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 1.2 or
13 any later version published by the Free Software Foundation; with the
14 Invariant Sections being ``GNU General Public License'' and ``Funding
15 Free Software'', the Front-Cover texts being (a) (see below), and with
16 the Back-Cover Texts being (b) (see below). A copy of the license is
17 included in the gfdl(7) man page.
19 (a) The FSF's Front-Cover Text is:
23 (b) The FSF's Back-Cover Text is:
25 You have freedom to copy and modify this GNU Manual, like GNU
26 software. Copies published by the Free Software Foundation raise
27 funds for GNU development.
29 @c Set file name and title for the man page.
31 @settitle GNU project C and C++ compiler
33 gcc [@option{-c}|@option{-S}|@option{-E}] [@option{-std=}@var{standard}]
34 [@option{-g}] [@option{-pg}] [@option{-O}@var{level}]
35 [@option{-W}@var{warn}@dots{}] [@option{-pedantic}]
36 [@option{-I}@var{dir}@dots{}] [@option{-L}@var{dir}@dots{}]
37 [@option{-D}@var{macro}[=@var{defn}]@dots{}] [@option{-U}@var{macro}]
38 [@option{-f}@var{option}@dots{}] [@option{-m}@var{machine-option}@dots{}]
39 [@option{-o} @var{outfile}] @var{infile}@dots{}
41 Only the most useful options are listed here; see below for the
42 remainder. @samp{g++} accepts mostly the same options as @samp{gcc}.
45 gpl(7), gfdl(7), fsf-funding(7),
46 cpp(1), gcov(1), 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
69 When you invoke GCC, it normally does preprocessing, compilation,
70 assembly and linking. The ``overall options'' allow you to stop this
71 process at an intermediate stage. For example, the @option{-c} option
72 says not to run the linker. Then the output consists of object files
73 output by the assembler.
75 Other options are passed on to one stage of processing. Some options
76 control the preprocessor and others the compiler itself. Yet other
77 options control the assembler and linker; most of these are not
78 documented here, since you rarely need to use any of them.
80 @cindex C compilation options
81 Most of the command line options that you can use with GCC are useful
82 for C programs; when an option is only useful with another language
83 (usually C++), the explanation says so explicitly. If the description
84 for a particular option does not mention a source language, you can use
85 that option with all supported languages.
87 @cindex C++ compilation options
88 @xref{Invoking G++,,Compiling C++ Programs}, for a summary of special
89 options for compiling C++ programs.
91 @cindex grouping options
92 @cindex options, grouping
93 The @command{gcc} program accepts options and file names as operands. Many
94 options have multi-letter names; therefore multiple single-letter options
95 may @emph{not} be grouped: @option{-dr} is very different from @w{@samp{-d
98 @cindex order of options
99 @cindex options, order
100 You can mix options and other arguments. For the most part, the order
101 you use doesn't matter. Order does matter when you use several options
102 of the same kind; for example, if you specify @option{-L} more than once,
103 the directories are searched in the order specified.
105 Many options have long names starting with @samp{-f} or with
106 @samp{-W}---for example, @option{-fforce-mem},
107 @option{-fstrength-reduce}, @option{-Wformat} and so on. Most of
108 these have both positive and negative forms; the negative form of
109 @option{-ffoo} would be @option{-fno-foo}. This manual documents
110 only one of these two forms, whichever one is not the default.
114 @xref{Option Index}, for an index to GCC's options.
117 * Option Summary:: Brief list of all options, without explanations.
118 * Overall Options:: Controlling the kind of output:
119 an executable, object files, assembler files,
120 or preprocessed source.
121 * Invoking G++:: Compiling C++ programs.
122 * C Dialect Options:: Controlling the variant of C language compiled.
123 * C++ Dialect Options:: Variations on C++.
124 * Objective-C Dialect Options:: Variations on Objective-C.
125 * Language Independent Options:: Controlling how diagnostics should be
127 * Warning Options:: How picky should the compiler be?
128 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
129 * Optimize Options:: How much optimization?
130 * Preprocessor Options:: Controlling header files and macro definitions.
131 Also, getting dependency information for Make.
132 * Assembler Options:: Passing options to the assembler.
133 * Link Options:: Specifying libraries and so on.
134 * Directory Options:: Where to find header files and libraries.
135 Where to find the compiler executable files.
136 * Spec Files:: How to pass switches to sub-processes.
137 * Target Options:: Running a cross-compiler, or an old version of GCC.
138 * Submodel Options:: Specifying minor hardware or convention variations,
139 such as 68010 vs 68020.
140 * Code Gen Options:: Specifying conventions for function calls, data layout
142 * Environment Variables:: Env vars that affect GCC.
143 * Precompiled Headers:: Compiling a header once, and using it many times.
144 * Running Protoize:: Automatically adding or removing function prototypes.
150 @section Option Summary
152 Here is a summary of all the options, grouped by type. Explanations are
153 in the following sections.
156 @item Overall Options
157 @xref{Overall Options,,Options Controlling the Kind of Output}.
158 @gccoptlist{-c -S -E -o @var{file} -pipe -pass-exit-codes @gol
159 -x @var{language} -v -### --help --target-help --version}
161 @item C Language Options
162 @xref{C Dialect Options,,Options Controlling C Dialect}.
163 @gccoptlist{-ansi -std=@var{standard} -aux-info @var{filename} @gol
164 -fno-asm -fno-builtin -fno-builtin-@var{function} @gol
165 -fhosted -ffreestanding -fms-extensions @gol
166 -trigraphs -no-integrated-cpp -traditional -traditional-cpp @gol
167 -fallow-single-precision -fcond-mismatch @gol
168 -fsigned-bitfields -fsigned-char @gol
169 -funsigned-bitfields -funsigned-char @gol
172 @item C++ Language Options
173 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}.
174 @gccoptlist{-fabi-version=@var{n} -fno-access-control -fcheck-new @gol
175 -fconserve-space -fno-const-strings @gol
176 -fno-elide-constructors @gol
177 -fno-enforce-eh-specs -fexternal-templates @gol
178 -falt-external-templates @gol
179 -ffor-scope -fno-for-scope -fno-gnu-keywords @gol
180 -fno-implicit-templates @gol
181 -fno-implicit-inline-templates @gol
182 -fno-implement-inlines -fms-extensions @gol
183 -fno-nonansi-builtins -fno-operator-names @gol
184 -fno-optional-diags -fpermissive @gol
185 -frepo -fno-rtti -fstats -ftemplate-depth-@var{n} @gol
186 -fuse-cxa-atexit -fno-weak -nostdinc++ @gol
187 -fno-default-inline -Wabi -Wctor-dtor-privacy @gol
188 -Wnon-virtual-dtor -Wreorder @gol
189 -Weffc++ -Wno-deprecated @gol
190 -Wno-non-template-friend -Wold-style-cast @gol
191 -Woverloaded-virtual -Wno-pmf-conversions @gol
192 -Wsign-promo -Wsynth}
194 @item Objective-C Language Options
195 @xref{Objective-C Dialect Options,,Options Controlling Objective-C Dialect}.
196 @gccoptlist{-fconstant-string-class=@var{class-name} @gol
197 -fgnu-runtime -fnext-runtime -gen-decls @gol
198 -Wno-protocol -Wselector -Wundeclared-selector}
200 @item Language Independent Options
201 @xref{Language Independent Options,,Options to Control Diagnostic Messages Formatting}.
202 @gccoptlist{-fmessage-length=@var{n} @gol
203 -fdiagnostics-show-location=@r{[}once@r{|}every-line@r{]}}
205 @item Warning Options
206 @xref{Warning Options,,Options to Request or Suppress Warnings}.
207 @gccoptlist{-fsyntax-only -pedantic -pedantic-errors @gol
208 -w -Wextra -Wall -Waggregate-return @gol
209 -Wcast-align -Wcast-qual -Wchar-subscripts -Wcomment @gol
210 -Wconversion -Wno-deprecated-declarations @gol
211 -Wdisabled-optimization -Wno-div-by-zero -Werror @gol
212 -Wfloat-equal -Wformat -Wformat=2 @gol
213 -Wformat-nonliteral -Wformat-security @gol
214 -Wimplicit -Wimplicit-int @gol
215 -Wimplicit-function-declaration @gol
216 -Werror-implicit-function-declaration @gol
217 -Winit-self Wimport -Winline -Winvalid-pch -Wno-endif-labels @gol
218 -Wno-invalid-offsetof @gol
219 -Wlarger-than-@var{len} -Wlong-long @gol
220 -Wmain -Wmissing-braces @gol
221 -Wmissing-format-attribute -Wmissing-noreturn @gol
222 -Wno-multichar -Wno-format-extra-args -Wno-format-y2k @gol
223 -Wno-import -Wnonnull -Wpacked -Wpadded @gol
224 -Wparentheses -Wpointer-arith -Wredundant-decls @gol
225 -Wreturn-type -Wsequence-point -Wshadow @gol
226 -Wsign-compare -Wstrict-aliasing @gol
227 -Wswitch -Wswitch-default -Wswitch-enum @gol
228 -Wsystem-headers -Wtrigraphs -Wundef -Wuninitialized @gol
229 -Wunknown-pragmas -Wunreachable-code @gol
230 -Wunused -Wunused-function -Wunused-label -Wunused-parameter @gol
231 -Wunused-value -Wunused-variable -Wwrite-strings}
233 @item C-only Warning Options
234 @gccoptlist{-Wbad-function-cast -Wmissing-declarations @gol
235 -Wmissing-prototypes -Wnested-externs @gol
236 -Wstrict-prototypes -Wtraditional @gol
237 -Wdeclaration-after-statement}
239 @item Debugging Options
240 @xref{Debugging Options,,Options for Debugging Your Program or GCC}.
241 @gccoptlist{-d@var{letters} -dumpspecs -dumpmachine -dumpversion @gol
242 -fdump-unnumbered -fdump-translation-unit@r{[}-@var{n}@r{]} @gol
243 -fdump-class-hierarchy@r{[}-@var{n}@r{]} @gol
244 -fdump-tree-original@r{[}-@var{n}@r{]} @gol
245 -fdump-tree-optimized@r{[}-@var{n}@r{]} @gol
246 -fdump-tree-inlined@r{[}-@var{n}@r{]} @gol
247 -feliminate-dwarf2-dups -feliminate-unused-debug-types @gol
248 -feliminate-unused-debug-symbols -fmem-report -fprofile-arcs @gol
249 -frandom-seed=@var{string} -fsched-verbose=@var{n} @gol
250 -ftest-coverage -ftime-report @gol
251 -g -g@var{level} -gcoff -gdwarf -gdwarf-1 -gdwarf-1+ -gdwarf-2 @gol
252 -ggdb -gstabs -gstabs+ -gvms -gxcoff -gxcoff+ @gol
253 -p -pg -print-file-name=@var{library} -print-libgcc-file-name @gol
254 -print-multi-directory -print-multi-lib @gol
255 -print-prog-name=@var{program} -print-search-dirs -Q @gol
258 @item Optimization Options
259 @xref{Optimize Options,,Options that Control Optimization}.
260 @gccoptlist{-falign-functions=@var{n} -falign-jumps=@var{n} @gol
261 -falign-labels=@var{n} -falign-loops=@var{n} @gol
262 -fbranch-probabilities -fprofile-values -fbranch-target-load-optimize @gol
263 -fbranch-target-load-optimize2 -fcaller-saves -fcprop-registers @gol
264 -fcse-follow-jumps -fcse-skip-blocks -fdata-sections @gol
265 -fdelayed-branch -fdelete-null-pointer-checks @gol
266 -fexpensive-optimizations -ffast-math -ffloat-store @gol
267 -fforce-addr -fforce-mem -ffunction-sections @gol
268 -fgcse -fgcse-lm -fgcse-sm -floop-optimize -fcrossjumping @gol
269 -fif-conversion -fif-conversion2 @gol
270 -finline-functions -finline-limit=@var{n} -fkeep-inline-functions @gol
271 -fkeep-static-consts -fmerge-constants -fmerge-all-constants @gol
272 -fmove-all-movables -fnew-ra -fno-branch-count-reg @gol
273 -fno-default-inline -fno-defer-pop @gol
274 -fno-function-cse -fno-guess-branch-probability @gol
275 -fno-inline -fno-math-errno -fno-peephole -fno-peephole2 @gol
276 -funsafe-math-optimizations -ffinite-math-only @gol
277 -fno-trapping-math -fno-zero-initialized-in-bss @gol
278 -fomit-frame-pointer -foptimize-register-move @gol
279 -foptimize-sibling-calls -fprefetch-loop-arrays @gol
280 -freduce-all-givs -fregmove -frename-registers @gol
281 -freorder-blocks -freorder-functions @gol
282 -frerun-cse-after-loop -frerun-loop-opt @gol
283 -fschedule-insns -fschedule-insns2 @gol
284 -fno-sched-interblock -fno-sched-spec -fsched-spec-load @gol
285 -fsched-spec-load-dangerous -fsched2-use-superblocks @gol
286 -fsched2-use-traces -fsignaling-nans @gol
287 -fsingle-precision-constant -fssa -fssa-ccp -fssa-dce @gol
288 -fstrength-reduce -fstrict-aliasing -ftracer -fthread-jumps @gol
289 -funroll-all-loops -funroll-loops -fpeel-loops @gol
290 -funswitch-loops -fold-unroll-loops -fold-unroll-all-loops @gol
291 --param @var{name}=@var{value}
292 -O -O0 -O1 -O2 -O3 -Os}
294 @item Preprocessor Options
295 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
296 @gccoptlist{-A@var{question}=@var{answer} @gol
297 -A-@var{question}@r{[}=@var{answer}@r{]} @gol
298 -C -dD -dI -dM -dN @gol
299 -D@var{macro}@r{[}=@var{defn}@r{]} -E -H @gol
300 -idirafter @var{dir} @gol
301 -include @var{file} -imacros @var{file} @gol
302 -iprefix @var{file} -iwithprefix @var{dir} @gol
303 -iwithprefixbefore @var{dir} -isystem @var{dir} @gol
304 -M -MM -MF -MG -MP -MQ -MT -nostdinc @gol
305 -P -fworking-directory -remap @gol
306 -trigraphs -undef -U@var{macro} -Wp,@var{option} @gol
307 -Xpreprocessor @var{option}}
309 @item Assembler Option
310 @xref{Assembler Options,,Passing Options to the Assembler}.
311 @gccoptlist{-Wa,@var{option} -Xassembler @var{option}}
314 @xref{Link Options,,Options for Linking}.
315 @gccoptlist{@var{object-file-name} -l@var{library} @gol
316 -nostartfiles -nodefaultlibs -nostdlib -pie @gol
317 -s -static -static-libgcc -shared -shared-libgcc -symbolic @gol
318 -Wl,@var{option} -Xlinker @var{option} @gol
321 @item Directory Options
322 @xref{Directory Options,,Options for Directory Search}.
323 @gccoptlist{-B@var{prefix} -I@var{dir} -I- -L@var{dir} -specs=@var{file}}
326 @c I wrote this xref this way to avoid overfull hbox. -- rms
327 @xref{Target Options}.
328 @gccoptlist{-V @var{version} -b @var{machine}}
330 @item Machine Dependent Options
331 @xref{Submodel Options,,Hardware Models and Configurations}.
333 @emph{M680x0 Options}
334 @gccoptlist{-m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
335 -m68060 -mcpu32 -m5200 -m68881 -mbitfield -mc68000 -mc68020 @gol
336 -mnobitfield -mrtd -mshort -msoft-float -mpcrel @gol
337 -malign-int -mstrict-align}
339 @emph{M68hc1x Options}
340 @gccoptlist{-m6811 -m6812 -m68hc11 -m68hc12 -m68hcs12 @gol
341 -mauto-incdec -minmax -mlong-calls -mshort @gol
342 -msoft-reg-count=@var{count}}
345 @gccoptlist{-mg -mgnu -munix}
348 @gccoptlist{-mcpu=@var{cpu-type} @gol
349 -mtune=@var{cpu-type} @gol
350 -mcmodel=@var{code-model} @gol
352 -mapp-regs -mbroken-saverestore -mcypress @gol
353 -mfaster-structs -mflat @gol
354 -mfpu -mhard-float -mhard-quad-float @gol
355 -mimpure-text -mlittle-endian -mlive-g0 -mno-app-regs @gol
356 -mno-faster-structs -mno-flat -mno-fpu @gol
357 -mno-impure-text -mno-stack-bias -mno-unaligned-doubles @gol
358 -msoft-float -msoft-quad-float -msparclite -mstack-bias @gol
359 -msupersparc -munaligned-doubles -mv8}
362 @gccoptlist{-mapcs-frame -mno-apcs-frame @gol
363 -mapcs-26 -mapcs-32 @gol
364 -mapcs-stack-check -mno-apcs-stack-check @gol
365 -mapcs-float -mno-apcs-float @gol
366 -mapcs-reentrant -mno-apcs-reentrant @gol
367 -msched-prolog -mno-sched-prolog @gol
368 -mlittle-endian -mbig-endian -mwords-little-endian @gol
369 -malignment-traps -mno-alignment-traps @gol
370 -msoft-float -mhard-float -mfpe @gol
371 -mthumb-interwork -mno-thumb-interwork @gol
372 -mcpu=@var{name} -march=@var{name} -mfpe=@var{name} @gol
373 -mstructure-size-boundary=@var{n} @gol
374 -mabort-on-noreturn @gol
375 -mlong-calls -mno-long-calls @gol
376 -msingle-pic-base -mno-single-pic-base @gol
377 -mpic-register=@var{reg} @gol
378 -mnop-fun-dllimport @gol
379 -mcirrus-fix-invalid-insns -mno-cirrus-fix-invalid-insns @gol
380 -mpoke-function-name @gol
382 -mtpcs-frame -mtpcs-leaf-frame @gol
383 -mcaller-super-interworking -mcallee-super-interworking}
385 @emph{MN10200 Options}
388 @emph{MN10300 Options}
389 @gccoptlist{-mmult-bug -mno-mult-bug @gol
390 -mam33 -mno-am33 @gol
391 -mam33-2 -mno-am33-2 @gol
394 @emph{M32R/D Options}
395 @gccoptlist{-m32rx -m32r -mcode-model=@var{model-type} @gol
396 -msdata=@var{sdata-type} -G @var{num}}
399 @gccoptlist{-m88000 -m88100 -m88110 -mbig-pic @gol
400 -mcheck-zero-division -mhandle-large-shift @gol
401 -midentify-revision -mno-check-zero-division @gol
402 -mno-ocs-debug-info -mno-ocs-frame-position @gol
403 -mno-optimize-arg-area -mno-serialize-volatile @gol
404 -mno-underscores -mocs-debug-info @gol
405 -mocs-frame-position -moptimize-arg-area @gol
406 -mserialize-volatile -mshort-data-@var{num} -msvr3 @gol
407 -msvr4 -mtrap-large-shift -muse-div-instruction @gol
408 -mversion-03.00 -mwarn-passed-structs}
410 @emph{RS/6000 and PowerPC Options}
411 @gccoptlist{-mcpu=@var{cpu-type} @gol
412 -mtune=@var{cpu-type} @gol
413 -mpower -mno-power -mpower2 -mno-power2 @gol
414 -mpowerpc -mpowerpc64 -mno-powerpc @gol
415 -maltivec -mno-altivec @gol
416 -mpowerpc-gpopt -mno-powerpc-gpopt @gol
417 -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol
418 -mnew-mnemonics -mold-mnemonics @gol
419 -mfull-toc -mminimal-toc -mno-fp-in-toc -mno-sum-in-toc @gol
420 -m64 -m32 -mxl-call -mno-xl-call -mpe @gol
421 -malign-power -malign-natural @gol
422 -msoft-float -mhard-float -mmultiple -mno-multiple @gol
423 -mstring -mno-string -mupdate -mno-update @gol
424 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol
425 -mstrict-align -mno-strict-align -mrelocatable @gol
426 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol
427 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol
428 -mdynamic-no-pic @gol
429 -mcall-sysv -mcall-netbsd @gol
430 -maix-struct-return -msvr4-struct-return @gol
431 -mabi=altivec -mabi=no-altivec @gol
432 -mabi=spe -mabi=no-spe @gol
433 -misel=yes -misel=no @gol
434 -mspe=yes -mspe=no @gol
435 -mfloat-gprs=yes -mfloat-gprs=no @gol
436 -mprototype -mno-prototype @gol
437 -msim -mmvme -mads -myellowknife -memb -msdata @gol
438 -msdata=@var{opt} -mvxworks -mwindiss -G @var{num} -pthread}
440 @emph{Darwin Options}
441 @gccoptlist{-all_load -allowable_client -arch -arch_errors_fatal @gol
442 -arch_only -bind_at_load -bundle -bundle_loader @gol
443 -client_name -compatibility_version -current_version @gol
444 -dependency-file -dylib_file -dylinker_install_name @gol
445 -dynamic -dynamiclib -exported_symbols_list @gol
446 -filelist -flat_namespace -force_cpusubtype_ALL @gol
447 -force_flat_namespace -headerpad_max_install_names @gol
448 -image_base -init -install_name -keep_private_externs @gol
449 -multi_module -multiply_defined -multiply_defined_unused @gol
450 -noall_load -nomultidefs -noprebind -noseglinkedit @gol
451 -pagezero_size -prebind -prebind_all_twolevel_modules @gol
452 -private_bundle -read_only_relocs -sectalign @gol
453 -sectobjectsymbols -whyload -seg1addr @gol
454 -sectcreate -sectobjectsymbols -sectorder @gol
455 -seg_addr_table -seg_addr_table_filename -seglinkedit @gol
456 -segprot -segs_read_only_addr -segs_read_write_addr @gol
457 -single_module -static -sub_library -sub_umbrella @gol
458 -twolevel_namespace -umbrella -undefined @gol
459 -unexported_symbols_list -weak_reference_mismatches @gol
463 @gccoptlist{-mcall-lib-mul -mfp-arg-in-fpregs -mfp-arg-in-gregs @gol
464 -mfull-fp-blocks -mhc-struct-return -min-line-mul @gol
465 -mminimum-fp-blocks -mnohc-struct-return}
468 @gccoptlist{-mabicalls -march=@var{cpu-type} -mtune=@var{cpu=type} @gol
469 -mcpu=@var{cpu-type} -membedded-data -muninit-const-in-rodata @gol
470 -membedded-pic -mfp32 -mfp64 -mfused-madd -mno-fused-madd @gol
471 -mgas -mgp32 -mgp64 -mhard-float -mint64 -mips1 @gol
472 -mips2 -mips3 -mips4 -mips32 -mips32r2 -mips64 @gol
473 -mlong64 -mlong32 -mlong-calls -mmemcpy @gol
474 -mmips-as -mmips-tfile -mno-abicalls -mxgot @gol
475 -mno-embedded-data -mno-uninit-const-in-rodata @gol
476 -mno-embedded-pic -mno-long-calls @gol
477 -mno-memcpy -mno-mips-tfile -mno-rnames @gol
478 -mrnames -msoft-float @gol
479 -m4650 -msingle-float -mmad @gol
480 -EL -EB -G @var{num} -nocpp @gol
481 -mabi=32 -mabi=n32 -mabi=64 -mabi=eabi -mabi-fake-default @gol
482 -mfix7000 -mfix-sb1 -mno-fix-sb1 @gol
483 -mno-crt0 -mflush-func=@var{func} -mno-flush-func @gol
484 -mbranch-likely -mno-branch-likely}
486 @emph{i386 and x86-64 Options}
487 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
488 -mfpmath=@var{unit} @gol
489 -masm=@var{dialect} -mno-fancy-math-387 @gol
490 -mno-fp-ret-in-387 -msoft-float -msvr3-shlib @gol
491 -mno-wide-multiply -mrtd -malign-double @gol
492 -mpreferred-stack-boundary=@var{num} @gol
493 -mmmx -msse -msse2 -mpni -m3dnow @gol
494 -mthreads -mno-align-stringops -minline-all-stringops @gol
495 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
496 -m96bit-long-double -mregparm=@var{num} -momit-leaf-frame-pointer @gol
497 -mno-red-zone -mno-tls-direct-seg-refs @gol
498 -mcmodel=@var{code-model} @gol
502 @gccoptlist{-march=@var{architecture-type} @gol
503 -mbig-switch -mdisable-fpregs -mdisable-indexing @gol
504 -mfast-indirect-calls -mgas -mgnu-ld -mhp-ld @gol
505 -mjump-in-delay -mlinker-opt -mlong-calls @gol
506 -mlong-load-store -mno-big-switch -mno-disable-fpregs @gol
507 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
508 -mno-jump-in-delay -mno-long-load-store @gol
509 -mno-portable-runtime -mno-soft-float @gol
510 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
511 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
512 -mschedule=@var{cpu-type} -mspace-regs -msio -mwsio @gol
513 -nolibdld -static -threads}
515 @emph{Intel 960 Options}
516 @gccoptlist{-m@var{cpu-type} -masm-compat -mclean-linkage @gol
517 -mcode-align -mcomplex-addr -mleaf-procedures @gol
518 -mic-compat -mic2.0-compat -mic3.0-compat @gol
519 -mintel-asm -mno-clean-linkage -mno-code-align @gol
520 -mno-complex-addr -mno-leaf-procedures @gol
521 -mno-old-align -mno-strict-align -mno-tail-call @gol
522 -mnumerics -mold-align -msoft-float -mstrict-align @gol
525 @emph{DEC Alpha Options}
526 @gccoptlist{-mno-fp-regs -msoft-float -malpha-as -mgas @gol
527 -mieee -mieee-with-inexact -mieee-conformant @gol
528 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol
529 -mtrap-precision=@var{mode} -mbuild-constants @gol
530 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} @gol
531 -mbwx -mmax -mfix -mcix @gol
532 -mfloat-vax -mfloat-ieee @gol
533 -mexplicit-relocs -msmall-data -mlarge-data @gol
534 -msmall-text -mlarge-text @gol
535 -mmemory-latency=@var{time}}
537 @emph{DEC Alpha/VMS Options}
538 @gccoptlist{-mvms-return-codes}
540 @emph{H8/300 Options}
541 @gccoptlist{-mrelax -mh -ms -mn -mint32 -malign-300}
544 @gccoptlist{-m1 -m2 -m2e -m3 -m3e @gol
545 -m4-nofpu -m4-single-only -m4-single -m4 @gol
546 -m5-64media -m5-64media-nofpu @gol
547 -m5-32media -m5-32media-nofpu @gol
548 -m5-compact -m5-compact-nofpu @gol
549 -mb -ml -mdalign -mrelax @gol
550 -mbigtable -mfmovd -mhitachi -mnomacsave @gol
551 -mieee -misize -mpadstruct -mspace @gol
552 -mprefergot -musermode}
554 @emph{System V Options}
555 @gccoptlist{-Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
558 @gccoptlist{-EB -EL @gol
559 -mmangle-cpu -mcpu=@var{cpu} -mtext=@var{text-section} @gol
560 -mdata=@var{data-section} -mrodata=@var{readonly-data-section}}
562 @emph{TMS320C3x/C4x Options}
563 @gccoptlist{-mcpu=@var{cpu} -mbig -msmall -mregparm -mmemparm @gol
564 -mfast-fix -mmpyi -mbk -mti -mdp-isr-reload @gol
565 -mrpts=@var{count} -mrptb -mdb -mloop-unsigned @gol
566 -mparallel-insns -mparallel-mpy -mpreserve-float}
569 @gccoptlist{-mlong-calls -mno-long-calls -mep -mno-ep @gol
570 -mprolog-function -mno-prolog-function -mspace @gol
571 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
572 -mapp-regs -mno-app-regs @gol
573 -mdisable-callt -mno-disable-callt @gol
578 @gccoptlist{-m32032 -m32332 -m32532 -m32081 -m32381 @gol
579 -mmult-add -mnomult-add -msoft-float -mrtd -mnortd @gol
580 -mregparam -mnoregparam -msb -mnosb @gol
581 -mbitfield -mnobitfield -mhimem -mnohimem}
584 @gccoptlist{-mmcu=@var{mcu} -msize -minit-stack=@var{n} -mno-interrupts @gol
585 -mcall-prologues -mno-tablejump -mtiny-stack}
588 @gccoptlist{-mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
589 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
590 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
591 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
592 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
595 @gccoptlist{-mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
596 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
597 -melf -mbranch-predict -mno-branch-predict -mbase-addresses @gol
598 -mno-base-addresses -msingle-exit -mno-single-exit}
601 @gccoptlist{-mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
602 -mvolatile-asm-stop -mb-step -mregister-names -mno-sdata @gol
603 -mconstant-gp -mauto-pic -minline-float-divide-min-latency @gol
604 -minline-float-divide-max-throughput @gol
605 -minline-int-divide-min-latency @gol
606 -minline-int-divide-max-throughput -mno-dwarf2-asm @gol
607 -mfixed-range=@var{register-range}}
610 @gccoptlist{-mextmem -mextmemory -monchip -mno-asm-optimize @gol
611 -masm-optimize -mbranch-cost=@var{n} -mcond-exec=@var{n}}
613 @emph{S/390 and zSeries Options}
614 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
615 -mhard-float -msoft-float -mbackchain -mno-backchain @gol
616 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
617 -m64 -m31 -mdebug -mno-debug -mesa -mzarch}
620 @gccoptlist{-mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol
621 -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol
622 -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol
623 -mstack-align -mdata-align -mconst-align @gol
624 -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol
625 -melf -maout -melinux -mlinux -sim -sim2}
627 @emph{PDP-11 Options}
628 @gccoptlist{-mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol
629 -mbcopy -mbcopy-builtin -mint32 -mno-int16 @gol
630 -mint16 -mno-int32 -mfloat32 -mno-float64 @gol
631 -mfloat64 -mno-float32 -mabshi -mno-abshi @gol
632 -mbranch-expensive -mbranch-cheap @gol
633 -msplit -mno-split -munix-asm -mdec-asm}
635 @emph{Xstormy16 Options}
638 @emph{Xtensa Options}
639 @gccoptlist{-mbig-endian -mlittle-endian @gol
640 -mdensity -mno-density @gol
641 -mconst16 -mno-const16 @gol
643 -maddx -mno-addx @gol
644 -mmac16 -mno-mac16 @gol
645 -mmul16 -mno-mul16 @gol
646 -mmul32 -mno-mul32 @gol
648 -mminmax -mno-minmax @gol
649 -msext -mno-sext @gol
650 -mbooleans -mno-booleans @gol
651 -mhard-float -msoft-float @gol
652 -mfused-madd -mno-fused-madd @gol
653 -mtext-section-literals -mno-text-section-literals @gol
654 -mtarget-align -mno-target-align @gol
655 -mlongcalls -mno-longcalls}
658 @gccoptlist{-mgpr-32 -mgpr-64 -mfpr-32 -mfpr-64 @gol
659 -mhard-float -msoft-float @gol
660 -malloc-cc -mfixed-cc -mdword -mno-dword @gol
661 -mdouble -mno-double @gol
662 -mmedia -mno-media -mmuladd -mno-muladd @gol
663 -mlibrary-pic -macc-4 -macc-8 @gol
664 -mpack -mno-pack -mno-eflags -mcond-move -mno-cond-move @gol
665 -mscc -mno-scc -mcond-exec -mno-cond-exec @gol
666 -mvliw-branch -mno-vliw-branch @gol
667 -mmulti-cond-exec -mno-multi-cond-exec -mnested-cond-exec @gol
668 -mno-nested-cond-exec -mtomcat-stats @gol
671 @item Code Generation Options
672 @xref{Code Gen Options,,Options for Code Generation Conventions}.
673 @gccoptlist{-fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
674 -ffixed-@var{reg} -fexceptions @gol
675 -fnon-call-exceptions -funwind-tables @gol
676 -fasynchronous-unwind-tables @gol
677 -finhibit-size-directive -finstrument-functions @gol
678 -fno-common -fno-ident -fno-gnu-linker @gol
679 -fpcc-struct-return -fpic -fPIC -fpie -fPIE @gol
680 -freg-struct-return -fshared-data -fshort-enums @gol
681 -fshort-double -fshort-wchar @gol
682 -fverbose-asm -fpack-struct -fstack-check @gol
683 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
684 -fargument-alias -fargument-noalias @gol
685 -fargument-noalias-global -fleading-underscore @gol
686 -ftls-model=@var{model} @gol
687 -ftrapv -fwrapv -fbounds-check}
691 * Overall Options:: Controlling the kind of output:
692 an executable, object files, assembler files,
693 or preprocessed source.
694 * C Dialect Options:: Controlling the variant of C language compiled.
695 * C++ Dialect Options:: Variations on C++.
696 * Objective-C Dialect Options:: Variations on Objective-C.
697 * Language Independent Options:: Controlling how diagnostics should be
699 * Warning Options:: How picky should the compiler be?
700 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
701 * Optimize Options:: How much optimization?
702 * Preprocessor Options:: Controlling header files and macro definitions.
703 Also, getting dependency information for Make.
704 * Assembler Options:: Passing options to the assembler.
705 * Link Options:: Specifying libraries and so on.
706 * Directory Options:: Where to find header files and libraries.
707 Where to find the compiler executable files.
708 * Spec Files:: How to pass switches to sub-processes.
709 * Target Options:: Running a cross-compiler, or an old version of GCC.
712 @node Overall Options
713 @section Options Controlling the Kind of Output
715 Compilation can involve up to four stages: preprocessing, compilation
716 proper, assembly and linking, always in that order. GCC is capable of
717 preprocessing and compiling several files either into several
718 assembler input files, or into one assembler input file; then each
719 assembler input file produces an object file, and linking combines all
720 the object files (those newly compiled, and those specified as input)
721 into an executable file.
723 @cindex file name suffix
724 For any given input file, the file name suffix determines what kind of
729 C source code which must be preprocessed.
732 C source code which should not be preprocessed.
735 C++ source code which should not be preprocessed.
738 Objective-C source code. Note that you must link with the library
739 @file{libobjc.a} to make an Objective-C program work.
742 Objective-C source code which should not be preprocessed.
745 C or C++ header file to be turned into a precompiled header.
749 @itemx @var{file}.cxx
750 @itemx @var{file}.cpp
751 @itemx @var{file}.CPP
752 @itemx @var{file}.c++
754 C++ source code which must be preprocessed. Note that in @samp{.cxx},
755 the last two letters must both be literally @samp{x}. Likewise,
756 @samp{.C} refers to a literal capital C@.
760 C++ header file to be turned into a precompiled header.
763 @itemx @var{file}.for
764 @itemx @var{file}.FOR
765 Fortran source code which should not be preprocessed.
768 @itemx @var{file}.fpp
769 @itemx @var{file}.FPP
770 Fortran source code which must be preprocessed (with the traditional
774 Fortran source code which must be preprocessed with a RATFOR
775 preprocessor (not included with GCC)@.
777 @xref{Overall Options,,Options Controlling the Kind of Output, g77,
778 Using and Porting GNU Fortran}, for more details of the handling of
781 @c FIXME: Descriptions of Java file types.
788 Ada source code file which contains a library unit declaration (a
789 declaration of a package, subprogram, or generic, or a generic
790 instantiation), or a library unit renaming declaration (a package,
791 generic, or subprogram renaming declaration). Such files are also
794 @itemx @var{file}.adb
795 Ada source code file containing a library unit body (a subprogram or
796 package body). Such files are also called @dfn{bodies}.
798 @c GCC also knows about some suffixes for languages not yet included:
807 Assembler code which must be preprocessed.
810 An object file to be fed straight into linking.
811 Any file name with no recognized suffix is treated this way.
815 You can specify the input language explicitly with the @option{-x} option:
818 @item -x @var{language}
819 Specify explicitly the @var{language} for the following input files
820 (rather than letting the compiler choose a default based on the file
821 name suffix). This option applies to all following input files until
822 the next @option{-x} option. Possible values for @var{language} are:
824 c c-header cpp-output
825 c++ c++-header c++-cpp-output
826 objective-c objective-c-header objc-cpp-output
827 assembler assembler-with-cpp
829 f77 f77-cpp-input ratfor
835 Turn off any specification of a language, so that subsequent files are
836 handled according to their file name suffixes (as they are if @option{-x}
837 has not been used at all).
839 @item -pass-exit-codes
840 @opindex pass-exit-codes
841 Normally the @command{gcc} program will exit with the code of 1 if any
842 phase of the compiler returns a non-success return code. If you specify
843 @option{-pass-exit-codes}, the @command{gcc} program will instead return with
844 numerically highest error produced by any phase that returned an error
848 If you only want some of the stages of compilation, you can use
849 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
850 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
851 @command{gcc} is to stop. Note that some combinations (for example,
852 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
857 Compile or assemble the source files, but do not link. The linking
858 stage simply is not done. The ultimate output is in the form of an
859 object file for each source file.
861 By default, the object file name for a source file is made by replacing
862 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
864 Unrecognized input files, not requiring compilation or assembly, are
869 Stop after the stage of compilation proper; do not assemble. The output
870 is in the form of an assembler code file for each non-assembler input
873 By default, the assembler file name for a source file is made by
874 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
876 Input files that don't require compilation are ignored.
880 Stop after the preprocessing stage; do not run the compiler proper. The
881 output is in the form of preprocessed source code, which is sent to the
884 Input files which don't require preprocessing are ignored.
886 @cindex output file option
889 Place output in file @var{file}. This applies regardless to whatever
890 sort of output is being produced, whether it be an executable file,
891 an object file, an assembler file or preprocessed C code.
893 If you specify @option{-o} when compiling more than one input file, or
894 you are producing an executable file as output, all the source files
895 on the command line will be compiled at once.
897 If @option{-o} is not specified, the default is to put an executable file
898 in @file{a.out}, the object file for @file{@var{source}.@var{suffix}} in
899 @file{@var{source}.o}, its assembler file in @file{@var{source}.s}, and
900 all preprocessed C source on standard output.
904 Print (on standard error output) the commands executed to run the stages
905 of compilation. Also print the version number of the compiler driver
906 program and of the preprocessor and the compiler proper.
910 Like @option{-v} except the commands are not executed and all command
911 arguments are quoted. This is useful for shell scripts to capture the
912 driver-generated command lines.
916 Use pipes rather than temporary files for communication between the
917 various stages of compilation. This fails to work on some systems where
918 the assembler is unable to read from a pipe; but the GNU assembler has
923 Print (on the standard output) a description of the command line options
924 understood by @command{gcc}. If the @option{-v} option is also specified
925 then @option{--help} will also be passed on to the various processes
926 invoked by @command{gcc}, so that they can display the command line options
927 they accept. If the @option{-Wextra} option is also specified then command
928 line options which have no documentation associated with them will also
933 Print (on the standard output) a description of target specific command
934 line options for each tool.
938 Display the version number and copyrights of the invoked GCC.
942 @section Compiling C++ Programs
944 @cindex suffixes for C++ source
945 @cindex C++ source file suffixes
946 C++ source files conventionally use one of the suffixes @samp{.C},
947 @samp{.cc}, @samp{.cpp}, @samp{.CPP}, @samp{.c++}, @samp{.cp}, or
948 @samp{.cxx}; C++ header files often use @samp{.hh} or @samp{.H}; and
949 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
950 files with these names and compiles them as C++ programs even if you
951 call the compiler the same way as for compiling C programs (usually
952 with the name @command{gcc}).
956 However, C++ programs often require class libraries as well as a
957 compiler that understands the C++ language---and under some
958 circumstances, you might want to compile programs or header files from
959 standard input, or otherwise without a suffix that flags them as C++
960 programs. You might also like to precompile a C header file with a
961 @samp{.h} extension to be used in C++ compilations. @command{g++} is a
962 program that calls GCC with the default language set to C++, and
963 automatically specifies linking against the C++ library. On many
964 systems, @command{g++} is also installed with the name @command{c++}.
966 @cindex invoking @command{g++}
967 When you compile C++ programs, you may specify many of the same
968 command-line options that you use for compiling programs in any
969 language; or command-line options meaningful for C and related
970 languages; or options that are meaningful only for C++ programs.
971 @xref{C Dialect Options,,Options Controlling C Dialect}, for
972 explanations of options for languages related to C@.
973 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
974 explanations of options that are meaningful only for C++ programs.
976 @node C Dialect Options
977 @section Options Controlling C Dialect
978 @cindex dialect options
979 @cindex language dialect options
980 @cindex options, dialect
982 The following options control the dialect of C (or languages derived
983 from C, such as C++ and Objective-C) that the compiler accepts:
990 In C mode, support all ISO C90 programs. In C++ mode,
991 remove GNU extensions that conflict with ISO C++.
993 This turns off certain features of GCC that are incompatible with ISO
994 C90 (when compiling C code), or of standard C++ (when compiling C++ code),
995 such as the @code{asm} and @code{typeof} keywords, and
996 predefined macros such as @code{unix} and @code{vax} that identify the
997 type of system you are using. It also enables the undesirable and
998 rarely used ISO trigraph feature. For the C compiler,
999 it disables recognition of C++ style @samp{//} comments as well as
1000 the @code{inline} keyword.
1002 The alternate keywords @code{__asm__}, @code{__extension__},
1003 @code{__inline__} and @code{__typeof__} continue to work despite
1004 @option{-ansi}. You would not want to use them in an ISO C program, of
1005 course, but it is useful to put them in header files that might be included
1006 in compilations done with @option{-ansi}. Alternate predefined macros
1007 such as @code{__unix__} and @code{__vax__} are also available, with or
1008 without @option{-ansi}.
1010 The @option{-ansi} option does not cause non-ISO programs to be
1011 rejected gratuitously. For that, @option{-pedantic} is required in
1012 addition to @option{-ansi}. @xref{Warning Options}.
1014 The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi}
1015 option is used. Some header files may notice this macro and refrain
1016 from declaring certain functions or defining certain macros that the
1017 ISO standard doesn't call for; this is to avoid interfering with any
1018 programs that might use these names for other things.
1020 Functions which would normally be built in but do not have semantics
1021 defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in
1022 functions with @option{-ansi} is used. @xref{Other Builtins,,Other
1023 built-in functions provided by GCC}, for details of the functions
1028 Determine the language standard. This option is currently only
1029 supported when compiling C or C++. A value for this option must be
1030 provided; possible values are
1035 ISO C90 (same as @option{-ansi}).
1037 @item iso9899:199409
1038 ISO C90 as modified in amendment 1.
1044 ISO C99. Note that this standard is not yet fully supported; see
1045 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The
1046 names @samp{c9x} and @samp{iso9899:199x} are deprecated.
1049 Default, ISO C90 plus GNU extensions (including some C99 features).
1053 ISO C99 plus GNU extensions. When ISO C99 is fully implemented in GCC,
1054 this will become the default. The name @samp{gnu9x} is deprecated.
1057 The 1998 ISO C++ standard plus amendments.
1060 The same as @option{-std=c++98} plus GNU extensions. This is the
1061 default for C++ code.
1064 Even when this option is not specified, you can still use some of the
1065 features of newer standards in so far as they do not conflict with
1066 previous C standards. For example, you may use @code{__restrict__} even
1067 when @option{-std=c99} is not specified.
1069 The @option{-std} options specifying some version of ISO C have the same
1070 effects as @option{-ansi}, except that features that were not in ISO C90
1071 but are in the specified version (for example, @samp{//} comments and
1072 the @code{inline} keyword in ISO C99) are not disabled.
1074 @xref{Standards,,Language Standards Supported by GCC}, for details of
1075 these standard versions.
1077 @item -aux-info @var{filename}
1079 Output to the given filename prototyped declarations for all functions
1080 declared and/or defined in a translation unit, including those in header
1081 files. This option is silently ignored in any language other than C@.
1083 Besides declarations, the file indicates, in comments, the origin of
1084 each declaration (source file and line), whether the declaration was
1085 implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or
1086 @samp{O} for old, respectively, in the first character after the line
1087 number and the colon), and whether it came from a declaration or a
1088 definition (@samp{C} or @samp{F}, respectively, in the following
1089 character). In the case of function definitions, a K&R-style list of
1090 arguments followed by their declarations is also provided, inside
1091 comments, after the declaration.
1095 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
1096 keyword, so that code can use these words as identifiers. You can use
1097 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
1098 instead. @option{-ansi} implies @option{-fno-asm}.
1100 In C++, this switch only affects the @code{typeof} keyword, since
1101 @code{asm} and @code{inline} are standard keywords. You may want to
1102 use the @option{-fno-gnu-keywords} flag instead, which has the same
1103 effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this
1104 switch only affects the @code{asm} and @code{typeof} keywords, since
1105 @code{inline} is a standard keyword in ISO C99.
1108 @itemx -fno-builtin-@var{function}
1109 @opindex fno-builtin
1110 @cindex built-in functions
1111 Don't recognize built-in functions that do not begin with
1112 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
1113 functions provided by GCC}, for details of the functions affected,
1114 including those which are not built-in functions when @option{-ansi} or
1115 @option{-std} options for strict ISO C conformance are used because they
1116 do not have an ISO standard meaning.
1118 GCC normally generates special code to handle certain built-in functions
1119 more efficiently; for instance, calls to @code{alloca} may become single
1120 instructions that adjust the stack directly, and calls to @code{memcpy}
1121 may become inline copy loops. The resulting code is often both smaller
1122 and faster, but since the function calls no longer appear as such, you
1123 cannot set a breakpoint on those calls, nor can you change the behavior
1124 of the functions by linking with a different library.
1126 With the @option{-fno-builtin-@var{function}} option
1127 only the built-in function @var{function} is
1128 disabled. @var{function} must not begin with @samp{__builtin_}. If a
1129 function is named this is not built-in in this version of GCC, this
1130 option is ignored. There is no corresponding
1131 @option{-fbuiltin-@var{function}} option; if you wish to enable
1132 built-in functions selectively when using @option{-fno-builtin} or
1133 @option{-ffreestanding}, you may define macros such as:
1136 #define abs(n) __builtin_abs ((n))
1137 #define strcpy(d, s) __builtin_strcpy ((d), (s))
1142 @cindex hosted environment
1144 Assert that compilation takes place in a hosted environment. This implies
1145 @option{-fbuiltin}. A hosted environment is one in which the
1146 entire standard library is available, and in which @code{main} has a return
1147 type of @code{int}. Examples are nearly everything except a kernel.
1148 This is equivalent to @option{-fno-freestanding}.
1150 @item -ffreestanding
1151 @opindex ffreestanding
1152 @cindex hosted environment
1154 Assert that compilation takes place in a freestanding environment. This
1155 implies @option{-fno-builtin}. A freestanding environment
1156 is one in which the standard library may not exist, and program startup may
1157 not necessarily be at @code{main}. The most obvious example is an OS kernel.
1158 This is equivalent to @option{-fno-hosted}.
1160 @xref{Standards,,Language Standards Supported by GCC}, for details of
1161 freestanding and hosted environments.
1163 @item -fms-extensions
1164 @opindex fms-extensions
1165 Accept some non-standard constructs used in Microsoft header files.
1169 Support ISO C trigraphs. The @option{-ansi} option (and @option{-std}
1170 options for strict ISO C conformance) implies @option{-trigraphs}.
1172 @item -no-integrated-cpp
1173 @opindex no-integrated-cpp
1174 Performs a compilation in two passes: preprocessing and compiling. This
1175 option allows a user supplied "cc1", "cc1plus", or "cc1obj" via the
1176 @option{-B} option. The user supplied compilation step can then add in
1177 an additional preprocessing step after normal preprocessing but before
1178 compiling. The default is to use the integrated cpp (internal cpp)
1180 The semantics of this option will change if "cc1", "cc1plus", and
1181 "cc1obj" are merged.
1183 @cindex traditional C language
1184 @cindex C language, traditional
1186 @itemx -traditional-cpp
1187 @opindex traditional-cpp
1188 @opindex traditional
1189 Formerly, these options caused GCC to attempt to emulate a pre-standard
1190 C compiler. They are now only supported with the @option{-E} switch.
1191 The preprocessor continues to support a pre-standard mode. See the GNU
1192 CPP manual for details.
1194 @item -fcond-mismatch
1195 @opindex fcond-mismatch
1196 Allow conditional expressions with mismatched types in the second and
1197 third arguments. The value of such an expression is void. This option
1198 is not supported for C++.
1200 @item -funsigned-char
1201 @opindex funsigned-char
1202 Let the type @code{char} be unsigned, like @code{unsigned char}.
1204 Each kind of machine has a default for what @code{char} should
1205 be. It is either like @code{unsigned char} by default or like
1206 @code{signed char} by default.
1208 Ideally, a portable program should always use @code{signed char} or
1209 @code{unsigned char} when it depends on the signedness of an object.
1210 But many programs have been written to use plain @code{char} and
1211 expect it to be signed, or expect it to be unsigned, depending on the
1212 machines they were written for. This option, and its inverse, let you
1213 make such a program work with the opposite default.
1215 The type @code{char} is always a distinct type from each of
1216 @code{signed char} or @code{unsigned char}, even though its behavior
1217 is always just like one of those two.
