1 @c Copyright (C) 1988,89,92,93,94,95,96,97,1998 Free Software Foundation, Inc.
2 @c This is part of the GCC manual.
3 @c For copying conditions, see the file gcc.texi.
6 @chapter GNU CC Command Options
7 @cindex GNU CC command options
8 @cindex command options
9 @cindex options, GNU CC command
11 When you invoke GNU CC, it normally does preprocessing, compilation,
12 assembly and linking. The ``overall options'' allow you to stop this
13 process at an intermediate stage. For example, the @samp{-c} option
14 says not to run the linker. Then the output consists of object files
15 output by the assembler.
17 Other options are passed on to one stage of processing. Some options
18 control the preprocessor and others the compiler itself. Yet other
19 options control the assembler and linker; most of these are not
20 documented here, since you rarely need to use any of them.
22 @cindex C compilation options
23 Most of the command line options that you can use with GNU CC are useful
24 for C programs; when an option is only useful with another language
25 (usually C++), the explanation says so explicitly. If the description
26 for a particular option does not mention a source language, you can use
27 that option with all supported languages.
29 @cindex C++ compilation options
30 @xref{Invoking G++,,Compiling C++ Programs}, for a summary of special
31 options for compiling C++ programs.
33 @cindex grouping options
34 @cindex options, grouping
35 The @code{gcc} program accepts options and file names as operands. Many
36 options have multiletter names; therefore multiple single-letter options
37 may @emph{not} be grouped: @samp{-dr} is very different from @w{@samp{-d
40 @cindex order of options
41 @cindex options, order
42 You can mix options and other arguments. For the most part, the order
43 you use doesn't matter. Order does matter when you use several options
44 of the same kind; for example, if you specify @samp{-L} more than once,
45 the directories are searched in the order specified.
47 Many options have long names starting with @samp{-f} or with
48 @samp{-W}---for example, @samp{-fforce-mem},
49 @samp{-fstrength-reduce}, @samp{-Wformat} and so on. Most of
50 these have both positive and negative forms; the negative form of
51 @samp{-ffoo} would be @samp{-fno-foo}. This manual documents
52 only one of these two forms, whichever one is not the default.
55 * Option Summary:: Brief list of all options, without explanations.
56 * Overall Options:: Controlling the kind of output:
57 an executable, object files, assembler files,
58 or preprocessed source.
59 * Invoking G++:: Compiling C++ programs.
60 * C Dialect Options:: Controlling the variant of C language compiled.
61 * C++ Dialect Options:: Variations on C++.
62 * Warning Options:: How picky should the compiler be?
63 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
64 * Optimize Options:: How much optimization?
65 * Preprocessor Options:: Controlling header files and macro definitions.
66 Also, getting dependency information for Make.
67 * Assembler Options:: Passing options to the assembler.
68 * Link Options:: Specifying libraries and so on.
69 * Directory Options:: Where to find header files and libraries.
70 Where to find the compiler executable files.
71 * Target Options:: Running a cross-compiler, or an old version of GNU CC.
72 * Submodel Options:: Specifying minor hardware or convention variations,
73 such as 68010 vs 68020.
74 * Code Gen Options:: Specifying conventions for function calls, data layout
76 * Environment Variables:: Env vars that affect GNU CC.
77 * Running Protoize:: Automatically adding or removing function prototypes.
81 @section Option Summary
83 Here is a summary of all the options, grouped by type. Explanations are
84 in the following sections.
88 @xref{Overall Options,,Options Controlling the Kind of Output}.
90 -c -S -E -o @var{file} -pipe -v --help -x @var{language}
93 @item C Language Options
94 @xref{C Dialect Options,,Options Controlling C Dialect}.
96 -ansi -flang-isoc9x -fallow-single-precision -fcond-mismatch -fno-asm
97 -fno-builtin -ffreestanding -fhosted -fsigned-bitfields -fsigned-char
98 -funsigned-bitfields -funsigned-char -fwritable-strings
99 -traditional -traditional-cpp -trigraphs
102 @item C++ Language Options
103 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}.
105 -fno-access-control -fcheck-new -fconserve-space -fdollars-in-identifiers
106 -fno-elide-constructors -fexternal-templates -ffor-scope
107 -fno-for-scope -fno-gnu-keywords -fguiding-decls -fhandle-signatures
108 -fhonor-std -fhuge-objects -fno-implicit-templates -finit-priority
109 -fno-implement-inlines -fname-mangling-version-@var{n} -fno-default-inline
110 -foperator-names -fno-optional-diags -frepo -fstrict-prototype
111 -fsquangle -ftemplate-depth-@var{n} -fthis-is-variable -fvtable-thunks
115 @item Warning Options
116 @xref{Warning Options,,Options to Request or Suppress Warnings}.
118 -fsyntax-only -pedantic -pedantic-errors
119 -w -W -Wall -Waggregate-return -Wbad-function-cast
120 -Wcast-align -Wcast-qual -Wchar-subscript -Wcomment
121 -Wconversion -Werror -Wformat
122 -Wid-clash-@var{len} -Wimplicit -Wimplicit-int
123 -Wimplicit-function-declaration -Wimport
124 -Werror-implicit-function-declaration -Winline
125 -Wlarger-than-@var{len} -Wlong-long
126 -Wmain -Wmissing-declarations -Wmissing-noreturn
127 -Wmissing-prototypes -Wmultichar -Wnested-externs -Wno-import
128 -Wno-non-template-friend -Wold-style-cast -Woverloaded-virtual
129 -Wparentheses -Wpointer-arith -Wredundant-decls -Wreorder
130 -Wreturn-type -Wshadow -Wsign-compare -Wstrict-prototypes
131 -Wswitch -Wsynth -Wtraditional
132 -Wtrigraphs -Wundef -Wuninitialized -Wunused -Wwrite-strings
136 @item Debugging Options
137 @xref{Debugging Options,,Options for Debugging Your Program or GCC}.
139 -a -ax -d@var{letters} -fdump-unnumbered -fpretend-float
140 -fprofile-arcs -ftest-coverage
141 -g -g@var{level} -gcoff -gdwarf -gdwarf-1 -gdwarf-1+ -gdwarf-2
142 -ggdb -gstabs -gstabs+ -gxcoff -gxcoff+
143 -p -pg -print-file-name=@var{library} -print-libgcc-file-name
144 -print-prog-name=@var{program} -print-search-dirs -save-temps
147 @item Optimization Options
148 @xref{Optimize Options,,Options that Control Optimization}.
150 -fbranch-probabilities -foptimize-register-moves
151 -fcaller-saves -fcse-follow-jumps -fcse-skip-blocks
152 -fdelayed-branch -fexpensive-optimizations
153 -ffast-math -ffloat-store -fforce-addr -fforce-mem
154 -fdata-sections -ffunction-sections -fgcse
155 -finline-functions -fkeep-inline-functions
156 -fno-default-inline -fno-defer-pop -fno-function-cse
157 -fno-inline -fno-peephole -fomit-frame-pointer -fregmove
158 -frerun-cse-after-loop -frerun-loop-opt -fschedule-insns
159 -fschedule-insns2 -fstrength-reduce -fthread-jumps
160 -funroll-all-loops -funroll-loops
161 -fmove-all-movables -freduce-all-givs -fstrict-aliasing
162 -O -O0 -O1 -O2 -O3 -Os
165 @item Preprocessor Options
166 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
168 -A@var{question}(@var{answer}) -C -dD -dM -dN
169 -D@var{macro}@r{[}=@var{defn}@r{]} -E -H
171 -include @var{file} -imacros @var{file}
172 -iprefix @var{file} -iwithprefix @var{dir}
173 -iwithprefixbefore @var{dir} -isystem @var{dir}
174 -M -MD -MM -MMD -MG -nostdinc -P -trigraphs
175 -undef -U@var{macro} -Wp,@var{option}
178 @item Assembler Option
179 @xref{Assembler Options,,Passing Options to the Assembler}.
185 @xref{Link Options,,Options for Linking}.
187 @var{object-file-name} -l@var{library}
188 -nostartfiles -nodefaultlibs -nostdlib
189 -s -static -shared -symbolic
190 -Wl,@var{option} -Xlinker @var{option}
194 @item Directory Options
195 @xref{Directory Options,,Options for Directory Search}.
197 -B@var{prefix} -I@var{dir} -I- -L@var{dir} -specs=@var{file}
201 @c I wrote this xref this way to avoid overfull hbox. -- rms
202 @xref{Target Options}.
204 -b @var{machine} -V @var{version}
207 @item Machine Dependent Options
208 @xref{Submodel Options,,Hardware Models and Configurations}.
210 @emph{M680x0 Options}
211 -m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040
212 -m68060 -mcpu32 -m5200 -m68881 -mbitfield -mc68000 -mc68020
213 -mfpa -mnobitfield -mrtd -mshort -msoft-float
221 -mtune=@var{cpu type}
222 -mcmodel=@var{code model}
223 -malign-jumps=@var{num} -malign-loops=@var{num}
224 -malign-functions=@var{num}
226 -mapp-regs -mbroken-saverestore -mcypress -mepilogue
227 -mflat -mfpu -mhard-float -mhard-quad-float
228 -mimpure-text -mlive-g0 -mno-app-regs -mno-epilogue
229 -mno-flat -mno-fpu -mno-impure-text
230 -mno-stack-bias -mno-unaligned-doubles
231 -msoft-float -msoft-quad-float -msparclite -mstack-bias
232 -msupersparc -munaligned-doubles -mv8
234 @emph{Convex Options}
235 -mc1 -mc2 -mc32 -mc34 -mc38
236 -margcount -mnoargcount
238 -mvolatile-cache -mvolatile-nocache
240 @emph{AMD29K Options}
241 -m29000 -m29050 -mbw -mnbw -mdw -mndw
242 -mlarge -mnormal -msmall
243 -mkernel-registers -mno-reuse-arg-regs
244 -mno-stack-check -mno-storem-bug
245 -mreuse-arg-regs -msoft-float -mstack-check
246 -mstorem-bug -muser-registers
249 -mapcs-frame -mno-apcs-frame
251 -mapcs-stack-check -mno-apcs-stack-check
252 -mapcs-float -mno-apcs-float
253 -mapcs-reentrant -mno-apcs-reentrant
254 -msched-prolog -mno-sched-prolog
255 -mlittle-endian -mbig-endian -mwords-little-endian
256 -mshort-load-bytes -mno-short-load-bytes -mshort-load-words -mno-short-load-words
257 -msoft-float -mhard-float -mfpe
258 -mthumb-interwork -mno-thumb-interwork
259 -mcpu= -march= -mfpe=
260 -mstructure-size-boundary=
261 -mbsd -mxopen -mno-symrename
264 -mtpcs-frame -mno-tpcs-frame
265 -mtpcs-leaf-frame -mno-tpcs-leaf-frame
266 -mlittle-endian -mbig-endian
267 -mthumb-interwork -mno-thumb-interwork
268 -mstructure-size-boundary=
270 @emph{MN10200 Options}
273 @emph{MN10300 Options}
278 @emph{M32R/D Options}
279 -mcode-model=@var{model type} -msdata=@var{sdata type}
283 -m88000 -m88100 -m88110 -mbig-pic
284 -mcheck-zero-division -mhandle-large-shift
285 -midentify-revision -mno-check-zero-division
286 -mno-ocs-debug-info -mno-ocs-frame-position
287 -mno-optimize-arg-area -mno-serialize-volatile
288 -mno-underscores -mocs-debug-info
289 -mocs-frame-position -moptimize-arg-area
290 -mserialize-volatile -mshort-data-@var{num} -msvr3
291 -msvr4 -mtrap-large-shift -muse-div-instruction
292 -mversion-03.00 -mwarn-passed-structs
294 @emph{RS/6000 and PowerPC Options}
296 -mtune=@var{cpu type}
297 -mpower -mno-power -mpower2 -mno-power2
298 -mpowerpc -mno-powerpc
299 -mpowerpc-gpopt -mno-powerpc-gpopt
300 -mpowerpc-gfxopt -mno-powerpc-gfxopt
301 -mnew-mnemonics -mno-new-mnemonics
302 -mfull-toc -mminimal-toc -mno-fop-in-toc -mno-sum-in-toc
303 -maix64 -maix32 -mxl-call -mno-xl-call -mthreads -mpe
304 -msoft-float -mhard-float -mmultiple -mno-multiple
305 -mstring -mno-string -mupdate -mno-update
306 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align
307 -mstrict-align -mno-strict-align -mrelocatable
308 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib
309 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian
310 -mcall-aix -mcall-sysv -mprototype -mno-prototype
311 -msim -mmvme -mads -myellowknife -memb -msdata
312 -msdata=@var{opt} -G @var{num}
315 -mcall-lib-mul -mfp-arg-in-fpregs -mfp-arg-in-gregs
316 -mfull-fp-blocks -mhc-struct-return -min-line-mul
317 -mminimum-fp-blocks -mnohc-struct-return
320 -mabicalls -mcpu=@var{cpu type} -membedded-data
321 -membedded-pic -mfp32 -mfp64 -mgas -mgp32 -mgp64
322 -mgpopt -mhalf-pic -mhard-float -mint64 -mips1
323 -mips2 -mips3 -mips4 -mlong64 -mlong-calls -mmemcpy
324 -mmips-as -mmips-tfile -mno-abicalls
325 -mno-embedded-data -mno-embedded-pic
326 -mno-gpopt -mno-long-calls
327 -mno-memcpy -mno-mips-tfile -mno-rnames -mno-stats
328 -mrnames -msoft-float
329 -m4650 -msingle-float -mmad
330 -mstats -EL -EB -G @var{num} -nocpp
331 -mabi=32 -mabi=n32 -mabi=64 -mabi=eabi
335 -march=@var{cpu type}
336 -mieee-fp -mno-fancy-math-387
337 -mno-fp-ret-in-387 -msoft-float -msvr3-shlib
338 -mno-wide-multiply -mrtd -malign-double
339 -mreg-alloc=@var{list} -mregparm=@var{num}
340 -malign-jumps=@var{num} -malign-loops=@var{num}
341 -malign-functions=@var{num}
344 -mbig-switch -mdisable-fpregs -mdisable-indexing
345 -mfast-indirect-calls -mgas -mjump-in-delay
346 -mlong-load-store -mno-big-switch -mno-disable-fpregs
347 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas
348 -mno-jump-in-delay -mno-long-load-store
349 -mno-portable-runtime -mno-soft-float -mno-space
350 -mno-space-regs -msoft-float -mpa-risc-1-0
351 -mpa-risc-1-1 -mportable-runtime
352 -mschedule=@var{list} -mspace -mspace-regs
354 @emph{Intel 960 Options}
355 -m@var{cpu type} -masm-compat -mclean-linkage
356 -mcode-align -mcomplex-addr -mleaf-procedures
357 -mic-compat -mic2.0-compat -mic3.0-compat
358 -mintel-asm -mno-clean-linkage -mno-code-align
359 -mno-complex-addr -mno-leaf-procedures
360 -mno-old-align -mno-strict-align -mno-tail-call
361 -mnumerics -mold-align -msoft-float -mstrict-align
364 @emph{DEC Alpha Options}
365 -mfp-regs -mno-fp-regs -mno-soft-float -msoft-float
367 -mieee -mieee-with-inexact -mieee-conformant
368 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode}
369 -mtrap-precision=@var{mode} -mbuild-constants
371 -mbwx -mno-bwx -mcix -mno-cix -mmax -mno-max
372 -mmemory-latency=@var{time}
374 @emph{Clipper Options}
377 @emph{H8/300 Options}
378 -mrelax -mh -ms -mint32 -malign-300
381 -m1 -m2 -m3 -m3e -mb -ml -mdalign -mrelax
383 @emph{System V Options}
384 -Qy -Qn -YP,@var{paths} -Ym,@var{dir}
388 -mmangle-cpu -mcpu=@var{cpu} -mtext=@var{text section}
389 -mdata=@var{data section} -mrodata=@var{readonly data section}
392 -mlong-calls -mno-long-calls -mep -mno-ep
393 -mprolog-function -mno-prolog-function -mspace
394 -mtda=@var{n} -msda=@var{n} -mzda=@var{n}
398 @item Code Generation Options
399 @xref{Code Gen Options,,Options for Code Generation Conventions}.
401 -fcall-saved-@var{reg} -fcall-used-@var{reg}
402 -fexceptions -ffixed-@var{reg} -finhibit-size-directive
403 -fcheck-memory-usage -fprefix-function-name
404 -fno-common -fno-ident -fno-gnu-linker
405 -fpcc-struct-return -fpic -fPIC
406 -freg-struct-return -fshared-data -fshort-enums
407 -fshort-double -fvolatile -fvolatile-global
408 -fverbose-asm -fpack-struct -fstack-check
409 -fargument-alias -fargument-noalias
410 -fargument-noalias-global
415 * Overall Options:: Controlling the kind of output:
416 an executable, object files, assembler files,
417 or preprocessed source.
418 * C Dialect Options:: Controlling the variant of C language compiled.
419 * C++ Dialect Options:: Variations on C++.
420 * Warning Options:: How picky should the compiler be?
421 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
422 * Optimize Options:: How much optimization?
423 * Preprocessor Options:: Controlling header files and macro definitions.
424 Also, getting dependency information for Make.
425 * Assembler Options:: Passing options to the assembler.
426 * Link Options:: Specifying libraries and so on.
427 * Directory Options:: Where to find header files and libraries.
428 Where to find the compiler executable files.
429 * Target Options:: Running a cross-compiler, or an old version of GNU CC.
432 @node Overall Options
433 @section Options Controlling the Kind of Output
435 Compilation can involve up to four stages: preprocessing, compilation
436 proper, assembly and linking, always in that order. The first three
437 stages apply to an individual source file, and end by producing an
438 object file; linking combines all the object files (those newly
439 compiled, and those specified as input) into an executable file.
441 @cindex file name suffix
442 For any given input file, the file name suffix determines what kind of
447 C source code which must be preprocessed.
450 C source code which should not be preprocessed.
453 C++ source code which should not be preprocessed.
456 Objective-C source code. Note that you must link with the library
457 @file{libobjc.a} to make an Objective-C program work.
460 C header file (not to be compiled or linked).
463 @itemx @var{file}.cxx
464 @itemx @var{file}.cpp
466 C++ source code which must be preprocessed. Note that in @samp{.cxx},
467 the last two letters must both be literally @samp{x}. Likewise,
468 @samp{.C} refers to a literal capital C.
474 Assembler code which must be preprocessed.
477 An object file to be fed straight into linking.
478 Any file name with no recognized suffix is treated this way.
481 You can specify the input language explicitly with the @samp{-x} option:
484 @item -x @var{language}
485 Specify explicitly the @var{language} for the following input files
486 (rather than letting the compiler choose a default based on the file
487 name suffix). This option applies to all following input files until
488 the next @samp{-x} option. Possible values for @var{language} are:
491 c-header cpp-output c++-cpp-output
492 assembler assembler-with-cpp
496 Turn off any specification of a language, so that subsequent files are
497 handled according to their file name suffixes (as they are if @samp{-x}
498 has not been used at all).
501 If you only want some of the stages of compilation, you can use
502 @samp{-x} (or filename suffixes) to tell @code{gcc} where to start, and
503 one of the options @samp{-c}, @samp{-S}, or @samp{-E} to say where
504 @code{gcc} is to stop. Note that some combinations (for example,
505 @samp{-x cpp-output -E} instruct @code{gcc} to do nothing at all.
509 Compile or assemble the source files, but do not link. The linking
510 stage simply is not done. The ultimate output is in the form of an
511 object file for each source file.
513 By default, the object file name for a source file is made by replacing
514 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
516 Unrecognized input files, not requiring compilation or assembly, are
520 Stop after the stage of compilation proper; do not assemble. The output
521 is in the form of an assembler code file for each non-assembler input
524 By default, the assembler file name for a source file is made by
525 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
527 Input files that don't require compilation are ignored.
530 Stop after the preprocessing stage; do not run the compiler proper. The
531 output is in the form of preprocessed source code, which is sent to the
534 Input files which don't require preprocessing are ignored.
536 @cindex output file option
538 Place output in file @var{file}. This applies regardless to whatever
539 sort of output is being produced, whether it be an executable file,
540 an object file, an assembler file or preprocessed C code.
542 Since only one output file can be specified, it does not make sense to
543 use @samp{-o} when compiling more than one input file, unless you are
544 producing an executable file as output.
546 If @samp{-o} is not specified, the default is to put an executable file
547 in @file{a.out}, the object file for @file{@var{source}.@var{suffix}} in
548 @file{@var{source}.o}, its assembler file in @file{@var{source}.s}, and
549 all preprocessed C source on standard output.@refill
552 Print (on standard error output) the commands executed to run the stages
553 of compilation. Also print the version number of the compiler driver
554 program and of the preprocessor and the compiler proper.
557 Use pipes rather than temporary files for communication between the
558 various stages of compilation. This fails to work on some systems where
559 the assembler is unable to read from a pipe; but the GNU assembler has
563 Print (on the standard output) a description of the command line options
564 understood by @code{gcc}. If the @code{-v} option is also specified
565 then @code{--help} will also be passed on to the various processes
566 invoked by @code{gcc}, so that they can display the command line options
567 they accept. If the @code{-W} option is also specified then command
568 line options which have no documentation associated with them will also
573 @section Compiling C++ Programs
575 @cindex suffixes for C++ source
576 @cindex C++ source file suffixes
577 C++ source files conventionally use one of the suffixes @samp{.C},
578 @samp{.cc}, @samp{cpp}, or @samp{.cxx}; preprocessed C++ files use the
579 suffix @samp{.ii}. GNU CC recognizes files with these names and
580 compiles them as C++ programs even if you call the compiler the same way
581 as for compiling C programs (usually with the name @code{gcc}).
585 However, C++ programs often require class libraries as well as a
586 compiler that understands the C++ language---and under some
587 circumstances, you might want to compile programs from standard input,
588 or otherwise without a suffix that flags them as C++ programs.
589 @code{g++} is a program that calls GNU CC with the default language
590 set to C++, and automatically specifies linking against the C++
592 @cindex @code{g++ 1.@var{xx}}
593 @cindex @code{g++}, separate compiler
594 @cindex @code{g++} older version
595 @footnote{Prior to release 2 of the compiler,
596 there was a separate @code{g++} compiler. That version was based on GNU
597 CC, but not integrated with it. Versions of @code{g++} with a
598 @samp{1.@var{xx}} version number---for example, @code{g++} version 1.37
599 or 1.42---are much less reliable than the versions integrated with GCC
600 2. Moreover, combining G++ @samp{1.@var{xx}} with a version 2 GCC will
601 simply not work.} On many systems, the script @code{g++} is also
602 installed with the name @code{c++}.
604 @cindex invoking @code{g++}
605 When you compile C++ programs, you may specify many of the same
606 command-line options that you use for compiling programs in any
607 language; or command-line options meaningful for C and related
608 languages; or options that are meaningful only for C++ programs.
609 @xref{C Dialect Options,,Options Controlling C Dialect}, for
610 explanations of options for languages related to C.
611 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
612 explanations of options that are meaningful only for C++ programs.
614 @node C Dialect Options
615 @section Options Controlling C Dialect
616 @cindex dialect options
617 @cindex language dialect options
618 @cindex options, dialect
620 The following options control the dialect of C (or languages derived
621 from C, such as C++ and Objective C) that the compiler accepts:
626 Support all ANSI standard C programs.
628 This turns off certain features of GNU C that are incompatible with ANSI
629 C, such as the @code{asm}, @code{inline} and @code{typeof} keywords, and
630 predefined macros such as @code{unix} and @code{vax} that identify the
631 type of system you are using. It also enables the undesirable and
632 rarely used ANSI trigraph feature, and it disables recognition of C++
633 style @samp{//} comments.
635 The alternate keywords @code{__asm__}, @code{__extension__},
636 @code{__inline__} and @code{__typeof__} continue to work despite
637 @samp{-ansi}. You would not want to use them in an ANSI C program, of
638 course, but it is useful to put them in header files that might be included
639 in compilations done with @samp{-ansi}. Alternate predefined macros
640 such as @code{__unix__} and @code{__vax__} are also available, with or
641 without @samp{-ansi}.
643 The @samp{-ansi} option does not cause non-ANSI programs to be
644 rejected gratuitously. For that, @samp{-pedantic} is required in
645 addition to @samp{-ansi}. @xref{Warning Options}.
647 The macro @code{__STRICT_ANSI__} is predefined when the @samp{-ansi}
648 option is used. Some header files may notice this macro and refrain
649 from declaring certain functions or defining certain macros that the
650 ANSI standard doesn't call for; this is to avoid interfering with any
651 programs that might use these names for other things.
653 The functions @code{alloca}, @code{abort}, @code{exit}, and
654 @code{_exit} are not builtin functions when @samp{-ansi} is used.
657 Enable support for features found in the C9X standard. In particular,
658 enable support for the C9X @code{restrict} keyword.
660 Even when this option is not specified, you can still use some C9X
661 features in so far as they do not conflict with previous C standards.
662 For example, you may use @code{__restrict__} even when -flang-isoc9x
666 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
667 keyword, so that code can use these words as identifiers. You can use
668 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
669 instead. @samp{-ansi} implies @samp{-fno-asm}.
671 In C++, this switch only affects the @code{typeof} keyword, since
672 @code{asm} and @code{inline} are standard keywords. You may want to
673 use the @samp{-fno-gnu-keywords} flag instead, as it also disables the
674 other, C++-specific, extension keywords such as @code{headof}.
677 @cindex builtin functions
693 Don't recognize builtin functions that do not begin with `__builtin_'
694 as prefix. Currently, the functions affected include @code{abort},
695 @code{abs}, @code{alloca}, @code{cos}, @code{exit}, @code{fabs},
696 @code{ffs}, @code{labs}, @code{memcmp}, @code{memcpy}, @code{sin},
697 @code{sqrt}, @code{strcmp}, @code{strcpy}, and @code{strlen}.
699 GCC normally generates special code to handle certain builtin functions
700 more efficiently; for instance, calls to @code{alloca} may become single
701 instructions that adjust the stack directly, and calls to @code{memcpy}
702 may become inline copy loops. The resulting code is often both smaller
703 and faster, but since the function calls no longer appear as such, you
704 cannot set a breakpoint on those calls, nor can you change the behavior
705 of the functions by linking with a different library.
