1 @c Copyright (C) 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
2 @c 2000, 2001, 2002, 2003 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 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 1.1 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), g77(1), as(1), ld(1), gdb(1), adb(1), dbx(1), sdb(1)
47 and the Info entries for @file{gcc}, @file{cpp}, @file{g77}, @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
70 When you invoke GCC, it normally does preprocessing, compilation,
71 assembly and linking. The ``overall options'' allow you to stop this
72 process at an intermediate stage. For example, the @option{-c} option
73 says not to run the linker. Then the output consists of object files
74 output by the assembler.
76 Other options are passed on to one stage of processing. Some options
77 control the preprocessor and others the compiler itself. Yet other
78 options control the assembler and linker; most of these are not
79 documented here, since you rarely need to use any of them.
81 @cindex C compilation options
82 Most of the command line options that you can use with GCC are useful
83 for C programs; when an option is only useful with another language
84 (usually C++), the explanation says so explicitly. If the description
85 for a particular option does not mention a source language, you can use
86 that option with all supported languages.
88 @cindex C++ compilation options
89 @xref{Invoking G++,,Compiling C++ Programs}, for a summary of special
90 options for compiling C++ programs.
92 @cindex grouping options
93 @cindex options, grouping
94 The @command{gcc} program accepts options and file names as operands. Many
95 options have multi-letter names; therefore multiple single-letter options
96 may @emph{not} be grouped: @option{-dr} is very different from @w{@samp{-d
99 @cindex order of options
100 @cindex options, order
101 You can mix options and other arguments. For the most part, the order
102 you use doesn't matter. Order does matter when you use several options
103 of the same kind; for example, if you specify @option{-L} more than once,
104 the directories are searched in the order specified.
106 Many options have long names starting with @samp{-f} or with
107 @samp{-W}---for example, @option{-fforce-mem},
108 @option{-fstrength-reduce}, @option{-Wformat} and so on. Most of
109 these have both positive and negative forms; the negative form of
110 @option{-ffoo} would be @option{-fno-foo}. This manual documents
111 only one of these two forms, whichever one is not the default.
115 @xref{Option Index}, for an index to GCC's options.
118 * Option Summary:: Brief list of all options, without explanations.
119 * Overall Options:: Controlling the kind of output:
120 an executable, object files, assembler files,
121 or preprocessed source.
122 * Invoking G++:: Compiling C++ programs.
123 * C Dialect Options:: Controlling the variant of C language compiled.
124 * C++ Dialect Options:: Variations on C++.
125 * Objective-C Dialect Options:: Variations on Objective-C.
126 * Language Independent Options:: Controlling how diagnostics should be
128 * Warning Options:: How picky should the compiler be?
129 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
130 * Optimize Options:: How much optimization?
131 * Preprocessor Options:: Controlling header files and macro definitions.
132 Also, getting dependency information for Make.
133 * Assembler Options:: Passing options to the assembler.
134 * Link Options:: Specifying libraries and so on.
135 * Directory Options:: Where to find header files and libraries.
136 Where to find the compiler executable files.
137 * Spec Files:: How to pass switches to sub-processes.
138 * Target Options:: Running a cross-compiler, or an old version of GCC.
139 * Submodel Options:: Specifying minor hardware or convention variations,
140 such as 68010 vs 68020.
141 * Code Gen Options:: Specifying conventions for function calls, data layout
143 * Environment Variables:: Env vars that affect GCC.
144 * Precompiled Headers:: Compiling a header once, and using it many times.
145 * Running Protoize:: Automatically adding or removing function prototypes.
151 @section Option Summary
153 Here is a summary of all the options, grouped by type. Explanations are
154 in the following sections.
157 @item Overall Options
158 @xref{Overall Options,,Options Controlling the Kind of Output}.
160 -c -S -E -o @var{file} -pipe -pass-exit-codes -x @var{language} @gol
161 -v -### --help --target-help --version}
163 @item C Language Options
164 @xref{C Dialect Options,,Options Controlling C Dialect}.
166 -ansi -std=@var{standard} -aux-info @var{filename} @gol
167 -fno-asm -fno-builtin -fno-builtin-@var{function} @gol
168 -fhosted -ffreestanding -fms-extensions @gol
169 -trigraphs -traditional -traditional-cpp @gol
170 -fallow-single-precision -fcond-mismatch @gol
171 -fsigned-bitfields -fsigned-char @gol
172 -funsigned-bitfields -funsigned-char @gol
175 @item C++ Language Options
176 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}.
178 -fabi-version=@var{n} -fno-access-control -fcheck-new @gol
179 -fconserve-space -fno-const-strings -fdollars-in-identifiers @gol
180 -fno-elide-constructors @gol
181 -fno-enforce-eh-specs -fexternal-templates @gol
182 -falt-external-templates @gol
183 -ffor-scope -fno-for-scope -fno-gnu-keywords @gol
184 -fno-implicit-templates @gol
185 -fno-implicit-inline-templates @gol
186 -fno-implement-inlines -fms-extensions @gol
187 -fno-nonansi-builtins -fno-operator-names @gol
188 -fno-optional-diags -fpermissive @gol
189 -frepo -fno-rtti -fstats -ftemplate-depth-@var{n} @gol
190 -fuse-cxa-atexit -fvtable-gc -fno-weak -nostdinc++ @gol
191 -fno-default-inline -Wabi -Wctor-dtor-privacy @gol
192 -Wnon-virtual-dtor -Wreorder @gol
193 -Weffc++ -Wno-deprecated @gol
194 -Wno-non-template-friend -Wold-style-cast @gol
195 -Woverloaded-virtual -Wno-pmf-conversions @gol
196 -Wsign-promo -Wsynth}
198 @item Objective-C Language Options
199 @xref{Objective-C Dialect Options,,Options Controlling Objective-C Dialect}.
201 -fconstant-string-class=@var{class-name} @gol
202 -fgnu-runtime -fnext-runtime -gen-decls @gol
203 -Wno-protocol -Wselector -Wundeclared-selector}
205 @item Language Independent Options
206 @xref{Language Independent Options,,Options to Control Diagnostic Messages Formatting}.
208 -fmessage-length=@var{n} @gol
209 -fdiagnostics-show-location=@r{[}once@r{|}every-line@r{]}}
211 @item Warning Options
212 @xref{Warning Options,,Options to Request or Suppress Warnings}.
214 -fsyntax-only -pedantic -pedantic-errors @gol
215 -w -W -Wall -Waggregate-return @gol
216 -Wcast-align -Wcast-qual -Wchar-subscripts -Wcomment @gol
217 -Wconversion -Wno-deprecated-declarations @gol
218 -Wdisabled-optimization -Wno-div-by-zero -Werror @gol
219 -Wfloat-equal -Wformat -Wformat=2 @gol
220 -Wformat-nonliteral -Wformat-security @gol
221 -Wimplicit -Wimplicit-int @gol
222 -Wimplicit-function-declaration @gol
223 -Werror-implicit-function-declaration @gol
224 -Wimport -Winline -Winvalid-pch -Wno-endif-labels @gol
225 -Wlarger-than-@var{len} -Wlong-long @gol
226 -Wmain -Wmissing-braces @gol
227 -Wmissing-format-attribute -Wmissing-noreturn @gol
228 -Wno-multichar -Wno-format-extra-args -Wno-format-y2k @gol
229 -Wno-import -Wnonnull -Wpacked -Wpadded @gol
230 -Wparentheses -Wpointer-arith -Wredundant-decls @gol
231 -Wreturn-type -Wsequence-point -Wshadow @gol
232 -Wsign-compare -Wstrict-aliasing @gol
233 -Wswitch -Wswitch-default -Wswitch-enum @gol
234 -Wsystem-headers -Wtrigraphs -Wundef -Wuninitialized @gol
235 -Wunknown-pragmas -Wunreachable-code @gol
236 -Wunused -Wunused-function -Wunused-label -Wunused-parameter @gol
237 -Wunused-value -Wunused-variable -Wwrite-strings}
239 @item C-only Warning Options
241 -Wbad-function-cast -Wmissing-declarations @gol
242 -Wmissing-prototypes -Wnested-externs @gol
243 -Wstrict-prototypes -Wtraditional}
245 @item Debugging Options
246 @xref{Debugging Options,,Options for Debugging Your Program or GCC}.
248 -d@var{letters} -dumpspecs -dumpmachine -dumpversion @gol
249 -fdump-unnumbered -fdump-translation-unit@r{[}-@var{n}@r{]} @gol
250 -fdump-class-hierarchy@r{[}-@var{n}@r{]} @gol
251 -fdump-tree-original@r{[}-@var{n}@r{]} -fdump-tree-optimized@r{[}-@var{n}@r{]} @gol
252 -fdump-tree-inlined@r{[}-@var{n}@r{]} @gol
253 -feliminate-dwarf2-dups -fmem-report @gol
254 -fprofile-arcs -fsched-verbose=@var{n} @gol
255 -ftest-coverage -ftime-report @gol
256 -g -g@var{level} -gcoff -gdwarf -gdwarf-1 -gdwarf-1+ -gdwarf-2 @gol
257 -ggdb -gstabs -gstabs+ -gvms -gxcoff -gxcoff+ @gol
258 -p -pg -print-file-name=@var{library} -print-libgcc-file-name @gol
259 -print-multi-directory -print-multi-lib @gol
260 -print-prog-name=@var{program} -print-search-dirs -Q @gol
263 @item Optimization Options
264 @xref{Optimize Options,,Options that Control Optimization}.
266 -falign-functions=@var{n} -falign-jumps=@var{n} @gol
267 -falign-labels=@var{n} -falign-loops=@var{n} @gol
269 -fbranch-probabilities -fcaller-saves -fcprop-registers @gol
270 -fcse-follow-jumps -fcse-skip-blocks -fdata-sections @gol
271 -fdelayed-branch -fdelete-null-pointer-checks @gol
272 -fexpensive-optimizations -ffast-math -ffloat-store @gol
273 -fforce-addr -fforce-mem -ffunction-sections @gol
274 -fgcse -fgcse-lm -fgcse-sm -floop-optimize -fcrossjumping @gol
275 -fif-conversion -fif-conversion2 @gol
276 -finline-functions -finline-limit=@var{n} -fkeep-inline-functions @gol
277 -fkeep-static-consts -fmerge-constants -fmerge-all-constants @gol
278 -fmove-all-movables -fnew-ra -fno-branch-count-reg @gol
279 -fno-default-inline -fno-defer-pop @gol
280 -fno-function-cse -fno-guess-branch-probability @gol
281 -fno-inline -fno-math-errno -fno-peephole -fno-peephole2 @gol
282 -funsafe-math-optimizations -ffinite-math-only @gol
283 -fno-trapping-math -fno-zero-initialized-in-bss @gol
284 -fomit-frame-pointer -foptimize-register-move @gol
285 -foptimize-sibling-calls -fprefetch-loop-arrays @gol
286 -freduce-all-givs -fregmove -frename-registers @gol
287 -freorder-blocks -freorder-functions @gol
288 -frerun-cse-after-loop -frerun-loop-opt @gol
289 -fschedule-insns -fschedule-insns2 @gol
290 -fno-sched-interblock -fno-sched-spec -fsched-spec-load @gol
291 -fsched-spec-load-dangerous -fsignaling-nans @gol
292 -fsingle-precision-constant -fssa -fssa-ccp -fssa-dce @gol
293 -fstrength-reduce -fstrict-aliasing -ftracer -fthread-jumps @gol
294 -ftrapv -funroll-all-loops -funroll-loops @gol
295 --param @var{name}=@var{value}
296 -O -O0 -O1 -O2 -O3 -Os}
298 @item Preprocessor Options
299 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
301 -$ -A@var{question}=@var{answer} -A-@var{question}@r{[}=@var{answer}@r{]} @gol
302 -C -dD -dI -dM -dN @gol
303 -D@var{macro}@r{[}=@var{defn}@r{]} -E -H @gol
304 -idirafter @var{dir} @gol
305 -include @var{file} -imacros @var{file} @gol
306 -iprefix @var{file} -iwithprefix @var{dir} @gol
307 -iwithprefixbefore @var{dir} -isystem @var{dir} @gol
308 -M -MM -MF -MG -MP -MQ -MT -nostdinc -P -remap @gol
309 -trigraphs -undef -U@var{macro} -Wp,@var{option} -Xpreprocessor @var{option}}
311 @item Assembler Option
312 @xref{Assembler Options,,Passing Options to the Assembler}.
314 -Wa,@var{option} -Xassembler @var{option}}
317 @xref{Link Options,,Options for Linking}.
319 @var{object-file-name} -l@var{library} @gol
320 -nostartfiles -nodefaultlibs -nostdlib @gol
321 -s -static -static-libgcc -shared -shared-libgcc -symbolic @gol
322 -Wl,@var{option} -Xlinker @var{option} @gol
325 @item Directory Options
326 @xref{Directory Options,,Options for Directory Search}.
328 -B@var{prefix} -I@var{dir} -I- -L@var{dir} -specs=@var{file}}
331 @c I wrote this xref this way to avoid overfull hbox. -- rms
332 @xref{Target Options}.
334 -V @var{version} -b @var{machine}}
336 @item Machine Dependent Options
337 @xref{Submodel Options,,Hardware Models and Configurations}.
339 @emph{M680x0 Options}
341 -m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
342 -m68060 -mcpu32 -m5200 -m68881 -mbitfield -mc68000 -mc68020 @gol
343 -mfpa -mnobitfield -mrtd -mshort -msoft-float -mpcrel @gol
344 -malign-int -mstrict-align}
346 @emph{M68hc1x Options}
348 -m6811 -m6812 -m68hc11 -m68hc12 -m68hcs12 @gol
349 -mauto-incdec -minmax -mlong-calls -mshort @gol
350 -msoft-reg-count=@var{count}}
358 -mcpu=@var{cpu-type} @gol
359 -mtune=@var{cpu-type} @gol
360 -mcmodel=@var{code-model} @gol
362 -mapp-regs -mbroken-saverestore -mcypress @gol
363 -mfaster-structs -mflat @gol
364 -mfpu -mhard-float -mhard-quad-float @gol
365 -mimpure-text -mlive-g0 -mno-app-regs @gol
366 -mno-faster-structs -mno-flat -mno-fpu @gol
367 -mno-impure-text -mno-stack-bias -mno-unaligned-doubles @gol
368 -msoft-float -msoft-quad-float -msparclite -mstack-bias @gol
369 -msupersparc -munaligned-doubles -mv8}
373 -mapcs-frame -mno-apcs-frame @gol
374 -mapcs-26 -mapcs-32 @gol
375 -mapcs-stack-check -mno-apcs-stack-check @gol
376 -mapcs-float -mno-apcs-float @gol
377 -mapcs-reentrant -mno-apcs-reentrant @gol
378 -msched-prolog -mno-sched-prolog @gol
379 -mlittle-endian -mbig-endian -mwords-little-endian @gol
380 -malignment-traps -mno-alignment-traps @gol
381 -msoft-float -mhard-float -mfpe @gol
382 -mthumb-interwork -mno-thumb-interwork @gol
383 -mcpu=@var{name} -march=@var{name} -mfpe=@var{name} @gol
384 -mstructure-size-boundary=@var{n} @gol
385 -mabort-on-noreturn @gol
386 -mlong-calls -mno-long-calls @gol
387 -msingle-pic-base -mno-single-pic-base @gol
388 -mpic-register=@var{reg} @gol
389 -mnop-fun-dllimport @gol
390 -mpoke-function-name @gol
392 -mtpcs-frame -mtpcs-leaf-frame @gol
393 -mcaller-super-interworking -mcallee-super-interworking }
395 @emph{MN10200 Options}
399 @emph{MN10300 Options}
401 -mmult-bug -mno-mult-bug @gol
402 -mam33 -mno-am33 @gol
405 @emph{M32R/D Options}
407 -m32rx -m32r -mcode-model=@var{model-type} -msdata=@var{sdata-type} @gol
412 -m88000 -m88100 -m88110 -mbig-pic @gol
413 -mcheck-zero-division -mhandle-large-shift @gol
414 -midentify-revision -mno-check-zero-division @gol
415 -mno-ocs-debug-info -mno-ocs-frame-position @gol
416 -mno-optimize-arg-area -mno-serialize-volatile @gol
417 -mno-underscores -mocs-debug-info @gol
418 -mocs-frame-position -moptimize-arg-area @gol
419 -mserialize-volatile -mshort-data-@var{num} -msvr3 @gol
420 -msvr4 -mtrap-large-shift -muse-div-instruction @gol
421 -mversion-03.00 -mwarn-passed-structs}
423 @emph{RS/6000 and PowerPC Options}
425 -mcpu=@var{cpu-type} @gol
426 -mtune=@var{cpu-type} @gol
427 -mpower -mno-power -mpower2 -mno-power2 @gol
428 -mpowerpc -mpowerpc64 -mno-powerpc @gol
429 -maltivec -mno-altivec @gol
430 -mpowerpc-gpopt -mno-powerpc-gpopt @gol
431 -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol
432 -mnew-mnemonics -mold-mnemonics @gol
433 -mfull-toc -mminimal-toc -mno-fp-in-toc -mno-sum-in-toc @gol
434 -m64 -m32 -mxl-call -mno-xl-call -mpe @gol
435 -msoft-float -mhard-float -mmultiple -mno-multiple @gol
436 -mstring -mno-string -mupdate -mno-update @gol
437 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol
438 -mstrict-align -mno-strict-align -mrelocatable @gol
439 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol
440 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol
441 -mcall-aix -mcall-sysv -mcall-netbsd @gol
442 -maix-struct-return -msvr4-struct-return @gol
443 -mabi=altivec -mabi=no-altivec @gol
444 -mabi=spe -mabi=no-spe @gol
445 -misel=yes -misel=no @gol
446 -mprototype -mno-prototype @gol
447 -msim -mmvme -mads -myellowknife -memb -msdata @gol
448 -msdata=@var{opt} -mvxworks -mwindiss -G @var{num} -pthread}
450 @emph{Darwin Options}
452 -all_load -allowable_client -arch -arch_errors_fatal @gol
453 -arch_only -bind_at_load -bundle -bundle_loader @gol
454 -client_name -compatibility_version -current_version @gol
455 -dependency-file -dylib_file -dylinker_install_name @gol
456 -dynamic -dynamiclib -exported_symbols_list @gol
457 -filelist -flat_namespace -force_cpusubtype_ALL @gol
458 -force_flat_namespace -headerpad_max_install_names @gol
459 -image_base -init -install_name -keep_private_externs @gol
460 -multi_module -multiply_defined -multiply_defined_unused @gol
461 -noall_load -nomultidefs -noprebind -noseglinkedit @gol
462 -pagezero_size -prebind -prebind_all_twolevel_modules @gol
463 -private_bundle -read_only_relocs -sectalign @gol
464 -sectobjectsymbols -whyload -seg1addr @gol
465 -sectcreate -sectobjectsymbols -sectorder @gol
466 -seg_addr_table -seg_addr_table_filename -seglinkedit @gol
467 -segprot -segs_read_only_addr -segs_read_write_addr @gol
468 -single_module -static -sub_library -sub_umbrella @gol
469 -twolevel_namespace -umbrella -undefined @gol
470 -unexported_symbols_list -weak_reference_mismatches -whatsloaded}
474 -mcall-lib-mul -mfp-arg-in-fpregs -mfp-arg-in-gregs @gol
475 -mfull-fp-blocks -mhc-struct-return -min-line-mul @gol
476 -mminimum-fp-blocks -mnohc-struct-return}
480 -mabicalls -march=@var{cpu-type} -mtune=@var{cpu=type} @gol
481 -mcpu=@var{cpu-type} -membedded-data -muninit-const-in-rodata @gol
482 -membedded-pic -mfp32 -mfp64 -mfused-madd -mno-fused-madd @gol
483 -mgas -mgp32 -mgp64 @gol
484 -mgpopt -mhalf-pic -mhard-float -mint64 -mips1 @gol
485 -mips2 -mips3 -mips4 -mips32 -mips32r2 -mips64 @gol
486 -mlong64 -mlong32 -mlong-calls -mmemcpy @gol
487 -mmips-as -mmips-tfile -mno-abicalls @gol
488 -mno-embedded-data -mno-uninit-const-in-rodata @gol
489 -mno-embedded-pic -mno-gpopt -mno-long-calls @gol
490 -mno-memcpy -mno-mips-tfile -mno-rnames -mno-stats @gol
491 -mrnames -msoft-float @gol
492 -m4650 -msingle-float -mmad @gol
493 -mstats -EL -EB -G @var{num} -nocpp @gol
494 -mabi=32 -mabi=n32 -mabi=64 -mabi=eabi @gol
495 -mfix7000 -mno-crt0 -mflush-func=@var{func} -mno-flush-func @gol
496 -mbranch-likely -mno-branch-likely}
498 @emph{i386 and x86-64 Options}
500 -mcpu=@var{cpu-type} -march=@var{cpu-type} -mfpmath=@var{unit} @gol
501 -masm=@var{dialect} -mno-fancy-math-387 @gol
502 -mno-fp-ret-in-387 -msoft-float -msvr3-shlib @gol
503 -mno-wide-multiply -mrtd -malign-double @gol
504 -mpreferred-stack-boundary=@var{num} @gol
505 -mmmx -msse -msse2 -m3dnow @gol
506 -mthreads -mno-align-stringops -minline-all-stringops @gol
507 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
508 -m96bit-long-double -mregparm=@var{num} -momit-leaf-frame-pointer @gol
510 -mcmodel=@var{code-model} @gol
515 -march=@var{architecture-type} @gol
516 -mbig-switch -mdisable-fpregs -mdisable-indexing @gol
517 -mfast-indirect-calls -mgas -mgnu-ld -mhp-ld @gol
518 -mjump-in-delay -mlinker-opt -mlong-calls @gol
519 -mlong-load-store -mno-big-switch -mno-disable-fpregs @gol
520 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
521 -mno-jump-in-delay -mno-long-load-store @gol
522 -mno-portable-runtime -mno-soft-float @gol
523 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
524 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
525 -mschedule=@var{cpu-type} -mspace-regs -msio -mwsio}
527 @emph{Intel 960 Options}
529 -m@var{cpu-type} -masm-compat -mclean-linkage @gol
530 -mcode-align -mcomplex-addr -mleaf-procedures @gol
531 -mic-compat -mic2.0-compat -mic3.0-compat @gol
532 -mintel-asm -mno-clean-linkage -mno-code-align @gol
533 -mno-complex-addr -mno-leaf-procedures @gol
534 -mno-old-align -mno-strict-align -mno-tail-call @gol
535 -mnumerics -mold-align -msoft-float -mstrict-align @gol
538 @emph{DEC Alpha Options}
540 -mno-fp-regs -msoft-float -malpha-as -mgas @gol
541 -mieee -mieee-with-inexact -mieee-conformant @gol
542 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol
543 -mtrap-precision=@var{mode} -mbuild-constants @gol
544 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} @gol
545 -mbwx -mmax -mfix -mcix @gol
546 -mfloat-vax -mfloat-ieee @gol
547 -mexplicit-relocs -msmall-data -mlarge-data -msmall-text -mlarge-text @gol
548 -mmemory-latency=@var{time}}
550 @emph{DEC Alpha/VMS Options}
554 @emph{H8/300 Options}
556 -mrelax -mh -ms -mn -mint32 -malign-300}
560 -m1 -m2 -m3 -m3e @gol
561 -m4-nofpu -m4-single-only -m4-single -m4 @gol
562 -m5-64media -m5-64media-nofpu @gol
563 -m5-32media -m5-32media-nofpu @gol
564 -m5-compact -m5-compact-nofpu @gol
565 -mb -ml -mdalign -mrelax @gol
566 -mbigtable -mfmovd -mhitachi -mnomacsave @gol
567 -mieee -misize -mpadstruct -mspace @gol
568 -mprefergot -musermode}
570 @emph{System V Options}
572 -Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
577 -mmangle-cpu -mcpu=@var{cpu} -mtext=@var{text-section} @gol
578 -mdata=@var{data-section} -mrodata=@var{readonly-data-section}}
580 @emph{TMS320C3x/C4x Options}
582 -mcpu=@var{cpu} -mbig -msmall -mregparm -mmemparm @gol
583 -mfast-fix -mmpyi -mbk -mti -mdp-isr-reload @gol
584 -mrpts=@var{count} -mrptb -mdb -mloop-unsigned @gol
585 -mparallel-insns -mparallel-mpy -mpreserve-float}
589 -mlong-calls -mno-long-calls -mep -mno-ep @gol
590 -mprolog-function -mno-prolog-function -mspace @gol
591 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
592 -mapp-regs -mno-app-regs @gol
593 -mdisable-callt -mno-disable-callt @gol
599 -m32032 -m32332 -m32532 -m32081 -m32381 @gol
600 -mmult-add -mnomult-add -msoft-float -mrtd -mnortd @gol
601 -mregparam -mnoregparam -msb -mnosb @gol
602 -mbitfield -mnobitfield -mhimem -mnohimem}
606 -mmcu=@var{mcu} -msize -minit-stack=@var{n} -mno-interrupts @gol
607 -mcall-prologues -mno-tablejump -mtiny-stack}
611 -mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
612 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
613 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
614 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
615 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
619 -mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
620 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
621 -melf -mbranch-predict -mno-branch-predict -mbase-addresses @gol
622 -mno-base-addresses -msingle-exit -mno-single-exit}
626 -mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
627 -mvolatile-asm-stop -mb-step -mregister-names -mno-sdata @gol
628 -mconstant-gp -mauto-pic -minline-float-divide-min-latency @gol
629 -minline-float-divide-max-throughput -minline-int-divide-min-latency @gol
630 -minline-int-divide-max-throughput -mno-dwarf2-asm @gol
631 -mfixed-range=@var{register-range}}
635 -mextmem -mextmemory -monchip -mno-asm-optimize @gol
636 -masm-optimize -mbranch-cost=@var{n} -mcond-exec=@var{n}}
638 @emph{S/390 and zSeries Options}
640 -mtune=@var{cpu-type} -march=@var{cpu-type} @gol
641 -mhard-float -msoft-float -mbackchain -mno-backchain @gol
642 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
643 -m64 -m31 -mdebug -mno-debug -mesa -mzarch}
647 -mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol
648 -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol
649 -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol
650 -mstack-align -mdata-align -mconst-align @gol
651 -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol
652 -melf -maout -melinux -mlinux -sim -sim2}
654 @emph{PDP-11 Options}
656 -mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol
657 -mbcopy -mbcopy-builtin -mint32 -mno-int16 @gol
658 -mint16 -mno-int32 -mfloat32 -mno-float64 @gol
659 -mfloat64 -mno-float32 -mabshi -mno-abshi @gol
660 -mbranch-expensive -mbranch-cheap @gol
661 -msplit -mno-split -munix-asm -mdec-asm}
663 @emph{Xstormy16 Options}
667 @emph{Xtensa Options}
669 -mbig-endian -mlittle-endian @gol
670 -mdensity -mno-density @gol
671 -mmac16 -mno-mac16 @gol
672 -mmul16 -mno-mul16 @gol
673 -mmul32 -mno-mul32 @gol
675 -mminmax -mno-minmax @gol
676 -msext -mno-sext @gol
677 -mbooleans -mno-booleans @gol
678 -mhard-float -msoft-float @gol
679 -mfused-madd -mno-fused-madd @gol
680 -mserialize-volatile -mno-serialize-volatile @gol
681 -mtext-section-literals -mno-text-section-literals @gol
682 -mtarget-align -mno-target-align @gol
683 -mlongcalls -mno-longcalls}
687 -mgpr-32 -mgpr-64 -mfpr-32 -mfpr-64 -mhard-float -msoft-float @gol
688 -malloc-cc -mfixed-cc -mdword -mno-dword -mdouble -mno-double @gol
689 -mmedia -mno-media -mmuladd -mno-muladd -mlibrary-pic -macc-4 @gol
690 -macc-8 -mpack -mno-pack -mno-eflags -mcond-move -mno-cond-move @gol
691 -mscc -mno-scc -mcond-exec -mno-cond-exec -mvliw-branch -mno-vliw-branch @gol
692 -mmulti-cond-exec -mno-multi-cond-exec -mnested-cond-exec @gol
693 -mno-nested-cond-exec -mtomcat-stats @gol
698 @item Code Generation Options
699 @xref{Code Gen Options,,Options for Code Generation Conventions}.
701 -fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
702 -ffixed-@var{reg} -fexceptions @gol
703 -fnon-call-exceptions -funwind-tables @gol
704 -fasynchronous-unwind-tables @gol
705 -finhibit-size-directive -finstrument-functions @gol
706 -fno-common -fno-ident -fno-gnu-linker @gol
707 -fpcc-struct-return -fpic -fPIC @gol
708 -freg-struct-return -fshared-data -fshort-enums @gol
709 -fshort-double -fshort-wchar -fvolatile @gol
710 -fvolatile-global -fvolatile-static @gol
711 -fverbose-asm -fpack-struct -fstack-check @gol
712 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
713 -fargument-alias -fargument-noalias @gol
714 -fargument-noalias-global -fleading-underscore @gol
715 -ftls-model=@var{model}}
719 * Overall Options:: Controlling the kind of output:
720 an executable, object files, assembler files,
721 or preprocessed source.
722 * C Dialect Options:: Controlling the variant of C language compiled.
723 * C++ Dialect Options:: Variations on C++.
724 * Objective-C Dialect Options:: Variations on Objective-C.
725 * Language Independent Options:: Controlling how diagnostics should be
727 * Warning Options:: How picky should the compiler be?
728 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
729 * Optimize Options:: How much optimization?
730 * Preprocessor Options:: Controlling header files and macro definitions.
731 Also, getting dependency information for Make.
732 * Assembler Options:: Passing options to the assembler.
733 * Link Options:: Specifying libraries and so on.
734 * Directory Options:: Where to find header files and libraries.
735 Where to find the compiler executable files.
736 * Spec Files:: How to pass switches to sub-processes.
737 * Target Options:: Running a cross-compiler, or an old version of GCC.
740 @node Overall Options
741 @section Options Controlling the Kind of Output
743 Compilation can involve up to four stages: preprocessing, compilation
744 proper, assembly and linking, always in that order. The first three
745 stages apply to an individual source file, and end by producing an
746 object file; linking combines all the object files (those newly
747 compiled, and those specified as input) into an executable file.
749 @cindex file name suffix
750 For any given input file, the file name suffix determines what kind of
755 C source code which must be preprocessed.
758 C source code which should not be preprocessed.
761 C++ source code which should not be preprocessed.
764 Objective-C source code. Note that you must link with the library
765 @file{libobjc.a} to make an Objective-C program work.
768 Objective-C source code which should not be preprocessed.
771 C or C++ header file to be turned into a precompiled header.
775 @itemx @var{file}.cxx
776 @itemx @var{file}.cpp
777 @itemx @var{file}.CPP
778 @itemx @var{file}.c++
780 C++ source code which must be preprocessed. Note that in @samp{.cxx},
781 the last two letters must both be literally @samp{x}. Likewise,
782 @samp{.C} refers to a literal capital C@.
786 C++ header file to be turned into a precompiled header.
789 @itemx @var{file}.for
790 @itemx @var{file}.FOR
791 Fortran source code which should not be preprocessed.
794 @itemx @var{file}.fpp
795 @itemx @var{file}.FPP
796 Fortran source code which must be preprocessed (with the traditional
800 Fortran source code which must be preprocessed with a RATFOR
801 preprocessor (not included with GCC)@.
803 @xref{Overall Options,,Options Controlling the Kind of Output, g77,
804 Using and Porting GNU Fortran}, for more details of the handling of
807 @c FIXME: Descriptions of Java file types.
814 Ada source code file which contains a library unit declaration (a
815 declaration of a package, subprogram, or generic, or a generic
816 instantiation), or a library unit renaming declaration (a package,
817 generic, or subprogram renaming declaration). Such files are also
820 @itemx @var{file}.adb
821 Ada source code file containing a library unit body (a subprogram or
822 package body). Such files are also called @dfn{bodies}.
824 @c GCC also knows about some suffixes for languages not yet included:
833 Assembler code which must be preprocessed.
836 An object file to be fed straight into linking.
837 Any file name with no recognized suffix is treated this way.
841 You can specify the input language explicitly with the @option{-x} option:
844 @item -x @var{language}
845 Specify explicitly the @var{language} for the following input files
846 (rather than letting the compiler choose a default based on the file
847 name suffix). This option applies to all following input files until
848 the next @option{-x} option. Possible values for @var{language} are:
850 c c-header cpp-output
851 c++ c++-header c++-cpp-output
852 objective-c objc-cpp-output
853 assembler assembler-with-cpp
855 f77 f77-cpp-input ratfor
861 Turn off any specification of a language, so that subsequent files are
862 handled according to their file name suffixes (as they are if @option{-x}
863 has not been used at all).
865 @item -pass-exit-codes
866 @opindex pass-exit-codes
867 Normally the @command{gcc} program will exit with the code of 1 if any
868 phase of the compiler returns a non-success return code. If you specify
869 @option{-pass-exit-codes}, the @command{gcc} program will instead return with
870 numerically highest error produced by any phase that returned an error
874 If you only want some of the stages of compilation, you can use
875 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
876 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
877 @command{gcc} is to stop. Note that some combinations (for example,
878 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
883 Compile or assemble the source files, but do not link. The linking
884 stage simply is not done. The ultimate output is in the form of an
885 object file for each source file.
887 By default, the object file name for a source file is made by replacing
888 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
890 Unrecognized input files, not requiring compilation or assembly, are
895 Stop after the stage of compilation proper; do not assemble. The output
896 is in the form of an assembler code file for each non-assembler input
899 By default, the assembler file name for a source file is made by
900 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
902 Input files that don't require compilation are ignored.
906 Stop after the preprocessing stage; do not run the compiler proper. The
907 output is in the form of preprocessed source code, which is sent to the
910 Input files which don't require preprocessing are ignored.
912 @cindex output file option
915 Place output in file @var{file}. This applies regardless to whatever
916 sort of output is being produced, whether it be an executable file,
917 an object file, an assembler file or preprocessed C code.
919 Since only one output file can be specified, it does not make sense to
920 use @option{-o} when compiling more than one input file, unless you are
921 producing an executable file as output.
923 If @option{-o} is not specified, the default is to put an executable file
924 in @file{a.out}, the object file for @file{@var{source}.@var{suffix}} in
925 @file{@var{source}.o}, its assembler file in @file{@var{source}.s}, and
926 all preprocessed C source on standard output.
930 Print (on standard error output) the commands executed to run the stages
931 of compilation. Also print the version number of the compiler driver
932 program and of the preprocessor and the compiler proper.
936 Like @option{-v} except the commands are not executed and all command
937 arguments are quoted. This is useful for shell scripts to capture the
938 driver-generated command lines.
942 Use pipes rather than temporary files for communication between the
943 various stages of compilation. This fails to work on some systems where
944 the assembler is unable to read from a pipe; but the GNU assembler has
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{-W} 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 @cindex traditional C language
1199 @cindex C language, traditional
1201 @itemx -traditional-cpp
1202 @opindex traditional-cpp
1203 @opindex traditional
1204 Formerly, these options caused GCC to attempt to emulate a pre-standard
1205 C compiler. They are now only supported with the @option{-E} switch.
1206 The preprocessor continues to support a pre-standard mode. See the GNU
1207 CPP manual for details.
1209 @item -fcond-mismatch
1210 @opindex fcond-mismatch
1211 Allow conditional expressions with mismatched types in the second and
1212 third arguments. The value of such an expression is void. This option
1213 is not supported for C++.
1215 @item -funsigned-char
1216 @opindex funsigned-char
1217 Let the type @code{char} be unsigned, like @code{unsigned char}.
1219 Each kind of machine has a default for what @code{char} should
1220 be. It is either like @code{unsigned char} by default or like
1221 @code{signed char} by default.
1223 Ideally, a portable program should always use @code{signed char} or
1224 @code{unsigned char} when it depends on the signedness of an object.
1225 But many programs have been written to use plain @code{char} and
1226 expect it to be signed, or expect it to be unsigned, depending on the
1227 machines they were written for. This option, and its inverse, let you
1228 make such a program work with the opposite default.
1230 The type @code{char} is always a distinct type from each of
1231 @code{signed char} or @code{unsigned char}, even though its behavior
1232 is always just like one of those two.
1235 @opindex fsigned-char
1236 Let the type @code{char} be signed, like @code{signed char}.
1238 Note that this is equivalent to @option{-fno-unsigned-char}, which is
1239 the negative form of @option{-funsigned-char}. Likewise, the option
1240 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
1242 @item -fsigned-bitfields
1243 @itemx -funsigned-bitfields
1244 @itemx -fno-signed-bitfields
1245 @itemx -fno-unsigned-bitfields
1246 @opindex fsigned-bitfields
1247 @opindex funsigned-bitfields
1248 @opindex fno-signed-bitfields
1249 @opindex fno-unsigned-bitfields
1250 These options control whether a bit-field is signed or unsigned, when the
1251 declaration does not use either @code{signed} or @code{unsigned}. By
1252 default, such a bit-field is signed, because this is consistent: the
1253 basic integer types such as @code{int} are signed types.
1255 @item -fwritable-strings
1256 @opindex fwritable-strings
1257 Store string constants in the writable data segment and don't uniquize
1258 them. This is for compatibility with old programs which assume they can
1259 write into string constants.
1261 Writing into string constants is a very bad idea; ``constants'' should
1265 @node C++ Dialect Options
1266 @section Options Controlling C++ Dialect
1268 @cindex compiler options, C++
1269 @cindex C++ options, command line
1270 @cindex options, C++
1271 This section describes the command-line options that are only meaningful
1272 for C++ programs; but you can also use most of the GNU compiler options
1273 regardless of what language your program is in. For example, you
1274 might compile a file @code{firstClass.C} like this:
1277 g++ -g -frepo -O -c firstClass.C
1281 In this example, only @option{-frepo} is an option meant
1282 only for C++ programs; you can use the other options with any
1283 language supported by GCC@.
1285 Here is a list of options that are @emph{only} for compiling C++ programs:
1289 @item -fabi-version=@var{n}
1290 @opindex fabi-version
1291 Use version @var{n} of the C++ ABI. Version 1 is the version of the C++
1292 ABI that first appeared in G++ 3.2. Version 0 will always be the
1293 version that conforms most closely to the C++ ABI specification.
1294 Therefore, the ABI obtained using version 0 will change as ABI bugs are
1297 The default is version 1.
1299 @item -fno-access-control
1300 @opindex fno-access-control
1301 Turn off all access checking. This switch is mainly useful for working
1302 around bugs in the access control code.
1306 Check that the pointer returned by @code{operator new} is non-null
1307 before attempting to modify the storage allocated. The current Working
1308 Paper requires that @code{operator new} never return a null pointer, so
1309 this check is normally unnecessary.
1311 An alternative to using this option is to specify that your
1312 @code{operator new} does not throw any exceptions; if you declare it
1313 @samp{throw()}, G++ will check the return value. See also @samp{new
1316 @item -fconserve-space
1317 @opindex fconserve-space
1318 Put uninitialized or runtime-initialized global variables into the
1319 common segment, as C does. This saves space in the executable at the
1320 cost of not diagnosing duplicate definitions. If you compile with this
1321 flag and your program mysteriously crashes after @code{main()} has
1322 completed, you may have an object that is being destroyed twice because
1323 two definitions were merged.
1325 This option is no longer useful on most targets, now that support has
1326 been added for putting variables into BSS without making them common.
1328 @item -fno-const-strings
1329 @opindex fno-const-strings
1330 Give string constants type @code{char *} instead of type @code{const
1331 char *}. By default, G++ uses type @code{const char *} as required by
1332 the standard. Even if you use @option{-fno-const-strings}, you cannot
1333 actually modify the value of a string constant, unless you also use
1334 @option{-fwritable-strings}.
1336 This option might be removed in a future release of G++. For maximum
1337 portability, you should structure your code so that it works with
1338 string constants that have type @code{const char *}.
1340 @item -fdollars-in-identifiers
1341 @opindex fdollars-in-identifiers
1342 Accept @samp{$} in identifiers. You can also explicitly prohibit use of
1343 @samp{$} with the option @option{-fno-dollars-in-identifiers}. (GNU C allows
1344 @samp{$} by default on most target systems, but there are a few exceptions.)
1345 Traditional C allowed the character @samp{$} to form part of
1346 identifiers. However, ISO C and C++ forbid @samp{$} in identifiers.
1348 @item -fno-elide-constructors
1349 @opindex fno-elide-constructors
1350 The C++ standard allows an implementation to omit creating a temporary
1351 which is only used to initialize another object of the same type.
1352 Specifying this option disables that optimization, and forces G++ to
1353 call the copy constructor in all cases.
1355 @item -fno-enforce-eh-specs
1356 @opindex fno-enforce-eh-specs
1357 Don't check for violation of exception specifications at runtime. This
1358 option violates the C++ standard, but may be useful for reducing code
1359 size in production builds, much like defining @samp{NDEBUG}. The compiler
1360 will still optimize based on the exception specifications.
1362 @item -fexternal-templates
1363 @opindex fexternal-templates
1365 Cause @samp{#pragma interface} and @samp{implementation} to apply to
1366 template instantiation; template instances are emitted or not according
1367 to the location of the template definition. @xref{Template
1368 Instantiation}, for more information.
1370 This option is deprecated.
1372 @item -falt-external-templates
1373 @opindex falt-external-templates
1374 Similar to @option{-fexternal-templates}, but template instances are
1375 emitted or not according to the place where they are first instantiated.
1376 @xref{Template Instantiation}, for more information.
1378 This option is deprecated.
1381 @itemx -fno-for-scope
1383 @opindex fno-for-scope
1384 If @option{-ffor-scope} is specified, the scope of variables declared in
1385 a @i{for-init-statement} is limited to the @samp{for} loop itself,
1386 as specified by the C++ standard.
1387 If @option{-fno-for-scope} is specified, the scope of variables declared in
1388 a @i{for-init-statement} extends to the end of the enclosing scope,
1389 as was the case in old versions of G++, and other (traditional)
1390 implementations of C++.
1392 The default if neither flag is given to follow the standard,
1393 but to allow and give a warning for old-style code that would
1394 otherwise be invalid, or have different behavior.
1396 @item -fno-gnu-keywords
1397 @opindex fno-gnu-keywords
1398 Do not recognize @code{typeof} as a keyword, so that code can use this
1399 word as an identifier. You can use the keyword @code{__typeof__} instead.
1400 @option{-ansi} implies @option{-fno-gnu-keywords}.
1402 @item -fno-implicit-templates
1403 @opindex fno-implicit-templates
1404 Never emit code for non-inline templates which are instantiated
1405 implicitly (i.e.@: by use); only emit code for explicit instantiations.
1406 @xref{Template Instantiation}, for more information.
1408 @item -fno-implicit-inline-templates
1409 @opindex fno-implicit-inline-templates
1410 Don't emit code for implicit instantiations of inline templates, either.
1411 The default is to handle inlines differently so that compiles with and
1412 without optimization will need the same set of explicit instantiations.
1414 @item -fno-implement-inlines
1415 @opindex fno-implement-inlines
1416 To save space, do not emit out-of-line copies of inline functions
1417 controlled by @samp{#pragma implementation}. This will cause linker
1418 errors if these functions are not inlined everywhere they are called.
1420 @item -fms-extensions
1421 @opindex fms-extensions
1422 Disable pedantic warnings about constructs used in MFC, such as implicit
1423 int and getting a pointer to member function via non-standard syntax.
1425 @item -fno-nonansi-builtins
1426 @opindex fno-nonansi-builtins
1427 Disable built-in declarations of functions that are not mandated by
1428 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
1429 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
1431 @item -fno-operator-names
1432 @opindex fno-operator-names
1433 Do not treat the operator name keywords @code{and}, @code{bitand},
1434 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
1435 synonyms as keywords.
1437 @item -fno-optional-diags
1438 @opindex fno-optional-diags
1439 Disable diagnostics that the standard says a compiler does not need to
1440 issue. Currently, the only such diagnostic issued by G++ is the one for
1441 a name having multiple meanings within a class.
1444 @opindex fpermissive
1445 Downgrade messages about nonconformant code from errors to warnings. By
1446 default, G++ effectively sets @option{-pedantic-errors} without
1447 @option{-pedantic}; this option reverses that. This behavior and this
1448 option are superseded by @option{-pedantic}, which works as it does for GNU C@.