1220 @opindex fsigned-char
1221 Let the type @code{char} be signed, like @code{signed char}.
1223 Note that this is equivalent to @option{-fno-unsigned-char}, which is
1224 the negative form of @option{-funsigned-char}. Likewise, the option
1225 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
1227 @item -fsigned-bitfields
1228 @itemx -funsigned-bitfields
1229 @itemx -fno-signed-bitfields
1230 @itemx -fno-unsigned-bitfields
1231 @opindex fsigned-bitfields
1232 @opindex funsigned-bitfields
1233 @opindex fno-signed-bitfields
1234 @opindex fno-unsigned-bitfields
1235 These options control whether a bit-field is signed or unsigned, when the
1236 declaration does not use either @code{signed} or @code{unsigned}. By
1237 default, such a bit-field is signed, because this is consistent: the
1238 basic integer types such as @code{int} are signed types.
1240 @item -fwritable-strings
1241 @opindex fwritable-strings
1242 Store string constants in the writable data segment and don't uniquize
1243 them. This is for compatibility with old programs which assume they can
1244 write into string constants.
1246 Writing into string constants is a very bad idea; ``constants'' should
1250 @node C++ Dialect Options
1251 @section Options Controlling C++ Dialect
1253 @cindex compiler options, C++
1254 @cindex C++ options, command line
1255 @cindex options, C++
1256 This section describes the command-line options that are only meaningful
1257 for C++ programs; but you can also use most of the GNU compiler options
1258 regardless of what language your program is in. For example, you
1259 might compile a file @code{firstClass.C} like this:
1262 g++ -g -frepo -O -c firstClass.C
1266 In this example, only @option{-frepo} is an option meant
1267 only for C++ programs; you can use the other options with any
1268 language supported by GCC@.
1270 Here is a list of options that are @emph{only} for compiling C++ programs:
1274 @item -fabi-version=@var{n}
1275 @opindex fabi-version
1276 Use version @var{n} of the C++ ABI. Version 1 is the version of the C++
1277 ABI that first appeared in G++ 3.2. Version 0 will always be the
1278 version that conforms most closely to the C++ ABI specification.
1279 Therefore, the ABI obtained using version 0 will change as ABI bugs are
1282 The default is version 1.
1284 @item -fno-access-control
1285 @opindex fno-access-control
1286 Turn off all access checking. This switch is mainly useful for working
1287 around bugs in the access control code.
1291 Check that the pointer returned by @code{operator new} is non-null
1292 before attempting to modify the storage allocated. This check is
1293 normally unnecessary because the C++ standard specifies that
1294 @code{operator new} will only return @code{0} if it is declared
1295 @samp{throw()}, in which case the compiler will always check the
1296 return value even without this option. In all other cases, when
1297 @code{operator new} has a non-empty exception specification, memory
1298 exhaustion is signalled by throwing @code{std::bad_alloc}. See also
1299 @samp{new (nothrow)}.
1301 @item -fconserve-space
1302 @opindex fconserve-space
1303 Put uninitialized or runtime-initialized global variables into the
1304 common segment, as C does. This saves space in the executable at the
1305 cost of not diagnosing duplicate definitions. If you compile with this
1306 flag and your program mysteriously crashes after @code{main()} has
1307 completed, you may have an object that is being destroyed twice because
1308 two definitions were merged.
1310 This option is no longer useful on most targets, now that support has
1311 been added for putting variables into BSS without making them common.
1313 @item -fno-const-strings
1314 @opindex fno-const-strings
1315 Give string constants type @code{char *} instead of type @code{const
1316 char *}. By default, G++ uses type @code{const char *} as required by
1317 the standard. Even if you use @option{-fno-const-strings}, you cannot
1318 actually modify the value of a string constant, unless you also use
1319 @option{-fwritable-strings}.
1321 This option might be removed in a future release of G++. For maximum
1322 portability, you should structure your code so that it works with
1323 string constants that have type @code{const char *}.
1325 @item -fno-elide-constructors
1326 @opindex fno-elide-constructors
1327 The C++ standard allows an implementation to omit creating a temporary
1328 which is only used to initialize another object of the same type.
1329 Specifying this option disables that optimization, and forces G++ to
1330 call the copy constructor in all cases.
1332 @item -fno-enforce-eh-specs
1333 @opindex fno-enforce-eh-specs
1334 Don't check for violation of exception specifications at runtime. This
1335 option violates the C++ standard, but may be useful for reducing code
1336 size in production builds, much like defining @samp{NDEBUG}. The compiler
1337 will still optimize based on the exception specifications.
1339 @item -fexternal-templates
1340 @opindex fexternal-templates
1342 Cause @samp{#pragma interface} and @samp{implementation} to apply to
1343 template instantiation; template instances are emitted or not according
1344 to the location of the template definition. @xref{Template
1345 Instantiation}, for more information.
1347 This option is deprecated.
1349 @item -falt-external-templates
1350 @opindex falt-external-templates
1351 Similar to @option{-fexternal-templates}, but template instances are
1352 emitted or not according to the place where they are first instantiated.
1353 @xref{Template Instantiation}, for more information.
1355 This option is deprecated.
1358 @itemx -fno-for-scope
1360 @opindex fno-for-scope
1361 If @option{-ffor-scope} is specified, the scope of variables declared in
1362 a @i{for-init-statement} is limited to the @samp{for} loop itself,
1363 as specified by the C++ standard.
1364 If @option{-fno-for-scope} is specified, the scope of variables declared in
1365 a @i{for-init-statement} extends to the end of the enclosing scope,
1366 as was the case in old versions of G++, and other (traditional)
1367 implementations of C++.
1369 The default if neither flag is given to follow the standard,
1370 but to allow and give a warning for old-style code that would
1371 otherwise be invalid, or have different behavior.
1373 @item -fno-gnu-keywords
1374 @opindex fno-gnu-keywords
1375 Do not recognize @code{typeof} as a keyword, so that code can use this
1376 word as an identifier. You can use the keyword @code{__typeof__} instead.
1377 @option{-ansi} implies @option{-fno-gnu-keywords}.
1379 @item -fno-implicit-templates
1380 @opindex fno-implicit-templates
1381 Never emit code for non-inline templates which are instantiated
1382 implicitly (i.e.@: by use); only emit code for explicit instantiations.
1383 @xref{Template Instantiation}, for more information.
1385 @item -fno-implicit-inline-templates
1386 @opindex fno-implicit-inline-templates
1387 Don't emit code for implicit instantiations of inline templates, either.
1388 The default is to handle inlines differently so that compiles with and
1389 without optimization will need the same set of explicit instantiations.
1391 @item -fno-implement-inlines
1392 @opindex fno-implement-inlines
1393 To save space, do not emit out-of-line copies of inline functions
1394 controlled by @samp{#pragma implementation}. This will cause linker
1395 errors if these functions are not inlined everywhere they are called.
1397 @item -fms-extensions
1398 @opindex fms-extensions
1399 Disable pedantic warnings about constructs used in MFC, such as implicit
1400 int and getting a pointer to member function via non-standard syntax.
1402 @item -fno-nonansi-builtins
1403 @opindex fno-nonansi-builtins
1404 Disable built-in declarations of functions that are not mandated by
1405 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
1406 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
1408 @item -fno-operator-names
1409 @opindex fno-operator-names
1410 Do not treat the operator name keywords @code{and}, @code{bitand},
1411 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
1412 synonyms as keywords.
1414 @item -fno-optional-diags
1415 @opindex fno-optional-diags
1416 Disable diagnostics that the standard says a compiler does not need to
1417 issue. Currently, the only such diagnostic issued by G++ is the one for
1418 a name having multiple meanings within a class.
1421 @opindex fpermissive
1422 Downgrade some diagnostics about nonconformant code from errors to
1423 warnings. Thus, using @option{-fpermissive} will allow some
1424 nonconforming code to compile.
1428 Enable automatic template instantiation at link time. This option also
1429 implies @option{-fno-implicit-templates}. @xref{Template
1430 Instantiation}, for more information.
1434 Disable generation of information about every class with virtual
1435 functions for use by the C++ runtime type identification features
1436 (@samp{dynamic_cast} and @samp{typeid}). If you don't use those parts
1437 of the language, you can save some space by using this flag. Note that
1438 exception handling uses the same information, but it will generate it as
1443 Emit statistics about front-end processing at the end of the compilation.
1444 This information is generally only useful to the G++ development team.
1446 @item -ftemplate-depth-@var{n}
1447 @opindex ftemplate-depth
1448 Set the maximum instantiation depth for template classes to @var{n}.
1449 A limit on the template instantiation depth is needed to detect
1450 endless recursions during template class instantiation. ANSI/ISO C++
1451 conforming programs must not rely on a maximum depth greater than 17.
1453 @item -fuse-cxa-atexit
1454 @opindex fuse-cxa-atexit
1455 Register destructors for objects with static storage duration with the
1456 @code{__cxa_atexit} function rather than the @code{atexit} function.
1457 This option is required for fully standards-compliant handling of static
1458 destructors, but will only work if your C library supports
1459 @code{__cxa_atexit}.
1463 Do not use weak symbol support, even if it is provided by the linker.
1464 By default, G++ will use weak symbols if they are available. This
1465 option exists only for testing, and should not be used by end-users;
1466 it will result in inferior code and has no benefits. This option may
1467 be removed in a future release of G++.
1471 Do not search for header files in the standard directories specific to
1472 C++, but do still search the other standard directories. (This option
1473 is used when building the C++ library.)
1476 In addition, these optimization, warning, and code generation options
1477 have meanings only for C++ programs:
1480 @item -fno-default-inline
1481 @opindex fno-default-inline
1482 Do not assume @samp{inline} for functions defined inside a class scope.
1483 @xref{Optimize Options,,Options That Control Optimization}. Note that these
1484 functions will have linkage like inline functions; they just won't be
1487 @item -Wabi @r{(C++ only)}
1489 Warn when G++ generates code that is probably not compatible with the
1490 vendor-neutral C++ ABI. Although an effort has been made to warn about
1491 all such cases, there are probably some cases that are not warned about,
1492 even though G++ is generating incompatible code. There may also be
1493 cases where warnings are emitted even though the code that is generated
1496 You should rewrite your code to avoid these warnings if you are
1497 concerned about the fact that code generated by G++ may not be binary
1498 compatible with code generated by other compilers.
1500 The known incompatibilities at this point include:
1505 Incorrect handling of tail-padding for bit-fields. G++ may attempt to
1506 pack data into the same byte as a base class. For example:
1509 struct A @{ virtual void f(); int f1 : 1; @};
1510 struct B : public A @{ int f2 : 1; @};
1514 In this case, G++ will place @code{B::f2} into the same byte
1515 as@code{A::f1}; other compilers will not. You can avoid this problem
1516 by explicitly padding @code{A} so that its size is a multiple of the
1517 byte size on your platform; that will cause G++ and other compilers to
1518 layout @code{B} identically.
1521 Incorrect handling of tail-padding for virtual bases. G++ does not use
1522 tail padding when laying out virtual bases. For example:
1525 struct A @{ virtual void f(); char c1; @};
1526 struct B @{ B(); char c2; @};
1527 struct C : public A, public virtual B @{@};
1531 In this case, G++ will not place @code{B} into the tail-padding for
1532 @code{A}; other compilers will. You can avoid this problem by
1533 explicitly padding @code{A} so that its size is a multiple of its
1534 alignment (ignoring virtual base classes); that will cause G++ and other
1535 compilers to layout @code{C} identically.
1538 Incorrect handling of bit-fields with declared widths greater than that
1539 of their underlying types, when the bit-fields appear in a union. For
1543 union U @{ int i : 4096; @};
1547 Assuming that an @code{int} does not have 4096 bits, G++ will make the
1548 union too small by the number of bits in an @code{int}.
1551 Empty classes can be placed at incorrect offsets. For example:
1561 struct C : public B, public A @{@};
1565 G++ will place the @code{A} base class of @code{C} at a nonzero offset;
1566 it should be placed at offset zero. G++ mistakenly believes that the
1567 @code{A} data member of @code{B} is already at offset zero.
1570 Names of template functions whose types involve @code{typename} or
1571 template template parameters can be mangled incorrectly.
1574 template <typename Q>
1575 void f(typename Q::X) @{@}
1577 template <template <typename> class Q>
1578 void f(typename Q<int>::X) @{@}
1582 Instantiations of these templates may be mangled incorrectly.
1586 @item -Wctor-dtor-privacy @r{(C++ only)}
1587 @opindex Wctor-dtor-privacy
1588 Warn when a class seems unusable because all the constructors or
1589 destructors in that class are private, and it has neither friends nor
1590 public static member functions.
1592 @item -Wnon-virtual-dtor @r{(C++ only)}
1593 @opindex Wnon-virtual-dtor
1594 Warn when a class appears to be polymorphic, thereby requiring a virtual
1595 destructor, yet it declares a non-virtual one.
1596 This warning is enabled by @option{-Wall}.
1598 @item -Wreorder @r{(C++ only)}
1600 @cindex reordering, warning
1601 @cindex warning for reordering of member initializers
1602 Warn when the order of member initializers given in the code does not
1603 match the order in which they must be executed. For instance:
1609 A(): j (0), i (1) @{ @}
1613 The compiler will rearrange the member initializers for @samp{i}
1614 and @samp{j} to match the declaration order of the members, emitting
1615 a warning to that effect. This warning is enabled by @option{-Wall}.
1618 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
1621 @item -Weffc++ @r{(C++ only)}
1623 Warn about violations of the following style guidelines from Scott Meyers'
1624 @cite{Effective C++} book:
1628 Item 11: Define a copy constructor and an assignment operator for classes
1629 with dynamically allocated memory.
1632 Item 12: Prefer initialization to assignment in constructors.
1635 Item 14: Make destructors virtual in base classes.
1638 Item 15: Have @code{operator=} return a reference to @code{*this}.
1641 Item 23: Don't try to return a reference when you must return an object.
1645 Also warn about violations of the following style guidelines from
1646 Scott Meyers' @cite{More Effective C++} book:
1650 Item 6: Distinguish between prefix and postfix forms of increment and
1651 decrement operators.
1654 Item 7: Never overload @code{&&}, @code{||}, or @code{,}.
1658 When selecting this option, be aware that the standard library
1659 headers do not obey all of these guidelines; use @samp{grep -v}
1660 to filter out those warnings.
1662 @item -Wno-deprecated @r{(C++ only)}
1663 @opindex Wno-deprecated
1664 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
1666 @item -Wno-non-template-friend @r{(C++ only)}
1667 @opindex Wno-non-template-friend
1668 Disable warnings when non-templatized friend functions are declared
1669 within a template. Since the advent of explicit template specification
1670 support in G++, if the name of the friend is an unqualified-id (i.e.,
1671 @samp{friend foo(int)}), the C++ language specification demands that the
1672 friend declare or define an ordinary, nontemplate function. (Section
1673 14.5.3). Before G++ implemented explicit specification, unqualified-ids
1674 could be interpreted as a particular specialization of a templatized
1675 function. Because this non-conforming behavior is no longer the default
1676 behavior for G++, @option{-Wnon-template-friend} allows the compiler to
1677 check existing code for potential trouble spots and is on by default.
1678 This new compiler behavior can be turned off with
1679 @option{-Wno-non-template-friend} which keeps the conformant compiler code
1680 but disables the helpful warning.
1682 @item -Wold-style-cast @r{(C++ only)}
1683 @opindex Wold-style-cast
1684 Warn if an old-style (C-style) cast to a non-void type is used within
1685 a C++ program. The new-style casts (@samp{static_cast},
1686 @samp{reinterpret_cast}, and @samp{const_cast}) are less vulnerable to
1687 unintended effects and much easier to search for.
1689 @item -Woverloaded-virtual @r{(C++ only)}
1690 @opindex Woverloaded-virtual
1691 @cindex overloaded virtual fn, warning
1692 @cindex warning for overloaded virtual fn
1693 Warn when a function declaration hides virtual functions from a
1694 base class. For example, in:
1701 struct B: public A @{
1706 the @code{A} class version of @code{f} is hidden in @code{B}, and code
1714 will fail to compile.
1716 @item -Wno-pmf-conversions @r{(C++ only)}
1717 @opindex Wno-pmf-conversions
1718 Disable the diagnostic for converting a bound pointer to member function
1721 @item -Wsign-promo @r{(C++ only)}
1722 @opindex Wsign-promo
1723 Warn when overload resolution chooses a promotion from unsigned or
1724 enumeral type to a signed type, over a conversion to an unsigned type of
1725 the same size. Previous versions of G++ would try to preserve
1726 unsignedness, but the standard mandates the current behavior.
1728 @item -Wsynth @r{(C++ only)}
1730 @cindex warning for synthesized methods
1731 @cindex synthesized methods, warning
1732 Warn when G++'s synthesis behavior does not match that of cfront. For
1738 A& operator = (int);
1748 In this example, G++ will synthesize a default @samp{A& operator =
1749 (const A&);}, while cfront will use the user-defined @samp{operator =}.
1752 @node Objective-C Dialect Options
1753 @section Options Controlling Objective-C Dialect
1755 @cindex compiler options, Objective-C
1756 @cindex Objective-C options, command line
1757 @cindex options, Objective-C
1758 This section describes the command-line options that are only meaningful
1759 for Objective-C programs, but you can also use most of the GNU compiler
1760 options regardless of what language your program is in. For example,
1761 you might compile a file @code{some_class.m} like this:
1764 gcc -g -fgnu-runtime -O -c some_class.m
1768 In this example, @option{-fgnu-runtime} is an option meant only for
1769 Objective-C programs; you can use the other options with any language
1772 Here is a list of options that are @emph{only} for compiling Objective-C
1776 @item -fconstant-string-class=@var{class-name}
1777 @opindex fconstant-string-class
1778 Use @var{class-name} as the name of the class to instantiate for each
1779 literal string specified with the syntax @code{@@"@dots{}"}. The default
1780 class name is @code{NXConstantString}.
1783 @opindex fgnu-runtime
1784 Generate object code compatible with the standard GNU Objective-C
1785 runtime. This is the default for most types of systems.
1787 @item -fnext-runtime
1788 @opindex fnext-runtime
1789 Generate output compatible with the NeXT runtime. This is the default
1790 for NeXT-based systems, including Darwin and Mac OS X@. The macro
1791 @code{__NEXT_RUNTIME__} is predefined if (and only if) this option is
1796 Dump interface declarations for all classes seen in the source file to a
1797 file named @file{@var{sourcename}.decl}.
1800 @opindex Wno-protocol
1801 If a class is declared to implement a protocol, a warning is issued for
1802 every method in the protocol that is not implemented by the class. The
1803 default behavior is to issue a warning for every method not explicitly
1804 implemented in the class, even if a method implementation is inherited
1805 from the superclass. If you use the @code{-Wno-protocol} option, then
1806 methods inherited from the superclass are considered to be implemented,
1807 and no warning is issued for them.
1811 Warn if multiple methods of different types for the same selector are
1812 found during compilation. The check is performed on the list of methods
1813 in the final stage of compilation. Additionally, a check is performed
1814 for each selector appearing in a @code{@@selector(@dots{})}
1815 expression, and a corresponding method for that selector has been found
1816 during compilation. Because these checks scan the method table only at
1817 the end of compilation, these warnings are not produced if the final
1818 stage of compilation is not reached, for example because an error is
1819 found during compilation, or because the @code{-fsyntax-only} option is
1822 @item -Wundeclared-selector
1823 @opindex Wundeclared-selector
1824 Warn if a @code{@@selector(@dots{})} expression referring to an
1825 undeclared selector is found. A selector is considered undeclared if no
1826 method with that name has been declared before the
1827 @code{@@selector(@dots{})} expression, either explicitly in an
1828 @code{@@interface} or @code{@@protocol} declaration, or implicitly in
1829 an @code{@@implementation} section. This option always performs its
1830 checks as soon as a @code{@@selector(@dots{})} expression is found,
1831 while @code{-Wselector} only performs its checks in the final stage of
1832 compilation. This also enforces the coding style convention
1833 that methods and selectors must be declared before being used.
1835 @c not documented because only avail via -Wp
1836 @c @item -print-objc-runtime-info
1840 @node Language Independent Options
1841 @section Options to Control Diagnostic Messages Formatting
1842 @cindex options to control diagnostics formatting
1843 @cindex diagnostic messages
1844 @cindex message formatting
1846 Traditionally, diagnostic messages have been formatted irrespective of
1847 the output device's aspect (e.g.@: its width, @dots{}). The options described
1848 below can be used to control the diagnostic messages formatting
1849 algorithm, e.g.@: how many characters per line, how often source location
1850 information should be reported. Right now, only the C++ front end can
1851 honor these options. However it is expected, in the near future, that
1852 the remaining front ends would be able to digest them correctly.
1855 @item -fmessage-length=@var{n}
1856 @opindex fmessage-length
1857 Try to format error messages so that they fit on lines of about @var{n}
1858 characters. The default is 72 characters for @command{g++} and 0 for the rest of
1859 the front ends supported by GCC@. If @var{n} is zero, then no
1860 line-wrapping will be done; each error message will appear on a single
1863 @opindex fdiagnostics-show-location
1864 @item -fdiagnostics-show-location=once
1865 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
1866 reporter to emit @emph{once} source location information; that is, in
1867 case the message is too long to fit on a single physical line and has to
1868 be wrapped, the source location won't be emitted (as prefix) again,
1869 over and over, in subsequent continuation lines. This is the default
1872 @item -fdiagnostics-show-location=every-line
1873 Only meaningful in line-wrapping mode. Instructs the diagnostic
1874 messages reporter to emit the same source location information (as
1875 prefix) for physical lines that result from the process of breaking
1876 a message which is too long to fit on a single line.
1880 @node Warning Options
1881 @section Options to Request or Suppress Warnings
1882 @cindex options to control warnings
1883 @cindex warning messages
1884 @cindex messages, warning
1885 @cindex suppressing warnings
1887 Warnings are diagnostic messages that report constructions which
1888 are not inherently erroneous but which are risky or suggest there
1889 may have been an error.
1891 You can request many specific warnings with options beginning @samp{-W},
1892 for example @option{-Wimplicit} to request warnings on implicit
1893 declarations. Each of these specific warning options also has a
1894 negative form beginning @samp{-Wno-} to turn off warnings;
1895 for example, @option{-Wno-implicit}. This manual lists only one of the
1896 two forms, whichever is not the default.
1898 The following options control the amount and kinds of warnings produced
1899 by GCC; for further, language-specific options also refer to
1900 @ref{C++ Dialect Options} and @ref{Objective-C Dialect Options}.
1903 @cindex syntax checking
1905 @opindex fsyntax-only
1906 Check the code for syntax errors, but don't do anything beyond that.
1910 Issue all the warnings demanded by strict ISO C and ISO C++;
1911 reject all programs that use forbidden extensions, and some other
1912 programs that do not follow ISO C and ISO C++. For ISO C, follows the
1913 version of the ISO C standard specified by any @option{-std} option used.
1915 Valid ISO C and ISO C++ programs should compile properly with or without
1916 this option (though a rare few will require @option{-ansi} or a
1917 @option{-std} option specifying the required version of ISO C)@. However,
1918 without this option, certain GNU extensions and traditional C and C++
1919 features are supported as well. With this option, they are rejected.
1921 @option{-pedantic} does not cause warning messages for use of the
1922 alternate keywords whose names begin and end with @samp{__}. Pedantic
1923 warnings are also disabled in the expression that follows
1924 @code{__extension__}. However, only system header files should use
1925 these escape routes; application programs should avoid them.
1926 @xref{Alternate Keywords}.
1928 Some users try to use @option{-pedantic} to check programs for strict ISO
1929 C conformance. They soon find that it does not do quite what they want:
1930 it finds some non-ISO practices, but not all---only those for which
1931 ISO C @emph{requires} a diagnostic, and some others for which
1932 diagnostics have been added.
1934 A feature to report any failure to conform to ISO C might be useful in
1935 some instances, but would require considerable additional work and would
1936 be quite different from @option{-pedantic}. We don't have plans to
1937 support such a feature in the near future.
1939 Where the standard specified with @option{-std} represents a GNU
1940 extended dialect of C, such as @samp{gnu89} or @samp{gnu99}, there is a
1941 corresponding @dfn{base standard}, the version of ISO C on which the GNU
1942 extended dialect is based. Warnings from @option{-pedantic} are given
1943 where they are required by the base standard. (It would not make sense
1944 for such warnings to be given only for features not in the specified GNU
1945 C dialect, since by definition the GNU dialects of C include all
1946 features the compiler supports with the given option, and there would be
1947 nothing to warn about.)
1949 @item -pedantic-errors
1950 @opindex pedantic-errors
1951 Like @option{-pedantic}, except that errors are produced rather than
1956 Inhibit all warning messages.
1960 Inhibit warning messages about the use of @samp{#import}.
1962 @item -Wchar-subscripts
1963 @opindex Wchar-subscripts
1964 Warn if an array subscript has type @code{char}. This is a common cause
1965 of error, as programmers often forget that this type is signed on some
1970 Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*}
1971 comment, or whenever a Backslash-Newline appears in a @samp{//} comment.
1975 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
1976 the arguments supplied have types appropriate to the format string
1977 specified, and that the conversions specified in the format string make
1978 sense. This includes standard functions, and others specified by format
1979 attributes (@pxref{Function Attributes}), in the @code{printf},
1980 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
1981 not in the C standard) families.
1983 The formats are checked against the format features supported by GNU
1984 libc version 2.2. These include all ISO C90 and C99 features, as well
1985 as features from the Single Unix Specification and some BSD and GNU
1986 extensions. Other library implementations may not support all these
1987 features; GCC does not support warning about features that go beyond a
1988 particular library's limitations. However, if @option{-pedantic} is used
1989 with @option{-Wformat}, warnings will be given about format features not
1990 in the selected standard version (but not for @code{strfmon} formats,
1991 since those are not in any version of the C standard). @xref{C Dialect
1992 Options,,Options Controlling C Dialect}.
1994 Since @option{-Wformat} also checks for null format arguments for
1995 several functions, @option{-Wformat} also implies @option{-Wnonnull}.
1997 @option{-Wformat} is included in @option{-Wall}. For more control over some
1998 aspects of format checking, the options @option{-Wno-format-y2k},
1999 @option{-Wno-format-extra-args}, @option{-Wno-format-zero-length},
2000 @option{-Wformat-nonliteral}, @option{-Wformat-security}, and
2001 @option{-Wformat=2} are available, but are not included in @option{-Wall}.
2003 @item -Wno-format-y2k
2004 @opindex Wno-format-y2k
2005 If @option{-Wformat} is specified, do not warn about @code{strftime}
2006 formats which may yield only a two-digit year.
2008 @item -Wno-format-extra-args
2009 @opindex Wno-format-extra-args
2010 If @option{-Wformat} is specified, do not warn about excess arguments to a
2011 @code{printf} or @code{scanf} format function. The C standard specifies
2012 that such arguments are ignored.
2014 Where the unused arguments lie between used arguments that are
2015 specified with @samp{$} operand number specifications, normally
2016 warnings are still given, since the implementation could not know what
2017 type to pass to @code{va_arg} to skip the unused arguments. However,
2018 in the case of @code{scanf} formats, this option will suppress the
2019 warning if the unused arguments are all pointers, since the Single
2020 Unix Specification says that such unused arguments are allowed.
2022 @item -Wno-format-zero-length
2023 @opindex Wno-format-zero-length
2024 If @option{-Wformat} is specified, do not warn about zero-length formats.
2025 The C standard specifies that zero-length formats are allowed.
2027 @item -Wformat-nonliteral
2028 @opindex Wformat-nonliteral
2029 If @option{-Wformat} is specified, also warn if the format string is not a
2030 string literal and so cannot be checked, unless the format function
2031 takes its format arguments as a @code{va_list}.
2033 @item -Wformat-security
2034 @opindex Wformat-security
2035 If @option{-Wformat} is specified, also warn about uses of format
2036 functions that represent possible security problems. At present, this
2037 warns about calls to @code{printf} and @code{scanf} functions where the
2038 format string is not a string literal and there are no format arguments,
2039 as in @code{printf (foo);}. This may be a security hole if the format
2040 string came from untrusted input and contains @samp{%n}. (This is
2041 currently a subset of what @option{-Wformat-nonliteral} warns about, but
2042 in future warnings may be added to @option{-Wformat-security} that are not
2043 included in @option{-Wformat-nonliteral}.)
2047 Enable @option{-Wformat} plus format checks not included in
2048 @option{-Wformat}. Currently equivalent to @samp{-Wformat
2049 -Wformat-nonliteral -Wformat-security}.
2053 Enable warning about passing a null pointer for arguments marked as
2054 requiring a non-null value by the @code{nonnull} function attribute.
2056 @option{-Wnonnull} is included in @option{-Wall} and @option{-Wformat}. It
2057 can be disabled with the @option{-Wno-nonnull} option.
2059 @item -Winit-self @r{(C, C++, and Objective-C only)}
2061 Enable warning about uninitialized variables which are initalized with themselves.
2062 Note this option can only be used with the @option{-Wuninitialized} option and
2063 that only works with @option{-O}.
2065 For an example, the following code will not warn about i being uninitialized
2066 without this option:
2077 @item -Wimplicit-int
2078 @opindex Wimplicit-int
2079 Warn when a declaration does not specify a type.
2081 @item -Wimplicit-function-declaration
2082 @itemx -Werror-implicit-function-declaration
2083 @opindex Wimplicit-function-declaration
2084 @opindex Werror-implicit-function-declaration
2085 Give a warning (or error) whenever a function is used before being
2090 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
2094 Warn if the type of @samp{main} is suspicious. @samp{main} should be a
2095 function with external linkage, returning int, taking either zero
2096 arguments, two, or three arguments of appropriate types.
2098 @item -Wmissing-braces
2099 @opindex Wmissing-braces
2100 Warn if an aggregate or union initializer is not fully bracketed. In
2101 the following example, the initializer for @samp{a} is not fully
2102 bracketed, but that for @samp{b} is fully bracketed.
2105 int a[2][2] = @{ 0, 1, 2, 3 @};
2106 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
2110 @opindex Wparentheses
2111 Warn if parentheses are omitted in certain contexts, such
2112 as when there is an assignment in a context where a truth value
2113 is expected, or when operators are nested whose precedence people
2114 often get confused about.
2116 Also warn about constructions where there may be confusion to which
2117 @code{if} statement an @code{else} branch belongs. Here is an example of
2132 In C, every @code{else} branch belongs to the innermost possible @code{if}
2133 statement, which in this example is @code{if (b)}. This is often not
2134 what the programmer expected, as illustrated in the above example by
2135 indentation the programmer chose. When there is the potential for this
2136 confusion, GCC will issue a warning when this flag is specified.
2137 To eliminate the warning, add explicit braces around the innermost
2138 @code{if} statement so there is no way the @code{else} could belong to
2139 the enclosing @code{if}. The resulting code would look like this:
2155 @item -Wsequence-point
2156 @opindex Wsequence-point
2157 Warn about code that may have undefined semantics because of violations
2158 of sequence point rules in the C standard.
2160 The C standard defines the order in which expressions in a C program are
2161 evaluated in terms of @dfn{sequence points}, which represent a partial
2162 ordering between the execution of parts of the program: those executed
2163 before the sequence point, and those executed after it. These occur
2164 after the evaluation of a full expression (one which is not part of a
2165 larger expression), after the evaluation of the first operand of a
2166 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
2167 function is called (but after the evaluation of its arguments and the
2168 expression denoting the called function), and in certain other places.
2169 Other than as expressed by the sequence point rules, the order of
2170 evaluation of subexpressions of an expression is not specified. All
2171 these rules describe only a partial order rather than a total order,
2172 since, for example, if two functions are called within one expression
2173 with no sequence point between them, the order in which the functions
2174 are called is not specified. However, the standards committee have
2175 ruled that function calls do not overlap.
2177 It is not specified when between sequence points modifications to the
2178 values of objects take effect. Programs whose behavior depends on this
2179 have undefined behavior; the C standard specifies that ``Between the
2180 previous and next sequence point an object shall have its stored value
2181 modified at most once by the evaluation of an expression. Furthermore,
2182 the prior value shall be read only to determine the value to be
2183 stored.''. If a program breaks these rules, the results on any
2184 particular implementation are entirely unpredictable.
2186 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
2187 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
2188 diagnosed by this option, and it may give an occasional false positive
2189 result, but in general it has been found fairly effective at detecting
2190 this sort of problem in programs.
2192 The present implementation of this option only works for C programs. A
2193 future implementation may also work for C++ programs.
2195 The C standard is worded confusingly, therefore there is some debate
2196 over the precise meaning of the sequence point rules in subtle cases.
2197 Links to discussions of the problem, including proposed formal
2198 definitions, may be found on our readings page, at
2199 @w{@uref{http://gcc.gnu.org/readings.html}}.
2202 @opindex Wreturn-type
2203 Warn whenever a function is defined with a return-type that defaults to
2204 @code{int}. Also warn about any @code{return} statement with no
2205 return-value in a function whose return-type is not @code{void}.
2207 For C++, a function without return type always produces a diagnostic
2208 message, even when @option{-Wno-return-type} is specified. The only
2209 exceptions are @samp{main} and functions defined in system headers.
2213 Warn whenever a @code{switch} statement has an index of enumeral type
2214 and lacks a @code{case} for one or more of the named codes of that
2215 enumeration. (The presence of a @code{default} label prevents this
2216 warning.) @code{case} labels outside the enumeration range also
2217 provoke warnings when this option is used.
2219 @item -Wswitch-default
2220 @opindex Wswitch-switch
2221 Warn whenever a @code{switch} statement does not have a @code{default}
2225 @opindex Wswitch-enum
2226 Warn whenever a @code{switch} statement has an index of enumeral type
2227 and lacks a @code{case} for one or more of the named codes of that
2228 enumeration. @code{case} labels outside the enumeration range also
2229 provoke warnings when this option is used.
2233 Warn if any trigraphs are encountered that might change the meaning of
2234 the program (trigraphs within comments are not warned about).
2236 @item -Wunused-function
2237 @opindex Wunused-function
2238 Warn whenever a static function is declared but not defined or a
2239 non\-inline static function is unused.
2241 @item -Wunused-label
2242 @opindex Wunused-label
2243 Warn whenever a label is declared but not used.
2245 To suppress this warning use the @samp{unused} attribute
2246 (@pxref{Variable Attributes}).
2248 @item -Wunused-parameter
2249 @opindex Wunused-parameter
2250 Warn whenever a function parameter is unused aside from its declaration.
2252 To suppress this warning use the @samp{unused} attribute
2253 (@pxref{Variable Attributes}).
2255 @item -Wunused-variable
2256 @opindex Wunused-variable
2257 Warn whenever a local variable or non-constant static variable is unused
2258 aside from its declaration
2260 To suppress this warning use the @samp{unused} attribute
2261 (@pxref{Variable Attributes}).
2263 @item -Wunused-value
2264 @opindex Wunused-value
2265 Warn whenever a statement computes a result that is explicitly not used.
2267 To suppress this warning cast the expression to @samp{void}.
2271 All the above @option{-Wunused} options combined.
2273 In order to get a warning about an unused function parameter, you must
2274 either specify @samp{-Wextra -Wunused} (note that @samp{-Wall} implies
2275 @samp{-Wunused}), or separately specify @option{-Wunused-parameter}.
2277 @item -Wuninitialized
2278 @opindex Wuninitialized
2279 Warn if an automatic variable is used without first being initialized or
2280 if a variable may be clobbered by a @code{setjmp} call.
2282 These warnings are possible only in optimizing compilation,
2283 because they require data flow information that is computed only
2284 when optimizing. If you don't specify @option{-O}, you simply won't
2287 If you want to warn about code which uses the uninitialized value of the
2288 variable in its own initializer, use the @option{-Winit-self} option.
2290 These warnings occur only for variables that are candidates for
2291 register allocation. Therefore, they do not occur for a variable that
2292 is declared @code{volatile}, or whose address is taken, or whose size
2293 is other than 1, 2, 4 or 8 bytes. Also, they do not occur for
2294 structures, unions or arrays, even when they are in registers.
2296 Note that there may be no warning about a variable that is used only
2297 to compute a value that itself is never used, because such
2298 computations may be deleted by data flow analysis before the warnings
2301 These warnings are made optional because GCC is not smart
2302 enough to see all the reasons why the code might be correct
2303 despite appearing to have an error. Here is one example of how
2324 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
2325 always initialized, but GCC doesn't know this. Here is
2326 another common case:
2331 if (change_y) save_y = y, y = new_y;
2333 if (change_y) y = save_y;
2338 This has no bug because @code{save_y} is used only if it is set.
2340 @cindex @code{longjmp} warnings
2341 This option also warns when a non-volatile automatic variable might be
2342 changed by a call to @code{longjmp}. These warnings as well are possible
2343 only in optimizing compilation.
2345 The compiler sees only the calls to @code{setjmp}. It cannot know
2346 where @code{longjmp} will be called; in fact, a signal handler could
2347 call it at any point in the code. As a result, you may get a warning
2348 even when there is in fact no problem because @code{longjmp} cannot
2349 in fact be called at the place which would cause a problem.
2351 Some spurious warnings can be avoided if you declare all the functions
2352 you use that never return as @code{noreturn}. @xref{Function
2355 @item -Wunknown-pragmas
2356 @opindex Wunknown-pragmas
2357 @cindex warning for unknown pragmas
2358 @cindex unknown pragmas, warning
2359 @cindex pragmas, warning of unknown
2360 Warn when a #pragma directive is encountered which is not understood by
2361 GCC@. If this command line option is used, warnings will even be issued
2362 for unknown pragmas in system header files. This is not the case if
2363 the warnings were only enabled by the @option{-Wall} command line option.
2365 @item -Wstrict-aliasing
2366 @opindex Wstrict-aliasing
2367 This option is only active when @option{-fstrict-aliasing} is active.
2368 It warns about code which might break the strict aliasing rules that the
2369 compiler is using for optimization. The warning does not catch all
2370 cases, but does attempt to catch the more common pitfalls. It is
2371 included in @option{-Wall}.
2375 All of the above @samp{-W} options combined. This enables all the
2376 warnings about constructions that some users consider questionable, and
2377 that are easy to avoid (or modify to prevent the warning), even in
2378 conjunction with macros. This also enables some language-specific
2379 warnings described in @ref{C++ Dialect Options} and
2380 @ref{Objective-C Dialect Options}.
2383 The following @option{-W@dots{}} options are not implied by @option{-Wall}.
2384 Some of them warn about constructions that users generally do not
2385 consider questionable, but which occasionally you might wish to check
2386 for; others warn about constructions that are necessary or hard to avoid
2387 in some cases, and there is no simple way to modify the code to suppress
2394 (This option used to be called @option{-W}. The older name is still
2395 supported, but the newer name is more descriptive.) Print extra warning
2396 messages for these events:
2400 A function can return either with or without a value. (Falling
2401 off the end of the function body is considered returning without
2402 a value.) For example, this function would evoke such a
2416 An expression-statement or the left-hand side of a comma expression
2417 contains no side effects.
2418 To suppress the warning, cast the unused expression to void.
2419 For example, an expression such as @samp{x[i,j]} will cause a warning,
2420 but @samp{x[(void)i,j]} will not.
2423 An unsigned value is compared against zero with @samp{<} or @samp{>=}.
2426 A comparison like @samp{x<=y<=z} appears; this is equivalent to
2427 @samp{(x<=y ? 1 : 0) <= z}, which is a different interpretation from
2428 that of ordinary mathematical notation.
2431 Storage-class specifiers like @code{static} are not the first things in
2432 a declaration. According to the C Standard, this usage is obsolescent.
2435 The return type of a function has a type qualifier such as @code{const}.
2436 Such a type qualifier has no effect, since the value returned by a
2437 function is not an lvalue. (But don't warn about the GNU extension of
2438 @code{volatile void} return types. That extension will be warned about
2439 if @option{-pedantic} is specified.)
2442 If @option{-Wall} or @option{-Wunused} is also specified, warn about unused
2446 A comparison between signed and unsigned values could produce an
2447 incorrect result when the signed value is converted to unsigned.
2448 (But don't warn if @option{-Wno-sign-compare} is also specified.)
2451 An aggregate has an initializer which does not initialize all members.
2452 For example, the following code would cause such a warning, because
2453 @code{x.h} would be implicitly initialized to zero:
2456 struct s @{ int f, g, h; @};
2457 struct s x = @{ 3, 4 @};
2461 A function parameter is declared without a type specifier in K&R-style
2469 An empty body occurs in an @samp{if} or @samp{else} statement.
2472 A pointer is compared against integer zero with @samp{<}, @samp{<=},
2473 @samp{>}, or @samp{>=}.
2476 A variable might be changed by @samp{longjmp} or @samp{vfork}.
2479 Any of several floating-point events that often indicate errors, such as
2480 overflow, underflow, loss of precision, etc.
2482 @item @r{(C++ only)}
2483 An enumerator and a non-enumerator both appear in a conditional expression.
2485 @item @r{(C++ only)}
2486 A non-static reference or non-static @samp{const} member appears in a
2487 class without constructors.
2489 @item @r{(C++ only)}
2490 Ambiguous virtual bases.
2492 @item @r{(C++ only)}
2493 Subscripting an array which has been declared @samp{register}.
2495 @item @r{(C++ only)}
2496 Taking the address of a variable which has been declared @samp{register}.
2498 @item @r{(C++ only)}
2499 A base class is not initialized in a derived class' copy constructor.
2502 @item -Wno-div-by-zero
2503 @opindex Wno-div-by-zero
2504 @opindex Wdiv-by-zero
2505 Do not warn about compile-time integer division by zero. Floating point
2506 division by zero is not warned about, as it can be a legitimate way of
2507 obtaining infinities and NaNs.
2509 @item -Wsystem-headers
2510 @opindex Wsystem-headers
2511 @cindex warnings from system headers
2512 @cindex system headers, warnings from
2513 Print warning messages for constructs found in system header files.
2514 Warnings from system headers are normally suppressed, on the assumption
2515 that they usually do not indicate real problems and would only make the
2516 compiler output harder to read. Using this command line option tells
2517 GCC to emit warnings from system headers as if they occurred in user
2518 code. However, note that using @option{-Wall} in conjunction with this
2519 option will @emph{not} warn about unknown pragmas in system
2520 headers---for that, @option{-Wunknown-pragmas} must also be used.
2523 @opindex Wfloat-equal
2524 Warn if floating point values are used in equality comparisons.
2526 The idea behind this is that sometimes it is convenient (for the
2527 programmer) to consider floating-point values as approximations to
2528 infinitely precise real numbers. If you are doing this, then you need
2529 to compute (by analyzing the code, or in some other way) the maximum or
2530 likely maximum error that the computation introduces, and allow for it
2531 when performing comparisons (and when producing output, but that's a
2532 different problem). In particular, instead of testing for equality, you
2533 would check to see whether the two values have ranges that overlap; and
2534 this is done with the relational operators, so equality comparisons are
2537 @item -Wtraditional @r{(C only)}
2538 @opindex Wtraditional
2539 Warn about certain constructs that behave differently in traditional and
2540 ISO C@. Also warn about ISO C constructs that have no traditional C
2541 equivalent, and/or problematic constructs which should be avoided.
2545 Macro parameters that appear within string literals in the macro body.
2546 In traditional C macro replacement takes place within string literals,
2547 but does not in ISO C@.
2550 In traditional C, some preprocessor directives did not exist.
2551 Traditional preprocessors would only consider a line to be a directive
2552 if the @samp{#} appeared in column 1 on the line. Therefore
2553 @option{-Wtraditional} warns about directives that traditional C
2554 understands but would ignore because the @samp{#} does not appear as the
2555 first character on the line. It also suggests you hide directives like
2556 @samp{#pragma} not understood by traditional C by indenting them. Some
2557 traditional implementations would not recognize @samp{#elif}, so it
2558 suggests avoiding it altogether.
2561 A function-like macro that appears without arguments.
2564 The unary plus operator.
2567 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating point
2568 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
2569 constants.) Note, these suffixes appear in macros defined in the system
2570 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
2571 Use of these macros in user code might normally lead to spurious
2572 warnings, however gcc's integrated preprocessor has enough context to
2573 avoid warning in these cases.
2576 A function declared external in one block and then used after the end of
2580 A @code{switch} statement has an operand of type @code{long}.
2583 A non-@code{static} function declaration follows a @code{static} one.
2584 This construct is not accepted by some traditional C compilers.
2587 The ISO type of an integer constant has a different width or
2588 signedness from its traditional type. This warning is only issued if
2589 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
2590 typically represent bit patterns, are not warned about.
2593 Usage of ISO string concatenation is detected.
2596 Initialization of automatic aggregates.
2599 Identifier conflicts with labels. Traditional C lacks a separate
2600 namespace for labels.
2603 Initialization of unions. If the initializer is zero, the warning is
2604 omitted. This is done under the assumption that the zero initializer in
2605 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
2606 initializer warnings and relies on default initialization to zero in the
2610 Conversions by prototypes between fixed/floating point values and vice
2611 versa. The absence of these prototypes when compiling with traditional
2612 C would cause serious problems. This is a subset of the possible
2613 conversion warnings, for the full set use @option{-Wconversion}.
2616 Use of ISO C style function definitions. This warning intentionally is
2617 @emph{not} issued for prototype declarations or variadic functions
2618 because these ISO C features will appear in your code when using
2619 libiberty's traditional C compatibility macros, @code{PARAMS} and
2620 @code{VPARAMS}. This warning is also bypassed for nested functions
2621 because that feature is already a gcc extension and thus not relevant to
2622 traditional C compatibility.
2625 @item -Wdeclaration-after-statement @r{(C only)}
2626 @opindex Wdeclaration-after-statement
2627 Warn when a declaration is found after a statement in a block. This
2628 construct, known from C++, was introduced with ISO C99 and is by default
2629 allowed in GCC@. It is not supported by ISO C90 and was not supported by
2630 GCC versions before GCC 3.0. @xref{Mixed Declarations}.
2634 Warn if an undefined identifier is evaluated in an @samp{#if} directive.
2636 @item -Wendif-labels
2637 @opindex Wendif-labels
2638 Warn whenever an @samp{#else} or an @samp{#endif} are followed by text.
2642 Warn whenever a local variable shadows another local variable, parameter or
2643 global variable or whenever a built-in function is shadowed.
2645 @item -Wlarger-than-@var{len}
2646 @opindex Wlarger-than
2647 Warn whenever an object of larger than @var{len} bytes is defined.
2649 @item -Wpointer-arith
2650 @opindex Wpointer-arith
2651 Warn about anything that depends on the ``size of'' a function type or
2652 of @code{void}. GNU C assigns these types a size of 1, for
2653 convenience in calculations with @code{void *} pointers and pointers
2656 @item -Wbad-function-cast @r{(C only)}
2657 @opindex Wbad-function-cast
2658 Warn whenever a function call is cast to a non-matching type.
2659 For example, warn if @code{int malloc()} is cast to @code{anything *}.
2663 Warn whenever a pointer is cast so as to remove a type qualifier from
2664 the target type. For example, warn if a @code{const char *} is cast
2665 to an ordinary @code{char *}.
2668 @opindex Wcast-align
2669 Warn whenever a pointer is cast such that the required alignment of the
2670 target is increased. For example, warn if a @code{char *} is cast to
2671 an @code{int *} on machines where integers can only be accessed at
2672 two- or four-byte boundaries.
2674 @item -Wwrite-strings
2675 @opindex Wwrite-strings
2676 When compiling C, give string constants the type @code{const
2677 char[@var{length}]} so that
2678 copying the address of one into a non-@code{const} @code{char *}
2679 pointer will get a warning; when compiling C++, warn about the
2680 deprecated conversion from string constants to @code{char *}.
2681 These warnings will help you find at
2682 compile time code that can try to write into a string constant, but
2683 only if you have been very careful about using @code{const} in
2684 declarations and prototypes. Otherwise, it will just be a nuisance;
2685 this is why we did not make @option{-Wall} request these warnings.
2688 @opindex Wconversion
2689 Warn if a prototype causes a type conversion that is different from what
2690 would happen to the same argument in the absence of a prototype. This
2691 includes conversions of fixed point to floating and vice versa, and
2692 conversions changing the width or signedness of a fixed point argument
2693 except when the same as the default promotion.
2695 Also, warn if a negative integer constant expression is implicitly
2696 converted to an unsigned type. For example, warn about the assignment
2697 @code{x = -1} if @code{x} is unsigned. But do not warn about explicit
2698 casts like @code{(unsigned) -1}.
2700 @item -Wsign-compare
2701 @opindex Wsign-compare
2702 @cindex warning for comparison of signed and unsigned values
2703 @cindex comparison of signed and unsigned values, warning
2704 @cindex signed and unsigned values, comparison warning
2705 Warn when a comparison between signed and unsigned values could produce
2706 an incorrect result when the signed value is converted to unsigned.
2707 This warning is also enabled by @option{-Wextra}; to get the other warnings
2708 of @option{-Wextra} without this warning, use @samp{-Wextra -Wno-sign-compare}.
2710 @item -Waggregate-return
2711 @opindex Waggregate-return
2712 Warn if any functions that return structures or unions are defined or
2713 called. (In languages where you can return an array, this also elicits
2716 @item -Wstrict-prototypes @r{(C only)}
2717 @opindex Wstrict-prototypes
2718 Warn if a function is declared or defined without specifying the
2719 argument types. (An old-style function definition is permitted without
2720 a warning if preceded by a declaration which specifies the argument
2723 @item -Wmissing-prototypes @r{(C only)}
2724 @opindex Wmissing-prototypes
2725 Warn if a global function is defined without a previous prototype
2726 declaration. This warning is issued even if the definition itself
2727 provides a prototype. The aim is to detect global functions that fail
2728 to be declared in header files.
