Specifically, the assembler translates a direct @code{CALL}
16995 instruction into an @code{L32R} followed by a @code{CALLX} instruction.
16996 The default is @option{-mno-longcalls}. This option should be used in
16997 programs where the call target can potentially be out of range. This
16998 option is implemented in the assembler, not the compiler, so the
16999 assembly code generated by GCC will still show direct call
17000 instructions---look at the disassembled object code to see the actual
17001 instructions. Note that the assembler will use an indirect call for
17002 every cross-file call, not just those that really will be out of range.
17005 @node zSeries Options
17006 @subsection zSeries Options
17007 @cindex zSeries options
17009 These are listed under @xref{S/390 and zSeries Options}.
17011 @node Code Gen Options
17012 @section Options for Code Generation Conventions
17013 @cindex code generation conventions
17014 @cindex options, code generation
17015 @cindex run-time options
17017 These machine-independent options control the interface conventions
17018 used in code generation.
17020 Most of them have both positive and negative forms; the negative form
17021 of @option{-ffoo} would be @option{-fno-foo}. In the table below, only
17022 one of the forms is listed---the one which is not the default. You
17023 can figure out the other form by either removing @samp{no-} or adding
17027 @item -fbounds-check
17028 @opindex fbounds-check
17029 For front-ends that support it, generate additional code to check that
17030 indices used to access arrays are within the declared range. This is
17031 currently only supported by the Java and Fortran front-ends, where
17032 this option defaults to true and false respectively.
17036 This option generates traps for signed overflow on addition, subtraction,
17037 multiplication operations.
17041 This option instructs the compiler to assume that signed arithmetic
17042 overflow of addition, subtraction and multiplication wraps around
17043 using twos-complement representation. This flag enables some optimizations
17044 and disables others. This option is enabled by default for the Java
17045 front-end, as required by the Java language specification.
17048 @opindex fexceptions
17049 Enable exception handling. Generates extra code needed to propagate
17050 exceptions. For some targets, this implies GCC will generate frame
17051 unwind information for all functions, which can produce significant data
17052 size overhead, although it does not affect execution. If you do not
17053 specify this option, GCC will enable it by default for languages like
17054 C++ which normally require exception handling, and disable it for
17055 languages like C that do not normally require it. However, you may need
17056 to enable this option when compiling C code that needs to interoperate
17057 properly with exception handlers written in C++. You may also wish to
17058 disable this option if you are compiling older C++ programs that don't
17059 use exception handling.
17061 @item -fnon-call-exceptions
17062 @opindex fnon-call-exceptions
17063 Generate code that allows trapping instructions to throw exceptions.
17064 Note that this requires platform-specific runtime support that does
17065 not exist everywhere. Moreover, it only allows @emph{trapping}
17066 instructions to throw exceptions, i.e.@: memory references or floating
17067 point instructions. It does not allow exceptions to be thrown from
17068 arbitrary signal handlers such as @code{SIGALRM}.
17070 @item -funwind-tables
17071 @opindex funwind-tables
17072 Similar to @option{-fexceptions}, except that it will just generate any needed
17073 static data, but will not affect the generated code in any other way.
17074 You will normally not enable this option; instead, a language processor
17075 that needs this handling would enable it on your behalf.
17077 @item -fasynchronous-unwind-tables
17078 @opindex fasynchronous-unwind-tables
17079 Generate unwind table in dwarf2 format, if supported by target machine. The
17080 table is exact at each instruction boundary, so it can be used for stack
17081 unwinding from asynchronous events (such as debugger or garbage collector).
17083 @item -fpcc-struct-return
17084 @opindex fpcc-struct-return
17085 Return ``short'' @code{struct} and @code{union} values in memory like
17086 longer ones, rather than in registers. This convention is less
17087 efficient, but it has the advantage of allowing intercallability between
17088 GCC-compiled files and files compiled with other compilers, particularly
17089 the Portable C Compiler (pcc).
17091 The precise convention for returning structures in memory depends
17092 on the target configuration macros.
17094 Short structures and unions are those whose size and alignment match
17095 that of some integer type.