707 The @samp{-ansi} option prevents @code{alloca} and @code{ffs} from being
708 builtin functions, since these functions do not have an ANSI standard
712 @cindex hosted environment
714 Assert that compilation takes place in a hosted environment. This implies
715 @samp{-fbuiltin}. A hosted environment is one in which the
716 entire standard library is available, and in which @code{main} has a return
717 type of @code{int}. Examples are nearly everything except a kernel.
718 This is equivalent to @samp{-fno-freestanding}.
721 @cindex hosted environment
723 Assert that compilation takes place in a freestanding environment. This
724 implies @samp{-fno-builtin}. A freestanding environment
725 is one in which the standard library may not exist, and program startup may
726 not necessarily be at @code{main}. The most obvious example is an OS kernel.
727 This is equivalent to @samp{-fno-hosted}.
730 Support ANSI C trigraphs. You don't want to know about this
731 brain-damage. The @samp{-ansi} option implies @samp{-trigraphs}.
733 @cindex traditional C language
734 @cindex C language, traditional
736 Attempt to support some aspects of traditional C compilers.
741 All @code{extern} declarations take effect globally even if they
742 are written inside of a function definition. This includes implicit
743 declarations of functions.
746 The newer keywords @code{typeof}, @code{inline}, @code{signed}, @code{const}
747 and @code{volatile} are not recognized. (You can still use the
748 alternative keywords such as @code{__typeof__}, @code{__inline__}, and
752 Comparisons between pointers and integers are always allowed.
755 Integer types @code{unsigned short} and @code{unsigned char} promote
756 to @code{unsigned int}.
759 Out-of-range floating point literals are not an error.
762 Certain constructs which ANSI regards as a single invalid preprocessing
763 number, such as @samp{0xe-0xd}, are treated as expressions instead.
766 String ``constants'' are not necessarily constant; they are stored in
767 writable space, and identical looking constants are allocated
768 separately. (This is the same as the effect of
769 @samp{-fwritable-strings}.)
771 @cindex @code{longjmp} and automatic variables
773 All automatic variables not declared @code{register} are preserved by
774 @code{longjmp}. Ordinarily, GNU C follows ANSI C: automatic variables
775 not declared @code{volatile} may be clobbered.
780 @cindex escape sequences, traditional
781 The character escape sequences @samp{\x} and @samp{\a} evaluate as the
782 literal characters @samp{x} and @samp{a} respectively. Without
783 @w{@samp{-traditional}}, @samp{\x} is a prefix for the hexadecimal
784 representation of a character, and @samp{\a} produces a bell.
787 You may wish to use @samp{-fno-builtin} as well as @samp{-traditional}
788 if your program uses names that are normally GNU C builtin functions for
789 other purposes of its own.
791 You cannot use @samp{-traditional} if you include any header files that
792 rely on ANSI C features. Some vendors are starting to ship systems with
793 ANSI C header files and you cannot use @samp{-traditional} on such
794 systems to compile files that include any system headers.
796 The @samp{-traditional} option also enables @samp{-traditional-cpp},
797 which is described next.
799 @item -traditional-cpp
800 Attempt to support some aspects of traditional C preprocessors.
805 Comments convert to nothing at all, rather than to a space. This allows
806 traditional token concatenation.
809 In a preprocessing directive, the @samp{#} symbol must appear as the first
813 Macro arguments are recognized within string constants in a macro
814 definition (and their values are stringified, though without additional
815 quote marks, when they appear in such a context). The preprocessor
816 always considers a string constant to end at a newline.
819 @cindex detecting @w{@samp{-traditional}}
820 The predefined macro @code{__STDC__} is not defined when you use
821 @samp{-traditional}, but @code{__GNUC__} is (since the GNU extensions
822 which @code{__GNUC__} indicates are not affected by
823 @samp{-traditional}). If you need to write header files that work
824 differently depending on whether @samp{-traditional} is in use, by
825 testing both of these predefined macros you can distinguish four
826 situations: GNU C, traditional GNU C, other ANSI C compilers, and other
827 old C compilers. The predefined macro @code{__STDC_VERSION__} is also
828 not defined when you use @samp{-traditional}. @xref{Standard
829 Predefined,,Standard Predefined Macros,cpp.info,The C Preprocessor},
830 for more discussion of these and other predefined macros.
833 @cindex string constants vs newline
834 @cindex newline vs string constants
835 The preprocessor considers a string constant to end at a newline (unless
836 the newline is escaped with @samp{\}). (Without @w{@samp{-traditional}},
837 string constants can contain the newline character as typed.)
840 @item -fcond-mismatch
841 Allow conditional expressions with mismatched types in the second and
842 third arguments. The value of such an expression is void.
844 @item -funsigned-char
845 Let the type @code{char} be unsigned, like @code{unsigned char}.
847 Each kind of machine has a default for what @code{char} should
848 be. It is either like @code{unsigned char} by default or like
849 @code{signed char} by default.
851 Ideally, a portable program should always use @code{signed char} or
852 @code{unsigned char} when it depends on the signedness of an object.
853 But many programs have been written to use plain @code{char} and
854 expect it to be signed, or expect it to be unsigned, depending on the
855 machines they were written for. This option, and its inverse, let you
856 make such a program work with the opposite default.
858 The type @code{char} is always a distinct type from each of
859 @code{signed char} or @code{unsigned char}, even though its behavior
860 is always just like one of those two.
863 Let the type @code{char} be signed, like @code{signed char}.
865 Note that this is equivalent to @samp{-fno-unsigned-char}, which is
866 the negative form of @samp{-funsigned-char}. Likewise, the option
867 @samp{-fno-signed-char} is equivalent to @samp{-funsigned-char}.
869 You may wish to use @samp{-fno-builtin} as well as @samp{-traditional}
870 if your program uses names that are normally GNU C builtin functions for
871 other purposes of its own.
873 You cannot use @samp{-traditional} if you include any header files that
874 rely on ANSI C features. Some vendors are starting to ship systems with
875 ANSI C header files and you cannot use @samp{-traditional} on such
876 systems to compile files that include any system headers.
878 @item -fsigned-bitfields
879 @itemx -funsigned-bitfields
880 @itemx -fno-signed-bitfields
881 @itemx -fno-unsigned-bitfields
882 These options control whether a bitfield is signed or unsigned, when the
883 declaration does not use either @code{signed} or @code{unsigned}. By
884 default, such a bitfield is signed, because this is consistent: the
885 basic integer types such as @code{int} are signed types.
887 However, when @samp{-traditional} is used, bitfields are all unsigned
890 @item -fwritable-strings
891 Store string constants in the writable data segment and don't uniquize
892 them. This is for compatibility with old programs which assume they can
893 write into string constants. The option @samp{-traditional} also has
896 Writing into string constants is a very bad idea; ``constants'' should
899 @item -fallow-single-precision
900 Do not promote single precision math operations to double precision,
901 even when compiling with @samp{-traditional}.
903 Traditional K&R C promotes all floating point operations to double
904 precision, regardless of the sizes of the operands. On the
905 architecture for which you are compiling, single precision may be faster
906 than double precision. If you must use @samp{-traditional}, but want
907 to use single precision operations when the operands are single
908 precision, use this option. This option has no effect when compiling
909 with ANSI or GNU C conventions (the default).
913 @node C++ Dialect Options
914 @section Options Controlling C++ Dialect
916 @cindex compiler options, C++
917 @cindex C++ options, command line
919 This section describes the command-line options that are only meaningful
920 for C++ programs; but you can also use most of the GNU compiler options
921 regardless of what language your program is in. For example, you
922 might compile a file @code{firstClass.C} like this:
925 g++ -g -frepo -O -c firstClass.C
929 In this example, only @samp{-frepo} is an option meant
930 only for C++ programs; you can use the other options with any
931 language supported by GNU CC.
933 Here is a list of options that are @emph{only} for compiling C++ programs:
936 @item -fno-access-control
937 Turn off all access checking. This switch is mainly useful for working
938 around bugs in the access control code.
941 Check that the pointer returned by @code{operator new} is non-null
942 before attempting to modify the storage allocated. The current Working
943 Paper requires that @code{operator new} never return a null pointer, so
944 this check is normally unnecessary.
946 An alternative to using this option is to specify that your
947 @code{operator new} does not throw any exceptions; if you declare it
948 @samp{throw()}, g++ will check the return value. See also @samp{new
951 @item -fconserve-space
952 Put uninitialized or runtime-initialized global variables into the
953 common segment, as C does. This saves space in the executable at the
954 cost of not diagnosing duplicate definitions. If you compile with this
955 flag and your program mysteriously crashes after @code{main()} has
956 completed, you may have an object that is being destroyed twice because
957 two definitions were merged.
959 This option is no longer useful on most targets, now that support has
960 been added for putting variables into BSS without making them common.
962 @item -fdollars-in-identifiers
963 Accept @samp{$} in identifiers. You can also explicitly prohibit use of
964 @samp{$} with the option @samp{-fno-dollars-in-identifiers}. (GNU C allows
965 @samp{$} by default on most target systems, but there are a few exceptions.)
966 Traditional C allowed the character @samp{$} to form part of
967 identifiers. However, ANSI C and C++ forbid @samp{$} in identifiers.
969 @item -fno-elide-constructors
970 The C++ standard allows an implementation to omit creating a temporary
971 which is only used to initialize another object of the same type.
972 Specifying this option disables that optimization, and forces g++ to
973 call the copy constructor in all cases.
975 @item -fexternal-templates
976 Cause template instantiations to obey @samp{#pragma interface} and
977 @samp{implementation}; template instances are emitted or not according
978 to the location of the template definition. @xref{Template
979 Instantiation}, for more information.
981 This option is deprecated.
983 @item -falt-external-templates
984 Similar to -fexternal-templates, but template instances are emitted or
985 not according to the place where they are first instantiated.
986 @xref{Template Instantiation}, for more information.
988 This option is deprecated.
991 @itemx -fno-for-scope
992 If -ffor-scope is specified, the scope of variables declared in
993 a @i{for-init-statement} is limited to the @samp{for} loop itself,
994 as specified by the draft C++ standard.
995 If -fno-for-scope is specified, the scope of variables declared in
996 a @i{for-init-statement} extends to the end of the enclosing scope,
997 as was the case in old versions of gcc, and other (traditional)
998 implementations of C++.
1000 The default if neither flag is given to follow the standard,
1001 but to allow and give a warning for old-style code that would
1002 otherwise be invalid, or have different behavior.
1004 @item -fno-gnu-keywords
1005 Do not recognize @code{classof}, @code{headof}, @code{signature},
1006 @code{sigof} or @code{typeof} as a keyword, so that code can use these
1007 words as identifiers. You can use the keywords @code{__classof__},
1008 @code{__headof__}, @code{__signature__}, @code{__sigof__}, and
1009 @code{__typeof__} instead. @samp{-ansi} implies
1010 @samp{-fno-gnu-keywords}.
1012 @item -fguiding-decls
1013 Treat a function declaration with the same type as a potential function
1014 template instantiation as though it declares that instantiation, not a
1015 normal function. If a definition is given for the function later in the
1016 translation unit (or another translation unit if the target supports
1017 weak symbols), that definition will be used; otherwise the template will
1018 be instantiated. This behavior reflects the C++ language prior to
1019 September 1996, when guiding declarations were removed.
1021 This option implies @samp{-fname-mangling-version-0}, and will not work
1022 with other name mangling versions. Like all options that change the
1023 ABI, all C++ code, @emph{including libgcc.a} must be built with the same
1024 setting of this option.
1026 @item -fhandle-signatures
1027 Recognize the @code{signature} and @code{sigof} keywords for specifying
1028 abstract types. The default (@samp{-fno-handle-signatures}) is not to
1029 recognize them. @xref{C++ Signatures, Type Abstraction using
1033 Treat the @code{namespace std} as a namespace, instead of ignoring
1034 it. For compatibility with earlier versions of g++, the compiler will,
1035 by default, ignore @code{namespace-declarations},
1036 @code{using-declarations}, @code{using-directives}, and
1037 @code{namespace-names}, if they involve @code{std}.
1039 @item -fhuge-objects
1040 Support virtual function calls for objects that exceed the size
1041 representable by a @samp{short int}. Users should not use this flag by
1042 default; if you need to use it, the compiler will tell you so.
1044 This flag is not useful when compiling with -fvtable-thunks.
1046 Like all options that change the ABI, all C++ code, @emph{including
1047 libgcc} must be built with the same setting of this option.
1049 @item -fno-implicit-templates
1050 Never emit code for templates which are instantiated implicitly (i.e. by
1051 use); only emit code for explicit instantiations. @xref{Template
1052 Instantiation}, for more information.
1054 @item -finit-priority
1055 Support @samp{__attribute__ ((init_priority (n)))} for controlling the
1056 order of initialization of file-scope objects. On ELF targets, this
1057 requires GNU ld 2.10 or later.
1059 @item -fno-implement-inlines
1060 To save space, do not emit out-of-line copies of inline functions
1061 controlled by @samp{#pragma implementation}. This will cause linker
1062 errors if these functions are not inlined everywhere they are called.
1064 @item -fname-mangling-version-@var{n}
1065 Control the way in which names are mangled. Version 0 is compatible
1066 with versions of g++ before 2.8. Version 1 is the default. Version 1
1067 will allow correct mangling of function templates. For example,
1068 version 0 mangling does not mangle foo<int, double> and foo<int, char>
1069 given this declaration:
1072 template <class T, class U> void foo(T t);
1075 @item -foperator-names
1076 Recognize the operator name keywords @code{and}, @code{bitand},
1077 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
1078 synonyms for the symbols they refer to. @samp{-ansi} implies
1079 @samp{-foperator-names}.
1081 @item -fno-optional-diags
1082 Disable diagnostics that the standard says a compiler does not need to
1083 issue. Currently, this means the diagnostic for a name having multiple
1084 meanings within a class.
1087 Enable automatic template instantiation. This option also implies
1088 @samp{-fno-implicit-templates}. @xref{Template Instantiation}, for more
1091 @item -fstrict-prototype
1092 Within an @samp{extern "C"} linkage specification, treat a function
1093 declaration with no arguments, such as @samp{int foo ();}, as declaring
1094 the function to take no arguments. Normally, such a declaration means
1095 that the function @code{foo} can take any combination of arguments, as
1096 in C. @samp{-pedantic} implies @samp{-fstrict-prototype} unless
1097 overridden with @samp{-fno-strict-prototype}.
1099 Specifying this option will also suppress implicit declarations of
1102 This flag no longer affects declarations with C++ linkage.
1105 @itemx -fno-squangle
1106 @samp{-fsquangle} will enable a compressed form of name mangling for
1107 identifiers. In particular, it helps to shorten very long names by recognizing
1108 types and class names which occur more than once, replacing them with special
1109 short ID codes. This option also requires any C++ libraries being used to
1110 be compiled with this option as well. The compiler has this disabled (the
1111 equivalent of @samp{-fno-squangle}) by default.
1113 Like all options that change the ABI, all C++ code, @emph{including
1114 libgcc.a} must be built with the same setting of this option.
1116 @item -ftemplate-depth-@var{n}
1117 Set the maximum instantiation depth for template classes to @var{n}.
1118 A limit on the template instantiation depth is needed to detect
1119 endless recursions during template class instantiation. ANSI/ISO C++
1120 conforming programs must not rely on a maximum depth greater than 17.
1122 @item -fthis-is-variable
1123 Permit assignment to @code{this}. The incorporation of user-defined
1124 free store management into C++ has made assignment to @samp{this} an
1125 anachronism. Therefore, by default it is invalid to assign to
1126 @code{this} within a class member function; that is, GNU C++ treats
1127 @samp{this} in a member function of class @code{X} as a non-lvalue of
1128 type @samp{X *}. However, for backwards compatibility, you can make it
1129 valid with @samp{-fthis-is-variable}.
1131 @item -fvtable-thunks
1132 Use @samp{thunks} to implement the virtual function dispatch table
1133 (@samp{vtable}). The traditional (cfront-style) approach to
1134 implementing vtables was to store a pointer to the function and two
1135 offsets for adjusting the @samp{this} pointer at the call site. Newer
1136 implementations store a single pointer to a @samp{thunk} function which
1137 does any necessary adjustment and then calls the target function.
1139 Like all options that change the ABI, all C++ code, @emph{including
1140 libgcc.a} must be built with the same setting of this option.
1143 Do not search for header files in the standard directories specific to
1144 C++, but do still search the other standard directories. (This option
1145 is used when building the C++ library.)
1148 In addition, these optimization, warning, and code generation options
1149 have meanings only for C++ programs:
1152 @item -fno-default-inline
1153 Do not assume @samp{inline} for functions defined inside a class scope.
1154 @xref{Optimize Options,,Options That Control Optimization}. Note that these
1155 functions will have linkage like inline functions; they just won't be
1158 @item -Wno-non-template-friend
1159 @xref{Warning Options,,Options to Request or Suppress Warnings}.
1160 @item -Wold-style-cast
1161 @itemx -Woverloaded-virtual
1162 Warnings that apply only to C++ programs. @xref{Warning
1163 Options,,Options to Request or Suppress Warnings}.
1166 Warn about violation of some style rules from Effective C++ by Scott Myers.
1169 @node Warning Options
1170 @section Options to Request or Suppress Warnings
1171 @cindex options to control warnings
1172 @cindex warning messages
1173 @cindex messages, warning
1174 @cindex suppressing warnings
1176 Warnings are diagnostic messages that report constructions which
1177 are not inherently erroneous but which are risky or suggest there
1178 may have been an error.
1180 You can request many specific warnings with options beginning @samp{-W},
1181 for example @samp{-Wimplicit} to request warnings on implicit
1182 declarations. Each of these specific warning options also has a
1183 negative form beginning @samp{-Wno-} to turn off warnings;
1184 for example, @samp{-Wno-implicit}. This manual lists only one of the
1185 two forms, whichever is not the default.
1187 These options control the amount and kinds of warnings produced by GNU
1191 @cindex syntax checking
1193 Check the code for syntax errors, but don't do anything beyond that.
1196 Issue all the warnings demanded by strict ANSI C and ISO C++;
1197 reject all programs that use forbidden extensions.
1199 Valid ANSI C and ISO C++ programs should compile properly with or without
1200 this option (though a rare few will require @samp{-ansi}). However,
1201 without this option, certain GNU extensions and traditional C and C++
1202 features are supported as well. With this option, they are rejected.
1204 @samp{-pedantic} does not cause warning messages for use of the
1205 alternate keywords whose names begin and end with @samp{__}. Pedantic
1206 warnings are also disabled in the expression that follows
1207 @code{__extension__}. However, only system header files should use
1208 these escape routes; application programs should avoid them.
1209 @xref{Alternate Keywords}.
1211 This option is not intended to be @i{useful}; it exists only to satisfy
1212 pedants who would otherwise claim that GNU CC fails to support the ANSI
1215 Some users try to use @samp{-pedantic} to check programs for strict ANSI
1216 C conformance. They soon find that it does not do quite what they want:
1217 it finds some non-ANSI practices, but not all---only those for which
1218 ANSI C @emph{requires} a diagnostic.
1220 A feature to report any failure to conform to ANSI C might be useful in
1221 some instances, but would require considerable additional work and would
1222 be quite different from @samp{-pedantic}. We recommend, rather, that
1223 users take advantage of the extensions of GNU C and disregard the
1224 limitations of other compilers. Aside from certain supercomputers and
1225 obsolete small machines, there is less and less reason ever to use any
1226 other C compiler other than for bootstrapping GNU CC.
1228 @item -pedantic-errors
1229 Like @samp{-pedantic}, except that errors are produced rather than
1233 Inhibit all warning messages.
1236 Inhibit warning messages about the use of @samp{#import}.
1238 @item -Wchar-subscripts
1239 Warn if an array subscript has type @code{char}. This is a common cause
1240 of error, as programmers often forget that this type is signed on some
1244 Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*}
1245 comment, or whenever a Backslash-Newline appears in a @samp{//} comment.
1248 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
1249 the arguments supplied have types appropriate to the format string
1252 @item -Wimplicit-int
1253 Warn when a declaration does not specify a type.
1255 @item -Wimplicit-function-declaration
1256 @itemx -Werror-implicit-function-declaration
1257 Give a warning (or error) whenever a function is used before being
1261 Same as @samp{-Wimplicit-int} and @samp{-Wimplicit-function-}@*
1265 Warn if the type of @samp{main} is suspicious. @samp{main} should be a
1266 function with external linkage, returning int, taking either zero
1267 arguments, two, or three arguments of appropriate types.
1270 Warn if a multicharacter constant (@samp{'FOOF'}) is used. Usually they
1271 indicate a typo in the user's code, as they have implementation-defined
1272 values, and should not be used in portable code.
1275 Warn if parentheses are omitted in certain contexts, such
1276 as when there is an assignment in a context where a truth value
1277 is expected, or when operators are nested whose precedence people
1278 often get confused about.
1280 Also warn about constructions where there may be confusion to which
1281 @code{if} statement an @code{else} branch belongs. Here is an example of
1294 In C, every @code{else} branch belongs to the innermost possible @code{if}
1295 statement, which in this example is @code{if (b)}. This is often not
1296 what the programmer expected, as illustrated in the above example by
1297 indentation the programmer chose. When there is the potential for this
1298 confusion, GNU C will issue a warning when this flag is specified.
1299 To eliminate the warning, add explicit braces around the innermost
1300 @code{if} statement so there is no way the @code{else} could belong to
1301 the enclosing @code{if}. The resulting code would look like this:
1316 Warn whenever a function is defined with a return-type that defaults
1317 to @code{int}. Also warn about any @code{return} statement with no
1318 return-value in a function whose return-type is not @code{void}.
1321 Warn whenever a @code{switch} statement has an index of enumeral type
1322 and lacks a @code{case} for one or more of the named codes of that
1323 enumeration. (The presence of a @code{default} label prevents this
1324 warning.) @code{case} labels outside the enumeration range also
1325 provoke warnings when this option is used.
1328 Warn if any trigraphs are encountered (assuming they are enabled).
1331 Warn whenever a variable is unused aside from its declaration,
1332 whenever a function is declared static but never defined, whenever a
1333 label is declared but not used, and whenever a statement computes a
1334 result that is explicitly not used.
1336 In order to get a warning about an unused function parameter, you must
1337 specify both @samp{-W} and @samp{-Wunused}.
1339 To suppress this warning for an expression, simply cast it to void. For
1340 unused variables and parameters, use the @samp{unused} attribute
1341 (@pxref{Variable Attributes}).
1343 @item -Wuninitialized
1344 An automatic variable is used without first being initialized.
1346 These warnings are possible only in optimizing compilation,
1347 because they require data flow information that is computed only
1348 when optimizing. If you don't specify @samp{-O}, you simply won't
1351 These warnings occur only for variables that are candidates for
1352 register allocation. Therefore, they do not occur for a variable that
1353 is declared @code{volatile}, or whose address is taken, or whose size
1354 is other than 1, 2, 4 or 8 bytes. Also, they do not occur for
1355 structures, unions or arrays, even when they are in registers.
1357 Note that there may be no warning about a variable that is used only
1358 to compute a value that itself is never used, because such
1359 computations may be deleted by data flow analysis before the warnings
1362 These warnings are made optional because GNU CC is not smart
1363 enough to see all the reasons why the code might be correct
1364 despite appearing to have an error. Here is one example of how
1383 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
1384 always initialized, but GNU CC doesn't know this. Here is
1385 another common case:
1390 if (change_y) save_y = y, y = new_y;
1392 if (change_y) y = save_y;
1397 This has no bug because @code{save_y} is used only if it is set.
1399 Some spurious warnings can be avoided if you declare all the functions
1400 you use that never return as @code{noreturn}. @xref{Function
1403 @item -Wreorder (C++ only)
1404 @cindex reordering, warning
1405 @cindex warning for reordering of member initializers
1406 Warn when the order of member initializers given in the code does not
1407 match the order in which they must be executed. For instance:
1413 A(): j (0), i (1) @{ @}
1417 Here the compiler will warn that the member initializers for @samp{i}
1418 and @samp{j} will be rearranged to match the declaration order of the
1421 @item -Wtemplate-debugging
1422 @cindex template debugging
1423 When using templates in a C++ program, warn if debugging is not yet
1424 fully available (C++ only).
1426 @item -Wunknown-pragmas
1427 @cindex warning for unknown pragmas
1428 @cindex unknown pragmas, warning
1429 @cindex pragmas, warning of unknown
1430 Warn when a #pragma directive is encountered which is not understood by
1431 GCC. If this command line option is used, warnings will even be issued
1432 for unknown pragmas in system header files. This is not the case if
1433 the warnings were only enabled by the @samp{-Wall} command line option.
1436 All of the above @samp{-W} options combined. This enables all the
1437 warnings about constructions that some users consider questionable, and
1438 that are easy to avoid (or modify to prevent the warning), even in
1439 conjunction with macros.
1442 The following @samp{-W@dots{}} options are not implied by @samp{-Wall}.
1443 Some of them warn about constructions that users generally do not
1444 consider questionable, but which occasionally you might wish to check
1445 for; others warn about constructions that are necessary or hard to avoid
1446 in some cases, and there is no simple way to modify the code to suppress
1451 Print extra warning messages for these events:
1454 @cindex @code{longjmp} warnings
1456 A nonvolatile automatic variable might be changed by a call to
1457 @code{longjmp}. These warnings as well are possible only in
1458 optimizing compilation.
1460 The compiler sees only the calls to @code{setjmp}. It cannot know
1461 where @code{longjmp} will be called; in fact, a signal handler could
1462 call it at any point in the code. As a result, you may get a warning
1463 even when there is in fact no problem because @code{longjmp} cannot
1464 in fact be called at the place which would cause a problem.
1467 A function can return either with or without a value. (Falling
1468 off the end of the function body is considered returning without
1469 a value.) For example, this function would evoke such a
1483 An expression-statement or the left-hand side of a comma expression
1484 contains no side effects.
1485 To suppress the warning, cast the unused expression to void.
1486 For example, an expression such as @samp{x[i,j]} will cause a warning,
1487 but @samp{x[(void)i,j]} will not.
1490 An unsigned value is compared against zero with @samp{<} or @samp{<=}.
1493 A comparison like @samp{x<=y<=z} appears; this is equivalent to
1494 @samp{(x<=y ? 1 : 0) <= z}, which is a different interpretation from
1495 that of ordinary mathematical notation.
1498 Storage-class specifiers like @code{static} are not the first things in
1499 a declaration. According to the C Standard, this usage is obsolescent.
1502 If @samp{-Wall} or @samp{-Wunused} is also specified, warn about unused
1506 A comparison between signed and unsigned values could produce an
1507 incorrect result when the signed value is converted to unsigned.