2730 @item -Wmissing-declarations @r{(C only)}
2731 @opindex Wmissing-declarations
2732 Warn if a global function is defined without a previous declaration.
2733 Do so even if the definition itself provides a prototype.
2734 Use this option to detect global functions that are not declared in
2737 @item -Wmissing-noreturn
2738 @opindex Wmissing-noreturn
2739 Warn about functions which might be candidates for attribute @code{noreturn}.
2740 Note these are only possible candidates, not absolute ones. Care should
2741 be taken to manually verify functions actually do not ever return before
2742 adding the @code{noreturn} attribute, otherwise subtle code generation
2743 bugs could be introduced. You will not get a warning for @code{main} in
2744 hosted C environments.
2746 @item -Wmissing-format-attribute
2747 @opindex Wmissing-format-attribute
2749 If @option{-Wformat} is enabled, also warn about functions which might be
2750 candidates for @code{format} attributes. Note these are only possible
2751 candidates, not absolute ones. GCC will guess that @code{format}
2752 attributes might be appropriate for any function that calls a function
2753 like @code{vprintf} or @code{vscanf}, but this might not always be the
2754 case, and some functions for which @code{format} attributes are
2755 appropriate may not be detected. This option has no effect unless
2756 @option{-Wformat} is enabled (possibly by @option{-Wall}).
2758 @item -Wno-multichar
2759 @opindex Wno-multichar
2761 Do not warn if a multicharacter constant (@samp{'FOOF'}) is used.
2762 Usually they indicate a typo in the user's code, as they have
2763 implementation-defined values, and should not be used in portable code.
2765 @item -Wno-deprecated-declarations
2766 @opindex Wno-deprecated-declarations
2767 Do not warn about uses of functions, variables, and types marked as
2768 deprecated by using the @code{deprecated} attribute.
2769 (@pxref{Function Attributes}, @pxref{Variable Attributes},
2770 @pxref{Type Attributes}.)
2774 Warn if a structure is given the packed attribute, but the packed
2775 attribute has no effect on the layout or size of the structure.
2776 Such structures may be mis-aligned for little benefit. For
2777 instance, in this code, the variable @code{f.x} in @code{struct bar}
2778 will be misaligned even though @code{struct bar} does not itself
2779 have the packed attribute:
2786 @} __attribute__((packed));
2796 Warn if padding is included in a structure, either to align an element
2797 of the structure or to align the whole structure. Sometimes when this
2798 happens it is possible to rearrange the fields of the structure to
2799 reduce the padding and so make the structure smaller.
2801 @item -Wredundant-decls
2802 @opindex Wredundant-decls
2803 Warn if anything is declared more than once in the same scope, even in
2804 cases where multiple declaration is valid and changes nothing.
2806 @item -Wnested-externs @r{(C only)}
2807 @opindex Wnested-externs
2808 Warn if an @code{extern} declaration is encountered within a function.
2810 @item -Wunreachable-code
2811 @opindex Wunreachable-code
2812 Warn if the compiler detects that code will never be executed.
2814 This option is intended to warn when the compiler detects that at
2815 least a whole line of source code will never be executed, because
2816 some condition is never satisfied or because it is after a
2817 procedure that never returns.
2819 It is possible for this option to produce a warning even though there
2820 are circumstances under which part of the affected line can be executed,
2821 so care should be taken when removing apparently-unreachable code.
2823 For instance, when a function is inlined, a warning may mean that the
2824 line is unreachable in only one inlined copy of the function.
2826 This option is not made part of @option{-Wall} because in a debugging
2827 version of a program there is often substantial code which checks
2828 correct functioning of the program and is, hopefully, unreachable
2829 because the program does work. Another common use of unreachable
2830 code is to provide behavior which is selectable at compile-time.
2834 Warn if a function can not be inlined and it was declared as inline.
2835 Even with this option, the compiler will not warn about failures to
2836 inline functions declared in system headers.
2838 The compiler uses a variety of heuristics to determine whether or not
2839 to inline a function. For example, the compiler takes into account
2840 the size of the function being inlined and the the amount of inlining
2841 that has already been done in the current function. Therefore,
2842 seemingly insignificant changes in the source program can cause the
2843 warnings produced by @option{-Winline} to appear or disappear.
2845 @item -Wno-invalid-offsetof @r{(C++ only)}
2846 @opindex Wno-invalid-offsetof
2847 Suppress warnings from applying the @samp{offsetof} macro to a non-POD
2848 type. According to the 1998 ISO C++ standard, applying @samp{offsetof}
2849 to a non-POD type is undefined. In existing C++ implementations,
2850 however, @samp{offsetof} typically gives meaningful results even when
2851 applied to certain kinds of non-POD types. (Such as a simple
2852 @samp{struct} that fails to be a POD type only by virtue of having a
2853 constructor.) This flag is for users who are aware that they are
2854 writing nonportable code and who have deliberately chosen to ignore the
2857 The restrictions on @samp{offsetof} may be relaxed in a future version
2858 of the C++ standard.
2861 @opindex Winvalid-pch
2862 Warn if a precompiled header (@pxref{Precompiled Headers}) is found in
2863 the search path but can't be used.
2867 @opindex Wno-long-long
2868 Warn if @samp{long long} type is used. This is default. To inhibit
2869 the warning messages, use @option{-Wno-long-long}. Flags
2870 @option{-Wlong-long} and @option{-Wno-long-long} are taken into account
2871 only when @option{-pedantic} flag is used.
2873 @item -Wdisabled-optimization
2874 @opindex Wdisabled-optimization
2875 Warn if a requested optimization pass is disabled. This warning does
2876 not generally indicate that there is anything wrong with your code; it
2877 merely indicates that GCC's optimizers were unable to handle the code
2878 effectively. Often, the problem is that your code is too big or too
2879 complex; GCC will refuse to optimize programs when the optimization
2880 itself is likely to take inordinate amounts of time.
2884 Make all warnings into errors.
2887 @node Debugging Options
2888 @section Options for Debugging Your Program or GCC
2889 @cindex options, debugging
2890 @cindex debugging information options
2892 GCC has various special options that are used for debugging
2893 either your program or GCC:
2898 Produce debugging information in the operating system's native format
2899 (stabs, COFF, XCOFF, or DWARF)@. GDB can work with this debugging
2902 On most systems that use stabs format, @option{-g} enables use of extra
2903 debugging information that only GDB can use; this extra information
2904 makes debugging work better in GDB but will probably make other debuggers
2906 refuse to read the program. If you want to control for certain whether
2907 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
2908 @option{-gxcoff+}, @option{-gxcoff}, @option{-gdwarf-1+}, @option{-gdwarf-1},
2909 or @option{-gvms} (see below).
2911 Unlike most other C compilers, GCC allows you to use @option{-g} with
2912 @option{-O}. The shortcuts taken by optimized code may occasionally
2913 produce surprising results: some variables you declared may not exist
2914 at all; flow of control may briefly move where you did not expect it;
2915 some statements may not be executed because they compute constant
2916 results or their values were already at hand; some statements may
2917 execute in different places because they were moved out of loops.
2919 Nevertheless it proves possible to debug optimized output. This makes
2920 it reasonable to use the optimizer for programs that might have bugs.
2922 The following options are useful when GCC is generated with the
2923 capability for more than one debugging format.
2927 Produce debugging information for use by GDB@. This means to use the
2928 most expressive format available (DWARF 2, stabs, or the native format
2929 if neither of those are supported), including GDB extensions if at all
2934 Produce debugging information in stabs format (if that is supported),
2935 without GDB extensions. This is the format used by DBX on most BSD
2936 systems. On MIPS, Alpha and System V Release 4 systems this option
2937 produces stabs debugging output which is not understood by DBX or SDB@.
2938 On System V Release 4 systems this option requires the GNU assembler.
2940 @item -feliminate-unused-debug-symbols
2941 @opindex feliminate-unused-debug-symbols
2942 Produce debugging information in stabs format (if that is supported),
2943 for only symbols that are actually used.
2947 Produce debugging information in stabs format (if that is supported),
2948 using GNU extensions understood only by the GNU debugger (GDB)@. The
2949 use of these extensions is likely to make other debuggers crash or
2950 refuse to read the program.
2954 Produce debugging information in COFF format (if that is supported).
2955 This is the format used by SDB on most System V systems prior to
2960 Produce debugging information in XCOFF format (if that is supported).
2961 This is the format used by the DBX debugger on IBM RS/6000 systems.
2965 Produce debugging information in XCOFF format (if that is supported),
2966 using GNU extensions understood only by the GNU debugger (GDB)@. The
2967 use of these extensions is likely to make other debuggers crash or
2968 refuse to read the program, and may cause assemblers other than the GNU
2969 assembler (GAS) to fail with an error.
2973 Produce debugging information in DWARF version 1 format (if that is
2974 supported). This is the format used by SDB on most System V Release 4
2977 This option is deprecated.
2981 Produce debugging information in DWARF version 1 format (if that is
2982 supported), using GNU extensions understood only by the GNU debugger
2983 (GDB)@. The use of these extensions is likely to make other debuggers
2984 crash or refuse to read the program.
2986 This option is deprecated.
2990 Produce debugging information in DWARF version 2 format (if that is
2991 supported). This is the format used by DBX on IRIX 6.
2995 Produce debugging information in VMS debug format (if that is
2996 supported). This is the format used by DEBUG on VMS systems.
2999 @itemx -ggdb@var{level}
3000 @itemx -gstabs@var{level}
3001 @itemx -gcoff@var{level}
3002 @itemx -gxcoff@var{level}
3003 @itemx -gvms@var{level}
3004 Request debugging information and also use @var{level} to specify how
3005 much information. The default level is 2.
3007 Level 1 produces minimal information, enough for making backtraces in
3008 parts of the program that you don't plan to debug. This includes
3009 descriptions of functions and external variables, but no information
3010 about local variables and no line numbers.
3012 Level 3 includes extra information, such as all the macro definitions
3013 present in the program. Some debuggers support macro expansion when
3014 you use @option{-g3}.
3016 Note that in order to avoid confusion between DWARF1 debug level 2,
3017 and DWARF2, neither @option{-gdwarf} nor @option{-gdwarf-2} accept
3018 a concatenated debug level. Instead use an additional @option{-g@var{level}}
3019 option to change the debug level for DWARF1 or DWARF2.
3021 @item -feliminate-dwarf2-dups
3022 @opindex feliminate-dwarf2-dups
3023 Compress DWARF2 debugging information by eliminating duplicated
3024 information about each symbol. This option only makes sense when
3025 generating DWARF2 debugging information with @option{-gdwarf-2}.
3027 @cindex @command{prof}
3030 Generate extra code to write profile information suitable for the
3031 analysis program @command{prof}. You must use this option when compiling
3032 the source files you want data about, and you must also use it when
3035 @cindex @command{gprof}
3038 Generate extra code to write profile information suitable for the
3039 analysis program @command{gprof}. You must use this option when compiling
3040 the source files you want data about, and you must also use it when
3045 Makes the compiler print out each function name as it is compiled, and
3046 print some statistics about each pass when it finishes.
3049 @opindex ftime-report
3050 Makes the compiler print some statistics about the time consumed by each
3051 pass when it finishes.
3054 @opindex fmem-report
3055 Makes the compiler print some statistics about permanent memory
3056 allocation when it finishes.
3058 @item -fprofile-arcs
3059 @opindex fprofile-arcs
3060 Add code so that program flow @dfn{arcs} are instrumented. During
3061 execution the program records how many times each branch and call is
3062 executed and how many times it is taken or returns. When the compiled
3063 program exits it saves this data to a file called
3064 @file{@var{auxname}.gcda} for each source file. The data may be used for
3065 profile-directed optimizations (@option{-fbranch-probabilities}), or for
3066 test coverage analysis (@option{-ftest-coverage}). Each object file's
3067 @var{auxname} is generated from the name of the output file, if
3068 explicitly specified and it is not the final executable, otherwise it is
3069 the basename of the source file. In both cases any suffix is removed
3070 (e.g. @file{foo.gcda} for input file @file{dir/foo.c}, or
3071 @file{dir/foo.gcda} for output file specified as @option{-o dir/foo.o}).
3076 Compile the source files with @option{-fprofile-arcs} plus optimization
3077 and code generation options. For test coverage analysis, use the
3078 additional @option{-ftest-coverage} option. You do not need to profile
3079 every source file in a program.
3082 Link your object files with @option{-lgcov} or @option{-fprofile-arcs}
3083 (the latter implies the former).
3086 Run the program on a representative workload to generate the arc profile
3087 information. This may be repeated any number of times. You can run
3088 concurrent instances of your program, and provided that the file system
3089 supports locking, the data files will be correctly updated. Also
3090 @code{fork} calls are detected and correctly handled (double counting
3094 For profile-directed optimizations, compile the source files again with
3095 the same optimization and code generation options plus
3096 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
3097 Control Optimization}).
3100 For test coverage analysis, use @command{gcov} to produce human readable
3101 information from the @file{.gcno} and @file{.gcda} files. Refer to the
3102 @command{gcov} documentation for further information.
3106 With @option{-fprofile-arcs}, for each function of your program GCC
3107 creates a program flow graph, then finds a spanning tree for the graph.
3108 Only arcs that are not on the spanning tree have to be instrumented: the
3109 compiler adds code to count the number of times that these arcs are
3110 executed. When an arc is the only exit or only entrance to a block, the
3111 instrumentation code can be added to the block; otherwise, a new basic
3112 block must be created to hold the instrumentation code.
3115 @item -ftest-coverage
3116 @opindex ftest-coverage
3117 Produce a notes file that the @command{gcov} code-coverage utility
3118 (@pxref{Gcov,, @command{gcov}---a Test Coverage Program}) can use to
3119 show program coverage. Each source file's note file is called
3120 @file{@var{auxname}.gcno}. Refer to the @option{-fprofile-arcs} option
3121 above for a description of @var{auxname} and instructions on how to
3122 generate test coverage data. Coverage data will match the source files
3123 more closely, if you do not optimize.
3125 @item -d@var{letters}
3127 Says to make debugging dumps during compilation at times specified by
3128 @var{letters}. This is used for debugging the compiler. The file names
3129 for most of the dumps are made by appending a pass number and a word to
3130 the @var{dumpname}. @var{dumpname} is generated from the name of the
3131 output file, if explicitly specified and it is not an executable,
3132 otherwise it is the basename of the source file. In both cases any
3133 suffix is removed (e.g. @file{foo.00.rtl} or @file{foo.01.sibling}).
3134 Here are the possible letters for use in @var{letters}, and their
3140 Annotate the assembler output with miscellaneous debugging information.
3143 Dump after computing branch probabilities, to @file{@var{file}.16.bp}.
3146 Dump after block reordering, to @file{@var{file}.32.bbro}.
3149 Dump after instruction combination, to the file @file{@var{file}.22.combine}.
3152 Dump after the first if conversion, to the file @file{@var{file}.17.ce1}.
3153 Also dump after the second if conversion, to the file @file{@var{file}.23.ce2}.
3156 Dump after branch target load optimization, to to @file{@var{file}.34.btl}.
3157 Also dump after delayed branch scheduling, to @file{@var{file}.37.dbr}.
3160 Dump all macro definitions, at the end of preprocessing, in addition to
3164 Dump after SSA optimizations, to @file{@var{file}.05.ssa} and
3165 @file{@var{file}.010.ussa}.
3168 Dump after the second if conversion, to @file{@var{file}.33.ce3}.
3171 Dump after control and data flow analysis, to @file{@var{file}.15.cfg}.
3172 Also dump after life analysis, to @file{@var{file}.21.life}.
3175 Dump after purging @code{ADDRESSOF} codes, to @file{@var{file}.11.addressof}.
3178 Dump after global register allocation, to @file{@var{file}.27.greg}.
3181 Dump after GCSE, to @file{@var{file}.12.gcse}.
3182 Also dump after jump bypassing and control flow optimizations, to
3183 @file{@var{file}.14.bypass}.
3186 Dump after finalization of EH handling code, to @file{@var{file}.03.eh}.
3189 Dump after sibling call optimizations, to @file{@var{file}.02.sibling}.
3192 Dump after the first jump optimization, to @file{@var{file}.04.jump}.
3195 Dump after conversion from registers to stack, to @file{@var{file}.36.stack}.
3198 Dump after local register allocation, to @file{@var{file}.26.lreg}.
3201 Dump after loop optimization passes, to @file{@var{file}.13.loop} and
3202 @file{@var{file}.19.loop2}.
3205 Dump after performing the machine dependent reorganization pass, to
3206 @file{@var{file}.37.mach}.
3209 Dump after register renumbering, to @file{@var{file}.31.rnreg}.
3212 Dump after the register move pass, to @file{@var{file}.24.regmove}.
3215 Dump after post-reload optimizations, to @file{@var{file}.28.postreload}.
3218 Dump after RTL generation, to @file{@var{file}.01.rtl}.
3221 Dump after the second scheduling pass, to @file{@var{file}.35.sched2}.
3224 Dump after CSE (including the jump optimization that sometimes follows
3225 CSE), to @file{@var{file}.019.cse}.
3228 Dump after the first scheduling pass, to @file{@var{file}.25.sched}.
3231 Dump after the second CSE pass (including the jump optimization that
3232 sometimes follows CSE), to @file{@var{file}.20.cse2}.
3235 Dump after running tracer, to @file{@var{file}.18.tracer}.
3238 Dump after null pointer elimination pass to @file{@var{file}.018.null}.
3241 Dump callgraph and unit-at-a-time optimization @file{@var{file}.00.unit}.
3244 Dump after the second flow pass, to @file{@var{file}.29.flow2}.
3247 Dump after SSA conditional constant propagation, to
3248 @file{@var{file}.06.ssaccp}.
3251 Dump after SSA dead code elimination, to @file{@var{file}.07.ssadce}.
3254 Dump after the peephole pass, to @file{@var{file}.30.peephole2}.
3257 Produce all the dumps listed above.
3260 Produce a core dump whenever an error occurs.
3263 Print statistics on memory usage, at the end of the run, to
3267 Annotate the assembler output with a comment indicating which
3268 pattern and alternative was used. The length of each instruction is
3272 Dump the RTL in the assembler output as a comment before each instruction.
3273 Also turns on @option{-dp} annotation.
3276 For each of the other indicated dump files (except for
3277 @file{@var{file}.01.rtl}), dump a representation of the control flow graph
3278 suitable for viewing with VCG to @file{@var{file}.@var{pass}.vcg}.
3281 Just generate RTL for a function instead of compiling it. Usually used
3285 Dump debugging information during parsing, to standard error.
3288 @item -fdump-unnumbered
3289 @opindex fdump-unnumbered
3290 When doing debugging dumps (see @option{-d} option above), suppress instruction
3291 numbers and line number note output. This makes it more feasible to
3292 use diff on debugging dumps for compiler invocations with different
3293 options, in particular with and without @option{-g}.
3295 @item -fdump-translation-unit @r{(C and C++ only)}
3296 @itemx -fdump-translation-unit-@var{options} @r{(C and C++ only)}
3297 @opindex fdump-translation-unit
3298 Dump a representation of the tree structure for the entire translation
3299 unit to a file. The file name is made by appending @file{.tu} to the
3300 source file name. If the @samp{-@var{options}} form is used, @var{options}
3301 controls the details of the dump as described for the
3302 @option{-fdump-tree} options.
3304 @item -fdump-class-hierarchy @r{(C++ only)}
3305 @itemx -fdump-class-hierarchy-@var{options} @r{(C++ only)}
3306 @opindex fdump-class-hierarchy
3307 Dump a representation of each class's hierarchy and virtual function
3308 table layout to a file. The file name is made by appending @file{.class}
3309 to the source file name. If the @samp{-@var{options}} form is used,
3310 @var{options} controls the details of the dump as described for the
3311 @option{-fdump-tree} options.
3313 @item -fdump-tree-@var{switch} @r{(C++ only)}
3314 @itemx -fdump-tree-@var{switch}-@var{options} @r{(C++ only)}
3316 Control the dumping at various stages of processing the intermediate
3317 language tree to a file. The file name is generated by appending a switch
3318 specific suffix to the source file name. If the @samp{-@var{options}}
3319 form is used, @var{options} is a list of @samp{-} separated options that
3320 control the details of the dump. Not all options are applicable to all
3321 dumps, those which are not meaningful will be ignored. The following
3322 options are available
3326 Print the address of each node. Usually this is not meaningful as it
3327 changes according to the environment and source file. Its primary use
3328 is for tying up a dump file with a debug environment.
3330 Inhibit dumping of members of a scope or body of a function merely
3331 because that scope has been reached. Only dump such items when they
3332 are directly reachable by some other path.
3334 Turn on all options.
3337 The following tree dumps are possible:
3340 Dump before any tree based optimization, to @file{@var{file}.original}.
3342 Dump after all tree based optimization, to @file{@var{file}.optimized}.
3344 Dump after function inlining, to @file{@var{file}.inlined}.
3347 @item -frandom-seed=@var{string}
3348 @opindex frandom-string
3349 This option provides a seed that GCC uses when it would otherwise use
3350 random numbers. It is used to generate certain symbol names
3351 that have to be different in every compiled file. It is also used to
3352 place unique stamps in coverage data files and the object files that
3353 produce them. You can use the @option{-frandom-seed} option to produce
3354 reproducibly identical object files.
3356 The @var{string} should be different for every file you compile.
3358 @item -fsched-verbose=@var{n}
3359 @opindex fsched-verbose
3360 On targets that use instruction scheduling, this option controls the
3361 amount of debugging output the scheduler prints. This information is
3362 written to standard error, unless @option{-dS} or @option{-dR} is
3363 specified, in which case it is output to the usual dump
3364 listing file, @file{.sched} or @file{.sched2} respectively. However
3365 for @var{n} greater than nine, the output is always printed to standard
3368 For @var{n} greater than zero, @option{-fsched-verbose} outputs the
3369 same information as @option{-dRS}. For @var{n} greater than one, it
3370 also output basic block probabilities, detailed ready list information
3371 and unit/insn info. For @var{n} greater than two, it includes RTL
3372 at abort point, control-flow and regions info. And for @var{n} over
3373 four, @option{-fsched-verbose} also includes dependence info.
3377 Store the usual ``temporary'' intermediate files permanently; place them
3378 in the current directory and name them based on the source file. Thus,
3379 compiling @file{foo.c} with @samp{-c -save-temps} would produce files
3380 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
3381 preprocessed @file{foo.i} output file even though the compiler now
3382 normally uses an integrated preprocessor.
3386 Report the CPU time taken by each subprocess in the compilation
3387 sequence. For C source files, this is the compiler proper and assembler
3388 (plus the linker if linking is done). The output looks like this:
3395 The first number on each line is the ``user time,'' that is time spent
3396 executing the program itself. The second number is ``system time,''
3397 time spent executing operating system routines on behalf of the program.
3398 Both numbers are in seconds.
3400 @item -print-file-name=@var{library}
3401 @opindex print-file-name
3402 Print the full absolute name of the library file @var{library} that
3403 would be used when linking---and don't do anything else. With this
3404 option, GCC does not compile or link anything; it just prints the
3407 @item -print-multi-directory
3408 @opindex print-multi-directory
3409 Print the directory name corresponding to the multilib selected by any
3410 other switches present in the command line. This directory is supposed
3411 to exist in @env{GCC_EXEC_PREFIX}.
3413 @item -print-multi-lib
3414 @opindex print-multi-lib
3415 Print the mapping from multilib directory names to compiler switches
3416 that enable them. The directory name is separated from the switches by
3417 @samp{;}, and each switch starts with an @samp{@@} instead of the
3418 @samp{-}, without spaces between multiple switches. This is supposed to
3419 ease shell-processing.
3421 @item -print-prog-name=@var{program}
3422 @opindex print-prog-name
3423 Like @option{-print-file-name}, but searches for a program such as @samp{cpp}.
3425 @item -print-libgcc-file-name
3426 @opindex print-libgcc-file-name
3427 Same as @option{-print-file-name=libgcc.a}.
3429 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
3430 but you do want to link with @file{libgcc.a}. You can do
3433 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
3436 @item -print-search-dirs
3437 @opindex print-search-dirs
3438 Print the name of the configured installation directory and a list of
3439 program and library directories gcc will search---and don't do anything else.
3441 This is useful when gcc prints the error message
3442 @samp{installation problem, cannot exec cpp0: No such file or directory}.
3443 To resolve this you either need to put @file{cpp0} and the other compiler
3444 components where gcc expects to find them, or you can set the environment
3445 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
3446 Don't forget the trailing '/'.
3447 @xref{Environment Variables}.
3450 @opindex dumpmachine
3451 Print the compiler's target machine (for example,
3452 @samp{i686-pc-linux-gnu})---and don't do anything else.
3455 @opindex dumpversion
3456 Print the compiler version (for example, @samp{3.0})---and don't do
3461 Print the compiler's built-in specs---and don't do anything else. (This
3462 is used when GCC itself is being built.) @xref{Spec Files}.
3464 @item -feliminate-unused-debug-types
3465 @opindex feliminate-unused-debug-types
3466 Normally, when producing DWARF2 output, GCC will emit debugging
3467 information for all types declared in a compilation
3468 unit, regardless of whether or not they are actually used
3469 in that compilation unit. Sometimes this is useful, such as
3470 if, in the debugger, you want to cast a value to a type that is
3471 not actually used in your program (but is declared). More often,
3472 however, this results in a significant amount of wasted space.
3473 With this option, GCC will avoid producing debug symbol output
3474 for types that are nowhere used in the source file being compiled.
3477 @node Optimize Options
3478 @section Options That Control Optimization
3479 @cindex optimize options
3480 @cindex options, optimization
3482 These options control various sorts of optimizations.
3484 Without any optimization option, the compiler's goal is to reduce the
3485 cost of compilation and to make debugging produce the expected
3486 results. Statements are independent: if you stop the program with a
3487 breakpoint between statements, you can then assign a new value to any
3488 variable or change the program counter to any other statement in the
3489 function and get exactly the results you would expect from the source
3492 Turning on optimization flags makes the compiler attempt to improve
3493 the performance and/or code size at the expense of compilation time
3494 and possibly the ability to debug the program.
3496 The compiler performs optimisation based on the knowledge it has of
3497 the program. Using the @option{-funit-at-a-time} flag will allow the
3498 compiler to consider information gained from later functions in the
3499 file when compiling a function. Compiling multiple files at once to a
3500 single output file (and using @option{-funit-at-a-time}) will allow
3501 the compiler to use information gained from all of the files when
3502 compiling each of them.
3504 Not all optimizations are controlled directly by a flag. Only
3505 optimizations that have a flag are listed.
3512 Optimize. Optimizing compilation takes somewhat more time, and a lot
3513 more memory for a large function.
3515 With @option{-O}, the compiler tries to reduce code size and execution
3516 time, without performing any optimizations that take a great deal of
3519 @option{-O} turns on the following optimization flags:
3520 @gccoptlist{-fdefer-pop @gol
3521 -fmerge-constants @gol
3523 -floop-optimize @gol
3525 -fif-conversion @gol
3526 -fif-conversion2 @gol
3527 -fdelayed-branch @gol
3528 -fguess-branch-probability @gol
3531 @option{-O} also turns on @option{-fomit-frame-pointer} on machines
3532 where doing so does not interfere with debugging.
3536 Optimize even more. GCC performs nearly all supported optimizations
3537 that do not involve a space-speed tradeoff. The compiler does not
3538 perform loop unrolling or function inlining when you specify @option{-O2}.
3539 As compared to @option{-O}, this option increases both compilation time
3540 and the performance of the generated code.
3542 @option{-O2} turns on all optimization flags specified by @option{-O}. It
3543 also turns on the following optimization flags:
3544 @gccoptlist{-fforce-mem @gol
3545 -foptimize-sibling-calls @gol
3546 -fstrength-reduce @gol
3547 -fcse-follow-jumps -fcse-skip-blocks @gol
3548 -frerun-cse-after-loop -frerun-loop-opt @gol
3549 -fgcse -fgcse-lm -fgcse-sm @gol
3550 -fdelete-null-pointer-checks @gol
3551 -fexpensive-optimizations @gol
3553 -fschedule-insns -fschedule-insns2 @gol
3554 -fsched-interblock -fsched-spec @gol
3557 -freorder-blocks -freorder-functions @gol
3558 -fstrict-aliasing @gol
3559 -falign-functions -falign-jumps @gol
3560 -falign-loops -falign-labels}
3562 Please note the warning under @option{-fgcse} about
3563 invoking @option{-O2} on programs that use computed gotos.
3567 Optimize yet more. @option{-O3} turns on all optimizations specified by
3568 @option{-O2} and also turns on the @option{-finline-functions},
3569 @option{-funit-at-a-time} and @option{-frename-registers} options.
3573 Do not optimize. This is the default.
3577 Optimize for size. @option{-Os} enables all @option{-O2} optimizations that
3578 do not typically increase code size. It also performs further
3579 optimizations designed to reduce code size.
3581 @option{-Os} disables the following optimization flags:
3582 @gccoptlist{-falign-functions -falign-jumps -falign-loops @gol
3583 -falign-labels -freorder-blocks -fprefetch-loop-arrays}
3585 If you use multiple @option{-O} options, with or without level numbers,
3586 the last such option is the one that is effective.
3589 Options of the form @option{-f@var{flag}} specify machine-independent
3590 flags. Most flags have both positive and negative forms; the negative
3591 form of @option{-ffoo} would be @option{-fno-foo}. In the table
3592 below, only one of the forms is listed---the one you typically will
3593 use. You can figure out the other form by either removing @samp{no-}
3596 The following options control specific optimizations. They are either
3597 activated by @option{-O} options or are related to ones that are. You
3598 can use the following flags in the rare cases when ``fine-tuning'' of
3599 optimizations to be performed is desired.
3602 @item -fno-default-inline
3603 @opindex fno-default-inline
3604 Do not make member functions inline by default merely because they are
3605 defined inside the class scope (C++ only). Otherwise, when you specify
3606 @w{@option{-O}}, member functions defined inside class scope are compiled
3607 inline by default; i.e., you don't need to add @samp{inline} in front of
3608 the member function name.
3610 @item -fno-defer-pop
3611 @opindex fno-defer-pop
3612 Always pop the arguments to each function call as soon as that function
3613 returns. For machines which must pop arguments after a function call,
3614 the compiler normally lets arguments accumulate on the stack for several
3615 function calls and pops them all at once.
3617 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
3621 Force memory operands to be copied into registers before doing
3622 arithmetic on them. This produces better code by making all memory
3623 references potential common subexpressions. When they are not common
3624 subexpressions, instruction combination should eliminate the separate
3627 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3630 @opindex fforce-addr
3631 Force memory address constants to be copied into registers before
3632 doing arithmetic on them. This may produce better code just as
3633 @option{-fforce-mem} may.
3635 @item -fomit-frame-pointer
3636 @opindex fomit-frame-pointer
3637 Don't keep the frame pointer in a register for functions that
3638 don't need one. This avoids the instructions to save, set up and
3639 restore frame pointers; it also makes an extra register available
3640 in many functions. @strong{It also makes debugging impossible on
3643 On some machines, such as the VAX, this flag has no effect, because
3644 the standard calling sequence automatically handles the frame pointer
3645 and nothing is saved by pretending it doesn't exist. The
3646 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
3647 whether a target machine supports this flag. @xref{Registers,,Register
3648 Usage, gccint, GNU Compiler Collection (GCC) Internals}.
3650 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
3652 @item -foptimize-sibling-calls
3653 @opindex foptimize-sibling-calls
3654 Optimize sibling and tail recursive calls.
3656 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3660 Don't pay attention to the @code{inline} keyword. Normally this option
3661 is used to keep the compiler from expanding any functions inline.
3662 Note that if you are not optimizing, no functions can be expanded inline.
3664 @item -finline-functions
3665 @opindex finline-functions
3666 Integrate all simple functions into their callers. The compiler
3667 heuristically decides which functions are simple enough to be worth
3668 integrating in this way.
3670 If all calls to a given function are integrated, and the function is
3671 declared @code{static}, then the function is normally not output as
3672 assembler code in its own right.
3674 Enabled at level @option{-O3}.
3676 @item -finline-limit=@var{n}
3677 @opindex finline-limit
3678 By default, gcc limits the size of functions that can be inlined. This flag
3679 allows the control of this limit for functions that are explicitly marked as
3680 inline (i.e., marked with the inline keyword or defined within the class
3681 definition in c++). @var{n} is the size of functions that can be inlined in
3682 number of pseudo instructions (not counting parameter handling). The default
3683 value of @var{n} is 600.
3684 Increasing this value can result in more inlined code at
3685 the cost of compilation time and memory consumption. Decreasing usually makes
3686 the compilation faster and less code will be inlined (which presumably
3687 means slower programs). This option is particularly useful for programs that
3688 use inlining heavily such as those based on recursive templates with C++.
3690 Inlining is actually controlled by a number of parameters, which may be
3691 specified individually by using @option{--param @var{name}=@var{value}}.
3692 The @option{-finline-limit=@var{n}} option sets some of these parameters
3696 @item max-inline-insns
3698 @item max-inline-insns-single
3699 is set to @var{n}/2.
3700 @item max-inline-insns-auto
3701 is set to @var{n}/2.
3702 @item min-inline-insns
3703 is set to 130 or @var{n}/4, whichever is smaller.
3704 @item max-inline-insns-rtl
3708 Using @option{-finline-limit=600} thus results in the default settings
3709 for these parameters. See below for a documentation of the individual
3710 parameters controlling inlining.
3712 @emph{Note:} pseudo instruction represents, in this particular context, an
3713 abstract measurement of function's size. In no way, it represents a count
3714 of assembly instructions and as such its exact meaning might change from one
3715 release to an another.
3717 @item -fkeep-inline-functions
3718 @opindex fkeep-inline-functions
3719 Even if all calls to a given function are integrated, and the function
3720 is declared @code{static}, nevertheless output a separate run-time
3721 callable version of the function. This switch does not affect
3722 @code{extern inline} functions.
3724 @item -fkeep-static-consts
3725 @opindex fkeep-static-consts
3726 Emit variables declared @code{static const} when optimization isn't turned
3727 on, even if the variables aren't referenced.
3729 GCC enables this option by default. If you want to force the compiler to
3730 check if the variable was referenced, regardless of whether or not
3731 optimization is turned on, use the @option{-fno-keep-static-consts} option.
3733 @item -fmerge-constants
3734 Attempt to merge identical constants (string constants and floating point
3735 constants) across compilation units.
3737 This option is the default for optimized compilation if the assembler and
3738 linker support it. Use @option{-fno-merge-constants} to inhibit this
3741 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
3743 @item -fmerge-all-constants
3744 Attempt to merge identical constants and identical variables.
3746 This option implies @option{-fmerge-constants}. In addition to
3747 @option{-fmerge-constants} this considers e.g. even constant initialized
3748 arrays or initialized constant variables with integral or floating point
3749 types. Languages like C or C++ require each non-automatic variable to
3750 have distinct location, so using this option will result in non-conforming
3755 Use a graph coloring register allocator. Currently this option is meant
3756 for testing, so we are interested to hear about miscompilations with
3759 @item -fno-branch-count-reg
3760 @opindex fno-branch-count-reg
3761 Do not use ``decrement and branch'' instructions on a count register,
3762 but instead generate a sequence of instructions that decrement a
3763 register, compare it against zero, then branch based upon the result.
3764 This option is only meaningful on architectures that support such
3765 instructions, which include x86, PowerPC, IA-64 and S/390.
3767 The default is @option{-fbranch-count-reg}, enabled when
3768 @option{-fstrength-reduce} is enabled.
3770 @item -fno-function-cse
3771 @opindex fno-function-cse
3772 Do not put function addresses in registers; make each instruction that
3773 calls a constant function contain the function's address explicitly.
3775 This option results in less efficient code, but some strange hacks
3776 that alter the assembler output may be confused by the optimizations
3777 performed when this option is not used.
3779 The default is @option{-ffunction-cse}
3781 @item -fno-zero-initialized-in-bss
3782 @opindex fno-zero-initialized-in-bss
3783 If the target supports a BSS section, GCC by default puts variables that
3784 are initialized to zero into BSS@. This can save space in the resulting
3787 This option turns off this behavior because some programs explicitly
3788 rely on variables going to the data section. E.g., so that the
3789 resulting executable can find the beginning of that section and/or make
3790 assumptions based on that.
3792 The default is @option{-fzero-initialized-in-bss}.
3794 @item -fstrength-reduce
3795 @opindex fstrength-reduce
3796 Perform the optimizations of loop strength reduction and
3797 elimination of iteration variables.
3799 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3801 @item -fthread-jumps
3802 @opindex fthread-jumps
3803 Perform optimizations where we check to see if a jump branches to a
3804 location where another comparison subsumed by the first is found. If
3805 so, the first branch is redirected to either the destination of the
3806 second branch or a point immediately following it, depending on whether
3807 the condition is known to be true or false.
3809 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
3811 @item -fcse-follow-jumps
3812 @opindex fcse-follow-jumps
3813 In common subexpression elimination, scan through jump instructions
3814 when the target of the jump is not reached by any other path. For
3815 example, when CSE encounters an @code{if} statement with an
3816 @code{else} clause, CSE will follow the jump when the condition
3819 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3821 @item -fcse-skip-blocks
3822 @opindex fcse-skip-blocks
3823 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
3824 follow jumps which conditionally skip over blocks. When CSE
3825 encounters a simple @code{if} statement with no else clause,
3826 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
3827 body of the @code{if}.
3829 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3831 @item -frerun-cse-after-loop
3832 @opindex frerun-cse-after-loop
3833 Re-run common subexpression elimination after loop optimizations has been
3836 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3838 @item -frerun-loop-opt
3839 @opindex frerun-loop-opt
3840 Run the loop optimizer twice.
3842 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3846 Perform a global common subexpression elimination pass.
3847 This pass also performs global constant and copy propagation.
3849 @emph{Note:} When compiling a program using computed gotos, a GCC
3850 extension, you may get better runtime performance if you disable
3851 the global common subexpression elimination pass by adding
3852 @option{-fno-gcse} to the command line.
3854 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3858 When @option{-fgcse-lm} is enabled, global common subexpression elimination will
3859 attempt to move loads which are only killed by stores into themselves. This
3860 allows a loop containing a load/store sequence to be changed to a load outside
3861 the loop, and a copy/store within the loop.
3863 Enabled by default when gcse is enabled.
3867 When @option{-fgcse-sm} is enabled, A store motion pass is run after global common
3868 subexpression elimination. This pass will attempt to move stores out of loops.
3869 When used in conjunction with @option{-fgcse-lm}, loops containing a load/store sequence
3870 can be changed to a load before the loop and a store after the loop.
3872 Enabled by default when gcse is enabled.
3874 @item -floop-optimize
3875 @opindex floop-optimize
3876 Perform loop optimizations: move constant expressions out of loops, simplify
3877 exit test conditions and optionally do strength-reduction and loop unrolling as
3880 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
3882 @item -fcrossjumping
3883 @opindex crossjumping
3884 Perform cross-jumping transformation. This transformation unifies equivalent code and save code size. The
3885 resulting code may or may not perform better than without cross-jumping.
3887 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
3889 @item -fif-conversion
3890 @opindex if-conversion
3891 Attempt to transform conditional jumps into branch-less equivalents. This
3892 include use of conditional moves, min, max, set flags and abs instructions, and
3893 some tricks doable by standard arithmetics. The use of conditional execution
3894 on chips where it is available is controlled by @code{if-conversion2}.
3896 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
3898 @item -fif-conversion2
3899 @opindex if-conversion2
3900 Use conditional execution (where available) to transform conditional jumps into
3901 branch-less equivalents.
3903 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
3905 @item -fdelete-null-pointer-checks
3906 @opindex fdelete-null-pointer-checks
3907 Use global dataflow analysis to identify and eliminate useless checks
3908 for null pointers. The compiler assumes that dereferencing a null
3909 pointer would have halted the program. If a pointer is checked after
3910 it has already been dereferenced, it cannot be null.
3912 In some environments, this assumption is not true, and programs can
3913 safely dereference null pointers. Use
3914 @option{-fno-delete-null-pointer-checks} to disable this optimization
3915 for programs which depend on that behavior.
3917 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3919 @item -fexpensive-optimizations
3920 @opindex fexpensive-optimizations
3921 Perform a number of minor optimizations that are relatively expensive.
3923 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3925 @item -foptimize-register-move
3927 @opindex foptimize-register-move
3929 Attempt to reassign register numbers in move instructions and as
3930 operands of other simple instructions in order to maximize the amount of
3931 register tying. This is especially helpful on machines with two-operand
3934 Note @option{-fregmove} and @option{-foptimize-register-move} are the same
3937 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3939 @item -fdelayed-branch
3940 @opindex fdelayed-branch
3941 If supported for the target machine, attempt to reorder instructions
3942 to exploit instruction slots available after delayed branch
3945 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
3947 @item -fschedule-insns
3948 @opindex fschedule-insns
3949 If supported for the target machine, attempt to reorder instructions to
3950 eliminate execution stalls due to required data being unavailable. This
3951 helps machines that have slow floating point or memory load instructions
3952 by allowing other instructions to be issued until the result of the load
3953 or floating point instruction is required.
3955 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3957 @item -fschedule-insns2
3958 @opindex fschedule-insns2
3959 Similar to @option{-fschedule-insns}, but requests an additional pass of
3960 instruction scheduling after register allocation has been done. This is
3961 especially useful on machines with a relatively small number of
3962 registers and where memory load instructions take more than one cycle.
3964 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
3966 @item -fno-sched-interblock
3967 @opindex fno-sched-interblock
3968 Don't schedule instructions across basic blocks. This is normally
3969 enabled by default when scheduling before register allocation, i.e.@:
3970 with @option{-fschedule-insns} or at @option{-O2} or higher.
3972 @item -fno-sched-spec
3973 @opindex fno-sched-spec
3974 Don't allow speculative motion of non-load instructions. This is normally
3975 enabled by default when scheduling before register allocation, i.e.@:
3976 with @option{-fschedule-insns} or at @option{-O2} or higher.
3978 @item -fsched-spec-load
3979 @opindex fsched-spec-load
3980 Allow speculative motion of some load instructions. This only makes
3981 sense when scheduling before register allocation, i.e.@: with
3982 @option{-fschedule-insns} or at @option{-O2} or higher.
3984 @item -fsched-spec-load-dangerous
3985 @opindex fsched-spec-load-dangerous
3986 Allow speculative motion of more load instructions. This only makes
3987 sense when scheduling before register allocation, i.e.@: with
3988 @option{-fschedule-insns} or at @option{-O2} or higher.
3990 @item -fsched2-use-superblocks
3991 @opindex fsched2-use-superblocks
3992 When scheduling after register allocation, do use superblock scheduling
3993 algorithm. Superblock scheduling allows motion across basic block boundaries
3994 resulting on faster schedules. This option is experimental, as not all machine
3995 descriptions used by GCC model the CPU closely enough to avoid unreliable
3996 results from the algorithm.
3998 This only makes sense when scheduling after register allocation, i.e.@: with
3999 @option{-fschedule-insns2} or at @option{-O2} or higher.
4001 @item -fsched2-use-traces
4002 @opindex fsched2-use-traces
4003 Use @option{-fsched2-use-superblocks} algorithm when scheduling after register
4004 allocation and additionally perform code duplication in order to increase the
4005 size of superblocks using tracer pass. See @option{-ftracer} for details on
4008 This mode should produce faster but significantly longer programs. Also
4009 without @code{-fbranch-probabilities} the traces constructed may not match the
4010 reality and hurt the performance. This only makes
4011 sense when scheduling after register allocation, i.e.@: with
4012 @option{-fschedule-insns2} or at @option{-O2} or higher.
4014 @item -fcaller-saves
4015 @opindex fcaller-saves
4016 Enable values to be allocated in registers that will be clobbered by
4017 function calls, by emitting extra instructions to save and restore the
4018 registers around such calls. Such allocation is done only when it
4019 seems to result in better code than would otherwise be produced.
4021 This option is always enabled by default on certain machines, usually
4022 those which have no call-preserved registers to use instead.
4024 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4026 @item -fmove-all-movables
4027 @opindex fmove-all-movables
4028 Forces all invariant computations in loops to be moved
4031 @item -freduce-all-givs
4032 @opindex freduce-all-givs
4033 Forces all general-induction variables in loops to be
4036 @emph{Note:} When compiling programs written in Fortran,
4037 @option{-fmove-all-movables} and @option{-freduce-all-givs} are enabled
4038 by default when you use the optimizer.
4040 These options may generate better or worse code; results are highly
4041 dependent on the structure of loops within the source code.
4043 These two options are intended to be removed someday, once
4044 they have helped determine the efficacy of various
4045 approaches to improving loop optimizations.
4047 Please let us (@w{@email{gcc@@gcc.gnu.org}} and @w{@email{fortran@@gnu.org}})
4048 know how use of these options affects
4049 the performance of your production code.
4050 We're very interested in code that runs @emph{slower}
4051 when these options are @emph{enabled}.
4054 @itemx -fno-peephole2
4055 @opindex fno-peephole
4056 @opindex fno-peephole2
4057 Disable any machine-specific peephole optimizations. The difference
4058 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
4059 are implemented in the compiler; some targets use one, some use the
4060 other, a few use both.
4062 @option{-fpeephole} is enabled by default.
4063 @option{-fpeephole2} enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4065 @item -fno-guess-branch-probability
4066 @opindex fno-guess-branch-probability
4067 Do not guess branch probabilities using a randomized model.
4069 Sometimes gcc will opt to use a randomized model to guess branch
4070 probabilities, when none are available from either profiling feedback
4071 (@option{-fprofile-arcs}) or @samp{__builtin_expect}. This means that
4072 different runs of the compiler on the same program may produce different
4075 In a hard real-time system, people don't want different runs of the
4076 compiler to produce code that has different behavior; minimizing
4077 non-determinism is of paramount import. This switch allows users to
4078 reduce non-determinism, possibly at the expense of inferior
4081 The default is @option{-fguess-branch-probability} at levels
4082 @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4084 @item -freorder-blocks
4085 @opindex freorder-blocks
4086 Reorder basic blocks in the compiled function in order to reduce number of
4087 taken branches and improve code locality.
4089 Enabled at levels @option{-O2}, @option{-O3}.
4091 @item -freorder-functions
4092 @opindex freorder-functions
4093 Reorder basic blocks in the compiled function in order to reduce number of
4094 taken branches and improve code locality. This is implemented by using special
4095 subsections @code{text.hot} for most frequently executed functions and
4096 @code{text.unlikely} for unlikely executed functions. Reordering is done by
4097 the linker so object file format must support named sections and linker must
4098 place them in a reasonable way.
4100 Also profile feedback must be available in to make this option effective. See
4101 @option{-fprofile-arcs} for details.
4103 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4105 @item -fstrict-aliasing
4106 @opindex fstrict-aliasing
4107 Allows the compiler to assume the strictest aliasing rules applicable to
4108 the language being compiled. For C (and C++), this activates
4109 optimizations based on the type of expressions. In particular, an
4110 object of one type is assumed never to reside at the same address as an
4111 object of a different type, unless the types are almost the same. For
4112 example, an @code{unsigned int} can alias an @code{int}, but not a
4113 @code{void*} or a @code{double}. A character type may alias any other
4116 Pay special attention to code like this:
4129 The practice of reading from a different union member than the one most
4130 recently written to (called ``type-punning'') is common. Even with
4131 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
4132 is accessed through the union type. So, the code above will work as
4133 expected. However, this code might not:
4144 Every language that wishes to perform language-specific alias analysis
4145 should define a function that computes, given an @code{tree}
4146 node, an alias set for the node. Nodes in different alias sets are not
4147 allowed to alias. For an example, see the C front-end function
4148 @code{c_get_alias_set}.
4150 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
4152 @item -falign-functions
4153 @itemx -falign-functions=@var{n}
4154 @opindex falign-functions
4155 Align the start of functions to the next power-of-two greater than
4156 @var{n}, skipping up to @var{n} bytes. For instance,
4157 @option{-falign-functions=32} aligns functions to the next 32-byte
4158 boundary, but @option{-falign-functions=24} would align to the next
4159 32-byte boundary only if this can be done by skipping 23 bytes or less.
4161 @option{-fno-align-functions} and @option{-falign-functions=1} are
4162 equivalent and mean that functions will not be aligned.
4164 Some assemblers only support this flag when @var{n} is a power of two;
4165 in that case, it is rounded up.
4167 If @var{n} is not specified or is zero, use a machine-dependent default.
4169 Enabled at levels @option{-O2}, @option{-O3}.
4171 @item -falign-labels
4172 @itemx -falign-labels=@var{n}
4173 @opindex falign-labels
4174 Align all branch targets to a power-of-two boundary, skipping up to
4175 @var{n} bytes like @option{-falign-functions}. This option can easily
4176 make code slower, because it must insert dummy operations for when the
4177 branch target is reached in the usual flow of the code.
4179 @option{-fno-align-labels} and @option{-falign-labels=1} are
4180 equivalent and mean that labels will not be aligned.
4182 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
4183 are greater than this value, then their values are used instead.