17097 @strong{Warning:} code compiled with the @option{-fpcc-struct-return}
17098 switch is not binary compatible with code compiled with the
17099 @option{-freg-struct-return} switch.
17100 Use it to conform to a non-default application binary interface.
17102 @item -freg-struct-return
17103 @opindex freg-struct-return
17104 Return @code{struct} and @code{union} values in registers when possible.
17105 This is more efficient for small structures than
17106 @option{-fpcc-struct-return}.
17108 If you specify neither @option{-fpcc-struct-return} nor
17109 @option{-freg-struct-return}, GCC defaults to whichever convention is
17110 standard for the target. If there is no standard convention, GCC
17111 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
17112 the principal compiler. In those cases, we can choose the standard, and
17113 we chose the more efficient register return alternative.
17115 @strong{Warning:} code compiled with the @option{-freg-struct-return}
17116 switch is not binary compatible with code compiled with the
17117 @option{-fpcc-struct-return} switch.
17118 Use it to conform to a non-default application binary interface.
17120 @item -fshort-enums
17121 @opindex fshort-enums
17122 Allocate to an @code{enum} type only as many bytes as it needs for the
17123 declared range of possible values. Specifically, the @code{enum} type
17124 will be equivalent to the smallest integer type which has enough room.
17126 @strong{Warning:} the @option{-fshort-enums} switch causes GCC to generate
17127 code that is not binary compatible with code generated without that switch.
17128 Use it to conform to a non-default application binary interface.
17130 @item -fshort-double
17131 @opindex fshort-double
17132 Use the same size for @code{double} as for @code{float}.
17134 @strong{Warning:} the @option{-fshort-double} switch causes GCC to generate
17135 code that is not binary compatible with code generated without that switch.
17136 Use it to conform to a non-default application binary interface.
17138 @item -fshort-wchar
17139 @opindex fshort-wchar
17140 Override the underlying type for @samp{wchar_t} to be @samp{short
17141 unsigned int} instead of the default for the target. This option is
17142 useful for building programs to run under WINE@.
17144 @strong{Warning:} the @option{-fshort-wchar} switch causes GCC to generate
17145 code that is not binary compatible with code generated without that switch.
17146 Use it to conform to a non-default application binary interface.
17149 @opindex fno-common
17150 In C code, controls the placement of uninitialized global variables.
17151 Unix C compilers have traditionally permitted multiple definitions of
17152 such variables in different compilation units by placing the variables
17154 This is the behavior specified by @option{-fcommon}, and is the default
17155 for GCC on most targets.
17156 On the other hand, this behavior is not required by ISO C, and on some
17157 targets may carry a speed or code size penalty on variable references.
17158 The @option{-fno-common} option specifies that the compiler should place
17159 uninitialized global variables in the data section of the object file,
17160 rather than generating them as common blocks.
17161 This has the effect that if the same variable is declared
17162 (without @code{extern}) in two different compilations,
17163 you will get a multiple-definition error when you link them.
17164 In this case, you must compile with @option{-fcommon} instead.
17165 Compiling with @option{-fno-common} is useful on targets for which
17166 it provides better performance, or if you wish to verify that the
17167 program will work on other systems which always treat uninitialized
17168 variable declarations this way.
17172 Ignore the @samp{#ident} directive.
17174 @item -finhibit-size-directive
17175 @opindex finhibit-size-directive
17176 Don't output a @code{.size} assembler directive, or anything else that
17177 would cause trouble if the function is split in the middle, and the
17178 two halves are placed at locations far apart in memory. This option is
17179 used when compiling @file{crtstuff.c}; you should not need to use it
17182 @item -fverbose-asm
17183 @opindex fverbose-asm
17184 Put extra commentary information in the generated assembly code to
17185 make it more readable. This option is generally only of use to those
17186 who actually need to read the generated assembly code (perhaps while
17187 debugging the compiler itself).
17189 @option{-fno-verbose-asm}, the default, causes the
17190 extra information to be omitted and is useful when comparing two assembler
17193 @item -frecord-gcc-switches
17194 @opindex frecord-gcc-switches
17195 This switch causes the command line that was used to invoke the
17196 compiler to be recorded into the object file that is being created.