1508 (But don't warn if @samp{-Wno-sign-compare} is also specified.)
1511 An aggregate has a partly bracketed initializer.
1512 For example, the following code would evoke such a warning,
1513 because braces are missing around the initializer for @code{x.h}:
1516 struct s @{ int f, g; @};
1517 struct t @{ struct s h; int i; @};
1518 struct t x = @{ 1, 2, 3 @};
1522 An aggregate has an initializer which does not initialize all members.
1523 For example, the following code would cause such a warning, because
1524 @code{x.h} would be implicitly initialized to zero:
1527 struct s @{ int f, g, h; @};
1528 struct s x = @{ 3, 4 @};
1533 Warn about certain constructs that behave differently in traditional and
1538 Macro arguments occurring within string constants in the macro body.
1539 These would substitute the argument in traditional C, but are part of
1540 the constant in ANSI C.
1543 A function declared external in one block and then used after the end of
1547 A @code{switch} statement has an operand of type @code{long}.
1551 Warn if an undefined identifier is evaluated in an @samp{#if} directive.
1554 Warn whenever a local variable shadows another local variable.
1556 @item -Wid-clash-@var{len}
1557 Warn whenever two distinct identifiers match in the first @var{len}
1558 characters. This may help you prepare a program that will compile
1559 with certain obsolete, brain-damaged compilers.
1561 @item -Wlarger-than-@var{len}
1562 Warn whenever an object of larger than @var{len} bytes is defined.
1564 @item -Wpointer-arith
1565 Warn about anything that depends on the ``size of'' a function type or
1566 of @code{void}. GNU C assigns these types a size of 1, for
1567 convenience in calculations with @code{void *} pointers and pointers
1570 @item -Wbad-function-cast
1571 Warn whenever a function call is cast to a non-matching type.
1572 For example, warn if @code{int malloc()} is cast to @code{anything *}.
1575 Warn whenever a pointer is cast so as to remove a type qualifier from
1576 the target type. For example, warn if a @code{const char *} is cast
1577 to an ordinary @code{char *}.
1580 Warn whenever a pointer is cast such that the required alignment of the
1581 target is increased. For example, warn if a @code{char *} is cast to
1582 an @code{int *} on machines where integers can only be accessed at
1583 two- or four-byte boundaries.
1585 @item -Wwrite-strings
1586 Give string constants the type @code{const char[@var{length}]} so that
1587 copying the address of one into a non-@code{const} @code{char *}
1588 pointer will get a warning. These warnings will help you find at
1589 compile time code that can try to write into a string constant, but
1590 only if you have been very careful about using @code{const} in
1591 declarations and prototypes. Otherwise, it will just be a nuisance;
1592 this is why we did not make @samp{-Wall} request these warnings.
1595 Warn if a prototype causes a type conversion that is different from what
1596 would happen to the same argument in the absence of a prototype. This
1597 includes conversions of fixed point to floating and vice versa, and
1598 conversions changing the width or signedness of a fixed point argument
1599 except when the same as the default promotion.
1601 Also, warn if a negative integer constant expression is implicitly
1602 converted to an unsigned type. For example, warn about the assignment
1603 @code{x = -1} if @code{x} is unsigned. But do not warn about explicit
1604 casts like @code{(unsigned) -1}.
1606 @item -Wsign-compare
1607 @cindex warning for comparison of signed and unsigned values
1608 @cindex comparison of signed and unsigned values, warning
1609 @cindex signed and unsigned values, comparison warning
1610 Warn when a comparison between signed and unsigned values could produce
1611 an incorrect result when the signed value is converted to unsigned.
1612 This warning is also enabled by @samp{-W}; to get the other warnings
1613 of @samp{-W} without this warning, use @samp{-W -Wno-sign-compare}.
1615 @item -Waggregate-return
1616 Warn if any functions that return structures or unions are defined or
1617 called. (In languages where you can return an array, this also elicits
1620 @item -Wstrict-prototypes
1621 Warn if a function is declared or defined without specifying the
1622 argument types. (An old-style function definition is permitted without
1623 a warning if preceded by a declaration which specifies the argument
1626 @item -Wmissing-prototypes
1627 Warn if a global function is defined without a previous prototype
1628 declaration. This warning is issued even if the definition itself
1629 provides a prototype. The aim is to detect global functions that fail
1630 to be declared in header files.
1632 @item -Wmissing-declarations
1633 Warn if a global function is defined without a previous declaration.
1634 Do so even if the definition itself provides a prototype.
1635 Use this option to detect global functions that are not declared in
1638 @item -Wmissing-noreturn
1639 Warn about functions which might be candidates for attribute @code{noreturn}.
1640 Note these are only possible candidates, not absolute ones. Care should
1641 be taken to manually verify functions actually do not ever return before
1642 adding the @code{noreturn} attribute, otherwise subtle code generation
1643 bugs could be introduced.
1645 @item -Wredundant-decls
1646 Warn if anything is declared more than once in the same scope, even in
1647 cases where multiple declaration is valid and changes nothing.
1649 @item -Wnested-externs
1650 Warn if an @code{extern} declaration is encountered within an function.
1652 @item -Wno-non-template-friend
1653 Disable warnings when non-templatized friend functions are declared
1654 within a template. With the advent of explicit template specification
1655 support in g++, if the name of the friend is an unqualified-id (ie,
1656 @samp{friend foo(int)}), the C++ language specification demands that the
1657 friend declare or define an ordinary, nontemplate function. (Section
1658 14.5.3). Before g++ implemented explicit specification, unqualified-ids
1659 could be interpreted as a particular specialization of a templatized
1660 function. Because this non-conforming behavior is no longer the default
1661 behavior for g++, @samp{-Wnon-template-friend} allows the compiler to
1662 check existing code for potential trouble spots, and is on by default.
1663 This new compiler behavior can also be turned off with the flag
1664 @samp{-fguiding-decls}, which activates the older, non-specification
1665 compiler code, or with @samp{-Wno-non-template-friend} which keeps the
1666 conformant compiler code but disables the helpful warning.
1669 Warn if a function can not be inlined, and either it was declared as inline,
1670 or else the @samp{-finline-functions} option was given.
1672 @item -Wold-style-cast
1673 Warn if an old-style (C-style) cast is used within a program.
1675 @item -Woverloaded-virtual
1676 @cindex overloaded virtual fn, warning
1677 @cindex warning for overloaded virtual fn
1678 Warn when a derived class function declaration may be an error in
1679 defining a virtual function (C++ only). In a derived class, the
1680 definitions of virtual functions must match the type signature of a
1681 virtual function declared in the base class. With this option, the
1682 compiler warns when you define a function with the same name as a
1683 virtual function, but with a type signature that does not match any
1684 declarations from the base class.
1686 @item -Wsynth (C++ only)
1687 @cindex warning for synthesized methods
1688 @cindex synthesized methods, warning
1689 Warn when g++'s synthesis behavior does not match that of cfront. For
1695 A& operator = (int);
1705 In this example, g++ will synthesize a default @samp{A& operator =
1706 (const A&);}, while cfront will use the user-defined @samp{operator =}.
1709 Warn if @samp{long long} type is used. This is default. To inhibit
1710 the warning messages, use @samp{-Wno-long-long}. Flags
1711 @samp{-Wlong-long} and @samp{-Wno-long-long} are taken into account
1712 only when @samp{-pedantic} flag is used.
1715 Make all warnings into errors.
1718 @node Debugging Options
1719 @section Options for Debugging Your Program or GNU CC
1720 @cindex options, debugging
1721 @cindex debugging information options
1723 GNU CC has various special options that are used for debugging
1724 either your program or GCC:
1728 Produce debugging information in the operating system's native format
1729 (stabs, COFF, XCOFF, or DWARF). GDB can work with this debugging
1732 On most systems that use stabs format, @samp{-g} enables use of extra
1733 debugging information that only GDB can use; this extra information
1734 makes debugging work better in GDB but will probably make other debuggers
1736 refuse to read the program. If you want to control for certain whether
1737 to generate the extra information, use @samp{-gstabs+}, @samp{-gstabs},
1738 @samp{-gxcoff+}, @samp{-gxcoff}, @samp{-gdwarf-1+}, or @samp{-gdwarf-1}
1741 Unlike most other C compilers, GNU CC allows you to use @samp{-g} with
1742 @samp{-O}. The shortcuts taken by optimized code may occasionally
1743 produce surprising results: some variables you declared may not exist
1744 at all; flow of control may briefly move where you did not expect it;
1745 some statements may not be executed because they compute constant
1746 results or their values were already at hand; some statements may
1747 execute in different places because they were moved out of loops.
1749 Nevertheless it proves possible to debug optimized output. This makes
1750 it reasonable to use the optimizer for programs that might have bugs.
1752 The following options are useful when GNU CC is generated with the
1753 capability for more than one debugging format.
1756 Produce debugging information for use by GDB. This means to use the
1757 most expressive format available (DWARF 2, stabs, or the native format
1758 if neither of those are supported), including GDB extensions if at all
1762 Produce debugging information in stabs format (if that is supported),
1763 without GDB extensions. This is the format used by DBX on most BSD
1764 systems. On MIPS, Alpha and System V Release 4 systems this option
1765 produces stabs debugging output which is not understood by DBX or SDB.
1766 On System V Release 4 systems this option requires the GNU assembler.
1769 Produce debugging information in stabs format (if that is supported),
1770 using GNU extensions understood only by the GNU debugger (GDB). The
1771 use of these extensions is likely to make other debuggers crash or
1772 refuse to read the program.
1775 Produce debugging information in COFF format (if that is supported).
1776 This is the format used by SDB on most System V systems prior to
1780 Produce debugging information in XCOFF format (if that is supported).
1781 This is the format used by the DBX debugger on IBM RS/6000 systems.
1784 Produce debugging information in XCOFF format (if that is supported),
1785 using GNU extensions understood only by the GNU debugger (GDB). The
1786 use of these extensions is likely to make other debuggers crash or
1787 refuse to read the program, and may cause assemblers other than the GNU
1788 assembler (GAS) to fail with an error.
1791 Produce debugging information in DWARF version 1 format (if that is
1792 supported). This is the format used by SDB on most System V Release 4
1796 Produce debugging information in DWARF version 1 format (if that is
1797 supported), using GNU extensions understood only by the GNU debugger
1798 (GDB). The use of these extensions is likely to make other debuggers
1799 crash or refuse to read the program.
1802 Produce debugging information in DWARF version 2 format (if that is
1803 supported). This is the format used by DBX on IRIX 6.
1806 @itemx -ggdb@var{level}
1807 @itemx -gstabs@var{level}
1808 @itemx -gcoff@var{level}
1809 @itemx -gxcoff@var{level}
1810 @itemx -gdwarf@var{level}
1811 @itemx -gdwarf-2@var{level}
1812 Request debugging information and also use @var{level} to specify how
1813 much information. The default level is 2.
1815 Level 1 produces minimal information, enough for making backtraces in
1816 parts of the program that you don't plan to debug. This includes
1817 descriptions of functions and external variables, but no information
1818 about local variables and no line numbers.
1820 Level 3 includes extra information, such as all the macro definitions
1821 present in the program. Some debuggers support macro expansion when
1826 Generate extra code to write profile information suitable for the
1827 analysis program @code{prof}. You must use this option when compiling
1828 the source files you want data about, and you must also use it when
1831 @cindex @code{gprof}
1833 Generate extra code to write profile information suitable for the
1834 analysis program @code{gprof}. You must use this option when compiling
1835 the source files you want data about, and you must also use it when
1840 Generate extra code to write profile information for basic blocks, which will
1841 record the number of times each basic block is executed, the basic block start
1842 address, and the function name containing the basic block. If @samp{-g} is
1843 used, the line number and filename of the start of the basic block will also be
1844 recorded. If not overridden by the machine description, the default action is
1845 to append to the text file @file{bb.out}.
1847 This data could be analyzed by a program like @code{tcov}. Note,
1848 however, that the format of the data is not what @code{tcov} expects.
1849 Eventually GNU @code{gprof} should be extended to process this data.
1852 Makes the compiler print out each function name as it is compiled, and
1853 print some statistics about each pass when it finishes.
1856 Generate extra code to profile basic blocks. Your executable will
1857 produce output that is a superset of that produced when @samp{-a} is
1858 used. Additional output is the source and target address of the basic
1859 blocks where a jump takes place, the number of times a jump is executed,
1860 and (optionally) the complete sequence of basic blocks being executed.
1861 The output is appended to file @file{bb.out}.
1863 You can examine different profiling aspects without recompilation. Your
1864 executable will read a list of function names from file @file{bb.in}.
1865 Profiling starts when a function on the list is entered and stops when
1866 that invocation is exited. To exclude a function from profiling, prefix
1867 its name with `-'. If a function name is not unique, you can
1868 disambiguate it by writing it in the form
1869 @samp{/path/filename.d:functionname}. Your executable will write the
1870 available paths and filenames in file @file{bb.out}.
1872 Several function names have a special meaning:
1875 Write source, target and frequency of jumps to file @file{bb.out}.
1876 @item __bb_hidecall__
1877 Exclude function calls from frequency count.
1878 @item __bb_showret__
1879 Include function returns in frequency count.
1881 Write the sequence of basic blocks executed to file @file{bbtrace.gz}.
1882 The file will be compressed using the program @samp{gzip}, which must
1883 exist in your @code{PATH}. On systems without the @samp{popen}
1884 function, the file will be named @file{bbtrace} and will not be
1885 compressed. @strong{Profiling for even a few seconds on these systems
1886 will produce a very large file.} Note: @code{__bb_hidecall__} and
1887 @code{__bb_showret__} will not affect the sequence written to
1891 Here's a short example using different profiling parameters
1892 in file @file{bb.in}. Assume function @code{foo} consists of basic blocks
1893 1 and 2 and is called twice from block 3 of function @code{main}. After
1894 the calls, block 3 transfers control to block 4 of @code{main}.
1896 With @code{__bb_trace__} and @code{main} contained in file @file{bb.in},
1897 the following sequence of blocks is written to file @file{bbtrace.gz}:
1898 0 3 1 2 1 2 4. The return from block 2 to block 3 is not shown, because
1899 the return is to a point inside the block and not to the top. The
1900 block address 0 always indicates, that control is transferred
1901 to the trace from somewhere outside the observed functions. With
1902 @samp{-foo} added to @file{bb.in}, the blocks of function
1903 @code{foo} are removed from the trace, so only 0 3 4 remains.
1905 With @code{__bb_jumps__} and @code{main} contained in file @file{bb.in},
1906 jump frequencies will be written to file @file{bb.out}. The
1907 frequencies are obtained by constructing a trace of blocks
1908 and incrementing a counter for every neighbouring pair of blocks
1909 in the trace. The trace 0 3 1 2 1 2 4 displays the following
1913 Jump from block 0x0 to block 0x3 executed 1 time(s)
1914 Jump from block 0x3 to block 0x1 executed 1 time(s)
1915 Jump from block 0x1 to block 0x2 executed 2 time(s)
1916 Jump from block 0x2 to block 0x1 executed 1 time(s)
1917 Jump from block 0x2 to block 0x4 executed 1 time(s)
1920 With @code{__bb_hidecall__}, control transfer due to call instructions
1921 is removed from the trace, that is the trace is cut into three parts: 0
1922 3 4, 0 1 2 and 0 1 2. With @code{__bb_showret__}, control transfer due
1923 to return instructions is added to the trace. The trace becomes: 0 3 1
1924 2 3 1 2 3 4. Note, that this trace is not the same, as the sequence
1925 written to @file{bbtrace.gz}. It is solely used for counting jump
1928 @item -fprofile-arcs
1929 Instrument @dfn{arcs} during compilation. For each function of your
1930 program, GNU CC creates a program flow graph, then finds a spanning tree
1931 for the graph. Only arcs that are not on the spanning tree have to be
1932 instrumented: the compiler adds code to count the number of times that these
1933 arcs are executed. When an arc is the only exit or only entrance to a
1934 block, the instrumentation code can be added to the block; otherwise, a
1935 new basic block must be created to hold the instrumentation code.
1937 Since not every arc in the program must be instrumented, programs
1938 compiled with this option run faster than programs compiled with
1939 @samp{-a}, which adds instrumentation code to every basic block in the
1940 program. The tradeoff: since @code{gcov} does not have
1941 execution counts for all branches, it must start with the execution
1942 counts for the instrumented branches, and then iterate over the program
1943 flow graph until the entire graph has been solved. Hence, @code{gcov}
1944 runs a little more slowly than a program which uses information from
1947 @samp{-fprofile-arcs} also makes it possible to estimate branch
1948 probabilities, and to calculate basic block execution counts. In
1949 general, basic block execution counts do not give enough information to
1950 estimate all branch probabilities. When the compiled program exits, it
1951 saves the arc execution counts to a file called
1952 @file{@var{sourcename}.da}. Use the compiler option
1953 @samp{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
1954 Control Optimization}) when recompiling, to optimize using estimated
1955 branch probabilities.
1958 @item -ftest-coverage
1959 Create data files for the @code{gcov} code-coverage utility
1960 (@pxref{Gcov,, @code{gcov}: a GNU CC Test Coverage Program}).
1961 The data file names begin with the name of your source file:
1964 @item @var{sourcename}.bb
1965 A mapping from basic blocks to line numbers, which @code{gcov} uses to
1966 associate basic block execution counts with line numbers.
1968 @item @var{sourcename}.bbg
1969 A list of all arcs in the program flow graph. This allows @code{gcov}
1970 to reconstruct the program flow graph, so that it can compute all basic
1971 block and arc execution counts from the information in the
1972 @code{@var{sourcename}.da} file (this last file is the output from
1973 @samp{-fprofile-arcs}).
1977 Makes the compiler print out each function name as it is compiled, and
1978 print some statistics about each pass when it finishes.
1980 @item -d@var{letters}
1981 Says to make debugging dumps during compilation at times specified by
1982 @var{letters}. This is used for debugging the compiler. The file names
1983 for most of the dumps are made by appending a word to the source file
1984 name (e.g. @file{foo.c.rtl} or @file{foo.c.jump}). Here are the
1985 possible letters for use in @var{letters}, and their meanings:
1989 Dump after computing branch probabilities, to @file{@var{file}.bp}.
1991 Dump after instruction combination, to the file @file{@var{file}.combine}.
1993 Dump after delayed branch scheduling, to @file{@var{file}.dbr}.
1995 Dump all macro definitions, at the end of preprocessing, in addition to
1998 Dump debugging information during parsing, to standard error.
2000 Dump after RTL generation, to @file{@var{file}.rtl}.
2002 Just generate RTL for a function instead of compiling it. Usually used
2005 Dump after first jump optimization, to @file{@var{file}.jump}.
2007 Dump after CSE (including the jump optimization that sometimes
2008 follows CSE), to @file{@var{file}.cse}.
2010 Dump after purging ADDRESSOF, to @file{@var{file}.addressof}.
2012 Dump after flow analysis, to @file{@var{file}.flow}.
2014 Dump after global register allocation, to @file{@var{file}.greg}.
2016 Dump after GCSE, to @file{@var{file}.gcse}.
2018 Dump after first jump optimization, to @file{@var{file}.jump}.
2020 Dump after last jump optimization, to @file{@var{file}.jump2}.
2022 Dump after conversion from registers to stack, to @file{@var{file}.stack}.
2024 Dump after local register allocation, to @file{@var{file}.lreg}.
2026 Dump after loop optimization, to @file{@var{file}.loop}.
2028 Dump after performing the machine dependent reorganisation pass, to
2029 @file{@var{file}.mach}.
2031 Dump after the register move pass, to @file{@var{file}.regmove}.
2033 Dump after RTL generation, to @file{@var{file}.rtl}.
2035 Dump after the second instruction scheduling pass, to @file{@var{file}.sched2}.
2037 Dump after CSE (including the jump optimization that sometimes follows
2038 CSE), to @file{@var{file}.cse}.
2040 Dump after the first instruction scheduling pass, to @file{@var{file}.sched}.
2042 Dump after the second CSE pass (including the jump optimization that
2043 sometimes follows CSE), to @file{@var{file}.cse2}.
2045 Just generate RTL for a function instead of compiling it. Usually used
2048 Produce all the dumps listed above.
2050 Print statistics on memory usage, at the end of the run, to
2053 Annotate the assembler output with a comment indicating which
2054 pattern and alternative was used.
2056 Dump debugging information during parsing, to standard error.
2058 Annotate the assembler output with miscellaneous debugging information.
2061 @item -fdump-unnumbered
2062 When doing debugging dumps (see -d option above), suppress instruction
2063 numbers and line number note output. This makes it more feasible to
2064 use diff on debugging dumps for compiler invokations with different
2065 options, in particular with and without -g.
2067 @item -fpretend-float
2068 When running a cross-compiler, pretend that the target machine uses the
2069 same floating point format as the host machine. This causes incorrect
2070 output of the actual floating constants, but the actual instruction
2071 sequence will probably be the same as GNU CC would make when running on
2075 Store the usual ``temporary'' intermediate files permanently; place them
2076 in the current directory and name them based on the source file. Thus,
2077 compiling @file{foo.c} with @samp{-c -save-temps} would produce files
2078 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}.
2080 @item -print-file-name=@var{library}
2081 Print the full absolute name of the library file @var{library} that
2082 would be used when linking---and don't do anything else. With this
2083 option, GNU CC does not compile or link anything; it just prints the
2086 @item -print-prog-name=@var{program}
2087 Like @samp{-print-file-name}, but searches for a program such as @samp{cpp}.
2089 @item -print-libgcc-file-name
2090 Same as @samp{-print-file-name=libgcc.a}.
2092 This is useful when you use @samp{-nostdlib} or @samp{-nodefaultlibs}
2093 but you do want to link with @file{libgcc.a}. You can do
2096 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
2099 @item -print-search-dirs
2100 Print the name of the configured installation directory and a list of
2101 program and library directories gcc will search---and don't do anything else.
2103 This is useful when gcc prints the error message
2104 @samp{installation problem, cannot exec cpp: No such file or directory}.
2105 To resolve this you either need to put @file{cpp} and the other compiler
2106 components where gcc expects to find them, or you can set the environment
2107 variable @code{GCC_EXEC_PREFIX} to the directory where you installed them.
2108 Don't forget the trailing '/'.
2109 @xref{Environment Variables}.
2112 @node Optimize Options
2113 @section Options That Control Optimization
2114 @cindex optimize options
2115 @cindex options, optimization
2117 These options control various sorts of optimizations:
2122 Optimize. Optimizing compilation takes somewhat more time, and a lot
2123 more memory for a large function.
2125 Without @samp{-O}, the compiler's goal is to reduce the cost of
2126 compilation and to make debugging produce the expected results.
2127 Statements are independent: if you stop the program with a breakpoint
2128 between statements, you can then assign a new value to any variable or
2129 change the program counter to any other statement in the function and
2130 get exactly the results you would expect from the source code.
2132 Without @samp{-O}, the compiler only allocates variables declared
2133 @code{register} in registers. The resulting compiled code is a little
2134 worse than produced by PCC without @samp{-O}.
2136 With @samp{-O}, the compiler tries to reduce code size and execution
2139 When you specify @samp{-O}, the compiler turns on @samp{-fthread-jumps}
2140 and @samp{-fdefer-pop} on all machines. The compiler turns on
2141 @samp{-fdelayed-branch} on machines that have delay slots, and
2142 @samp{-fomit-frame-pointer} on machines that can support debugging even
2143 without a frame pointer. On some machines the compiler also turns
2144 on other flags.@refill
2147 Optimize even more. GNU CC performs nearly all supported optimizations
2148 that do not involve a space-speed tradeoff. The compiler does not
2149 perform loop unrolling or function inlining when you specify @samp{-O2}.
2150 As compared to @samp{-O}, this option increases both compilation time
2151 and the performance of the generated code.
2153 @samp{-O2} turns on all optional optimizations except for loop unrolling
2154 and function inlining. It also turns on the @samp{-fforce-mem} option
2155 on all machines and frame pointer elimination on machines where doing so
2156 does not interfere with debugging.
2159 Optimize yet more. @samp{-O3} turns on all optimizations specified by
2160 @samp{-O2} and also turns on the @samp{inline-functions} option.
2166 Optimize for size. @samp{-Os} enables all @samp{-O2} optimizations that
2167 do not typically increase code size. It also performs further
2168 optimizations designed to reduce code size.
2170 If you use multiple @samp{-O} options, with or without level numbers,
2171 the last such option is the one that is effective.
2174 Options of the form @samp{-f@var{flag}} specify machine-independent
2175 flags. Most flags have both positive and negative forms; the negative
2176 form of @samp{-ffoo} would be @samp{-fno-foo}. In the table below,
2177 only one of the forms is listed---the one which is not the default.
2178 You can figure out the other form by either removing @samp{no-} or
2183 Do not store floating point variables in registers, and inhibit other
2184 options that might change whether a floating point value is taken from a
2187 @cindex floating point precision
2188 This option prevents undesirable excess precision on machines such as
2189 the 68000 where the floating registers (of the 68881) keep more
2190 precision than a @code{double} is supposed to have. Similarly for the
2191 x86 architecture. For most programs, the excess precision does only
2192 good, but a few programs rely on the precise definition of IEEE floating
2193 point. Use @samp{-ffloat-store} for such programs, after modifying
2194 them to store all pertinent intermediate computations into variables.
2196 @item -fno-default-inline
2197 Do not make member functions inline by default merely because they are
2198 defined inside the class scope (C++ only). Otherwise, when you specify
2199 @w{@samp{-O}}, member functions defined inside class scope are compiled
2200 inline by default; i.e., you don't need to add @samp{inline} in front of
2201 the member function name.
2203 @item -fno-defer-pop
2204 Always pop the arguments to each function call as soon as that function
2205 returns. For machines which must pop arguments after a function call,
2206 the compiler normally lets arguments accumulate on the stack for several
2207 function calls and pops them all at once.
2210 Force memory operands to be copied into registers before doing
2211 arithmetic on them. This produces better code by making all memory
2212 references potential common subexpressions. When they are not common
2213 subexpressions, instruction combination should eliminate the separate
2214 register-load. The @samp{-O2} option turns on this option.
2217 Force memory address constants to be copied into registers before
2218 doing arithmetic on them. This may produce better code just as
2219 @samp{-fforce-mem} may.