4185 If @var{n} is not specified or is zero, use a machine-dependent default
4186 which is very likely to be @samp{1}, meaning no alignment.
4188 Enabled at levels @option{-O2}, @option{-O3}.
4191 @itemx -falign-loops=@var{n}
4192 @opindex falign-loops
4193 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
4194 like @option{-falign-functions}. The hope is that the loop will be
4195 executed many times, which will make up for any execution of the dummy
4198 @option{-fno-align-loops} and @option{-falign-loops=1} are
4199 equivalent and mean that loops will not be aligned.
4201 If @var{n} is not specified or is zero, use a machine-dependent default.
4203 Enabled at levels @option{-O2}, @option{-O3}.
4206 @itemx -falign-jumps=@var{n}
4207 @opindex falign-jumps
4208 Align branch targets to a power-of-two boundary, for branch targets
4209 where the targets can only be reached by jumping, skipping up to @var{n}
4210 bytes like @option{-falign-functions}. In this case, no dummy operations
4213 @option{-fno-align-jumps} and @option{-falign-jumps=1} are
4214 equivalent and mean that loops will not be aligned.
4216 If @var{n} is not specified or is zero, use a machine-dependent default.
4218 Enabled at levels @option{-O2}, @option{-O3}.
4220 @item -frename-registers
4221 @opindex frename-registers
4222 Attempt to avoid false dependencies in scheduled code by making use
4223 of registers left over after register allocation. This optimization
4224 will most benefit processors with lots of registers. It can, however,
4225 make debugging impossible, since variables will no longer stay in
4226 a ``home register''.
4228 Enabled at levels @option{-O3}.
4230 @item -fno-cprop-registers
4231 @opindex fno-cprop-registers
4232 After register allocation and post-register allocation instruction splitting,
4233 we perform a copy-propagation pass to try to reduce scheduling dependencies
4234 and occasionally eliminate the copy.
4236 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
4240 The following options control compiler behavior regarding floating
4241 point arithmetic. These options trade off between speed and
4242 correctness. All must be specifically enabled.
4246 @opindex ffloat-store
4247 Do not store floating point variables in registers, and inhibit other
4248 options that might change whether a floating point value is taken from a
4251 @cindex floating point precision
4252 This option prevents undesirable excess precision on machines such as
4253 the 68000 where the floating registers (of the 68881) keep more
4254 precision than a @code{double} is supposed to have. Similarly for the
4255 x86 architecture. For most programs, the excess precision does only
4256 good, but a few programs rely on the precise definition of IEEE floating
4257 point. Use @option{-ffloat-store} for such programs, after modifying
4258 them to store all pertinent intermediate computations into variables.
4262 Sets @option{-fno-math-errno}, @option{-funsafe-math-optimizations}, @*
4263 @option{-fno-trapping-math}, @option{-ffinite-math-only} and @*
4264 @option{-fno-signaling-nans}.
4266 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
4268 This option should never be turned on by any @option{-O} option since
4269 it can result in incorrect output for programs which depend on
4270 an exact implementation of IEEE or ISO rules/specifications for
4273 @item -fno-math-errno
4274 @opindex fno-math-errno
4275 Do not set ERRNO after calling math functions that are executed
4276 with a single instruction, e.g., sqrt. A program that relies on
4277 IEEE exceptions for math error handling may want to use this flag
4278 for speed while maintaining IEEE arithmetic compatibility.
4280 This option should never be turned on by any @option{-O} option since
4281 it can result in incorrect output for programs which depend on
4282 an exact implementation of IEEE or ISO rules/specifications for
4285 The default is @option{-fmath-errno}.
4287 @item -funsafe-math-optimizations
4288 @opindex funsafe-math-optimizations
4289 Allow optimizations for floating-point arithmetic that (a) assume
4290 that arguments and results are valid and (b) may violate IEEE or
4291 ANSI standards. When used at link-time, it may include libraries
4292 or startup files that change the default FPU control word or other
4293 similar optimizations.
4295 This option should never be turned on by any @option{-O} option since
4296 it can result in incorrect output for programs which depend on
4297 an exact implementation of IEEE or ISO rules/specifications for
4300 The default is @option{-fno-unsafe-math-optimizations}.
4302 @item -ffinite-math-only
4303 @opindex ffinite-math-only
4304 Allow optimizations for floating-point arithmetic that assume
4305 that arguments and results are not NaNs or +-Infs.
4307 This option should never be turned on by any @option{-O} option since
4308 it can result in incorrect output for programs which depend on
4309 an exact implementation of IEEE or ISO rules/specifications.
4311 The default is @option{-fno-finite-math-only}.
4313 @item -fno-trapping-math
4314 @opindex fno-trapping-math
4315 Compile code assuming that floating-point operations cannot generate
4316 user-visible traps. These traps include division by zero, overflow,
4317 underflow, inexact result and invalid operation. This option implies
4318 @option{-fno-signaling-nans}. Setting this option may allow faster
4319 code if one relies on ``non-stop'' IEEE arithmetic, for example.
4321 This option should never be turned on by any @option{-O} option since
4322 it can result in incorrect output for programs which depend on
4323 an exact implementation of IEEE or ISO rules/specifications for
4326 The default is @option{-ftrapping-math}.
4328 @item -fsignaling-nans
4329 @opindex fsignaling-nans
4330 Compile code assuming that IEEE signaling NaNs may generate user-visible
4331 traps during floating-point operations. Setting this option disables
4332 optimizations that may change the number of exceptions visible with
4333 signaling NaNs. This option implies @option{-ftrapping-math}.
4335 This option causes the preprocessor macro @code{__SUPPORT_SNAN__} to
4338 The default is @option{-fno-signaling-nans}.
4340 This option is experimental and does not currently guarantee to
4341 disable all GCC optimizations that affect signaling NaN behavior.
4343 @item -fsingle-precision-constant
4344 @opindex fsingle-precision-constant
4345 Treat floating point constant as single precision constant instead of
4346 implicitly converting it to double precision constant.
4351 The following options control optimizations that may improve
4352 performance, but are not enabled by any @option{-O} options. This
4353 section includes experimental options that may produce broken code.
4356 @item -fbranch-probabilities
4357 @opindex fbranch-probabilities
4358 After running a program compiled with @option{-fprofile-arcs}
4359 (@pxref{Debugging Options,, Options for Debugging Your Program or
4360 @command{gcc}}), you can compile it a second time using
4361 @option{-fbranch-probabilities}, to improve optimizations based on
4362 the number of times each branch was taken. When the program
4363 compiled with @option{-fprofile-arcs} exits it saves arc execution
4364 counts to a file called @file{@var{sourcename}.gcda} for each source
4365 file The information in this data file is very dependent on the
4366 structure of the generated code, so you must use the same source code
4367 and the same optimization options for both compilations.
4369 With @option{-fbranch-probabilities}, GCC puts a
4370 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
4371 These can be used to improve optimization. Currently, they are only
4372 used in one place: in @file{reorg.c}, instead of guessing which path a
4373 branch is mostly to take, the @samp{REG_BR_PROB} values are used to
4374 exactly determine which path is taken more often.
4376 @item -fprofile-values
4377 @opindex fprofile-values
4378 If combined with @option{-fprofile-arcs}, it adds code so that some
4379 data about values of expressions in the program is gathered.
4381 With @option{-fbranch-probabilities}, it reads back the data gathered
4382 from profiling values of expressions and adds @samp{REG_VALUE_PROFILE}
4383 notes to instructions for their later usage in optimizations.
4387 Use a graph coloring register allocator. Currently this option is meant
4388 for testing, so we are interested to hear about miscompilations with
4393 Perform tail duplication to enlarge superblock size. This transformation
4394 simplifies the control flow of the function allowing other optimizations to do
4397 @item -funit-at-a-time
4398 @opindex funit-at-a-time
4399 Parse the whole compilation unit before starting to produce code.
4400 This allows some extra optimizations to take place but consumes more
4403 @item -funroll-loops
4404 @opindex funroll-loops
4405 Unroll loops whose number of iterations can be determined at compile time or
4406 upon entry to the loop. @option{-funroll-loops} implies
4407 @option{-frerun-cse-after-loop}. It also turns on complete loop peeling
4408 (i.e. complete removal of loops with small constant number of iterations).
4409 This option makes code larger, and may or may not make it run faster.
4411 @item -funroll-all-loops
4412 @opindex funroll-all-loops
4413 Unroll all loops, even if their number of iterations is uncertain when
4414 the loop is entered. This usually makes programs run more slowly.
4415 @option{-funroll-all-loops} implies the same options as
4416 @option{-funroll-loops}.
4419 @opindex fpeel-loops
4420 Peels the loops for that there is enough information that they do not
4421 roll much (from profile feedback). It also turns on complete loop peeling
4422 (i.e. complete removal of loops with small constant number of iterations).
4424 @item -funswitch-loops
4425 @opindex funswitch-loops
4426 Move branches with loop invariant conditions out of the loop, with duplicates
4427 of the loop on both branches (modified according to result of the condition).
4429 @item -fold-unroll-loops
4430 @opindex fold-unroll-loops
4431 Unroll loops whose number of iterations can be determined at compile
4432 time or upon entry to the loop, using the old loop unroller whose loop
4433 recognition is based on notes from frontend. @option{-fold-unroll-loops} implies
4434 both @option{-fstrength-reduce} and @option{-frerun-cse-after-loop}. This
4435 option makes code larger, and may or may not make it run faster.
4437 @item -fold-unroll-all-loops
4438 @opindex fold-unroll-all-loops
4439 Unroll all loops, even if their number of iterations is uncertain when
4440 the loop is entered. This is done using the old loop unroller whose loop
4441 recognition is based on notes from frontend. This usually makes programs run more slowly.
4442 @option{-fold-unroll-all-loops} implies the same options as
4443 @option{-fold-unroll-loops}.
4445 @item -funswitch-loops
4446 @opindex funswitch-loops
4447 Move branches with loop invariant conditions out of the loop, with duplicates
4448 of the loop on both branches (modified according to result of the condition).
4450 @item -funswitch-loops
4451 @opindex funswitch-loops
4452 Move branches with loop invariant conditions out of the loop, with duplicates
4453 of the loop on both branches (modified according to result of the condition).
4455 @item -fprefetch-loop-arrays
4456 @opindex fprefetch-loop-arrays
4457 If supported by the target machine, generate instructions to prefetch
4458 memory to improve the performance of loops that access large arrays.
4460 Disabled at level @option{-Os}.
4462 @item -ffunction-sections
4463 @itemx -fdata-sections
4464 @opindex ffunction-sections
4465 @opindex fdata-sections
4466 Place each function or data item into its own section in the output
4467 file if the target supports arbitrary sections. The name of the
4468 function or the name of the data item determines the section's name
4471 Use these options on systems where the linker can perform optimizations
4472 to improve locality of reference in the instruction space. Most systems
4473 using the ELF object format and SPARC processors running Solaris 2 have
4474 linkers with such optimizations. AIX may have these optimizations in
4477 Only use these options when there are significant benefits from doing
4478 so. When you specify these options, the assembler and linker will
4479 create larger object and executable files and will also be slower.
4480 You will not be able to use @code{gprof} on all systems if you
4481 specify this option and you may have problems with debugging if
4482 you specify both this option and @option{-g}.
4486 Perform optimizations in static single assignment form. Each function's
4487 flow graph is translated into SSA form, optimizations are performed, and
4488 the flow graph is translated back from SSA form. Users should not
4489 specify this option, since it is not yet ready for production use.
4493 Perform Sparse Conditional Constant Propagation in SSA form. Requires
4494 @option{-fssa}. Like @option{-fssa}, this is an experimental feature.
4498 Perform aggressive dead-code elimination in SSA form. Requires @option{-fssa}.
4499 Like @option{-fssa}, this is an experimental feature.
4501 @item -fbranch-target-load-optimize
4502 @opindex fbranch-target-load-optimize
4503 Perform branch target register load optimization before prologue / epilogue
4505 The use of target registers can typically be exposed only during reload,
4506 thus hoisting loads out of loops and doing inter-block scheduling needs
4507 a separate optimization pass.
4509 @item -fbranch-target-load-optimize2
4510 @opindex fbranch-target-load-optimize2
4511 Perform branch target register load optimization after prologue / epilogue
4517 @item --param @var{name}=@var{value}
4519 In some places, GCC uses various constants to control the amount of
4520 optimization that is done. For example, GCC will not inline functions
4521 that contain more that a certain number of instructions. You can
4522 control some of these constants on the command-line using the
4523 @option{--param} option.
4525 In each case, the @var{value} is an integer. The allowable choices for
4526 @var{name} are given in the following table:
4529 @item max-crossjump-edges
4530 The maximum number of incoming edges to consider for crossjumping.
4531 The algorithm used by @option{-fcrossjumping} is @math{O(N^2)} in
4532 the number of edges incoming to each block. Increasing values mean
4533 more aggressive optimization, making the compile time increase with
4534 probably small improvement in executable size.
4536 @item max-delay-slot-insn-search
4537 The maximum number of instructions to consider when looking for an
4538 instruction to fill a delay slot. If more than this arbitrary number of
4539 instructions is searched, the time savings from filling the delay slot
4540 will be minimal so stop searching. Increasing values mean more
4541 aggressive optimization, making the compile time increase with probably
4542 small improvement in executable run time.
4544 @item max-delay-slot-live-search
4545 When trying to fill delay slots, the maximum number of instructions to
4546 consider when searching for a block with valid live register
4547 information. Increasing this arbitrarily chosen value means more
4548 aggressive optimization, increasing the compile time. This parameter
4549 should be removed when the delay slot code is rewritten to maintain the
4552 @item max-gcse-memory
4553 The approximate maximum amount of memory that will be allocated in
4554 order to perform the global common subexpression elimination
4555 optimization. If more memory than specified is required, the
4556 optimization will not be done.
4558 @item max-gcse-passes
4559 The maximum number of passes of GCSE to run.
4561 @item max-pending-list-length
4562 The maximum number of pending dependencies scheduling will allow
4563 before flushing the current state and starting over. Large functions
4564 with few branches or calls can create excessively large lists which
4565 needlessly consume memory and resources.
4567 @item max-inline-insns-single
4568 Several parameters control the tree inliner used in gcc.
4569 This number sets the maximum number of instructions (counted in gcc's
4570 internal representation) in a single function that the tree inliner
4571 will consider for inlining. This only affects functions declared
4572 inline and methods implemented in a class declaration (C++).
4573 The default value is 100.
4575 @item max-inline-insns-auto
4576 When you use @option{-finline-functions} (included in @option{-O3}),
4577 a lot of functions that would otherwise not be considered for inlining
4578 by the compiler will be investigated. To those functions, a different
4579 (more restrictive) limit compared to functions declared inline can
4581 The default value is 100.
4583 @item max-inline-insns
4584 The tree inliner does decrease the allowable size for single functions
4585 to be inlined after we already inlined the number of instructions
4586 given here by repeated inlining. This number should be a factor of
4587 two or more larger than the single function limit.
4588 Higher numbers result in better runtime performance, but incur higher
4589 compile-time resource (CPU time, memory) requirements and result in
4590 larger binaries. Very high values are not advisable, as too large
4591 binaries may adversely affect runtime performance.
4592 The default value is 200.
4594 @item max-inline-slope
4595 After exceeding the maximum number of inlined instructions by repeated
4596 inlining, a linear function is used to decrease the allowable size
4597 for single functions. The slope of that function is the negative
4598 reciprocal of the number specified here.
4599 This parameter is ignored when @option{-funit-at-a-time} is used.
4600 The default value is 32.
4602 @item min-inline-insns
4603 The repeated inlining is throttled more and more by the linear function
4604 after exceeding the limit. To avoid too much throttling, a minimum for
4605 this function is specified here to allow repeated inlining for very small
4606 functions even when a lot of repeated inlining already has been done.
4607 This parameter is ignored when @option{-funit-at-a-time} is used.
4608 The default value is 10.
4610 @item large-function-insns
4611 The limit specifying really large functions. For functions greater than this
4612 limit inlining is constrained by @option{--param large-function-growth}.
4613 This parameter is usefull primarily to avoid extreme compilation time caused by non-linear
4614 algorithms used by the backend.
4615 This parameter is ignored when @option{-funit-at-a-time} is not used.
4616 The default value is 30000.
4618 @item large-function-growth
4619 Specifies maximal growth of large functtion caused by inlining in percents.
4620 This parameter is ignored when @option{-funit-at-a-time} is not used.
4621 The default value is 200.
4623 @item inline-unit-growth
4624 Specifies maximal overall growth of the compilation unit caused by inlining.
4625 This parameter is ignored when @option{-funit-at-a-time} is not used.
4626 The default value is 150.
4628 @item max-inline-insns-rtl
4629 For languages that use the RTL inliner (this happens at a later stage
4630 than tree inlining), you can set the maximum allowable size (counted
4631 in RTL instructions) for the RTL inliner with this parameter.
4632 The default value is 600.
4635 @item max-unrolled-insns
4636 The maximum number of instructions that a loop should have if that loop
4637 is unrolled, and if the loop is unrolled, it determines how many times
4638 the loop code is unrolled.
4640 @item max-average-unrolled-insns
4641 The maximum number of instructions biased by probabilities of their execution
4642 that a loop should have if that loop is unrolled, and if the loop is unrolled,
4643 it determines how many times the loop code is unrolled.
4645 @item max-unroll-times
4646 The maximum number of unrollings of a single loop.
4648 @item max-peeled-insns
4649 The maximum number of instructions that a loop should have if that loop
4650 is peeled, and if the loop is peeled, it determines how many times
4651 the loop code is peeled.
4653 @item max-peel-times
4654 The maximum number of peelings of a single loop.
4656 @item max-completely-peeled-insns
4657 The maximum number of insns of a completely peeled loop.
4659 @item max-completely-peel-times
4660 The maximum number of iterations of a loop to be suitable for complete peeling.
4662 @item max-unswitch-insns
4663 The maximum number of insns of an unswitched loop.
4665 @item max-unswitch-level
4666 The maximum number of branches unswitched in a single loop.
4668 @item hot-bb-count-fraction
4669 Select fraction of the maximal count of repetitions of basic block in program
4670 given basic block needs to have to be considered hot.
4672 @item hot-bb-frequency-fraction
4673 Select fraction of the maximal frequency of executions of basic block in
4674 function given basic block needs to have to be considered hot
4676 @item tracer-dynamic-coverage
4677 @itemx tracer-dynamic-coverage-feedback
4679 This value is used to limit superblock formation once the given percentage of
4680 executed instructions is covered. This limits unnecessary code size
4683 The @option{tracer-dynamic-coverage-feedback} is used only when profile
4684 feedback is available. The real profiles (as opposed to statically estimated
4685 ones) are much less balanced allowing the threshold to be larger value.
4687 @item tracer-max-code-growth
4688 Stop tail duplication once code growth has reached given percentage. This is
4689 rather hokey argument, as most of the duplicates will be eliminated later in
4690 cross jumping, so it may be set to much higher values than is the desired code
4693 @item tracer-min-branch-ratio
4695 Stop reverse growth when the reverse probability of best edge is less than this
4696 threshold (in percent).
4698 @item tracer-min-branch-ratio
4699 @itemx tracer-min-branch-ratio-feedback
4701 Stop forward growth if the best edge do have probability lower than this
4704 Similarly to @option{tracer-dynamic-coverage} two values are present, one for
4705 compilation for profile feedback and one for compilation without. The value
4706 for compilation with profile feedback needs to be more conservative (higher) in
4707 order to make tracer effective.
4709 @item max-cse-path-length
4711 Maximum number of basic blocks on path that cse considers.
4713 @item ggc-min-expand
4715 GCC uses a garbage collector to manage its own memory allocation. This
4716 parameter specifies the minimum percentage by which the garbage
4717 collector's heap should be allowed to expand between collections.
4718 Tuning this may improve compilation speed; it has no effect on code
4721 The default is 30% + 70% * (RAM/1GB) with an upper bound of 100% when
4722 RAM >= 1GB. If @code{getrlimit} is available, the notion of "RAM" is
4723 the smallest of actual RAM, RLIMIT_RSS, RLIMIT_DATA and RLIMIT_AS. If
4724 GCC is not able to calculate RAM on a particular platform, the lower
4725 bound of 30% is used. Setting this parameter and
4726 @option{ggc-min-heapsize} to zero causes a full collection to occur at
4727 every opportunity. This is extremely slow, but can be useful for
4730 @item ggc-min-heapsize
4732 Minimum size of the garbage collector's heap before it begins bothering
4733 to collect garbage. The first collection occurs after the heap expands
4734 by @option{ggc-min-expand}% beyond @option{ggc-min-heapsize}. Again,
4735 tuning this may improve compilation speed, and has no effect on code
4738 The default is RAM/8, with a lower bound of 4096 (four megabytes) and an
4739 upper bound of 131072 (128 megabytes). If @code{getrlimit} is
4740 available, the notion of "RAM" is the smallest of actual RAM,
4741 RLIMIT_RSS, RLIMIT_DATA and RLIMIT_AS. If GCC is not able to calculate
4742 RAM on a particular platform, the lower bound is used. Setting this
4743 parameter very large effectively disables garbage collection. Setting
4744 this parameter and @option{ggc-min-expand} to zero causes a full
4745 collection to occur at every opportunity.
4747 @item reorder-blocks-duplicate
4748 @itemx reorder-blocks-duplicate-feedback
4750 Used by basic block reordering pass to decide whether to use unconditional
4751 branch or duplicate the code on its destination. Code is duplicated when its
4752 estimated size is smaller than this value multiplied by the estimated size of
4753 unconditional jump in the hot spots of the program.
4755 The @option{reorder-block-duplicate-feedback} is used only when profile
4756 feedback is available and may be set to higher values than
4757 @option{reorder-block-duplicate} since information about the hot spots is more
4762 @node Preprocessor Options
4763 @section Options Controlling the Preprocessor
4764 @cindex preprocessor options
4765 @cindex options, preprocessor
4767 These options control the C preprocessor, which is run on each C source
4768 file before actual compilation.
4770 If you use the @option{-E} option, nothing is done except preprocessing.
4771 Some of these options make sense only together with @option{-E} because
4772 they cause the preprocessor output to be unsuitable for actual
4777 You can use @option{-Wp,@var{option}} to bypass the compiler driver
4778 and pass @var{option} directly through to the preprocessor. If
4779 @var{option} contains commas, it is split into multiple options at the
4780 commas. However, many options are modified, translated or interpreted
4781 by the compiler driver before being passed to the preprocessor, and
4782 @option{-Wp} forcibly bypasses this phase. The preprocessor's direct
4783 interface is undocumented and subject to change, so whenever possible
4784 you should avoid using @option{-Wp} and let the driver handle the
4787 @item -Xpreprocessor @var{option}
4788 @opindex preprocessor
4789 Pass @var{option} as an option to the preprocessor. You can use this to
4790 supply system-specific preprocessor options which GCC does not know how to
4793 If you want to pass an option that takes an argument, you must use
4794 @option{-Xpreprocessor} twice, once for the option and once for the argument.
4797 @include cppopts.texi
4799 @node Assembler Options
4800 @section Passing Options to the Assembler
4802 @c prevent bad page break with this line
4803 You can pass options to the assembler.
4806 @item -Wa,@var{option}
4808 Pass @var{option} as an option to the assembler. If @var{option}
4809 contains commas, it is split into multiple options at the commas.
4811 @item -Xassembler @var{option}
4813 Pass @var{option} as an option to the assembler. You can use this to
4814 supply system-specific assembler options which GCC does not know how to
4817 If you want to pass an option that takes an argument, you must use
4818 @option{-Xassembler} twice, once for the option and once for the argument.
4823 @section Options for Linking
4824 @cindex link options
4825 @cindex options, linking
4827 These options come into play when the compiler links object files into
4828 an executable output file. They are meaningless if the compiler is
4829 not doing a link step.
4833 @item @var{object-file-name}
4834 A file name that does not end in a special recognized suffix is
4835 considered to name an object file or library. (Object files are
4836 distinguished from libraries by the linker according to the file
4837 contents.) If linking is done, these object files are used as input
4846 If any of these options is used, then the linker is not run, and
4847 object file names should not be used as arguments. @xref{Overall
4851 @item -l@var{library}
4852 @itemx -l @var{library}
4854 Search the library named @var{library} when linking. (The second
4855 alternative with the library as a separate argument is only for
4856 POSIX compliance and is not recommended.)
4858 It makes a difference where in the command you write this option; the
4859 linker searches and processes libraries and object files in the order they
4860 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
4861 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
4862 to functions in @samp{z}, those functions may not be loaded.
4864 The linker searches a standard list of directories for the library,
4865 which is actually a file named @file{lib@var{library}.a}. The linker
4866 then uses this file as if it had been specified precisely by name.
4868 The directories searched include several standard system directories
4869 plus any that you specify with @option{-L}.
4871 Normally the files found this way are library files---archive files
4872 whose members are object files. The linker handles an archive file by
4873 scanning through it for members which define symbols that have so far
4874 been referenced but not defined. But if the file that is found is an
4875 ordinary object file, it is linked in the usual fashion. The only
4876 difference between using an @option{-l} option and specifying a file name
4877 is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
4878 and searches several directories.
4882 You need this special case of the @option{-l} option in order to
4883 link an Objective-C program.
4886 @opindex nostartfiles
4887 Do not use the standard system startup files when linking.
4888 The standard system libraries are used normally, unless @option{-nostdlib}
4889 or @option{-nodefaultlibs} is used.
4891 @item -nodefaultlibs
4892 @opindex nodefaultlibs
4893 Do not use the standard system libraries when linking.
4894 Only the libraries you specify will be passed to the linker.
4895 The standard startup files are used normally, unless @option{-nostartfiles}
4896 is used. The compiler may generate calls to memcmp, memset, and memcpy
4897 for System V (and ISO C) environments or to bcopy and bzero for
4898 BSD environments. These entries are usually resolved by entries in
4899 libc. These entry points should be supplied through some other
4900 mechanism when this option is specified.
4904 Do not use the standard system startup files or libraries when linking.
4905 No startup files and only the libraries you specify will be passed to
4906 the linker. The compiler may generate calls to memcmp, memset, and memcpy
4907 for System V (and ISO C) environments or to bcopy and bzero for
4908 BSD environments. These entries are usually resolved by entries in
4909 libc. These entry points should be supplied through some other
4910 mechanism when this option is specified.
4912 @cindex @option{-lgcc}, use with @option{-nostdlib}
4913 @cindex @option{-nostdlib} and unresolved references
4914 @cindex unresolved references and @option{-nostdlib}
4915 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
4916 @cindex @option{-nodefaultlibs} and unresolved references
4917 @cindex unresolved references and @option{-nodefaultlibs}
4918 One of the standard libraries bypassed by @option{-nostdlib} and
4919 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
4920 that GCC uses to overcome shortcomings of particular machines, or special
4921 needs for some languages.
4922 (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler
4923 Collection (GCC) Internals},
4924 for more discussion of @file{libgcc.a}.)
4925 In most cases, you need @file{libgcc.a} even when you want to avoid
4926 other standard libraries. In other words, when you specify @option{-nostdlib}
4927 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
4928 This ensures that you have no unresolved references to internal GCC
4929 library subroutines. (For example, @samp{__main}, used to ensure C++
4930 constructors will be called; @pxref{Collect2,,@code{collect2}, gccint,
4931 GNU Compiler Collection (GCC) Internals}.)
4935 Produce a position independent executable on targets which support it.
4936 For predictable results, you must also specify the same set of options
4937 that were used to generate code (@option{-fpie}, @option{-fPIE},
4938 or model suboptions) when you specify this option.
4942 Remove all symbol table and relocation information from the executable.
4946 On systems that support dynamic linking, this prevents linking with the shared
4947 libraries. On other systems, this option has no effect.
4951 Produce a shared object which can then be linked with other objects to
4952 form an executable. Not all systems support this option. For predictable
4953 results, you must also specify the same set of options that were used to
4954 generate code (@option{-fpic}, @option{-fPIC}, or model suboptions)
4955 when you specify this option.@footnote{On some systems, @samp{gcc -shared}
4956 needs to build supplementary stub code for constructors to work. On
4957 multi-libbed systems, @samp{gcc -shared} must select the correct support
4958 libraries to link against. Failing to supply the correct flags may lead
4959 to subtle defects. Supplying them in cases where they are not necessary
4962 @item -shared-libgcc
4963 @itemx -static-libgcc
4964 @opindex shared-libgcc
4965 @opindex static-libgcc
4966 On systems that provide @file{libgcc} as a shared library, these options
4967 force the use of either the shared or static version respectively.
4968 If no shared version of @file{libgcc} was built when the compiler was
4969 configured, these options have no effect.
4971 There are several situations in which an application should use the
4972 shared @file{libgcc} instead of the static version. The most common
4973 of these is when the application wishes to throw and catch exceptions
4974 across different shared libraries. In that case, each of the libraries
4975 as well as the application itself should use the shared @file{libgcc}.
4977 Therefore, the G++ and GCJ drivers automatically add
4978 @option{-shared-libgcc} whenever you build a shared library or a main
4979 executable, because C++ and Java programs typically use exceptions, so
4980 this is the right thing to do.
4982 If, instead, you use the GCC driver to create shared libraries, you may
4983 find that they will not always be linked with the shared @file{libgcc}.
4984 If GCC finds, at its configuration time, that you have a GNU linker that
4985 does not support option @option{--eh-frame-hdr}, it will link the shared
4986 version of @file{libgcc} into shared libraries by default. Otherwise,
4987 it will take advantage of the linker and optimize away the linking with
4988 the shared version of @file{libgcc}, linking with the static version of
4989 libgcc by default. This allows exceptions to propagate through such
4990 shared libraries, without incurring relocation costs at library load
4993 However, if a library or main executable is supposed to throw or catch
4994 exceptions, you must link it using the G++ or GCJ driver, as appropriate
4995 for the languages used in the program, or using the option
4996 @option{-shared-libgcc}, such that it is linked with the shared
5001 Bind references to global symbols when building a shared object. Warn
5002 about any unresolved references (unless overridden by the link editor
5003 option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support
5006 @item -Xlinker @var{option}
5008 Pass @var{option} as an option to the linker. You can use this to
5009 supply system-specific linker options which GCC does not know how to
5012 If you want to pass an option that takes an argument, you must use
5013 @option{-Xlinker} twice, once for the option and once for the argument.
5014 For example, to pass @option{-assert definitions}, you must write
5015 @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write
5016 @option{-Xlinker "-assert definitions"}, because this passes the entire
5017 string as a single argument, which is not what the linker expects.
5019 @item -Wl,@var{option}
5021 Pass @var{option} as an option to the linker. If @var{option} contains
5022 commas, it is split into multiple options at the commas.
5024 @item -u @var{symbol}
5026 Pretend the symbol @var{symbol} is undefined, to force linking of
5027 library modules to define it. You can use @option{-u} multiple times with
5028 different symbols to force loading of additional library modules.
5031 @node Directory Options
5032 @section Options for Directory Search
5033 @cindex directory options
5034 @cindex options, directory search
5037 These options specify directories to search for header files, for
5038 libraries and for parts of the compiler:
5043 Add the directory @var{dir} to the head of the list of directories to be
5044 searched for header files. This can be used to override a system header
5045 file, substituting your own version, since these directories are
5046 searched before the system header file directories. However, you should
5047 not use this option to add directories that contain vendor-supplied
5048 system header files (use @option{-isystem} for that). If you use more than
5049 one @option{-I} option, the directories are scanned in left-to-right
5050 order; the standard system directories come after.
5052 If a standard system include directory, or a directory specified with
5053 @option{-isystem}, is also specified with @option{-I}, the @option{-I}
5054 option will be ignored. The directory will still be searched but as a
5055 system directory at its normal position in the system include chain.
5056 This is to ensure that GCC's procedure to fix buggy system headers and
5057 the ordering for the include_next directive are not inadvertently changed.
5058 If you really need to change the search order for system directories,
5059 use the @option{-nostdinc} and/or @option{-isystem} options.
5063 Any directories you specify with @option{-I} options before the @option{-I-}
5064 option are searched only for the case of @samp{#include "@var{file}"};
5065 they are not searched for @samp{#include <@var{file}>}.
5067 If additional directories are specified with @option{-I} options after
5068 the @option{-I-}, these directories are searched for all @samp{#include}
5069 directives. (Ordinarily @emph{all} @option{-I} directories are used
5072 In addition, the @option{-I-} option inhibits the use of the current
5073 directory (where the current input file came from) as the first search
5074 directory for @samp{#include "@var{file}"}. There is no way to
5075 override this effect of @option{-I-}. With @option{-I.} you can specify
5076 searching the directory which was current when the compiler was
5077 invoked. That is not exactly the same as what the preprocessor does
5078 by default, but it is often satisfactory.
5080 @option{-I-} does not inhibit the use of the standard system directories
5081 for header files. Thus, @option{-I-} and @option{-nostdinc} are
5086 Add directory @var{dir} to the list of directories to be searched
5089 @item -B@var{prefix}
5091 This option specifies where to find the executables, libraries,
5092 include files, and data files of the compiler itself.
5094 The compiler driver program runs one or more of the subprograms
5095 @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries
5096 @var{prefix} as a prefix for each program it tries to run, both with and
5097 without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}).
5099 For each subprogram to be run, the compiler driver first tries the
5100 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
5101 was not specified, the driver tries two standard prefixes, which are
5102 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc/}. If neither of
5103 those results in a file name that is found, the unmodified program
5104 name is searched for using the directories specified in your
5105 @env{PATH} environment variable.
5107 The compiler will check to see if the path provided by the @option{-B}
5108 refers to a directory, and if necessary it will add a directory
5109 separator character at the end of the path.
5111 @option{-B} prefixes that effectively specify directory names also apply
5112 to libraries in the linker, because the compiler translates these
5113 options into @option{-L} options for the linker. They also apply to
5114 includes files in the preprocessor, because the compiler translates these
5115 options into @option{-isystem} options for the preprocessor. In this case,
5116 the compiler appends @samp{include} to the prefix.
5118 The run-time support file @file{libgcc.a} can also be searched for using
5119 the @option{-B} prefix, if needed. If it is not found there, the two
5120 standard prefixes above are tried, and that is all. The file is left
5121 out of the link if it is not found by those means.
5123 Another way to specify a prefix much like the @option{-B} prefix is to use
5124 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
5127 As a special kludge, if the path provided by @option{-B} is
5128 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
5129 9, then it will be replaced by @file{[dir/]include}. This is to help
5130 with boot-strapping the compiler.
5132 @item -specs=@var{file}
5134 Process @var{file} after the compiler reads in the standard @file{specs}
5135 file, in order to override the defaults that the @file{gcc} driver
5136 program uses when determining what switches to pass to @file{cc1},
5137 @file{cc1plus}, @file{as}, @file{ld}, etc. More than one
5138 @option{-specs=@var{file}} can be specified on the command line, and they
5139 are processed in order, from left to right.
5145 @section Specifying subprocesses and the switches to pass to them
5148 @command{gcc} is a driver program. It performs its job by invoking a
5149 sequence of other programs to do the work of compiling, assembling and
5150 linking. GCC interprets its command-line parameters and uses these to
5151 deduce which programs it should invoke, and which command-line options
5152 it ought to place on their command lines. This behavior is controlled
5153 by @dfn{spec strings}. In most cases there is one spec string for each
5154 program that GCC can invoke, but a few programs have multiple spec
5155 strings to control their behavior. The spec strings built into GCC can
5156 be overridden by using the @option{-specs=} command-line switch to specify
5159 @dfn{Spec files} are plaintext files that are used to construct spec
5160 strings. They consist of a sequence of directives separated by blank
5161 lines. The type of directive is determined by the first non-whitespace
5162 character on the line and it can be one of the following:
5165 @item %@var{command}
5166 Issues a @var{command} to the spec file processor. The commands that can
5170 @item %include <@var{file}>
5172 Search for @var{file} and insert its text at the current point in the
5175 @item %include_noerr <@var{file}>
5176 @cindex %include_noerr
5177 Just like @samp{%include}, but do not generate an error message if the include
5178 file cannot be found.
5180 @item %rename @var{old_name} @var{new_name}
5182 Rename the spec string @var{old_name} to @var{new_name}.
5186 @item *[@var{spec_name}]:
5187 This tells the compiler to create, override or delete the named spec
5188 string. All lines after this directive up to the next directive or
5189 blank line are considered to be the text for the spec string. If this
5190 results in an empty string then the spec will be deleted. (Or, if the
5191 spec did not exist, then nothing will happened.) Otherwise, if the spec
5192 does not currently exist a new spec will be created. If the spec does
5193 exist then its contents will be overridden by the text of this
5194 directive, unless the first character of that text is the @samp{+}
5195 character, in which case the text will be appended to the spec.
5197 @item [@var{suffix}]:
5198 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
5199 and up to the next directive or blank line are considered to make up the
5200 spec string for the indicated suffix. When the compiler encounters an
5201 input file with the named suffix, it will processes the spec string in
5202 order to work out how to compile that file. For example:
5209 This says that any input file whose name ends in @samp{.ZZ} should be
5210 passed to the program @samp{z-compile}, which should be invoked with the
5211 command-line switch @option{-input} and with the result of performing the
5212 @samp{%i} substitution. (See below.)
5214 As an alternative to providing a spec string, the text that follows a
5215 suffix directive can be one of the following:
5218 @item @@@var{language}
5219 This says that the suffix is an alias for a known @var{language}. This is
5220 similar to using the @option{-x} command-line switch to GCC to specify a
5221 language explicitly. For example:
5228 Says that .ZZ files are, in fact, C++ source files.
5231 This causes an error messages saying:
5234 @var{name} compiler not installed on this system.
5238 GCC already has an extensive list of suffixes built into it.
5239 This directive will add an entry to the end of the list of suffixes, but
5240 since the list is searched from the end backwards, it is effectively
5241 possible to override earlier entries using this technique.
5245 GCC has the following spec strings built into it. Spec files can
5246 override these strings or create their own. Note that individual
5247 targets can also add their own spec strings to this list.
5250 asm Options to pass to the assembler
5251 asm_final Options to pass to the assembler post-processor
5252 cpp Options to pass to the C preprocessor
5253 cc1 Options to pass to the C compiler
5254 cc1plus Options to pass to the C++ compiler
5255 endfile Object files to include at the end of the link
5256 link Options to pass to the linker
5257 lib Libraries to include on the command line to the linker
5258 libgcc Decides which GCC support library to pass to the linker
5259 linker Sets the name of the linker
5260 predefines Defines to be passed to the C preprocessor
5261 signed_char Defines to pass to CPP to say whether @code{char} is signed
5263 startfile Object files to include at the start of the link
5266 Here is a small example of a spec file:
5272 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
5275 This example renames the spec called @samp{lib} to @samp{old_lib} and
5276 then overrides the previous definition of @samp{lib} with a new one.
5277 The new definition adds in some extra command-line options before
5278 including the text of the old definition.
5280 @dfn{Spec strings} are a list of command-line options to be passed to their
5281 corresponding program. In addition, the spec strings can contain
5282 @samp{%}-prefixed sequences to substitute variable text or to
5283 conditionally insert text into the command line. Using these constructs
5284 it is possible to generate quite complex command lines.
5286 Here is a table of all defined @samp{%}-sequences for spec
5287 strings. Note that spaces are not generated automatically around the
5288 results of expanding these sequences. Therefore you can concatenate them
5289 together or combine them with constant text in a single argument.
5293 Substitute one @samp{%} into the program name or argument.
5296 Substitute the name of the input file being processed.
5299 Substitute the basename of the input file being processed.
5300 This is the substring up to (and not including) the last period
5301 and not including the directory.
5304 This is the same as @samp{%b}, but include the file suffix (text after
5308 Marks the argument containing or following the @samp{%d} as a
5309 temporary file name, so that that file will be deleted if GCC exits
5310 successfully. Unlike @samp{%g}, this contributes no text to the
5313 @item %g@var{suffix}
5314 Substitute a file name that has suffix @var{suffix} and is chosen
5315 once per compilation, and mark the argument in the same way as
5316 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
5317 name is now chosen in a way that is hard to predict even when previously
5318 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
5319 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
5320 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
5321 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
5322 was simply substituted with a file name chosen once per compilation,
5323 without regard to any appended suffix (which was therefore treated
5324 just like ordinary text), making such attacks more likely to succeed.
5326 @item %u@var{suffix}
5327 Like @samp{%g}, but generates a new temporary file name even if
5328 @samp{%u@var{suffix}} was already seen.
5330 @item %U@var{suffix}
5331 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
5332 new one if there is no such last file name. In the absence of any
5333 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
5334 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
5335 would involve the generation of two distinct file names, one
5336 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
5337 simply substituted with a file name chosen for the previous @samp{%u},
5338 without regard to any appended suffix.
5340 @item %j@var{suffix}
5341 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
5342 writable, and if save-temps is off; otherwise, substitute the name
5343 of a temporary file, just like @samp{%u}. This temporary file is not
5344 meant for communication between processes, but rather as a junk
5347 @item %|@var{suffix}
5348 @itemx %m@var{suffix}
5349 Like @samp{%g}, except if @option{-pipe} is in effect. In that case
5350 @samp{%|} substitutes a single dash and @samp{%m} substitutes nothing at
5351 all. These are the two most common ways to instruct a program that it
5352 should read from standard input or write to standard output. If you
5353 need something more elaborate you can use an @samp{%@{pipe:@code{X}@}}
5354 construct: see for example @file{f/lang-specs.h}.
5356 @item %.@var{SUFFIX}
5357 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
5358 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
5359 terminated by the next space or %.
5362 Marks the argument containing or following the @samp{%w} as the
5363 designated output file of this compilation. This puts the argument
5364 into the sequence of arguments that @samp{%o} will substitute later.
5367 Substitutes the names of all the output files, with spaces
5368 automatically placed around them. You should write spaces
5369 around the @samp{%o} as well or the results are undefined.
5370 @samp{%o} is for use in the specs for running the linker.
5371 Input files whose names have no recognized suffix are not compiled
5372 at all, but they are included among the output files, so they will
5376 Substitutes the suffix for object files. Note that this is
5377 handled specially when it immediately follows @samp{%g, %u, or %U},
5378 because of the need for those to form complete file names. The
5379 handling is such that @samp{%O} is treated exactly as if it had already
5380 been substituted, except that @samp{%g, %u, and %U} do not currently
5381 support additional @var{suffix} characters following @samp{%O} as they would
5382 following, for example, @samp{.o}.
5385 Substitutes the standard macro predefinitions for the
5386 current target machine. Use this when running @code{cpp}.
5389 Like @samp{%p}, but puts @samp{__} before and after the name of each
5390 predefined macro, except for macros that start with @samp{__} or with
5391 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
5395 Substitute any of @option{-iprefix} (made from @env{GCC_EXEC_PREFIX}),
5396 @option{-isysroot} (made from @env{TARGET_SYSTEM_ROOT}), and
5397 @option{-isystem} (made from @env{COMPILER_PATH} and @option{-B} options)
5401 Current argument is the name of a library or startup file of some sort.
5402 Search for that file in a standard list of directories and substitute
5403 the full name found.
5406 Print @var{str} as an error message. @var{str} is terminated by a newline.
5407 Use this when inconsistent options are detected.
5410 Substitute the contents of spec string @var{name} at this point.
5413 Like @samp{%(@dots{})} but put @samp{__} around @option{-D} arguments.
5415 @item %x@{@var{option}@}
5416 Accumulate an option for @samp{%X}.
5419 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
5423 Output the accumulated assembler options specified by @option{-Wa}.
5426 Output the accumulated preprocessor options specified by @option{-Wp}.
5429 Process the @code{asm} spec. This is used to compute the
5430 switches to be passed to the assembler.
5433 Process the @code{asm_final} spec. This is a spec string for
5434 passing switches to an assembler post-processor, if such a program is
5438 Process the @code{link} spec. This is the spec for computing the
5439 command line passed to the linker. Typically it will make use of the
5440 @samp{%L %G %S %D and %E} sequences.
5443 Dump out a @option{-L} option for each directory that GCC believes might
5444 contain startup files. If the target supports multilibs then the
5445 current multilib directory will be prepended to each of these paths.
5448 Output the multilib directory with directory separators replaced with
5449 @samp{_}. If multilib directories are not set, or the multilib directory is
5450 @file{.} then this option emits nothing.
5453 Process the @code{lib} spec. This is a spec string for deciding which
5454 libraries should be included on the command line to the linker.
5457 Process the @code{libgcc} spec. This is a spec string for deciding
5458 which GCC support library should be included on the command line to the linker.
5461 Process the @code{startfile} spec. This is a spec for deciding which
5462 object files should be the first ones passed to the linker. Typically
5463 this might be a file named @file{crt0.o}.
5466 Process the @code{endfile} spec. This is a spec string that specifies
5467 the last object files that will be passed to the linker.
5470 Process the @code{cpp} spec. This is used to construct the arguments
5471 to be passed to the C preprocessor.
5474 Process the @code{signed_char} spec. This is intended to be used
5475 to tell cpp whether a char is signed. It typically has the definition:
5477 %@{funsigned-char:-D__CHAR_UNSIGNED__@}
5481 Process the @code{cc1} spec. This is used to construct the options to be
5482 passed to the actual C compiler (@samp{cc1}).
5485 Process the @code{cc1plus} spec. This is used to construct the options to be
5486 passed to the actual C++ compiler (@samp{cc1plus}).
5489 Substitute the variable part of a matched option. See below.
5490 Note that each comma in the substituted string is replaced by
5494 Remove all occurrences of @code{-S} from the command line. Note---this
5495 command is position dependent. @samp{%} commands in the spec string
5496 before this one will see @code{-S}, @samp{%} commands in the spec string
5497 after this one will not.
5499 @item %:@var{function}(@var{args})
5500 Call the named function @var{function}, passing it @var{args}.
5501 @var{args} is first processed as a nested spec string, then split
5502 into an argument vector in the usual fashion. The function returns
5503 a string which is processed as if it had appeared literally as part
5504 of the current spec.
5506 The following built-in spec functions are provided:
5509 @item @code{if-exists}
5510 The @code{if-exists} spec function takes one argument, an absolute
5511 pathname to a file. If the file exists, @code{if-exists} returns the
5512 pathname. Here is a small example of its usage:
5516 crt0%O%s %:if-exists(crti%O%s) crtbegin%O%s
5519 @item @code{if-exists-else}
5520 The @code{if-exists-else} spec function is similar to the @code{if-exists}
5521 spec function, except that it takes two arguments. The first argument is
5522 an absolute pathname to a file. If the file exists, @code{if-exists-else}
5523 returns the pathname. If it does not exist, it returns the second argument.
5524 This way, @code{if-exists-else} can be used to select one file or another,
5525 based on the existence of the first. Here is a small example of its usage:
5529 crt0%O%s %:if-exists(crti%O%s) \
5530 %:if-exists-else(crtbeginT%O%s crtbegin%O%s)
5535 Substitutes the @code{-S} switch, if that switch was given to GCC@.
5536 If that switch was not specified, this substitutes nothing. Note that
5537 the leading dash is omitted when specifying this option, and it is
5538 automatically inserted if the substitution is performed. Thus the spec
5539 string @samp{%@{foo@}} would match the command-line option @option{-foo}
5540 and would output the command line option @option{-foo}.
5542 @item %W@{@code{S}@}
5543 Like %@{@code{S}@} but mark last argument supplied within as a file to be
5546 @item %@{@code{S}*@}
5547 Substitutes all the switches specified to GCC whose names start
5548 with @code{-S}, but which also take an argument. This is used for
5549 switches like @option{-o}, @option{-D}, @option{-I}, etc.
5550 GCC considers @option{-o foo} as being
5551 one switch whose names starts with @samp{o}. %@{o*@} would substitute this
5552 text, including the space. Thus two arguments would be generated.
5554 @item %@{@code{S}*&@code{T}*@}
5555 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
5556 (the order of @code{S} and @code{T} in the spec is not significant).
5557 There can be any number of ampersand-separated variables; for each the
5558 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
5560 @item %@{@code{S}:@code{X}@}
5561 Substitutes @code{X}, if the @samp{-S} switch was given to GCC@.
5563 @item %@{!@code{S}:@code{X}@}
5564 Substitutes @code{X}, if the @samp{-S} switch was @emph{not} given to GCC@.
5566 @item %@{@code{S}*:@code{X}@}
5567 Substitutes @code{X} if one or more switches whose names start with
5568 @code{-S} are specified to GCC@. Normally @code{X} is substituted only
5569 once, no matter how many such switches appeared. However, if @code{%*}
5570 appears somewhere in @code{X}, then @code{X} will be substituted once
5571 for each matching switch, with the @code{%*} replaced by the part of
5572 that switch that matched the @code{*}.
5574 @item %@{.@code{S}:@code{X}@}
5575 Substitutes @code{X}, if processing a file with suffix @code{S}.
5577 @item %@{!.@code{S}:@code{X}@}
5578 Substitutes @code{X}, if @emph{not} processing a file with suffix @code{S}.
5580 @item %@{@code{S}|@code{P}:@code{X}@}
5581 Substitutes @code{X} if either @code{-S} or @code{-P} was given to GCC@.
5582 This may be combined with @samp{!}, @samp{.}, and @code{*} sequences as well,
5583 although they have a stronger binding than the @samp{|}. If @code{%*}
5584 appears in @code{X}, all of the alternatives must be starred, and only
5585 the first matching alternative is substituted.
