1 @c Copyright (C) 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
2 @c 2000, 2001, 2002, 2003, 2004 Free Software Foundation, Inc.
3 @c This is part of the GCC manual.
4 @c For copying conditions, see the file gcc.texi.
8 Copyright @copyright{} 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997,
9 1998, 1999, 2000, 2001, 2002, 2003, 2004 Free Software Foundation, Inc.
11 Permission is granted to copy, distribute and/or modify this document
12 under the terms of the GNU Free Documentation License, Version 1.2 or
13 any later version published by the Free Software Foundation; with the
14 Invariant Sections being ``GNU General Public License'' and ``Funding
15 Free Software'', the Front-Cover texts being (a) (see below), and with
16 the Back-Cover Texts being (b) (see below). A copy of the license is
17 included in the gfdl(7) man page.
19 (a) The FSF's Front-Cover Text is:
23 (b) The FSF's Back-Cover Text is:
25 You have freedom to copy and modify this GNU Manual, like GNU
26 software. Copies published by the Free Software Foundation raise
27 funds for GNU development.
29 @c Set file name and title for the man page.
31 @settitle GNU project C and C++ compiler
33 gcc [@option{-c}|@option{-S}|@option{-E}] [@option{-std=}@var{standard}]
34 [@option{-g}] [@option{-pg}] [@option{-O}@var{level}]
35 [@option{-W}@var{warn}@dots{}] [@option{-pedantic}]
36 [@option{-I}@var{dir}@dots{}] [@option{-L}@var{dir}@dots{}]
37 [@option{-D}@var{macro}[=@var{defn}]@dots{}] [@option{-U}@var{macro}]
38 [@option{-f}@var{option}@dots{}] [@option{-m}@var{machine-option}@dots{}]
39 [@option{-o} @var{outfile}] @var{infile}@dots{}
41 Only the most useful options are listed here; see below for the
42 remainder. @samp{g++} accepts mostly the same options as @samp{gcc}.
45 gpl(7), gfdl(7), fsf-funding(7),
46 cpp(1), gcov(1), as(1), ld(1), gdb(1), adb(1), dbx(1), sdb(1)
47 and the Info entries for @file{gcc}, @file{cpp}, @file{as},
48 @file{ld}, @file{binutils} and @file{gdb}.
51 For instructions on reporting bugs, see
52 @w{@uref{http://gcc.gnu.org/bugs.html}}. Use of the @command{gccbug}
53 script to report bugs is recommended.
56 See the Info entry for @command{gcc}, or
57 @w{@uref{http://gcc.gnu.org/onlinedocs/gcc/Contributors.html}},
58 for contributors to GCC@.
63 @chapter GCC Command Options
64 @cindex GCC command options
65 @cindex command options
66 @cindex options, GCC command
68 @c man begin DESCRIPTION
69 When you invoke GCC, it normally does preprocessing, compilation,
70 assembly and linking. The ``overall options'' allow you to stop this
71 process at an intermediate stage. For example, the @option{-c} option
72 says not to run the linker. Then the output consists of object files
73 output by the assembler.
75 Other options are passed on to one stage of processing. Some options
76 control the preprocessor and others the compiler itself. Yet other
77 options control the assembler and linker; most of these are not
78 documented here, since you rarely need to use any of them.
80 @cindex C compilation options
81 Most of the command line options that you can use with GCC are useful
82 for C programs; when an option is only useful with another language
83 (usually C++), the explanation says so explicitly. If the description
84 for a particular option does not mention a source language, you can use
85 that option with all supported languages.
87 @cindex C++ compilation options
88 @xref{Invoking G++,,Compiling C++ Programs}, for a summary of special
89 options for compiling C++ programs.
91 @cindex grouping options
92 @cindex options, grouping
93 The @command{gcc} program accepts options and file names as operands. Many
94 options have multi-letter names; therefore multiple single-letter options
95 may @emph{not} be grouped: @option{-dr} is very different from @w{@samp{-d
98 @cindex order of options
99 @cindex options, order
100 You can mix options and other arguments. For the most part, the order
101 you use doesn't matter. Order does matter when you use several options
102 of the same kind; for example, if you specify @option{-L} more than once,
103 the directories are searched in the order specified.
105 Many options have long names starting with @samp{-f} or with
106 @samp{-W}---for example, @option{-fforce-mem},
107 @option{-fstrength-reduce}, @option{-Wformat} and so on. Most of
108 these have both positive and negative forms; the negative form of
109 @option{-ffoo} would be @option{-fno-foo}. This manual documents
110 only one of these two forms, whichever one is not the default.
114 @xref{Option Index}, for an index to GCC's options.
117 * Option Summary:: Brief list of all options, without explanations.
118 * Overall Options:: Controlling the kind of output:
119 an executable, object files, assembler files,
120 or preprocessed source.
121 * Invoking G++:: Compiling C++ programs.
122 * C Dialect Options:: Controlling the variant of C language compiled.
123 * C++ Dialect Options:: Variations on C++.
124 * Objective-C Dialect Options:: Variations on Objective-C.
125 * Language Independent Options:: Controlling how diagnostics should be
127 * Warning Options:: How picky should the compiler be?
128 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
129 * Optimize Options:: How much optimization?
130 * Preprocessor Options:: Controlling header files and macro definitions.
131 Also, getting dependency information for Make.
132 * Assembler Options:: Passing options to the assembler.
133 * Link Options:: Specifying libraries and so on.
134 * Directory Options:: Where to find header files and libraries.
135 Where to find the compiler executable files.
136 * Spec Files:: How to pass switches to sub-processes.
137 * Target Options:: Running a cross-compiler, or an old version of GCC.
138 * Submodel Options:: Specifying minor hardware or convention variations,
139 such as 68010 vs 68020.
140 * Code Gen Options:: Specifying conventions for function calls, data layout
142 * Environment Variables:: Env vars that affect GCC.
143 * Precompiled Headers:: Compiling a header once, and using it many times.
144 * Running Protoize:: Automatically adding or removing function prototypes.
150 @section Option Summary
152 Here is a summary of all the options, grouped by type. Explanations are
153 in the following sections.
156 @item Overall Options
157 @xref{Overall Options,,Options Controlling the Kind of Output}.
158 @gccoptlist{-c -S -E -o @var{file} -combine -pipe -pass-exit-codes @gol
159 -x @var{language} -v -### --help --target-help --version}
161 @item C Language Options
162 @xref{C Dialect Options,,Options Controlling C Dialect}.
163 @gccoptlist{-ansi -std=@var{standard} -aux-info @var{filename} @gol
164 -fno-asm -fno-builtin -fno-builtin-@var{function} @gol
165 -fhosted -ffreestanding -fms-extensions @gol
166 -trigraphs -no-integrated-cpp -traditional -traditional-cpp @gol
167 -fallow-single-precision -fcond-mismatch @gol
168 -fsigned-bitfields -fsigned-char @gol
169 -funsigned-bitfields -funsigned-char}
171 @item C++ Language Options
172 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}.
173 @gccoptlist{-fabi-version=@var{n} -fno-access-control -fcheck-new @gol
174 -fconserve-space -fno-const-strings @gol
175 -fno-elide-constructors @gol
176 -fno-enforce-eh-specs @gol
177 -ffor-scope -fno-for-scope -fno-gnu-keywords @gol
178 -fno-implicit-templates @gol
179 -fno-implicit-inline-templates @gol
180 -fno-implement-inlines -fms-extensions @gol
181 -fno-nonansi-builtins -fno-operator-names @gol
182 -fno-optional-diags -fpermissive @gol
183 -frepo -fno-rtti -fstats -ftemplate-depth-@var{n} @gol
184 -fuse-cxa-atexit -fno-weak -nostdinc++ @gol
185 -fno-default-inline -Wabi -Wctor-dtor-privacy @gol
186 -Wnon-virtual-dtor -Wreorder @gol
187 -Weffc++ -Wno-deprecated @gol
188 -Wno-non-template-friend -Wold-style-cast @gol
189 -Woverloaded-virtual -Wno-pmf-conversions @gol
190 -Wsign-promo -Wsynth}
192 @item Objective-C Language Options
193 @xref{Objective-C Dialect Options,,Options Controlling Objective-C Dialect}.
195 -fconstant-string-class=@var{class-name} @gol
196 -fgnu-runtime -fnext-runtime @gol
197 -fno-nil-receivers @gol
198 -fobjc-exceptions @gol
199 -freplace-objc-classes @gol
202 -Wno-protocol -Wselector -Wundeclared-selector}
204 @item Language Independent Options
205 @xref{Language Independent Options,,Options to Control Diagnostic Messages Formatting}.
206 @gccoptlist{-fmessage-length=@var{n} @gol
207 -fdiagnostics-show-location=@r{[}once@r{|}every-line@r{]}}
209 @item Warning Options
210 @xref{Warning Options,,Options to Request or Suppress Warnings}.
211 @gccoptlist{-fsyntax-only -pedantic -pedantic-errors @gol
212 -w -Wextra -Wall -Waggregate-return @gol
213 -Wcast-align -Wcast-qual -Wchar-subscripts -Wcomment @gol
214 -Wconversion -Wno-deprecated-declarations @gol
215 -Wdisabled-optimization -Wno-div-by-zero -Wendif-labels @gol
216 -Werror -Werror-implicit-function-declaration @gol
217 -Wfatal-errors -Wfloat-equal -Wformat -Wformat=2 @gol
218 -Wno-format-extra-args -Wformat-nonliteral @gol
219 -Wformat-security -Wformat-y2k @gol
220 -Wimplicit -Wimplicit-function-declaration -Wimplicit-int @gol
221 -Wimport -Wno-import -Winit-self -Winline @gol
222 -Wno-invalid-offsetof -Winvalid-pch @gol
223 -Wlarger-than-@var{len} -Wlong-long @gol
224 -Wmain -Wmissing-braces @gol
225 -Wmissing-format-attribute -Wmissing-include-dirs @gol
226 -Wmissing-noreturn @gol
227 -Wno-multichar -Wnonnull -Wpacked -Wpadded @gol
228 -Wparentheses -Wpointer-arith -Wredundant-decls @gol
229 -Wreturn-type -Wsequence-point -Wshadow @gol
230 -Wsign-compare -Wstrict-aliasing -Wstrict-aliasing=2 @gol
231 -Wswitch -Wswitch-default -Wswitch-enum @gol
232 -Wsystem-headers -Wtrigraphs -Wundef -Wuninitialized @gol
233 -Wunknown-pragmas -Wunreachable-code @gol
234 -Wunused -Wunused-function -Wunused-label -Wunused-parameter @gol
235 -Wunused-value -Wunused-variable -Wwrite-strings @gol
238 @item C-only Warning Options
239 @gccoptlist{-Wbad-function-cast -Wmissing-declarations @gol
240 -Wmissing-prototypes -Wnested-externs -Wold-style-definition @gol
241 -Wstrict-prototypes -Wtraditional @gol
242 -Wdeclaration-after-statement}
244 @item Debugging Options
245 @xref{Debugging Options,,Options for Debugging Your Program or GCC}.
246 @gccoptlist{-d@var{letters} -dumpspecs -dumpmachine -dumpversion @gol
247 -fdump-unnumbered -fdump-translation-unit@r{[}-@var{n}@r{]} @gol
248 -fdump-class-hierarchy@r{[}-@var{n}@r{]} @gol
250 -fdump-tree-original@r{[}-@var{n}@r{]} @gol
251 -fdump-tree-optimized@r{[}-@var{n}@r{]} @gol
252 -fdump-tree-inlined@r{[}-@var{n}@r{]} @gol
253 -fdump-tree-cfg -fdump-tree-vcg -fdump-tree-alias @gol
255 -fdump-tree-ssa@r{[}-@var{n}@r{]} -fdump-tree-pre@r{[}-@var{n}@r{]} @gol
256 -fdump-tree-ccp@r{[}-@var{n}@r{]} -fdump-tree-dce@r{[}-@var{n}@r{]} @gol
257 -fdump-tree-gimple@r{[}-raw@r{]} -fdump-tree-mudflap@r{[}-@var{n}@r{]} @gol
258 -fdump-tree-dom@r{[}-@var{n}@r{]} @gol
259 -fdump-tree-dse@r{[}-@var{n}@r{]} @gol
260 -fdump-tree-phiopt@r{[}-@var{n}@r{]} @gol
261 -fdump-tree-forwprop@r{[}-@var{n}@r{]} @gol
262 -fdump-tree-copyrename@r{[}-@var{n}@r{]} @gol
264 -fdump-tree-sra@r{[}-@var{n}@r{]} @gol
265 -fdump-tree-fre@r{[}-@var{n}@r{]} @gol
266 -feliminate-dwarf2-dups -feliminate-unused-debug-types @gol
267 -feliminate-unused-debug-symbols -fmem-report -fprofile-arcs -ftree-based-profiling @gol
268 -frandom-seed=@var{string} -fsched-verbose=@var{n} @gol
269 -ftest-coverage -ftime-report -fvar-tracking @gol
270 -g -g@var{level} -gcoff -gdwarf-2 @gol
271 -ggdb -gstabs -gstabs+ -gvms -gxcoff -gxcoff+ @gol
272 -p -pg -print-file-name=@var{library} -print-libgcc-file-name @gol
273 -print-multi-directory -print-multi-lib @gol
274 -print-prog-name=@var{program} -print-search-dirs -Q @gol
277 @item Optimization Options
278 @xref{Optimize Options,,Options that Control Optimization}.
279 @gccoptlist{-falign-functions=@var{n} -falign-jumps=@var{n} @gol
280 -falign-labels=@var{n} -falign-loops=@var{n} @gol
281 -fbounds-check -fmudflap -fmudflapth -fmudflapir @gol
282 -fbranch-probabilities -fprofile-values -fvpt -fbranch-target-load-optimize @gol
283 -fbranch-target-load-optimize2 -fbtr-bb-exclusive @gol
284 -fcaller-saves -fcprop-registers @gol
285 -fcse-follow-jumps -fcse-skip-blocks -fdata-sections @gol
286 -fdelayed-branch -fdelete-null-pointer-checks @gol
287 -fexpensive-optimizations -ffast-math -ffloat-store @gol
288 -fforce-addr -fforce-mem -ffunction-sections @gol
289 -fgcse -fgcse-lm -fgcse-sm -fgcse-las -fgcse-after-reload @gol
290 -floop-optimize -fcrossjumping -fif-conversion -fif-conversion2 @gol
291 -finline-functions -finline-limit=@var{n} -fkeep-inline-functions @gol
292 -fkeep-static-consts -fmerge-constants -fmerge-all-constants @gol
293 -fmodulo-sched -fmove-all-movables -fnew-ra -fno-branch-count-reg @gol
294 -fno-default-inline -fno-defer-pop -floop-optimize2 -fmove-loop-invariants @gol
295 -fno-function-cse -fno-guess-branch-probability @gol
296 -fno-inline -fno-math-errno -fno-peephole -fno-peephole2 @gol
297 -funsafe-math-optimizations -ffinite-math-only @gol
298 -fno-trapping-math -fno-zero-initialized-in-bss @gol
299 -fomit-frame-pointer -foptimize-register-move @gol
300 -foptimize-sibling-calls -fprefetch-loop-arrays @gol
301 -fprofile-generate -fprofile-use @gol
302 -freduce-all-givs -fregmove -frename-registers @gol
303 -freorder-blocks -freorder-blocks-and-partition -freorder-functions @gol
304 -frerun-cse-after-loop -frerun-loop-opt @gol
305 -frounding-math -fschedule-insns -fschedule-insns2 @gol
306 -fno-sched-interblock -fno-sched-spec -fsched-spec-load @gol
307 -fsched-spec-load-dangerous @gol
308 -fsched-stalled-insns=@var{n} -sched-stalled-insns-dep=@var{n} @gol
309 -fsched2-use-superblocks @gol
310 -fsched2-use-traces -fsignaling-nans @gol
311 -fsingle-precision-constant @gol
312 -fstrength-reduce -fstrict-aliasing -ftracer -fthread-jumps @gol
313 -funroll-all-loops -funroll-loops -fpeel-loops @gol
314 -funswitch-loops -fold-unroll-loops -fold-unroll-all-loops @gol
315 -ftree-pre -ftree-ccp -ftree-dce -ftree-loop-optimize @gol
317 -ftree-dominator-opts -ftree-dse -ftree-copyrename @gol
318 -ftree-ch -ftree-sra -ftree-ter -ftree-lrs -ftree-fre @gol
319 --param @var{name}=@var{value}
320 -O -O0 -O1 -O2 -O3 -Os}
322 @item Preprocessor Options
323 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
324 @gccoptlist{-A@var{question}=@var{answer} @gol
325 -A-@var{question}@r{[}=@var{answer}@r{]} @gol
326 -C -dD -dI -dM -dN @gol
327 -D@var{macro}@r{[}=@var{defn}@r{]} -E -H @gol
328 -idirafter @var{dir} @gol
329 -include @var{file} -imacros @var{file} @gol
330 -iprefix @var{file} -iwithprefix @var{dir} @gol
331 -iwithprefixbefore @var{dir} -isystem @var{dir} @gol
332 -M -MM -MF -MG -MP -MQ -MT -nostdinc @gol
333 -P -fworking-directory -remap @gol
334 -trigraphs -undef -U@var{macro} -Wp,@var{option} @gol
335 -Xpreprocessor @var{option}}
337 @item Assembler Option
338 @xref{Assembler Options,,Passing Options to the Assembler}.
339 @gccoptlist{-Wa,@var{option} -Xassembler @var{option}}
342 @xref{Link Options,,Options for Linking}.
343 @gccoptlist{@var{object-file-name} -l@var{library} @gol
344 -nostartfiles -nodefaultlibs -nostdlib -pie @gol
345 -s -static -static-libgcc -shared -shared-libgcc -symbolic @gol
346 -Wl,@var{option} -Xlinker @var{option} @gol
349 @item Directory Options
350 @xref{Directory Options,,Options for Directory Search}.
351 @gccoptlist{-B@var{prefix} -I@var{dir} -iquote@var{dir} -L@var{dir} -specs=@var{file} -I-}
354 @c I wrote this xref this way to avoid overfull hbox. -- rms
355 @xref{Target Options}.
356 @gccoptlist{-V @var{version} -b @var{machine}}
358 @item Machine Dependent Options
359 @xref{Submodel Options,,Hardware Models and Configurations}.
360 @c This list is ordered alphanumerically by subsection name.
361 @c Try and put the significant identifier (CPU or system) first,
362 @c so users have a clue at guessing where the ones they want will be.
365 @gccoptlist{-EB -EL @gol
366 -mmangle-cpu -mcpu=@var{cpu} -mtext=@var{text-section} @gol
367 -mdata=@var{data-section} -mrodata=@var{readonly-data-section}}
370 @gccoptlist{-mapcs-frame -mno-apcs-frame @gol
371 -mabi=@var{name} @gol
372 -mapcs-stack-check -mno-apcs-stack-check @gol
373 -mapcs-float -mno-apcs-float @gol
374 -mapcs-reentrant -mno-apcs-reentrant @gol
375 -msched-prolog -mno-sched-prolog @gol
376 -mlittle-endian -mbig-endian -mwords-little-endian @gol
377 -mfloat-abi=@var{name} -msoft-float -mhard-float -mfpe @gol
378 -mthumb-interwork -mno-thumb-interwork @gol
379 -mcpu=@var{name} -march=@var{name} -mfpu=@var{name} @gol
380 -mstructure-size-boundary=@var{n} @gol
381 -mabort-on-noreturn @gol
382 -mlong-calls -mno-long-calls @gol
383 -msingle-pic-base -mno-single-pic-base @gol
384 -mpic-register=@var{reg} @gol
385 -mnop-fun-dllimport @gol
386 -mcirrus-fix-invalid-insns -mno-cirrus-fix-invalid-insns @gol
387 -mpoke-function-name @gol
389 -mtpcs-frame -mtpcs-leaf-frame @gol
390 -mcaller-super-interworking -mcallee-super-interworking}
393 @gccoptlist{-mmcu=@var{mcu} -msize -minit-stack=@var{n} -mno-interrupts @gol
394 -mcall-prologues -mno-tablejump -mtiny-stack -mint8}
397 @gccoptlist{-mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol
398 -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol
399 -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol
400 -mstack-align -mdata-align -mconst-align @gol
401 -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol
402 -melf -maout -melinux -mlinux -sim -sim2 @gol
403 -mmul-bug-workaround -mno-mul-bug-workaround}
405 @emph{Darwin Options}
406 @gccoptlist{-all_load -allowable_client -arch -arch_errors_fatal @gol
407 -arch_only -bind_at_load -bundle -bundle_loader @gol
408 -client_name -compatibility_version -current_version @gol
409 -dependency-file -dylib_file -dylinker_install_name @gol
410 -dynamic -dynamiclib -exported_symbols_list @gol
411 -filelist -flat_namespace -force_cpusubtype_ALL @gol
412 -force_flat_namespace -headerpad_max_install_names @gol
413 -image_base -init -install_name -keep_private_externs @gol
414 -multi_module -multiply_defined -multiply_defined_unused @gol
415 -noall_load -nofixprebinding -nomultidefs -noprebind -noseglinkedit @gol
416 -pagezero_size -prebind -prebind_all_twolevel_modules @gol
417 -private_bundle -read_only_relocs -sectalign @gol
418 -sectobjectsymbols -whyload -seg1addr @gol
419 -sectcreate -sectobjectsymbols -sectorder @gol
420 -seg_addr_table -seg_addr_table_filename -seglinkedit @gol
421 -segprot -segs_read_only_addr -segs_read_write_addr @gol
422 -single_module -static -sub_library -sub_umbrella @gol
423 -twolevel_namespace -umbrella -undefined @gol
424 -unexported_symbols_list -weak_reference_mismatches @gol
425 -whatsloaded -F -gused -gfull}
427 @emph{DEC Alpha Options}
428 @gccoptlist{-mno-fp-regs -msoft-float -malpha-as -mgas @gol
429 -mieee -mieee-with-inexact -mieee-conformant @gol
430 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol
431 -mtrap-precision=@var{mode} -mbuild-constants @gol
432 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} @gol
433 -mbwx -mmax -mfix -mcix @gol
434 -mfloat-vax -mfloat-ieee @gol
435 -mexplicit-relocs -msmall-data -mlarge-data @gol
436 -msmall-text -mlarge-text @gol
437 -mmemory-latency=@var{time}}
439 @emph{DEC Alpha/VMS Options}
440 @gccoptlist{-mvms-return-codes}
443 @gccoptlist{-mgpr-32 -mgpr-64 -mfpr-32 -mfpr-64 @gol
444 -mhard-float -msoft-float @gol
445 -malloc-cc -mfixed-cc -mdword -mno-dword @gol
446 -mdouble -mno-double @gol
447 -mmedia -mno-media -mmuladd -mno-muladd @gol
448 -mfdpic -minline-plt -mgprel-ro -multilib-library-pic -mlinked-fp @gol
449 -mlibrary-pic -macc-4 -macc-8 @gol
450 -mpack -mno-pack -mno-eflags -mcond-move -mno-cond-move @gol
451 -mscc -mno-scc -mcond-exec -mno-cond-exec @gol
452 -mvliw-branch -mno-vliw-branch @gol
453 -mmulti-cond-exec -mno-multi-cond-exec -mnested-cond-exec @gol
454 -mno-nested-cond-exec -mtomcat-stats @gol
457 @emph{H8/300 Options}
458 @gccoptlist{-mrelax -mh -ms -mn -mint32 -malign-300}
461 @gccoptlist{-march=@var{architecture-type} @gol
462 -mbig-switch -mdisable-fpregs -mdisable-indexing @gol
463 -mfast-indirect-calls -mgas -mgnu-ld -mhp-ld @gol
464 -mfixed-range=@var{register-range} @gol
465 -mjump-in-delay -mlinker-opt -mlong-calls @gol
466 -mlong-load-store -mno-big-switch -mno-disable-fpregs @gol
467 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
468 -mno-jump-in-delay -mno-long-load-store @gol
469 -mno-portable-runtime -mno-soft-float @gol
470 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
471 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
472 -mschedule=@var{cpu-type} -mspace-regs -msio -mwsio @gol
473 -nolibdld -static -threads}
475 @emph{i386 and x86-64 Options}
476 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
477 -mfpmath=@var{unit} @gol
478 -masm=@var{dialect} -mno-fancy-math-387 @gol
479 -mno-fp-ret-in-387 -msoft-float -msvr3-shlib @gol
480 -mno-wide-multiply -mrtd -malign-double @gol
481 -mpreferred-stack-boundary=@var{num} @gol
482 -mmmx -msse -msse2 -msse3 -m3dnow @gol
483 -mthreads -mno-align-stringops -minline-all-stringops @gol
484 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
485 -m96bit-long-double -mregparm=@var{num} -momit-leaf-frame-pointer @gol
486 -mno-red-zone -mno-tls-direct-seg-refs @gol
487 -mcmodel=@var{code-model} @gol
491 @gccoptlist{-mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
492 -mvolatile-asm-stop -mb-step -mregister-names -mno-sdata @gol
493 -mconstant-gp -mauto-pic -minline-float-divide-min-latency @gol
494 -minline-float-divide-max-throughput @gol
495 -minline-int-divide-min-latency @gol
496 -minline-int-divide-max-throughput -mno-dwarf2-asm @gol
497 -mfixed-range=@var{register-range}}
499 @emph{M32R/D Options}
500 @gccoptlist{-m32r2 -m32rx -m32r @gol
502 -malign-loops -mno-align-loops @gol
503 -missue-rate=@var{number} @gol
504 -mbranch-cost=@var{number} @gol
505 -mmodel=@var{code-size-model-type} @gol
506 -msdata=@var{sdata-type} @gol
507 -mno-flush-func -mflush-func=@var{name} @gol
508 -mno-flush-trap -mflush-trap=@var{number} @gol
511 @emph{M680x0 Options}
512 @gccoptlist{-m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
513 -m68060 -mcpu32 -m5200 -m68881 -mbitfield -mc68000 -mc68020 @gol
514 -mnobitfield -mrtd -mshort -msoft-float -mpcrel @gol
515 -malign-int -mstrict-align -msep-data -mno-sep-data @gol
516 -mshared-library-id=n -mid-shared-library -mno-id-shared-library}
518 @emph{M68hc1x Options}
519 @gccoptlist{-m6811 -m6812 -m68hc11 -m68hc12 -m68hcs12 @gol
520 -mauto-incdec -minmax -mlong-calls -mshort @gol
521 -msoft-reg-count=@var{count}}
524 @gccoptlist{-mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
525 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
526 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
527 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
528 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
531 @gccoptlist{-EL -EB -march=@var{arch} -mtune=@var{arch} @gol
532 -mips1 -mips2 -mips3 -mips4 -mips32 -mips32r2 -mips64 @gol
533 -mips16 -mno-mips16 -mabi=@var{abi} -mabicalls -mno-abicalls @gol
534 -mxgot -mno-xgot -mgp32 -mgp64 -mfp32 -mfp64 @gol
535 -mhard-float -msoft-float -msingle-float -mdouble-float @gol
536 -mint64 -mlong64 -mlong32 @gol
537 -G@var{num} -membedded-data -mno-embedded-data @gol
538 -muninit-const-in-rodata -mno-uninit-const-in-rodata @gol
539 -msplit-addresses -mno-split-addresses @gol
540 -mexplicit-relocs -mno-explicit-relocs @gol
541 -mcheck-zero-division -mno-check-zero-division @gol
542 -mmemcpy -mno-memcpy -mlong-calls -mno-long-calls @gol
543 -mmad -mno-mad -mfused-madd -mno-fused-madd -nocpp @gol
544 -mfix-r4000 -mno-fix-r4000 -mfix-r4400 -mno-fix-r4400 @gol
545 -mfix-vr4120 -mno-fix-vr4120 -mfix-sb1 -mno-fix-sb1 @gol
546 -mflush-func=@var{func} -mno-flush-func @gol
547 -mbranch-likely -mno-branch-likely @gol
548 -mfp-exceptions -mno-fp-exceptions @gol
549 -mvr4130-align -mno-vr4130-align}
552 @gccoptlist{-mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
553 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
554 -melf -mbranch-predict -mno-branch-predict -mbase-addresses @gol
555 -mno-base-addresses -msingle-exit -mno-single-exit}
557 @emph{MN10300 Options}
558 @gccoptlist{-mmult-bug -mno-mult-bug @gol
559 -mam33 -mno-am33 @gol
560 -mam33-2 -mno-am33-2 @gol
564 @gccoptlist{-m32032 -m32332 -m32532 -m32081 -m32381 @gol
565 -mmult-add -mnomult-add -msoft-float -mrtd -mnortd @gol
566 -mregparam -mnoregparam -msb -mnosb @gol
567 -mbitfield -mnobitfield -mhimem -mnohimem}
569 @emph{PDP-11 Options}
570 @gccoptlist{-mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol
571 -mbcopy -mbcopy-builtin -mint32 -mno-int16 @gol
572 -mint16 -mno-int32 -mfloat32 -mno-float64 @gol
573 -mfloat64 -mno-float32 -mabshi -mno-abshi @gol
574 -mbranch-expensive -mbranch-cheap @gol
575 -msplit -mno-split -munix-asm -mdec-asm}
577 @emph{PowerPC Options}
578 See RS/6000 and PowerPC Options.
580 @emph{RS/6000 and PowerPC Options}
581 @gccoptlist{-mcpu=@var{cpu-type} @gol
582 -mtune=@var{cpu-type} @gol
583 -mpower -mno-power -mpower2 -mno-power2 @gol
584 -mpowerpc -mpowerpc64 -mno-powerpc @gol
585 -maltivec -mno-altivec @gol
586 -mpowerpc-gpopt -mno-powerpc-gpopt @gol
587 -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol
588 -mnew-mnemonics -mold-mnemonics @gol
589 -mfull-toc -mminimal-toc -mno-fp-in-toc -mno-sum-in-toc @gol
590 -m64 -m32 -mxl-call -mno-xl-call -mpe @gol
591 -malign-power -malign-natural @gol
592 -msoft-float -mhard-float -mmultiple -mno-multiple @gol
593 -mstring -mno-string -mupdate -mno-update @gol
594 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol
595 -mstrict-align -mno-strict-align -mrelocatable @gol
596 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol
597 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol
598 -mdynamic-no-pic @gol
599 -mprioritize-restricted-insns=@var{priority} @gol
600 -msched-costly-dep=@var{dependence_type} @gol
601 -minsert-sched-nops=@var{scheme} @gol
602 -mcall-sysv -mcall-netbsd @gol
603 -maix-struct-return -msvr4-struct-return @gol
604 -mabi=altivec -mabi=no-altivec @gol
605 -mabi=spe -mabi=no-spe @gol
606 -misel=yes -misel=no @gol
607 -mspe=yes -mspe=no @gol
608 -mfloat-gprs=yes -mfloat-gprs=no @gol
609 -mprototype -mno-prototype @gol
610 -msim -mmvme -mads -myellowknife -memb -msdata @gol
611 -msdata=@var{opt} -mvxworks -mwindiss -G @var{num} -pthread}
613 @emph{S/390 and zSeries Options}
614 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
615 -mhard-float -msoft-float -mbackchain -mno-backchain @gol
616 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
617 -m64 -m31 -mdebug -mno-debug -mesa -mzarch @gol
618 -mtpf-trace -mno-tpf-trace -mfused-madd -mno-fused-madd}
621 @gccoptlist{-m1 -m2 -m2e -m3 -m3e @gol
622 -m4-nofpu -m4-single-only -m4-single -m4 @gol
623 -m5-64media -m5-64media-nofpu @gol
624 -m5-32media -m5-32media-nofpu @gol
625 -m5-compact -m5-compact-nofpu @gol
626 -mb -ml -mdalign -mrelax @gol
627 -mbigtable -mfmovd -mhitachi -mnomacsave @gol
628 -mieee -misize -mpadstruct -mspace @gol
629 -mprefergot -musermode}
632 @gccoptlist{-mcpu=@var{cpu-type} @gol
633 -mtune=@var{cpu-type} @gol
634 -mcmodel=@var{code-model} @gol
635 -m32 -m64 -mapp-regs -mno-app-regs @gol
636 -mfaster-structs -mno-faster-structs @gol
637 -mfpu -mno-fpu -mhard-float -msoft-float @gol
638 -mhard-quad-float -msoft-quad-float @gol
639 -mimpure-text -mno-impure-text -mlittle-endian @gol
640 -mstack-bias -mno-stack-bias @gol
641 -munaligned-doubles -mno-unaligned-doubles @gol
642 -mv8plus -mno-v8plus -mvis -mno-vis
645 @emph{System V Options}
646 @gccoptlist{-Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
648 @emph{TMS320C3x/C4x Options}
649 @gccoptlist{-mcpu=@var{cpu} -mbig -msmall -mregparm -mmemparm @gol
650 -mfast-fix -mmpyi -mbk -mti -mdp-isr-reload @gol
651 -mrpts=@var{count} -mrptb -mdb -mloop-unsigned @gol
652 -mparallel-insns -mparallel-mpy -mpreserve-float}
655 @gccoptlist{-mlong-calls -mno-long-calls -mep -mno-ep @gol
656 -mprolog-function -mno-prolog-function -mspace @gol
657 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
658 -mapp-regs -mno-app-regs @gol
659 -mdisable-callt -mno-disable-callt @gol
665 @gccoptlist{-mg -mgnu -munix}
667 @emph{x86-64 Options}
668 See i386 and x86-64 Options.
670 @emph{Xstormy16 Options}
673 @emph{Xtensa Options}
674 @gccoptlist{-mconst16 -mno-const16 @gol
675 -mfused-madd -mno-fused-madd @gol
676 -mtext-section-literals -mno-text-section-literals @gol
677 -mtarget-align -mno-target-align @gol
678 -mlongcalls -mno-longcalls}
680 @emph{zSeries Options}
681 See S/390 and zSeries Options.
683 @item Code Generation Options
684 @xref{Code Gen Options,,Options for Code Generation Conventions}.
685 @gccoptlist{-fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
686 -ffixed-@var{reg} -fexceptions @gol
687 -fnon-call-exceptions -funwind-tables @gol
688 -fasynchronous-unwind-tables @gol
689 -finhibit-size-directive -finstrument-functions @gol
690 -fno-common -fno-ident @gol
691 -fpcc-struct-return -fpic -fPIC -fpie -fPIE @gol
692 -freg-struct-return -fshared-data -fshort-enums @gol
693 -fshort-double -fshort-wchar @gol
694 -fverbose-asm -fpack-struct -fstack-check @gol
695 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
696 -fargument-alias -fargument-noalias @gol
697 -fargument-noalias-global -fleading-underscore @gol
698 -ftls-model=@var{model} @gol
699 -ftrapv -fwrapv -fbounds-check}
703 * Overall Options:: Controlling the kind of output:
704 an executable, object files, assembler files,
705 or preprocessed source.
706 * C Dialect Options:: Controlling the variant of C language compiled.
707 * C++ Dialect Options:: Variations on C++.
708 * Objective-C Dialect Options:: Variations on Objective-C.
709 * Language Independent Options:: Controlling how diagnostics should be
711 * Warning Options:: How picky should the compiler be?
712 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
713 * Optimize Options:: How much optimization?
714 * Preprocessor Options:: Controlling header files and macro definitions.
715 Also, getting dependency information for Make.
716 * Assembler Options:: Passing options to the assembler.
717 * Link Options:: Specifying libraries and so on.
718 * Directory Options:: Where to find header files and libraries.
719 Where to find the compiler executable files.
720 * Spec Files:: How to pass switches to sub-processes.
721 * Target Options:: Running a cross-compiler, or an old version of GCC.
724 @node Overall Options
725 @section Options Controlling the Kind of Output
727 Compilation can involve up to four stages: preprocessing, compilation
728 proper, assembly and linking, always in that order. GCC is capable of
729 preprocessing and compiling several files either into several
730 assembler input files, or into one assembler input file; then each
731 assembler input file produces an object file, and linking combines all
732 the object files (those newly compiled, and those specified as input)
733 into an executable file.
735 @cindex file name suffix
736 For any given input file, the file name suffix determines what kind of
741 C source code which must be preprocessed.
744 C source code which should not be preprocessed.
747 C++ source code which should not be preprocessed.
750 Objective-C source code. Note that you must link with the library
751 @file{libobjc.a} to make an Objective-C program work.
754 Objective-C source code which should not be preprocessed.
757 C or C++ header file to be turned into a precompiled header.
761 @itemx @var{file}.cxx
762 @itemx @var{file}.cpp
763 @itemx @var{file}.CPP
764 @itemx @var{file}.c++
766 C++ source code which must be preprocessed. Note that in @samp{.cxx},
767 the last two letters must both be literally @samp{x}. Likewise,
768 @samp{.C} refers to a literal capital C@.
772 C++ header file to be turned into a precompiled header.
775 @itemx @var{file}.for
776 @itemx @var{file}.FOR
777 Fortran source code which should not be preprocessed.
780 @itemx @var{file}.fpp
781 @itemx @var{file}.FPP
782 Fortran source code which must be preprocessed (with the traditional
786 Fortran source code which must be preprocessed with a RATFOR
787 preprocessor (not included with GCC)@.
790 @itemx @var{file}.f95
791 Fortran 90/95 source code which should not be preprocessed.
793 @c FIXME: Descriptions of Java file types.
800 Ada source code file which contains a library unit declaration (a
801 declaration of a package, subprogram, or generic, or a generic
802 instantiation), or a library unit renaming declaration (a package,
803 generic, or subprogram renaming declaration). Such files are also
806 @itemx @var{file}.adb
807 Ada source code file containing a library unit body (a subprogram or
808 package body). Such files are also called @dfn{bodies}.
810 @c GCC also knows about some suffixes for languages not yet included:
819 Assembler code which must be preprocessed.
822 An object file to be fed straight into linking.
823 Any file name with no recognized suffix is treated this way.
827 You can specify the input language explicitly with the @option{-x} option:
830 @item -x @var{language}
831 Specify explicitly the @var{language} for the following input files
832 (rather than letting the compiler choose a default based on the file
833 name suffix). This option applies to all following input files until
834 the next @option{-x} option. Possible values for @var{language} are:
836 c c-header cpp-output
837 c++ c++-header c++-cpp-output
838 objective-c objective-c-header objc-cpp-output
839 assembler assembler-with-cpp
841 f77 f77-cpp-input ratfor
848 Turn off any specification of a language, so that subsequent files are
849 handled according to their file name suffixes (as they are if @option{-x}
850 has not been used at all).
852 @item -pass-exit-codes
853 @opindex pass-exit-codes
854 Normally the @command{gcc} program will exit with the code of 1 if any
855 phase of the compiler returns a non-success return code. If you specify
856 @option{-pass-exit-codes}, the @command{gcc} program will instead return with
857 numerically highest error produced by any phase that returned an error
861 If you only want some of the stages of compilation, you can use
862 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
863 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
864 @command{gcc} is to stop. Note that some combinations (for example,
865 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
870 Compile or assemble the source files, but do not link. The linking
871 stage simply is not done. The ultimate output is in the form of an
872 object file for each source file.
874 By default, the object file name for a source file is made by replacing
875 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
877 Unrecognized input files, not requiring compilation or assembly, are
882 Stop after the stage of compilation proper; do not assemble. The output
883 is in the form of an assembler code file for each non-assembler input
886 By default, the assembler file name for a source file is made by
887 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
889 Input files that don't require compilation are ignored.
893 Stop after the preprocessing stage; do not run the compiler proper. The
894 output is in the form of preprocessed source code, which is sent to the
897 Input files which don't require preprocessing are ignored.
899 @cindex output file option
902 Place output in file @var{file}. This applies regardless to whatever
903 sort of output is being produced, whether it be an executable file,
904 an object file, an assembler file or preprocessed C code.
906 If @option{-o} is not specified, the default is to put an executable
907 file in @file{a.out}, the object file for
908 @file{@var{source}.@var{suffix}} in @file{@var{source}.o}, its
909 assembler file in @file{@var{source}.s}, a precompiled header file in
910 @file{@var{source}.@var{suffix}.gch}, and all preprocessed C source on
915 Print (on standard error output) the commands executed to run the stages
916 of compilation. Also print the version number of the compiler driver
917 program and of the preprocessor and the compiler proper.
921 Like @option{-v} except the commands are not executed and all command
922 arguments are quoted. This is useful for shell scripts to capture the
923 driver-generated command lines.
927 Use pipes rather than temporary files for communication between the
928 various stages of compilation. This fails to work on some systems where
929 the assembler is unable to read from a pipe; but the GNU assembler has
934 If you are compiling multiple source files, this option tells the driver
935 to pass all the source files to the compiler at once (for those
936 languages for which the compiler can handle this). This will allow
937 intermodule analysis (IMA) to be performed by the compiler. Currently the only
938 language for which this is supported is C. If you pass source files for
939 multiple languages to the driver, using this option, the driver will invoke
940 the compiler(s) that support IMA once each, passing each compiler all the
941 source files appropriate for it. For those languages that do not support
942 IMA this option will be ignored, and the compiler will be invoked once for
943 each source file in that language. If you use this option in conjunction
944 with -save-temps, the compiler will generate multiple pre-processed files
945 (one for each source file), but only one (combined) .o or .s file.
949 Print (on the standard output) a description of the command line options
950 understood by @command{gcc}. If the @option{-v} option is also specified
951 then @option{--help} will also be passed on to the various processes
952 invoked by @command{gcc}, so that they can display the command line options
953 they accept. If the @option{-Wextra} option is also specified then command
954 line options which have no documentation associated with them will also
959 Print (on the standard output) a description of target specific command
960 line options for each tool.
964 Display the version number and copyrights of the invoked GCC.
968 @section Compiling C++ Programs
970 @cindex suffixes for C++ source
971 @cindex C++ source file suffixes
972 C++ source files conventionally use one of the suffixes @samp{.C},
973 @samp{.cc}, @samp{.cpp}, @samp{.CPP}, @samp{.c++}, @samp{.cp}, or
974 @samp{.cxx}; C++ header files often use @samp{.hh} or @samp{.H}; and
975 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
976 files with these names and compiles them as C++ programs even if you
977 call the compiler the same way as for compiling C programs (usually
978 with the name @command{gcc}).
982 However, C++ programs often require class libraries as well as a
983 compiler that understands the C++ language---and under some
984 circumstances, you might want to compile programs or header files from
985 standard input, or otherwise without a suffix that flags them as C++
986 programs. You might also like to precompile a C header file with a
987 @samp{.h} extension to be used in C++ compilations. @command{g++} is a
988 program that calls GCC with the default language set to C++, and
989 automatically specifies linking against the C++ library. On many
990 systems, @command{g++} is also installed with the name @command{c++}.
992 @cindex invoking @command{g++}
993 When you compile C++ programs, you may specify many of the same
994 command-line options that you use for compiling programs in any
995 language; or command-line options meaningful for C and related
996 languages; or options that are meaningful only for C++ programs.
997 @xref{C Dialect Options,,Options Controlling C Dialect}, for
998 explanations of options for languages related to C@.
999 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
1000 explanations of options that are meaningful only for C++ programs.
1002 @node C Dialect Options
1003 @section Options Controlling C Dialect
1004 @cindex dialect options
1005 @cindex language dialect options
1006 @cindex options, dialect
1008 The following options control the dialect of C (or languages derived
1009 from C, such as C++ and Objective-C) that the compiler accepts:
1012 @cindex ANSI support
1016 In C mode, support all ISO C90 programs. In C++ mode,
1017 remove GNU extensions that conflict with ISO C++.
1019 This turns off certain features of GCC that are incompatible with ISO
1020 C90 (when compiling C code), or of standard C++ (when compiling C++ code),
1021 such as the @code{asm} and @code{typeof} keywords, and
1022 predefined macros such as @code{unix} and @code{vax} that identify the
1023 type of system you are using. It also enables the undesirable and
1024 rarely used ISO trigraph feature. For the C compiler,
1025 it disables recognition of C++ style @samp{//} comments as well as
1026 the @code{inline} keyword.
1028 The alternate keywords @code{__asm__}, @code{__extension__},
1029 @code{__inline__} and @code{__typeof__} continue to work despite
1030 @option{-ansi}. You would not want to use them in an ISO C program, of
1031 course, but it is useful to put them in header files that might be included
1032 in compilations done with @option{-ansi}. Alternate predefined macros
1033 such as @code{__unix__} and @code{__vax__} are also available, with or
1034 without @option{-ansi}.
1036 The @option{-ansi} option does not cause non-ISO programs to be
1037 rejected gratuitously. For that, @option{-pedantic} is required in
1038 addition to @option{-ansi}. @xref{Warning Options}.
1040 The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi}
1041 option is used. Some header files may notice this macro and refrain
1042 from declaring certain functions or defining certain macros that the
1043 ISO standard doesn't call for; this is to avoid interfering with any
1044 programs that might use these names for other things.
1046 Functions which would normally be built in but do not have semantics
1047 defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in
1048 functions with @option{-ansi} is used. @xref{Other Builtins,,Other
1049 built-in functions provided by GCC}, for details of the functions
1054 Determine the language standard. This option is currently only
1055 supported when compiling C or C++. A value for this option must be
1056 provided; possible values are
1061 ISO C90 (same as @option{-ansi}).
1063 @item iso9899:199409
1064 ISO C90 as modified in amendment 1.
1070 ISO C99. Note that this standard is not yet fully supported; see
1071 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The
1072 names @samp{c9x} and @samp{iso9899:199x} are deprecated.
1075 Default, ISO C90 plus GNU extensions (including some C99 features).
1079 ISO C99 plus GNU extensions. When ISO C99 is fully implemented in GCC,
1080 this will become the default. The name @samp{gnu9x} is deprecated.
1083 The 1998 ISO C++ standard plus amendments.
1086 The same as @option{-std=c++98} plus GNU extensions. This is the
1087 default for C++ code.
1090 Even when this option is not specified, you can still use some of the
1091 features of newer standards in so far as they do not conflict with
1092 previous C standards. For example, you may use @code{__restrict__} even
1093 when @option{-std=c99} is not specified.
1095 The @option{-std} options specifying some version of ISO C have the same
1096 effects as @option{-ansi}, except that features that were not in ISO C90
1097 but are in the specified version (for example, @samp{//} comments and
1098 the @code{inline} keyword in ISO C99) are not disabled.
1100 @xref{Standards,,Language Standards Supported by GCC}, for details of
1101 these standard versions.
1103 @item -aux-info @var{filename}
1105 Output to the given filename prototyped declarations for all functions
1106 declared and/or defined in a translation unit, including those in header
1107 files. This option is silently ignored in any language other than C@.
1109 Besides declarations, the file indicates, in comments, the origin of
1110 each declaration (source file and line), whether the declaration was
1111 implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or
1112 @samp{O} for old, respectively, in the first character after the line
1113 number and the colon), and whether it came from a declaration or a
1114 definition (@samp{C} or @samp{F}, respectively, in the following
1115 character). In the case of function definitions, a K&R-style list of
1116 arguments followed by their declarations is also provided, inside
1117 comments, after the declaration.
1121 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
1122 keyword, so that code can use these words as identifiers. You can use
1123 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
1124 instead. @option{-ansi} implies @option{-fno-asm}.
1126 In C++, this switch only affects the @code{typeof} keyword, since
1127 @code{asm} and @code{inline} are standard keywords. You may want to
1128 use the @option{-fno-gnu-keywords} flag instead, which has the same
1129 effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this
1130 switch only affects the @code{asm} and @code{typeof} keywords, since
1131 @code{inline} is a standard keyword in ISO C99.
1134 @itemx -fno-builtin-@var{function}
1135 @opindex fno-builtin
1136 @cindex built-in functions
1137 Don't recognize built-in functions that do not begin with
1138 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
1139 functions provided by GCC}, for details of the functions affected,
1140 including those which are not built-in functions when @option{-ansi} or
1141 @option{-std} options for strict ISO C conformance are used because they
1142 do not have an ISO standard meaning.
1144 GCC normally generates special code to handle certain built-in functions
1145 more efficiently; for instance, calls to @code{alloca} may become single
1146 instructions that adjust the stack directly, and calls to @code{memcpy}
1147 may become inline copy loops. The resulting code is often both smaller
1148 and faster, but since the function calls no longer appear as such, you
1149 cannot set a breakpoint on those calls, nor can you change the behavior
1150 of the functions by linking with a different library.
1152 With the @option{-fno-builtin-@var{function}} option
1153 only the built-in function @var{function} is
1154 disabled. @var{function} must not begin with @samp{__builtin_}. If a
1155 function is named this is not built-in in this version of GCC, this
1156 option is ignored. There is no corresponding
1157 @option{-fbuiltin-@var{function}} option; if you wish to enable
1158 built-in functions selectively when using @option{-fno-builtin} or
1159 @option{-ffreestanding}, you may define macros such as:
1162 #define abs(n) __builtin_abs ((n))
1163 #define strcpy(d, s) __builtin_strcpy ((d), (s))
1168 @cindex hosted environment
1170 Assert that compilation takes place in a hosted environment. This implies
1171 @option{-fbuiltin}. A hosted environment is one in which the
1172 entire standard library is available, and in which @code{main} has a return
1173 type of @code{int}. Examples are nearly everything except a kernel.
1174 This is equivalent to @option{-fno-freestanding}.
1176 @item -ffreestanding
1177 @opindex ffreestanding
1178 @cindex hosted environment
1180 Assert that compilation takes place in a freestanding environment. This
1181 implies @option{-fno-builtin}. A freestanding environment
1182 is one in which the standard library may not exist, and program startup may
1183 not necessarily be at @code{main}. The most obvious example is an OS kernel.
1184 This is equivalent to @option{-fno-hosted}.
1186 @xref{Standards,,Language Standards Supported by GCC}, for details of
1187 freestanding and hosted environments.
1189 @item -fms-extensions
1190 @opindex fms-extensions
1191 Accept some non-standard constructs used in Microsoft header files.
1195 Support ISO C trigraphs. The @option{-ansi} option (and @option{-std}
1196 options for strict ISO C conformance) implies @option{-trigraphs}.
1198 @item -no-integrated-cpp
1199 @opindex no-integrated-cpp
1200 Performs a compilation in two passes: preprocessing and compiling. This
1201 option allows a user supplied "cc1", "cc1plus", or "cc1obj" via the
1202 @option{-B} option. The user supplied compilation step can then add in
1203 an additional preprocessing step after normal preprocessing but before
1204 compiling. The default is to use the integrated cpp (internal cpp)
1206 The semantics of this option will change if "cc1", "cc1plus", and
1207 "cc1obj" are merged.
1209 @cindex traditional C language
1210 @cindex C language, traditional
1212 @itemx -traditional-cpp
1213 @opindex traditional-cpp
1214 @opindex traditional
1215 Formerly, these options caused GCC to attempt to emulate a pre-standard
1216 C compiler. They are now only supported with the @option{-E} switch.
1217 The preprocessor continues to support a pre-standard mode. See the GNU
1218 CPP manual for details.
1220 @item -fcond-mismatch
1221 @opindex fcond-mismatch
1222 Allow conditional expressions with mismatched types in the second and
1223 third arguments. The value of such an expression is void. This option
1224 is not supported for C++.
1226 @item -funsigned-char
1227 @opindex funsigned-char
1228 Let the type @code{char} be unsigned, like @code{unsigned char}.
1230 Each kind of machine has a default for what @code{char} should
1231 be. It is either like @code{unsigned char} by default or like
1232 @code{signed char} by default.