2221 @item -fomit-frame-pointer
2222 Don't keep the frame pointer in a register for functions that
2223 don't need one. This avoids the instructions to save, set up and
2224 restore frame pointers; it also makes an extra register available
2225 in many functions. @strong{It also makes debugging impossible on
2229 On some machines, such as the Vax, this flag has no effect, because
2230 the standard calling sequence automatically handles the frame pointer
2231 and nothing is saved by pretending it doesn't exist. The
2232 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
2233 whether a target machine supports this flag. @xref{Registers}.@refill
2236 On some machines, such as the Vax, this flag has no effect, because
2237 the standard calling sequence automatically handles the frame pointer
2238 and nothing is saved by pretending it doesn't exist. The
2239 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
2240 whether a target machine supports this flag. @xref{Registers,,Register
2241 Usage, gcc.info, Using and Porting GCC}.@refill
2245 Don't pay attention to the @code{inline} keyword. Normally this option
2246 is used to keep the compiler from expanding any functions inline.
2247 Note that if you are not optimizing, no functions can be expanded inline.
2249 @item -finline-functions
2250 Integrate all simple functions into their callers. The compiler
2251 heuristically decides which functions are simple enough to be worth
2252 integrating in this way.
2254 If all calls to a given function are integrated, and the function is
2255 declared @code{static}, then the function is normally not output as
2256 assembler code in its own right.
2258 @item -fkeep-inline-functions
2259 Even if all calls to a given function are integrated, and the function
2260 is declared @code{static}, nevertheless output a separate run-time
2261 callable version of the function. This switch does not affect
2262 @code{extern inline} functions.
2264 @item -fkeep-static-consts
2265 Emit variables declared @code{static const} when optimization isn't turned
2266 on, even if the variables aren't referenced.
2268 GNU CC enables this option by default. If you want to force the compiler to
2269 check if the variable was referenced, regardless of whether or not
2270 optimization is turned on, use the @samp{-fno-keep-static-consts} option.
2272 @item -fno-function-cse
2273 Do not put function addresses in registers; make each instruction that
2274 calls a constant function contain the function's address explicitly.
2276 This option results in less efficient code, but some strange hacks
2277 that alter the assembler output may be confused by the optimizations
2278 performed when this option is not used.
2281 This option allows GCC to violate some ANSI or IEEE rules and/or
2282 specifications in the interest of optimizing code for speed. For
2283 example, it allows the compiler to assume arguments to the @code{sqrt}
2284 function are non-negative numbers and that no floating-point values
2287 This option should never be turned on by any @samp{-O} option since
2288 it can result in incorrect output for programs which depend on
2289 an exact implementation of IEEE or ANSI rules/specifications for
2293 @c following causes underfulls.. they don't look great, but we deal.
2295 The following options control specific optimizations. The @samp{-O2}
2296 option turns on all of these optimizations except @samp{-funroll-loops}
2297 and @samp{-funroll-all-loops}. On most machines, the @samp{-O} option
2298 turns on the @samp{-fthread-jumps} and @samp{-fdelayed-branch} options,
2299 but specific machines may handle it differently.
2301 You can use the following flags in the rare cases when ``fine-tuning''
2302 of optimizations to be performed is desired.
2305 @item -fstrength-reduce
2306 Perform the optimizations of loop strength reduction and
2307 elimination of iteration variables.
2309 @item -fthread-jumps
2310 Perform optimizations where we check to see if a jump branches to a
2311 location where another comparison subsumed by the first is found. If
2312 so, the first branch is redirected to either the destination of the
2313 second branch or a point immediately following it, depending on whether
2314 the condition is known to be true or false.
2316 @item -fcse-follow-jumps
2317 In common subexpression elimination, scan through jump instructions
2318 when the target of the jump is not reached by any other path. For
2319 example, when CSE encounters an @code{if} statement with an
2320 @code{else} clause, CSE will follow the jump when the condition
2323 @item -fcse-skip-blocks
2324 This is similar to @samp{-fcse-follow-jumps}, but causes CSE to
2325 follow jumps which conditionally skip over blocks. When CSE
2326 encounters a simple @code{if} statement with no else clause,
2327 @samp{-fcse-skip-blocks} causes CSE to follow the jump around the
2328 body of the @code{if}.
2330 @item -frerun-cse-after-loop
2331 Re-run common subexpression elimination after loop optimizations has been
2334 @item -frerun-loop-opt
2335 Run the loop optimizer twice.
2338 Perform a global common subexpression elimination pass.
2339 This pass also performs global constant and copy propagation.
2341 @item -fexpensive-optimizations
2342 Perform a number of minor optimizations that are relatively expensive.
2344 @item -foptimize-register-moves
2346 Attempt to reassign register numbers in move instructions and as
2347 operands of other simple instructions in order to maximize the amount of
2348 register tying. This is especially helpful on machines with two-operand
2349 instructions. GNU CC enables this optimization by default with @samp{-O2}
2352 Note @code{-fregmove} and @code{-foptimize-register-moves} are the same
2355 @item -fdelayed-branch
2356 If supported for the target machine, attempt to reorder instructions
2357 to exploit instruction slots available after delayed branch
2360 @item -fschedule-insns
2361 If supported for the target machine, attempt to reorder instructions to
2362 eliminate execution stalls due to required data being unavailable. This
2363 helps machines that have slow floating point or memory load instructions
2364 by allowing other instructions to be issued until the result of the load
2365 or floating point instruction is required.
2367 @item -fschedule-insns2
2368 Similar to @samp{-fschedule-insns}, but requests an additional pass of
2369 instruction scheduling after register allocation has been done. This is
2370 especially useful on machines with a relatively small number of
2371 registers and where memory load instructions take more than one cycle.
2373 @item -ffunction-sections
2374 @item -fdata-sections
2375 Place each function or data item into its own section in the output
2376 file if the target supports arbitrary sections. The name of the
2377 function or the name of the data item determines the section's name
2380 Use these options on systems where the linker can perform optimizations
2381 to improve locality of reference in the instruction space. HPPA
2382 processors running HP-UX and Sparc processors running Solaris 2 have
2383 linkers with such optimizations. Other systems using the ELF object format
2384 as well as AIX may have these optimizations in the future.
2386 Only use these options when there are significant benefits from doing
2387 so. When you specify these options, the assembler and linker will
2388 create larger object and executable files and will also be slower.
2389 You will not be able to use @code{gprof} on all systems if you
2390 specify this option and you may have problems with debugging if
2391 you specify both this option and @samp{-g}.
2393 @item -fcaller-saves
2394 Enable values to be allocated in registers that will be clobbered by
2395 function calls, by emitting extra instructions to save and restore the
2396 registers around such calls. Such allocation is done only when it
2397 seems to result in better code than would otherwise be produced.
2399 This option is always enabled by default on certain machines, usually
2400 those which have no call-preserved registers to use instead.
2402 For all machines, optimization level 2 and higher enables this flag by
2405 @item -funroll-loops
2406 Perform the optimization of loop unrolling. This is only done for loops
2407 whose number of iterations can be determined at compile time or run time.
2408 @samp{-funroll-loop} implies both @samp{-fstrength-reduce} and
2409 @samp{-frerun-cse-after-loop}.
2411 @item -funroll-all-loops
2412 Perform the optimization of loop unrolling. This is done for all loops
2413 and usually makes programs run more slowly. @samp{-funroll-all-loops}
2414 implies @samp{-fstrength-reduce} as well as @samp{-frerun-cse-after-loop}.
2416 @item -fmove-all-movables
2417 Forces all invariant computations in loops to be moved
2420 @item -freduce-all-givs
2421 Forces all general-induction variables in loops to be
2424 @emph{Note:} When compiling programs written in Fortran,
2425 @samp{-fmove-all-moveables} and @samp{-freduce-all-givs} are enabled
2426 by default when you use the optimizer.
2428 These options may generate better or worse code; results are highly
2429 dependent on the structure of loops within the source code.
2431 These two options are intended to be removed someday, once
2432 they have helped determine the efficacy of various
2433 approaches to improving loop optimizations.
2435 Please let us (@code{egcs@@cygnus.com} and @code{fortran@@gnu.org})
2436 know how use of these options affects
2437 the performance of your production code.
2438 We're very interested in code that runs @emph{slower}
2439 when these options are @emph{enabled}.
2442 Disable any machine-specific peephole optimizations.
2444 @item -fbranch-probabilities
2445 After running a program compiled with @samp{-fprofile-arcs}
2446 (@pxref{Debugging Options,, Options for Debugging Your Program or
2447 @code{gcc}}), you can compile it a second time using
2448 @samp{-fbranch-probabilities}, to improve optimizations based on
2449 guessing the path a branch might take.
2452 With @samp{-fbranch-probabilities}, GCC puts a @samp{REG_EXEC_COUNT}
2453 note on the first instruction of each basic block, and a
2454 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
2455 These can be used to improve optimization. Currently, they are only
2456 used in one place: in @file{reorg.c}, instead of guessing which path a
2457 branch is mostly to take, the @samp{REG_BR_PROB} values are used to
2458 exactly determine which path is taken more often.
2461 @item -fstrict-aliasing
2462 Allows the compiler to assume the strictest aliasing rules applicable to
2463 the language being compiled. For C (and C++), this activates
2464 optimizations based on the type of expressions. In particular, an
2465 object of one type is assumed never to reside at the same address as an
2466 object of a different type, unless the types are almost the same. For
2467 example, an @code{unsigned int} can alias an @code{int}, but not a
2468 @code{void*} or a @code{double}. A character type may alias any other
2471 Pay special attention to code like this:
2484 The practice of reading from a different union member than the one most
2485 recently written to (called ``type-punning'') is common. Even with
2486 @samp{-fstrict-aliasing}, type-punning is allowed, provided the memory
2487 is accessed through the union type. So, the code above will work as
2488 expected. However, this code might not:
2500 Every language that wishes to perform language-specific alias analysis
2501 should define a function that computes, given an @code{tree}
2502 node, an alias set for the node. Nodes in different alias sets are not
2503 allowed to alias. For an example, see the C front-end function
2504 @code{c_get_alias_set}.
2509 @node Preprocessor Options
2510 @section Options Controlling the Preprocessor
2511 @cindex preprocessor options
2512 @cindex options, preprocessor
2514 These options control the C preprocessor, which is run on each C source
2515 file before actual compilation.
2517 If you use the @samp{-E} option, nothing is done except preprocessing.
2518 Some of these options make sense only together with @samp{-E} because
2519 they cause the preprocessor output to be unsuitable for actual
2523 @item -include @var{file}
2524 Process @var{file} as input before processing the regular input file.
2525 In effect, the contents of @var{file} are compiled first. Any @samp{-D}
2526 and @samp{-U} options on the command line are always processed before
2527 @samp{-include @var{file}}, regardless of the order in which they are
2528 written. All the @samp{-include} and @samp{-imacros} options are
2529 processed in the order in which they are written.
2531 @item -imacros @var{file}
2532 Process @var{file} as input, discarding the resulting output, before
2533 processing the regular input file. Because the output generated from
2534 @var{file} is discarded, the only effect of @samp{-imacros @var{file}}
2535 is to make the macros defined in @var{file} available for use in the
2538 Any @samp{-D} and @samp{-U} options on the command line are always
2539 processed before @samp{-imacros @var{file}}, regardless of the order in
2540 which they are written. All the @samp{-include} and @samp{-imacros}
2541 options are processed in the order in which they are written.
2543 @item -idirafter @var{dir}
2544 @cindex second include path
2545 Add the directory @var{dir} to the second include path. The directories
2546 on the second include path are searched when a header file is not found
2547 in any of the directories in the main include path (the one that
2550 @item -iprefix @var{prefix}
2551 Specify @var{prefix} as the prefix for subsequent @samp{-iwithprefix}
2554 @item -iwithprefix @var{dir}
2555 Add a directory to the second include path. The directory's name is
2556 made by concatenating @var{prefix} and @var{dir}, where @var{prefix} was
2557 specified previously with @samp{-iprefix}. If you have not specified a
2558 prefix yet, the directory containing the installed passes of the
2559 compiler is used as the default.
2561 @item -iwithprefixbefore @var{dir}
2562 Add a directory to the main include path. The directory's name is made
2563 by concatenating @var{prefix} and @var{dir}, as in the case of
2564 @samp{-iwithprefix}.
2566 @item -isystem @var{dir}
2567 Add a directory to the beginning of the second include path, marking it
2568 as a system directory, so that it gets the same special treatment as
2569 is applied to the standard system directories.
2572 Do not search the standard system directories for header files. Only
2573 the directories you have specified with @samp{-I} options (and the
2574 current directory, if appropriate) are searched. @xref{Directory
2575 Options}, for information on @samp{-I}.
2577 By using both @samp{-nostdinc} and @samp{-I-}, you can limit the include-file
2578 search path to only those directories you specify explicitly.
2581 Do not predefine any nonstandard macros. (Including architecture flags).
2584 Run only the C preprocessor. Preprocess all the C source files
2585 specified and output the results to standard output or to the
2586 specified output file.
2589 Tell the preprocessor not to discard comments. Used with the
2593 Tell the preprocessor not to generate @samp{#line} directives.
2594 Used with the @samp{-E} option.
2597 @cindex dependencies, make
2599 Tell the preprocessor to output a rule suitable for @code{make}
2600 describing the dependencies of each object file. For each source file,
2601 the preprocessor outputs one @code{make}-rule whose target is the object
2602 file name for that source file and whose dependencies are all the
2603 @code{#include} header files it uses. This rule may be a single line or
2604 may be continued with @samp{\}-newline if it is long. The list of rules
2605 is printed on standard output instead of the preprocessed C program.
2607 @samp{-M} implies @samp{-E}.
2609 Another way to specify output of a @code{make} rule is by setting
2610 the environment variable @code{DEPENDENCIES_OUTPUT} (@pxref{Environment
2614 Like @samp{-M} but the output mentions only the user header files
2615 included with @samp{#include "@var{file}"}. System header files
2616 included with @samp{#include <@var{file}>} are omitted.
2619 Like @samp{-M} but the dependency information is written to a file made by
2620 replacing ".c" with ".d" at the end of the input file names.
2621 This is in addition to compiling the file as specified---@samp{-MD} does
2622 not inhibit ordinary compilation the way @samp{-M} does.
2624 In Mach, you can use the utility @code{md} to merge multiple dependency
2625 files into a single dependency file suitable for using with the @samp{make}
2629 Like @samp{-MD} except mention only user header files, not system
2633 Treat missing header files as generated files and assume they live in the
2634 same directory as the source file. If you specify @samp{-MG}, you
2635 must also specify either @samp{-M} or @samp{-MM}. @samp{-MG} is not
2636 supported with @samp{-MD} or @samp{-MMD}.
2639 Print the name of each header file used, in addition to other normal
2642 @item -A@var{question}(@var{answer})
2643 Assert the answer @var{answer} for @var{question}, in case it is tested
2644 with a preprocessing conditional such as @samp{#if
2645 #@var{question}(@var{answer})}. @samp{-A-} disables the standard
2646 assertions that normally describe the target machine.
2649 Define macro @var{macro} with the string @samp{1} as its definition.
2651 @item -D@var{macro}=@var{defn}
2652 Define macro @var{macro} as @var{defn}. All instances of @samp{-D} on
2653 the command line are processed before any @samp{-U} options.
2656 Undefine macro @var{macro}. @samp{-U} options are evaluated after all
2657 @samp{-D} options, but before any @samp{-include} and @samp{-imacros}
2661 Tell the preprocessor to output only a list of the macro definitions
2662 that are in effect at the end of preprocessing. Used with the @samp{-E}
2666 Tell the preprocessing to pass all macro definitions into the output, in
2667 their proper sequence in the rest of the output.
2670 Like @samp{-dD} except that the macro arguments and contents are omitted.
2671 Only @samp{#define @var{name}} is included in the output.
2674 Support ANSI C trigraphs. The @samp{-ansi} option also has this effect.
2676 @item -Wp,@var{option}
2677 Pass @var{option} as an option to the preprocessor. If @var{option}
2678 contains commas, it is split into multiple options at the commas.
2681 @node Assembler Options
2682 @section Passing Options to the Assembler
2684 @c prevent bad page break with this line
2685 You can pass options to the assembler.
2688 @item -Wa,@var{option}
2689 Pass @var{option} as an option to the assembler. If @var{option}
2690 contains commas, it is split into multiple options at the commas.
2694 @section Options for Linking
2695 @cindex link options
2696 @cindex options, linking
2698 These options come into play when the compiler links object files into
2699 an executable output file. They are meaningless if the compiler is
2700 not doing a link step.
2704 @item @var{object-file-name}
2705 A file name that does not end in a special recognized suffix is
2706 considered to name an object file or library. (Object files are
2707 distinguished from libraries by the linker according to the file
2708 contents.) If linking is done, these object files are used as input
2714 If any of these options is used, then the linker is not run, and
2715 object file names should not be used as arguments. @xref{Overall
2719 @item -l@var{library}
2720 Search the library named @var{library} when linking.
2722 It makes a difference where in the command you write this option; the
2723 linker searches processes libraries and object files in the order they
2724 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
2725 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
2726 to functions in @samp{z}, those functions may not be loaded.
2728 The linker searches a standard list of directories for the library,
2729 which is actually a file named @file{lib@var{library}.a}. The linker
2730 then uses this file as if it had been specified precisely by name.
2732 The directories searched include several standard system directories
2733 plus any that you specify with @samp{-L}.
2735 Normally the files found this way are library files---archive files
2736 whose members are object files. The linker handles an archive file by
2737 scanning through it for members which define symbols that have so far
2738 been referenced but not defined. But if the file that is found is an
2739 ordinary object file, it is linked in the usual fashion. The only
2740 difference between using an @samp{-l} option and specifying a file name
2741 is that @samp{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
2742 and searches several directories.
2745 You need this special case of the @samp{-l} option in order to
2746 link an Objective C program.
2749 Do not use the standard system startup files when linking.
2750 The standard system libraries are used normally, unless @code{-nostdlib}
2751 or @code{-nodefaultlibs} is used.
2753 @item -nodefaultlibs
2754 Do not use the standard system libraries when linking.
2755 Only the libraries you specify will be passed to the linker.
2756 The standard startup files are used normally, unless @code{-nostartfiles}
2757 is used. The compiler may generate calls to memcmp, memset, and memcpy
2758 for System V (and ANSI C) environments or to bcopy and bzero for
2759 BSD environments. These entries are usually resolved by entries in
2760 libc. These entry points should be supplied through some other
2761 mechanism when this option is specified.
2764 Do not use the standard system startup files or libraries when linking.
2765 No startup files and only the libraries you specify will be passed to
2766 the linker. The compiler may generate calls to memcmp, memset, and memcpy
2767 for System V (and ANSI C) environments or to bcopy and bzero for
2768 BSD environments. These entries are usually resolved by entries in
2769 libc. These entry points should be supplied through some other
2770 mechanism when this option is specified.
2772 @cindex @code{-lgcc}, use with @code{-nostdlib}
2773 @cindex @code{-nostdlib} and unresolved references
2774 @cindex unresolved references and @code{-nostdlib}
2775 @cindex @code{-lgcc}, use with @code{-nodefaultlibs}
2776 @cindex @code{-nodefaultlibs} and unresolved references
2777 @cindex unresolved references and @code{-nodefaultlibs}
2778 One of the standard libraries bypassed by @samp{-nostdlib} and
2779 @samp{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
2780 that GNU CC uses to overcome shortcomings of particular machines, or special
2781 needs for some languages.
2783 (@xref{Interface,,Interfacing to GNU CC Output}, for more discussion of
2787 (@xref{Interface,,Interfacing to GNU CC Output,gcc.info,Porting GNU CC},
2788 for more discussion of @file{libgcc.a}.)
2790 In most cases, you need @file{libgcc.a} even when you want to avoid
2791 other standard libraries. In other words, when you specify @samp{-nostdlib}
2792 or @samp{-nodefaultlibs} you should usually specify @samp{-lgcc} as well.
2793 This ensures that you have no unresolved references to internal GNU CC
2794 library subroutines. (For example, @samp{__main}, used to ensure C++
2795 constructors will be called; @pxref{Collect2,,@code{collect2}}.)
2798 Remove all symbol table and relocation information from the executable.
2801 On systems that support dynamic linking, this prevents linking with the shared
2802 libraries. On other systems, this option has no effect.
2805 Produce a shared object which can then be linked with other objects to
2806 form an executable. Not all systems support this option. You must
2807 also specify @samp{-fpic} or @samp{-fPIC} on some systems when
2808 you specify this option.
2811 Bind references to global symbols when building a shared object. Warn
2812 about any unresolved references (unless overridden by the link editor
2813 option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support
2816 @item -Xlinker @var{option}
2817 Pass @var{option} as an option to the linker. You can use this to
2818 supply system-specific linker options which GNU CC does not know how to
2821 If you want to pass an option that takes an argument, you must use
2822 @samp{-Xlinker} twice, once for the option and once for the argument.
2823 For example, to pass @samp{-assert definitions}, you must write
2824 @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write
2825 @samp{-Xlinker "-assert definitions"}, because this passes the entire
2826 string as a single argument, which is not what the linker expects.
2828 @item -Wl,@var{option}
2829 Pass @var{option} as an option to the linker. If @var{option} contains
2830 commas, it is split into multiple options at the commas.
2832 @item -u @var{symbol}
2833 Pretend the symbol @var{symbol} is undefined, to force linking of
2834 library modules to define it. You can use @samp{-u} multiple times with
2835 different symbols to force loading of additional library modules.
2838 @node Directory Options
2839 @section Options for Directory Search
2840 @cindex directory options
2841 @cindex options, directory search
2844 These options specify directories to search for header files, for
2845 libraries and for parts of the compiler:
2849 Add the directory @var{dir} to the head of the list of directories to be
2850 searched for header files. This can be used to override a system header
2851 file, substituting your own version, since these directories are
2852 searched before the system header file directories. If you use more
2853 than one @samp{-I} option, the directories are scanned in left-to-right
2854 order; the standard system directories come after.
2857 Any directories you specify with @samp{-I} options before the @samp{-I-}
2858 option are searched only for the case of @samp{#include "@var{file}"};
2859 they are not searched for @samp{#include <@var{file}>}.
2861 If additional directories are specified with @samp{-I} options after
2862 the @samp{-I-}, these directories are searched for all @samp{#include}
2863 directives. (Ordinarily @emph{all} @samp{-I} directories are used
2866 In addition, the @samp{-I-} option inhibits the use of the current
2867 directory (where the current input file came from) as the first search
2868 directory for @samp{#include "@var{file}"}. There is no way to
2869 override this effect of @samp{-I-}. With @samp{-I.} you can specify
2870 searching the directory which was current when the compiler was
2871 invoked. That is not exactly the same as what the preprocessor does
2872 by default, but it is often satisfactory.
2874 @samp{-I-} does not inhibit the use of the standard system directories
2875 for header files. Thus, @samp{-I-} and @samp{-nostdinc} are
2879 Add directory @var{dir} to the list of directories to be searched
2882 @item -B@var{prefix}
2883 This option specifies where to find the executables, libraries,
2884 include files, and data files of the compiler itself.
2886 The compiler driver program runs one or more of the subprograms
2887 @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries
2888 @var{prefix} as a prefix for each program it tries to run, both with and
2889 without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}).
2891 For each subprogram to be run, the compiler driver first tries the
2892 @samp{-B} prefix, if any. If that name is not found, or if @samp{-B}
2893 was not specified, the driver tries two standard prefixes, which are
2894 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc-lib/}. If neither of
2895 those results in a file name that is found, the unmodified program
2896 name is searched for using the directories specified in your
2897 @samp{PATH} environment variable.
2899 @samp{-B} prefixes that effectively specify directory names also apply
2900 to libraries in the linker, because the compiler translates these
2901 options into @samp{-L} options for the linker. They also apply to
2902 includes files in the preprocessor, because the compiler translates these
2903 options into @samp{-isystem} options for the preprocessor. In this case,
2904 the compiler appends @samp{include} to the prefix.
2906 The run-time support file @file{libgcc.a} can also be searched for using
2907 the @samp{-B} prefix, if needed. If it is not found there, the two
2908 standard prefixes above are tried, and that is all. The file is left
2909 out of the link if it is not found by those means.
2911 Another way to specify a prefix much like the @samp{-B} prefix is to use
2912 the environment variable @code{GCC_EXEC_PREFIX}. @xref{Environment
2915 @item -specs=@var{file}
2916 Process @var{file} after the compiler reads in the standard @file{specs}
2917 file, in order to override the defaults that the @file{gcc} driver
2918 program uses when determining what switches to pass to @file{cc1},
2919 @file{cc1plus}, @file{as}, @file{ld}, etc. More than one
2920 @samp{-specs=}@var{file} can be specified on the command line, and they
2921 are processed in order, from left to right.
2924 @node Target Options
2925 @section Specifying Target Machine and Compiler Version
2926 @cindex target options
2927 @cindex cross compiling
2928 @cindex specifying machine version
2929 @cindex specifying compiler version and target machine
2930 @cindex compiler version, specifying
2931 @cindex target machine, specifying
2933 By default, GNU CC compiles code for the same type of machine that you
2934 are using. However, it can also be installed as a cross-compiler, to
2935 compile for some other type of machine. In fact, several different
2936 configurations of GNU CC, for different target machines, can be
2937 installed side by side. Then you specify which one to use with the
2940 In addition, older and newer versions of GNU CC can be installed side
2941 by side. One of them (probably the newest) will be the default, but
2942 you may sometimes wish to use another.
2945 @item -b @var{machine}
2946 The argument @var{machine} specifies the target machine for compilation.
2947 This is useful when you have installed GNU CC as a cross-compiler.
2949 The value to use for @var{machine} is the same as was specified as the
2950 machine type when configuring GNU CC as a cross-compiler. For
2951 example, if a cross-compiler was configured with @samp{configure
2952 i386v}, meaning to compile for an 80386 running System V, then you
2953 would specify @samp{-b i386v} to run that cross compiler.
2955 When you do not specify @samp{-b}, it normally means to compile for
2956 the same type of machine that you are using.
2958 @item -V @var{version}
2959 The argument @var{version} specifies which version of GNU CC to run.
2960 This is useful when multiple versions are installed. For example,
2961 @var{version} might be @samp{2.0}, meaning to run GNU CC version 2.0.
2963 The default version, when you do not specify @samp{-V}, is the last
2964 version of GNU CC that you installed.
2967 The @samp{-b} and @samp{-V} options actually work by controlling part of
2968 the file name used for the executable files and libraries used for
2969 compilation. A given version of GNU CC, for a given target machine, is
2970 normally kept in the directory @file{/usr/local/lib/gcc-lib/@var{machine}/@var{version}}.@refill
2972 Thus, sites can customize the effect of @samp{-b} or @samp{-V} either by
2973 changing the names of these directories or adding alternate names (or
2974 symbolic links). If in directory @file{/usr/local/lib/gcc-lib/} the
2975 file @file{80386} is a link to the file @file{i386v}, then @samp{-b
2976 80386} becomes an alias for @samp{-b i386v}.