17197 This switch is only implemented on some targets and the exact format
17198 of the recording is target and binary file format dependent, but it
17199 usually takes the form of a section containing ASCII text. This
17200 switch is related to the @option{-fverbose-asm} switch, but that
17201 switch only records information in the assembler output file as
17202 comments, so it never reaches the object file.
17206 @cindex global offset table
17208 Generate position-independent code (PIC) suitable for use in a shared
17209 library, if supported for the target machine. Such code accesses all
17210 constant addresses through a global offset table (GOT)@. The dynamic
17211 loader resolves the GOT entries when the program starts (the dynamic
17212 loader is not part of GCC; it is part of the operating system). If
17213 the GOT size for the linked executable exceeds a machine-specific
17214 maximum size, you get an error message from the linker indicating that
17215 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
17216 instead. (These maximums are 8k on the SPARC and 32k
17217 on the m68k and RS/6000. The 386 has no such limit.)
17219 Position-independent code requires special support, and therefore works
17220 only on certain machines. For the 386, GCC supports PIC for System V
17221 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
17222 position-independent.
17224 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
17229 If supported for the target machine, emit position-independent code,
17230 suitable for dynamic linking and avoiding any limit on the size of the
17231 global offset table. This option makes a difference on the m68k,
17232 PowerPC and SPARC@.
17234 Position-independent code requires special support, and therefore works
17235 only on certain machines.
17237 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
17244 These options are similar to @option{-fpic} and @option{-fPIC}, but
17245 generated position independent code can be only linked into executables.
17246 Usually these options are used when @option{-pie} GCC option will be
17247 used during linking.
17249 @option{-fpie} and @option{-fPIE} both define the macros
17250 @code{__pie__} and @code{__PIE__}. The macros have the value 1
17251 for @option{-fpie} and 2 for @option{-fPIE}.
17253 @item -fno-jump-tables
17254 @opindex fno-jump-tables
17255 Do not use jump tables for switch statements even where it would be
17256 more efficient than other code generation strategies. This option is
17257 of use in conjunction with @option{-fpic} or @option{-fPIC} for
17258 building code which forms part of a dynamic linker and cannot
17259 reference the address of a jump table. On some targets, jump tables
17260 do not require a GOT and this option is not needed.
17262 @item -ffixed-@var{reg}
17264 Treat the register named @var{reg} as a fixed register; generated code
17265 should never refer to it (except perhaps as a stack pointer, frame
17266 pointer or in some other fixed role).
17268 @var{reg} must be the name of a register. The register names accepted
17269 are machine-specific and are defined in the @code{REGISTER_NAMES}
17270 macro in the machine description macro file.
17272 This flag does not have a negative form, because it specifies a
17275 @item -fcall-used-@var{reg}
17276 @opindex fcall-used
17277 Treat the register named @var{reg} as an allocable register that is
17278 clobbered by function calls. It may be allocated for temporaries or
17279 variables that do not live across a call. Functions compiled this way
17280 will not save and restore the register @var{reg}.
17282 It is an error to used this flag with the frame pointer or stack pointer.
17283 Use of this flag for other registers that have fixed pervasive roles in
17284 the machine's execution model will produce disastrous results.
17286 This flag does not have a negative form, because it specifies a
17289 @item -fcall-saved-@var{reg}
17290 @opindex fcall-saved
17291 Treat the register named @var{reg} as an allocable register saved by
17292 functions. It may be allocated even for temporaries or variables that
17293 live across a call. Functions compiled this way will save and restore
17294 the register @var{reg} if they use it.
17296 It is an error to used this flag with the frame pointer or stack pointer.
17297 Use of this flag for other registers that have fixed pervasive roles in
17298 the machine's execution model will produce disastrous results.
17300 A different sort of disaster will result from the use of this flag for
17301 a register in which function values may be returned.
17303 This flag does not have a negative form, because it specifies a
17306 @item -fpack-struct[=@var{n}]
17307 @opindex fpack-struct
17308 Without a value specified, pack all structure members together without
17309 holes. When a value is specified (which must be a small power of two), pack
17310 structure members according to this value, representing the maximum
17311 alignment (that is, objects with default alignment requirements larger than
17312 this will be output potentially unaligned at the next fitting location.