1234 Ideally, a portable program should always use @code{signed char} or
1235 @code{unsigned char} when it depends on the signedness of an object.
1236 But many programs have been written to use plain @code{char} and
1237 expect it to be signed, or expect it to be unsigned, depending on the
1238 machines they were written for. This option, and its inverse, let you
1239 make such a program work with the opposite default.
1241 The type @code{char} is always a distinct type from each of
1242 @code{signed char} or @code{unsigned char}, even though its behavior
1243 is always just like one of those two.
1246 @opindex fsigned-char
1247 Let the type @code{char} be signed, like @code{signed char}.
1249 Note that this is equivalent to @option{-fno-unsigned-char}, which is
1250 the negative form of @option{-funsigned-char}. Likewise, the option
1251 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
1253 @item -fsigned-bitfields
1254 @itemx -funsigned-bitfields
1255 @itemx -fno-signed-bitfields
1256 @itemx -fno-unsigned-bitfields
1257 @opindex fsigned-bitfields
1258 @opindex funsigned-bitfields
1259 @opindex fno-signed-bitfields
1260 @opindex fno-unsigned-bitfields
1261 These options control whether a bit-field is signed or unsigned, when the
1262 declaration does not use either @code{signed} or @code{unsigned}. By
1263 default, such a bit-field is signed, because this is consistent: the
1264 basic integer types such as @code{int} are signed types.
1267 @node C++ Dialect Options
1268 @section Options Controlling C++ Dialect
1270 @cindex compiler options, C++
1271 @cindex C++ options, command line
1272 @cindex options, C++
1273 This section describes the command-line options that are only meaningful
1274 for C++ programs; but you can also use most of the GNU compiler options
1275 regardless of what language your program is in. For example, you
1276 might compile a file @code{firstClass.C} like this:
1279 g++ -g -frepo -O -c firstClass.C
1283 In this example, only @option{-frepo} is an option meant
1284 only for C++ programs; you can use the other options with any
1285 language supported by GCC@.
1287 Here is a list of options that are @emph{only} for compiling C++ programs:
1291 @item -fabi-version=@var{n}
1292 @opindex fabi-version
1293 Use version @var{n} of the C++ ABI. Version 2 is the version of the
1294 C++ ABI that first appeared in G++ 3.4. Version 1 is the version of
1295 the C++ ABI that first appeared in G++ 3.2. Version 0 will always be
1296 the version that conforms most closely to the C++ ABI specification.
1297 Therefore, the ABI obtained using version 0 will change as ABI bugs
1300 The default is version 2.
1302 @item -fno-access-control
1303 @opindex fno-access-control
1304 Turn off all access checking. This switch is mainly useful for working
1305 around bugs in the access control code.
1309 Check that the pointer returned by @code{operator new} is non-null
1310 before attempting to modify the storage allocated. This check is
1311 normally unnecessary because the C++ standard specifies that
1312 @code{operator new} will only return @code{0} if it is declared
1313 @samp{throw()}, in which case the compiler will always check the
1314 return value even without this option. In all other cases, when
1315 @code{operator new} has a non-empty exception specification, memory
1316 exhaustion is signalled by throwing @code{std::bad_alloc}. See also
1317 @samp{new (nothrow)}.
1319 @item -fconserve-space
1320 @opindex fconserve-space
1321 Put uninitialized or runtime-initialized global variables into the
1322 common segment, as C does. This saves space in the executable at the
1323 cost of not diagnosing duplicate definitions. If you compile with this
1324 flag and your program mysteriously crashes after @code{main()} has
1325 completed, you may have an object that is being destroyed twice because
1326 two definitions were merged.
1328 This option is no longer useful on most targets, now that support has
1329 been added for putting variables into BSS without making them common.
1331 @item -fno-const-strings
1332 @opindex fno-const-strings
1333 Give string constants type @code{char *} instead of type @code{const
1334 char *}. By default, G++ uses type @code{const char *} as required by
1335 the standard. Even if you use @option{-fno-const-strings}, you cannot
1336 actually modify the value of a string constant.
1338 This option might be removed in a future release of G++. For maximum
1339 portability, you should structure your code so that it works with
1340 string constants that have type @code{const char *}.
1342 @item -fno-elide-constructors
1343 @opindex fno-elide-constructors
1344 The C++ standard allows an implementation to omit creating a temporary
1345 which is only used to initialize another object of the same type.
1346 Specifying this option disables that optimization, and forces G++ to
1347 call the copy constructor in all cases.
1349 @item -fno-enforce-eh-specs
1350 @opindex fno-enforce-eh-specs
1351 Don't check for violation of exception specifications at runtime. This
1352 option violates the C++ standard, but may be useful for reducing code
1353 size in production builds, much like defining @samp{NDEBUG}. The compiler
1354 will still optimize based on the exception specifications.
1357 @itemx -fno-for-scope
1359 @opindex fno-for-scope
1360 If @option{-ffor-scope} is specified, the scope of variables declared in
1361 a @i{for-init-statement} is limited to the @samp{for} loop itself,
1362 as specified by the C++ standard.
1363 If @option{-fno-for-scope} is specified, the scope of variables declared in
1364 a @i{for-init-statement} extends to the end of the enclosing scope,
1365 as was the case in old versions of G++, and other (traditional)
1366 implementations of C++.
1368 The default if neither flag is given to follow the standard,
1369 but to allow and give a warning for old-style code that would
1370 otherwise be invalid, or have different behavior.
1372 @item -fno-gnu-keywords
1373 @opindex fno-gnu-keywords
1374 Do not recognize @code{typeof} as a keyword, so that code can use this
1375 word as an identifier. You can use the keyword @code{__typeof__} instead.
1376 @option{-ansi} implies @option{-fno-gnu-keywords}.
1378 @item -fno-implicit-templates
1379 @opindex fno-implicit-templates
1380 Never emit code for non-inline templates which are instantiated
1381 implicitly (i.e.@: by use); only emit code for explicit instantiations.
1382 @xref{Template Instantiation}, for more information.
1384 @item -fno-implicit-inline-templates
1385 @opindex fno-implicit-inline-templates
1386 Don't emit code for implicit instantiations of inline templates, either.
1387 The default is to handle inlines differently so that compiles with and
1388 without optimization will need the same set of explicit instantiations.
1390 @item -fno-implement-inlines
1391 @opindex fno-implement-inlines
1392 To save space, do not emit out-of-line copies of inline functions
1393 controlled by @samp{#pragma implementation}. This will cause linker
1394 errors if these functions are not inlined everywhere they are called.
1396 @item -fms-extensions
1397 @opindex fms-extensions
1398 Disable pedantic warnings about constructs used in MFC, such as implicit
1399 int and getting a pointer to member function via non-standard syntax.
1401 @item -fno-nonansi-builtins
1402 @opindex fno-nonansi-builtins
1403 Disable built-in declarations of functions that are not mandated by
1404 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
1405 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
1407 @item -fno-operator-names
1408 @opindex fno-operator-names
1409 Do not treat the operator name keywords @code{and}, @code{bitand},
1410 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
1411 synonyms as keywords.
1413 @item -fno-optional-diags
1414 @opindex fno-optional-diags
1415 Disable diagnostics that the standard says a compiler does not need to
1416 issue. Currently, the only such diagnostic issued by G++ is the one for
1417 a name having multiple meanings within a class.
1420 @opindex fpermissive
1421 Downgrade some diagnostics about nonconformant code from errors to
1422 warnings. Thus, using @option{-fpermissive} will allow some
1423 nonconforming code to compile.
1427 Enable automatic template instantiation at link time. This option also
1428 implies @option{-fno-implicit-templates}. @xref{Template
1429 Instantiation}, for more information.
1433 Disable generation of information about every class with virtual
1434 functions for use by the C++ runtime type identification features
1435 (@samp{dynamic_cast} and @samp{typeid}). If you don't use those parts
1436 of the language, you can save some space by using this flag. Note that
1437 exception handling uses the same information, but it will generate it as
1442 Emit statistics about front-end processing at the end of the compilation.
1443 This information is generally only useful to the G++ development team.
1445 @item -ftemplate-depth-@var{n}
1446 @opindex ftemplate-depth
1447 Set the maximum instantiation depth for template classes to @var{n}.
1448 A limit on the template instantiation depth is needed to detect
1449 endless recursions during template class instantiation. ANSI/ISO C++
1450 conforming programs must not rely on a maximum depth greater than 17.
1452 @item -fuse-cxa-atexit
1453 @opindex fuse-cxa-atexit
1454 Register destructors for objects with static storage duration with the
1455 @code{__cxa_atexit} function rather than the @code{atexit} function.
1456 This option is required for fully standards-compliant handling of static
1457 destructors, but will only work if your C library supports
1458 @code{__cxa_atexit}.
1462 Do not use weak symbol support, even if it is provided by the linker.
1463 By default, G++ will use weak symbols if they are available. This
1464 option exists only for testing, and should not be used by end-users;
1465 it will result in inferior code and has no benefits. This option may
1466 be removed in a future release of G++.
1470 Do not search for header files in the standard directories specific to
1471 C++, but do still search the other standard directories. (This option
1472 is used when building the C++ library.)
1475 In addition, these optimization, warning, and code generation options
1476 have meanings only for C++ programs:
1479 @item -fno-default-inline
1480 @opindex fno-default-inline
1481 Do not assume @samp{inline} for functions defined inside a class scope.
1482 @xref{Optimize Options,,Options That Control Optimization}. Note that these
1483 functions will have linkage like inline functions; they just won't be
1486 @item -Wabi @r{(C++ only)}
1488 Warn when G++ generates code that is probably not compatible with the
1489 vendor-neutral C++ ABI. Although an effort has been made to warn about
1490 all such cases, there are probably some cases that are not warned about,
1491 even though G++ is generating incompatible code. There may also be
1492 cases where warnings are emitted even though the code that is generated
1495 You should rewrite your code to avoid these warnings if you are
1496 concerned about the fact that code generated by G++ may not be binary
1497 compatible with code generated by other compilers.
1499 The known incompatibilities at this point include:
1504 Incorrect handling of tail-padding for bit-fields. G++ may attempt to
1505 pack data into the same byte as a base class. For example:
1508 struct A @{ virtual void f(); int f1 : 1; @};
1509 struct B : public A @{ int f2 : 1; @};
1513 In this case, G++ will place @code{B::f2} into the same byte
1514 as@code{A::f1}; other compilers will not. You can avoid this problem
1515 by explicitly padding @code{A} so that its size is a multiple of the
1516 byte size on your platform; that will cause G++ and other compilers to
1517 layout @code{B} identically.
1520 Incorrect handling of tail-padding for virtual bases. G++ does not use
1521 tail padding when laying out virtual bases. For example:
1524 struct A @{ virtual void f(); char c1; @};
1525 struct B @{ B(); char c2; @};
1526 struct C : public A, public virtual B @{@};
1530 In this case, G++ will not place @code{B} into the tail-padding for
1531 @code{A}; other compilers will. You can avoid this problem by
1532 explicitly padding @code{A} so that its size is a multiple of its
1533 alignment (ignoring virtual base classes); that will cause G++ and other
1534 compilers to layout @code{C} identically.
1537 Incorrect handling of bit-fields with declared widths greater than that
1538 of their underlying types, when the bit-fields appear in a union. For
1542 union U @{ int i : 4096; @};
1546 Assuming that an @code{int} does not have 4096 bits, G++ will make the
1547 union too small by the number of bits in an @code{int}.
1550 Empty classes can be placed at incorrect offsets. For example:
1560 struct C : public B, public A @{@};
1564 G++ will place the @code{A} base class of @code{C} at a nonzero offset;
1565 it should be placed at offset zero. G++ mistakenly believes that the
1566 @code{A} data member of @code{B} is already at offset zero.
1569 Names of template functions whose types involve @code{typename} or
1570 template template parameters can be mangled incorrectly.
1573 template <typename Q>
1574 void f(typename Q::X) @{@}
1576 template <template <typename> class Q>
1577 void f(typename Q<int>::X) @{@}
1581 Instantiations of these templates may be mangled incorrectly.
1585 @item -Wctor-dtor-privacy @r{(C++ only)}
1586 @opindex Wctor-dtor-privacy
1587 Warn when a class seems unusable because all the constructors or
1588 destructors in that class are private, and it has neither friends nor
1589 public static member functions.
1591 @item -Wnon-virtual-dtor @r{(C++ only)}
1592 @opindex Wnon-virtual-dtor
1593 Warn when a class appears to be polymorphic, thereby requiring a virtual
1594 destructor, yet it declares a non-virtual one.
1595 This warning is enabled by @option{-Wall}.
1597 @item -Wreorder @r{(C++ only)}
1599 @cindex reordering, warning
1600 @cindex warning for reordering of member initializers
1601 Warn when the order of member initializers given in the code does not
1602 match the order in which they must be executed. For instance:
1608 A(): j (0), i (1) @{ @}
1612 The compiler will rearrange the member initializers for @samp{i}
1613 and @samp{j} to match the declaration order of the members, emitting
1614 a warning to that effect. This warning is enabled by @option{-Wall}.
1617 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
1620 @item -Weffc++ @r{(C++ only)}
1622 Warn about violations of the following style guidelines from Scott Meyers'
1623 @cite{Effective C++} book:
1627 Item 11: Define a copy constructor and an assignment operator for classes
1628 with dynamically allocated memory.
1631 Item 12: Prefer initialization to assignment in constructors.
1634 Item 14: Make destructors virtual in base classes.
1637 Item 15: Have @code{operator=} return a reference to @code{*this}.
1640 Item 23: Don't try to return a reference when you must return an object.
1644 Also warn about violations of the following style guidelines from
1645 Scott Meyers' @cite{More Effective C++} book:
1649 Item 6: Distinguish between prefix and postfix forms of increment and
1650 decrement operators.
1653 Item 7: Never overload @code{&&}, @code{||}, or @code{,}.
1657 When selecting this option, be aware that the standard library
1658 headers do not obey all of these guidelines; use @samp{grep -v}
1659 to filter out those warnings.
1661 @item -Wno-deprecated @r{(C++ only)}
1662 @opindex Wno-deprecated
1663 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
1665 @item -Wno-non-template-friend @r{(C++ only)}
1666 @opindex Wno-non-template-friend
1667 Disable warnings when non-templatized friend functions are declared
1668 within a template. Since the advent of explicit template specification
1669 support in G++, if the name of the friend is an unqualified-id (i.e.,
1670 @samp{friend foo(int)}), the C++ language specification demands that the
1671 friend declare or define an ordinary, nontemplate function. (Section
1672 14.5.3). Before G++ implemented explicit specification, unqualified-ids
1673 could be interpreted as a particular specialization of a templatized
1674 function. Because this non-conforming behavior is no longer the default
1675 behavior for G++, @option{-Wnon-template-friend} allows the compiler to
1676 check existing code for potential trouble spots and is on by default.
1677 This new compiler behavior can be turned off with
1678 @option{-Wno-non-template-friend} which keeps the conformant compiler code
1679 but disables the helpful warning.
1681 @item -Wold-style-cast @r{(C++ only)}
1682 @opindex Wold-style-cast
1683 Warn if an old-style (C-style) cast to a non-void type is used within
1684 a C++ program. The new-style casts (@samp{static_cast},
1685 @samp{reinterpret_cast}, and @samp{const_cast}) are less vulnerable to
1686 unintended effects and much easier to search for.
1688 @item -Woverloaded-virtual @r{(C++ only)}
1689 @opindex Woverloaded-virtual
1690 @cindex overloaded virtual fn, warning
1691 @cindex warning for overloaded virtual fn
1692 Warn when a function declaration hides virtual functions from a
1693 base class. For example, in:
1700 struct B: public A @{
1705 the @code{A} class version of @code{f} is hidden in @code{B}, and code
1713 will fail to compile.
1715 @item -Wno-pmf-conversions @r{(C++ only)}
1716 @opindex Wno-pmf-conversions
1717 Disable the diagnostic for converting a bound pointer to member function
1720 @item -Wsign-promo @r{(C++ only)}
1721 @opindex Wsign-promo
1722 Warn when overload resolution chooses a promotion from unsigned or
1723 enumerated type to a signed type, over a conversion to an unsigned type of
1724 the same size. Previous versions of G++ would try to preserve
1725 unsignedness, but the standard mandates the current behavior.
1727 @item -Wsynth @r{(C++ only)}
1729 @cindex warning for synthesized methods
1730 @cindex synthesized methods, warning
1731 Warn when G++'s synthesis behavior does not match that of cfront. For
1737 A& operator = (int);
1747 In this example, G++ will synthesize a default @samp{A& operator =
1748 (const A&);}, while cfront will use the user-defined @samp{operator =}.
1751 @node Objective-C Dialect Options
1752 @section Options Controlling Objective-C Dialect
1754 @cindex compiler options, Objective-C
1755 @cindex Objective-C options, command line
1756 @cindex options, Objective-C
1757 (NOTE: This manual does not describe the Objective-C language itself. See
1758 @w{@uref{http://gcc.gnu.org/readings.html}} for references.)
1760 This section describes the command-line options that are only meaningful
1761 for Objective-C programs, but you can also use most of the GNU compiler
1762 options regardless of what language your program is in. For example,
1763 you might compile a file @code{some_class.m} like this:
1766 gcc -g -fgnu-runtime -O -c some_class.m
1770 In this example, @option{-fgnu-runtime} is an option meant only for
1771 Objective-C programs; you can use the other options with any language
1774 Here is a list of options that are @emph{only} for compiling Objective-C
1778 @item -fconstant-string-class=@var{class-name}
1779 @opindex fconstant-string-class
1780 Use @var{class-name} as the name of the class to instantiate for each
1781 literal string specified with the syntax @code{@@"@dots{}"}. The default
1782 class name is @code{NXConstantString} if the GNU runtime is being used, and
1783 @code{NSConstantString} if the NeXT runtime is being used (see below). The
1784 @option{-fconstant-cfstrings} option, if also present, will override the
1785 @option{-fconstant-string-class} setting and cause @code{@@"@dots{}"} literals
1786 to be laid out as constant CoreFoundation strings.
1789 @opindex fgnu-runtime
1790 Generate object code compatible with the standard GNU Objective-C
1791 runtime. This is the default for most types of systems.
1793 @item -fnext-runtime
1794 @opindex fnext-runtime
1795 Generate output compatible with the NeXT runtime. This is the default
1796 for NeXT-based systems, including Darwin and Mac OS X@. The macro
1797 @code{__NEXT_RUNTIME__} is predefined if (and only if) this option is
1800 @item -fno-nil-receivers
1801 @opindex fno-nil-receivers
1802 Assume that all Objective-C message dispatches (e.g.,
1803 @code{[receiver message:arg]}) in this translation unit ensure that the receiver
1804 is not @code{nil}. This allows for more efficient entry points in the runtime to be
1805 used. Currently, this option is only available in conjunction with
1806 the NeXT runtime on Mac OS X 10.3 and later.
1808 @item -fobjc-exceptions
1809 @opindex fobjc-exceptions
1810 Enable syntactic support for structured exception handling in Objective-C,
1811 similar to what is offered by C++ and Java. Currently, this option is only
1812 available in conjunction with the NeXT runtime on Mac OS X 10.3 and later.
1820 @@catch (AnObjCClass *exc) @{
1827 @@catch (AnotherClass *exc) @{
1830 @@catch (id allOthers) @{
1840 The @code{@@throw} statement may appear anywhere in an Objective-C or
1841 Objective-C++ program; when used inside of a @code{@@catch} block, the
1842 @code{@@throw} may appear without an argument (as shown above), in which case
1843 the object caught by the @code{@@catch} will be rethrown.
1845 Note that only (pointers to) Objective-C objects may be thrown and
1846 caught using this scheme. When an object is thrown, it will be caught
1847 by the nearest @code{@@catch} clause capable of handling objects of that type,
1848 analogously to how @code{catch} blocks work in C++ and Java. A
1849 @code{@@catch(id @dots{})} clause (as shown above) may also be provided to catch
1850 any and all Objective-C exceptions not caught by previous @code{@@catch}
1853 The @code{@@finally} clause, if present, will be executed upon exit from the
1854 immediately preceding @code{@@try @dots{} @@catch} section. This will happen
1855 regardless of whether any exceptions are thrown, caught or rethrown
1856 inside the @code{@@try @dots{} @@catch} section, analogously to the behavior
1857 of the @code{finally} clause in Java.
1859 There are several caveats to using the new exception mechanism:
1863 Although currently designed to be binary compatible with @code{NS_HANDLER}-style
1864 idioms provided by the @code{NSException} class, the new
1865 exceptions can only be used on Mac OS X 10.3 (Panther) and later
1866 systems, due to additional functionality needed in the (NeXT) Objective-C
1870 As mentioned above, the new exceptions do not support handling
1871 types other than Objective-C objects. Furthermore, when used from
1872 Objective-C++, the Objective-C exception model does not interoperate with C++
1873 exceptions at this time. This means you cannot @code{@@throw} an exception
1874 from Objective-C and @code{catch} it in C++, or vice versa
1875 (i.e., @code{throw @dots{} @@catch}).
1878 The @option{-fobjc-exceptions} switch also enables the use of synchronization
1879 blocks for thread-safe execution:
1882 @@synchronized (ObjCClass *guard) @{
1887 Upon entering the @code{@@synchronized} block, a thread of execution shall
1888 first check whether a lock has been placed on the corresponding @code{guard}
1889 object by another thread. If it has, the current thread shall wait until
1890 the other thread relinquishes its lock. Once @code{guard} becomes available,
1891 the current thread will place its own lock on it, execute the code contained in
1892 the @code{@@synchronized} block, and finally relinquish the lock (thereby
1893 making @code{guard} available to other threads).
1895 Unlike Java, Objective-C does not allow for entire methods to be marked
1896 @code{@@synchronized}. Note that throwing exceptions out of
1897 @code{@@synchronized} blocks is allowed, and will cause the guarding object
1898 to be unlocked properly.
1900 @item -freplace-objc-classes
1901 @opindex freplace-objc-classes
1902 Emit a special marker instructing @command{ld(1)} not to statically link in
1903 the resulting object file, and allow @command{dyld(1)} to load it in at
1904 run time instead. This is used in conjunction with the Fix-and-Continue
1905 debugging mode, where the object file in question may be recompiled and
1906 dynamically reloaded in the course of program execution, without the need
1907 to restart the program itself. Currently, Fix-and-Continue functionality
1908 is only available in conjunction with the NeXT runtime on Mac OS X 10.3
1913 When compiling for the NeXT runtime, the compiler ordinarily replaces calls
1914 to @code{objc_getClass("@dots{}")} (when the name of the class is known at
1915 compile time) with static class references that get initialized at load time,
1916 which improves run-time performance. Specifying the @option{-fzero-link} flag
1917 suppresses this behavior and causes calls to @code{objc_getClass("@dots{}")}
1918 to be retained. This is useful in Zero-Link debugging mode, since it allows
1919 for individual class implementations to be modified during program execution.
1923 Dump interface declarations for all classes seen in the source file to a
1924 file named @file{@var{sourcename}.decl}.
1927 @opindex Wno-protocol
1928 If a class is declared to implement a protocol, a warning is issued for
1929 every method in the protocol that is not implemented by the class. The
1930 default behavior is to issue a warning for every method not explicitly
1931 implemented in the class, even if a method implementation is inherited
1932 from the superclass. If you use the @code{-Wno-protocol} option, then
1933 methods inherited from the superclass are considered to be implemented,
1934 and no warning is issued for them.
1938 Warn if multiple methods of different types for the same selector are
1939 found during compilation. The check is performed on the list of methods
1940 in the final stage of compilation. Additionally, a check is performed
1941 for each selector appearing in a @code{@@selector(@dots{})}
1942 expression, and a corresponding method for that selector has been found
1943 during compilation. Because these checks scan the method table only at
1944 the end of compilation, these warnings are not produced if the final
1945 stage of compilation is not reached, for example because an error is
1946 found during compilation, or because the @code{-fsyntax-only} option is
1949 @item -Wundeclared-selector
1950 @opindex Wundeclared-selector
1951 Warn if a @code{@@selector(@dots{})} expression referring to an
1952 undeclared selector is found. A selector is considered undeclared if no
1953 method with that name has been declared before the
1954 @code{@@selector(@dots{})} expression, either explicitly in an
1955 @code{@@interface} or @code{@@protocol} declaration, or implicitly in
1956 an @code{@@implementation} section. This option always performs its
1957 checks as soon as a @code{@@selector(@dots{})} expression is found,
1958 while @code{-Wselector} only performs its checks in the final stage of
1959 compilation. This also enforces the coding style convention
1960 that methods and selectors must be declared before being used.
1962 @item -print-objc-runtime-info
1963 @opindex print-objc-runtime-info
1964 Generate C header describing the largest structure that is passed by
1969 @node Language Independent Options
1970 @section Options to Control Diagnostic Messages Formatting
1971 @cindex options to control diagnostics formatting
1972 @cindex diagnostic messages
1973 @cindex message formatting
1975 Traditionally, diagnostic messages have been formatted irrespective of
1976 the output device's aspect (e.g.@: its width, @dots{}). The options described
1977 below can be used to control the diagnostic messages formatting
1978 algorithm, e.g.@: how many characters per line, how often source location
1979 information should be reported. Right now, only the C++ front end can
1980 honor these options. However it is expected, in the near future, that
1981 the remaining front ends would be able to digest them correctly.
1984 @item -fmessage-length=@var{n}
1985 @opindex fmessage-length
1986 Try to format error messages so that they fit on lines of about @var{n}
1987 characters. The default is 72 characters for @command{g++} and 0 for the rest of
1988 the front ends supported by GCC@. If @var{n} is zero, then no
1989 line-wrapping will be done; each error message will appear on a single
1992 @opindex fdiagnostics-show-location
1993 @item -fdiagnostics-show-location=once
1994 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
1995 reporter to emit @emph{once} source location information; that is, in
1996 case the message is too long to fit on a single physical line and has to
1997 be wrapped, the source location won't be emitted (as prefix) again,
1998 over and over, in subsequent continuation lines. This is the default
2001 @item -fdiagnostics-show-location=every-line
2002 Only meaningful in line-wrapping mode. Instructs the diagnostic
2003 messages reporter to emit the same source location information (as
2004 prefix) for physical lines that result from the process of breaking
2005 a message which is too long to fit on a single line.
2009 @node Warning Options
2010 @section Options to Request or Suppress Warnings
2011 @cindex options to control warnings
2012 @cindex warning messages
2013 @cindex messages, warning
2014 @cindex suppressing warnings
2016 Warnings are diagnostic messages that report constructions which
2017 are not inherently erroneous but which are risky or suggest there
2018 may have been an error.
2020 You can request many specific warnings with options beginning @samp{-W},
2021 for example @option{-Wimplicit} to request warnings on implicit
2022 declarations. Each of these specific warning options also has a
2023 negative form beginning @samp{-Wno-} to turn off warnings;
2024 for example, @option{-Wno-implicit}. This manual lists only one of the
2025 two forms, whichever is not the default.
2027 The following options control the amount and kinds of warnings produced
2028 by GCC; for further, language-specific options also refer to
2029 @ref{C++ Dialect Options} and @ref{Objective-C Dialect Options}.
2032 @cindex syntax checking
2034 @opindex fsyntax-only
2035 Check the code for syntax errors, but don't do anything beyond that.
2039 Issue all the warnings demanded by strict ISO C and ISO C++;
2040 reject all programs that use forbidden extensions, and some other
2041 programs that do not follow ISO C and ISO C++. For ISO C, follows the
2042 version of the ISO C standard specified by any @option{-std} option used.
2044 Valid ISO C and ISO C++ programs should compile properly with or without
2045 this option (though a rare few will require @option{-ansi} or a
2046 @option{-std} option specifying the required version of ISO C)@. However,
2047 without this option, certain GNU extensions and traditional C and C++
2048 features are supported as well. With this option, they are rejected.
2050 @option{-pedantic} does not cause warning messages for use of the
2051 alternate keywords whose names begin and end with @samp{__}. Pedantic
2052 warnings are also disabled in the expression that follows
2053 @code{__extension__}. However, only system header files should use
2054 these escape routes; application programs should avoid them.
2055 @xref{Alternate Keywords}.
2057 Some users try to use @option{-pedantic} to check programs for strict ISO
2058 C conformance. They soon find that it does not do quite what they want:
2059 it finds some non-ISO practices, but not all---only those for which
2060 ISO C @emph{requires} a diagnostic, and some others for which
2061 diagnostics have been added.
2063 A feature to report any failure to conform to ISO C might be useful in
2064 some instances, but would require considerable additional work and would
2065 be quite different from @option{-pedantic}. We don't have plans to
2066 support such a feature in the near future.
2068 Where the standard specified with @option{-std} represents a GNU
2069 extended dialect of C, such as @samp{gnu89} or @samp{gnu99}, there is a
2070 corresponding @dfn{base standard}, the version of ISO C on which the GNU
2071 extended dialect is based. Warnings from @option{-pedantic} are given
2072 where they are required by the base standard. (It would not make sense
2073 for such warnings to be given only for features not in the specified GNU
2074 C dialect, since by definition the GNU dialects of C include all
2075 features the compiler supports with the given option, and there would be
2076 nothing to warn about.)
2078 @item -pedantic-errors
2079 @opindex pedantic-errors
2080 Like @option{-pedantic}, except that errors are produced rather than
2085 Inhibit all warning messages.
2089 Inhibit warning messages about the use of @samp{#import}.
2091 @item -Wchar-subscripts
2092 @opindex Wchar-subscripts
2093 Warn if an array subscript has type @code{char}. This is a common cause
2094 of error, as programmers often forget that this type is signed on some
2099 Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*}
2100 comment, or whenever a Backslash-Newline appears in a @samp{//} comment.
2102 @item -Wfatal-errors
2103 @opindex Wfatal-errors
2104 This option causes the compiler to abort compilation on the first error
2105 occurred rather than trying to keep going and printing further error
2110 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
2111 the arguments supplied have types appropriate to the format string
2112 specified, and that the conversions specified in the format string make
2113 sense. This includes standard functions, and others specified by format
2114 attributes (@pxref{Function Attributes}), in the @code{printf},
2115 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
2116 not in the C standard) families.
2118 The formats are checked against the format features supported by GNU
2119 libc version 2.2. These include all ISO C90 and C99 features, as well
2120 as features from the Single Unix Specification and some BSD and GNU
2121 extensions. Other library implementations may not support all these
2122 features; GCC does not support warning about features that go beyond a
2123 particular library's limitations. However, if @option{-pedantic} is used
2124 with @option{-Wformat}, warnings will be given about format features not
2125 in the selected standard version (but not for @code{strfmon} formats,
2126 since those are not in any version of the C standard). @xref{C Dialect
2127 Options,,Options Controlling C Dialect}.
2129 Since @option{-Wformat} also checks for null format arguments for
2130 several functions, @option{-Wformat} also implies @option{-Wnonnull}.
2132 @option{-Wformat} is included in @option{-Wall}. For more control over some
2133 aspects of format checking, the options @option{-Wformat-y2k},
2134 @option{-Wno-format-extra-args}, @option{-Wno-format-zero-length},
2135 @option{-Wformat-nonliteral}, @option{-Wformat-security}, and
2136 @option{-Wformat=2} are available, but are not included in @option{-Wall}.
2139 @opindex Wformat-y2k
2140 If @option{-Wformat} is specified, also warn about @code{strftime}
2141 formats which may yield only a two-digit year.
2143 @item -Wno-format-extra-args
2144 @opindex Wno-format-extra-args
2145 If @option{-Wformat} is specified, do not warn about excess arguments to a
2146 @code{printf} or @code{scanf} format function. The C standard specifies
2147 that such arguments are ignored.
2149 Where the unused arguments lie between used arguments that are
2150 specified with @samp{$} operand number specifications, normally
2151 warnings are still given, since the implementation could not know what
2152 type to pass to @code{va_arg} to skip the unused arguments. However,
2153 in the case of @code{scanf} formats, this option will suppress the
2154 warning if the unused arguments are all pointers, since the Single
2155 Unix Specification says that such unused arguments are allowed.
2157 @item -Wno-format-zero-length
2158 @opindex Wno-format-zero-length
2159 If @option{-Wformat} is specified, do not warn about zero-length formats.
2160 The C standard specifies that zero-length formats are allowed.
2162 @item -Wformat-nonliteral
2163 @opindex Wformat-nonliteral
2164 If @option{-Wformat} is specified, also warn if the format string is not a
2165 string literal and so cannot be checked, unless the format function
2166 takes its format arguments as a @code{va_list}.
2168 @item -Wformat-security
2169 @opindex Wformat-security
2170 If @option{-Wformat} is specified, also warn about uses of format
2171 functions that represent possible security problems. At present, this
2172 warns about calls to @code{printf} and @code{scanf} functions where the
2173 format string is not a string literal and there are no format arguments,
2174 as in @code{printf (foo);}. This may be a security hole if the format
2175 string came from untrusted input and contains @samp{%n}. (This is
2176 currently a subset of what @option{-Wformat-nonliteral} warns about, but
2177 in future warnings may be added to @option{-Wformat-security} that are not
2178 included in @option{-Wformat-nonliteral}.)
2182 Enable @option{-Wformat} plus format checks not included in
2183 @option{-Wformat}. Currently equivalent to @samp{-Wformat
2184 -Wformat-nonliteral -Wformat-security -Wformat-y2k}.
2188 Warn about passing a null pointer for arguments marked as
2189 requiring a non-null value by the @code{nonnull} function attribute.
2191 @option{-Wnonnull} is included in @option{-Wall} and @option{-Wformat}. It
2192 can be disabled with the @option{-Wno-nonnull} option.
2194 @item -Winit-self @r{(C, C++, and Objective-C only)}
2196 Warn about uninitialized variables which are initialized with themselves.
2197 Note this option can only be used with the @option{-Wuninitialized} option,
2198 which in turn only works with @option{-O1} and above.
2200 For example, GCC will warn about @code{i} being uninitialized in the
2201 following snippet only when @option{-Winit-self} has been specified:
2212 @item -Wimplicit-int
2213 @opindex Wimplicit-int
2214 Warn when a declaration does not specify a type.
2216 @item -Wimplicit-function-declaration
2217 @itemx -Werror-implicit-function-declaration
2218 @opindex Wimplicit-function-declaration
2219 @opindex Werror-implicit-function-declaration
2220 Give a warning (or error) whenever a function is used before being
2225 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
2229 Warn if the type of @samp{main} is suspicious. @samp{main} should be a
2230 function with external linkage, returning int, taking either zero
2231 arguments, two, or three arguments of appropriate types.
2233 @item -Wmissing-braces
2234 @opindex Wmissing-braces
2235 Warn if an aggregate or union initializer is not fully bracketed. In
2236 the following example, the initializer for @samp{a} is not fully
2237 bracketed, but that for @samp{b} is fully bracketed.
2240 int a[2][2] = @{ 0, 1, 2, 3 @};
2241 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
2244 @item -Wmissing-include-dirs @r{(C, C++, and Objective-C only)}
2245 @opindex Wmissing-include-dirs
2246 Warn if a user-supplied include directory does not exist.
2249 @opindex Wparentheses
2250 Warn if parentheses are omitted in certain contexts, such
2251 as when there is an assignment in a context where a truth value
2252 is expected, or when operators are nested whose precedence people
2253 often get confused about.
2255 Also warn about constructions where there may be confusion to which
2256 @code{if} statement an @code{else} branch belongs. Here is an example of
2271 In C, every @code{else} branch belongs to the innermost possible @code{if}
2272 statement, which in this example is @code{if (b)}. This is often not
2273 what the programmer expected, as illustrated in the above example by
2274 indentation the programmer chose. When there is the potential for this
2275 confusion, GCC will issue a warning when this flag is specified.
2276 To eliminate the warning, add explicit braces around the innermost
2277 @code{if} statement so there is no way the @code{else} could belong to
2278 the enclosing @code{if}. The resulting code would look like this:
2294 @item -Wsequence-point
2295 @opindex Wsequence-point
2296 Warn about code that may have undefined semantics because of violations
2297 of sequence point rules in the C standard.
2299 The C standard defines the order in which expressions in a C program are
2300 evaluated in terms of @dfn{sequence points}, which represent a partial
2301 ordering between the execution of parts of the program: those executed
2302 before the sequence point, and those executed after it. These occur
2303 after the evaluation of a full expression (one which is not part of a
2304 larger expression), after the evaluation of the first operand of a
2305 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
2306 function is called (but after the evaluation of its arguments and the
2307 expression denoting the called function), and in certain other places.
2308 Other than as expressed by the sequence point rules, the order of
2309 evaluation of subexpressions of an expression is not specified. All
2310 these rules describe only a partial order rather than a total order,
2311 since, for example, if two functions are called within one expression
2312 with no sequence point between them, the order in which the functions
2313 are called is not specified. However, the standards committee have
2314 ruled that function calls do not overlap.
2316 It is not specified when between sequence points modifications to the
2317 values of objects take effect. Programs whose behavior depends on this
2318 have undefined behavior; the C standard specifies that ``Between the
2319 previous and next sequence point an object shall have its stored value
2320 modified at most once by the evaluation of an expression. Furthermore,
2321 the prior value shall be read only to determine the value to be
2322 stored.''. If a program breaks these rules, the results on any
2323 particular implementation are entirely unpredictable.
2325 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
2326 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
2327 diagnosed by this option, and it may give an occasional false positive
2328 result, but in general it has been found fairly effective at detecting
2329 this sort of problem in programs.
2331 The present implementation of this option only works for C programs. A
2332 future implementation may also work for C++ programs.
2334 The C standard is worded confusingly, therefore there is some debate
2335 over the precise meaning of the sequence point rules in subtle cases.
2336 Links to discussions of the problem, including proposed formal
2337 definitions, may be found on the GCC readings page, at
2338 @w{@uref{http://gcc.gnu.org/readings.html}}.
2341 @opindex Wreturn-type
2342 Warn whenever a function is defined with a return-type that defaults to
2343 @code{int}. Also warn about any @code{return} statement with no
2344 return-value in a function whose return-type is not @code{void}.
2346 For C++, a function without return type always produces a diagnostic
2347 message, even when @option{-Wno-return-type} is specified. The only
2348 exceptions are @samp{main} and functions defined in system headers.
2352 Warn whenever a @code{switch} statement has an index of enumerated type
2353 and lacks a @code{case} for one or more of the named codes of that
2354 enumeration. (The presence of a @code{default} label prevents this
2355 warning.) @code{case} labels outside the enumeration range also
2356 provoke warnings when this option is used.
2358 @item -Wswitch-default
2359 @opindex Wswitch-switch
2360 Warn whenever a @code{switch} statement does not have a @code{default}
2364 @opindex Wswitch-enum
2365 Warn whenever a @code{switch} statement has an index of enumerated type
2366 and lacks a @code{case} for one or more of the named codes of that
2367 enumeration. @code{case} labels outside the enumeration range also
2368 provoke warnings when this option is used.
2372 Warn if any trigraphs are encountered that might change the meaning of
2373 the program (trigraphs within comments are not warned about).
2375 @item -Wunused-function
2376 @opindex Wunused-function
2377 Warn whenever a static function is declared but not defined or a
2378 non\-inline static function is unused.
2380 @item -Wunused-label
2381 @opindex Wunused-label
2382 Warn whenever a label is declared but not used.
2384 To suppress this warning use the @samp{unused} attribute
2385 (@pxref{Variable Attributes}).
2387 @item -Wunused-parameter
2388 @opindex Wunused-parameter
2389 Warn whenever a function parameter is unused aside from its declaration.
2391 To suppress this warning use the @samp{unused} attribute
2392 (@pxref{Variable Attributes}).
2394 @item -Wunused-variable
2395 @opindex Wunused-variable
2396 Warn whenever a local variable or non-constant static variable is unused
2397 aside from its declaration
2399 To suppress this warning use the @samp{unused} attribute
2400 (@pxref{Variable Attributes}).
2402 @item -Wunused-value
2403 @opindex Wunused-value
2404 Warn whenever a statement computes a result that is explicitly not used.
2406 To suppress this warning cast the expression to @samp{void}.
2410 All the above @option{-Wunused} options combined.
2412 In order to get a warning about an unused function parameter, you must
2413 either specify @samp{-Wextra -Wunused} (note that @samp{-Wall} implies
2414 @samp{-Wunused}), or separately specify @option{-Wunused-parameter}.
2416 @item -Wuninitialized
2417 @opindex Wuninitialized
2418 Warn if an automatic variable is used without first being initialized or
2419 if a variable may be clobbered by a @code{setjmp} call.
2421 These warnings are possible only in optimizing compilation,
2422 because they require data flow information that is computed only
2423 when optimizing. If you don't specify @option{-O}, you simply won't
2426 If you want to warn about code which uses the uninitialized value of the
2427 variable in its own initializer, use the @option{-Winit-self} option.
2429 These warnings occur only for variables that are candidates for
2430 register allocation. Therefore, they do not occur for a variable that
2431 is declared @code{volatile}, or whose address is taken, or whose size
2432 is other than 1, 2, 4 or 8 bytes. Also, they do not occur for
2433 structures, unions or arrays, even when they are in registers.
2435 Note that there may be no warning about a variable that is used only
2436 to compute a value that itself is never used, because such
2437 computations may be deleted by data flow analysis before the warnings
2440 These warnings are made optional because GCC is not smart
2441 enough to see all the reasons why the code might be correct
2442 despite appearing to have an error. Here is one example of how
2463 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
2464 always initialized, but GCC doesn't know this. Here is
2465 another common case:
2470 if (change_y) save_y = y, y = new_y;
2472 if (change_y) y = save_y;
2477 This has no bug because @code{save_y} is used only if it is set.
2479 @cindex @code{longjmp} warnings
2480 This option also warns when a non-volatile automatic variable might be
2481 changed by a call to @code{longjmp}. These warnings as well are possible
2482 only in optimizing compilation.
2484 The compiler sees only the calls to @code{setjmp}. It cannot know
2485 where @code{longjmp} will be called; in fact, a signal handler could
2486 call it at any point in the code. As a result, you may get a warning
2487 even when there is in fact no problem because @code{longjmp} cannot
2488 in fact be called at the place which would cause a problem.
2490 Some spurious warnings can be avoided if you declare all the functions
2491 you use that never return as @code{noreturn}. @xref{Function
2494 @item -Wunknown-pragmas
2495 @opindex Wunknown-pragmas
2496 @cindex warning for unknown pragmas
2497 @cindex unknown pragmas, warning
2498 @cindex pragmas, warning of unknown
2499 Warn when a #pragma directive is encountered which is not understood by
2500 GCC@. If this command line option is used, warnings will even be issued
2501 for unknown pragmas in system header files. This is not the case if
2502 the warnings were only enabled by the @option{-Wall} command line option.
2504 @item -Wstrict-aliasing
2505 @opindex Wstrict-aliasing
2506 This option is only active when @option{-fstrict-aliasing} is active.
2507 It warns about code which might break the strict aliasing rules that the
2508 compiler is using for optimization. The warning does not catch all
2509 cases, but does attempt to catch the more common pitfalls. It is
2510 included in @option{-Wall}.
2512 @item -Wstrict-aliasing=2
2513 @opindex Wstrict-aliasing=2
2514 This option is only active when @option{-fstrict-aliasing} is active.
2515 It warns about all code which might break the strict aliasing rules that the
2516 compiler is using for optimization. This warning catches all cases, but
2517 it will also give a warning for some ambiguous cases that are safe.
2521 All of the above @samp{-W} options combined. This enables all the
2522 warnings about constructions that some users consider questionable, and
2523 that are easy to avoid (or modify to prevent the warning), even in
2524 conjunction with macros. This also enables some language-specific
2525 warnings described in @ref{C++ Dialect Options} and
2526 @ref{Objective-C Dialect Options}.
2529 The following @option{-W@dots{}} options are not implied by @option{-Wall}.
2530 Some of them warn about constructions that users generally do not
2531 consider questionable, but which occasionally you might wish to check
2532 for; others warn about constructions that are necessary or hard to avoid
2533 in some cases, and there is no simple way to modify the code to suppress
2540 (This option used to be called @option{-W}. The older name is still
2541 supported, but the newer name is more descriptive.) Print extra warning
2542 messages for these events:
2546 A function can return either with or without a value. (Falling
2547 off the end of the function body is considered returning without
2548 a value.) For example, this function would evoke such a
2562 An expression-statement or the left-hand side of a comma expression
2563 contains no side effects.
2564 To suppress the warning, cast the unused expression to void.
2565 For example, an expression such as @samp{x[i,j]} will cause a warning,
2566 but @samp{x[(void)i,j]} will not.
2569 An unsigned value is compared against zero with @samp{<} or @samp{>=}.
2572 A comparison like @samp{x<=y<=z} appears; this is equivalent to
2573 @samp{(x<=y ? 1 : 0) <= z}, which is a different interpretation from
2574 that of ordinary mathematical notation.
2577 Storage-class specifiers like @code{static} are not the first things in
2578 a declaration. According to the C Standard, this usage is obsolescent.
2581 The return type of a function has a type qualifier such as @code{const}.
2582 Such a type qualifier has no effect, since the value returned by a
2583 function is not an lvalue. (But don't warn about the GNU extension of
2584 @code{volatile void} return types. That extension will be warned about
2585 if @option{-pedantic} is specified.)
2588 If @option{-Wall} or @option{-Wunused} is also specified, warn about unused
2592 A comparison between signed and unsigned values could produce an
2593 incorrect result when the signed value is converted to unsigned.
2594 (But don't warn if @option{-Wno-sign-compare} is also specified.)
2597 An aggregate has an initializer which does not initialize all members.
2598 For example, the following code would cause such a warning, because
2599 @code{x.h} would be implicitly initialized to zero:
2602 struct s @{ int f, g, h; @};
2603 struct s x = @{ 3, 4 @};
2607 A function parameter is declared without a type specifier in K&R-style
2615 An empty body occurs in an @samp{if} or @samp{else} statement.
2618 A pointer is compared against integer zero with @samp{<}, @samp{<=},
2619 @samp{>}, or @samp{>=}.
2622 A variable might be changed by @samp{longjmp} or @samp{vfork}.
2625 Any of several floating-point events that often indicate errors, such as
2626 overflow, underflow, loss of precision, etc.
2628 @item @r{(C++ only)}
2629 An enumerator and a non-enumerator both appear in a conditional expression.
2631 @item @r{(C++ only)}
2632 A non-static reference or non-static @samp{const} member appears in a
2633 class without constructors.
2635 @item @r{(C++ only)}
2636 Ambiguous virtual bases.
2638 @item @r{(C++ only)}
2639 Subscripting an array which has been declared @samp{register}.
2641 @item @r{(C++ only)}
2642 Taking the address of a variable which has been declared @samp{register}.
2644 @item @r{(C++ only)}
2645 A base class is not initialized in a derived class' copy constructor.
2648 @item -Wno-div-by-zero
2649 @opindex Wno-div-by-zero
2650 @opindex Wdiv-by-zero
2651 Do not warn about compile-time integer division by zero. Floating point
2652 division by zero is not warned about, as it can be a legitimate way of
2653 obtaining infinities and NaNs.
2655 @item -Wsystem-headers
2656 @opindex Wsystem-headers
2657 @cindex warnings from system headers
2658 @cindex system headers, warnings from
2659 Print warning messages for constructs found in system header files.
2660 Warnings from system headers are normally suppressed, on the assumption
2661 that they usually do not indicate real problems and would only make the
2662 compiler output harder to read. Using this command line option tells
2663 GCC to emit warnings from system headers as if they occurred in user
2664 code. However, note that using @option{-Wall} in conjunction with this
2665 option will @emph{not} warn about unknown pragmas in system
2666 headers---for that, @option{-Wunknown-pragmas} must also be used.
2669 @opindex Wfloat-equal
2670 Warn if floating point values are used in equality comparisons.
2672 The idea behind this is that sometimes it is convenient (for the
2673 programmer) to consider floating-point values as approximations to
2674 infinitely precise real numbers. If you are doing this, then you need
2675 to compute (by analyzing the code, or in some other way) the maximum or
2676 likely maximum error that the computation introduces, and allow for it
2677 when performing comparisons (and when producing output, but that's a
2678 different problem). In particular, instead of testing for equality, you
2679 would check to see whether the two values have ranges that overlap; and
2680 this is done with the relational operators, so equality comparisons are
2683 @item -Wtraditional @r{(C only)}
2684 @opindex Wtraditional
2685 Warn about certain constructs that behave differently in traditional and
2686 ISO C@. Also warn about ISO C constructs that have no traditional C
2687 equivalent, and/or problematic constructs which should be avoided.
2691 Macro parameters that appear within string literals in the macro body.
2692 In traditional C macro replacement takes place within string literals,
2693 but does not in ISO C@.
2696 In traditional C, some preprocessor directives did not exist.
2697 Traditional preprocessors would only consider a line to be a directive
2698 if the @samp{#} appeared in column 1 on the line. Therefore
2699 @option{-Wtraditional} warns about directives that traditional C
2700 understands but would ignore because the @samp{#} does not appear as the
2701 first character on the line. It also suggests you hide directives like
2702 @samp{#pragma} not understood by traditional C by indenting them. Some
2703 traditional implementations would not recognize @samp{#elif}, so it
2704 suggests avoiding it altogether.
2707 A function-like macro that appears without arguments.
2710 The unary plus operator.
2713 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating point
2714 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
2715 constants.) Note, these suffixes appear in macros defined in the system
2716 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
2717 Use of these macros in user code might normally lead to spurious
2718 warnings, however GCC's integrated preprocessor has enough context to
2719 avoid warning in these cases.
2722 A function declared external in one block and then used after the end of
2726 A @code{switch} statement has an operand of type @code{long}.
2729 A non-@code{static} function declaration follows a @code{static} one.
2730 This construct is not accepted by some traditional C compilers.
2733 The ISO type of an integer constant has a different width or
2734 signedness from its traditional type. This warning is only issued if
2735 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
2736 typically represent bit patterns, are not warned about.
2739 Usage of ISO string concatenation is detected.
2742 Initialization of automatic aggregates.
2745 Identifier conflicts with labels. Traditional C lacks a separate
2746 namespace for labels.
2749 Initialization of unions. If the initializer is zero, the warning is
2750 omitted. This is done under the assumption that the zero initializer in
2751 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
2752 initializer warnings and relies on default initialization to zero in the
2756 Conversions by prototypes between fixed/floating point values and vice
2757 versa. The absence of these prototypes when compiling with traditional
2758 C would cause serious problems. This is a subset of the possible
2759 conversion warnings, for the full set use @option{-Wconversion}.
2762 Use of ISO C style function definitions. This warning intentionally is
2763 @emph{not} issued for prototype declarations or variadic functions
2764 because these ISO C features will appear in your code when using
2765 libiberty's traditional C compatibility macros, @code{PARAMS} and
2766 @code{VPARAMS}. This warning is also bypassed for nested functions
2767 because that feature is already a GCC extension and thus not relevant to
2768 traditional C compatibility.
2771 @item -Wdeclaration-after-statement @r{(C only)}
2772 @opindex Wdeclaration-after-statement
2773 Warn when a declaration is found after a statement in a block. This
2774 construct, known from C++, was introduced with ISO C99 and is by default
2775 allowed in GCC@. It is not supported by ISO C90 and was not supported by
2776 GCC versions before GCC 3.0. @xref{Mixed Declarations}.
2780 Warn if an undefined identifier is evaluated in an @samp{#if} directive.
2782 @item -Wendif-labels
2783 @opindex Wendif-labels
2784 Warn whenever an @samp{#else} or an @samp{#endif} are followed by text.
2788 Warn whenever a local variable shadows another local variable, parameter or
2789 global variable or whenever a built-in function is shadowed.