2978 In one respect, the @samp{-b} or @samp{-V} do not completely change
2979 to a different compiler: the top-level driver program @code{gcc}
2980 that you originally invoked continues to run and invoke the other
2981 executables (preprocessor, compiler per se, assembler and linker)
2982 that do the real work. However, since no real work is done in the
2983 driver program, it usually does not matter that the driver program
2984 in use is not the one for the specified target and version.
2986 The only way that the driver program depends on the target machine is
2987 in the parsing and handling of special machine-specific options.
2988 However, this is controlled by a file which is found, along with the
2989 other executables, in the directory for the specified version and
2990 target machine. As a result, a single installed driver program adapts
2991 to any specified target machine and compiler version.
2993 The driver program executable does control one significant thing,
2994 however: the default version and target machine. Therefore, you can
2995 install different instances of the driver program, compiled for
2996 different targets or versions, under different names.
2998 For example, if the driver for version 2.0 is installed as @code{ogcc}
2999 and that for version 2.1 is installed as @code{gcc}, then the command
3000 @code{gcc} will use version 2.1 by default, while @code{ogcc} will use
3001 2.0 by default. However, you can choose either version with either
3002 command with the @samp{-V} option.
3004 @node Submodel Options
3005 @section Hardware Models and Configurations
3006 @cindex submodel options
3007 @cindex specifying hardware config
3008 @cindex hardware models and configurations, specifying
3009 @cindex machine dependent options
3011 Earlier we discussed the standard option @samp{-b} which chooses among
3012 different installed compilers for completely different target
3013 machines, such as Vax vs. 68000 vs. 80386.
3015 In addition, each of these target machine types can have its own
3016 special options, starting with @samp{-m}, to choose among various
3017 hardware models or configurations---for example, 68010 vs 68020,
3018 floating coprocessor or none. A single installed version of the
3019 compiler can compile for any model or configuration, according to the
3022 Some configurations of the compiler also support additional special
3023 options, usually for compatibility with other compilers on the same
3027 These options are defined by the macro @code{TARGET_SWITCHES} in the
3028 machine description. The default for the options is also defined by
3029 that macro, which enables you to change the defaults.
3044 * RS/6000 and PowerPC Options::
3049 * Intel 960 Options::
3050 * DEC Alpha Options::
3054 * System V Options::
3059 @node M680x0 Options
3060 @subsection M680x0 Options
3061 @cindex M680x0 options
3063 These are the @samp{-m} options defined for the 68000 series. The default
3064 values for these options depends on which style of 68000 was selected when
3065 the compiler was configured; the defaults for the most common choices are
3071 Generate output for a 68000. This is the default
3072 when the compiler is configured for 68000-based systems.
3074 Use this option for microcontrollers with a 68000 or EC000 core,
3075 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
3079 Generate output for a 68020. This is the default
3080 when the compiler is configured for 68020-based systems.
3083 Generate output containing 68881 instructions for floating point.
3084 This is the default for most 68020 systems unless @samp{-nfp} was
3085 specified when the compiler was configured.
3088 Generate output for a 68030. This is the default when the compiler is
3089 configured for 68030-based systems.
3092 Generate output for a 68040. This is the default when the compiler is
3093 configured for 68040-based systems.
3095 This option inhibits the use of 68881/68882 instructions that have to be
3096 emulated by software on the 68040. Use this option if your 68040 does not
3097 have code to emulate those instructions.
3100 Generate output for a 68060. This is the default when the compiler is
3101 configured for 68060-based systems.
3103 This option inhibits the use of 68020 and 68881/68882 instructions that
3104 have to be emulated by software on the 68060. Use this option if your 68060
3105 does not have code to emulate those instructions.
3108 Generate output for a CPU32. This is the default
3109 when the compiler is configured for CPU32-based systems.
3111 Use this option for microcontrollers with a
3112 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
3113 68336, 68340, 68341, 68349 and 68360.
3116 Generate output for a 520X "coldfire" family cpu. This is the default
3117 when the compiler is configured for 520X-based systems.
3119 Use this option for microcontroller with a 5200 core, including
3120 the MCF5202, MCF5203, MCF5204 and MCF5202.
3124 Generate output for a 68040, without using any of the new instructions.
3125 This results in code which can run relatively efficiently on either a
3126 68020/68881 or a 68030 or a 68040. The generated code does use the
3127 68881 instructions that are emulated on the 68040.
3130 Generate output for a 68060, without using any of the new instructions.
3131 This results in code which can run relatively efficiently on either a
3132 68020/68881 or a 68030 or a 68040. The generated code does use the
3133 68881 instructions that are emulated on the 68060.
3136 Generate output containing Sun FPA instructions for floating point.
3139 Generate output containing library calls for floating point.
3140 @strong{Warning:} the requisite libraries are not available for all m68k
3141 targets. Normally the facilities of the machine's usual C compiler are
3142 used, but this can't be done directly in cross-compilation. You must
3143 make your own arrangements to provide suitable library functions for
3144 cross-compilation. The embedded targets @samp{m68k-*-aout} and
3145 @samp{m68k-*-coff} do provide software floating point support.
3148 Consider type @code{int} to be 16 bits wide, like @code{short int}.
3151 Do not use the bit-field instructions. The @samp{-m68000}, @samp{-mcpu32}
3152 and @samp{-m5200} options imply @w{@samp{-mnobitfield}}.
3155 Do use the bit-field instructions. The @samp{-m68020} option implies
3156 @samp{-mbitfield}. This is the default if you use a configuration
3157 designed for a 68020.
3160 Use a different function-calling convention, in which functions
3161 that take a fixed number of arguments return with the @code{rtd}
3162 instruction, which pops their arguments while returning. This
3163 saves one instruction in the caller since there is no need to pop
3164 the arguments there.
3166 This calling convention is incompatible with the one normally
3167 used on Unix, so you cannot use it if you need to call libraries
3168 compiled with the Unix compiler.
3170 Also, you must provide function prototypes for all functions that
3171 take variable numbers of arguments (including @code{printf});
3172 otherwise incorrect code will be generated for calls to those
3175 In addition, seriously incorrect code will result if you call a
3176 function with too many arguments. (Normally, extra arguments are
3177 harmlessly ignored.)
3179 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
3180 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
3183 @itemx -mno-align-int
3184 Control whether GNU CC aligns @code{int}, @code{long}, @code{long long},
3185 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
3186 boundary (@samp{-malign-int}) or a 16-bit boundary (@samp{-mno-align-int}).
3187 Aligning variables on 32-bit boundaries produces code that runs somewhat
3188 faster on processors with 32-bit busses at the expense of more memory.
3190 @strong{Warning:} if you use the @samp{-malign-int} switch, GNU CC will
3191 align structures containing the above types differently than
3192 most published application binary interface specifications for the m68k.
3197 @subsection VAX Options
3200 These @samp{-m} options are defined for the Vax:
3204 Do not output certain jump instructions (@code{aobleq} and so on)
3205 that the Unix assembler for the Vax cannot handle across long
3209 Do output those jump instructions, on the assumption that you
3210 will assemble with the GNU assembler.
3213 Output code for g-format floating point numbers instead of d-format.
3217 @subsection SPARC Options
3218 @cindex SPARC options
3220 These @samp{-m} switches are supported on the SPARC:
3225 Specify @samp{-mapp-regs} to generate output using the global registers
3226 2 through 4, which the SPARC SVR4 ABI reserves for applications. This
3229 To be fully SVR4 ABI compliant at the cost of some performance loss,
3230 specify @samp{-mno-app-regs}. You should compile libraries and system
3231 software with this option.
3235 Generate output containing floating point instructions. This is the
3240 Generate output containing library calls for floating point.
3241 @strong{Warning:} the requisite libraries are not available for all SPARC
3242 targets. Normally the facilities of the machine's usual C compiler are
3243 used, but this cannot be done directly in cross-compilation. You must make
3244 your own arrangements to provide suitable library functions for
3245 cross-compilation. The embedded targets @samp{sparc-*-aout} and
3246 @samp{sparclite-*-*} do provide software floating point support.
3248 @samp{-msoft-float} changes the calling convention in the output file;
3249 therefore, it is only useful if you compile @emph{all} of a program with
3250 this option. In particular, you need to compile @file{libgcc.a}, the
3251 library that comes with GNU CC, with @samp{-msoft-float} in order for
3254 @item -mhard-quad-float
3255 Generate output containing quad-word (long double) floating point
3258 @item -msoft-quad-float
3259 Generate output containing library calls for quad-word (long double)
3260 floating point instructions. The functions called are those specified
3261 in the SPARC ABI. This is the default.
3263 As of this writing, there are no sparc implementations that have hardware
3264 support for the quad-word floating point instructions. They all invoke
3265 a trap handler for one of these instructions, and then the trap handler
3266 emulates the effect of the instruction. Because of the trap handler overhead,
3267 this is much slower than calling the ABI library routines. Thus the
3268 @samp{-msoft-quad-float} option is the default.
3272 With @samp{-mepilogue} (the default), the compiler always emits code for
3273 function exit at the end of each function. Any function exit in
3274 the middle of the function (such as a return statement in C) will
3275 generate a jump to the exit code at the end of the function.
3277 With @samp{-mno-epilogue}, the compiler tries to emit exit code inline
3278 at every function exit.
3282 With @samp{-mflat}, the compiler does not generate save/restore instructions
3283 and will use a "flat" or single register window calling convention.
3284 This model uses %i7 as the frame pointer and is compatible with the normal
3285 register window model. Code from either may be intermixed.
3286 The local registers and the input registers (0-5) are still treated as
3287 "call saved" registers and will be saved on the stack as necessary.
3289 With @samp{-mno-flat} (the default), the compiler emits save/restore
3290 instructions (except for leaf functions) and is the normal mode of operation.
3292 @item -mno-unaligned-doubles
3293 @itemx -munaligned-doubles
3294 Assume that doubles have 8 byte alignment. This is the default.
3296 With @samp{-munaligned-doubles}, GNU CC assumes that doubles have 8 byte
3297 alignment only if they are contained in another type, or if they have an
3298 absolute address. Otherwise, it assumes they have 4 byte alignment.
3299 Specifying this option avoids some rare compatibility problems with code
3300 generated by other compilers. It is not the default because it results
3301 in a performance loss, especially for floating point code.
3305 These two options select variations on the SPARC architecture.
3307 By default (unless specifically configured for the Fujitsu SPARClite),
3308 GCC generates code for the v7 variant of the SPARC architecture.
3310 @samp{-mv8} will give you SPARC v8 code. The only difference from v7
3311 code is that the compiler emits the integer multiply and integer
3312 divide instructions which exist in SPARC v8 but not in SPARC v7.
3314 @samp{-msparclite} will give you SPARClite code. This adds the integer
3315 multiply, integer divide step and scan (@code{ffs}) instructions which
3316 exist in SPARClite but not in SPARC v7.
3318 These options are deprecated and will be deleted in GNU CC 2.9.
3319 They have been replaced with @samp{-mcpu=xxx}.
3323 These two options select the processor for which the code is optimised.
3325 With @samp{-mcypress} (the default), the compiler optimizes code for the
3326 Cypress CY7C602 chip, as used in the SparcStation/SparcServer 3xx series.
3327 This is also appropriate for the older SparcStation 1, 2, IPX etc.
3329 With @samp{-msupersparc} the compiler optimizes code for the SuperSparc cpu, as
3330 used in the SparcStation 10, 1000 and 2000 series. This flag also enables use
3331 of the full SPARC v8 instruction set.
3333 These options are deprecated and will be deleted in GNU CC 2.9.
3334 They have been replaced with @samp{-mcpu=xxx}.
3336 @item -mcpu=@var{cpu_type}
3337 Set the instruction set, register set, and instruction scheduling parameters
3338 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
3339 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite},
3340 @samp{f930}, @samp{f934}, @samp{sparclet}, @samp{tsc701}, @samp{v9}, and
3343 Default instruction scheduling parameters are used for values that select
3344 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
3345 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
3347 Here is a list of each supported architecture and their supported
3353 sparclite: f930, f934
3358 @item -mtune=@var{cpu_type}
3359 Set the instruction scheduling parameters for machine type
3360 @var{cpu_type}, but do not set the instruction set or register set that the
3361 option @samp{-mcpu=}@var{cpu_type} would.
3363 The same values for @samp{-mcpu=}@var{cpu_type} are used for
3364 @samp{-mtune=}@*@var{cpu_type}, though the only useful values are those that
3365 select a particular cpu implementation: @samp{cypress}, @samp{supersparc},
3366 @samp{f930}, @samp{f934}, @samp{tsc701}, @samp{ultrasparc}.
3368 @item -malign-loops=@var{num}
3369 Align loops to a 2 raised to a @var{num} byte boundary. If
3370 @samp{-malign-loops} is not specified, the default is 2.
3372 @item -malign-jumps=@var{num}
3373 Align instructions that are only jumped to to a 2 raised to a @var{num}
3374 byte boundary. If @samp{-malign-jumps} is not specified, the default is 2.
3376 @item -malign-functions=@var{num}
3377 Align the start of functions to a 2 raised to @var{num} byte boundary.
3378 If @samp{-malign-functions} is not specified, the default is 2 if compiling
3379 for 32 bit sparc, and 5 if compiling for 64 bit sparc.
3383 These @samp{-m} switches are supported in addition to the above
3384 on the SPARCLET processor.
3387 @item -mlittle-endian
3388 Generate code for a processor running in little-endian mode.
3391 Treat register @code{%g0} as a normal register.
3392 GCC will continue to clobber it as necessary but will not assume
3393 it always reads as 0.
3395 @item -mbroken-saverestore
3396 Generate code that does not use non-trivial forms of the @code{save} and
3397 @code{restore} instructions. Early versions of the SPARCLET processor do
3398 not correctly handle @code{save} and @code{restore} instructions used with
3399 arguments. They correctly handle them used without arguments. A @code{save}
3400 instruction used without arguments increments the current window pointer
3401 but does not allocate a new stack frame. It is assumed that the window
3402 overflow trap handler will properly handle this case as will interrupt
3406 These @samp{-m} switches are supported in addition to the above
3407 on SPARC V9 processors in 64 bit environments.
3410 @item -mlittle-endian
3411 Generate code for a processor running in little-endian mode.
3415 Generate code for a 32 bit or 64 bit environment.
3416 The 32 bit environment sets int, long and pointer to 32 bits.
3417 The 64 bit environment sets int to 32 bits and long and pointer
3420 @item -mcmodel=medlow
3421 Generate code for the Medium/Low code model: the program must be linked
3422 in the low 32 bits of the address space. Pointers are 64 bits.
3423 Programs can be statically or dynamically linked.
3425 @item -mcmodel=medmid
3426 Generate code for the Medium/Middle code model: the program must be linked
3427 in the low 44 bits of the address space, the text segment must be less than
3428 2G bytes, and data segment must be within 2G of the text segment.
3429 Pointers are 64 bits.
3431 @item -mcmodel=medany
3432 Generate code for the Medium/Anywhere code model: the program may be linked
3433 anywhere in the address space, the text segment must be less than
3434 2G bytes, and data segment must be within 2G of the text segment.
3435 Pointers are 64 bits.
3437 @item -mcmodel=embmedany
3438 Generate code for the Medium/Anywhere code model for embedded systems:
3439 assume a 32 bit text and a 32 bit data segment, both starting anywhere
3440 (determined at link time). Register %g4 points to the base of the
3441 data segment. Pointers still 64 bits.
3442 Programs are statically linked, PIC is not supported.
3445 @itemx -mno-stack-bias
3446 With @samp{-mstack-bias}, GNU CC assumes that the stack pointer, and
3447 frame pointer if present, are offset by -2047 which must be added back
3448 when making stack frame references.
3449 Otherwise, assume no such offset is present.
3452 @node Convex Options
3453 @subsection Convex Options
3454 @cindex Convex options
3456 These @samp{-m} options are defined for Convex:
3460 Generate output for C1. The code will run on any Convex machine.
3461 The preprocessor symbol @code{__convex__c1__} is defined.
3464 Generate output for C2. Uses instructions not available on C1.
3465 Scheduling and other optimizations are chosen for max performance on C2.
3466 The preprocessor symbol @code{__convex_c2__} is defined.
3469 Generate output for C32xx. Uses instructions not available on C1.
3470 Scheduling and other optimizations are chosen for max performance on C32.
3471 The preprocessor symbol @code{__convex_c32__} is defined.
3474 Generate output for C34xx. Uses instructions not available on C1.
3475 Scheduling and other optimizations are chosen for max performance on C34.
3476 The preprocessor symbol @code{__convex_c34__} is defined.
3479 Generate output for C38xx. Uses instructions not available on C1.
3480 Scheduling and other optimizations are chosen for max performance on C38.
3481 The preprocessor symbol @code{__convex_c38__} is defined.
3484 Generate code which puts an argument count in the word preceding each
3485 argument list. This is compatible with regular CC, and a few programs
3486 may need the argument count word. GDB and other source-level debuggers
3487 do not need it; this info is in the symbol table.
3490 Omit the argument count word. This is the default.
3492 @item -mvolatile-cache
3493 Allow volatile references to be cached. This is the default.
3495 @item -mvolatile-nocache
3496 Volatile references bypass the data cache, going all the way to memory.
3497 This is only needed for multi-processor code that does not use standard
3498 synchronization instructions. Making non-volatile references to volatile
3499 locations will not necessarily work.
3502 Type long is 32 bits, the same as type int. This is the default.
3505 Type long is 64 bits, the same as type long long. This option is useless,
3506 because no library support exists for it.
3509 @node AMD29K Options
3510 @subsection AMD29K Options
3511 @cindex AMD29K options
3513 These @samp{-m} options are defined for the AMD Am29000:
3518 @cindex DW bit (29k)
3519 Generate code that assumes the @code{DW} bit is set, i.e., that byte and
3520 halfword operations are directly supported by the hardware. This is the
3525 Generate code that assumes the @code{DW} bit is not set.
3529 @cindex byte writes (29k)
3530 Generate code that assumes the system supports byte and halfword write
3531 operations. This is the default.
3535 Generate code that assumes the systems does not support byte and
3536 halfword write operations. @samp{-mnbw} implies @samp{-mndw}.
3540 @cindex memory model (29k)
3541 Use a small memory model that assumes that all function addresses are
3542 either within a single 256 KB segment or at an absolute address of less
3543 than 256k. This allows the @code{call} instruction to be used instead
3544 of a @code{const}, @code{consth}, @code{calli} sequence.
3548 Use the normal memory model: Generate @code{call} instructions only when
3549 calling functions in the same file and @code{calli} instructions
3550 otherwise. This works if each file occupies less than 256 KB but allows
3551 the entire executable to be larger than 256 KB. This is the default.
3554 Always use @code{calli} instructions. Specify this option if you expect
3555 a single file to compile into more than 256 KB of code.
3559 @cindex processor selection (29k)
3560 Generate code for the Am29050.
3564 Generate code for the Am29000. This is the default.
3566 @item -mkernel-registers
3567 @kindex -mkernel-registers
3568 @cindex kernel and user registers (29k)
3569 Generate references to registers @code{gr64-gr95} instead of to
3570 registers @code{gr96-gr127}. This option can be used when compiling
3571 kernel code that wants a set of global registers disjoint from that used
3574 Note that when this option is used, register names in @samp{-f} flags
3575 must use the normal, user-mode, names.
3577 @item -muser-registers
3578 @kindex -muser-registers
3579 Use the normal set of global registers, @code{gr96-gr127}. This is the
3583 @itemx -mno-stack-check
3584 @kindex -mstack-check
3585 @cindex stack checks (29k)
3586 Insert (or do not insert) a call to @code{__msp_check} after each stack
3587 adjustment. This is often used for kernel code.
3590 @itemx -mno-storem-bug
3591 @kindex -mstorem-bug
3592 @cindex storem bug (29k)
3593 @samp{-mstorem-bug} handles 29k processors which cannot handle the
3594 separation of a mtsrim insn and a storem instruction (most 29000 chips
3595 to date, but not the 29050).
3597 @item -mno-reuse-arg-regs
3598 @itemx -mreuse-arg-regs
3599 @kindex -mreuse-arg-regs
3600 @samp{-mno-reuse-arg-regs} tells the compiler to only use incoming argument
3601 registers for copying out arguments. This helps detect calling a function
3602 with fewer arguments than it was declared with.
3604 @item -mno-impure-text
3605 @itemx -mimpure-text
3606 @kindex -mimpure-text
3607 @samp{-mimpure-text}, used in addition to @samp{-shared}, tells the compiler to
3608 not pass @samp{-assert pure-text} to the linker when linking a shared object.
3611 @kindex -msoft-float
3612 Generate output containing library calls for floating point.
3613 @strong{Warning:} the requisite libraries are not part of GNU CC.
3614 Normally the facilities of the machine's usual C compiler are used, but
3615 this can't be done directly in cross-compilation. You must make your
3616 own arrangements to provide suitable library functions for
3621 @subsection ARM Options
3624 These @samp{-m} options are defined for Advanced RISC Machines (ARM)
3629 @kindex -mapcs-frame
3630 Generate a stack frame that is compliant with the ARM Procedure Call
3631 Standard for all functions, even if this is not strictly necessary for
3632 correct execution of the code. Specifying @samp{-fomit-frame-pointer}
3633 with this option will cause the stack frames not to be generated for
3634 leaf functions. The default is @samp{-mno-apcs-frame}.
3638 This is a synonym for @samp{-mapcs-frame}.
3642 Generate code for a processor running with a 26-bit program counter,
3643 and conforming to the function calling standards for the APCS 26-bit
3644 option. This option replaces the @samp{-m2} and @samp{-m3} options
3645 of previous releases of the compiler.
3649 Generate code for a processor running with a 32-bit program counter,
3650 and conforming to the function calling standards for the APCS 32-bit
3651 option. This option replaces the @samp{-m6} option of previous releases
3654 @item -mapcs-stack-check
3655 @kindex -mapcs-stack-check
3656 @kindex -mno-apcs-stack-check
3657 Generate code to check the amount of stack space available upon entry to
3658 every function (that actually uses some stack space). If there is
3659 insufficient space available then either the function
3660 @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be
3661 called, depending upon the amount of stack space required. The run time
3662 system is required to provide these functions. The default is
3663 @samp{-mno-apcs-stack-check}, since this produces smaller code.
3666 @kindex -mapcs-float
3667 @kindex -mno-apcs-float
3668 Pass floating point arguments using the float point registers. This is
3669 one of the variants of the APCS. This option is reccommended if the
3670 target hardware has a floating point unit or if a lot of floating point
3671 arithmetic is going to be performed by the code. The default is
3672 @samp{-mno-apcs-float}, since integer only code is slightly increased in
3673 size if @samp{-mapcs-float} is used.
3675 @item -mapcs-reentrant
3676 @kindex -mapcs-reentrant
3677 @kindex -mno-apcs-reentrant
3678 Generate reentrant, position independent code. This is the equivalent
3679 to specifying the @samp{-fpic} option. The default is
3680 @samp{-mno-apcs-reentrant}.
3682 @item -mthumb-interwork
3683 @kindex -mthumb-interwork
3684 @kindex -mno-thumb-interwork
3685 Generate code which supports calling between the ARM and THUMB
3686 instruction sets. Without this option the two instruction sets cannot
3687 be reliably used inside one program. The default is
3688 @samp{-mno-thumb-interwork}, since slightly larger code is generated
3689 when @samp{-mthumb-interwork} is specified.
3691 @item -mno-sched-prolog
3692 @kindex -mno-sched-prolog
3693 @kindex -msched-prolog
3694 Prevent the reordering of instructions in the function prolog, or the
3695 merging of those instruction with the instructions in the function's
3696 body. This means that all functions will start with a recognisable set
3697 of instructions (or in fact one of a chioce from a small set of
3698 different function prologues), and this information can be used to
3699 locate the start if functions inside an executable piece of code. The
3700 default is @samp{-msched-prolog}.
3703 Generate output containing floating point instructions. This is the
3707 Generate output containing library calls for floating point.
3708 @strong{Warning:} the requisite libraries are not available for all ARM
3709 targets. Normally the facilities of the machine's usual C compiler are
3710 used, but this cannot be done directly in cross-compilation. You must make
3711 your own arrangements to provide suitable library functions for
3714 @samp{-msoft-float} changes the calling convention in the output file;
3715 therefore, it is only useful if you compile @emph{all} of a program with
3716 this option. In particular, you need to compile @file{libgcc.a}, the
3717 library that comes with GNU CC, with @samp{-msoft-float} in order for
3720 @item -mlittle-endian
3721 Generate code for a processor running in little-endian mode. This is
3722 the default for all standard configurations.
3725 Generate code for a processor running in big-endian mode; the default is
3726 to compile code for a little-endian processor.
3728 @item -mwords-little-endian
3729 This option only applies when generating code for big-endian processors.
3730 Generate code for a little-endian word order but a big-endian byte
3731 order. That is, a byte order of the form @samp{32107654}. Note: this
3732 option should only be used if you require compatibility with code for
3733 big-endian ARM processors generated by versions of the compiler prior to
3736 @item -mshort-load-bytes
3737 @kindex -mshort-load-bytes
3738 Do not try to load half-words (eg @samp{short}s) by loading a word from
3739 an unaligned address. For some targets the MMU is configured to trap
3740 unaligned loads; use this option to generate code that is safe in these
3743 @item -mno-short-load-bytes
3744 @kindex -mno-short-load-bytes
3745 Use unaligned word loads to load half-words (eg @samp{short}s). This
3746 option produces more efficient code, but the MMU is sometimes configured
3747 to trap these instructions.
3749 @item -mshort-load-words
3750 @kindex -mshort-load-words
3751 This is a synonym for the @samp{-mno-short-load-bytes}.
3753 @item -mno-short-load-words
3754 @kindex -mno-short-load-words
3755 This is a synonym for the @samp{-mshort-load-bytes}.
3759 This option only applies to RISC iX. Emulate the native BSD-mode
3760 compiler. This is the default if @samp{-ansi} is not specified.
3764 This option only applies to RISC iX. Emulate the native X/Open-mode
3767 @item -mno-symrename
3768 @kindex -mno-symrename
3769 This option only applies to RISC iX. Do not run the assembler
3770 post-processor, @samp{symrename}, after code has been assembled.