5587 For example, a spec string like this:
5590 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
5593 will output the following command-line options from the following input
5594 command-line options:
5599 -d fred.c -foo -baz -boggle
5600 -d jim.d -bar -baz -boggle
5603 @item %@{S:X; T:Y; :D@}
5605 If @code{S} was given to GCC, substitutes @code{X}; else if @code{T} was
5606 given to GCC, substitutes @code{Y}; else substitutes @code{D}. There can
5607 be as many clauses as you need. This may be combined with @code{.},
5608 @code{!}, @code{|}, and @code{*} as needed.
5613 The conditional text @code{X} in a %@{@code{S}:@code{X}@} or similar
5614 construct may contain other nested @samp{%} constructs or spaces, or
5615 even newlines. They are processed as usual, as described above.
5616 Trailing white space in @code{X} is ignored. White space may also
5617 appear anywhere on the left side of the colon in these constructs,
5618 except between @code{.} or @code{*} and the corresponding word.
5620 The @option{-O}, @option{-f}, @option{-m}, and @option{-W} switches are
5621 handled specifically in these constructs. If another value of
5622 @option{-O} or the negated form of a @option{-f}, @option{-m}, or
5623 @option{-W} switch is found later in the command line, the earlier
5624 switch value is ignored, except with @{@code{S}*@} where @code{S} is
5625 just one letter, which passes all matching options.
5627 The character @samp{|} at the beginning of the predicate text is used to
5628 indicate that a command should be piped to the following command, but
5629 only if @option{-pipe} is specified.
5631 It is built into GCC which switches take arguments and which do not.
5632 (You might think it would be useful to generalize this to allow each
5633 compiler's spec to say which switches take arguments. But this cannot
5634 be done in a consistent fashion. GCC cannot even decide which input
5635 files have been specified without knowing which switches take arguments,
5636 and it must know which input files to compile in order to tell which
5639 GCC also knows implicitly that arguments starting in @option{-l} are to be
5640 treated as compiler output files, and passed to the linker in their
5641 proper position among the other output files.
5643 @c man begin OPTIONS
5645 @node Target Options
5646 @section Specifying Target Machine and Compiler Version
5647 @cindex target options
5648 @cindex cross compiling
5649 @cindex specifying machine version
5650 @cindex specifying compiler version and target machine
5651 @cindex compiler version, specifying
5652 @cindex target machine, specifying
5654 The usual way to run GCC is to run the executable called @file{gcc}, or
5655 @file{<machine>-gcc} when cross-compiling, or
5656 @file{<machine>-gcc-<version>} to run a version other than the one that
5657 was installed last. Sometimes this is inconvenient, so GCC provides
5658 options that will switch to another cross-compiler or version.
5661 @item -b @var{machine}
5663 The argument @var{machine} specifies the target machine for compilation.
5665 The value to use for @var{machine} is the same as was specified as the
5666 machine type when configuring GCC as a cross-compiler. For
5667 example, if a cross-compiler was configured with @samp{configure
5668 i386v}, meaning to compile for an 80386 running System V, then you
5669 would specify @option{-b i386v} to run that cross compiler.
5671 @item -V @var{version}
5673 The argument @var{version} specifies which version of GCC to run.
5674 This is useful when multiple versions are installed. For example,
5675 @var{version} might be @samp{2.0}, meaning to run GCC version 2.0.
5678 The @option{-V} and @option{-b} options work by running the
5679 @file{<machine>-gcc-<version>} executable, so there's no real reason to
5680 use them if you can just run that directly.
5682 @node Submodel Options
5683 @section Hardware Models and Configurations
5684 @cindex submodel options
5685 @cindex specifying hardware config
5686 @cindex hardware models and configurations, specifying
5687 @cindex machine dependent options
5689 Earlier we discussed the standard option @option{-b} which chooses among
5690 different installed compilers for completely different target
5691 machines, such as VAX vs.@: 68000 vs.@: 80386.
5693 In addition, each of these target machine types can have its own
5694 special options, starting with @samp{-m}, to choose among various
5695 hardware models or configurations---for example, 68010 vs 68020,
5696 floating coprocessor or none. A single installed version of the
5697 compiler can compile for any model or configuration, according to the
5700 Some configurations of the compiler also support additional special
5701 options, usually for compatibility with other compilers on the same
5704 These options are defined by the macro @code{TARGET_SWITCHES} in the
5705 machine description. The default for the options is also defined by
5706 that macro, which enables you to change the defaults.
5718 * RS/6000 and PowerPC Options::
5722 * i386 and x86-64 Options::
5724 * Intel 960 Options::
5725 * DEC Alpha Options::
5726 * DEC Alpha/VMS Options::
5729 * System V Options::
5730 * TMS320C3x/C4x Options::
5738 * S/390 and zSeries Options::
5742 * Xstormy16 Options::
5747 @node M680x0 Options
5748 @subsection M680x0 Options
5749 @cindex M680x0 options
5751 These are the @samp{-m} options defined for the 68000 series. The default
5752 values for these options depends on which style of 68000 was selected when
5753 the compiler was configured; the defaults for the most common choices are
5761 Generate output for a 68000. This is the default
5762 when the compiler is configured for 68000-based systems.
5764 Use this option for microcontrollers with a 68000 or EC000 core,
5765 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
5771 Generate output for a 68020. This is the default
5772 when the compiler is configured for 68020-based systems.
5776 Generate output containing 68881 instructions for floating point.
5777 This is the default for most 68020 systems unless @option{--nfp} was
5778 specified when the compiler was configured.
5782 Generate output for a 68030. This is the default when the compiler is
5783 configured for 68030-based systems.
5787 Generate output for a 68040. This is the default when the compiler is
5788 configured for 68040-based systems.
5790 This option inhibits the use of 68881/68882 instructions that have to be
5791 emulated by software on the 68040. Use this option if your 68040 does not
5792 have code to emulate those instructions.
5796 Generate output for a 68060. This is the default when the compiler is
5797 configured for 68060-based systems.
5799 This option inhibits the use of 68020 and 68881/68882 instructions that
5800 have to be emulated by software on the 68060. Use this option if your 68060
5801 does not have code to emulate those instructions.
5805 Generate output for a CPU32. This is the default
5806 when the compiler is configured for CPU32-based systems.
5808 Use this option for microcontrollers with a
5809 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
5810 68336, 68340, 68341, 68349 and 68360.
5814 Generate output for a 520X ``coldfire'' family cpu. This is the default
5815 when the compiler is configured for 520X-based systems.
5817 Use this option for microcontroller with a 5200 core, including
5818 the MCF5202, MCF5203, MCF5204 and MCF5202.
5823 Generate output for a 68040, without using any of the new instructions.
5824 This results in code which can run relatively efficiently on either a
5825 68020/68881 or a 68030 or a 68040. The generated code does use the
5826 68881 instructions that are emulated on the 68040.
5830 Generate output for a 68060, without using any of the new instructions.
5831 This results in code which can run relatively efficiently on either a
5832 68020/68881 or a 68030 or a 68040. The generated code does use the
5833 68881 instructions that are emulated on the 68060.
5836 @opindex msoft-float
5837 Generate output containing library calls for floating point.
5838 @strong{Warning:} the requisite libraries are not available for all m68k
5839 targets. Normally the facilities of the machine's usual C compiler are
5840 used, but this can't be done directly in cross-compilation. You must
5841 make your own arrangements to provide suitable library functions for
5842 cross-compilation. The embedded targets @samp{m68k-*-aout} and
5843 @samp{m68k-*-coff} do provide software floating point support.
5847 Consider type @code{int} to be 16 bits wide, like @code{short int}.
5850 @opindex mnobitfield
5851 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
5852 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
5856 Do use the bit-field instructions. The @option{-m68020} option implies
5857 @option{-mbitfield}. This is the default if you use a configuration
5858 designed for a 68020.
5862 Use a different function-calling convention, in which functions
5863 that take a fixed number of arguments return with the @code{rtd}
5864 instruction, which pops their arguments while returning. This
5865 saves one instruction in the caller since there is no need to pop
5866 the arguments there.
5868 This calling convention is incompatible with the one normally
5869 used on Unix, so you cannot use it if you need to call libraries
5870 compiled with the Unix compiler.
5872 Also, you must provide function prototypes for all functions that
5873 take variable numbers of arguments (including @code{printf});
5874 otherwise incorrect code will be generated for calls to those
5877 In addition, seriously incorrect code will result if you call a
5878 function with too many arguments. (Normally, extra arguments are
5879 harmlessly ignored.)
5881 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
5882 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
5885 @itemx -mno-align-int
5887 @opindex mno-align-int
5888 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
5889 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
5890 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
5891 Aligning variables on 32-bit boundaries produces code that runs somewhat
5892 faster on processors with 32-bit busses at the expense of more memory.
5894 @strong{Warning:} if you use the @option{-malign-int} switch, GCC will
5895 align structures containing the above types differently than
5896 most published application binary interface specifications for the m68k.
5900 Use the pc-relative addressing mode of the 68000 directly, instead of
5901 using a global offset table. At present, this option implies @option{-fpic},
5902 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
5903 not presently supported with @option{-mpcrel}, though this could be supported for
5904 68020 and higher processors.
5906 @item -mno-strict-align
5907 @itemx -mstrict-align
5908 @opindex mno-strict-align
5909 @opindex mstrict-align
5910 Do not (do) assume that unaligned memory references will be handled by
5915 @node M68hc1x Options
5916 @subsection M68hc1x Options
5917 @cindex M68hc1x options
5919 These are the @samp{-m} options defined for the 68hc11 and 68hc12
5920 microcontrollers. The default values for these options depends on
5921 which style of microcontroller was selected when the compiler was configured;
5922 the defaults for the most common choices are given below.
5929 Generate output for a 68HC11. This is the default
5930 when the compiler is configured for 68HC11-based systems.
5936 Generate output for a 68HC12. This is the default
5937 when the compiler is configured for 68HC12-based systems.
5943 Generate output for a 68HCS12.
5946 @opindex mauto-incdec
5947 Enable the use of 68HC12 pre and post auto-increment and auto-decrement
5954 Enable the use of 68HC12 min and max instructions.
5957 @itemx -mno-long-calls
5958 @opindex mlong-calls
5959 @opindex mno-long-calls
5960 Treat all calls as being far away (near). If calls are assumed to be
5961 far away, the compiler will use the @code{call} instruction to
5962 call a function and the @code{rtc} instruction for returning.
5966 Consider type @code{int} to be 16 bits wide, like @code{short int}.
5968 @item -msoft-reg-count=@var{count}
5969 @opindex msoft-reg-count
5970 Specify the number of pseudo-soft registers which are used for the
5971 code generation. The maximum number is 32. Using more pseudo-soft
5972 register may or may not result in better code depending on the program.
5973 The default is 4 for 68HC11 and 2 for 68HC12.
5978 @subsection VAX Options
5981 These @samp{-m} options are defined for the VAX:
5986 Do not output certain jump instructions (@code{aobleq} and so on)
5987 that the Unix assembler for the VAX cannot handle across long
5992 Do output those jump instructions, on the assumption that you
5993 will assemble with the GNU assembler.
5997 Output code for g-format floating point numbers instead of d-format.
6001 @subsection SPARC Options
6002 @cindex SPARC options
6004 These @samp{-m} switches are supported on the SPARC:
6009 @opindex mno-app-regs
6011 Specify @option{-mapp-regs} to generate output using the global registers
6012 2 through 4, which the SPARC SVR4 ABI reserves for applications. This
6015 To be fully SVR4 ABI compliant at the cost of some performance loss,
6016 specify @option{-mno-app-regs}. You should compile libraries and system
6017 software with this option.
6022 @opindex mhard-float
6023 Generate output containing floating point instructions. This is the
6029 @opindex msoft-float
6030 Generate output containing library calls for floating point.
6031 @strong{Warning:} the requisite libraries are not available for all SPARC
6032 targets. Normally the facilities of the machine's usual C compiler are
6033 used, but this cannot be done directly in cross-compilation. You must make
6034 your own arrangements to provide suitable library functions for
6035 cross-compilation. The embedded targets @samp{sparc-*-aout} and
6036 @samp{sparclite-*-*} do provide software floating point support.
6038 @option{-msoft-float} changes the calling convention in the output file;
6039 therefore, it is only useful if you compile @emph{all} of a program with
6040 this option. In particular, you need to compile @file{libgcc.a}, the
6041 library that comes with GCC, with @option{-msoft-float} in order for
6044 @item -mhard-quad-float
6045 @opindex mhard-quad-float
6046 Generate output containing quad-word (long double) floating point
6049 @item -msoft-quad-float
6050 @opindex msoft-quad-float
6051 Generate output containing library calls for quad-word (long double)
6052 floating point instructions. The functions called are those specified
6053 in the SPARC ABI@. This is the default.
6055 As of this writing, there are no sparc implementations that have hardware
6056 support for the quad-word floating point instructions. They all invoke
6057 a trap handler for one of these instructions, and then the trap handler
6058 emulates the effect of the instruction. Because of the trap handler overhead,
6059 this is much slower than calling the ABI library routines. Thus the
6060 @option{-msoft-quad-float} option is the default.
6066 With @option{-mflat}, the compiler does not generate save/restore instructions
6067 and will use a ``flat'' or single register window calling convention.
6068 This model uses %i7 as the frame pointer and is compatible with the normal
6069 register window model. Code from either may be intermixed.
6070 The local registers and the input registers (0--5) are still treated as
6071 ``call saved'' registers and will be saved on the stack as necessary.
6073 With @option{-mno-flat} (the default), the compiler emits save/restore
6074 instructions (except for leaf functions) and is the normal mode of operation.
6076 @item -mno-unaligned-doubles
6077 @itemx -munaligned-doubles
6078 @opindex mno-unaligned-doubles
6079 @opindex munaligned-doubles
6080 Assume that doubles have 8 byte alignment. This is the default.
6082 With @option{-munaligned-doubles}, GCC assumes that doubles have 8 byte
6083 alignment only if they are contained in another type, or if they have an
6084 absolute address. Otherwise, it assumes they have 4 byte alignment.
6085 Specifying this option avoids some rare compatibility problems with code
6086 generated by other compilers. It is not the default because it results
6087 in a performance loss, especially for floating point code.
6089 @item -mno-faster-structs
6090 @itemx -mfaster-structs
6091 @opindex mno-faster-structs
6092 @opindex mfaster-structs
6093 With @option{-mfaster-structs}, the compiler assumes that structures
6094 should have 8 byte alignment. This enables the use of pairs of
6095 @code{ldd} and @code{std} instructions for copies in structure
6096 assignment, in place of twice as many @code{ld} and @code{st} pairs.
6097 However, the use of this changed alignment directly violates the SPARC
6098 ABI@. Thus, it's intended only for use on targets where the developer
6099 acknowledges that their resulting code will not be directly in line with
6100 the rules of the ABI@.
6103 @opindex mimpure-text
6104 @option{-mimpure-text}, used in addition to @option{-shared}, tells
6105 the compiler to not pass @option{-z text} to the linker when linking a
6106 shared object. Using this option, you can link position-dependent
6107 code into a shared object.
6109 @option{-mimpure-text} suppresses the ``relocations remain against
6110 allocatable but non-writable sections'' linker error message.
6111 However, the necessary relocations will trigger copy-on-write, and the
6112 shared object is not actually shared across processes. Instead of
6113 using @option{-mimpure-text}, you should compile all source code with
6114 @option{-fpic} or @option{-fPIC}.
6116 This option is only available on SunOS and Solaris.
6122 These two options select variations on the SPARC architecture.
6124 By default (unless specifically configured for the Fujitsu SPARClite),
6125 GCC generates code for the v7 variant of the SPARC architecture.
6127 @option{-mv8} will give you SPARC v8 code. The only difference from v7
6128 code is that the compiler emits the integer multiply and integer
6129 divide instructions which exist in SPARC v8 but not in SPARC v7.
6131 @option{-msparclite} will give you SPARClite code. This adds the integer
6132 multiply, integer divide step and scan (@code{ffs}) instructions which
6133 exist in SPARClite but not in SPARC v7.
6135 These options are deprecated and will be deleted in a future GCC release.
6136 They have been replaced with @option{-mcpu=xxx}.
6141 @opindex msupersparc
6142 These two options select the processor for which the code is optimized.
6144 With @option{-mcypress} (the default), the compiler optimizes code for the
6145 Cypress CY7C602 chip, as used in the SPARCStation/SPARCServer 3xx series.
6146 This is also appropriate for the older SPARCStation 1, 2, IPX etc.
6148 With @option{-msupersparc} the compiler optimizes code for the SuperSPARC cpu, as
6149 used in the SPARCStation 10, 1000 and 2000 series. This flag also enables use
6150 of the full SPARC v8 instruction set.
6152 These options are deprecated and will be deleted in a future GCC release.
6153 They have been replaced with @option{-mcpu=xxx}.
6155 @item -mcpu=@var{cpu_type}
6157 Set the instruction set, register set, and instruction scheduling parameters
6158 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
6159 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite},
6160 @samp{hypersparc}, @samp{sparclite86x}, @samp{f930}, @samp{f934},
6161 @samp{sparclet}, @samp{tsc701}, @samp{v9}, @samp{ultrasparc}, and
6164 Default instruction scheduling parameters are used for values that select
6165 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
6166 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
6168 Here is a list of each supported architecture and their supported
6173 v8: supersparc, hypersparc
6174 sparclite: f930, f934, sparclite86x
6176 v9: ultrasparc, ultrasparc3
6179 @item -mtune=@var{cpu_type}
6181 Set the instruction scheduling parameters for machine type
6182 @var{cpu_type}, but do not set the instruction set or register set that the
6183 option @option{-mcpu=@var{cpu_type}} would.
6185 The same values for @option{-mcpu=@var{cpu_type}} can be used for
6186 @option{-mtune=@var{cpu_type}}, but the only useful values are those
6187 that select a particular cpu implementation. Those are @samp{cypress},
6188 @samp{supersparc}, @samp{hypersparc}, @samp{f930}, @samp{f934},
6189 @samp{sparclite86x}, @samp{tsc701}, @samp{ultrasparc}, and
6194 These @samp{-m} switches are supported in addition to the above
6195 on the SPARCLET processor.
6198 @item -mlittle-endian
6199 @opindex mlittle-endian
6200 Generate code for a processor running in little-endian mode.
6204 Treat register @code{%g0} as a normal register.
6205 GCC will continue to clobber it as necessary but will not assume
6206 it always reads as 0.
6208 @item -mbroken-saverestore
6209 @opindex mbroken-saverestore
6210 Generate code that does not use non-trivial forms of the @code{save} and
6211 @code{restore} instructions. Early versions of the SPARCLET processor do
6212 not correctly handle @code{save} and @code{restore} instructions used with
6213 arguments. They correctly handle them used without arguments. A @code{save}
6214 instruction used without arguments increments the current window pointer
6215 but does not allocate a new stack frame. It is assumed that the window
6216 overflow trap handler will properly handle this case as will interrupt
6220 These @samp{-m} switches are supported in addition to the above
6221 on SPARC V9 processors in 64-bit environments.
6224 @item -mlittle-endian
6225 @opindex mlittle-endian
6226 Generate code for a processor running in little-endian mode. It is only
6227 available for a few configurations and most notably not on Solaris.
6233 Generate code for a 32-bit or 64-bit environment.
6234 The 32-bit environment sets int, long and pointer to 32 bits.
6235 The 64-bit environment sets int to 32 bits and long and pointer
6238 @item -mcmodel=medlow
6239 @opindex mcmodel=medlow
6240 Generate code for the Medium/Low code model: the program must be linked
6241 in the low 32 bits of the address space. Pointers are 64 bits.
6242 Programs can be statically or dynamically linked.
6244 @item -mcmodel=medmid
6245 @opindex mcmodel=medmid
6246 Generate code for the Medium/Middle code model: the program must be linked
6247 in the low 44 bits of the address space, the text segment must be less than
6248 2G bytes, and data segment must be within 2G of the text segment.
6249 Pointers are 64 bits.
6251 @item -mcmodel=medany
6252 @opindex mcmodel=medany
6253 Generate code for the Medium/Anywhere code model: the program may be linked
6254 anywhere in the address space, the text segment must be less than
6255 2G bytes, and data segment must be within 2G of the text segment.
6256 Pointers are 64 bits.
6258 @item -mcmodel=embmedany
6259 @opindex mcmodel=embmedany
6260 Generate code for the Medium/Anywhere code model for embedded systems:
6261 assume a 32-bit text and a 32-bit data segment, both starting anywhere
6262 (determined at link time). Register %g4 points to the base of the
6263 data segment. Pointers are still 64 bits.
6264 Programs are statically linked, PIC is not supported.
6267 @itemx -mno-stack-bias
6268 @opindex mstack-bias
6269 @opindex mno-stack-bias
6270 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
6271 frame pointer if present, are offset by @minus{}2047 which must be added back
6272 when making stack frame references.
6273 Otherwise, assume no such offset is present.
6277 @subsection ARM Options
6280 These @samp{-m} options are defined for Advanced RISC Machines (ARM)
6285 @opindex mapcs-frame
6286 Generate a stack frame that is compliant with the ARM Procedure Call
6287 Standard for all functions, even if this is not strictly necessary for
6288 correct execution of the code. Specifying @option{-fomit-frame-pointer}
6289 with this option will cause the stack frames not to be generated for
6290 leaf functions. The default is @option{-mno-apcs-frame}.
6294 This is a synonym for @option{-mapcs-frame}.
6298 Generate code for a processor running with a 26-bit program counter,
6299 and conforming to the function calling standards for the APCS 26-bit
6300 option. This option replaces the @option{-m2} and @option{-m3} options
6301 of previous releases of the compiler.
6305 Generate code for a processor running with a 32-bit program counter,
6306 and conforming to the function calling standards for the APCS 32-bit
6307 option. This option replaces the @option{-m6} option of previous releases
6311 @c not currently implemented
6312 @item -mapcs-stack-check
6313 @opindex mapcs-stack-check
6314 Generate code to check the amount of stack space available upon entry to
6315 every function (that actually uses some stack space). If there is
6316 insufficient space available then either the function
6317 @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be
6318 called, depending upon the amount of stack space required. The run time
6319 system is required to provide these functions. The default is
6320 @option{-mno-apcs-stack-check}, since this produces smaller code.
6322 @c not currently implemented
6324 @opindex mapcs-float
6325 Pass floating point arguments using the float point registers. This is
6326 one of the variants of the APCS@. This option is recommended if the
6327 target hardware has a floating point unit or if a lot of floating point
6328 arithmetic is going to be performed by the code. The default is
6329 @option{-mno-apcs-float}, since integer only code is slightly increased in
6330 size if @option{-mapcs-float} is used.
6332 @c not currently implemented
6333 @item -mapcs-reentrant
6334 @opindex mapcs-reentrant
6335 Generate reentrant, position independent code. The default is
6336 @option{-mno-apcs-reentrant}.
6339 @item -mthumb-interwork
6340 @opindex mthumb-interwork
6341 Generate code which supports calling between the ARM and Thumb
6342 instruction sets. Without this option the two instruction sets cannot
6343 be reliably used inside one program. The default is
6344 @option{-mno-thumb-interwork}, since slightly larger code is generated
6345 when @option{-mthumb-interwork} is specified.
6347 @item -mno-sched-prolog
6348 @opindex mno-sched-prolog
6349 Prevent the reordering of instructions in the function prolog, or the
6350 merging of those instruction with the instructions in the function's
6351 body. This means that all functions will start with a recognizable set
6352 of instructions (or in fact one of a choice from a small set of
6353 different function prologues), and this information can be used to
6354 locate the start if functions inside an executable piece of code. The
6355 default is @option{-msched-prolog}.
6358 @opindex mhard-float
6359 Generate output containing floating point instructions. This is the
6363 @opindex msoft-float
6364 Generate output containing library calls for floating point.
6365 @strong{Warning:} the requisite libraries are not available for all ARM
6366 targets. Normally the facilities of the machine's usual C compiler are
6367 used, but this cannot be done directly in cross-compilation. You must make
6368 your own arrangements to provide suitable library functions for
6371 @option{-msoft-float} changes the calling convention in the output file;
6372 therefore, it is only useful if you compile @emph{all} of a program with
6373 this option. In particular, you need to compile @file{libgcc.a}, the
6374 library that comes with GCC, with @option{-msoft-float} in order for
6377 @item -mlittle-endian
6378 @opindex mlittle-endian
6379 Generate code for a processor running in little-endian mode. This is
6380 the default for all standard configurations.
6383 @opindex mbig-endian
6384 Generate code for a processor running in big-endian mode; the default is
6385 to compile code for a little-endian processor.
6387 @item -mwords-little-endian
6388 @opindex mwords-little-endian
6389 This option only applies when generating code for big-endian processors.
6390 Generate code for a little-endian word order but a big-endian byte
6391 order. That is, a byte order of the form @samp{32107654}. Note: this
6392 option should only be used if you require compatibility with code for
6393 big-endian ARM processors generated by versions of the compiler prior to
6396 @item -malignment-traps
6397 @opindex malignment-traps
6398 Generate code that will not trap if the MMU has alignment traps enabled.
6399 On ARM architectures prior to ARMv4, there were no instructions to
6400 access half-word objects stored in memory. However, when reading from
6401 memory a feature of the ARM architecture allows a word load to be used,
6402 even if the address is unaligned, and the processor core will rotate the
6403 data as it is being loaded. This option tells the compiler that such
6404 misaligned accesses will cause a MMU trap and that it should instead
6405 synthesize the access as a series of byte accesses. The compiler can
6406 still use word accesses to load half-word data if it knows that the
6407 address is aligned to a word boundary.
6409 This option is ignored when compiling for ARM architecture 4 or later,
6410 since these processors have instructions to directly access half-word
6413 @item -mno-alignment-traps
6414 @opindex mno-alignment-traps
6415 Generate code that assumes that the MMU will not trap unaligned
6416 accesses. This produces better code when the target instruction set
6417 does not have half-word memory operations (i.e.@: implementations prior to
6420 Note that you cannot use this option to access unaligned word objects,
6421 since the processor will only fetch one 32-bit aligned object from
6424 The default setting for most targets is @option{-mno-alignment-traps}, since
6425 this produces better code when there are no half-word memory
6426 instructions available.
6428 @item -mshort-load-bytes
6429 @itemx -mno-short-load-words
6430 @opindex mshort-load-bytes
6431 @opindex mno-short-load-words
6432 These are deprecated aliases for @option{-malignment-traps}.
6434 @item -mno-short-load-bytes
6435 @itemx -mshort-load-words
6436 @opindex mno-short-load-bytes
6437 @opindex mshort-load-words
6438 This are deprecated aliases for @option{-mno-alignment-traps}.
6440 @item -mcpu=@var{name}
6442 This specifies the name of the target ARM processor. GCC uses this name
6443 to determine what kind of instructions it can emit when generating
6444 assembly code. Permissible names are: @samp{arm2}, @samp{arm250},
6445 @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610},
6446 @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm},
6447 @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700},
6448 @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100},
6449 @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm8},
6450 @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100},
6451 @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920},
6452 @samp{arm920t}, @samp{arm940t}, @samp{arm9tdmi}, @samp{arm10tdmi},
6453 @samp{arm1020t}, @samp{xscale}, @samp{iwmmxt}, @samp{ep9312}.
6455 @itemx -mtune=@var{name}
6457 This option is very similar to the @option{-mcpu=} option, except that
6458 instead of specifying the actual target processor type, and hence
6459 restricting which instructions can be used, it specifies that GCC should
6460 tune the performance of the code as if the target were of the type
6461 specified in this option, but still choosing the instructions that it
6462 will generate based on the cpu specified by a @option{-mcpu=} option.
6463 For some ARM implementations better performance can be obtained by using
6466 @item -march=@var{name}
6468 This specifies the name of the target ARM architecture. GCC uses this
6469 name to determine what kind of instructions it can emit when generating
6470 assembly code. This option can be used in conjunction with or instead
6471 of the @option{-mcpu=} option. Permissible names are: @samp{armv2},
6472 @samp{armv2a}, @samp{armv3}, @samp{armv3m}, @samp{armv4}, @samp{armv4t},
6473 @samp{armv5}, @samp{armv5t}, @samp{armv5te}, @samp{iwmmxt}, @samp{ep9312}.
6475 @item -mfpe=@var{number}
6476 @itemx -mfp=@var{number}
6479 This specifies the version of the floating point emulation available on
6480 the target. Permissible values are 2 and 3. @option{-mfp=} is a synonym
6481 for @option{-mfpe=}, for compatibility with older versions of GCC@.
6483 @item -mstructure-size-boundary=@var{n}
6484 @opindex mstructure-size-boundary
6485 The size of all structures and unions will be rounded up to a multiple
6486 of the number of bits set by this option. Permissible values are 8 and
6487 32. The default value varies for different toolchains. For the COFF
6488 targeted toolchain the default value is 8. Specifying the larger number
6489 can produce faster, more efficient code, but can also increase the size
6490 of the program. The two values are potentially incompatible. Code
6491 compiled with one value cannot necessarily expect to work with code or
6492 libraries compiled with the other value, if they exchange information
6493 using structures or unions.
6495 @item -mabort-on-noreturn
6496 @opindex mabort-on-noreturn
6497 Generate a call to the function @code{abort} at the end of a
6498 @code{noreturn} function. It will be executed if the function tries to
6502 @itemx -mno-long-calls
6503 @opindex mlong-calls
6504 @opindex mno-long-calls
6505 Tells the compiler to perform function calls by first loading the
6506 address of the function into a register and then performing a subroutine
6507 call on this register. This switch is needed if the target function
6508 will lie outside of the 64 megabyte addressing range of the offset based
6509 version of subroutine call instruction.
6511 Even if this switch is enabled, not all function calls will be turned
6512 into long calls. The heuristic is that static functions, functions
6513 which have the @samp{short-call} attribute, functions that are inside
6514 the scope of a @samp{#pragma no_long_calls} directive and functions whose
6515 definitions have already been compiled within the current compilation
6516 unit, will not be turned into long calls. The exception to this rule is
6517 that weak function definitions, functions with the @samp{long-call}
6518 attribute or the @samp{section} attribute, and functions that are within
6519 the scope of a @samp{#pragma long_calls} directive, will always be
6520 turned into long calls.
6522 This feature is not enabled by default. Specifying
6523 @option{-mno-long-calls} will restore the default behavior, as will
6524 placing the function calls within the scope of a @samp{#pragma
6525 long_calls_off} directive. Note these switches have no effect on how
6526 the compiler generates code to handle function calls via function
6529 @item -mnop-fun-dllimport
6530 @opindex mnop-fun-dllimport
6531 Disable support for the @code{dllimport} attribute.
6533 @item -msingle-pic-base
6534 @opindex msingle-pic-base
6535 Treat the register used for PIC addressing as read-only, rather than
6536 loading it in the prologue for each function. The run-time system is
6537 responsible for initializing this register with an appropriate value
6538 before execution begins.
6540 @item -mpic-register=@var{reg}
6541 @opindex mpic-register
6542 Specify the register to be used for PIC addressing. The default is R10
6543 unless stack-checking is enabled, when R9 is used.
6545 @item -mcirrus-fix-invalid-insns
6546 @opindex mcirrus-fix-invalid-insns
6547 @opindex mno-cirrus-fix-invalid-insns
6548 Insert NOPs into the instruction stream to in order to work around
6549 problems with invalid Maverick instruction combinations. This option
6550 is only valid if the @option{-mcpu=ep9312} option has been used to
6551 enable generation of instructions for the Cirrus Maverick floating
6552 point co-processor. This option is not enabled by default, since the
6553 problem is only present in older Maverick implementations. The default
6554 can be re-enabled by use of the @option{-mno-cirrus-fix-invalid-insns}
6557 @item -mpoke-function-name
6558 @opindex mpoke-function-name
6559 Write the name of each function into the text section, directly
6560 preceding the function prologue. The generated code is similar to this:
6564 .ascii "arm_poke_function_name", 0
6567 .word 0xff000000 + (t1 - t0)
6568 arm_poke_function_name
6570 stmfd sp!, @{fp, ip, lr, pc@}
6574 When performing a stack backtrace, code can inspect the value of
6575 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
6576 location @code{pc - 12} and the top 8 bits are set, then we know that
6577 there is a function name embedded immediately preceding this location
6578 and has length @code{((pc[-3]) & 0xff000000)}.
6582 Generate code for the 16-bit Thumb instruction set. The default is to
6583 use the 32-bit ARM instruction set.
6586 @opindex mtpcs-frame
6587 Generate a stack frame that is compliant with the Thumb Procedure Call
6588 Standard for all non-leaf functions. (A leaf function is one that does
6589 not call any other functions.) The default is @option{-mno-tpcs-frame}.
6591 @item -mtpcs-leaf-frame
6592 @opindex mtpcs-leaf-frame
6593 Generate a stack frame that is compliant with the Thumb Procedure Call
6594 Standard for all leaf functions. (A leaf function is one that does
6595 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
6597 @item -mcallee-super-interworking
6598 @opindex mcallee-super-interworking
6599 Gives all externally visible functions in the file being compiled an ARM
6600 instruction set header which switches to Thumb mode before executing the
6601 rest of the function. This allows these functions to be called from
6602 non-interworking code.
6604 @item -mcaller-super-interworking
6605 @opindex mcaller-super-interworking
6606 Allows calls via function pointers (including virtual functions) to
6607 execute correctly regardless of whether the target code has been
6608 compiled for interworking or not. There is a small overhead in the cost
6609 of executing a function pointer if this option is enabled.
6613 @node MN10200 Options
6614 @subsection MN10200 Options
6615 @cindex MN10200 options
6617 These @option{-m} options are defined for Matsushita MN10200 architectures:
6622 Indicate to the linker that it should perform a relaxation optimization pass
6623 to shorten branches, calls and absolute memory addresses. This option only
6624 has an effect when used on the command line for the final link step.
6626 This option makes symbolic debugging impossible.
6629 @node MN10300 Options
6630 @subsection MN10300 Options
6631 @cindex MN10300 options
6633 These @option{-m} options are defined for Matsushita MN10300 architectures:
6638 Generate code to avoid bugs in the multiply instructions for the MN10300
6639 processors. This is the default.
6642 @opindex mno-mult-bug
6643 Do not generate code to avoid bugs in the multiply instructions for the
6648 Generate code which uses features specific to the AM33 processor.
6652 Do not generate code which uses features specific to the AM33 processor. This
6657 Do not link in the C run-time initialization object file.
6661 Indicate to the linker that it should perform a relaxation optimization pass
6662 to shorten branches, calls and absolute memory addresses. This option only
6663 has an effect when used on the command line for the final link step.
6665 This option makes symbolic debugging impossible.
6669 @node M32R/D Options
6670 @subsection M32R/D Options
6671 @cindex M32R/D options
6673 These @option{-m} options are defined for Mitsubishi M32R/D architectures:
6678 Generate code for the M32R/X@.
6682 Generate code for the M32R@. This is the default.
6684 @item -mcode-model=small
6685 @opindex mcode-model=small
6686 Assume all objects live in the lower 16MB of memory (so that their addresses
6687 can be loaded with the @code{ld24} instruction), and assume all subroutines
6688 are reachable with the @code{bl} instruction.
6689 This is the default.
6691 The addressability of a particular object can be set with the
6692 @code{model} attribute.
6694 @item -mcode-model=medium
6695 @opindex mcode-model=medium
6696 Assume objects may be anywhere in the 32-bit address space (the compiler
6697 will generate @code{seth/add3} instructions to load their addresses), and
6698 assume all subroutines are reachable with the @code{bl} instruction.
6700 @item -mcode-model=large
6701 @opindex mcode-model=large
6702 Assume objects may be anywhere in the 32-bit address space (the compiler
6703 will generate @code{seth/add3} instructions to load their addresses), and
6704 assume subroutines may not be reachable with the @code{bl} instruction
6705 (the compiler will generate the much slower @code{seth/add3/jl}
6706 instruction sequence).
6709 @opindex msdata=none
6710 Disable use of the small data area. Variables will be put into
6711 one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the
6712 @code{section} attribute has been specified).
6713 This is the default.
6715 The small data area consists of sections @samp{.sdata} and @samp{.sbss}.
6716 Objects may be explicitly put in the small data area with the
6717 @code{section} attribute using one of these sections.
6720 @opindex msdata=sdata
6721 Put small global and static data in the small data area, but do not
6722 generate special code to reference them.
6726 Put small global and static data in the small data area, and generate
6727 special instructions to reference them.
6731 @cindex smaller data references
6732 Put global and static objects less than or equal to @var{num} bytes
6733 into the small data or bss sections instead of the normal data or bss
6734 sections. The default value of @var{num} is 8.
6735 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
6736 for this option to have any effect.
6738 All modules should be compiled with the same @option{-G @var{num}} value.
6739 Compiling with different values of @var{num} may or may not work; if it
6740 doesn't the linker will give an error message---incorrect code will not be
6746 @subsection M88K Options
6747 @cindex M88k options
6749 These @samp{-m} options are defined for Motorola 88k architectures:
6754 Generate code that works well on both the m88100 and the
6759 Generate code that works best for the m88100, but that also
6764 Generate code that works best for the m88110, and may not run
6769 Obsolete option to be removed from the next revision.
6772 @item -midentify-revision
6773 @opindex midentify-revision
6774 @cindex identifying source, compiler (88k)
6775 Include an @code{ident} directive in the assembler output recording the
6776 source file name, compiler name and version, timestamp, and compilation
6779 @item -mno-underscores
6780 @opindex mno-underscores
6781 @cindex underscores, avoiding (88k)
6782 In assembler output, emit symbol names without adding an underscore
6783 character at the beginning of each name. The default is to use an
6784 underscore as prefix on each name.
6786 @item -mocs-debug-info
6787 @itemx -mno-ocs-debug-info
6788 @opindex mocs-debug-info
6789 @opindex mno-ocs-debug-info
6791 @cindex debugging, 88k OCS
6792 Include (or omit) additional debugging information (about registers used
6793 in each stack frame) as specified in the 88open Object Compatibility
6794 Standard, ``OCS''@. This extra information allows debugging of code that
6795 has had the frame pointer eliminated. The default for SVr4 and Delta 88
6796 SVr3.2 is to include this information; other 88k configurations omit this
6797 information by default.
6799 @item -mocs-frame-position
6800 @opindex mocs-frame-position
6801 @cindex register positions in frame (88k)
6802 When emitting COFF debugging information for automatic variables and
6803 parameters stored on the stack, use the offset from the canonical frame
6804 address, which is the stack pointer (register 31) on entry to the
6805 function. The SVr4 and Delta88 SVr3.2, and BCS configurations use
6806 @option{-mocs-frame-position}; other 88k configurations have the default
6807 @option{-mno-ocs-frame-position}.
6809 @item -mno-ocs-frame-position
6810 @opindex mno-ocs-frame-position
6811 @cindex register positions in frame (88k)
6812 When emitting COFF debugging information for automatic variables and
6813 parameters stored on the stack, use the offset from the frame pointer
6814 register (register 30). When this option is in effect, the frame
6815 pointer is not eliminated when debugging information is selected by the
6818 @item -moptimize-arg-area
6819 @opindex moptimize-arg-area
6820 @cindex arguments in frame (88k)
6821 Save space by reorganizing the stack frame. This option generates code
6822 that does not agree with the 88open specifications, but uses less
6825 @itemx -mno-optimize-arg-area
6826 @opindex mno-optimize-arg-area
6827 Do not reorganize the stack frame to save space. This is the default.
6828 The generated conforms to the specification, but uses more memory.
6830 @item -mshort-data-@var{num}
6831 @opindex mshort-data
6832 @cindex smaller data references (88k)
6833 @cindex r0-relative references (88k)
6834 Generate smaller data references by making them relative to @code{r0},
6835 which allows loading a value using a single instruction (rather than the
6836 usual two). You control which data references are affected by
6837 specifying @var{num} with this option. For example, if you specify
6838 @option{-mshort-data-512}, then the data references affected are those
6839 involving displacements of less than 512 bytes.
6840 @option{-mshort-data-@var{num}} is not effective for @var{num} greater
6843 @item -mserialize-volatile
6844 @opindex mserialize-volatile
6845 @itemx -mno-serialize-volatile
6846 @opindex mno-serialize-volatile
6847 @cindex sequential consistency on 88k
6848 Do, or don't, generate code to guarantee sequential consistency
6849 of volatile memory references. By default, consistency is
6852 The order of memory references made by the MC88110 processor does
6853 not always match the order of the instructions requesting those
6854 references. In particular, a load instruction may execute before
6855 a preceding store instruction. Such reordering violates
6856 sequential consistency of volatile memory references, when there
6857 are multiple processors. When consistency must be guaranteed,
6858 GCC generates special instructions, as needed, to force
6859 execution in the proper order.
6861 The MC88100 processor does not reorder memory references and so
6862 always provides sequential consistency. However, by default, GCC
6863 generates the special instructions to guarantee consistency
6864 even when you use @option{-m88100}, so that the code may be run on an
6865 MC88110 processor. If you intend to run your code only on the
6866 MC88100 processor, you may use @option{-mno-serialize-volatile}.
6868 The extra code generated to guarantee consistency may affect the
6869 performance of your application. If you know that you can safely
6870 forgo this guarantee, you may use @option{-mno-serialize-volatile}.
6876 @cindex assembler syntax, 88k
6878 Turn on (@option{-msvr4}) or off (@option{-msvr3}) compiler extensions
6879 related to System V release 4 (SVr4). This controls the following:
6883 Which variant of the assembler syntax to emit.
6885 @option{-msvr4} makes the C preprocessor recognize @samp{#pragma weak}
6886 that is used on System V release 4.
6888 @option{-msvr4} makes GCC issue additional declaration directives used in
6892 @option{-msvr4} is the default for the m88k-motorola-sysv4 configuration.
6893 @option{-msvr3} is the default for all other m88k configurations.
6895 @item -mversion-03.00
6896 @opindex mversion-03.00
6897 This option is obsolete, and is ignored.
6898 @c ??? which asm syntax better for GAS? option there too?
6900 @item -mno-check-zero-division
6901 @itemx -mcheck-zero-division
6902 @opindex mno-check-zero-division
6903 @opindex mcheck-zero-division
6904 @cindex zero division on 88k
6905 Do, or don't, generate code to guarantee that integer division by
6906 zero will be detected. By default, detection is guaranteed.
6908 Some models of the MC88100 processor fail to trap upon integer
6909 division by zero under certain conditions. By default, when
6910 compiling code that might be run on such a processor, GCC
6911 generates code that explicitly checks for zero-valued divisors
6912 and traps with exception number 503 when one is detected. Use of
6913 @option{-mno-check-zero-division} suppresses such checking for code
6914 generated to run on an MC88100 processor.
6916 GCC assumes that the MC88110 processor correctly detects all instances
6917 of integer division by zero. When @option{-m88110} is specified, no
6918 explicit checks for zero-valued divisors are generated, and both
6919 @option{-mcheck-zero-division} and @option{-mno-check-zero-division} are
6922 @item -muse-div-instruction
6923 @opindex muse-div-instruction
6924 @cindex divide instruction, 88k
6925 Use the div instruction for signed integer division on the
6926 MC88100 processor. By default, the div instruction is not used.
6928 On the MC88100 processor the signed integer division instruction
6929 div) traps to the operating system on a negative operand. The
6930 operating system transparently completes the operation, but at a
6931 large cost in execution time. By default, when compiling code
6932 that might be run on an MC88100 processor, GCC emulates signed
6933 integer division using the unsigned integer division instruction
6934 divu), thereby avoiding the large penalty of a trap to the
6935 operating system. Such emulation has its own, smaller, execution
6936 cost in both time and space. To the extent that your code's
6937 important signed integer division operations are performed on two
6938 nonnegative operands, it may be desirable to use the div
6939 instruction directly.
6941 On the MC88110 processor the div instruction (also known as the
6942 divs instruction) processes negative operands without trapping to
6943 the operating system. When @option{-m88110} is specified,
6944 @option{-muse-div-instruction} is ignored, and the div instruction is used
6945 for signed integer division.
6947 Note that the result of dividing @code{INT_MIN} by @minus{}1 is undefined. In
6948 particular, the behavior of such a division with and without
6949 @option{-muse-div-instruction} may differ.
6951 @item -mtrap-large-shift
6952 @itemx -mhandle-large-shift
6953 @opindex mtrap-large-shift
6954 @opindex mhandle-large-shift
6955 @cindex bit shift overflow (88k)
6956 @cindex large bit shifts (88k)
6957 Include code to detect bit-shifts of more than 31 bits; respectively,
6958 trap such shifts or emit code to handle them properly. By default GCC
6959 makes no special provision for large bit shifts.
6961 @item -mwarn-passed-structs
6962 @opindex mwarn-passed-structs
6963 @cindex structure passing (88k)
6964 Warn when a function passes a struct as an argument or result.
6965 Structure-passing conventions have changed during the evolution of the C
6966 language, and are often the source of portability problems. By default,
6967 GCC issues no such warning.
6970 @c break page here to avoid unsightly interparagraph stretch.
6974 @node RS/6000 and PowerPC Options
6975 @subsection IBM RS/6000 and PowerPC Options
6976 @cindex RS/6000 and PowerPC Options
6977 @cindex IBM RS/6000 and PowerPC Options
6979 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
6987 @itemx -mpowerpc-gpopt
6988 @itemx -mno-powerpc-gpopt
6989 @itemx -mpowerpc-gfxopt
6990 @itemx -mno-powerpc-gfxopt
6992 @itemx -mno-powerpc64
6998 @opindex mno-powerpc
6999 @opindex mpowerpc-gpopt
7000 @opindex mno-powerpc-gpopt
7001 @opindex mpowerpc-gfxopt
7002 @opindex mno-powerpc-gfxopt
7004 @opindex mno-powerpc64
7005 GCC supports two related instruction set architectures for the
7006 RS/6000 and PowerPC@. The @dfn{POWER} instruction set are those
7007 instructions supported by the @samp{rios} chip set used in the original
7008 RS/6000 systems and the @dfn{PowerPC} instruction set is the
7009 architecture of the Motorola MPC5xx, MPC6xx, MPC8xx microprocessors, and
7010 the IBM 4xx microprocessors.
7012 Neither architecture is a subset of the other. However there is a
7013 large common subset of instructions supported by both. An MQ
7014 register is included in processors supporting the POWER architecture.
7016 You use these options to specify which instructions are available on the
7017 processor you are using. The default value of these options is
7018 determined when configuring GCC@. Specifying the
7019 @option{-mcpu=@var{cpu_type}} overrides the specification of these
7020 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
7021 rather than the options listed above.
7023 The @option{-mpower} option allows GCC to generate instructions that
7024 are found only in the POWER architecture and to use the MQ register.
7025 Specifying @option{-mpower2} implies @option{-power} and also allows GCC
7026 to generate instructions that are present in the POWER2 architecture but
7027 not the original POWER architecture.
7029 The @option{-mpowerpc} option allows GCC to generate instructions that
7030 are found only in the 32-bit subset of the PowerPC architecture.
7031 Specifying @option{-mpowerpc-gpopt} implies @option{-mpowerpc} and also allows
7032 GCC to use the optional PowerPC architecture instructions in the
7033 General Purpose group, including floating-point square root. Specifying
7034 @option{-mpowerpc-gfxopt} implies @option{-mpowerpc} and also allows GCC to
7035 use the optional PowerPC architecture instructions in the Graphics
7036 group, including floating-point select.
7038 The @option{-mpowerpc64} option allows GCC to generate the additional
7039 64-bit instructions that are found in the full PowerPC64 architecture
7040 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
7041 @option{-mno-powerpc64}.
7043 If you specify both @option{-mno-power} and @option{-mno-powerpc}, GCC
7044 will use only the instructions in the common subset of both
7045 architectures plus some special AIX common-mode calls, and will not use
7046 the MQ register. Specifying both @option{-mpower} and @option{-mpowerpc}
7047 permits GCC to use any instruction from either architecture and to
7048 allow use of the MQ register; specify this for the Motorola MPC601.
7050 @item -mnew-mnemonics
7051 @itemx -mold-mnemonics
7052 @opindex mnew-mnemonics
7053 @opindex mold-mnemonics
7054 Select which mnemonics to use in the generated assembler code. With
7055 @option{-mnew-mnemonics}, GCC uses the assembler mnemonics defined for
7056 the PowerPC architecture. With @option{-mold-mnemonics} it uses the
7057 assembler mnemonics defined for the POWER architecture. Instructions
7058 defined in only one architecture have only one mnemonic; GCC uses that
7059 mnemonic irrespective of which of these options is specified.
7061 GCC defaults to the mnemonics appropriate for the architecture in
7062 use. Specifying @option{-mcpu=@var{cpu_type}} sometimes overrides the
7063 value of these option. Unless you are building a cross-compiler, you
7064 should normally not specify either @option{-mnew-mnemonics} or
7065 @option{-mold-mnemonics}, but should instead accept the default.
7067 @item -mcpu=@var{cpu_type}
7069 Set architecture type, register usage, choice of mnemonics, and
7070 instruction scheduling parameters for machine type @var{cpu_type}.