17314 @strong{Warning:} the @option{-fpack-struct} switch causes GCC to generate
17315 code that is not binary compatible with code generated without that switch.
17316 Additionally, it makes the code suboptimal.
17317 Use it to conform to a non-default application binary interface.
17319 @item -finstrument-functions
17320 @opindex finstrument-functions
17321 Generate instrumentation calls for entry and exit to functions. Just
17322 after function entry and just before function exit, the following
17323 profiling functions will be called with the address of the current
17324 function and its call site. (On some platforms,
17325 @code{__builtin_return_address} does not work beyond the current
17326 function, so the call site information may not be available to the
17327 profiling functions otherwise.)
17330 void __cyg_profile_func_enter (void *this_fn,
17332 void __cyg_profile_func_exit (void *this_fn,
17336 The first argument is the address of the start of the current function,
17337 which may be looked up exactly in the symbol table.
17339 This instrumentation is also done for functions expanded inline in other
17340 functions. The profiling calls will indicate where, conceptually, the
17341 inline function is entered and exited. This means that addressable
17342 versions of such functions must be available. If all your uses of a
17343 function are expanded inline, this may mean an additional expansion of
17344 code size. If you use @samp{extern inline} in your C code, an
17345 addressable version of such functions must be provided. (This is
17346 normally the case anyways, but if you get lucky and the optimizer always
17347 expands the functions inline, you might have gotten away without
17348 providing static copies.)
17350 A function may be given the attribute @code{no_instrument_function}, in
17351 which case this instrumentation will not be done. This can be used, for
17352 example, for the profiling functions listed above, high-priority
17353 interrupt routines, and any functions from which the profiling functions
17354 cannot safely be called (perhaps signal handlers, if the profiling
17355 routines generate output or allocate memory).
17357 @item -finstrument-functions-exclude-file-list=@var{file},@var{file},@dots{}
17358 @opindex finstrument-functions-exclude-file-list
17360 Set the list of functions that are excluded from instrumentation (see
17361 the description of @code{-finstrument-functions}). If the file that
17362 contains a function definition matches with one of @var{file}, then
17363 that function is not instrumented. The match is done on substrings:
17364 if the @var{file} parameter is a substring of the file name, it is
17365 considered to be a match.
17368 @code{-finstrument-functions-exclude-file-list=/bits/stl,include/sys}
17369 will exclude any inline function defined in files whose pathnames
17370 contain @code{/bits/stl} or @code{include/sys}.
17372 If, for some reason, you want to include letter @code{','} in one of
17373 @var{sym}, write @code{'\,'}. For example,
17374 @code{-finstrument-functions-exclude-file-list='\,\,tmp'}
17375 (note the single quote surrounding the option).
17377 @item -finstrument-functions-exclude-function-list=@var{sym},@var{sym},@dots{}
17378 @opindex finstrument-functions-exclude-function-list
17380 This is similar to @code{-finstrument-functions-exclude-file-list},
17381 but this option sets the list of function names to be excluded from
17382 instrumentation. The function name to be matched is its user-visible
17383 name, such as @code{vector<int> blah(const vector<int> &)}, not the
17384 internal mangled name (e.g., @code{_Z4blahRSt6vectorIiSaIiEE}). The
17385 match is done on substrings: if the @var{sym} parameter is a substring
17386 of the function name, it is considered to be a match. For C99 and C++
17387 extended identifiers, the function name must be given in UTF-8, not
17388 using universal character names.
17390 @item -fstack-check
17391 @opindex fstack-check
17392 Generate code to verify that you do not go beyond the boundary of the
17393 stack. You should specify this flag if you are running in an
17394 environment with multiple threads, but only rarely need to specify it in
17395 a single-threaded environment since stack overflow is automatically
17396 detected on nearly all systems if there is only one stack.
17398 Note that this switch does not actually cause checking to be done; the
17399 operating system or the language runtime must do that. The switch causes
17400 generation of code to ensure that they see the stack being extended.