3771 Normally it is necessary to modify some of the standard symbols in
3772 preparation for linking with the RISC iX C library; this option
3773 suppresses this pass. The post-processor is never run when the
3774 compiler is built for cross-compilation.
3778 This specifies the name of the target ARM processor. GCC uses this name
3779 to determine what kind of instructions it can use when generating
3780 assembly code. Permissable names are: arm2, arm250, arm3, arm6, arm60,
3781 arm600, arm610, arm620, arm7, arm7m, arm7d, arm7dm, arm7di, arm7dmi,
3782 arm70, arm700, arm700i, arm710, arm710c, arm7100, arm7500, arm7500fe,
3783 arm7tdmi, arm8, strongarm, strongarm110
3787 This specifies the name of the target ARM architecture. GCC uses this
3788 name to determine what kind of instructions it can use when generating
3789 assembly code. This option can be used in conjunction with or instead
3790 of the @samp{-mcpu=} option. Permissable names are: armv2, armv2a,
3791 armv3, armv3m, armv4, armv4t
3793 @item -mfpe=<number>
3795 This specifes the version of the floating point emulation available on
3796 the target. Permissable values are 2 and 3.
3798 @item -mstructure-size-boundary=<n>
3799 @kindex -mstructure-size-boundary
3800 The size of all structures and unions will be rounded up to a multiple
3801 of the number of bits set by this option. Permissable values are 8 and
3802 32. The default value varies for different toolchains. For the COFF
3803 targeted toolchain the default value is 8. Specifying the larger number
3804 can produced faster, more efficient code, but can also increase the size
3805 of the program. The two values are potentially incompatible. Code
3806 compiled with one value cannot necessarily expect to work with code or
3807 libraries compiled with the other value, if they exchange information
3808 using structures or unions. Programmers are encouraged to use the 32
3809 value as future versions of the toolchain may default to this value.
3814 @subsection Thumb Options
3815 @cindex Thumb Options
3819 @item -mthumb-interwork
3820 @kindex -mthumb-interwork
3821 @kindex -mno-thumb-interwork
3822 Generate code which supports calling between the THUMB and ARM
3823 instruction sets. Without this option the two instruction sets cannot
3824 be reliably used inside one program. The default is
3825 @samp{-mno-thumb-interwork}, since slightly smaller code is generated
3829 @kindex -mtpcs-frame
3830 @kindex -mno-tpcs-frame
3831 Generate a stack frame that is compliant with the Thumb Procedure Call
3832 Standard for all non-leaf functions. (A leaf function is one that does
3833 not call any other functions). The default is @samp{-mno-apcs-frame}.
3835 @item -mtpcs-leaf-frame
3836 @kindex -mtpcs-leaf-frame
3837 @kindex -mno-tpcs-leaf-frame
3838 Generate a stack frame that is compliant with the Thumb Procedure Call
3839 Standard for all leaf functions. (A leaf function is one that does
3840 not call any other functions). The default is @samp{-mno-apcs-leaf-frame}.
3842 @item -mlittle-endian
3843 @kindex -mlittle-endian
3844 Generate code for a processor running in little-endian mode. This is
3845 the default for all standard configurations.
3848 @kindex -mbig-endian
3849 Generate code for a processor running in big-endian mode.
3851 @item -mstructure-size-boundary=<n>
3852 @kindex -mstructure-size-boundary
3853 The size of all structures and unions will be rounded up to a multiple
3854 of the number of bits set by this option. Permissable values are 8 and
3855 32. The default value varies for different toolchains. For the COFF
3856 targeted toolchain the default value is 8. Specifying the larger number
3857 can produced faster, more efficient code, but can also increase the size
3858 of the program. The two values are potentially incompatible. Code
3859 compiled with one value cannot necessarily expect to work with code or
3860 libraries compiled with the other value, if they exchange information
3861 using structures or unions. Programmers are encouraged to use the 32
3862 value as future versions of the toolchain may default to this value.
3866 @node MN10200 Options
3867 @subsection MN10200 Options
3868 @cindex MN10200 options
3869 These @samp{-m} options are defined for Matsushita MN10200 architectures:
3873 Indicate to the linker that it should perform a relaxation optimization pass
3874 to shorten branches, calls and absolute memory addresses. This option only
3875 has an effect when used on the command line for the final link step.
3877 This option makes symbolic debugging impossible.
3880 @node MN10300 Options
3881 @subsection MN10300 Options
3882 @cindex MN10300 options
3883 These @samp{-m} options are defined for Matsushita MN10300 architectures:
3887 Generate code to avoid bugs in the multiply instructions for the MN10300
3888 processors. This is the default.
3891 Do not generate code to avoid bugs in the multiply instructions for the
3895 Indicate to the linker that it should perform a relaxation optimization pass
3896 to shorten branches, calls and absolute memory addresses. This option only
3897 has an effect when used on the command line for the final link step.
3899 This option makes symbolic debugging impossible.
3903 @node M32R/D Options
3904 @subsection M32R/D Options
3905 @cindex M32R/D options
3907 These @samp{-m} options are defined for Mitsubishi M32R/D architectures:
3910 @item -mcode-model=small
3911 Assume all objects live in the lower 16MB of memory (so that their addresses
3912 can be loaded with the @code{ld24} instruction), and assume all subroutines
3913 are reachable with the @code{bl} instruction.
3914 This is the default.
3916 The addressability of a particular object can be set with the
3917 @code{model} attribute.
3919 @item -mcode-model=medium
3920 Assume objects may be anywhere in the 32 bit address space (the compiler
3921 will generate @code{seth/add3} instructions to load their addresses), and
3922 assume all subroutines are reachable with the @code{bl} instruction.
3924 @item -mcode-model=large
3925 Assume objects may be anywhere in the 32 bit address space (the compiler
3926 will generate @code{seth/add3} instructions to load their addresses), and
3927 assume subroutines may not be reachable with the @code{bl} instruction
3928 (the compiler will generate the much slower @code{seth/add3/jl}
3929 instruction sequence).
3932 Disable use of the small data area. Variables will be put into
3933 one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the
3934 @code{section} attribute has been specified).
3935 This is the default.
3937 The small data area consists of sections @samp{.sdata} and @samp{.sbss}.
3938 Objects may be explicitly put in the small data area with the
3939 @code{section} attribute using one of these sections.
3942 Put small global and static data in the small data area, but do not
3943 generate special code to reference them.
3946 Put small global and static data in the small data area, and generate
3947 special instructions to reference them.
3950 @cindex smaller data references
3951 Put global and static objects less than or equal to @var{num} bytes
3952 into the small data or bss sections instead of the normal data or bss
3953 sections. The default value of @var{num} is 8.
3954 The @samp{-msdata} option must be set to one of @samp{sdata} or @samp{use}
3955 for this option to have any effect.
3957 All modules should be compiled with the same @samp{-G @var{num}} value.
3958 Compiling with different values of @var{num} may or may not work; if it
3959 doesn't the linker will give an error message - incorrect code will not be
3965 @subsection M88K Options
3966 @cindex M88k options
3968 These @samp{-m} options are defined for Motorola 88k architectures:
3973 Generate code that works well on both the m88100 and the
3978 Generate code that works best for the m88100, but that also
3983 Generate code that works best for the m88110, and may not run
3988 Obsolete option to be removed from the next revision.
3991 @item -midentify-revision
3992 @kindex -midentify-revision
3994 @cindex identifying source, compiler (88k)
3995 Include an @code{ident} directive in the assembler output recording the
3996 source file name, compiler name and version, timestamp, and compilation
3999 @item -mno-underscores
4000 @kindex -mno-underscores
4001 @cindex underscores, avoiding (88k)
4002 In assembler output, emit symbol names without adding an underscore
4003 character at the beginning of each name. The default is to use an
4004 underscore as prefix on each name.
4006 @item -mocs-debug-info
4007 @itemx -mno-ocs-debug-info
4008 @kindex -mocs-debug-info
4009 @kindex -mno-ocs-debug-info
4011 @cindex debugging, 88k OCS
4012 Include (or omit) additional debugging information (about registers used
4013 in each stack frame) as specified in the 88open Object Compatibility
4014 Standard, ``OCS''. This extra information allows debugging of code that
4015 has had the frame pointer eliminated. The default for DG/UX, SVr4, and
4016 Delta 88 SVr3.2 is to include this information; other 88k configurations
4017 omit this information by default.
4019 @item -mocs-frame-position
4020 @kindex -mocs-frame-position
4021 @cindex register positions in frame (88k)
4022 When emitting COFF debugging information for automatic variables and
4023 parameters stored on the stack, use the offset from the canonical frame
4024 address, which is the stack pointer (register 31) on entry to the
4025 function. The DG/UX, SVr4, Delta88 SVr3.2, and BCS configurations use
4026 @samp{-mocs-frame-position}; other 88k configurations have the default
4027 @samp{-mno-ocs-frame-position}.
4029 @item -mno-ocs-frame-position
4030 @kindex -mno-ocs-frame-position
4031 @cindex register positions in frame (88k)
4032 When emitting COFF debugging information for automatic variables and
4033 parameters stored on the stack, use the offset from the frame pointer
4034 register (register 30). When this option is in effect, the frame
4035 pointer is not eliminated when debugging information is selected by the
4038 @item -moptimize-arg-area
4039 @itemx -mno-optimize-arg-area
4040 @kindex -moptimize-arg-area
4041 @kindex -mno-optimize-arg-area
4042 @cindex arguments in frame (88k)
4043 Control how function arguments are stored in stack frames.
4044 @samp{-moptimize-arg-area} saves space by optimizing them, but this
4045 conflicts with the 88open specifications. The opposite alternative,
4046 @samp{-mno-optimize-arg-area}, agrees with 88open standards. By default
4047 GNU CC does not optimize the argument area.
4049 @item -mshort-data-@var{num}
4050 @kindex -mshort-data-@var{num}
4051 @cindex smaller data references (88k)
4052 @cindex r0-relative references (88k)
4053 Generate smaller data references by making them relative to @code{r0},
4054 which allows loading a value using a single instruction (rather than the
4055 usual two). You control which data references are affected by
4056 specifying @var{num} with this option. For example, if you specify
4057 @samp{-mshort-data-512}, then the data references affected are those
4058 involving displacements of less than 512 bytes.
4059 @samp{-mshort-data-@var{num}} is not effective for @var{num} greater
4062 @item -mserialize-volatile
4063 @kindex -mserialize-volatile
4064 @itemx -mno-serialize-volatile
4065 @kindex -mno-serialize-volatile
4066 @cindex sequential consistency on 88k
4067 Do, or don't, generate code to guarantee sequential consistency
4068 of volatile memory references. By default, consistency is
4071 The order of memory references made by the MC88110 processor does
4072 not always match the order of the instructions requesting those
4073 references. In particular, a load instruction may execute before
4074 a preceding store instruction. Such reordering violates
4075 sequential consistency of volatile memory references, when there
4076 are multiple processors. When consistency must be guaranteed,
4077 GNU C generates special instructions, as needed, to force
4078 execution in the proper order.
4080 The MC88100 processor does not reorder memory references and so
4081 always provides sequential consistency. However, by default, GNU
4082 C generates the special instructions to guarantee consistency
4083 even when you use @samp{-m88100}, so that the code may be run on an
4084 MC88110 processor. If you intend to run your code only on the
4085 MC88100 processor, you may use @samp{-mno-serialize-volatile}.
4087 The extra code generated to guarantee consistency may affect the
4088 performance of your application. If you know that you can safely
4089 forgo this guarantee, you may use @samp{-mno-serialize-volatile}.
4095 @cindex assembler syntax, 88k
4097 Turn on (@samp{-msvr4}) or off (@samp{-msvr3}) compiler extensions
4098 related to System V release 4 (SVr4). This controls the following:
4102 Which variant of the assembler syntax to emit.
4104 @samp{-msvr4} makes the C preprocessor recognize @samp{#pragma weak}
4105 that is used on System V release 4.
4107 @samp{-msvr4} makes GNU CC issue additional declaration directives used in
4111 @samp{-msvr4} is the default for the m88k-motorola-sysv4 and
4112 m88k-dg-dgux m88k configurations. @samp{-msvr3} is the default for all
4113 other m88k configurations.
4115 @item -mversion-03.00
4116 @kindex -mversion-03.00
4117 This option is obsolete, and is ignored.
4118 @c ??? which asm syntax better for GAS? option there too?
4120 @item -mno-check-zero-division
4121 @itemx -mcheck-zero-division
4122 @kindex -mno-check-zero-division
4123 @kindex -mcheck-zero-division
4124 @cindex zero division on 88k
4125 Do, or don't, generate code to guarantee that integer division by
4126 zero will be detected. By default, detection is guaranteed.
4128 Some models of the MC88100 processor fail to trap upon integer
4129 division by zero under certain conditions. By default, when
4130 compiling code that might be run on such a processor, GNU C
4131 generates code that explicitly checks for zero-valued divisors
4132 and traps with exception number 503 when one is detected. Use of
4133 mno-check-zero-division suppresses such checking for code
4134 generated to run on an MC88100 processor.
4136 GNU C assumes that the MC88110 processor correctly detects all
4137 instances of integer division by zero. When @samp{-m88110} is
4138 specified, both @samp{-mcheck-zero-division} and
4139 @samp{-mno-check-zero-division} are ignored, and no explicit checks for
4140 zero-valued divisors are generated.
4142 @item -muse-div-instruction
4143 @kindex -muse-div-instruction
4144 @cindex divide instruction, 88k
4145 Use the div instruction for signed integer division on the
4146 MC88100 processor. By default, the div instruction is not used.
4148 On the MC88100 processor the signed integer division instruction
4149 div) traps to the operating system on a negative operand. The
4150 operating system transparently completes the operation, but at a
4151 large cost in execution time. By default, when compiling code
4152 that might be run on an MC88100 processor, GNU C emulates signed
4153 integer division using the unsigned integer division instruction
4154 divu), thereby avoiding the large penalty of a trap to the
4155 operating system. Such emulation has its own, smaller, execution
4156 cost in both time and space. To the extent that your code's
4157 important signed integer division operations are performed on two
4158 nonnegative operands, it may be desirable to use the div
4159 instruction directly.
4161 On the MC88110 processor the div instruction (also known as the
4162 divs instruction) processes negative operands without trapping to
4163 the operating system. When @samp{-m88110} is specified,
4164 @samp{-muse-div-instruction} is ignored, and the div instruction is used
4165 for signed integer division.
4167 Note that the result of dividing INT_MIN by -1 is undefined. In
4168 particular, the behavior of such a division with and without
4169 @samp{-muse-div-instruction} may differ.
4171 @item -mtrap-large-shift
4172 @itemx -mhandle-large-shift
4173 @kindex -mtrap-large-shift
4174 @kindex -mhandle-large-shift
4175 @cindex bit shift overflow (88k)
4176 @cindex large bit shifts (88k)
4177 Include code to detect bit-shifts of more than 31 bits; respectively,
4178 trap such shifts or emit code to handle them properly. By default GNU CC
4179 makes no special provision for large bit shifts.
4181 @item -mwarn-passed-structs
4182 @kindex -mwarn-passed-structs
4183 @cindex structure passing (88k)
4184 Warn when a function passes a struct as an argument or result.
4185 Structure-passing conventions have changed during the evolution of the C
4186 language, and are often the source of portability problems. By default,
4187 GNU CC issues no such warning.
4190 @node RS/6000 and PowerPC Options
4191 @subsection IBM RS/6000 and PowerPC Options
4192 @cindex RS/6000 and PowerPC Options
4193 @cindex IBM RS/6000 and PowerPC Options
4195 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
4203 @itemx -mpowerpc-gpopt
4204 @itemx -mno-powerpc-gpopt
4205 @itemx -mpowerpc-gfxopt
4206 @itemx -mno-powerpc-gfxopt
4208 @itemx -mno-powerpc64
4212 @kindex -mpowerpc-gpopt
4213 @kindex -mpowerpc-gfxopt
4215 GNU CC supports two related instruction set architectures for the
4216 RS/6000 and PowerPC. The @dfn{POWER} instruction set are those
4217 instructions supported by the @samp{rios} chip set used in the original
4218 RS/6000 systems and the @dfn{PowerPC} instruction set is the
4219 architecture of the Motorola MPC5xx, MPC6xx, MPC8xx microprocessors, and
4220 the IBM 4xx microprocessors.
4222 Neither architecture is a subset of the other. However there is a
4223 large common subset of instructions supported by both. An MQ
4224 register is included in processors supporting the POWER architecture.
4226 You use these options to specify which instructions are available on the
4227 processor you are using. The default value of these options is
4228 determined when configuring GNU CC. Specifying the
4229 @samp{-mcpu=@var{cpu_type}} overrides the specification of these
4230 options. We recommend you use the @samp{-mcpu=@var{cpu_type}} option
4231 rather than the options listed above.
4233 The @samp{-mpower} option allows GNU CC to generate instructions that
4234 are found only in the POWER architecture and to use the MQ register.
4235 Specifying @samp{-mpower2} implies @samp{-power} and also allows GNU CC
4236 to generate instructions that are present in the POWER2 architecture but
4237 not the original POWER architecture.
4239 The @samp{-mpowerpc} option allows GNU CC to generate instructions that
4240 are found only in the 32-bit subset of the PowerPC architecture.
4241 Specifying @samp{-mpowerpc-gpopt} implies @samp{-mpowerpc} and also allows
4242 GNU CC to use the optional PowerPC architecture instructions in the
4243 General Purpose group, including floating-point square root. Specifying
4244 @samp{-mpowerpc-gfxopt} implies @samp{-mpowerpc} and also allows GNU CC to
4245 use the optional PowerPC architecture instructions in the Graphics
4246 group, including floating-point select.
4248 The @samp{-mpowerpc64} option allows GNU CC to generate the additional
4249 64-bit instructions that are found in the full PowerPC64 architecture
4250 and to treat GPRs as 64-bit, doubleword quantities. GNU CC defaults to
4251 @samp{-mno-powerpc64}.
4253 If you specify both @samp{-mno-power} and @samp{-mno-powerpc}, GNU CC
4254 will use only the instructions in the common subset of both
4255 architectures plus some special AIX common-mode calls, and will not use
4256 the MQ register. Specifying both @samp{-mpower} and @samp{-mpowerpc}
4257 permits GNU CC to use any instruction from either architecture and to
4258 allow use of the MQ register; specify this for the Motorola MPC601.
4260 @item -mnew-mnemonics
4261 @itemx -mold-mnemonics
4262 @kindex -mnew-mnemonics
4263 @kindex -mold-mnemonics
4264 Select which mnemonics to use in the generated assembler code.
4265 @samp{-mnew-mnemonics} requests output that uses the assembler mnemonics
4266 defined for the PowerPC architecture, while @samp{-mold-mnemonics}
4267 requests the assembler mnemonics defined for the POWER architecture.
4268 Instructions defined in only one architecture have only one mnemonic;
4269 GNU CC uses that mnemonic irrespective of which of these options is
4272 GNU CC defaults to the mnemonics appropriate for the architecture in
4273 use. Specifying @samp{-mcpu=@var{cpu_type}} sometimes overrides the
4274 value of these option. Unless you are building a cross-compiler, you
4275 should normally not specify either @samp{-mnew-mnemonics} or
4276 @samp{-mold-mnemonics}, but should instead accept the default.
4278 @item -mcpu=@var{cpu_type}
4280 Set architecture type, register usage, choice of mnemonics, and
4281 instruction scheduling parameters for machine type @var{cpu_type}.
4282 Supported values for @var{cpu_type} are @samp{rs6000}, @samp{rios1},
4283 @samp{rios2}, @samp{rsc}, @samp{601}, @samp{602}, @samp{603},
4284 @samp{603e}, @samp{604}, @samp{604e}, @samp{620}, @samp{power},
4285 @samp{power2}, @samp{powerpc}, @samp{403}, @samp{505}, @samp{801},
4286 @samp{821}, @samp{823}, and @samp{860} and @samp{common}.
4287 @samp{-mcpu=power}, @samp{-mcpu=power2}, and @samp{-mcpu=powerpc}
4288 specify generic POWER, POWER2 and pure PowerPC (i.e., not MPC601)
4289 architecture machine types, with an appropriate, generic processor model
4290 assumed for scheduling purposes.@refill
4292 @c overfull hbox here --bob 22 jul96
4293 @c original text between ignore ... end ignore
4295 Specifying any of the @samp{-mcpu=rios1}, @samp{-mcpu=rios2},
4296 @samp{-mcpu=rsc}, @samp{-mcpu=power}, or @samp{-mcpu=power2} options
4297 enables the @samp{-mpower} option and disables the @samp{-mpowerpc}
4298 option; @samp{-mcpu=601} enables both the @samp{-mpower} and
4299 @samp{-mpowerpc} options; all of @samp{-mcpu=602}, @samp{-mcpu=603},
4300 @samp{-mcpu=603e}, @samp{-mcpu=604}, @samp{-mcpu=604e},
4301 @samp{-mcpu=620}, @samp{-mcpu=403}, @samp{-mcpu=505}, @samp{-mcpu=801},
4302 @samp{-mcpu=821}, @samp{-mcpu=823}, @samp{-mcpu=860} and
4303 @samp{-mcpu=powerpc} enable the @samp{-mpowerpc} option and disable the
4304 @samp{-mpower} option; @samp{-mcpu=common} disables both the
4305 @samp{-mpower} and @samp{-mpowerpc} options.@refill
4307 @c changed paragraph
4308 Specifying any of the following options:
4309 @samp{-mcpu=rios1}, @samp{-mcpu=rios2}, @samp{-mcpu=rsc},
4310 @samp{-mcpu=power}, or @samp{-mcpu=power2}
4311 enables the @samp{-mpower} option and disables the @samp{-mpowerpc} option;
4312 @samp{-mcpu=601} enables both the @samp{-mpower} and @samp{-mpowerpc} options.
4313 All of @samp{-mcpu=602}, @samp{-mcpu=603}, @samp{-mcpu=603e},
4314 @samp{-mcpu=604}, @samp{-mcpu=620},
4315 enable the @samp{-mpowerpc} option and disable the @samp{-mpower} option.
4316 Exactly similarly, all of @samp{-mcpu=403},
4317 @samp{-mcpu=505}, @samp{-mcpu=821}, @samp{-mcpu=860} and @samp{-mcpu=powerpc}
4318 enable the @samp{-mpowerpc} option and disable the @samp{-mpower} option.
4319 @samp{-mcpu=common} disables both the
4320 @samp{-mpower} and @samp{-mpowerpc} options.@refill
4321 @c end changes to prevent overfull hboxes
4323 AIX versions 4 or greater selects @samp{-mcpu=common} by default, so
4324 that code will operate on all members of the RS/6000 and PowerPC
4325 families. In that case, GNU CC will use only the instructions in the
4326 common subset of both architectures plus some special AIX common-mode
4327 calls, and will not use the MQ register. GNU CC assumes a generic
4328 processor model for scheduling purposes.
4330 Specifying any of the options @samp{-mcpu=rios1}, @samp{-mcpu=rios2},
4331 @samp{-mcpu=rsc}, @samp{-mcpu=power}, or @samp{-mcpu=power2} also
4332 disables the @samp{new-mnemonics} option. Specifying @samp{-mcpu=601},
4333 @samp{-mcpu=602}, @samp{-mcpu=603}, @samp{-mcpu=603e}, @samp{-mcpu=604},
4334 @samp{620}, @samp{403}, or @samp{-mcpu=powerpc} also enables the
4335 @samp{new-mnemonics} option.@refill
4337 Specifying @samp{-mcpu=403}, @samp{-mcpu=821}, or @samp{-mcpu=860} also
4338 enables the @samp{-msoft-float} option.
4340 @item -mtune=@var{cpu_type}
4341 Set the instruction scheduling parameters for machine type
4342 @var{cpu_type}, but do not set the architecture type, register usage,
4343 choice of mnemonics like @samp{-mcpu=}@var{cpu_type} would. The same
4344 values for @var{cpu_type} are used for @samp{-mtune=}@var{cpu_type} as
4345 for @samp{-mcpu=}@var{cpu_type}. The @samp{-mtune=}@var{cpu_type}
4346 option overrides the @samp{-mcpu=}@var{cpu_type} option in terms of
4347 instruction scheduling parameters.
4350 @itemx -mno-fp-in-toc
4351 @itemx -mno-sum-in-toc
4352 @itemx -mminimal-toc
4353 @kindex -mminimal-toc
4354 Modify generation of the TOC (Table Of Contents), which is created for
4355 every executable file. The @samp{-mfull-toc} option is selected by
4356 default. In that case, GNU CC will allocate at least one TOC entry for
4357 each unique non-automatic variable reference in your program. GNU CC
4358 will also place floating-point constants in the TOC. However, only
4359 16,384 entries are available in the TOC.
4361 If you receive a linker error message that saying you have overflowed
4362 the available TOC space, you can reduce the amount of TOC space used
4363 with the @samp{-mno-fp-in-toc} and @samp{-mno-sum-in-toc} options.
4364 @samp{-mno-fp-in-toc} prevents GNU CC from putting floating-point
4365 constants in the TOC and @samp{-mno-sum-in-toc} forces GNU CC to
4366 generate code to calculate the sum of an address and a constant at
4367 run-time instead of putting that sum into the TOC. You may specify one
4368 or both of these options. Each causes GNU CC to produce very slightly
4369 slower and larger code at the expense of conserving TOC space.
4371 If you still run out of space in the TOC even when you specify both of
4372 these options, specify @samp{-mminimal-toc} instead. This option causes
4373 GNU CC to make only one TOC entry for every file. When you specify this
4374 option, GNU CC will produce code that is slower and larger but which
4375 uses extremely little TOC space. You may wish to use this option
4376 only on files that contain less frequently executed code. @refill
4382 Enable AIX 64-bit ABI and calling convention: 64-bit pointers, 64-bit
4383 @code{long} type, and the infrastructure needed to support them.
4384 Specifying @samp{-maix64} implies @samp{-mpowerpc64} and
4385 @samp{-mpowerpc}, while @samp{-maix32} disables the 64-bit ABI and
4386 implies @samp{-mno-powerpc64}. GNU CC defaults to @samp{-maix32}.