2791 @item -Wlarger-than-@var{len}
2792 @opindex Wlarger-than
2793 Warn whenever an object of larger than @var{len} bytes is defined.
2795 @item -Wpointer-arith
2796 @opindex Wpointer-arith
2797 Warn about anything that depends on the ``size of'' a function type or
2798 of @code{void}. GNU C assigns these types a size of 1, for
2799 convenience in calculations with @code{void *} pointers and pointers
2802 @item -Wbad-function-cast @r{(C only)}
2803 @opindex Wbad-function-cast
2804 Warn whenever a function call is cast to a non-matching type.
2805 For example, warn if @code{int malloc()} is cast to @code{anything *}.
2809 Warn whenever a pointer is cast so as to remove a type qualifier from
2810 the target type. For example, warn if a @code{const char *} is cast
2811 to an ordinary @code{char *}.
2814 @opindex Wcast-align
2815 Warn whenever a pointer is cast such that the required alignment of the
2816 target is increased. For example, warn if a @code{char *} is cast to
2817 an @code{int *} on machines where integers can only be accessed at
2818 two- or four-byte boundaries.
2820 @item -Wwrite-strings
2821 @opindex Wwrite-strings
2822 When compiling C, give string constants the type @code{const
2823 char[@var{length}]} so that
2824 copying the address of one into a non-@code{const} @code{char *}
2825 pointer will get a warning; when compiling C++, warn about the
2826 deprecated conversion from string constants to @code{char *}.
2827 These warnings will help you find at
2828 compile time code that can try to write into a string constant, but
2829 only if you have been very careful about using @code{const} in
2830 declarations and prototypes. Otherwise, it will just be a nuisance;
2831 this is why we did not make @option{-Wall} request these warnings.
2834 @opindex Wconversion
2835 Warn if a prototype causes a type conversion that is different from what
2836 would happen to the same argument in the absence of a prototype. This
2837 includes conversions of fixed point to floating and vice versa, and
2838 conversions changing the width or signedness of a fixed point argument
2839 except when the same as the default promotion.
2841 Also, warn if a negative integer constant expression is implicitly
2842 converted to an unsigned type. For example, warn about the assignment
2843 @code{x = -1} if @code{x} is unsigned. But do not warn about explicit
2844 casts like @code{(unsigned) -1}.
2846 @item -Wsign-compare
2847 @opindex Wsign-compare
2848 @cindex warning for comparison of signed and unsigned values
2849 @cindex comparison of signed and unsigned values, warning
2850 @cindex signed and unsigned values, comparison warning
2851 Warn when a comparison between signed and unsigned values could produce
2852 an incorrect result when the signed value is converted to unsigned.
2853 This warning is also enabled by @option{-Wextra}; to get the other warnings
2854 of @option{-Wextra} without this warning, use @samp{-Wextra -Wno-sign-compare}.
2856 @item -Waggregate-return
2857 @opindex Waggregate-return
2858 Warn if any functions that return structures or unions are defined or
2859 called. (In languages where you can return an array, this also elicits
2862 @item -Wstrict-prototypes @r{(C only)}
2863 @opindex Wstrict-prototypes
2864 Warn if a function is declared or defined without specifying the
2865 argument types. (An old-style function definition is permitted without
2866 a warning if preceded by a declaration which specifies the argument
2869 @item -Wold-style-definition @r{(C only)}
2870 @opindex Wold-style-definition
2871 Warn if an old-style function definition is used. A warning is given
2872 even if there is a previous prototype.
2874 @item -Wmissing-prototypes @r{(C only)}
2875 @opindex Wmissing-prototypes
2876 Warn if a global function is defined without a previous prototype
2877 declaration. This warning is issued even if the definition itself
2878 provides a prototype. The aim is to detect global functions that fail
2879 to be declared in header files.
2881 @item -Wmissing-declarations @r{(C only)}
2882 @opindex Wmissing-declarations
2883 Warn if a global function is defined without a previous declaration.
2884 Do so even if the definition itself provides a prototype.
2885 Use this option to detect global functions that are not declared in
2888 @item -Wmissing-noreturn
2889 @opindex Wmissing-noreturn
2890 Warn about functions which might be candidates for attribute @code{noreturn}.
2891 Note these are only possible candidates, not absolute ones. Care should
2892 be taken to manually verify functions actually do not ever return before
2893 adding the @code{noreturn} attribute, otherwise subtle code generation
2894 bugs could be introduced. You will not get a warning for @code{main} in
2895 hosted C environments.
2897 @item -Wmissing-format-attribute
2898 @opindex Wmissing-format-attribute
2900 If @option{-Wformat} is enabled, also warn about functions which might be
2901 candidates for @code{format} attributes. Note these are only possible
2902 candidates, not absolute ones. GCC will guess that @code{format}
2903 attributes might be appropriate for any function that calls a function
2904 like @code{vprintf} or @code{vscanf}, but this might not always be the
2905 case, and some functions for which @code{format} attributes are
2906 appropriate may not be detected. This option has no effect unless
2907 @option{-Wformat} is enabled (possibly by @option{-Wall}).
2909 @item -Wno-multichar
2910 @opindex Wno-multichar
2912 Do not warn if a multicharacter constant (@samp{'FOOF'}) is used.
2913 Usually they indicate a typo in the user's code, as they have
2914 implementation-defined values, and should not be used in portable code.
2916 @item -Wno-deprecated-declarations
2917 @opindex Wno-deprecated-declarations
2918 Do not warn about uses of functions, variables, and types marked as
2919 deprecated by using the @code{deprecated} attribute.
2920 (@pxref{Function Attributes}, @pxref{Variable Attributes},
2921 @pxref{Type Attributes}.)
2925 Warn if a structure is given the packed attribute, but the packed
2926 attribute has no effect on the layout or size of the structure.
2927 Such structures may be mis-aligned for little benefit. For
2928 instance, in this code, the variable @code{f.x} in @code{struct bar}
2929 will be misaligned even though @code{struct bar} does not itself
2930 have the packed attribute:
2937 @} __attribute__((packed));
2947 Warn if padding is included in a structure, either to align an element
2948 of the structure or to align the whole structure. Sometimes when this
2949 happens it is possible to rearrange the fields of the structure to
2950 reduce the padding and so make the structure smaller.
2952 @item -Wredundant-decls
2953 @opindex Wredundant-decls
2954 Warn if anything is declared more than once in the same scope, even in
2955 cases where multiple declaration is valid and changes nothing.
2957 @item -Wnested-externs @r{(C only)}
2958 @opindex Wnested-externs
2959 Warn if an @code{extern} declaration is encountered within a function.
2961 @item -Wunreachable-code
2962 @opindex Wunreachable-code
2963 Warn if the compiler detects that code will never be executed.
2965 This option is intended to warn when the compiler detects that at
2966 least a whole line of source code will never be executed, because
2967 some condition is never satisfied or because it is after a
2968 procedure that never returns.
2970 It is possible for this option to produce a warning even though there
2971 are circumstances under which part of the affected line can be executed,
2972 so care should be taken when removing apparently-unreachable code.
2974 For instance, when a function is inlined, a warning may mean that the
2975 line is unreachable in only one inlined copy of the function.
2977 This option is not made part of @option{-Wall} because in a debugging
2978 version of a program there is often substantial code which checks
2979 correct functioning of the program and is, hopefully, unreachable
2980 because the program does work. Another common use of unreachable
2981 code is to provide behavior which is selectable at compile-time.
2985 Warn if a function can not be inlined and it was declared as inline.
2986 Even with this option, the compiler will not warn about failures to
2987 inline functions declared in system headers.
2989 The compiler uses a variety of heuristics to determine whether or not
2990 to inline a function. For example, the compiler takes into account
2991 the size of the function being inlined and the the amount of inlining
2992 that has already been done in the current function. Therefore,
2993 seemingly insignificant changes in the source program can cause the
2994 warnings produced by @option{-Winline} to appear or disappear.
2996 @item -Wno-invalid-offsetof @r{(C++ only)}
2997 @opindex Wno-invalid-offsetof
2998 Suppress warnings from applying the @samp{offsetof} macro to a non-POD
2999 type. According to the 1998 ISO C++ standard, applying @samp{offsetof}
3000 to a non-POD type is undefined. In existing C++ implementations,
3001 however, @samp{offsetof} typically gives meaningful results even when
3002 applied to certain kinds of non-POD types. (Such as a simple
3003 @samp{struct} that fails to be a POD type only by virtue of having a
3004 constructor.) This flag is for users who are aware that they are
3005 writing nonportable code and who have deliberately chosen to ignore the
3008 The restrictions on @samp{offsetof} may be relaxed in a future version
3009 of the C++ standard.
3012 @opindex Winvalid-pch
3013 Warn if a precompiled header (@pxref{Precompiled Headers}) is found in
3014 the search path but can't be used.
3018 @opindex Wno-long-long
3019 Warn if @samp{long long} type is used. This is default. To inhibit
3020 the warning messages, use @option{-Wno-long-long}. Flags
3021 @option{-Wlong-long} and @option{-Wno-long-long} are taken into account
3022 only when @option{-pedantic} flag is used.
3024 @item -Wvariadic-macros
3025 @opindex Wvariadic-macros
3026 @opindex Wno-variadic-macros
3027 Warn if variadic macros are used in pedantic ISO C90 mode, or the GNU
3028 alternate syntax when in pedantic ISO C99 mode. This is default.
3029 To inhibit the warning messages, use @option{-Wno-variadic-macros}.
3031 @item -Wdisabled-optimization
3032 @opindex Wdisabled-optimization
3033 Warn if a requested optimization pass is disabled. This warning does
3034 not generally indicate that there is anything wrong with your code; it
3035 merely indicates that GCC's optimizers were unable to handle the code
3036 effectively. Often, the problem is that your code is too big or too
3037 complex; GCC will refuse to optimize programs when the optimization
3038 itself is likely to take inordinate amounts of time.
3042 Make all warnings into errors.
3045 @node Debugging Options
3046 @section Options for Debugging Your Program or GCC
3047 @cindex options, debugging
3048 @cindex debugging information options
3050 GCC has various special options that are used for debugging
3051 either your program or GCC:
3056 Produce debugging information in the operating system's native format
3057 (stabs, COFF, XCOFF, or DWARF)@. GDB can work with this debugging
3060 On most systems that use stabs format, @option{-g} enables use of extra
3061 debugging information that only GDB can use; this extra information
3062 makes debugging work better in GDB but will probably make other debuggers
3064 refuse to read the program. If you want to control for certain whether
3065 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
3066 @option{-gxcoff+}, @option{-gxcoff}, or @option{-gvms} (see below).
3068 Unlike most other C compilers, GCC allows you to use @option{-g} with
3069 @option{-O}. The shortcuts taken by optimized code may occasionally
3070 produce surprising results: some variables you declared may not exist
3071 at all; flow of control may briefly move where you did not expect it;
3072 some statements may not be executed because they compute constant
3073 results or their values were already at hand; some statements may
3074 execute in different places because they were moved out of loops.
3076 Nevertheless it proves possible to debug optimized output. This makes
3077 it reasonable to use the optimizer for programs that might have bugs.
3079 The following options are useful when GCC is generated with the
3080 capability for more than one debugging format.
3084 Produce debugging information for use by GDB@. This means to use the
3085 most expressive format available (DWARF 2, stabs, or the native format
3086 if neither of those are supported), including GDB extensions if at all
3091 Produce debugging information in stabs format (if that is supported),
3092 without GDB extensions. This is the format used by DBX on most BSD
3093 systems. On MIPS, Alpha and System V Release 4 systems this option
3094 produces stabs debugging output which is not understood by DBX or SDB@.
3095 On System V Release 4 systems this option requires the GNU assembler.
3097 @item -feliminate-unused-debug-symbols
3098 @opindex feliminate-unused-debug-symbols
3099 Produce debugging information in stabs format (if that is supported),
3100 for only symbols that are actually used.
3104 Produce debugging information in stabs format (if that is supported),
3105 using GNU extensions understood only by the GNU debugger (GDB)@. The
3106 use of these extensions is likely to make other debuggers crash or
3107 refuse to read the program.
3111 Produce debugging information in COFF format (if that is supported).
3112 This is the format used by SDB on most System V systems prior to
3117 Produce debugging information in XCOFF format (if that is supported).
3118 This is the format used by the DBX debugger on IBM RS/6000 systems.
3122 Produce debugging information in XCOFF format (if that is supported),
3123 using GNU extensions understood only by the GNU debugger (GDB)@. The
3124 use of these extensions is likely to make other debuggers crash or
3125 refuse to read the program, and may cause assemblers other than the GNU
3126 assembler (GAS) to fail with an error.
3130 Produce debugging information in DWARF version 2 format (if that is
3131 supported). This is the format used by DBX on IRIX 6.
3135 Produce debugging information in VMS debug format (if that is
3136 supported). This is the format used by DEBUG on VMS systems.
3139 @itemx -ggdb@var{level}
3140 @itemx -gstabs@var{level}
3141 @itemx -gcoff@var{level}
3142 @itemx -gxcoff@var{level}
3143 @itemx -gvms@var{level}
3144 Request debugging information and also use @var{level} to specify how
3145 much information. The default level is 2.
3147 Level 1 produces minimal information, enough for making backtraces in
3148 parts of the program that you don't plan to debug. This includes
3149 descriptions of functions and external variables, but no information
3150 about local variables and no line numbers.
3152 Level 3 includes extra information, such as all the macro definitions
3153 present in the program. Some debuggers support macro expansion when
3154 you use @option{-g3}.
3156 Note that in order to avoid confusion between DWARF1 debug level 2,
3157 and DWARF2 @option{-gdwarf-2} does not accept a concatenated debug
3158 level. Instead use an additional @option{-g@var{level}} option to
3159 change the debug level for DWARF2.
3161 @item -feliminate-dwarf2-dups
3162 @opindex feliminate-dwarf2-dups
3163 Compress DWARF2 debugging information by eliminating duplicated
3164 information about each symbol. This option only makes sense when
3165 generating DWARF2 debugging information with @option{-gdwarf-2}.
3167 @cindex @command{prof}
3170 Generate extra code to write profile information suitable for the
3171 analysis program @command{prof}. You must use this option when compiling
3172 the source files you want data about, and you must also use it when
3175 @cindex @command{gprof}
3178 Generate extra code to write profile information suitable for the
3179 analysis program @command{gprof}. You must use this option when compiling
3180 the source files you want data about, and you must also use it when
3185 Makes the compiler print out each function name as it is compiled, and
3186 print some statistics about each pass when it finishes.
3189 @opindex ftime-report
3190 Makes the compiler print some statistics about the time consumed by each
3191 pass when it finishes.
3194 @opindex fmem-report
3195 Makes the compiler print some statistics about permanent memory
3196 allocation when it finishes.
3198 @item -fprofile-arcs
3199 @opindex fprofile-arcs
3200 Add code so that program flow @dfn{arcs} are instrumented. During
3201 execution the program records how many times each branch and call is
3202 executed and how many times it is taken or returns. When the compiled
3203 program exits it saves this data to a file called
3204 @file{@var{auxname}.gcda} for each source file. The data may be used for
3205 profile-directed optimizations (@option{-fbranch-probabilities}), or for
3206 test coverage analysis (@option{-ftest-coverage}). Each object file's
3207 @var{auxname} is generated from the name of the output file, if
3208 explicitly specified and it is not the final executable, otherwise it is
3209 the basename of the source file. In both cases any suffix is removed
3210 (e.g. @file{foo.gcda} for input file @file{dir/foo.c}, or
3211 @file{dir/foo.gcda} for output file specified as @option{-o dir/foo.o}).
3216 Compile the source files with @option{-fprofile-arcs} plus optimization
3217 and code generation options. For test coverage analysis, use the
3218 additional @option{-ftest-coverage} option. You do not need to profile
3219 every source file in a program.
3222 Link your object files with @option{-lgcov} or @option{-fprofile-arcs}
3223 (the latter implies the former).
3226 Run the program on a representative workload to generate the arc profile
3227 information. This may be repeated any number of times. You can run
3228 concurrent instances of your program, and provided that the file system
3229 supports locking, the data files will be correctly updated. Also
3230 @code{fork} calls are detected and correctly handled (double counting
3234 For profile-directed optimizations, compile the source files again with
3235 the same optimization and code generation options plus
3236 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
3237 Control Optimization}).
3240 For test coverage analysis, use @command{gcov} to produce human readable
3241 information from the @file{.gcno} and @file{.gcda} files. Refer to the
3242 @command{gcov} documentation for further information.
3246 With @option{-fprofile-arcs}, for each function of your program GCC
3247 creates a program flow graph, then finds a spanning tree for the graph.
3248 Only arcs that are not on the spanning tree have to be instrumented: the
3249 compiler adds code to count the number of times that these arcs are
3250 executed. When an arc is the only exit or only entrance to a block, the
3251 instrumentation code can be added to the block; otherwise, a new basic
3252 block must be created to hold the instrumentation code.
3254 @item -ftree-based-profiling
3255 @opindex ftree-based-profiling
3256 This option is used in addition to @option{-fprofile-arcs} or
3257 @option{-fbranch-probabilities} to control whether those optimizations
3258 are performed on a tree-based or rtl-based internal representation.
3259 If you use this option when compiling with @option{-fprofile-arcs},
3260 you must also use it when compiling later with @option{-fbranch-probabilities}.
3261 Currently the tree-based optimization is in an early stage of
3262 development, and this option is recommended only for those people
3263 working on improving it.
3266 @item -ftest-coverage
3267 @opindex ftest-coverage
3268 Produce a notes file that the @command{gcov} code-coverage utility
3269 (@pxref{Gcov,, @command{gcov}---a Test Coverage Program}) can use to
3270 show program coverage. Each source file's note file is called
3271 @file{@var{auxname}.gcno}. Refer to the @option{-fprofile-arcs} option
3272 above for a description of @var{auxname} and instructions on how to
3273 generate test coverage data. Coverage data will match the source files
3274 more closely, if you do not optimize.
3276 @item -d@var{letters}
3278 Says to make debugging dumps during compilation at times specified by
3279 @var{letters}. This is used for debugging the compiler. The file names
3280 for most of the dumps are made by appending a pass number and a word to
3281 the @var{dumpname}. @var{dumpname} is generated from the name of the
3282 output file, if explicitly specified and it is not an executable,
3283 otherwise it is the basename of the source file. In both cases any
3284 suffix is removed (e.g. @file{foo.01.rtl} or @file{foo.02.sibling}).
3285 Here are the possible letters for use in @var{letters}, and their
3291 Annotate the assembler output with miscellaneous debugging information.
3294 Dump after computing branch probabilities, to @file{@var{file}.12.bp}.
3297 Dump after block reordering, to @file{@var{file}.32.bbro}.
3300 Dump after instruction combination, to the file @file{@var{file}.20.combine}.
3303 Dump after the first if conversion, to the file @file{@var{file}.14.ce1}.
3304 Also dump after the second if conversion, to the file @file{@var{file}.21.ce2}.
3307 Dump after branch target load optimization, to to @file{@var{file}.33.btl}.
3308 Also dump after delayed branch scheduling, to @file{@var{file}.37.dbr}.
3311 Dump all macro definitions, at the end of preprocessing, in addition to
3315 Dump after the third if conversion, to @file{@var{file}.31.ce3}.
3318 Dump after control and data flow analysis, to @file{@var{file}.11.cfg}.
3319 Also dump after life analysis, to @file{@var{file}.19.life}.
3322 Dump after global register allocation, to @file{@var{file}.26.greg}.
3325 Dump after GCSE, to @file{@var{file}.08.gcse}.
3326 Also dump after jump bypassing and control flow optimizations, to
3327 @file{@var{file}.10.bypass}.
3330 Dump after finalization of EH handling code, to @file{@var{file}.03.eh}.
3333 Dump after sibling call optimizations, to @file{@var{file}.02.sibling}.
3336 Dump after the first jump optimization, to @file{@var{file}.04.jump}.
3339 Dump after conversion from registers to stack, to @file{@var{file}.35.stack}.
3342 Dump after local register allocation, to @file{@var{file}.25.lreg}.
3345 Dump after loop optimization passes, to @file{@var{file}.09.loop} and
3346 @file{@var{file}.16.loop2}.
3349 Dump after modulo scheduling, to @file{@var{file}.23.sms}.
3352 Dump after performing the machine dependent reorganization pass, to
3353 @file{@var{file}.36.mach}.
3356 Dump after register renumbering, to @file{@var{file}.30.rnreg}.
3359 Dump after the register move pass, to @file{@var{file}.22.regmove}.
3362 Dump after post-reload optimizations, to @file{@var{file}.27.postreload}.
3365 Dump after RTL generation, to @file{@var{file}.01.rtl}.
3368 Dump after the second scheduling pass, to @file{@var{file}.34.sched2}.
3371 Dump after CSE (including the jump optimization that sometimes follows
3372 CSE), to @file{@var{file}.06.cse}.
3375 Dump after the first scheduling pass, to @file{@var{file}.24.sched}.
3378 Dump after the second CSE pass (including the jump optimization that
3379 sometimes follows CSE), to @file{@var{file}.18.cse2}.
3382 Dump after running tracer, to @file{@var{file}.15.tracer}.
3385 Dump after null pointer elimination pass to @file{@var{file}.05.null}.
3388 Dump callgraph and unit-at-a-time optimization @file{@var{file}.00.unit}.
3391 Dump after the value profile transformations, to @file{@var{file}.13.vpt}.
3392 Also dump after variable tracking, to @file{@var{file}.35.vartrack}.
3395 Dump after the second flow pass, to @file{@var{file}.28.flow2}.
3398 Dump after the peephole pass, to @file{@var{file}.29.peephole2}.
3401 Dump after constructing the web, to @file{@var{file}.17.web}.
3404 Produce all the dumps listed above.
3407 Produce a core dump whenever an error occurs.
3410 Print statistics on memory usage, at the end of the run, to
3414 Annotate the assembler output with a comment indicating which
3415 pattern and alternative was used. The length of each instruction is
3419 Dump the RTL in the assembler output as a comment before each instruction.
3420 Also turns on @option{-dp} annotation.
3423 For each of the other indicated dump files (except for
3424 @file{@var{file}.01.rtl}), dump a representation of the control flow graph
3425 suitable for viewing with VCG to @file{@var{file}.@var{pass}.vcg}.
3428 Just generate RTL for a function instead of compiling it. Usually used
3432 Dump debugging information during parsing, to standard error.
3435 @item -fdump-unnumbered
3436 @opindex fdump-unnumbered
3437 When doing debugging dumps (see @option{-d} option above), suppress instruction
3438 numbers and line number note output. This makes it more feasible to
3439 use diff on debugging dumps for compiler invocations with different
3440 options, in particular with and without @option{-g}.
3442 @item -fdump-translation-unit @r{(C and C++ only)}
3443 @itemx -fdump-translation-unit-@var{options} @r{(C and C++ only)}
3444 @opindex fdump-translation-unit
3445 Dump a representation of the tree structure for the entire translation
3446 unit to a file. The file name is made by appending @file{.tu} to the
3447 source file name. If the @samp{-@var{options}} form is used, @var{options}
3448 controls the details of the dump as described for the
3449 @option{-fdump-tree} options.
3451 @item -fdump-class-hierarchy @r{(C++ only)}
3452 @itemx -fdump-class-hierarchy-@var{options} @r{(C++ only)}
3453 @opindex fdump-class-hierarchy
3454 Dump a representation of each class's hierarchy and virtual function
3455 table layout to a file. The file name is made by appending @file{.class}
3456 to the source file name. If the @samp{-@var{options}} form is used,
3457 @var{options} controls the details of the dump as described for the
3458 @option{-fdump-tree} options.
3460 @item -fdump-tree-@var{switch} @r{(C and C++ only)}
3461 @itemx -fdump-tree-@var{switch}-@var{options} @r{(C and C++ only)}
3463 Control the dumping at various stages of processing the intermediate
3464 language tree to a file. The file name is generated by appending a switch
3465 specific suffix to the source file name. If the @samp{-@var{options}}
3466 form is used, @var{options} is a list of @samp{-} separated options that
3467 control the details of the dump. Not all options are applicable to all
3468 dumps, those which are not meaningful will be ignored. The following
3469 options are available
3473 Print the address of each node. Usually this is not meaningful as it
3474 changes according to the environment and source file. Its primary use
3475 is for tying up a dump file with a debug environment.
3477 Inhibit dumping of members of a scope or body of a function merely
3478 because that scope has been reached. Only dump such items when they
3479 are directly reachable by some other path. When dumping pretty-printed
3480 trees, this option inhibits dumping the bodies of control structures.
3482 Print a raw representation of the tree. By default, trees are
3483 pretty-printed into a C-like representation.
3485 Enable more detailed dumps (not honored by every dump option).
3487 Enable dumping various statistics about the pass (not honored by every dump
3490 Enable showing basic block boundaries (disabled in raw dumps).
3492 Enable showing virtual operands for every statement.
3494 Enable showing line numbers for statements.
3496 Enable showing the unique ID (@code{DECL_UID}) for each variable.
3498 Turn on all options, except @option{raw}, @option{slim} and @option{lineno}.
3501 The following tree dumps are possible:
3505 Dump before any tree based optimization, to @file{@var{file}.original}.
3508 Dump after all tree based optimization, to @file{@var{file}.optimized}.
3511 Dump after function inlining, to @file{@var{file}.inlined}.
3514 @opindex fdump-tree-gimple
3515 Dump each function before and after the gimplification pass to a file. The
3516 file name is made by appending @file{.gimple} to the source file name.
3519 @opindex fdump-tree-cfg
3520 Dump the control flow graph of each function to a file. The file name is
3521 made by appending @file{.cfg} to the source file name.
3524 @opindex fdump-tree-vcg
3525 Dump the control flow graph of each function to a file in VCG format. The
3526 file name is made by appending @file{.vcg} to the source file name. Note
3527 that if the file contains more than one function, the generated file cannot
3528 be used directly by VCG. You will need to cut and paste each function's
3529 graph into its own separate file first.
3532 @opindex fdump-tree-ch
3533 Dump each function after copying loop headers. The file name is made by
3534 appending @file{.ch} to the source file name.
3537 @opindex fdump-tree-ssa
3538 Dump SSA related information to a file. The file name is made by appending
3539 @file{.ssa} to the source file name.
3542 @opindex fdump-tree-alias
3543 Dump aliasing information for each function. The file name is made by
3544 appending @file{.alias} to the source file name.
3547 @opindex fdump-tree-ccp
3548 Dump each function after CCP. The file name is made by appending
3549 @file{.ccp} to the source file name.
3552 @opindex fdump-tree-pre
3553 Dump trees after partial redundancy elimination. The file name is made
3554 by appending @file{.pre} to the source file name.
3557 @opindex fdump-tree-fre
3558 Dump trees after full redundancy elimination. The file name is made
3559 by appending @file{.fre} to the source file name.
3562 @opindex fdump-tree-dce
3563 Dump each function after dead code elimination. The file name is made by
3564 appending @file{.dce} to the source file name.
3567 @opindex fdump-tree-mudflap
3568 Dump each function after adding mudflap instrumentation. The file name is
3569 made by appending @file{.mudflap} to the source file name.
3572 @opindex fdump-tree-sra
3573 Dump each function after performing scalar replacement of aggregates. The
3574 file name is made by appending @file{.sra} to the source file name.
3577 @opindex fdump-tree-dom
3578 Dump each function after applying dominator tree optimizations. The file
3579 name is made by appending @file{.dom} to the source file name.
3582 @opindex fdump-tree-dse
3583 Dump each function after applying dead store elimination. The file
3584 name is made by appending @file{.dse} to the source file name.
3587 @opindex fdump-tree-phiopt
3588 Dump each function after optimizing PHI nodes into straightline code. The file
3589 name is made by appending @file{.phiopt} to the source file name.
3592 @opindex fdump-tree-forwprop
3593 Dump each function after forward propagating single use variables. The file
3594 name is made by appending @file{.forwprop} to the source file name.
3597 @opindex fdump-tree-copyrename
3598 Dump each function after applying the copy rename optimization. The file
3599 name is made by appending @file{.copyrename} to the source file name.
3602 @opindex fdump-tree-nrv
3603 Dump each function after applying the named return value optimization on
3604 generic trees. The file name is made by appending @file{.nrv} to the source
3608 @opindex fdump-tree-all
3609 Enable all the available tree dumps with the flags provided in this option.
3612 @item -frandom-seed=@var{string}
3613 @opindex frandom-string
3614 This option provides a seed that GCC uses when it would otherwise use
3615 random numbers. It is used to generate certain symbol names
3616 that have to be different in every compiled file. It is also used to
3617 place unique stamps in coverage data files and the object files that
3618 produce them. You can use the @option{-frandom-seed} option to produce
3619 reproducibly identical object files.
3621 The @var{string} should be different for every file you compile.
3623 @item -fsched-verbose=@var{n}
3624 @opindex fsched-verbose
3625 On targets that use instruction scheduling, this option controls the
3626 amount of debugging output the scheduler prints. This information is
3627 written to standard error, unless @option{-dS} or @option{-dR} is
3628 specified, in which case it is output to the usual dump
3629 listing file, @file{.sched} or @file{.sched2} respectively. However
3630 for @var{n} greater than nine, the output is always printed to standard
3633 For @var{n} greater than zero, @option{-fsched-verbose} outputs the
3634 same information as @option{-dRS}. For @var{n} greater than one, it
3635 also output basic block probabilities, detailed ready list information
3636 and unit/insn info. For @var{n} greater than two, it includes RTL
3637 at abort point, control-flow and regions info. And for @var{n} over
3638 four, @option{-fsched-verbose} also includes dependence info.
3642 Store the usual ``temporary'' intermediate files permanently; place them
3643 in the current directory and name them based on the source file. Thus,
3644 compiling @file{foo.c} with @samp{-c -save-temps} would produce files
3645 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
3646 preprocessed @file{foo.i} output file even though the compiler now
3647 normally uses an integrated preprocessor.
3651 Report the CPU time taken by each subprocess in the compilation
3652 sequence. For C source files, this is the compiler proper and assembler
3653 (plus the linker if linking is done). The output looks like this:
3660 The first number on each line is the ``user time,'' that is time spent
3661 executing the program itself. The second number is ``system time,''
3662 time spent executing operating system routines on behalf of the program.
3663 Both numbers are in seconds.
3665 @item -fvar-tracking
3666 @opindex fvar-tracking
3667 Run variable tracking pass. It computes where variables are stored at each
3668 position in code. Better debugging information is then generated
3669 (if the debugging information format supports this information).
3671 It is enabled by default when compiling with optimization (@option{-Os},
3672 @option{-O}, @option{-O2}, ...), debugging information (@option{-g}) and
3673 the debug info format supports it.
3675 @item -print-file-name=@var{library}
3676 @opindex print-file-name
3677 Print the full absolute name of the library file @var{library} that
3678 would be used when linking---and don't do anything else. With this
3679 option, GCC does not compile or link anything; it just prints the
3682 @item -print-multi-directory
3683 @opindex print-multi-directory
3684 Print the directory name corresponding to the multilib selected by any
3685 other switches present in the command line. This directory is supposed
3686 to exist in @env{GCC_EXEC_PREFIX}.
3688 @item -print-multi-lib
3689 @opindex print-multi-lib
3690 Print the mapping from multilib directory names to compiler switches
3691 that enable them. The directory name is separated from the switches by
3692 @samp{;}, and each switch starts with an @samp{@@} instead of the
3693 @samp{-}, without spaces between multiple switches. This is supposed to
3694 ease shell-processing.
3696 @item -print-prog-name=@var{program}
3697 @opindex print-prog-name
3698 Like @option{-print-file-name}, but searches for a program such as @samp{cpp}.
3700 @item -print-libgcc-file-name
3701 @opindex print-libgcc-file-name
3702 Same as @option{-print-file-name=libgcc.a}.
3704 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
3705 but you do want to link with @file{libgcc.a}. You can do
3708 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
3711 @item -print-search-dirs
3712 @opindex print-search-dirs
3713 Print the name of the configured installation directory and a list of
3714 program and library directories @command{gcc} will search---and don't do anything else.
3716 This is useful when @command{gcc} prints the error message
3717 @samp{installation problem, cannot exec cpp0: No such file or directory}.
3718 To resolve this you either need to put @file{cpp0} and the other compiler
3719 components where @command{gcc} expects to find them, or you can set the environment
3720 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
3721 Don't forget the trailing '/'.
3722 @xref{Environment Variables}.
3725 @opindex dumpmachine
3726 Print the compiler's target machine (for example,
3727 @samp{i686-pc-linux-gnu})---and don't do anything else.
3730 @opindex dumpversion
3731 Print the compiler version (for example, @samp{3.0})---and don't do
3736 Print the compiler's built-in specs---and don't do anything else. (This
3737 is used when GCC itself is being built.) @xref{Spec Files}.
3739 @item -feliminate-unused-debug-types
3740 @opindex feliminate-unused-debug-types
3741 Normally, when producing DWARF2 output, GCC will emit debugging
3742 information for all types declared in a compilation
3743 unit, regardless of whether or not they are actually used
3744 in that compilation unit. Sometimes this is useful, such as
3745 if, in the debugger, you want to cast a value to a type that is
3746 not actually used in your program (but is declared). More often,
3747 however, this results in a significant amount of wasted space.
3748 With this option, GCC will avoid producing debug symbol output
3749 for types that are nowhere used in the source file being compiled.
3752 @node Optimize Options
3753 @section Options That Control Optimization
3754 @cindex optimize options
3755 @cindex options, optimization
3757 These options control various sorts of optimizations.
3759 Without any optimization option, the compiler's goal is to reduce the
3760 cost of compilation and to make debugging produce the expected
3761 results. Statements are independent: if you stop the program with a
3762 breakpoint between statements, you can then assign a new value to any
3763 variable or change the program counter to any other statement in the
3764 function and get exactly the results you would expect from the source
3767 Turning on optimization flags makes the compiler attempt to improve
3768 the performance and/or code size at the expense of compilation time
3769 and possibly the ability to debug the program.
3771 The compiler performs optimization based on the knowledge it has of
3772 the program. Optimization levels @option{-O2} and above, in
3773 particular, enable @emph{unit-at-a-time} mode, which allows the
3774 compiler to consider information gained from later functions in
3775 the file when compiling a function. Compiling multiple files at
3776 once to a single output file in @emph{unit-at-a-time} mode allows
3777 the compiler to use information gained from all of the files when
3778 compiling each of them.
3780 Not all optimizations are controlled directly by a flag. Only
3781 optimizations that have a flag are listed.
3788 Optimize. Optimizing compilation takes somewhat more time, and a lot
3789 more memory for a large function.
3791 With @option{-O}, the compiler tries to reduce code size and execution
3792 time, without performing any optimizations that take a great deal of
3795 @option{-O} turns on the following optimization flags:
3796 @gccoptlist{-fdefer-pop @gol
3797 -fmerge-constants @gol
3799 -floop-optimize @gol
3800 -fif-conversion @gol
3801 -fif-conversion2 @gol
3802 -fdelayed-branch @gol
3803 -fguess-branch-probability @gol
3806 @option{-O} also turns on @option{-fomit-frame-pointer} on machines
3807 where doing so does not interfere with debugging.
3811 Optimize even more. GCC performs nearly all supported optimizations
3812 that do not involve a space-speed tradeoff. The compiler does not
3813 perform loop unrolling or function inlining when you specify @option{-O2}.
3814 As compared to @option{-O}, this option increases both compilation time
3815 and the performance of the generated code.
3817 @option{-O2} turns on all optimization flags specified by @option{-O}. It
3818 also turns on the following optimization flags:
3819 @gccoptlist{-fforce-mem @gol
3820 -foptimize-sibling-calls @gol
3821 -fstrength-reduce @gol
3822 -fcse-follow-jumps -fcse-skip-blocks @gol
3823 -frerun-cse-after-loop -frerun-loop-opt @gol
3824 -fgcse -fgcse-lm -fgcse-sm -fgcse-las @gol
3825 -fdelete-null-pointer-checks @gol
3826 -fexpensive-optimizations @gol
3828 -fschedule-insns -fschedule-insns2 @gol
3829 -fsched-interblock -fsched-spec @gol
3832 -freorder-blocks -freorder-functions @gol
3833 -fstrict-aliasing @gol
3834 -funit-at-a-time @gol
3835 -falign-functions -falign-jumps @gol
3836 -falign-loops -falign-labels @gol
3839 Please note the warning under @option{-fgcse} about
3840 invoking @option{-O2} on programs that use computed gotos.
3844 Optimize yet more. @option{-O3} turns on all optimizations specified by
3845 @option{-O2} and also turns on the @option{-finline-functions},
3846 @option{-fweb} and @option{-fgcse-after-reload} options.
3850 Do not optimize. This is the default.
3854 Optimize for size. @option{-Os} enables all @option{-O2} optimizations that
3855 do not typically increase code size. It also performs further
3856 optimizations designed to reduce code size.
3858 @option{-Os} disables the following optimization flags:
3859 @gccoptlist{-falign-functions -falign-jumps -falign-loops @gol
3860 -falign-labels -freorder-blocks -freorder-blocks-and-partition -fprefetch-loop-arrays}
3862 If you use multiple @option{-O} options, with or without level numbers,
3863 the last such option is the one that is effective.
3866 Options of the form @option{-f@var{flag}} specify machine-independent
3867 flags. Most flags have both positive and negative forms; the negative
3868 form of @option{-ffoo} would be @option{-fno-foo}. In the table
3869 below, only one of the forms is listed---the one you typically will
3870 use. You can figure out the other form by either removing @samp{no-}
3873 The following options control specific optimizations. They are either
3874 activated by @option{-O} options or are related to ones that are. You
3875 can use the following flags in the rare cases when ``fine-tuning'' of
3876 optimizations to be performed is desired.
3879 @item -fno-default-inline
3880 @opindex fno-default-inline
3881 Do not make member functions inline by default merely because they are
3882 defined inside the class scope (C++ only). Otherwise, when you specify
3883 @w{@option{-O}}, member functions defined inside class scope are compiled
3884 inline by default; i.e., you don't need to add @samp{inline} in front of
3885 the member function name.
3887 @item -fno-defer-pop
3888 @opindex fno-defer-pop
3889 Always pop the arguments to each function call as soon as that function
3890 returns. For machines which must pop arguments after a function call,
3891 the compiler normally lets arguments accumulate on the stack for several
3892 function calls and pops them all at once.
3894 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
3898 Force memory operands to be copied into registers before doing
3899 arithmetic on them. This produces better code by making all memory
3900 references potential common subexpressions. When they are not common
3901 subexpressions, instruction combination should eliminate the separate
3904 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3907 @opindex fforce-addr
3908 Force memory address constants to be copied into registers before
3909 doing arithmetic on them. This may produce better code just as
3910 @option{-fforce-mem} may.
3912 @item -fomit-frame-pointer
3913 @opindex fomit-frame-pointer
3914 Don't keep the frame pointer in a register for functions that
3915 don't need one. This avoids the instructions to save, set up and
3916 restore frame pointers; it also makes an extra register available
3917 in many functions. @strong{It also makes debugging impossible on
3920 On some machines, such as the VAX, this flag has no effect, because
3921 the standard calling sequence automatically handles the frame pointer
3922 and nothing is saved by pretending it doesn't exist. The
3923 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
3924 whether a target machine supports this flag. @xref{Registers,,Register
3925 Usage, gccint, GNU Compiler Collection (GCC) Internals}.
3927 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
3929 @item -foptimize-sibling-calls
3930 @opindex foptimize-sibling-calls
3931 Optimize sibling and tail recursive calls.
3933 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3937 Don't pay attention to the @code{inline} keyword. Normally this option
3938 is used to keep the compiler from expanding any functions inline.
3939 Note that if you are not optimizing, no functions can be expanded inline.
3941 @item -finline-functions
3942 @opindex finline-functions
3943 Integrate all simple functions into their callers. The compiler
3944 heuristically decides which functions are simple enough to be worth
3945 integrating in this way.
3947 If all calls to a given function are integrated, and the function is
3948 declared @code{static}, then the function is normally not output as
3949 assembler code in its own right.
3951 Enabled at level @option{-O3}.
3953 @item -finline-limit=@var{n}
3954 @opindex finline-limit
3955 By default, GCC limits the size of functions that can be inlined. This flag
3956 allows the control of this limit for functions that are explicitly marked as
3957 inline (i.e., marked with the inline keyword or defined within the class
3958 definition in c++). @var{n} is the size of functions that can be inlined in
3959 number of pseudo instructions (not counting parameter handling). The default
3960 value of @var{n} is 600.
3961 Increasing this value can result in more inlined code at
3962 the cost of compilation time and memory consumption. Decreasing usually makes
3963 the compilation faster and less code will be inlined (which presumably
3964 means slower programs). This option is particularly useful for programs that
3965 use inlining heavily such as those based on recursive templates with C++.
3967 Inlining is actually controlled by a number of parameters, which may be
3968 specified individually by using @option{--param @var{name}=@var{value}}.
3969 The @option{-finline-limit=@var{n}} option sets some of these parameters
3973 @item max-inline-insns-single
3974 is set to @var{n}/2.
3975 @item max-inline-insns-auto
3976 is set to @var{n}/2.
3977 @item min-inline-insns
3978 is set to 130 or @var{n}/4, whichever is smaller.
3979 @item max-inline-insns-rtl
3983 See below for a documentation of the individual
3984 parameters controlling inlining.
3986 @emph{Note:} pseudo instruction represents, in this particular context, an
3987 abstract measurement of function's size. In no way, it represents a count
3988 of assembly instructions and as such its exact meaning might change from one
3989 release to an another.
3991 @item -fkeep-inline-functions
3992 @opindex fkeep-inline-functions
3993 Even if all calls to a given function are integrated, and the function
3994 is declared @code{static}, nevertheless output a separate run-time
3995 callable version of the function. This switch does not affect
3996 @code{extern inline} functions.
3998 @item -fkeep-static-consts
3999 @opindex fkeep-static-consts
4000 Emit variables declared @code{static const} when optimization isn't turned
4001 on, even if the variables aren't referenced.
4003 GCC enables this option by default. If you want to force the compiler to
4004 check if the variable was referenced, regardless of whether or not
4005 optimization is turned on, use the @option{-fno-keep-static-consts} option.
4007 @item -fmerge-constants
4008 Attempt to merge identical constants (string constants and floating point
4009 constants) across compilation units.
4011 This option is the default for optimized compilation if the assembler and
4012 linker support it. Use @option{-fno-merge-constants} to inhibit this
4015 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4017 @item -fmerge-all-constants
4018 Attempt to merge identical constants and identical variables.
4020 This option implies @option{-fmerge-constants}. In addition to
4021 @option{-fmerge-constants} this considers e.g. even constant initialized
4022 arrays or initialized constant variables with integral or floating point
4023 types. Languages like C or C++ require each non-automatic variable to
4024 have distinct location, so using this option will result in non-conforming
4027 @item -fmodulo-sched
4028 @opindex fmodulo-sched
4029 Perform swing modulo scheduling immediately before the first scheduling
4030 pass. This pass looks at innermost loops and reorders their
4031 instructions by overlapping different iterations.
4035 Use a graph coloring register allocator. Currently this option is meant
4036 only for testing. Users should not specify this option, since it is not
4037 yet ready for production use.
4039 @item -fno-branch-count-reg
4040 @opindex fno-branch-count-reg
4041 Do not use ``decrement and branch'' instructions on a count register,
4042 but instead generate a sequence of instructions that decrement a
4043 register, compare it against zero, then branch based upon the result.
4044 This option is only meaningful on architectures that support such
4045 instructions, which include x86, PowerPC, IA-64 and S/390.
4047 The default is @option{-fbranch-count-reg}, enabled when
4048 @option{-fstrength-reduce} is enabled.
4050 @item -fno-function-cse
4051 @opindex fno-function-cse
4052 Do not put function addresses in registers; make each instruction that
4053 calls a constant function contain the function's address explicitly.
4055 This option results in less efficient code, but some strange hacks
4056 that alter the assembler output may be confused by the optimizations
4057 performed when this option is not used.
4059 The default is @option{-ffunction-cse}
4061 @item -fno-zero-initialized-in-bss
4062 @opindex fno-zero-initialized-in-bss
4063 If the target supports a BSS section, GCC by default puts variables that
4064 are initialized to zero into BSS@. This can save space in the resulting
4067 This option turns off this behavior because some programs explicitly
4068 rely on variables going to the data section. E.g., so that the
4069 resulting executable can find the beginning of that section and/or make
4070 assumptions based on that.
4072 The default is @option{-fzero-initialized-in-bss}.
4074 @item -fbounds-check
4075 @opindex fbounds-check
4076 For front-ends that support it, generate additional code to check that
4077 indices used to access arrays are within the declared range. This is
4078 currently only supported by the Java and Fortran front-ends, where
4079 this option defaults to true and false respectively.
4081 @item -fmudflap -fmudflapth -fmudflapir
4085 @cindex bounds checking
4087 For front-ends that support it (C and C++), instrument all risky
4088 pointer/array dereferencing operations, some standard library
4089 string/heap functions, and some other associated constructs with
4090 range/validity tests. Modules so instrumented should be immune to
4091 buffer overflows, invalid heap use, and some other classes of C/C++
4092 programming errors. The instrumentation relies on a separate runtime
4093 library (@file{libmudflap}), which will be linked into a program if
4094 @option{-fmudflap} is given at link time. Run-time behavior of the
4095 instrumented program is controlled by the @env{MUDFLAP_OPTIONS}
4096 environment variable. See @code{env MUDFLAP_OPTIONS=-help a.out}
4099 Use @option{-fmudflapth} instead of @option{-fmudflap} to compile and to
4100 link if your program is multi-threaded. Use @option{-fmudflapir}, in
4101 addition to @option{-fmudflap} or @option{-fmudflapth}, if
4102 instrumentation should ignore pointer reads. This produces less
4103 instrumentation (and therefore faster execution) and still provides
4104 some protection against outright memory corrupting writes, but allows
4105 erroneously read data to propagate within a program.
4107 @item -fstrength-reduce
4108 @opindex fstrength-reduce
4109 Perform the optimizations of loop strength reduction and
4110 elimination of iteration variables.
4112 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4114 @item -fthread-jumps
4115 @opindex fthread-jumps
4116 Perform optimizations where we check to see if a jump branches to a
4117 location where another comparison subsumed by the first is found. If
4118 so, the first branch is redirected to either the destination of the
4119 second branch or a point immediately following it, depending on whether
4120 the condition is known to be true or false.
4122 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4124 @item -fcse-follow-jumps
4125 @opindex fcse-follow-jumps
4126 In common subexpression elimination, scan through jump instructions
4127 when the target of the jump is not reached by any other path. For
4128 example, when CSE encounters an @code{if} statement with an
4129 @code{else} clause, CSE will follow the jump when the condition
4132 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4134 @item -fcse-skip-blocks
4135 @opindex fcse-skip-blocks
4136 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
4137 follow jumps which conditionally skip over blocks. When CSE
4138 encounters a simple @code{if} statement with no else clause,
4139 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
4140 body of the @code{if}.
4142 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4144 @item -frerun-cse-after-loop
4145 @opindex frerun-cse-after-loop
4146 Re-run common subexpression elimination after loop optimizations has been
4149 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4151 @item -frerun-loop-opt
4152 @opindex frerun-loop-opt
4153 Run the loop optimizer twice.
4155 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4159 Perform a global common subexpression elimination pass.
4160 This pass also performs global constant and copy propagation.
4162 @emph{Note:} When compiling a program using computed gotos, a GCC
4163 extension, you may get better runtime performance if you disable
4164 the global common subexpression elimination pass by adding
4165 @option{-fno-gcse} to the command line.
4167 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4171 When @option{-fgcse-lm} is enabled, global common subexpression elimination will
4172 attempt to move loads which are only killed by stores into themselves. This
4173 allows a loop containing a load/store sequence to be changed to a load outside
4174 the loop, and a copy/store within the loop.
4176 Enabled by default when gcse is enabled.
4180 When @option{-fgcse-sm} is enabled, a store motion pass is run after
4181 global common subexpression elimination. This pass will attempt to move
4182 stores out of loops. When used in conjunction with @option{-fgcse-lm},
4183 loops containing a load/store sequence can be changed to a load before
4184 the loop and a store after the loop.
4186 Enabled by default when gcse is enabled.
4190 When @option{-fgcse-las} is enabled, the global common subexpression
4191 elimination pass eliminates redundant loads that come after stores to the
4192 same memory location (both partial and full redundancies).
4194 Enabled by default when gcse is enabled.
4196 @item -fgcse-after-reload
4197 @opindex fgcse-after-reload
4198 When @option{-fgcse-after-reload} is enabled, a redundant load elimination
4199 pass is performed after reload. The purpose of this pass is to cleanup
4202 @item -floop-optimize
4203 @opindex floop-optimize
4204 Perform loop optimizations: move constant expressions out of loops, simplify
4205 exit test conditions and optionally do strength-reduction and loop unrolling as
4208 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4210 @item -floop-optimize2
4211 @opindex floop-optimize2
4212 Perform loop optimizations using the new loop optimizer. The optimizations
4213 (loop unrolling, peeling and unswitching, loop invariant motion) are enabled
4216 @item -fcrossjumping
4217 @opindex crossjumping
4218 Perform cross-jumping transformation. This transformation unifies equivalent code and save code size. The
4219 resulting code may or may not perform better than without cross-jumping.
4221 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4223 @item -fif-conversion
4224 @opindex if-conversion
4225 Attempt to transform conditional jumps into branch-less equivalents. This
4226 include use of conditional moves, min, max, set flags and abs instructions, and
4227 some tricks doable by standard arithmetics. The use of conditional execution
4228 on chips where it is available is controlled by @code{if-conversion2}.
4230 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4232 @item -fif-conversion2
4233 @opindex if-conversion2
4234 Use conditional execution (where available) to transform conditional jumps into
4235 branch-less equivalents.
4237 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4239 @item -fdelete-null-pointer-checks
4240 @opindex fdelete-null-pointer-checks
4241 Use global dataflow analysis to identify and eliminate useless checks
4242 for null pointers. The compiler assumes that dereferencing a null
4243 pointer would have halted the program. If a pointer is checked after
4244 it has already been dereferenced, it cannot be null.
4246 In some environments, this assumption is not true, and programs can
4247 safely dereference null pointers. Use
4248 @option{-fno-delete-null-pointer-checks} to disable this optimization
4249 for programs which depend on that behavior.
4251 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4253 @item -fexpensive-optimizations
4254 @opindex fexpensive-optimizations
4255 Perform a number of minor optimizations that are relatively expensive.
4257 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4259 @item -foptimize-register-move
4261 @opindex foptimize-register-move
4263 Attempt to reassign register numbers in move instructions and as
4264 operands of other simple instructions in order to maximize the amount of
4265 register tying. This is especially helpful on machines with two-operand
4268 Note @option{-fregmove} and @option{-foptimize-register-move} are the same
4271 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4273 @item -fdelayed-branch
4274 @opindex fdelayed-branch
4275 If supported for the target machine, attempt to reorder instructions
4276 to exploit instruction slots available after delayed branch
4279 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4281 @item -fschedule-insns
4282 @opindex fschedule-insns
4283 If supported for the target machine, attempt to reorder instructions to
4284 eliminate execution stalls due to required data being unavailable. This
4285 helps machines that have slow floating point or memory load instructions
4286 by allowing other instructions to be issued until the result of the load
4287 or floating point instruction is required.