7071 Supported values for @var{cpu_type} are @samp{rios}, @samp{rios1},
7072 @samp{rsc}, @samp{rios2}, @samp{rs64a}, @samp{601}, @samp{602},
7073 @samp{603}, @samp{603e}, @samp{604}, @samp{604e}, @samp{620},
7074 @samp{630}, @samp{740}, @samp{7400}, @samp{7450}, @samp{750},
7075 @samp{power}, @samp{power2}, @samp{powerpc}, @samp{403}, @samp{505},
7076 @samp{801}, @samp{821}, @samp{823}, and @samp{860} and @samp{common}.
7078 @option{-mcpu=common} selects a completely generic processor. Code
7079 generated under this option will run on any POWER or PowerPC processor.
7080 GCC will use only the instructions in the common subset of both
7081 architectures, and will not use the MQ register. GCC assumes a generic
7082 processor model for scheduling purposes.
7084 @option{-mcpu=power}, @option{-mcpu=power2}, @option{-mcpu=powerpc}, and
7085 @option{-mcpu=powerpc64} specify generic POWER, POWER2, pure 32-bit
7086 PowerPC (i.e., not MPC601), and 64-bit PowerPC architecture machine
7087 types, with an appropriate, generic processor model assumed for
7088 scheduling purposes.
7090 The other options specify a specific processor. Code generated under
7091 those options will run best on that processor, and may not run at all on
7094 The @option{-mcpu} options automatically enable or disable other
7095 @option{-m} options as follows:
7099 @option{-mno-power}, @option{-mno-powerpc}
7106 @option{-mpower}, @option{-mno-powerpc}, @option{-mno-new-mnemonics}
7121 @option{-mno-power}, @option{-mpowerpc}, @option{-mnew-mnemonics}
7124 @option{-mpower}, @option{-mpowerpc}, @option{-mnew-mnemonics}
7129 @option{-mno-power}, @option{-mpowerpc}, @option{-mnew-mnemonics}, @option{-msoft-float}
7132 @item -mtune=@var{cpu_type}
7134 Set the instruction scheduling parameters for machine type
7135 @var{cpu_type}, but do not set the architecture type, register usage, or
7136 choice of mnemonics, as @option{-mcpu=@var{cpu_type}} would. The same
7137 values for @var{cpu_type} are used for @option{-mtune} as for
7138 @option{-mcpu}. If both are specified, the code generated will use the
7139 architecture, registers, and mnemonics set by @option{-mcpu}, but the
7140 scheduling parameters set by @option{-mtune}.
7145 @opindex mno-altivec
7146 These switches enable or disable the use of built-in functions that
7147 allow access to the AltiVec instruction set. You may also need to set
7148 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
7153 Extend the current ABI with SPE ABI extensions. This does not change
7154 the default ABI, instead it adds the SPE ABI extensions to the current
7158 @opindex mabi=no-spe
7159 Disable Booke SPE ABI extensions for the current ABI.
7161 @item -misel=@var{yes/no}
7164 This switch enables or disables the generation of ISEL instructions.
7166 @item -mspe=@var{yes/no}
7169 This switch enables or disables the generation of SPE simd
7172 @item -mfloat-gprs=@var{yes/no}
7174 @opindex mfloat-gprs
7175 This switch enables or disables the generation of floating point
7176 operations on the general purpose registers for architectures that
7177 support it. This option is currently only available on the MPC8540.
7180 @itemx -mno-fp-in-toc
7181 @itemx -mno-sum-in-toc
7182 @itemx -mminimal-toc
7184 @opindex mno-fp-in-toc
7185 @opindex mno-sum-in-toc
7186 @opindex mminimal-toc
7187 Modify generation of the TOC (Table Of Contents), which is created for
7188 every executable file. The @option{-mfull-toc} option is selected by
7189 default. In that case, GCC will allocate at least one TOC entry for
7190 each unique non-automatic variable reference in your program. GCC
7191 will also place floating-point constants in the TOC@. However, only
7192 16,384 entries are available in the TOC@.
7194 If you receive a linker error message that saying you have overflowed
7195 the available TOC space, you can reduce the amount of TOC space used
7196 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
7197 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
7198 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
7199 generate code to calculate the sum of an address and a constant at
7200 run-time instead of putting that sum into the TOC@. You may specify one
7201 or both of these options. Each causes GCC to produce very slightly
7202 slower and larger code at the expense of conserving TOC space.
7204 If you still run out of space in the TOC even when you specify both of
7205 these options, specify @option{-mminimal-toc} instead. This option causes
7206 GCC to make only one TOC entry for every file. When you specify this
7207 option, GCC will produce code that is slower and larger but which
7208 uses extremely little TOC space. You may wish to use this option
7209 only on files that contain less frequently executed code.
7215 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
7216 @code{long} type, and the infrastructure needed to support them.
7217 Specifying @option{-maix64} implies @option{-mpowerpc64} and
7218 @option{-mpowerpc}, while @option{-maix32} disables the 64-bit ABI and
7219 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
7224 @opindex mno-xl-call
7225 On AIX, pass floating-point arguments to prototyped functions beyond the
7226 register save area (RSA) on the stack in addition to argument FPRs. The
7227 AIX calling convention was extended but not initially documented to
7228 handle an obscure K&R C case of calling a function that takes the
7229 address of its arguments with fewer arguments than declared. AIX XL
7230 compilers access floating point arguments which do not fit in the
7231 RSA from the stack when a subroutine is compiled without
7232 optimization. Because always storing floating-point arguments on the
7233 stack is inefficient and rarely needed, this option is not enabled by
7234 default and only is necessary when calling subroutines compiled by AIX
7235 XL compilers without optimization.
7239 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
7240 application written to use message passing with special startup code to
7241 enable the application to run. The system must have PE installed in the
7242 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
7243 must be overridden with the @option{-specs=} option to specify the
7244 appropriate directory location. The Parallel Environment does not
7245 support threads, so the @option{-mpe} option and the @option{-pthread}
7246 option are incompatible.
7248 @item -malign-natural
7249 @itemx -malign-power
7250 @opindex malign-natural
7251 @opindex malign-power
7252 On AIX, Darwin, and 64-bit PowerPC Linux, the option
7253 @option{-malign-natural} overrides the ABI-defined alignment of larger
7254 types, such as floating-point doubles, on their natural size-based boundary.
7255 The option @option{-malign-power} instructs GCC to follow the ABI-specified
7256 alignment rules. GCC defaults to the standard alignment defined in the ABI.
7260 @opindex msoft-float
7261 @opindex mhard-float
7262 Generate code that does not use (uses) the floating-point register set.
7263 Software floating point emulation is provided if you use the
7264 @option{-msoft-float} option, and pass the option to GCC when linking.
7267 @itemx -mno-multiple
7269 @opindex mno-multiple
7270 Generate code that uses (does not use) the load multiple word
7271 instructions and the store multiple word instructions. These
7272 instructions are generated by default on POWER systems, and not
7273 generated on PowerPC systems. Do not use @option{-mmultiple} on little
7274 endian PowerPC systems, since those instructions do not work when the
7275 processor is in little endian mode. The exceptions are PPC740 and
7276 PPC750 which permit the instructions usage in little endian mode.
7282 Generate code that uses (does not use) the load string instructions
7283 and the store string word instructions to save multiple registers and
7284 do small block moves. These instructions are generated by default on
7285 POWER systems, and not generated on PowerPC systems. Do not use
7286 @option{-mstring} on little endian PowerPC systems, since those
7287 instructions do not work when the processor is in little endian mode.
7288 The exceptions are PPC740 and PPC750 which permit the instructions
7289 usage in little endian mode.
7295 Generate code that uses (does not use) the load or store instructions
7296 that update the base register to the address of the calculated memory
7297 location. These instructions are generated by default. If you use
7298 @option{-mno-update}, there is a small window between the time that the
7299 stack pointer is updated and the address of the previous frame is
7300 stored, which means code that walks the stack frame across interrupts or
7301 signals may get corrupted data.
7304 @itemx -mno-fused-madd
7305 @opindex mfused-madd
7306 @opindex mno-fused-madd
7307 Generate code that uses (does not use) the floating point multiply and
7308 accumulate instructions. These instructions are generated by default if
7309 hardware floating is used.
7311 @item -mno-bit-align
7313 @opindex mno-bit-align
7315 On System V.4 and embedded PowerPC systems do not (do) force structures
7316 and unions that contain bit-fields to be aligned to the base type of the
7319 For example, by default a structure containing nothing but 8
7320 @code{unsigned} bit-fields of length 1 would be aligned to a 4 byte
7321 boundary and have a size of 4 bytes. By using @option{-mno-bit-align},
7322 the structure would be aligned to a 1 byte boundary and be one byte in
7325 @item -mno-strict-align
7326 @itemx -mstrict-align
7327 @opindex mno-strict-align
7328 @opindex mstrict-align
7329 On System V.4 and embedded PowerPC systems do not (do) assume that
7330 unaligned memory references will be handled by the system.
7333 @itemx -mno-relocatable
7334 @opindex mrelocatable
7335 @opindex mno-relocatable
7336 On embedded PowerPC systems generate code that allows (does not allow)
7337 the program to be relocated to a different address at runtime. If you
7338 use @option{-mrelocatable} on any module, all objects linked together must
7339 be compiled with @option{-mrelocatable} or @option{-mrelocatable-lib}.
7341 @item -mrelocatable-lib
7342 @itemx -mno-relocatable-lib
7343 @opindex mrelocatable-lib
7344 @opindex mno-relocatable-lib
7345 On embedded PowerPC systems generate code that allows (does not allow)
7346 the program to be relocated to a different address at runtime. Modules
7347 compiled with @option{-mrelocatable-lib} can be linked with either modules
7348 compiled without @option{-mrelocatable} and @option{-mrelocatable-lib} or
7349 with modules compiled with the @option{-mrelocatable} options.
7355 On System V.4 and embedded PowerPC systems do not (do) assume that
7356 register 2 contains a pointer to a global area pointing to the addresses
7357 used in the program.
7360 @itemx -mlittle-endian
7362 @opindex mlittle-endian
7363 On System V.4 and embedded PowerPC systems compile code for the
7364 processor in little endian mode. The @option{-mlittle-endian} option is
7365 the same as @option{-mlittle}.
7370 @opindex mbig-endian
7371 On System V.4 and embedded PowerPC systems compile code for the
7372 processor in big endian mode. The @option{-mbig-endian} option is
7373 the same as @option{-mbig}.
7375 @item -mdynamic-no-pic
7376 @opindex mdynamic-no-pic
7377 On Darwin and Mac OS X systems, compile code so that it is not
7378 relocatable, but that its external references are relocatable. The
7379 resulting code is suitable for applications, but not shared
7384 On System V.4 and embedded PowerPC systems compile code using calling
7385 conventions that adheres to the March 1995 draft of the System V
7386 Application Binary Interface, PowerPC processor supplement. This is the
7387 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
7389 @item -mcall-sysv-eabi
7390 @opindex mcall-sysv-eabi
7391 Specify both @option{-mcall-sysv} and @option{-meabi} options.
7393 @item -mcall-sysv-noeabi
7394 @opindex mcall-sysv-noeabi
7395 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
7397 @item -mcall-solaris
7398 @opindex mcall-solaris
7399 On System V.4 and embedded PowerPC systems compile code for the Solaris
7403 @opindex mcall-linux
7404 On System V.4 and embedded PowerPC systems compile code for the
7405 Linux-based GNU system.
7409 On System V.4 and embedded PowerPC systems compile code for the
7410 Hurd-based GNU system.
7413 @opindex mcall-netbsd
7414 On System V.4 and embedded PowerPC systems compile code for the
7415 NetBSD operating system.
7417 @item -maix-struct-return
7418 @opindex maix-struct-return
7419 Return all structures in memory (as specified by the AIX ABI)@.
7421 @item -msvr4-struct-return
7422 @opindex msvr4-struct-return
7423 Return structures smaller than 8 bytes in registers (as specified by the
7427 @opindex mabi=altivec
7428 Extend the current ABI with AltiVec ABI extensions. This does not
7429 change the default ABI, instead it adds the AltiVec ABI extensions to
7432 @item -mabi=no-altivec
7433 @opindex mabi=no-altivec
7434 Disable AltiVec ABI extensions for the current ABI.
7437 @itemx -mno-prototype
7439 @opindex mno-prototype
7440 On System V.4 and embedded PowerPC systems assume that all calls to
7441 variable argument functions are properly prototyped. Otherwise, the
7442 compiler must insert an instruction before every non prototyped call to
7443 set or clear bit 6 of the condition code register (@var{CR}) to
7444 indicate whether floating point values were passed in the floating point
7445 registers in case the function takes a variable arguments. With
7446 @option{-mprototype}, only calls to prototyped variable argument functions
7447 will set or clear the bit.
7451 On embedded PowerPC systems, assume that the startup module is called
7452 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
7453 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}.
7458 On embedded PowerPC systems, assume that the startup module is called
7459 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
7464 On embedded PowerPC systems, assume that the startup module is called
7465 @file{crt0.o} and the standard C libraries are @file{libads.a} and
7469 @opindex myellowknife
7470 On embedded PowerPC systems, assume that the startup module is called
7471 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
7476 On System V.4 and embedded PowerPC systems, specify that you are
7477 compiling for a VxWorks system.
7481 Specify that you are compiling for the WindISS simulation environment.
7485 On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags
7486 header to indicate that @samp{eabi} extended relocations are used.
7492 On System V.4 and embedded PowerPC systems do (do not) adhere to the
7493 Embedded Applications Binary Interface (eabi) which is a set of
7494 modifications to the System V.4 specifications. Selecting @option{-meabi}
7495 means that the stack is aligned to an 8 byte boundary, a function
7496 @code{__eabi} is called to from @code{main} to set up the eabi
7497 environment, and the @option{-msdata} option can use both @code{r2} and
7498 @code{r13} to point to two separate small data areas. Selecting
7499 @option{-mno-eabi} means that the stack is aligned to a 16 byte boundary,
7500 do not call an initialization function from @code{main}, and the
7501 @option{-msdata} option will only use @code{r13} to point to a single
7502 small data area. The @option{-meabi} option is on by default if you
7503 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
7506 @opindex msdata=eabi
7507 On System V.4 and embedded PowerPC systems, put small initialized
7508 @code{const} global and static data in the @samp{.sdata2} section, which
7509 is pointed to by register @code{r2}. Put small initialized
7510 non-@code{const} global and static data in the @samp{.sdata} section,
7511 which is pointed to by register @code{r13}. Put small uninitialized
7512 global and static data in the @samp{.sbss} section, which is adjacent to
7513 the @samp{.sdata} section. The @option{-msdata=eabi} option is
7514 incompatible with the @option{-mrelocatable} option. The
7515 @option{-msdata=eabi} option also sets the @option{-memb} option.
7518 @opindex msdata=sysv
7519 On System V.4 and embedded PowerPC systems, put small global and static
7520 data in the @samp{.sdata} section, which is pointed to by register
7521 @code{r13}. Put small uninitialized global and static data in the
7522 @samp{.sbss} section, which is adjacent to the @samp{.sdata} section.
7523 The @option{-msdata=sysv} option is incompatible with the
7524 @option{-mrelocatable} option.
7526 @item -msdata=default
7528 @opindex msdata=default
7530 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
7531 compile code the same as @option{-msdata=eabi}, otherwise compile code the
7532 same as @option{-msdata=sysv}.
7535 @opindex msdata-data
7536 On System V.4 and embedded PowerPC systems, put small global and static
7537 data in the @samp{.sdata} section. Put small uninitialized global and
7538 static data in the @samp{.sbss} section. Do not use register @code{r13}
7539 to address small data however. This is the default behavior unless
7540 other @option{-msdata} options are used.
7544 @opindex msdata=none
7546 On embedded PowerPC systems, put all initialized global and static data
7547 in the @samp{.data} section, and all uninitialized data in the
7548 @samp{.bss} section.
7552 @cindex smaller data references (PowerPC)
7553 @cindex .sdata/.sdata2 references (PowerPC)
7554 On embedded PowerPC systems, put global and static items less than or
7555 equal to @var{num} bytes into the small data or bss sections instead of
7556 the normal data or bss section. By default, @var{num} is 8. The
7557 @option{-G @var{num}} switch is also passed to the linker.
7558 All modules should be compiled with the same @option{-G @var{num}} value.
7561 @itemx -mno-regnames
7563 @opindex mno-regnames
7564 On System V.4 and embedded PowerPC systems do (do not) emit register
7565 names in the assembly language output using symbolic forms.
7568 @itemx -mno-longcall
7570 @opindex mno-longcall
7571 Default to making all function calls via pointers, so that functions
7572 which reside further than 64 megabytes (67,108,864 bytes) from the
7573 current location can be called. This setting can be overridden by the
7574 @code{shortcall} function attribute, or by @code{#pragma longcall(0)}.
7576 Some linkers are capable of detecting out-of-range calls and generating
7577 glue code on the fly. On these systems, long calls are unnecessary and
7578 generate slower code. As of this writing, the AIX linker can do this,
7579 as can the GNU linker for PowerPC/64. It is planned to add this feature
7580 to the GNU linker for 32-bit PowerPC systems as well.
7582 In the future, we may cause GCC to ignore all longcall specifications
7583 when the linker is known to generate glue.
7587 Adds support for multithreading with the @dfn{pthreads} library.
7588 This option sets flags for both the preprocessor and linker.
7592 @node Darwin Options
7593 @subsection Darwin Options
7594 @cindex Darwin options
7596 These options are defined for all architectures running the Darwin operating
7597 system. They are useful for compatibility with other Mac OS compilers.
7602 Loads all members of static archive libraries.
7603 See man ld(1) for more information.
7605 @item -arch_errors_fatal
7606 @opindex arch_errors_fatal
7607 Cause the errors having to do with files that have the wrong architecture
7611 @opindex bind_at_load
7612 Causes the output file to be marked such that the dynamic linker will
7613 bind all undefined references when the file is loaded or launched.
7617 Produce a Mach-o bundle format file.
7618 See man ld(1) for more information.
7620 @item -bundle_loader @var{executable}
7621 @opindex bundle_loader
7622 This specifies the @var{executable} that will be loading the build
7623 output file being linked. See man ld(1) for more information.
7625 @item -allowable_client @var{client_name}
7629 @item -compatibility_version
7630 @item -current_version
7631 @item -dependency-file
7633 @item -dylinker_install_name
7636 @item -exported_symbols_list
7638 @item -flat_namespace
7639 @item -force_cpusubtype_ALL
7640 @item -force_flat_namespace
7641 @item -headerpad_max_install_names
7645 @item -keep_private_externs
7647 @item -multiply_defined
7648 @item -multiply_defined_unused
7652 @item -noseglinkedit
7653 @item -pagezero_size
7655 @item -prebind_all_twolevel_modules
7656 @item -private_bundle
7657 @item -read_only_relocs
7659 @item -sectobjectsymbols
7663 @item -sectobjectsymbols
7665 @item -seg_addr_table
7666 @item -seg_addr_table_filename
7669 @item -segs_read_only_addr
7670 @item -segs_read_write_addr
7671 @item -single_module
7675 @item -twolevel_namespace
7678 @item -unexported_symbols_list
7679 @item -weak_reference_mismatches
7682 @opindex allowable_client
7684 @opindex client_name
7685 @opindex compatibility_version
7686 @opindex current_version
7687 @opindex dependency-file
7689 @opindex dylinker_install_name
7692 @opindex exported_symbols_list
7694 @opindex flat_namespace
7695 @opindex force_cpusubtype_ALL
7696 @opindex force_flat_namespace
7697 @opindex headerpad_max_install_names
7700 @opindex install_name
7701 @opindex keep_private_externs
7702 @opindex multi_module
7703 @opindex multiply_defined
7704 @opindex multiply_defined_unused
7706 @opindex nomultidefs
7708 @opindex noseglinkedit
7709 @opindex pagezero_size
7711 @opindex prebind_all_twolevel_modules
7712 @opindex private_bundle
7713 @opindex read_only_relocs
7715 @opindex sectobjectsymbols
7719 @opindex sectobjectsymbols
7721 @opindex seg_addr_table
7722 @opindex seg_addr_table_filename
7723 @opindex seglinkedit
7725 @opindex segs_read_only_addr
7726 @opindex segs_read_write_addr
7727 @opindex single_module
7729 @opindex sub_library
7730 @opindex sub_umbrella
7731 @opindex twolevel_namespace
7734 @opindex unexported_symbols_list
7735 @opindex weak_reference_mismatches
7736 @opindex whatsloaded
7738 This options are available for Darwin linker. Darwin linker man page
7739 describes them in detail.
7744 @subsection IBM RT Options
7746 @cindex IBM RT options
7748 These @samp{-m} options are defined for the IBM RT PC:
7752 @opindex min-line-mul
7753 Use an in-line code sequence for integer multiplies. This is the
7756 @item -mcall-lib-mul
7757 @opindex mcall-lib-mul
7758 Call @code{lmul$$} for integer multiples.
7760 @item -mfull-fp-blocks
7761 @opindex mfull-fp-blocks
7762 Generate full-size floating point data blocks, including the minimum
7763 amount of scratch space recommended by IBM@. This is the default.
7765 @item -mminimum-fp-blocks
7766 @opindex mminimum-fp-blocks
7767 Do not include extra scratch space in floating point data blocks. This
7768 results in smaller code, but slower execution, since scratch space must
7769 be allocated dynamically.
7771 @cindex @file{stdarg.h} and RT PC
7772 @item -mfp-arg-in-fpregs
7773 @opindex mfp-arg-in-fpregs
7774 Use a calling sequence incompatible with the IBM calling convention in
7775 which floating point arguments are passed in floating point registers.
7776 Note that @code{stdarg.h} will not work with floating point operands
7777 if this option is specified.
7779 @item -mfp-arg-in-gregs
7780 @opindex mfp-arg-in-gregs
7781 Use the normal calling convention for floating point arguments. This is
7784 @item -mhc-struct-return
7785 @opindex mhc-struct-return
7786 Return structures of more than one word in memory, rather than in a
7787 register. This provides compatibility with the MetaWare HighC (hc)
7788 compiler. Use the option @option{-fpcc-struct-return} for compatibility
7789 with the Portable C Compiler (pcc).
7791 @item -mnohc-struct-return
7792 @opindex mnohc-struct-return
7793 Return some structures of more than one word in registers, when
7794 convenient. This is the default. For compatibility with the
7795 IBM-supplied compilers, use the option @option{-fpcc-struct-return} or the
7796 option @option{-mhc-struct-return}.
7800 @subsection MIPS Options
7801 @cindex MIPS options
7803 These @samp{-m} options are defined for the MIPS family of computers:
7807 @item -march=@var{arch}
7809 Generate code that will run on @var{arch}, which can be the name of a
7810 generic MIPS ISA, or the name of a particular processor.
7812 @samp{mips1}, @samp{mips2}, @samp{mips3}, @samp{mips4},
7813 @samp{mips32}, @samp{mips32r2}, and @samp{mips64}.
7814 The processor names are:
7815 @samp{4kc}, @samp{4kp}, @samp{5kc}, @samp{20kc},
7817 @samp{r2000}, @samp{r3000}, @samp{r3900}, @samp{r4000}, @samp{r4400},
7818 @samp{r4600}, @samp{r4650}, @samp{r6000}, @samp{r8000}, @samp{rm7000},
7822 @samp{vr4100}, @samp{vr4111}, @samp{vr4120}, @samp{vr4300},
7823 @samp{vr5000}, @samp{vr5400} and @samp{vr5500}.
7824 The special value @samp{from-abi} selects the
7825 most compatible architecture for the selected ABI (that is,
7826 @samp{mips1} for 32-bit ABIs and @samp{mips3} for 64-bit ABIs)@.
7828 In processor names, a final @samp{000} can be abbreviated as @samp{k}
7829 (for example, @samp{-march=r2k}). Prefixes are optional, and
7830 @samp{vr} may be written @samp{r}.
7832 GCC defines two macros based on the value of this option. The first
7833 is @samp{_MIPS_ARCH}, which gives the name of target architecture, as
7834 a string. The second has the form @samp{_MIPS_ARCH_@var{foo}},
7835 where @var{foo} is the capitalized value of @samp{_MIPS_ARCH}@.
7836 For example, @samp{-march=r2000} will set @samp{_MIPS_ARCH}
7837 to @samp{"r2000"} and define the macro @samp{_MIPS_ARCH_R2000}.
7839 Note that the @samp{_MIPS_ARCH} macro uses the processor names given
7840 above. In other words, it will have the full prefix and will not
7841 abbreviate @samp{000} as @samp{k}. In the case of @samp{from-abi},
7842 the macro names the resolved architecture (either @samp{"mips1"} or
7843 @samp{"mips3"}). It names the default architecture when no
7844 @option{-march} option is given.
7846 @item -mtune=@var{arch}
7848 Optimize for @var{arch}. Among other things, this option controls
7849 the way instructions are scheduled, and the perceived cost of arithmetic
7850 operations. The list of @var{arch} values is the same as for
7853 When this option is not used, GCC will optimize for the processor
7854 specified by @option{-march}. By using @option{-march} and
7855 @option{-mtune} together, it is possible to generate code that will
7856 run on a family of processors, but optimize the code for one
7857 particular member of that family.
7859 @samp{-mtune} defines the macros @samp{_MIPS_TUNE} and
7860 @samp{_MIPS_TUNE_@var{foo}}, which work in the same way as the
7861 @samp{-march} ones described above.
7865 Equivalent to @samp{-march=mips1}.
7869 Equivalent to @samp{-march=mips2}.
7873 Equivalent to @samp{-march=mips3}.
7877 Equivalent to @samp{-march=mips4}.
7881 Equivalent to @samp{-march=mips32}.
7885 Equivalent to @samp{-march=mips32r2}.
7889 Equivalent to @samp{-march=mips64}.
7892 @itemx -mno-fused-madd
7893 @opindex mfused-madd
7894 @opindex mno-fused-madd
7895 Generate code that uses (does not use) the floating point multiply and
7896 accumulate instructions, when they are available. These instructions
7897 are generated by default if they are available, but this may be
7898 undesirable if the extra precision causes problems or on certain chips
7899 in the mode where denormals are rounded to zero where denormals
7900 generated by multiply and accumulate instructions cause exceptions
7905 Assume that floating point registers are 32 bits wide.
7909 Assume that floating point registers are 64 bits wide.
7913 Assume that general purpose registers are 32 bits wide.
7917 Assume that general purpose registers are 64 bits wide.
7921 Force int and long types to be 64 bits wide. See @option{-mlong32} for an
7922 explanation of the default, and the width of pointers.
7926 Force long types to be 64 bits wide. See @option{-mlong32} for an
7927 explanation of the default, and the width of pointers.
7931 Force long, int, and pointer types to be 32 bits wide.
7933 The default size of ints, longs and pointers depends on the ABI@. All
7934 the supported ABIs use 32-bit ints. The n64 ABI uses 64-bit longs, as
7935 does the 64-bit Cygnus EABI; the others use 32-bit longs. Pointers
7936 are the same size as longs, or the same size as integer registers,
7937 whichever is smaller.
7951 Generate code for the given ABI@.
7953 Note that there are two embedded ABIs: @option{-mabi=eabi}
7954 selects the one defined by Cygnus while @option{-meabi=meabi}
7955 selects the one defined by MIPS@. Both these ABIs have
7956 32-bit and 64-bit variants. Normally, GCC will generate
7957 64-bit code when you select a 64-bit architecture, but you
7958 can use @option{-mgp32} to get 32-bit code instead.
7960 @item -mabi-fake-default
7961 @opindex mabi-fake-default
7962 You don't want to know what this option does. No, really. I mean
7963 it. Move on to the next option.
7965 What? You're still here? Oh, well@enddots{} Ok, here's the deal. GCC
7966 wants the default set of options to get the root of the multilib tree,
7967 and the shared library SONAMEs without any multilib-indicating
7968 suffixes. This is not convenience for @samp{mips64-linux-gnu}, since
7969 we want to default to the N32 ABI, while still being binary-compatible
7970 with @samp{mips-linux-gnu} if you stick to the O32 ABI@. Being
7971 binary-compatible means shared libraries should have the same SONAMEs,
7972 and libraries should live in the same location. Having O32 libraries
7973 in a sub-directory named say @file{o32} is not acceptable.
7975 So we trick GCC into believing that O32 is the default ABI, except
7976 that we override the default with some internal command-line
7977 processing magic. Problem is, if we stopped at that, and you then
7978 created a multilib-aware package that used the output of @command{gcc
7979 -print-multi-lib} to decide which multilibs to build, and how, and
7980 you'd find yourself in an awkward situation when you found out that
7981 some of the options listed ended up mapping to the same multilib, and
7982 none of your libraries was actually built for the multilib that
7983 @option{-print-multi-lib} claims to be the default. So we added this
7984 option that disables the default switcher, falling back to GCC's
7985 original notion of the default library. Confused yet?
7987 For short: don't ever use this option, unless you find it in the list
7988 of additional options to be used when building for multilibs, in the
7989 output of @option{gcc -print-multi-lib}.
7993 Generate code for the MIPS assembler, and invoke @file{mips-tfile} to
7994 add normal debug information. This is the default for all
7995 platforms except for the OSF/1 reference platform, using the OSF/rose
7996 object format. If the either of the @option{-gstabs} or @option{-gstabs+}
7997 switches are used, the @file{mips-tfile} program will encapsulate the
7998 stabs within MIPS ECOFF@.
8002 Generate code for the GNU assembler. This is the default on the OSF/1
8003 reference platform, using the OSF/rose object format. Also, this is
8004 the default if the configure option @option{--with-gnu-as} is used.
8006 @item -msplit-addresses
8007 @itemx -mno-split-addresses
8008 @opindex msplit-addresses
8009 @opindex mno-split-addresses
8010 Generate code to load the high and low parts of address constants separately.
8011 This allows GCC to optimize away redundant loads of the high order
8012 bits of addresses. This optimization requires GNU as and GNU ld.
8013 This optimization is enabled by default for some embedded targets where
8014 GNU as and GNU ld are standard.
8020 The @option{-mrnames} switch says to output code using the MIPS software
8021 names for the registers, instead of the hardware names (ie, @var{a0}
8022 instead of @var{$4}). The only known assembler that supports this option
8023 is the Algorithmics assembler.
8029 The @option{-mmemcpy} switch makes all block moves call the appropriate
8030 string function (@samp{memcpy} or @samp{bcopy}) instead of possibly
8031 generating inline code.
8034 @itemx -mno-mips-tfile
8035 @opindex mmips-tfile
8036 @opindex mno-mips-tfile
8037 The @option{-mno-mips-tfile} switch causes the compiler not
8038 postprocess the object file with the @file{mips-tfile} program,
8039 after the MIPS assembler has generated it to add debug support. If
8040 @file{mips-tfile} is not run, then no local variables will be
8041 available to the debugger. In addition, @file{stage2} and
8042 @file{stage3} objects will have the temporary file names passed to the
8043 assembler embedded in the object file, which means the objects will
8044 not compare the same. The @option{-mno-mips-tfile} switch should only
8045 be used when there are bugs in the @file{mips-tfile} program that
8046 prevents compilation.
8049 @opindex msoft-float
8050 Generate output containing library calls for floating point.
8051 @strong{Warning:} the requisite libraries are not part of GCC@.
8052 Normally the facilities of the machine's usual C compiler are used, but
8053 this can't be done directly in cross-compilation. You must make your
8054 own arrangements to provide suitable library functions for
8058 @opindex mhard-float
8059 Generate output containing floating point instructions. This is the
8060 default if you use the unmodified sources.
8063 @itemx -mno-abicalls
8065 @opindex mno-abicalls
8066 Emit (or do not emit) the pseudo operations @samp{.abicalls},
8067 @samp{.cpload}, and @samp{.cprestore} that some System V.4 ports use for
8068 position independent code.
8074 Lift (or do not lift) the usual restrictions on the size of the global
8077 GCC normally uses a single instruction to load values from the GOT.
8078 While this is relatively efficient, it will only work if the GOT
8079 is smaller than about 64k. Anything larger will cause the linker
8080 to report an error such as:
8082 @cindex relocation truncated to fit (MIPS)
8084 relocation truncated to fit: R_MIPS_GOT16 foobar
8087 If this happens, you should recompile your code with @option{-mxgot}.
8088 It should then work with very large GOTs, although it will also be
8089 less efficient, since it will take three instructions to fetch the
8090 value of a global symbol.
8092 Note that some linkers can create multiple GOTs. If you have such a
8093 linker, you should only need to use @option{-mxgot} when a single object
8094 file accesses more than 64k's worth of GOT entries. Very few do.
8096 These options have no effect unless GCC is generating position
8100 @itemx -mno-long-calls
8101 @opindex mlong-calls
8102 @opindex mno-long-calls
8103 Do all calls with the @samp{JALR} instruction, which requires
8104 loading up a function's address into a register before the call.
8105 You need to use this switch, if you call outside of the current
8106 512 megabyte segment to functions that are not through pointers.
8108 @item -membedded-pic
8109 @itemx -mno-embedded-pic
8110 @opindex membedded-pic
8111 @opindex mno-embedded-pic
8112 Generate PIC code suitable for some embedded systems. All calls are
8113 made using PC relative address, and all data is addressed using the $gp
8114 register. No more than 65536 bytes of global data may be used. This
8115 requires GNU as and GNU ld which do most of the work. This currently
8116 only works on targets which use ECOFF; it does not work with ELF@.
8118 @item -membedded-data
8119 @itemx -mno-embedded-data
8120 @opindex membedded-data
8121 @opindex mno-embedded-data
8122 Allocate variables to the read-only data section first if possible, then
8123 next in the small data section if possible, otherwise in data. This gives
8124 slightly slower code than the default, but reduces the amount of RAM required
8125 when executing, and thus may be preferred for some embedded systems.
8127 @item -muninit-const-in-rodata
8128 @itemx -mno-uninit-const-in-rodata
8129 @opindex muninit-const-in-rodata
8130 @opindex mno-uninit-const-in-rodata
8131 When used together with @option{-membedded-data}, it will always store uninitialized
8132 const variables in the read-only data section.
8134 @item -msingle-float
8135 @itemx -mdouble-float
8136 @opindex msingle-float
8137 @opindex mdouble-float
8138 The @option{-msingle-float} switch tells gcc to assume that the floating
8139 point coprocessor only supports single precision operations, as on the
8140 @samp{r4650} chip. The @option{-mdouble-float} switch permits gcc to use
8141 double precision operations. This is the default.
8147 Permit use of the @samp{mad}, @samp{madu} and @samp{mul} instructions,
8148 as on the @samp{r4650} chip.
8152 Turns on @option{-msingle-float}, @option{-mmad}, and, at least for now,
8153 @option{-mcpu=r4650}.
8159 Enable 16-bit instructions.
8163 Use the entry and exit pseudo ops. This option can only be used with
8168 Compile code for the processor in little endian mode.
8169 The requisite libraries are assumed to exist.
8173 Compile code for the processor in big endian mode.
8174 The requisite libraries are assumed to exist.
8178 @cindex smaller data references (MIPS)
8179 @cindex gp-relative references (MIPS)
8180 Put global and static items less than or equal to @var{num} bytes into
8181 the small data or bss sections instead of the normal data or bss
8182 section. This allows the assembler to emit one word memory reference
8183 instructions based on the global pointer (@var{gp} or @var{$28}),
8184 instead of the normal two words used. By default, @var{num} is 8 when
8185 the MIPS assembler is used, and 0 when the GNU assembler is used. The
8186 @option{-G @var{num}} switch is also passed to the assembler and linker.
8187 All modules should be compiled with the same @option{-G @var{num}}
8192 Tell the MIPS assembler to not run its preprocessor over user
8193 assembler files (with a @samp{.s} suffix) when assembling them.
8197 Pass an option to gas which will cause nops to be inserted if
8198 the read of the destination register of an mfhi or mflo instruction
8199 occurs in the following two instructions.
8204 Work around certain SB-1 CPU core errata.
8205 (This flag currently works around the SB-1 revision 2
8206 ``F1'' and ``F2'' floating point errata.)
8210 Do not include the default crt0.
8212 @item -mflush-func=@var{func}
8213 @itemx -mno-flush-func
8214 @opindex mflush-func
8215 Specifies the function to call to flush the I and D caches, or to not
8216 call any such function. If called, the function must take the same
8217 arguments as the common @code{_flush_func()}, that is, the address of the
8218 memory range for which the cache is being flushed, the size of the
8219 memory range, and the number 3 (to flush both caches). The default
8220 depends on the target gcc was configured for, but commonly is either
8221 @samp{_flush_func} or @samp{__cpu_flush}.
8223 @item -mbranch-likely
8224 @itemx -mno-branch-likely
8225 @opindex mbranch-likely
8226 @opindex mno-branch-likely
8227 Enable or disable use of Branch Likely instructions, regardless of the
8228 default for the selected architecture. By default, Branch Likely
8229 instructions may be generated if they are supported by the selected
8230 architecture. An exception is for the MIPS32 and MIPS64 architectures
8231 and processors which implement those architectures; for those, Branch
8232 Likely instructions will not be generated by default because the MIPS32
8233 and MIPS64 architectures specifically deprecate their use.
8236 @node i386 and x86-64 Options
8237 @subsection Intel 386 and AMD x86-64 Options
8238 @cindex i386 Options
8239 @cindex x86-64 Options
8240 @cindex Intel 386 Options
8241 @cindex AMD x86-64 Options
8243 These @samp{-m} options are defined for the i386 and x86-64 family of
8247 @item -mtune=@var{cpu-type}
8249 Tune to @var{cpu-type} everything applicable about the generated code, except
8250 for the ABI and the set of available instructions. The choices for
8251 @var{cpu-type} are @samp{i386}, @samp{i486}, @samp{i586}, @samp{i686},
8252 @samp{pentium}, @samp{pentium-mmx}, @samp{pentiumpro}, @samp{pentium2},
8253 @samp{pentium3}, @samp{pentium4}, @samp{k6}, @samp{k6-2}, @samp{k6-3},
8254 @samp{athlon}, @samp{athlon-tbird}, @samp{athlon-4}, @samp{athlon-xp},
8255 @samp{athlon-mp}, @samp{winchip-c6}, @samp{winchip2}, @samp{k8}, @samp{c3}
8258 While picking a specific @var{cpu-type} will schedule things appropriately
8259 for that particular chip, the compiler will not generate any code that
8260 does not run on the i386 without the @option{-march=@var{cpu-type}} option
8261 being used. @samp{i586} is equivalent to @samp{pentium} and @samp{i686}
8262 is equivalent to @samp{pentiumpro}. @samp{k6} and @samp{athlon} are the
8263 AMD chips as opposed to the Intel ones.
8265 @item -march=@var{cpu-type}
8267 Generate instructions for the machine type @var{cpu-type}. The choices
8268 for @var{cpu-type} are the same as for @option{-mtune}. Moreover,
8269 specifying @option{-march=@var{cpu-type}} implies @option{-mtune=@var{cpu-type}}.
8271 @item -mcpu=@var{cpu-type}
8273 A deprecated synonym for @option{-mtune}.
8282 @opindex mpentiumpro
8283 These options are synonyms for @option{-mtune=i386}, @option{-mtune=i486},
8284 @option{-mtune=pentium}, and @option{-mtune=pentiumpro} respectively.
8285 These synonyms are deprecated.
8287 @item -mfpmath=@var{unit}
8289 generate floating point arithmetics for selected unit @var{unit}. the choices
8294 Use the standard 387 floating point coprocessor present majority of chips and
8295 emulated otherwise. Code compiled with this option will run almost everywhere.
8296 The temporary results are computed in 80bit precision instead of precision
8297 specified by the type resulting in slightly different results compared to most
8298 of other chips. See @option{-ffloat-store} for more detailed description.
8300 This is the default choice for i386 compiler.
8303 Use scalar floating point instructions present in the SSE instruction set.
8304 This instruction set is supported by Pentium3 and newer chips, in the AMD line
8305 by Athlon-4, Athlon-xp and Athlon-mp chips. The earlier version of SSE
8306 instruction set supports only single precision arithmetics, thus the double and
8307 extended precision arithmetics is still done using 387. Later version, present
8308 only in Pentium4 and the future AMD x86-64 chips supports double precision
8311 For i387 you need to use @option{-march=@var{cpu-type}}, @option{-msse} or
8312 @option{-msse2} switches to enable SSE extensions and make this option
8313 effective. For x86-64 compiler, these extensions are enabled by default.
8315 The resulting code should be considerably faster in majority of cases and avoid
8316 the numerical instability problems of 387 code, but may break some existing
8317 code that expects temporaries to be 80bit.
8319 This is the default choice for x86-64 compiler.
8322 Use all SSE extensions enabled by @option{-msse2} as well as the new
8323 SSE extensions in Prescott New Instructions. @option{-mpni} also
8324 enables 2 builtin functions, @code{__builtin_ia32_monitor} and
8325 @code{__builtin_ia32_mwait}, for new instructions @code{monitor} and
8329 Attempt to utilize both instruction sets at once. This effectively double the
8330 amount of available registers and on chips with separate execution units for
8331 387 and SSE the execution resources too. Use this option with care, as it is
8332 still experimental, because gcc register allocator does not model separate
8333 functional units well resulting in instable performance.
8336 @item -masm=@var{dialect}
8337 @opindex masm=@var{dialect}
8338 Output asm instructions using selected @var{dialect}. Supported choices are
8339 @samp{intel} or @samp{att} (the default one).
8344 @opindex mno-ieee-fp
8345 Control whether or not the compiler uses IEEE floating point
8346 comparisons. These handle correctly the case where the result of a
8347 comparison is unordered.
8350 @opindex msoft-float
8351 Generate output containing library calls for floating point.
8352 @strong{Warning:} the requisite libraries are not part of GCC@.
8353 Normally the facilities of the machine's usual C compiler are used, but
8354 this can't be done directly in cross-compilation. You must make your
8355 own arrangements to provide suitable library functions for
8358 On machines where a function returns floating point results in the 80387
8359 register stack, some floating point opcodes may be emitted even if
8360 @option{-msoft-float} is used.
8362 @item -mno-fp-ret-in-387
8363 @opindex mno-fp-ret-in-387
8364 Do not use the FPU registers for return values of functions.
8366 The usual calling convention has functions return values of types
8367 @code{float} and @code{double} in an FPU register, even if there
8368 is no FPU@. The idea is that the operating system should emulate
8371 The option @option{-mno-fp-ret-in-387} causes such values to be returned
8372 in ordinary CPU registers instead.
8374 @item -mno-fancy-math-387
8375 @opindex mno-fancy-math-387
8376 Some 387 emulators do not support the @code{sin}, @code{cos} and
8377 @code{sqrt} instructions for the 387. Specify this option to avoid
8378 generating those instructions. This option is the default on FreeBSD,
8379 OpenBSD and NetBSD@. This option is overridden when @option{-march}
8380 indicates that the target cpu will always have an FPU and so the
8381 instruction will not need emulation. As of revision 2.6.1, these
8382 instructions are not generated unless you also use the
8383 @option{-funsafe-math-optimizations} switch.
8385 @item -malign-double
8386 @itemx -mno-align-double
8387 @opindex malign-double
8388 @opindex mno-align-double
8389 Control whether GCC aligns @code{double}, @code{long double}, and
8390 @code{long long} variables on a two word boundary or a one word
8391 boundary. Aligning @code{double} variables on a two word boundary will
8392 produce code that runs somewhat faster on a @samp{Pentium} at the
8393 expense of more memory.
8395 @strong{Warning:} if you use the @option{-malign-double} switch,
8396 structures containing the above types will be aligned differently than
8397 the published application binary interface specifications for the 386
8398 and will not be binary compatible with structures in code compiled
8399 without that switch.
8401 @item -m96bit-long-double
8402 @item -m128bit-long-double
8403 @opindex m96bit-long-double
8404 @opindex m128bit-long-double
8405 These switches control the size of @code{long double} type. The i386
8406 application binary interface specifies the size to be 96 bits,
8407 so @option{-m96bit-long-double} is the default in 32 bit mode.
8409 Modern architectures (Pentium and newer) would prefer @code{long double}
8410 to be aligned to an 8 or 16 byte boundary. In arrays or structures
8411 conforming to the ABI, this would not be possible. So specifying a
8412 @option{-m128bit-long-double} will align @code{long double}
8413 to a 16 byte boundary by padding the @code{long double} with an additional
8416 In the x86-64 compiler, @option{-m128bit-long-double} is the default choice as
8417 its ABI specifies that @code{long double} is to be aligned on 16 byte boundary.
8419 Notice that neither of these options enable any extra precision over the x87
8420 standard of 80 bits for a @code{long double}.
8422 @strong{Warning:} if you override the default value for your target ABI, the
8423 structures and arrays containing @code{long double} will change their size as
8424 well as function calling convention for function taking @code{long double}
8425 will be modified. Hence they will not be binary compatible with arrays or
8426 structures in code compiled without that switch.
8430 @itemx -mno-svr3-shlib
8431 @opindex msvr3-shlib
8432 @opindex mno-svr3-shlib
8433 Control whether GCC places uninitialized local variables into the
8434 @code{bss} or @code{data} segments. @option{-msvr3-shlib} places them
8435 into @code{bss}. These options are meaningful only on System V Release 3.
8439 Use a different function-calling convention, in which functions that
8440 take a fixed number of arguments return with the @code{ret} @var{num}
8441 instruction, which pops their arguments while returning. This saves one
8442 instruction in the caller since there is no need to pop the arguments
8445 You can specify that an individual function is called with this calling
8446 sequence with the function attribute @samp{stdcall}. You can also
8447 override the @option{-mrtd} option by using the function attribute
8448 @samp{cdecl}. @xref{Function Attributes}.
8450 @strong{Warning:} this calling convention is incompatible with the one
8451 normally used on Unix, so you cannot use it if you need to call
8452 libraries compiled with the Unix compiler.
8454 Also, you must provide function prototypes for all functions that
8455 take variable numbers of arguments (including @code{printf});
8456 otherwise incorrect code will be generated for calls to those
8459 In addition, seriously incorrect code will result if you call a
8460 function with too many arguments. (Normally, extra arguments are
8461 harmlessly ignored.)
8463 @item -mregparm=@var{num}
8465 Control how many registers are used to pass integer arguments. By
8466 default, no registers are used to pass arguments, and at most 3
8467 registers can be used. You can control this behavior for a specific
8468 function by using the function attribute @samp{regparm}.
8469 @xref{Function Attributes}.
8471 @strong{Warning:} if you use this switch, and
8472 @var{num} is nonzero, then you must build all modules with the same
8473 value, including any libraries. This includes the system libraries and
8476 @item -mpreferred-stack-boundary=@var{num}
8477 @opindex mpreferred-stack-boundary
8478 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
8479 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
8480 the default is 4 (16 bytes or 128 bits), except when optimizing for code
8481 size (@option{-Os}), in which case the default is the minimum correct
8482 alignment (4 bytes for x86, and 8 bytes for x86-64).
8484 On Pentium and PentiumPro, @code{double} and @code{long double} values
8485 should be aligned to an 8 byte boundary (see @option{-malign-double}) or
8486 suffer significant run time performance penalties. On Pentium III, the
8487 Streaming SIMD Extension (SSE) data type @code{__m128} suffers similar
8488 penalties if it is not 16 byte aligned.
8490 To ensure proper alignment of this values on the stack, the stack boundary
8491 must be as aligned as that required by any value stored on the stack.
8492 Further, every function must be generated such that it keeps the stack
8493 aligned. Thus calling a function compiled with a higher preferred
8494 stack boundary from a function compiled with a lower preferred stack
8495 boundary will most likely misalign the stack. It is recommended that
8496 libraries that use callbacks always use the default setting.
8498 This extra alignment does consume extra stack space, and generally
8499 increases code size. Code that is sensitive to stack space usage, such
8500 as embedded systems and operating system kernels, may want to reduce the
8501 preferred alignment to @option{-mpreferred-stack-boundary=2}.
8519 These switches enable or disable the use of built-in functions that allow
8520 direct access to the MMX, SSE and 3Dnow extensions of the instruction set.
8522 @xref{X86 Built-in Functions}, for details of the functions enabled
8523 and disabled by these switches.
8525 To have SSE/SSE2 instructions generated automatically from floating-point
8526 code, see @option{-mfpmath=sse}.
8529 @itemx -mno-push-args
8531 @opindex mno-push-args
8532 Use PUSH operations to store outgoing parameters. This method is shorter
8533 and usually equally fast as method using SUB/MOV operations and is enabled
8534 by default. In some cases disabling it may improve performance because of
8535 improved scheduling and reduced dependencies.
8537 @item -maccumulate-outgoing-args
8538 @opindex maccumulate-outgoing-args
8539 If enabled, the maximum amount of space required for outgoing arguments will be
8540 computed in the function prologue. This is faster on most modern CPUs
8541 because of reduced dependencies, improved scheduling and reduced stack usage
8542 when preferred stack boundary is not equal to 2. The drawback is a notable
8543 increase in code size. This switch implies @option{-mno-push-args}.
8547 Support thread-safe exception handling on @samp{Mingw32}. Code that relies
8548 on thread-safe exception handling must compile and link all code with the
8549 @option{-mthreads} option. When compiling, @option{-mthreads} defines
8550 @option{-D_MT}; when linking, it links in a special thread helper library
8551 @option{-lmingwthrd} which cleans up per thread exception handling data.