1452 Enable automatic template instantiation at link time. This option also
1453 implies @option{-fno-implicit-templates}. @xref{Template
1454 Instantiation}, for more information.
1458 Disable generation of information about every class with virtual
1459 functions for use by the C++ runtime type identification features
1460 (@samp{dynamic_cast} and @samp{typeid}). If you don't use those parts
1461 of the language, you can save some space by using this flag. Note that
1462 exception handling uses the same information, but it will generate it as
1467 Emit statistics about front-end processing at the end of the compilation.
1468 This information is generally only useful to the G++ development team.
1470 @item -ftemplate-depth-@var{n}
1471 @opindex ftemplate-depth
1472 Set the maximum instantiation depth for template classes to @var{n}.
1473 A limit on the template instantiation depth is needed to detect
1474 endless recursions during template class instantiation. ANSI/ISO C++
1475 conforming programs must not rely on a maximum depth greater than 17.
1477 @item -fuse-cxa-atexit
1478 @opindex fuse-cxa-atexit
1479 Register destructors for objects with static storage duration with the
1480 @code{__cxa_atexit} function rather than the @code{atexit} function.
1481 This option is required for fully standards-compliant handling of static
1482 destructors, but will only work if your C library supports
1483 @code{__cxa_atexit}.
1487 Emit special relocations for vtables and virtual function references
1488 so that the linker can identify unused virtual functions and zero out
1489 vtable slots that refer to them. This is most useful with
1490 @option{-ffunction-sections} and @option{-Wl,--gc-sections}, in order to
1491 also discard the functions themselves.
1493 This optimization requires GNU as and GNU ld. Not all systems support
1494 this option. @option{-Wl,--gc-sections} is ignored without @option{-static}.
1498 Do not use weak symbol support, even if it is provided by the linker.
1499 By default, G++ will use weak symbols if they are available. This
1500 option exists only for testing, and should not be used by end-users;
1501 it will result in inferior code and has no benefits. This option may
1502 be removed in a future release of G++.
1506 Do not search for header files in the standard directories specific to
1507 C++, but do still search the other standard directories. (This option
1508 is used when building the C++ library.)
1511 In addition, these optimization, warning, and code generation options
1512 have meanings only for C++ programs:
1515 @item -fno-default-inline
1516 @opindex fno-default-inline
1517 Do not assume @samp{inline} for functions defined inside a class scope.
1518 @xref{Optimize Options,,Options That Control Optimization}. Note that these
1519 functions will have linkage like inline functions; they just won't be
1522 @item -Wabi @r{(C++ only)}
1524 Warn when G++ generates code that is probably not compatible with the
1525 vendor-neutral C++ ABI. Although an effort has been made to warn about
1526 all such cases, there are probably some cases that are not warned about,
1527 even though G++ is generating incompatible code. There may also be
1528 cases where warnings are emitted even though the code that is generated
1531 You should rewrite your code to avoid these warnings if you are
1532 concerned about the fact that code generated by G++ may not be binary
1533 compatible with code generated by other compilers.
1535 The known incompatibilities at this point include:
1540 Incorrect handling of tail-padding for bit-fields. G++ may attempt to
1541 pack data into the same byte as a base class. For example:
1544 struct A @{ virtual void f(); int f1 : 1; @};
1545 struct B : public A @{ int f2 : 1; @};
1549 In this case, G++ will place @code{B::f2} into the same byte
1550 as@code{A::f1}; other compilers will not. You can avoid this problem
1551 by explicitly padding @code{A} so that its size is a multiple of the
1552 byte size on your platform; that will cause G++ and other compilers to
1553 layout @code{B} identically.
1556 Incorrect handling of tail-padding for virtual bases. G++ does not use
1557 tail padding when laying out virtual bases. For example:
1560 struct A @{ virtual void f(); char c1; @};
1561 struct B @{ B(); char c2; @};
1562 struct C : public A, public virtual B @{@};
1566 In this case, G++ will not place @code{B} into the tail-padding for
1567 @code{A}; other compilers will. You can avoid this problem by
1568 explicitly padding @code{A} so that its size is a multiple of its
1569 alignment (ignoring virtual base classes); that will cause G++ and other
1570 compilers to layout @code{C} identically.
1573 Incorrect handling of bit-fields with declared widths greater than that
1574 of their underlying types, when the bit-fields appear in a union. For
1578 union U @{ int i : 4096; @};
1582 Assuming that an @code{int} does not have 4096 bits, G++ will make the
1583 union too small by the number of bits in an @code{int}.
1586 Empty classes can be placed at incorrect offsets. For example:
1596 struct C : public B, public A @{@};
1600 G++ will place the @code{A} base class of @code{C} at a nonzero offset;
1601 it should be placed at offset zero. G++ mistakenly believes that the
1602 @code{A} data member of @code{B} is already at offset zero.
1605 Names of template functions whose types involve @code{typename} or
1606 template template parameters can be mangled incorrectly.
1609 template <typename Q>
1610 void f(typename Q::X) @{@}
1612 template <template <typename> class Q>
1613 void f(typename Q<int>::X) @{@}
1617 Instantiations of these templates may be mangled incorrectly.
1621 @item -Wctor-dtor-privacy @r{(C++ only)}
1622 @opindex Wctor-dtor-privacy
1623 Warn when a class seems unusable, because all the constructors or
1624 destructors in a class are private and the class has no friends or
1625 public static member functions. This warning is enabled by default.
1627 @item -Wnon-virtual-dtor @r{(C++ only)}
1628 @opindex Wnon-virtual-dtor
1629 Warn when a class declares a non-virtual destructor that should probably
1630 be virtual, because it looks like the class will be used polymorphically.
1631 This warning is enabled by @option{-Wall}.
1633 @item -Wreorder @r{(C++ only)}
1635 @cindex reordering, warning
1636 @cindex warning for reordering of member initializers
1637 Warn when the order of member initializers given in the code does not
1638 match the order in which they must be executed. For instance:
1644 A(): j (0), i (1) @{ @}
1648 Here the compiler will warn that the member initializers for @samp{i}
1649 and @samp{j} will be rearranged to match the declaration order of the
1650 members. This warning is enabled by @option{-Wall}.
1653 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
1656 @item -Weffc++ @r{(C++ only)}
1658 Warn about violations of the following style guidelines from Scott Meyers'
1659 @cite{Effective C++} book:
1663 Item 11: Define a copy constructor and an assignment operator for classes
1664 with dynamically allocated memory.
1667 Item 12: Prefer initialization to assignment in constructors.
1670 Item 14: Make destructors virtual in base classes.
1673 Item 15: Have @code{operator=} return a reference to @code{*this}.
1676 Item 23: Don't try to return a reference when you must return an object.
1680 and about violations of the following style guidelines from Scott Meyers'
1681 @cite{More Effective C++} book:
1685 Item 6: Distinguish between prefix and postfix forms of increment and
1686 decrement operators.
1689 Item 7: Never overload @code{&&}, @code{||}, or @code{,}.
1693 If you use this option, you should be aware that the standard library
1694 headers do not obey all of these guidelines; you can use @samp{grep -v}
1695 to filter out those warnings.
1697 @item -Wno-deprecated @r{(C++ only)}
1698 @opindex Wno-deprecated
1699 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
1701 @item -Wno-non-template-friend @r{(C++ only)}
1702 @opindex Wno-non-template-friend
1703 Disable warnings when non-templatized friend functions are declared
1704 within a template. With the advent of explicit template specification
1705 support in G++, if the name of the friend is an unqualified-id (i.e.,
1706 @samp{friend foo(int)}), the C++ language specification demands that the
1707 friend declare or define an ordinary, nontemplate function. (Section
1708 14.5.3). Before G++ implemented explicit specification, unqualified-ids
1709 could be interpreted as a particular specialization of a templatized
1710 function. Because this non-conforming behavior is no longer the default
1711 behavior for G++, @option{-Wnon-template-friend} allows the compiler to
1712 check existing code for potential trouble spots, and is on by default.
1713 This new compiler behavior can be turned off with
1714 @option{-Wno-non-template-friend} which keeps the conformant compiler code
1715 but disables the helpful warning.
1717 @item -Wold-style-cast @r{(C++ only)}
1718 @opindex Wold-style-cast
1719 Warn if an old-style (C-style) cast to a non-void type is used within
1720 a C++ program. The new-style casts (@samp{static_cast},
1721 @samp{reinterpret_cast}, and @samp{const_cast}) are less vulnerable to
1722 unintended effects, and much easier to grep for.
1724 @item -Woverloaded-virtual @r{(C++ only)}
1725 @opindex Woverloaded-virtual
1726 @cindex overloaded virtual fn, warning
1727 @cindex warning for overloaded virtual fn
1728 Warn when a function declaration hides virtual functions from a
1729 base class. For example, in:
1736 struct B: public A @{
1741 the @code{A} class version of @code{f} is hidden in @code{B}, and code
1749 will fail to compile.
1751 @item -Wno-pmf-conversions @r{(C++ only)}
1752 @opindex Wno-pmf-conversions
1753 Disable the diagnostic for converting a bound pointer to member function
1756 @item -Wsign-promo @r{(C++ only)}
1757 @opindex Wsign-promo
1758 Warn when overload resolution chooses a promotion from unsigned or
1759 enumeral type to a signed type over a conversion to an unsigned type of
1760 the same size. Previous versions of G++ would try to preserve
1761 unsignedness, but the standard mandates the current behavior.
1763 @item -Wsynth @r{(C++ only)}
1765 @cindex warning for synthesized methods
1766 @cindex synthesized methods, warning
1767 Warn when G++'s synthesis behavior does not match that of cfront. For
1773 A& operator = (int);
1783 In this example, G++ will synthesize a default @samp{A& operator =
1784 (const A&);}, while cfront will use the user-defined @samp{operator =}.
1787 @node Objective-C Dialect Options
1788 @section Options Controlling Objective-C Dialect
1790 @cindex compiler options, Objective-C
1791 @cindex Objective-C options, command line
1792 @cindex options, Objective-C
1793 This section describes the command-line options that are only meaningful
1794 for Objective-C programs; but you can also use most of the GNU compiler
1795 options regardless of what language your program is in. For example,
1796 you might compile a file @code{some_class.m} like this:
1799 gcc -g -fgnu-runtime -O -c some_class.m
1803 In this example, only @option{-fgnu-runtime} is an option meant only for
1804 Objective-C programs; you can use the other options with any language
1807 Here is a list of options that are @emph{only} for compiling Objective-C
1811 @item -fconstant-string-class=@var{class-name}
1812 @opindex fconstant-string-class
1813 Use @var{class-name} as the name of the class to instantiate for each
1814 literal string specified with the syntax @code{@@"@dots{}"}. The default
1815 class name is @code{NXConstantString}.
1818 @opindex fgnu-runtime
1819 Generate object code compatible with the standard GNU Objective-C
1820 runtime. This is the default for most types of systems.
1822 @item -fnext-runtime
1823 @opindex fnext-runtime
1824 Generate output compatible with the NeXT runtime. This is the default
1825 for NeXT-based systems, including Darwin and Mac OS X@. The macro
1826 @code{__NEXT_RUNTIME__} is predefined if (and only if) this option is
1831 Dump interface declarations for all classes seen in the source file to a
1832 file named @file{@var{sourcename}.decl}.
1835 @opindex Wno-protocol
1836 If a class is declared to implement a protocol, a warning is issued for
1837 every method in the protocol that is not implemented by the class. The
1838 default behavior is to issue a warning for every method not explicitly
1839 implemented in the class, even if a method implementation is inherited
1840 from the superclass. If you use the @code{-Wno-protocol} option, then
1841 methods inherited from the superclass are considered to be implemented,
1842 and no warning is issued for them.
1846 Warn if multiple methods of different types for the same selector are
1847 found during compilation. The check is performed on the list of methods
1848 in the final stage of compilation. Additionally, a check is performed
1849 that for each selector appearing in a @code{@@selector(@dots{})}
1850 expression, a corresponding method with that selector has been found
1851 during compilation. Because these checks scan the method table only at
1852 the end of compilation, these warnings are not produced if the final
1853 stage of compilation is not reached, for example because an error is
1854 found during compilation, or because the @code{-fsyntax-only} option is
1857 @item -Wundeclared-selector
1858 @opindex Wundeclared-selector
1859 Warn if a @code{@@selector(@dots{})} expression referring to an
1860 undeclared selector is found. A selector is considered undeclared if no
1861 method with that name has been declared (explicitly, in an
1862 @code{@@interface} or @code{@@protocol} declaration, or implicitly, in
1863 an @code{@@implementation} section) before the
1864 @code{@@selector(@dots{})} expression. This option always performs its
1865 checks as soon as a @code{@@selector(@dots{})} expression is found
1866 (while @code{-Wselector} only performs its checks in the final stage of
1867 compilation), and so additionally enforces the coding style convention
1868 that methods and selectors must be declared before being used.
1870 @c not documented because only avail via -Wp
1871 @c @item -print-objc-runtime-info
1875 @node Language Independent Options
1876 @section Options to Control Diagnostic Messages Formatting
1877 @cindex options to control diagnostics formatting
1878 @cindex diagnostic messages
1879 @cindex message formatting
1881 Traditionally, diagnostic messages have been formatted irrespective of
1882 the output device's aspect (e.g.@: its width, @dots{}). The options described
1883 below can be used to control the diagnostic messages formatting
1884 algorithm, e.g.@: how many characters per line, how often source location
1885 information should be reported. Right now, only the C++ front end can
1886 honor these options. However it is expected, in the near future, that
1887 the remaining front ends would be able to digest them correctly.
1890 @item -fmessage-length=@var{n}
1891 @opindex fmessage-length
1892 Try to format error messages so that they fit on lines of about @var{n}
1893 characters. The default is 72 characters for @command{g++} and 0 for the rest of
1894 the front ends supported by GCC@. If @var{n} is zero, then no
1895 line-wrapping will be done; each error message will appear on a single
1898 @opindex fdiagnostics-show-location
1899 @item -fdiagnostics-show-location=once
1900 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
1901 reporter to emit @emph{once} source location information; that is, in
1902 case the message is too long to fit on a single physical line and has to
1903 be wrapped, the source location won't be emitted (as prefix) again,
1904 over and over, in subsequent continuation lines. This is the default
1907 @item -fdiagnostics-show-location=every-line
1908 Only meaningful in line-wrapping mode. Instructs the diagnostic
1909 messages reporter to emit the same source location information (as
1910 prefix) for physical lines that result from the process of breaking
1911 a message which is too long to fit on a single line.
1915 @node Warning Options
1916 @section Options to Request or Suppress Warnings
1917 @cindex options to control warnings
1918 @cindex warning messages
1919 @cindex messages, warning
1920 @cindex suppressing warnings
1922 Warnings are diagnostic messages that report constructions which
1923 are not inherently erroneous but which are risky or suggest there
1924 may have been an error.
1926 You can request many specific warnings with options beginning @samp{-W},
1927 for example @option{-Wimplicit} to request warnings on implicit
1928 declarations. Each of these specific warning options also has a
1929 negative form beginning @samp{-Wno-} to turn off warnings;
1930 for example, @option{-Wno-implicit}. This manual lists only one of the
1931 two forms, whichever is not the default.
1933 The following options control the amount and kinds of warnings produced
1934 by GCC; for further, language-specific options also refer to
1935 @ref{C++ Dialect Options} and @ref{Objective-C Dialect Options}.
1938 @cindex syntax checking
1940 @opindex fsyntax-only
1941 Check the code for syntax errors, but don't do anything beyond that.
1945 Issue all the warnings demanded by strict ISO C and ISO C++;
1946 reject all programs that use forbidden extensions, and some other
1947 programs that do not follow ISO C and ISO C++. For ISO C, follows the
1948 version of the ISO C standard specified by any @option{-std} option used.
1950 Valid ISO C and ISO C++ programs should compile properly with or without
1951 this option (though a rare few will require @option{-ansi} or a
1952 @option{-std} option specifying the required version of ISO C)@. However,
1953 without this option, certain GNU extensions and traditional C and C++
1954 features are supported as well. With this option, they are rejected.
1956 @option{-pedantic} does not cause warning messages for use of the
1957 alternate keywords whose names begin and end with @samp{__}. Pedantic
1958 warnings are also disabled in the expression that follows
1959 @code{__extension__}. However, only system header files should use
1960 these escape routes; application programs should avoid them.
1961 @xref{Alternate Keywords}.
1963 Some users try to use @option{-pedantic} to check programs for strict ISO
1964 C conformance. They soon find that it does not do quite what they want:
1965 it finds some non-ISO practices, but not all---only those for which
1966 ISO C @emph{requires} a diagnostic, and some others for which
1967 diagnostics have been added.
1969 A feature to report any failure to conform to ISO C might be useful in
1970 some instances, but would require considerable additional work and would
1971 be quite different from @option{-pedantic}. We don't have plans to
1972 support such a feature in the near future.
1974 Where the standard specified with @option{-std} represents a GNU
1975 extended dialect of C, such as @samp{gnu89} or @samp{gnu99}, there is a
1976 corresponding @dfn{base standard}, the version of ISO C on which the GNU
1977 extended dialect is based. Warnings from @option{-pedantic} are given
1978 where they are required by the base standard. (It would not make sense
1979 for such warnings to be given only for features not in the specified GNU
1980 C dialect, since by definition the GNU dialects of C include all
1981 features the compiler supports with the given option, and there would be
1982 nothing to warn about.)
1984 @item -pedantic-errors
1985 @opindex pedantic-errors
1986 Like @option{-pedantic}, except that errors are produced rather than
1991 Inhibit all warning messages.
1995 Inhibit warning messages about the use of @samp{#import}.
1997 @item -Wchar-subscripts
1998 @opindex Wchar-subscripts
1999 Warn if an array subscript has type @code{char}. This is a common cause
2000 of error, as programmers often forget that this type is signed on some
2005 Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*}
2006 comment, or whenever a Backslash-Newline appears in a @samp{//} comment.
2010 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
2011 the arguments supplied have types appropriate to the format string
2012 specified, and that the conversions specified in the format string make
2013 sense. This includes standard functions, and others specified by format
2014 attributes (@pxref{Function Attributes}), in the @code{printf},
2015 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
2016 not in the C standard) families.
2018 The formats are checked against the format features supported by GNU
2019 libc version 2.2. These include all ISO C90 and C99 features, as well
2020 as features from the Single Unix Specification and some BSD and GNU
2021 extensions. Other library implementations may not support all these
2022 features; GCC does not support warning about features that go beyond a
2023 particular library's limitations. However, if @option{-pedantic} is used
2024 with @option{-Wformat}, warnings will be given about format features not
2025 in the selected standard version (but not for @code{strfmon} formats,
2026 since those are not in any version of the C standard). @xref{C Dialect
2027 Options,,Options Controlling C Dialect}.
2029 Since @option{-Wformat} also checks for null format arguments for
2030 several functions, @option{-Wformat} also implies @option{-Wnonnull}.
2032 @option{-Wformat} is included in @option{-Wall}. For more control over some
2033 aspects of format checking, the options @option{-Wno-format-y2k},
2034 @option{-Wno-format-extra-args}, @option{-Wno-format-zero-length},
2035 @option{-Wformat-nonliteral}, @option{-Wformat-security}, and
2036 @option{-Wformat=2} are available, but are not included in @option{-Wall}.
2038 @item -Wno-format-y2k
2039 @opindex Wno-format-y2k
2040 If @option{-Wformat} is specified, do not warn about @code{strftime}
2041 formats which may yield only a two-digit year.
2043 @item -Wno-format-extra-args
2044 @opindex Wno-format-extra-args
2045 If @option{-Wformat} is specified, do not warn about excess arguments to a
2046 @code{printf} or @code{scanf} format function. The C standard specifies
2047 that such arguments are ignored.
2049 Where the unused arguments lie between used arguments that are
2050 specified with @samp{$} operand number specifications, normally
2051 warnings are still given, since the implementation could not know what
2052 type to pass to @code{va_arg} to skip the unused arguments. However,
2053 in the case of @code{scanf} formats, this option will suppress the
2054 warning if the unused arguments are all pointers, since the Single
2055 Unix Specification says that such unused arguments are allowed.
2057 @item -Wno-format-zero-length
2058 @opindex Wno-format-zero-length
2059 If @option{-Wformat} is specified, do not warn about zero-length formats.
2060 The C standard specifies that zero-length formats are allowed.
2062 @item -Wformat-nonliteral
2063 @opindex Wformat-nonliteral
2064 If @option{-Wformat} is specified, also warn if the format string is not a
2065 string literal and so cannot be checked, unless the format function
2066 takes its format arguments as a @code{va_list}.
2068 @item -Wformat-security
2069 @opindex Wformat-security
2070 If @option{-Wformat} is specified, also warn about uses of format
2071 functions that represent possible security problems. At present, this
2072 warns about calls to @code{printf} and @code{scanf} functions where the
2073 format string is not a string literal and there are no format arguments,
2074 as in @code{printf (foo);}. This may be a security hole if the format
2075 string came from untrusted input and contains @samp{%n}. (This is
2076 currently a subset of what @option{-Wformat-nonliteral} warns about, but
2077 in future warnings may be added to @option{-Wformat-security} that are not
2078 included in @option{-Wformat-nonliteral}.)
2082 Enable @option{-Wformat} plus format checks not included in
2083 @option{-Wformat}. Currently equivalent to @samp{-Wformat
2084 -Wformat-nonliteral -Wformat-security}.
2088 Enable warning about passing a null pointer for arguments marked as
2089 requiring a non-null value by the @code{nonnull} function attribute.
2091 @option{-Wnonnull} is included in @option{-Wall} and @option{-Wformat}. It
2092 can be disabled with the @option{-Wno-nonnull} option.
2094 @item -Wimplicit-int
2095 @opindex Wimplicit-int
2096 Warn when a declaration does not specify a type.
2098 @item -Wimplicit-function-declaration
2099 @itemx -Werror-implicit-function-declaration
2100 @opindex Wimplicit-function-declaration
2101 @opindex Werror-implicit-function-declaration
2102 Give a warning (or error) whenever a function is used before being
2107 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
2111 Warn if the type of @samp{main} is suspicious. @samp{main} should be a
2112 function with external linkage, returning int, taking either zero
2113 arguments, two, or three arguments of appropriate types.
2115 @item -Wmissing-braces
2116 @opindex Wmissing-braces
2117 Warn if an aggregate or union initializer is not fully bracketed. In
2118 the following example, the initializer for @samp{a} is not fully
2119 bracketed, but that for @samp{b} is fully bracketed.
2122 int a[2][2] = @{ 0, 1, 2, 3 @};
2123 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
2127 @opindex Wparentheses
2128 Warn if parentheses are omitted in certain contexts, such
2129 as when there is an assignment in a context where a truth value
2130 is expected, or when operators are nested whose precedence people
2131 often get confused about.
2133 Also warn about constructions where there may be confusion to which
2134 @code{if} statement an @code{else} branch belongs. Here is an example of
2149 In C, every @code{else} branch belongs to the innermost possible @code{if}
2150 statement, which in this example is @code{if (b)}. This is often not
2151 what the programmer expected, as illustrated in the above example by
2152 indentation the programmer chose. When there is the potential for this
2153 confusion, GCC will issue a warning when this flag is specified.
2154 To eliminate the warning, add explicit braces around the innermost
2155 @code{if} statement so there is no way the @code{else} could belong to
2156 the enclosing @code{if}. The resulting code would look like this:
2172 @item -Wsequence-point
2173 @opindex Wsequence-point
2174 Warn about code that may have undefined semantics because of violations
2175 of sequence point rules in the C standard.
2177 The C standard defines the order in which expressions in a C program are
2178 evaluated in terms of @dfn{sequence points}, which represent a partial
2179 ordering between the execution of parts of the program: those executed
2180 before the sequence point, and those executed after it. These occur
2181 after the evaluation of a full expression (one which is not part of a
2182 larger expression), after the evaluation of the first operand of a
2183 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
2184 function is called (but after the evaluation of its arguments and the
2185 expression denoting the called function), and in certain other places.
2186 Other than as expressed by the sequence point rules, the order of
2187 evaluation of subexpressions of an expression is not specified. All
2188 these rules describe only a partial order rather than a total order,
2189 since, for example, if two functions are called within one expression
2190 with no sequence point between them, the order in which the functions
2191 are called is not specified. However, the standards committee have
2192 ruled that function calls do not overlap.
2194 It is not specified when between sequence points modifications to the
2195 values of objects take effect. Programs whose behavior depends on this
2196 have undefined behavior; the C standard specifies that ``Between the
2197 previous and next sequence point an object shall have its stored value
2198 modified at most once by the evaluation of an expression. Furthermore,
2199 the prior value shall be read only to determine the value to be
2200 stored.''. If a program breaks these rules, the results on any
2201 particular implementation are entirely unpredictable.
2203 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
2204 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
2205 diagnosed by this option, and it may give an occasional false positive
2206 result, but in general it has been found fairly effective at detecting
2207 this sort of problem in programs.
2209 The present implementation of this option only works for C programs. A
2210 future implementation may also work for C++ programs.
2212 The C standard is worded confusingly, therefore there is some debate
2213 over the precise meaning of the sequence point rules in subtle cases.
2214 Links to discussions of the problem, including proposed formal
2215 definitions, may be found on our readings page, at
2216 @w{@uref{http://gcc.gnu.org/readings.html}}.
2219 @opindex Wreturn-type
2220 Warn whenever a function is defined with a return-type that defaults to
2221 @code{int}. Also warn about any @code{return} statement with no
2222 return-value in a function whose return-type is not @code{void}.
2224 For C++, a function without return type always produces a diagnostic
2225 message, even when @option{-Wno-return-type} is specified. The only
2226 exceptions are @samp{main} and functions defined in system headers.
2230 Warn whenever a @code{switch} statement has an index of enumeral type
2231 and lacks a @code{case} for one or more of the named codes of that
2232 enumeration. (The presence of a @code{default} label prevents this
2233 warning.) @code{case} labels outside the enumeration range also
2234 provoke warnings when this option is used.
2236 @item -Wswitch-default
2237 @opindex Wswitch-switch
2238 Warn whenever a @code{switch} statement does not have a @code{default}
2242 @opindex Wswitch-enum
2243 Warn whenever a @code{switch} statement has an index of enumeral type
2244 and lacks a @code{case} for one or more of the named codes of that
2245 enumeration. @code{case} labels outside the enumeration range also
2246 provoke warnings when this option is used.
2250 Warn if any trigraphs are encountered that might change the meaning of
2251 the program (trigraphs within comments are not warned about).
2253 @item -Wunused-function
2254 @opindex Wunused-function
2255 Warn whenever a static function is declared but not defined or a
2256 non\-inline static function is unused.
2258 @item -Wunused-label
2259 @opindex Wunused-label
2260 Warn whenever a label is declared but not used.
2262 To suppress this warning use the @samp{unused} attribute
2263 (@pxref{Variable Attributes}).
2265 @item -Wunused-parameter
2266 @opindex Wunused-parameter
2267 Warn whenever a function parameter is unused aside from its declaration.
2269 To suppress this warning use the @samp{unused} attribute
2270 (@pxref{Variable Attributes}).
2272 @item -Wunused-variable
2273 @opindex Wunused-variable
2274 Warn whenever a local variable or non-constant static variable is unused
2275 aside from its declaration
2277 To suppress this warning use the @samp{unused} attribute
2278 (@pxref{Variable Attributes}).
2280 @item -Wunused-value
2281 @opindex Wunused-value
2282 Warn whenever a statement computes a result that is explicitly not used.
2284 To suppress this warning cast the expression to @samp{void}.
2288 All the above @option{-Wunused} options combined.
2290 In order to get a warning about an unused function parameter, you must
2291 either specify @samp{-W -Wunused} or separately specify
2292 @option{-Wunused-parameter}.
2294 @item -Wuninitialized
2295 @opindex Wuninitialized
2296 Warn if an automatic variable is used without first being initialized or
2297 if a variable may be clobbered by a @code{setjmp} call.
2299 These warnings are possible only in optimizing compilation,
2300 because they require data flow information that is computed only
2301 when optimizing. If you don't specify @option{-O}, you simply won't
2304 These warnings occur only for variables that are candidates for
2305 register allocation. Therefore, they do not occur for a variable that
2306 is declared @code{volatile}, or whose address is taken, or whose size
2307 is other than 1, 2, 4 or 8 bytes. Also, they do not occur for
2308 structures, unions or arrays, even when they are in registers.
2310 Note that there may be no warning about a variable that is used only
2311 to compute a value that itself is never used, because such
2312 computations may be deleted by data flow analysis before the warnings
2315 These warnings are made optional because GCC is not smart
2316 enough to see all the reasons why the code might be correct
2317 despite appearing to have an error. Here is one example of how
2338 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
2339 always initialized, but GCC doesn't know this. Here is
2340 another common case:
2345 if (change_y) save_y = y, y = new_y;
2347 if (change_y) y = save_y;
2352 This has no bug because @code{save_y} is used only if it is set.
2354 @cindex @code{longjmp} warnings
2355 This option also warns when a non-volatile automatic variable might be
2356 changed by a call to @code{longjmp}. These warnings as well are possible
2357 only in optimizing compilation.
2359 The compiler sees only the calls to @code{setjmp}. It cannot know
2360 where @code{longjmp} will be called; in fact, a signal handler could
2361 call it at any point in the code. As a result, you may get a warning
2362 even when there is in fact no problem because @code{longjmp} cannot
2363 in fact be called at the place which would cause a problem.
2365 Some spurious warnings can be avoided if you declare all the functions
2366 you use that never return as @code{noreturn}. @xref{Function
2369 @item -Wunknown-pragmas
2370 @opindex Wunknown-pragmas
2371 @cindex warning for unknown pragmas
2372 @cindex unknown pragmas, warning
2373 @cindex pragmas, warning of unknown
2374 Warn when a #pragma directive is encountered which is not understood by
2375 GCC@. If this command line option is used, warnings will even be issued
2376 for unknown pragmas in system header files. This is not the case if
2377 the warnings were only enabled by the @option{-Wall} command line option.
2379 @item -Wstrict-aliasing
2380 @opindex Wstrict-aliasing
2381 This option is only active when @option{-fstrict-aliasing} is active.
2382 It warns about code which might break the strict aliasing rules that the
2383 compiler is using for optimization. The warning does not catch all
2384 cases, but does attempt to catch the more common pitfalls. It is
2385 included in @option{-Wall}.
2389 All of the above @samp{-W} options combined. This enables all the
2390 warnings about constructions that some users consider questionable, and
2391 that are easy to avoid (or modify to prevent the warning), even in
2392 conjunction with macros. This also enables some language-specific
2393 warnings described in @ref{C++ Dialect Options} and
2394 @ref{Objective-C Dialect Options}.
2397 The following @option{-W@dots{}} options are not implied by @option{-Wall}.
2398 Some of them warn about constructions that users generally do not
2399 consider questionable, but which occasionally you might wish to check
2400 for; others warn about constructions that are necessary or hard to avoid
2401 in some cases, and there is no simple way to modify the code to suppress
2407 Print extra warning messages for these events:
2411 A function can return either with or without a value. (Falling
2412 off the end of the function body is considered returning without
2413 a value.) For example, this function would evoke such a
2427 An expression-statement or the left-hand side of a comma expression
2428 contains no side effects.
2429 To suppress the warning, cast the unused expression to void.
2430 For example, an expression such as @samp{x[i,j]} will cause a warning,
2431 but @samp{x[(void)i,j]} will not.
2434 An unsigned value is compared against zero with @samp{<} or @samp{<=}.
2437 A comparison like @samp{x<=y<=z} appears; this is equivalent to
2438 @samp{(x<=y ? 1 : 0) <= z}, which is a different interpretation from
2439 that of ordinary mathematical notation.
2442 Storage-class specifiers like @code{static} are not the first things in
2443 a declaration. According to the C Standard, this usage is obsolescent.
2446 The return type of a function has a type qualifier such as @code{const}.
2447 Such a type qualifier has no effect, since the value returned by a
2448 function is not an lvalue. (But don't warn about the GNU extension of
2449 @code{volatile void} return types. That extension will be warned about
2450 if @option{-pedantic} is specified.)
2453 If @option{-Wall} or @option{-Wunused} is also specified, warn about unused
2457 A comparison between signed and unsigned values could produce an
2458 incorrect result when the signed value is converted to unsigned.
2459 (But don't warn if @option{-Wno-sign-compare} is also specified.)
2462 An aggregate has a partly bracketed initializer.
2463 For example, the following code would evoke such a warning,
2464 because braces are missing around the initializer for @code{x.h}:
2467 struct s @{ int f, g; @};
2468 struct t @{ struct s h; int i; @};
2469 struct t x = @{ 1, 2, 3 @};
2473 An aggregate has an initializer which does not initialize all members.
2474 For example, the following code would cause such a warning, because
2475 @code{x.h} would be implicitly initialized to zero:
2478 struct s @{ int f, g, h; @};
2479 struct s x = @{ 3, 4 @};
2483 @item -Wno-div-by-zero
2484 @opindex Wno-div-by-zero
2485 @opindex Wdiv-by-zero
2486 Do not warn about compile-time integer division by zero. Floating point
2487 division by zero is not warned about, as it can be a legitimate way of
2488 obtaining infinities and NaNs.
2490 @item -Wsystem-headers
2491 @opindex Wsystem-headers
2492 @cindex warnings from system headers
2493 @cindex system headers, warnings from
2494 Print warning messages for constructs found in system header files.
2495 Warnings from system headers are normally suppressed, on the assumption
2496 that they usually do not indicate real problems and would only make the
2497 compiler output harder to read. Using this command line option tells
2498 GCC to emit warnings from system headers as if they occurred in user
2499 code. However, note that using @option{-Wall} in conjunction with this
2500 option will @emph{not} warn about unknown pragmas in system
2501 headers---for that, @option{-Wunknown-pragmas} must also be used.
2504 @opindex Wfloat-equal
2505 Warn if floating point values are used in equality comparisons.
2507 The idea behind this is that sometimes it is convenient (for the
2508 programmer) to consider floating-point values as approximations to
2509 infinitely precise real numbers. If you are doing this, then you need
2510 to compute (by analyzing the code, or in some other way) the maximum or
2511 likely maximum error that the computation introduces, and allow for it
2512 when performing comparisons (and when producing output, but that's a
2513 different problem). In particular, instead of testing for equality, you
2514 would check to see whether the two values have ranges that overlap; and
2515 this is done with the relational operators, so equality comparisons are
2518 @item -Wtraditional @r{(C only)}
2519 @opindex Wtraditional
2520 Warn about certain constructs that behave differently in traditional and
2521 ISO C@. Also warn about ISO C constructs that have no traditional C
2522 equivalent, and/or problematic constructs which should be avoided.
2526 Macro parameters that appear within string literals in the macro body.
2527 In traditional C macro replacement takes place within string literals,
2528 but does not in ISO C@.
2531 In traditional C, some preprocessor directives did not exist.
2532 Traditional preprocessors would only consider a line to be a directive
2533 if the @samp{#} appeared in column 1 on the line. Therefore
2534 @option{-Wtraditional} warns about directives that traditional C
2535 understands but would ignore because the @samp{#} does not appear as the
2536 first character on the line. It also suggests you hide directives like
2537 @samp{#pragma} not understood by traditional C by indenting them. Some
2538 traditional implementations would not recognize @samp{#elif}, so it
2539 suggests avoiding it altogether.
2542 A function-like macro that appears without arguments.
2545 The unary plus operator.
2548 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating point
2549 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
2550 constants.) Note, these suffixes appear in macros defined in the system
2551 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
2552 Use of these macros in user code might normally lead to spurious
2553 warnings, however gcc's integrated preprocessor has enough context to
2554 avoid warning in these cases.
2557 A function declared external in one block and then used after the end of
2561 A @code{switch} statement has an operand of type @code{long}.
2564 A non-@code{static} function declaration follows a @code{static} one.
2565 This construct is not accepted by some traditional C compilers.
2568 The ISO type of an integer constant has a different width or
2569 signedness from its traditional type. This warning is only issued if
2570 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
2571 typically represent bit patterns, are not warned about.
2574 Usage of ISO string concatenation is detected.
2577 Initialization of automatic aggregates.
2580 Identifier conflicts with labels. Traditional C lacks a separate
2581 namespace for labels.
2584 Initialization of unions. If the initializer is zero, the warning is
2585 omitted. This is done under the assumption that the zero initializer in
2586 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
2587 initializer warnings and relies on default initialization to zero in the
2591 Conversions by prototypes between fixed/floating point values and vice
2592 versa. The absence of these prototypes when compiling with traditional
2593 C would cause serious problems. This is a subset of the possible
2594 conversion warnings, for the full set use @option{-Wconversion}.
2597 Use of ISO C style function definitions. This warning intentionally is
2598 @emph{not} issued for prototype declarations or variadic functions
2599 because these ISO C features will appear in your code when using
2600 libiberty's traditional C compatibility macros, @code{PARAMS} and
2601 @code{VPARAMS}. This warning is also bypassed for nested functions
2602 because that feature is already a gcc extension and thus not relevant to
2603 traditional C compatibility.
2608 Warn if an undefined identifier is evaluated in an @samp{#if} directive.
2610 @item -Wendif-labels
2611 @opindex Wendif-labels
2612 Warn whenever an @samp{#else} or an @samp{#endif} are followed by text.
2616 Warn whenever a local variable shadows another local variable, parameter or
2617 global variable or whenever a built-in function is shadowed.
2619 @item -Wlarger-than-@var{len}
2620 @opindex Wlarger-than
2621 Warn whenever an object of larger than @var{len} bytes is defined.
2623 @item -Wpointer-arith
2624 @opindex Wpointer-arith
2625 Warn about anything that depends on the ``size of'' a function type or
2626 of @code{void}. GNU C assigns these types a size of 1, for
2627 convenience in calculations with @code{void *} pointers and pointers
2630 @item -Wbad-function-cast @r{(C only)}
2631 @opindex Wbad-function-cast
2632 Warn whenever a function call is cast to a non-matching type.
2633 For example, warn if @code{int malloc()} is cast to @code{anything *}.
2637 Warn whenever a pointer is cast so as to remove a type qualifier from
2638 the target type. For example, warn if a @code{const char *} is cast
2639 to an ordinary @code{char *}.
2642 @opindex Wcast-align
2643 Warn whenever a pointer is cast such that the required alignment of the
2644 target is increased. For example, warn if a @code{char *} is cast to
2645 an @code{int *} on machines where integers can only be accessed at
2646 two- or four-byte boundaries.
2648 @item -Wwrite-strings
2649 @opindex Wwrite-strings
2650 When compiling C, give string constants the type @code{const
2651 char[@var{length}]} so that
2652 copying the address of one into a non-@code{const} @code{char *}
2653 pointer will get a warning; when compiling C++, warn about the
2654 deprecated conversion from string constants to @code{char *}.
2655 These warnings will help you find at
2656 compile time code that can try to write into a string constant, but
2657 only if you have been very careful about using @code{const} in
2658 declarations and prototypes. Otherwise, it will just be a nuisance;
2659 this is why we did not make @option{-Wall} request these warnings.
2662 @opindex Wconversion
2663 Warn if a prototype causes a type conversion that is different from what
2664 would happen to the same argument in the absence of a prototype. This
2665 includes conversions of fixed point to floating and vice versa, and
2666 conversions changing the width or signedness of a fixed point argument
2667 except when the same as the default promotion.
2669 Also, warn if a negative integer constant expression is implicitly
2670 converted to an unsigned type. For example, warn about the assignment
2671 @code{x = -1} if @code{x} is unsigned. But do not warn about explicit
2672 casts like @code{(unsigned) -1}.
2674 @item -Wsign-compare
2675 @opindex Wsign-compare
2676 @cindex warning for comparison of signed and unsigned values
2677 @cindex comparison of signed and unsigned values, warning
2678 @cindex signed and unsigned values, comparison warning
2679 Warn when a comparison between signed and unsigned values could produce
2680 an incorrect result when the signed value is converted to unsigned.
2681 This warning is also enabled by @option{-W}; to get the other warnings
2682 of @option{-W} without this warning, use @samp{-W -Wno-sign-compare}.
2684 @item -Waggregate-return
2685 @opindex Waggregate-return
2686 Warn if any functions that return structures or unions are defined or
2687 called. (In languages where you can return an array, this also elicits
2690 @item -Wstrict-prototypes @r{(C only)}
2691 @opindex Wstrict-prototypes
2692 Warn if a function is declared or defined without specifying the
2693 argument types. (An old-style function definition is permitted without
2694 a warning if preceded by a declaration which specifies the argument
2697 @item -Wmissing-prototypes @r{(C only)}
2698 @opindex Wmissing-prototypes
2699 Warn if a global function is defined without a previous prototype
2700 declaration. This warning is issued even if the definition itself
2701 provides a prototype. The aim is to detect global functions that fail
2702 to be declared in header files.
2704 @item -Wmissing-declarations
2705 @opindex Wmissing-declarations
2706 Warn if a global function is defined without a previous declaration.
2707 Do so even if the definition itself provides a prototype.
2708 Use this option to detect global functions that are not declared in
2711 @item -Wmissing-noreturn
2712 @opindex Wmissing-noreturn
2713 Warn about functions which might be candidates for attribute @code{noreturn}.
2714 Note these are only possible candidates, not absolute ones. Care should
2715 be taken to manually verify functions actually do not ever return before
2716 adding the @code{noreturn} attribute, otherwise subtle code generation
2717 bugs could be introduced. You will not get a warning for @code{main} in
2718 hosted C environments.
2720 @item -Wmissing-format-attribute
2721 @opindex Wmissing-format-attribute
2723 If @option{-Wformat} is enabled, also warn about functions which might be
2724 candidates for @code{format} attributes. Note these are only possible
2725 candidates, not absolute ones. GCC will guess that @code{format}
2726 attributes might be appropriate for any function that calls a function
2727 like @code{vprintf} or @code{vscanf}, but this might not always be the
2728 case, and some functions for which @code{format} attributes are
2729 appropriate may not be detected. This option has no effect unless
2730 @option{-Wformat} is enabled (possibly by @option{-Wall}).
2732 @item -Wno-multichar
2733 @opindex Wno-multichar
2735 Do not warn if a multicharacter constant (@samp{'FOOF'}) is used.
2736 Usually they indicate a typo in the user's code, as they have
2737 implementation-defined values, and should not be used in portable code.
2739 @item -Wno-deprecated-declarations
2740 @opindex Wno-deprecated-declarations
2741 Do not warn about uses of functions, variables, and types marked as
2742 deprecated by using the @code{deprecated} attribute.
2743 (@pxref{Function Attributes}, @pxref{Variable Attributes},
2744 @pxref{Type Attributes}.)
2748 Warn if a structure is given the packed attribute, but the packed
2749 attribute has no effect on the layout or size of the structure.
2750 Such structures may be mis-aligned for little benefit. For
2751 instance, in this code, the variable @code{f.x} in @code{struct bar}
2752 will be misaligned even though @code{struct bar} does not itself
2753 have the packed attribute:
2760 @} __attribute__((packed));
2770 Warn if padding is included in a structure, either to align an element
2771 of the structure or to align the whole structure. Sometimes when this
2772 happens it is possible to rearrange the fields of the structure to
2773 reduce the padding and so make the structure smaller.
2775 @item -Wredundant-decls
2776 @opindex Wredundant-decls
2777 Warn if anything is declared more than once in the same scope, even in
2778 cases where multiple declaration is valid and changes nothing.
2780 @item -Wnested-externs @r{(C only)}
2781 @opindex Wnested-externs
2782 Warn if an @code{extern} declaration is encountered within a function.
2784 @item -Wunreachable-code
2785 @opindex Wunreachable-code
2786 Warn if the compiler detects that code will never be executed.
2788 This option is intended to warn when the compiler detects that at
2789 least a whole line of source code will never be executed, because
2790 some condition is never satisfied or because it is after a
2791 procedure that never returns.
2793 It is possible for this option to produce a warning even though there
2794 are circumstances under which part of the affected line can be executed,
2795 so care should be taken when removing apparently-unreachable code.
2797 For instance, when a function is inlined, a warning may mean that the
2798 line is unreachable in only one inlined copy of the function.
2800 This option is not made part of @option{-Wall} because in a debugging
2801 version of a program there is often substantial code which checks
2802 correct functioning of the program and is, hopefully, unreachable
2803 because the program does work. Another common use of unreachable
2804 code is to provide behavior which is selectable at compile-time.
2808 Warn if a function can not be inlined and it was declared as inline.
2811 @opindex Winvalid-pch
2812 Warn if a precompiled header (@pxref{Precompiled Headers}) is found in
2813 the search path but can't be used.
2817 @opindex Wno-long-long
2818 Warn if @samp{long long} type is used. This is default. To inhibit
2819 the warning messages, use @option{-Wno-long-long}. Flags
2820 @option{-Wlong-long} and @option{-Wno-long-long} are taken into account
2821 only when @option{-pedantic} flag is used.