4391 On AIX, pass floating-point arguments to prototyped functions beyond the
4392 register save area (RSA) on the stack in addition to argument FPRs. The
4393 AIX calling convention was extended but not initially documented to
4394 handle an obscure K&R C case of calling a function that takes the
4395 address of its arguments with fewer arguments than declared. AIX XL
4396 compilers access floating point arguments which do not fit in the
4397 RSA from the stack when a subroutine is compiled without
4398 optimization. Because always storing floating-point arguments on the
4399 stack is inefficient and rarely needed, this option is not enabled by
4400 default and only is necessary when calling subroutines compiled by AIX
4401 XL compilers without optimization.
4405 Support @dfn{AIX Threads}. Link an application written to use
4406 @dfn{pthreads} with special libraries and startup code to enable the
4411 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE). Link an
4412 application written to use message passing with special startup code to
4413 enable the application to run. The system must have PE installed in the
4414 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
4415 must be overridden with the @samp{-specs=} option to specify the
4416 appropriate directory location. The Parallel Environment does not
4417 support threads, so the @samp{-mpe} option and the @samp{-mthreads}
4418 option are incompatible.
4422 @kindex -msoft-float
4423 Generate code that does not use (uses) the floating-point register set.
4424 Software floating point emulation is provided if you use the
4425 @samp{-msoft-float} option, and pass the option to GNU CC when linking.
4428 @itemx -mno-multiple
4429 Generate code that uses (does not use) the load multiple word
4430 instructions and the store multiple word instructions. These
4431 instructions are generated by default on POWER systems, and not
4432 generated on PowerPC systems. Do not use @samp{-mmultiple} on little
4433 endian PowerPC systems, since those instructions do not work when the
4434 processor is in little endian mode.
4439 Generate code that uses (does not use) the load string instructions and the
4440 store string word instructions to save multiple registers and do small block
4441 moves. These instructions are generated by default on POWER systems, and not
4442 generated on PowerPC systems. Do not use @samp{-mstring} on little endian
4443 PowerPC systems, since those instructions do not work when the processor is in
4449 Generate code that uses (does not use) the load or store instructions
4450 that update the base register to the address of the calculated memory
4451 location. These instructions are generated by default. If you use
4452 @samp{-mno-update}, there is a small window between the time that the
4453 stack pointer is updated and the address of the previous frame is
4454 stored, which means code that walks the stack frame across interrupts or
4455 signals may get corrupted data.
4458 @itemx -mno-fused-madd
4459 @kindex -mfused-madd
4460 Generate code that uses (does not use) the floating point multiply and
4461 accumulate instructions. These instructions are generated by default if
4462 hardware floating is used.
4464 @item -mno-bit-align
4467 On System V.4 and embedded PowerPC systems do not (do) force structures
4468 and unions that contain bit fields to be aligned to the base type of the
4471 For example, by default a structure containing nothing but 8
4472 @code{unsigned} bitfields of length 1 would be aligned to a 4 byte
4473 boundary and have a size of 4 bytes. By using @samp{-mno-bit-align},
4474 the structure would be aligned to a 1 byte boundary and be one byte in
4477 @item -mno-strict-align
4478 @itemx -mstrict-align
4479 @kindex -mstrict-align
4480 On System V.4 and embedded PowerPC systems do not (do) assume that
4481 unaligned memory references will be handled by the system.
4484 @itemx -mno-relocatable
4485 @kindex -mrelocatable
4486 On embedded PowerPC systems generate code that allows (does not allow)
4487 the program to be relocated to a different address at runtime. If you
4488 use @samp{-mrelocatable} on any module, all objects linked together must
4489 be compiled with @samp{-mrelocatable} or @samp{-mrelocatable-lib}.
4491 @item -mrelocatable-lib
4492 @itemx -mno-relocatable-lib
4493 On embedded PowerPC systems generate code that allows (does not allow)
4494 the program to be relocated to a different address at runtime. Modules
4495 compiled with @samp{-mrelocatable-lib} can be linked with either modules
4496 compiled without @samp{-mrelocatable} and @samp{-mrelocatable-lib} or
4497 with modules compiled with the @samp{-mrelocatable} options.
4501 On System V.4 and embedded PowerPC systems do not (do) assume that
4502 register 2 contains a pointer to a global area pointing to the addresses
4503 used in the program.
4506 @itemx -mlittle-endian
4507 On System V.4 and embedded PowerPC systems compile code for the
4508 processor in little endian mode. The @samp{-mlittle-endian} option is
4509 the same as @samp{-mlittle}.
4513 On System V.4 and embedded PowerPC systems compile code for the
4514 processor in big endian mode. The @samp{-mbig-endian} option is
4515 the same as @samp{-mbig}.
4518 On System V.4 and embedded PowerPC systems compile code using calling
4519 conventions that adheres to the March 1995 draft of the System V
4520 Application Binary Interface, PowerPC processor supplement. This is the
4521 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
4523 @item -mcall-sysv-eabi
4524 Specify both @samp{-mcall-sysv} and @samp{-meabi} options.
4526 @item -mcall-sysv-noeabi
4527 Specify both @samp{-mcall-sysv} and @samp{-mno-eabi} options.
4530 On System V.4 and embedded PowerPC systems compile code using calling
4531 conventions that are similar to those used on AIX. This is the
4532 default if you configured GCC using @samp{powerpc-*-eabiaix}.
4534 @item -mcall-solaris
4535 On System V.4 and embedded PowerPC systems compile code for the Solaris
4539 On System V.4 and embedded PowerPC systems compile code for the
4540 Linux-based GNU system.
4543 @itemx -mno-prototype
4544 On System V.4 and embedded PowerPC systems assume that all calls to
4545 variable argument functions are properly prototyped. Otherwise, the
4546 compiler must insert an instruction before every non prototyped call to
4547 set or clear bit 6 of the condition code register (@var{CR}) to
4548 indicate whether floating point values were passed in the floating point
4549 registers in case the function takes a variable arguments. With
4550 @samp{-mprototype}, only calls to prototyped variable argument functions
4551 will set or clear the bit.
4554 On embedded PowerPC systems, assume that the startup module is called
4555 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
4556 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}.
4560 On embedded PowerPC systems, assume that the startup module is called
4561 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
4565 On embedded PowerPC systems, assume that the startup module is called
4566 @file{crt0.o} and the standard C libraries are @file{libads.a} and
4570 On embedded PowerPC systems, assume that the startup module is called
4571 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
4575 On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags
4576 header to indicate that @samp{eabi} extended relocations are used.
4580 On System V.4 and embedded PowerPC systems do (do not) adhere to the
4581 Embedded Applications Binary Interface (eabi) which is a set of
4582 modifications to the System V.4 specifications. Selecting @code{-meabi}
4583 means that the stack is aligned to an 8 byte boundary, a function
4584 @code{__eabi} is called to from @code{main} to set up the eabi
4585 environment, and the @samp{-msdata} option can use both @code{r2} and
4586 @code{r13} to point to two separate small data areas. Selecting
4587 @code{-mno-eabi} means that the stack is aligned to a 16 byte boundary,
4588 do not call an initialization function from @code{main}, and the
4589 @samp{-msdata} option will only use @code{r13} to point to a single
4590 small data area. The @samp{-meabi} option is on by default if you
4591 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
4594 On System V.4 and embedded PowerPC systems, put small initialized
4595 @code{const} global and static data in the @samp{.sdata2} section, which
4596 is pointed to by register @code{r2}. Put small initialized
4597 non-@code{const} global and static data in the @samp{.sdata} section,
4598 which is pointed to by register @code{r13}. Put small uninitialized
4599 global and static data in the @samp{.sbss} section, which is adjacent to
4600 the @samp{.sdata} section. The @samp{-msdata=eabi} option is
4601 incompatible with the @samp{-mrelocatable} option. The
4602 @samp{-msdata=eabi} option also sets the @samp{-memb} option.
4605 On System V.4 and embedded PowerPC systems, put small global and static
4606 data in the @samp{.sdata} section, which is pointed to by register
4607 @code{r13}. Put small uninitialized global and static data in the
4608 @samp{.sbss} section, which is adjacent to the @samp{.sdata} section.
4609 The @samp{-msdata=sysv} option is incompatible with the
4610 @samp{-mrelocatable} option.
4612 @item -msdata=default
4614 On System V.4 and embedded PowerPC systems, if @samp{-meabi} is used,
4615 compile code the same as @samp{-msdata=eabi}, otherwise compile code the
4616 same as @samp{-msdata=sysv}.
4619 On System V.4 and embedded PowerPC systems, put small global and static
4620 data in the @samp{.sdata} section. Put small uninitialized global and
4621 static data in the @samp{.sbss} section. Do not use register @code{r13}
4622 to address small data however. This is the default behavior unless
4623 other @samp{-msdata} options are used.
4627 On embedded PowerPC systems, put all initialized global and static data
4628 in the @samp{.data} section, and all uninitialized data in the
4629 @samp{.bss} section.
4632 @cindex smaller data references (PowerPC)
4633 @cindex .sdata/.sdata2 references (PowerPC)
4634 On embedded PowerPC systems, put global and static items less than or
4635 equal to @var{num} bytes into the small data or bss sections instead of
4636 the normal data or bss section. By default, @var{num} is 8. The
4637 @samp{-G @var{num}} switch is also passed to the linker.
4638 All modules should be compiled with the same @samp{-G @var{num}} value.
4641 @itemx -mno-regnames
4642 On System V.4 and embedded PowerPC systems do (do not) emit register
4643 names in the assembly language output using symbolic forms.
4646 @subsection IBM RT Options
4648 @cindex IBM RT options
4650 These @samp{-m} options are defined for the IBM RT PC:
4654 Use an in-line code sequence for integer multiplies. This is the
4657 @item -mcall-lib-mul
4658 Call @code{lmul$$} for integer multiples.
4660 @item -mfull-fp-blocks
4661 Generate full-size floating point data blocks, including the minimum
4662 amount of scratch space recommended by IBM. This is the default.
4664 @item -mminimum-fp-blocks
4665 Do not include extra scratch space in floating point data blocks. This
4666 results in smaller code, but slower execution, since scratch space must
4667 be allocated dynamically.
4669 @cindex @file{varargs.h} and RT PC
4670 @cindex @file{stdarg.h} and RT PC
4671 @item -mfp-arg-in-fpregs
4672 Use a calling sequence incompatible with the IBM calling convention in
4673 which floating point arguments are passed in floating point registers.
4674 Note that @code{varargs.h} and @code{stdargs.h} will not work with
4675 floating point operands if this option is specified.
4677 @item -mfp-arg-in-gregs
4678 Use the normal calling convention for floating point arguments. This is
4681 @item -mhc-struct-return
4682 Return structures of more than one word in memory, rather than in a
4683 register. This provides compatibility with the MetaWare HighC (hc)
4684 compiler. Use the option @samp{-fpcc-struct-return} for compatibility
4685 with the Portable C Compiler (pcc).
4687 @item -mnohc-struct-return
4688 Return some structures of more than one word in registers, when
4689 convenient. This is the default. For compatibility with the
4690 IBM-supplied compilers, use the option @samp{-fpcc-struct-return} or the
4691 option @samp{-mhc-struct-return}.
4695 @subsection MIPS Options
4696 @cindex MIPS options
4698 These @samp{-m} options are defined for the MIPS family of computers:
4701 @item -mcpu=@var{cpu type}
4702 Assume the defaults for the machine type @var{cpu type} when scheduling
4703 instructions. The choices for @var{cpu type} are @samp{r2000}, @samp{r3000},
4704 @samp{r4000}, @samp{r4400}, @samp{r4600}, and @samp{r6000}. While picking a
4705 specific @var{cpu type} will schedule things appropriately for that
4706 particular chip, the compiler will not generate any code that does not
4707 meet level 1 of the MIPS ISA (instruction set architecture) without
4708 the @samp{-mips2} or @samp{-mips3} switches being used.
4711 Issue instructions from level 1 of the MIPS ISA. This is the default.
4712 @samp{r3000} is the default @var{cpu type} at this ISA level.
4715 Issue instructions from level 2 of the MIPS ISA (branch likely, square
4716 root instructions). @samp{r6000} is the default @var{cpu type} at this
4720 Issue instructions from level 3 of the MIPS ISA (64 bit instructions).
4721 @samp{r4000} is the default @var{cpu type} at this ISA level.
4722 This option does not change the sizes of any of the C data types.
4725 Issue instructions from level 4 of the MIPS ISA. @samp{r8000} is the
4726 default @var{cpu type} at this ISA level.
4729 Assume that 32 32-bit floating point registers are available. This is
4733 Assume that 32 64-bit floating point registers are available. This is
4734 the default when the @samp{-mips3} option is used.
4737 Assume that 32 32-bit general purpose registers are available. This is
4741 Assume that 32 64-bit general purpose registers are available. This is
4742 the default when the @samp{-mips3} option is used.
4745 Types long, int, and pointer are 64 bits. This works only if @samp{-mips3}
4749 Types long and pointer are 64 bits, and type int is 32 bits.
4750 This works only if @samp{-mips3} is also specified.
4756 Generate code for the indicated ABI.
4759 Generate code for the MIPS assembler, and invoke @file{mips-tfile} to
4760 add normal debug information. This is the default for all
4761 platforms except for the OSF/1 reference platform, using the OSF/rose
4762 object format. If the either of the @samp{-gstabs} or @samp{-gstabs+}
4763 switches are used, the @file{mips-tfile} program will encapsulate the
4764 stabs within MIPS ECOFF.
4767 Generate code for the GNU assembler. This is the default on the OSF/1
4768 reference platform, using the OSF/rose object format. Also, this is
4769 the default if the configure option @samp{--with-gnu-as} is used.
4771 @item -msplit-addresses
4772 @itemx -mno-split-addresses
4773 Generate code to load the high and low parts of address constants separately.
4774 This allows @code{gcc} to optimize away redundant loads of the high order
4775 bits of addresses. This optimization requires GNU as and GNU ld.
4776 This optimization is enabled by default for some embedded targets where
4777 GNU as and GNU ld are standard.
4781 The @samp{-mrnames} switch says to output code using the MIPS software
4782 names for the registers, instead of the hardware names (ie, @var{a0}
4783 instead of @var{$4}). The only known assembler that supports this option
4784 is the Algorithmics assembler.
4788 The @samp{-mgpopt} switch says to write all of the data declarations
4789 before the instructions in the text section, this allows the MIPS
4790 assembler to generate one word memory references instead of using two
4791 words for short global or static data items. This is on by default if
4792 optimization is selected.
4796 For each non-inline function processed, the @samp{-mstats} switch
4797 causes the compiler to emit one line to the standard error file to
4798 print statistics about the program (number of registers saved, stack
4803 The @samp{-mmemcpy} switch makes all block moves call the appropriate
4804 string function (@samp{memcpy} or @samp{bcopy}) instead of possibly
4805 generating inline code.
4808 @itemx -mno-mips-tfile
4809 The @samp{-mno-mips-tfile} switch causes the compiler not
4810 postprocess the object file with the @file{mips-tfile} program,
4811 after the MIPS assembler has generated it to add debug support. If
4812 @file{mips-tfile} is not run, then no local variables will be
4813 available to the debugger. In addition, @file{stage2} and
4814 @file{stage3} objects will have the temporary file names passed to the
4815 assembler embedded in the object file, which means the objects will
4816 not compare the same. The @samp{-mno-mips-tfile} switch should only
4817 be used when there are bugs in the @file{mips-tfile} program that
4818 prevents compilation.
4821 Generate output containing library calls for floating point.
4822 @strong{Warning:} the requisite libraries are not part of GNU CC.
4823 Normally the facilities of the machine's usual C compiler are used, but
4824 this can't be done directly in cross-compilation. You must make your
4825 own arrangements to provide suitable library functions for
4829 Generate output containing floating point instructions. This is the
4830 default if you use the unmodified sources.
4833 @itemx -mno-abicalls
4834 Emit (or do not emit) the pseudo operations @samp{.abicalls},
4835 @samp{.cpload}, and @samp{.cprestore} that some System V.4 ports use for
4836 position independent code.
4839 @itemx -mno-long-calls
4840 Do all calls with the @samp{JALR} instruction, which requires
4841 loading up a function's address into a register before the call.
4842 You need to use this switch, if you call outside of the current
4843 512 megabyte segment to functions that are not through pointers.
4846 @itemx -mno-half-pic
4847 Put pointers to extern references into the data section and load them
4848 up, rather than put the references in the text section.
4850 @item -membedded-pic
4851 @itemx -mno-embedded-pic
4852 Generate PIC code suitable for some embedded systems. All calls are
4853 made using PC relative address, and all data is addressed using the $gp
4854 register. No more than 65536 bytes of global data may be used. This
4855 requires GNU as and GNU ld which do most of the work. This currently
4856 only works on targets which use ECOFF; it does not work with ELF.
4858 @item -membedded-data
4859 @itemx -mno-embedded-data
4860 Allocate variables to the read-only data section first if possible, then
4861 next in the small data section if possible, otherwise in data. This gives
4862 slightly slower code than the default, but reduces the amount of RAM required
4863 when executing, and thus may be preferred for some embedded systems.
4865 @item -msingle-float
4866 @itemx -mdouble-float
4867 The @samp{-msingle-float} switch tells gcc to assume that the floating
4868 point coprocessor only supports single precision operations, as on the
4869 @samp{r4650} chip. The @samp{-mdouble-float} switch permits gcc to use
4870 double precision operations. This is the default.
4874 Permit use of the @samp{mad}, @samp{madu} and @samp{mul} instructions,
4875 as on the @samp{r4650} chip.
4878 Turns on @samp{-msingle-float}, @samp{-mmad}, and, at least for now,
4882 Compile code for the processor in little endian mode.
4883 The requisite libraries are assumed to exist.
4886 Compile code for the processor in big endian mode.
4887 The requisite libraries are assumed to exist.
4890 @cindex smaller data references (MIPS)
4891 @cindex gp-relative references (MIPS)
4892 Put global and static items less than or equal to @var{num} bytes into
4893 the small data or bss sections instead of the normal data or bss
4894 section. This allows the assembler to emit one word memory reference
4895 instructions based on the global pointer (@var{gp} or @var{$28}),
4896 instead of the normal two words used. By default, @var{num} is 8 when
4897 the MIPS assembler is used, and 0 when the GNU assembler is used. The
4898 @samp{-G @var{num}} switch is also passed to the assembler and linker.
4899 All modules should be compiled with the same @samp{-G @var{num}}
4903 Tell the MIPS assembler to not run its preprocessor over user
4904 assembler files (with a @samp{.s} suffix) when assembling them.
4908 These options are defined by the macro
4909 @code{TARGET_SWITCHES} in the machine description. The default for the
4910 options is also defined by that macro, which enables you to change the
4915 @subsection Intel 386 Options
4916 @cindex i386 Options
4917 @cindex Intel 386 Options
4919 These @samp{-m} options are defined for the i386 family of computers:
4922 @item -mcpu=@var{cpu type}
4923 Assume the defaults for the machine type @var{cpu type} when scheduling
4924 instructions. The choices for @var{cpu type} are: @samp{i386},
4925 @samp{i486}, @samp{i586} (@samp{pentium}), @samp{pentium}, @samp{i686}
4926 (@samp{pentiumpro}) and @samp{pentiumpro}. While picking a specific
4927 @var{cpu type} will schedule things appropriately for that particular
4928 chip, the compiler will not generate any code that does not run on the
4929 i386 without the @samp{-march=@var{cpu type}} option being used.
4931 @item -march=@var{cpu type}
4932 Generate instructions for the machine type @var{cpu type}. The choices
4933 for @var{cpu type} are: @samp{i386}, @samp{i486}, @samp{pentium}, and
4934 @samp{pentiumpro}. Specifying @samp{-march=@var{cpu type}} implies
4935 @samp{-mcpu=@var{cpu type}}.
4941 Synonyms for -mcpu=i386, -mcpu=i486, -mcpu=pentium, and -mcpu=pentiumpro
4946 Control whether or not the compiler uses IEEE floating point
4947 comparisons. These handle correctly the case where the result of a
4948 comparison is unordered.
4951 Generate output containing library calls for floating point.
4952 @strong{Warning:} the requisite libraries are not part of GNU CC.
4953 Normally the facilities of the machine's usual C compiler are used, but
4954 this can't be done directly in cross-compilation. You must make your
4955 own arrangements to provide suitable library functions for
4958 On machines where a function returns floating point results in the 80387
4959 register stack, some floating point opcodes may be emitted even if
4960 @samp{-msoft-float} is used.
4962 @item -mno-fp-ret-in-387
4963 Do not use the FPU registers for return values of functions.
4965 The usual calling convention has functions return values of types
4966 @code{float} and @code{double} in an FPU register, even if there
4967 is no FPU. The idea is that the operating system should emulate
4970 The option @samp{-mno-fp-ret-in-387} causes such values to be returned
4971 in ordinary CPU registers instead.
4973 @item -mno-fancy-math-387
4974 Some 387 emulators do not support the @code{sin}, @code{cos} and
4975 @code{sqrt} instructions for the 387. Specify this option to avoid
4976 generating those instructions. This option is the default on FreeBSD.
4977 As of revision 2.6.1, these instructions are not generated unless you
4978 also use the @samp{-ffast-math} switch.
4980 @item -malign-double
4981 @itemx -mno-align-double
4982 Control whether GNU CC aligns @code{double}, @code{long double}, and
4983 @code{long long} variables on a two word boundary or a one word
4984 boundary. Aligning @code{double} variables on a two word boundary will
4985 produce code that runs somewhat faster on a @samp{Pentium} at the
4986 expense of more memory.
4988 @strong{Warning:} if you use the @samp{-malign-double} switch,
4989 structures containing the above types will be aligned differently than
4990 the published application binary interface specifications for the 386.
4993 @itemx -mno-svr3-shlib
4994 Control whether GNU CC places uninitialized locals into @code{bss} or
4995 @code{data}. @samp{-msvr3-shlib} places these locals into @code{bss}.
4996 These options are meaningful only on System V Release 3.
4998 @item -mno-wide-multiply
4999 @itemx -mwide-multiply
5000 Control whether GNU CC uses the @code{mul} and @code{imul} that produce
5001 64 bit results in @code{eax:edx} from 32 bit operands to do @code{long
5002 long} multiplies and 32-bit division by constants.
5005 Use a different function-calling convention, in which functions that
5006 take a fixed number of arguments return with the @code{ret} @var{num}
5007 instruction, which pops their arguments while returning. This saves one
5008 instruction in the caller since there is no need to pop the arguments
5011 You can specify that an individual function is called with this calling
5012 sequence with the function attribute @samp{stdcall}. You can also
5013 override the @samp{-mrtd} option by using the function attribute
5014 @samp{cdecl}. @xref{Function Attributes}
5016 @strong{Warning:} this calling convention is incompatible with the one
5017 normally used on Unix, so you cannot use it if you need to call
5018 libraries compiled with the Unix compiler.
5020 Also, you must provide function prototypes for all functions that
5021 take variable numbers of arguments (including @code{printf});
5022 otherwise incorrect code will be generated for calls to those
5025 In addition, seriously incorrect code will result if you call a
5026 function with too many arguments. (Normally, extra arguments are
5027 harmlessly ignored.)
5029 @item -mreg-alloc=@var{regs}
5030 Control the default allocation order of integer registers. The
5031 string @var{regs} is a series of letters specifying a register. The
5032 supported letters are: @code{a} allocate EAX; @code{b} allocate EBX;
5033 @code{c} allocate ECX; @code{d} allocate EDX; @code{S} allocate ESI;
5034 @code{D} allocate EDI; @code{B} allocate EBP.
5036 @item -mregparm=@var{num}
5037 Control how many registers are used to pass integer arguments. By
5038 default, no registers are used to pass arguments, and at most 3
5039 registers can be used. You can control this behavior for a specific
5040 function by using the function attribute @samp{regparm}. @xref{Function Attributes}
5042 @strong{Warning:} if you use this switch, and
5043 @var{num} is nonzero, then you must build all modules with the same
5044 value, including any libraries. This includes the system libraries and
5047 @item -malign-loops=@var{num}
5048 Align loops to a 2 raised to a @var{num} byte boundary. If
5049 @samp{-malign-loops} is not specified, the default is 2 unless
5050 gas 2.8 (or later) is being used in which case the default is
5051 to align the loop on a 16 byte boundary if it is less than 8
5054 @item -malign-jumps=@var{num}
5055 Align instructions that are only jumped to to a 2 raised to a @var{num}
5056 byte boundary. If @samp{-malign-jumps} is not specified, the default is
5057 2 if optimizing for a 386, and 4 if optimizing for a 486 unless
5058 gas 2.8 (or later) is being used in which case the default is
5059 to align the instruction on a 16 byte boundary if it is less
5062 @item -malign-functions=@var{num}
5063 Align the start of functions to a 2 raised to @var{num} byte boundary.
5064 If @samp{-malign-functions} is not specified, the default is 2 if optimizing
5065 for a 386, and 4 if optimizing for a 486.
5069 @subsection HPPA Options
5070 @cindex HPPA Options
5072 These @samp{-m} options are defined for the HPPA family of computers:
5076 Generate code for a PA 1.0 processor.
5079 Generate code for a PA 1.1 processor.
5082 Generate code suitable for big switch tables. Use this option only if
5083 the assembler/linker complain about out of range branches within a switch
5086 @item -mjump-in-delay
5087 Fill delay slots of function calls with unconditional jump instructions
5088 by modifying the return pointer for the function call to be the target
5089 of the conditional jump.
5091 @item -mdisable-fpregs
5092 Prevent floating point registers from being used in any manner. This is
5093 necessary for compiling kernels which perform lazy context switching of
5094 floating point registers. If you use this option and attempt to perform
5095 floating point operations, the compiler will abort.
5097 @item -mdisable-indexing
5098 Prevent the compiler from using indexing address modes. This avoids some
5099 rather obscure problems when compiling MIG generated code under MACH.
5101 @item -mno-space-regs
5102 Generate code that assumes the target has no space registers. This allows
5103 GCC to generate faster indirect calls and use unscaled index address modes.
5105 Such code is suitable for level 0 PA systems and kernels.
5107 @item -mfast-indirect-calls
5108 Generate code that assumes calls never cross space boundaries. This
5109 allows GCC to emit code which performs faster indirect calls.
5111 This option will not work in the presense of shared libraries or nested
5115 Optimize for space rather than execution time. Currently this only
5116 enables out of line function prologues and epilogues. This option is
5117 incompatible with PIC code generation and profiling.
5119 @item -mlong-load-store
5120 Generate 3-instruction load and store sequences as sometimes required by
5121 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
5124 @item -mportable-runtime
5125 Use the portable calling conventions proposed by HP for ELF systems.
5128 Enable the use of assembler directives only GAS understands.
5130 @item -mschedule=@var{cpu type}
5131 Schedule code according to the constraints for the machine type
5132 @var{cpu type}. The choices for @var{cpu type} are @samp{700} for
5133 7@var{n}0 machines, @samp{7100} for 7@var{n}5 machines, and @samp{7100LC}
5134 for 7@var{n}2 machines. @samp{7100} is the default for @var{cpu type}.