4289 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4291 @item -fschedule-insns2
4292 @opindex fschedule-insns2
4293 Similar to @option{-fschedule-insns}, but requests an additional pass of
4294 instruction scheduling after register allocation has been done. This is
4295 especially useful on machines with a relatively small number of
4296 registers and where memory load instructions take more than one cycle.
4298 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4300 @item -fno-sched-interblock
4301 @opindex fno-sched-interblock
4302 Don't schedule instructions across basic blocks. This is normally
4303 enabled by default when scheduling before register allocation, i.e.@:
4304 with @option{-fschedule-insns} or at @option{-O2} or higher.
4306 @item -fno-sched-spec
4307 @opindex fno-sched-spec
4308 Don't allow speculative motion of non-load instructions. This is normally
4309 enabled by default when scheduling before register allocation, i.e.@:
4310 with @option{-fschedule-insns} or at @option{-O2} or higher.
4312 @item -fsched-spec-load
4313 @opindex fsched-spec-load
4314 Allow speculative motion of some load instructions. This only makes
4315 sense when scheduling before register allocation, i.e.@: with
4316 @option{-fschedule-insns} or at @option{-O2} or higher.
4318 @item -fsched-spec-load-dangerous
4319 @opindex fsched-spec-load-dangerous
4320 Allow speculative motion of more load instructions. This only makes
4321 sense when scheduling before register allocation, i.e.@: with
4322 @option{-fschedule-insns} or at @option{-O2} or higher.
4324 @item -fsched-stalled-insns=@var{n}
4325 @opindex fsched-stalled-insns
4326 Define how many insns (if any) can be moved prematurely from the queue
4327 of stalled insns into the ready list, during the second scheduling pass.
4329 @item -fsched-stalled-insns-dep=@var{n}
4330 @opindex fsched-stalled-insns-dep
4331 Define how many insn groups (cycles) will be examined for a dependency
4332 on a stalled insn that is candidate for premature removal from the queue
4333 of stalled insns. Has an effect only during the second scheduling pass,
4334 and only if @option{-fsched-stalled-insns} is used and its value is not zero.
4336 @item -fsched2-use-superblocks
4337 @opindex fsched2-use-superblocks
4338 When scheduling after register allocation, do use superblock scheduling
4339 algorithm. Superblock scheduling allows motion across basic block boundaries
4340 resulting on faster schedules. This option is experimental, as not all machine
4341 descriptions used by GCC model the CPU closely enough to avoid unreliable
4342 results from the algorithm.
4344 This only makes sense when scheduling after register allocation, i.e.@: with
4345 @option{-fschedule-insns2} or at @option{-O2} or higher.
4347 @item -fsched2-use-traces
4348 @opindex fsched2-use-traces
4349 Use @option{-fsched2-use-superblocks} algorithm when scheduling after register
4350 allocation and additionally perform code duplication in order to increase the
4351 size of superblocks using tracer pass. See @option{-ftracer} for details on
4354 This mode should produce faster but significantly longer programs. Also
4355 without @code{-fbranch-probabilities} the traces constructed may not match the
4356 reality and hurt the performance. This only makes
4357 sense when scheduling after register allocation, i.e.@: with
4358 @option{-fschedule-insns2} or at @option{-O2} or higher.
4360 @item -fcaller-saves
4361 @opindex fcaller-saves
4362 Enable values to be allocated in registers that will be clobbered by
4363 function calls, by emitting extra instructions to save and restore the
4364 registers around such calls. Such allocation is done only when it
4365 seems to result in better code than would otherwise be produced.
4367 This option is always enabled by default on certain machines, usually
4368 those which have no call-preserved registers to use instead.
4370 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4373 Perform Partial Redundancy Elimination (PRE) on trees. This flag is
4374 enabled by default at -O and higher.
4377 Perform Full Redundancy Elimination (FRE) on trees. The difference
4378 between FRE and PRE is that FRE only considers expressions
4379 that are computed on all paths leading to the redundant computation.
4380 This analysis faster than PRE, though it exposes fewer redundancies.
4381 This flag is enabled by default at -O and higher.
4384 Perform sparse conditional constant propagation (CCP) on trees. This flag
4385 is enabled by default at -O and higher.
4388 Perform dead code elimination (DCE) on trees. This flag is enabled by
4389 default at -O and higher.
4391 @item -ftree-dominator-opts
4392 Perform dead code elimination (DCE) on trees. This flag is enabled by
4393 default at -O and higher.
4396 Perform loop header copying on trees. This is beneficial since it increases
4397 effectivity of code motion optimizations. It also saves one jump. This flag
4398 is enabled by default at -O and higher. It is not enabled for -Os, since it
4399 usually increases code size.
4401 @item -ftree-loop-optimize
4402 Perform loop optimizations on trees. This flag is enabled by default at -O
4406 Perform loop invariant motion on trees. This pass moves only invartiants that
4407 would be hard to handle on rtl level (function calls, operations that expand to
4408 nontrivial sequences of insns). With @option{-funswitch-loops} it also moves
4409 operands of conditions that are invariant out of the loop, so that we can use
4410 just trivial invariantness analysis in loop unswitching. The pass also includes
4414 Perform scalar replacement of aggregates. This pass replaces structure
4415 references with scalars to prevent committing structures to memory too
4416 early. This flag is enabled by default at -O and higher.
4418 @item -ftree-copyrename
4419 Perform copy renaming on trees. This pass attempts to rename compiler
4420 temporaries to other variables at copy locations, usually resulting in
4421 variable names which more closely resemble the original variables. This flag
4422 is enabled by default at -O and higher.
4425 Perform temporary expression replacement during the SSA->normal phase. Single
4426 use/single def temporaries are replaced at their use location with their
4427 defining expression. This results in non-GIMPLE code, but gives the expanders
4428 much more complex trees to work on resulting in better RTL generation. This is
4429 enabled by default at -O and higher.
4432 Perform live range splitting during the SSA->normal phase. Distinct live
4433 ranges of a variable are split into unique variables, allowing for better
4434 optimization later. This is enabled by default at -O and higher.
4438 Perform tail duplication to enlarge superblock size. This transformation
4439 simplifies the control flow of the function allowing other optimizations to do
4442 @item -funroll-loops
4443 @opindex funroll-loops
4444 Unroll loops whose number of iterations can be determined at compile
4445 time or upon entry to the loop. @option{-funroll-loops} implies both
4446 @option{-fstrength-reduce} and @option{-frerun-cse-after-loop}. This
4447 option makes code larger, and may or may not make it run faster.
4449 @item -funroll-all-loops
4450 @opindex funroll-all-loops
4451 Unroll all loops, even if their number of iterations is uncertain when
4452 the loop is entered. This usually makes programs run more slowly.
4453 @option{-funroll-all-loops} implies the same options as
4454 @option{-funroll-loops},
4456 @item -fprefetch-loop-arrays
4457 @opindex fprefetch-loop-arrays
4458 If supported by the target machine, generate instructions to prefetch
4459 memory to improve the performance of loops that access large arrays.
4461 @item -fmove-all-movables
4462 @opindex fmove-all-movables
4463 Forces all invariant computations in loops to be moved
4466 @item -freduce-all-givs
4467 @opindex freduce-all-givs
4468 Forces all general-induction variables in loops to be
4471 @emph{Note:} When compiling programs written in Fortran,
4472 @option{-fmove-all-movables} and @option{-freduce-all-givs} are enabled
4473 by default when you use the optimizer.
4475 These options may generate better or worse code; results are highly
4476 dependent on the structure of loops within the source code.
4478 These two options are intended to be removed someday, once
4479 they have helped determine the efficacy of various
4480 approaches to improving loop optimizations.
4482 Please contact @w{@email{gcc@@gcc.gnu.org}}, and describe how use of
4483 these options affects the performance of your production code.
4484 Examples of code that runs @emph{slower} when these options are
4485 @emph{enabled} are very valuable.
4488 @itemx -fno-peephole2
4489 @opindex fno-peephole
4490 @opindex fno-peephole2
4491 Disable any machine-specific peephole optimizations. The difference
4492 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
4493 are implemented in the compiler; some targets use one, some use the
4494 other, a few use both.
4496 @option{-fpeephole} is enabled by default.
4497 @option{-fpeephole2} enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4499 @item -fno-guess-branch-probability
4500 @opindex fno-guess-branch-probability
4501 Do not guess branch probabilities using a randomized model.
4503 Sometimes GCC will opt to use a randomized model to guess branch
4504 probabilities, when none are available from either profiling feedback
4505 (@option{-fprofile-arcs}) or @samp{__builtin_expect}. This means that
4506 different runs of the compiler on the same program may produce different
4509 In a hard real-time system, people don't want different runs of the
4510 compiler to produce code that has different behavior; minimizing
4511 non-determinism is of paramount import. This switch allows users to
4512 reduce non-determinism, possibly at the expense of inferior
4515 The default is @option{-fguess-branch-probability} at levels
4516 @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4518 @item -freorder-blocks
4519 @opindex freorder-blocks
4520 Reorder basic blocks in the compiled function in order to reduce number of
4521 taken branches and improve code locality.
4523 Enabled at levels @option{-O2}, @option{-O3}.
4525 @item -freorder-blocks-and-partition
4526 @opindex freorder-blocks-and-partition
4527 In addition to reordering basic blocks in the compiled function, in order
4528 to reduce number of taken branches, partitions hot and cold basic blocks
4529 into separate sections of the assembly and .o files, to improve
4530 paging and cache locality performance.
4532 @item -freorder-functions
4533 @opindex freorder-functions
4534 Reorder basic blocks in the compiled function in order to reduce number of
4535 taken branches and improve code locality. This is implemented by using special
4536 subsections @code{.text.hot} for most frequently executed functions and
4537 @code{.text.unlikely} for unlikely executed functions. Reordering is done by
4538 the linker so object file format must support named sections and linker must
4539 place them in a reasonable way.
4541 Also profile feedback must be available in to make this option effective. See
4542 @option{-fprofile-arcs} for details.
4544 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4546 @item -fstrict-aliasing
4547 @opindex fstrict-aliasing
4548 Allows the compiler to assume the strictest aliasing rules applicable to
4549 the language being compiled. For C (and C++), this activates
4550 optimizations based on the type of expressions. In particular, an
4551 object of one type is assumed never to reside at the same address as an
4552 object of a different type, unless the types are almost the same. For
4553 example, an @code{unsigned int} can alias an @code{int}, but not a
4554 @code{void*} or a @code{double}. A character type may alias any other
4557 Pay special attention to code like this:
4570 The practice of reading from a different union member than the one most
4571 recently written to (called ``type-punning'') is common. Even with
4572 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
4573 is accessed through the union type. So, the code above will work as
4574 expected. However, this code might not:
4585 Every language that wishes to perform language-specific alias analysis
4586 should define a function that computes, given an @code{tree}
4587 node, an alias set for the node. Nodes in different alias sets are not
4588 allowed to alias. For an example, see the C front-end function
4589 @code{c_get_alias_set}.
4591 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4593 @item -falign-functions
4594 @itemx -falign-functions=@var{n}
4595 @opindex falign-functions
4596 Align the start of functions to the next power-of-two greater than
4597 @var{n}, skipping up to @var{n} bytes. For instance,
4598 @option{-falign-functions=32} aligns functions to the next 32-byte
4599 boundary, but @option{-falign-functions=24} would align to the next
4600 32-byte boundary only if this can be done by skipping 23 bytes or less.
4602 @option{-fno-align-functions} and @option{-falign-functions=1} are
4603 equivalent and mean that functions will not be aligned.
4605 Some assemblers only support this flag when @var{n} is a power of two;
4606 in that case, it is rounded up.
4608 If @var{n} is not specified or is zero, use a machine-dependent default.
4610 Enabled at levels @option{-O2}, @option{-O3}.
4612 @item -falign-labels
4613 @itemx -falign-labels=@var{n}
4614 @opindex falign-labels
4615 Align all branch targets to a power-of-two boundary, skipping up to
4616 @var{n} bytes like @option{-falign-functions}. This option can easily
4617 make code slower, because it must insert dummy operations for when the
4618 branch target is reached in the usual flow of the code.
4620 @option{-fno-align-labels} and @option{-falign-labels=1} are
4621 equivalent and mean that labels will not be aligned.
4623 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
4624 are greater than this value, then their values are used instead.
4626 If @var{n} is not specified or is zero, use a machine-dependent default
4627 which is very likely to be @samp{1}, meaning no alignment.
4629 Enabled at levels @option{-O2}, @option{-O3}.
4632 @itemx -falign-loops=@var{n}
4633 @opindex falign-loops
4634 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
4635 like @option{-falign-functions}. The hope is that the loop will be
4636 executed many times, which will make up for any execution of the dummy
4639 @option{-fno-align-loops} and @option{-falign-loops=1} are
4640 equivalent and mean that loops will not be aligned.
4642 If @var{n} is not specified or is zero, use a machine-dependent default.
4644 Enabled at levels @option{-O2}, @option{-O3}.
4647 @itemx -falign-jumps=@var{n}
4648 @opindex falign-jumps
4649 Align branch targets to a power-of-two boundary, for branch targets
4650 where the targets can only be reached by jumping, skipping up to @var{n}
4651 bytes like @option{-falign-functions}. In this case, no dummy operations
4654 @option{-fno-align-jumps} and @option{-falign-jumps=1} are
4655 equivalent and mean that loops will not be aligned.
4657 If @var{n} is not specified or is zero, use a machine-dependent default.
4659 Enabled at levels @option{-O2}, @option{-O3}.
4661 @item -funit-at-a-time
4662 @opindex funit-at-a-time
4663 Parse the whole compilation unit before starting to produce code.
4664 This allows some extra optimizations to take place but consumes
4665 more memory (in general). There are some compatibility issues
4666 with @emph{unit-at-at-time} mode:
4669 enabling @emph{unit-at-a-time} mode may change the order
4670 in which functions, variables, and top-level @code{asm} statements
4671 are emitted, and will likely break code relying on some particular
4672 ordering. The majority of such top-level @code{asm} statements,
4673 though, can be replaced by @code{section} attributes.
4676 @emph{unit-at-a-time} mode removes unreferenced static variables
4677 and functions are removed. This may result in undefined references
4678 when an @code{asm} statement refers directly to variables or functions
4679 that are otherwise unused. In that case either the variable/function
4680 shall be listed as an operand of the @code{asm} statement operand or,
4681 in the case of top-level @code{asm} statements the attribute @code{used}
4682 shall be used on the declaration.
4685 Static functions now can use non-standard passing conventions that
4686 may break @code{asm} statements calling functions directly. Again,
4687 attribute @code{used} will prevent this behavior.
4690 As a temporary workaround, @option{-fno-unit-at-a-time} can be used,
4691 but this scheme may not be supported by future releases of GCC.
4693 Enabled at levels @option{-O2}, @option{-O3}.
4697 Constructs webs as commonly used for register allocation purposes and assign
4698 each web individual pseudo register. This allows the register allocation pass
4699 to operate on pseudos directly, but also strengthens several other optimization
4700 passes, such as CSE, loop optimizer and trivial dead code remover. It can,
4701 however, make debugging impossible, since variables will no longer stay in a
4704 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os},
4705 on targets where the default format for debugging information supports
4708 @item -fno-cprop-registers
4709 @opindex fno-cprop-registers
4710 After register allocation and post-register allocation instruction splitting,
4711 we perform a copy-propagation pass to try to reduce scheduling dependencies
4712 and occasionally eliminate the copy.
4714 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4716 @item -fprofile-generate
4717 @opindex fprofile-generate
4719 Enable options usually used for instrumenting application to produce
4720 profile useful for later recompilation with profile feedback based
4721 optimization. You must use @code{-fprofile-generate} both when
4722 compiling and when linking your program.
4724 The following options are enabled: @code{-fprofile-arcs}, @code{-fprofile-values}, @code{-fvpt}.
4727 @opindex fprofile-use
4728 Enable profile feedback directed optimizations, and optimizations
4729 generally profitable only with profile feedback available.
4731 The following options are enabled: @code{-fbranch-probabilities},
4732 @code{-fvpt}, @code{-funroll-loops}, @code{-fpeel-loops}, @code{-ftracer}.
4736 The following options control compiler behavior regarding floating
4737 point arithmetic. These options trade off between speed and
4738 correctness. All must be specifically enabled.
4742 @opindex ffloat-store
4743 Do not store floating point variables in registers, and inhibit other
4744 options that might change whether a floating point value is taken from a
4747 @cindex floating point precision
4748 This option prevents undesirable excess precision on machines such as
4749 the 68000 where the floating registers (of the 68881) keep more
4750 precision than a @code{double} is supposed to have. Similarly for the
4751 x86 architecture. For most programs, the excess precision does only
4752 good, but a few programs rely on the precise definition of IEEE floating
4753 point. Use @option{-ffloat-store} for such programs, after modifying
4754 them to store all pertinent intermediate computations into variables.
4758 Sets @option{-fno-math-errno}, @option{-funsafe-math-optimizations}, @*
4759 @option{-fno-trapping-math}, @option{-ffinite-math-only},
4760 @option{-fno-rounding-math} and @option{-fno-signaling-nans}.
4762 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
4764 This option should never be turned on by any @option{-O} option since
4765 it can result in incorrect output for programs which depend on
4766 an exact implementation of IEEE or ISO rules/specifications for
4769 @item -fno-math-errno
4770 @opindex fno-math-errno
4771 Do not set ERRNO after calling math functions that are executed
4772 with a single instruction, e.g., sqrt. A program that relies on
4773 IEEE exceptions for math error handling may want to use this flag
4774 for speed while maintaining IEEE arithmetic compatibility.
4776 This option should never be turned on by any @option{-O} option since
4777 it can result in incorrect output for programs which depend on
4778 an exact implementation of IEEE or ISO rules/specifications for
4781 The default is @option{-fmath-errno}.
4783 @item -funsafe-math-optimizations
4784 @opindex funsafe-math-optimizations
4785 Allow optimizations for floating-point arithmetic that (a) assume
4786 that arguments and results are valid and (b) may violate IEEE or
4787 ANSI standards. When used at link-time, it may include libraries
4788 or startup files that change the default FPU control word or other
4789 similar optimizations.
4791 This option should never be turned on by any @option{-O} option since
4792 it can result in incorrect output for programs which depend on
4793 an exact implementation of IEEE or ISO rules/specifications for
4796 The default is @option{-fno-unsafe-math-optimizations}.
4798 @item -ffinite-math-only
4799 @opindex ffinite-math-only
4800 Allow optimizations for floating-point arithmetic that assume
4801 that arguments and results are not NaNs or +-Infs.
4803 This option should never be turned on by any @option{-O} option since
4804 it can result in incorrect output for programs which depend on
4805 an exact implementation of IEEE or ISO rules/specifications.
4807 The default is @option{-fno-finite-math-only}.
4809 @item -fno-trapping-math
4810 @opindex fno-trapping-math
4811 Compile code assuming that floating-point operations cannot generate
4812 user-visible traps. These traps include division by zero, overflow,
4813 underflow, inexact result and invalid operation. This option implies
4814 @option{-fno-signaling-nans}. Setting this option may allow faster
4815 code if one relies on ``non-stop'' IEEE arithmetic, for example.
4817 This option should never be turned on by any @option{-O} option since
4818 it can result in incorrect output for programs which depend on
4819 an exact implementation of IEEE or ISO rules/specifications for
4822 The default is @option{-ftrapping-math}.
4824 @item -frounding-math
4825 @opindex frounding-math
4826 Disable transformations and optimizations that assume default floating
4827 point rounding behavior. This is round-to-zero for all floating point
4828 to integer conversions, and round-to-nearest for all other arithmetic
4829 truncations. This option should be specified for programs that change
4830 the FP rounding mode dynamically, or that may be executed with a
4831 non-default rounding mode. This option disables constant folding of
4832 floating point expressions at compile-time (which may be affected by
4833 rounding mode) and arithmetic transformations that are unsafe in the
4834 presence of sign-dependent rounding modes.
4836 The default is @option{-fno-rounding-math}.
4838 This option is experimental and does not currently guarantee to
4839 disable all GCC optimizations that are affected by rounding mode.
4840 Future versions of GCC may provide finer control of this setting
4841 using C99's @code{FENV_ACCESS} pragma. This command line option
4842 will be used to specify the default state for @code{FENV_ACCESS}.
4844 @item -fsignaling-nans
4845 @opindex fsignaling-nans
4846 Compile code assuming that IEEE signaling NaNs may generate user-visible
4847 traps during floating-point operations. Setting this option disables
4848 optimizations that may change the number of exceptions visible with
4849 signaling NaNs. This option implies @option{-ftrapping-math}.
4851 This option causes the preprocessor macro @code{__SUPPORT_SNAN__} to
4854 The default is @option{-fno-signaling-nans}.
4856 This option is experimental and does not currently guarantee to
4857 disable all GCC optimizations that affect signaling NaN behavior.
4859 @item -fsingle-precision-constant
4860 @opindex fsingle-precision-constant
4861 Treat floating point constant as single precision constant instead of
4862 implicitly converting it to double precision constant.
4867 The following options control optimizations that may improve
4868 performance, but are not enabled by any @option{-O} options. This
4869 section includes experimental options that may produce broken code.
4872 @item -fbranch-probabilities
4873 @opindex fbranch-probabilities
4874 After running a program compiled with @option{-fprofile-arcs}
4875 (@pxref{Debugging Options,, Options for Debugging Your Program or
4876 @command{gcc}}), you can compile it a second time using
4877 @option{-fbranch-probabilities}, to improve optimizations based on
4878 the number of times each branch was taken. When the program
4879 compiled with @option{-fprofile-arcs} exits it saves arc execution
4880 counts to a file called @file{@var{sourcename}.gcda} for each source
4881 file The information in this data file is very dependent on the
4882 structure of the generated code, so you must use the same source code
4883 and the same optimization options for both compilations.
4885 With @option{-fbranch-probabilities}, GCC puts a
4886 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
4887 These can be used to improve optimization. Currently, they are only
4888 used in one place: in @file{reorg.c}, instead of guessing which path a
4889 branch is mostly to take, the @samp{REG_BR_PROB} values are used to
4890 exactly determine which path is taken more often.
4892 @item -fprofile-values
4893 @opindex fprofile-values
4894 If combined with @option{-fprofile-arcs}, it adds code so that some
4895 data about values of expressions in the program is gathered.
4897 With @option{-fbranch-probabilities}, it reads back the data gathered
4898 from profiling values of expressions and adds @samp{REG_VALUE_PROFILE}
4899 notes to instructions for their later usage in optimizations.
4901 Enabled with @option{-fprofile-generate} and @option{-fprofile-use}.
4905 If combined with @option{-fprofile-arcs}, it instructs the compiler to add
4906 a code to gather information about values of expressions.
4908 With @option{-fbranch-probabilities}, it reads back the data gathered
4909 and actually performs the optimizations based on them.
4910 Currently the optimizations include specialization of division operation
4911 using the knowledge about the value of the denominator.
4913 Enabled with @option{-fprofile-generate} and @option{-fprofile-use}.
4915 @item -frename-registers
4916 @opindex frename-registers
4917 Attempt to avoid false dependencies in scheduled code by making use
4918 of registers left over after register allocation. This optimization
4919 will most benefit processors with lots of registers. Depending on the
4920 debug information format adopted by the target, however, it can
4921 make debugging impossible, since variables will no longer stay in
4922 a ``home register''.
4924 Not enabled by default at any level because it has known bugs.
4928 Use a graph coloring register allocator. Currently this option is meant
4929 for testing, so we are interested to hear about miscompilations with
4934 Perform tail duplication to enlarge superblock size. This transformation
4935 simplifies the control flow of the function allowing other optimizations to do
4938 Enabled with @option{-fprofile-use}.
4940 @item -funroll-loops
4941 @opindex funroll-loops
4942 Unroll loops whose number of iterations can be determined at compile time or
4943 upon entry to the loop. @option{-funroll-loops} implies
4944 @option{-frerun-cse-after-loop}. It also turns on complete loop peeling
4945 (i.e. complete removal of loops with small constant number of iterations).
4946 This option makes code larger, and may or may not make it run faster.
4948 Enabled with @option{-fprofile-use}.
4950 @item -funroll-all-loops
4951 @opindex funroll-all-loops
4952 Unroll all loops, even if their number of iterations is uncertain when
4953 the loop is entered. This usually makes programs run more slowly.
4954 @option{-funroll-all-loops} implies the same options as
4955 @option{-funroll-loops}.
4958 @opindex fpeel-loops
4959 Peels the loops for that there is enough information that they do not
4960 roll much (from profile feedback). It also turns on complete loop peeling
4961 (i.e. complete removal of loops with small constant number of iterations).
4963 Enabled with @option{-fprofile-use}.
4965 @item -fmove-loop-invariants
4966 @opindex fmove-loop-invariants
4967 Enables the loop invariant motion pass in the new loop optimizer. Enabled
4968 at level @option{-O1}
4970 @item -funswitch-loops
4971 @opindex funswitch-loops
4972 Move branches with loop invariant conditions out of the loop, with duplicates
4973 of the loop on both branches (modified according to result of the condition).
4975 @item -fold-unroll-loops
4976 @opindex fold-unroll-loops
4977 Unroll loops whose number of iterations can be determined at compile
4978 time or upon entry to the loop, using the old loop unroller whose loop
4979 recognition is based on notes from frontend. @option{-fold-unroll-loops} implies
4980 both @option{-fstrength-reduce} and @option{-frerun-cse-after-loop}. This
4981 option makes code larger, and may or may not make it run faster.
4983 @item -fold-unroll-all-loops
4984 @opindex fold-unroll-all-loops
4985 Unroll all loops, even if their number of iterations is uncertain when
4986 the loop is entered. This is done using the old loop unroller whose loop
4987 recognition is based on notes from frontend. This usually makes programs run more slowly.
4988 @option{-fold-unroll-all-loops} implies the same options as
4989 @option{-fold-unroll-loops}.
4991 @item -fprefetch-loop-arrays
4992 @opindex fprefetch-loop-arrays
4993 If supported by the target machine, generate instructions to prefetch
4994 memory to improve the performance of loops that access large arrays.
4996 Disabled at level @option{-Os}.
4998 @item -ffunction-sections
4999 @itemx -fdata-sections
5000 @opindex ffunction-sections
5001 @opindex fdata-sections
5002 Place each function or data item into its own section in the output
5003 file if the target supports arbitrary sections. The name of the
5004 function or the name of the data item determines the section's name
5007 Use these options on systems where the linker can perform optimizations
5008 to improve locality of reference in the instruction space. Most systems
5009 using the ELF object format and SPARC processors running Solaris 2 have
5010 linkers with such optimizations. AIX may have these optimizations in
5013 Only use these options when there are significant benefits from doing
5014 so. When you specify these options, the assembler and linker will
5015 create larger object and executable files and will also be slower.
5016 You will not be able to use @code{gprof} on all systems if you
5017 specify this option and you may have problems with debugging if
5018 you specify both this option and @option{-g}.
5020 @item -fbranch-target-load-optimize
5021 @opindex fbranch-target-load-optimize
5022 Perform branch target register load optimization before prologue / epilogue
5024 The use of target registers can typically be exposed only during reload,
5025 thus hoisting loads out of loops and doing inter-block scheduling needs
5026 a separate optimization pass.
5028 @item -fbranch-target-load-optimize2
5029 @opindex fbranch-target-load-optimize2
5030 Perform branch target register load optimization after prologue / epilogue
5033 @item -fbtr-bb-exclusive
5034 @opindex fbtr-bb-exclusive
5035 When performing branch target register load optimization, don't reuse
5036 branch target registers in within any basic block.
5038 @item --param @var{name}=@var{value}
5040 In some places, GCC uses various constants to control the amount of
5041 optimization that is done. For example, GCC will not inline functions
5042 that contain more that a certain number of instructions. You can
5043 control some of these constants on the command-line using the
5044 @option{--param} option.
5046 The names of specific parameters, and the meaning of the values, are
5047 tied to the internals of the compiler, and are subject to change
5048 without notice in future releases.
5050 In each case, the @var{value} is an integer. The allowable choices for
5051 @var{name} are given in the following table:
5054 @item max-crossjump-edges
5055 The maximum number of incoming edges to consider for crossjumping.
5056 The algorithm used by @option{-fcrossjumping} is @math{O(N^2)} in
5057 the number of edges incoming to each block. Increasing values mean
5058 more aggressive optimization, making the compile time increase with
5059 probably small improvement in executable size.
5061 @item max-delay-slot-insn-search
5062 The maximum number of instructions to consider when looking for an
5063 instruction to fill a delay slot. If more than this arbitrary number of
5064 instructions is searched, the time savings from filling the delay slot
5065 will be minimal so stop searching. Increasing values mean more
5066 aggressive optimization, making the compile time increase with probably
5067 small improvement in executable run time.
5069 @item max-delay-slot-live-search
5070 When trying to fill delay slots, the maximum number of instructions to
5071 consider when searching for a block with valid live register
5072 information. Increasing this arbitrarily chosen value means more
5073 aggressive optimization, increasing the compile time. This parameter
5074 should be removed when the delay slot code is rewritten to maintain the
5077 @item max-gcse-memory
5078 The approximate maximum amount of memory that will be allocated in
5079 order to perform the global common subexpression elimination
5080 optimization. If more memory than specified is required, the
5081 optimization will not be done.
5083 @item max-gcse-passes
5084 The maximum number of passes of GCSE to run. The default is 1.
5086 @item max-pending-list-length
5087 The maximum number of pending dependencies scheduling will allow
5088 before flushing the current state and starting over. Large functions
5089 with few branches or calls can create excessively large lists which
5090 needlessly consume memory and resources.
5092 @item max-inline-insns-single
5093 Several parameters control the tree inliner used in gcc.
5094 This number sets the maximum number of instructions (counted in GCC's
5095 internal representation) in a single function that the tree inliner
5096 will consider for inlining. This only affects functions declared
5097 inline and methods implemented in a class declaration (C++).
5098 The default value is 500.
5100 @item max-inline-insns-auto
5101 When you use @option{-finline-functions} (included in @option{-O3}),
5102 a lot of functions that would otherwise not be considered for inlining
5103 by the compiler will be investigated. To those functions, a different
5104 (more restrictive) limit compared to functions declared inline can
5106 The default value is 120.
5108 @item large-function-insns
5109 The limit specifying really large functions. For functions greater than this
5110 limit inlining is constrained by @option{--param large-function-growth}.
5111 This parameter is useful primarily to avoid extreme compilation time caused by non-linear
5112 algorithms used by the backend.
5113 This parameter is ignored when @option{-funit-at-a-time} is not used.
5114 The default value is 3000.
5116 @item large-function-growth
5117 Specifies maximal growth of large function caused by inlining in percents.
5118 This parameter is ignored when @option{-funit-at-a-time} is not used.
5119 The default value is 200.
5121 @item inline-unit-growth
5122 Specifies maximal overall growth of the compilation unit caused by inlining.
5123 This parameter is ignored when @option{-funit-at-a-time} is not used.
5124 The default value is 150.
5126 @item max-inline-insns-recursive
5127 @itemx max-inline-insns-recursive-auto
5128 Specifies maximum number of instructions out-of-line copy of self recursive inline
5129 function can grow into by performing recursive inlining.
5131 For functions declared inline @option{--param max-inline-insns-recursive} is
5132 taken into acount. For function not declared inline, recursive inlining
5133 happens only when @option{-finline-functions} (included in @option{-O3}) is
5134 enabled and @option{--param max-inline-insns-recursive-auto} is used. The
5135 default value is 500.
5137 @item max-inline-recursive-depth
5138 @itemx max-inline-recursive-depth-auto
5139 Specifies maximum recursion depth used by the recursive inlining.
5141 For functions declared inline @option{--param max-inline-recursive-depth} is
5142 taken into acount. For function not declared inline, recursive inlining
5143 happens only when @option{-finline-functions} (included in @option{-O3}) is
5144 enabled and @option{--param max-inline-recursive-depth-auto} is used. The
5145 default value is 500.
5147 @item max-inline-insns-rtl
5148 For languages that use the RTL inliner (this happens at a later stage
5149 than tree inlining), you can set the maximum allowable size (counted
5150 in RTL instructions) for the RTL inliner with this parameter.
5151 The default value is 600.
5153 @item max-unrolled-insns
5154 The maximum number of instructions that a loop should have if that loop
5155 is unrolled, and if the loop is unrolled, it determines how many times
5156 the loop code is unrolled.
5158 @item max-average-unrolled-insns
5159 The maximum number of instructions biased by probabilities of their execution
5160 that a loop should have if that loop is unrolled, and if the loop is unrolled,
5161 it determines how many times the loop code is unrolled.
5163 @item max-unroll-times
5164 The maximum number of unrollings of a single loop.
5166 @item max-peeled-insns
5167 The maximum number of instructions that a loop should have if that loop
5168 is peeled, and if the loop is peeled, it determines how many times
5169 the loop code is peeled.
5171 @item max-peel-times
5172 The maximum number of peelings of a single loop.
5174 @item max-completely-peeled-insns
5175 The maximum number of insns of a completely peeled loop.
5177 @item max-completely-peel-times
5178 The maximum number of iterations of a loop to be suitable for complete peeling.
5180 @item max-unswitch-insns
5181 The maximum number of insns of an unswitched loop.
5183 @item max-unswitch-level
5184 The maximum number of branches unswitched in a single loop.
5187 The minimum cost of an expensive expression in the loop invariant motion.
5189 @item max-iterations-to-track
5191 The maximum number of iterations of a loop the brute force algorithm
5192 for analysis of # of iterations of the loop tries to evaluate.
5194 @item hot-bb-count-fraction
5195 Select fraction of the maximal count of repetitions of basic block in program
5196 given basic block needs to have to be considered hot.
5198 @item hot-bb-frequency-fraction
5199 Select fraction of the maximal frequency of executions of basic block in
5200 function given basic block needs to have to be considered hot
5202 @item tracer-dynamic-coverage
5203 @itemx tracer-dynamic-coverage-feedback
5205 This value is used to limit superblock formation once the given percentage of
5206 executed instructions is covered. This limits unnecessary code size
5209 The @option{tracer-dynamic-coverage-feedback} is used only when profile
5210 feedback is available. The real profiles (as opposed to statically estimated
5211 ones) are much less balanced allowing the threshold to be larger value.
5213 @item tracer-max-code-growth
5214 Stop tail duplication once code growth has reached given percentage. This is
5215 rather hokey argument, as most of the duplicates will be eliminated later in
5216 cross jumping, so it may be set to much higher values than is the desired code
5219 @item tracer-min-branch-ratio
5221 Stop reverse growth when the reverse probability of best edge is less than this
5222 threshold (in percent).
5224 @item tracer-min-branch-ratio
5225 @itemx tracer-min-branch-ratio-feedback
5227 Stop forward growth if the best edge do have probability lower than this
5230 Similarly to @option{tracer-dynamic-coverage} two values are present, one for
5231 compilation for profile feedback and one for compilation without. The value
5232 for compilation with profile feedback needs to be more conservative (higher) in
5233 order to make tracer effective.
5235 @item max-cse-path-length
5237 Maximum number of basic blocks on path that cse considers. The default is 10.
5239 @item global-var-threshold
5241 Counts the number of function calls (N) and the number of
5242 call-clobbered variables (V). If NxV is larger than this limit, a
5243 single artificial variable will be created to represent all the
5244 call-clobbered variables at function call sites. This artificial
5245 variable will then be made to alias every call-clobbered variable.
5246 (done as int * size_t on the host machine; beware overflow).
5248 @item max-aliased-vops
5250 Maxiumum number of virtual operands allowed to represent aliases
5251 before triggering the alias grouping heuristic. Alias grouping
5252 reduces compile times and memory consumption needed for aliasing at
5253 the expense of precision loss in alias information.
5255 @item ggc-min-expand
5257 GCC uses a garbage collector to manage its own memory allocation. This
5258 parameter specifies the minimum percentage by which the garbage
5259 collector's heap should be allowed to expand between collections.
5260 Tuning this may improve compilation speed; it has no effect on code
5263 The default is 30% + 70% * (RAM/1GB) with an upper bound of 100% when
5264 RAM >= 1GB. If @code{getrlimit} is available, the notion of "RAM" is
5265 the smallest of actual RAM, RLIMIT_RSS, RLIMIT_DATA and RLIMIT_AS. If
5266 GCC is not able to calculate RAM on a particular platform, the lower
5267 bound of 30% is used. Setting this parameter and
5268 @option{ggc-min-heapsize} to zero causes a full collection to occur at
5269 every opportunity. This is extremely slow, but can be useful for
5272 @item ggc-min-heapsize
5274 Minimum size of the garbage collector's heap before it begins bothering
5275 to collect garbage. The first collection occurs after the heap expands
5276 by @option{ggc-min-expand}% beyond @option{ggc-min-heapsize}. Again,
5277 tuning this may improve compilation speed, and has no effect on code
5280 The default is RAM/8, with a lower bound of 4096 (four megabytes) and an
5281 upper bound of 131072 (128 megabytes). If @code{getrlimit} is
5282 available, the notion of "RAM" is the smallest of actual RAM,
5283 RLIMIT_RSS, RLIMIT_DATA and RLIMIT_AS. If GCC is not able to calculate
5284 RAM on a particular platform, the lower bound is used. Setting this
5285 parameter very large effectively disables garbage collection. Setting
5286 this parameter and @option{ggc-min-expand} to zero causes a full
5287 collection to occur at every opportunity.
5289 @item max-reload-search-insns
5290 The maximum number of instruction reload should look backward for equivalent
5291 register. Increasing values mean more aggressive optimization, making the
5292 compile time increase with probably slightly better performance. The default
5295 @item max-cselib-memory-location
5296 The maximum number of memory locations cselib should take into acount.
5297 Increasing values mean more aggressive optimization, making the compile time
5298 increase with probably slightly better performance. The default value is 500.
5300 @item reorder-blocks-duplicate
5301 @itemx reorder-blocks-duplicate-feedback
5303 Used by basic block reordering pass to decide whether to use unconditional
5304 branch or duplicate the code on its destination. Code is duplicated when its
5305 estimated size is smaller than this value multiplied by the estimated size of
5306 unconditional jump in the hot spots of the program.
5308 The @option{reorder-block-duplicate-feedback} is used only when profile
5309 feedback is available and may be set to higher values than
5310 @option{reorder-block-duplicate} since information about the hot spots is more
5313 @item max-sched-region-blocks
5314 The maximum number of blocks in a region to be considered for
5315 interblock scheduling. The default value is 10.
5317 @item max-sched-region-insns
5318 The maximum number of insns in a region to be considered for
5319 interblock scheduling. The default value is 100.
5323 @node Preprocessor Options
5324 @section Options Controlling the Preprocessor
5325 @cindex preprocessor options
5326 @cindex options, preprocessor
5328 These options control the C preprocessor, which is run on each C source
5329 file before actual compilation.
5331 If you use the @option{-E} option, nothing is done except preprocessing.
5332 Some of these options make sense only together with @option{-E} because
5333 they cause the preprocessor output to be unsuitable for actual
5338 You can use @option{-Wp,@var{option}} to bypass the compiler driver
5339 and pass @var{option} directly through to the preprocessor. If
5340 @var{option} contains commas, it is split into multiple options at the
5341 commas. However, many options are modified, translated or interpreted
5342 by the compiler driver before being passed to the preprocessor, and
5343 @option{-Wp} forcibly bypasses this phase. The preprocessor's direct
5344 interface is undocumented and subject to change, so whenever possible
5345 you should avoid using @option{-Wp} and let the driver handle the
5348 @item -Xpreprocessor @var{option}
5349 @opindex preprocessor
5350 Pass @var{option} as an option to the preprocessor. You can use this to
5351 supply system-specific preprocessor options which GCC does not know how to
5354 If you want to pass an option that takes an argument, you must use
5355 @option{-Xpreprocessor} twice, once for the option and once for the argument.
5358 @include cppopts.texi
5360 @node Assembler Options
5361 @section Passing Options to the Assembler
5363 @c prevent bad page break with this line
5364 You can pass options to the assembler.
5367 @item -Wa,@var{option}
5369 Pass @var{option} as an option to the assembler. If @var{option}
5370 contains commas, it is split into multiple options at the commas.
5372 @item -Xassembler @var{option}
5374 Pass @var{option} as an option to the assembler. You can use this to
5375 supply system-specific assembler options which GCC does not know how to
5378 If you want to pass an option that takes an argument, you must use
5379 @option{-Xassembler} twice, once for the option and once for the argument.
5384 @section Options for Linking
5385 @cindex link options
5386 @cindex options, linking
5388 These options come into play when the compiler links object files into
5389 an executable output file. They are meaningless if the compiler is
5390 not doing a link step.
5394 @item @var{object-file-name}
5395 A file name that does not end in a special recognized suffix is
5396 considered to name an object file or library. (Object files are
5397 distinguished from libraries by the linker according to the file
5398 contents.) If linking is done, these object files are used as input
5407 If any of these options is used, then the linker is not run, and
5408 object file names should not be used as arguments. @xref{Overall
5412 @item -l@var{library}
5413 @itemx -l @var{library}
5415 Search the library named @var{library} when linking. (The second
5416 alternative with the library as a separate argument is only for
5417 POSIX compliance and is not recommended.)
5419 It makes a difference where in the command you write this option; the
5420 linker searches and processes libraries and object files in the order they
5421 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
5422 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
5423 to functions in @samp{z}, those functions may not be loaded.
5425 The linker searches a standard list of directories for the library,
5426 which is actually a file named @file{lib@var{library}.a}. The linker
5427 then uses this file as if it had been specified precisely by name.
5429 The directories searched include several standard system directories
5430 plus any that you specify with @option{-L}.
5432 Normally the files found this way are library files---archive files
5433 whose members are object files. The linker handles an archive file by
5434 scanning through it for members which define symbols that have so far
5435 been referenced but not defined. But if the file that is found is an
5436 ordinary object file, it is linked in the usual fashion. The only
5437 difference between using an @option{-l} option and specifying a file name
5438 is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
5439 and searches several directories.
5443 You need this special case of the @option{-l} option in order to
5444 link an Objective-C program.
5447 @opindex nostartfiles
5448 Do not use the standard system startup files when linking.
5449 The standard system libraries are used normally, unless @option{-nostdlib}
5450 or @option{-nodefaultlibs} is used.
5452 @item -nodefaultlibs
5453 @opindex nodefaultlibs
5454 Do not use the standard system libraries when linking.
5455 Only the libraries you specify will be passed to the linker.
5456 The standard startup files are used normally, unless @option{-nostartfiles}
5457 is used. The compiler may generate calls to @code{memcmp},
5458 @code{memset}, @code{memcpy} and @code{memmove}.
5459 These entries are usually resolved by entries in
5460 libc. These entry points should be supplied through some other
5461 mechanism when this option is specified.
5465 Do not use the standard system startup files or libraries when linking.
5466 No startup files and only the libraries you specify will be passed to
5467 the linker. The compiler may generate calls to @code{memcmp}, @code{memset},
5468 @code{memcpy} and @code{memmove}.
5469 These entries are usually resolved by entries in
5470 libc. These entry points should be supplied through some other
5471 mechanism when this option is specified.
5473 @cindex @option{-lgcc}, use with @option{-nostdlib}
5474 @cindex @option{-nostdlib} and unresolved references
5475 @cindex unresolved references and @option{-nostdlib}
5476 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
5477 @cindex @option{-nodefaultlibs} and unresolved references
5478 @cindex unresolved references and @option{-nodefaultlibs}
5479 One of the standard libraries bypassed by @option{-nostdlib} and
5480 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
5481 that GCC uses to overcome shortcomings of particular machines, or special
5482 needs for some languages.
5483 (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler
5484 Collection (GCC) Internals},
5485 for more discussion of @file{libgcc.a}.)
5486 In most cases, you need @file{libgcc.a} even when you want to avoid
5487 other standard libraries. In other words, when you specify @option{-nostdlib}
5488 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
5489 This ensures that you have no unresolved references to internal GCC
5490 library subroutines. (For example, @samp{__main}, used to ensure C++
5491 constructors will be called; @pxref{Collect2,,@code{collect2}, gccint,
5492 GNU Compiler Collection (GCC) Internals}.)
5496 Produce a position independent executable on targets which support it.
5497 For predictable results, you must also specify the same set of options
5498 that were used to generate code (@option{-fpie}, @option{-fPIE},
5499 or model suboptions) when you specify this option.
5503 Remove all symbol table and relocation information from the executable.
5507 On systems that support dynamic linking, this prevents linking with the shared
5508 libraries. On other systems, this option has no effect.
5512 Produce a shared object which can then be linked with other objects to
5513 form an executable. Not all systems support this option. For predictable
5514 results, you must also specify the same set of options that were used to
5515 generate code (@option{-fpic}, @option{-fPIC}, or model suboptions)
5516 when you specify this option.@footnote{On some systems, @samp{gcc -shared}
5517 needs to build supplementary stub code for constructors to work. On
5518 multi-libbed systems, @samp{gcc -shared} must select the correct support
5519 libraries to link against. Failing to supply the correct flags may lead
5520 to subtle defects. Supplying them in cases where they are not necessary
5523 @item -shared-libgcc
5524 @itemx -static-libgcc
5525 @opindex shared-libgcc
5526 @opindex static-libgcc
5527 On systems that provide @file{libgcc} as a shared library, these options
5528 force the use of either the shared or static version respectively.
5529 If no shared version of @file{libgcc} was built when the compiler was
5530 configured, these options have no effect.
5532 There are several situations in which an application should use the
5533 shared @file{libgcc} instead of the static version. The most common
5534 of these is when the application wishes to throw and catch exceptions
5535 across different shared libraries. In that case, each of the libraries
5536 as well as the application itself should use the shared @file{libgcc}.
5538 Therefore, the G++ and GCJ drivers automatically add
5539 @option{-shared-libgcc} whenever you build a shared library or a main
5540 executable, because C++ and Java programs typically use exceptions, so
5541 this is the right thing to do.
5543 If, instead, you use the GCC driver to create shared libraries, you may
5544 find that they will not always be linked with the shared @file{libgcc}.
5545 If GCC finds, at its configuration time, that you have a non-GNU linker
5546 or a GNU linker that does not support option @option{--eh-frame-hdr},
5547 it will link the shared version of @file{libgcc} into shared libraries
5548 by default. Otherwise, it will take advantage of the linker and optimize
5549 away the linking with the shared version of @file{libgcc}, linking with
5550 the static version of libgcc by default. This allows exceptions to
5551 propagate through such shared libraries, without incurring relocation
5552 costs at library load time.
5554 However, if a library or main executable is supposed to throw or catch
5555 exceptions, you must link it using the G++ or GCJ driver, as appropriate
5556 for the languages used in the program, or using the option
5557 @option{-shared-libgcc}, such that it is linked with the shared
5562 Bind references to global symbols when building a shared object. Warn
5563 about any unresolved references (unless overridden by the link editor
5564 option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support
5567 @item -Xlinker @var{option}
5569 Pass @var{option} as an option to the linker. You can use this to
5570 supply system-specific linker options which GCC does not know how to
5573 If you want to pass an option that takes an argument, you must use
5574 @option{-Xlinker} twice, once for the option and once for the argument.
5575 For example, to pass @option{-assert definitions}, you must write
5576 @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write
5577 @option{-Xlinker "-assert definitions"}, because this passes the entire
5578 string as a single argument, which is not what the linker expects.
5580 @item -Wl,@var{option}
5582 Pass @var{option} as an option to the linker. If @var{option} contains
5583 commas, it is split into multiple options at the commas.
5585 @item -u @var{symbol}
5587 Pretend the symbol @var{symbol} is undefined, to force linking of
5588 library modules to define it. You can use @option{-u} multiple times with
5589 different symbols to force loading of additional library modules.
5592 @node Directory Options
5593 @section Options for Directory Search
5594 @cindex directory options
5595 @cindex options, directory search
5598 These options specify directories to search for header files, for
5599 libraries and for parts of the compiler:
5604 Add the directory @var{dir} to the head of the list of directories to be
5605 searched for header files. This can be used to override a system header
5606 file, substituting your own version, since these directories are
5607 searched before the system header file directories. However, you should
5608 not use this option to add directories that contain vendor-supplied
5609 system header files (use @option{-isystem} for that). If you use more than
5610 one @option{-I} option, the directories are scanned in left-to-right
5611 order; the standard system directories come after.
5613 If a standard system include directory, or a directory specified with
5614 @option{-isystem}, is also specified with @option{-I}, the @option{-I}
5615 option will be ignored. The directory will still be searched but as a
5616 system directory at its normal position in the system include chain.
5617 This is to ensure that GCC's procedure to fix buggy system headers and
5618 the ordering for the include_next directive are not inadvertently changed.
5619 If you really need to change the search order for system directories,
5620 use the @option{-nostdinc} and/or @option{-isystem} options.
5622 @item -iquote@var{dir}
5624 Add the directory @var{dir} to the head of the list of directories to
5625 be searched for header files only for the case of @samp{#include
5626 "@var{file}"}; they are not searched for @samp{#include <@var{file}>},
5627 otherwise just like @option{-I}.
5631 Add directory @var{dir} to the list of directories to be searched
5634 @item -B@var{prefix}
5636 This option specifies where to find the executables, libraries,
5637 include files, and data files of the compiler itself.
5639 The compiler driver program runs one or more of the subprograms
5640 @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries
5641 @var{prefix} as a prefix for each program it tries to run, both with and
5642 without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}).
5644 For each subprogram to be run, the compiler driver first tries the
5645 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
5646 was not specified, the driver tries two standard prefixes, which are
5647 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc/}. If neither of
5648 those results in a file name that is found, the unmodified program
5649 name is searched for using the directories specified in your
5650 @env{PATH} environment variable.
5652 The compiler will check to see if the path provided by the @option{-B}
5653 refers to a directory, and if necessary it will add a directory
5654 separator character at the end of the path.
5656 @option{-B} prefixes that effectively specify directory names also apply
5657 to libraries in the linker, because the compiler translates these
5658 options into @option{-L} options for the linker. They also apply to
5659 includes files in the preprocessor, because the compiler translates these
5660 options into @option{-isystem} options for the preprocessor. In this case,
5661 the compiler appends @samp{include} to the prefix.
5663 The run-time support file @file{libgcc.a} can also be searched for using
5664 the @option{-B} prefix, if needed. If it is not found there, the two
5665 standard prefixes above are tried, and that is all. The file is left
5666 out of the link if it is not found by those means.
5668 Another way to specify a prefix much like the @option{-B} prefix is to use
5669 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
5672 As a special kludge, if the path provided by @option{-B} is
5673 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
5674 9, then it will be replaced by @file{[dir/]include}. This is to help
5675 with boot-strapping the compiler.
5677 @item -specs=@var{file}
5679 Process @var{file} after the compiler reads in the standard @file{specs}
5680 file, in order to override the defaults that the @file{gcc} driver
5681 program uses when determining what switches to pass to @file{cc1},
5682 @file{cc1plus}, @file{as}, @file{ld}, etc. More than one
5683 @option{-specs=@var{file}} can be specified on the command line, and they
5684 are processed in order, from left to right.
5688 This option has been deprecated. Please use @option{-iquote} instead for
5689 @option{-I} directories before the @option{-I-} and remove the @option{-I-}.