2823 @item -Wdisabled-optimization
2824 @opindex Wdisabled-optimization
2825 Warn if a requested optimization pass is disabled. This warning does
2826 not generally indicate that there is anything wrong with your code; it
2827 merely indicates that GCC's optimizers were unable to handle the code
2828 effectively. Often, the problem is that your code is too big or too
2829 complex; GCC will refuse to optimize programs when the optimization
2830 itself is likely to take inordinate amounts of time.
2834 Make all warnings into errors.
2837 @node Debugging Options
2838 @section Options for Debugging Your Program or GCC
2839 @cindex options, debugging
2840 @cindex debugging information options
2842 GCC has various special options that are used for debugging
2843 either your program or GCC:
2848 Produce debugging information in the operating system's native format
2849 (stabs, COFF, XCOFF, or DWARF)@. GDB can work with this debugging
2852 On most systems that use stabs format, @option{-g} enables use of extra
2853 debugging information that only GDB can use; this extra information
2854 makes debugging work better in GDB but will probably make other debuggers
2856 refuse to read the program. If you want to control for certain whether
2857 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
2858 @option{-gxcoff+}, @option{-gxcoff}, @option{-gdwarf-1+}, @option{-gdwarf-1},
2859 or @option{-gvms} (see below).
2861 Unlike most other C compilers, GCC allows you to use @option{-g} with
2862 @option{-O}. The shortcuts taken by optimized code may occasionally
2863 produce surprising results: some variables you declared may not exist
2864 at all; flow of control may briefly move where you did not expect it;
2865 some statements may not be executed because they compute constant
2866 results or their values were already at hand; some statements may
2867 execute in different places because they were moved out of loops.
2869 Nevertheless it proves possible to debug optimized output. This makes
2870 it reasonable to use the optimizer for programs that might have bugs.
2872 The following options are useful when GCC is generated with the
2873 capability for more than one debugging format.
2877 Produce debugging information for use by GDB@. This means to use the
2878 most expressive format available (DWARF 2, stabs, or the native format
2879 if neither of those are supported), including GDB extensions if at all
2884 Produce debugging information in stabs format (if that is supported),
2885 without GDB extensions. This is the format used by DBX on most BSD
2886 systems. On MIPS, Alpha and System V Release 4 systems this option
2887 produces stabs debugging output which is not understood by DBX or SDB@.
2888 On System V Release 4 systems this option requires the GNU assembler.
2892 Produce debugging information in stabs format (if that is supported),
2893 using GNU extensions understood only by the GNU debugger (GDB)@. The
2894 use of these extensions is likely to make other debuggers crash or
2895 refuse to read the program.
2899 Produce debugging information in COFF format (if that is supported).
2900 This is the format used by SDB on most System V systems prior to
2905 Produce debugging information in XCOFF format (if that is supported).
2906 This is the format used by the DBX debugger on IBM RS/6000 systems.
2910 Produce debugging information in XCOFF format (if that is supported),
2911 using GNU extensions understood only by the GNU debugger (GDB)@. The
2912 use of these extensions is likely to make other debuggers crash or
2913 refuse to read the program, and may cause assemblers other than the GNU
2914 assembler (GAS) to fail with an error.
2918 Produce debugging information in DWARF version 1 format (if that is
2919 supported). This is the format used by SDB on most System V Release 4
2922 This option is deprecated.
2926 Produce debugging information in DWARF version 1 format (if that is
2927 supported), using GNU extensions understood only by the GNU debugger
2928 (GDB)@. The use of these extensions is likely to make other debuggers
2929 crash or refuse to read the program.
2931 This option is deprecated.
2935 Produce debugging information in DWARF version 2 format (if that is
2936 supported). This is the format used by DBX on IRIX 6.
2940 Produce debugging information in VMS debug format (if that is
2941 supported). This is the format used by DEBUG on VMS systems.
2944 @itemx -ggdb@var{level}
2945 @itemx -gstabs@var{level}
2946 @itemx -gcoff@var{level}
2947 @itemx -gxcoff@var{level}
2948 @itemx -gvms@var{level}
2949 Request debugging information and also use @var{level} to specify how
2950 much information. The default level is 2.
2952 Level 1 produces minimal information, enough for making backtraces in
2953 parts of the program that you don't plan to debug. This includes
2954 descriptions of functions and external variables, but no information
2955 about local variables and no line numbers.
2957 Level 3 includes extra information, such as all the macro definitions
2958 present in the program. Some debuggers support macro expansion when
2959 you use @option{-g3}.
2961 Note that in order to avoid confusion between DWARF1 debug level 2,
2962 and DWARF2, neither @option{-gdwarf} nor @option{-gdwarf-2} accept
2963 a concatenated debug level. Instead use an additional @option{-g@var{level}}
2964 option to change the debug level for DWARF1 or DWARF2.
2966 @item -feliminate-dwarf2-dups
2967 @opindex feliminate-dwarf2-dups
2968 Compress DWARF2 debugging information by eliminating duplicated
2969 information about each symbol. This option only makes sense when
2970 generating DWARF2 debugging information with @option{-gdwarf-2}.
2975 Generate extra code to write profile information suitable for the
2976 analysis program @code{prof}. You must use this option when compiling
2977 the source files you want data about, and you must also use it when
2980 @cindex @code{gprof}
2983 Generate extra code to write profile information suitable for the
2984 analysis program @code{gprof}. You must use this option when compiling
2985 the source files you want data about, and you must also use it when
2990 Makes the compiler print out each function name as it is compiled, and
2991 print some statistics about each pass when it finishes.
2994 @opindex ftime-report
2995 Makes the compiler print some statistics about the time consumed by each
2996 pass when it finishes.
2999 @opindex fmem-report
3000 Makes the compiler print some statistics about permanent memory
3001 allocation when it finishes.
3003 @item -fprofile-arcs
3004 @opindex fprofile-arcs
3005 Instrument @dfn{arcs} during compilation to generate coverage data or
3006 for profile-directed block ordering. During execution the program
3007 records how many times each branch is executed and how many times it is
3008 taken. When the compiled program exits it saves this data to a file
3009 called @file{@var{auxname}.da} for each source file. @var{auxname} is
3010 generated from the name of the output file, if explicitly specified and
3011 it is not the final executable, otherwise it is the basename of the
3012 source file. In both cases any suffix is removed (e.g. @file{foo.da}
3013 for input file @file{dir/foo.c}, or @file{dir/foo.da} for output file
3014 specified as @option{-o dir/foo.o}).
3016 For profile-directed block ordering, compile the program with
3017 @option{-fprofile-arcs} plus optimization and code generation options,
3018 generate the arc profile information by running the program on a
3019 selected workload, and then compile the program again with the same
3020 optimization and code generation options plus
3021 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
3022 Control Optimization}).
3024 The other use of @option{-fprofile-arcs} is for use with @code{gcov},
3025 when it is used with the @option{-ftest-coverage} option.
3027 With @option{-fprofile-arcs}, for each function of your program GCC
3028 creates a program flow graph, then finds a spanning tree for the graph.
3029 Only arcs that are not on the spanning tree have to be instrumented: the
3030 compiler adds code to count the number of times that these arcs are
3031 executed. When an arc is the only exit or only entrance to a block, the
3032 instrumentation code can be added to the block; otherwise, a new basic
3033 block must be created to hold the instrumentation code.
3036 @item -ftest-coverage
3037 @opindex ftest-coverage
3038 Create data files for the @code{gcov} code-coverage utility
3039 (@pxref{Gcov,, @code{gcov}: a GCC Test Coverage Program}). See
3040 @option{-fprofile-arcs} option above for a description of @var{auxname}.
3043 @item @var{auxname}.bb
3044 A mapping from basic blocks to line numbers, which @code{gcov} uses to
3045 associate basic block execution counts with line numbers.
3047 @item @var{auxname}.bbg
3048 A list of all arcs in the program flow graph. This allows @code{gcov}
3049 to reconstruct the program flow graph, so that it can compute all basic
3050 block and arc execution counts from the information in the
3051 @file{@var{auxname}.da} file.
3054 Use @option{-ftest-coverage} with @option{-fprofile-arcs}; the latter
3055 option adds instrumentation to the program, which then writes
3056 execution counts to another data file:
3059 @item @var{auxname}.da
3060 Runtime arc execution counts, used in conjunction with the arc
3061 information in the file @file{@var{auxname}.bbg}.
3064 Coverage data will map better to the source files if
3065 @option{-ftest-coverage} is used without optimization.
3067 @item -d@var{letters}
3069 Says to make debugging dumps during compilation at times specified by
3070 @var{letters}. This is used for debugging the compiler. The file names
3071 for most of the dumps are made by appending a pass number and a word to
3072 the @var{dumpname}. @var{dumpname} is generated from the name of the
3073 output file, if explicitly specified and it is not an executable,
3074 otherwise it is the basename of the source file. In both cases any
3075 suffix is removed (e.g. @file{foo.00.rtl} or @file{foo.01.sibling}).
3076 Here are the possible letters for use in @var{letters}, and their
3082 Annotate the assembler output with miscellaneous debugging information.
3085 Dump after computing branch probabilities, to @file{@var{file}.14.bp}.
3088 Dump after block reordering, to @file{@var{file}.32.bbro}.
3091 Dump after instruction combination, to the file @file{@var{file}.19.combine}.
3094 Dump after the first if conversion, to the file @file{@var{file}.15.ce1}.
3097 Dump after delayed branch scheduling, to @file{@var{file}.34.dbr}.
3100 Dump all macro definitions, at the end of preprocessing, in addition to
3104 Dump after SSA optimizations, to @file{@var{file}.04.ssa} and
3105 @file{@var{file}.07.ussa}.
3108 Dump after the second if conversion, to @file{@var{file}.29.ce3}.
3111 Dump after life analysis, to @file{@var{file}.18.life}.
3114 Dump after purging @code{ADDRESSOF} codes, to @file{@var{file}.10.addressof}.
3117 Dump after global register allocation, to @file{@var{file}.24.greg}.
3120 Dump after finalization of EH handling code, to @file{@var{file}.02.eh}.
3123 Dump after reg-to-stack conversion, to @file{@var{file}.31.stack}.
3126 Dump after post-reload optimizations, to @file{@var{file}.25.postreload}.
3129 Dump after GCSE, to @file{@var{file}.11.gcse}.
3132 Dump after sibling call optimizations, to @file{@var{file}.01.sibling}.
3135 Dump after the first jump optimization, to @file{@var{file}.03.jump}.
3138 Dump after conversion from registers to stack, to @file{@var{file}.31.stack}.
3141 Dump after local register allocation, to @file{@var{file}.23.lreg}.
3144 Dump after loop optimization, to @file{@var{file}.12.loop}.
3147 Dump after performing the machine dependent reorganization pass, to
3148 @file{@var{file}.33.mach}.
3151 Dump after register renumbering, to @file{@var{file}.28.rnreg}.
3154 Dump after the register move pass, to @file{@var{file}.21.regmove}.
3157 Dump after RTL generation, to @file{@var{file}.00.rtl}.
3160 Dump after the second scheduling pass, to @file{@var{file}.30.sched2}.
3163 Dump after CSE (including the jump optimization that sometimes follows
3164 CSE), to @file{@var{file}.09.cse}.
3167 Dump after the first scheduling pass, to @file{@var{file}.22.sched}.
3170 Dump after the second CSE pass (including the jump optimization that
3171 sometimes follows CSE), to @file{@var{file}.17.cse2}.
3174 Dump after null pointer elimination pass to @file{@var{file}.08.null}.
3177 Dump after the second flow pass, to @file{@var{file}.26.flow2}.
3180 Dump after SSA dead code elimination, to @file{@var{file}.06.ssadce}.
3183 Dump after the peephole pass, to @file{@var{file}.27.peephole2}.
3186 Produce all the dumps listed above.
3189 Print statistics on memory usage, at the end of the run, to
3193 Annotate the assembler output with a comment indicating which
3194 pattern and alternative was used. The length of each instruction is
3198 Dump the RTL in the assembler output as a comment before each instruction.
3199 Also turns on @option{-dp} annotation.
3202 For each of the other indicated dump files (except for
3203 @file{@var{file}.00.rtl}), dump a representation of the control flow graph
3204 suitable for viewing with VCG to @file{@var{file}.@var{pass}.vcg}.
3207 Just generate RTL for a function instead of compiling it. Usually used
3211 Dump debugging information during parsing, to standard error.
3214 @item -fdump-unnumbered
3215 @opindex fdump-unnumbered
3216 When doing debugging dumps (see @option{-d} option above), suppress instruction
3217 numbers and line number note output. This makes it more feasible to
3218 use diff on debugging dumps for compiler invocations with different
3219 options, in particular with and without @option{-g}.
3221 @item -fdump-translation-unit @r{(C and C++ only)}
3222 @itemx -fdump-translation-unit-@var{options} @r{(C and C++ only)}
3223 @opindex fdump-translation-unit
3224 Dump a representation of the tree structure for the entire translation
3225 unit to a file. The file name is made by appending @file{.tu} to the
3226 source file name. If the @samp{-@var{options}} form is used, @var{options}
3227 controls the details of the dump as described for the
3228 @option{-fdump-tree} options.
3230 @item -fdump-class-hierarchy @r{(C++ only)}
3231 @itemx -fdump-class-hierarchy-@var{options} @r{(C++ only)}
3232 @opindex fdump-class-hierarchy
3233 Dump a representation of each class's hierarchy and virtual function
3234 table layout to a file. The file name is made by appending @file{.class}
3235 to the source file name. If the @samp{-@var{options}} form is used,
3236 @var{options} controls the details of the dump as described for the
3237 @option{-fdump-tree} options.
3239 @item -fdump-tree-@var{switch} @r{(C++ only)}
3240 @itemx -fdump-tree-@var{switch}-@var{options} @r{(C++ only)}
3242 Control the dumping at various stages of processing the intermediate
3243 language tree to a file. The file name is generated by appending a switch
3244 specific suffix to the source file name. If the @samp{-@var{options}}
3245 form is used, @var{options} is a list of @samp{-} separated options that
3246 control the details of the dump. Not all options are applicable to all
3247 dumps, those which are not meaningful will be ignored. The following
3248 options are available
3252 Print the address of each node. Usually this is not meaningful as it
3253 changes according to the environment and source file. Its primary use
3254 is for tying up a dump file with a debug environment.
3256 Inhibit dumping of members of a scope or body of a function merely
3257 because that scope has been reached. Only dump such items when they
3258 are directly reachable by some other path.
3260 Turn on all options.
3263 The following tree dumps are possible:
3266 Dump before any tree based optimization, to @file{@var{file}.original}.
3268 Dump after all tree based optimization, to @file{@var{file}.optimized}.
3270 Dump after function inlining, to @file{@var{file}.inlined}.
3273 @item -fsched-verbose=@var{n}
3274 @opindex fsched-verbose
3275 On targets that use instruction scheduling, this option controls the
3276 amount of debugging output the scheduler prints. This information is
3277 written to standard error, unless @option{-dS} or @option{-dR} is
3278 specified, in which case it is output to the usual dump
3279 listing file, @file{.sched} or @file{.sched2} respectively. However
3280 for @var{n} greater than nine, the output is always printed to standard
3283 For @var{n} greater than zero, @option{-fsched-verbose} outputs the
3284 same information as @option{-dRS}. For @var{n} greater than one, it
3285 also output basic block probabilities, detailed ready list information
3286 and unit/insn info. For @var{n} greater than two, it includes RTL
3287 at abort point, control-flow and regions info. And for @var{n} over
3288 four, @option{-fsched-verbose} also includes dependence info.
3292 Store the usual ``temporary'' intermediate files permanently; place them
3293 in the current directory and name them based on the source file. Thus,
3294 compiling @file{foo.c} with @samp{-c -save-temps} would produce files
3295 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
3296 preprocessed @file{foo.i} output file even though the compiler now
3297 normally uses an integrated preprocessor.
3301 Report the CPU time taken by each subprocess in the compilation
3302 sequence. For C source files, this is the compiler proper and assembler
3303 (plus the linker if linking is done). The output looks like this:
3310 The first number on each line is the ``user time,'' that is time spent
3311 executing the program itself. The second number is ``system time,''
3312 time spent executing operating system routines on behalf of the program.
3313 Both numbers are in seconds.
3315 @item -print-file-name=@var{library}
3316 @opindex print-file-name
3317 Print the full absolute name of the library file @var{library} that
3318 would be used when linking---and don't do anything else. With this
3319 option, GCC does not compile or link anything; it just prints the
3322 @item -print-multi-directory
3323 @opindex print-multi-directory
3324 Print the directory name corresponding to the multilib selected by any
3325 other switches present in the command line. This directory is supposed
3326 to exist in @env{GCC_EXEC_PREFIX}.
3328 @item -print-multi-lib
3329 @opindex print-multi-lib
3330 Print the mapping from multilib directory names to compiler switches
3331 that enable them. The directory name is separated from the switches by
3332 @samp{;}, and each switch starts with an @samp{@@} instead of the
3333 @samp{-}, without spaces between multiple switches. This is supposed to
3334 ease shell-processing.
3336 @item -print-prog-name=@var{program}
3337 @opindex print-prog-name
3338 Like @option{-print-file-name}, but searches for a program such as @samp{cpp}.
3340 @item -print-libgcc-file-name
3341 @opindex print-libgcc-file-name
3342 Same as @option{-print-file-name=libgcc.a}.
3344 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
3345 but you do want to link with @file{libgcc.a}. You can do
3348 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
3351 @item -print-search-dirs
3352 @opindex print-search-dirs
3353 Print the name of the configured installation directory and a list of
3354 program and library directories gcc will search---and don't do anything else.
3356 This is useful when gcc prints the error message
3357 @samp{installation problem, cannot exec cpp0: No such file or directory}.
3358 To resolve this you either need to put @file{cpp0} and the other compiler
3359 components where gcc expects to find them, or you can set the environment
3360 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
3361 Don't forget the trailing '/'.
3362 @xref{Environment Variables}.
3365 @opindex dumpmachine
3366 Print the compiler's target machine (for example,
3367 @samp{i686-pc-linux-gnu})---and don't do anything else.
3370 @opindex dumpversion
3371 Print the compiler version (for example, @samp{3.0})---and don't do
3376 Print the compiler's built-in specs---and don't do anything else. (This
3377 is used when GCC itself is being built.) @xref{Spec Files}.
3380 @node Optimize Options
3381 @section Options That Control Optimization
3382 @cindex optimize options
3383 @cindex options, optimization
3385 These options control various sorts of optimizations.
3387 Without any optimization option, the compiler's goal is to reduce the
3388 cost of compilation and to make debugging produce the expected
3389 results. Statements are independent: if you stop the program with a
3390 breakpoint between statements, you can then assign a new value to any
3391 variable or change the program counter to any other statement in the
3392 function and get exactly the results you would expect from the source
3395 Turning on optimization flags makes the compiler attempt to improve
3396 the performance and/or code size at the expense of compilation time
3397 and possibly the ability to debug the program.
3399 Not all optimizations are controlled directly by a flag. Only
3400 optimizations that have a flag are listed.
3407 Optimize. Optimizing compilation takes somewhat more time, and a lot
3408 more memory for a large function.
3410 With @option{-O}, the compiler tries to reduce code size and execution
3411 time, without performing any optimizations that take a great deal of
3414 @option{-O} turns on the following optimization flags:
3415 @gccoptlist{-fdefer-pop
3423 -fguess-branch-probability
3426 @option{-O} also turns on @option{-fomit-frame-pointer} on machines
3427 where doing so does not interfere with debugging.
3431 Optimize even more. GCC performs nearly all supported optimizations
3432 that do not involve a space-speed tradeoff. The compiler does not
3433 perform loop unrolling or function inlining when you specify @option{-O2}.
3434 As compared to @option{-O}, this option increases both compilation time
3435 and the performance of the generated code.
3437 @option{-O2} turns on all optimization flags specified by @option{-O}. It
3438 also turns on the following optimization flags:
3439 @gccoptlist{-fforce-mem
3440 -foptimize-sibling-calls
3442 -fcse-follow-jumps -fcse-skip-blocks
3443 -frerun-cse-after-loop -frerun-loop-opt
3444 -fgcse -fgcse-lm -fgcse-sm
3445 -fdelete-null-pointer-checks
3446 -fexpensive-optimizations
3448 -fschedule-insns -fschedule-insns2
3449 -fsched-interblock -fsched-spec
3452 -freorder-blocks -freorder-functions
3454 -falign-functions -falign-jumps
3455 -falign-loops -falign-labels}
3457 Please note the warning under @option{-fgcse} about
3458 invoking @option{-O2} on programs that use computed gotos.
3462 Optimize yet more. @option{-O3} turns on all optimizations specified by
3463 @option{-O2} and also turns on the @option{-finline-functions} and
3464 @option{-frename-registers} options.
3468 Do not optimize. This is the default.
3472 Optimize for size. @option{-Os} enables all @option{-O2} optimizations that
3473 do not typically increase code size. It also performs further
3474 optimizations designed to reduce code size.
3476 @option{-Os} disables the following optimization flags:
3477 @gccoptlist{-falign-functions -falign-jumps -falign-loops
3478 -falign-labels -freorder-blocks -fprefetch-loop-arrays}
3480 If you use multiple @option{-O} options, with or without level numbers,
3481 the last such option is the one that is effective.
3484 Options of the form @option{-f@var{flag}} specify machine-independent
3485 flags. Most flags have both positive and negative forms; the negative
3486 form of @option{-ffoo} would be @option{-fno-foo}. In the table
3487 below, only one of the forms is listed---the one you typically will
3488 use. You can figure out the other form by either removing @samp{no-}
3491 The following options control specific optimizations. They are either
3492 activated by @option{-O} options or are related to ones that are. You
3493 can use the following flags in the rare cases when ``fine-tuning'' of
3494 optimizations to be performed is desired.
3498 @opindex ffloat-store
3499 Do not store floating point variables in registers, and inhibit other
3500 options that might change whether a floating point value is taken from a
3503 @cindex floating point precision
3504 This option prevents undesirable excess precision on machines such as
3505 the 68000 where the floating registers (of the 68881) keep more
3506 precision than a @code{double} is supposed to have. Similarly for the
3507 x86 architecture. For most programs, the excess precision does only
3508 good, but a few programs rely on the precise definition of IEEE floating
3509 point. Use @option{-ffloat-store} for such programs, after modifying
3510 them to store all pertinent intermediate computations into variables.
3512 @item -fno-default-inline
3513 @opindex fno-default-inline
3514 Do not make member functions inline by default merely because they are
3515 defined inside the class scope (C++ only). Otherwise, when you specify
3516 @w{@option{-O}}, member functions defined inside class scope are compiled
3517 inline by default; i.e., you don't need to add @samp{inline} in front of
3518 the member function name.
3520 @item -fno-defer-pop
3521 @opindex fno-defer-pop
3522 Always pop the arguments to each function call as soon as that function
3523 returns. For machines which must pop arguments after a function call,
3524 the compiler normally lets arguments accumulate on the stack for several
3525 function calls and pops them all at once.
3529 Force memory operands to be copied into registers before doing
3530 arithmetic on them. This produces better code by making all memory
3531 references potential common subexpressions. When they are not common
3532 subexpressions, instruction combination should eliminate the separate
3533 register-load. The @option{-O2} option turns on this option.
3536 @opindex fforce-addr
3537 Force memory address constants to be copied into registers before
3538 doing arithmetic on them. This may produce better code just as
3539 @option{-fforce-mem} may.
3541 @item -fomit-frame-pointer
3542 @opindex fomit-frame-pointer
3543 Don't keep the frame pointer in a register for functions that
3544 don't need one. This avoids the instructions to save, set up and
3545 restore frame pointers; it also makes an extra register available
3546 in many functions. @strong{It also makes debugging impossible on
3549 On some machines, such as the VAX, this flag has no effect, because
3550 the standard calling sequence automatically handles the frame pointer
3551 and nothing is saved by pretending it doesn't exist. The
3552 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
3553 whether a target machine supports this flag. @xref{Registers,,Register
3554 Usage, gccint, GNU Compiler Collection (GCC) Internals}.
3556 @item -foptimize-sibling-calls
3557 @opindex foptimize-sibling-calls
3558 Optimize sibling and tail recursive calls.
3562 This option generates traps for signed overflow on addition, subtraction,
3563 multiplication operations.
3567 Don't pay attention to the @code{inline} keyword. Normally this option
3568 is used to keep the compiler from expanding any functions inline.
3569 Note that if you are not optimizing, no functions can be expanded inline.
3571 @item -finline-functions
3572 @opindex finline-functions
3573 Integrate all simple functions into their callers. The compiler
3574 heuristically decides which functions are simple enough to be worth
3575 integrating in this way.
3577 If all calls to a given function are integrated, and the function is
3578 declared @code{static}, then the function is normally not output as
3579 assembler code in its own right.
3581 @item -finline-limit=@var{n}
3582 @opindex finline-limit
3583 By default, gcc limits the size of functions that can be inlined. This flag
3584 allows the control of this limit for functions that are explicitly marked as
3585 inline (i.e., marked with the inline keyword or defined within the class
3586 definition in c++). @var{n} is the size of functions that can be inlined in
3587 number of pseudo instructions (not counting parameter handling). The default
3588 value of @var{n} is 600.
3589 Increasing this value can result in more inlined code at
3590 the cost of compilation time and memory consumption. Decreasing usually makes
3591 the compilation faster and less code will be inlined (which presumably
3592 means slower programs). This option is particularly useful for programs that
3593 use inlining heavily such as those based on recursive templates with C++.
3595 @emph{Note:} pseudo instruction represents, in this particular context, an
3596 abstract measurement of function's size. In no way, it represents a count
3597 of assembly instructions and as such its exact meaning might change from one
3598 release to an another.
3600 @item -fkeep-inline-functions
3601 @opindex fkeep-inline-functions
3602 Even if all calls to a given function are integrated, and the function
3603 is declared @code{static}, nevertheless output a separate run-time
3604 callable version of the function. This switch does not affect
3605 @code{extern inline} functions.
3607 @item -fkeep-static-consts
3608 @opindex fkeep-static-consts
3609 Emit variables declared @code{static const} when optimization isn't turned
3610 on, even if the variables aren't referenced.
3612 GCC enables this option by default. If you want to force the compiler to
3613 check if the variable was referenced, regardless of whether or not
3614 optimization is turned on, use the @option{-fno-keep-static-consts} option.
3616 @item -fmerge-constants
3617 Attempt to merge identical constants (string constants and floating point
3618 constants) across compilation units.
3620 This option is the default for optimized compilation if the assembler and
3621 linker support it. Use @option{-fno-merge-constants} to inhibit this
3624 @item -fmerge-all-constants
3625 Attempt to merge identical constants and identical variables.
3627 This option implies @option{-fmerge-constants}. In addition to
3628 @option{-fmerge-constants} this considers e.g. even constant initialized
3629 arrays or initialized constant variables with integral or floating point
3630 types. Languages like C or C++ require each non-automatic variable to
3631 have distinct location, so using this option will result in non-conforming
3636 Use a graph coloring register allocator. Currently this option is meant
3637 for testing, so we are interested to hear about miscompilations with
3640 @item -fno-branch-count-reg
3641 @opindex fno-branch-count-reg
3642 Do not use ``decrement and branch'' instructions on a count register,
3643 but instead generate a sequence of instructions that decrement a
3644 register, compare it against zero, then branch based upon the result.
3645 This option is only meaningful on architectures that support such
3646 instructions, which include x86, PowerPC, IA-64 and S/390.
3648 @item -fno-function-cse
3649 @opindex fno-function-cse
3650 Do not put function addresses in registers; make each instruction that
3651 calls a constant function contain the function's address explicitly.
3653 This option results in less efficient code, but some strange hacks
3654 that alter the assembler output may be confused by the optimizations
3655 performed when this option is not used.
3659 Sets @option{-fno-math-errno}, @option{-funsafe-math-optimizations}, @*
3660 @option{-fno-trapping-math}, @option{-ffinite-math-only} and @*
3661 @option{-fno-signaling-nans}.
3663 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
3665 This option should never be turned on by any @option{-O} option since
3666 it can result in incorrect output for programs which depend on
3667 an exact implementation of IEEE or ISO rules/specifications for
3670 @item -fno-math-errno
3671 @opindex fno-math-errno
3672 Do not set ERRNO after calling math functions that are executed
3673 with a single instruction, e.g., sqrt. A program that relies on
3674 IEEE exceptions for math error handling may want to use this flag
3675 for speed while maintaining IEEE arithmetic compatibility.
3677 This option should never be turned on by any @option{-O} option since
3678 it can result in incorrect output for programs which depend on
3679 an exact implementation of IEEE or ISO rules/specifications for
3682 The default is @option{-fmath-errno}.
3684 @item -funsafe-math-optimizations
3685 @opindex funsafe-math-optimizations
3686 Allow optimizations for floating-point arithmetic that (a) assume
3687 that arguments and results are valid and (b) may violate IEEE or
3688 ANSI standards. When used at link-time, it may include libraries
3689 or startup files that change the default FPU control word or other
3690 similar optimizations.
3692 This option should never be turned on by any @option{-O} option since
3693 it can result in incorrect output for programs which depend on
3694 an exact implementation of IEEE or ISO rules/specifications for
3697 The default is @option{-fno-unsafe-math-optimizations}.
3699 @item -ffinite-math-only
3700 @opindex ffinite-math-only
3701 Allow optimizations for floating-point arithmetic that assume
3702 that arguments and results are not NaNs or +-Infs.
3704 This option should never be turned on by any @option{-O} option since
3705 it can result in incorrect output for programs which depend on
3706 an exact implementation of IEEE or ISO rules/specifications.
3708 The default is @option{-fno-finite-math-only}.
3710 @item -fno-trapping-math
3711 @opindex fno-trapping-math
3712 Compile code assuming that floating-point operations cannot generate
3713 user-visible traps. These traps include division by zero, overflow,
3714 underflow, inexact result and invalid operation. This option implies
3715 @option{-fno-signaling-nans}. Setting this option may allow faster
3716 code if one relies on ``non-stop'' IEEE arithmetic, for example.
3718 This option should never be turned on by any @option{-O} option since
3719 it can result in incorrect output for programs which depend on
3720 an exact implementation of IEEE or ISO rules/specifications for
3723 The default is @option{-ftrapping-math}.
3725 @item -fsignaling-nans
3726 @opindex fsignaling-nans
3727 Compile code assuming that IEEE signaling NaNs may generate user-visible
3728 traps during floating-point operations. Setting this option disables
3729 optimizations that may change the number of exceptions visible with
3730 signaling NaNs. This option implies @option{-ftrapping-math}.
3732 This option causes the preprocessor macro @code{__SUPPORT_SNAN__} to
3735 The default is @option{-fno-signaling-nans}.
3737 This option is experimental and does not currently guarantee to
3738 disable all GCC optimizations that affect signaling NaN behavior.
3740 @item -fno-zero-initialized-in-bss
3741 @opindex fno-zero-initialized-in-bss
3742 If the target supports a BSS section, GCC by default puts variables that
3743 are initialized to zero into BSS@. This can save space in the resulting
3746 This option turns off this behavior because some programs explicitly
3747 rely on variables going to the data section. E.g., so that the
3748 resulting executable can find the beginning of that section and/or make
3749 assumptions based on that.
3751 The default is @option{-fzero-initialized-in-bss}.
3753 @item -fbounds-check
3754 @opindex fbounds-check
3755 For front-ends that support it, generate additional code to check that
3756 indices used to access arrays are within the declared range. This is
3757 currently only supported by the Java and Fortran 77 front-ends, where
3758 this option defaults to true and false respectively.
3760 @item -fstrength-reduce
3761 @opindex fstrength-reduce
3762 Perform the optimizations of loop strength reduction and
3763 elimination of iteration variables.
3765 @item -fthread-jumps
3766 @opindex fthread-jumps
3767 Perform optimizations where we check to see if a jump branches to a
3768 location where another comparison subsumed by the first is found. If
3769 so, the first branch is redirected to either the destination of the
3770 second branch or a point immediately following it, depending on whether
3771 the condition is known to be true or false.
3773 @item -fcse-follow-jumps
3774 @opindex fcse-follow-jumps
3775 In common subexpression elimination, scan through jump instructions
3776 when the target of the jump is not reached by any other path. For
3777 example, when CSE encounters an @code{if} statement with an
3778 @code{else} clause, CSE will follow the jump when the condition
3781 @item -fcse-skip-blocks
3782 @opindex fcse-skip-blocks
3783 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
3784 follow jumps which conditionally skip over blocks. When CSE
3785 encounters a simple @code{if} statement with no else clause,
3786 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
3787 body of the @code{if}.
3789 @item -frerun-cse-after-loop
3790 @opindex frerun-cse-after-loop
3791 Re-run common subexpression elimination after loop optimizations has been
3794 @item -frerun-loop-opt
3795 @opindex frerun-loop-opt
3796 Run the loop optimizer twice.
3800 Perform a global common subexpression elimination pass.
3801 This pass also performs global constant and copy propagation.
3803 @emph{Note:} When compiling a program using computed gotos, a GCC
3804 extension, you may get better runtime performance if you disable
3805 the global common subexpression elimination pass by adding
3806 @option{-fno-gcse} to the command line.
3810 When @option{-fgcse-lm} is enabled, global common subexpression elimination will
3811 attempt to move loads which are only killed by stores into themselves. This
3812 allows a loop containing a load/store sequence to be changed to a load outside
3813 the loop, and a copy/store within the loop.
3817 When @option{-fgcse-sm} is enabled, A store motion pass is run after global common
3818 subexpression elimination. This pass will attempt to move stores out of loops.
3819 When used in conjunction with @option{-fgcse-lm}, loops containing a load/store sequence
3820 can be changed to a load before the loop and a store after the loop.
3822 @item -floop-optimize
3823 @opindex floop-optimize
3824 Perform loop optimizations: move constant expressions out of loops, simplify
3825 exit test conditions and optionally do strength-reduction and loop unrolling as
3828 @item -fcrossjumping
3829 @opindex crossjumping
3830 Perform cross-jumping transformation. This transformation unifies equivalent code and save code size. The
3831 resulting code may or may not perform better than without cross-jumping.
3833 @item -fif-conversion
3834 @opindex if-conversion
3835 Attempt to transform conditional jumps into branch-less equivalents. This
3836 include use of conditional moves, min, max, set flags and abs instructions, and
3837 some tricks doable by standard arithmetics. The use of conditional execution
3838 on chips where it is available is controlled by @code{if-conversion2}.
3840 @item -fif-conversion2
3841 @opindex if-conversion2
3842 Use conditional execution (where available) to transform conditional jumps into
3843 branch-less equivalents.
3845 @item -fdelete-null-pointer-checks
3846 @opindex fdelete-null-pointer-checks
3847 Use global dataflow analysis to identify and eliminate useless checks
3848 for null pointers. The compiler assumes that dereferencing a null
3849 pointer would have halted the program. If a pointer is checked after
3850 it has already been dereferenced, it cannot be null.
3852 In some environments, this assumption is not true, and programs can
3853 safely dereference null pointers. Use
3854 @option{-fno-delete-null-pointer-checks} to disable this optimization
3855 for programs which depend on that behavior.
3857 @item -fexpensive-optimizations
3858 @opindex fexpensive-optimizations
3859 Perform a number of minor optimizations that are relatively expensive.
3861 @item -foptimize-register-move
3863 @opindex foptimize-register-move
3865 Attempt to reassign register numbers in move instructions and as
3866 operands of other simple instructions in order to maximize the amount of
3867 register tying. This is especially helpful on machines with two-operand
3868 instructions. GCC enables this optimization by default with @option{-O2}
3871 Note @option{-fregmove} and @option{-foptimize-register-move} are the same
3874 @item -fdelayed-branch
3875 @opindex fdelayed-branch
3876 If supported for the target machine, attempt to reorder instructions
3877 to exploit instruction slots available after delayed branch
3880 @item -fschedule-insns
3881 @opindex fschedule-insns
3882 If supported for the target machine, attempt to reorder instructions to
3883 eliminate execution stalls due to required data being unavailable. This
3884 helps machines that have slow floating point or memory load instructions
3885 by allowing other instructions to be issued until the result of the load
3886 or floating point instruction is required.
3888 @item -fschedule-insns2
3889 @opindex fschedule-insns2
3890 Similar to @option{-fschedule-insns}, but requests an additional pass of
3891 instruction scheduling after register allocation has been done. This is
3892 especially useful on machines with a relatively small number of
3893 registers and where memory load instructions take more than one cycle.
3895 @item -fno-sched-interblock
3896 @opindex fno-sched-interblock
3897 Don't schedule instructions across basic blocks. This is normally
3898 enabled by default when scheduling before register allocation, i.e.@:
3899 with @option{-fschedule-insns} or at @option{-O2} or higher.
3901 @item -fno-sched-spec
3902 @opindex fno-sched-spec
3903 Don't allow speculative motion of non-load instructions. This is normally
3904 enabled by default when scheduling before register allocation, i.e.@:
3905 with @option{-fschedule-insns} or at @option{-O2} or higher.
3907 @item -fsched-spec-load
3908 @opindex fsched-spec-load
3909 Allow speculative motion of some load instructions. This only makes
3910 sense when scheduling before register allocation, i.e.@: with
3911 @option{-fschedule-insns} or at @option{-O2} or higher.
3913 @item -fsched-spec-load-dangerous
3914 @opindex fsched-spec-load-dangerous
3915 Allow speculative motion of more load instructions. This only makes
3916 sense when scheduling before register allocation, i.e.@: with
3917 @option{-fschedule-insns} or at @option{-O2} or higher.
3919 @item -ffunction-sections
3920 @itemx -fdata-sections
3921 @opindex ffunction-sections
3922 @opindex fdata-sections
3923 Place each function or data item into its own section in the output
3924 file if the target supports arbitrary sections. The name of the
3925 function or the name of the data item determines the section's name
3928 Use these options on systems where the linker can perform optimizations
3929 to improve locality of reference in the instruction space. HPPA
3930 processors running HP-UX and SPARC processors running Solaris 2 have
3931 linkers with such optimizations. Other systems using the ELF object format
3932 as well as AIX may have these optimizations in the future.
3934 Only use these options when there are significant benefits from doing
3935 so. When you specify these options, the assembler and linker will
3936 create larger object and executable files and will also be slower.
3937 You will not be able to use @code{gprof} on all systems if you
3938 specify this option and you may have problems with debugging if
3939 you specify both this option and @option{-g}.
3941 @item -fcaller-saves
3942 @opindex fcaller-saves
3943 Enable values to be allocated in registers that will be clobbered by
3944 function calls, by emitting extra instructions to save and restore the
3945 registers around such calls. Such allocation is done only when it
3946 seems to result in better code than would otherwise be produced.
3948 This option is always enabled by default on certain machines, usually
3949 those which have no call-preserved registers to use instead.
3951 For all machines, optimization level 2 and higher enables this flag by
3956 Perform tail duplication to enlarge superblock size. This transformation
3957 simplifies the control flow of the function allowing other optimizations to do
3960 @item -funroll-loops
3961 @opindex funroll-loops
3962 Unroll loops whose number of iterations can be determined at compile
3963 time or upon entry to the loop. @option{-funroll-loops} implies both
3964 @option{-fstrength-reduce} and @option{-frerun-cse-after-loop}. This
3965 option makes code larger, and may or may not make it run faster.
3967 @item -funroll-all-loops
3968 @opindex funroll-all-loops
3969 Unroll all loops, even if their number of iterations is uncertain when
3970 the loop is entered. This usually makes programs run more slowly.
3971 @option{-funroll-all-loops} implies the same options as
3972 @option{-funroll-loops},
3974 @item -fprefetch-loop-arrays
3975 @opindex fprefetch-loop-arrays
3976 If supported by the target machine, generate instructions to prefetch
3977 memory to improve the performance of loops that access large arrays.
3979 @item -fmove-all-movables
3980 @opindex fmove-all-movables
3981 Forces all invariant computations in loops to be moved
3984 @item -freduce-all-givs
3985 @opindex freduce-all-givs
3986 Forces all general-induction variables in loops to be
3989 @emph{Note:} When compiling programs written in Fortran,
3990 @option{-fmove-all-movables} and @option{-freduce-all-givs} are enabled
3991 by default when you use the optimizer.
3993 These options may generate better or worse code; results are highly
3994 dependent on the structure of loops within the source code.
3996 These two options are intended to be removed someday, once
3997 they have helped determine the efficacy of various
3998 approaches to improving loop optimizations.
4000 Please let us (@w{@email{gcc@@gcc.gnu.org}} and @w{@email{fortran@@gnu.org}})
4001 know how use of these options affects
4002 the performance of your production code.
4003 We're very interested in code that runs @emph{slower}
4004 when these options are @emph{enabled}.
4007 @itemx -fno-peephole2
4008 @opindex fno-peephole
4009 @opindex fno-peephole2
4010 Disable any machine-specific peephole optimizations. The difference
4011 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
4012 are implemented in the compiler; some targets use one, some use the
4013 other, a few use both.
4015 @item -fbranch-probabilities
4016 @opindex fbranch-probabilities
4017 After running a program compiled with @option{-fprofile-arcs}
4018 (@pxref{Debugging Options,, Options for Debugging Your Program or
4019 @command{gcc}}), you can compile it a second time using
4020 @option{-fbranch-probabilities}, to improve optimizations based on
4021 the number of times each branch was taken. When the program
4022 compiled with @option{-fprofile-arcs} exits it saves arc execution
4023 counts to a file called @file{@var{sourcename}.da} for each source
4024 file The information in this data file is very dependent on the
4025 structure of the generated code, so you must use the same source code
4026 and the same optimization options for both compilations.
4028 With @option{-fbranch-probabilities}, GCC puts a @samp{REG_EXEC_COUNT}
4029 note on the first instruction of each basic block, and a
4030 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
4031 These can be used to improve optimization. Currently, they are only
4032 used in one place: in @file{reorg.c}, instead of guessing which path a
4033 branch is mostly to take, the @samp{REG_BR_PROB} values are used to
4034 exactly determine which path is taken more often.
4036 @item -fno-guess-branch-probability
4037 @opindex fno-guess-branch-probability
4038 Do not guess branch probabilities using a randomized model.
4040 Sometimes gcc will opt to use a randomized model to guess branch
4041 probabilities, when none are available from either profiling feedback
4042 (@option{-fprofile-arcs}) or @samp{__builtin_expect}. This means that
4043 different runs of the compiler on the same program may produce different
4046 In a hard real-time system, people don't want different runs of the
4047 compiler to produce code that has different behavior; minimizing
4048 non-determinism is of paramount import. This switch allows users to
4049 reduce non-determinism, possibly at the expense of inferior
4052 @item -freorder-blocks
4053 @opindex freorder-blocks
4054 Reorder basic blocks in the compiled function in order to reduce number of
4055 taken branches and improve code locality.
4057 @item -freorder-functions
4058 @opindex freorder-functions
4059 Reorder basic blocks in the compiled function in order to reduce number of
4060 taken branches and improve code locality. This is implemented by using special
4061 subsections @code{text.hot} for most frequently executed functions and
4062 @code{text.unlikely} for unlikely executed functions. Reordering is done by
4063 the linker so object file format must support named sections and linker must
4064 place them in a reasonable way.
4066 Also profile feedback must be available in to make this option effective. See
4067 @option{-fprofile-arcs} for details.
4069 @item -fstrict-aliasing
4070 @opindex fstrict-aliasing
4071 Allows the compiler to assume the strictest aliasing rules applicable to
4072 the language being compiled. For C (and C++), this activates
4073 optimizations based on the type of expressions. In particular, an
4074 object of one type is assumed never to reside at the same address as an
4075 object of a different type, unless the types are almost the same. For
4076 example, an @code{unsigned int} can alias an @code{int}, but not a
4077 @code{void*} or a @code{double}. A character type may alias any other
4080 Pay special attention to code like this:
4093 The practice of reading from a different union member than the one most
4094 recently written to (called ``type-punning'') is common. Even with
4095 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
4096 is accessed through the union type. So, the code above will work as
4097 expected. However, this code might not:
4108 Every language that wishes to perform language-specific alias analysis
4109 should define a function that computes, given an @code{tree}
4110 node, an alias set for the node. Nodes in different alias sets are not
4111 allowed to alias. For an example, see the C front-end function
4112 @code{c_get_alias_set}.
4114 For all machines, optimization level 2 and higher enables this flag by
4117 @item -falign-functions
4118 @itemx -falign-functions=@var{n}
4119 @opindex falign-functions
4120 Align the start of functions to the next power-of-two greater than
4121 @var{n}, skipping up to @var{n} bytes. For instance,
4122 @option{-falign-functions=32} aligns functions to the next 32-byte
4123 boundary, but @option{-falign-functions=24} would align to the next
4124 32-byte boundary only if this can be done by skipping 23 bytes or less.
4126 @option{-fno-align-functions} and @option{-falign-functions=1} are
4127 equivalent and mean that functions will not be aligned.