8553 @item -mno-align-stringops
8554 @opindex mno-align-stringops
8555 Do not align destination of inlined string operations. This switch reduces
8556 code size and improves performance in case the destination is already aligned,
8557 but gcc don't know about it.
8559 @item -minline-all-stringops
8560 @opindex minline-all-stringops
8561 By default GCC inlines string operations only when destination is known to be
8562 aligned at least to 4 byte boundary. This enables more inlining, increase code
8563 size, but may improve performance of code that depends on fast memcpy, strlen
8564 and memset for short lengths.
8566 @item -momit-leaf-frame-pointer
8567 @opindex momit-leaf-frame-pointer
8568 Don't keep the frame pointer in a register for leaf functions. This
8569 avoids the instructions to save, set up and restore frame pointers and
8570 makes an extra register available in leaf functions. The option
8571 @option{-fomit-frame-pointer} removes the frame pointer for all functions
8572 which might make debugging harder.
8574 @item -mtls-direct-seg-refs
8575 @itemx -mno-tls-direct-seg-refs
8576 @opindex mtls-direct-seg-refs
8577 Controls whether TLS variables may be accessed with offsets from the
8578 TLS segment register (@code{%gs} for 32-bit, @code{%fs} for 64-bit),
8579 or whether the thread base pointer must be added. Whether or not this
8580 is legal depends on the operating system, and whether it maps the
8581 segment to cover the entire TLS area.
8583 For systems that use GNU libc, the default is on.
8586 These @samp{-m} switches are supported in addition to the above
8587 on AMD x86-64 processors in 64-bit environments.
8594 Generate code for a 32-bit or 64-bit environment.
8595 The 32-bit environment sets int, long and pointer to 32 bits and
8596 generates code that runs on any i386 system.
8597 The 64-bit environment sets int to 32 bits and long and pointer
8598 to 64 bits and generates code for AMD's x86-64 architecture.
8601 @opindex no-red-zone
8602 Do not use a so called red zone for x86-64 code. The red zone is mandated
8603 by the x86-64 ABI, it is a 128-byte area beyond the location of the
8604 stack pointer that will not be modified by signal or interrupt handlers
8605 and therefore can be used for temporary data without adjusting the stack
8606 pointer. The flag @option{-mno-red-zone} disables this red zone.
8608 @item -mcmodel=small
8609 @opindex mcmodel=small
8610 Generate code for the small code model: the program and its symbols must
8611 be linked in the lower 2 GB of the address space. Pointers are 64 bits.
8612 Programs can be statically or dynamically linked. This is the default
8615 @item -mcmodel=kernel
8616 @opindex mcmodel=kernel
8617 Generate code for the kernel code model. The kernel runs in the
8618 negative 2 GB of the address space.
8619 This model has to be used for Linux kernel code.
8621 @item -mcmodel=medium
8622 @opindex mcmodel=medium
8623 Generate code for the medium model: The program is linked in the lower 2
8624 GB of the address space but symbols can be located anywhere in the
8625 address space. Programs can be statically or dynamically linked, but
8626 building of shared libraries are not supported with the medium model.
8628 @item -mcmodel=large
8629 @opindex mcmodel=large
8630 Generate code for the large model: This model makes no assumptions
8631 about addresses and sizes of sections. Currently GCC does not implement
8636 @subsection HPPA Options
8637 @cindex HPPA Options
8639 These @samp{-m} options are defined for the HPPA family of computers:
8642 @item -march=@var{architecture-type}
8644 Generate code for the specified architecture. The choices for
8645 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
8646 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
8647 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
8648 architecture option for your machine. Code compiled for lower numbered
8649 architectures will run on higher numbered architectures, but not the
8652 PA 2.0 support currently requires gas snapshot 19990413 or later. The
8653 next release of binutils (current is 2.9.1) will probably contain PA 2.0
8657 @itemx -mpa-risc-1-1
8658 @itemx -mpa-risc-2-0
8659 @opindex mpa-risc-1-0
8660 @opindex mpa-risc-1-1
8661 @opindex mpa-risc-2-0
8662 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
8665 @opindex mbig-switch
8666 Generate code suitable for big switch tables. Use this option only if
8667 the assembler/linker complain about out of range branches within a switch
8670 @item -mjump-in-delay
8671 @opindex mjump-in-delay
8672 Fill delay slots of function calls with unconditional jump instructions
8673 by modifying the return pointer for the function call to be the target
8674 of the conditional jump.
8676 @item -mdisable-fpregs
8677 @opindex mdisable-fpregs
8678 Prevent floating point registers from being used in any manner. This is
8679 necessary for compiling kernels which perform lazy context switching of
8680 floating point registers. If you use this option and attempt to perform
8681 floating point operations, the compiler will abort.
8683 @item -mdisable-indexing
8684 @opindex mdisable-indexing
8685 Prevent the compiler from using indexing address modes. This avoids some
8686 rather obscure problems when compiling MIG generated code under MACH@.
8688 @item -mno-space-regs
8689 @opindex mno-space-regs
8690 Generate code that assumes the target has no space registers. This allows
8691 GCC to generate faster indirect calls and use unscaled index address modes.
8693 Such code is suitable for level 0 PA systems and kernels.
8695 @item -mfast-indirect-calls
8696 @opindex mfast-indirect-calls
8697 Generate code that assumes calls never cross space boundaries. This
8698 allows GCC to emit code which performs faster indirect calls.
8700 This option will not work in the presence of shared libraries or nested
8703 @item -mlong-load-store
8704 @opindex mlong-load-store
8705 Generate 3-instruction load and store sequences as sometimes required by
8706 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
8709 @item -mportable-runtime
8710 @opindex mportable-runtime
8711 Use the portable calling conventions proposed by HP for ELF systems.
8715 Enable the use of assembler directives only GAS understands.
8717 @item -mschedule=@var{cpu-type}
8719 Schedule code according to the constraints for the machine type
8720 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
8721 @samp{7100}, @samp{7100LC}, @samp{7200}, @samp{7300} and @samp{8000}. Refer
8722 to @file{/usr/lib/sched.models} on an HP-UX system to determine the
8723 proper scheduling option for your machine. The default scheduling is
8727 @opindex mlinker-opt
8728 Enable the optimization pass in the HP-UX linker. Note this makes symbolic
8729 debugging impossible. It also triggers a bug in the HP-UX 8 and HP-UX 9
8730 linkers in which they give bogus error messages when linking some programs.
8733 @opindex msoft-float
8734 Generate output containing library calls for floating point.
8735 @strong{Warning:} the requisite libraries are not available for all HPPA
8736 targets. Normally the facilities of the machine's usual C compiler are
8737 used, but this cannot be done directly in cross-compilation. You must make
8738 your own arrangements to provide suitable library functions for
8739 cross-compilation. The embedded target @samp{hppa1.1-*-pro}
8740 does provide software floating point support.
8742 @option{-msoft-float} changes the calling convention in the output file;
8743 therefore, it is only useful if you compile @emph{all} of a program with
8744 this option. In particular, you need to compile @file{libgcc.a}, the
8745 library that comes with GCC, with @option{-msoft-float} in order for
8750 Generate the predefine, @code{_SIO}, for server IO. The default is
8751 @option{-mwsio}. This generates the predefines, @code{__hp9000s700},
8752 @code{__hp9000s700__} and @code{_WSIO}, for workstation IO. These
8753 options are available under HP-UX and HI-UX.
8757 Use GNU ld specific options. This passes @option{-shared} to ld when
8758 building a shared library. It is the default when GCC is configured,
8759 explicitly or implicitly, with the GNU linker. This option does not
8760 have any affect on which ld is called, it only changes what parameters
8761 are passed to that ld. The ld that is called is determined by the
8762 @option{--with-ld} configure option, gcc's program search path, and
8763 finally by the user's @env{PATH}. The linker used by GCC can be printed
8764 using @samp{which `gcc -print-prog-name=ld`}.
8768 Use HP ld specific options. This passes @option{-b} to ld when building
8769 a shared library and passes @option{+Accept TypeMismatch} to ld on all
8770 links. It is the default when GCC is configured, explicitly or
8771 implicitly, with the HP linker. This option does not have any affect on
8772 which ld is called, it only changes what parameters are passed to that
8773 ld. The ld that is called is determined by the @option{--with-ld}
8774 configure option, gcc's program search path, and finally by the user's
8775 @env{PATH}. The linker used by GCC can be printed using @samp{which
8776 `gcc -print-prog-name=ld`}.
8779 @opindex mno-long-calls
8780 Generate code that uses long call sequences. This ensures that a call
8781 is always able to reach linker generated stubs. The default is to generate
8782 long calls only when the distance from the call site to the beginning
8783 of the function or translation unit, as the case may be, exceeds a
8784 predefined limit set by the branch type being used. The limits for
8785 normal calls are 7,600,000 and 240,000 bytes, respectively for the
8786 PA 2.0 and PA 1.X architectures. Sibcalls are always limited at
8789 Distances are measured from the beginning of functions when using the
8790 @option{-ffunction-sections} option, or when using the @option{-mgas}
8791 and @option{-mno-portable-runtime} options together under HP-UX with
8794 It is normally not desirable to use this option as it will degrade
8795 performance. However, it may be useful in large applications,
8796 particularly when partial linking is used to build the application.
8798 The types of long calls used depends on the capabilities of the
8799 assembler and linker, and the type of code being generated. The
8800 impact on systems that support long absolute calls, and long pic
8801 symbol-difference or pc-relative calls should be relatively small.
8802 However, an indirect call is used on 32-bit ELF systems in pic code
8803 and it is quite long.
8807 Suppress the generation of link options to search libdld.sl when the
8808 @option{-static} option is specified on HP-UX 10 and later.
8812 The HP-UX implementation of setlocale in libc has a dependency on
8813 libdld.sl. There isn't an archive version of libdld.sl. Thus,
8814 when the @option{-static} option is specified, special link options
8815 are needed to resolve this dependency.
8817 On HP-UX 10 and later, the GCC driver adds the necessary options to
8818 link with libdld.sl when the @option{-static} option is specified.
8819 This causes the resulting binary to be dynamic. On the 64-bit port,
8820 the linkers generate dynamic binaries by default in any case. The
8821 @option{-nolibdld} option can be used to prevent the GCC driver from
8822 adding these link options.
8826 Add support for multithreading with the @dfn{dce thread} library
8827 under HP-UX. This option sets flags for both the preprocessor and
8831 @node Intel 960 Options
8832 @subsection Intel 960 Options
8834 These @samp{-m} options are defined for the Intel 960 implementations:
8837 @item -m@var{cpu-type}
8845 Assume the defaults for the machine type @var{cpu-type} for some of
8846 the other options, including instruction scheduling, floating point
8847 support, and addressing modes. The choices for @var{cpu-type} are
8848 @samp{ka}, @samp{kb}, @samp{mc}, @samp{ca}, @samp{cf},
8849 @samp{sa}, and @samp{sb}.
8856 @opindex msoft-float
8857 The @option{-mnumerics} option indicates that the processor does support
8858 floating-point instructions. The @option{-msoft-float} option indicates
8859 that floating-point support should not be assumed.
8861 @item -mleaf-procedures
8862 @itemx -mno-leaf-procedures
8863 @opindex mleaf-procedures
8864 @opindex mno-leaf-procedures
8865 Do (or do not) attempt to alter leaf procedures to be callable with the
8866 @code{bal} instruction as well as @code{call}. This will result in more
8867 efficient code for explicit calls when the @code{bal} instruction can be
8868 substituted by the assembler or linker, but less efficient code in other
8869 cases, such as calls via function pointers, or using a linker that doesn't
8870 support this optimization.
8873 @itemx -mno-tail-call
8875 @opindex mno-tail-call
8876 Do (or do not) make additional attempts (beyond those of the
8877 machine-independent portions of the compiler) to optimize tail-recursive
8878 calls into branches. You may not want to do this because the detection of
8879 cases where this is not valid is not totally complete. The default is
8880 @option{-mno-tail-call}.
8882 @item -mcomplex-addr
8883 @itemx -mno-complex-addr
8884 @opindex mcomplex-addr
8885 @opindex mno-complex-addr
8886 Assume (or do not assume) that the use of a complex addressing mode is a
8887 win on this implementation of the i960. Complex addressing modes may not
8888 be worthwhile on the K-series, but they definitely are on the C-series.
8889 The default is currently @option{-mcomplex-addr} for all processors except
8893 @itemx -mno-code-align
8894 @opindex mcode-align
8895 @opindex mno-code-align
8896 Align code to 8-byte boundaries for faster fetching (or don't bother).
8897 Currently turned on by default for C-series implementations only.
8900 @item -mclean-linkage
8901 @itemx -mno-clean-linkage
8902 @opindex mclean-linkage
8903 @opindex mno-clean-linkage
8904 These options are not fully implemented.
8908 @itemx -mic2.0-compat
8909 @itemx -mic3.0-compat
8911 @opindex mic2.0-compat
8912 @opindex mic3.0-compat
8913 Enable compatibility with iC960 v2.0 or v3.0.
8917 @opindex masm-compat
8919 Enable compatibility with the iC960 assembler.
8921 @item -mstrict-align
8922 @itemx -mno-strict-align
8923 @opindex mstrict-align
8924 @opindex mno-strict-align
8925 Do not permit (do permit) unaligned accesses.
8929 Enable structure-alignment compatibility with Intel's gcc release version
8930 1.3 (based on gcc 1.37). This option implies @option{-mstrict-align}.
8932 @item -mlong-double-64
8933 @opindex mlong-double-64
8934 Implement type @samp{long double} as 64-bit floating point numbers.
8935 Without the option @samp{long double} is implemented by 80-bit
8936 floating point numbers. The only reason we have it because there is
8937 no 128-bit @samp{long double} support in @samp{fp-bit.c} yet. So it
8938 is only useful for people using soft-float targets. Otherwise, we
8939 should recommend against use of it.
8943 @node DEC Alpha Options
8944 @subsection DEC Alpha Options
8946 These @samp{-m} options are defined for the DEC Alpha implementations:
8949 @item -mno-soft-float
8951 @opindex mno-soft-float
8952 @opindex msoft-float
8953 Use (do not use) the hardware floating-point instructions for
8954 floating-point operations. When @option{-msoft-float} is specified,
8955 functions in @file{libgcc.a} will be used to perform floating-point
8956 operations. Unless they are replaced by routines that emulate the
8957 floating-point operations, or compiled in such a way as to call such
8958 emulations routines, these routines will issue floating-point
8959 operations. If you are compiling for an Alpha without floating-point
8960 operations, you must ensure that the library is built so as not to call
8963 Note that Alpha implementations without floating-point operations are
8964 required to have floating-point registers.
8969 @opindex mno-fp-regs
8970 Generate code that uses (does not use) the floating-point register set.
8971 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
8972 register set is not used, floating point operands are passed in integer
8973 registers as if they were integers and floating-point results are passed
8974 in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence,
8975 so any function with a floating-point argument or return value called by code
8976 compiled with @option{-mno-fp-regs} must also be compiled with that
8979 A typical use of this option is building a kernel that does not use,
8980 and hence need not save and restore, any floating-point registers.
8984 The Alpha architecture implements floating-point hardware optimized for
8985 maximum performance. It is mostly compliant with the IEEE floating
8986 point standard. However, for full compliance, software assistance is
8987 required. This option generates code fully IEEE compliant code
8988 @emph{except} that the @var{inexact-flag} is not maintained (see below).
8989 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
8990 defined during compilation. The resulting code is less efficient but is
8991 able to correctly support denormalized numbers and exceptional IEEE
8992 values such as not-a-number and plus/minus infinity. Other Alpha
8993 compilers call this option @option{-ieee_with_no_inexact}.
8995 @item -mieee-with-inexact
8996 @opindex mieee-with-inexact
8997 This is like @option{-mieee} except the generated code also maintains
8998 the IEEE @var{inexact-flag}. Turning on this option causes the
8999 generated code to implement fully-compliant IEEE math. In addition to
9000 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
9001 macro. On some Alpha implementations the resulting code may execute
9002 significantly slower than the code generated by default. Since there is
9003 very little code that depends on the @var{inexact-flag}, you should
9004 normally not specify this option. Other Alpha compilers call this
9005 option @option{-ieee_with_inexact}.
9007 @item -mfp-trap-mode=@var{trap-mode}
9008 @opindex mfp-trap-mode
9009 This option controls what floating-point related traps are enabled.
9010 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
9011 The trap mode can be set to one of four values:
9015 This is the default (normal) setting. The only traps that are enabled
9016 are the ones that cannot be disabled in software (e.g., division by zero
9020 In addition to the traps enabled by @samp{n}, underflow traps are enabled
9024 Like @samp{su}, but the instructions are marked to be safe for software
9025 completion (see Alpha architecture manual for details).
9028 Like @samp{su}, but inexact traps are enabled as well.
9031 @item -mfp-rounding-mode=@var{rounding-mode}
9032 @opindex mfp-rounding-mode
9033 Selects the IEEE rounding mode. Other Alpha compilers call this option
9034 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
9039 Normal IEEE rounding mode. Floating point numbers are rounded towards
9040 the nearest machine number or towards the even machine number in case
9044 Round towards minus infinity.
9047 Chopped rounding mode. Floating point numbers are rounded towards zero.
9050 Dynamic rounding mode. A field in the floating point control register
9051 (@var{fpcr}, see Alpha architecture reference manual) controls the
9052 rounding mode in effect. The C library initializes this register for
9053 rounding towards plus infinity. Thus, unless your program modifies the
9054 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
9057 @item -mtrap-precision=@var{trap-precision}
9058 @opindex mtrap-precision
9059 In the Alpha architecture, floating point traps are imprecise. This
9060 means without software assistance it is impossible to recover from a
9061 floating trap and program execution normally needs to be terminated.
9062 GCC can generate code that can assist operating system trap handlers
9063 in determining the exact location that caused a floating point trap.
9064 Depending on the requirements of an application, different levels of
9065 precisions can be selected:
9069 Program precision. This option is the default and means a trap handler
9070 can only identify which program caused a floating point exception.
9073 Function precision. The trap handler can determine the function that
9074 caused a floating point exception.
9077 Instruction precision. The trap handler can determine the exact
9078 instruction that caused a floating point exception.
9081 Other Alpha compilers provide the equivalent options called
9082 @option{-scope_safe} and @option{-resumption_safe}.
9084 @item -mieee-conformant
9085 @opindex mieee-conformant
9086 This option marks the generated code as IEEE conformant. You must not
9087 use this option unless you also specify @option{-mtrap-precision=i} and either
9088 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
9089 is to emit the line @samp{.eflag 48} in the function prologue of the
9090 generated assembly file. Under DEC Unix, this has the effect that
9091 IEEE-conformant math library routines will be linked in.
9093 @item -mbuild-constants
9094 @opindex mbuild-constants
9095 Normally GCC examines a 32- or 64-bit integer constant to
9096 see if it can construct it from smaller constants in two or three
9097 instructions. If it cannot, it will output the constant as a literal and
9098 generate code to load it from the data segment at runtime.
9100 Use this option to require GCC to construct @emph{all} integer constants
9101 using code, even if it takes more instructions (the maximum is six).
9103 You would typically use this option to build a shared library dynamic
9104 loader. Itself a shared library, it must relocate itself in memory
9105 before it can find the variables and constants in its own data segment.
9111 Select whether to generate code to be assembled by the vendor-supplied
9112 assembler (@option{-malpha-as}) or by the GNU assembler @option{-mgas}.
9130 Indicate whether GCC should generate code to use the optional BWX,
9131 CIX, FIX and MAX instruction sets. The default is to use the instruction
9132 sets supported by the CPU type specified via @option{-mcpu=} option or that
9133 of the CPU on which GCC was built if none was specified.
9138 @opindex mfloat-ieee
9139 Generate code that uses (does not use) VAX F and G floating point
9140 arithmetic instead of IEEE single and double precision.
9142 @item -mexplicit-relocs
9143 @itemx -mno-explicit-relocs
9144 @opindex mexplicit-relocs
9145 @opindex mno-explicit-relocs
9146 Older Alpha assemblers provided no way to generate symbol relocations
9147 except via assembler macros. Use of these macros does not allow
9148 optimal instruction scheduling. GNU binutils as of version 2.12
9149 supports a new syntax that allows the compiler to explicitly mark
9150 which relocations should apply to which instructions. This option
9151 is mostly useful for debugging, as GCC detects the capabilities of
9152 the assembler when it is built and sets the default accordingly.
9156 @opindex msmall-data
9157 @opindex mlarge-data
9158 When @option{-mexplicit-relocs} is in effect, static data is
9159 accessed via @dfn{gp-relative} relocations. When @option{-msmall-data}
9160 is used, objects 8 bytes long or smaller are placed in a @dfn{small data area}
9161 (the @code{.sdata} and @code{.sbss} sections) and are accessed via
9162 16-bit relocations off of the @code{$gp} register. This limits the
9163 size of the small data area to 64KB, but allows the variables to be
9164 directly accessed via a single instruction.
9166 The default is @option{-mlarge-data}. With this option the data area
9167 is limited to just below 2GB. Programs that require more than 2GB of
9168 data must use @code{malloc} or @code{mmap} to allocate the data in the
9169 heap instead of in the program's data segment.
9171 When generating code for shared libraries, @option{-fpic} implies
9172 @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}.
9176 @opindex msmall-text
9177 @opindex mlarge-text
9178 When @option{-msmall-text} is used, the compiler assumes that the
9179 code of the entire program (or shared library) fits in 4MB, and is
9180 thus reachable with a branch instruction. When @option{-msmall-data}
9181 is used, the compiler can assume that all local symbols share the
9182 same @code{$gp} value, and thus reduce the number of instructions
9183 required for a function call from 4 to 1.
9185 The default is @option{-mlarge-text}.
9187 @item -mcpu=@var{cpu_type}
9189 Set the instruction set and instruction scheduling parameters for
9190 machine type @var{cpu_type}. You can specify either the @samp{EV}
9191 style name or the corresponding chip number. GCC supports scheduling
9192 parameters for the EV4, EV5 and EV6 family of processors and will
9193 choose the default values for the instruction set from the processor
9194 you specify. If you do not specify a processor type, GCC will default
9195 to the processor on which the compiler was built.
9197 Supported values for @var{cpu_type} are
9203 Schedules as an EV4 and has no instruction set extensions.
9207 Schedules as an EV5 and has no instruction set extensions.
9211 Schedules as an EV5 and supports the BWX extension.
9216 Schedules as an EV5 and supports the BWX and MAX extensions.
9220 Schedules as an EV6 and supports the BWX, FIX, and MAX extensions.
9224 Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions.
9227 @item -mtune=@var{cpu_type}
9229 Set only the instruction scheduling parameters for machine type
9230 @var{cpu_type}. The instruction set is not changed.
9232 @item -mmemory-latency=@var{time}
9233 @opindex mmemory-latency
9234 Sets the latency the scheduler should assume for typical memory
9235 references as seen by the application. This number is highly
9236 dependent on the memory access patterns used by the application
9237 and the size of the external cache on the machine.
9239 Valid options for @var{time} are
9243 A decimal number representing clock cycles.
9249 The compiler contains estimates of the number of clock cycles for
9250 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
9251 (also called Dcache, Scache, and Bcache), as well as to main memory.
9252 Note that L3 is only valid for EV5.
9257 @node DEC Alpha/VMS Options
9258 @subsection DEC Alpha/VMS Options
9260 These @samp{-m} options are defined for the DEC Alpha/VMS implementations:
9263 @item -mvms-return-codes
9264 @opindex mvms-return-codes
9265 Return VMS condition codes from main. The default is to return POSIX
9266 style condition (e.g.@ error) codes.
9269 @node H8/300 Options
9270 @subsection H8/300 Options
9272 These @samp{-m} options are defined for the H8/300 implementations:
9277 Shorten some address references at link time, when possible; uses the
9278 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
9279 ld.info, Using ld}, for a fuller description.
9283 Generate code for the H8/300H@.
9287 Generate code for the H8S@.
9291 Generate code for the H8S and H8/300H in the normal mode. This switch
9292 must be used either with -mh or -ms.
9296 Generate code for the H8S/2600. This switch must be used with @option{-ms}.
9300 Make @code{int} data 32 bits by default.
9304 On the H8/300H and H8S, use the same alignment rules as for the H8/300.
9305 The default for the H8/300H and H8S is to align longs and floats on 4
9307 @option{-malign-300} causes them to be aligned on 2 byte boundaries.
9308 This option has no effect on the H8/300.
9312 @subsection SH Options
9314 These @samp{-m} options are defined for the SH implementations:
9319 Generate code for the SH1.
9323 Generate code for the SH2.
9326 Generate code for the SH2e.
9330 Generate code for the SH3.
9334 Generate code for the SH3e.
9338 Generate code for the SH4 without a floating-point unit.
9340 @item -m4-single-only
9341 @opindex m4-single-only
9342 Generate code for the SH4 with a floating-point unit that only
9343 supports single-precision arithmetic.
9347 Generate code for the SH4 assuming the floating-point unit is in
9348 single-precision mode by default.
9352 Generate code for the SH4.
9356 Compile code for the processor in big endian mode.
9360 Compile code for the processor in little endian mode.
9364 Align doubles at 64-bit boundaries. Note that this changes the calling
9365 conventions, and thus some functions from the standard C library will
9366 not work unless you recompile it first with @option{-mdalign}.
9370 Shorten some address references at link time, when possible; uses the
9371 linker option @option{-relax}.
9375 Use 32-bit offsets in @code{switch} tables. The default is to use
9380 Enable the use of the instruction @code{fmovd}.
9384 Comply with the calling conventions defined by Renesas.
9388 Mark the @code{MAC} register as call-clobbered, even if
9389 @option{-mhitachi} is given.
9393 Increase IEEE-compliance of floating-point code.
9397 Dump instruction size and location in the assembly code.
9401 This option is deprecated. It pads structures to multiple of 4 bytes,
9402 which is incompatible with the SH ABI@.
9406 Optimize for space instead of speed. Implied by @option{-Os}.
9410 When generating position-independent code, emit function calls using
9411 the Global Offset Table instead of the Procedure Linkage Table.
9415 Generate a library function call to invalidate instruction cache
9416 entries, after fixing up a trampoline. This library function call
9417 doesn't assume it can write to the whole memory address space. This
9418 is the default when the target is @code{sh-*-linux*}.
9421 @node System V Options
9422 @subsection Options for System V
9424 These additional options are available on System V Release 4 for
9425 compatibility with other compilers on those systems:
9430 Create a shared object.
9431 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
9435 Identify the versions of each tool used by the compiler, in a
9436 @code{.ident} assembler directive in the output.
9440 Refrain from adding @code{.ident} directives to the output file (this is
9443 @item -YP,@var{dirs}
9445 Search the directories @var{dirs}, and no others, for libraries
9446 specified with @option{-l}.
9450 Look in the directory @var{dir} to find the M4 preprocessor.
9451 The assembler uses this option.
9452 @c This is supposed to go with a -Yd for predefined M4 macro files, but
9453 @c the generic assembler that comes with Solaris takes just -Ym.
9456 @node TMS320C3x/C4x Options
9457 @subsection TMS320C3x/C4x Options
9458 @cindex TMS320C3x/C4x Options
9460 These @samp{-m} options are defined for TMS320C3x/C4x implementations:
9464 @item -mcpu=@var{cpu_type}
9466 Set the instruction set, register set, and instruction scheduling
9467 parameters for machine type @var{cpu_type}. Supported values for
9468 @var{cpu_type} are @samp{c30}, @samp{c31}, @samp{c32}, @samp{c40}, and
9469 @samp{c44}. The default is @samp{c40} to generate code for the
9474 @itemx -msmall-memory
9476 @opindex mbig-memory
9478 @opindex msmall-memory
9480 Generates code for the big or small memory model. The small memory
9481 model assumed that all data fits into one 64K word page. At run-time
9482 the data page (DP) register must be set to point to the 64K page
9483 containing the .bss and .data program sections. The big memory model is
9484 the default and requires reloading of the DP register for every direct
9491 Allow (disallow) allocation of general integer operands into the block
9498 Enable (disable) generation of code using decrement and branch,
9499 DBcond(D), instructions. This is enabled by default for the C4x. To be
9500 on the safe side, this is disabled for the C3x, since the maximum
9501 iteration count on the C3x is @math{2^{23} + 1} (but who iterates loops more than
9502 @math{2^{23}} times on the C3x?). Note that GCC will try to reverse a loop so
9503 that it can utilize the decrement and branch instruction, but will give
9504 up if there is more than one memory reference in the loop. Thus a loop
9505 where the loop counter is decremented can generate slightly more
9506 efficient code, in cases where the RPTB instruction cannot be utilized.
9508 @item -mdp-isr-reload
9510 @opindex mdp-isr-reload
9512 Force the DP register to be saved on entry to an interrupt service
9513 routine (ISR), reloaded to point to the data section, and restored on
9514 exit from the ISR@. This should not be required unless someone has
9515 violated the small memory model by modifying the DP register, say within
9522 For the C3x use the 24-bit MPYI instruction for integer multiplies
9523 instead of a library call to guarantee 32-bit results. Note that if one
9524 of the operands is a constant, then the multiplication will be performed
9525 using shifts and adds. If the @option{-mmpyi} option is not specified for the C3x,
9526 then squaring operations are performed inline instead of a library call.
9529 @itemx -mno-fast-fix
9531 @opindex mno-fast-fix
9532 The C3x/C4x FIX instruction to convert a floating point value to an
9533 integer value chooses the nearest integer less than or equal to the
9534 floating point value rather than to the nearest integer. Thus if the
9535 floating point number is negative, the result will be incorrectly
9536 truncated an additional code is necessary to detect and correct this
9537 case. This option can be used to disable generation of the additional
9538 code required to correct the result.
9544 Enable (disable) generation of repeat block sequences using the RPTB
9545 instruction for zero overhead looping. The RPTB construct is only used
9546 for innermost loops that do not call functions or jump across the loop
9547 boundaries. There is no advantage having nested RPTB loops due to the
9548 overhead required to save and restore the RC, RS, and RE registers.
9549 This is enabled by default with @option{-O2}.
9551 @item -mrpts=@var{count}
9555 Enable (disable) the use of the single instruction repeat instruction
9556 RPTS@. If a repeat block contains a single instruction, and the loop
9557 count can be guaranteed to be less than the value @var{count}, GCC will
9558 emit a RPTS instruction instead of a RPTB@. If no value is specified,
9559 then a RPTS will be emitted even if the loop count cannot be determined
9560 at compile time. Note that the repeated instruction following RPTS does
9561 not have to be reloaded from memory each iteration, thus freeing up the
9562 CPU buses for operands. However, since interrupts are blocked by this
9563 instruction, it is disabled by default.
9565 @item -mloop-unsigned
9566 @itemx -mno-loop-unsigned
9567 @opindex mloop-unsigned
9568 @opindex mno-loop-unsigned
9569 The maximum iteration count when using RPTS and RPTB (and DB on the C40)
9570 is @math{2^{31} + 1} since these instructions test if the iteration count is
9571 negative to terminate the loop. If the iteration count is unsigned
9572 there is a possibility than the @math{2^{31} + 1} maximum iteration count may be
9573 exceeded. This switch allows an unsigned iteration count.
9577 Try to emit an assembler syntax that the TI assembler (asm30) is happy
9578 with. This also enforces compatibility with the API employed by the TI
9579 C3x C compiler. For example, long doubles are passed as structures
9580 rather than in floating point registers.
9586 Generate code that uses registers (stack) for passing arguments to functions.
9587 By default, arguments are passed in registers where possible rather
9588 than by pushing arguments on to the stack.
9590 @item -mparallel-insns
9591 @itemx -mno-parallel-insns
9592 @opindex mparallel-insns
9593 @opindex mno-parallel-insns
9594 Allow the generation of parallel instructions. This is enabled by
9595 default with @option{-O2}.
9597 @item -mparallel-mpy
9598 @itemx -mno-parallel-mpy
9599 @opindex mparallel-mpy
9600 @opindex mno-parallel-mpy
9601 Allow the generation of MPY||ADD and MPY||SUB parallel instructions,
9602 provided @option{-mparallel-insns} is also specified. These instructions have
9603 tight register constraints which can pessimize the code generation
9609 @subsection V850 Options
9610 @cindex V850 Options
9612 These @samp{-m} options are defined for V850 implementations:
9616 @itemx -mno-long-calls
9617 @opindex mlong-calls
9618 @opindex mno-long-calls
9619 Treat all calls as being far away (near). If calls are assumed to be
9620 far away, the compiler will always load the functions address up into a
9621 register, and call indirect through the pointer.
9627 Do not optimize (do optimize) basic blocks that use the same index
9628 pointer 4 or more times to copy pointer into the @code{ep} register, and
9629 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
9630 option is on by default if you optimize.
9632 @item -mno-prolog-function
9633 @itemx -mprolog-function
9634 @opindex mno-prolog-function
9635 @opindex mprolog-function
9636 Do not use (do use) external functions to save and restore registers
9637 at the prologue and epilogue of a function. The external functions
9638 are slower, but use less code space if more than one function saves
9639 the same number of registers. The @option{-mprolog-function} option
9640 is on by default if you optimize.
9644 Try to make the code as small as possible. At present, this just turns
9645 on the @option{-mep} and @option{-mprolog-function} options.
9649 Put static or global variables whose size is @var{n} bytes or less into
9650 the tiny data area that register @code{ep} points to. The tiny data
9651 area can hold up to 256 bytes in total (128 bytes for byte references).
9655 Put static or global variables whose size is @var{n} bytes or less into
9656 the small data area that register @code{gp} points to. The small data
9657 area can hold up to 64 kilobytes.
9661 Put static or global variables whose size is @var{n} bytes or less into
9662 the first 32 kilobytes of memory.
9666 Specify that the target processor is the V850.
9669 @opindex mbig-switch
9670 Generate code suitable for big switch tables. Use this option only if
9671 the assembler/linker complain about out of range branches within a switch
9676 This option will cause r2 and r5 to be used in the code generated by
9677 the compiler. This setting is the default.
9680 @opindex mno-app-regs
9681 This option will cause r2 and r5 to be treated as fixed registers.
9685 Specify that the target processor is the V850E. The preprocessor
9686 constant @samp{__v850e__} will be defined if this option is used.
9688 If neither @option{-mv850} nor @option{-mv850e} are defined
9689 then a default target processor will be chosen and the relevant
9690 @samp{__v850*__} preprocessor constant will be defined.
9692 The preprocessor constants @samp{__v850} and @samp{__v851__} are always
9693 defined, regardless of which processor variant is the target.
9695 @item -mdisable-callt
9696 @opindex mdisable-callt
9697 This option will suppress generation of the CALLT instruction for the
9698 v850e flavors of the v850 architecture. The default is
9699 @option{-mno-disable-callt} which allows the CALLT instruction to be used.
9704 @subsection ARC Options
9707 These options are defined for ARC implementations:
9712 Compile code for little endian mode. This is the default.
9716 Compile code for big endian mode.
9719 @opindex mmangle-cpu
9720 Prepend the name of the cpu to all public symbol names.
9721 In multiple-processor systems, there are many ARC variants with different
9722 instruction and register set characteristics. This flag prevents code
9723 compiled for one cpu to be linked with code compiled for another.
9724 No facility exists for handling variants that are ``almost identical''.
9725 This is an all or nothing option.
9727 @item -mcpu=@var{cpu}
9729 Compile code for ARC variant @var{cpu}.
9730 Which variants are supported depend on the configuration.
9731 All variants support @option{-mcpu=base}, this is the default.
9733 @item -mtext=@var{text-section}
9734 @itemx -mdata=@var{data-section}
9735 @itemx -mrodata=@var{readonly-data-section}
9739 Put functions, data, and readonly data in @var{text-section},
9740 @var{data-section}, and @var{readonly-data-section} respectively
9741 by default. This can be overridden with the @code{section} attribute.
9742 @xref{Variable Attributes}.
9747 @subsection NS32K Options
9748 @cindex NS32K options
9750 These are the @samp{-m} options defined for the 32000 series. The default
9751 values for these options depends on which style of 32000 was selected when
9752 the compiler was configured; the defaults for the most common choices are
9760 Generate output for a 32032. This is the default
9761 when the compiler is configured for 32032 and 32016 based systems.
9767 Generate output for a 32332. This is the default
9768 when the compiler is configured for 32332-based systems.
9774 Generate output for a 32532. This is the default
9775 when the compiler is configured for 32532-based systems.
9779 Generate output containing 32081 instructions for floating point.
9780 This is the default for all systems.
9784 Generate output containing 32381 instructions for floating point. This
9785 also implies @option{-m32081}. The 32381 is only compatible with the 32332
9786 and 32532 cpus. This is the default for the pc532-netbsd configuration.
9790 Try and generate multiply-add floating point instructions @code{polyF}
9791 and @code{dotF}. This option is only available if the @option{-m32381}
9792 option is in effect. Using these instructions requires changes to
9793 register allocation which generally has a negative impact on
9794 performance. This option should only be enabled when compiling code
9795 particularly likely to make heavy use of multiply-add instructions.
9798 @opindex mnomulti-add
9799 Do not try and generate multiply-add floating point instructions
9800 @code{polyF} and @code{dotF}. This is the default on all platforms.
9803 @opindex msoft-float
9804 Generate output containing library calls for floating point.
9805 @strong{Warning:} the requisite libraries may not be available.
9807 @item -mieee-compare
9808 @itemx -mno-ieee-compare
9809 @opindex mieee-compare
9810 @opindex mno-ieee-compare
9811 Control whether or not the compiler uses IEEE floating point
9812 comparisons. These handle correctly the case where the result of a
9813 comparison is unordered.
9814 @strong{Warning:} the requisite kernel support may not be available.
9817 @opindex mnobitfield
9818 Do not use the bit-field instructions. On some machines it is faster to
9819 use shifting and masking operations. This is the default for the pc532.
9823 Do use the bit-field instructions. This is the default for all platforms
9828 Use a different function-calling convention, in which functions
9829 that take a fixed number of arguments return pop their
9830 arguments on return with the @code{ret} instruction.
9832 This calling convention is incompatible with the one normally
9833 used on Unix, so you cannot use it if you need to call libraries
9834 compiled with the Unix compiler.
9836 Also, you must provide function prototypes for all functions that
9837 take variable numbers of arguments (including @code{printf});
9838 otherwise incorrect code will be generated for calls to those
9841 In addition, seriously incorrect code will result if you call a
9842 function with too many arguments. (Normally, extra arguments are
9843 harmlessly ignored.)
9845 This option takes its name from the 680x0 @code{rtd} instruction.
9850 Use a different function-calling convention where the first two arguments
9851 are passed in registers.
9853 This calling convention is incompatible with the one normally
9854 used on Unix, so you cannot use it if you need to call libraries
9855 compiled with the Unix compiler.
9858 @opindex mnoregparam
9859 Do not pass any arguments in registers. This is the default for all
9864 It is OK to use the sb as an index register which is always loaded with
9865 zero. This is the default for the pc532-netbsd target.
9869 The sb register is not available for use or has not been initialized to
9870 zero by the run time system. This is the default for all targets except
9871 the pc532-netbsd. It is also implied whenever @option{-mhimem} or
9872 @option{-fpic} is set.
9876 Many ns32000 series addressing modes use displacements of up to 512MB@.
9877 If an address is above 512MB then displacements from zero can not be used.
9878 This option causes code to be generated which can be loaded above 512MB@.
9879 This may be useful for operating systems or ROM code.
9883 Assume code will be loaded in the first 512MB of virtual address space.
9884 This is the default for all platforms.
9890 @subsection AVR Options
9893 These options are defined for AVR implementations:
9896 @item -mmcu=@var{mcu}
9898 Specify ATMEL AVR instruction set or MCU type.
9900 Instruction set avr1 is for the minimal AVR core, not supported by the C
9901 compiler, only for assembler programs (MCU types: at90s1200, attiny10,
9902 attiny11, attiny12, attiny15, attiny28).
9904 Instruction set avr2 (default) is for the classic AVR core with up to
9905 8K program memory space (MCU types: at90s2313, at90s2323, attiny22,
9906 at90s2333, at90s2343, at90s4414, at90s4433, at90s4434, at90s8515,
9907 at90c8534, at90s8535).
9909 Instruction set avr3 is for the classic AVR core with up to 128K program
9910 memory space (MCU types: atmega103, atmega603, at43usb320, at76c711).
9912 Instruction set avr4 is for the enhanced AVR core with up to 8K program
9913 memory space (MCU types: atmega8, atmega83, atmega85).
9915 Instruction set avr5 is for the enhanced AVR core with up to 128K program
9916 memory space (MCU types: atmega16, atmega161, atmega163, atmega32, atmega323,
9917 atmega64, atmega128, at43usb355, at94k).
9921 Output instruction sizes to the asm file.
9923 @item -minit-stack=@var{N}
9924 @opindex minit-stack
9925 Specify the initial stack address, which may be a symbol or numeric value,
9926 @samp{__stack} is the default.
9928 @item -mno-interrupts
9929 @opindex mno-interrupts
9930 Generated code is not compatible with hardware interrupts.
9931 Code size will be smaller.
9933 @item -mcall-prologues
9934 @opindex mcall-prologues
9935 Functions prologues/epilogues expanded as call to appropriate
9936 subroutines. Code size will be smaller.
9938 @item -mno-tablejump
9939 @opindex mno-tablejump
9940 Do not generate tablejump insns which sometimes increase code size.
9943 @opindex mtiny-stack
9944 Change only the low 8 bits of the stack pointer.
9948 @subsection MCore Options
9949 @cindex MCore options
9951 These are the @samp{-m} options defined for the Motorola M*Core
9959 @opindex mno-hardlit
9960 Inline constants into the code stream if it can be done in two
9961 instructions or less.
9967 Use the divide instruction. (Enabled by default).
9969 @item -mrelax-immediate
9970 @itemx -mno-relax-immediate
9971 @opindex mrelax-immediate
9972 @opindex mno-relax-immediate
9973 Allow arbitrary sized immediates in bit operations.
9975 @item -mwide-bitfields
9976 @itemx -mno-wide-bitfields
9977 @opindex mwide-bitfields
9978 @opindex mno-wide-bitfields
9979 Always treat bit-fields as int-sized.
9981 @item -m4byte-functions
9982 @itemx -mno-4byte-functions
9983 @opindex m4byte-functions
9984 @opindex mno-4byte-functions
9985 Force all functions to be aligned to a four byte boundary.
9987 @item -mcallgraph-data
9988 @itemx -mno-callgraph-data
9989 @opindex mcallgraph-data
9990 @opindex mno-callgraph-data
9991 Emit callgraph information.
9994 @itemx -mno-slow-bytes
9995 @opindex mslow-bytes
9996 @opindex mno-slow-bytes
9997 Prefer word access when reading byte quantities.
9999 @item -mlittle-endian
10000 @itemx -mbig-endian
10001 @opindex mlittle-endian
10002 @opindex mbig-endian
10003 Generate code for a little endian target.
10009 Generate code for the 210 processor.
10012 @node IA-64 Options
10013 @subsection IA-64 Options
10014 @cindex IA-64 Options
10016 These are the @samp{-m} options defined for the Intel IA-64 architecture.
10020 @opindex mbig-endian
10021 Generate code for a big endian target. This is the default for HP-UX@.
10023 @item -mlittle-endian
10024 @opindex mlittle-endian
10025 Generate code for a little endian target. This is the default for AIX5
10031 @opindex mno-gnu-as
10032 Generate (or don't) code for the GNU assembler. This is the default.
10033 @c Also, this is the default if the configure option @option{--with-gnu-as}
10039 @opindex mno-gnu-ld
10040 Generate (or don't) code for the GNU linker. This is the default.
10041 @c Also, this is the default if the configure option @option{--with-gnu-ld}
10046 Generate code that does not use a global pointer register. The result
10047 is not position independent code, and violates the IA-64 ABI@.
10049 @item -mvolatile-asm-stop
10050 @itemx -mno-volatile-asm-stop
10051 @opindex mvolatile-asm-stop
10052 @opindex mno-volatile-asm-stop
10053 Generate (or don't) a stop bit immediately before and after volatile asm
10058 Generate code that works around Itanium B step errata.
10060 @item -mregister-names
10061 @itemx -mno-register-names
10062 @opindex mregister-names
10063 @opindex mno-register-names
10064 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
10065 the stacked registers. This may make assembler output more readable.
10071 Disable (or enable) optimizations that use the small data section. This may
10072 be useful for working around optimizer bugs.
10074 @item -mconstant-gp
10075 @opindex mconstant-gp
10076 Generate code that uses a single constant global pointer value. This is
10077 useful when compiling kernel code.
10081 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
10082 This is useful when compiling firmware code.
10084 @item -minline-float-divide-min-latency
10085 @opindex minline-float-divide-min-latency
10086 Generate code for inline divides of floating point values
10087 using the minimum latency algorithm.
10089 @item -minline-float-divide-max-throughput
10090 @opindex minline-float-divide-max-throughput
10091 Generate code for inline divides of floating point values
10092 using the maximum throughput algorithm.
10094 @item -minline-int-divide-min-latency
10095 @opindex minline-int-divide-min-latency
10096 Generate code for inline divides of integer values
10097 using the minimum latency algorithm.
10099 @item -minline-int-divide-max-throughput
10100 @opindex minline-int-divide-max-throughput
10101 Generate code for inline divides of integer values
10102 using the maximum throughput algorithm.
10104 @item -mno-dwarf2-asm
10105 @itemx -mdwarf2-asm
10106 @opindex mno-dwarf2-asm
10107 @opindex mdwarf2-asm
10108 Don't (or do) generate assembler code for the DWARF2 line number debugging
10109 info. This may be useful when not using the GNU assembler.
10111 @item -mfixed-range=@var{register-range}
10112 @opindex mfixed-range
10113 Generate code treating the given register range as fixed registers.
10114 A fixed register is one that the register allocator can not use. This is
10115 useful when compiling kernel code. A register range is specified as
10116 two registers separated by a dash. Multiple register ranges can be
10117 specified separated by a comma.
10119 @item -mearly-stop-bits
10120 @itemx -mno-early-stop-bits
10121 @opindex mearly-stop-bits
10122 @opindex mno-early-stop-bits
10123 Allow stop bits to be placed earlier than immediately preceding the
10124 instruction that triggered the stop bit. This can improve instruction
10125 scheduling, but does not always do so.
10129 @subsection D30V Options
10130 @cindex D30V Options
10132 These @samp{-m} options are defined for D30V implementations:
10137 Link the @samp{.text}, @samp{.data}, @samp{.bss}, @samp{.strings},
10138 @samp{.rodata}, @samp{.rodata1}, @samp{.data1} sections into external
10139 memory, which starts at location @code{0x80000000}.
10142 @opindex mextmemory
10143 Same as the @option{-mextmem} switch.
10147 Link the @samp{.text} section into onchip text memory, which starts at
10148 location @code{0x0}. Also link @samp{.data}, @samp{.bss},
10149 @samp{.strings}, @samp{.rodata}, @samp{.rodata1}, @samp{.data1} sections
10150 into onchip data memory, which starts at location @code{0x20000000}.
10152 @item -mno-asm-optimize
10153 @itemx -masm-optimize
10154 @opindex mno-asm-optimize
10155 @opindex masm-optimize
10156 Disable (enable) passing @option{-O} to the assembler when optimizing.
10157 The assembler uses the @option{-O} option to automatically parallelize
10158 adjacent short instructions where possible.
10160 @item -mbranch-cost=@var{n}
10161 @opindex mbranch-cost
10162 Increase the internal costs of branches to @var{n}. Higher costs means
10163 that the compiler will issue more instructions to avoid doing a branch.
10166 @item -mcond-exec=@var{n}
10167 @opindex mcond-exec
10168 Specify the maximum number of conditionally executed instructions that
10169 replace a branch. The default is 4.
10172 @node S/390 and zSeries Options
10173 @subsection S/390 and zSeries Options
10174 @cindex S/390 and zSeries Options
10176 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
10180 @itemx -msoft-float
10181 @opindex mhard-float
10182 @opindex msoft-float
10183 Use (do not use) the hardware floating-point instructions and registers
10184 for floating-point operations. When @option{-msoft-float} is specified,
10185 functions in @file{libgcc.a} will be used to perform floating-point
10186 operations. When @option{-mhard-float} is specified, the compiler
10187 generates IEEE floating-point instructions. This is the default.
10190 @itemx -mno-backchain
10191 @opindex mbackchain
10192 @opindex mno-backchain
10193 Generate (or do not generate) code which maintains an explicit
10194 backchain within the stack frame that points to the caller's frame.
10195 This is currently needed to allow debugging. The default is to
10196 generate the backchain.
10199 @itemx -mno-small-exec
10200 @opindex msmall-exec
10201 @opindex mno-small-exec
10202 Generate (or do not generate) code using the @code{bras} instruction
10203 to do subroutine calls.
10204 This only works reliably if the total executable size does not
10205 exceed 64k. The default is to use the @code{basr} instruction instead,
10206 which does not have this limitation.
10212 When @option{-m31} is specified, generate code compliant to the
10213 Linux for S/390 ABI@. When @option{-m64} is specified, generate
10214 code compliant to the Linux for zSeries ABI@. This allows GCC in
10215 particular to generate 64-bit instructions. For the @samp{s390}
10216 targets, the default is @option{-m31}, while the @samp{s390x}
10217 targets default to @option{-m64}.