5690 Any directories you specify with @option{-I} options before the @option{-I-}
5691 option are searched only for the case of @samp{#include "@var{file}"};
5692 they are not searched for @samp{#include <@var{file}>}.
5694 If additional directories are specified with @option{-I} options after
5695 the @option{-I-}, these directories are searched for all @samp{#include}
5696 directives. (Ordinarily @emph{all} @option{-I} directories are used
5699 In addition, the @option{-I-} option inhibits the use of the current
5700 directory (where the current input file came from) as the first search
5701 directory for @samp{#include "@var{file}"}. There is no way to
5702 override this effect of @option{-I-}. With @option{-I.} you can specify
5703 searching the directory which was current when the compiler was
5704 invoked. That is not exactly the same as what the preprocessor does
5705 by default, but it is often satisfactory.
5707 @option{-I-} does not inhibit the use of the standard system directories
5708 for header files. Thus, @option{-I-} and @option{-nostdinc} are
5715 @section Specifying subprocesses and the switches to pass to them
5718 @command{gcc} is a driver program. It performs its job by invoking a
5719 sequence of other programs to do the work of compiling, assembling and
5720 linking. GCC interprets its command-line parameters and uses these to
5721 deduce which programs it should invoke, and which command-line options
5722 it ought to place on their command lines. This behavior is controlled
5723 by @dfn{spec strings}. In most cases there is one spec string for each
5724 program that GCC can invoke, but a few programs have multiple spec
5725 strings to control their behavior. The spec strings built into GCC can
5726 be overridden by using the @option{-specs=} command-line switch to specify
5729 @dfn{Spec files} are plaintext files that are used to construct spec
5730 strings. They consist of a sequence of directives separated by blank
5731 lines. The type of directive is determined by the first non-whitespace
5732 character on the line and it can be one of the following:
5735 @item %@var{command}
5736 Issues a @var{command} to the spec file processor. The commands that can
5740 @item %include <@var{file}>
5742 Search for @var{file} and insert its text at the current point in the
5745 @item %include_noerr <@var{file}>
5746 @cindex %include_noerr
5747 Just like @samp{%include}, but do not generate an error message if the include
5748 file cannot be found.
5750 @item %rename @var{old_name} @var{new_name}
5752 Rename the spec string @var{old_name} to @var{new_name}.
5756 @item *[@var{spec_name}]:
5757 This tells the compiler to create, override or delete the named spec
5758 string. All lines after this directive up to the next directive or
5759 blank line are considered to be the text for the spec string. If this
5760 results in an empty string then the spec will be deleted. (Or, if the
5761 spec did not exist, then nothing will happened.) Otherwise, if the spec
5762 does not currently exist a new spec will be created. If the spec does
5763 exist then its contents will be overridden by the text of this
5764 directive, unless the first character of that text is the @samp{+}
5765 character, in which case the text will be appended to the spec.
5767 @item [@var{suffix}]:
5768 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
5769 and up to the next directive or blank line are considered to make up the
5770 spec string for the indicated suffix. When the compiler encounters an
5771 input file with the named suffix, it will processes the spec string in
5772 order to work out how to compile that file. For example:
5779 This says that any input file whose name ends in @samp{.ZZ} should be
5780 passed to the program @samp{z-compile}, which should be invoked with the
5781 command-line switch @option{-input} and with the result of performing the
5782 @samp{%i} substitution. (See below.)
5784 As an alternative to providing a spec string, the text that follows a
5785 suffix directive can be one of the following:
5788 @item @@@var{language}
5789 This says that the suffix is an alias for a known @var{language}. This is
5790 similar to using the @option{-x} command-line switch to GCC to specify a
5791 language explicitly. For example:
5798 Says that .ZZ files are, in fact, C++ source files.
5801 This causes an error messages saying:
5804 @var{name} compiler not installed on this system.
5808 GCC already has an extensive list of suffixes built into it.
5809 This directive will add an entry to the end of the list of suffixes, but
5810 since the list is searched from the end backwards, it is effectively
5811 possible to override earlier entries using this technique.
5815 GCC has the following spec strings built into it. Spec files can
5816 override these strings or create their own. Note that individual
5817 targets can also add their own spec strings to this list.
5820 asm Options to pass to the assembler
5821 asm_final Options to pass to the assembler post-processor
5822 cpp Options to pass to the C preprocessor
5823 cc1 Options to pass to the C compiler
5824 cc1plus Options to pass to the C++ compiler
5825 endfile Object files to include at the end of the link
5826 link Options to pass to the linker
5827 lib Libraries to include on the command line to the linker
5828 libgcc Decides which GCC support library to pass to the linker
5829 linker Sets the name of the linker
5830 predefines Defines to be passed to the C preprocessor
5831 signed_char Defines to pass to CPP to say whether @code{char} is signed
5833 startfile Object files to include at the start of the link
5836 Here is a small example of a spec file:
5842 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
5845 This example renames the spec called @samp{lib} to @samp{old_lib} and
5846 then overrides the previous definition of @samp{lib} with a new one.
5847 The new definition adds in some extra command-line options before
5848 including the text of the old definition.
5850 @dfn{Spec strings} are a list of command-line options to be passed to their
5851 corresponding program. In addition, the spec strings can contain
5852 @samp{%}-prefixed sequences to substitute variable text or to
5853 conditionally insert text into the command line. Using these constructs
5854 it is possible to generate quite complex command lines.
5856 Here is a table of all defined @samp{%}-sequences for spec
5857 strings. Note that spaces are not generated automatically around the
5858 results of expanding these sequences. Therefore you can concatenate them
5859 together or combine them with constant text in a single argument.
5863 Substitute one @samp{%} into the program name or argument.
5866 Substitute the name of the input file being processed.
5869 Substitute the basename of the input file being processed.
5870 This is the substring up to (and not including) the last period
5871 and not including the directory.
5874 This is the same as @samp{%b}, but include the file suffix (text after
5878 Marks the argument containing or following the @samp{%d} as a
5879 temporary file name, so that that file will be deleted if GCC exits
5880 successfully. Unlike @samp{%g}, this contributes no text to the
5883 @item %g@var{suffix}
5884 Substitute a file name that has suffix @var{suffix} and is chosen
5885 once per compilation, and mark the argument in the same way as
5886 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
5887 name is now chosen in a way that is hard to predict even when previously
5888 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
5889 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
5890 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
5891 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
5892 was simply substituted with a file name chosen once per compilation,
5893 without regard to any appended suffix (which was therefore treated
5894 just like ordinary text), making such attacks more likely to succeed.
5896 @item %u@var{suffix}
5897 Like @samp{%g}, but generates a new temporary file name even if
5898 @samp{%u@var{suffix}} was already seen.
5900 @item %U@var{suffix}
5901 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
5902 new one if there is no such last file name. In the absence of any
5903 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
5904 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
5905 would involve the generation of two distinct file names, one
5906 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
5907 simply substituted with a file name chosen for the previous @samp{%u},
5908 without regard to any appended suffix.
5910 @item %j@var{suffix}
5911 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
5912 writable, and if save-temps is off; otherwise, substitute the name
5913 of a temporary file, just like @samp{%u}. This temporary file is not
5914 meant for communication between processes, but rather as a junk
5917 @item %|@var{suffix}
5918 @itemx %m@var{suffix}
5919 Like @samp{%g}, except if @option{-pipe} is in effect. In that case
5920 @samp{%|} substitutes a single dash and @samp{%m} substitutes nothing at
5921 all. These are the two most common ways to instruct a program that it
5922 should read from standard input or write to standard output. If you
5923 need something more elaborate you can use an @samp{%@{pipe:@code{X}@}}
5924 construct: see for example @file{f/lang-specs.h}.
5926 @item %.@var{SUFFIX}
5927 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
5928 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
5929 terminated by the next space or %.
5932 Marks the argument containing or following the @samp{%w} as the
5933 designated output file of this compilation. This puts the argument
5934 into the sequence of arguments that @samp{%o} will substitute later.
5937 Substitutes the names of all the output files, with spaces
5938 automatically placed around them. You should write spaces
5939 around the @samp{%o} as well or the results are undefined.
5940 @samp{%o} is for use in the specs for running the linker.
5941 Input files whose names have no recognized suffix are not compiled
5942 at all, but they are included among the output files, so they will
5946 Substitutes the suffix for object files. Note that this is
5947 handled specially when it immediately follows @samp{%g, %u, or %U},
5948 because of the need for those to form complete file names. The
5949 handling is such that @samp{%O} is treated exactly as if it had already
5950 been substituted, except that @samp{%g, %u, and %U} do not currently
5951 support additional @var{suffix} characters following @samp{%O} as they would
5952 following, for example, @samp{.o}.
5955 Substitutes the standard macro predefinitions for the
5956 current target machine. Use this when running @code{cpp}.
5959 Like @samp{%p}, but puts @samp{__} before and after the name of each
5960 predefined macro, except for macros that start with @samp{__} or with
5961 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
5965 Substitute any of @option{-iprefix} (made from @env{GCC_EXEC_PREFIX}),
5966 @option{-isysroot} (made from @env{TARGET_SYSTEM_ROOT}), and
5967 @option{-isystem} (made from @env{COMPILER_PATH} and @option{-B} options)
5971 Current argument is the name of a library or startup file of some sort.
5972 Search for that file in a standard list of directories and substitute
5973 the full name found.
5976 Print @var{str} as an error message. @var{str} is terminated by a newline.
5977 Use this when inconsistent options are detected.
5980 Substitute the contents of spec string @var{name} at this point.
5983 Like @samp{%(@dots{})} but put @samp{__} around @option{-D} arguments.
5985 @item %x@{@var{option}@}
5986 Accumulate an option for @samp{%X}.
5989 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
5993 Output the accumulated assembler options specified by @option{-Wa}.
5996 Output the accumulated preprocessor options specified by @option{-Wp}.
5999 Process the @code{asm} spec. This is used to compute the
6000 switches to be passed to the assembler.
6003 Process the @code{asm_final} spec. This is a spec string for
6004 passing switches to an assembler post-processor, if such a program is
6008 Process the @code{link} spec. This is the spec for computing the
6009 command line passed to the linker. Typically it will make use of the
6010 @samp{%L %G %S %D and %E} sequences.
6013 Dump out a @option{-L} option for each directory that GCC believes might
6014 contain startup files. If the target supports multilibs then the
6015 current multilib directory will be prepended to each of these paths.
6018 Output the multilib directory with directory separators replaced with
6019 @samp{_}. If multilib directories are not set, or the multilib directory is
6020 @file{.} then this option emits nothing.
6023 Process the @code{lib} spec. This is a spec string for deciding which
6024 libraries should be included on the command line to the linker.
6027 Process the @code{libgcc} spec. This is a spec string for deciding
6028 which GCC support library should be included on the command line to the linker.
6031 Process the @code{startfile} spec. This is a spec for deciding which
6032 object files should be the first ones passed to the linker. Typically
6033 this might be a file named @file{crt0.o}.
6036 Process the @code{endfile} spec. This is a spec string that specifies
6037 the last object files that will be passed to the linker.
6040 Process the @code{cpp} spec. This is used to construct the arguments
6041 to be passed to the C preprocessor.
6044 Process the @code{cc1} spec. This is used to construct the options to be
6045 passed to the actual C compiler (@samp{cc1}).
6048 Process the @code{cc1plus} spec. This is used to construct the options to be
6049 passed to the actual C++ compiler (@samp{cc1plus}).
6052 Substitute the variable part of a matched option. See below.
6053 Note that each comma in the substituted string is replaced by
6057 Remove all occurrences of @code{-S} from the command line. Note---this
6058 command is position dependent. @samp{%} commands in the spec string
6059 before this one will see @code{-S}, @samp{%} commands in the spec string
6060 after this one will not.
6062 @item %:@var{function}(@var{args})
6063 Call the named function @var{function}, passing it @var{args}.
6064 @var{args} is first processed as a nested spec string, then split
6065 into an argument vector in the usual fashion. The function returns
6066 a string which is processed as if it had appeared literally as part
6067 of the current spec.
6069 The following built-in spec functions are provided:
6072 @item @code{if-exists}
6073 The @code{if-exists} spec function takes one argument, an absolute
6074 pathname to a file. If the file exists, @code{if-exists} returns the
6075 pathname. Here is a small example of its usage:
6079 crt0%O%s %:if-exists(crti%O%s) crtbegin%O%s
6082 @item @code{if-exists-else}
6083 The @code{if-exists-else} spec function is similar to the @code{if-exists}
6084 spec function, except that it takes two arguments. The first argument is
6085 an absolute pathname to a file. If the file exists, @code{if-exists-else}
6086 returns the pathname. If it does not exist, it returns the second argument.
6087 This way, @code{if-exists-else} can be used to select one file or another,
6088 based on the existence of the first. Here is a small example of its usage:
6092 crt0%O%s %:if-exists(crti%O%s) \
6093 %:if-exists-else(crtbeginT%O%s crtbegin%O%s)
6098 Substitutes the @code{-S} switch, if that switch was given to GCC@.
6099 If that switch was not specified, this substitutes nothing. Note that
6100 the leading dash is omitted when specifying this option, and it is
6101 automatically inserted if the substitution is performed. Thus the spec
6102 string @samp{%@{foo@}} would match the command-line option @option{-foo}
6103 and would output the command line option @option{-foo}.
6105 @item %W@{@code{S}@}
6106 Like %@{@code{S}@} but mark last argument supplied within as a file to be
6109 @item %@{@code{S}*@}
6110 Substitutes all the switches specified to GCC whose names start
6111 with @code{-S}, but which also take an argument. This is used for
6112 switches like @option{-o}, @option{-D}, @option{-I}, etc.
6113 GCC considers @option{-o foo} as being
6114 one switch whose names starts with @samp{o}. %@{o*@} would substitute this
6115 text, including the space. Thus two arguments would be generated.
6117 @item %@{@code{S}*&@code{T}*@}
6118 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
6119 (the order of @code{S} and @code{T} in the spec is not significant).
6120 There can be any number of ampersand-separated variables; for each the
6121 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
6123 @item %@{@code{S}:@code{X}@}
6124 Substitutes @code{X}, if the @samp{-S} switch was given to GCC@.
6126 @item %@{!@code{S}:@code{X}@}
6127 Substitutes @code{X}, if the @samp{-S} switch was @emph{not} given to GCC@.
6129 @item %@{@code{S}*:@code{X}@}
6130 Substitutes @code{X} if one or more switches whose names start with
6131 @code{-S} are specified to GCC@. Normally @code{X} is substituted only
6132 once, no matter how many such switches appeared. However, if @code{%*}
6133 appears somewhere in @code{X}, then @code{X} will be substituted once
6134 for each matching switch, with the @code{%*} replaced by the part of
6135 that switch that matched the @code{*}.
6137 @item %@{.@code{S}:@code{X}@}
6138 Substitutes @code{X}, if processing a file with suffix @code{S}.
6140 @item %@{!.@code{S}:@code{X}@}
6141 Substitutes @code{X}, if @emph{not} processing a file with suffix @code{S}.
6143 @item %@{@code{S}|@code{P}:@code{X}@}
6144 Substitutes @code{X} if either @code{-S} or @code{-P} was given to GCC@.
6145 This may be combined with @samp{!}, @samp{.}, and @code{*} sequences as well,
6146 although they have a stronger binding than the @samp{|}. If @code{%*}
6147 appears in @code{X}, all of the alternatives must be starred, and only
6148 the first matching alternative is substituted.
6150 For example, a spec string like this:
6153 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
6156 will output the following command-line options from the following input
6157 command-line options:
6162 -d fred.c -foo -baz -boggle
6163 -d jim.d -bar -baz -boggle
6166 @item %@{S:X; T:Y; :D@}
6168 If @code{S} was given to GCC, substitutes @code{X}; else if @code{T} was
6169 given to GCC, substitutes @code{Y}; else substitutes @code{D}. There can
6170 be as many clauses as you need. This may be combined with @code{.},
6171 @code{!}, @code{|}, and @code{*} as needed.
6176 The conditional text @code{X} in a %@{@code{S}:@code{X}@} or similar
6177 construct may contain other nested @samp{%} constructs or spaces, or
6178 even newlines. They are processed as usual, as described above.
6179 Trailing white space in @code{X} is ignored. White space may also
6180 appear anywhere on the left side of the colon in these constructs,
6181 except between @code{.} or @code{*} and the corresponding word.
6183 The @option{-O}, @option{-f}, @option{-m}, and @option{-W} switches are
6184 handled specifically in these constructs. If another value of
6185 @option{-O} or the negated form of a @option{-f}, @option{-m}, or
6186 @option{-W} switch is found later in the command line, the earlier
6187 switch value is ignored, except with @{@code{S}*@} where @code{S} is
6188 just one letter, which passes all matching options.
6190 The character @samp{|} at the beginning of the predicate text is used to
6191 indicate that a command should be piped to the following command, but
6192 only if @option{-pipe} is specified.
6194 It is built into GCC which switches take arguments and which do not.
6195 (You might think it would be useful to generalize this to allow each
6196 compiler's spec to say which switches take arguments. But this cannot
6197 be done in a consistent fashion. GCC cannot even decide which input
6198 files have been specified without knowing which switches take arguments,
6199 and it must know which input files to compile in order to tell which
6202 GCC also knows implicitly that arguments starting in @option{-l} are to be
6203 treated as compiler output files, and passed to the linker in their
6204 proper position among the other output files.
6206 @c man begin OPTIONS
6208 @node Target Options
6209 @section Specifying Target Machine and Compiler Version
6210 @cindex target options
6211 @cindex cross compiling
6212 @cindex specifying machine version
6213 @cindex specifying compiler version and target machine
6214 @cindex compiler version, specifying
6215 @cindex target machine, specifying
6217 The usual way to run GCC is to run the executable called @file{gcc}, or
6218 @file{<machine>-gcc} when cross-compiling, or
6219 @file{<machine>-gcc-<version>} to run a version other than the one that
6220 was installed last. Sometimes this is inconvenient, so GCC provides
6221 options that will switch to another cross-compiler or version.
6224 @item -b @var{machine}
6226 The argument @var{machine} specifies the target machine for compilation.
6228 The value to use for @var{machine} is the same as was specified as the
6229 machine type when configuring GCC as a cross-compiler. For
6230 example, if a cross-compiler was configured with @samp{configure
6231 i386v}, meaning to compile for an 80386 running System V, then you
6232 would specify @option{-b i386v} to run that cross compiler.
6234 @item -V @var{version}
6236 The argument @var{version} specifies which version of GCC to run.
6237 This is useful when multiple versions are installed. For example,
6238 @var{version} might be @samp{2.0}, meaning to run GCC version 2.0.
6241 The @option{-V} and @option{-b} options work by running the
6242 @file{<machine>-gcc-<version>} executable, so there's no real reason to
6243 use them if you can just run that directly.
6245 @node Submodel Options
6246 @section Hardware Models and Configurations
6247 @cindex submodel options
6248 @cindex specifying hardware config
6249 @cindex hardware models and configurations, specifying
6250 @cindex machine dependent options
6252 Earlier we discussed the standard option @option{-b} which chooses among
6253 different installed compilers for completely different target
6254 machines, such as VAX vs.@: 68000 vs.@: 80386.
6256 In addition, each of these target machine types can have its own
6257 special options, starting with @samp{-m}, to choose among various
6258 hardware models or configurations---for example, 68010 vs 68020,
6259 floating coprocessor or none. A single installed version of the
6260 compiler can compile for any model or configuration, according to the
6263 Some configurations of the compiler also support additional special
6264 options, usually for compatibility with other compilers on the same
6267 These options are defined by the macro @code{TARGET_SWITCHES} in the
6268 machine description. The default for the options is also defined by
6269 that macro, which enables you to change the defaults.
6271 @c This list is ordered alphanumerically by subsection name.
6272 @c It should be the same order and spelling as these options are listed
6273 @c in Machine Dependent Options
6281 * DEC Alpha Options::
6282 * DEC Alpha/VMS Options::
6286 * i386 and x86-64 Options::
6298 * RS/6000 and PowerPC Options::
6299 * S/390 and zSeries Options::
6302 * System V Options::
6303 * TMS320C3x/C4x Options::
6307 * Xstormy16 Options::
6313 @subsection ARC Options
6316 These options are defined for ARC implementations:
6321 Compile code for little endian mode. This is the default.
6325 Compile code for big endian mode.
6328 @opindex mmangle-cpu
6329 Prepend the name of the cpu to all public symbol names.
6330 In multiple-processor systems, there are many ARC variants with different
6331 instruction and register set characteristics. This flag prevents code
6332 compiled for one cpu to be linked with code compiled for another.
6333 No facility exists for handling variants that are ``almost identical''.
6334 This is an all or nothing option.
6336 @item -mcpu=@var{cpu}
6338 Compile code for ARC variant @var{cpu}.
6339 Which variants are supported depend on the configuration.
6340 All variants support @option{-mcpu=base}, this is the default.
6342 @item -mtext=@var{text-section}
6343 @itemx -mdata=@var{data-section}
6344 @itemx -mrodata=@var{readonly-data-section}
6348 Put functions, data, and readonly data in @var{text-section},
6349 @var{data-section}, and @var{readonly-data-section} respectively
6350 by default. This can be overridden with the @code{section} attribute.
6351 @xref{Variable Attributes}.
6356 @subsection ARM Options
6359 These @samp{-m} options are defined for Advanced RISC Machines (ARM)
6363 @item -mabi=@var{name}
6365 Generate code for the specified ABI. Permissible values are: @samp{apcs-gnu},
6366 @samp{atpcs}, @samp{aapcs} and @samp{iwmmxt}.
6369 @opindex mapcs-frame
6370 Generate a stack frame that is compliant with the ARM Procedure Call
6371 Standard for all functions, even if this is not strictly necessary for
6372 correct execution of the code. Specifying @option{-fomit-frame-pointer}
6373 with this option will cause the stack frames not to be generated for
6374 leaf functions. The default is @option{-mno-apcs-frame}.
6378 This is a synonym for @option{-mapcs-frame}.
6381 @c not currently implemented
6382 @item -mapcs-stack-check
6383 @opindex mapcs-stack-check
6384 Generate code to check the amount of stack space available upon entry to
6385 every function (that actually uses some stack space). If there is
6386 insufficient space available then either the function
6387 @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be
6388 called, depending upon the amount of stack space required. The run time
6389 system is required to provide these functions. The default is
6390 @option{-mno-apcs-stack-check}, since this produces smaller code.
6392 @c not currently implemented
6394 @opindex mapcs-float
6395 Pass floating point arguments using the float point registers. This is
6396 one of the variants of the APCS@. This option is recommended if the
6397 target hardware has a floating point unit or if a lot of floating point
6398 arithmetic is going to be performed by the code. The default is
6399 @option{-mno-apcs-float}, since integer only code is slightly increased in
6400 size if @option{-mapcs-float} is used.
6402 @c not currently implemented
6403 @item -mapcs-reentrant
6404 @opindex mapcs-reentrant
6405 Generate reentrant, position independent code. The default is
6406 @option{-mno-apcs-reentrant}.
6409 @item -mthumb-interwork
6410 @opindex mthumb-interwork
6411 Generate code which supports calling between the ARM and Thumb
6412 instruction sets. Without this option the two instruction sets cannot
6413 be reliably used inside one program. The default is
6414 @option{-mno-thumb-interwork}, since slightly larger code is generated
6415 when @option{-mthumb-interwork} is specified.
6417 @item -mno-sched-prolog
6418 @opindex mno-sched-prolog
6419 Prevent the reordering of instructions in the function prolog, or the
6420 merging of those instruction with the instructions in the function's
6421 body. This means that all functions will start with a recognizable set
6422 of instructions (or in fact one of a choice from a small set of
6423 different function prologues), and this information can be used to
6424 locate the start if functions inside an executable piece of code. The
6425 default is @option{-msched-prolog}.
6428 @opindex mhard-float
6429 Generate output containing floating point instructions. This is the
6433 @opindex msoft-float
6434 Generate output containing library calls for floating point.
6435 @strong{Warning:} the requisite libraries are not available for all ARM
6436 targets. Normally the facilities of the machine's usual C compiler are
6437 used, but this cannot be done directly in cross-compilation. You must make
6438 your own arrangements to provide suitable library functions for
6441 @option{-msoft-float} changes the calling convention in the output file;
6442 therefore, it is only useful if you compile @emph{all} of a program with
6443 this option. In particular, you need to compile @file{libgcc.a}, the
6444 library that comes with GCC, with @option{-msoft-float} in order for
6447 @item -mfloat-abi=@var{name}
6449 Specifies which ABI to use for floating point values. Permissible values
6450 are: @samp{soft}, @samp{softfp} and @samp{hard}.
6452 @samp{soft} and @samp{hard} are equivalent to @option{-msoft-float}
6453 and @option{-mhard-float} respectively. @samp{softfp} allows the generation
6454 of floating point instructions, but still uses the soft-float calling
6457 @item -mlittle-endian
6458 @opindex mlittle-endian
6459 Generate code for a processor running in little-endian mode. This is
6460 the default for all standard configurations.
6463 @opindex mbig-endian
6464 Generate code for a processor running in big-endian mode; the default is
6465 to compile code for a little-endian processor.
6467 @item -mwords-little-endian
6468 @opindex mwords-little-endian
6469 This option only applies when generating code for big-endian processors.
6470 Generate code for a little-endian word order but a big-endian byte
6471 order. That is, a byte order of the form @samp{32107654}. Note: this
6472 option should only be used if you require compatibility with code for
6473 big-endian ARM processors generated by versions of the compiler prior to
6476 @item -mcpu=@var{name}
6478 This specifies the name of the target ARM processor. GCC uses this name
6479 to determine what kind of instructions it can emit when generating
6480 assembly code. Permissible names are: @samp{arm2}, @samp{arm250},
6481 @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610},
6482 @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm},
6483 @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700},
6484 @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100},
6485 @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm8},
6486 @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100},
6487 @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920},
6488 @samp{arm920t}, @samp{arm926ejs}, @samp{arm940t}, @samp{arm9tdmi},
6489 @samp{arm10tdmi}, @samp{arm1020t}, @samp{arm1026ejs},
6490 @samp{arm1136js}, @samp{arm1136jfs} ,@samp{xscale}, @samp{iwmmxt},
6493 @itemx -mtune=@var{name}
6495 This option is very similar to the @option{-mcpu=} option, except that
6496 instead of specifying the actual target processor type, and hence
6497 restricting which instructions can be used, it specifies that GCC should
6498 tune the performance of the code as if the target were of the type
6499 specified in this option, but still choosing the instructions that it
6500 will generate based on the cpu specified by a @option{-mcpu=} option.
6501 For some ARM implementations better performance can be obtained by using
6504 @item -march=@var{name}
6506 This specifies the name of the target ARM architecture. GCC uses this
6507 name to determine what kind of instructions it can emit when generating
6508 assembly code. This option can be used in conjunction with or instead
6509 of the @option{-mcpu=} option. Permissible names are: @samp{armv2},
6510 @samp{armv2a}, @samp{armv3}, @samp{armv3m}, @samp{armv4}, @samp{armv4t},
6511 @samp{armv5}, @samp{armv5t}, @samp{armv5te}, @samp{armv6}, @samp{armv6j},
6512 @samp{iwmmxt}, @samp{ep9312}.
6514 @item -mfpu=@var{name}
6515 @itemx -mfpe=@var{number}
6516 @itemx -mfp=@var{number}
6520 This specifies what floating point hardware (or hardware emulation) is
6521 available on the target. Permissible names are: @samp{fpa}, @samp{fpe2},
6522 @samp{fpe3}, @samp{maverick}, @samp{vfp}. @option{-mfp} and @option{-mfpe}
6523 are synonyms for @option{-mfpu}=@samp{fpe}@var{number}, for compatibility
6524 with older versions of GCC@.
6526 If @option{-msoft-float} is specified this specifies the format of
6527 floating point values.
6529 @item -mstructure-size-boundary=@var{n}
6530 @opindex mstructure-size-boundary
6531 The size of all structures and unions will be rounded up to a multiple
6532 of the number of bits set by this option. Permissible values are 8, 32
6533 and 64. The default value varies for different toolchains. For the COFF
6534 targeted toolchain the default value is 8. A value of 64 is only allowed
6535 if the underlying ABI supports it.
6537 Specifying the larger number can produce faster, more efficient code, but
6538 can also increase the size of the program. Different values are potentially
6539 incompatible. Code compiled with one value cannot necessarily expect to
6540 work with code or libraries compiled with another value, if they exchange
6541 information using structures or unions.
6543 @item -mabort-on-noreturn
6544 @opindex mabort-on-noreturn
6545 Generate a call to the function @code{abort} at the end of a
6546 @code{noreturn} function. It will be executed if the function tries to
6550 @itemx -mno-long-calls
6551 @opindex mlong-calls
6552 @opindex mno-long-calls
6553 Tells the compiler to perform function calls by first loading the
6554 address of the function into a register and then performing a subroutine
6555 call on this register. This switch is needed if the target function
6556 will lie outside of the 64 megabyte addressing range of the offset based
6557 version of subroutine call instruction.
6559 Even if this switch is enabled, not all function calls will be turned
6560 into long calls. The heuristic is that static functions, functions
6561 which have the @samp{short-call} attribute, functions that are inside
6562 the scope of a @samp{#pragma no_long_calls} directive and functions whose
6563 definitions have already been compiled within the current compilation
6564 unit, will not be turned into long calls. The exception to this rule is
6565 that weak function definitions, functions with the @samp{long-call}
6566 attribute or the @samp{section} attribute, and functions that are within
6567 the scope of a @samp{#pragma long_calls} directive, will always be
6568 turned into long calls.
6570 This feature is not enabled by default. Specifying
6571 @option{-mno-long-calls} will restore the default behavior, as will
6572 placing the function calls within the scope of a @samp{#pragma
6573 long_calls_off} directive. Note these switches have no effect on how
6574 the compiler generates code to handle function calls via function
6577 @item -mnop-fun-dllimport
6578 @opindex mnop-fun-dllimport
6579 Disable support for the @code{dllimport} attribute.
6581 @item -msingle-pic-base
6582 @opindex msingle-pic-base
6583 Treat the register used for PIC addressing as read-only, rather than
6584 loading it in the prologue for each function. The run-time system is
6585 responsible for initializing this register with an appropriate value
6586 before execution begins.
6588 @item -mpic-register=@var{reg}
6589 @opindex mpic-register
6590 Specify the register to be used for PIC addressing. The default is R10
6591 unless stack-checking is enabled, when R9 is used.
6593 @item -mcirrus-fix-invalid-insns
6594 @opindex mcirrus-fix-invalid-insns
6595 @opindex mno-cirrus-fix-invalid-insns
6596 Insert NOPs into the instruction stream to in order to work around
6597 problems with invalid Maverick instruction combinations. This option
6598 is only valid if the @option{-mcpu=ep9312} option has been used to
6599 enable generation of instructions for the Cirrus Maverick floating
6600 point co-processor. This option is not enabled by default, since the
6601 problem is only present in older Maverick implementations. The default
6602 can be re-enabled by use of the @option{-mno-cirrus-fix-invalid-insns}
6605 @item -mpoke-function-name
6606 @opindex mpoke-function-name
6607 Write the name of each function into the text section, directly
6608 preceding the function prologue. The generated code is similar to this:
6612 .ascii "arm_poke_function_name", 0
6615 .word 0xff000000 + (t1 - t0)
6616 arm_poke_function_name
6618 stmfd sp!, @{fp, ip, lr, pc@}
6622 When performing a stack backtrace, code can inspect the value of
6623 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
6624 location @code{pc - 12} and the top 8 bits are set, then we know that
6625 there is a function name embedded immediately preceding this location
6626 and has length @code{((pc[-3]) & 0xff000000)}.
6630 Generate code for the 16-bit Thumb instruction set. The default is to
6631 use the 32-bit ARM instruction set.
6634 @opindex mtpcs-frame
6635 Generate a stack frame that is compliant with the Thumb Procedure Call
6636 Standard for all non-leaf functions. (A leaf function is one that does
6637 not call any other functions.) The default is @option{-mno-tpcs-frame}.
6639 @item -mtpcs-leaf-frame
6640 @opindex mtpcs-leaf-frame
6641 Generate a stack frame that is compliant with the Thumb Procedure Call
6642 Standard for all leaf functions. (A leaf function is one that does
6643 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
6645 @item -mcallee-super-interworking
6646 @opindex mcallee-super-interworking
6647 Gives all externally visible functions in the file being compiled an ARM
6648 instruction set header which switches to Thumb mode before executing the
6649 rest of the function. This allows these functions to be called from
6650 non-interworking code.
6652 @item -mcaller-super-interworking
6653 @opindex mcaller-super-interworking
6654 Allows calls via function pointers (including virtual functions) to
6655 execute correctly regardless of whether the target code has been
6656 compiled for interworking or not. There is a small overhead in the cost
6657 of executing a function pointer if this option is enabled.
6662 @subsection AVR Options
6665 These options are defined for AVR implementations:
6668 @item -mmcu=@var{mcu}
6670 Specify ATMEL AVR instruction set or MCU type.
6672 Instruction set avr1 is for the minimal AVR core, not supported by the C
6673 compiler, only for assembler programs (MCU types: at90s1200, attiny10,
6674 attiny11, attiny12, attiny15, attiny28).
6676 Instruction set avr2 (default) is for the classic AVR core with up to
6677 8K program memory space (MCU types: at90s2313, at90s2323, attiny22,
6678 at90s2333, at90s2343, at90s4414, at90s4433, at90s4434, at90s8515,
6679 at90c8534, at90s8535).
6681 Instruction set avr3 is for the classic AVR core with up to 128K program
6682 memory space (MCU types: atmega103, atmega603, at43usb320, at76c711).
6684 Instruction set avr4 is for the enhanced AVR core with up to 8K program
6685 memory space (MCU types: atmega8, atmega83, atmega85).
6687 Instruction set avr5 is for the enhanced AVR core with up to 128K program
6688 memory space (MCU types: atmega16, atmega161, atmega163, atmega32, atmega323,
6689 atmega64, atmega128, at43usb355, at94k).
6693 Output instruction sizes to the asm file.
6695 @item -minit-stack=@var{N}
6696 @opindex minit-stack
6697 Specify the initial stack address, which may be a symbol or numeric value,
6698 @samp{__stack} is the default.
6700 @item -mno-interrupts
6701 @opindex mno-interrupts
6702 Generated code is not compatible with hardware interrupts.
6703 Code size will be smaller.
6705 @item -mcall-prologues
6706 @opindex mcall-prologues
6707 Functions prologues/epilogues expanded as call to appropriate
6708 subroutines. Code size will be smaller.
6710 @item -mno-tablejump
6711 @opindex mno-tablejump
6712 Do not generate tablejump insns which sometimes increase code size.
6715 @opindex mtiny-stack
6716 Change only the low 8 bits of the stack pointer.
6720 Assume int to be 8 bit integer. This affects the sizes of all types: A
6721 char will be 1 byte, an int will be 1 byte, an long will be 2 bytes
6722 and long long will be 4 bytes. Please note that this option does not
6723 comply to the C standards, but it will provide you with smaller code
6728 @subsection CRIS Options
6729 @cindex CRIS Options
6731 These options are defined specifically for the CRIS ports.
6734 @item -march=@var{architecture-type}
6735 @itemx -mcpu=@var{architecture-type}
6738 Generate code for the specified architecture. The choices for
6739 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
6740 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX.
6741 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
6744 @item -mtune=@var{architecture-type}
6746 Tune to @var{architecture-type} everything applicable about the generated
6747 code, except for the ABI and the set of available instructions. The
6748 choices for @var{architecture-type} are the same as for
6749 @option{-march=@var{architecture-type}}.
6751 @item -mmax-stack-frame=@var{n}
6752 @opindex mmax-stack-frame
6753 Warn when the stack frame of a function exceeds @var{n} bytes.
6755 @item -melinux-stacksize=@var{n}
6756 @opindex melinux-stacksize
6757 Only available with the @samp{cris-axis-aout} target. Arranges for
6758 indications in the program to the kernel loader that the stack of the
6759 program should be set to @var{n} bytes.
6765 The options @option{-metrax4} and @option{-metrax100} are synonyms for
6766 @option{-march=v3} and @option{-march=v8} respectively.
6768 @item -mmul-bug-workaround
6769 @itemx -mno-mul-bug-workaround
6770 @opindex mmul-bug-workaround
6771 @opindex mno-mul-bug-workaround
6772 Work around a bug in the @code{muls} and @code{mulu} instructions for CPU
6773 models where it applies. This option is active by default.
6777 Enable CRIS-specific verbose debug-related information in the assembly
6778 code. This option also has the effect to turn off the @samp{#NO_APP}
6779 formatted-code indicator to the assembler at the beginning of the
6784 Do not use condition-code results from previous instruction; always emit
6785 compare and test instructions before use of condition codes.
6787 @item -mno-side-effects
6788 @opindex mno-side-effects
6789 Do not emit instructions with side-effects in addressing modes other than
6793 @itemx -mno-stack-align
6795 @itemx -mno-data-align
6796 @itemx -mconst-align
6797 @itemx -mno-const-align
6798 @opindex mstack-align
6799 @opindex mno-stack-align
6800 @opindex mdata-align
6801 @opindex mno-data-align
6802 @opindex mconst-align
6803 @opindex mno-const-align
6804 These options (no-options) arranges (eliminate arrangements) for the
6805 stack-frame, individual data and constants to be aligned for the maximum
6806 single data access size for the chosen CPU model. The default is to
6807 arrange for 32-bit alignment. ABI details such as structure layout are
6808 not affected by these options.
6816 Similar to the stack- data- and const-align options above, these options
6817 arrange for stack-frame, writable data and constants to all be 32-bit,
6818 16-bit or 8-bit aligned. The default is 32-bit alignment.
6820 @item -mno-prologue-epilogue
6821 @itemx -mprologue-epilogue
6822 @opindex mno-prologue-epilogue
6823 @opindex mprologue-epilogue
6824 With @option{-mno-prologue-epilogue}, the normal function prologue and
6825 epilogue that sets up the stack-frame are omitted and no return
6826 instructions or return sequences are generated in the code. Use this
6827 option only together with visual inspection of the compiled code: no
6828 warnings or errors are generated when call-saved registers must be saved,
6829 or storage for local variable needs to be allocated.
6835 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
6836 instruction sequences that load addresses for functions from the PLT part
6837 of the GOT rather than (traditional on other architectures) calls to the
6838 PLT. The default is @option{-mgotplt}.
6842 Legacy no-op option only recognized with the cris-axis-aout target.
6846 Legacy no-op option only recognized with the cris-axis-elf and
6847 cris-axis-linux-gnu targets.
6851 Only recognized with the cris-axis-aout target, where it selects a
6852 GNU/linux-like multilib, include files and instruction set for
6857 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
6861 This option, recognized for the cris-axis-aout and cris-axis-elf arranges
6862 to link with input-output functions from a simulator library. Code,
6863 initialized data and zero-initialized data are allocated consecutively.
6867 Like @option{-sim}, but pass linker options to locate initialized data at
6868 0x40000000 and zero-initialized data at 0x80000000.
6871 @node Darwin Options
6872 @subsection Darwin Options
6873 @cindex Darwin options
6875 These options are defined for all architectures running the Darwin operating
6876 system. They are useful for compatibility with other Mac OS compilers.
6881 Add the framework directory @var{dir} to the head of the list of
6882 directories to be searched for header files. These directories are
6883 interleaved with those specified by @option{-I} options and are
6884 scanned in a left-to-right order.
6886 A framework directory is a directory with frameworks in it. A
6887 framework is a directory with a @samp{"Headers"} and/or
6888 @samp{"PrivateHeaders"} directory contained directly in it that ends
6889 in @samp{".framework"}. The name of a framework is the name of this
6890 directory excluding the @samp{".framework"}. Headers associated with
6891 the framework are found in one of those two directories, with
6892 @samp{"Headers"} being searched first. A subframework is a framework
6893 directory that is in a framework's @samp{"Frameworks"} directory.
6894 Includes of subframework headers can only appear in a header of a
6895 framework that contains the subframework, or in a sibling subframework
6896 header. Two subframeworks are siblings if they occur in the same
6897 framework. A subframework should not have the same name as a
6898 framework, a warning will be issued if this is violated. Currently a
6899 subframework cannot have subframeworks, in the future, the mechanism
6900 may be extended to support this. The standard frameworks can be found
6901 in @samp{"/System/Library/Frameworks"}, @samp{"/Library/Frameworks"}
6902 and @samp{"/Local/Library/Frameworks"}. An example include looks like
6903 @code{#include <Framework/header.h>}, where @samp{Framework} denotes
6904 the name of the framework and header.h is found in the
6905 @samp{"PrivateHeaders"} or @samp{"Headers"} directory.
6909 Emit debugging information for symbols that are used. For STABS
6910 debugging format, this enables @option{-feliminate-unused-debug-symbols}.
6911 This is by default ON.
6915 Emit debugging information for all symbols and types.
6919 Loads all members of static archive libraries.
6920 See man ld(1) for more information.
6922 @item -arch_errors_fatal
6923 @opindex arch_errors_fatal
6924 Cause the errors having to do with files that have the wrong architecture
6928 @opindex bind_at_load
6929 Causes the output file to be marked such that the dynamic linker will
6930 bind all undefined references when the file is loaded or launched.
6934 Produce a Mach-o bundle format file.
6935 See man ld(1) for more information.
6937 @item -bundle_loader @var{executable}
6938 @opindex bundle_loader
6939 This specifies the @var{executable} that will be loading the build
6940 output file being linked. See man ld(1) for more information.
6942 @item -allowable_client @var{client_name}
6946 @itemx -compatibility_version
6947 @itemx -current_version
6948 @itemx -dependency-file
6950 @itemx -dylinker_install_name
6953 @itemx -exported_symbols_list
6955 @itemx -flat_namespace
6956 @itemx -force_cpusubtype_ALL
6957 @itemx -force_flat_namespace
6958 @itemx -headerpad_max_install_names
6961 @itemx -install_name
6962 @itemx -keep_private_externs
6963 @itemx -multi_module
6964 @itemx -multiply_defined
6965 @itemx -multiply_defined_unused
6967 @itemx -nofixprebinding
6970 @itemx -noseglinkedit
6971 @itemx -pagezero_size
6973 @itemx -prebind_all_twolevel_modules
6974 @itemx -private_bundle
6975 @itemx -read_only_relocs
6977 @itemx -sectobjectsymbols
6981 @itemx -sectobjectsymbols
6983 @itemx -seg_addr_table
6984 @itemx -seg_addr_table_filename
6987 @itemx -segs_read_only_addr
6988 @itemx -segs_read_write_addr
6989 @itemx -single_module
6992 @itemx -sub_umbrella
6993 @itemx -twolevel_namespace
6996 @itemx -unexported_symbols_list
6997 @itemx -weak_reference_mismatches
7000 @opindex allowable_client
7002 @opindex client_name
7003 @opindex compatibility_version
7004 @opindex current_version
7005 @opindex dependency-file
7007 @opindex dylinker_install_name
7010 @opindex exported_symbols_list
7012 @opindex flat_namespace
7013 @opindex force_cpusubtype_ALL
7014 @opindex force_flat_namespace
7015 @opindex headerpad_max_install_names
7018 @opindex install_name
7019 @opindex keep_private_externs
7020 @opindex multi_module
7021 @opindex multiply_defined
7022 @opindex multiply_defined_unused
7024 @opindex nofixprebinding
7025 @opindex nomultidefs
7027 @opindex noseglinkedit
7028 @opindex pagezero_size
7030 @opindex prebind_all_twolevel_modules
7031 @opindex private_bundle
7032 @opindex read_only_relocs
7034 @opindex sectobjectsymbols
7038 @opindex sectobjectsymbols
7040 @opindex seg_addr_table
7041 @opindex seg_addr_table_filename
7042 @opindex seglinkedit
7044 @opindex segs_read_only_addr
7045 @opindex segs_read_write_addr
7046 @opindex single_module
7048 @opindex sub_library
7049 @opindex sub_umbrella
7050 @opindex twolevel_namespace
7053 @opindex unexported_symbols_list
7054 @opindex weak_reference_mismatches
7055 @opindex whatsloaded
7057 These options are available for Darwin linker. Darwin linker man page
7058 describes them in detail.
7061 @node DEC Alpha Options
7062 @subsection DEC Alpha Options
7064 These @samp{-m} options are defined for the DEC Alpha implementations:
7067 @item -mno-soft-float
7069 @opindex mno-soft-float
7070 @opindex msoft-float
7071 Use (do not use) the hardware floating-point instructions for
7072 floating-point operations. When @option{-msoft-float} is specified,
7073 functions in @file{libgcc.a} will be used to perform floating-point
7074 operations. Unless they are replaced by routines that emulate the
7075 floating-point operations, or compiled in such a way as to call such
7076 emulations routines, these routines will issue floating-point
7077 operations. If you are compiling for an Alpha without floating-point
7078 operations, you must ensure that the library is built so as not to call
7081 Note that Alpha implementations without floating-point operations are
7082 required to have floating-point registers.
7087 @opindex mno-fp-regs
7088 Generate code that uses (does not use) the floating-point register set.
7089 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
7090 register set is not used, floating point operands are passed in integer
7091 registers as if they were integers and floating-point results are passed
7092 in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence,
7093 so any function with a floating-point argument or return value called by code
7094 compiled with @option{-mno-fp-regs} must also be compiled with that
7097 A typical use of this option is building a kernel that does not use,
7098 and hence need not save and restore, any floating-point registers.
7102 The Alpha architecture implements floating-point hardware optimized for
7103 maximum performance. It is mostly compliant with the IEEE floating
7104 point standard. However, for full compliance, software assistance is
7105 required. This option generates code fully IEEE compliant code
7106 @emph{except} that the @var{inexact-flag} is not maintained (see below).
7107 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
7108 defined during compilation. The resulting code is less efficient but is
7109 able to correctly support denormalized numbers and exceptional IEEE
7110 values such as not-a-number and plus/minus infinity. Other Alpha
7111 compilers call this option @option{-ieee_with_no_inexact}.
7113 @item -mieee-with-inexact
7114 @opindex mieee-with-inexact
7115 This is like @option{-mieee} except the generated code also maintains
7116 the IEEE @var{inexact-flag}. Turning on this option causes the
7117 generated code to implement fully-compliant IEEE math. In addition to
7118 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
7119 macro. On some Alpha implementations the resulting code may execute
7120 significantly slower than the code generated by default. Since there is
7121 very little code that depends on the @var{inexact-flag}, you should
7122 normally not specify this option. Other Alpha compilers call this
7123 option @option{-ieee_with_inexact}.
7125 @item -mfp-trap-mode=@var{trap-mode}
7126 @opindex mfp-trap-mode
7127 This option controls what floating-point related traps are enabled.
7128 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
7129 The trap mode can be set to one of four values:
7133 This is the default (normal) setting. The only traps that are enabled
7134 are the ones that cannot be disabled in software (e.g., division by zero
7138 In addition to the traps enabled by @samp{n}, underflow traps are enabled
7142 Like @samp{su}, but the instructions are marked to be safe for software
7143 completion (see Alpha architecture manual for details).
7146 Like @samp{su}, but inexact traps are enabled as well.
7149 @item -mfp-rounding-mode=@var{rounding-mode}
7150 @opindex mfp-rounding-mode
7151 Selects the IEEE rounding mode. Other Alpha compilers call this option
7152 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
7157 Normal IEEE rounding mode. Floating point numbers are rounded towards
7158 the nearest machine number or towards the even machine number in case
7162 Round towards minus infinity.
7165 Chopped rounding mode. Floating point numbers are rounded towards zero.
7168 Dynamic rounding mode. A field in the floating point control register
7169 (@var{fpcr}, see Alpha architecture reference manual) controls the
7170 rounding mode in effect. The C library initializes this register for
7171 rounding towards plus infinity. Thus, unless your program modifies the
7172 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
7175 @item -mtrap-precision=@var{trap-precision}
7176 @opindex mtrap-precision
7177 In the Alpha architecture, floating point traps are imprecise. This
7178 means without software assistance it is impossible to recover from a
7179 floating trap and program execution normally needs to be terminated.
7180 GCC can generate code that can assist operating system trap handlers
7181 in determining the exact location that caused a floating point trap.
7182 Depending on the requirements of an application, different levels of
7183 precisions can be selected:
7187 Program precision. This option is the default and means a trap handler
7188 can only identify which program caused a floating point exception.
7191 Function precision. The trap handler can determine the function that
7192 caused a floating point exception.
7195 Instruction precision. The trap handler can determine the exact
7196 instruction that caused a floating point exception.
7199 Other Alpha compilers provide the equivalent options called
7200 @option{-scope_safe} and @option{-resumption_safe}.
7202 @item -mieee-conformant
7203 @opindex mieee-conformant
7204 This option marks the generated code as IEEE conformant. You must not
7205 use this option unless you also specify @option{-mtrap-precision=i} and either
7206 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
7207 is to emit the line @samp{.eflag 48} in the function prologue of the
7208 generated assembly file. Under DEC Unix, this has the effect that
7209 IEEE-conformant math library routines will be linked in.
7211 @item -mbuild-constants
7212 @opindex mbuild-constants
7213 Normally GCC examines a 32- or 64-bit integer constant to
7214 see if it can construct it from smaller constants in two or three
7215 instructions. If it cannot, it will output the constant as a literal and
7216 generate code to load it from the data segment at runtime.
7218 Use this option to require GCC to construct @emph{all} integer constants
7219 using code, even if it takes more instructions (the maximum is six).
7221 You would typically use this option to build a shared library dynamic
7222 loader. Itself a shared library, it must relocate itself in memory
7223 before it can find the variables and constants in its own data segment.
7229 Select whether to generate code to be assembled by the vendor-supplied
7230 assembler (@option{-malpha-as}) or by the GNU assembler @option{-mgas}.
7248 Indicate whether GCC should generate code to use the optional BWX,
7249 CIX, FIX and MAX instruction sets. The default is to use the instruction
7250 sets supported by the CPU type specified via @option{-mcpu=} option or that
7251 of the CPU on which GCC was built if none was specified.
7256 @opindex mfloat-ieee
7257 Generate code that uses (does not use) VAX F and G floating point
7258 arithmetic instead of IEEE single and double precision.
7260 @item -mexplicit-relocs
7261 @itemx -mno-explicit-relocs
7262 @opindex mexplicit-relocs
7263 @opindex mno-explicit-relocs
7264 Older Alpha assemblers provided no way to generate symbol relocations
7265 except via assembler macros. Use of these macros does not allow
7266 optimal instruction scheduling. GNU binutils as of version 2.12
7267 supports a new syntax that allows the compiler to explicitly mark
7268 which relocations should apply to which instructions. This option
7269 is mostly useful for debugging, as GCC detects the capabilities of
7270 the assembler when it is built and sets the default accordingly.
7274 @opindex msmall-data
7275 @opindex mlarge-data
7276 When @option{-mexplicit-relocs} is in effect, static data is
7277 accessed via @dfn{gp-relative} relocations. When @option{-msmall-data}
7278 is used, objects 8 bytes long or smaller are placed in a @dfn{small data area}
7279 (the @code{.sdata} and @code{.sbss} sections) and are accessed via
7280 16-bit relocations off of the @code{$gp} register. This limits the
7281 size of the small data area to 64KB, but allows the variables to be
7282 directly accessed via a single instruction.