4129 Some assemblers only support this flag when @var{n} is a power of two;
4130 in that case, it is rounded up.
4132 If @var{n} is not specified, use a machine-dependent default.
4134 @item -falign-labels
4135 @itemx -falign-labels=@var{n}
4136 @opindex falign-labels
4137 Align all branch targets to a power-of-two boundary, skipping up to
4138 @var{n} bytes like @option{-falign-functions}. This option can easily
4139 make code slower, because it must insert dummy operations for when the
4140 branch target is reached in the usual flow of the code.
4142 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
4143 are greater than this value, then their values are used instead.
4145 If @var{n} is not specified, use a machine-dependent default which is
4146 very likely to be @samp{1}, meaning no alignment.
4149 @itemx -falign-loops=@var{n}
4150 @opindex falign-loops
4151 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
4152 like @option{-falign-functions}. The hope is that the loop will be
4153 executed many times, which will make up for any execution of the dummy
4156 If @var{n} is not specified, use a machine-dependent default.
4159 @itemx -falign-jumps=@var{n}
4160 @opindex falign-jumps
4161 Align branch targets to a power-of-two boundary, for branch targets
4162 where the targets can only be reached by jumping, skipping up to @var{n}
4163 bytes like @option{-falign-functions}. In this case, no dummy operations
4166 If @var{n} is not specified, use a machine-dependent default.
4170 Perform optimizations in static single assignment form. Each function's
4171 flow graph is translated into SSA form, optimizations are performed, and
4172 the flow graph is translated back from SSA form. Users should not
4173 specify this option, since it is not yet ready for production use.
4177 Perform Sparse Conditional Constant Propagation in SSA form. Requires
4178 @option{-fssa}. Like @option{-fssa}, this is an experimental feature.
4182 Perform aggressive dead-code elimination in SSA form. Requires @option{-fssa}.
4183 Like @option{-fssa}, this is an experimental feature.
4185 @item -fsingle-precision-constant
4186 @opindex fsingle-precision-constant
4187 Treat floating point constant as single precision constant instead of
4188 implicitly converting it to double precision constant.
4190 @item -frename-registers
4191 @opindex frename-registers
4192 Attempt to avoid false dependencies in scheduled code by making use
4193 of registers left over after register allocation. This optimization
4194 will most benefit processors with lots of registers. It can, however,
4195 make debugging impossible, since variables will no longer stay in
4196 a ``home register''.
4198 @item -fno-cprop-registers
4199 @opindex fno-cprop-registers
4200 After register allocation and post-register allocation instruction splitting,
4201 we perform a copy-propagation pass to try to reduce scheduling dependencies
4202 and occasionally eliminate the copy.
4204 @item --param @var{name}=@var{value}
4206 In some places, GCC uses various constants to control the amount of
4207 optimization that is done. For example, GCC will not inline functions
4208 that contain more that a certain number of instructions. You can
4209 control some of these constants on the command-line using the
4210 @option{--param} option.
4212 In each case, the @var{value} is an integer. The allowable choices for
4213 @var{name} are given in the following table:
4216 @item max-delay-slot-insn-search
4217 The maximum number of instructions to consider when looking for an
4218 instruction to fill a delay slot. If more than this arbitrary number of
4219 instructions is searched, the time savings from filling the delay slot
4220 will be minimal so stop searching. Increasing values mean more
4221 aggressive optimization, making the compile time increase with probably
4222 small improvement in executable run time.
4224 @item max-delay-slot-live-search
4225 When trying to fill delay slots, the maximum number of instructions to
4226 consider when searching for a block with valid live register
4227 information. Increasing this arbitrarily chosen value means more
4228 aggressive optimization, increasing the compile time. This parameter
4229 should be removed when the delay slot code is rewritten to maintain the
4232 @item max-gcse-memory
4233 The approximate maximum amount of memory that will be allocated in
4234 order to perform the global common subexpression elimination
4235 optimization. If more memory than specified is required, the
4236 optimization will not be done.
4238 @item max-gcse-passes
4239 The maximum number of passes of GCSE to run.
4241 @item max-pending-list-length
4242 The maximum number of pending dependencies scheduling will allow
4243 before flushing the current state and starting over. Large functions
4244 with few branches or calls can create excessively large lists which
4245 needlessly consume memory and resources.
4247 @item max-inline-insns
4248 If an function contains more than this many instructions, it
4249 will not be inlined. This option is precisely equivalent to
4250 @option{-finline-limit}.
4252 @item max-unrolled-insns
4253 The maximum number of instructions that a loop should have if that loop
4254 is unrolled, and if the loop is unrolled, it determines how many times
4255 the loop code is unrolled.
4257 @item hot-bb-count-fraction
4258 Select fraction of the maximal count of repetitions of basic block in program
4259 given basic block needs to have to be considered hot.
4261 @item hot-bb-frequency-fraction
4262 Select fraction of the maximal frequency of executions of basic block in
4263 function given basic block needs to have to be considered hot
4265 @item tracer-dynamic-coverage
4266 @itemx tracer-dynamic-coverage-feedback
4268 This value is used to limit superblock formation once the given percentage of
4269 executed instructions is covered. This limits unnecessary code size
4272 The @option{tracer-dynamic-coverage-feedback} is used only when profile
4273 feedback is available. The real profiles (as opposed to statically estimated
4274 ones) are much less balanced allowing the threshold to be larger value.
4276 @item tracer-max-code-growth
4277 Stop tail duplication once code growth has reached given percentage. This is
4278 rather hokey argument, as most of the duplicates will be eliminated later in
4279 cross jumping, so it may be set to much higher values than is the desired code
4282 @item tracer-min-branch-ratio
4284 Stop reverse growth when the reverse probability of best edge is less than this
4285 threshold (in percent).
4287 @item tracer-min-branch-ratio
4288 @itemx tracer-min-branch-ratio-feedback
4290 Stop forward growth if the best edge do have probability lower than this
4293 Similarly to @option{tracer-dynamic-coverage} two values are present, one for
4294 compilation for profile feedback and one for compilation without. The value
4295 for compilation with profile feedback needs to be more conservative (higher) in
4296 order to make tracer effective.
4298 @item ggc-min-expand
4300 GCC uses a garbage collector to manage its own memory allocation. This
4301 parameter specifies the minimum percentage by which the garbage
4302 collector's heap should be allowed to expand between collections.
4303 Tuning this may improve compilation speed; it has no effect on code
4306 The default is 30%. Setting this parameter and
4307 @option{ggc-min-heapsize} to zero causes a full collection to occur at
4308 every opportunity. This is extremely slow, but can be useful for
4311 @item ggc-min-heapsize
4313 Minimum size of the garbage collector's heap before it begins bothering
4314 to collect garbage. The first collection occurs after the heap expands
4315 by @option{ggc-min-expand}% beyond @option{ggc-min-heapsize}. Again,
4316 tuning this may improve compilation speed, and has no effect on code
4319 The default is 4096 (four megabytes). Setting this parameter very large
4320 effectively disables garbage collection. Setting this parameter and
4321 @option{ggc-min-expand} to zero causes a full collection to occur at
4327 @node Preprocessor Options
4328 @section Options Controlling the Preprocessor
4329 @cindex preprocessor options
4330 @cindex options, preprocessor
4332 These options control the C preprocessor, which is run on each C source
4333 file before actual compilation.
4335 If you use the @option{-E} option, nothing is done except preprocessing.
4336 Some of these options make sense only together with @option{-E} because
4337 they cause the preprocessor output to be unsuitable for actual
4342 You can use @option{-Wp,@var{option}} to bypass the compiler driver
4343 and pass @var{option} directly through to the preprocessor. If
4344 @var{option} contains commas, it is split into multiple options at the
4345 commas. However, many options are modified, translated or interpreted
4346 by the compiler driver before being passed to the preprocessor, and
4347 @option{-Wp} forcibly bypasses this phase. The preprocessor's direct
4348 interface is undocumented and subject to change, so whenever possible
4349 you should avoid using @option{-Wp} and let the driver handle the
4352 @item -Xpreprocessor @var{option}
4353 @opindex preprocessor
4354 Pass @var{option} as an option to the preprocessor. You can use this to
4355 supply system-specific preprocessor options which GCC does not know how to
4358 If you want to pass an option that takes an argument, you must use
4359 @option{-Xpreprocessor} twice, once for the option and once for the argument.
4362 @include cppopts.texi
4364 @node Assembler Options
4365 @section Passing Options to the Assembler
4367 @c prevent bad page break with this line
4368 You can pass options to the assembler.
4371 @item -Wa,@var{option}
4373 Pass @var{option} as an option to the assembler. If @var{option}
4374 contains commas, it is split into multiple options at the commas.
4376 @item -Xassembler @var{option}
4378 Pass @var{option} as an option to the assembler. You can use this to
4379 supply system-specific assembler options which GCC does not know how to
4382 If you want to pass an option that takes an argument, you must use
4383 @option{-Xassembler} twice, once for the option and once for the argument.
4388 @section Options for Linking
4389 @cindex link options
4390 @cindex options, linking
4392 These options come into play when the compiler links object files into
4393 an executable output file. They are meaningless if the compiler is
4394 not doing a link step.
4398 @item @var{object-file-name}
4399 A file name that does not end in a special recognized suffix is
4400 considered to name an object file or library. (Object files are
4401 distinguished from libraries by the linker according to the file
4402 contents.) If linking is done, these object files are used as input
4411 If any of these options is used, then the linker is not run, and
4412 object file names should not be used as arguments. @xref{Overall
4416 @item -l@var{library}
4417 @itemx -l @var{library}
4419 Search the library named @var{library} when linking. (The second
4420 alternative with the library as a separate argument is only for
4421 POSIX compliance and is not recommended.)
4423 It makes a difference where in the command you write this option; the
4424 linker searches and processes libraries and object files in the order they
4425 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
4426 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
4427 to functions in @samp{z}, those functions may not be loaded.
4429 The linker searches a standard list of directories for the library,
4430 which is actually a file named @file{lib@var{library}.a}. The linker
4431 then uses this file as if it had been specified precisely by name.
4433 The directories searched include several standard system directories
4434 plus any that you specify with @option{-L}.
4436 Normally the files found this way are library files---archive files
4437 whose members are object files. The linker handles an archive file by
4438 scanning through it for members which define symbols that have so far
4439 been referenced but not defined. But if the file that is found is an
4440 ordinary object file, it is linked in the usual fashion. The only
4441 difference between using an @option{-l} option and specifying a file name
4442 is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
4443 and searches several directories.
4447 You need this special case of the @option{-l} option in order to
4448 link an Objective-C program.
4451 @opindex nostartfiles
4452 Do not use the standard system startup files when linking.
4453 The standard system libraries are used normally, unless @option{-nostdlib}
4454 or @option{-nodefaultlibs} is used.
4456 @item -nodefaultlibs
4457 @opindex nodefaultlibs
4458 Do not use the standard system libraries when linking.
4459 Only the libraries you specify will be passed to the linker.
4460 The standard startup files are used normally, unless @option{-nostartfiles}
4461 is used. The compiler may generate calls to memcmp, memset, and memcpy
4462 for System V (and ISO C) environments or to bcopy and bzero for
4463 BSD environments. These entries are usually resolved by entries in
4464 libc. These entry points should be supplied through some other
4465 mechanism when this option is specified.
4469 Do not use the standard system startup files or libraries when linking.
4470 No startup files and only the libraries you specify will be passed to
4471 the linker. The compiler may generate calls to memcmp, memset, and memcpy
4472 for System V (and ISO C) environments or to bcopy and bzero for
4473 BSD environments. These entries are usually resolved by entries in
4474 libc. These entry points should be supplied through some other
4475 mechanism when this option is specified.
4477 @cindex @option{-lgcc}, use with @option{-nostdlib}
4478 @cindex @option{-nostdlib} and unresolved references
4479 @cindex unresolved references and @option{-nostdlib}
4480 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
4481 @cindex @option{-nodefaultlibs} and unresolved references
4482 @cindex unresolved references and @option{-nodefaultlibs}
4483 One of the standard libraries bypassed by @option{-nostdlib} and
4484 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
4485 that GCC uses to overcome shortcomings of particular machines, or special
4486 needs for some languages.
4487 (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler
4488 Collection (GCC) Internals},
4489 for more discussion of @file{libgcc.a}.)
4490 In most cases, you need @file{libgcc.a} even when you want to avoid
4491 other standard libraries. In other words, when you specify @option{-nostdlib}
4492 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
4493 This ensures that you have no unresolved references to internal GCC
4494 library subroutines. (For example, @samp{__main}, used to ensure C++
4495 constructors will be called; @pxref{Collect2,,@code{collect2}, gccint,
4496 GNU Compiler Collection (GCC) Internals}.)
4500 Remove all symbol table and relocation information from the executable.
4504 On systems that support dynamic linking, this prevents linking with the shared
4505 libraries. On other systems, this option has no effect.
4509 Produce a shared object which can then be linked with other objects to
4510 form an executable. Not all systems support this option. For predictable
4511 results, you must also specify the same set of options that were used to
4512 generate code (@option{-fpic}, @option{-fPIC}, or model suboptions)
4513 when you specify this option.@footnote{On some systems, @samp{gcc -shared}
4514 needs to build supplementary stub code for constructors to work. On
4515 multi-libbed systems, @samp{gcc -shared} must select the correct support
4516 libraries to link against. Failing to supply the correct flags may lead
4517 to subtle defects. Supplying them in cases where they are not necessary
4520 @item -shared-libgcc
4521 @itemx -static-libgcc
4522 @opindex shared-libgcc
4523 @opindex static-libgcc
4524 On systems that provide @file{libgcc} as a shared library, these options
4525 force the use of either the shared or static version respectively.
4526 If no shared version of @file{libgcc} was built when the compiler was
4527 configured, these options have no effect.
4529 There are several situations in which an application should use the
4530 shared @file{libgcc} instead of the static version. The most common
4531 of these is when the application wishes to throw and catch exceptions
4532 across different shared libraries. In that case, each of the libraries
4533 as well as the application itself should use the shared @file{libgcc}.
4535 Therefore, the G++ and GCJ drivers automatically add
4536 @option{-shared-libgcc} whenever you build a shared library or a main
4537 executable, because C++ and Java programs typically use exceptions, so
4538 this is the right thing to do.
4540 If, instead, you use the GCC driver to create shared libraries, you may
4541 find that they will not always be linked with the shared @file{libgcc}.
4542 If GCC finds, at its configuration time, that you have a GNU linker that
4543 does not support option @option{--eh-frame-hdr}, it will link the shared
4544 version of @file{libgcc} into shared libraries by default. Otherwise,
4545 it will take advantage of the linker and optimize away the linking with
4546 the shared version of @file{libgcc}, linking with the static version of
4547 libgcc by default. This allows exceptions to propagate through such
4548 shared libraries, without incurring relocation costs at library load
4551 However, if a library or main executable is supposed to throw or catch
4552 exceptions, you must link it using the G++ or GCJ driver, as appropriate
4553 for the languages used in the program, or using the option
4554 @option{-shared-libgcc}, such that it is linked with the shared
4559 Bind references to global symbols when building a shared object. Warn
4560 about any unresolved references (unless overridden by the link editor
4561 option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support
4564 @item -Xlinker @var{option}
4566 Pass @var{option} as an option to the linker. You can use this to
4567 supply system-specific linker options which GCC does not know how to
4570 If you want to pass an option that takes an argument, you must use
4571 @option{-Xlinker} twice, once for the option and once for the argument.
4572 For example, to pass @option{-assert definitions}, you must write
4573 @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write
4574 @option{-Xlinker "-assert definitions"}, because this passes the entire
4575 string as a single argument, which is not what the linker expects.
4577 @item -Wl,@var{option}
4579 Pass @var{option} as an option to the linker. If @var{option} contains
4580 commas, it is split into multiple options at the commas.
4582 @item -u @var{symbol}
4584 Pretend the symbol @var{symbol} is undefined, to force linking of
4585 library modules to define it. You can use @option{-u} multiple times with
4586 different symbols to force loading of additional library modules.
4589 @node Directory Options
4590 @section Options for Directory Search
4591 @cindex directory options
4592 @cindex options, directory search
4595 These options specify directories to search for header files, for
4596 libraries and for parts of the compiler:
4601 Add the directory @var{dir} to the head of the list of directories to be
4602 searched for header files. This can be used to override a system header
4603 file, substituting your own version, since these directories are
4604 searched before the system header file directories. However, you should
4605 not use this option to add directories that contain vendor-supplied
4606 system header files (use @option{-isystem} for that). If you use more than
4607 one @option{-I} option, the directories are scanned in left-to-right
4608 order; the standard system directories come after.
4610 If a standard system include directory, or a directory specified with
4611 @option{-isystem}, is also specified with @option{-I}, the @option{-I}
4612 option will be ignored. The directory will still be searched but as a
4613 system directory at its normal position in the system include chain.
4614 This is to ensure that GCC's procedure to fix buggy system headers and
4615 the ordering for the include_next directive are not inadvertently changed.
4616 If you really need to change the search order for system directories,
4617 use the @option{-nostdinc} and/or @option{-isystem} options.
4621 Any directories you specify with @option{-I} options before the @option{-I-}
4622 option are searched only for the case of @samp{#include "@var{file}"};
4623 they are not searched for @samp{#include <@var{file}>}.
4625 If additional directories are specified with @option{-I} options after
4626 the @option{-I-}, these directories are searched for all @samp{#include}
4627 directives. (Ordinarily @emph{all} @option{-I} directories are used
4630 In addition, the @option{-I-} option inhibits the use of the current
4631 directory (where the current input file came from) as the first search
4632 directory for @samp{#include "@var{file}"}. There is no way to
4633 override this effect of @option{-I-}. With @option{-I.} you can specify
4634 searching the directory which was current when the compiler was
4635 invoked. That is not exactly the same as what the preprocessor does
4636 by default, but it is often satisfactory.
4638 @option{-I-} does not inhibit the use of the standard system directories
4639 for header files. Thus, @option{-I-} and @option{-nostdinc} are
4644 Add directory @var{dir} to the list of directories to be searched
4647 @item -B@var{prefix}
4649 This option specifies where to find the executables, libraries,
4650 include files, and data files of the compiler itself.
4652 The compiler driver program runs one or more of the subprograms
4653 @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries
4654 @var{prefix} as a prefix for each program it tries to run, both with and
4655 without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}).
4657 For each subprogram to be run, the compiler driver first tries the
4658 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
4659 was not specified, the driver tries two standard prefixes, which are
4660 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc-lib/}. If neither of
4661 those results in a file name that is found, the unmodified program
4662 name is searched for using the directories specified in your
4663 @env{PATH} environment variable.
4665 The compiler will check to see if the path provided by the @option{-B}
4666 refers to a directory, and if necessary it will add a directory
4667 separator character at the end of the path.
4669 @option{-B} prefixes that effectively specify directory names also apply
4670 to libraries in the linker, because the compiler translates these
4671 options into @option{-L} options for the linker. They also apply to
4672 includes files in the preprocessor, because the compiler translates these
4673 options into @option{-isystem} options for the preprocessor. In this case,
4674 the compiler appends @samp{include} to the prefix.
4676 The run-time support file @file{libgcc.a} can also be searched for using
4677 the @option{-B} prefix, if needed. If it is not found there, the two
4678 standard prefixes above are tried, and that is all. The file is left
4679 out of the link if it is not found by those means.
4681 Another way to specify a prefix much like the @option{-B} prefix is to use
4682 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
4685 As a special kludge, if the path provided by @option{-B} is
4686 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
4687 9, then it will be replaced by @file{[dir/]include}. This is to help
4688 with boot-strapping the compiler.
4690 @item -specs=@var{file}
4692 Process @var{file} after the compiler reads in the standard @file{specs}
4693 file, in order to override the defaults that the @file{gcc} driver
4694 program uses when determining what switches to pass to @file{cc1},
4695 @file{cc1plus}, @file{as}, @file{ld}, etc. More than one
4696 @option{-specs=@var{file}} can be specified on the command line, and they
4697 are processed in order, from left to right.
4703 @section Specifying subprocesses and the switches to pass to them
4705 @command{gcc} is a driver program. It performs its job by invoking a
4706 sequence of other programs to do the work of compiling, assembling and
4707 linking. GCC interprets its command-line parameters and uses these to
4708 deduce which programs it should invoke, and which command-line options
4709 it ought to place on their command lines. This behavior is controlled
4710 by @dfn{spec strings}. In most cases there is one spec string for each
4711 program that GCC can invoke, but a few programs have multiple spec
4712 strings to control their behavior. The spec strings built into GCC can
4713 be overridden by using the @option{-specs=} command-line switch to specify
4716 @dfn{Spec files} are plaintext files that are used to construct spec
4717 strings. They consist of a sequence of directives separated by blank
4718 lines. The type of directive is determined by the first non-whitespace
4719 character on the line and it can be one of the following:
4722 @item %@var{command}
4723 Issues a @var{command} to the spec file processor. The commands that can
4727 @item %include <@var{file}>
4729 Search for @var{file} and insert its text at the current point in the
4732 @item %include_noerr <@var{file}>
4733 @cindex %include_noerr
4734 Just like @samp{%include}, but do not generate an error message if the include
4735 file cannot be found.
4737 @item %rename @var{old_name} @var{new_name}
4739 Rename the spec string @var{old_name} to @var{new_name}.
4743 @item *[@var{spec_name}]:
4744 This tells the compiler to create, override or delete the named spec
4745 string. All lines after this directive up to the next directive or
4746 blank line are considered to be the text for the spec string. If this
4747 results in an empty string then the spec will be deleted. (Or, if the
4748 spec did not exist, then nothing will happened.) Otherwise, if the spec
4749 does not currently exist a new spec will be created. If the spec does
4750 exist then its contents will be overridden by the text of this
4751 directive, unless the first character of that text is the @samp{+}
4752 character, in which case the text will be appended to the spec.
4754 @item [@var{suffix}]:
4755 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
4756 and up to the next directive or blank line are considered to make up the
4757 spec string for the indicated suffix. When the compiler encounters an
4758 input file with the named suffix, it will processes the spec string in
4759 order to work out how to compile that file. For example:
4766 This says that any input file whose name ends in @samp{.ZZ} should be
4767 passed to the program @samp{z-compile}, which should be invoked with the
4768 command-line switch @option{-input} and with the result of performing the
4769 @samp{%i} substitution. (See below.)
4771 As an alternative to providing a spec string, the text that follows a
4772 suffix directive can be one of the following:
4775 @item @@@var{language}
4776 This says that the suffix is an alias for a known @var{language}. This is
4777 similar to using the @option{-x} command-line switch to GCC to specify a
4778 language explicitly. For example:
4785 Says that .ZZ files are, in fact, C++ source files.
4788 This causes an error messages saying:
4791 @var{name} compiler not installed on this system.
4795 GCC already has an extensive list of suffixes built into it.
4796 This directive will add an entry to the end of the list of suffixes, but
4797 since the list is searched from the end backwards, it is effectively
4798 possible to override earlier entries using this technique.
4802 GCC has the following spec strings built into it. Spec files can
4803 override these strings or create their own. Note that individual
4804 targets can also add their own spec strings to this list.
4807 asm Options to pass to the assembler
4808 asm_final Options to pass to the assembler post-processor
4809 cpp Options to pass to the C preprocessor
4810 cc1 Options to pass to the C compiler
4811 cc1plus Options to pass to the C++ compiler
4812 endfile Object files to include at the end of the link
4813 link Options to pass to the linker
4814 lib Libraries to include on the command line to the linker
4815 libgcc Decides which GCC support library to pass to the linker
4816 linker Sets the name of the linker
4817 predefines Defines to be passed to the C preprocessor
4818 signed_char Defines to pass to CPP to say whether @code{char} is signed
4820 startfile Object files to include at the start of the link
4823 Here is a small example of a spec file:
4829 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
4832 This example renames the spec called @samp{lib} to @samp{old_lib} and
4833 then overrides the previous definition of @samp{lib} with a new one.
4834 The new definition adds in some extra command-line options before
4835 including the text of the old definition.
4837 @dfn{Spec strings} are a list of command-line options to be passed to their
4838 corresponding program. In addition, the spec strings can contain
4839 @samp{%}-prefixed sequences to substitute variable text or to
4840 conditionally insert text into the command line. Using these constructs
4841 it is possible to generate quite complex command lines.
4843 Here is a table of all defined @samp{%}-sequences for spec
4844 strings. Note that spaces are not generated automatically around the
4845 results of expanding these sequences. Therefore you can concatenate them
4846 together or combine them with constant text in a single argument.
4850 Substitute one @samp{%} into the program name or argument.
4853 Substitute the name of the input file being processed.
4856 Substitute the basename of the input file being processed.
4857 This is the substring up to (and not including) the last period
4858 and not including the directory.
4861 This is the same as @samp{%b}, but include the file suffix (text after
4865 Marks the argument containing or following the @samp{%d} as a
4866 temporary file name, so that that file will be deleted if GCC exits
4867 successfully. Unlike @samp{%g}, this contributes no text to the
4870 @item %g@var{suffix}
4871 Substitute a file name that has suffix @var{suffix} and is chosen
4872 once per compilation, and mark the argument in the same way as
4873 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
4874 name is now chosen in a way that is hard to predict even when previously
4875 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
4876 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
4877 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
4878 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
4879 was simply substituted with a file name chosen once per compilation,
4880 without regard to any appended suffix (which was therefore treated
4881 just like ordinary text), making such attacks more likely to succeed.
4883 @item %u@var{suffix}
4884 Like @samp{%g}, but generates a new temporary file name even if
4885 @samp{%u@var{suffix}} was already seen.
4887 @item %U@var{suffix}
4888 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
4889 new one if there is no such last file name. In the absence of any
4890 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
4891 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
4892 would involve the generation of two distinct file names, one
4893 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
4894 simply substituted with a file name chosen for the previous @samp{%u},
4895 without regard to any appended suffix.
4897 @item %j@var{suffix}
4898 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
4899 writable, and if save-temps is off; otherwise, substitute the name
4900 of a temporary file, just like @samp{%u}. This temporary file is not
4901 meant for communication between processes, but rather as a junk
4904 @item %|@var{suffix}
4905 @itemx %m@var{suffix}
4906 Like @samp{%g}, except if @option{-pipe} is in effect. In that case
4907 @samp{%|} substitutes a single dash and @samp{%m} substitutes nothing at
4908 all. These are the two most common ways to instruct a program that it
4909 should read from standard input or write to standard output. If you
4910 need something more elaborate you can use an @samp{%@{pipe:@code{X}@}}
4911 construct: see for example @file{f/lang-specs.h}.
4913 @item %.@var{SUFFIX}
4914 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
4915 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
4916 terminated by the next space or %.
4919 Marks the argument containing or following the @samp{%w} as the
4920 designated output file of this compilation. This puts the argument
4921 into the sequence of arguments that @samp{%o} will substitute later.
4924 Substitutes the names of all the output files, with spaces
4925 automatically placed around them. You should write spaces
4926 around the @samp{%o} as well or the results are undefined.
4927 @samp{%o} is for use in the specs for running the linker.
4928 Input files whose names have no recognized suffix are not compiled
4929 at all, but they are included among the output files, so they will
4933 Substitutes the suffix for object files. Note that this is
4934 handled specially when it immediately follows @samp{%g, %u, or %U},
4935 because of the need for those to form complete file names. The
4936 handling is such that @samp{%O} is treated exactly as if it had already
4937 been substituted, except that @samp{%g, %u, and %U} do not currently
4938 support additional @var{suffix} characters following @samp{%O} as they would
4939 following, for example, @samp{.o}.
4942 Substitutes the standard macro predefinitions for the
4943 current target machine. Use this when running @code{cpp}.
4946 Like @samp{%p}, but puts @samp{__} before and after the name of each
4947 predefined macro, except for macros that start with @samp{__} or with
4948 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
4952 Substitute a @option{-iprefix} option made from @env{GCC_EXEC_PREFIX}.
4955 Current argument is the name of a library or startup file of some sort.
4956 Search for that file in a standard list of directories and substitute
4957 the full name found.
4960 Print @var{str} as an error message. @var{str} is terminated by a newline.
4961 Use this when inconsistent options are detected.
4964 Substitute the contents of spec string @var{name} at this point.
4967 Like @samp{%(@dots{})} but put @samp{__} around @option{-D} arguments.
4969 @item %x@{@var{option}@}
4970 Accumulate an option for @samp{%X}.
4973 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
4977 Output the accumulated assembler options specified by @option{-Wa}.
4980 Output the accumulated preprocessor options specified by @option{-Wp}.
4983 Substitute the major version number of GCC@.
4984 (For version 2.9.5, this is 2.)
4987 Substitute the minor version number of GCC@.
4988 (For version 2.9.5, this is 9.)
4991 Substitute the patch level number of GCC@.
4992 (For version 2.9.5, this is 5.)
4995 Process the @code{asm} spec. This is used to compute the
4996 switches to be passed to the assembler.
4999 Process the @code{asm_final} spec. This is a spec string for
5000 passing switches to an assembler post-processor, if such a program is
5004 Process the @code{link} spec. This is the spec for computing the
5005 command line passed to the linker. Typically it will make use of the
5006 @samp{%L %G %S %D and %E} sequences.
5009 Dump out a @option{-L} option for each directory that GCC believes might
5010 contain startup files. If the target supports multilibs then the
5011 current multilib directory will be prepended to each of these paths.
5014 Output the multilib directory with directory separators replaced with
5015 @samp{_}. If multilib directories are not set, or the multilib directory is
5016 @file{.} then this option emits nothing.
5019 Process the @code{lib} spec. This is a spec string for deciding which
5020 libraries should be included on the command line to the linker.
5023 Process the @code{libgcc} spec. This is a spec string for deciding
5024 which GCC support library should be included on the command line to the linker.
5027 Process the @code{startfile} spec. This is a spec for deciding which
5028 object files should be the first ones passed to the linker. Typically
5029 this might be a file named @file{crt0.o}.
5032 Process the @code{endfile} spec. This is a spec string that specifies
5033 the last object files that will be passed to the linker.
5036 Process the @code{cpp} spec. This is used to construct the arguments
5037 to be passed to the C preprocessor.
5040 Process the @code{signed_char} spec. This is intended to be used
5041 to tell cpp whether a char is signed. It typically has the definition:
5043 %@{funsigned-char:-D__CHAR_UNSIGNED__@}
5047 Process the @code{cc1} spec. This is used to construct the options to be
5048 passed to the actual C compiler (@samp{cc1}).
5051 Process the @code{cc1plus} spec. This is used to construct the options to be
5052 passed to the actual C++ compiler (@samp{cc1plus}).
5055 Substitute the variable part of a matched option. See below.
5056 Note that each comma in the substituted string is replaced by
5060 Remove all occurrences of @code{-S} from the command line. Note---this
5061 command is position dependent. @samp{%} commands in the spec string
5062 before this one will see @code{-S}, @samp{%} commands in the spec string
5063 after this one will not.
5065 @item %:@var{function}(@var{args})
5066 Call the named function @var{function}, passing it @var{args}.
5067 @var{args} is first processed as a nested spec string, then split
5068 into an argument vector in the usual fashion. The function returns
5069 a string which is processed as if it had appeared literally as part
5070 of the current spec.
5072 The following built-in spec functions are provided:
5075 @item @code{if-exists}
5076 The @code{if-exists} spec function takes one argument, an absolute
5077 pathname to a file. If the file exists, @code{if-exists} returns the
5078 pathname. Here is a small example of its usage:
5082 crt0%O%s %:if-exists(crti%O%s) crtbegin%O%s
5085 @item @code{if-exists-else}
5086 The @code{if-exists-else} spec function is similar to the @code{if-exists}
5087 spec function, except that it takes two arguments. The first argument is
5088 an absolute pathname to a file. If the file exists, @code{if-exists-else}
5089 returns the pathname. If it does not exist, it returns the second argument.
5090 This way, @code{if-exists-else} can be used to select one file or another,
5091 based on the existence of the first. Here is a small example of its usage:
5095 crt0%O%s %:if-exists(crti%O%s) %:if-exists-else(crtbeginT%O%s crtbegin%O%s)
5100 Substitutes the @code{-S} switch, if that switch was given to GCC@.
5101 If that switch was not specified, this substitutes nothing. Note that
5102 the leading dash is omitted when specifying this option, and it is
5103 automatically inserted if the substitution is performed. Thus the spec
5104 string @samp{%@{foo@}} would match the command-line option @option{-foo}
5105 and would output the command line option @option{-foo}.
5107 @item %W@{@code{S}@}
5108 Like %@{@code{S}@} but mark last argument supplied within as a file to be
5111 @item %@{@code{S}*@}
5112 Substitutes all the switches specified to GCC whose names start
5113 with @code{-S}, but which also take an argument. This is used for
5114 switches like @option{-o}, @option{-D}, @option{-I}, etc.
5115 GCC considers @option{-o foo} as being
5116 one switch whose names starts with @samp{o}. %@{o*@} would substitute this
5117 text, including the space. Thus two arguments would be generated.
5119 @item %@{@code{S}*&@code{T}*@}
5120 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
5121 (the order of @code{S} and @code{T} in the spec is not significant).
5122 There can be any number of ampersand-separated variables; for each the
5123 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
5125 @item %@{@code{S}:@code{X}@}
5126 Substitutes @code{X}, if the @samp{-S} switch was given to GCC@.
5128 @item %@{!@code{S}:@code{X}@}
5129 Substitutes @code{X}, if the @samp{-S} switch was @emph{not} given to GCC@.
5131 @item %@{@code{S}*:@code{X}@}
5132 Substitutes @code{X} if one or more switches whose names start with
5133 @code{-S} are specified to GCC@. Normally @code{X} is substituted only
5134 once, no matter how many such switches appeared. However, if @code{%*}
5135 appears somewhere in @code{X}, then @code{X} will be substituted once
5136 for each matching switch, with the @code{%*} replaced by the part of
5137 that switch that matched the @code{*}.
5139 @item %@{.@code{S}:@code{X}@}
5140 Substitutes @code{X}, if processing a file with suffix @code{S}.
5142 @item %@{!.@code{S}:@code{X}@}
5143 Substitutes @code{X}, if @emph{not} processing a file with suffix @code{S}.
5145 @item %@{@code{S}|@code{P}:@code{X}@}
5146 Substitutes @code{X} if either @code{-S} or @code{-P} was given to GCC@.
5147 This may be combined with @samp{!}, @samp{.}, and @code{*} sequences as well,
5148 although they have a stronger binding than the @samp{|}. If @code{%*}
5149 appears in @code{X}, all of the alternatives must be starred, and only
5150 the first matching alternative is substituted.
5152 For example, a spec string like this:
5155 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
5158 will output the following command-line options from the following input
5159 command-line options:
5164 -d fred.c -foo -baz -boggle
5165 -d jim.d -bar -baz -boggle
5168 @item %@{S:X; T:Y; :D@}
5170 If @code{S} was given to GCC, substitues @code{X}; else if @code{T} was
5171 given to GCC, substitues @code{Y}; else substitutes @code{D}. There can
5172 be as many clauses as you need. This may be combined with @code{.},
5173 @code{!}, @code{|}, and @code{*} as needed.
5178 The conditional text @code{X} in a %@{@code{S}:@code{X}@} or similar
5179 construct may contain other nested @samp{%} constructs or spaces, or
5180 even newlines. They are processed as usual, as described above.
5181 Trailing white space in @code{X} is ignored. White space may also
5182 appear anywhere on the left side of the colon in these constructs,
5183 except between @code{.} or @code{*} and the corresponding word.
5185 The @option{-O}, @option{-f}, @option{-m}, and @option{-W} switches are
5186 handled specifically in these constructs. If another value of
5187 @option{-O} or the negated form of a @option{-f}, @option{-m}, or
5188 @option{-W} switch is found later in the command line, the earlier
5189 switch value is ignored, except with @{@code{S}*@} where @code{S} is
5190 just one letter, which passes all matching options.
5192 The character @samp{|} at the beginning of the predicate text is used to
5193 indicate that a command should be piped to the following command, but
5194 only if @option{-pipe} is specified.
5196 It is built into GCC which switches take arguments and which do not.
5197 (You might think it would be useful to generalize this to allow each
5198 compiler's spec to say which switches take arguments. But this cannot
5199 be done in a consistent fashion. GCC cannot even decide which input
5200 files have been specified without knowing which switches take arguments,
5201 and it must know which input files to compile in order to tell which
5204 GCC also knows implicitly that arguments starting in @option{-l} are to be
5205 treated as compiler output files, and passed to the linker in their
5206 proper position among the other output files.
5208 @c man begin OPTIONS
5210 @node Target Options
5211 @section Specifying Target Machine and Compiler Version
5212 @cindex target options
5213 @cindex cross compiling
5214 @cindex specifying machine version
5215 @cindex specifying compiler version and target machine
5216 @cindex compiler version, specifying
5217 @cindex target machine, specifying
5219 The usual way to run GCC is to run the executable called @file{gcc}, or
5220 @file{<machine>-gcc} when cross-compiling, or
5221 @file{<machine>-gcc-<version>} to run a version other than the one that
5222 was installed last. Sometimes this is inconvenient, so GCC provides
5223 options that will switch to another cross-compiler or version.
5226 @item -b @var{machine}
5228 The argument @var{machine} specifies the target machine for compilation.
5230 The value to use for @var{machine} is the same as was specified as the
5231 machine type when configuring GCC as a cross-compiler. For
5232 example, if a cross-compiler was configured with @samp{configure
5233 i386v}, meaning to compile for an 80386 running System V, then you
5234 would specify @option{-b i386v} to run that cross compiler.
5236 @item -V @var{version}
5238 The argument @var{version} specifies which version of GCC to run.
5239 This is useful when multiple versions are installed. For example,
5240 @var{version} might be @samp{2.0}, meaning to run GCC version 2.0.
5243 The @option{-V} and @option{-b} options work by running the
5244 @file{<machine>-gcc-<version>} executable, so there's no real reason to
5245 use them if you can just run that directly.
5247 @node Submodel Options
5248 @section Hardware Models and Configurations
5249 @cindex submodel options
5250 @cindex specifying hardware config
5251 @cindex hardware models and configurations, specifying
5252 @cindex machine dependent options
5254 Earlier we discussed the standard option @option{-b} which chooses among
5255 different installed compilers for completely different target
5256 machines, such as VAX vs.@: 68000 vs.@: 80386.
5258 In addition, each of these target machine types can have its own
5259 special options, starting with @samp{-m}, to choose among various
5260 hardware models or configurations---for example, 68010 vs 68020,
5261 floating coprocessor or none. A single installed version of the
5262 compiler can compile for any model or configuration, according to the
5265 Some configurations of the compiler also support additional special
5266 options, usually for compatibility with other compilers on the same
5269 These options are defined by the macro @code{TARGET_SWITCHES} in the
5270 machine description. The default for the options is also defined by
5271 that macro, which enables you to change the defaults.
5283 * RS/6000 and PowerPC Options::
5287 * i386 and x86-64 Options::
5289 * Intel 960 Options::
5290 * DEC Alpha Options::
5291 * DEC Alpha/VMS Options::
5294 * System V Options::
5295 * TMS320C3x/C4x Options::
5303 * S/390 and zSeries Options::
5307 * Xstormy16 Options::
5312 @node M680x0 Options
5313 @subsection M680x0 Options
5314 @cindex M680x0 options
5316 These are the @samp{-m} options defined for the 68000 series. The default
5317 values for these options depends on which style of 68000 was selected when
5318 the compiler was configured; the defaults for the most common choices are
5326 Generate output for a 68000. This is the default
5327 when the compiler is configured for 68000-based systems.
5329 Use this option for microcontrollers with a 68000 or EC000 core,
5330 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
5336 Generate output for a 68020. This is the default
5337 when the compiler is configured for 68020-based systems.
5341 Generate output containing 68881 instructions for floating point.
5342 This is the default for most 68020 systems unless @option{--nfp} was
5343 specified when the compiler was configured.
5347 Generate output for a 68030. This is the default when the compiler is
5348 configured for 68030-based systems.
5352 Generate output for a 68040. This is the default when the compiler is
5353 configured for 68040-based systems.
5355 This option inhibits the use of 68881/68882 instructions that have to be
5356 emulated by software on the 68040. Use this option if your 68040 does not
5357 have code to emulate those instructions.
5361 Generate output for a 68060. This is the default when the compiler is
5362 configured for 68060-based systems.
5364 This option inhibits the use of 68020 and 68881/68882 instructions that
5365 have to be emulated by software on the 68060. Use this option if your 68060
5366 does not have code to emulate those instructions.
5370 Generate output for a CPU32. This is the default
5371 when the compiler is configured for CPU32-based systems.
5373 Use this option for microcontrollers with a
5374 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
5375 68336, 68340, 68341, 68349 and 68360.
5379 Generate output for a 520X ``coldfire'' family cpu. This is the default
5380 when the compiler is configured for 520X-based systems.
5382 Use this option for microcontroller with a 5200 core, including
5383 the MCF5202, MCF5203, MCF5204 and MCF5202.
5388 Generate output for a 68040, without using any of the new instructions.
5389 This results in code which can run relatively efficiently on either a
5390 68020/68881 or a 68030 or a 68040. The generated code does use the
5391 68881 instructions that are emulated on the 68040.
5395 Generate output for a 68060, without using any of the new instructions.
5396 This results in code which can run relatively efficiently on either a
5397 68020/68881 or a 68030 or a 68040. The generated code does use the
5398 68881 instructions that are emulated on the 68060.
5402 Generate output containing Sun FPA instructions for floating point.
5405 @opindex msoft-float
5406 Generate output containing library calls for floating point.
5407 @strong{Warning:} the requisite libraries are not available for all m68k
5408 targets. Normally the facilities of the machine's usual C compiler are
5409 used, but this can't be done directly in cross-compilation. You must
5410 make your own arrangements to provide suitable library functions for
5411 cross-compilation. The embedded targets @samp{m68k-*-aout} and
5412 @samp{m68k-*-coff} do provide software floating point support.
5416 Consider type @code{int} to be 16 bits wide, like @code{short int}.
5419 @opindex mnobitfield
5420 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
5421 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
5425 Do use the bit-field instructions. The @option{-m68020} option implies
5426 @option{-mbitfield}. This is the default if you use a configuration
5427 designed for a 68020.
5431 Use a different function-calling convention, in which functions
5432 that take a fixed number of arguments return with the @code{rtd}
5433 instruction, which pops their arguments while returning. This
5434 saves one instruction in the caller since there is no need to pop
5435 the arguments there.
5437 This calling convention is incompatible with the one normally
5438 used on Unix, so you cannot use it if you need to call libraries
5439 compiled with the Unix compiler.
5441 Also, you must provide function prototypes for all functions that
5442 take variable numbers of arguments (including @code{printf});
5443 otherwise incorrect code will be generated for calls to those
5446 In addition, seriously incorrect code will result if you call a
5447 function with too many arguments. (Normally, extra arguments are
5448 harmlessly ignored.)
5450 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
5451 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
5454 @itemx -mno-align-int
5456 @opindex mno-align-int
5457 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
5458 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
5459 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
5460 Aligning variables on 32-bit boundaries produces code that runs somewhat
5461 faster on processors with 32-bit busses at the expense of more memory.
5463 @strong{Warning:} if you use the @option{-malign-int} switch, GCC will
5464 align structures containing the above types differently than
5465 most published application binary interface specifications for the m68k.
5469 Use the pc-relative addressing mode of the 68000 directly, instead of
5470 using a global offset table. At present, this option implies @option{-fpic},
5471 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
5472 not presently supported with @option{-mpcrel}, though this could be supported for
5473 68020 and higher processors.
5475 @item -mno-strict-align
5476 @itemx -mstrict-align
5477 @opindex mno-strict-align
5478 @opindex mstrict-align
5479 Do not (do) assume that unaligned memory references will be handled by
5484 @node M68hc1x Options
5485 @subsection M68hc1x Options
5486 @cindex M68hc1x options
5488 These are the @samp{-m} options defined for the 68hc11 and 68hc12
5489 microcontrollers. The default values for these options depends on
5490 which style of microcontroller was selected when the compiler was configured;
5491 the defaults for the most common choices are given below.
5498 Generate output for a 68HC11. This is the default
5499 when the compiler is configured for 68HC11-based systems.
5505 Generate output for a 68HC12. This is the default
5506 when the compiler is configured for 68HC12-based systems.
5512 Generate output for a 68HCS12.
5515 @opindex mauto-incdec
5516 Enable the use of 68HC12 pre and post auto-increment and auto-decrement
5523 Enable the use of 68HC12 min and max instructions.
5526 @itemx -mno-long-calls
5527 @opindex mlong-calls
5528 @opindex mno-long-calls
5529 Treat all calls as being far away (near). If calls are assumed to be
5530 far away, the compiler will use the @code{call} instruction to
5531 call a function and the @code{rtc} instruction for returning.