17402 You can additionally specify a string parameter: @code{no} means no
17403 checking, @code{generic} means force the use of old-style checking,
17404 @code{specific} means use the best checking method and is equivalent
17405 to bare @option{-fstack-check}.
17407 Old-style checking is a generic mechanism that requires no specific
17408 target support in the compiler but comes with the following drawbacks:
17412 Modified allocation strategy for large objects: they will always be
17413 allocated dynamically if their size exceeds a fixed threshold.
17416 Fixed limit on the size of the static frame of functions: when it is
17417 topped by a particular function, stack checking is not reliable and
17418 a warning is issued by the compiler.
17421 Inefficiency: because of both the modified allocation strategy and the
17422 generic implementation, the performances of the code are hampered.
17425 Note that old-style stack checking is also the fallback method for
17426 @code{specific} if no target support has been added in the compiler.
17428 @item -fstack-limit-register=@var{reg}
17429 @itemx -fstack-limit-symbol=@var{sym}
17430 @itemx -fno-stack-limit
17431 @opindex fstack-limit-register
17432 @opindex fstack-limit-symbol
17433 @opindex fno-stack-limit
17434 Generate code to ensure that the stack does not grow beyond a certain value,
17435 either the value of a register or the address of a symbol. If the stack
17436 would grow beyond the value, a signal is raised. For most targets,
17437 the signal is raised before the stack overruns the boundary, so
17438 it is possible to catch the signal without taking special precautions.
17440 For instance, if the stack starts at absolute address @samp{0x80000000}
17441 and grows downwards, you can use the flags
17442 @option{-fstack-limit-symbol=__stack_limit} and
17443 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
17444 of 128KB@. Note that this may only work with the GNU linker.
17446 @item -fleading-underscore
17447 @opindex fleading-underscore
17448 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
17449 change the way C symbols are represented in the object file. One use
17450 is to help link with legacy assembly code.
17452 @strong{Warning:} the @option{-fleading-underscore} switch causes GCC to
17453 generate code that is not binary compatible with code generated without that
17454 switch. Use it to conform to a non-default application binary interface.
17455 Not all targets provide complete support for this switch.
17457 @item -ftls-model=@var{model}
17458 @opindex ftls-model
17459 Alter the thread-local storage model to be used (@pxref{Thread-Local}).
17460 The @var{model} argument should be one of @code{global-dynamic},
17461 @code{local-dynamic}, @code{initial-exec} or @code{local-exec}.
17463 The default without @option{-fpic} is @code{initial-exec}; with
17464 @option{-fpic} the default is @code{global-dynamic}.
17466 @item -fvisibility=@var{default|internal|hidden|protected}
17467 @opindex fvisibility
17468 Set the default ELF image symbol visibility to the specified option---all
17469 symbols will be marked with this unless overridden within the code.
17470 Using this feature can very substantially improve linking and
17471 load times of shared object libraries, produce more optimized
17472 code, provide near-perfect API export and prevent symbol clashes.
17473 It is @strong{strongly} recommended that you use this in any shared objects
17476 Despite the nomenclature, @code{default} always means public ie;
17477 available to be linked against from outside the shared object.
17478 @code{protected} and @code{internal} are pretty useless in real-world
17479 usage so the only other commonly used option will be @code{hidden}.
17480 The default if @option{-fvisibility} isn't specified is
17481 @code{default}, i.e., make every
17482 symbol public---this causes the same behavior as previous versions of
17485 A good explanation of the benefits offered by ensuring ELF
17486 symbols have the correct visibility is given by ``How To Write
17487 Shared Libraries'' by Ulrich Drepper (which can be found at
17488 @w{@uref{http://people.redhat.com/~drepper/}})---however a superior
17489 solution made possible by this option to marking things hidden when
17490 the default is public is to make the default hidden and mark things
17491 public. This is the norm with DLL's on Windows and with @option{-fvisibility=hidden}
17492 and @code{__attribute__ ((visibility("default")))} instead of
17493 @code{__declspec(dllexport)} you get almost identical semantics with
17494 identical syntax. This is a great boon to those working with
17495 cross-platform projects.