7284 The default is @option{-mlarge-data}. With this option the data area
7285 is limited to just below 2GB. Programs that require more than 2GB of
7286 data must use @code{malloc} or @code{mmap} to allocate the data in the
7287 heap instead of in the program's data segment.
7289 When generating code for shared libraries, @option{-fpic} implies
7290 @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}.
7294 @opindex msmall-text
7295 @opindex mlarge-text
7296 When @option{-msmall-text} is used, the compiler assumes that the
7297 code of the entire program (or shared library) fits in 4MB, and is
7298 thus reachable with a branch instruction. When @option{-msmall-data}
7299 is used, the compiler can assume that all local symbols share the
7300 same @code{$gp} value, and thus reduce the number of instructions
7301 required for a function call from 4 to 1.
7303 The default is @option{-mlarge-text}.
7305 @item -mcpu=@var{cpu_type}
7307 Set the instruction set and instruction scheduling parameters for
7308 machine type @var{cpu_type}. You can specify either the @samp{EV}
7309 style name or the corresponding chip number. GCC supports scheduling
7310 parameters for the EV4, EV5 and EV6 family of processors and will
7311 choose the default values for the instruction set from the processor
7312 you specify. If you do not specify a processor type, GCC will default
7313 to the processor on which the compiler was built.
7315 Supported values for @var{cpu_type} are
7321 Schedules as an EV4 and has no instruction set extensions.
7325 Schedules as an EV5 and has no instruction set extensions.
7329 Schedules as an EV5 and supports the BWX extension.
7334 Schedules as an EV5 and supports the BWX and MAX extensions.
7338 Schedules as an EV6 and supports the BWX, FIX, and MAX extensions.
7342 Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions.
7345 @item -mtune=@var{cpu_type}
7347 Set only the instruction scheduling parameters for machine type
7348 @var{cpu_type}. The instruction set is not changed.
7350 @item -mmemory-latency=@var{time}
7351 @opindex mmemory-latency
7352 Sets the latency the scheduler should assume for typical memory
7353 references as seen by the application. This number is highly
7354 dependent on the memory access patterns used by the application
7355 and the size of the external cache on the machine.
7357 Valid options for @var{time} are
7361 A decimal number representing clock cycles.
7367 The compiler contains estimates of the number of clock cycles for
7368 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
7369 (also called Dcache, Scache, and Bcache), as well as to main memory.
7370 Note that L3 is only valid for EV5.
7375 @node DEC Alpha/VMS Options
7376 @subsection DEC Alpha/VMS Options
7378 These @samp{-m} options are defined for the DEC Alpha/VMS implementations:
7381 @item -mvms-return-codes
7382 @opindex mvms-return-codes
7383 Return VMS condition codes from main. The default is to return POSIX
7384 style condition (e.g.@ error) codes.
7388 @subsection FRV Options
7395 Only use the first 32 general purpose registers.
7400 Use all 64 general purpose registers.
7405 Use only the first 32 floating point registers.
7410 Use all 64 floating point registers
7413 @opindex mhard-float
7415 Use hardware instructions for floating point operations.
7418 @opindex msoft-float
7420 Use library routines for floating point operations.
7425 Dynamically allocate condition code registers.
7430 Do not try to dynamically allocate condition code registers, only
7431 use @code{icc0} and @code{fcc0}.
7436 Change ABI to use double word insns.
7441 Do not use double word instructions.
7446 Use floating point double instructions.
7451 Do not use floating point double instructions.
7456 Use media instructions.
7461 Do not use media instructions.
7466 Use multiply and add/subtract instructions.
7471 Do not use multiply and add/subtract instructions.
7474 @opindex mlibrary-pic
7476 Generate position-independent EABI code.
7481 Use only the first four media accumulator registers.
7486 Use all eight media accumulator registers.
7491 Pack VLIW instructions.
7496 Do not pack VLIW instructions.
7501 Do not mark ABI switches in e_flags.
7506 Enable the use of conditional-move instructions (default).
7508 This switch is mainly for debugging the compiler and will likely be removed
7509 in a future version.
7511 @item -mno-cond-move
7512 @opindex mno-cond-move
7514 Disable the use of conditional-move instructions.
7516 This switch is mainly for debugging the compiler and will likely be removed
7517 in a future version.
7522 Enable the use of conditional set instructions (default).
7524 This switch is mainly for debugging the compiler and will likely be removed
7525 in a future version.
7530 Disable the use of conditional set instructions.
7532 This switch is mainly for debugging the compiler and will likely be removed
7533 in a future version.
7538 Enable the use of conditional execution (default).
7540 This switch is mainly for debugging the compiler and will likely be removed
7541 in a future version.
7543 @item -mno-cond-exec
7544 @opindex mno-cond-exec
7546 Disable the use of conditional execution.
7548 This switch is mainly for debugging the compiler and will likely be removed
7549 in a future version.
7552 @opindex mvliw-branch
7554 Run a pass to pack branches into VLIW instructions (default).
7556 This switch is mainly for debugging the compiler and will likely be removed
7557 in a future version.
7559 @item -mno-vliw-branch
7560 @opindex mno-vliw-branch
7562 Do not run a pass to pack branches into VLIW instructions.
7564 This switch is mainly for debugging the compiler and will likely be removed
7565 in a future version.
7567 @item -mmulti-cond-exec
7568 @opindex mmulti-cond-exec
7570 Enable optimization of @code{&&} and @code{||} in conditional execution
7573 This switch is mainly for debugging the compiler and will likely be removed
7574 in a future version.
7576 @item -mno-multi-cond-exec
7577 @opindex mno-multi-cond-exec
7579 Disable optimization of @code{&&} and @code{||} in conditional execution.
7581 This switch is mainly for debugging the compiler and will likely be removed
7582 in a future version.
7584 @item -mnested-cond-exec
7585 @opindex mnested-cond-exec
7587 Enable nested conditional execution optimizations (default).
7589 This switch is mainly for debugging the compiler and will likely be removed
7590 in a future version.
7592 @item -mno-nested-cond-exec
7593 @opindex mno-nested-cond-exec
7595 Disable nested conditional execution optimizations.
7597 This switch is mainly for debugging the compiler and will likely be removed
7598 in a future version.
7600 @item -mtomcat-stats
7601 @opindex mtomcat-stats
7603 Cause gas to print out tomcat statistics.
7605 @item -mcpu=@var{cpu}
7608 Select the processor type for which to generate code. Possible values are
7609 @samp{simple}, @samp{tomcat}, @samp{fr500}, @samp{fr400}, @samp{fr300},
7614 @node H8/300 Options
7615 @subsection H8/300 Options
7617 These @samp{-m} options are defined for the H8/300 implementations:
7622 Shorten some address references at link time, when possible; uses the
7623 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
7624 ld, Using ld}, for a fuller description.
7628 Generate code for the H8/300H@.
7632 Generate code for the H8S@.
7636 Generate code for the H8S and H8/300H in the normal mode. This switch
7637 must be used either with -mh or -ms.
7641 Generate code for the H8S/2600. This switch must be used with @option{-ms}.
7645 Make @code{int} data 32 bits by default.
7649 On the H8/300H and H8S, use the same alignment rules as for the H8/300.
7650 The default for the H8/300H and H8S is to align longs and floats on 4
7652 @option{-malign-300} causes them to be aligned on 2 byte boundaries.
7653 This option has no effect on the H8/300.
7657 @subsection HPPA Options
7658 @cindex HPPA Options
7660 These @samp{-m} options are defined for the HPPA family of computers:
7663 @item -march=@var{architecture-type}
7665 Generate code for the specified architecture. The choices for
7666 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
7667 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
7668 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
7669 architecture option for your machine. Code compiled for lower numbered
7670 architectures will run on higher numbered architectures, but not the
7673 PA 2.0 support currently requires gas snapshot 19990413 or later. The
7674 next release of binutils (current is 2.9.1) will probably contain PA 2.0
7678 @itemx -mpa-risc-1-1
7679 @itemx -mpa-risc-2-0
7680 @opindex mpa-risc-1-0
7681 @opindex mpa-risc-1-1
7682 @opindex mpa-risc-2-0
7683 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
7686 @opindex mbig-switch
7687 Generate code suitable for big switch tables. Use this option only if
7688 the assembler/linker complain about out of range branches within a switch
7691 @item -mjump-in-delay
7692 @opindex mjump-in-delay
7693 Fill delay slots of function calls with unconditional jump instructions
7694 by modifying the return pointer for the function call to be the target
7695 of the conditional jump.
7697 @item -mdisable-fpregs
7698 @opindex mdisable-fpregs
7699 Prevent floating point registers from being used in any manner. This is
7700 necessary for compiling kernels which perform lazy context switching of
7701 floating point registers. If you use this option and attempt to perform
7702 floating point operations, the compiler will abort.
7704 @item -mdisable-indexing
7705 @opindex mdisable-indexing
7706 Prevent the compiler from using indexing address modes. This avoids some
7707 rather obscure problems when compiling MIG generated code under MACH@.
7709 @item -mno-space-regs
7710 @opindex mno-space-regs
7711 Generate code that assumes the target has no space registers. This allows
7712 GCC to generate faster indirect calls and use unscaled index address modes.
7714 Such code is suitable for level 0 PA systems and kernels.
7716 @item -mfast-indirect-calls
7717 @opindex mfast-indirect-calls
7718 Generate code that assumes calls never cross space boundaries. This
7719 allows GCC to emit code which performs faster indirect calls.
7721 This option will not work in the presence of shared libraries or nested
7724 @item -mfixed-range=@var{register-range}
7725 @opindex mfixed-range
7726 Generate code treating the given register range as fixed registers.
7727 A fixed register is one that the register allocator can not use. This is
7728 useful when compiling kernel code. A register range is specified as
7729 two registers separated by a dash. Multiple register ranges can be
7730 specified separated by a comma.
7732 @item -mlong-load-store
7733 @opindex mlong-load-store
7734 Generate 3-instruction load and store sequences as sometimes required by
7735 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
7738 @item -mportable-runtime
7739 @opindex mportable-runtime
7740 Use the portable calling conventions proposed by HP for ELF systems.
7744 Enable the use of assembler directives only GAS understands.
7746 @item -mschedule=@var{cpu-type}
7748 Schedule code according to the constraints for the machine type
7749 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
7750 @samp{7100}, @samp{7100LC}, @samp{7200}, @samp{7300} and @samp{8000}. Refer
7751 to @file{/usr/lib/sched.models} on an HP-UX system to determine the
7752 proper scheduling option for your machine. The default scheduling is
7756 @opindex mlinker-opt
7757 Enable the optimization pass in the HP-UX linker. Note this makes symbolic
7758 debugging impossible. It also triggers a bug in the HP-UX 8 and HP-UX 9
7759 linkers in which they give bogus error messages when linking some programs.
7762 @opindex msoft-float
7763 Generate output containing library calls for floating point.
7764 @strong{Warning:} the requisite libraries are not available for all HPPA
7765 targets. Normally the facilities of the machine's usual C compiler are
7766 used, but this cannot be done directly in cross-compilation. You must make
7767 your own arrangements to provide suitable library functions for
7768 cross-compilation. The embedded target @samp{hppa1.1-*-pro}
7769 does provide software floating point support.
7771 @option{-msoft-float} changes the calling convention in the output file;
7772 therefore, it is only useful if you compile @emph{all} of a program with
7773 this option. In particular, you need to compile @file{libgcc.a}, the
7774 library that comes with GCC, with @option{-msoft-float} in order for
7779 Generate the predefine, @code{_SIO}, for server IO. The default is
7780 @option{-mwsio}. This generates the predefines, @code{__hp9000s700},
7781 @code{__hp9000s700__} and @code{_WSIO}, for workstation IO. These
7782 options are available under HP-UX and HI-UX.
7786 Use GNU ld specific options. This passes @option{-shared} to ld when
7787 building a shared library. It is the default when GCC is configured,
7788 explicitly or implicitly, with the GNU linker. This option does not
7789 have any affect on which ld is called, it only changes what parameters
7790 are passed to that ld. The ld that is called is determined by the
7791 @option{--with-ld} configure option, GCC's program search path, and
7792 finally by the user's @env{PATH}. The linker used by GCC can be printed
7793 using @samp{which `gcc -print-prog-name=ld`}.
7797 Use HP ld specific options. This passes @option{-b} to ld when building
7798 a shared library and passes @option{+Accept TypeMismatch} to ld on all
7799 links. It is the default when GCC is configured, explicitly or
7800 implicitly, with the HP linker. This option does not have any affect on
7801 which ld is called, it only changes what parameters are passed to that
7802 ld. The ld that is called is determined by the @option{--with-ld}
7803 configure option, GCC's program search path, and finally by the user's
7804 @env{PATH}. The linker used by GCC can be printed using @samp{which
7805 `gcc -print-prog-name=ld`}.
7810 Select the FDPIC ABI, that uses function descriptors to represent
7811 pointers to functions. Without any PIC/PIE-related options, it
7812 implies @option{-fPIE}. With @option{-fpic} or @option{-fpie}, it
7813 assumes GOT entries and small data are within a 12-bit range from the
7814 GOT base address; with @option{-fPIC} or @option{-fPIE}, GOT offsets
7815 are computed with 32 bits.
7818 @opindex minline-plt
7820 Enable inlining of PLT entries in function calls to functions that are
7821 not known to bind locally. It has no effect without @option{-mfdpic}.
7822 It's enabled by default if optimizing for speed and compiling for
7823 shared libraries (i.e., @option{-fPIC} or @option{-fpic}), or when an
7824 optimization option such as @option{-O3} or above is present in the
7830 Enable the use of @code{GPREL} relocations in the FDPIC ABI for data
7831 that is known to be in read-only sections. It's enabled by default,
7832 except for @option{-fpic} or @option{-fpie}: even though it may help
7833 make the global offset table smaller, it trades 1 instruction for 4.
7834 With @option{-fPIC} or @option{-fPIE}, it trades 3 instructions for 4,
7835 one of which may be shared by multiple symbols, and it avoids the need
7836 for a GOT entry for the referenced symbol, so it's more likely to be a
7837 win. If it is not, @option{-mno-gprel-ro} can be used to disable it.
7839 @item -multilib-library-pic
7840 @opindex multilib-library-pic
7842 Link with the (library, not FD) pic libraries. It's implied by
7843 @option{-mlibrary-pic}, as well as by @option{-fPIC} and
7844 @option{-fpic} without @option{-mfdpic}. You should never have to use
7850 Follow the EABI requirement of always creating a frame pointer whenever
7851 a stack frame is allocated. This option is enabled by default and can
7852 be disabled with @option{-mno-linked-fp}.
7855 @opindex mno-long-calls
7856 Generate code that uses long call sequences. This ensures that a call
7857 is always able to reach linker generated stubs. The default is to generate
7858 long calls only when the distance from the call site to the beginning
7859 of the function or translation unit, as the case may be, exceeds a
7860 predefined limit set by the branch type being used. The limits for
7861 normal calls are 7,600,000 and 240,000 bytes, respectively for the
7862 PA 2.0 and PA 1.X architectures. Sibcalls are always limited at
7865 Distances are measured from the beginning of functions when using the
7866 @option{-ffunction-sections} option, or when using the @option{-mgas}
7867 and @option{-mno-portable-runtime} options together under HP-UX with
7870 It is normally not desirable to use this option as it will degrade
7871 performance. However, it may be useful in large applications,
7872 particularly when partial linking is used to build the application.
7874 The types of long calls used depends on the capabilities of the
7875 assembler and linker, and the type of code being generated. The
7876 impact on systems that support long absolute calls, and long pic
7877 symbol-difference or pc-relative calls should be relatively small.
7878 However, an indirect call is used on 32-bit ELF systems in pic code
7879 and it is quite long.
7883 Suppress the generation of link options to search libdld.sl when the
7884 @option{-static} option is specified on HP-UX 10 and later.
7888 The HP-UX implementation of setlocale in libc has a dependency on
7889 libdld.sl. There isn't an archive version of libdld.sl. Thus,
7890 when the @option{-static} option is specified, special link options
7891 are needed to resolve this dependency.
7893 On HP-UX 10 and later, the GCC driver adds the necessary options to
7894 link with libdld.sl when the @option{-static} option is specified.
7895 This causes the resulting binary to be dynamic. On the 64-bit port,
7896 the linkers generate dynamic binaries by default in any case. The
7897 @option{-nolibdld} option can be used to prevent the GCC driver from
7898 adding these link options.
7902 Add support for multithreading with the @dfn{dce thread} library
7903 under HP-UX. This option sets flags for both the preprocessor and
7907 @node i386 and x86-64 Options
7908 @subsection Intel 386 and AMD x86-64 Options
7909 @cindex i386 Options
7910 @cindex x86-64 Options
7911 @cindex Intel 386 Options
7912 @cindex AMD x86-64 Options
7914 These @samp{-m} options are defined for the i386 and x86-64 family of
7918 @item -mtune=@var{cpu-type}
7920 Tune to @var{cpu-type} everything applicable about the generated code, except
7921 for the ABI and the set of available instructions. The choices for
7925 Original Intel's i386 CPU.
7927 Intel's i486 CPU. (No scheduling is implemented for this chip.)
7929 Intel Pentium CPU with no MMX support.
7931 Intel PentiumMMX CPU based on Pentium core with MMX instruction set support.
7932 @item i686, pentiumpro
7933 Intel PentiumPro CPU.
7935 Intel Pentium2 CPU based on PentiumPro core with MMX instruction set support.
7936 @item pentium3, pentium3m
7937 Intel Pentium3 CPU based on PentiumPro core with MMX and SSE instruction set
7940 Low power version of Intel Pentium3 CPU with MMX, SSE and SSE2 instruction set
7941 support. Used by Centrino notebooks.
7942 @item pentium4, pentium4m
7943 Intel Pentium4 CPU with MMX, SSE and SSE2 instruction set support.
7945 Improved version of Intel Pentium4 CPU with MMX, SSE, SSE2 and SSE3 instruction
7948 Improved version of Intel Pentium4 CPU with 64-bit extensions, MMX, SSE,
7949 SSE2 and SSE3 instruction set support.
7951 AMD K6 CPU with MMX instruction set support.
7953 Improved versions of AMD K6 CPU with MMX and 3dNOW! instruction set support.
7954 @item athlon, athlon-tbird
7955 AMD Athlon CPU with MMX, 3dNOW!, enhanced 3dNOW! and SSE prefetch instructions
7957 @item athlon-4, athlon-xp, athlon-mp
7958 Improved AMD Athlon CPU with MMX, 3dNOW!, enhanced 3dNOW! and full SSE
7959 instruction set support.
7960 @item k8, opteron, athlon64, athlon-fx
7961 AMD K8 core based CPUs with x86-64 instruction set support. (This supersets
7962 MMX, SSE, SSE2, 3dNOW!, enhanced 3dNOW! and 64-bit instruction set extensions.)
7964 IDT Winchip C6 CPU, dealt in same way as i486 with additional MMX instruction
7967 IDT Winchip2 CPU, dealt in same way as i486 with additional MMX and 3dNOW!
7968 instruction set support.
7970 Via C3 CPU with MMX and 3dNOW! instruction set support. (No scheduling is
7971 implemented for this chip.)
7973 Via C3-2 CPU with MMX and SSE instruction set support. (No scheduling is
7974 implemented for this chip.)
7977 While picking a specific @var{cpu-type} will schedule things appropriately
7978 for that particular chip, the compiler will not generate any code that
7979 does not run on the i386 without the @option{-march=@var{cpu-type}} option
7982 @item -march=@var{cpu-type}
7984 Generate instructions for the machine type @var{cpu-type}. The choices
7985 for @var{cpu-type} are the same as for @option{-mtune}. Moreover,
7986 specifying @option{-march=@var{cpu-type}} implies @option{-mtune=@var{cpu-type}}.
7988 @item -mcpu=@var{cpu-type}
7990 A deprecated synonym for @option{-mtune}.
7999 @opindex mpentiumpro
8000 These options are synonyms for @option{-mtune=i386}, @option{-mtune=i486},
8001 @option{-mtune=pentium}, and @option{-mtune=pentiumpro} respectively.
8002 These synonyms are deprecated.
8004 @item -mfpmath=@var{unit}
8006 Generate floating point arithmetics for selected unit @var{unit}. The choices
8011 Use the standard 387 floating point coprocessor present majority of chips and
8012 emulated otherwise. Code compiled with this option will run almost everywhere.
8013 The temporary results are computed in 80bit precision instead of precision
8014 specified by the type resulting in slightly different results compared to most
8015 of other chips. See @option{-ffloat-store} for more detailed description.
8017 This is the default choice for i386 compiler.
8020 Use scalar floating point instructions present in the SSE instruction set.
8021 This instruction set is supported by Pentium3 and newer chips, in the AMD line
8022 by Athlon-4, Athlon-xp and Athlon-mp chips. The earlier version of SSE
8023 instruction set supports only single precision arithmetics, thus the double and
8024 extended precision arithmetics is still done using 387. Later version, present
8025 only in Pentium4 and the future AMD x86-64 chips supports double precision
8028 For i387 you need to use @option{-march=@var{cpu-type}}, @option{-msse} or
8029 @option{-msse2} switches to enable SSE extensions and make this option
8030 effective. For x86-64 compiler, these extensions are enabled by default.
8032 The resulting code should be considerably faster in the majority of cases and avoid
8033 the numerical instability problems of 387 code, but may break some existing
8034 code that expects temporaries to be 80bit.
8036 This is the default choice for the x86-64 compiler.
8039 Attempt to utilize both instruction sets at once. This effectively double the
8040 amount of available registers and on chips with separate execution units for
8041 387 and SSE the execution resources too. Use this option with care, as it is
8042 still experimental, because the GCC register allocator does not model separate
8043 functional units well resulting in instable performance.
8046 @item -masm=@var{dialect}
8047 @opindex masm=@var{dialect}
8048 Output asm instructions using selected @var{dialect}. Supported choices are
8049 @samp{intel} or @samp{att} (the default one).
8054 @opindex mno-ieee-fp
8055 Control whether or not the compiler uses IEEE floating point
8056 comparisons. These handle correctly the case where the result of a
8057 comparison is unordered.
8060 @opindex msoft-float
8061 Generate output containing library calls for floating point.
8062 @strong{Warning:} the requisite libraries are not part of GCC@.
8063 Normally the facilities of the machine's usual C compiler are used, but
8064 this can't be done directly in cross-compilation. You must make your
8065 own arrangements to provide suitable library functions for
8068 On machines where a function returns floating point results in the 80387
8069 register stack, some floating point opcodes may be emitted even if
8070 @option{-msoft-float} is used.
8072 @item -mno-fp-ret-in-387
8073 @opindex mno-fp-ret-in-387
8074 Do not use the FPU registers for return values of functions.
8076 The usual calling convention has functions return values of types
8077 @code{float} and @code{double} in an FPU register, even if there
8078 is no FPU@. The idea is that the operating system should emulate
8081 The option @option{-mno-fp-ret-in-387} causes such values to be returned
8082 in ordinary CPU registers instead.
8084 @item -mno-fancy-math-387
8085 @opindex mno-fancy-math-387
8086 Some 387 emulators do not support the @code{sin}, @code{cos} and
8087 @code{sqrt} instructions for the 387. Specify this option to avoid
8088 generating those instructions. This option is the default on FreeBSD,
8089 OpenBSD and NetBSD@. This option is overridden when @option{-march}
8090 indicates that the target cpu will always have an FPU and so the
8091 instruction will not need emulation. As of revision 2.6.1, these
8092 instructions are not generated unless you also use the
8093 @option{-funsafe-math-optimizations} switch.
8095 @item -malign-double
8096 @itemx -mno-align-double
8097 @opindex malign-double
8098 @opindex mno-align-double
8099 Control whether GCC aligns @code{double}, @code{long double}, and
8100 @code{long long} variables on a two word boundary or a one word
8101 boundary. Aligning @code{double} variables on a two word boundary will
8102 produce code that runs somewhat faster on a @samp{Pentium} at the
8103 expense of more memory.
8105 @strong{Warning:} if you use the @option{-malign-double} switch,
8106 structures containing the above types will be aligned differently than
8107 the published application binary interface specifications for the 386
8108 and will not be binary compatible with structures in code compiled
8109 without that switch.
8111 @item -m96bit-long-double
8112 @itemx -m128bit-long-double
8113 @opindex m96bit-long-double
8114 @opindex m128bit-long-double
8115 These switches control the size of @code{long double} type. The i386
8116 application binary interface specifies the size to be 96 bits,
8117 so @option{-m96bit-long-double} is the default in 32 bit mode.
8119 Modern architectures (Pentium and newer) would prefer @code{long double}
8120 to be aligned to an 8 or 16 byte boundary. In arrays or structures
8121 conforming to the ABI, this would not be possible. So specifying a
8122 @option{-m128bit-long-double} will align @code{long double}
8123 to a 16 byte boundary by padding the @code{long double} with an additional
8126 In the x86-64 compiler, @option{-m128bit-long-double} is the default choice as
8127 its ABI specifies that @code{long double} is to be aligned on 16 byte boundary.
8129 Notice that neither of these options enable any extra precision over the x87
8130 standard of 80 bits for a @code{long double}.
8132 @strong{Warning:} if you override the default value for your target ABI, the
8133 structures and arrays containing @code{long double} variables will change
8134 their size as well as function calling convention for function taking
8135 @code{long double} will be modified. Hence they will not be binary
8136 compatible with arrays or structures in code compiled without that switch.
8140 @itemx -mno-svr3-shlib
8141 @opindex msvr3-shlib
8142 @opindex mno-svr3-shlib
8143 Control whether GCC places uninitialized local variables into the
8144 @code{bss} or @code{data} segments. @option{-msvr3-shlib} places them
8145 into @code{bss}. These options are meaningful only on System V Release 3.
8149 Use a different function-calling convention, in which functions that
8150 take a fixed number of arguments return with the @code{ret} @var{num}
8151 instruction, which pops their arguments while returning. This saves one
8152 instruction in the caller since there is no need to pop the arguments
8155 You can specify that an individual function is called with this calling
8156 sequence with the function attribute @samp{stdcall}. You can also
8157 override the @option{-mrtd} option by using the function attribute
8158 @samp{cdecl}. @xref{Function Attributes}.
8160 @strong{Warning:} this calling convention is incompatible with the one
8161 normally used on Unix, so you cannot use it if you need to call
8162 libraries compiled with the Unix compiler.
8164 Also, you must provide function prototypes for all functions that
8165 take variable numbers of arguments (including @code{printf});
8166 otherwise incorrect code will be generated for calls to those
8169 In addition, seriously incorrect code will result if you call a
8170 function with too many arguments. (Normally, extra arguments are
8171 harmlessly ignored.)
8173 @item -mregparm=@var{num}
8175 Control how many registers are used to pass integer arguments. By
8176 default, no registers are used to pass arguments, and at most 3
8177 registers can be used. You can control this behavior for a specific
8178 function by using the function attribute @samp{regparm}.
8179 @xref{Function Attributes}.
8181 @strong{Warning:} if you use this switch, and
8182 @var{num} is nonzero, then you must build all modules with the same
8183 value, including any libraries. This includes the system libraries and
8186 @item -mpreferred-stack-boundary=@var{num}
8187 @opindex mpreferred-stack-boundary
8188 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
8189 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
8190 the default is 4 (16 bytes or 128 bits), except when optimizing for code
8191 size (@option{-Os}), in which case the default is the minimum correct
8192 alignment (4 bytes for x86, and 8 bytes for x86-64).
8194 On Pentium and PentiumPro, @code{double} and @code{long double} values
8195 should be aligned to an 8 byte boundary (see @option{-malign-double}) or
8196 suffer significant run time performance penalties. On Pentium III, the
8197 Streaming SIMD Extension (SSE) data type @code{__m128} suffers similar
8198 penalties if it is not 16 byte aligned.
8200 To ensure proper alignment of this values on the stack, the stack boundary
8201 must be as aligned as that required by any value stored on the stack.
8202 Further, every function must be generated such that it keeps the stack
8203 aligned. Thus calling a function compiled with a higher preferred
8204 stack boundary from a function compiled with a lower preferred stack
8205 boundary will most likely misalign the stack. It is recommended that
8206 libraries that use callbacks always use the default setting.
8208 This extra alignment does consume extra stack space, and generally
8209 increases code size. Code that is sensitive to stack space usage, such
8210 as embedded systems and operating system kernels, may want to reduce the
8211 preferred alignment to @option{-mpreferred-stack-boundary=2}.
8229 These switches enable or disable the use of built-in functions that allow
8230 direct access to the MMX, SSE, SSE2, SSE3 and 3Dnow extensions of the
8233 @xref{X86 Built-in Functions}, for details of the functions enabled
8234 and disabled by these switches.
8236 To have SSE/SSE2 instructions generated automatically from floating-point
8237 code, see @option{-mfpmath=sse}.
8240 @itemx -mno-push-args
8242 @opindex mno-push-args
8243 Use PUSH operations to store outgoing parameters. This method is shorter
8244 and usually equally fast as method using SUB/MOV operations and is enabled
8245 by default. In some cases disabling it may improve performance because of
8246 improved scheduling and reduced dependencies.
8248 @item -maccumulate-outgoing-args
8249 @opindex maccumulate-outgoing-args
8250 If enabled, the maximum amount of space required for outgoing arguments will be
8251 computed in the function prologue. This is faster on most modern CPUs
8252 because of reduced dependencies, improved scheduling and reduced stack usage
8253 when preferred stack boundary is not equal to 2. The drawback is a notable
8254 increase in code size. This switch implies @option{-mno-push-args}.
8258 Support thread-safe exception handling on @samp{Mingw32}. Code that relies
8259 on thread-safe exception handling must compile and link all code with the
8260 @option{-mthreads} option. When compiling, @option{-mthreads} defines
8261 @option{-D_MT}; when linking, it links in a special thread helper library
8262 @option{-lmingwthrd} which cleans up per thread exception handling data.
8264 @item -mno-align-stringops
8265 @opindex mno-align-stringops
8266 Do not align destination of inlined string operations. This switch reduces
8267 code size and improves performance in case the destination is already aligned,
8268 but GCC doesn't know about it.
8270 @item -minline-all-stringops
8271 @opindex minline-all-stringops
8272 By default GCC inlines string operations only when destination is known to be
8273 aligned at least to 4 byte boundary. This enables more inlining, increase code
8274 size, but may improve performance of code that depends on fast memcpy, strlen
8275 and memset for short lengths.
8277 @item -momit-leaf-frame-pointer
8278 @opindex momit-leaf-frame-pointer
8279 Don't keep the frame pointer in a register for leaf functions. This
8280 avoids the instructions to save, set up and restore frame pointers and
8281 makes an extra register available in leaf functions. The option
8282 @option{-fomit-frame-pointer} removes the frame pointer for all functions
8283 which might make debugging harder.
8285 @item -mtls-direct-seg-refs
8286 @itemx -mno-tls-direct-seg-refs
8287 @opindex mtls-direct-seg-refs
8288 Controls whether TLS variables may be accessed with offsets from the
8289 TLS segment register (@code{%gs} for 32-bit, @code{%fs} for 64-bit),
8290 or whether the thread base pointer must be added. Whether or not this
8291 is legal depends on the operating system, and whether it maps the
8292 segment to cover the entire TLS area.
8294 For systems that use GNU libc, the default is on.
8297 These @samp{-m} switches are supported in addition to the above
8298 on AMD x86-64 processors in 64-bit environments.
8305 Generate code for a 32-bit or 64-bit environment.
8306 The 32-bit environment sets int, long and pointer to 32 bits and
8307 generates code that runs on any i386 system.
8308 The 64-bit environment sets int to 32 bits and long and pointer
8309 to 64 bits and generates code for AMD's x86-64 architecture.
8312 @opindex no-red-zone
8313 Do not use a so called red zone for x86-64 code. The red zone is mandated
8314 by the x86-64 ABI, it is a 128-byte area beyond the location of the
8315 stack pointer that will not be modified by signal or interrupt handlers
8316 and therefore can be used for temporary data without adjusting the stack
8317 pointer. The flag @option{-mno-red-zone} disables this red zone.
8319 @item -mcmodel=small
8320 @opindex mcmodel=small
8321 Generate code for the small code model: the program and its symbols must
8322 be linked in the lower 2 GB of the address space. Pointers are 64 bits.
8323 Programs can be statically or dynamically linked. This is the default
8326 @item -mcmodel=kernel
8327 @opindex mcmodel=kernel
8328 Generate code for the kernel code model. The kernel runs in the
8329 negative 2 GB of the address space.
8330 This model has to be used for Linux kernel code.
8332 @item -mcmodel=medium
8333 @opindex mcmodel=medium
8334 Generate code for the medium model: The program is linked in the lower 2
8335 GB of the address space but symbols can be located anywhere in the
8336 address space. Programs can be statically or dynamically linked, but
8337 building of shared libraries are not supported with the medium model.
8339 @item -mcmodel=large
8340 @opindex mcmodel=large
8341 Generate code for the large model: This model makes no assumptions
8342 about addresses and sizes of sections. Currently GCC does not implement
8347 @subsection IA-64 Options
8348 @cindex IA-64 Options
8350 These are the @samp{-m} options defined for the Intel IA-64 architecture.
8354 @opindex mbig-endian
8355 Generate code for a big endian target. This is the default for HP-UX@.
8357 @item -mlittle-endian
8358 @opindex mlittle-endian
8359 Generate code for a little endian target. This is the default for AIX5
8366 Generate (or don't) code for the GNU assembler. This is the default.
8367 @c Also, this is the default if the configure option @option{--with-gnu-as}
8374 Generate (or don't) code for the GNU linker. This is the default.
8375 @c Also, this is the default if the configure option @option{--with-gnu-ld}
8380 Generate code that does not use a global pointer register. The result
8381 is not position independent code, and violates the IA-64 ABI@.
8383 @item -mvolatile-asm-stop
8384 @itemx -mno-volatile-asm-stop
8385 @opindex mvolatile-asm-stop
8386 @opindex mno-volatile-asm-stop
8387 Generate (or don't) a stop bit immediately before and after volatile asm
8392 Generate code that works around Itanium B step errata.
8394 @item -mregister-names
8395 @itemx -mno-register-names
8396 @opindex mregister-names
8397 @opindex mno-register-names
8398 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
8399 the stacked registers. This may make assembler output more readable.
8405 Disable (or enable) optimizations that use the small data section. This may
8406 be useful for working around optimizer bugs.
8409 @opindex mconstant-gp
8410 Generate code that uses a single constant global pointer value. This is
8411 useful when compiling kernel code.
8415 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
8416 This is useful when compiling firmware code.
8418 @item -minline-float-divide-min-latency
8419 @opindex minline-float-divide-min-latency
8420 Generate code for inline divides of floating point values
8421 using the minimum latency algorithm.
8423 @item -minline-float-divide-max-throughput
8424 @opindex minline-float-divide-max-throughput
8425 Generate code for inline divides of floating point values
8426 using the maximum throughput algorithm.
8428 @item -minline-int-divide-min-latency
8429 @opindex minline-int-divide-min-latency
8430 Generate code for inline divides of integer values
8431 using the minimum latency algorithm.
8433 @item -minline-int-divide-max-throughput
8434 @opindex minline-int-divide-max-throughput
8435 Generate code for inline divides of integer values
8436 using the maximum throughput algorithm.
8438 @item -mno-dwarf2-asm
8440 @opindex mno-dwarf2-asm
8441 @opindex mdwarf2-asm
8442 Don't (or do) generate assembler code for the DWARF2 line number debugging
8443 info. This may be useful when not using the GNU assembler.
8445 @item -mfixed-range=@var{register-range}
8446 @opindex mfixed-range
8447 Generate code treating the given register range as fixed registers.
8448 A fixed register is one that the register allocator can not use. This is
8449 useful when compiling kernel code. A register range is specified as
8450 two registers separated by a dash. Multiple register ranges can be
8451 specified separated by a comma.
8453 @item -mearly-stop-bits
8454 @itemx -mno-early-stop-bits
8455 @opindex mearly-stop-bits
8456 @opindex mno-early-stop-bits
8457 Allow stop bits to be placed earlier than immediately preceding the
8458 instruction that triggered the stop bit. This can improve instruction
8459 scheduling, but does not always do so.
8462 @node M32R/D Options
8463 @subsection M32R/D Options
8464 @cindex M32R/D options
8466 These @option{-m} options are defined for Renesas M32R/D architectures:
8471 Generate code for the M32R/2@.
8475 Generate code for the M32R/X@.
8479 Generate code for the M32R@. This is the default.
8482 @opindex mmodel=small
8483 Assume all objects live in the lower 16MB of memory (so that their addresses
8484 can be loaded with the @code{ld24} instruction), and assume all subroutines
8485 are reachable with the @code{bl} instruction.
8486 This is the default.
8488 The addressability of a particular object can be set with the
8489 @code{model} attribute.
8491 @item -mmodel=medium
8492 @opindex mmodel=medium
8493 Assume objects may be anywhere in the 32-bit address space (the compiler
8494 will generate @code{seth/add3} instructions to load their addresses), and
8495 assume all subroutines are reachable with the @code{bl} instruction.
8498 @opindex mmodel=large
8499 Assume objects may be anywhere in the 32-bit address space (the compiler
8500 will generate @code{seth/add3} instructions to load their addresses), and
8501 assume subroutines may not be reachable with the @code{bl} instruction
8502 (the compiler will generate the much slower @code{seth/add3/jl}
8503 instruction sequence).
8506 @opindex msdata=none
8507 Disable use of the small data area. Variables will be put into
8508 one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the
8509 @code{section} attribute has been specified).
8510 This is the default.
8512 The small data area consists of sections @samp{.sdata} and @samp{.sbss}.
8513 Objects may be explicitly put in the small data area with the
8514 @code{section} attribute using one of these sections.
8517 @opindex msdata=sdata
8518 Put small global and static data in the small data area, but do not
8519 generate special code to reference them.
8523 Put small global and static data in the small data area, and generate
8524 special instructions to reference them.
8528 @cindex smaller data references
8529 Put global and static objects less than or equal to @var{num} bytes
8530 into the small data or bss sections instead of the normal data or bss
8531 sections. The default value of @var{num} is 8.
8532 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
8533 for this option to have any effect.
8535 All modules should be compiled with the same @option{-G @var{num}} value.
8536 Compiling with different values of @var{num} may or may not work; if it
8537 doesn't the linker will give an error message---incorrect code will not be
8542 Makes the M32R specific code in the compiler display some statistics
8543 that might help in debugging programs.
8546 @opindex malign-loops
8547 Align all loops to a 32-byte boundary.
8549 @item -mno-align-loops
8550 @opindex mno-align-loops
8551 Do not enforce a 32-byte alignment for loops. This is the default.
8553 @item -missue-rate=@var{number}
8554 @opindex missue-rate=@var{number}
8555 Issue @var{number} instructions per cycle. @var{number} can only be 1
8558 @item -mbranch-cost=@var{number}
8559 @opindex mbranch-cost=@var{number}
8560 @var{number} can only be 1 or 2. If it is 1 then branches will be
8561 preferred over conditional code, if it is 2, then the opposite will
8564 @item -mflush-trap=@var{number}
8565 @opindex mflush-trap=@var{number}
8566 Specifies the trap number to use to flush the cache. The default is
8567 12. Valid numbers are between 0 and 15 inclusive.
8569 @item -mno-flush-trap
8570 @opindex mno-flush-trap
8571 Specifies that the cache cannot be flushed by using a trap.
8573 @item -mflush-func=@var{name}
8574 @opindex mflush-func=@var{name}
8575 Specifies the name of the operating system function to call to flush
8576 the cache. The default is @emph{_flush_cache}, but a function call
8577 will only be used if a trap is not available.
8579 @item -mno-flush-func
8580 @opindex mno-flush-func
8581 Indicates that there is no OS function for flushing the cache.
8585 @node M680x0 Options
8586 @subsection M680x0 Options
8587 @cindex M680x0 options
8589 These are the @samp{-m} options defined for the 68000 series. The default
8590 values for these options depends on which style of 68000 was selected when
8591 the compiler was configured; the defaults for the most common choices are
8599 Generate output for a 68000. This is the default
8600 when the compiler is configured for 68000-based systems.
8602 Use this option for microcontrollers with a 68000 or EC000 core,
8603 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
8609 Generate output for a 68020. This is the default
8610 when the compiler is configured for 68020-based systems.
8614 Generate output containing 68881 instructions for floating point.
8615 This is the default for most 68020 systems unless @option{--nfp} was
8616 specified when the compiler was configured.
8620 Generate output for a 68030. This is the default when the compiler is
8621 configured for 68030-based systems.
8625 Generate output for a 68040. This is the default when the compiler is
8626 configured for 68040-based systems.
8628 This option inhibits the use of 68881/68882 instructions that have to be
8629 emulated by software on the 68040. Use this option if your 68040 does not
8630 have code to emulate those instructions.
8634 Generate output for a 68060. This is the default when the compiler is
8635 configured for 68060-based systems.
8637 This option inhibits the use of 68020 and 68881/68882 instructions that
8638 have to be emulated by software on the 68060. Use this option if your 68060
8639 does not have code to emulate those instructions.
8643 Generate output for a CPU32. This is the default
8644 when the compiler is configured for CPU32-based systems.
8646 Use this option for microcontrollers with a
8647 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
8648 68336, 68340, 68341, 68349 and 68360.
8652 Generate output for a 520X ``coldfire'' family cpu. This is the default
8653 when the compiler is configured for 520X-based systems.
8655 Use this option for microcontroller with a 5200 core, including
8656 the MCF5202, MCF5203, MCF5204 and MCF5202.
8661 Generate output for a 68040, without using any of the new instructions.
8662 This results in code which can run relatively efficiently on either a
8663 68020/68881 or a 68030 or a 68040. The generated code does use the
8664 68881 instructions that are emulated on the 68040.
8668 Generate output for a 68060, without using any of the new instructions.
8669 This results in code which can run relatively efficiently on either a
8670 68020/68881 or a 68030 or a 68040. The generated code does use the
8671 68881 instructions that are emulated on the 68060.
8674 @opindex msoft-float
8675 Generate output containing library calls for floating point.
8676 @strong{Warning:} the requisite libraries are not available for all m68k
8677 targets. Normally the facilities of the machine's usual C compiler are
8678 used, but this can't be done directly in cross-compilation. You must
8679 make your own arrangements to provide suitable library functions for
8680 cross-compilation. The embedded targets @samp{m68k-*-aout} and
8681 @samp{m68k-*-coff} do provide software floating point support.
8685 Consider type @code{int} to be 16 bits wide, like @code{short int}.
8686 Additionally, parameters passed on the stack are also aligned to a
8687 16-bit boundary even on targets whose API mandates promotion to 32-bit.
8690 @opindex mnobitfield
8691 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
8692 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
8696 Do use the bit-field instructions. The @option{-m68020} option implies
8697 @option{-mbitfield}. This is the default if you use a configuration
8698 designed for a 68020.
8702 Use a different function-calling convention, in which functions
8703 that take a fixed number of arguments return with the @code{rtd}
8704 instruction, which pops their arguments while returning. This
8705 saves one instruction in the caller since there is no need to pop
8706 the arguments there.
8708 This calling convention is incompatible with the one normally
8709 used on Unix, so you cannot use it if you need to call libraries
8710 compiled with the Unix compiler.
8712 Also, you must provide function prototypes for all functions that
8713 take variable numbers of arguments (including @code{printf});
8714 otherwise incorrect code will be generated for calls to those
8717 In addition, seriously incorrect code will result if you call a
8718 function with too many arguments. (Normally, extra arguments are
8719 harmlessly ignored.)
8721 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
8722 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
8725 @itemx -mno-align-int
8727 @opindex mno-align-int
8728 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
8729 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
8730 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
8731 Aligning variables on 32-bit boundaries produces code that runs somewhat
8732 faster on processors with 32-bit busses at the expense of more memory.
8734 @strong{Warning:} if you use the @option{-malign-int} switch, GCC will
8735 align structures containing the above types differently than
8736 most published application binary interface specifications for the m68k.
8740 Use the pc-relative addressing mode of the 68000 directly, instead of
8741 using a global offset table. At present, this option implies @option{-fpic},
8742 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
8743 not presently supported with @option{-mpcrel}, though this could be supported for
8744 68020 and higher processors.
8746 @item -mno-strict-align
8747 @itemx -mstrict-align
8748 @opindex mno-strict-align
8749 @opindex mstrict-align
8750 Do not (do) assume that unaligned memory references will be handled by
8754 Generate code that allows the data segment to be located in a different
8755 area of memory from the text segment. This allows for execute in place in
8756 an environment without virtual memory management. This option implies -fPIC.
8759 Generate code that assumes that the data segment follows the text segment.
8760 This is the default.
8762 @item -mid-shared-library
8763 Generate code that supports shared libraries via the library ID method.
8764 This allows for execute in place and shared libraries in an environment
8765 without virtual memory management. This option implies -fPIC.
8767 @item -mno-id-shared-library
8768 Generate code that doesn't assume ID based shared libraries are being used.
8769 This is the default.
8771 @item -mshared-library-id=n
8772 Specified the identification number of the ID based shared library being
8773 compiled. Specifying a value of 0 will generate more compact code, specifying
8774 other values will force the allocation of that number to the current
8775 library but is no more space or time efficient than omitting this option.
8779 @node M68hc1x Options
8780 @subsection M68hc1x Options
8781 @cindex M68hc1x options
8783 These are the @samp{-m} options defined for the 68hc11 and 68hc12
8784 microcontrollers. The default values for these options depends on
8785 which style of microcontroller was selected when the compiler was configured;
8786 the defaults for the most common choices are given below.
8793 Generate output for a 68HC11. This is the default
8794 when the compiler is configured for 68HC11-based systems.
8800 Generate output for a 68HC12. This is the default
8801 when the compiler is configured for 68HC12-based systems.
8807 Generate output for a 68HCS12.
8810 @opindex mauto-incdec
8811 Enable the use of 68HC12 pre and post auto-increment and auto-decrement
8818 Enable the use of 68HC12 min and max instructions.
8821 @itemx -mno-long-calls
8822 @opindex mlong-calls
8823 @opindex mno-long-calls
8824 Treat all calls as being far away (near). If calls are assumed to be
8825 far away, the compiler will use the @code{call} instruction to
8826 call a function and the @code{rtc} instruction for returning.
8830 Consider type @code{int} to be 16 bits wide, like @code{short int}.
8832 @item -msoft-reg-count=@var{count}
8833 @opindex msoft-reg-count
8834 Specify the number of pseudo-soft registers which are used for the
8835 code generation. The maximum number is 32. Using more pseudo-soft
8836 register may or may not result in better code depending on the program.
8837 The default is 4 for 68HC11 and 2 for 68HC12.
8842 @subsection MCore Options
8843 @cindex MCore options
8845 These are the @samp{-m} options defined for the Motorola M*Core
8853 @opindex mno-hardlit
8854 Inline constants into the code stream if it can be done in two
8855 instructions or less.
8861 Use the divide instruction. (Enabled by default).
8863 @item -mrelax-immediate
8864 @itemx -mno-relax-immediate
8865 @opindex mrelax-immediate
8866 @opindex mno-relax-immediate
8867 Allow arbitrary sized immediates in bit operations.
8869 @item -mwide-bitfields
8870 @itemx -mno-wide-bitfields
8871 @opindex mwide-bitfields
8872 @opindex mno-wide-bitfields
8873 Always treat bit-fields as int-sized.
8875 @item -m4byte-functions
8876 @itemx -mno-4byte-functions
8877 @opindex m4byte-functions
8878 @opindex mno-4byte-functions
8879 Force all functions to be aligned to a four byte boundary.
8881 @item -mcallgraph-data
8882 @itemx -mno-callgraph-data
8883 @opindex mcallgraph-data
8884 @opindex mno-callgraph-data
8885 Emit callgraph information.
8888 @itemx -mno-slow-bytes
8889 @opindex mslow-bytes
8890 @opindex mno-slow-bytes
8891 Prefer word access when reading byte quantities.
8893 @item -mlittle-endian
8895 @opindex mlittle-endian
8896 @opindex mbig-endian
8897 Generate code for a little endian target.
8903 Generate code for the 210 processor.
8907 @subsection MIPS Options
8908 @cindex MIPS options
8914 Generate big-endian code.
8918 Generate little-endian code. This is the default for @samp{mips*el-*-*}
8921 @item -march=@var{arch}
8923 Generate code that will run on @var{arch}, which can be the name of a
8924 generic MIPS ISA, or the name of a particular processor.
8926 @samp{mips1}, @samp{mips2}, @samp{mips3}, @samp{mips4},
8927 @samp{mips32}, @samp{mips32r2}, and @samp{mips64}.
8928 The processor names are:
8929 @samp{4kc}, @samp{4kp}, @samp{5kc}, @samp{20kc},
8931 @samp{r2000}, @samp{r3000}, @samp{r3900}, @samp{r4000}, @samp{r4400},
8932 @samp{r4600}, @samp{r4650}, @samp{r6000}, @samp{r8000}, @samp{rm7000},
8936 @samp{vr4100}, @samp{vr4111}, @samp{vr4120}, @samp{vr4130}, @samp{vr4300},
8937 @samp{vr5000}, @samp{vr5400} and @samp{vr5500}.
8938 The special value @samp{from-abi} selects the
8939 most compatible architecture for the selected ABI (that is,
8940 @samp{mips1} for 32-bit ABIs and @samp{mips3} for 64-bit ABIs)@.
8942 In processor names, a final @samp{000} can be abbreviated as @samp{k}
8943 (for example, @samp{-march=r2k}). Prefixes are optional, and
8944 @samp{vr} may be written @samp{r}.
8946 GCC defines two macros based on the value of this option. The first
8947 is @samp{_MIPS_ARCH}, which gives the name of target architecture, as
8948 a string. The second has the form @samp{_MIPS_ARCH_@var{foo}},
8949 where @var{foo} is the capitalized value of @samp{_MIPS_ARCH}@.
8950 For example, @samp{-march=r2000} will set @samp{_MIPS_ARCH}
8951 to @samp{"r2000"} and define the macro @samp{_MIPS_ARCH_R2000}.
8953 Note that the @samp{_MIPS_ARCH} macro uses the processor names given
8954 above. In other words, it will have the full prefix and will not
8955 abbreviate @samp{000} as @samp{k}. In the case of @samp{from-abi},
8956 the macro names the resolved architecture (either @samp{"mips1"} or
8957 @samp{"mips3"}). It names the default architecture when no
8958 @option{-march} option is given.
8960 @item -mtune=@var{arch}
8962 Optimize for @var{arch}. Among other things, this option controls
8963 the way instructions are scheduled, and the perceived cost of arithmetic
8964 operations. The list of @var{arch} values is the same as for
8967 When this option is not used, GCC will optimize for the processor
8968 specified by @option{-march}. By using @option{-march} and
8969 @option{-mtune} together, it is possible to generate code that will
8970 run on a family of processors, but optimize the code for one
8971 particular member of that family.
8973 @samp{-mtune} defines the macros @samp{_MIPS_TUNE} and
8974 @samp{_MIPS_TUNE_@var{foo}}, which work in the same way as the
8975 @samp{-march} ones described above.