5535 Consider type @code{int} to be 16 bits wide, like @code{short int}.
5537 @item -msoft-reg-count=@var{count}
5538 @opindex msoft-reg-count
5539 Specify the number of pseudo-soft registers which are used for the
5540 code generation. The maximum number is 32. Using more pseudo-soft
5541 register may or may not result in better code depending on the program.
5542 The default is 4 for 68HC11 and 2 for 68HC12.
5547 @subsection VAX Options
5550 These @samp{-m} options are defined for the VAX:
5555 Do not output certain jump instructions (@code{aobleq} and so on)
5556 that the Unix assembler for the VAX cannot handle across long
5561 Do output those jump instructions, on the assumption that you
5562 will assemble with the GNU assembler.
5566 Output code for g-format floating point numbers instead of d-format.
5570 @subsection SPARC Options
5571 @cindex SPARC options
5573 These @samp{-m} switches are supported on the SPARC:
5578 @opindex mno-app-regs
5580 Specify @option{-mapp-regs} to generate output using the global registers
5581 2 through 4, which the SPARC SVR4 ABI reserves for applications. This
5584 To be fully SVR4 ABI compliant at the cost of some performance loss,
5585 specify @option{-mno-app-regs}. You should compile libraries and system
5586 software with this option.
5591 @opindex mhard-float
5592 Generate output containing floating point instructions. This is the
5598 @opindex msoft-float
5599 Generate output containing library calls for floating point.
5600 @strong{Warning:} the requisite libraries are not available for all SPARC
5601 targets. Normally the facilities of the machine's usual C compiler are
5602 used, but this cannot be done directly in cross-compilation. You must make
5603 your own arrangements to provide suitable library functions for
5604 cross-compilation. The embedded targets @samp{sparc-*-aout} and
5605 @samp{sparclite-*-*} do provide software floating point support.
5607 @option{-msoft-float} changes the calling convention in the output file;
5608 therefore, it is only useful if you compile @emph{all} of a program with
5609 this option. In particular, you need to compile @file{libgcc.a}, the
5610 library that comes with GCC, with @option{-msoft-float} in order for
5613 @item -mhard-quad-float
5614 @opindex mhard-quad-float
5615 Generate output containing quad-word (long double) floating point
5618 @item -msoft-quad-float
5619 @opindex msoft-quad-float
5620 Generate output containing library calls for quad-word (long double)
5621 floating point instructions. The functions called are those specified
5622 in the SPARC ABI@. This is the default.
5624 As of this writing, there are no sparc implementations that have hardware
5625 support for the quad-word floating point instructions. They all invoke
5626 a trap handler for one of these instructions, and then the trap handler
5627 emulates the effect of the instruction. Because of the trap handler overhead,
5628 this is much slower than calling the ABI library routines. Thus the
5629 @option{-msoft-quad-float} option is the default.
5635 With @option{-mflat}, the compiler does not generate save/restore instructions
5636 and will use a ``flat'' or single register window calling convention.
5637 This model uses %i7 as the frame pointer and is compatible with the normal
5638 register window model. Code from either may be intermixed.
5639 The local registers and the input registers (0--5) are still treated as
5640 ``call saved'' registers and will be saved on the stack as necessary.
5642 With @option{-mno-flat} (the default), the compiler emits save/restore
5643 instructions (except for leaf functions) and is the normal mode of operation.
5645 @item -mno-unaligned-doubles
5646 @itemx -munaligned-doubles
5647 @opindex mno-unaligned-doubles
5648 @opindex munaligned-doubles
5649 Assume that doubles have 8 byte alignment. This is the default.
5651 With @option{-munaligned-doubles}, GCC assumes that doubles have 8 byte
5652 alignment only if they are contained in another type, or if they have an
5653 absolute address. Otherwise, it assumes they have 4 byte alignment.
5654 Specifying this option avoids some rare compatibility problems with code
5655 generated by other compilers. It is not the default because it results
5656 in a performance loss, especially for floating point code.
5658 @item -mno-faster-structs
5659 @itemx -mfaster-structs
5660 @opindex mno-faster-structs
5661 @opindex mfaster-structs
5662 With @option{-mfaster-structs}, the compiler assumes that structures
5663 should have 8 byte alignment. This enables the use of pairs of
5664 @code{ldd} and @code{std} instructions for copies in structure
5665 assignment, in place of twice as many @code{ld} and @code{st} pairs.
5666 However, the use of this changed alignment directly violates the SPARC
5667 ABI@. Thus, it's intended only for use on targets where the developer
5668 acknowledges that their resulting code will not be directly in line with
5669 the rules of the ABI@.
5675 These two options select variations on the SPARC architecture.
5677 By default (unless specifically configured for the Fujitsu SPARClite),
5678 GCC generates code for the v7 variant of the SPARC architecture.
5680 @option{-mv8} will give you SPARC v8 code. The only difference from v7
5681 code is that the compiler emits the integer multiply and integer
5682 divide instructions which exist in SPARC v8 but not in SPARC v7.
5684 @option{-msparclite} will give you SPARClite code. This adds the integer
5685 multiply, integer divide step and scan (@code{ffs}) instructions which
5686 exist in SPARClite but not in SPARC v7.
5688 These options are deprecated and will be deleted in a future GCC release.
5689 They have been replaced with @option{-mcpu=xxx}.
5694 @opindex msupersparc
5695 These two options select the processor for which the code is optimized.
5697 With @option{-mcypress} (the default), the compiler optimizes code for the
5698 Cypress CY7C602 chip, as used in the SPARCStation/SPARCServer 3xx series.
5699 This is also appropriate for the older SPARCStation 1, 2, IPX etc.
5701 With @option{-msupersparc} the compiler optimizes code for the SuperSPARC cpu, as
5702 used in the SPARCStation 10, 1000 and 2000 series. This flag also enables use
5703 of the full SPARC v8 instruction set.
5705 These options are deprecated and will be deleted in a future GCC release.
5706 They have been replaced with @option{-mcpu=xxx}.
5708 @item -mcpu=@var{cpu_type}
5710 Set the instruction set, register set, and instruction scheduling parameters
5711 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
5712 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite},
5713 @samp{hypersparc}, @samp{sparclite86x}, @samp{f930}, @samp{f934},
5714 @samp{sparclet}, @samp{tsc701}, @samp{v9}, @samp{ultrasparc}, and
5717 Default instruction scheduling parameters are used for values that select
5718 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
5719 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
5721 Here is a list of each supported architecture and their supported
5726 v8: supersparc, hypersparc
5727 sparclite: f930, f934, sparclite86x
5729 v9: ultrasparc, ultrasparc3
5732 @item -mtune=@var{cpu_type}
5734 Set the instruction scheduling parameters for machine type
5735 @var{cpu_type}, but do not set the instruction set or register set that the
5736 option @option{-mcpu=@var{cpu_type}} would.
5738 The same values for @option{-mcpu=@var{cpu_type}} can be used for
5739 @option{-mtune=@var{cpu_type}}, but the only useful values are those
5740 that select a particular cpu implementation. Those are @samp{cypress},
5741 @samp{supersparc}, @samp{hypersparc}, @samp{f930}, @samp{f934},
5742 @samp{sparclite86x}, @samp{tsc701}, @samp{ultrasparc}, and
5747 These @samp{-m} switches are supported in addition to the above
5748 on the SPARCLET processor.
5751 @item -mlittle-endian
5752 @opindex mlittle-endian
5753 Generate code for a processor running in little-endian mode.
5757 Treat register @code{%g0} as a normal register.
5758 GCC will continue to clobber it as necessary but will not assume
5759 it always reads as 0.
5761 @item -mbroken-saverestore
5762 @opindex mbroken-saverestore
5763 Generate code that does not use non-trivial forms of the @code{save} and
5764 @code{restore} instructions. Early versions of the SPARCLET processor do
5765 not correctly handle @code{save} and @code{restore} instructions used with
5766 arguments. They correctly handle them used without arguments. A @code{save}
5767 instruction used without arguments increments the current window pointer
5768 but does not allocate a new stack frame. It is assumed that the window
5769 overflow trap handler will properly handle this case as will interrupt
5773 These @samp{-m} switches are supported in addition to the above
5774 on SPARC V9 processors in 64-bit environments.
5777 @item -mlittle-endian
5778 @opindex mlittle-endian
5779 Generate code for a processor running in little-endian mode.
5785 Generate code for a 32-bit or 64-bit environment.
5786 The 32-bit environment sets int, long and pointer to 32 bits.
5787 The 64-bit environment sets int to 32 bits and long and pointer
5790 @item -mcmodel=medlow
5791 @opindex mcmodel=medlow
5792 Generate code for the Medium/Low code model: the program must be linked
5793 in the low 32 bits of the address space. Pointers are 64 bits.
5794 Programs can be statically or dynamically linked.
5796 @item -mcmodel=medmid
5797 @opindex mcmodel=medmid
5798 Generate code for the Medium/Middle code model: the program must be linked
5799 in the low 44 bits of the address space, the text segment must be less than
5800 2G bytes, and data segment must be within 2G of the text segment.
5801 Pointers are 64 bits.
5803 @item -mcmodel=medany
5804 @opindex mcmodel=medany
5805 Generate code for the Medium/Anywhere code model: the program may be linked
5806 anywhere in the address space, the text segment must be less than
5807 2G bytes, and data segment must be within 2G of the text segment.
5808 Pointers are 64 bits.
5810 @item -mcmodel=embmedany
5811 @opindex mcmodel=embmedany
5812 Generate code for the Medium/Anywhere code model for embedded systems:
5813 assume a 32-bit text and a 32-bit data segment, both starting anywhere
5814 (determined at link time). Register %g4 points to the base of the
5815 data segment. Pointers are still 64 bits.
5816 Programs are statically linked, PIC is not supported.
5819 @itemx -mno-stack-bias
5820 @opindex mstack-bias
5821 @opindex mno-stack-bias
5822 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
5823 frame pointer if present, are offset by @minus{}2047 which must be added back
5824 when making stack frame references.
5825 Otherwise, assume no such offset is present.
5829 @subsection ARM Options
5832 These @samp{-m} options are defined for Advanced RISC Machines (ARM)
5837 @opindex mapcs-frame
5838 Generate a stack frame that is compliant with the ARM Procedure Call
5839 Standard for all functions, even if this is not strictly necessary for
5840 correct execution of the code. Specifying @option{-fomit-frame-pointer}
5841 with this option will cause the stack frames not to be generated for
5842 leaf functions. The default is @option{-mno-apcs-frame}.
5846 This is a synonym for @option{-mapcs-frame}.
5850 Generate code for a processor running with a 26-bit program counter,
5851 and conforming to the function calling standards for the APCS 26-bit
5852 option. This option replaces the @option{-m2} and @option{-m3} options
5853 of previous releases of the compiler.
5857 Generate code for a processor running with a 32-bit program counter,
5858 and conforming to the function calling standards for the APCS 32-bit
5859 option. This option replaces the @option{-m6} option of previous releases
5863 @c not currently implemented
5864 @item -mapcs-stack-check
5865 @opindex mapcs-stack-check
5866 Generate code to check the amount of stack space available upon entry to
5867 every function (that actually uses some stack space). If there is
5868 insufficient space available then either the function
5869 @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be
5870 called, depending upon the amount of stack space required. The run time
5871 system is required to provide these functions. The default is
5872 @option{-mno-apcs-stack-check}, since this produces smaller code.
5874 @c not currently implemented
5876 @opindex mapcs-float
5877 Pass floating point arguments using the float point registers. This is
5878 one of the variants of the APCS@. This option is recommended if the
5879 target hardware has a floating point unit or if a lot of floating point
5880 arithmetic is going to be performed by the code. The default is
5881 @option{-mno-apcs-float}, since integer only code is slightly increased in
5882 size if @option{-mapcs-float} is used.
5884 @c not currently implemented
5885 @item -mapcs-reentrant
5886 @opindex mapcs-reentrant
5887 Generate reentrant, position independent code. The default is
5888 @option{-mno-apcs-reentrant}.
5891 @item -mthumb-interwork
5892 @opindex mthumb-interwork
5893 Generate code which supports calling between the ARM and Thumb
5894 instruction sets. Without this option the two instruction sets cannot
5895 be reliably used inside one program. The default is
5896 @option{-mno-thumb-interwork}, since slightly larger code is generated
5897 when @option{-mthumb-interwork} is specified.
5899 @item -mno-sched-prolog
5900 @opindex mno-sched-prolog
5901 Prevent the reordering of instructions in the function prolog, or the
5902 merging of those instruction with the instructions in the function's
5903 body. This means that all functions will start with a recognizable set
5904 of instructions (or in fact one of a choice from a small set of
5905 different function prologues), and this information can be used to
5906 locate the start if functions inside an executable piece of code. The
5907 default is @option{-msched-prolog}.
5910 @opindex mhard-float
5911 Generate output containing floating point instructions. This is the
5915 @opindex msoft-float
5916 Generate output containing library calls for floating point.
5917 @strong{Warning:} the requisite libraries are not available for all ARM
5918 targets. Normally the facilities of the machine's usual C compiler are
5919 used, but this cannot be done directly in cross-compilation. You must make
5920 your own arrangements to provide suitable library functions for
5923 @option{-msoft-float} changes the calling convention in the output file;
5924 therefore, it is only useful if you compile @emph{all} of a program with
5925 this option. In particular, you need to compile @file{libgcc.a}, the
5926 library that comes with GCC, with @option{-msoft-float} in order for
5929 @item -mlittle-endian
5930 @opindex mlittle-endian
5931 Generate code for a processor running in little-endian mode. This is
5932 the default for all standard configurations.
5935 @opindex mbig-endian
5936 Generate code for a processor running in big-endian mode; the default is
5937 to compile code for a little-endian processor.
5939 @item -mwords-little-endian
5940 @opindex mwords-little-endian
5941 This option only applies when generating code for big-endian processors.
5942 Generate code for a little-endian word order but a big-endian byte
5943 order. That is, a byte order of the form @samp{32107654}. Note: this
5944 option should only be used if you require compatibility with code for
5945 big-endian ARM processors generated by versions of the compiler prior to
5948 @item -malignment-traps
5949 @opindex malignment-traps
5950 Generate code that will not trap if the MMU has alignment traps enabled.
5951 On ARM architectures prior to ARMv4, there were no instructions to
5952 access half-word objects stored in memory. However, when reading from
5953 memory a feature of the ARM architecture allows a word load to be used,
5954 even if the address is unaligned, and the processor core will rotate the
5955 data as it is being loaded. This option tells the compiler that such
5956 misaligned accesses will cause a MMU trap and that it should instead
5957 synthesize the access as a series of byte accesses. The compiler can
5958 still use word accesses to load half-word data if it knows that the
5959 address is aligned to a word boundary.
5961 This option is ignored when compiling for ARM architecture 4 or later,
5962 since these processors have instructions to directly access half-word
5965 @item -mno-alignment-traps
5966 @opindex mno-alignment-traps
5967 Generate code that assumes that the MMU will not trap unaligned
5968 accesses. This produces better code when the target instruction set
5969 does not have half-word memory operations (i.e.@: implementations prior to
5972 Note that you cannot use this option to access unaligned word objects,
5973 since the processor will only fetch one 32-bit aligned object from
5976 The default setting for most targets is @option{-mno-alignment-traps}, since
5977 this produces better code when there are no half-word memory
5978 instructions available.
5980 @item -mshort-load-bytes
5981 @itemx -mno-short-load-words
5982 @opindex mshort-load-bytes
5983 @opindex mno-short-load-words
5984 These are deprecated aliases for @option{-malignment-traps}.
5986 @item -mno-short-load-bytes
5987 @itemx -mshort-load-words
5988 @opindex mno-short-load-bytes
5989 @opindex mshort-load-words
5990 This are deprecated aliases for @option{-mno-alignment-traps}.
5992 @item -mcpu=@var{name}
5994 This specifies the name of the target ARM processor. GCC uses this name
5995 to determine what kind of instructions it can emit when generating
5996 assembly code. Permissible names are: @samp{arm2}, @samp{arm250},
5997 @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610},
5998 @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm},
5999 @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700},
6000 @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100},
6001 @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm8},
6002 @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100},
6003 @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920},
6004 @samp{arm920t}, @samp{arm940t}, @samp{arm9tdmi}, @samp{arm10tdmi},
6005 @samp{arm1020t}, @samp{xscale}.
6007 @itemx -mtune=@var{name}
6009 This option is very similar to the @option{-mcpu=} option, except that
6010 instead of specifying the actual target processor type, and hence
6011 restricting which instructions can be used, it specifies that GCC should
6012 tune the performance of the code as if the target were of the type
6013 specified in this option, but still choosing the instructions that it
6014 will generate based on the cpu specified by a @option{-mcpu=} option.
6015 For some ARM implementations better performance can be obtained by using
6018 @item -march=@var{name}
6020 This specifies the name of the target ARM architecture. GCC uses this
6021 name to determine what kind of instructions it can emit when generating
6022 assembly code. This option can be used in conjunction with or instead
6023 of the @option{-mcpu=} option. Permissible names are: @samp{armv2},
6024 @samp{armv2a}, @samp{armv3}, @samp{armv3m}, @samp{armv4}, @samp{armv4t},
6025 @samp{armv5}, @samp{armv5t}, @samp{armv5te}.
6027 @item -mfpe=@var{number}
6028 @itemx -mfp=@var{number}
6031 This specifies the version of the floating point emulation available on
6032 the target. Permissible values are 2 and 3. @option{-mfp=} is a synonym
6033 for @option{-mfpe=}, for compatibility with older versions of GCC@.
6035 @item -mstructure-size-boundary=@var{n}
6036 @opindex mstructure-size-boundary
6037 The size of all structures and unions will be rounded up to a multiple
6038 of the number of bits set by this option. Permissible values are 8 and
6039 32. The default value varies for different toolchains. For the COFF
6040 targeted toolchain the default value is 8. Specifying the larger number
6041 can produce faster, more efficient code, but can also increase the size
6042 of the program. The two values are potentially incompatible. Code
6043 compiled with one value cannot necessarily expect to work with code or
6044 libraries compiled with the other value, if they exchange information
6045 using structures or unions.
6047 @item -mabort-on-noreturn
6048 @opindex mabort-on-noreturn
6049 Generate a call to the function @code{abort} at the end of a
6050 @code{noreturn} function. It will be executed if the function tries to
6054 @itemx -mno-long-calls
6055 @opindex mlong-calls
6056 @opindex mno-long-calls
6057 Tells the compiler to perform function calls by first loading the
6058 address of the function into a register and then performing a subroutine
6059 call on this register. This switch is needed if the target function
6060 will lie outside of the 64 megabyte addressing range of the offset based
6061 version of subroutine call instruction.
6063 Even if this switch is enabled, not all function calls will be turned
6064 into long calls. The heuristic is that static functions, functions
6065 which have the @samp{short-call} attribute, functions that are inside
6066 the scope of a @samp{#pragma no_long_calls} directive and functions whose
6067 definitions have already been compiled within the current compilation
6068 unit, will not be turned into long calls. The exception to this rule is
6069 that weak function definitions, functions with the @samp{long-call}
6070 attribute or the @samp{section} attribute, and functions that are within
6071 the scope of a @samp{#pragma long_calls} directive, will always be
6072 turned into long calls.
6074 This feature is not enabled by default. Specifying
6075 @option{-mno-long-calls} will restore the default behavior, as will
6076 placing the function calls within the scope of a @samp{#pragma
6077 long_calls_off} directive. Note these switches have no effect on how
6078 the compiler generates code to handle function calls via function
6081 @item -mnop-fun-dllimport
6082 @opindex mnop-fun-dllimport
6083 Disable support for the @code{dllimport} attribute.
6085 @item -msingle-pic-base
6086 @opindex msingle-pic-base
6087 Treat the register used for PIC addressing as read-only, rather than
6088 loading it in the prologue for each function. The run-time system is
6089 responsible for initializing this register with an appropriate value
6090 before execution begins.
6092 @item -mpic-register=@var{reg}
6093 @opindex mpic-register
6094 Specify the register to be used for PIC addressing. The default is R10
6095 unless stack-checking is enabled, when R9 is used.
6097 @item -mpoke-function-name
6098 @opindex mpoke-function-name
6099 Write the name of each function into the text section, directly
6100 preceding the function prologue. The generated code is similar to this:
6104 .ascii "arm_poke_function_name", 0
6107 .word 0xff000000 + (t1 - t0)
6108 arm_poke_function_name
6110 stmfd sp!, @{fp, ip, lr, pc@}
6114 When performing a stack backtrace, code can inspect the value of
6115 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
6116 location @code{pc - 12} and the top 8 bits are set, then we know that
6117 there is a function name embedded immediately preceding this location
6118 and has length @code{((pc[-3]) & 0xff000000)}.
6122 Generate code for the 16-bit Thumb instruction set. The default is to
6123 use the 32-bit ARM instruction set.
6126 @opindex mtpcs-frame
6127 Generate a stack frame that is compliant with the Thumb Procedure Call
6128 Standard for all non-leaf functions. (A leaf function is one that does
6129 not call any other functions.) The default is @option{-mno-tpcs-frame}.
6131 @item -mtpcs-leaf-frame
6132 @opindex mtpcs-leaf-frame
6133 Generate a stack frame that is compliant with the Thumb Procedure Call
6134 Standard for all leaf functions. (A leaf function is one that does
6135 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
6137 @item -mcallee-super-interworking
6138 @opindex mcallee-super-interworking
6139 Gives all externally visible functions in the file being compiled an ARM
6140 instruction set header which switches to Thumb mode before executing the
6141 rest of the function. This allows these functions to be called from
6142 non-interworking code.
6144 @item -mcaller-super-interworking
6145 @opindex mcaller-super-interworking
6146 Allows calls via function pointers (including virtual functions) to
6147 execute correctly regardless of whether the target code has been
6148 compiled for interworking or not. There is a small overhead in the cost
6149 of executing a function pointer if this option is enabled.
6153 @node MN10200 Options
6154 @subsection MN10200 Options
6155 @cindex MN10200 options
6156 These @option{-m} options are defined for Matsushita MN10200 architectures:
6161 Indicate to the linker that it should perform a relaxation optimization pass
6162 to shorten branches, calls and absolute memory addresses. This option only
6163 has an effect when used on the command line for the final link step.
6165 This option makes symbolic debugging impossible.
6168 @node MN10300 Options
6169 @subsection MN10300 Options
6170 @cindex MN10300 options
6171 These @option{-m} options are defined for Matsushita MN10300 architectures:
6176 Generate code to avoid bugs in the multiply instructions for the MN10300
6177 processors. This is the default.
6180 @opindex mno-mult-bug
6181 Do not generate code to avoid bugs in the multiply instructions for the
6186 Generate code which uses features specific to the AM33 processor.
6190 Do not generate code which uses features specific to the AM33 processor. This
6195 Do not link in the C run-time initialization object file.
6199 Indicate to the linker that it should perform a relaxation optimization pass
6200 to shorten branches, calls and absolute memory addresses. This option only
6201 has an effect when used on the command line for the final link step.
6203 This option makes symbolic debugging impossible.
6207 @node M32R/D Options
6208 @subsection M32R/D Options
6209 @cindex M32R/D options
6211 These @option{-m} options are defined for Mitsubishi M32R/D architectures:
6216 Generate code for the M32R/X@.
6220 Generate code for the M32R@. This is the default.
6222 @item -mcode-model=small
6223 @opindex mcode-model=small
6224 Assume all objects live in the lower 16MB of memory (so that their addresses
6225 can be loaded with the @code{ld24} instruction), and assume all subroutines
6226 are reachable with the @code{bl} instruction.
6227 This is the default.
6229 The addressability of a particular object can be set with the
6230 @code{model} attribute.
6232 @item -mcode-model=medium
6233 @opindex mcode-model=medium
6234 Assume objects may be anywhere in the 32-bit address space (the compiler
6235 will generate @code{seth/add3} instructions to load their addresses), and
6236 assume all subroutines are reachable with the @code{bl} instruction.
6238 @item -mcode-model=large
6239 @opindex mcode-model=large
6240 Assume objects may be anywhere in the 32-bit address space (the compiler
6241 will generate @code{seth/add3} instructions to load their addresses), and
6242 assume subroutines may not be reachable with the @code{bl} instruction
6243 (the compiler will generate the much slower @code{seth/add3/jl}
6244 instruction sequence).
6247 @opindex msdata=none
6248 Disable use of the small data area. Variables will be put into
6249 one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the
6250 @code{section} attribute has been specified).
6251 This is the default.
6253 The small data area consists of sections @samp{.sdata} and @samp{.sbss}.
6254 Objects may be explicitly put in the small data area with the
6255 @code{section} attribute using one of these sections.
6258 @opindex msdata=sdata
6259 Put small global and static data in the small data area, but do not
6260 generate special code to reference them.
6264 Put small global and static data in the small data area, and generate
6265 special instructions to reference them.
6269 @cindex smaller data references
6270 Put global and static objects less than or equal to @var{num} bytes
6271 into the small data or bss sections instead of the normal data or bss
6272 sections. The default value of @var{num} is 8.
6273 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
6274 for this option to have any effect.
6276 All modules should be compiled with the same @option{-G @var{num}} value.
6277 Compiling with different values of @var{num} may or may not work; if it
6278 doesn't the linker will give an error message---incorrect code will not be
6284 @subsection M88K Options
6285 @cindex M88k options
6287 These @samp{-m} options are defined for Motorola 88k architectures:
6292 Generate code that works well on both the m88100 and the
6297 Generate code that works best for the m88100, but that also
6302 Generate code that works best for the m88110, and may not run
6307 Obsolete option to be removed from the next revision.
6310 @item -midentify-revision
6311 @opindex midentify-revision
6312 @cindex identifying source, compiler (88k)
6313 Include an @code{ident} directive in the assembler output recording the
6314 source file name, compiler name and version, timestamp, and compilation
6317 @item -mno-underscores
6318 @opindex mno-underscores
6319 @cindex underscores, avoiding (88k)
6320 In assembler output, emit symbol names without adding an underscore
6321 character at the beginning of each name. The default is to use an
6322 underscore as prefix on each name.
6324 @item -mocs-debug-info
6325 @itemx -mno-ocs-debug-info
6326 @opindex mocs-debug-info
6327 @opindex mno-ocs-debug-info
6329 @cindex debugging, 88k OCS
6330 Include (or omit) additional debugging information (about registers used
6331 in each stack frame) as specified in the 88open Object Compatibility
6332 Standard, ``OCS''@. This extra information allows debugging of code that
6333 has had the frame pointer eliminated. The default for SVr4 and Delta 88
6334 SVr3.2 is to include this information; other 88k configurations omit this
6335 information by default.
6337 @item -mocs-frame-position
6338 @opindex mocs-frame-position
6339 @cindex register positions in frame (88k)
6340 When emitting COFF debugging information for automatic variables and
6341 parameters stored on the stack, use the offset from the canonical frame
6342 address, which is the stack pointer (register 31) on entry to the
6343 function. The SVr4 and Delta88 SVr3.2, and BCS configurations use
6344 @option{-mocs-frame-position}; other 88k configurations have the default
6345 @option{-mno-ocs-frame-position}.
6347 @item -mno-ocs-frame-position
6348 @opindex mno-ocs-frame-position
6349 @cindex register positions in frame (88k)
6350 When emitting COFF debugging information for automatic variables and
6351 parameters stored on the stack, use the offset from the frame pointer
6352 register (register 30). When this option is in effect, the frame
6353 pointer is not eliminated when debugging information is selected by the
6356 @item -moptimize-arg-area
6357 @opindex moptimize-arg-area
6358 @cindex arguments in frame (88k)
6359 Save space by reorganizing the stack frame. This option generates code
6360 that does not agree with the 88open specifications, but uses less
6363 @itemx -mno-optimize-arg-area
6364 @opindex mno-optimize-arg-area
6365 Do not reorganize the stack frame to save space. This is the default.
6366 The generated conforms to the specification, but uses more memory.
6368 @item -mshort-data-@var{num}
6369 @opindex mshort-data
6370 @cindex smaller data references (88k)
6371 @cindex r0-relative references (88k)
6372 Generate smaller data references by making them relative to @code{r0},
6373 which allows loading a value using a single instruction (rather than the
6374 usual two). You control which data references are affected by
6375 specifying @var{num} with this option. For example, if you specify
6376 @option{-mshort-data-512}, then the data references affected are those
6377 involving displacements of less than 512 bytes.
6378 @option{-mshort-data-@var{num}} is not effective for @var{num} greater
6381 @item -mserialize-volatile
6382 @opindex mserialize-volatile
6383 @itemx -mno-serialize-volatile
6384 @opindex mno-serialize-volatile
6385 @cindex sequential consistency on 88k
6386 Do, or don't, generate code to guarantee sequential consistency
6387 of volatile memory references. By default, consistency is
6390 The order of memory references made by the MC88110 processor does
6391 not always match the order of the instructions requesting those
6392 references. In particular, a load instruction may execute before
6393 a preceding store instruction. Such reordering violates
6394 sequential consistency of volatile memory references, when there
6395 are multiple processors. When consistency must be guaranteed,
6396 GCC generates special instructions, as needed, to force
6397 execution in the proper order.
6399 The MC88100 processor does not reorder memory references and so
6400 always provides sequential consistency. However, by default, GCC
6401 generates the special instructions to guarantee consistency
6402 even when you use @option{-m88100}, so that the code may be run on an
6403 MC88110 processor. If you intend to run your code only on the
6404 MC88100 processor, you may use @option{-mno-serialize-volatile}.
6406 The extra code generated to guarantee consistency may affect the
6407 performance of your application. If you know that you can safely
6408 forgo this guarantee, you may use @option{-mno-serialize-volatile}.
6414 @cindex assembler syntax, 88k
6416 Turn on (@option{-msvr4}) or off (@option{-msvr3}) compiler extensions
6417 related to System V release 4 (SVr4). This controls the following:
6421 Which variant of the assembler syntax to emit.
6423 @option{-msvr4} makes the C preprocessor recognize @samp{#pragma weak}
6424 that is used on System V release 4.
6426 @option{-msvr4} makes GCC issue additional declaration directives used in
6430 @option{-msvr4} is the default for the m88k-motorola-sysv4 configuration.
6431 @option{-msvr3} is the default for all other m88k configurations.
6433 @item -mversion-03.00
6434 @opindex mversion-03.00
6435 This option is obsolete, and is ignored.
6436 @c ??? which asm syntax better for GAS? option there too?
6438 @item -mno-check-zero-division
6439 @itemx -mcheck-zero-division
6440 @opindex mno-check-zero-division
6441 @opindex mcheck-zero-division
6442 @cindex zero division on 88k
6443 Do, or don't, generate code to guarantee that integer division by
6444 zero will be detected. By default, detection is guaranteed.
6446 Some models of the MC88100 processor fail to trap upon integer
6447 division by zero under certain conditions. By default, when
6448 compiling code that might be run on such a processor, GCC
6449 generates code that explicitly checks for zero-valued divisors
6450 and traps with exception number 503 when one is detected. Use of
6451 @option{-mno-check-zero-division} suppresses such checking for code
6452 generated to run on an MC88100 processor.
6454 GCC assumes that the MC88110 processor correctly detects all instances
6455 of integer division by zero. When @option{-m88110} is specified, no
6456 explicit checks for zero-valued divisors are generated, and both
6457 @option{-mcheck-zero-division} and @option{-mno-check-zero-division} are
6460 @item -muse-div-instruction
6461 @opindex muse-div-instruction
6462 @cindex divide instruction, 88k
6463 Use the div instruction for signed integer division on the
6464 MC88100 processor. By default, the div instruction is not used.
6466 On the MC88100 processor the signed integer division instruction
6467 div) traps to the operating system on a negative operand. The
6468 operating system transparently completes the operation, but at a
6469 large cost in execution time. By default, when compiling code
6470 that might be run on an MC88100 processor, GCC emulates signed
6471 integer division using the unsigned integer division instruction
6472 divu), thereby avoiding the large penalty of a trap to the
6473 operating system. Such emulation has its own, smaller, execution
6474 cost in both time and space. To the extent that your code's
6475 important signed integer division operations are performed on two
6476 nonnegative operands, it may be desirable to use the div
6477 instruction directly.
6479 On the MC88110 processor the div instruction (also known as the
6480 divs instruction) processes negative operands without trapping to
6481 the operating system. When @option{-m88110} is specified,
6482 @option{-muse-div-instruction} is ignored, and the div instruction is used
6483 for signed integer division.
6485 Note that the result of dividing @code{INT_MIN} by @minus{}1 is undefined. In
6486 particular, the behavior of such a division with and without
6487 @option{-muse-div-instruction} may differ.
6489 @item -mtrap-large-shift
6490 @itemx -mhandle-large-shift
6491 @opindex mtrap-large-shift
6492 @opindex mhandle-large-shift
6493 @cindex bit shift overflow (88k)
6494 @cindex large bit shifts (88k)
6495 Include code to detect bit-shifts of more than 31 bits; respectively,
6496 trap such shifts or emit code to handle them properly. By default GCC
6497 makes no special provision for large bit shifts.
6499 @item -mwarn-passed-structs
6500 @opindex mwarn-passed-structs
6501 @cindex structure passing (88k)
6502 Warn when a function passes a struct as an argument or result.
6503 Structure-passing conventions have changed during the evolution of the C
6504 language, and are often the source of portability problems. By default,
6505 GCC issues no such warning.
6508 @c break page here to avoid unsightly interparagraph stretch.
6512 @node RS/6000 and PowerPC Options
6513 @subsection IBM RS/6000 and PowerPC Options
6514 @cindex RS/6000 and PowerPC Options
6515 @cindex IBM RS/6000 and PowerPC Options
6517 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
6525 @itemx -mpowerpc-gpopt
6526 @itemx -mno-powerpc-gpopt
6527 @itemx -mpowerpc-gfxopt
6528 @itemx -mno-powerpc-gfxopt
6530 @itemx -mno-powerpc64
6536 @opindex mno-powerpc
6537 @opindex mpowerpc-gpopt
6538 @opindex mno-powerpc-gpopt
6539 @opindex mpowerpc-gfxopt
6540 @opindex mno-powerpc-gfxopt
6542 @opindex mno-powerpc64
6543 GCC supports two related instruction set architectures for the
6544 RS/6000 and PowerPC@. The @dfn{POWER} instruction set are those
6545 instructions supported by the @samp{rios} chip set used in the original
6546 RS/6000 systems and the @dfn{PowerPC} instruction set is the
6547 architecture of the Motorola MPC5xx, MPC6xx, MPC8xx microprocessors, and
6548 the IBM 4xx microprocessors.
6550 Neither architecture is a subset of the other. However there is a
6551 large common subset of instructions supported by both. An MQ
6552 register is included in processors supporting the POWER architecture.
6554 You use these options to specify which instructions are available on the
6555 processor you are using. The default value of these options is
6556 determined when configuring GCC@. Specifying the
6557 @option{-mcpu=@var{cpu_type}} overrides the specification of these
6558 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
6559 rather than the options listed above.
6561 The @option{-mpower} option allows GCC to generate instructions that
6562 are found only in the POWER architecture and to use the MQ register.
6563 Specifying @option{-mpower2} implies @option{-power} and also allows GCC
6564 to generate instructions that are present in the POWER2 architecture but
6565 not the original POWER architecture.
6567 The @option{-mpowerpc} option allows GCC to generate instructions that
6568 are found only in the 32-bit subset of the PowerPC architecture.
6569 Specifying @option{-mpowerpc-gpopt} implies @option{-mpowerpc} and also allows
6570 GCC to use the optional PowerPC architecture instructions in the
6571 General Purpose group, including floating-point square root. Specifying
6572 @option{-mpowerpc-gfxopt} implies @option{-mpowerpc} and also allows GCC to
6573 use the optional PowerPC architecture instructions in the Graphics
6574 group, including floating-point select.
6576 The @option{-mpowerpc64} option allows GCC to generate the additional
6577 64-bit instructions that are found in the full PowerPC64 architecture
6578 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
6579 @option{-mno-powerpc64}.
6581 If you specify both @option{-mno-power} and @option{-mno-powerpc}, GCC
6582 will use only the instructions in the common subset of both
6583 architectures plus some special AIX common-mode calls, and will not use
6584 the MQ register. Specifying both @option{-mpower} and @option{-mpowerpc}
6585 permits GCC to use any instruction from either architecture and to
6586 allow use of the MQ register; specify this for the Motorola MPC601.
6588 @item -mnew-mnemonics
6589 @itemx -mold-mnemonics
6590 @opindex mnew-mnemonics
6591 @opindex mold-mnemonics
6592 Select which mnemonics to use in the generated assembler code. With
6593 @option{-mnew-mnemonics}, GCC uses the assembler mnemonics defined for
6594 the PowerPC architecture. With @option{-mold-mnemonics} it uses the
6595 assembler mnemonics defined for the POWER architecture. Instructions
6596 defined in only one architecture have only one mnemonic; GCC uses that
6597 mnemonic irrespective of which of these options is specified.
6599 GCC defaults to the mnemonics appropriate for the architecture in
6600 use. Specifying @option{-mcpu=@var{cpu_type}} sometimes overrides the
6601 value of these option. Unless you are building a cross-compiler, you
6602 should normally not specify either @option{-mnew-mnemonics} or
6603 @option{-mold-mnemonics}, but should instead accept the default.
6605 @item -mcpu=@var{cpu_type}
6607 Set architecture type, register usage, choice of mnemonics, and
6608 instruction scheduling parameters for machine type @var{cpu_type}.
6609 Supported values for @var{cpu_type} are @samp{rios}, @samp{rios1},
6610 @samp{rsc}, @samp{rios2}, @samp{rs64a}, @samp{601}, @samp{602},
6611 @samp{603}, @samp{603e}, @samp{604}, @samp{604e}, @samp{620},
6612 @samp{630}, @samp{740}, @samp{7400}, @samp{7450}, @samp{750},
6613 @samp{power}, @samp{power2}, @samp{powerpc}, @samp{403}, @samp{505},
6614 @samp{801}, @samp{821}, @samp{823}, and @samp{860} and @samp{common}.
6616 @option{-mcpu=common} selects a completely generic processor. Code
6617 generated under this option will run on any POWER or PowerPC processor.
6618 GCC will use only the instructions in the common subset of both
6619 architectures, and will not use the MQ register. GCC assumes a generic
6620 processor model for scheduling purposes.
6622 @option{-mcpu=power}, @option{-mcpu=power2}, @option{-mcpu=powerpc}, and
6623 @option{-mcpu=powerpc64} specify generic POWER, POWER2, pure 32-bit
6624 PowerPC (i.e., not MPC601), and 64-bit PowerPC architecture machine
6625 types, with an appropriate, generic processor model assumed for
6626 scheduling purposes.
6628 The other options specify a specific processor. Code generated under
6629 those options will run best on that processor, and may not run at all on
6632 The @option{-mcpu} options automatically enable or disable other
6633 @option{-m} options as follows:
6637 @option{-mno-power}, @option{-mno-powerpc}
6644 @option{-mpower}, @option{-mno-powerpc}, @option{-mno-new-mnemonics}
6659 @option{-mno-power}, @option{-mpowerpc}, @option{-mnew-mnemonics}
6662 @option{-mpower}, @option{-mpowerpc}, @option{-mnew-mnemonics}
6667 @option{-mno-power}, @option{-mpowerpc}, @option{-mnew-mnemonics}, @option{-msoft-float}
6670 @item -mtune=@var{cpu_type}
6672 Set the instruction scheduling parameters for machine type
6673 @var{cpu_type}, but do not set the architecture type, register usage, or
6674 choice of mnemonics, as @option{-mcpu=@var{cpu_type}} would. The same
6675 values for @var{cpu_type} are used for @option{-mtune} as for
6676 @option{-mcpu}. If both are specified, the code generated will use the
6677 architecture, registers, and mnemonics set by @option{-mcpu}, but the
6678 scheduling parameters set by @option{-mtune}.
6683 @opindex mno-altivec
6684 These switches enable or disable the use of built-in functions that
6685 allow access to the AltiVec instruction set. You may also need to set
6686 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
6691 Extend the current ABI with SPE ABI extensions. This does not change
6692 the default ABI, instead it adds the SPE ABI extensions to the current
6696 @opindex mabi=no-spe
6697 Disable Booke SPE ABI extensions for the current ABI.
6699 @item -misel=@var{yes/no}
6702 This switch enables or disables the generation of ISEL instructions.
6705 @itemx -mno-fp-in-toc
6706 @itemx -mno-sum-in-toc
6707 @itemx -mminimal-toc
6709 @opindex mno-fp-in-toc
6710 @opindex mno-sum-in-toc
6711 @opindex mminimal-toc
6712 Modify generation of the TOC (Table Of Contents), which is created for
6713 every executable file. The @option{-mfull-toc} option is selected by
6714 default. In that case, GCC will allocate at least one TOC entry for
6715 each unique non-automatic variable reference in your program. GCC
6716 will also place floating-point constants in the TOC@. However, only
6717 16,384 entries are available in the TOC@.
6719 If you receive a linker error message that saying you have overflowed
6720 the available TOC space, you can reduce the amount of TOC space used
6721 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
6722 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
6723 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
6724 generate code to calculate the sum of an address and a constant at
6725 run-time instead of putting that sum into the TOC@. You may specify one
6726 or both of these options. Each causes GCC to produce very slightly
6727 slower and larger code at the expense of conserving TOC space.
6729 If you still run out of space in the TOC even when you specify both of
6730 these options, specify @option{-mminimal-toc} instead. This option causes
6731 GCC to make only one TOC entry for every file. When you specify this
6732 option, GCC will produce code that is slower and larger but which
6733 uses extremely little TOC space. You may wish to use this option
6734 only on files that contain less frequently executed code.
6740 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
6741 @code{long} type, and the infrastructure needed to support them.
6742 Specifying @option{-maix64} implies @option{-mpowerpc64} and
6743 @option{-mpowerpc}, while @option{-maix32} disables the 64-bit ABI and
6744 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
6749 @opindex mno-xl-call
6750 On AIX, pass floating-point arguments to prototyped functions beyond the
6751 register save area (RSA) on the stack in addition to argument FPRs. The
6752 AIX calling convention was extended but not initially documented to
6753 handle an obscure K&R C case of calling a function that takes the
6754 address of its arguments with fewer arguments than declared. AIX XL
6755 compilers access floating point arguments which do not fit in the
6756 RSA from the stack when a subroutine is compiled without
6757 optimization. Because always storing floating-point arguments on the
6758 stack is inefficient and rarely needed, this option is not enabled by
6759 default and only is necessary when calling subroutines compiled by AIX
6760 XL compilers without optimization.
6764 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
6765 application written to use message passing with special startup code to
6766 enable the application to run. The system must have PE installed in the
6767 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
6768 must be overridden with the @option{-specs=} option to specify the
6769 appropriate directory location. The Parallel Environment does not
6770 support threads, so the @option{-mpe} option and the @option{-pthread}
6771 option are incompatible.
6775 @opindex msoft-float
6776 @opindex mhard-float
6777 Generate code that does not use (uses) the floating-point register set.
6778 Software floating point emulation is provided if you use the
6779 @option{-msoft-float} option, and pass the option to GCC when linking.
6782 @itemx -mno-multiple
6784 @opindex mno-multiple
6785 Generate code that uses (does not use) the load multiple word
6786 instructions and the store multiple word instructions. These
6787 instructions are generated by default on POWER systems, and not
6788 generated on PowerPC systems. Do not use @option{-mmultiple} on little
6789 endian PowerPC systems, since those instructions do not work when the
6790 processor is in little endian mode. The exceptions are PPC740 and
6791 PPC750 which permit the instructions usage in little endian mode.
6797 Generate code that uses (does not use) the load string instructions
6798 and the store string word instructions to save multiple registers and
6799 do small block moves. These instructions are generated by default on
6800 POWER systems, and not generated on PowerPC systems. Do not use
6801 @option{-mstring} on little endian PowerPC systems, since those
6802 instructions do not work when the processor is in little endian mode.
6803 The exceptions are PPC740 and PPC750 which permit the instructions
6804 usage in little endian mode.
6810 Generate code that uses (does not use) the load or store instructions
6811 that update the base register to the address of the calculated memory
6812 location. These instructions are generated by default. If you use
6813 @option{-mno-update}, there is a small window between the time that the
6814 stack pointer is updated and the address of the previous frame is
6815 stored, which means code that walks the stack frame across interrupts or
6816 signals may get corrupted data.
6819 @itemx -mno-fused-madd
6820 @opindex mfused-madd
6821 @opindex mno-fused-madd
6822 Generate code that uses (does not use) the floating point multiply and
6823 accumulate instructions. These instructions are generated by default if
6824 hardware floating is used.