17497 For those adding visibility support to existing code, you may find
17498 @samp{#pragma GCC visibility} of use. This works by you enclosing
17499 the declarations you wish to set visibility for with (for example)
17500 @samp{#pragma GCC visibility push(hidden)} and
17501 @samp{#pragma GCC visibility pop}.
17502 Bear in mind that symbol visibility should be viewed @strong{as
17503 part of the API interface contract} and thus all new code should
17504 always specify visibility when it is not the default ie; declarations
17505 only for use within the local DSO should @strong{always} be marked explicitly
17506 as hidden as so to avoid PLT indirection overheads---making this
17507 abundantly clear also aids readability and self-documentation of the code.
17508 Note that due to ISO C++ specification requirements, operator new and
17509 operator delete must always be of default visibility.
17511 Be aware that headers from outside your project, in particular system
17512 headers and headers from any other library you use, may not be
17513 expecting to be compiled with visibility other than the default. You
17514 may need to explicitly say @samp{#pragma GCC visibility push(default)}
17515 before including any such headers.
17517 @samp{extern} declarations are not affected by @samp{-fvisibility}, so
17518 a lot of code can be recompiled with @samp{-fvisibility=hidden} with
17519 no modifications. However, this means that calls to @samp{extern}
17520 functions with no explicit visibility will use the PLT, so it is more
17521 effective to use @samp{__attribute ((visibility))} and/or
17522 @samp{#pragma GCC visibility} to tell the compiler which @samp{extern}
17523 declarations should be treated as hidden.
17525 Note that @samp{-fvisibility} does affect C++ vague linkage
17526 entities. This means that, for instance, an exception class that will
17527 be thrown between DSOs must be explicitly marked with default
17528 visibility so that the @samp{type_info} nodes will be unified between
17531 An overview of these techniques, their benefits and how to use them
17532 is at @w{@uref{http://gcc.gnu.org/wiki/Visibility}}.
17538 @node Environment Variables
17539 @section Environment Variables Affecting GCC
17540 @cindex environment variables
17542 @c man begin ENVIRONMENT
17543 This section describes several environment variables that affect how GCC
17544 operates. Some of them work by specifying directories or prefixes to use
17545 when searching for various kinds of files. Some are used to specify other
17546 aspects of the compilation environment.
17548 Note that you can also specify places to search using options such as
17549 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
17550 take precedence over places specified using environment variables, which
17551 in turn take precedence over those specified by the configuration of GCC@.
17552 @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint,
17553 GNU Compiler Collection (GCC) Internals}.
17558 @c @itemx LC_COLLATE
17560 @c @itemx LC_MONETARY
17561 @c @itemx LC_NUMERIC
17566 @c @findex LC_COLLATE
17567 @findex LC_MESSAGES
17568 @c @findex LC_MONETARY
17569 @c @findex LC_NUMERIC
17573 These environment variables control the way that GCC uses
17574 localization information that allow GCC to work with different
17575 national conventions. GCC inspects the locale categories
17576 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
17577 so. These locale categories can be set to any value supported by your
17578 installation. A typical value is @samp{en_GB.UTF-8} for English in the United
17579 Kingdom encoded in UTF-8.
17581 The @env{LC_CTYPE} environment variable specifies character
17582 classification. GCC uses it to determine the character boundaries in
17583 a string; this is needed for some multibyte encodings that contain quote
17584 and escape characters that would otherwise be interpreted as a string
17587 The @env{LC_MESSAGES} environment variable specifies the language to
17588 use in diagnostic messages.
17590 If the @env{LC_ALL} environment variable is set, it overrides the value
17591 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
17592 and @env{LC_MESSAGES} default to the value of the @env{LANG}
17593 environment variable. If none of these variables are set, GCC
17594 defaults to traditional C English behavior.
17598 If @env{TMPDIR} is set, it specifies the directory to use for temporary
17599 files. GCC uses temporary files to hold the output of one stage of
17600 compilation which is to be used as input to the next stage: for example,
17601 the output of the preprocessor, which is the input to the compiler
17604 @item GCC_EXEC_PREFIX
17605 @findex GCC_EXEC_PREFIX
17606 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
17607 names of the subprograms executed by the compiler. No slash is added
17608 when this prefix is combined with the name of a subprogram, but you can
17609 specify a prefix that ends with a slash if you wish.