8979 Equivalent to @samp{-march=mips1}.
8983 Equivalent to @samp{-march=mips2}.
8987 Equivalent to @samp{-march=mips3}.
8991 Equivalent to @samp{-march=mips4}.
8995 Equivalent to @samp{-march=mips32}.
8999 Equivalent to @samp{-march=mips32r2}.
9003 Equivalent to @samp{-march=mips64}.
9009 Use (do not use) the MIPS16 ISA.
9021 Generate code for the given ABI@.
9023 Note that the EABI has a 32-bit and a 64-bit variant. GCC normally
9024 generates 64-bit code when you select a 64-bit architecture, but you
9025 can use @option{-mgp32} to get 32-bit code instead.
9027 For information about the O64 ABI, see
9028 @w{@uref{http://gcc.gnu.org/projects/mipso64-abi.html}}.
9031 @itemx -mno-abicalls
9033 @opindex mno-abicalls
9034 Generate (do not generate) SVR4-style position-independent code.
9035 @option{-mabicalls} is the default for SVR4-based systems.
9041 Lift (do not lift) the usual restrictions on the size of the global
9044 GCC normally uses a single instruction to load values from the GOT.
9045 While this is relatively efficient, it will only work if the GOT
9046 is smaller than about 64k. Anything larger will cause the linker
9047 to report an error such as:
9049 @cindex relocation truncated to fit (MIPS)
9051 relocation truncated to fit: R_MIPS_GOT16 foobar
9054 If this happens, you should recompile your code with @option{-mxgot}.
9055 It should then work with very large GOTs, although it will also be
9056 less efficient, since it will take three instructions to fetch the
9057 value of a global symbol.
9059 Note that some linkers can create multiple GOTs. If you have such a
9060 linker, you should only need to use @option{-mxgot} when a single object
9061 file accesses more than 64k's worth of GOT entries. Very few do.
9063 These options have no effect unless GCC is generating position
9068 Assume that general-purpose registers are 32 bits wide.
9072 Assume that general-purpose registers are 64 bits wide.
9076 Assume that floating-point registers are 32 bits wide.
9080 Assume that floating-point registers are 64 bits wide.
9083 @opindex mhard-float
9084 Use floating-point coprocessor instructions.
9087 @opindex msoft-float
9088 Do not use floating-point coprocessor instructions. Implement
9089 floating-point calculations using library calls instead.
9091 @item -msingle-float
9092 @opindex msingle-float
9093 Assume that the floating-point coprocessor only supports single-precision
9096 @itemx -mdouble-float
9097 @opindex mdouble-float
9098 Assume that the floating-point coprocessor supports double-precision
9099 operations. This is the default.
9103 Force @code{int} and @code{long} types to be 64 bits wide. See
9104 @option{-mlong32} for an explanation of the default and the way
9105 that the pointer size is determined.
9109 Force @code{long} types to be 64 bits wide. See @option{-mlong32} for
9110 an explanation of the default and the way that the pointer size is
9115 Force @code{long}, @code{int}, and pointer types to be 32 bits wide.
9117 The default size of @code{int}s, @code{long}s and pointers depends on
9118 the ABI@. All the supported ABIs use 32-bit @code{int}s. The n64 ABI
9119 uses 64-bit @code{long}s, as does the 64-bit EABI; the others use
9120 32-bit @code{long}s. Pointers are the same size as @code{long}s,
9121 or the same size as integer registers, whichever is smaller.
9125 @cindex smaller data references (MIPS)
9126 @cindex gp-relative references (MIPS)
9127 Put global and static items less than or equal to @var{num} bytes into
9128 the small data or bss section instead of the normal data or bss section.
9129 This allows the data to be accessed using a single instruction.
9131 All modules should be compiled with the same @option{-G @var{num}}
9134 @item -membedded-data
9135 @itemx -mno-embedded-data
9136 @opindex membedded-data
9137 @opindex mno-embedded-data
9138 Allocate variables to the read-only data section first if possible, then
9139 next in the small data section if possible, otherwise in data. This gives
9140 slightly slower code than the default, but reduces the amount of RAM required
9141 when executing, and thus may be preferred for some embedded systems.
9143 @item -muninit-const-in-rodata
9144 @itemx -mno-uninit-const-in-rodata
9145 @opindex muninit-const-in-rodata
9146 @opindex mno-uninit-const-in-rodata
9147 Put uninitialized @code{const} variables in the read-only data section.
9148 This option is only meaningful in conjunction with @option{-membedded-data}.
9150 @item -msplit-addresses
9151 @itemx -mno-split-addresses
9152 @opindex msplit-addresses
9153 @opindex mno-split-addresses
9154 Enable (disable) use of the @code{%hi()} and @code{%lo()} assembler
9155 relocation operators. This option has been superceded by
9156 @option{-mexplicit-relocs} but is retained for backwards compatibility.
9158 @item -mexplicit-relocs
9159 @itemx -mno-explicit-relocs
9160 @opindex mexplicit-relocs
9161 @opindex mno-explicit-relocs
9162 Use (do not use) assembler relocation operators when dealing with symbolic
9163 addresses. The alternative, selected by @option{-mno-explicit-relocs},
9164 is to use assembler macros instead.
9166 @option{-mexplicit-relocs} is usually the default if GCC was configured
9167 to use an assembler that supports relocation operators. However, the
9168 combination of @option{-mabicalls} and @option{-fno-unit-at-a-time}
9169 implies @option{-mno-explicit-relocs} unless explicitly overridden.
9170 This is because, when generating abicalls, the choice of relocation
9171 depends on whether a symbol is local or global. In some rare cases,
9172 GCC will not be able to decide this until the whole compilation unit
9175 @item -mcheck-zero-division
9176 @itemx -mno-check-zero-division
9177 @opindex mcheck-zero-division
9178 @opindex mno-check-zero-division
9179 Trap (do not trap) on integer division by zero. The default is
9180 @option{-mcheck-zero-division}.
9186 Force (do not force) the use of @code{memcpy()} for non-trivial block
9187 moves. The default is @option{-mno-memcpy}, which allows GCC to inline
9188 most constant-sized copies.
9191 @itemx -mno-long-calls
9192 @opindex mlong-calls
9193 @opindex mno-long-calls
9194 Disable (do not disable) use of the @code{jal} instruction. Calling
9195 functions using @code{jal} is more efficient but requires the caller
9196 and callee to be in the same 256 megabyte segment.
9198 This option has no effect on abicalls code. The default is
9199 @option{-mno-long-calls}.
9205 Enable (disable) use of the @code{mad}, @code{madu} and @code{mul}
9206 instructions, as provided by the R4650 ISA.
9209 @itemx -mno-fused-madd
9210 @opindex mfused-madd
9211 @opindex mno-fused-madd
9212 Enable (disable) use of the floating point multiply-accumulate
9213 instructions, when they are available. The default is
9214 @option{-mfused-madd}.
9216 When multiply-accumulate instructions are used, the intermediate
9217 product is calculated to infinite precision and is not subject to
9218 the FCSR Flush to Zero bit. This may be undesirable in some
9223 Tell the MIPS assembler to not run its preprocessor over user
9224 assembler files (with a @samp{.s} suffix) when assembling them.
9227 @itemx -mno-fix-r4000
9229 @opindex mno-fix-r4000
9230 Work around certain R4000 CPU errata:
9233 A double-word or a variable shift may give an incorrect result if executed
9234 immediately after starting an integer division.
9236 A double-word or a variable shift may give an incorrect result if executed
9237 while an integer multiplication is in progress.
9239 An integer division may give an incorrect result if started in a delay slot
9240 of a taken branch or a jump.
9244 @itemx -mno-fix-r4400
9246 @opindex mno-fix-r4400
9247 Work around certain R4400 CPU errata:
9250 A double-word or a variable shift may give an incorrect result if executed
9251 immediately after starting an integer division.
9255 @itemx -mno-fix-vr4120
9256 @opindex mfix-vr4120
9257 Work around certain VR4120 errata:
9260 @code{dmultu} does not always produce the correct result.
9262 @code{div} and @code{ddiv} do not always produce the correct result if one
9263 of the operands is negative.
9265 The workarounds for the division errata rely on special functions in
9266 @file{libgcc.a}. At present, these functions are only provided by
9267 the @code{mips64vr*-elf} configurations.
9269 Other VR4120 errata require a nop to be inserted between certain pairs of
9270 instructions. These errata are handled by the assembler, not by GCC itself.
9275 Work around certain SB-1 CPU core errata.
9276 (This flag currently works around the SB-1 revision 2
9277 ``F1'' and ``F2'' floating point errata.)
9279 @item -mflush-func=@var{func}
9280 @itemx -mno-flush-func
9281 @opindex mflush-func
9282 Specifies the function to call to flush the I and D caches, or to not
9283 call any such function. If called, the function must take the same
9284 arguments as the common @code{_flush_func()}, that is, the address of the
9285 memory range for which the cache is being flushed, the size of the
9286 memory range, and the number 3 (to flush both caches). The default
9287 depends on the target GCC was configured for, but commonly is either
9288 @samp{_flush_func} or @samp{__cpu_flush}.
9290 @item -mbranch-likely
9291 @itemx -mno-branch-likely
9292 @opindex mbranch-likely
9293 @opindex mno-branch-likely
9294 Enable or disable use of Branch Likely instructions, regardless of the
9295 default for the selected architecture. By default, Branch Likely
9296 instructions may be generated if they are supported by the selected
9297 architecture. An exception is for the MIPS32 and MIPS64 architectures
9298 and processors which implement those architectures; for those, Branch
9299 Likely instructions will not be generated by default because the MIPS32
9300 and MIPS64 architectures specifically deprecate their use.
9302 @item -mfp-exceptions
9303 @itemx -mno-fp-exceptions
9304 @opindex mfp-exceptions
9305 Specifies whether FP exceptions are enabled. This affects how we schedule
9306 FP instructions for some processors. The default is that FP exceptions are
9309 For instance, on the SB-1, if FP exceptions are disabled, and we are emitting
9310 64-bit code, then we can use both FP pipes. Otherwise, we can only use one
9313 @item -mvr4130-align
9314 @itemx -mno-vr4130-align
9315 @opindex mvr4130-align
9316 The VR4130 pipeline is two-way superscalar, but can only issue two
9317 instructions together if the first one is 8-byte aligned. When this
9318 option is enabled, GCC will align pairs of instructions that it
9319 thinks should execute in parallel.
9321 This option only has an effect when optimizing for the VR4130.
9322 It normally makes code faster, but at the expense of making it bigger.
9323 It is enabled by default at optimization level @option{-O3}.
9327 @subsection MMIX Options
9328 @cindex MMIX Options
9330 These options are defined for the MMIX:
9334 @itemx -mno-libfuncs
9336 @opindex mno-libfuncs
9337 Specify that intrinsic library functions are being compiled, passing all
9338 values in registers, no matter the size.
9343 @opindex mno-epsilon
9344 Generate floating-point comparison instructions that compare with respect
9345 to the @code{rE} epsilon register.
9347 @item -mabi=mmixware
9349 @opindex mabi-mmixware
9351 Generate code that passes function parameters and return values that (in
9352 the called function) are seen as registers @code{$0} and up, as opposed to
9353 the GNU ABI which uses global registers @code{$231} and up.
9356 @itemx -mno-zero-extend
9357 @opindex mzero-extend
9358 @opindex mno-zero-extend
9359 When reading data from memory in sizes shorter than 64 bits, use (do not
9360 use) zero-extending load instructions by default, rather than
9361 sign-extending ones.
9364 @itemx -mno-knuthdiv
9366 @opindex mno-knuthdiv
9367 Make the result of a division yielding a remainder have the same sign as
9368 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
9369 remainder follows the sign of the dividend. Both methods are
9370 arithmetically valid, the latter being almost exclusively used.
9372 @item -mtoplevel-symbols
9373 @itemx -mno-toplevel-symbols
9374 @opindex mtoplevel-symbols
9375 @opindex mno-toplevel-symbols
9376 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
9377 code can be used with the @code{PREFIX} assembly directive.
9381 Generate an executable in the ELF format, rather than the default
9382 @samp{mmo} format used by the @command{mmix} simulator.
9384 @item -mbranch-predict
9385 @itemx -mno-branch-predict
9386 @opindex mbranch-predict
9387 @opindex mno-branch-predict
9388 Use (do not use) the probable-branch instructions, when static branch
9389 prediction indicates a probable branch.
9391 @item -mbase-addresses
9392 @itemx -mno-base-addresses
9393 @opindex mbase-addresses
9394 @opindex mno-base-addresses
9395 Generate (do not generate) code that uses @emph{base addresses}. Using a
9396 base address automatically generates a request (handled by the assembler
9397 and the linker) for a constant to be set up in a global register. The
9398 register is used for one or more base address requests within the range 0
9399 to 255 from the value held in the register. The generally leads to short
9400 and fast code, but the number of different data items that can be
9401 addressed is limited. This means that a program that uses lots of static
9402 data may require @option{-mno-base-addresses}.
9405 @itemx -mno-single-exit
9406 @opindex msingle-exit
9407 @opindex mno-single-exit
9408 Force (do not force) generated code to have a single exit point in each
9412 @node MN10300 Options
9413 @subsection MN10300 Options
9414 @cindex MN10300 options
9416 These @option{-m} options are defined for Matsushita MN10300 architectures:
9421 Generate code to avoid bugs in the multiply instructions for the MN10300
9422 processors. This is the default.
9425 @opindex mno-mult-bug
9426 Do not generate code to avoid bugs in the multiply instructions for the
9431 Generate code which uses features specific to the AM33 processor.
9435 Do not generate code which uses features specific to the AM33 processor. This
9440 Do not link in the C run-time initialization object file.
9444 Indicate to the linker that it should perform a relaxation optimization pass
9445 to shorten branches, calls and absolute memory addresses. This option only
9446 has an effect when used on the command line for the final link step.
9448 This option makes symbolic debugging impossible.
9452 @subsection NS32K Options
9453 @cindex NS32K options
9455 These are the @samp{-m} options defined for the 32000 series. The default
9456 values for these options depends on which style of 32000 was selected when
9457 the compiler was configured; the defaults for the most common choices are
9465 Generate output for a 32032. This is the default
9466 when the compiler is configured for 32032 and 32016 based systems.
9472 Generate output for a 32332. This is the default
9473 when the compiler is configured for 32332-based systems.
9479 Generate output for a 32532. This is the default
9480 when the compiler is configured for 32532-based systems.
9484 Generate output containing 32081 instructions for floating point.
9485 This is the default for all systems.
9489 Generate output containing 32381 instructions for floating point. This
9490 also implies @option{-m32081}. The 32381 is only compatible with the 32332
9491 and 32532 cpus. This is the default for the pc532-netbsd configuration.
9495 Try and generate multiply-add floating point instructions @code{polyF}
9496 and @code{dotF}. This option is only available if the @option{-m32381}
9497 option is in effect. Using these instructions requires changes to
9498 register allocation which generally has a negative impact on
9499 performance. This option should only be enabled when compiling code
9500 particularly likely to make heavy use of multiply-add instructions.
9503 @opindex mnomulti-add
9504 Do not try and generate multiply-add floating point instructions
9505 @code{polyF} and @code{dotF}. This is the default on all platforms.
9508 @opindex msoft-float
9509 Generate output containing library calls for floating point.
9510 @strong{Warning:} the requisite libraries may not be available.
9512 @item -mieee-compare
9513 @itemx -mno-ieee-compare
9514 @opindex mieee-compare
9515 @opindex mno-ieee-compare
9516 Control whether or not the compiler uses IEEE floating point
9517 comparisons. These handle correctly the case where the result of a
9518 comparison is unordered.
9519 @strong{Warning:} the requisite kernel support may not be available.
9522 @opindex mnobitfield
9523 Do not use the bit-field instructions. On some machines it is faster to
9524 use shifting and masking operations. This is the default for the pc532.
9528 Do use the bit-field instructions. This is the default for all platforms
9533 Use a different function-calling convention, in which functions
9534 that take a fixed number of arguments return pop their
9535 arguments on return with the @code{ret} instruction.
9537 This calling convention is incompatible with the one normally
9538 used on Unix, so you cannot use it if you need to call libraries
9539 compiled with the Unix compiler.
9541 Also, you must provide function prototypes for all functions that
9542 take variable numbers of arguments (including @code{printf});
9543 otherwise incorrect code will be generated for calls to those
9546 In addition, seriously incorrect code will result if you call a
9547 function with too many arguments. (Normally, extra arguments are
9548 harmlessly ignored.)
9550 This option takes its name from the 680x0 @code{rtd} instruction.
9555 Use a different function-calling convention where the first two arguments
9556 are passed in registers.
9558 This calling convention is incompatible with the one normally
9559 used on Unix, so you cannot use it if you need to call libraries
9560 compiled with the Unix compiler.
9563 @opindex mnoregparam
9564 Do not pass any arguments in registers. This is the default for all
9569 It is OK to use the sb as an index register which is always loaded with
9570 zero. This is the default for the pc532-netbsd target.
9574 The sb register is not available for use or has not been initialized to
9575 zero by the run time system. This is the default for all targets except
9576 the pc532-netbsd. It is also implied whenever @option{-mhimem} or
9577 @option{-fpic} is set.
9581 Many ns32000 series addressing modes use displacements of up to 512MB@.
9582 If an address is above 512MB then displacements from zero can not be used.
9583 This option causes code to be generated which can be loaded above 512MB@.
9584 This may be useful for operating systems or ROM code.
9588 Assume code will be loaded in the first 512MB of virtual address space.
9589 This is the default for all platforms.
9593 @node PDP-11 Options
9594 @subsection PDP-11 Options
9595 @cindex PDP-11 Options
9597 These options are defined for the PDP-11:
9602 Use hardware FPP floating point. This is the default. (FIS floating
9603 point on the PDP-11/40 is not supported.)
9606 @opindex msoft-float
9607 Do not use hardware floating point.
9611 Return floating-point results in ac0 (fr0 in Unix assembler syntax).
9615 Return floating-point results in memory. This is the default.
9619 Generate code for a PDP-11/40.
9623 Generate code for a PDP-11/45. This is the default.
9627 Generate code for a PDP-11/10.
9629 @item -mbcopy-builtin
9630 @opindex bcopy-builtin
9631 Use inline @code{movmemhi} patterns for copying memory. This is the
9636 Do not use inline @code{movmemhi} patterns for copying memory.
9642 Use 16-bit @code{int}. This is the default.
9648 Use 32-bit @code{int}.
9653 @opindex mno-float32
9654 Use 64-bit @code{float}. This is the default.
9659 @opindex mno-float64
9660 Use 32-bit @code{float}.
9664 Use @code{abshi2} pattern. This is the default.
9668 Do not use @code{abshi2} pattern.
9670 @item -mbranch-expensive
9671 @opindex mbranch-expensive
9672 Pretend that branches are expensive. This is for experimenting with
9673 code generation only.
9675 @item -mbranch-cheap
9676 @opindex mbranch-cheap
9677 Do not pretend that branches are expensive. This is the default.
9681 Generate code for a system with split I&D.
9685 Generate code for a system without split I&D. This is the default.
9689 Use Unix assembler syntax. This is the default when configured for
9694 Use DEC assembler syntax. This is the default when configured for any
9695 PDP-11 target other than @samp{pdp11-*-bsd}.
9698 @node PowerPC Options
9699 @subsection PowerPC Options
9700 @cindex PowerPC options
9702 These are listed under @xref{RS/6000 and PowerPC Options}.
9704 @node RS/6000 and PowerPC Options
9705 @subsection IBM RS/6000 and PowerPC Options
9706 @cindex RS/6000 and PowerPC Options
9707 @cindex IBM RS/6000 and PowerPC Options
9709 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
9717 @itemx -mpowerpc-gpopt
9718 @itemx -mno-powerpc-gpopt
9719 @itemx -mpowerpc-gfxopt
9720 @itemx -mno-powerpc-gfxopt
9722 @itemx -mno-powerpc64
9728 @opindex mno-powerpc
9729 @opindex mpowerpc-gpopt
9730 @opindex mno-powerpc-gpopt
9731 @opindex mpowerpc-gfxopt
9732 @opindex mno-powerpc-gfxopt
9734 @opindex mno-powerpc64
9735 GCC supports two related instruction set architectures for the
9736 RS/6000 and PowerPC@. The @dfn{POWER} instruction set are those
9737 instructions supported by the @samp{rios} chip set used in the original
9738 RS/6000 systems and the @dfn{PowerPC} instruction set is the
9739 architecture of the Motorola MPC5xx, MPC6xx, MPC8xx microprocessors, and
9740 the IBM 4xx microprocessors.
9742 Neither architecture is a subset of the other. However there is a
9743 large common subset of instructions supported by both. An MQ
9744 register is included in processors supporting the POWER architecture.
9746 You use these options to specify which instructions are available on the
9747 processor you are using. The default value of these options is
9748 determined when configuring GCC@. Specifying the
9749 @option{-mcpu=@var{cpu_type}} overrides the specification of these
9750 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
9751 rather than the options listed above.
9753 The @option{-mpower} option allows GCC to generate instructions that
9754 are found only in the POWER architecture and to use the MQ register.
9755 Specifying @option{-mpower2} implies @option{-power} and also allows GCC
9756 to generate instructions that are present in the POWER2 architecture but
9757 not the original POWER architecture.
9759 The @option{-mpowerpc} option allows GCC to generate instructions that
9760 are found only in the 32-bit subset of the PowerPC architecture.
9761 Specifying @option{-mpowerpc-gpopt} implies @option{-mpowerpc} and also allows
9762 GCC to use the optional PowerPC architecture instructions in the
9763 General Purpose group, including floating-point square root. Specifying
9764 @option{-mpowerpc-gfxopt} implies @option{-mpowerpc} and also allows GCC to
9765 use the optional PowerPC architecture instructions in the Graphics
9766 group, including floating-point select.
9768 The @option{-mpowerpc64} option allows GCC to generate the additional
9769 64-bit instructions that are found in the full PowerPC64 architecture
9770 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
9771 @option{-mno-powerpc64}.
9773 If you specify both @option{-mno-power} and @option{-mno-powerpc}, GCC
9774 will use only the instructions in the common subset of both
9775 architectures plus some special AIX common-mode calls, and will not use
9776 the MQ register. Specifying both @option{-mpower} and @option{-mpowerpc}
9777 permits GCC to use any instruction from either architecture and to
9778 allow use of the MQ register; specify this for the Motorola MPC601.
9780 @item -mnew-mnemonics
9781 @itemx -mold-mnemonics
9782 @opindex mnew-mnemonics
9783 @opindex mold-mnemonics
9784 Select which mnemonics to use in the generated assembler code. With
9785 @option{-mnew-mnemonics}, GCC uses the assembler mnemonics defined for
9786 the PowerPC architecture. With @option{-mold-mnemonics} it uses the
9787 assembler mnemonics defined for the POWER architecture. Instructions
9788 defined in only one architecture have only one mnemonic; GCC uses that
9789 mnemonic irrespective of which of these options is specified.
9791 GCC defaults to the mnemonics appropriate for the architecture in
9792 use. Specifying @option{-mcpu=@var{cpu_type}} sometimes overrides the
9793 value of these option. Unless you are building a cross-compiler, you
9794 should normally not specify either @option{-mnew-mnemonics} or
9795 @option{-mold-mnemonics}, but should instead accept the default.
9797 @item -mcpu=@var{cpu_type}
9799 Set architecture type, register usage, choice of mnemonics, and
9800 instruction scheduling parameters for machine type @var{cpu_type}.
9801 Supported values for @var{cpu_type} are @samp{401}, @samp{403},
9802 @samp{405}, @samp{405fp}, @samp{440}, @samp{440fp}, @samp{505},
9803 @samp{601}, @samp{602}, @samp{603}, @samp{603e}, @samp{604},
9804 @samp{604e}, @samp{620}, @samp{630}, @samp{740}, @samp{7400},
9805 @samp{7450}, @samp{750}, @samp{801}, @samp{821}, @samp{823},
9806 @samp{860}, @samp{970}, @samp{common}, @samp{ec603e}, @samp{G3},
9807 @samp{G4}, @samp{G5}, @samp{power}, @samp{power2}, @samp{power3},
9808 @samp{power4}, @samp{power5}, @samp{powerpc}, @samp{powerpc64},
9809 @samp{rios}, @samp{rios1}, @samp{rios2}, @samp{rsc}, and @samp{rs64a}.
9811 @option{-mcpu=common} selects a completely generic processor. Code
9812 generated under this option will run on any POWER or PowerPC processor.
9813 GCC will use only the instructions in the common subset of both
9814 architectures, and will not use the MQ register. GCC assumes a generic
9815 processor model for scheduling purposes.
9817 @option{-mcpu=power}, @option{-mcpu=power2}, @option{-mcpu=powerpc}, and
9818 @option{-mcpu=powerpc64} specify generic POWER, POWER2, pure 32-bit
9819 PowerPC (i.e., not MPC601), and 64-bit PowerPC architecture machine
9820 types, with an appropriate, generic processor model assumed for
9821 scheduling purposes.
9823 The other options specify a specific processor. Code generated under
9824 those options will run best on that processor, and may not run at all on
9827 The @option{-mcpu} options automatically enable or disable the
9828 following options: @option{-maltivec}, @option{-mhard-float},
9829 @option{-mmfcrf}, @option{-mmultiple}, @option{-mnew-mnemonics},
9830 @option{-mpower}, @option{-mpower2}, @option{-mpowerpc64},
9831 @option{-mpowerpc-gpopt}, @option{-mpowerpc-gfxopt},
9832 @option{-mstring}. The particular options set for any particular CPU
9833 will vary between compiler versions, depending on what setting seems
9834 to produce optimal code for that CPU; it doesn't necessarily reflect
9835 the actual hardware's capabilities. If you wish to set an individual
9836 option to a particular value, you may specify it after the
9837 @option{-mcpu} option, like @samp{-mcpu=970 -mno-altivec}.
9839 On AIX, the @option{-maltivec} and @option{-mpowerpc64} options are
9840 not enabled or disabled by the @option{-mcpu} option at present, since
9841 AIX does not have full support for these options. You may still
9842 enable or disable them individually if you're sure it'll work in your
9845 @item -mtune=@var{cpu_type}
9847 Set the instruction scheduling parameters for machine type
9848 @var{cpu_type}, but do not set the architecture type, register usage, or
9849 choice of mnemonics, as @option{-mcpu=@var{cpu_type}} would. The same
9850 values for @var{cpu_type} are used for @option{-mtune} as for
9851 @option{-mcpu}. If both are specified, the code generated will use the
9852 architecture, registers, and mnemonics set by @option{-mcpu}, but the
9853 scheduling parameters set by @option{-mtune}.
9858 @opindex mno-altivec
9859 These switches enable or disable the use of built-in functions that
9860 allow access to the AltiVec instruction set. You may also need to set
9861 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
9866 Extend the current ABI with SPE ABI extensions. This does not change
9867 the default ABI, instead it adds the SPE ABI extensions to the current
9871 @opindex mabi=no-spe
9872 Disable Booke SPE ABI extensions for the current ABI.
9874 @item -misel=@var{yes/no}
9877 This switch enables or disables the generation of ISEL instructions.
9879 @item -mspe=@var{yes/no}
9882 This switch enables or disables the generation of SPE simd
9885 @item -mfloat-gprs=@var{yes/no}
9887 @opindex mfloat-gprs
9888 This switch enables or disables the generation of floating point
9889 operations on the general purpose registers for architectures that
9890 support it. This option is currently only available on the MPC8540.
9893 @itemx -mno-fp-in-toc
9894 @itemx -mno-sum-in-toc
9895 @itemx -mminimal-toc
9897 @opindex mno-fp-in-toc
9898 @opindex mno-sum-in-toc
9899 @opindex mminimal-toc
9900 Modify generation of the TOC (Table Of Contents), which is created for
9901 every executable file. The @option{-mfull-toc} option is selected by
9902 default. In that case, GCC will allocate at least one TOC entry for
9903 each unique non-automatic variable reference in your program. GCC
9904 will also place floating-point constants in the TOC@. However, only
9905 16,384 entries are available in the TOC@.
9907 If you receive a linker error message that saying you have overflowed
9908 the available TOC space, you can reduce the amount of TOC space used
9909 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
9910 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
9911 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
9912 generate code to calculate the sum of an address and a constant at
9913 run-time instead of putting that sum into the TOC@. You may specify one
9914 or both of these options. Each causes GCC to produce very slightly
9915 slower and larger code at the expense of conserving TOC space.
9917 If you still run out of space in the TOC even when you specify both of
9918 these options, specify @option{-mminimal-toc} instead. This option causes
9919 GCC to make only one TOC entry for every file. When you specify this
9920 option, GCC will produce code that is slower and larger but which
9921 uses extremely little TOC space. You may wish to use this option
9922 only on files that contain less frequently executed code.
9928 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
9929 @code{long} type, and the infrastructure needed to support them.
9930 Specifying @option{-maix64} implies @option{-mpowerpc64} and
9931 @option{-mpowerpc}, while @option{-maix32} disables the 64-bit ABI and
9932 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
9937 @opindex mno-xl-call
9938 On AIX, pass floating-point arguments to prototyped functions beyond the
9939 register save area (RSA) on the stack in addition to argument FPRs. The
9940 AIX calling convention was extended but not initially documented to
9941 handle an obscure K&R C case of calling a function that takes the
9942 address of its arguments with fewer arguments than declared. AIX XL
9943 compilers access floating point arguments which do not fit in the
9944 RSA from the stack when a subroutine is compiled without
9945 optimization. Because always storing floating-point arguments on the
9946 stack is inefficient and rarely needed, this option is not enabled by
9947 default and only is necessary when calling subroutines compiled by AIX
9948 XL compilers without optimization.
9952 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
9953 application written to use message passing with special startup code to
9954 enable the application to run. The system must have PE installed in the
9955 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
9956 must be overridden with the @option{-specs=} option to specify the
9957 appropriate directory location. The Parallel Environment does not
9958 support threads, so the @option{-mpe} option and the @option{-pthread}
9959 option are incompatible.
9961 @item -malign-natural
9962 @itemx -malign-power
9963 @opindex malign-natural
9964 @opindex malign-power
9965 On AIX, Darwin, and 64-bit PowerPC GNU/Linux, the option
9966 @option{-malign-natural} overrides the ABI-defined alignment of larger
9967 types, such as floating-point doubles, on their natural size-based boundary.
9968 The option @option{-malign-power} instructs GCC to follow the ABI-specified
9969 alignment rules. GCC defaults to the standard alignment defined in the ABI.
9973 @opindex msoft-float
9974 @opindex mhard-float
9975 Generate code that does not use (uses) the floating-point register set.
9976 Software floating point emulation is provided if you use the
9977 @option{-msoft-float} option, and pass the option to GCC when linking.
9980 @itemx -mno-multiple
9982 @opindex mno-multiple
9983 Generate code that uses (does not use) the load multiple word
9984 instructions and the store multiple word instructions. These
9985 instructions are generated by default on POWER systems, and not
9986 generated on PowerPC systems. Do not use @option{-mmultiple} on little
9987 endian PowerPC systems, since those instructions do not work when the
9988 processor is in little endian mode. The exceptions are PPC740 and
9989 PPC750 which permit the instructions usage in little endian mode.
9995 Generate code that uses (does not use) the load string instructions
9996 and the store string word instructions to save multiple registers and
9997 do small block moves. These instructions are generated by default on
9998 POWER systems, and not generated on PowerPC systems. Do not use
9999 @option{-mstring} on little endian PowerPC systems, since those
10000 instructions do not work when the processor is in little endian mode.
10001 The exceptions are PPC740 and PPC750 which permit the instructions
10002 usage in little endian mode.
10007 @opindex mno-update
10008 Generate code that uses (does not use) the load or store instructions
10009 that update the base register to the address of the calculated memory
10010 location. These instructions are generated by default. If you use
10011 @option{-mno-update}, there is a small window between the time that the
10012 stack pointer is updated and the address of the previous frame is
10013 stored, which means code that walks the stack frame across interrupts or
10014 signals may get corrupted data.
10017 @itemx -mno-fused-madd
10018 @opindex mfused-madd
10019 @opindex mno-fused-madd
10020 Generate code that uses (does not use) the floating point multiply and
10021 accumulate instructions. These instructions are generated by default if
10022 hardware floating is used.
10024 @item -mno-bit-align
10026 @opindex mno-bit-align
10027 @opindex mbit-align
10028 On System V.4 and embedded PowerPC systems do not (do) force structures
10029 and unions that contain bit-fields to be aligned to the base type of the
10032 For example, by default a structure containing nothing but 8
10033 @code{unsigned} bit-fields of length 1 would be aligned to a 4 byte
10034 boundary and have a size of 4 bytes. By using @option{-mno-bit-align},
10035 the structure would be aligned to a 1 byte boundary and be one byte in
10038 @item -mno-strict-align
10039 @itemx -mstrict-align
10040 @opindex mno-strict-align
10041 @opindex mstrict-align
10042 On System V.4 and embedded PowerPC systems do not (do) assume that
10043 unaligned memory references will be handled by the system.
10045 @item -mrelocatable
10046 @itemx -mno-relocatable
10047 @opindex mrelocatable
10048 @opindex mno-relocatable
10049 On embedded PowerPC systems generate code that allows (does not allow)
10050 the program to be relocated to a different address at runtime. If you
10051 use @option{-mrelocatable} on any module, all objects linked together must
10052 be compiled with @option{-mrelocatable} or @option{-mrelocatable-lib}.
10054 @item -mrelocatable-lib
10055 @itemx -mno-relocatable-lib
10056 @opindex mrelocatable-lib
10057 @opindex mno-relocatable-lib
10058 On embedded PowerPC systems generate code that allows (does not allow)
10059 the program to be relocated to a different address at runtime. Modules
10060 compiled with @option{-mrelocatable-lib} can be linked with either modules
10061 compiled without @option{-mrelocatable} and @option{-mrelocatable-lib} or
10062 with modules compiled with the @option{-mrelocatable} options.
10068 On System V.4 and embedded PowerPC systems do not (do) assume that
10069 register 2 contains a pointer to a global area pointing to the addresses
10070 used in the program.
10073 @itemx -mlittle-endian
10075 @opindex mlittle-endian
10076 On System V.4 and embedded PowerPC systems compile code for the
10077 processor in little endian mode. The @option{-mlittle-endian} option is
10078 the same as @option{-mlittle}.
10081 @itemx -mbig-endian
10083 @opindex mbig-endian
10084 On System V.4 and embedded PowerPC systems compile code for the
10085 processor in big endian mode. The @option{-mbig-endian} option is
10086 the same as @option{-mbig}.
10088 @item -mdynamic-no-pic
10089 @opindex mdynamic-no-pic
10090 On Darwin and Mac OS X systems, compile code so that it is not
10091 relocatable, but that its external references are relocatable. The
10092 resulting code is suitable for applications, but not shared
10095 @item -mprioritize-restricted-insns=@var{priority}
10096 @opindex mprioritize-restricted-insns
10097 This option controls the priority that is assigned to
10098 dispatch-slot restricted instructions during the second scheduling
10099 pass. The argument @var{priority} takes the value @var{0/1/2} to assign
10100 @var{no/highest/second-highest} priority to dispatch slot restricted
10103 @item -msched-costly-dep=@var{dependence_type}
10104 @opindex msched-costly-dep
10105 This option controls which dependences are considered costly
10106 by the target during instruction scheduling. The argument
10107 @var{dependence_type} takes one of the following values:
10108 @var{no}: no dependence is costly,
10109 @var{all}: all dependences are costly,
10110 @var{true_store_to_load}: a true dependence from store to load is costly,
10111 @var{store_to_load}: any dependence from store to load is costly,
10112 @var{number}: any dependence which latency >= @var{number} is costly.
10114 @item -minsert-sched-nops=@var{scheme}
10115 @opindex minsert-sched-nops
10116 This option controls which nop insertion scheme will be used during
10117 the second scheduling pass. The argument @var{scheme} takes one of the
10119 @var{no}: Don't insert nops.
10120 @var{pad}: Pad with nops any dispatch group which has vacant issue slots,
10121 according to the scheduler's grouping.
10122 @var{regroup_exact}: Insert nops to force costly dependent insns into
10123 separate groups. Insert exactly as many nops as needed to force an insn
10124 to a new group, according to the estimated processor grouping.
10125 @var{number}: Insert nops to force costly dependent insns into
10126 separate groups. Insert @var{number} nops to force an insn to a new group.
10129 @opindex mcall-sysv
10130 On System V.4 and embedded PowerPC systems compile code using calling
10131 conventions that adheres to the March 1995 draft of the System V
10132 Application Binary Interface, PowerPC processor supplement. This is the
10133 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
10135 @item -mcall-sysv-eabi
10136 @opindex mcall-sysv-eabi
10137 Specify both @option{-mcall-sysv} and @option{-meabi} options.
10139 @item -mcall-sysv-noeabi
10140 @opindex mcall-sysv-noeabi
10141 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
10143 @item -mcall-solaris
10144 @opindex mcall-solaris
10145 On System V.4 and embedded PowerPC systems compile code for the Solaris
10149 @opindex mcall-linux
10150 On System V.4 and embedded PowerPC systems compile code for the
10151 Linux-based GNU system.
10155 On System V.4 and embedded PowerPC systems compile code for the
10156 Hurd-based GNU system.
10158 @item -mcall-netbsd
10159 @opindex mcall-netbsd
10160 On System V.4 and embedded PowerPC systems compile code for the
10161 NetBSD operating system.
10163 @item -maix-struct-return
10164 @opindex maix-struct-return
10165 Return all structures in memory (as specified by the AIX ABI)@.
10167 @item -msvr4-struct-return
10168 @opindex msvr4-struct-return
10169 Return structures smaller than 8 bytes in registers (as specified by the
10172 @item -mabi=altivec
10173 @opindex mabi=altivec
10174 Extend the current ABI with AltiVec ABI extensions. This does not
10175 change the default ABI, instead it adds the AltiVec ABI extensions to
10178 @item -mabi=no-altivec
10179 @opindex mabi=no-altivec
10180 Disable AltiVec ABI extensions for the current ABI.
10183 @itemx -mno-prototype
10184 @opindex mprototype
10185 @opindex mno-prototype
10186 On System V.4 and embedded PowerPC systems assume that all calls to
10187 variable argument functions are properly prototyped. Otherwise, the
10188 compiler must insert an instruction before every non prototyped call to
10189 set or clear bit 6 of the condition code register (@var{CR}) to
10190 indicate whether floating point values were passed in the floating point
10191 registers in case the function takes a variable arguments. With
10192 @option{-mprototype}, only calls to prototyped variable argument functions
10193 will set or clear the bit.
10197 On embedded PowerPC systems, assume that the startup module is called
10198 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
10199 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}.
10204 On embedded PowerPC systems, assume that the startup module is called
10205 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
10210 On embedded PowerPC systems, assume that the startup module is called
10211 @file{crt0.o} and the standard C libraries are @file{libads.a} and
10214 @item -myellowknife
10215 @opindex myellowknife
10216 On embedded PowerPC systems, assume that the startup module is called
10217 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
10222 On System V.4 and embedded PowerPC systems, specify that you are
10223 compiling for a VxWorks system.
10227 Specify that you are compiling for the WindISS simulation environment.
10231 On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags
10232 header to indicate that @samp{eabi} extended relocations are used.
10238 On System V.4 and embedded PowerPC systems do (do not) adhere to the
10239 Embedded Applications Binary Interface (eabi) which is a set of
10240 modifications to the System V.4 specifications. Selecting @option{-meabi}
10241 means that the stack is aligned to an 8 byte boundary, a function
10242 @code{__eabi} is called to from @code{main} to set up the eabi
10243 environment, and the @option{-msdata} option can use both @code{r2} and
10244 @code{r13} to point to two separate small data areas. Selecting
10245 @option{-mno-eabi} means that the stack is aligned to a 16 byte boundary,
10246 do not call an initialization function from @code{main}, and the
10247 @option{-msdata} option will only use @code{r13} to point to a single
10248 small data area. The @option{-meabi} option is on by default if you
10249 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
10252 @opindex msdata=eabi
10253 On System V.4 and embedded PowerPC systems, put small initialized
10254 @code{const} global and static data in the @samp{.sdata2} section, which
10255 is pointed to by register @code{r2}. Put small initialized
10256 non-@code{const} global and static data in the @samp{.sdata} section,
10257 which is pointed to by register @code{r13}. Put small uninitialized
10258 global and static data in the @samp{.sbss} section, which is adjacent to
10259 the @samp{.sdata} section. The @option{-msdata=eabi} option is
10260 incompatible with the @option{-mrelocatable} option. The
10261 @option{-msdata=eabi} option also sets the @option{-memb} option.
10264 @opindex msdata=sysv
10265 On System V.4 and embedded PowerPC systems, put small global and static
10266 data in the @samp{.sdata} section, which is pointed to by register
10267 @code{r13}. Put small uninitialized global and static data in the
10268 @samp{.sbss} section, which is adjacent to the @samp{.sdata} section.
10269 The @option{-msdata=sysv} option is incompatible with the
10270 @option{-mrelocatable} option.
10272 @item -msdata=default
10274 @opindex msdata=default
10276 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
10277 compile code the same as @option{-msdata=eabi}, otherwise compile code the
10278 same as @option{-msdata=sysv}.
10281 @opindex msdata-data
10282 On System V.4 and embedded PowerPC systems, put small global and static
10283 data in the @samp{.sdata} section. Put small uninitialized global and
10284 static data in the @samp{.sbss} section. Do not use register @code{r13}
10285 to address small data however. This is the default behavior unless
10286 other @option{-msdata} options are used.
10290 @opindex msdata=none
10292 On embedded PowerPC systems, put all initialized global and static data
10293 in the @samp{.data} section, and all uninitialized data in the
10294 @samp{.bss} section.
10298 @cindex smaller data references (PowerPC)
10299 @cindex .sdata/.sdata2 references (PowerPC)
10300 On embedded PowerPC systems, put global and static items less than or
10301 equal to @var{num} bytes into the small data or bss sections instead of
10302 the normal data or bss section. By default, @var{num} is 8. The
10303 @option{-G @var{num}} switch is also passed to the linker.
10304 All modules should be compiled with the same @option{-G @var{num}} value.
10307 @itemx -mno-regnames
10309 @opindex mno-regnames
10310 On System V.4 and embedded PowerPC systems do (do not) emit register
10311 names in the assembly language output using symbolic forms.
10314 @itemx -mno-longcall
10316 @opindex mno-longcall
10317 Default to making all function calls indirectly, using a register, so
10318 that functions which reside further than 32 megabytes (33,554,432
10319 bytes) from the current location can be called. This setting can be
10320 overridden by the @code{shortcall} function attribute, or by
10321 @code{#pragma longcall(0)}.
10323 Some linkers are capable of detecting out-of-range calls and generating
10324 glue code on the fly. On these systems, long calls are unnecessary and
10325 generate slower code. As of this writing, the AIX linker can do this,
10326 as can the GNU linker for PowerPC/64. It is planned to add this feature
10327 to the GNU linker for 32-bit PowerPC systems as well.
10329 On Darwin/PPC systems, @code{#pragma longcall} will generate ``jbsr
10330 callee, L42'', plus a ``branch island'' (glue code). The two target
10331 addresses represent the callee and the ``branch island.'' The
10332 Darwin/PPC linker will prefer the first address and generate a ``bl
10333 callee'' if the PPC ``bl'' instruction will reach the callee directly;
10334 otherwise, the linker will generate ``bl L42'' to call the ``branch
10335 island.'' The ``branch island'' is appended to the body of the
10336 calling function; it computes the full 32-bit address of the callee
10339 On Mach-O (Darwin) systems, this option directs the compiler emit to
10340 the glue for every direct call, and the Darwin linker decides whether
10341 to use or discard it.
10343 In the future, we may cause GCC to ignore all longcall specifications
10344 when the linker is known to generate glue.
10348 Adds support for multithreading with the @dfn{pthreads} library.
10349 This option sets flags for both the preprocessor and linker.
10353 @node S/390 and zSeries Options
10354 @subsection S/390 and zSeries Options
10355 @cindex S/390 and zSeries Options
10357 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
10361 @itemx -msoft-float
10362 @opindex mhard-float
10363 @opindex msoft-float
10364 Use (do not use) the hardware floating-point instructions and registers
10365 for floating-point operations. When @option{-msoft-float} is specified,
10366 functions in @file{libgcc.a} will be used to perform floating-point
10367 operations. When @option{-mhard-float} is specified, the compiler
10368 generates IEEE floating-point instructions. This is the default.
10371 @itemx -mno-backchain
10372 @opindex mbackchain
10373 @opindex mno-backchain
10374 Generate (or do not generate) code which maintains an explicit
10375 backchain within the stack frame that points to the caller's frame.
10376 This may be needed to allow debugging using tools that do not understand
10377 DWARF-2 call frame information. The default is not to generate the
10381 @itemx -mno-small-exec
10382 @opindex msmall-exec
10383 @opindex mno-small-exec
10384 Generate (or do not generate) code using the @code{bras} instruction
10385 to do subroutine calls.
10386 This only works reliably if the total executable size does not
10387 exceed 64k. The default is to use the @code{basr} instruction instead,
10388 which does not have this limitation.
10394 When @option{-m31} is specified, generate code compliant to the
10395 GNU/Linux for S/390 ABI@. When @option{-m64} is specified, generate
10396 code compliant to the GNU/Linux for zSeries ABI@. This allows GCC in
10397 particular to generate 64-bit instructions. For the @samp{s390}
10398 targets, the default is @option{-m31}, while the @samp{s390x}
10399 targets default to @option{-m64}.
10405 When @option{-mzarch} is specified, generate code using the
10406 instructions available on z/Architecture.
10407 When @option{-mesa} is specified, generate code using the
10408 instructions available on ESA/390. Note that @option{-mesa} is
10409 not possible with @option{-m64}.
10410 When generating code compliant to the GNU/Linux for S/390 ABI,
10411 the default is @option{-mesa}. When generating code compliant
10412 to the GNU/Linux for zSeries ABI, the default is @option{-mzarch}.
10418 Generate (or do not generate) code using the @code{mvcle} instruction
10419 to perform block moves. When @option{-mno-mvcle} is specified,
10420 use a @code{mvc} loop instead. This is the default.
10426 Print (or do not print) additional debug information when compiling.
10427 The default is to not print debug information.
10429 @item -march=@var{cpu-type}
10431 Generate code that will run on @var{cpu-type}, which is the name of a system
10432 representing a certain processor type. Possible values for
10433 @var{cpu-type} are @samp{g5}, @samp{g6}, @samp{z900}, and @samp{z990}.
10434 When generating code using the instructions available on z/Architecture,
10435 the default is @option{-march=z900}. Otherwise, the default is
10436 @option{-march=g5}.
10438 @item -mtune=@var{cpu-type}
10440 Tune to @var{cpu-type} everything applicable about the generated code,
10441 except for the ABI and the set of available instructions.
10442 The list of @var{cpu-type} values is the same as for @option{-march}.
10443 The default is the value used for @option{-march}.
10446 @itemx -mno-tpf-trace
10447 @opindex mtpf-trace
10448 @opindex mno-tpf-trace
10449 Generate code that adds (does not add) in TPF OS specific branches to trace
10450 routines in the operating system. This option is off by default, even
10451 when compiling for the TPF OS.
10454 @itemx -mno-fused-madd
10455 @opindex mfused-madd
10456 @opindex mno-fused-madd
10457 Generate code that uses (does not use) the floating point multiply and
10458 accumulate instructions. These instructions are generated by default if
10459 hardware floating point is used.
10463 @subsection SH Options
10465 These @samp{-m} options are defined for the SH implementations:
10470 Generate code for the SH1.
10474 Generate code for the SH2.
10477 Generate code for the SH2e.
10481 Generate code for the SH3.
10485 Generate code for the SH3e.
10489 Generate code for the SH4 without a floating-point unit.
10491 @item -m4-single-only
10492 @opindex m4-single-only
10493 Generate code for the SH4 with a floating-point unit that only
10494 supports single-precision arithmetic.
10498 Generate code for the SH4 assuming the floating-point unit is in
10499 single-precision mode by default.
10503 Generate code for the SH4.
10507 Compile code for the processor in big endian mode.
10511 Compile code for the processor in little endian mode.
10515 Align doubles at 64-bit boundaries. Note that this changes the calling
10516 conventions, and thus some functions from the standard C library will
10517 not work unless you recompile it first with @option{-mdalign}.
10521 Shorten some address references at link time, when possible; uses the
10522 linker option @option{-relax}.
10526 Use 32-bit offsets in @code{switch} tables. The default is to use
10531 Enable the use of the instruction @code{fmovd}.
10535 Comply with the calling conventions defined by Renesas.
10538 @opindex mnomacsave
10539 Mark the @code{MAC} register as call-clobbered, even if
10540 @option{-mhitachi} is given.
10544 Increase IEEE-compliance of floating-point code.
10548 Dump instruction size and location in the assembly code.
10551 @opindex mpadstruct
10552 This option is deprecated. It pads structures to multiple of 4 bytes,
10553 which is incompatible with the SH ABI@.
10557 Optimize for space instead of speed. Implied by @option{-Os}.
10560 @opindex mprefergot
10561 When generating position-independent code, emit function calls using
10562 the Global Offset Table instead of the Procedure Linkage Table.
10566 Generate a library function call to invalidate instruction cache
10567 entries, after fixing up a trampoline. This library function call
10568 doesn't assume it can write to the whole memory address space. This
10569 is the default when the target is @code{sh-*-linux*}.
10572 @node SPARC Options
10573 @subsection SPARC Options
10574 @cindex SPARC options
10576 These @samp{-m} options are supported on the SPARC:
10579 @item -mno-app-regs
10581 @opindex mno-app-regs
10583 Specify @option{-mapp-regs} to generate output using the global registers
10584 2 through 4, which the SPARC SVR4 ABI reserves for applications. This
10587 To be fully SVR4 ABI compliant at the cost of some performance loss,
10588 specify @option{-mno-app-regs}. You should compile libraries and system
10589 software with this option.
10592 @itemx -mhard-float
10594 @opindex mhard-float
10595 Generate output containing floating point instructions. This is the
10599 @itemx -msoft-float
10601 @opindex msoft-float
10602 Generate output containing library calls for floating point.
10603 @strong{Warning:} the requisite libraries are not available for all SPARC
10604 targets. Normally the facilities of the machine's usual C compiler are
10605 used, but this cannot be done directly in cross-compilation. You must make
10606 your own arrangements to provide suitable library functions for
10607 cross-compilation. The embedded targets @samp{sparc-*-aout} and
10608 @samp{sparclite-*-*} do provide software floating point support.
10610 @option{-msoft-float} changes the calling convention in the output file;
10611 therefore, it is only useful if you compile @emph{all} of a program with
10612 this option. In particular, you need to compile @file{libgcc.a}, the
10613 library that comes with GCC, with @option{-msoft-float} in order for
10616 @item -mhard-quad-float
10617 @opindex mhard-quad-float
10618 Generate output containing quad-word (long double) floating point
10621 @item -msoft-quad-float
10622 @opindex msoft-quad-float
10623 Generate output containing library calls for quad-word (long double)
10624 floating point instructions. The functions called are those specified
10625 in the SPARC ABI@. This is the default.
10627 As of this writing, there are no SPARC implementations that have hardware
10628 support for the quad-word floating point instructions. They all invoke
10629 a trap handler for one of these instructions, and then the trap handler
10630 emulates the effect of the instruction. Because of the trap handler overhead,
10631 this is much slower than calling the ABI library routines. Thus the
10632 @option{-msoft-quad-float} option is the default.