6826 @item -mno-bit-align
6828 @opindex mno-bit-align
6830 On System V.4 and embedded PowerPC systems do not (do) force structures
6831 and unions that contain bit-fields to be aligned to the base type of the
6834 For example, by default a structure containing nothing but 8
6835 @code{unsigned} bit-fields of length 1 would be aligned to a 4 byte
6836 boundary and have a size of 4 bytes. By using @option{-mno-bit-align},
6837 the structure would be aligned to a 1 byte boundary and be one byte in
6840 @item -mno-strict-align
6841 @itemx -mstrict-align
6842 @opindex mno-strict-align
6843 @opindex mstrict-align
6844 On System V.4 and embedded PowerPC systems do not (do) assume that
6845 unaligned memory references will be handled by the system.
6848 @itemx -mno-relocatable
6849 @opindex mrelocatable
6850 @opindex mno-relocatable
6851 On embedded PowerPC systems generate code that allows (does not allow)
6852 the program to be relocated to a different address at runtime. If you
6853 use @option{-mrelocatable} on any module, all objects linked together must
6854 be compiled with @option{-mrelocatable} or @option{-mrelocatable-lib}.
6856 @item -mrelocatable-lib
6857 @itemx -mno-relocatable-lib
6858 @opindex mrelocatable-lib
6859 @opindex mno-relocatable-lib
6860 On embedded PowerPC systems generate code that allows (does not allow)
6861 the program to be relocated to a different address at runtime. Modules
6862 compiled with @option{-mrelocatable-lib} can be linked with either modules
6863 compiled without @option{-mrelocatable} and @option{-mrelocatable-lib} or
6864 with modules compiled with the @option{-mrelocatable} options.
6870 On System V.4 and embedded PowerPC systems do not (do) assume that
6871 register 2 contains a pointer to a global area pointing to the addresses
6872 used in the program.
6875 @itemx -mlittle-endian
6877 @opindex mlittle-endian
6878 On System V.4 and embedded PowerPC systems compile code for the
6879 processor in little endian mode. The @option{-mlittle-endian} option is
6880 the same as @option{-mlittle}.
6885 @opindex mbig-endian
6886 On System V.4 and embedded PowerPC systems compile code for the
6887 processor in big endian mode. The @option{-mbig-endian} option is
6888 the same as @option{-mbig}.
6892 On System V.4 and embedded PowerPC systems compile code using calling
6893 conventions that adheres to the March 1995 draft of the System V
6894 Application Binary Interface, PowerPC processor supplement. This is the
6895 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
6897 @item -mcall-sysv-eabi
6898 @opindex mcall-sysv-eabi
6899 Specify both @option{-mcall-sysv} and @option{-meabi} options.
6901 @item -mcall-sysv-noeabi
6902 @opindex mcall-sysv-noeabi
6903 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
6907 On System V.4 and embedded PowerPC systems compile code using calling
6908 conventions that are similar to those used on AIX@. This is the
6909 default if you configured GCC using @samp{powerpc-*-eabiaix}.
6911 @item -mcall-solaris
6912 @opindex mcall-solaris
6913 On System V.4 and embedded PowerPC systems compile code for the Solaris
6917 @opindex mcall-linux
6918 On System V.4 and embedded PowerPC systems compile code for the
6919 Linux-based GNU system.
6923 On System V.4 and embedded PowerPC systems compile code for the
6924 Hurd-based GNU system.
6927 @opindex mcall-netbsd
6928 On System V.4 and embedded PowerPC systems compile code for the
6929 NetBSD operating system.
6931 @item -maix-struct-return
6932 @opindex maix-struct-return
6933 Return all structures in memory (as specified by the AIX ABI)@.
6935 @item -msvr4-struct-return
6936 @opindex msvr4-struct-return
6937 Return structures smaller than 8 bytes in registers (as specified by the
6941 @opindex mabi=altivec
6942 Extend the current ABI with AltiVec ABI extensions. This does not
6943 change the default ABI, instead it adds the AltiVec ABI extensions to
6946 @item -mabi=no-altivec
6947 @opindex mabi=no-altivec
6948 Disable AltiVec ABI extensions for the current ABI.
6951 @itemx -mno-prototype
6953 @opindex mno-prototype
6954 On System V.4 and embedded PowerPC systems assume that all calls to
6955 variable argument functions are properly prototyped. Otherwise, the
6956 compiler must insert an instruction before every non prototyped call to
6957 set or clear bit 6 of the condition code register (@var{CR}) to
6958 indicate whether floating point values were passed in the floating point
6959 registers in case the function takes a variable arguments. With
6960 @option{-mprototype}, only calls to prototyped variable argument functions
6961 will set or clear the bit.
6965 On embedded PowerPC systems, assume that the startup module is called
6966 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
6967 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}.
6972 On embedded PowerPC systems, assume that the startup module is called
6973 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
6978 On embedded PowerPC systems, assume that the startup module is called
6979 @file{crt0.o} and the standard C libraries are @file{libads.a} and
6983 @opindex myellowknife
6984 On embedded PowerPC systems, assume that the startup module is called
6985 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
6990 On System V.4 and embedded PowerPC systems, specify that you are
6991 compiling for a VxWorks system.
6995 Specify that you are compiling for the WindISS simulation environment.
6999 On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags
7000 header to indicate that @samp{eabi} extended relocations are used.
7006 On System V.4 and embedded PowerPC systems do (do not) adhere to the
7007 Embedded Applications Binary Interface (eabi) which is a set of
7008 modifications to the System V.4 specifications. Selecting @option{-meabi}
7009 means that the stack is aligned to an 8 byte boundary, a function
7010 @code{__eabi} is called to from @code{main} to set up the eabi
7011 environment, and the @option{-msdata} option can use both @code{r2} and
7012 @code{r13} to point to two separate small data areas. Selecting
7013 @option{-mno-eabi} means that the stack is aligned to a 16 byte boundary,
7014 do not call an initialization function from @code{main}, and the
7015 @option{-msdata} option will only use @code{r13} to point to a single
7016 small data area. The @option{-meabi} option is on by default if you
7017 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
7020 @opindex msdata=eabi
7021 On System V.4 and embedded PowerPC systems, put small initialized
7022 @code{const} global and static data in the @samp{.sdata2} section, which
7023 is pointed to by register @code{r2}. Put small initialized
7024 non-@code{const} global and static data in the @samp{.sdata} section,
7025 which is pointed to by register @code{r13}. Put small uninitialized
7026 global and static data in the @samp{.sbss} section, which is adjacent to
7027 the @samp{.sdata} section. The @option{-msdata=eabi} option is
7028 incompatible with the @option{-mrelocatable} option. The
7029 @option{-msdata=eabi} option also sets the @option{-memb} option.
7032 @opindex msdata=sysv
7033 On System V.4 and embedded PowerPC systems, put small global and static
7034 data in the @samp{.sdata} section, which is pointed to by register
7035 @code{r13}. Put small uninitialized global and static data in the
7036 @samp{.sbss} section, which is adjacent to the @samp{.sdata} section.
7037 The @option{-msdata=sysv} option is incompatible with the
7038 @option{-mrelocatable} option.
7040 @item -msdata=default
7042 @opindex msdata=default
7044 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
7045 compile code the same as @option{-msdata=eabi}, otherwise compile code the
7046 same as @option{-msdata=sysv}.
7049 @opindex msdata-data
7050 On System V.4 and embedded PowerPC systems, put small global and static
7051 data in the @samp{.sdata} section. Put small uninitialized global and
7052 static data in the @samp{.sbss} section. Do not use register @code{r13}
7053 to address small data however. This is the default behavior unless
7054 other @option{-msdata} options are used.
7058 @opindex msdata=none
7060 On embedded PowerPC systems, put all initialized global and static data
7061 in the @samp{.data} section, and all uninitialized data in the
7062 @samp{.bss} section.
7066 @cindex smaller data references (PowerPC)
7067 @cindex .sdata/.sdata2 references (PowerPC)
7068 On embedded PowerPC systems, put global and static items less than or
7069 equal to @var{num} bytes into the small data or bss sections instead of
7070 the normal data or bss section. By default, @var{num} is 8. The
7071 @option{-G @var{num}} switch is also passed to the linker.
7072 All modules should be compiled with the same @option{-G @var{num}} value.
7075 @itemx -mno-regnames
7077 @opindex mno-regnames
7078 On System V.4 and embedded PowerPC systems do (do not) emit register
7079 names in the assembly language output using symbolic forms.
7082 @itemx -mno-longcall
7084 @opindex mno-longcall
7085 Default to making all function calls via pointers, so that functions
7086 which reside further than 64 megabytes (67,108,864 bytes) from the
7087 current location can be called. This setting can be overridden by the
7088 @code{shortcall} function attribute, or by @code{#pragma longcall(0)}.
7090 Some linkers are capable of detecting out-of-range calls and generating
7091 glue code on the fly. On these systems, long calls are unnecessary and
7092 generate slower code. As of this writing, the AIX linker can do this,
7093 as can the GNU linker for PowerPC/64. It is planned to add this feature
7094 to the GNU linker for 32-bit PowerPC systems as well.
7096 In the future, we may cause GCC to ignore all longcall specifications
7097 when the linker is known to generate glue.
7101 Adds support for multithreading with the @dfn{pthreads} library.
7102 This option sets flags for both the preprocessor and linker.
7106 @node Darwin Options
7107 @subsection Darwin Options
7108 @cindex Darwin options
7112 Loads all members of static archive libraries.
7113 See man ld(1) for more information.
7115 @item -arch_errors_fatal
7116 @opindex arch_errors_fatal
7117 Cause the errors having to do with files that have the wrong architecture
7121 @opindex bind_at_load
7122 Causes the output file to be marked such that the dynamic linker will
7123 bind all undefined references when the file is loaded or launched.
7127 Produce a Mach-o bundle format file.
7128 See man ld(1) for more information.
7130 @item -bundle_loader @var{executable}
7131 @opindex bundle_loader
7132 This specifies the @var{executable} that will be loading the build
7133 output file being linked. See man ld(1) for more information.
7135 @item -allowable_client @var{client_name}
7139 @item -compatibility_version
7140 @item -current_version
7141 @item -dependency-file
7143 @item -dylinker_install_name
7146 @item -exported_symbols_list
7148 @item -flat_namespace
7149 @item -force_cpusubtype_ALL
7150 @item -force_flat_namespace
7151 @item -headerpad_max_install_names
7155 @item -keep_private_externs
7157 @item -multiply_defined
7158 @item -multiply_defined_unused
7162 @item -noseglinkedit
7163 @item -pagezero_size
7165 @item -prebind_all_twolevel_modules
7166 @item -private_bundle
7167 @item -read_only_relocs
7169 @item -sectobjectsymbols
7173 @item -sectobjectsymbols
7175 @item -seg_addr_table
7176 @item -seg_addr_table_filename
7179 @item -segs_read_only_addr
7180 @item -segs_read_write_addr
7181 @item -single_module
7185 @item -twolevel_namespace
7188 @item -unexported_symbols_list
7189 @item -weak_reference_mismatches
7192 @opindex allowable_client
7194 @opindex client_name
7195 @opindex compatibility_version
7196 @opindex current_version
7197 @opindex dependency-file
7199 @opindex dylinker_install_name
7202 @opindex exported_symbols_list
7204 @opindex flat_namespace
7205 @opindex force_cpusubtype_ALL
7206 @opindex force_flat_namespace
7207 @opindex headerpad_max_install_names
7210 @opindex install_name
7211 @opindex keep_private_externs
7212 @opindex multi_module
7213 @opindex multiply_defined
7214 @opindex multiply_defined_unused
7216 @opindex nomultidefs
7218 @opindex noseglinkedit
7219 @opindex pagezero_size
7221 @opindex prebind_all_twolevel_modules
7222 @opindex private_bundle
7223 @opindex read_only_relocs
7225 @opindex sectobjectsymbols
7229 @opindex sectobjectsymbols
7231 @opindex seg_addr_table
7232 @opindex seg_addr_table_filename
7233 @opindex seglinkedit
7235 @opindex segs_read_only_addr
7236 @opindex segs_read_write_addr
7237 @opindex single_module
7239 @opindex sub_library
7240 @opindex sub_umbrella
7241 @opindex twolevel_namespace
7244 @opindex unexported_symbols_list
7245 @opindex weak_reference_mismatches
7246 @opindex whatsloaded
7248 This options are available for Darwin linker. Darwin linker man page
7249 describes them in detail.
7254 @subsection IBM RT Options
7256 @cindex IBM RT options
7258 These @samp{-m} options are defined for the IBM RT PC:
7262 @opindex min-line-mul
7263 Use an in-line code sequence for integer multiplies. This is the
7266 @item -mcall-lib-mul
7267 @opindex mcall-lib-mul
7268 Call @code{lmul$$} for integer multiples.
7270 @item -mfull-fp-blocks
7271 @opindex mfull-fp-blocks
7272 Generate full-size floating point data blocks, including the minimum
7273 amount of scratch space recommended by IBM@. This is the default.
7275 @item -mminimum-fp-blocks
7276 @opindex mminimum-fp-blocks
7277 Do not include extra scratch space in floating point data blocks. This
7278 results in smaller code, but slower execution, since scratch space must
7279 be allocated dynamically.
7281 @cindex @file{stdarg.h} and RT PC
7282 @item -mfp-arg-in-fpregs
7283 @opindex mfp-arg-in-fpregs
7284 Use a calling sequence incompatible with the IBM calling convention in
7285 which floating point arguments are passed in floating point registers.
7286 Note that @code{stdarg.h} will not work with floating point operands
7287 if this option is specified.
7289 @item -mfp-arg-in-gregs
7290 @opindex mfp-arg-in-gregs
7291 Use the normal calling convention for floating point arguments. This is
7294 @item -mhc-struct-return
7295 @opindex mhc-struct-return
7296 Return structures of more than one word in memory, rather than in a
7297 register. This provides compatibility with the MetaWare HighC (hc)
7298 compiler. Use the option @option{-fpcc-struct-return} for compatibility
7299 with the Portable C Compiler (pcc).
7301 @item -mnohc-struct-return
7302 @opindex mnohc-struct-return
7303 Return some structures of more than one word in registers, when
7304 convenient. This is the default. For compatibility with the
7305 IBM-supplied compilers, use the option @option{-fpcc-struct-return} or the
7306 option @option{-mhc-struct-return}.
7310 @subsection MIPS Options
7311 @cindex MIPS options
7313 These @samp{-m} options are defined for the MIPS family of computers:
7317 @item -march=@var{arch}
7319 Generate code that will run on @var{arch}, which can be the name of a
7320 generic MIPS ISA, or the name of a particular processor.
7322 @samp{mips1}, @samp{mips2}, @samp{mips3}, @samp{mips4},
7323 @samp{mips32}, @samp{mips32r2}, and @samp{mips64}.
7324 The processor names are:
7325 @samp{4kc}, @samp{4kp}, @samp{5kc}, @samp{20kc},
7327 @samp{r2000}, @samp{r3000}, @samp{r3900}, @samp{r4000}, @samp{r4400},
7328 @samp{r4600}, @samp{r4650}, @samp{r6000}, @samp{r8000},
7331 @samp{vr4100}, @samp{vr4300}, and @samp{vr5000}.
7332 The special value @samp{from-abi} selects the
7333 most compatible architecture for the selected ABI (that is,
7334 @samp{mips1} for 32-bit ABIs and @samp{mips3} for 64-bit ABIs)@.
7336 In processor names, a final @samp{000} can be abbreviated as @samp{k}
7337 (for example, @samp{-march=r2k}). Prefixes are optional, and
7338 @samp{vr} may be written @samp{r}.
7340 GCC defines two macros based on the value of this option. The first
7341 is @samp{_MIPS_ARCH}, which gives the name of target architecture, as
7342 a string. The second has the form @samp{_MIPS_ARCH_@var{foo}},
7343 where @var{foo} is the capitalized value of @samp{_MIPS_ARCH}@.
7344 For example, @samp{-march=r2000} will set @samp{_MIPS_ARCH}
7345 to @samp{"r2000"} and define the macro @samp{_MIPS_ARCH_R2000}.
7347 Note that the @samp{_MIPS_ARCH} macro uses the processor names given
7348 above. In other words, it will have the full prefix and will not
7349 abbreviate @samp{000} as @samp{k}. In the case of @samp{from-abi},
7350 the macro names the resolved architecture (either @samp{"mips1"} or
7351 @samp{"mips3"}). It names the default architecture when no
7352 @option{-march} option is given.
7354 @item -mtune=@var{arch}
7356 Optimize for @var{arch}. Among other things, this option controls
7357 the way instructions are scheduled, and the perceived cost of arithmetic
7358 operations. The list of @var{arch} values is the same as for
7361 When this option is not used, GCC will optimize for the processor
7362 specified by @option{-march}. By using @option{-march} and
7363 @option{-mtune} together, it is possible to generate code that will
7364 run on a family of processors, but optimize the code for one
7365 particular member of that family.
7367 @samp{-mtune} defines the macros @samp{_MIPS_TUNE} and
7368 @samp{_MIPS_TUNE_@var{foo}}, which work in the same way as the
7369 @samp{-march} ones described above.
7373 Equivalent to @samp{-march=mips1}.
7377 Equivalent to @samp{-march=mips2}.
7381 Equivalent to @samp{-march=mips3}.
7385 Equivalent to @samp{-march=mips4}.
7389 Equivalent to @samp{-march=mips32}.
7393 Equivalent to @samp{-march=mips32r2}.
7397 Equivalent to @samp{-march=mips64}.
7400 @itemx -mno-fused-madd
7401 @opindex mfused-madd
7402 @opindex mno-fused-madd
7403 Generate code that uses (does not use) the floating point multiply and
7404 accumulate instructions, when they are available. These instructions
7405 are generated by default if they are available, but this may be
7406 undesirable if the extra precision causes problems or on certain chips
7407 in the mode where denormals are rounded to zero where denormals
7408 generated by multiply and accumulate instructions cause exceptions
7413 Assume that floating point registers are 32 bits wide.
7417 Assume that floating point registers are 64 bits wide.
7421 Assume that general purpose registers are 32 bits wide.
7425 Assume that general purpose registers are 64 bits wide.
7429 Force int and long types to be 64 bits wide. See @option{-mlong32} for an
7430 explanation of the default, and the width of pointers.
7434 Force long types to be 64 bits wide. See @option{-mlong32} for an
7435 explanation of the default, and the width of pointers.
7439 Force long, int, and pointer types to be 32 bits wide.
7441 The default size of ints, longs and pointers depends on the ABI@. All
7442 the supported ABIs use 32-bit ints. The n64 ABI uses 64-bit longs, as
7443 does the 64-bit Cygnus EABI; the others use 32-bit longs. Pointers
7444 are the same size as longs, or the same size as integer registers,
7445 whichever is smaller.
7459 Generate code for the given ABI@.
7461 Note that there are two embedded ABIs: @option{-mabi=eabi}
7462 selects the one defined by Cygnus while @option{-meabi=meabi}
7463 selects the one defined by MIPS@. Both these ABIs have
7464 32-bit and 64-bit variants. Normally, GCC will generate
7465 64-bit code when you select a 64-bit architecture, but you
7466 can use @option{-mgp32} to get 32-bit code instead.
7470 Generate code for the MIPS assembler, and invoke @file{mips-tfile} to
7471 add normal debug information. This is the default for all
7472 platforms except for the OSF/1 reference platform, using the OSF/rose
7473 object format. If the either of the @option{-gstabs} or @option{-gstabs+}
7474 switches are used, the @file{mips-tfile} program will encapsulate the
7475 stabs within MIPS ECOFF@.
7479 Generate code for the GNU assembler. This is the default on the OSF/1
7480 reference platform, using the OSF/rose object format. Also, this is
7481 the default if the configure option @option{--with-gnu-as} is used.
7483 @item -msplit-addresses
7484 @itemx -mno-split-addresses
7485 @opindex msplit-addresses
7486 @opindex mno-split-addresses
7487 Generate code to load the high and low parts of address constants separately.
7488 This allows GCC to optimize away redundant loads of the high order
7489 bits of addresses. This optimization requires GNU as and GNU ld.
7490 This optimization is enabled by default for some embedded targets where
7491 GNU as and GNU ld are standard.
7497 The @option{-mrnames} switch says to output code using the MIPS software
7498 names for the registers, instead of the hardware names (ie, @var{a0}
7499 instead of @var{$4}). The only known assembler that supports this option
7500 is the Algorithmics assembler.
7506 The @option{-mgpopt} switch says to write all of the data declarations
7507 before the instructions in the text section, this allows the MIPS
7508 assembler to generate one word memory references instead of using two
7509 words for short global or static data items. This is on by default if
7510 optimization is selected.
7516 For each non-inline function processed, the @option{-mstats} switch
7517 causes the compiler to emit one line to the standard error file to
7518 print statistics about the program (number of registers saved, stack
7525 The @option{-mmemcpy} switch makes all block moves call the appropriate
7526 string function (@samp{memcpy} or @samp{bcopy}) instead of possibly
7527 generating inline code.
7530 @itemx -mno-mips-tfile
7531 @opindex mmips-tfile
7532 @opindex mno-mips-tfile
7533 The @option{-mno-mips-tfile} switch causes the compiler not
7534 postprocess the object file with the @file{mips-tfile} program,
7535 after the MIPS assembler has generated it to add debug support. If
7536 @file{mips-tfile} is not run, then no local variables will be
7537 available to the debugger. In addition, @file{stage2} and
7538 @file{stage3} objects will have the temporary file names passed to the
7539 assembler embedded in the object file, which means the objects will
7540 not compare the same. The @option{-mno-mips-tfile} switch should only
7541 be used when there are bugs in the @file{mips-tfile} program that
7542 prevents compilation.
7545 @opindex msoft-float
7546 Generate output containing library calls for floating point.
7547 @strong{Warning:} the requisite libraries are not part of GCC@.
7548 Normally the facilities of the machine's usual C compiler are used, but
7549 this can't be done directly in cross-compilation. You must make your
7550 own arrangements to provide suitable library functions for
7554 @opindex mhard-float
7555 Generate output containing floating point instructions. This is the
7556 default if you use the unmodified sources.
7559 @itemx -mno-abicalls
7561 @opindex mno-abicalls
7562 Emit (or do not emit) the pseudo operations @samp{.abicalls},
7563 @samp{.cpload}, and @samp{.cprestore} that some System V.4 ports use for
7564 position independent code.
7567 @itemx -mno-long-calls
7568 @opindex mlong-calls
7569 @opindex mno-long-calls
7570 Do all calls with the @samp{JALR} instruction, which requires
7571 loading up a function's address into a register before the call.
7572 You need to use this switch, if you call outside of the current
7573 512 megabyte segment to functions that are not through pointers.
7576 @itemx -mno-half-pic
7578 @opindex mno-half-pic
7579 Put pointers to extern references into the data section and load them
7580 up, rather than put the references in the text section.
7582 @item -membedded-pic
7583 @itemx -mno-embedded-pic
7584 @opindex membedded-pic
7585 @opindex mno-embedded-pic
7586 Generate PIC code suitable for some embedded systems. All calls are
7587 made using PC relative address, and all data is addressed using the $gp
7588 register. No more than 65536 bytes of global data may be used. This
7589 requires GNU as and GNU ld which do most of the work. This currently
7590 only works on targets which use ECOFF; it does not work with ELF@.
7592 @item -membedded-data
7593 @itemx -mno-embedded-data
7594 @opindex membedded-data
7595 @opindex mno-embedded-data
7596 Allocate variables to the read-only data section first if possible, then
7597 next in the small data section if possible, otherwise in data. This gives
7598 slightly slower code than the default, but reduces the amount of RAM required
7599 when executing, and thus may be preferred for some embedded systems.
7601 @item -muninit-const-in-rodata
7602 @itemx -mno-uninit-const-in-rodata
7603 @opindex muninit-const-in-rodata
7604 @opindex mno-uninit-const-in-rodata
7605 When used together with @option{-membedded-data}, it will always store uninitialized
7606 const variables in the read-only data section.
7608 @item -msingle-float
7609 @itemx -mdouble-float
7610 @opindex msingle-float
7611 @opindex mdouble-float
7612 The @option{-msingle-float} switch tells gcc to assume that the floating
7613 point coprocessor only supports single precision operations, as on the
7614 @samp{r4650} chip. The @option{-mdouble-float} switch permits gcc to use
7615 double precision operations. This is the default.
7621 Permit use of the @samp{mad}, @samp{madu} and @samp{mul} instructions,
7622 as on the @samp{r4650} chip.
7626 Turns on @option{-msingle-float}, @option{-mmad}, and, at least for now,
7627 @option{-mcpu=r4650}.
7633 Enable 16-bit instructions.
7637 Use the entry and exit pseudo ops. This option can only be used with
7642 Compile code for the processor in little endian mode.
7643 The requisite libraries are assumed to exist.
7647 Compile code for the processor in big endian mode.
7648 The requisite libraries are assumed to exist.
7652 @cindex smaller data references (MIPS)
7653 @cindex gp-relative references (MIPS)
7654 Put global and static items less than or equal to @var{num} bytes into
7655 the small data or bss sections instead of the normal data or bss
7656 section. This allows the assembler to emit one word memory reference
7657 instructions based on the global pointer (@var{gp} or @var{$28}),
7658 instead of the normal two words used. By default, @var{num} is 8 when
7659 the MIPS assembler is used, and 0 when the GNU assembler is used. The
7660 @option{-G @var{num}} switch is also passed to the assembler and linker.
7661 All modules should be compiled with the same @option{-G @var{num}}
7666 Tell the MIPS assembler to not run its preprocessor over user
7667 assembler files (with a @samp{.s} suffix) when assembling them.
7671 Pass an option to gas which will cause nops to be inserted if
7672 the read of the destination register of an mfhi or mflo instruction
7673 occurs in the following two instructions.
7677 Do not include the default crt0.
7679 @item -mflush-func=@var{func}
7680 @itemx -mno-flush-func
7681 @opindex mflush-func
7682 Specifies the function to call to flush the I and D caches, or to not
7683 call any such function. If called, the function must take the same
7684 arguments as the common @code{_flush_func()}, that is, the address of the
7685 memory range for which the cache is being flushed, the size of the
7686 memory range, and the number 3 (to flush both caches). The default
7687 depends on the target gcc was configured for, but commonly is either
7688 @samp{_flush_func} or @samp{__cpu_flush}.
7690 @item -mbranch-likely
7691 @itemx -mno-branch-likely
7692 @opindex mbranch-likely
7693 @opindex mno-branch-likely
7694 Enable or disable use of Branch Likely instructions, regardless of the
7695 default for the selected architecture. By default, Branch Likely
7696 instructions may be generated if they are supported by the selected
7697 architecture. An exception is for the MIPS32 and MIPS64 architectures
7698 and processors which implement those architectures; for those, Branch
7699 Likely instructions will not be generated by default because the MIPS32
7700 and MIPS64 architectures specifically deprecate their use.
7703 @node i386 and x86-64 Options
7704 @subsection Intel 386 and AMD x86-64 Options
7705 @cindex i386 Options
7706 @cindex x86-64 Options
7707 @cindex Intel 386 Options
7708 @cindex AMD x86-64 Options
7710 These @samp{-m} options are defined for the i386 and x86-64 family of
7714 @item -mcpu=@var{cpu-type}
7716 Tune to @var{cpu-type} everything applicable about the generated code, except
7717 for the ABI and the set of available instructions. The choices for
7718 @var{cpu-type} are @samp{i386}, @samp{i486}, @samp{i586}, @samp{i686},
7719 @samp{pentium}, @samp{pentium-mmx}, @samp{pentiumpro}, @samp{pentium2},
7720 @samp{pentium3}, @samp{pentium4}, @samp{k6}, @samp{k6-2}, @samp{k6-3},
7721 @samp{athlon}, @samp{athlon-tbird}, @samp{athlon-4}, @samp{athlon-xp},
7722 @samp{athlon-mp}, @samp{winchip-c6}, @samp{winchip2}, @samp{k8} and @samp{c3}.
7724 While picking a specific @var{cpu-type} will schedule things appropriately
7725 for that particular chip, the compiler will not generate any code that
7726 does not run on the i386 without the @option{-march=@var{cpu-type}} option
7727 being used. @samp{i586} is equivalent to @samp{pentium} and @samp{i686}
7728 is equivalent to @samp{pentiumpro}. @samp{k6} and @samp{athlon} are the
7729 AMD chips as opposed to the Intel ones.
7731 @item -march=@var{cpu-type}
7733 Generate instructions for the machine type @var{cpu-type}. The choices
7734 for @var{cpu-type} are the same as for @option{-mcpu}. Moreover,
7735 specifying @option{-march=@var{cpu-type}} implies @option{-mcpu=@var{cpu-type}}.
7744 @opindex mpentiumpro
7745 These options are synonyms for @option{-mcpu=i386}, @option{-mcpu=i486},
7746 @option{-mcpu=pentium}, and @option{-mcpu=pentiumpro} respectively.
7747 These synonyms are deprecated.
7749 @item -mfpmath=@var{unit}
7751 generate floating point arithmetics for selected unit @var{unit}. the choices
7756 Use the standard 387 floating point coprocessor present majority of chips and
7757 emulated otherwise. Code compiled with this option will run almost everywhere.
7758 The temporary results are computed in 80bit precision instead of precision
7759 specified by the type resulting in slightly different results compared to most
7760 of other chips. See @option{-ffloat-store} for more detailed description.
7762 This is the default choice for i386 compiler.
7765 Use scalar floating point instructions present in the SSE instruction set.
7766 This instruction set is supported by Pentium3 and newer chips, in the AMD line
7767 by Athlon-4, Athlon-xp and Athlon-mp chips. The earlier version of SSE
7768 instruction set supports only single precision arithmetics, thus the double and
7769 extended precision arithmetics is still done using 387. Later version, present
7770 only in Pentium4 and the future AMD x86-64 chips supports double precision
7773 For i387 you need to use @option{-march=@var{cpu-type}}, @option{-msse} or
7774 @option{-msse2} switches to enable SSE extensions and make this option
7775 effective. For x86-64 compiler, these extensions are enabled by default.
7777 The resulting code should be considerably faster in majority of cases and avoid
7778 the numerical instability problems of 387 code, but may break some existing
7779 code that expects temporaries to be 80bit.
7781 This is the default choice for x86-64 compiler.
7784 Attempt to utilize both instruction sets at once. This effectively double the
7785 amount of available registers and on chips with separate execution units for
7786 387 and SSE the execution resources too. Use this option with care, as it is
7787 still experimental, because gcc register allocator does not model separate
7788 functional units well resulting in instable performance.
7791 @item -masm=@var{dialect}
7792 @opindex masm=@var{dialect}
7793 Output asm instructions using selected @var{dialect}. Supported choices are
7794 @samp{intel} or @samp{att} (the default one).
7799 @opindex mno-ieee-fp
7800 Control whether or not the compiler uses IEEE floating point
7801 comparisons. These handle correctly the case where the result of a
7802 comparison is unordered.
7805 @opindex msoft-float
7806 Generate output containing library calls for floating point.
7807 @strong{Warning:} the requisite libraries are not part of GCC@.
7808 Normally the facilities of the machine's usual C compiler are used, but
7809 this can't be done directly in cross-compilation. You must make your
7810 own arrangements to provide suitable library functions for
7813 On machines where a function returns floating point results in the 80387
7814 register stack, some floating point opcodes may be emitted even if
7815 @option{-msoft-float} is used.
7817 @item -mno-fp-ret-in-387
7818 @opindex mno-fp-ret-in-387
7819 Do not use the FPU registers for return values of functions.
7821 The usual calling convention has functions return values of types
7822 @code{float} and @code{double} in an FPU register, even if there
7823 is no FPU@. The idea is that the operating system should emulate
7826 The option @option{-mno-fp-ret-in-387} causes such values to be returned
7827 in ordinary CPU registers instead.
7829 @item -mno-fancy-math-387
7830 @opindex mno-fancy-math-387
7831 Some 387 emulators do not support the @code{sin}, @code{cos} and
7832 @code{sqrt} instructions for the 387. Specify this option to avoid
7833 generating those instructions. This option is the default on FreeBSD,
7834 OpenBSD and NetBSD@. This option is overridden when @option{-march}
7835 indicates that the target cpu will always have an FPU and so the
7836 instruction will not need emulation. As of revision 2.6.1, these
7837 instructions are not generated unless you also use the
7838 @option{-funsafe-math-optimizations} switch.
7840 @item -malign-double
7841 @itemx -mno-align-double
7842 @opindex malign-double
7843 @opindex mno-align-double
7844 Control whether GCC aligns @code{double}, @code{long double}, and
7845 @code{long long} variables on a two word boundary or a one word
7846 boundary. Aligning @code{double} variables on a two word boundary will
7847 produce code that runs somewhat faster on a @samp{Pentium} at the
7848 expense of more memory.
7850 @strong{Warning:} if you use the @option{-malign-double} switch,
7851 structures containing the above types will be aligned differently than
7852 the published application binary interface specifications for the 386
7853 and will not be binary compatible with structures in code compiled
7854 without that switch.
7856 @item -m128bit-long-double
7857 @opindex m128bit-long-double
7858 Control the size of @code{long double} type. i386 application binary interface
7859 specify the size to be 12 bytes, while modern architectures (Pentium and newer)
7860 prefer @code{long double} aligned to 8 or 16 byte boundary. This is
7861 impossible to reach with 12 byte long doubles in the array accesses.
7863 @strong{Warning:} if you use the @option{-m128bit-long-double} switch, the
7864 structures and arrays containing @code{long double} will change their size as
7865 well as function calling convention for function taking @code{long double}
7868 @item -m96bit-long-double
7869 @opindex m96bit-long-double
7870 Set the size of @code{long double} to 96 bits as required by the i386
7871 application binary interface. This is the default.
7874 @itemx -mno-svr3-shlib
7875 @opindex msvr3-shlib
7876 @opindex mno-svr3-shlib
7877 Control whether GCC places uninitialized local variables into the
7878 @code{bss} or @code{data} segments. @option{-msvr3-shlib} places them
7879 into @code{bss}. These options are meaningful only on System V Release 3.
7883 Use a different function-calling convention, in which functions that
7884 take a fixed number of arguments return with the @code{ret} @var{num}
7885 instruction, which pops their arguments while returning. This saves one
7886 instruction in the caller since there is no need to pop the arguments
7889 You can specify that an individual function is called with this calling
7890 sequence with the function attribute @samp{stdcall}. You can also
7891 override the @option{-mrtd} option by using the function attribute
7892 @samp{cdecl}. @xref{Function Attributes}.
7894 @strong{Warning:} this calling convention is incompatible with the one
7895 normally used on Unix, so you cannot use it if you need to call
7896 libraries compiled with the Unix compiler.
7898 Also, you must provide function prototypes for all functions that
7899 take variable numbers of arguments (including @code{printf});
7900 otherwise incorrect code will be generated for calls to those
7903 In addition, seriously incorrect code will result if you call a
7904 function with too many arguments. (Normally, extra arguments are
7905 harmlessly ignored.)
7907 @item -mregparm=@var{num}
7909 Control how many registers are used to pass integer arguments. By
7910 default, no registers are used to pass arguments, and at most 3
7911 registers can be used. You can control this behavior for a specific
7912 function by using the function attribute @samp{regparm}.
7913 @xref{Function Attributes}.
7915 @strong{Warning:} if you use this switch, and
7916 @var{num} is nonzero, then you must build all modules with the same
7917 value, including any libraries. This includes the system libraries and
7920 @item -mpreferred-stack-boundary=@var{num}
7921 @opindex mpreferred-stack-boundary
7922 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
7923 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
7924 the default is 4 (16 bytes or 128 bits), except when optimizing for code
7925 size (@option{-Os}), in which case the default is the minimum correct
7926 alignment (4 bytes for x86, and 8 bytes for x86-64).
7928 On Pentium and PentiumPro, @code{double} and @code{long double} values
7929 should be aligned to an 8 byte boundary (see @option{-malign-double}) or
7930 suffer significant run time performance penalties. On Pentium III, the
7931 Streaming SIMD Extension (SSE) data type @code{__m128} suffers similar
7932 penalties if it is not 16 byte aligned.
7934 To ensure proper alignment of this values on the stack, the stack boundary
7935 must be as aligned as that required by any value stored on the stack.
7936 Further, every function must be generated such that it keeps the stack
7937 aligned. Thus calling a function compiled with a higher preferred
7938 stack boundary from a function compiled with a lower preferred stack
7939 boundary will most likely misalign the stack. It is recommended that
7940 libraries that use callbacks always use the default setting.
7942 This extra alignment does consume extra stack space, and generally
7943 increases code size. Code that is sensitive to stack space usage, such
7944 as embedded systems and operating system kernels, may want to reduce the
7945 preferred alignment to @option{-mpreferred-stack-boundary=2}.
7961 These switches enable or disable the use of built-in functions that allow
7962 direct access to the MMX, SSE and 3Dnow extensions of the instruction set.
7964 @xref{X86 Built-in Functions}, for details of the functions enabled
7965 and disabled by these switches.
7967 To have SSE/SSE2 instructions generated automatically from floating-point
7968 code, see @option{-mfpmath=sse}.
7971 @itemx -mno-push-args
7973 @opindex mno-push-args
7974 Use PUSH operations to store outgoing parameters. This method is shorter
7975 and usually equally fast as method using SUB/MOV operations and is enabled
7976 by default. In some cases disabling it may improve performance because of
7977 improved scheduling and reduced dependencies.
7979 @item -maccumulate-outgoing-args
7980 @opindex maccumulate-outgoing-args
7981 If enabled, the maximum amount of space required for outgoing arguments will be
7982 computed in the function prologue. This is faster on most modern CPUs
7983 because of reduced dependencies, improved scheduling and reduced stack usage
7984 when preferred stack boundary is not equal to 2. The drawback is a notable
7985 increase in code size. This switch implies @option{-mno-push-args}.
7989 Support thread-safe exception handling on @samp{Mingw32}. Code that relies
7990 on thread-safe exception handling must compile and link all code with the
7991 @option{-mthreads} option. When compiling, @option{-mthreads} defines
7992 @option{-D_MT}; when linking, it links in a special thread helper library
7993 @option{-lmingwthrd} which cleans up per thread exception handling data.
7995 @item -mno-align-stringops
7996 @opindex mno-align-stringops
7997 Do not align destination of inlined string operations. This switch reduces
7998 code size and improves performance in case the destination is already aligned,
7999 but gcc don't know about it.
8001 @item -minline-all-stringops
8002 @opindex minline-all-stringops
8003 By default GCC inlines string operations only when destination is known to be
8004 aligned at least to 4 byte boundary. This enables more inlining, increase code
8005 size, but may improve performance of code that depends on fast memcpy, strlen
8006 and memset for short lengths.
8008 @item -momit-leaf-frame-pointer
8009 @opindex momit-leaf-frame-pointer
8010 Don't keep the frame pointer in a register for leaf functions. This
8011 avoids the instructions to save, set up and restore frame pointers and
8012 makes an extra register available in leaf functions. The option
8013 @option{-fomit-frame-pointer} removes the frame pointer for all functions
8014 which might make debugging harder.
8017 These @samp{-m} switches are supported in addition to the above
8018 on AMD x86-64 processors in 64-bit environments.
8025 Generate code for a 32-bit or 64-bit environment.
8026 The 32-bit environment sets int, long and pointer to 32 bits and
8027 generates code that runs on any i386 system.
8028 The 64-bit environment sets int to 32 bits and long and pointer
8029 to 64 bits and generates code for AMD's x86-64 architecture.
8032 @opindex no-red-zone
8033 Do not use a so called red zone for x86-64 code. The red zone is mandated
8034 by the x86-64 ABI, it is a 128-byte area beyond the location of the
8035 stack pointer that will not be modified by signal or interrupt handlers
8036 and therefore can be used for temporary data without adjusting the stack
8037 pointer. The flag @option{-mno-red-zone} disables this red zone.
8039 @item -mcmodel=small
8040 @opindex mcmodel=small
8041 Generate code for the small code model: the program and its symbols must
8042 be linked in the lower 2 GB of the address space. Pointers are 64 bits.
8043 Programs can be statically or dynamically linked. This is the default
8046 @item -mcmodel=kernel
8047 @opindex mcmodel=kernel
8048 Generate code for the kernel code model. The kernel runs in the
8049 negative 2 GB of the address space.
8050 This model has to be used for Linux kernel code.
8052 @item -mcmodel=medium
8053 @opindex mcmodel=medium
8054 Generate code for the medium model: The program is linked in the lower 2
8055 GB of the address space but symbols can be located anywhere in the
8056 address space. Programs can be statically or dynamically linked, but
8057 building of shared libraries are not supported with the medium model.
8059 @item -mcmodel=large
8060 @opindex mcmodel=large
8061 Generate code for the large model: This model makes no assumptions
8062 about addresses and sizes of sections. Currently GCC does not implement
8067 @subsection HPPA Options
8068 @cindex HPPA Options
8070 These @samp{-m} options are defined for the HPPA family of computers:
8073 @item -march=@var{architecture-type}
8075 Generate code for the specified architecture. The choices for
8076 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
8077 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
8078 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
8079 architecture option for your machine. Code compiled for lower numbered
8080 architectures will run on higher numbered architectures, but not the
8083 PA 2.0 support currently requires gas snapshot 19990413 or later. The
8084 next release of binutils (current is 2.9.1) will probably contain PA 2.0
8088 @itemx -mpa-risc-1-1
8089 @itemx -mpa-risc-2-0
8090 @opindex mpa-risc-1-0
8091 @opindex mpa-risc-1-1
8092 @opindex mpa-risc-2-0
8093 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
8096 @opindex mbig-switch
8097 Generate code suitable for big switch tables. Use this option only if
8098 the assembler/linker complain about out of range branches within a switch
8101 @item -mjump-in-delay
8102 @opindex mjump-in-delay
8103 Fill delay slots of function calls with unconditional jump instructions
8104 by modifying the return pointer for the function call to be the target
8105 of the conditional jump.
8107 @item -mdisable-fpregs
8108 @opindex mdisable-fpregs
8109 Prevent floating point registers from being used in any manner. This is
8110 necessary for compiling kernels which perform lazy context switching of
8111 floating point registers. If you use this option and attempt to perform
8112 floating point operations, the compiler will abort.
8114 @item -mdisable-indexing
8115 @opindex mdisable-indexing
8116 Prevent the compiler from using indexing address modes. This avoids some
8117 rather obscure problems when compiling MIG generated code under MACH@.
8119 @item -mno-space-regs
8120 @opindex mno-space-regs
8121 Generate code that assumes the target has no space registers. This allows
8122 GCC to generate faster indirect calls and use unscaled index address modes.
8124 Such code is suitable for level 0 PA systems and kernels.
8126 @item -mfast-indirect-calls
8127 @opindex mfast-indirect-calls
8128 Generate code that assumes calls never cross space boundaries. This
8129 allows GCC to emit code which performs faster indirect calls.
8131 This option will not work in the presence of shared libraries or nested
8134 @item -mlong-load-store
8135 @opindex mlong-load-store
8136 Generate 3-instruction load and store sequences as sometimes required by
8137 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
8140 @item -mportable-runtime
8141 @opindex mportable-runtime
8142 Use the portable calling conventions proposed by HP for ELF systems.
8146 Enable the use of assembler directives only GAS understands.
8148 @item -mschedule=@var{cpu-type}
8150 Schedule code according to the constraints for the machine type
8151 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
8152 @samp{7100}, @samp{7100LC}, @samp{7200}, @samp{7300} and @samp{8000}. Refer
8153 to @file{/usr/lib/sched.models} on an HP-UX system to determine the
8154 proper scheduling option for your machine. The default scheduling is
8158 @opindex mlinker-opt
8159 Enable the optimization pass in the HP-UX linker. Note this makes symbolic
8160 debugging impossible. It also triggers a bug in the HP-UX 8 and HP-UX 9
8161 linkers in which they give bogus error messages when linking some programs.
8164 @opindex msoft-float
8165 Generate output containing library calls for floating point.
8166 @strong{Warning:} the requisite libraries are not available for all HPPA
8167 targets. Normally the facilities of the machine's usual C compiler are
8168 used, but this cannot be done directly in cross-compilation. You must make
8169 your own arrangements to provide suitable library functions for
8170 cross-compilation. The embedded target @samp{hppa1.1-*-pro}
8171 does provide software floating point support.
8173 @option{-msoft-float} changes the calling convention in the output file;
8174 therefore, it is only useful if you compile @emph{all} of a program with
8175 this option. In particular, you need to compile @file{libgcc.a}, the
8176 library that comes with GCC, with @option{-msoft-float} in order for
8181 Generate the predefine, @code{_SIO}, for server IO. The default is
8182 @option{-mwsio}. This generates the predefines, @code{__hp9000s700},
8183 @code{__hp9000s700__} and @code{_WSIO}, for workstation IO. These
8184 options are available under HP-UX and HI-UX.