17611 If @env{GCC_EXEC_PREFIX} is not set, GCC will attempt to figure out
17612 an appropriate prefix to use based on the pathname it was invoked with.
17614 If GCC cannot find the subprogram using the specified prefix, it
17615 tries looking in the usual places for the subprogram.
17617 The default value of @env{GCC_EXEC_PREFIX} is
17618 @file{@var{prefix}/lib/gcc/} where @var{prefix} is the prefix to
17619 the installed compiler. In many cases @var{prefix} is the value
17620 of @code{prefix} when you ran the @file{configure} script.
17622 Other prefixes specified with @option{-B} take precedence over this prefix.
17624 This prefix is also used for finding files such as @file{crt0.o} that are
17627 In addition, the prefix is used in an unusual way in finding the
17628 directories to search for header files. For each of the standard
17629 directories whose name normally begins with @samp{/usr/local/lib/gcc}
17630 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
17631 replacing that beginning with the specified prefix to produce an
17632 alternate directory name. Thus, with @option{-Bfoo/}, GCC will search
17633 @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}.
17634 These alternate directories are searched first; the standard directories
17635 come next. If a standard directory begins with the configured
17636 @var{prefix} then the value of @var{prefix} is replaced by
17637 @env{GCC_EXEC_PREFIX} when looking for header files.
17639 @item COMPILER_PATH
17640 @findex COMPILER_PATH
17641 The value of @env{COMPILER_PATH} is a colon-separated list of
17642 directories, much like @env{PATH}. GCC tries the directories thus
17643 specified when searching for subprograms, if it can't find the
17644 subprograms using @env{GCC_EXEC_PREFIX}.
17647 @findex LIBRARY_PATH
17648 The value of @env{LIBRARY_PATH} is a colon-separated list of
17649 directories, much like @env{PATH}. When configured as a native compiler,
17650 GCC tries the directories thus specified when searching for special
17651 linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking
17652 using GCC also uses these directories when searching for ordinary
17653 libraries for the @option{-l} option (but directories specified with
17654 @option{-L} come first).
17658 @cindex locale definition
17659 This variable is used to pass locale information to the compiler. One way in
17660 which this information is used is to determine the character set to be used
17661 when character literals, string literals and comments are parsed in C and C++.
17662 When the compiler is configured to allow multibyte characters,
17663 the following values for @env{LANG} are recognized:
17667 Recognize JIS characters.
17669 Recognize SJIS characters.
17671 Recognize EUCJP characters.
17674 If @env{LANG} is not defined, or if it has some other value, then the
17675 compiler will use mblen and mbtowc as defined by the default locale to
17676 recognize and translate multibyte characters.
17680 Some additional environments variables affect the behavior of the
17683 @include cppenv.texi
17687 @node Precompiled Headers
17688 @section Using Precompiled Headers
17689 @cindex precompiled headers
17690 @cindex speed of compilation
17692 Often large projects have many header files that are included in every
17693 source file. The time the compiler takes to process these header files
17694 over and over again can account for nearly all of the time required to
17695 build the project. To make builds faster, GCC allows users to
17696 `precompile' a header file; then, if builds can use the precompiled
17697 header file they will be much faster.
17699 To create a precompiled header file, simply compile it as you would any
17700 other file, if necessary using the @option{-x} option to make the driver
17701 treat it as a C or C++ header file. You will probably want to use a
17702 tool like @command{make} to keep the precompiled header up-to-date when
17703 the headers it contains change.
17705 A precompiled header file will be searched for when @code{#include} is
17706 seen in the compilation. As it searches for the included file
17707 (@pxref{Search Path,,Search Path,cpp,The C Preprocessor}) the
17708 compiler looks for a precompiled header in each directory just before it
17709 looks for the include file in that directory. The name searched for is
17710 the name specified in the @code{#include} with @samp{.gch} appended. If
17711 the precompiled header file can't be used, it is ignored.