10223 When @option{-mzarch} is specified, generate code using the
10224 instructions available on z/Architecture.
10225 When @option{-mesa} is specified, generate code using the
10226 instructions available on ESA/390. Note that @option{-mesa} is
10227 not possible with @option{-m64}.
10228 When generating code compliant to the Linux for S/390 ABI,
10229 the default is @option{-mesa}. When generating code compliant
10230 to the Linux for zSeries ABI, the default is @option{-mzarch}.
10236 Generate (or do not generate) code using the @code{mvcle} instruction
10237 to perform block moves. When @option{-mno-mvcle} is specified,
10238 use a @code{mvc} loop instead. This is the default.
10244 Print (or do not print) additional debug information when compiling.
10245 The default is to not print debug information.
10247 @item -march=@var{cpu-type}
10249 Generate code that will run on @var{cpu-type}, which is the name of a system
10250 representing a certain processor type. Possible values for
10251 @var{cpu-type} are @samp{g5}, @samp{g6}, @samp{z900}, and @samp{z990}.
10252 When generating code using the instructions available on z/Architecture,
10253 the default is @option{-march=z900}. Otherwise, the default is
10254 @option{-march=g5}.
10256 @item -mtune=@var{cpu-type}
10258 Tune to @var{cpu-type} everything applicable about the generated code,
10259 except for the ABI and the set of available instructions.
10260 The list of @var{cpu-type} values is the same as for @option{-march}.
10261 The default is the value used for @option{-march}.
10266 @subsection CRIS Options
10267 @cindex CRIS Options
10269 These options are defined specifically for the CRIS ports.
10272 @item -march=@var{architecture-type}
10273 @itemx -mcpu=@var{architecture-type}
10276 Generate code for the specified architecture. The choices for
10277 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
10278 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX.
10279 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
10282 @item -mtune=@var{architecture-type}
10284 Tune to @var{architecture-type} everything applicable about the generated
10285 code, except for the ABI and the set of available instructions. The
10286 choices for @var{architecture-type} are the same as for
10287 @option{-march=@var{architecture-type}}.
10289 @item -mmax-stack-frame=@var{n}
10290 @opindex mmax-stack-frame
10291 Warn when the stack frame of a function exceeds @var{n} bytes.
10293 @item -melinux-stacksize=@var{n}
10294 @opindex melinux-stacksize
10295 Only available with the @samp{cris-axis-aout} target. Arranges for
10296 indications in the program to the kernel loader that the stack of the
10297 program should be set to @var{n} bytes.
10303 The options @option{-metrax4} and @option{-metrax100} are synonyms for
10304 @option{-march=v3} and @option{-march=v8} respectively.
10308 Enable CRIS-specific verbose debug-related information in the assembly
10309 code. This option also has the effect to turn off the @samp{#NO_APP}
10310 formatted-code indicator to the assembler at the beginning of the
10315 Do not use condition-code results from previous instruction; always emit
10316 compare and test instructions before use of condition codes.
10318 @item -mno-side-effects
10319 @opindex mno-side-effects
10320 Do not emit instructions with side-effects in addressing modes other than
10323 @item -mstack-align
10324 @itemx -mno-stack-align
10325 @itemx -mdata-align
10326 @itemx -mno-data-align
10327 @itemx -mconst-align
10328 @itemx -mno-const-align
10329 @opindex mstack-align
10330 @opindex mno-stack-align
10331 @opindex mdata-align
10332 @opindex mno-data-align
10333 @opindex mconst-align
10334 @opindex mno-const-align
10335 These options (no-options) arranges (eliminate arrangements) for the
10336 stack-frame, individual data and constants to be aligned for the maximum
10337 single data access size for the chosen CPU model. The default is to
10338 arrange for 32-bit alignment. ABI details such as structure layout are
10339 not affected by these options.
10347 Similar to the stack- data- and const-align options above, these options
10348 arrange for stack-frame, writable data and constants to all be 32-bit,
10349 16-bit or 8-bit aligned. The default is 32-bit alignment.
10351 @item -mno-prologue-epilogue
10352 @itemx -mprologue-epilogue
10353 @opindex mno-prologue-epilogue
10354 @opindex mprologue-epilogue
10355 With @option{-mno-prologue-epilogue}, the normal function prologue and
10356 epilogue that sets up the stack-frame are omitted and no return
10357 instructions or return sequences are generated in the code. Use this
10358 option only together with visual inspection of the compiled code: no
10359 warnings or errors are generated when call-saved registers must be saved,
10360 or storage for local variable needs to be allocated.
10364 @opindex mno-gotplt
10366 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
10367 instruction sequences that load addresses for functions from the PLT part
10368 of the GOT rather than (traditional on other architectures) calls to the
10369 PLT. The default is @option{-mgotplt}.
10373 Legacy no-op option only recognized with the cris-axis-aout target.
10377 Legacy no-op option only recognized with the cris-axis-elf and
10378 cris-axis-linux-gnu targets.
10382 Only recognized with the cris-axis-aout target, where it selects a
10383 GNU/linux-like multilib, include files and instruction set for
10384 @option{-march=v8}.
10388 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
10392 This option, recognized for the cris-axis-aout and cris-axis-elf arranges
10393 to link with input-output functions from a simulator library. Code,
10394 initialized data and zero-initialized data are allocated consecutively.
10398 Like @option{-sim}, but pass linker options to locate initialized data at
10399 0x40000000 and zero-initialized data at 0x80000000.
10403 @subsection MMIX Options
10404 @cindex MMIX Options
10406 These options are defined for the MMIX:
10410 @itemx -mno-libfuncs
10412 @opindex mno-libfuncs
10413 Specify that intrinsic library functions are being compiled, passing all
10414 values in registers, no matter the size.
10417 @itemx -mno-epsilon
10419 @opindex mno-epsilon
10420 Generate floating-point comparison instructions that compare with respect
10421 to the @code{rE} epsilon register.
10423 @item -mabi=mmixware
10425 @opindex mabi-mmixware
10427 Generate code that passes function parameters and return values that (in
10428 the called function) are seen as registers @code{$0} and up, as opposed to
10429 the GNU ABI which uses global registers @code{$231} and up.
10431 @item -mzero-extend
10432 @itemx -mno-zero-extend
10433 @opindex mzero-extend
10434 @opindex mno-zero-extend
10435 When reading data from memory in sizes shorter than 64 bits, use (do not
10436 use) zero-extending load instructions by default, rather than
10437 sign-extending ones.
10440 @itemx -mno-knuthdiv
10442 @opindex mno-knuthdiv
10443 Make the result of a division yielding a remainder have the same sign as
10444 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
10445 remainder follows the sign of the dividend. Both methods are
10446 arithmetically valid, the latter being almost exclusively used.
10448 @item -mtoplevel-symbols
10449 @itemx -mno-toplevel-symbols
10450 @opindex mtoplevel-symbols
10451 @opindex mno-toplevel-symbols
10452 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
10453 code can be used with the @code{PREFIX} assembly directive.
10457 Generate an executable in the ELF format, rather than the default
10458 @samp{mmo} format used by the @command{mmix} simulator.
10460 @item -mbranch-predict
10461 @itemx -mno-branch-predict
10462 @opindex mbranch-predict
10463 @opindex mno-branch-predict
10464 Use (do not use) the probable-branch instructions, when static branch
10465 prediction indicates a probable branch.
10467 @item -mbase-addresses
10468 @itemx -mno-base-addresses
10469 @opindex mbase-addresses
10470 @opindex mno-base-addresses
10471 Generate (do not generate) code that uses @emph{base addresses}. Using a
10472 base address automatically generates a request (handled by the assembler
10473 and the linker) for a constant to be set up in a global register. The
10474 register is used for one or more base address requests within the range 0
10475 to 255 from the value held in the register. The generally leads to short
10476 and fast code, but the number of different data items that can be
10477 addressed is limited. This means that a program that uses lots of static
10478 data may require @option{-mno-base-addresses}.
10480 @item -msingle-exit
10481 @itemx -mno-single-exit
10482 @opindex msingle-exit
10483 @opindex mno-single-exit
10484 Force (do not force) generated code to have a single exit point in each
10488 @node PDP-11 Options
10489 @subsection PDP-11 Options
10490 @cindex PDP-11 Options
10492 These options are defined for the PDP-11:
10497 Use hardware FPP floating point. This is the default. (FIS floating
10498 point on the PDP-11/40 is not supported.)
10501 @opindex msoft-float
10502 Do not use hardware floating point.
10506 Return floating-point results in ac0 (fr0 in Unix assembler syntax).
10510 Return floating-point results in memory. This is the default.
10514 Generate code for a PDP-11/40.
10518 Generate code for a PDP-11/45. This is the default.
10522 Generate code for a PDP-11/10.
10524 @item -mbcopy-builtin
10525 @opindex bcopy-builtin
10526 Use inline @code{movstrhi} patterns for copying memory. This is the
10531 Do not use inline @code{movstrhi} patterns for copying memory.
10537 Use 16-bit @code{int}. This is the default.
10543 Use 32-bit @code{int}.
10546 @itemx -mno-float32
10548 @opindex mno-float32
10549 Use 64-bit @code{float}. This is the default.
10554 @opindex mno-float64
10555 Use 32-bit @code{float}.
10559 Use @code{abshi2} pattern. This is the default.
10563 Do not use @code{abshi2} pattern.
10565 @item -mbranch-expensive
10566 @opindex mbranch-expensive
10567 Pretend that branches are expensive. This is for experimenting with
10568 code generation only.
10570 @item -mbranch-cheap
10571 @opindex mbranch-cheap
10572 Do not pretend that branches are expensive. This is the default.
10576 Generate code for a system with split I&D.
10580 Generate code for a system without split I&D. This is the default.
10584 Use Unix assembler syntax. This is the default when configured for
10585 @samp{pdp11-*-bsd}.
10589 Use DEC assembler syntax. This is the default when configured for any
10590 PDP-11 target other than @samp{pdp11-*-bsd}.
10593 @node Xstormy16 Options
10594 @subsection Xstormy16 Options
10595 @cindex Xstormy16 Options
10597 These options are defined for Xstormy16:
10602 Choose startup files and linker script suitable for the simulator.
10606 @subsection FRV Options
10607 @cindex FRV Options
10613 Only use the first 32 general purpose registers.
10618 Use all 64 general purpose registers.
10623 Use only the first 32 floating point registers.
10628 Use all 64 floating point registers
10631 @opindex mhard-float
10633 Use hardware instructions for floating point operations.
10636 @opindex msoft-float
10638 Use library routines for floating point operations.
10643 Dynamically allocate condition code registers.
10648 Do not try to dynamically allocate condition code registers, only
10649 use @code{icc0} and @code{fcc0}.
10654 Change ABI to use double word insns.
10659 Do not use double word instructions.
10664 Use floating point double instructions.
10667 @opindex mno-double
10669 Do not use floating point double instructions.
10674 Use media instructions.
10679 Do not use media instructions.
10684 Use multiply and add/subtract instructions.
10687 @opindex mno-muladd
10689 Do not use multiply and add/subtract instructions.
10691 @item -mlibrary-pic
10692 @opindex mlibrary-pic
10694 Enable PIC support for building libraries
10699 Use only the first four media accumulator registers.
10704 Use all eight media accumulator registers.
10709 Pack VLIW instructions.
10714 Do not pack VLIW instructions.
10717 @opindex mno-eflags
10719 Do not mark ABI switches in e_flags.
10722 @opindex mcond-move
10724 Enable the use of conditional-move instructions (default).
10726 This switch is mainly for debugging the compiler and will likely be removed
10727 in a future version.
10729 @item -mno-cond-move
10730 @opindex mno-cond-move
10732 Disable the use of conditional-move instructions.
10734 This switch is mainly for debugging the compiler and will likely be removed
10735 in a future version.
10740 Enable the use of conditional set instructions (default).
10742 This switch is mainly for debugging the compiler and will likely be removed
10743 in a future version.
10748 Disable the use of conditional set instructions.
10750 This switch is mainly for debugging the compiler and will likely be removed
10751 in a future version.
10754 @opindex mcond-exec
10756 Enable the use of conditional execution (default).
10758 This switch is mainly for debugging the compiler and will likely be removed
10759 in a future version.
10761 @item -mno-cond-exec
10762 @opindex mno-cond-exec
10764 Disable the use of conditional execution.
10766 This switch is mainly for debugging the compiler and will likely be removed
10767 in a future version.
10769 @item -mvliw-branch
10770 @opindex mvliw-branch
10772 Run a pass to pack branches into VLIW instructions (default).
10774 This switch is mainly for debugging the compiler and will likely be removed
10775 in a future version.
10777 @item -mno-vliw-branch
10778 @opindex mno-vliw-branch
10780 Do not run a pass to pack branches into VLIW instructions.
10782 This switch is mainly for debugging the compiler and will likely be removed
10783 in a future version.
10785 @item -mmulti-cond-exec
10786 @opindex mmulti-cond-exec
10788 Enable optimization of @code{&&} and @code{||} in conditional execution
10791 This switch is mainly for debugging the compiler and will likely be removed
10792 in a future version.
10794 @item -mno-multi-cond-exec
10795 @opindex mno-multi-cond-exec
10797 Disable optimization of @code{&&} and @code{||} in conditional execution.
10799 This switch is mainly for debugging the compiler and will likely be removed
10800 in a future version.
10802 @item -mnested-cond-exec
10803 @opindex mnested-cond-exec
10805 Enable nested conditional execution optimizations (default).
10807 This switch is mainly for debugging the compiler and will likely be removed
10808 in a future version.
10810 @item -mno-nested-cond-exec
10811 @opindex mno-nested-cond-exec
10813 Disable nested conditional execution optimizations.
10815 This switch is mainly for debugging the compiler and will likely be removed
10816 in a future version.
10818 @item -mtomcat-stats
10819 @opindex mtomcat-stats
10821 Cause gas to print out tomcat statistics.
10823 @item -mcpu=@var{cpu}
10826 Select the processor type for which to generate code. Possible values are
10827 @samp{simple}, @samp{tomcat}, @samp{fr500}, @samp{fr400}, @samp{fr300},
10832 @node Xtensa Options
10833 @subsection Xtensa Options
10834 @cindex Xtensa Options
10836 The Xtensa architecture is designed to support many different
10837 configurations. The compiler's default options can be set to match a
10838 particular Xtensa configuration by copying a configuration file into the
10839 GCC sources when building GCC@. The options below may be used to
10840 override the default options.
10844 @itemx -mlittle-endian
10845 @opindex mbig-endian
10846 @opindex mlittle-endian
10847 Specify big-endian or little-endian byte ordering for the target Xtensa
10851 @itemx -mno-density
10853 @opindex mno-density
10854 Enable or disable use of the optional Xtensa code density instructions.
10857 @itemx -mno-const16
10859 @opindex mno-const16
10860 Enable or disable use of @code{CONST16} instructions for loading
10861 constant values. The @code{CONST16} instruction is currently not a
10862 standard option from Tensilica. When enabled, @code{CONST16}
10863 instructions are always used in place of the standard @code{L32R}
10864 instructions. The use of @code{CONST16} is enabled by default only if
10865 the @code{L32R} instruction is not available.
10871 Enable or disable use of the Xtensa @code{ABS} instruction for absolute
10878 Enable or disable use of the Xtensa @code{ADDX} and @code{SUBX}
10885 Enable or disable use of the Xtensa MAC16 option. When enabled, GCC
10886 will generate MAC16 instructions from standard C code, with the
10887 limitation that it will use neither the MR register file nor any
10888 instruction that operates on the MR registers. When this option is
10889 disabled, GCC will translate 16-bit multiply/accumulate operations to a
10890 combination of core instructions and library calls, depending on whether
10891 any other multiplier options are enabled.
10897 Enable or disable use of the 16-bit integer multiplier option. When
10898 enabled, the compiler will generate 16-bit multiply instructions for
10899 multiplications of 16 bits or smaller in standard C code. When this
10900 option is disabled, the compiler will either use 32-bit multiply or
10901 MAC16 instructions if they are available or generate library calls to
10902 perform the multiply operations using shifts and adds.
10908 Enable or disable use of the 32-bit integer multiplier option. When
10909 enabled, the compiler will generate 32-bit multiply instructions for
10910 multiplications of 32 bits or smaller in standard C code. When this
10911 option is disabled, the compiler will generate library calls to perform
10912 the multiply operations using either shifts and adds or 16-bit multiply
10913 instructions if they are available.
10919 Enable or disable use of the optional normalization shift amount
10920 (@code{NSA}) instructions to implement the built-in @code{ffs} function.
10925 @opindex mno-minmax
10926 Enable or disable use of the optional minimum and maximum value
10933 Enable or disable use of the optional sign extend (@code{SEXT})
10937 @itemx -mno-booleans
10939 @opindex mno-booleans
10940 Enable or disable support for the boolean register file used by Xtensa
10941 coprocessors. This is not typically useful by itself but may be
10942 required for other options that make use of the boolean registers (e.g.,
10943 the floating-point option).
10946 @itemx -msoft-float
10947 @opindex mhard-float
10948 @opindex msoft-float
10949 Enable or disable use of the floating-point option. When enabled, GCC
10950 generates floating-point instructions for 32-bit @code{float}
10951 operations. When this option is disabled, GCC generates library calls
10952 to emulate 32-bit floating-point operations using integer instructions.
10953 Regardless of this option, 64-bit @code{double} operations are always
10954 emulated with calls to library functions.
10957 @itemx -mno-fused-madd
10958 @opindex mfused-madd
10959 @opindex mno-fused-madd
10960 Enable or disable use of fused multiply/add and multiply/subtract
10961 instructions in the floating-point option. This has no effect if the
10962 floating-point option is not also enabled. Disabling fused multiply/add
10963 and multiply/subtract instructions forces the compiler to use separate
10964 instructions for the multiply and add/subtract operations. This may be
10965 desirable in some cases where strict IEEE 754-compliant results are
10966 required: the fused multiply add/subtract instructions do not round the
10967 intermediate result, thereby producing results with @emph{more} bits of
10968 precision than specified by the IEEE standard. Disabling fused multiply
10969 add/subtract instructions also ensures that the program output is not
10970 sensitive to the compiler's ability to combine multiply and add/subtract
10973 @item -mtext-section-literals
10974 @itemx -mno-text-section-literals
10975 @opindex mtext-section-literals
10976 @opindex mno-text-section-literals
10977 Control the treatment of literal pools. The default is
10978 @option{-mno-text-section-literals}, which places literals in a separate
10979 section in the output file. This allows the literal pool to be placed
10980 in a data RAM/ROM, and it also allows the linker to combine literal
10981 pools from separate object files to remove redundant literals and
10982 improve code size. With @option{-mtext-section-literals}, the literals
10983 are interspersed in the text section in order to keep them as close as
10984 possible to their references. This may be necessary for large assembly
10987 @item -mtarget-align
10988 @itemx -mno-target-align
10989 @opindex mtarget-align
10990 @opindex mno-target-align
10991 When this option is enabled, GCC instructs the assembler to
10992 automatically align instructions to reduce branch penalties at the
10993 expense of some code density. The assembler attempts to widen density
10994 instructions to align branch targets and the instructions following call
10995 instructions. If there are not enough preceding safe density
10996 instructions to align a target, no widening will be performed. The
10997 default is @option{-mtarget-align}. These options do not affect the
10998 treatment of auto-aligned instructions like @code{LOOP}, which the
10999 assembler will always align, either by widening density instructions or
11000 by inserting no-op instructions.
11003 @itemx -mno-longcalls
11004 @opindex mlongcalls
11005 @opindex mno-longcalls
11006 When this option is enabled, GCC instructs the assembler to translate
11007 direct calls to indirect calls unless it can determine that the target
11008 of a direct call is in the range allowed by the call instruction. This
11009 translation typically occurs for calls to functions in other source
11010 files. Specifically, the assembler translates a direct @code{CALL}
11011 instruction into an @code{L32R} followed by a @code{CALLX} instruction.
11012 The default is @option{-mno-longcalls}. This option should be used in
11013 programs where the call target can potentially be out of range. This
11014 option is implemented in the assembler, not the compiler, so the
11015 assembly code generated by GCC will still show direct call
11016 instructions---look at the disassembled object code to see the actual
11017 instructions. Note that the assembler will use an indirect call for
11018 every cross-file call, not just those that really will be out of range.
11021 @node Code Gen Options
11022 @section Options for Code Generation Conventions
11023 @cindex code generation conventions
11024 @cindex options, code generation
11025 @cindex run-time options
11027 These machine-independent options control the interface conventions
11028 used in code generation.
11030 Most of them have both positive and negative forms; the negative form
11031 of @option{-ffoo} would be @option{-fno-foo}. In the table below, only
11032 one of the forms is listed---the one which is not the default. You
11033 can figure out the other form by either removing @samp{no-} or adding
11037 @item -fbounds-check
11038 @opindex fbounds-check
11039 For front-ends that support it, generate additional code to check that
11040 indices used to access arrays are within the declared range. This is
11041 currently only supported by the Java and Fortran 77 front-ends, where
11042 this option defaults to true and false respectively.
11046 This option generates traps for signed overflow on addition, subtraction,
11047 multiplication operations.
11051 This option instructs the compiler to assume that signed arithmetic
11052 overflow of addition, subtraction and multiplication wraps around
11053 using twos-complement representation. This flag enables some optimizations
11054 and disables other. This option is enabled by default for the Java
11055 front-end, as required by the Java language specification.
11058 @opindex fexceptions
11059 Enable exception handling. Generates extra code needed to propagate
11060 exceptions. For some targets, this implies GCC will generate frame
11061 unwind information for all functions, which can produce significant data
11062 size overhead, although it does not affect execution. If you do not
11063 specify this option, GCC will enable it by default for languages like
11064 C++ which normally require exception handling, and disable it for
11065 languages like C that do not normally require it. However, you may need
11066 to enable this option when compiling C code that needs to interoperate
11067 properly with exception handlers written in C++. You may also wish to
11068 disable this option if you are compiling older C++ programs that don't
11069 use exception handling.
11071 @item -fnon-call-exceptions
11072 @opindex fnon-call-exceptions
11073 Generate code that allows trapping instructions to throw exceptions.
11074 Note that this requires platform-specific runtime support that does
11075 not exist everywhere. Moreover, it only allows @emph{trapping}
11076 instructions to throw exceptions, i.e.@: memory references or floating
11077 point instructions. It does not allow exceptions to be thrown from
11078 arbitrary signal handlers such as @code{SIGALRM}.
11080 @item -funwind-tables
11081 @opindex funwind-tables
11082 Similar to @option{-fexceptions}, except that it will just generate any needed
11083 static data, but will not affect the generated code in any other way.
11084 You will normally not enable this option; instead, a language processor
11085 that needs this handling would enable it on your behalf.
11087 @item -fasynchronous-unwind-tables
11088 @opindex funwind-tables
11089 Generate unwind table in dwarf2 format, if supported by target machine. The
11090 table is exact at each instruction boundary, so it can be used for stack
11091 unwinding from asynchronous events (such as debugger or garbage collector).
11093 @item -fpcc-struct-return
11094 @opindex fpcc-struct-return
11095 Return ``short'' @code{struct} and @code{union} values in memory like
11096 longer ones, rather than in registers. This convention is less
11097 efficient, but it has the advantage of allowing intercallability between
11098 GCC-compiled files and files compiled with other compilers, particularly
11099 the Portable C Compiler (pcc).
11101 The precise convention for returning structures in memory depends
11102 on the target configuration macros.
11104 Short structures and unions are those whose size and alignment match
11105 that of some integer type.
11107 @strong{Warning:} code compiled with the @option{-fpcc-struct-return}
11108 switch is not binary compatible with code compiled with the
11109 @option{-freg-struct-return} switch.
11110 Use it to conform to a non-default application binary interface.
11112 @item -freg-struct-return
11113 @opindex freg-struct-return
11114 Return @code{struct} and @code{union} values in registers when possible.
11115 This is more efficient for small structures than
11116 @option{-fpcc-struct-return}.
11118 If you specify neither @option{-fpcc-struct-return} nor
11119 @option{-freg-struct-return}, GCC defaults to whichever convention is
11120 standard for the target. If there is no standard convention, GCC
11121 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
11122 the principal compiler. In those cases, we can choose the standard, and
11123 we chose the more efficient register return alternative.
11125 @strong{Warning:} code compiled with the @option{-freg-struct-return}
11126 switch is not binary compatible with code compiled with the
11127 @option{-fpcc-struct-return} switch.
11128 Use it to conform to a non-default application binary interface.
11130 @item -fshort-enums
11131 @opindex fshort-enums
11132 Allocate to an @code{enum} type only as many bytes as it needs for the
11133 declared range of possible values. Specifically, the @code{enum} type
11134 will be equivalent to the smallest integer type which has enough room.
11136 @strong{Warning:} the @option{-fshort-enums} switch causes GCC to generate
11137 code that is not binary compatible with code generated without that switch.
11138 Use it to conform to a non-default application binary interface.
11140 @item -fshort-double
11141 @opindex fshort-double
11142 Use the same size for @code{double} as for @code{float}.
11144 @strong{Warning:} the @option{-fshort-double} switch causes GCC to generate
11145 code that is not binary compatible with code generated without that switch.
11146 Use it to conform to a non-default application binary interface.
11148 @item -fshort-wchar
11149 @opindex fshort-wchar
11150 Override the underlying type for @samp{wchar_t} to be @samp{short
11151 unsigned int} instead of the default for the target. This option is
11152 useful for building programs to run under WINE@.
11154 @strong{Warning:} the @option{-fshort-wchar} switch causes GCC to generate
11155 code that is not binary compatible with code generated without that switch.
11156 Use it to conform to a non-default application binary interface.
11158 @item -fshared-data
11159 @opindex fshared-data
11160 Requests that the data and non-@code{const} variables of this
11161 compilation be shared data rather than private data. The distinction
11162 makes sense only on certain operating systems, where shared data is
11163 shared between processes running the same program, while private data
11164 exists in one copy per process.
11167 @opindex fno-common
11168 In C, allocate even uninitialized global variables in the data section of the
11169 object file, rather than generating them as common blocks. This has the
11170 effect that if the same variable is declared (without @code{extern}) in
11171 two different compilations, you will get an error when you link them.
11172 The only reason this might be useful is if you wish to verify that the
11173 program will work on other systems which always work this way.
11177 Ignore the @samp{#ident} directive.
11179 @item -fno-gnu-linker
11180 @opindex fno-gnu-linker
11181 Do not output global initializations (such as C++ constructors and
11182 destructors) in the form used by the GNU linker (on systems where the GNU
11183 linker is the standard method of handling them). Use this option when
11184 you want to use a non-GNU linker, which also requires using the
11185 @command{collect2} program to make sure the system linker includes
11186 constructors and destructors. (@command{collect2} is included in the GCC
11187 distribution.) For systems which @emph{must} use @command{collect2}, the
11188 compiler driver @command{gcc} is configured to do this automatically.
11190 @item -finhibit-size-directive
11191 @opindex finhibit-size-directive
11192 Don't output a @code{.size} assembler directive, or anything else that
11193 would cause trouble if the function is split in the middle, and the
11194 two halves are placed at locations far apart in memory. This option is
11195 used when compiling @file{crtstuff.c}; you should not need to use it
11198 @item -fverbose-asm
11199 @opindex fverbose-asm
11200 Put extra commentary information in the generated assembly code to
11201 make it more readable. This option is generally only of use to those
11202 who actually need to read the generated assembly code (perhaps while
11203 debugging the compiler itself).
11205 @option{-fno-verbose-asm}, the default, causes the
11206 extra information to be omitted and is useful when comparing two assembler
11211 @cindex global offset table
11213 Generate position-independent code (PIC) suitable for use in a shared
11214 library, if supported for the target machine. Such code accesses all
11215 constant addresses through a global offset table (GOT)@. The dynamic
11216 loader resolves the GOT entries when the program starts (the dynamic
11217 loader is not part of GCC; it is part of the operating system). If
11218 the GOT size for the linked executable exceeds a machine-specific
11219 maximum size, you get an error message from the linker indicating that
11220 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
11221 instead. (These maximums are 16k on the m88k, 8k on the SPARC, and 32k
11222 on the m68k and RS/6000. The 386 has no such limit.)
11224 Position-independent code requires special support, and therefore works
11225 only on certain machines. For the 386, GCC supports PIC for System V
11226 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
11227 position-independent.
11231 If supported for the target machine, emit position-independent code,
11232 suitable for dynamic linking and avoiding any limit on the size of the
11233 global offset table. This option makes a difference on the m68k, m88k,
11236 Position-independent code requires special support, and therefore works
11237 only on certain machines.
11243 These options are similar to @option{-fpic} and @option{-fPIC}, but
11244 generated position independent code can be only linked into executables.
11245 Usually these options are used when @option{-pie} GCC option will be
11246 used during linking.
11248 @item -ffixed-@var{reg}
11250 Treat the register named @var{reg} as a fixed register; generated code
11251 should never refer to it (except perhaps as a stack pointer, frame
11252 pointer or in some other fixed role).
11254 @var{reg} must be the name of a register. The register names accepted
11255 are machine-specific and are defined in the @code{REGISTER_NAMES}
11256 macro in the machine description macro file.
11258 This flag does not have a negative form, because it specifies a
11261 @item -fcall-used-@var{reg}
11262 @opindex fcall-used
11263 Treat the register named @var{reg} as an allocable register that is
11264 clobbered by function calls. It may be allocated for temporaries or
11265 variables that do not live across a call. Functions compiled this way
11266 will not save and restore the register @var{reg}.
11268 It is an error to used this flag with the frame pointer or stack pointer.
11269 Use of this flag for other registers that have fixed pervasive roles in
11270 the machine's execution model will produce disastrous results.
11272 This flag does not have a negative form, because it specifies a
11275 @item -fcall-saved-@var{reg}
11276 @opindex fcall-saved
11277 Treat the register named @var{reg} as an allocable register saved by
11278 functions. It may be allocated even for temporaries or variables that
11279 live across a call. Functions compiled this way will save and restore
11280 the register @var{reg} if they use it.
11282 It is an error to used this flag with the frame pointer or stack pointer.
11283 Use of this flag for other registers that have fixed pervasive roles in
11284 the machine's execution model will produce disastrous results.
11286 A different sort of disaster will result from the use of this flag for
11287 a register in which function values may be returned.
11289 This flag does not have a negative form, because it specifies a
11292 @item -fpack-struct
11293 @opindex fpack-struct
11294 Pack all structure members together without holes.
11296 @strong{Warning:} the @option{-fpack-struct} switch causes GCC to generate
11297 code that is not binary compatible with code generated without that switch.
11298 Additionally, it makes the code suboptimal.
11299 Use it to conform to a non-default application binary interface.
11301 @item -finstrument-functions
11302 @opindex finstrument-functions
11303 Generate instrumentation calls for entry and exit to functions. Just
11304 after function entry and just before function exit, the following
11305 profiling functions will be called with the address of the current
11306 function and its call site. (On some platforms,
11307 @code{__builtin_return_address} does not work beyond the current
11308 function, so the call site information may not be available to the
11309 profiling functions otherwise.)
11312 void __cyg_profile_func_enter (void *this_fn,
11314 void __cyg_profile_func_exit (void *this_fn,
11318 The first argument is the address of the start of the current function,
11319 which may be looked up exactly in the symbol table.
11321 This instrumentation is also done for functions expanded inline in other
11322 functions. The profiling calls will indicate where, conceptually, the
11323 inline function is entered and exited. This means that addressable
11324 versions of such functions must be available. If all your uses of a
11325 function are expanded inline, this may mean an additional expansion of
11326 code size. If you use @samp{extern inline} in your C code, an
11327 addressable version of such functions must be provided. (This is
11328 normally the case anyways, but if you get lucky and the optimizer always
11329 expands the functions inline, you might have gotten away without
11330 providing static copies.)
11332 A function may be given the attribute @code{no_instrument_function}, in
11333 which case this instrumentation will not be done. This can be used, for
11334 example, for the profiling functions listed above, high-priority
11335 interrupt routines, and any functions from which the profiling functions
11336 cannot safely be called (perhaps signal handlers, if the profiling
11337 routines generate output or allocate memory).
11339 @item -fstack-check
11340 @opindex fstack-check
11341 Generate code to verify that you do not go beyond the boundary of the
11342 stack. You should specify this flag if you are running in an
11343 environment with multiple threads, but only rarely need to specify it in
11344 a single-threaded environment since stack overflow is automatically
11345 detected on nearly all systems if there is only one stack.
11347 Note that this switch does not actually cause checking to be done; the
11348 operating system must do that. The switch causes generation of code
11349 to ensure that the operating system sees the stack being extended.
11351 @item -fstack-limit-register=@var{reg}
11352 @itemx -fstack-limit-symbol=@var{sym}
11353 @itemx -fno-stack-limit
11354 @opindex fstack-limit-register
11355 @opindex fstack-limit-symbol
11356 @opindex fno-stack-limit
11357 Generate code to ensure that the stack does not grow beyond a certain value,
11358 either the value of a register or the address of a symbol. If the stack
11359 would grow beyond the value, a signal is raised. For most targets,
11360 the signal is raised before the stack overruns the boundary, so
11361 it is possible to catch the signal without taking special precautions.
11363 For instance, if the stack starts at absolute address @samp{0x80000000}
11364 and grows downwards, you can use the flags
11365 @option{-fstack-limit-symbol=__stack_limit} and
11366 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
11367 of 128KB@. Note that this may only work with the GNU linker.
11369 @cindex aliasing of parameters
11370 @cindex parameters, aliased
11371 @item -fargument-alias
11372 @itemx -fargument-noalias
11373 @itemx -fargument-noalias-global
11374 @opindex fargument-alias
11375 @opindex fargument-noalias
11376 @opindex fargument-noalias-global
11377 Specify the possible relationships among parameters and between
11378 parameters and global data.
11380 @option{-fargument-alias} specifies that arguments (parameters) may
11381 alias each other and may alias global storage.@*
11382 @option{-fargument-noalias} specifies that arguments do not alias
11383 each other, but may alias global storage.@*
11384 @option{-fargument-noalias-global} specifies that arguments do not
11385 alias each other and do not alias global storage.
11387 Each language will automatically use whatever option is required by
11388 the language standard. You should not need to use these options yourself.
11390 @item -fleading-underscore
11391 @opindex fleading-underscore
11392 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
11393 change the way C symbols are represented in the object file. One use
11394 is to help link with legacy assembly code.
11396 @strong{Warning:} the @option{-fleading-underscore} switch causes GCC to
11397 generate code that is not binary compatible with code generated without that
11398 switch. Use it to conform to a non-default application binary interface.
11399 Not all targets provide complete support for this switch.
11401 @item -ftls-model=@var{model}
11402 Alter the thread-local storage model to be used (@pxref{Thread-Local}).
11403 The @var{model} argument should be one of @code{global-dynamic},
11404 @code{local-dynamic}, @code{initial-exec} or @code{local-exec}.
11406 The default without @option{-fpic} is @code{initial-exec}; with
11407 @option{-fpic} the default is @code{global-dynamic}.
11412 @node Environment Variables
11413 @section Environment Variables Affecting GCC
11414 @cindex environment variables
11416 @c man begin ENVIRONMENT
11417 This section describes several environment variables that affect how GCC
11418 operates. Some of them work by specifying directories or prefixes to use
11419 when searching for various kinds of files. Some are used to specify other
11420 aspects of the compilation environment.
11422 Note that you can also specify places to search using options such as
11423 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
11424 take precedence over places specified using environment variables, which
11425 in turn take precedence over those specified by the configuration of GCC@.
11426 @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint,
11427 GNU Compiler Collection (GCC) Internals}.
11432 @c @itemx LC_COLLATE
11434 @c @itemx LC_MONETARY
11435 @c @itemx LC_NUMERIC
11440 @c @findex LC_COLLATE
11441 @findex LC_MESSAGES
11442 @c @findex LC_MONETARY
11443 @c @findex LC_NUMERIC
11447 These environment variables control the way that GCC uses
11448 localization information that allow GCC to work with different
11449 national conventions. GCC inspects the locale categories
11450 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
11451 so. These locale categories can be set to any value supported by your
11452 installation. A typical value is @samp{en_UK} for English in the United
11455 The @env{LC_CTYPE} environment variable specifies character
11456 classification. GCC uses it to determine the character boundaries in
11457 a string; this is needed for some multibyte encodings that contain quote
11458 and escape characters that would otherwise be interpreted as a string
11461 The @env{LC_MESSAGES} environment variable specifies the language to
11462 use in diagnostic messages.
11464 If the @env{LC_ALL} environment variable is set, it overrides the value
11465 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
11466 and @env{LC_MESSAGES} default to the value of the @env{LANG}
11467 environment variable. If none of these variables are set, GCC
11468 defaults to traditional C English behavior.
11472 If @env{TMPDIR} is set, it specifies the directory to use for temporary
11473 files. GCC uses temporary files to hold the output of one stage of
11474 compilation which is to be used as input to the next stage: for example,
11475 the output of the preprocessor, which is the input to the compiler
11478 @item GCC_EXEC_PREFIX
11479 @findex GCC_EXEC_PREFIX
11480 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
11481 names of the subprograms executed by the compiler. No slash is added
11482 when this prefix is combined with the name of a subprogram, but you can
11483 specify a prefix that ends with a slash if you wish.
11485 If @env{GCC_EXEC_PREFIX} is not set, GCC will attempt to figure out
11486 an appropriate prefix to use based on the pathname it was invoked with.
11488 If GCC cannot find the subprogram using the specified prefix, it
11489 tries looking in the usual places for the subprogram.
11491 The default value of @env{GCC_EXEC_PREFIX} is
11492 @file{@var{prefix}/lib/gcc/} where @var{prefix} is the value
11493 of @code{prefix} when you ran the @file{configure} script.
11495 Other prefixes specified with @option{-B} take precedence over this prefix.
11497 This prefix is also used for finding files such as @file{crt0.o} that are
11500 In addition, the prefix is used in an unusual way in finding the
11501 directories to search for header files. For each of the standard
11502 directories whose name normally begins with @samp{/usr/local/lib/gcc}
11503 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
11504 replacing that beginning with the specified prefix to produce an
11505 alternate directory name. Thus, with @option{-Bfoo/}, GCC will search
11506 @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}.
11507 These alternate directories are searched first; the standard directories
11510 @item COMPILER_PATH
11511 @findex COMPILER_PATH
11512 The value of @env{COMPILER_PATH} is a colon-separated list of
11513 directories, much like @env{PATH}. GCC tries the directories thus
11514 specified when searching for subprograms, if it can't find the
11515 subprograms using @env{GCC_EXEC_PREFIX}.
11518 @findex LIBRARY_PATH
11519 The value of @env{LIBRARY_PATH} is a colon-separated list of
11520 directories, much like @env{PATH}. When configured as a native compiler,
11521 GCC tries the directories thus specified when searching for special
11522 linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking
11523 using GCC also uses these directories when searching for ordinary
11524 libraries for the @option{-l} option (but directories specified with
11525 @option{-L} come first).
11529 @cindex locale definition
11530 This variable is used to pass locale information to the compiler. One way in
11531 which this information is used is to determine the character set to be used
11532 when character literals, string literals and comments are parsed in C and C++.
11533 When the compiler is configured to allow multibyte characters,
11534 the following values for @env{LANG} are recognized:
11538 Recognize JIS characters.
11540 Recognize SJIS characters.
11542 Recognize EUCJP characters.
11545 If @env{LANG} is not defined, or if it has some other value, then the
11546 compiler will use mblen and mbtowc as defined by the default locale to
11547 recognize and translate multibyte characters.
11551 Some additional environments variables affect the behavior of the
11554 @include cppenv.texi
11558 @node Precompiled Headers
11559 @section Using Precompiled Headers
11560 @cindex precompiled headers
11561 @cindex speed of compilation
11563 Often large projects have many header files that are included in every
11564 source file. The time the compiler takes to process these header files
11565 over and over again can account for nearly all of the time required to
11566 build the project. To make builds faster, GCC allows users to
11567 `precompile' a header file; then, if builds can use the precompiled
11568 header file they will be much faster.
11570 To create a precompiled header file, simply compile it as you would any
11571 other file, if necessary using the @option{-x} option to make the driver
11572 treat it as a C or C++ header file. You will probably want to use a
11573 tool like @command{make} to keep the precompiled header up-to-date when
11574 the headers it contains change.
11576 A precompiled header file will be searched for when @code{#include} is
11577 seen in the compilation. As it searches for the included file
11578 (@pxref{Search Path,,Search Path,cpp.info,The C Preprocessor}) the
11579 compiler looks for a precompiled header in each directory just before it
11580 looks for the include file in that directory. The name searched for is
11581 the name specified in the @code{#include} with @samp{.gch} appended. If
11582 the precompiled header file can't be used, it is ignored.
11584 For instance, if you have @code{#include "all.h"}, and you have
11585 @file{all.h.gch} in the same directory as @file{all.h}, then the
11586 precompiled header file will be used if possible, and the original
11587 header will be used otherwise.
11589 Alternatively, you might decide to put the precompiled header file in a
11590 directory and use @option{-I} to ensure that directory is searched
11591 before (or instead of) the directory containing the original header.
11592 Then, if you want to check that the precompiled header file is always
11593 used, you can put a file of the same name as the original header in this
11594 directory containing an @code{#error} command.
11596 This also works with @option{-include}. So yet another way to use
11597 precompiled headers, good for projects not designed with precompiled
11598 header files in mind, is to simply take most of the header files used by
11599 a project, include them from another header file, precompile that header
11600 file, and @option{-include} the precompiled header. If the header files
11601 have guards against multiple inclusion, they will be skipped because
11602 they've already been included (in the precompiled header).
11604 If you need to precompile the same header file for different
11605 languages, targets, or compiler options, you can instead make a
11606 @emph{directory} named like @file{all.h.gch}, and put each precompiled
11607 header in the directory. (It doesn't matter what you call the files
11608 in the directory, every precompiled header in the directory will be
11609 considered.) The first precompiled header encountered in the
11610 directory that is valid for this compilation will be used; they're
11611 searched in no particular order.
11613 There are many other possibilities, limited only by your imagination,
11614 good sense, and the constraints of your build system.
11616 A precompiled header file can be used only when these conditions apply:
11620 Only one precompiled header can be used in a particular compilation.
11622 A precompiled header can't be used once the first C token is seen. You
11623 can have preprocessor directives before a precompiled header; you can
11624 even include a precompiled header from inside another header, so long as
11625 there are no C tokens before the @code{#include}.
11627 The precompiled header file must be produced for the same language as
11628 the current compilation. You can't use a C precompiled header for a C++
11631 The precompiled header file must be produced by the same compiler
11632 version and configuration as the current compilation is using.
11633 The easiest way to guarantee this is to use the same compiler binary
11634 for creating and using precompiled headers.
11636 Any macros defined before the precompiled header (including with
11637 @option{-D}) must either be defined in the same way as when the
11638 precompiled header was generated, or must not affect the precompiled
11639 header, which usually means that the they don't appear in the
11640 precompiled header at all.
11642 Certain command-line options must be defined in the same way as when the
11643 precompiled header was generated. At present, it's not clear which
11644 options are safe to change and which are not; the safest choice is to
11645 use exactly the same options when generating and using the precompiled
11649 For all of these but the last, the compiler will automatically ignore
11650 the precompiled header if the conditions aren't met. For the last item,
11651 some option changes will cause the precompiled header to be rejected,
11652 but not all incompatible option combinations have yet been found. If
11653 you find a new incompatible combination, please consider filing a bug
11654 report, see @ref{Bugs}.
11656 @node Running Protoize
11657 @section Running Protoize
11659 The program @code{protoize} is an optional part of GCC@. You can use
11660 it to add prototypes to a program, thus converting the program to ISO
11661 C in one respect. The companion program @code{unprotoize} does the
11662 reverse: it removes argument types from any prototypes that are found.
11664 When you run these programs, you must specify a set of source files as
11665 command line arguments. The conversion programs start out by compiling
11666 these files to see what functions they define. The information gathered
11667 about a file @var{foo} is saved in a file named @file{@var{foo}.X}.
11669 After scanning comes actual conversion. The specified files are all
11670 eligible to be converted; any files they include (whether sources or
11671 just headers) are eligible as well.
11673 But not all the eligible files are converted. By default,
11674 @code{protoize} and @code{unprotoize} convert only source and header
11675 files in the current directory. You can specify additional directories
11676 whose files should be converted with the @option{-d @var{directory}}
11677 option. You can also specify particular files to exclude with the
11678 @option{-x @var{file}} option. A file is converted if it is eligible, its
11679 directory name matches one of the specified directory names, and its
11680 name within the directory has not been excluded.
11682 Basic conversion with @code{protoize} consists of rewriting most
11683 function definitions and function declarations to specify the types of
11684 the arguments. The only ones not rewritten are those for varargs
11687 @code{protoize} optionally inserts prototype declarations at the
11688 beginning of the source file, to make them available for any calls that
11689 precede the function's definition. Or it can insert prototype
11690 declarations with block scope in the blocks where undeclared functions
11693 Basic conversion with @code{unprotoize} consists of rewriting most
11694 function declarations to remove any argument types, and rewriting
11695 function definitions to the old-style pre-ISO form.
11697 Both conversion programs print a warning for any function declaration or
11698 definition that they can't convert. You can suppress these warnings
11701 The output from @code{protoize} or @code{unprotoize} replaces the
11702 original source file. The original file is renamed to a name ending
11703 with @samp{.save} (for DOS, the saved filename ends in @samp{.sav}
11704 without the original @samp{.c} suffix). If the @samp{.save} (@samp{.sav}
11705 for DOS) file already exists, then the source file is simply discarded.
11707 @code{protoize} and @code{unprotoize} both depend on GCC itself to
11708 scan the program and collect information about the functions it uses.
11709 So neither of these programs will work until GCC is installed.
11711 Here is a table of the options you can use with @code{protoize} and
11712 @code{unprotoize}. Each option works with both programs unless
11716 @item -B @var{directory}
11717 Look for the file @file{SYSCALLS.c.X} in @var{directory}, instead of the
11718 usual directory (normally @file{/usr/local/lib}). This file contains
11719 prototype information about standard system functions. This option
11720 applies only to @code{protoize}.
11722 @item -c @var{compilation-options}
11723 Use @var{compilation-options} as the options when running @command{gcc} to
11724 produce the @samp{.X} files. The special option @option{-aux-info} is
11725 always passed in addition, to tell @command{gcc} to write a @samp{.X} file.
11727 Note that the compilation options must be given as a single argument to
11728 @code{protoize} or @code{unprotoize}. If you want to specify several
11729 @command{gcc} options, you must quote the entire set of compilation options
11730 to make them a single word in the shell.
11732 There are certain @command{gcc} arguments that you cannot use, because they
11733 would produce the wrong kind of output. These include @option{-g},
11734 @option{-O}, @option{-c}, @option{-S}, and @option{-o} If you include these in
11735 the @var{compilation-options}, they are ignored.
11738 Rename files to end in @samp{.C} (@samp{.cc} for DOS-based file
11739 systems) instead of @samp{.c}. This is convenient if you are converting
11740 a C program to C++. This option applies only to @code{protoize}.
11743 Add explicit global declarations. This means inserting explicit
11744 declarations at the beginning of each source file for each function
11745 that is called in the file and was not declared. These declarations
11746 precede the first function definition that contains a call to an
11747 undeclared function. This option applies only to @code{protoize}.
11749 @item -i @var{string}
11750 Indent old-style parameter declarations with the string @var{string}.
11751 This option applies only to @code{protoize}.
11753 @code{unprotoize} converts prototyped function definitions to old-style
11754 function definitions, where the arguments are declared between the
11755 argument list and the initial @samp{@{}. By default, @code{unprotoize}
11756 uses five spaces as the indentation. If you want to indent with just
11757 one space instead, use @option{-i " "}.
11760 Keep the @samp{.X} files. Normally, they are deleted after conversion
11764 Add explicit local declarations. @code{protoize} with @option{-l} inserts
11765 a prototype declaration for each function in each block which calls the
11766 function without any declaration. This option applies only to
11770 Make no real changes. This mode just prints information about the conversions
11771 that would have been done without @option{-n}.
11774 Make no @samp{.save} files. The original files are simply deleted.
11775 Use this option with caution.
11777 @item -p @var{program}
11778 Use the program @var{program} as the compiler. Normally, the name
11779 @file{gcc} is used.
11782 Work quietly. Most warnings are suppressed.
11785 Print the version number, just like @option{-v} for @command{gcc}.
11788 If you need special compiler options to compile one of your program's
11789 source files, then you should generate that file's @samp{.X} file
11790 specially, by running @command{gcc} on that source file with the
11791 appropriate options and the option @option{-aux-info}. Then run
11792 @code{protoize} on the entire set of files. @code{protoize} will use
11793 the existing @samp{.X} file because it is newer than the source file.
11797 gcc -Dfoo=bar file1.c -aux-info file1.X
11802 You need to include the special files along with the rest in the
11803 @code{protoize} command, even though their @samp{.X} files already
11804 exist, because otherwise they won't get converted.
11806 @xref{Protoize Caveats}, for more information on how to use
11807 @code{protoize} successfully.