8188 Use GNU ld specific options. This passes @option{-shared} to ld when
8189 building a shared library. It is the default when GCC is configured,
8190 explicitly or implicitly, with the GNU linker. This option does not
8191 have any affect on which ld is called, it only changes what parameters
8192 are passed to that ld. The ld that is called is determined by the
8193 @option{--with-ld} configure option, gcc's program search path, and
8194 finally by the user's @env{PATH}. The linker used by GCC can be printed
8195 using @samp{which `gcc -print-prog-name=ld`}.
8199 Use HP ld specific options. This passes @option{-b} to ld when building
8200 a shared library and passes @option{+Accept TypeMismatch} to ld on all
8201 links. It is the default when GCC is configured, explicitly or
8202 implicitly, with the HP linker. This option does not have any affect on
8203 which ld is called, it only changes what parameters are passed to that
8204 ld. The ld that is called is determined by the @option{--with-ld}
8205 configure option, gcc's program search path, and finally by the user's
8206 @env{PATH}. The linker used by GCC can be printed using @samp{which
8207 `gcc -print-prog-name=ld`}.
8210 @opindex mno-long-calls
8211 Generate code that uses long call sequences. This ensures that a call
8212 is always able to reach linker generated stubs. The default is to generate
8213 long calls only when the distance from the call site to the beginning
8214 of the function or translation unit, as the case may be, exceeds a
8215 predefined limit set by the branch type being used. The limits for
8216 normal calls are 7,600,000 and 240,000 bytes, respectively for the
8217 PA 2.0 and PA 1.X architectures. Sibcalls are always limited at
8220 Distances are measured from the beginning of functions when using the
8221 @option{-ffunction-sections} option, or when using the @option{-mgas}
8222 and @option{-mno-portable-runtime} options together under HP-UX with
8225 It is normally not desirable to use this option as it will degrade
8226 performance. However, it may be useful in large applications,
8227 particularly when partial linking is used to build the application.
8229 The types of long calls used depends on the capabilities of the
8230 assembler and linker, and the type of code being generated. The
8231 impact on systems that support long absolute calls, and long pic
8232 symbol-difference or pc-relative calls should be relatively small.
8233 However, an indirect call is used on 32-bit ELF systems in pic code
8234 and it is quite long.
8238 @node Intel 960 Options
8239 @subsection Intel 960 Options
8241 These @samp{-m} options are defined for the Intel 960 implementations:
8244 @item -m@var{cpu-type}
8252 Assume the defaults for the machine type @var{cpu-type} for some of
8253 the other options, including instruction scheduling, floating point
8254 support, and addressing modes. The choices for @var{cpu-type} are
8255 @samp{ka}, @samp{kb}, @samp{mc}, @samp{ca}, @samp{cf},
8256 @samp{sa}, and @samp{sb}.
8263 @opindex msoft-float
8264 The @option{-mnumerics} option indicates that the processor does support
8265 floating-point instructions. The @option{-msoft-float} option indicates
8266 that floating-point support should not be assumed.
8268 @item -mleaf-procedures
8269 @itemx -mno-leaf-procedures
8270 @opindex mleaf-procedures
8271 @opindex mno-leaf-procedures
8272 Do (or do not) attempt to alter leaf procedures to be callable with the
8273 @code{bal} instruction as well as @code{call}. This will result in more
8274 efficient code for explicit calls when the @code{bal} instruction can be
8275 substituted by the assembler or linker, but less efficient code in other
8276 cases, such as calls via function pointers, or using a linker that doesn't
8277 support this optimization.
8280 @itemx -mno-tail-call
8282 @opindex mno-tail-call
8283 Do (or do not) make additional attempts (beyond those of the
8284 machine-independent portions of the compiler) to optimize tail-recursive
8285 calls into branches. You may not want to do this because the detection of
8286 cases where this is not valid is not totally complete. The default is
8287 @option{-mno-tail-call}.
8289 @item -mcomplex-addr
8290 @itemx -mno-complex-addr
8291 @opindex mcomplex-addr
8292 @opindex mno-complex-addr
8293 Assume (or do not assume) that the use of a complex addressing mode is a
8294 win on this implementation of the i960. Complex addressing modes may not
8295 be worthwhile on the K-series, but they definitely are on the C-series.
8296 The default is currently @option{-mcomplex-addr} for all processors except
8300 @itemx -mno-code-align
8301 @opindex mcode-align
8302 @opindex mno-code-align
8303 Align code to 8-byte boundaries for faster fetching (or don't bother).
8304 Currently turned on by default for C-series implementations only.
8307 @item -mclean-linkage
8308 @itemx -mno-clean-linkage
8309 @opindex mclean-linkage
8310 @opindex mno-clean-linkage
8311 These options are not fully implemented.
8315 @itemx -mic2.0-compat
8316 @itemx -mic3.0-compat
8318 @opindex mic2.0-compat
8319 @opindex mic3.0-compat
8320 Enable compatibility with iC960 v2.0 or v3.0.
8324 @opindex masm-compat
8326 Enable compatibility with the iC960 assembler.
8328 @item -mstrict-align
8329 @itemx -mno-strict-align
8330 @opindex mstrict-align
8331 @opindex mno-strict-align
8332 Do not permit (do permit) unaligned accesses.
8336 Enable structure-alignment compatibility with Intel's gcc release version
8337 1.3 (based on gcc 1.37). This option implies @option{-mstrict-align}.
8339 @item -mlong-double-64
8340 @opindex mlong-double-64
8341 Implement type @samp{long double} as 64-bit floating point numbers.
8342 Without the option @samp{long double} is implemented by 80-bit
8343 floating point numbers. The only reason we have it because there is
8344 no 128-bit @samp{long double} support in @samp{fp-bit.c} yet. So it
8345 is only useful for people using soft-float targets. Otherwise, we
8346 should recommend against use of it.
8350 @node DEC Alpha Options
8351 @subsection DEC Alpha Options
8353 These @samp{-m} options are defined for the DEC Alpha implementations:
8356 @item -mno-soft-float
8358 @opindex mno-soft-float
8359 @opindex msoft-float
8360 Use (do not use) the hardware floating-point instructions for
8361 floating-point operations. When @option{-msoft-float} is specified,
8362 functions in @file{libgcc.a} will be used to perform floating-point
8363 operations. Unless they are replaced by routines that emulate the
8364 floating-point operations, or compiled in such a way as to call such
8365 emulations routines, these routines will issue floating-point
8366 operations. If you are compiling for an Alpha without floating-point
8367 operations, you must ensure that the library is built so as not to call
8370 Note that Alpha implementations without floating-point operations are
8371 required to have floating-point registers.
8376 @opindex mno-fp-regs
8377 Generate code that uses (does not use) the floating-point register set.
8378 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
8379 register set is not used, floating point operands are passed in integer
8380 registers as if they were integers and floating-point results are passed
8381 in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence,
8382 so any function with a floating-point argument or return value called by code
8383 compiled with @option{-mno-fp-regs} must also be compiled with that
8386 A typical use of this option is building a kernel that does not use,
8387 and hence need not save and restore, any floating-point registers.
8391 The Alpha architecture implements floating-point hardware optimized for
8392 maximum performance. It is mostly compliant with the IEEE floating
8393 point standard. However, for full compliance, software assistance is
8394 required. This option generates code fully IEEE compliant code
8395 @emph{except} that the @var{inexact-flag} is not maintained (see below).
8396 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
8397 defined during compilation. The resulting code is less efficient but is
8398 able to correctly support denormalized numbers and exceptional IEEE
8399 values such as not-a-number and plus/minus infinity. Other Alpha
8400 compilers call this option @option{-ieee_with_no_inexact}.
8402 @item -mieee-with-inexact
8403 @opindex mieee-with-inexact
8404 This is like @option{-mieee} except the generated code also maintains
8405 the IEEE @var{inexact-flag}. Turning on this option causes the
8406 generated code to implement fully-compliant IEEE math. In addition to
8407 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
8408 macro. On some Alpha implementations the resulting code may execute
8409 significantly slower than the code generated by default. Since there is
8410 very little code that depends on the @var{inexact-flag}, you should
8411 normally not specify this option. Other Alpha compilers call this
8412 option @option{-ieee_with_inexact}.
8414 @item -mfp-trap-mode=@var{trap-mode}
8415 @opindex mfp-trap-mode
8416 This option controls what floating-point related traps are enabled.
8417 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
8418 The trap mode can be set to one of four values:
8422 This is the default (normal) setting. The only traps that are enabled
8423 are the ones that cannot be disabled in software (e.g., division by zero
8427 In addition to the traps enabled by @samp{n}, underflow traps are enabled
8431 Like @samp{su}, but the instructions are marked to be safe for software
8432 completion (see Alpha architecture manual for details).
8435 Like @samp{su}, but inexact traps are enabled as well.
8438 @item -mfp-rounding-mode=@var{rounding-mode}
8439 @opindex mfp-rounding-mode
8440 Selects the IEEE rounding mode. Other Alpha compilers call this option
8441 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
8446 Normal IEEE rounding mode. Floating point numbers are rounded towards
8447 the nearest machine number or towards the even machine number in case
8451 Round towards minus infinity.
8454 Chopped rounding mode. Floating point numbers are rounded towards zero.
8457 Dynamic rounding mode. A field in the floating point control register
8458 (@var{fpcr}, see Alpha architecture reference manual) controls the
8459 rounding mode in effect. The C library initializes this register for
8460 rounding towards plus infinity. Thus, unless your program modifies the
8461 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
8464 @item -mtrap-precision=@var{trap-precision}
8465 @opindex mtrap-precision
8466 In the Alpha architecture, floating point traps are imprecise. This
8467 means without software assistance it is impossible to recover from a
8468 floating trap and program execution normally needs to be terminated.
8469 GCC can generate code that can assist operating system trap handlers
8470 in determining the exact location that caused a floating point trap.
8471 Depending on the requirements of an application, different levels of
8472 precisions can be selected:
8476 Program precision. This option is the default and means a trap handler
8477 can only identify which program caused a floating point exception.
8480 Function precision. The trap handler can determine the function that
8481 caused a floating point exception.
8484 Instruction precision. The trap handler can determine the exact
8485 instruction that caused a floating point exception.
8488 Other Alpha compilers provide the equivalent options called
8489 @option{-scope_safe} and @option{-resumption_safe}.
8491 @item -mieee-conformant
8492 @opindex mieee-conformant
8493 This option marks the generated code as IEEE conformant. You must not
8494 use this option unless you also specify @option{-mtrap-precision=i} and either
8495 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
8496 is to emit the line @samp{.eflag 48} in the function prologue of the
8497 generated assembly file. Under DEC Unix, this has the effect that
8498 IEEE-conformant math library routines will be linked in.
8500 @item -mbuild-constants
8501 @opindex mbuild-constants
8502 Normally GCC examines a 32- or 64-bit integer constant to
8503 see if it can construct it from smaller constants in two or three
8504 instructions. If it cannot, it will output the constant as a literal and
8505 generate code to load it from the data segment at runtime.
8507 Use this option to require GCC to construct @emph{all} integer constants
8508 using code, even if it takes more instructions (the maximum is six).
8510 You would typically use this option to build a shared library dynamic
8511 loader. Itself a shared library, it must relocate itself in memory
8512 before it can find the variables and constants in its own data segment.
8518 Select whether to generate code to be assembled by the vendor-supplied
8519 assembler (@option{-malpha-as}) or by the GNU assembler @option{-mgas}.
8537 Indicate whether GCC should generate code to use the optional BWX,
8538 CIX, FIX and MAX instruction sets. The default is to use the instruction
8539 sets supported by the CPU type specified via @option{-mcpu=} option or that
8540 of the CPU on which GCC was built if none was specified.
8545 @opindex mfloat-ieee
8546 Generate code that uses (does not use) VAX F and G floating point
8547 arithmetic instead of IEEE single and double precision.
8549 @item -mexplicit-relocs
8550 @itemx -mno-explicit-relocs
8551 @opindex mexplicit-relocs
8552 @opindex mno-explicit-relocs
8553 Older Alpha assemblers provided no way to generate symbol relocations
8554 except via assembler macros. Use of these macros does not allow
8555 optimal instruction scheduling. GNU binutils as of version 2.12
8556 supports a new syntax that allows the compiler to explicitly mark
8557 which relocations should apply to which instructions. This option
8558 is mostly useful for debugging, as GCC detects the capabilities of
8559 the assembler when it is built and sets the default accordingly.
8563 @opindex msmall-data
8564 @opindex mlarge-data
8565 When @option{-mexplicit-relocs} is in effect, static data is
8566 accessed via @dfn{gp-relative} relocations. When @option{-msmall-data}
8567 is used, objects 8 bytes long or smaller are placed in a @dfn{small data area}
8568 (the @code{.sdata} and @code{.sbss} sections) and are accessed via
8569 16-bit relocations off of the @code{$gp} register. This limits the
8570 size of the small data area to 64KB, but allows the variables to be
8571 directly accessed via a single instruction.
8573 The default is @option{-mlarge-data}. With this option the data area
8574 is limited to just below 2GB. Programs that require more than 2GB of
8575 data must use @code{malloc} or @code{mmap} to allocate the data in the
8576 heap instead of in the program's data segment.
8578 When generating code for shared libraries, @option{-fpic} implies
8579 @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}.
8583 @opindex msmall-text
8584 @opindex mlarge-text
8585 When @option{-msmall-text} is used, the compiler assumes that the
8586 code of the entire program (or shared library) fits in 4MB, and is
8587 thus reachable with a branch instruction. When @option{-msmall-data}
8588 is used, the compiler can assume that all local symbols share the
8589 same @code{$gp} value, and thus reduce the number of instructions
8590 required for a function call from 4 to 1.
8592 The default is @option{-mlarge-text}.
8594 @item -mcpu=@var{cpu_type}
8596 Set the instruction set and instruction scheduling parameters for
8597 machine type @var{cpu_type}. You can specify either the @samp{EV}
8598 style name or the corresponding chip number. GCC supports scheduling
8599 parameters for the EV4, EV5 and EV6 family of processors and will
8600 choose the default values for the instruction set from the processor
8601 you specify. If you do not specify a processor type, GCC will default
8602 to the processor on which the compiler was built.
8604 Supported values for @var{cpu_type} are
8610 Schedules as an EV4 and has no instruction set extensions.
8614 Schedules as an EV5 and has no instruction set extensions.
8618 Schedules as an EV5 and supports the BWX extension.
8623 Schedules as an EV5 and supports the BWX and MAX extensions.
8627 Schedules as an EV6 and supports the BWX, FIX, and MAX extensions.
8631 Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions.
8634 @item -mtune=@var{cpu_type}
8636 Set only the instruction scheduling parameters for machine type
8637 @var{cpu_type}. The instruction set is not changed.
8639 @item -mmemory-latency=@var{time}
8640 @opindex mmemory-latency
8641 Sets the latency the scheduler should assume for typical memory
8642 references as seen by the application. This number is highly
8643 dependent on the memory access patterns used by the application
8644 and the size of the external cache on the machine.
8646 Valid options for @var{time} are
8650 A decimal number representing clock cycles.
8656 The compiler contains estimates of the number of clock cycles for
8657 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
8658 (also called Dcache, Scache, and Bcache), as well as to main memory.
8659 Note that L3 is only valid for EV5.
8664 @node DEC Alpha/VMS Options
8665 @subsection DEC Alpha/VMS Options
8667 These @samp{-m} options are defined for the DEC Alpha/VMS implementations:
8670 @item -mvms-return-codes
8671 @opindex mvms-return-codes
8672 Return VMS condition codes from main. The default is to return POSIX
8673 style condition (e.g.@ error) codes.
8676 @node H8/300 Options
8677 @subsection H8/300 Options
8679 These @samp{-m} options are defined for the H8/300 implementations:
8684 Shorten some address references at link time, when possible; uses the
8685 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
8686 ld.info, Using ld}, for a fuller description.
8690 Generate code for the H8/300H@.
8694 Generate code for the H8S@.
8698 Generate code for the H8S and H8/300H in the normal mode. This switch
8699 must be used either with -mh or -ms.
8703 Generate code for the H8S/2600. This switch must be used with @option{-ms}.
8707 Make @code{int} data 32 bits by default.
8711 On the H8/300H and H8S, use the same alignment rules as for the H8/300.
8712 The default for the H8/300H and H8S is to align longs and floats on 4
8714 @option{-malign-300} causes them to be aligned on 2 byte boundaries.
8715 This option has no effect on the H8/300.
8719 @subsection SH Options
8721 These @samp{-m} options are defined for the SH implementations:
8726 Generate code for the SH1.
8730 Generate code for the SH2.
8734 Generate code for the SH3.
8738 Generate code for the SH3e.
8742 Generate code for the SH4 without a floating-point unit.
8744 @item -m4-single-only
8745 @opindex m4-single-only
8746 Generate code for the SH4 with a floating-point unit that only
8747 supports single-precision arithmetic.
8751 Generate code for the SH4 assuming the floating-point unit is in
8752 single-precision mode by default.
8756 Generate code for the SH4.
8760 Compile code for the processor in big endian mode.
8764 Compile code for the processor in little endian mode.
8768 Align doubles at 64-bit boundaries. Note that this changes the calling
8769 conventions, and thus some functions from the standard C library will
8770 not work unless you recompile it first with @option{-mdalign}.
8774 Shorten some address references at link time, when possible; uses the
8775 linker option @option{-relax}.
8779 Use 32-bit offsets in @code{switch} tables. The default is to use
8784 Enable the use of the instruction @code{fmovd}.
8788 Comply with the calling conventions defined by Hitachi.
8792 Mark the @code{MAC} register as call-clobbered, even if
8793 @option{-mhitachi} is given.
8797 Increase IEEE-compliance of floating-point code.
8801 Dump instruction size and location in the assembly code.
8805 This option is deprecated. It pads structures to multiple of 4 bytes,
8806 which is incompatible with the SH ABI@.
8810 Optimize for space instead of speed. Implied by @option{-Os}.
8814 When generating position-independent code, emit function calls using
8815 the Global Offset Table instead of the Procedure Linkage Table.
8819 Generate a library function call to invalidate instruction cache
8820 entries, after fixing up a trampoline. This library function call
8821 doesn't assume it can write to the whole memory address space. This
8822 is the default when the target is @code{sh-*-linux*}.
8825 @node System V Options
8826 @subsection Options for System V
8828 These additional options are available on System V Release 4 for
8829 compatibility with other compilers on those systems:
8834 Create a shared object.
8835 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
8839 Identify the versions of each tool used by the compiler, in a
8840 @code{.ident} assembler directive in the output.
8844 Refrain from adding @code{.ident} directives to the output file (this is
8847 @item -YP,@var{dirs}
8849 Search the directories @var{dirs}, and no others, for libraries
8850 specified with @option{-l}.
8854 Look in the directory @var{dir} to find the M4 preprocessor.
8855 The assembler uses this option.
8856 @c This is supposed to go with a -Yd for predefined M4 macro files, but
8857 @c the generic assembler that comes with Solaris takes just -Ym.
8860 @node TMS320C3x/C4x Options
8861 @subsection TMS320C3x/C4x Options
8862 @cindex TMS320C3x/C4x Options
8864 These @samp{-m} options are defined for TMS320C3x/C4x implementations:
8868 @item -mcpu=@var{cpu_type}
8870 Set the instruction set, register set, and instruction scheduling
8871 parameters for machine type @var{cpu_type}. Supported values for
8872 @var{cpu_type} are @samp{c30}, @samp{c31}, @samp{c32}, @samp{c40}, and
8873 @samp{c44}. The default is @samp{c40} to generate code for the
8878 @itemx -msmall-memory
8880 @opindex mbig-memory
8882 @opindex msmall-memory
8884 Generates code for the big or small memory model. The small memory
8885 model assumed that all data fits into one 64K word page. At run-time
8886 the data page (DP) register must be set to point to the 64K page
8887 containing the .bss and .data program sections. The big memory model is
8888 the default and requires reloading of the DP register for every direct
8895 Allow (disallow) allocation of general integer operands into the block
8902 Enable (disable) generation of code using decrement and branch,
8903 DBcond(D), instructions. This is enabled by default for the C4x. To be
8904 on the safe side, this is disabled for the C3x, since the maximum
8905 iteration count on the C3x is @math{2^{23} + 1} (but who iterates loops more than
8906 @math{2^{23}} times on the C3x?). Note that GCC will try to reverse a loop so
8907 that it can utilize the decrement and branch instruction, but will give
8908 up if there is more than one memory reference in the loop. Thus a loop
8909 where the loop counter is decremented can generate slightly more
8910 efficient code, in cases where the RPTB instruction cannot be utilized.
8912 @item -mdp-isr-reload
8914 @opindex mdp-isr-reload
8916 Force the DP register to be saved on entry to an interrupt service
8917 routine (ISR), reloaded to point to the data section, and restored on
8918 exit from the ISR@. This should not be required unless someone has
8919 violated the small memory model by modifying the DP register, say within
8926 For the C3x use the 24-bit MPYI instruction for integer multiplies
8927 instead of a library call to guarantee 32-bit results. Note that if one
8928 of the operands is a constant, then the multiplication will be performed
8929 using shifts and adds. If the @option{-mmpyi} option is not specified for the C3x,
8930 then squaring operations are performed inline instead of a library call.
8933 @itemx -mno-fast-fix
8935 @opindex mno-fast-fix
8936 The C3x/C4x FIX instruction to convert a floating point value to an
8937 integer value chooses the nearest integer less than or equal to the
8938 floating point value rather than to the nearest integer. Thus if the
8939 floating point number is negative, the result will be incorrectly
8940 truncated an additional code is necessary to detect and correct this
8941 case. This option can be used to disable generation of the additional
8942 code required to correct the result.
8948 Enable (disable) generation of repeat block sequences using the RPTB
8949 instruction for zero overhead looping. The RPTB construct is only used
8950 for innermost loops that do not call functions or jump across the loop
8951 boundaries. There is no advantage having nested RPTB loops due to the
8952 overhead required to save and restore the RC, RS, and RE registers.
8953 This is enabled by default with @option{-O2}.
8955 @item -mrpts=@var{count}
8959 Enable (disable) the use of the single instruction repeat instruction
8960 RPTS@. If a repeat block contains a single instruction, and the loop
8961 count can be guaranteed to be less than the value @var{count}, GCC will
8962 emit a RPTS instruction instead of a RPTB@. If no value is specified,
8963 then a RPTS will be emitted even if the loop count cannot be determined
8964 at compile time. Note that the repeated instruction following RPTS does
8965 not have to be reloaded from memory each iteration, thus freeing up the
8966 CPU buses for operands. However, since interrupts are blocked by this
8967 instruction, it is disabled by default.
8969 @item -mloop-unsigned
8970 @itemx -mno-loop-unsigned
8971 @opindex mloop-unsigned
8972 @opindex mno-loop-unsigned
8973 The maximum iteration count when using RPTS and RPTB (and DB on the C40)
8974 is @math{2^{31} + 1} since these instructions test if the iteration count is
8975 negative to terminate the loop. If the iteration count is unsigned
8976 there is a possibility than the @math{2^{31} + 1} maximum iteration count may be
8977 exceeded. This switch allows an unsigned iteration count.
8981 Try to emit an assembler syntax that the TI assembler (asm30) is happy
8982 with. This also enforces compatibility with the API employed by the TI
8983 C3x C compiler. For example, long doubles are passed as structures
8984 rather than in floating point registers.
8990 Generate code that uses registers (stack) for passing arguments to functions.
8991 By default, arguments are passed in registers where possible rather
8992 than by pushing arguments on to the stack.
8994 @item -mparallel-insns
8995 @itemx -mno-parallel-insns
8996 @opindex mparallel-insns
8997 @opindex mno-parallel-insns
8998 Allow the generation of parallel instructions. This is enabled by
8999 default with @option{-O2}.
9001 @item -mparallel-mpy
9002 @itemx -mno-parallel-mpy
9003 @opindex mparallel-mpy
9004 @opindex mno-parallel-mpy
9005 Allow the generation of MPY||ADD and MPY||SUB parallel instructions,
9006 provided @option{-mparallel-insns} is also specified. These instructions have
9007 tight register constraints which can pessimize the code generation
9013 @subsection V850 Options
9014 @cindex V850 Options
9016 These @samp{-m} options are defined for V850 implementations:
9020 @itemx -mno-long-calls
9021 @opindex mlong-calls
9022 @opindex mno-long-calls
9023 Treat all calls as being far away (near). If calls are assumed to be
9024 far away, the compiler will always load the functions address up into a
9025 register, and call indirect through the pointer.
9031 Do not optimize (do optimize) basic blocks that use the same index
9032 pointer 4 or more times to copy pointer into the @code{ep} register, and
9033 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
9034 option is on by default if you optimize.
9036 @item -mno-prolog-function
9037 @itemx -mprolog-function
9038 @opindex mno-prolog-function
9039 @opindex mprolog-function
9040 Do not use (do use) external functions to save and restore registers at
9041 the prolog and epilog of a function. The external functions are slower,
9042 but use less code space if more than one function saves the same number
9043 of registers. The @option{-mprolog-function} option is on by default if
9048 Try to make the code as small as possible. At present, this just turns
9049 on the @option{-mep} and @option{-mprolog-function} options.
9053 Put static or global variables whose size is @var{n} bytes or less into
9054 the tiny data area that register @code{ep} points to. The tiny data
9055 area can hold up to 256 bytes in total (128 bytes for byte references).
9059 Put static or global variables whose size is @var{n} bytes or less into
9060 the small data area that register @code{gp} points to. The small data
9061 area can hold up to 64 kilobytes.
9065 Put static or global variables whose size is @var{n} bytes or less into
9066 the first 32 kilobytes of memory.
9070 Specify that the target processor is the V850.
9073 @opindex mbig-switch
9074 Generate code suitable for big switch tables. Use this option only if
9075 the assembler/linker complain about out of range branches within a switch
9080 This option will cause r2 and r5 to be used in the code generated by
9081 the compiler. This setting is the default.
9084 @opindex -mno-app-regs
9085 This option will cause r2 and r5 to be treated as fixed registers.
9089 Specify that the target processor is the V850E. The preprocessor
9090 constant @samp{__v850e__} will be defined if this option is used.
9092 If neither @option{-mv850} nor @option{-mv850e} are defined
9093 then a default target processor will be chosen and the relevant
9094 @samp{__v850*__} preprocessor constant will be defined.
9096 The preprocessor constants @samp{__v850} and @samp{__v851__} are always
9097 defined, regardless of which processor variant is the target.
9099 @item -mdisable-callt
9100 @opindex -mdisable-callt
9101 This option will suppress generation of the CALLT instruction for the
9102 v850e flavors of the v850 architecture. The default is
9103 @option{-mno-disable-callt} which allows the CALLT instruction to be used.
9108 @subsection ARC Options
9111 These options are defined for ARC implementations:
9116 Compile code for little endian mode. This is the default.
9120 Compile code for big endian mode.
9123 @opindex mmangle-cpu
9124 Prepend the name of the cpu to all public symbol names.
9125 In multiple-processor systems, there are many ARC variants with different
9126 instruction and register set characteristics. This flag prevents code
9127 compiled for one cpu to be linked with code compiled for another.
9128 No facility exists for handling variants that are ``almost identical''.
9129 This is an all or nothing option.
9131 @item -mcpu=@var{cpu}
9133 Compile code for ARC variant @var{cpu}.
9134 Which variants are supported depend on the configuration.
9135 All variants support @option{-mcpu=base}, this is the default.
9137 @item -mtext=@var{text-section}
9138 @itemx -mdata=@var{data-section}
9139 @itemx -mrodata=@var{readonly-data-section}
9143 Put functions, data, and readonly data in @var{text-section},
9144 @var{data-section}, and @var{readonly-data-section} respectively
9145 by default. This can be overridden with the @code{section} attribute.
9146 @xref{Variable Attributes}.
9151 @subsection NS32K Options
9152 @cindex NS32K options
9154 These are the @samp{-m} options defined for the 32000 series. The default
9155 values for these options depends on which style of 32000 was selected when
9156 the compiler was configured; the defaults for the most common choices are
9164 Generate output for a 32032. This is the default
9165 when the compiler is configured for 32032 and 32016 based systems.
9171 Generate output for a 32332. This is the default
9172 when the compiler is configured for 32332-based systems.
9178 Generate output for a 32532. This is the default
9179 when the compiler is configured for 32532-based systems.
9183 Generate output containing 32081 instructions for floating point.
9184 This is the default for all systems.
9188 Generate output containing 32381 instructions for floating point. This
9189 also implies @option{-m32081}. The 32381 is only compatible with the 32332
9190 and 32532 cpus. This is the default for the pc532-netbsd configuration.
9194 Try and generate multiply-add floating point instructions @code{polyF}
9195 and @code{dotF}. This option is only available if the @option{-m32381}
9196 option is in effect. Using these instructions requires changes to
9197 register allocation which generally has a negative impact on
9198 performance. This option should only be enabled when compiling code
9199 particularly likely to make heavy use of multiply-add instructions.
9202 @opindex mnomulti-add
9203 Do not try and generate multiply-add floating point instructions
9204 @code{polyF} and @code{dotF}. This is the default on all platforms.
9207 @opindex msoft-float
9208 Generate output containing library calls for floating point.
9209 @strong{Warning:} the requisite libraries may not be available.
9211 @item -mieee-compare
9212 @itemx -mno-ieee-compare
9213 @opindex mieee-compare
9214 @opindex mno-ieee-compare
9215 Control whether or not the compiler uses IEEE floating point
9216 comparisons. These handle correctly the case where the result of a
9217 comparison is unordered.
9218 @strong{Warning:} the requisite kernel support may not be available.
9221 @opindex mnobitfield
9222 Do not use the bit-field instructions. On some machines it is faster to
9223 use shifting and masking operations. This is the default for the pc532.
9227 Do use the bit-field instructions. This is the default for all platforms
9232 Use a different function-calling convention, in which functions
9233 that take a fixed number of arguments return pop their
9234 arguments on return with the @code{ret} instruction.
9236 This calling convention is incompatible with the one normally
9237 used on Unix, so you cannot use it if you need to call libraries
9238 compiled with the Unix compiler.
9240 Also, you must provide function prototypes for all functions that
9241 take variable numbers of arguments (including @code{printf});
9242 otherwise incorrect code will be generated for calls to those
9245 In addition, seriously incorrect code will result if you call a
9246 function with too many arguments. (Normally, extra arguments are
9247 harmlessly ignored.)
9249 This option takes its name from the 680x0 @code{rtd} instruction.
9254 Use a different function-calling convention where the first two arguments
9255 are passed in registers.
9257 This calling convention is incompatible with the one normally
9258 used on Unix, so you cannot use it if you need to call libraries
9259 compiled with the Unix compiler.
9262 @opindex mnoregparam
9263 Do not pass any arguments in registers. This is the default for all
9268 It is OK to use the sb as an index register which is always loaded with
9269 zero. This is the default for the pc532-netbsd target.
9273 The sb register is not available for use or has not been initialized to
9274 zero by the run time system. This is the default for all targets except
9275 the pc532-netbsd. It is also implied whenever @option{-mhimem} or
9276 @option{-fpic} is set.
9280 Many ns32000 series addressing modes use displacements of up to 512MB@.
9281 If an address is above 512MB then displacements from zero can not be used.
9282 This option causes code to be generated which can be loaded above 512MB@.
9283 This may be useful for operating systems or ROM code.
9287 Assume code will be loaded in the first 512MB of virtual address space.
9288 This is the default for all platforms.
9294 @subsection AVR Options
9297 These options are defined for AVR implementations:
9300 @item -mmcu=@var{mcu}
9302 Specify ATMEL AVR instruction set or MCU type.
9304 Instruction set avr1 is for the minimal AVR core, not supported by the C
9305 compiler, only for assembler programs (MCU types: at90s1200, attiny10,
9306 attiny11, attiny12, attiny15, attiny28).
9308 Instruction set avr2 (default) is for the classic AVR core with up to
9309 8K program memory space (MCU types: at90s2313, at90s2323, attiny22,
9310 at90s2333, at90s2343, at90s4414, at90s4433, at90s4434, at90s8515,
9311 at90c8534, at90s8535).
9313 Instruction set avr3 is for the classic AVR core with up to 128K program
9314 memory space (MCU types: atmega103, atmega603, at43usb320, at76c711).
9316 Instruction set avr4 is for the enhanced AVR core with up to 8K program
9317 memory space (MCU types: atmega8, atmega83, atmega85).
9319 Instruction set avr5 is for the enhanced AVR core with up to 128K program
9320 memory space (MCU types: atmega16, atmega161, atmega163, atmega32, atmega323,
9321 atmega64, atmega128, at43usb355, at94k).
9325 Output instruction sizes to the asm file.
9327 @item -minit-stack=@var{N}
9328 @opindex minit-stack
9329 Specify the initial stack address, which may be a symbol or numeric value,
9330 @samp{__stack} is the default.
9332 @item -mno-interrupts
9333 @opindex mno-interrupts
9334 Generated code is not compatible with hardware interrupts.
9335 Code size will be smaller.
9337 @item -mcall-prologues
9338 @opindex mcall-prologues
9339 Functions prologues/epilogues expanded as call to appropriate
9340 subroutines. Code size will be smaller.
9342 @item -mno-tablejump
9343 @opindex mno-tablejump
9344 Do not generate tablejump insns which sometimes increase code size.
9347 @opindex mtiny-stack
9348 Change only the low 8 bits of the stack pointer.
9352 @subsection MCore Options
9353 @cindex MCore options
9355 These are the @samp{-m} options defined for the Motorola M*Core
9365 @opindex mno-hardlit
9366 Inline constants into the code stream if it can be done in two
9367 instructions or less.
9375 Use the divide instruction. (Enabled by default).
9377 @item -mrelax-immediate
9378 @itemx -mrelax-immediate
9379 @itemx -mno-relax-immediate
9380 @opindex mrelax-immediate
9381 @opindex mrelax-immediate
9382 @opindex mno-relax-immediate
9383 Allow arbitrary sized immediates in bit operations.
9385 @item -mwide-bitfields
9386 @itemx -mwide-bitfields
9387 @itemx -mno-wide-bitfields
9388 @opindex mwide-bitfields
9389 @opindex mwide-bitfields
9390 @opindex mno-wide-bitfields
9391 Always treat bit-fields as int-sized.
9393 @item -m4byte-functions
9394 @itemx -m4byte-functions
9395 @itemx -mno-4byte-functions
9396 @opindex m4byte-functions
9397 @opindex m4byte-functions
9398 @opindex mno-4byte-functions
9399 Force all functions to be aligned to a four byte boundary.
9401 @item -mcallgraph-data
9402 @itemx -mcallgraph-data
9403 @itemx -mno-callgraph-data
9404 @opindex mcallgraph-data
9405 @opindex mcallgraph-data
9406 @opindex mno-callgraph-data
9407 Emit callgraph information.
9411 @itemx -mno-slow-bytes
9412 @opindex mslow-bytes
9413 @opindex mslow-bytes
9414 @opindex mno-slow-bytes
9415 Prefer word access when reading byte quantities.
9417 @item -mlittle-endian
9418 @itemx -mlittle-endian
9420 @opindex mlittle-endian
9421 @opindex mlittle-endian
9422 @opindex mbig-endian
9423 Generate code for a little endian target.
9431 Generate code for the 210 processor.
9435 @subsection IA-64 Options
9436 @cindex IA-64 Options
9438 These are the @samp{-m} options defined for the Intel IA-64 architecture.
9442 @opindex mbig-endian
9443 Generate code for a big endian target. This is the default for HP-UX@.
9445 @item -mlittle-endian
9446 @opindex mlittle-endian
9447 Generate code for a little endian target. This is the default for AIX5
9454 Generate (or don't) code for the GNU assembler. This is the default.
9455 @c Also, this is the default if the configure option @option{--with-gnu-as}
9462 Generate (or don't) code for the GNU linker. This is the default.
9463 @c Also, this is the default if the configure option @option{--with-gnu-ld}
9468 Generate code that does not use a global pointer register. The result
9469 is not position independent code, and violates the IA-64 ABI@.
9471 @item -mvolatile-asm-stop
9472 @itemx -mno-volatile-asm-stop
9473 @opindex mvolatile-asm-stop
9474 @opindex mno-volatile-asm-stop
9475 Generate (or don't) a stop bit immediately before and after volatile asm
9480 Generate code that works around Itanium B step errata.
9482 @item -mregister-names
9483 @itemx -mno-register-names
9484 @opindex mregister-names
9485 @opindex mno-register-names
9486 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
9487 the stacked registers. This may make assembler output more readable.
9493 Disable (or enable) optimizations that use the small data section. This may
9494 be useful for working around optimizer bugs.
9497 @opindex mconstant-gp
9498 Generate code that uses a single constant global pointer value. This is
9499 useful when compiling kernel code.
9503 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
9504 This is useful when compiling firmware code.
9506 @item -minline-float-divide-min-latency
9507 @opindex minline-float-divide-min-latency
9508 Generate code for inline divides of floating point values
9509 using the minimum latency algorithm.
9511 @item -minline-float-divide-max-throughput
9512 @opindex minline-float-divide-max-throughput
9513 Generate code for inline divides of floating point values
9514 using the maximum throughput algorithm.
9516 @item -minline-int-divide-min-latency
9517 @opindex minline-int-divide-min-latency
9518 Generate code for inline divides of integer values
9519 using the minimum latency algorithm.
9521 @item -minline-int-divide-max-throughput
9522 @opindex minline-int-divide-max-throughput
9523 Generate code for inline divides of integer values
9524 using the maximum throughput algorithm.
9526 @item -mno-dwarf2-asm
9528 @opindex mno-dwarf2-asm
9529 @opindex mdwarf2-asm
9530 Don't (or do) generate assembler code for the DWARF2 line number debugging
9531 info. This may be useful when not using the GNU assembler.
9533 @item -mfixed-range=@var{register-range}
9534 @opindex mfixed-range
9535 Generate code treating the given register range as fixed registers.
9536 A fixed register is one that the register allocator can not use. This is
9537 useful when compiling kernel code. A register range is specified as
9538 two registers separated by a dash. Multiple register ranges can be
9539 specified separated by a comma.
9541 @item -mearly-stop-bits
9542 @itemx -mno-early-stop-bits
9543 @opindex mearly-stop-bits
9544 @opindex mno-early-stop-bits
9545 Allow stop bits to be placed earlier than immediately preceding the
9546 instruction that triggered the stop bit. This can improve instruction
9547 scheduling, but does not always do so.
9551 @subsection D30V Options
9552 @cindex D30V Options
9554 These @samp{-m} options are defined for D30V implementations:
9559 Link the @samp{.text}, @samp{.data}, @samp{.bss}, @samp{.strings},
9560 @samp{.rodata}, @samp{.rodata1}, @samp{.data1} sections into external
9561 memory, which starts at location @code{0x80000000}.
9565 Same as the @option{-mextmem} switch.
9569 Link the @samp{.text} section into onchip text memory, which starts at
9570 location @code{0x0}. Also link @samp{.data}, @samp{.bss},
9571 @samp{.strings}, @samp{.rodata}, @samp{.rodata1}, @samp{.data1} sections
9572 into onchip data memory, which starts at location @code{0x20000000}.
9574 @item -mno-asm-optimize
9575 @itemx -masm-optimize
9576 @opindex mno-asm-optimize
9577 @opindex masm-optimize
9578 Disable (enable) passing @option{-O} to the assembler when optimizing.
9579 The assembler uses the @option{-O} option to automatically parallelize
9580 adjacent short instructions where possible.
9582 @item -mbranch-cost=@var{n}
9583 @opindex mbranch-cost
9584 Increase the internal costs of branches to @var{n}. Higher costs means
9585 that the compiler will issue more instructions to avoid doing a branch.
9588 @item -mcond-exec=@var{n}
9590 Specify the maximum number of conditionally executed instructions that
9591 replace a branch. The default is 4.
9594 @node S/390 and zSeries Options
9595 @subsection S/390 and zSeries Options
9596 @cindex S/390 and zSeries Options
9598 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
9603 @opindex mhard-float
9604 @opindex msoft-float
9605 Use (do not use) the hardware floating-point instructions and registers
9606 for floating-point operations. When @option{-msoft-float} is specified,
9607 functions in @file{libgcc.a} will be used to perform floating-point
9608 operations. When @option{-mhard-float} is specified, the compiler
9609 generates IEEE floating-point instructions. This is the default.
9612 @itemx -mno-backchain
9614 @opindex mno-backchain
9615 Generate (or do not generate) code which maintains an explicit
9616 backchain within the stack frame that points to the caller's frame.
9617 This is currently needed to allow debugging. The default is to
9618 generate the backchain.
9621 @itemx -mno-small-exec
9622 @opindex msmall-exec
9623 @opindex mno-small-exec
9624 Generate (or do not generate) code using the @code{bras} instruction
9625 to do subroutine calls.
9626 This only works reliably if the total executable size does not
9627 exceed 64k. The default is to use the @code{basr} instruction instead,
9628 which does not have this limitation.
9634 When @option{-m31} is specified, generate code compliant to the
9635 Linux for S/390 ABI@. When @option{-m64} is specified, generate
9636 code compliant to the Linux for zSeries ABI@. This allows GCC in
9637 particular to generate 64-bit instructions. For the @samp{s390}
9638 targets, the default is @option{-m31}, while the @samp{s390x}
9639 targets default to @option{-m64}.
9645 When @option{-mzarch} is specified, generate code using the
9646 instructions available on z/Architecture.
9647 When @option{-mesa} is specified, generate code using the
9648 instructions available on ESA/390. Note that @option{-mesa} is
9649 not possible with @option{-m64}.
9650 For the @samp{s390} targets, the default is @option{-mesa},
9651 while the @samp{s390x} targets default to @option{-mzarch}.
9657 Generate (or do not generate) code using the @code{mvcle} instruction
9658 to perform block moves. When @option{-mno-mvcle} is specified,
9659 use a @code{mvc} loop instead. This is the default.
9665 Print (or do not print) additional debug information when compiling.
9666 The default is to not print debug information.
9668 @item -march=@var{arch}
9670 Generate code that will run on @var{arch}, which is the name of system
9671 representing a certain processor type. Possible values for
9672 @var{cpu-type} are @samp{g5}, @samp{g6} and @samp{z900}.
9674 @item -mtune=@var{arch}
9676 Tune to @var{cpu-type} everything applicable about the generated code,
9677 except for the ABI and the set of available instructions.
9678 The list of @var{arch} values is the same as for @option{-march}.
9683 @subsection CRIS Options
9684 @cindex CRIS Options
9686 These options are defined specifically for the CRIS ports.
9689 @item -march=@var{architecture-type}
9690 @itemx -mcpu=@var{architecture-type}
9693 Generate code for the specified architecture. The choices for
9694 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
9695 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX.
9696 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
9699 @item -mtune=@var{architecture-type}
9701 Tune to @var{architecture-type} everything applicable about the generated
9702 code, except for the ABI and the set of available instructions. The
9703 choices for @var{architecture-type} are the same as for
9704 @option{-march=@var{architecture-type}}.
9706 @item -mmax-stack-frame=@var{n}
9707 @opindex mmax-stack-frame
9708 Warn when the stack frame of a function exceeds @var{n} bytes.
9710 @item -melinux-stacksize=@var{n}
9711 @opindex melinux-stacksize
9712 Only available with the @samp{cris-axis-aout} target. Arranges for
9713 indications in the program to the kernel loader that the stack of the
9714 program should be set to @var{n} bytes.
9720 The options @option{-metrax4} and @option{-metrax100} are synonyms for
9721 @option{-march=v3} and @option{-march=v8} respectively.
9725 Enable CRIS-specific verbose debug-related information in the assembly
9726 code. This option also has the effect to turn off the @samp{#NO_APP}
9727 formatted-code indicator to the assembler at the beginning of the
9732 Do not use condition-code results from previous instruction; always emit
9733 compare and test instructions before use of condition codes.
9735 @item -mno-side-effects
9736 @opindex mno-side-effects
9737 Do not emit instructions with side-effects in addressing modes other than
9741 @itemx -mno-stack-align
9743 @itemx -mno-data-align
9744 @itemx -mconst-align
9745 @itemx -mno-const-align
9746 @opindex mstack-align
9747 @opindex mno-stack-align
9748 @opindex mdata-align
9749 @opindex mno-data-align
9750 @opindex mconst-align
9751 @opindex mno-const-align
9752 These options (no-options) arranges (eliminate arrangements) for the
9753 stack-frame, individual data and constants to be aligned for the maximum
9754 single data access size for the chosen CPU model. The default is to
9755 arrange for 32-bit alignment. ABI details such as structure layout are
9756 not affected by these options.