17713 For instance, if you have @code{#include "all.h"}, and you have
17714 @file{all.h.gch} in the same directory as @file{all.h}, then the
17715 precompiled header file will be used if possible, and the original
17716 header will be used otherwise.
17718 Alternatively, you might decide to put the precompiled header file in a
17719 directory and use @option{-I} to ensure that directory is searched
17720 before (or instead of) the directory containing the original header.
17721 Then, if you want to check that the precompiled header file is always
17722 used, you can put a file of the same name as the original header in this
17723 directory containing an @code{#error} command.
17725 This also works with @option{-include}. So yet another way to use
17726 precompiled headers, good for projects not designed with precompiled
17727 header files in mind, is to simply take most of the header files used by
17728 a project, include them from another header file, precompile that header
17729 file, and @option{-include} the precompiled header. If the header files
17730 have guards against multiple inclusion, they will be skipped because
17731 they've already been included (in the precompiled header).
17733 If you need to precompile the same header file for different
17734 languages, targets, or compiler options, you can instead make a
17735 @emph{directory} named like @file{all.h.gch}, and put each precompiled
17736 header in the directory, perhaps using @option{-o}. It doesn't matter
17737 what you call the files in the directory, every precompiled header in
17738 the directory will be considered. The first precompiled header
17739 encountered in the directory that is valid for this compilation will
17740 be used; they're searched in no particular order.
17742 There are many other possibilities, limited only by your imagination,
17743 good sense, and the constraints of your build system.
17745 A precompiled header file can be used only when these conditions apply:
17749 Only one precompiled header can be used in a particular compilation.
17752 A precompiled header can't be used once the first C token is seen. You
17753 can have preprocessor directives before a precompiled header; you can
17754 even include a precompiled header from inside another header, so long as
17755 there are no C tokens before the @code{#include}.
17758 The precompiled header file must be produced for the same language as
17759 the current compilation. You can't use a C precompiled header for a C++
17763 The precompiled header file must have been produced by the same compiler
17764 binary as the current compilation is using.
17767 Any macros defined before the precompiled header is included must
17768 either be defined in the same way as when the precompiled header was
17769 generated, or must not affect the precompiled header, which usually
17770 means that they don't appear in the precompiled header at all.
17772 The @option{-D} option is one way to define a macro before a
17773 precompiled header is included; using a @code{#define} can also do it.
17774 There are also some options that define macros implicitly, like
17775 @option{-O} and @option{-Wdeprecated}; the same rule applies to macros
17778 @item If debugging information is output when using the precompiled
17779 header, using @option{-g} or similar, the same kind of debugging information
17780 must have been output when building the precompiled header. However,
17781 a precompiled header built using @option{-g} can be used in a compilation
17782 when no debugging information is being output.
17784 @item The same @option{-m} options must generally be used when building
17785 and using the precompiled header. @xref{Submodel Options},
17786 for any cases where this rule is relaxed.
17788 @item Each of the following options must be the same when building and using
17789 the precompiled header:
17791 @gccoptlist{-fexceptions}
17794 Some other command-line options starting with @option{-f},
17795 @option{-p}, or @option{-O} must be defined in the same way as when
17796 the precompiled header was generated. At present, it's not clear
17797 which options are safe to change and which are not; the safest choice
17798 is to use exactly the same options when generating and using the
17799 precompiled header. The following are known to be safe:
17801 @gccoptlist{-fmessage-length= -fpreprocessed -fsched-interblock @gol
17802 -fsched-spec -fsched-spec-load -fsched-spec-load-dangerous @gol
17803 -fsched-verbose=<number> -fschedule-insns -fvisibility= @gol
17808 For all of these except the last, the compiler will automatically
17809 ignore the precompiled header if the conditions aren't met. If you
17810 find an option combination that doesn't work and doesn't cause the
17811 precompiled header to be ignored, please consider filing a bug report,
17814 If you do use differing options when generating and using the
17815 precompiled header, the actual behavior will be a mixture of the
17816 behavior for the options. For instance, if you use @option{-g} to
17817 generate the precompiled header but not when using it, you may or may
17818 not get debugging information for routines in the precompiled header.