10634 @item -mno-unaligned-doubles
10635 @itemx -munaligned-doubles
10636 @opindex mno-unaligned-doubles
10637 @opindex munaligned-doubles
10638 Assume that doubles have 8 byte alignment. This is the default.
10640 With @option{-munaligned-doubles}, GCC assumes that doubles have 8 byte
10641 alignment only if they are contained in another type, or if they have an
10642 absolute address. Otherwise, it assumes they have 4 byte alignment.
10643 Specifying this option avoids some rare compatibility problems with code
10644 generated by other compilers. It is not the default because it results
10645 in a performance loss, especially for floating point code.
10647 @item -mno-faster-structs
10648 @itemx -mfaster-structs
10649 @opindex mno-faster-structs
10650 @opindex mfaster-structs
10651 With @option{-mfaster-structs}, the compiler assumes that structures
10652 should have 8 byte alignment. This enables the use of pairs of
10653 @code{ldd} and @code{std} instructions for copies in structure
10654 assignment, in place of twice as many @code{ld} and @code{st} pairs.
10655 However, the use of this changed alignment directly violates the SPARC
10656 ABI@. Thus, it's intended only for use on targets where the developer
10657 acknowledges that their resulting code will not be directly in line with
10658 the rules of the ABI@.
10660 @item -mimpure-text
10661 @opindex mimpure-text
10662 @option{-mimpure-text}, used in addition to @option{-shared}, tells
10663 the compiler to not pass @option{-z text} to the linker when linking a
10664 shared object. Using this option, you can link position-dependent
10665 code into a shared object.
10667 @option{-mimpure-text} suppresses the ``relocations remain against
10668 allocatable but non-writable sections'' linker error message.
10669 However, the necessary relocations will trigger copy-on-write, and the
10670 shared object is not actually shared across processes. Instead of
10671 using @option{-mimpure-text}, you should compile all source code with
10672 @option{-fpic} or @option{-fPIC}.
10674 This option is only available on SunOS and Solaris.
10676 @item -mcpu=@var{cpu_type}
10678 Set the instruction set, register set, and instruction scheduling parameters
10679 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
10680 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite},
10681 @samp{f930}, @samp{f934}, @samp{hypersparc}, @samp{sparclite86x},
10682 @samp{sparclet}, @samp{tsc701}, @samp{v9}, @samp{ultrasparc}, and
10683 @samp{ultrasparc3}.
10685 Default instruction scheduling parameters are used for values that select
10686 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
10687 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
10689 Here is a list of each supported architecture and their supported
10694 v8: supersparc, hypersparc
10695 sparclite: f930, f934, sparclite86x
10697 v9: ultrasparc, ultrasparc3
10700 By default (unless configured otherwise), GCC generates code for the V7
10701 variant of the SPARC architecture. With @option{-mcpu=cypress}, the compiler
10702 additionally optimizes it for the Cypress CY7C602 chip, as used in the
10703 SPARCStation/SPARCServer 3xx series. This is also appropriate for the older
10704 SPARCStation 1, 2, IPX etc.
10706 With @option{-mcpu=v8}, GCC generates code for the V8 variant of the SPARC
10707 architecture. The only difference from V7 code is that the compiler emits
10708 the integer multiply and integer divide instructions which exist in SPARC-V8
10709 but not in SPARC-V7. With @option{-mcpu=supersparc}, the compiler additionally
10710 optimizes it for the SuperSPARC chip, as used in the SPARCStation 10, 1000 and
10713 With @option{-mcpu=sparclite}, GCC generates code for the SPARClite variant of
10714 the SPARC architecture. This adds the integer multiply, integer divide step
10715 and scan (@code{ffs}) instructions which exist in SPARClite but not in SPARC-V7.
10716 With @option{-mcpu=f930}, the compiler additionally optimizes it for the
10717 Fujitsu MB86930 chip, which is the original SPARClite, with no FPU. With
10718 @option{-mcpu=f934}, the compiler additionally optimizes it for the Fujitsu
10719 MB86934 chip, which is the more recent SPARClite with FPU.
10721 With @option{-mcpu=sparclet}, GCC generates code for the SPARClet variant of
10722 the SPARC architecture. This adds the integer multiply, multiply/accumulate,
10723 integer divide step and scan (@code{ffs}) instructions which exist in SPARClet
10724 but not in SPARC-V7. With @option{-mcpu=tsc701}, the compiler additionally
10725 optimizes it for the TEMIC SPARClet chip.
10727 With @option{-mcpu=v9}, GCC generates code for the V9 variant of the SPARC
10728 architecture. This adds 64-bit integer and floating-point move instructions,
10729 3 additional floating-point condition code registers and conditional move
10730 instructions. With @option{-mcpu=ultrasparc}, the compiler additionally
10731 optimizes it for the Sun UltraSPARC I/II chips. With
10732 @option{-mcpu=ultrasparc3}, the compiler additionally optimizes it for the
10733 Sun UltraSPARC III chip.
10735 @item -mtune=@var{cpu_type}
10737 Set the instruction scheduling parameters for machine type
10738 @var{cpu_type}, but do not set the instruction set or register set that the
10739 option @option{-mcpu=@var{cpu_type}} would.
10741 The same values for @option{-mcpu=@var{cpu_type}} can be used for
10742 @option{-mtune=@var{cpu_type}}, but the only useful values are those
10743 that select a particular cpu implementation. Those are @samp{cypress},
10744 @samp{supersparc}, @samp{hypersparc}, @samp{f930}, @samp{f934},
10745 @samp{sparclite86x}, @samp{tsc701}, @samp{ultrasparc}, and
10746 @samp{ultrasparc3}.
10751 @opindex mno-v8plus
10752 With @option{-mv8plus}, GCC generates code for the SPARC-V8+ ABI. The
10753 difference from the V8 ABI is that the global and out registers are
10754 considered 64-bit wide. This is enabled by default on Solaris in 32-bit
10755 mode for all SPARC-V9 processors.
10761 With @option{-mvis}, GCC generates code that takes advantage of the UltraSPARC
10762 Visual Instruction Set extensions. The default is @option{-mno-vis}.
10765 These @samp{-m} options are supported in addition to the above
10766 on SPARC-V9 processors in 64-bit environments:
10769 @item -mlittle-endian
10770 @opindex mlittle-endian
10771 Generate code for a processor running in little-endian mode. It is only
10772 available for a few configurations and most notably not on Solaris.
10778 Generate code for a 32-bit or 64-bit environment.
10779 The 32-bit environment sets int, long and pointer to 32 bits.
10780 The 64-bit environment sets int to 32 bits and long and pointer
10783 @item -mcmodel=medlow
10784 @opindex mcmodel=medlow
10785 Generate code for the Medium/Low code model: 64-bit addresses, programs
10786 must be linked in the low 32 bits of memory. Programs can be statically
10787 or dynamically linked.
10789 @item -mcmodel=medmid
10790 @opindex mcmodel=medmid
10791 Generate code for the Medium/Middle code model: 64-bit addresses, programs
10792 must be linked in the low 44 bits of memory, the text and data segments must
10793 be less than 2GB in size and the data segment must be located within 2GB of
10796 @item -mcmodel=medany
10797 @opindex mcmodel=medany
10798 Generate code for the Medium/Anywhere code model: 64-bit addresses, programs
10799 may be linked anywhere in memory, the text and data segments must be less
10800 than 2GB in size and the data segment must be located within 2GB of the
10803 @item -mcmodel=embmedany
10804 @opindex mcmodel=embmedany
10805 Generate code for the Medium/Anywhere code model for embedded systems:
10806 64-bit addresses, the text and data segments must be less than 2GB in
10807 size, both starting anywhere in memory (determined at link time). The
10808 global register %g4 points to the base of the data segment. Programs
10809 are statically linked and PIC is not supported.
10812 @itemx -mno-stack-bias
10813 @opindex mstack-bias
10814 @opindex mno-stack-bias
10815 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
10816 frame pointer if present, are offset by @minus{}2047 which must be added back
10817 when making stack frame references. This is the default in 64-bit mode.
10818 Otherwise, assume no such offset is present.
10821 These switches are supported in addition to the above on Solaris:
10826 Add support for multithreading using the Solaris threads library. This
10827 option sets flags for both the preprocessor and linker. This option does
10828 not affect the thread safety of object code produced by the compiler or
10829 that of libraries supplied with it.
10833 Add support for multithreading using the POSIX threads library. This
10834 option sets flags for both the preprocessor and linker. This option does
10835 not affect the thread safety of object code produced by the compiler or
10836 that of libraries supplied with it.
10839 @node System V Options
10840 @subsection Options for System V
10842 These additional options are available on System V Release 4 for
10843 compatibility with other compilers on those systems:
10848 Create a shared object.
10849 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
10853 Identify the versions of each tool used by the compiler, in a
10854 @code{.ident} assembler directive in the output.
10858 Refrain from adding @code{.ident} directives to the output file (this is
10861 @item -YP,@var{dirs}
10863 Search the directories @var{dirs}, and no others, for libraries
10864 specified with @option{-l}.
10866 @item -Ym,@var{dir}
10868 Look in the directory @var{dir} to find the M4 preprocessor.
10869 The assembler uses this option.
10870 @c This is supposed to go with a -Yd for predefined M4 macro files, but
10871 @c the generic assembler that comes with Solaris takes just -Ym.
10874 @node TMS320C3x/C4x Options
10875 @subsection TMS320C3x/C4x Options
10876 @cindex TMS320C3x/C4x Options
10878 These @samp{-m} options are defined for TMS320C3x/C4x implementations:
10882 @item -mcpu=@var{cpu_type}
10884 Set the instruction set, register set, and instruction scheduling
10885 parameters for machine type @var{cpu_type}. Supported values for
10886 @var{cpu_type} are @samp{c30}, @samp{c31}, @samp{c32}, @samp{c40}, and
10887 @samp{c44}. The default is @samp{c40} to generate code for the
10892 @itemx -msmall-memory
10894 @opindex mbig-memory
10896 @opindex msmall-memory
10898 Generates code for the big or small memory model. The small memory
10899 model assumed that all data fits into one 64K word page. At run-time
10900 the data page (DP) register must be set to point to the 64K page
10901 containing the .bss and .data program sections. The big memory model is
10902 the default and requires reloading of the DP register for every direct
10909 Allow (disallow) allocation of general integer operands into the block
10910 count register BK@.
10916 Enable (disable) generation of code using decrement and branch,
10917 DBcond(D), instructions. This is enabled by default for the C4x. To be
10918 on the safe side, this is disabled for the C3x, since the maximum
10919 iteration count on the C3x is @math{2^{23} + 1} (but who iterates loops more than
10920 @math{2^{23}} times on the C3x?). Note that GCC will try to reverse a loop so
10921 that it can utilize the decrement and branch instruction, but will give
10922 up if there is more than one memory reference in the loop. Thus a loop
10923 where the loop counter is decremented can generate slightly more
10924 efficient code, in cases where the RPTB instruction cannot be utilized.
10926 @item -mdp-isr-reload
10928 @opindex mdp-isr-reload
10930 Force the DP register to be saved on entry to an interrupt service
10931 routine (ISR), reloaded to point to the data section, and restored on
10932 exit from the ISR@. This should not be required unless someone has
10933 violated the small memory model by modifying the DP register, say within
10940 For the C3x use the 24-bit MPYI instruction for integer multiplies
10941 instead of a library call to guarantee 32-bit results. Note that if one
10942 of the operands is a constant, then the multiplication will be performed
10943 using shifts and adds. If the @option{-mmpyi} option is not specified for the C3x,
10944 then squaring operations are performed inline instead of a library call.
10947 @itemx -mno-fast-fix
10949 @opindex mno-fast-fix
10950 The C3x/C4x FIX instruction to convert a floating point value to an
10951 integer value chooses the nearest integer less than or equal to the
10952 floating point value rather than to the nearest integer. Thus if the
10953 floating point number is negative, the result will be incorrectly
10954 truncated an additional code is necessary to detect and correct this
10955 case. This option can be used to disable generation of the additional
10956 code required to correct the result.
10962 Enable (disable) generation of repeat block sequences using the RPTB
10963 instruction for zero overhead looping. The RPTB construct is only used
10964 for innermost loops that do not call functions or jump across the loop
10965 boundaries. There is no advantage having nested RPTB loops due to the
10966 overhead required to save and restore the RC, RS, and RE registers.
10967 This is enabled by default with @option{-O2}.
10969 @item -mrpts=@var{count}
10973 Enable (disable) the use of the single instruction repeat instruction
10974 RPTS@. If a repeat block contains a single instruction, and the loop
10975 count can be guaranteed to be less than the value @var{count}, GCC will
10976 emit a RPTS instruction instead of a RPTB@. If no value is specified,
10977 then a RPTS will be emitted even if the loop count cannot be determined
10978 at compile time. Note that the repeated instruction following RPTS does
10979 not have to be reloaded from memory each iteration, thus freeing up the
10980 CPU buses for operands. However, since interrupts are blocked by this
10981 instruction, it is disabled by default.
10983 @item -mloop-unsigned
10984 @itemx -mno-loop-unsigned
10985 @opindex mloop-unsigned
10986 @opindex mno-loop-unsigned
10987 The maximum iteration count when using RPTS and RPTB (and DB on the C40)
10988 is @math{2^{31} + 1} since these instructions test if the iteration count is
10989 negative to terminate the loop. If the iteration count is unsigned
10990 there is a possibility than the @math{2^{31} + 1} maximum iteration count may be
10991 exceeded. This switch allows an unsigned iteration count.
10995 Try to emit an assembler syntax that the TI assembler (asm30) is happy
10996 with. This also enforces compatibility with the API employed by the TI
10997 C3x C compiler. For example, long doubles are passed as structures
10998 rather than in floating point registers.
11004 Generate code that uses registers (stack) for passing arguments to functions.
11005 By default, arguments are passed in registers where possible rather
11006 than by pushing arguments on to the stack.
11008 @item -mparallel-insns
11009 @itemx -mno-parallel-insns
11010 @opindex mparallel-insns
11011 @opindex mno-parallel-insns
11012 Allow the generation of parallel instructions. This is enabled by
11013 default with @option{-O2}.
11015 @item -mparallel-mpy
11016 @itemx -mno-parallel-mpy
11017 @opindex mparallel-mpy
11018 @opindex mno-parallel-mpy
11019 Allow the generation of MPY||ADD and MPY||SUB parallel instructions,
11020 provided @option{-mparallel-insns} is also specified. These instructions have
11021 tight register constraints which can pessimize the code generation
11022 of large functions.
11027 @subsection V850 Options
11028 @cindex V850 Options
11030 These @samp{-m} options are defined for V850 implementations:
11034 @itemx -mno-long-calls
11035 @opindex mlong-calls
11036 @opindex mno-long-calls
11037 Treat all calls as being far away (near). If calls are assumed to be
11038 far away, the compiler will always load the functions address up into a
11039 register, and call indirect through the pointer.
11045 Do not optimize (do optimize) basic blocks that use the same index
11046 pointer 4 or more times to copy pointer into the @code{ep} register, and
11047 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
11048 option is on by default if you optimize.
11050 @item -mno-prolog-function
11051 @itemx -mprolog-function
11052 @opindex mno-prolog-function
11053 @opindex mprolog-function
11054 Do not use (do use) external functions to save and restore registers
11055 at the prologue and epilogue of a function. The external functions
11056 are slower, but use less code space if more than one function saves
11057 the same number of registers. The @option{-mprolog-function} option
11058 is on by default if you optimize.
11062 Try to make the code as small as possible. At present, this just turns
11063 on the @option{-mep} and @option{-mprolog-function} options.
11065 @item -mtda=@var{n}
11067 Put static or global variables whose size is @var{n} bytes or less into
11068 the tiny data area that register @code{ep} points to. The tiny data
11069 area can hold up to 256 bytes in total (128 bytes for byte references).
11071 @item -msda=@var{n}
11073 Put static or global variables whose size is @var{n} bytes or less into
11074 the small data area that register @code{gp} points to. The small data
11075 area can hold up to 64 kilobytes.
11077 @item -mzda=@var{n}
11079 Put static or global variables whose size is @var{n} bytes or less into
11080 the first 32 kilobytes of memory.
11084 Specify that the target processor is the V850.
11087 @opindex mbig-switch
11088 Generate code suitable for big switch tables. Use this option only if
11089 the assembler/linker complain about out of range branches within a switch
11094 This option will cause r2 and r5 to be used in the code generated by
11095 the compiler. This setting is the default.
11097 @item -mno-app-regs
11098 @opindex mno-app-regs
11099 This option will cause r2 and r5 to be treated as fixed registers.
11103 Specify that the target processor is the V850E1. The preprocessor
11104 constants @samp{__v850e1__} and @samp{__v850e__} will be defined if
11105 this option is used.
11109 Specify that the target processor is the V850E. The preprocessor
11110 constant @samp{__v850e__} will be defined if this option is used.
11112 If neither @option{-mv850} nor @option{-mv850e} nor @option{-mv850e1}
11113 are defined then a default target processor will be chosen and the
11114 relevant @samp{__v850*__} preprocessor constant will be defined.
11116 The preprocessor constants @samp{__v850} and @samp{__v851__} are always
11117 defined, regardless of which processor variant is the target.
11119 @item -mdisable-callt
11120 @opindex mdisable-callt
11121 This option will suppress generation of the CALLT instruction for the
11122 v850e and v850e1 flavors of the v850 architecture. The default is
11123 @option{-mno-disable-callt} which allows the CALLT instruction to be used.
11128 @subsection VAX Options
11129 @cindex VAX options
11131 These @samp{-m} options are defined for the VAX:
11136 Do not output certain jump instructions (@code{aobleq} and so on)
11137 that the Unix assembler for the VAX cannot handle across long
11142 Do output those jump instructions, on the assumption that you
11143 will assemble with the GNU assembler.
11147 Output code for g-format floating point numbers instead of d-format.
11150 @node x86-64 Options
11151 @subsection x86-64 Options
11152 @cindex x86-64 options
11154 These are listed under @xref{i386 and x86-64 Options}.
11156 @node Xstormy16 Options
11157 @subsection Xstormy16 Options
11158 @cindex Xstormy16 Options
11160 These options are defined for Xstormy16:
11165 Choose startup files and linker script suitable for the simulator.
11168 @node Xtensa Options
11169 @subsection Xtensa Options
11170 @cindex Xtensa Options
11172 These options are supported for Xtensa targets:
11176 @itemx -mno-const16
11178 @opindex mno-const16
11179 Enable or disable use of @code{CONST16} instructions for loading
11180 constant values. The @code{CONST16} instruction is currently not a
11181 standard option from Tensilica. When enabled, @code{CONST16}
11182 instructions are always used in place of the standard @code{L32R}
11183 instructions. The use of @code{CONST16} is enabled by default only if
11184 the @code{L32R} instruction is not available.
11187 @itemx -mno-fused-madd
11188 @opindex mfused-madd
11189 @opindex mno-fused-madd
11190 Enable or disable use of fused multiply/add and multiply/subtract
11191 instructions in the floating-point option. This has no effect if the
11192 floating-point option is not also enabled. Disabling fused multiply/add
11193 and multiply/subtract instructions forces the compiler to use separate
11194 instructions for the multiply and add/subtract operations. This may be
11195 desirable in some cases where strict IEEE 754-compliant results are
11196 required: the fused multiply add/subtract instructions do not round the
11197 intermediate result, thereby producing results with @emph{more} bits of
11198 precision than specified by the IEEE standard. Disabling fused multiply
11199 add/subtract instructions also ensures that the program output is not
11200 sensitive to the compiler's ability to combine multiply and add/subtract
11203 @item -mtext-section-literals
11204 @itemx -mno-text-section-literals
11205 @opindex mtext-section-literals
11206 @opindex mno-text-section-literals
11207 Control the treatment of literal pools. The default is
11208 @option{-mno-text-section-literals}, which places literals in a separate
11209 section in the output file. This allows the literal pool to be placed
11210 in a data RAM/ROM, and it also allows the linker to combine literal
11211 pools from separate object files to remove redundant literals and
11212 improve code size. With @option{-mtext-section-literals}, the literals
11213 are interspersed in the text section in order to keep them as close as
11214 possible to their references. This may be necessary for large assembly
11217 @item -mtarget-align
11218 @itemx -mno-target-align
11219 @opindex mtarget-align
11220 @opindex mno-target-align
11221 When this option is enabled, GCC instructs the assembler to
11222 automatically align instructions to reduce branch penalties at the
11223 expense of some code density. The assembler attempts to widen density
11224 instructions to align branch targets and the instructions following call
11225 instructions. If there are not enough preceding safe density
11226 instructions to align a target, no widening will be performed. The
11227 default is @option{-mtarget-align}. These options do not affect the
11228 treatment of auto-aligned instructions like @code{LOOP}, which the
11229 assembler will always align, either by widening density instructions or
11230 by inserting no-op instructions.
11233 @itemx -mno-longcalls
11234 @opindex mlongcalls
11235 @opindex mno-longcalls
11236 When this option is enabled, GCC instructs the assembler to translate
11237 direct calls to indirect calls unless it can determine that the target
11238 of a direct call is in the range allowed by the call instruction. This
11239 translation typically occurs for calls to functions in other source
11240 files. Specifically, the assembler translates a direct @code{CALL}
11241 instruction into an @code{L32R} followed by a @code{CALLX} instruction.
11242 The default is @option{-mno-longcalls}. This option should be used in
11243 programs where the call target can potentially be out of range. This
11244 option is implemented in the assembler, not the compiler, so the
11245 assembly code generated by GCC will still show direct call
11246 instructions---look at the disassembled object code to see the actual
11247 instructions. Note that the assembler will use an indirect call for
11248 every cross-file call, not just those that really will be out of range.
11251 @node zSeries Options
11252 @subsection zSeries Options
11253 @cindex zSeries options
11255 These are listed under @xref{S/390 and zSeries Options}.
11257 @node Code Gen Options
11258 @section Options for Code Generation Conventions
11259 @cindex code generation conventions
11260 @cindex options, code generation
11261 @cindex run-time options
11263 These machine-independent options control the interface conventions
11264 used in code generation.
11266 Most of them have both positive and negative forms; the negative form
11267 of @option{-ffoo} would be @option{-fno-foo}. In the table below, only
11268 one of the forms is listed---the one which is not the default. You
11269 can figure out the other form by either removing @samp{no-} or adding
11273 @item -fbounds-check
11274 @opindex fbounds-check
11275 For front-ends that support it, generate additional code to check that
11276 indices used to access arrays are within the declared range. This is
11277 currently only supported by the Java and Fortran 77 front-ends, where
11278 this option defaults to true and false respectively.
11282 This option generates traps for signed overflow on addition, subtraction,
11283 multiplication operations.
11287 This option instructs the compiler to assume that signed arithmetic
11288 overflow of addition, subtraction and multiplication wraps around
11289 using twos-complement representation. This flag enables some optimizations
11290 and disables other. This option is enabled by default for the Java
11291 front-end, as required by the Java language specification.
11294 @opindex fexceptions
11295 Enable exception handling. Generates extra code needed to propagate
11296 exceptions. For some targets, this implies GCC will generate frame
11297 unwind information for all functions, which can produce significant data
11298 size overhead, although it does not affect execution. If you do not
11299 specify this option, GCC will enable it by default for languages like
11300 C++ which normally require exception handling, and disable it for
11301 languages like C that do not normally require it. However, you may need
11302 to enable this option when compiling C code that needs to interoperate
11303 properly with exception handlers written in C++. You may also wish to
11304 disable this option if you are compiling older C++ programs that don't
11305 use exception handling.
11307 @item -fnon-call-exceptions
11308 @opindex fnon-call-exceptions
11309 Generate code that allows trapping instructions to throw exceptions.
11310 Note that this requires platform-specific runtime support that does
11311 not exist everywhere. Moreover, it only allows @emph{trapping}
11312 instructions to throw exceptions, i.e.@: memory references or floating
11313 point instructions. It does not allow exceptions to be thrown from
11314 arbitrary signal handlers such as @code{SIGALRM}.
11316 @item -funwind-tables
11317 @opindex funwind-tables
11318 Similar to @option{-fexceptions}, except that it will just generate any needed
11319 static data, but will not affect the generated code in any other way.
11320 You will normally not enable this option; instead, a language processor
11321 that needs this handling would enable it on your behalf.
11323 @item -fasynchronous-unwind-tables
11324 @opindex fasynchronous-unwind-tables
11325 Generate unwind table in dwarf2 format, if supported by target machine. The
11326 table is exact at each instruction boundary, so it can be used for stack
11327 unwinding from asynchronous events (such as debugger or garbage collector).
11329 @item -fpcc-struct-return
11330 @opindex fpcc-struct-return
11331 Return ``short'' @code{struct} and @code{union} values in memory like
11332 longer ones, rather than in registers. This convention is less
11333 efficient, but it has the advantage of allowing intercallability between
11334 GCC-compiled files and files compiled with other compilers, particularly
11335 the Portable C Compiler (pcc).
11337 The precise convention for returning structures in memory depends
11338 on the target configuration macros.
11340 Short structures and unions are those whose size and alignment match
11341 that of some integer type.
11343 @strong{Warning:} code compiled with the @option{-fpcc-struct-return}
11344 switch is not binary compatible with code compiled with the
11345 @option{-freg-struct-return} switch.
11346 Use it to conform to a non-default application binary interface.
11348 @item -freg-struct-return
11349 @opindex freg-struct-return
11350 Return @code{struct} and @code{union} values in registers when possible.
11351 This is more efficient for small structures than
11352 @option{-fpcc-struct-return}.
11354 If you specify neither @option{-fpcc-struct-return} nor
11355 @option{-freg-struct-return}, GCC defaults to whichever convention is
11356 standard for the target. If there is no standard convention, GCC
11357 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
11358 the principal compiler. In those cases, we can choose the standard, and
11359 we chose the more efficient register return alternative.
11361 @strong{Warning:} code compiled with the @option{-freg-struct-return}
11362 switch is not binary compatible with code compiled with the
11363 @option{-fpcc-struct-return} switch.
11364 Use it to conform to a non-default application binary interface.
11366 @item -fshort-enums
11367 @opindex fshort-enums
11368 Allocate to an @code{enum} type only as many bytes as it needs for the
11369 declared range of possible values. Specifically, the @code{enum} type
11370 will be equivalent to the smallest integer type which has enough room.
11372 @strong{Warning:} the @option{-fshort-enums} switch causes GCC to generate
11373 code that is not binary compatible with code generated without that switch.
11374 Use it to conform to a non-default application binary interface.
11376 @item -fshort-double
11377 @opindex fshort-double
11378 Use the same size for @code{double} as for @code{float}.
11380 @strong{Warning:} the @option{-fshort-double} switch causes GCC to generate
11381 code that is not binary compatible with code generated without that switch.
11382 Use it to conform to a non-default application binary interface.
11384 @item -fshort-wchar
11385 @opindex fshort-wchar
11386 Override the underlying type for @samp{wchar_t} to be @samp{short
11387 unsigned int} instead of the default for the target. This option is
11388 useful for building programs to run under WINE@.
11390 @strong{Warning:} the @option{-fshort-wchar} switch causes GCC to generate
11391 code that is not binary compatible with code generated without that switch.
11392 Use it to conform to a non-default application binary interface.
11394 @item -fshared-data
11395 @opindex fshared-data
11396 Requests that the data and non-@code{const} variables of this
11397 compilation be shared data rather than private data. The distinction
11398 makes sense only on certain operating systems, where shared data is
11399 shared between processes running the same program, while private data
11400 exists in one copy per process.
11403 @opindex fno-common
11404 In C, allocate even uninitialized global variables in the data section of the
11405 object file, rather than generating them as common blocks. This has the
11406 effect that if the same variable is declared (without @code{extern}) in
11407 two different compilations, you will get an error when you link them.
11408 The only reason this might be useful is if you wish to verify that the
11409 program will work on other systems which always work this way.
11413 Ignore the @samp{#ident} directive.
11415 @item -finhibit-size-directive
11416 @opindex finhibit-size-directive
11417 Don't output a @code{.size} assembler directive, or anything else that
11418 would cause trouble if the function is split in the middle, and the
11419 two halves are placed at locations far apart in memory. This option is
11420 used when compiling @file{crtstuff.c}; you should not need to use it
11423 @item -fverbose-asm
11424 @opindex fverbose-asm
11425 Put extra commentary information in the generated assembly code to
11426 make it more readable. This option is generally only of use to those
11427 who actually need to read the generated assembly code (perhaps while
11428 debugging the compiler itself).
11430 @option{-fno-verbose-asm}, the default, causes the
11431 extra information to be omitted and is useful when comparing two assembler
11436 @cindex global offset table
11438 Generate position-independent code (PIC) suitable for use in a shared
11439 library, if supported for the target machine. Such code accesses all
11440 constant addresses through a global offset table (GOT)@. The dynamic
11441 loader resolves the GOT entries when the program starts (the dynamic
11442 loader is not part of GCC; it is part of the operating system). If
11443 the GOT size for the linked executable exceeds a machine-specific
11444 maximum size, you get an error message from the linker indicating that
11445 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
11446 instead. (These maximums are 8k on the SPARC and 32k
11447 on the m68k and RS/6000. The 386 has no such limit.)
11449 Position-independent code requires special support, and therefore works
11450 only on certain machines. For the 386, GCC supports PIC for System V
11451 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
11452 position-independent.
11456 If supported for the target machine, emit position-independent code,
11457 suitable for dynamic linking and avoiding any limit on the size of the
11458 global offset table. This option makes a difference on the m68k
11461 Position-independent code requires special support, and therefore works
11462 only on certain machines.
11468 These options are similar to @option{-fpic} and @option{-fPIC}, but
11469 generated position independent code can be only linked into executables.
11470 Usually these options are used when @option{-pie} GCC option will be
11471 used during linking.
11473 @item -ffixed-@var{reg}
11475 Treat the register named @var{reg} as a fixed register; generated code
11476 should never refer to it (except perhaps as a stack pointer, frame
11477 pointer or in some other fixed role).
11479 @var{reg} must be the name of a register. The register names accepted
11480 are machine-specific and are defined in the @code{REGISTER_NAMES}
11481 macro in the machine description macro file.
11483 This flag does not have a negative form, because it specifies a
11486 @item -fcall-used-@var{reg}
11487 @opindex fcall-used
11488 Treat the register named @var{reg} as an allocable register that is
11489 clobbered by function calls. It may be allocated for temporaries or
11490 variables that do not live across a call. Functions compiled this way
11491 will not save and restore the register @var{reg}.
11493 It is an error to used this flag with the frame pointer or stack pointer.
11494 Use of this flag for other registers that have fixed pervasive roles in
11495 the machine's execution model will produce disastrous results.
11497 This flag does not have a negative form, because it specifies a
11500 @item -fcall-saved-@var{reg}
11501 @opindex fcall-saved
11502 Treat the register named @var{reg} as an allocable register saved by
11503 functions. It may be allocated even for temporaries or variables that
11504 live across a call. Functions compiled this way will save and restore
11505 the register @var{reg} if they use it.
11507 It is an error to used this flag with the frame pointer or stack pointer.
11508 Use of this flag for other registers that have fixed pervasive roles in
11509 the machine's execution model will produce disastrous results.
11511 A different sort of disaster will result from the use of this flag for
11512 a register in which function values may be returned.
11514 This flag does not have a negative form, because it specifies a
11517 @item -fpack-struct
11518 @opindex fpack-struct
11519 Pack all structure members together without holes.
11521 @strong{Warning:} the @option{-fpack-struct} switch causes GCC to generate
11522 code that is not binary compatible with code generated without that switch.
11523 Additionally, it makes the code suboptimal.
11524 Use it to conform to a non-default application binary interface.
11526 @item -finstrument-functions
11527 @opindex finstrument-functions
11528 Generate instrumentation calls for entry and exit to functions. Just
11529 after function entry and just before function exit, the following
11530 profiling functions will be called with the address of the current
11531 function and its call site. (On some platforms,
11532 @code{__builtin_return_address} does not work beyond the current
11533 function, so the call site information may not be available to the
11534 profiling functions otherwise.)
11537 void __cyg_profile_func_enter (void *this_fn,
11539 void __cyg_profile_func_exit (void *this_fn,
11543 The first argument is the address of the start of the current function,
11544 which may be looked up exactly in the symbol table.
11546 This instrumentation is also done for functions expanded inline in other
11547 functions. The profiling calls will indicate where, conceptually, the
11548 inline function is entered and exited. This means that addressable
11549 versions of such functions must be available. If all your uses of a
11550 function are expanded inline, this may mean an additional expansion of
11551 code size. If you use @samp{extern inline} in your C code, an
11552 addressable version of such functions must be provided. (This is
11553 normally the case anyways, but if you get lucky and the optimizer always
11554 expands the functions inline, you might have gotten away without
11555 providing static copies.)
11557 A function may be given the attribute @code{no_instrument_function}, in
11558 which case this instrumentation will not be done. This can be used, for
11559 example, for the profiling functions listed above, high-priority
11560 interrupt routines, and any functions from which the profiling functions
11561 cannot safely be called (perhaps signal handlers, if the profiling
11562 routines generate output or allocate memory).
11564 @item -fstack-check
11565 @opindex fstack-check
11566 Generate code to verify that you do not go beyond the boundary of the
11567 stack. You should specify this flag if you are running in an
11568 environment with multiple threads, but only rarely need to specify it in
11569 a single-threaded environment since stack overflow is automatically
11570 detected on nearly all systems if there is only one stack.
11572 Note that this switch does not actually cause checking to be done; the
11573 operating system must do that. The switch causes generation of code
11574 to ensure that the operating system sees the stack being extended.
11576 @item -fstack-limit-register=@var{reg}
11577 @itemx -fstack-limit-symbol=@var{sym}
11578 @itemx -fno-stack-limit
11579 @opindex fstack-limit-register
11580 @opindex fstack-limit-symbol
11581 @opindex fno-stack-limit
11582 Generate code to ensure that the stack does not grow beyond a certain value,
11583 either the value of a register or the address of a symbol. If the stack
11584 would grow beyond the value, a signal is raised. For most targets,
11585 the signal is raised before the stack overruns the boundary, so
11586 it is possible to catch the signal without taking special precautions.
11588 For instance, if the stack starts at absolute address @samp{0x80000000}
11589 and grows downwards, you can use the flags
11590 @option{-fstack-limit-symbol=__stack_limit} and
11591 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
11592 of 128KB@. Note that this may only work with the GNU linker.
11594 @cindex aliasing of parameters
11595 @cindex parameters, aliased
11596 @item -fargument-alias
11597 @itemx -fargument-noalias
11598 @itemx -fargument-noalias-global
11599 @opindex fargument-alias
11600 @opindex fargument-noalias
11601 @opindex fargument-noalias-global
11602 Specify the possible relationships among parameters and between
11603 parameters and global data.
11605 @option{-fargument-alias} specifies that arguments (parameters) may
11606 alias each other and may alias global storage.@*
11607 @option{-fargument-noalias} specifies that arguments do not alias
11608 each other, but may alias global storage.@*
11609 @option{-fargument-noalias-global} specifies that arguments do not
11610 alias each other and do not alias global storage.
11612 Each language will automatically use whatever option is required by
11613 the language standard. You should not need to use these options yourself.
11615 @item -fleading-underscore
11616 @opindex fleading-underscore
11617 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
11618 change the way C symbols are represented in the object file. One use
11619 is to help link with legacy assembly code.
11621 @strong{Warning:} the @option{-fleading-underscore} switch causes GCC to
11622 generate code that is not binary compatible with code generated without that
11623 switch. Use it to conform to a non-default application binary interface.
11624 Not all targets provide complete support for this switch.
11626 @item -ftls-model=@var{model}
11627 Alter the thread-local storage model to be used (@pxref{Thread-Local}).
11628 The @var{model} argument should be one of @code{global-dynamic},
11629 @code{local-dynamic}, @code{initial-exec} or @code{local-exec}.
11631 The default without @option{-fpic} is @code{initial-exec}; with
11632 @option{-fpic} the default is @code{global-dynamic}.
11637 @node Environment Variables
11638 @section Environment Variables Affecting GCC
11639 @cindex environment variables
11641 @c man begin ENVIRONMENT
11642 This section describes several environment variables that affect how GCC
11643 operates. Some of them work by specifying directories or prefixes to use
11644 when searching for various kinds of files. Some are used to specify other
11645 aspects of the compilation environment.
11647 Note that you can also specify places to search using options such as
11648 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
11649 take precedence over places specified using environment variables, which
11650 in turn take precedence over those specified by the configuration of GCC@.
11651 @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint,
11652 GNU Compiler Collection (GCC) Internals}.
11657 @c @itemx LC_COLLATE
11659 @c @itemx LC_MONETARY
11660 @c @itemx LC_NUMERIC
11665 @c @findex LC_COLLATE
11666 @findex LC_MESSAGES
11667 @c @findex LC_MONETARY
11668 @c @findex LC_NUMERIC
11672 These environment variables control the way that GCC uses
11673 localization information that allow GCC to work with different
11674 national conventions. GCC inspects the locale categories
11675 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
11676 so. These locale categories can be set to any value supported by your
11677 installation. A typical value is @samp{en_UK} for English in the United
11680 The @env{LC_CTYPE} environment variable specifies character
11681 classification. GCC uses it to determine the character boundaries in
11682 a string; this is needed for some multibyte encodings that contain quote
11683 and escape characters that would otherwise be interpreted as a string
11686 The @env{LC_MESSAGES} environment variable specifies the language to
11687 use in diagnostic messages.
11689 If the @env{LC_ALL} environment variable is set, it overrides the value
11690 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
11691 and @env{LC_MESSAGES} default to the value of the @env{LANG}
11692 environment variable. If none of these variables are set, GCC
11693 defaults to traditional C English behavior.
11697 If @env{TMPDIR} is set, it specifies the directory to use for temporary
11698 files. GCC uses temporary files to hold the output of one stage of
11699 compilation which is to be used as input to the next stage: for example,
11700 the output of the preprocessor, which is the input to the compiler
11703 @item GCC_EXEC_PREFIX
11704 @findex GCC_EXEC_PREFIX
11705 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
11706 names of the subprograms executed by the compiler. No slash is added
11707 when this prefix is combined with the name of a subprogram, but you can
11708 specify a prefix that ends with a slash if you wish.
11710 If @env{GCC_EXEC_PREFIX} is not set, GCC will attempt to figure out
11711 an appropriate prefix to use based on the pathname it was invoked with.
11713 If GCC cannot find the subprogram using the specified prefix, it
11714 tries looking in the usual places for the subprogram.
11716 The default value of @env{GCC_EXEC_PREFIX} is
11717 @file{@var{prefix}/lib/gcc/} where @var{prefix} is the value
11718 of @code{prefix} when you ran the @file{configure} script.
11720 Other prefixes specified with @option{-B} take precedence over this prefix.
11722 This prefix is also used for finding files such as @file{crt0.o} that are
11725 In addition, the prefix is used in an unusual way in finding the
11726 directories to search for header files. For each of the standard
11727 directories whose name normally begins with @samp{/usr/local/lib/gcc}
11728 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
11729 replacing that beginning with the specified prefix to produce an
11730 alternate directory name. Thus, with @option{-Bfoo/}, GCC will search
11731 @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}.
11732 These alternate directories are searched first; the standard directories
11735 @item COMPILER_PATH
11736 @findex COMPILER_PATH
11737 The value of @env{COMPILER_PATH} is a colon-separated list of
11738 directories, much like @env{PATH}. GCC tries the directories thus
11739 specified when searching for subprograms, if it can't find the
11740 subprograms using @env{GCC_EXEC_PREFIX}.
11743 @findex LIBRARY_PATH
11744 The value of @env{LIBRARY_PATH} is a colon-separated list of
11745 directories, much like @env{PATH}. When configured as a native compiler,
11746 GCC tries the directories thus specified when searching for special
11747 linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking
11748 using GCC also uses these directories when searching for ordinary
11749 libraries for the @option{-l} option (but directories specified with
11750 @option{-L} come first).
11754 @cindex locale definition
11755 This variable is used to pass locale information to the compiler. One way in
11756 which this information is used is to determine the character set to be used
11757 when character literals, string literals and comments are parsed in C and C++.
11758 When the compiler is configured to allow multibyte characters,
11759 the following values for @env{LANG} are recognized:
11763 Recognize JIS characters.
11765 Recognize SJIS characters.
11767 Recognize EUCJP characters.
11770 If @env{LANG} is not defined, or if it has some other value, then the
11771 compiler will use mblen and mbtowc as defined by the default locale to
11772 recognize and translate multibyte characters.
11776 Some additional environments variables affect the behavior of the
11779 @include cppenv.texi
11783 @node Precompiled Headers
11784 @section Using Precompiled Headers
11785 @cindex precompiled headers
11786 @cindex speed of compilation
11788 Often large projects have many header files that are included in every
11789 source file. The time the compiler takes to process these header files
11790 over and over again can account for nearly all of the time required to
11791 build the project. To make builds faster, GCC allows users to
11792 `precompile' a header file; then, if builds can use the precompiled
11793 header file they will be much faster.
11795 @strong{Caution:} There are a few known situations where GCC will
11796 crash when trying to use a precompiled header. If you have trouble
11797 with a precompiled header, you should remove the precompiled header
11798 and compile without it. In addition, please use GCC's on-line
11799 defect-tracking system to report any problems you encounter with
11800 precompiled headers. @xref{Bugs}.
11802 To create a precompiled header file, simply compile it as you would any
11803 other file, if necessary using the @option{-x} option to make the driver
11804 treat it as a C or C++ header file. You will probably want to use a
11805 tool like @command{make} to keep the precompiled header up-to-date when
11806 the headers it contains change.
11808 A precompiled header file will be searched for when @code{#include} is
11809 seen in the compilation. As it searches for the included file
11810 (@pxref{Search Path,,Search Path,cpp,The C Preprocessor}) the
11811 compiler looks for a precompiled header in each directory just before it
11812 looks for the include file in that directory. The name searched for is
11813 the name specified in the @code{#include} with @samp{.gch} appended. If
11814 the precompiled header file can't be used, it is ignored.
11816 For instance, if you have @code{#include "all.h"}, and you have
11817 @file{all.h.gch} in the same directory as @file{all.h}, then the
11818 precompiled header file will be used if possible, and the original
11819 header will be used otherwise.
11821 Alternatively, you might decide to put the precompiled header file in a
11822 directory and use @option{-I} to ensure that directory is searched
11823 before (or instead of) the directory containing the original header.
11824 Then, if you want to check that the precompiled header file is always
11825 used, you can put a file of the same name as the original header in this
11826 directory containing an @code{#error} command.
11828 This also works with @option{-include}. So yet another way to use
11829 precompiled headers, good for projects not designed with precompiled
11830 header files in mind, is to simply take most of the header files used by
11831 a project, include them from another header file, precompile that header
11832 file, and @option{-include} the precompiled header. If the header files
11833 have guards against multiple inclusion, they will be skipped because
11834 they've already been included (in the precompiled header).
11836 If you need to precompile the same header file for different
11837 languages, targets, or compiler options, you can instead make a
11838 @emph{directory} named like @file{all.h.gch}, and put each precompiled
11839 header in the directory, perhaps using @option{-o}. It doesn't matter
11840 what you call the files in the directory, every precompiled header in
11841 the directory will be considered. The first precompiled header
11842 encountered in the directory that is valid for this compilation will
11843 be used; they're searched in no particular order.
11845 There are many other possibilities, limited only by your imagination,
11846 good sense, and the constraints of your build system.
11848 A precompiled header file can be used only when these conditions apply:
11852 Only one precompiled header can be used in a particular compilation.
11855 A precompiled header can't be used once the first C token is seen. You
11856 can have preprocessor directives before a precompiled header; you can
11857 even include a precompiled header from inside another header, so long as
11858 there are no C tokens before the @code{#include}.
11861 The precompiled header file must be produced for the same language as
11862 the current compilation. You can't use a C precompiled header for a C++
11866 The precompiled header file must be produced by the same compiler
11867 version and configuration as the current compilation is using.
11868 The easiest way to guarantee this is to use the same compiler binary
11869 for creating and using precompiled headers.
11872 Any macros defined before the precompiled header is included must
11873 either be defined in the same way as when the precompiled header was
11874 generated, or must not affect the precompiled header, which usually
11875 means that the they don't appear in the precompiled header at all.
11877 The @option{-D} option is one way to define a macro before a
11878 precompiled header is included; using a @code{#define} can also do it.
11879 There are also some options that define macros implicitly, like
11880 @option{-O} and @option{-Wdeprecated}; the same rule applies to macros
11883 @item If debugging information is output when using the precompiled
11884 header, using @option{-g} or similar, the same kind of debugging information
11885 must have been output when building the precompiled header. However,
11886 a precompiled header built using @option{-g} can be used in a compilation
11887 when no debugging information is being output.
11889 @item The same @option{-m} options must generally be used when building
11890 and using the precompiled header. @xref{Submodel Options},
11891 for any cases where this rule is relaxed.
11893 @item Each of the following options must be the same when building and using
11894 the precompiled header:
11896 @gccoptlist{-fexceptions -funit-at-a-time}
11899 Some other command-line options starting with @option{-f},
11900 @option{-p}, or @option{-O} must be defined in the same way as when
11901 the precompiled header was generated. At present, it's not clear
11902 which options are safe to change and which are not; the safest choice
11903 is to use exactly the same options when generating and using the
11904 precompiled header. The following are known to be safe:
11906 @gccoptlist{-fpreprocessed -pedantic-errors}
11910 For all of these except the last, the compiler will automatically
11911 ignore the precompiled header if the conditions aren't met. If you
11912 find an option combination that doesn't work and doesn't cause the
11913 precompiled header to be ignored, please consider filing a bug report,
11916 If you do use differing options when generating and using the
11917 precompiled header, the actual behaviour will be a mixture of the
11918 behaviour for the options. For instance, if you use @option{-g} to
11919 generate the precompiled header but not when using it, you may or may
11920 not get debugging information for routines in the precompiled header.
11922 @node Running Protoize
11923 @section Running Protoize
11925 The program @code{protoize} is an optional part of GCC@. You can use
11926 it to add prototypes to a program, thus converting the program to ISO
11927 C in one respect. The companion program @code{unprotoize} does the
11928 reverse: it removes argument types from any prototypes that are found.
11930 When you run these programs, you must specify a set of source files as
11931 command line arguments. The conversion programs start out by compiling
11932 these files to see what functions they define. The information gathered
11933 about a file @var{foo} is saved in a file named @file{@var{foo}.X}.
11935 After scanning comes actual conversion. The specified files are all
11936 eligible to be converted; any files they include (whether sources or
11937 just headers) are eligible as well.
11939 But not all the eligible files are converted. By default,
11940 @code{protoize} and @code{unprotoize} convert only source and header
11941 files in the current directory. You can specify additional directories
11942 whose files should be converted with the @option{-d @var{directory}}
11943 option. You can also specify particular files to exclude with the
11944 @option{-x @var{file}} option. A file is converted if it is eligible, its
11945 directory name matches one of the specified directory names, and its
11946 name within the directory has not been excluded.
11948 Basic conversion with @code{protoize} consists of rewriting most
11949 function definitions and function declarations to specify the types of
11950 the arguments. The only ones not rewritten are those for varargs
11953 @code{protoize} optionally inserts prototype declarations at the
11954 beginning of the source file, to make them available for any calls that
11955 precede the function's definition. Or it can insert prototype
11956 declarations with block scope in the blocks where undeclared functions
11959 Basic conversion with @code{unprotoize} consists of rewriting most
11960 function declarations to remove any argument types, and rewriting
11961 function definitions to the old-style pre-ISO form.
11963 Both conversion programs print a warning for any function declaration or
11964 definition that they can't convert. You can suppress these warnings
11967 The output from @code{protoize} or @code{unprotoize} replaces the
11968 original source file. The original file is renamed to a name ending
11969 with @samp{.save} (for DOS, the saved filename ends in @samp{.sav}
11970 without the original @samp{.c} suffix). If the @samp{.save} (@samp{.sav}
11971 for DOS) file already exists, then the source file is simply discarded.
11973 @code{protoize} and @code{unprotoize} both depend on GCC itself to
11974 scan the program and collect information about the functions it uses.
11975 So neither of these programs will work until GCC is installed.
11977 Here is a table of the options you can use with @code{protoize} and
11978 @code{unprotoize}. Each option works with both programs unless
11982 @item -B @var{directory}
11983 Look for the file @file{SYSCALLS.c.X} in @var{directory}, instead of the
11984 usual directory (normally @file{/usr/local/lib}). This file contains
11985 prototype information about standard system functions. This option
11986 applies only to @code{protoize}.
11988 @item -c @var{compilation-options}
11989 Use @var{compilation-options} as the options when running @command{gcc} to
11990 produce the @samp{.X} files. The special option @option{-aux-info} is
11991 always passed in addition, to tell @command{gcc} to write a @samp{.X} file.
11993 Note that the compilation options must be given as a single argument to
11994 @code{protoize} or @code{unprotoize}. If you want to specify several
11995 @command{gcc} options, you must quote the entire set of compilation options
11996 to make them a single word in the shell.
11998 There are certain @command{gcc} arguments that you cannot use, because they
11999 would produce the wrong kind of output. These include @option{-g},
12000 @option{-O}, @option{-c}, @option{-S}, and @option{-o} If you include these in
12001 the @var{compilation-options}, they are ignored.
12004 Rename files to end in @samp{.C} (@samp{.cc} for DOS-based file
12005 systems) instead of @samp{.c}. This is convenient if you are converting
12006 a C program to C++. This option applies only to @code{protoize}.
12009 Add explicit global declarations. This means inserting explicit
12010 declarations at the beginning of each source file for each function
12011 that is called in the file and was not declared. These declarations
12012 precede the first function definition that contains a call to an
12013 undeclared function. This option applies only to @code{protoize}.
12015 @item -i @var{string}
12016 Indent old-style parameter declarations with the string @var{string}.
12017 This option applies only to @code{protoize}.
12019 @code{unprotoize} converts prototyped function definitions to old-style
12020 function definitions, where the arguments are declared between the
12021 argument list and the initial @samp{@{}. By default, @code{unprotoize}
12022 uses five spaces as the indentation. If you want to indent with just
12023 one space instead, use @option{-i " "}.
12026 Keep the @samp{.X} files. Normally, they are deleted after conversion
12030 Add explicit local declarations. @code{protoize} with @option{-l} inserts
12031 a prototype declaration for each function in each block which calls the
12032 function without any declaration. This option applies only to
12036 Make no real changes. This mode just prints information about the conversions
12037 that would have been done without @option{-n}.
12040 Make no @samp{.save} files. The original files are simply deleted.
12041 Use this option with caution.
12043 @item -p @var{program}
12044 Use the program @var{program} as the compiler. Normally, the name
12045 @file{gcc} is used.
12048 Work quietly. Most warnings are suppressed.
12051 Print the version number, just like @option{-v} for @command{gcc}.
12054 If you need special compiler options to compile one of your program's
12055 source files, then you should generate that file's @samp{.X} file
12056 specially, by running @command{gcc} on that source file with the
12057 appropriate options and the option @option{-aux-info}. Then run
12058 @code{protoize} on the entire set of files. @code{protoize} will use
12059 the existing @samp{.X} file because it is newer than the source file.
12063 gcc -Dfoo=bar file1.c -aux-info file1.X
12068 You need to include the special files along with the rest in the
12069 @code{protoize} command, even though their @samp{.X} files already
12070 exist, because otherwise they won't get converted.
12072 @xref{Protoize Caveats}, for more information on how to use
12073 @code{protoize} successfully.