9764 Similar to the stack- data- and const-align options above, these options
9765 arrange for stack-frame, writable data and constants to all be 32-bit,
9766 16-bit or 8-bit aligned. The default is 32-bit alignment.
9768 @item -mno-prologue-epilogue
9769 @itemx -mprologue-epilogue
9770 @opindex mno-prologue-epilogue
9771 @opindex mprologue-epilogue
9772 With @option{-mno-prologue-epilogue}, the normal function prologue and
9773 epilogue that sets up the stack-frame are omitted and no return
9774 instructions or return sequences are generated in the code. Use this
9775 option only together with visual inspection of the compiled code: no
9776 warnings or errors are generated when call-saved registers must be saved,
9777 or storage for local variable needs to be allocated.
9783 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
9784 instruction sequences that load addresses for functions from the PLT part
9785 of the GOT rather than (traditional on other architectures) calls to the
9786 PLT. The default is @option{-mgotplt}.
9790 Legacy no-op option only recognized with the cris-axis-aout target.
9794 Legacy no-op option only recognized with the cris-axis-elf and
9795 cris-axis-linux-gnu targets.
9799 Only recognized with the cris-axis-aout target, where it selects a
9800 GNU/linux-like multilib, include files and instruction set for
9805 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
9809 This option, recognized for the cris-axis-aout and cris-axis-elf arranges
9810 to link with input-output functions from a simulator library. Code,
9811 initialized data and zero-initialized data are allocated consecutively.
9815 Like @option{-sim}, but pass linker options to locate initialized data at
9816 0x40000000 and zero-initialized data at 0x80000000.
9820 @subsection MMIX Options
9821 @cindex MMIX Options
9823 These options are defined for the MMIX:
9827 @itemx -mno-libfuncs
9829 @opindex mno-libfuncs
9830 Specify that intrinsic library functions are being compiled, passing all
9831 values in registers, no matter the size.
9836 @opindex mno-epsilon
9837 Generate floating-point comparison instructions that compare with respect
9838 to the @code{rE} epsilon register.
9840 @item -mabi=mmixware
9842 @opindex mabi-mmixware
9844 Generate code that passes function parameters and return values that (in
9845 the called function) are seen as registers @code{$0} and up, as opposed to
9846 the GNU ABI which uses global registers @code{$231} and up.
9849 @itemx -mno-zero-extend
9850 @opindex mzero-extend
9851 @opindex mno-zero-extend
9852 When reading data from memory in sizes shorter than 64 bits, use (do not
9853 use) zero-extending load instructions by default, rather than
9854 sign-extending ones.
9857 @itemx -mno-knuthdiv
9859 @opindex mno-knuthdiv
9860 Make the result of a division yielding a remainder have the same sign as
9861 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
9862 remainder follows the sign of the dividend. Both methods are
9863 arithmetically valid, the latter being almost exclusively used.
9865 @item -mtoplevel-symbols
9866 @itemx -mno-toplevel-symbols
9867 @opindex mtoplevel-symbols
9868 @opindex mno-toplevel-symbols
9869 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
9870 code can be used with the @code{PREFIX} assembly directive.
9874 Generate an executable in the ELF format, rather than the default
9875 @samp{mmo} format used by the @command{mmix} simulator.
9877 @item -mbranch-predict
9878 @itemx -mno-branch-predict
9879 @opindex mbranch-predict
9880 @opindex mno-branch-predict
9881 Use (do not use) the probable-branch instructions, when static branch
9882 prediction indicates a probable branch.
9884 @item -mbase-addresses
9885 @itemx -mno-base-addresses
9886 @opindex mbase-addresses
9887 @opindex mno-base-addresses
9888 Generate (do not generate) code that uses @emph{base addresses}. Using a
9889 base address automatically generates a request (handled by the assembler
9890 and the linker) for a constant to be set up in a global register. The
9891 register is used for one or more base address requests within the range 0
9892 to 255 from the value held in the register. The generally leads to short
9893 and fast code, but the number of different data items that can be
9894 addressed is limited. This means that a program that uses lots of static
9895 data may require @option{-mno-base-addresses}.
9898 @itemx -mno-single-exit
9899 @opindex msingle-exit
9900 @opindex mno-single-exit
9901 Force (do not force) generated code to have a single exit point in each
9905 @node PDP-11 Options
9906 @subsection PDP-11 Options
9907 @cindex PDP-11 Options
9909 These options are defined for the PDP-11:
9914 Use hardware FPP floating point. This is the default. (FIS floating
9915 point on the PDP-11/40 is not supported.)
9918 @opindex msoft-float
9919 Do not use hardware floating point.
9923 Return floating-point results in ac0 (fr0 in Unix assembler syntax).
9927 Return floating-point results in memory. This is the default.
9931 Generate code for a PDP-11/40.
9935 Generate code for a PDP-11/45. This is the default.
9939 Generate code for a PDP-11/10.
9941 @item -mbcopy-builtin
9942 @opindex bcopy-builtin
9943 Use inline @code{movstrhi} patterns for copying memory. This is the
9948 Do not use inline @code{movstrhi} patterns for copying memory.
9954 Use 16-bit @code{int}. This is the default.
9960 Use 32-bit @code{int}.
9965 @opindex mno-float32
9966 Use 64-bit @code{float}. This is the default.
9971 @opindex mno-float64
9972 Use 32-bit @code{float}.
9976 Use @code{abshi2} pattern. This is the default.
9980 Do not use @code{abshi2} pattern.
9982 @item -mbranch-expensive
9983 @opindex mbranch-expensive
9984 Pretend that branches are expensive. This is for experimenting with
9985 code generation only.
9987 @item -mbranch-cheap
9988 @opindex mbranch-cheap
9989 Do not pretend that branches are expensive. This is the default.
9993 Generate code for a system with split I&D.
9997 Generate code for a system without split I&D. This is the default.
10001 Use Unix assembler syntax. This is the default when configured for
10002 @samp{pdp11-*-bsd}.
10006 Use DEC assembler syntax. This is the default when configured for any
10007 PDP-11 target other than @samp{pdp11-*-bsd}.
10010 @node Xstormy16 Options
10011 @subsection Xstormy16 Options
10012 @cindex Xstormy16 Options
10014 These options are defined for Xstormy16:
10019 Choose startup files and linker script suitable for the simulator.
10023 @subsection FRV Options
10024 @cindex FRV Options
10030 Only use the first 32 general purpose registers.
10035 Use all 64 general purpose registers.
10040 Use only the first 32 floating point registers.
10045 Use all 64 floating point registers
10048 @opindex mhard-float
10050 Use hardware instructions for floating point operations.
10053 @opindex msoft-float
10055 Use library routines for floating point operations.
10060 Dynamically allocate condition code registers.
10065 Do not try to dynamically allocate condition code registers, only
10066 use @code{icc0} and @code{fcc0}.
10071 Change ABI to use double word insns.
10076 Do not use double word instructions.
10081 Use floating point double instructions.
10084 @opindex mno-double
10086 Do not use floating point double instructions.
10091 Use media instructions.
10096 Do not use media instructions.
10101 Use multiply and add/subtract instructions.
10104 @opindex mno-muladd
10106 Do not use multiply and add/subtract instructions.
10108 @item -mlibrary-pic
10109 @opindex mlibrary-pic
10111 Enable PIC support for building libraries
10116 Use only the first four media accumulator registers.
10121 Use all eight media accumulator registers.
10126 Pack VLIW instructions.
10131 Do not pack VLIW instructions.
10134 @opindex mno-eflags
10136 Do not mark ABI switches in e_flags.
10139 @opindex mcond-move
10141 Enable the use of conditional-move instructions (default).
10143 This switch is mainly for debugging the compiler and will likely be removed
10144 in a future version.
10146 @item -mno-cond-move
10147 @opindex mno-cond-move
10149 Disable the use of conditional-move instructions.
10151 This switch is mainly for debugging the compiler and will likely be removed
10152 in a future version.
10157 Enable the use of conditional set instructions (default).
10159 This switch is mainly for debugging the compiler and will likely be removed
10160 in a future version.
10165 Disable the use of conditional set instructions.
10167 This switch is mainly for debugging the compiler and will likely be removed
10168 in a future version.
10171 @opindex mcond-exec
10173 Enable the use of conditional execution (default).
10175 This switch is mainly for debugging the compiler and will likely be removed
10176 in a future version.
10178 @item -mno-cond-exec
10179 @opindex mno-cond-exec
10181 Disable the use of conditional execution.
10183 This switch is mainly for debugging the compiler and will likely be removed
10184 in a future version.
10186 @item -mvliw-branch
10187 @opindex mvliw-branch
10189 Run a pass to pack branches into VLIW instructions (default).
10191 This switch is mainly for debugging the compiler and will likely be removed
10192 in a future version.
10194 @item -mno-vliw-branch
10195 @opindex mno-vliw-branch
10197 Do not run a pass to pack branches into VLIW instructions.
10199 This switch is mainly for debugging the compiler and will likely be removed
10200 in a future version.
10202 @item -mmulti-cond-exec
10203 @opindex mmulti-cond-exec
10205 Enable optimization of @code{&&} and @code{||} in conditional execution
10208 This switch is mainly for debugging the compiler and will likely be removed
10209 in a future version.
10211 @item -mno-multi-cond-exec
10212 @opindex mno-multi-cond-exec
10214 Disable optimization of @code{&&} and @code{||} in conditional execution.
10216 This switch is mainly for debugging the compiler and will likely be removed
10217 in a future version.
10219 @item -mnested-cond-exec
10220 @opindex mnested-cond-exec
10222 Enable nested conditional execution optimizations (default).
10224 This switch is mainly for debugging the compiler and will likely be removed
10225 in a future version.
10227 @item -mno-nested-cond-exec
10228 @opindex mno-nested-cond-exec
10230 Disable nested conditional execution optimizations.
10232 This switch is mainly for debugging the compiler and will likely be removed
10233 in a future version.
10235 @item -mtomcat-stats
10236 @opindex mtomcat-stats
10238 Cause gas to print out tomcat statistics.
10240 @item -mcpu=@var{cpu}
10243 Select the processor type for which to generate code. Possible values are
10244 @samp{simple}, @samp{tomcat}, @samp{fr500}, @samp{fr400}, @samp{fr300},
10249 @node Xtensa Options
10250 @subsection Xtensa Options
10251 @cindex Xtensa Options
10253 The Xtensa architecture is designed to support many different
10254 configurations. The compiler's default options can be set to match a
10255 particular Xtensa configuration by copying a configuration file into the
10256 GCC sources when building GCC@. The options below may be used to
10257 override the default options.
10261 @itemx -mlittle-endian
10262 @opindex mbig-endian
10263 @opindex mlittle-endian
10264 Specify big-endian or little-endian byte ordering for the target Xtensa
10268 @itemx -mno-density
10270 @opindex mno-density
10271 Enable or disable use of the optional Xtensa code density instructions.
10277 Enable or disable use of the Xtensa MAC16 option. When enabled, GCC
10278 will generate MAC16 instructions from standard C code, with the
10279 limitation that it will use neither the MR register file nor any
10280 instruction that operates on the MR registers. When this option is
10281 disabled, GCC will translate 16-bit multiply/accumulate operations to a
10282 combination of core instructions and library calls, depending on whether
10283 any other multiplier options are enabled.
10289 Enable or disable use of the 16-bit integer multiplier option. When
10290 enabled, the compiler will generate 16-bit multiply instructions for
10291 multiplications of 16 bits or smaller in standard C code. When this
10292 option is disabled, the compiler will either use 32-bit multiply or
10293 MAC16 instructions if they are available or generate library calls to
10294 perform the multiply operations using shifts and adds.
10300 Enable or disable use of the 32-bit integer multiplier option. When
10301 enabled, the compiler will generate 32-bit multiply instructions for
10302 multiplications of 32 bits or smaller in standard C code. When this
10303 option is disabled, the compiler will generate library calls to perform
10304 the multiply operations using either shifts and adds or 16-bit multiply
10305 instructions if they are available.
10311 Enable or disable use of the optional normalization shift amount
10312 (@code{NSA}) instructions to implement the built-in @code{ffs} function.
10317 @opindex mno-minmax
10318 Enable or disable use of the optional minimum and maximum value
10325 Enable or disable use of the optional sign extend (@code{SEXT})
10329 @itemx -mno-booleans
10331 @opindex mno-booleans
10332 Enable or disable support for the boolean register file used by Xtensa
10333 coprocessors. This is not typically useful by itself but may be
10334 required for other options that make use of the boolean registers (e.g.,
10335 the floating-point option).
10338 @itemx -msoft-float
10339 @opindex mhard-float
10340 @opindex msoft-float
10341 Enable or disable use of the floating-point option. When enabled, GCC
10342 generates floating-point instructions for 32-bit @code{float}
10343 operations. When this option is disabled, GCC generates library calls
10344 to emulate 32-bit floating-point operations using integer instructions.
10345 Regardless of this option, 64-bit @code{double} operations are always
10346 emulated with calls to library functions.
10349 @itemx -mno-fused-madd
10350 @opindex mfused-madd
10351 @opindex mno-fused-madd
10352 Enable or disable use of fused multiply/add and multiply/subtract
10353 instructions in the floating-point option. This has no effect if the
10354 floating-point option is not also enabled. Disabling fused multiply/add
10355 and multiply/subtract instructions forces the compiler to use separate
10356 instructions for the multiply and add/subtract operations. This may be
10357 desirable in some cases where strict IEEE 754-compliant results are
10358 required: the fused multiply add/subtract instructions do not round the
10359 intermediate result, thereby producing results with @emph{more} bits of
10360 precision than specified by the IEEE standard. Disabling fused multiply
10361 add/subtract instructions also ensures that the program output is not
10362 sensitive to the compiler's ability to combine multiply and add/subtract
10365 @item -mserialize-volatile
10366 @itemx -mno-serialize-volatile
10367 @opindex mserialize-volatile
10368 @opindex mno-serialize-volatile
10369 When this option is enabled, GCC inserts @code{MEMW} instructions before
10370 @code{volatile} memory references to guarantee sequential consistency.
10371 The default is @option{-mserialize-volatile}. Use
10372 @option{-mno-serialize-volatile} to omit the @code{MEMW} instructions.
10374 @item -mtext-section-literals
10375 @itemx -mno-text-section-literals
10376 @opindex mtext-section-literals
10377 @opindex mno-text-section-literals
10378 Control the treatment of literal pools. The default is
10379 @option{-mno-text-section-literals}, which places literals in a separate
10380 section in the output file. This allows the literal pool to be placed
10381 in a data RAM/ROM, and it also allows the linker to combine literal
10382 pools from separate object files to remove redundant literals and
10383 improve code size. With @option{-mtext-section-literals}, the literals
10384 are interspersed in the text section in order to keep them as close as
10385 possible to their references. This may be necessary for large assembly
10388 @item -mtarget-align
10389 @itemx -mno-target-align
10390 @opindex mtarget-align
10391 @opindex mno-target-align
10392 When this option is enabled, GCC instructs the assembler to
10393 automatically align instructions to reduce branch penalties at the
10394 expense of some code density. The assembler attempts to widen density
10395 instructions to align branch targets and the instructions following call
10396 instructions. If there are not enough preceding safe density
10397 instructions to align a target, no widening will be performed. The
10398 default is @option{-mtarget-align}. These options do not affect the
10399 treatment of auto-aligned instructions like @code{LOOP}, which the
10400 assembler will always align, either by widening density instructions or
10401 by inserting no-op instructions.
10404 @itemx -mno-longcalls
10405 @opindex mlongcalls
10406 @opindex mno-longcalls
10407 When this option is enabled, GCC instructs the assembler to translate
10408 direct calls to indirect calls unless it can determine that the target
10409 of a direct call is in the range allowed by the call instruction. This
10410 translation typically occurs for calls to functions in other source
10411 files. Specifically, the assembler translates a direct @code{CALL}
10412 instruction into an @code{L32R} followed by a @code{CALLX} instruction.
10413 The default is @option{-mno-longcalls}. This option should be used in
10414 programs where the call target can potentially be out of range. This
10415 option is implemented in the assembler, not the compiler, so the
10416 assembly code generated by GCC will still show direct call
10417 instructions---look at the disassembled object code to see the actual
10418 instructions. Note that the assembler will use an indirect call for
10419 every cross-file call, not just those that really will be out of range.
10422 @node Code Gen Options
10423 @section Options for Code Generation Conventions
10424 @cindex code generation conventions
10425 @cindex options, code generation
10426 @cindex run-time options
10428 These machine-independent options control the interface conventions
10429 used in code generation.
10431 Most of them have both positive and negative forms; the negative form
10432 of @option{-ffoo} would be @option{-fno-foo}. In the table below, only
10433 one of the forms is listed---the one which is not the default. You
10434 can figure out the other form by either removing @samp{no-} or adding
10439 @opindex fexceptions
10440 Enable exception handling. Generates extra code needed to propagate
10441 exceptions. For some targets, this implies GCC will generate frame
10442 unwind information for all functions, which can produce significant data
10443 size overhead, although it does not affect execution. If you do not
10444 specify this option, GCC will enable it by default for languages like
10445 C++ which normally require exception handling, and disable it for
10446 languages like C that do not normally require it. However, you may need
10447 to enable this option when compiling C code that needs to interoperate
10448 properly with exception handlers written in C++. You may also wish to
10449 disable this option if you are compiling older C++ programs that don't
10450 use exception handling.
10452 @item -fnon-call-exceptions
10453 @opindex fnon-call-exceptions
10454 Generate code that allows trapping instructions to throw exceptions.
10455 Note that this requires platform-specific runtime support that does
10456 not exist everywhere. Moreover, it only allows @emph{trapping}
10457 instructions to throw exceptions, i.e.@: memory references or floating
10458 point instructions. It does not allow exceptions to be thrown from
10459 arbitrary signal handlers such as @code{SIGALRM}.
10461 @item -funwind-tables
10462 @opindex funwind-tables
10463 Similar to @option{-fexceptions}, except that it will just generate any needed
10464 static data, but will not affect the generated code in any other way.
10465 You will normally not enable this option; instead, a language processor
10466 that needs this handling would enable it on your behalf.
10468 @item -fasynchronous-unwind-tables
10469 @opindex funwind-tables
10470 Generate unwind table in dwarf2 format, if supported by target machine. The
10471 table is exact at each instruction boundary, so it can be used for stack
10472 unwinding from asynchronous events (such as debugger or garbage collector).
10474 @item -fpcc-struct-return
10475 @opindex fpcc-struct-return
10476 Return ``short'' @code{struct} and @code{union} values in memory like
10477 longer ones, rather than in registers. This convention is less
10478 efficient, but it has the advantage of allowing intercallability between
10479 GCC-compiled files and files compiled with other compilers, particularly
10480 the Portable C Compiler (pcc).
10482 The precise convention for returning structures in memory depends
10483 on the target configuration macros.
10485 Short structures and unions are those whose size and alignment match
10486 that of some integer type.
10488 @strong{Warning:} code compiled with the @option{-fpcc-struct-return}
10489 switch is not binary compatible with code compiled with the
10490 @option{-freg-struct-return} switch.
10491 Use it to conform to a non-default application binary interface.
10493 @item -freg-struct-return
10494 @opindex freg-struct-return
10495 Return @code{struct} and @code{union} values in registers when possible.
10496 This is more efficient for small structures than
10497 @option{-fpcc-struct-return}.
10499 If you specify neither @option{-fpcc-struct-return} nor
10500 @option{-freg-struct-return}, GCC defaults to whichever convention is
10501 standard for the target. If there is no standard convention, GCC
10502 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
10503 the principal compiler. In those cases, we can choose the standard, and
10504 we chose the more efficient register return alternative.
10506 @strong{Warning:} code compiled with the @option{-freg-struct-return}
10507 switch is not binary compatible with code compiled with the
10508 @option{-fpcc-struct-return} switch.
10509 Use it to conform to a non-default application binary interface.
10511 @item -fshort-enums
10512 @opindex fshort-enums
10513 Allocate to an @code{enum} type only as many bytes as it needs for the
10514 declared range of possible values. Specifically, the @code{enum} type
10515 will be equivalent to the smallest integer type which has enough room.
10517 @strong{Warning:} the @option{-fshort-enums} switch causes GCC to generate
10518 code that is not binary compatible with code generated without that switch.
10519 Use it to conform to a non-default application binary interface.
10521 @item -fshort-double
10522 @opindex fshort-double
10523 Use the same size for @code{double} as for @code{float}.
10525 @strong{Warning:} the @option{-fshort-double} switch causes GCC to generate
10526 code that is not binary compatible with code generated without that switch.
10527 Use it to conform to a non-default application binary interface.
10529 @item -fshort-wchar
10530 @opindex fshort-wchar
10531 Override the underlying type for @samp{wchar_t} to be @samp{short
10532 unsigned int} instead of the default for the target. This option is
10533 useful for building programs to run under WINE@.
10535 @strong{Warning:} the @option{-fshort-wchar} switch causes GCC to generate
10536 code that is not binary compatible with code generated without that switch.
10537 Use it to conform to a non-default application binary interface.
10539 @item -fshared-data
10540 @opindex fshared-data
10541 Requests that the data and non-@code{const} variables of this
10542 compilation be shared data rather than private data. The distinction
10543 makes sense only on certain operating systems, where shared data is
10544 shared between processes running the same program, while private data
10545 exists in one copy per process.
10548 @opindex fno-common
10549 In C, allocate even uninitialized global variables in the data section of the
10550 object file, rather than generating them as common blocks. This has the
10551 effect that if the same variable is declared (without @code{extern}) in
10552 two different compilations, you will get an error when you link them.
10553 The only reason this might be useful is if you wish to verify that the
10554 program will work on other systems which always work this way.
10558 Ignore the @samp{#ident} directive.
10560 @item -fno-gnu-linker
10561 @opindex fno-gnu-linker
10562 Do not output global initializations (such as C++ constructors and
10563 destructors) in the form used by the GNU linker (on systems where the GNU
10564 linker is the standard method of handling them). Use this option when
10565 you want to use a non-GNU linker, which also requires using the
10566 @command{collect2} program to make sure the system linker includes
10567 constructors and destructors. (@command{collect2} is included in the GCC
10568 distribution.) For systems which @emph{must} use @command{collect2}, the
10569 compiler driver @command{gcc} is configured to do this automatically.
10571 @item -finhibit-size-directive
10572 @opindex finhibit-size-directive
10573 Don't output a @code{.size} assembler directive, or anything else that
10574 would cause trouble if the function is split in the middle, and the
10575 two halves are placed at locations far apart in memory. This option is
10576 used when compiling @file{crtstuff.c}; you should not need to use it
10579 @item -fverbose-asm
10580 @opindex fverbose-asm
10581 Put extra commentary information in the generated assembly code to
10582 make it more readable. This option is generally only of use to those
10583 who actually need to read the generated assembly code (perhaps while
10584 debugging the compiler itself).
10586 @option{-fno-verbose-asm}, the default, causes the
10587 extra information to be omitted and is useful when comparing two assembler
10592 Consider all memory references through pointers to be volatile.
10594 @item -fvolatile-global
10595 @opindex fvolatile-global
10596 Consider all memory references to extern and global data items to
10597 be volatile. GCC does not consider static data items to be volatile
10598 because of this switch.
10600 @item -fvolatile-static
10601 @opindex fvolatile-static
10602 Consider all memory references to static data to be volatile.
10606 @cindex global offset table
10608 Generate position-independent code (PIC) suitable for use in a shared
10609 library, if supported for the target machine. Such code accesses all
10610 constant addresses through a global offset table (GOT)@. The dynamic
10611 loader resolves the GOT entries when the program starts (the dynamic
10612 loader is not part of GCC; it is part of the operating system). If
10613 the GOT size for the linked executable exceeds a machine-specific
10614 maximum size, you get an error message from the linker indicating that
10615 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
10616 instead. (These maximums are 16k on the m88k, 8k on the SPARC, and 32k
10617 on the m68k and RS/6000. The 386 has no such limit.)
10619 Position-independent code requires special support, and therefore works
10620 only on certain machines. For the 386, GCC supports PIC for System V
10621 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
10622 position-independent.
10626 If supported for the target machine, emit position-independent code,
10627 suitable for dynamic linking and avoiding any limit on the size of the
10628 global offset table. This option makes a difference on the m68k, m88k,
10631 Position-independent code requires special support, and therefore works
10632 only on certain machines.
10634 @item -ffixed-@var{reg}
10636 Treat the register named @var{reg} as a fixed register; generated code
10637 should never refer to it (except perhaps as a stack pointer, frame
10638 pointer or in some other fixed role).
10640 @var{reg} must be the name of a register. The register names accepted
10641 are machine-specific and are defined in the @code{REGISTER_NAMES}
10642 macro in the machine description macro file.
10644 This flag does not have a negative form, because it specifies a
10647 @item -fcall-used-@var{reg}
10648 @opindex fcall-used
10649 Treat the register named @var{reg} as an allocable register that is
10650 clobbered by function calls. It may be allocated for temporaries or
10651 variables that do not live across a call. Functions compiled this way
10652 will not save and restore the register @var{reg}.
10654 It is an error to used this flag with the frame pointer or stack pointer.
10655 Use of this flag for other registers that have fixed pervasive roles in
10656 the machine's execution model will produce disastrous results.
10658 This flag does not have a negative form, because it specifies a
10661 @item -fcall-saved-@var{reg}
10662 @opindex fcall-saved
10663 Treat the register named @var{reg} as an allocable register saved by
10664 functions. It may be allocated even for temporaries or variables that
10665 live across a call. Functions compiled this way will save and restore
10666 the register @var{reg} if they use it.
10668 It is an error to used this flag with the frame pointer or stack pointer.
10669 Use of this flag for other registers that have fixed pervasive roles in
10670 the machine's execution model will produce disastrous results.
10672 A different sort of disaster will result from the use of this flag for
10673 a register in which function values may be returned.
10675 This flag does not have a negative form, because it specifies a
10678 @item -fpack-struct
10679 @opindex fpack-struct
10680 Pack all structure members together without holes.
10682 @strong{Warning:} the @option{-fpack-struct} switch causes GCC to generate
10683 code that is not binary compatible with code generated without that switch.
10684 Additionally, it makes the code suboptimal.
10685 Use it to conform to a non-default application binary interface.
10687 @item -finstrument-functions
10688 @opindex finstrument-functions
10689 Generate instrumentation calls for entry and exit to functions. Just
10690 after function entry and just before function exit, the following
10691 profiling functions will be called with the address of the current
10692 function and its call site. (On some platforms,
10693 @code{__builtin_return_address} does not work beyond the current
10694 function, so the call site information may not be available to the
10695 profiling functions otherwise.)
10698 void __cyg_profile_func_enter (void *this_fn,
10700 void __cyg_profile_func_exit (void *this_fn,
10704 The first argument is the address of the start of the current function,
10705 which may be looked up exactly in the symbol table.
10707 This instrumentation is also done for functions expanded inline in other
10708 functions. The profiling calls will indicate where, conceptually, the
10709 inline function is entered and exited. This means that addressable
10710 versions of such functions must be available. If all your uses of a
10711 function are expanded inline, this may mean an additional expansion of
10712 code size. If you use @samp{extern inline} in your C code, an
10713 addressable version of such functions must be provided. (This is
10714 normally the case anyways, but if you get lucky and the optimizer always
10715 expands the functions inline, you might have gotten away without
10716 providing static copies.)
10718 A function may be given the attribute @code{no_instrument_function}, in
10719 which case this instrumentation will not be done. This can be used, for
10720 example, for the profiling functions listed above, high-priority
10721 interrupt routines, and any functions from which the profiling functions
10722 cannot safely be called (perhaps signal handlers, if the profiling
10723 routines generate output or allocate memory).
10725 @item -fstack-check
10726 @opindex fstack-check
10727 Generate code to verify that you do not go beyond the boundary of the
10728 stack. You should specify this flag if you are running in an
10729 environment with multiple threads, but only rarely need to specify it in
10730 a single-threaded environment since stack overflow is automatically
10731 detected on nearly all systems if there is only one stack.
10733 Note that this switch does not actually cause checking to be done; the
10734 operating system must do that. The switch causes generation of code
10735 to ensure that the operating system sees the stack being extended.
10737 @item -fstack-limit-register=@var{reg}
10738 @itemx -fstack-limit-symbol=@var{sym}
10739 @itemx -fno-stack-limit
10740 @opindex fstack-limit-register
10741 @opindex fstack-limit-symbol
10742 @opindex fno-stack-limit
10743 Generate code to ensure that the stack does not grow beyond a certain value,
10744 either the value of a register or the address of a symbol. If the stack
10745 would grow beyond the value, a signal is raised. For most targets,
10746 the signal is raised before the stack overruns the boundary, so
10747 it is possible to catch the signal without taking special precautions.
10749 For instance, if the stack starts at absolute address @samp{0x80000000}
10750 and grows downwards, you can use the flags
10751 @option{-fstack-limit-symbol=__stack_limit} and
10752 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
10753 of 128KB@. Note that this may only work with the GNU linker.
10755 @cindex aliasing of parameters
10756 @cindex parameters, aliased
10757 @item -fargument-alias
10758 @itemx -fargument-noalias
10759 @itemx -fargument-noalias-global
10760 @opindex fargument-alias
10761 @opindex fargument-noalias
10762 @opindex fargument-noalias-global
10763 Specify the possible relationships among parameters and between
10764 parameters and global data.
10766 @option{-fargument-alias} specifies that arguments (parameters) may
10767 alias each other and may alias global storage.@*
10768 @option{-fargument-noalias} specifies that arguments do not alias
10769 each other, but may alias global storage.@*
10770 @option{-fargument-noalias-global} specifies that arguments do not
10771 alias each other and do not alias global storage.
10773 Each language will automatically use whatever option is required by
10774 the language standard. You should not need to use these options yourself.
10776 @item -fleading-underscore
10777 @opindex fleading-underscore
10778 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
10779 change the way C symbols are represented in the object file. One use
10780 is to help link with legacy assembly code.
10782 @strong{Warning:} the @option{-fleading-underscore} switch causes GCC to
10783 generate code that is not binary compatible with code generated without that
10784 switch. Use it to conform to a non-default application binary interface.
10785 Not all targets provide complete support for this switch.
10787 @item -ftls-model=@var{model}
10788 Alter the thread-local storage model to be used (@pxref{Thread-Local}).
10789 The @var{model} argument should be one of @code{global-dynamic},
10790 @code{local-dynamic}, @code{initial-exec} or @code{local-exec}.
10792 The default without @option{-fpic} is @code{initial-exec}; with
10793 @option{-fpic} the default is @code{global-dynamic}.
10798 @node Environment Variables
10799 @section Environment Variables Affecting GCC
10800 @cindex environment variables
10802 @c man begin ENVIRONMENT
10804 This section describes several environment variables that affect how GCC
10805 operates. Some of them work by specifying directories or prefixes to use
10806 when searching for various kinds of files. Some are used to specify other
10807 aspects of the compilation environment.
10809 Note that you can also specify places to search using options such as
10810 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
10811 take precedence over places specified using environment variables, which
10812 in turn take precedence over those specified by the configuration of GCC@.
10813 @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint,
10814 GNU Compiler Collection (GCC) Internals}.
10819 @c @itemx LC_COLLATE
10821 @c @itemx LC_MONETARY
10822 @c @itemx LC_NUMERIC
10827 @c @findex LC_COLLATE
10828 @findex LC_MESSAGES
10829 @c @findex LC_MONETARY
10830 @c @findex LC_NUMERIC
10834 These environment variables control the way that GCC uses
10835 localization information that allow GCC to work with different
10836 national conventions. GCC inspects the locale categories
10837 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
10838 so. These locale categories can be set to any value supported by your
10839 installation. A typical value is @samp{en_UK} for English in the United
10842 The @env{LC_CTYPE} environment variable specifies character
10843 classification. GCC uses it to determine the character boundaries in
10844 a string; this is needed for some multibyte encodings that contain quote
10845 and escape characters that would otherwise be interpreted as a string
10848 The @env{LC_MESSAGES} environment variable specifies the language to
10849 use in diagnostic messages.
10851 If the @env{LC_ALL} environment variable is set, it overrides the value
10852 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
10853 and @env{LC_MESSAGES} default to the value of the @env{LANG}
10854 environment variable. If none of these variables are set, GCC
10855 defaults to traditional C English behavior.
10859 If @env{TMPDIR} is set, it specifies the directory to use for temporary
10860 files. GCC uses temporary files to hold the output of one stage of
10861 compilation which is to be used as input to the next stage: for example,
10862 the output of the preprocessor, which is the input to the compiler
10865 @item GCC_EXEC_PREFIX
10866 @findex GCC_EXEC_PREFIX
10867 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
10868 names of the subprograms executed by the compiler. No slash is added
10869 when this prefix is combined with the name of a subprogram, but you can
10870 specify a prefix that ends with a slash if you wish.
10872 If @env{GCC_EXEC_PREFIX} is not set, GCC will attempt to figure out
10873 an appropriate prefix to use based on the pathname it was invoked with.
10875 If GCC cannot find the subprogram using the specified prefix, it
10876 tries looking in the usual places for the subprogram.
10878 The default value of @env{GCC_EXEC_PREFIX} is
10879 @file{@var{prefix}/lib/gcc-lib/} where @var{prefix} is the value
10880 of @code{prefix} when you ran the @file{configure} script.
10882 Other prefixes specified with @option{-B} take precedence over this prefix.
10884 This prefix is also used for finding files such as @file{crt0.o} that are
10887 In addition, the prefix is used in an unusual way in finding the
10888 directories to search for header files. For each of the standard
10889 directories whose name normally begins with @samp{/usr/local/lib/gcc-lib}
10890 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
10891 replacing that beginning with the specified prefix to produce an
10892 alternate directory name. Thus, with @option{-Bfoo/}, GCC will search
10893 @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}.
10894 These alternate directories are searched first; the standard directories
10897 @item COMPILER_PATH
10898 @findex COMPILER_PATH
10899 The value of @env{COMPILER_PATH} is a colon-separated list of
10900 directories, much like @env{PATH}. GCC tries the directories thus
10901 specified when searching for subprograms, if it can't find the
10902 subprograms using @env{GCC_EXEC_PREFIX}.
10905 @findex LIBRARY_PATH
10906 The value of @env{LIBRARY_PATH} is a colon-separated list of
10907 directories, much like @env{PATH}. When configured as a native compiler,
10908 GCC tries the directories thus specified when searching for special
10909 linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking
10910 using GCC also uses these directories when searching for ordinary
10911 libraries for the @option{-l} option (but directories specified with
10912 @option{-L} come first).
10916 @cindex locale definition
10917 This variable is used to pass locale information to the compiler. One way in
10918 which this information is used is to determine the character set to be used
10919 when character literals, string literals and comments are parsed in C and C++.
10920 When the compiler is configured to allow multibyte characters,
10921 the following values for @env{LANG} are recognized:
10925 Recognize JIS characters.
10927 Recognize SJIS characters.
10929 Recognize EUCJP characters.
10932 If @env{LANG} is not defined, or if it has some other value, then the
10933 compiler will use mblen and mbtowc as defined by the default locale to
10934 recognize and translate multibyte characters.
10938 Some additional environments variables affect the behavior of the
10941 @include cppenv.texi
10945 @node Precompiled Headers
10946 @section Using Precompiled Headers
10947 @cindex precompiled headers
10948 @cindex speed of compilation
10950 Often large projects have many header files that are included in every
10951 source file. The time the compiler takes to process these header files
10952 over and over again can account for nearly all of the time required to
10953 build the project. To make builds faster, GCC allows users to
10954 `precompile' a header file; then, if builds can use the precompiled
10955 header file they will be much faster.
10957 To create a precompiled header file, simply compile it as you would any
10958 other file, if necessary using the @option{-x} option to make the driver
10959 treat it as a C or C++ header file. You will probably want to use a
10960 tool like @command{make} to keep the precompiled header up-to-date when
10961 the headers it contains change.
10963 A precompiled header file will be searched for when @code{#include} is
10964 seen in the compilation. As it searches for the included file
10965 (@pxref{Search Path,,Search Path,cpp.info,The C Preprocessor}) the
10966 compiler looks for a precompiled header in each directory just before it
10967 looks for the include file in that directory. The name searched for is
10968 the name specified in the @code{#include} with @samp{.pch} appended. If
10969 the precompiled header file can't be used, it is ignored.
10971 For instance, if you have @code{#include "all.h"}, and you have
10972 @file{all.h.pch} in the same directory as @file{all.h}, then the
10973 precompiled header file will be used if possible, and the original
10974 header will be used otherwise.
10976 Alternatively, you might decide to put the precompiled header file in a
10977 directory and use @option{-I} to ensure that directory is searched
10978 before (or instead of) the directory containing the original header.
10979 Then, if you want to check that the precompiled header file is always
10980 used, you can put a file of the same name as the original header in this
10981 directory containing an @code{#error} command.
10983 This also works with @option{-include}. So yet another way to use
10984 precompiled headers, good for projects not designed with precompiled
10985 header files in mind, is to simply take most of the header files used by
10986 a project, include them from another header file, precompile that header
10987 file, and @option{-include} the precompiled header. If the header files
10988 have guards against multiple inclusion, they will be skipped because
10989 they've already been included (in the precompiled header).
10991 If you need to precompile the same header file for different
10992 languages, targets, or compiler options, you can instead make a
10993 @emph{directory} named like @file{all.h.pch}, and put each precompiled
10994 header in the directory. (It doesn't matter what you call the files
10995 in the directory, every precompiled header in the directory will be
10996 considered.) The first precompiled header encountered in the
10997 directory that is valid for this compilation will be used; they're
10998 searched in no particular order.
11000 There are many other possibilities, limited only by your imagination,
11001 good sense, and the constraints of your build system.
11003 A precompiled header file can be used only when these conditions apply:
11007 Only one precompiled header can be used in a particular compilation.
11009 A precompiled header can't be used once the first C token is seen. You
11010 can have preprocessor directives before a precompiled header; you can
11011 even include a precompiled header from inside another header, so long as
11012 there are no C tokens before the @code{#include}.
11014 The precompiled header file must be produced for the same language as
11015 the current compilation. You can't use a C precompiled header for a C++
11018 The precompiled header file must be produced by the same compiler
11019 version and configuration as the current compilation is using.
11020 The easiest way to guarantee this is to use the same compiler binary
11021 for creating and using precompiled headers.
11023 Any macros defined before the precompiled header (including with
11024 @option{-D}) must either be defined in the same way as when the
11025 precompiled header was generated, or must not affect the precompiled
11026 header, which usually means that the they don't appear in the
11027 precompiled header at all.
11029 Certain command-line options must be defined in the same way as when the
11030 precompiled header was generated. At present, it's not clear which
11031 options are safe to change and which are not; the safest choice is to
11032 use exactly the same options when generating and using the precompiled
11036 For all of these but the last, the compiler will automatically ignore
11037 the precompiled header if the conditions aren't met. For the last item,
11038 some option changes will cause the precompiled header to be rejected,
11039 but not all incompatible option combinations have yet been found. If
11040 you find a new incompatible combination, please consider filing a bug
11041 report, see @ref{Bugs}.
11043 @node Running Protoize
11044 @section Running Protoize
11046 The program @code{protoize} is an optional part of GCC@. You can use
11047 it to add prototypes to a program, thus converting the program to ISO
11048 C in one respect. The companion program @code{unprotoize} does the
11049 reverse: it removes argument types from any prototypes that are found.
11051 When you run these programs, you must specify a set of source files as
11052 command line arguments. The conversion programs start out by compiling
11053 these files to see what functions they define. The information gathered
11054 about a file @var{foo} is saved in a file named @file{@var{foo}.X}.
11056 After scanning comes actual conversion. The specified files are all
11057 eligible to be converted; any files they include (whether sources or
11058 just headers) are eligible as well.
11060 But not all the eligible files are converted. By default,
11061 @code{protoize} and @code{unprotoize} convert only source and header
11062 files in the current directory. You can specify additional directories
11063 whose files should be converted with the @option{-d @var{directory}}
11064 option. You can also specify particular files to exclude with the
11065 @option{-x @var{file}} option. A file is converted if it is eligible, its
11066 directory name matches one of the specified directory names, and its
11067 name within the directory has not been excluded.
11069 Basic conversion with @code{protoize} consists of rewriting most
11070 function definitions and function declarations to specify the types of
11071 the arguments. The only ones not rewritten are those for varargs
11074 @code{protoize} optionally inserts prototype declarations at the
11075 beginning of the source file, to make them available for any calls that
11076 precede the function's definition. Or it can insert prototype
11077 declarations with block scope in the blocks where undeclared functions
11080 Basic conversion with @code{unprotoize} consists of rewriting most
11081 function declarations to remove any argument types, and rewriting
11082 function definitions to the old-style pre-ISO form.
11084 Both conversion programs print a warning for any function declaration or
11085 definition that they can't convert. You can suppress these warnings
11088 The output from @code{protoize} or @code{unprotoize} replaces the
11089 original source file. The original file is renamed to a name ending
11090 with @samp{.save} (for DOS, the saved filename ends in @samp{.sav}
11091 without the original @samp{.c} suffix). If the @samp{.save} (@samp{.sav}
11092 for DOS) file already exists, then the source file is simply discarded.
11094 @code{protoize} and @code{unprotoize} both depend on GCC itself to
11095 scan the program and collect information about the functions it uses.
11096 So neither of these programs will work until GCC is installed.
11098 Here is a table of the options you can use with @code{protoize} and
11099 @code{unprotoize}. Each option works with both programs unless
11103 @item -B @var{directory}
11104 Look for the file @file{SYSCALLS.c.X} in @var{directory}, instead of the
11105 usual directory (normally @file{/usr/local/lib}). This file contains
11106 prototype information about standard system functions. This option
11107 applies only to @code{protoize}.
11109 @item -c @var{compilation-options}
11110 Use @var{compilation-options} as the options when running @code{gcc} to
11111 produce the @samp{.X} files. The special option @option{-aux-info} is
11112 always passed in addition, to tell @code{gcc} to write a @samp{.X} file.
11114 Note that the compilation options must be given as a single argument to
11115 @code{protoize} or @code{unprotoize}. If you want to specify several
11116 @code{gcc} options, you must quote the entire set of compilation options
11117 to make them a single word in the shell.
11119 There are certain @code{gcc} arguments that you cannot use, because they
11120 would produce the wrong kind of output. These include @option{-g},
11121 @option{-O}, @option{-c}, @option{-S}, and @option{-o} If you include these in
11122 the @var{compilation-options}, they are ignored.
11125 Rename files to end in @samp{.C} (@samp{.cc} for DOS-based file
11126 systems) instead of @samp{.c}. This is convenient if you are converting
11127 a C program to C++. This option applies only to @code{protoize}.
11130 Add explicit global declarations. This means inserting explicit
11131 declarations at the beginning of each source file for each function
11132 that is called in the file and was not declared. These declarations
11133 precede the first function definition that contains a call to an
11134 undeclared function. This option applies only to @code{protoize}.
11136 @item -i @var{string}
11137 Indent old-style parameter declarations with the string @var{string}.
11138 This option applies only to @code{protoize}.
11140 @code{unprotoize} converts prototyped function definitions to old-style
11141 function definitions, where the arguments are declared between the
11142 argument list and the initial @samp{@{}. By default, @code{unprotoize}
11143 uses five spaces as the indentation. If you want to indent with just
11144 one space instead, use @option{-i " "}.
11147 Keep the @samp{.X} files. Normally, they are deleted after conversion
11151 Add explicit local declarations. @code{protoize} with @option{-l} inserts
11152 a prototype declaration for each function in each block which calls the
11153 function without any declaration. This option applies only to
11157 Make no real changes. This mode just prints information about the conversions
11158 that would have been done without @option{-n}.
11161 Make no @samp{.save} files. The original files are simply deleted.
11162 Use this option with caution.
11164 @item -p @var{program}
11165 Use the program @var{program} as the compiler. Normally, the name
11166 @file{gcc} is used.
11169 Work quietly. Most warnings are suppressed.
11172 Print the version number, just like @option{-v} for @code{gcc}.
11175 If you need special compiler options to compile one of your program's
11176 source files, then you should generate that file's @samp{.X} file
11177 specially, by running @code{gcc} on that source file with the
11178 appropriate options and the option @option{-aux-info}. Then run
11179 @code{protoize} on the entire set of files. @code{protoize} will use
11180 the existing @samp{.X} file because it is newer than the source file.
11184 gcc -Dfoo=bar file1.c -aux-info file1.X
11189 You need to include the special files along with the rest in the
11190 @code{protoize} command, even though their @samp{.X} files already
11191 exist, because otherwise they won't get converted.
11193 @xref{Protoize Caveats}, for more information on how to use
11194 @code{protoize} successfully.