5136 Note the @samp{7100LC} scheduling information is incomplete and using
5137 @samp{7100LC} often leads to bad schedules. For now it's probably best
5138 to use @samp{7100} instead of @samp{7100LC} for the 7@var{n}2 machines.
5141 Enable the optimization pass in the HPUX linker. Note this makes symbolic
5142 debugging impossible. It also triggers a bug in the HPUX 8 and HPUX 9 linkers
5143 in which they give bogus error messages when linking some programs.
5146 Generate output containing library calls for floating point.
5147 @strong{Warning:} the requisite libraries are not available for all HPPA
5148 targets. Normally the facilities of the machine's usual C compiler are
5149 used, but this cannot be done directly in cross-compilation. You must make
5150 your own arrangements to provide suitable library functions for
5151 cross-compilation. The embedded target @samp{hppa1.1-*-pro}
5152 does provide software floating point support.
5154 @samp{-msoft-float} changes the calling convention in the output file;
5155 therefore, it is only useful if you compile @emph{all} of a program with
5156 this option. In particular, you need to compile @file{libgcc.a}, the
5157 library that comes with GNU CC, with @samp{-msoft-float} in order for
5161 @node Intel 960 Options
5162 @subsection Intel 960 Options
5164 These @samp{-m} options are defined for the Intel 960 implementations:
5167 @item -m@var{cpu type}
5168 Assume the defaults for the machine type @var{cpu type} for some of
5169 the other options, including instruction scheduling, floating point
5170 support, and addressing modes. The choices for @var{cpu type} are
5171 @samp{ka}, @samp{kb}, @samp{mc}, @samp{ca}, @samp{cf},
5172 @samp{sa}, and @samp{sb}.
5178 The @samp{-mnumerics} option indicates that the processor does support
5179 floating-point instructions. The @samp{-msoft-float} option indicates
5180 that floating-point support should not be assumed.
5182 @item -mleaf-procedures
5183 @itemx -mno-leaf-procedures
5184 Do (or do not) attempt to alter leaf procedures to be callable with the
5185 @code{bal} instruction as well as @code{call}. This will result in more
5186 efficient code for explicit calls when the @code{bal} instruction can be
5187 substituted by the assembler or linker, but less efficient code in other
5188 cases, such as calls via function pointers, or using a linker that doesn't
5189 support this optimization.
5192 @itemx -mno-tail-call
5193 Do (or do not) make additional attempts (beyond those of the
5194 machine-independent portions of the compiler) to optimize tail-recursive
5195 calls into branches. You may not want to do this because the detection of
5196 cases where this is not valid is not totally complete. The default is
5197 @samp{-mno-tail-call}.
5199 @item -mcomplex-addr
5200 @itemx -mno-complex-addr
5201 Assume (or do not assume) that the use of a complex addressing mode is a
5202 win on this implementation of the i960. Complex addressing modes may not
5203 be worthwhile on the K-series, but they definitely are on the C-series.
5204 The default is currently @samp{-mcomplex-addr} for all processors except
5208 @itemx -mno-code-align
5209 Align code to 8-byte boundaries for faster fetching (or don't bother).
5210 Currently turned on by default for C-series implementations only.
5213 @item -mclean-linkage
5214 @itemx -mno-clean-linkage
5215 These options are not fully implemented.
5219 @itemx -mic2.0-compat
5220 @itemx -mic3.0-compat
5221 Enable compatibility with iC960 v2.0 or v3.0.
5225 Enable compatibility with the iC960 assembler.
5227 @item -mstrict-align
5228 @itemx -mno-strict-align
5229 Do not permit (do permit) unaligned accesses.
5232 Enable structure-alignment compatibility with Intel's gcc release version
5233 1.3 (based on gcc 1.37). This option implies @samp{-mstrict-align}.
5236 @node DEC Alpha Options
5237 @subsection DEC Alpha Options
5239 These @samp{-m} options are defined for the DEC Alpha implementations:
5242 @item -mno-soft-float
5244 Use (do not use) the hardware floating-point instructions for
5245 floating-point operations. When @code{-msoft-float} is specified,
5246 functions in @file{libgcc1.c} will be used to perform floating-point
5247 operations. Unless they are replaced by routines that emulate the
5248 floating-point operations, or compiled in such a way as to call such
5249 emulations routines, these routines will issue floating-point
5250 operations. If you are compiling for an Alpha without floating-point
5251 operations, you must ensure that the library is built so as not to call
5254 Note that Alpha implementations without floating-point operations are
5255 required to have floating-point registers.
5259 Generate code that uses (does not use) the floating-point register set.
5260 @code{-mno-fp-regs} implies @code{-msoft-float}. If the floating-point
5261 register set is not used, floating point operands are passed in integer
5262 registers as if they were integers and floating-point results are passed
5263 in $0 instead of $f0. This is a non-standard calling sequence, so any
5264 function with a floating-point argument or return value called by code
5265 compiled with @code{-mno-fp-regs} must also be compiled with that
5268 A typical use of this option is building a kernel that does not use,
5269 and hence need not save and restore, any floating-point registers.
5272 The Alpha architecture implements floating-point hardware optimized for
5273 maximum performance. It is mostly compliant with the IEEE floating
5274 point standard. However, for full compliance, software assistance is
5275 required. This option generates code fully IEEE compliant code
5276 @emph{except} that the @var{inexact flag} is not maintained (see below).
5277 If this option is turned on, the CPP macro @code{_IEEE_FP} is defined
5278 during compilation. The option is a shorthand for: @samp{-D_IEEE_FP
5279 -mfp-trap-mode=su -mtrap-precision=i -mieee-conformant}. The resulting
5280 code is less efficient but is able to correctly support denormalized
5281 numbers and exceptional IEEE values such as not-a-number and plus/minus
5282 infinity. Other Alpha compilers call this option
5283 @code{-ieee_with_no_inexact}.
5285 @item -mieee-with-inexact
5286 @c overfull hbox here --bob 22 jul96
5287 @c original text between ignore ... end ignore
5289 This is like @samp{-mieee} except the generated code also maintains the
5290 IEEE @var{inexact flag}. Turning on this option causes the generated
5291 code to implement fully-compliant IEEE math. The option is a shorthand
5292 for @samp{-D_IEEE_FP -D_IEEE_FP_INEXACT} plus @samp{-mieee-conformant},
5293 @samp{-mfp-trap-mode=sui}, and @samp{-mtrap-precision=i}. On some Alpha
5294 implementations the resulting code may execute significantly slower than
5295 the code generated by default. Since there is very little code that
5296 depends on the @var{inexact flag}, you should normally not specify this
5297 option. Other Alpha compilers call this option
5298 @samp{-ieee_with_inexact}.
5300 @c changed paragraph
5301 This is like @samp{-mieee} except the generated code also maintains the
5302 IEEE @var{inexact flag}. Turning on this option causes the generated
5303 code to implement fully-compliant IEEE math. The option is a shorthand
5304 for @samp{-D_IEEE_FP -D_IEEE_FP_INEXACT} plus the three following:
5305 @samp{-mieee-conformant},
5306 @samp{-mfp-trap-mode=sui},
5307 and @samp{-mtrap-precision=i}.
5308 On some Alpha implementations the resulting code may execute
5309 significantly slower than the code generated by default. Since there
5310 is very little code that depends on the @var{inexact flag}, you should
5311 normally not specify this option. Other Alpha compilers call this
5312 option @samp{-ieee_with_inexact}.
5313 @c end changes to prevent overfull hboxes
5315 @item -mfp-trap-mode=@var{trap mode}
5316 This option controls what floating-point related traps are enabled.
5317 Other Alpha compilers call this option @samp{-fptm }@var{trap mode}.
5318 The trap mode can be set to one of four values:
5322 This is the default (normal) setting. The only traps that are enabled
5323 are the ones that cannot be disabled in software (e.g., division by zero
5327 In addition to the traps enabled by @samp{n}, underflow traps are enabled
5331 Like @samp{su}, but the instructions are marked to be safe for software
5332 completion (see Alpha architecture manual for details).
5335 Like @samp{su}, but inexact traps are enabled as well.
5338 @item -mfp-rounding-mode=@var{rounding mode}
5339 Selects the IEEE rounding mode. Other Alpha compilers call this option
5340 @samp{-fprm }@var{rounding mode}. The @var{rounding mode} can be one
5345 Normal IEEE rounding mode. Floating point numbers are rounded towards
5346 the nearest machine number or towards the even machine number in case
5350 Round towards minus infinity.
5353 Chopped rounding mode. Floating point numbers are rounded towards zero.
5356 Dynamic rounding mode. A field in the floating point control register
5357 (@var{fpcr}, see Alpha architecture reference manual) controls the
5358 rounding mode in effect. The C library initializes this register for
5359 rounding towards plus infinity. Thus, unless your program modifies the
5360 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.@end table
5362 @item -mtrap-precision=@var{trap precision}
5363 In the Alpha architecture, floating point traps are imprecise. This
5364 means without software assistance it is impossible to recover from a
5365 floating trap and program execution normally needs to be terminated.
5366 GNU CC can generate code that can assist operating system trap handlers
5367 in determining the exact location that caused a floating point trap.
5368 Depending on the requirements of an application, different levels of
5369 precisions can be selected:
5373 Program precision. This option is the default and means a trap handler
5374 can only identify which program caused a floating point exception.
5377 Function precision. The trap handler can determine the function that
5378 caused a floating point exception.
5381 Instruction precision. The trap handler can determine the exact
5382 instruction that caused a floating point exception.
5385 Other Alpha compilers provide the equivalent options called
5386 @samp{-scope_safe} and @samp{-resumption_safe}.
5388 @item -mieee-conformant
5389 This option marks the generated code as IEEE conformant. You must not
5390 use this option unless you also specify @samp{-mtrap-precision=i} and either
5391 @samp{-mfp-trap-mode=su} or @samp{-mfp-trap-mode=sui}. Its only effect
5392 is to emit the line @samp{.eflag 48} in the function prologue of the
5393 generated assembly file. Under DEC Unix, this has the effect that
5394 IEEE-conformant math library routines will be linked in.
5396 @item -mbuild-constants
5397 Normally GNU CC examines a 32- or 64-bit integer constant to
5398 see if it can construct it from smaller constants in two or three
5399 instructions. If it cannot, it will output the constant as a literal and
5400 generate code to load it from the data segment at runtime.
5402 Use this option to require GNU CC to construct @emph{all} integer constants
5403 using code, even if it takes more instructions (the maximum is six).
5405 You would typically use this option to build a shared library dynamic
5406 loader. Itself a shared library, it must relocate itself in memory
5407 before it can find the variables and constants in its own data segment.
5411 Select whether to generate code to be assembled by the vendor-supplied
5412 assembler (@samp{-malpha-as}) or by the GNU assembler @samp{-mgas}.
5420 Indicate whether GNU CC should generate code to use the optional BWX,
5421 CIX, and MAX instruction sets. The default is to use the instruction sets
5422 supported by the CPU type specified via @samp{-mcpu=} option or that
5423 of the CPU on which GNU CC was built if none was specified.
5425 @item -mcpu=@var{cpu_type}
5426 Set the instruction set, register set, and instruction scheduling
5427 parameters for machine type @var{cpu_type}. You can specify either the
5428 @samp{EV} style name or the corresponding chip number. GNU CC
5429 supports scheduling parameters for the EV4 and EV5 family of processors
5430 and will choose the default values for the instruction set from
5431 the processor you specify. If you do not specify a processor type,
5432 GNU CC will default to the processor on which the compiler was built.
5434 Supported values for @var{cpu_type} are
5439 Schedules as an EV4 and has no instruction set extensions.
5443 Schedules as an EV5 and has no instruction set extensions.
5447 Schedules as an EV5 and supports the BWX extension.
5452 Schedules as an EV5 and supports the BWX and MAX extensions.
5456 Schedules as an EV5 (until Digital releases the scheduling parameters
5457 for the EV6) and supports the BWX, CIX, and MAX extensions.
5460 @item -mmemory-latency=@var{time}
5461 Sets the latency the scheduler should assume for typical memory
5462 references as seen by the application. This number is highly
5463 dependant on the memory access patterns used by the application
5464 and the size of the external cache on the machine.
5466 Valid options for @var{time} are
5470 A decimal number representing clock cycles.
5476 The compiler contains estimates of the number of clock cycles for
5477 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
5478 (also called Dcache, Scache, and Bcache), as well as to main memory.
5479 Note that L3 is only valid for EV5.
5484 @node Clipper Options
5485 @subsection Clipper Options
5487 These @samp{-m} options are defined for the Clipper implementations:
5491 Produce code for a C300 Clipper processor. This is the default.
5494 Produce code for a C400 Clipper processor i.e. use floating point
5498 @node H8/300 Options
5499 @subsection H8/300 Options
5501 These @samp{-m} options are defined for the H8/300 implementations:
5505 Shorten some address references at link time, when possible; uses the
5506 linker option @samp{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
5507 ld.info, Using ld}, for a fuller description.
5510 Generate code for the H8/300H.
5513 Generate code for the H8/S.
5516 Make @code{int} data 32 bits by default.
5519 On the h8/300h, use the same alignment rules as for the h8/300.
5520 The default for the h8/300h is to align longs and floats on 4 byte boundaries.
5521 @samp{-malign-300} causes them to be aligned on 2 byte boundaries.
5522 This option has no effect on the h8/300.
5526 @subsection SH Options
5528 These @samp{-m} options are defined for the SH implementations:
5532 Generate code for the SH1.
5535 Generate code for the SH2.
5538 Generate code for the SH3.
5541 Generate code for the SH3e.
5544 Compile code for the processor in big endian mode.
5547 Compile code for the processor in little endian mode.
5550 Align doubles at 64 bit boundaries. Note that this changes the calling
5551 conventions, and thus some functions from the standard C library will
5552 not work unless you recompile it first with -mdalign.
5555 Shorten some address references at link time, when possible; uses the
5556 linker option @samp{-relax}.
5559 @node System V Options
5560 @subsection Options for System V
5562 These additional options are available on System V Release 4 for
5563 compatibility with other compilers on those systems:
5567 Create a shared object.
5568 It is recommended that @samp{-symbolic} or @samp{-shared} be used instead.
5571 Identify the versions of each tool used by the compiler, in a
5572 @code{.ident} assembler directive in the output.
5575 Refrain from adding @code{.ident} directives to the output file (this is
5578 @item -YP,@var{dirs}
5579 Search the directories @var{dirs}, and no others, for libraries
5580 specified with @samp{-l}.
5583 Look in the directory @var{dir} to find the M4 preprocessor.
5584 The assembler uses this option.
5585 @c This is supposed to go with a -Yd for predefined M4 macro files, but
5586 @c the generic assembler that comes with Solaris takes just -Ym.
5590 @subsection V850 Options
5591 @cindex V850 Options
5593 These @samp{-m} options are defined for V850 implementations:
5597 @itemx -mno-long-calls
5598 Treat all calls as being far away (near). If calls are assumed to be
5599 far away, the compiler will always load the functions address up into a
5600 register, and call indirect through the pointer.
5604 Do not optimize (do optimize) basic blocks that use the same index
5605 pointer 4 or more times to copy pointer into the @code{ep} register, and
5606 use the shorter @code{sld} and @code{sst} instructions. The @samp{-mep}
5607 option is on by default if you optimize.
5609 @item -mno-prolog-function
5610 @itemx -mprolog-function
5611 Do not use (do use) external functions to save and restore registers at
5612 the prolog and epilog of a function. The external functions are slower,
5613 but use less code space if more than one function saves the same number
5614 of registers. The @samp{-mprolog-function} option is on by default if
5618 Try to make the code as small as possible. At present, this just turns
5619 on the @samp{-mep} and @samp{-mprolog-function} options.
5622 Put static or global variables whose size is @var{n} bytes or less into
5623 the tiny data area that register @code{ep} points to. The tiny data
5624 area can hold up to 256 bytes in total (128 bytes for byte references).
5627 Put static or global variables whose size is @var{n} bytes or less into
5628 the small data area that register @code{gp} points to. The small data
5629 area can hold up to 64 kilobytes.
5632 Put static or global variables whose size is @var{n} bytes or less into
5633 the first 32 kilobytes of memory.
5636 Specify that the target processor is the V850.
5639 Generate code suitable for big switch tables. Use this option only if
5640 the assembler/linker complain about out of range branches within a switch
5645 @subsection ARC Options
5648 These options are defined for ARC implementations:
5652 Compile code for little endian mode. This is the default.
5655 Compile code for big endian mode.
5658 Prepend the name of the cpu to all public symbol names.
5659 In multiple-processor systems, there are many ARC variants with different
5660 instruction and register set characteristics. This flag prevents code
5661 compiled for one cpu to be linked with code compiled for another.
5662 No facility exists for handling variants that are "almost identical".
5663 This is an all or nothing option.
5665 @item -mcpu=@var{cpu}
5666 Compile code for ARC variant @var{cpu}.
5667 Which variants are supported depend on the configuration.
5668 All variants support @samp{-mcpu=base}, this is the default.
5670 @item -mtext=@var{text section}
5671 @item -mdata=@var{data section}
5672 @item -mrodata=@var{readonly data section}
5673 Put functions, data, and readonly data in @var{text section},
5674 @var{data section}, and @var{readonly data section} respectively
5675 by default. This can be overridden with the @code{section} attribute.
5676 @xref{Variable Attributes}
5681 @node Code Gen Options
5682 @section Options for Code Generation Conventions
5683 @cindex code generation conventions
5684 @cindex options, code generation
5685 @cindex run-time options
5687 These machine-independent options control the interface conventions
5688 used in code generation.
5690 Most of them have both positive and negative forms; the negative form
5691 of @samp{-ffoo} would be @samp{-fno-foo}. In the table below, only
5692 one of the forms is listed---the one which is not the default. You
5693 can figure out the other form by either removing @samp{no-} or adding
5698 Enable exception handling. Generates extra code needed to propagate
5699 exceptions. For some targets, this implies generation of frame unwind
5700 information for all functions. This can produce significant data size
5701 overhead, although it does not affect execution.
5702 If you do not specify this option, it is enabled by
5703 default for languages like C++ which normally require exception handling,
5704 and disabled for languages like C that do not normally require it.
5705 However, when compiling C code that needs to interoperate properly with
5706 exception handlers written in C++, you may need to enable this option.
5707 You may also wish to disable this option is you are compiling older C++
5708 programs that don't use exception handling.
5710 @item -fpcc-struct-return
5711 Return ``short'' @code{struct} and @code{union} values in memory like
5712 longer ones, rather than in registers. This convention is less
5713 efficient, but it has the advantage of allowing intercallability between
5714 GNU CC-compiled files and files compiled with other compilers.
5716 The precise convention for returning structures in memory depends
5717 on the target configuration macros.
5719 Short structures and unions are those whose size and alignment match
5720 that of some integer type.
5722 @item -freg-struct-return
5723 Use the convention that @code{struct} and @code{union} values are
5724 returned in registers when possible. This is more efficient for small
5725 structures than @samp{-fpcc-struct-return}.
5727 If you specify neither @samp{-fpcc-struct-return} nor its contrary
5728 @samp{-freg-struct-return}, GNU CC defaults to whichever convention is
5729 standard for the target. If there is no standard convention, GNU CC
5730 defaults to @samp{-fpcc-struct-return}, except on targets where GNU CC
5731 is the principal compiler. In those cases, we can choose the standard,
5732 and we chose the more efficient register return alternative.
5735 Allocate to an @code{enum} type only as many bytes as it needs for the
5736 declared range of possible values. Specifically, the @code{enum} type
5737 will be equivalent to the smallest integer type which has enough room.
5739 @item -fshort-double
5740 Use the same size for @code{double} as for @code{float}.
5743 Requests that the data and non-@code{const} variables of this
5744 compilation be shared data rather than private data. The distinction
5745 makes sense only on certain operating systems, where shared data is
5746 shared between processes running the same program, while private data
5747 exists in one copy per process.
5750 Allocate even uninitialized global variables in the bss section of the
5751 object file, rather than generating them as common blocks. This has the
5752 effect that if the same variable is declared (without @code{extern}) in
5753 two different compilations, you will get an error when you link them.
5754 The only reason this might be useful is if you wish to verify that the
5755 program will work on other systems which always work this way.
5758 Ignore the @samp{#ident} directive.
5760 @item -fno-gnu-linker
5761 Do not output global initializations (such as C++ constructors and
5762 destructors) in the form used by the GNU linker (on systems where the GNU
5763 linker is the standard method of handling them). Use this option when
5764 you want to use a non-GNU linker, which also requires using the
5765 @code{collect2} program to make sure the system linker includes
5766 constructors and destructors. (@code{collect2} is included in the GNU CC
5767 distribution.) For systems which @emph{must} use @code{collect2}, the
5768 compiler driver @code{gcc} is configured to do this automatically.
5770 @item -finhibit-size-directive
5771 Don't output a @code{.size} assembler directive, or anything else that
5772 would cause trouble if the function is split in the middle, and the
5773 two halves are placed at locations far apart in memory. This option is
5774 used when compiling @file{crtstuff.c}; you should not need to use it
5778 Put extra commentary information in the generated assembly code to
5779 make it more readable. This option is generally only of use to those
5780 who actually need to read the generated assembly code (perhaps while
5781 debugging the compiler itself).
5783 @samp{-fno-verbose-asm}, the default, causes the
5784 extra information to be omitted and is useful when comparing two assembler
5788 Consider all memory references through pointers to be volatile.
5790 @item -fvolatile-global
5791 Consider all memory references to extern and global data items to
5795 @cindex global offset table
5797 Generate position-independent code (PIC) suitable for use in a shared
5798 library, if supported for the target machine. Such code accesses all
5799 constant addresses through a global offset table (GOT). The dynamic
5800 loader resolves the GOT entries when the program starts (the dynamic
5801 loader is not part of GNU CC; it is part of the operating system). If
5802 the GOT size for the linked executable exceeds a machine-specific
5803 maximum size, you get an error message from the linker indicating that
5804 @samp{-fpic} does not work; in that case, recompile with @samp{-fPIC}
5805 instead. (These maximums are 16k on the m88k, 8k on the Sparc, and 32k
5806 on the m68k and RS/6000. The 386 has no such limit.)
5808 Position-independent code requires special support, and therefore works
5809 only on certain machines. For the 386, GNU CC supports PIC for System V
5810 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
5811 position-independent.
5814 If supported for the target machine, emit position-independent code,
5815 suitable for dynamic linking and avoiding any limit on the size of the
5816 global offset table. This option makes a difference on the m68k, m88k,
5819 Position-independent code requires special support, and therefore works
5820 only on certain machines.
5822 @item -ffixed-@var{reg}
5823 Treat the register named @var{reg} as a fixed register; generated code
5824 should never refer to it (except perhaps as a stack pointer, frame
5825 pointer or in some other fixed role).
5827 @var{reg} must be the name of a register. The register names accepted
5828 are machine-specific and are defined in the @code{REGISTER_NAMES}
5829 macro in the machine description macro file.
5831 This flag does not have a negative form, because it specifies a
5834 @item -fcall-used-@var{reg}
5835 Treat the register named @var{reg} as an allocable register that is
5836 clobbered by function calls. It may be allocated for temporaries or
5837 variables that do not live across a call. Functions compiled this way
5838 will not save and restore the register @var{reg}.
5840 It is an error to used this flag with the frame pointer or stack pointer.
5841 Use of this flag for other registers that have fixed pervasive roles in
5842 the machine's execution model will produce disastrous results.
5844 This flag does not have a negative form, because it specifies a
5847 @item -fcall-saved-@var{reg}
5848 Treat the register named @var{reg} as an allocable register saved by
5849 functions. It may be allocated even for temporaries or variables that
5850 live across a call. Functions compiled this way will save and restore
5851 the register @var{reg} if they use it.
5853 It is an error to used this flag with the frame pointer or stack pointer.
5854 Use of this flag for other registers that have fixed pervasive roles in
5855 the machine's execution model will produce disastrous results.
5857 A different sort of disaster will result from the use of this flag for
5858 a register in which function values may be returned.
5860 This flag does not have a negative form, because it specifies a
5864 Pack all structure members together without holes. Usually you would
5865 not want to use this option, since it makes the code suboptimal, and
5866 the offsets of structure members won't agree with system libraries.
5868 @item -fcheck-memory-usage
5869 Generate extra code to check each memory access. GNU CC will generate
5870 code that is suitable for a detector of bad memory accesses such as
5873 You must also specify this option when you compile functions you call that
5874 have side effects. If you do not, you may get erroneous messages from
5875 the detector. Normally, you should compile all your code with this option.
5876 If you use functions from a library that have side-effects (such as
5877 @code{read}), you may not be able to recompile the library and
5878 specify this option. In that case, you can enable the
5879 @samp{-fprefix-function-name} option, which requests GNU CC to encapsulate
5880 your code and make other functions look as if they were compiled with
5881 @samp{-fcheck-memory-usage}. This is done by calling ``stubs'',
5882 which are provided by the detector. If you cannot find or build
5883 stubs for every function you call, you may have to specify
5884 @samp{-fcheck-memory-usage} without @samp{-fprefix-function-name}.
5886 If you specify this option, you can not use the @code{asm} or
5887 @code{__asm__} keywords in functions with memory checking enabled. The
5888 compiler cannot understand what the @code{asm} statement will do, and
5889 therefore cannot generate the appropriate code, so it is rejected.
5890 However, the function attribute @code{no_check_memory_usage} will
5891 disable memory checking within a function, and @code{asm} statements can
5892 be put inside such functions. Inline expansion of a non-checked
5893 function within a checked function is permitted; the inline function's
5894 memory accesses won't be checked, but the rest will.
5896 If you move your @code{asm} statements to non-checked inline functions,
5897 but they do access memory, you can add calls to the support code in your
5898 inline function, to indicate any reads, writes, or copies being done.
5899 These calls would be similar to those done in the stubs described above.
5901 @c FIXME: The support-routine interface is defined by the compiler and
5902 @c should be documented!
5904 @item -fprefix-function-name
5905 Request GNU CC to add a prefix to the symbols generated for function names.
5906 GNU CC adds a prefix to the names of functions defined as well as
5907 functions called. Code compiled with this option and code compiled
5908 without the option can't be linked together, unless or stubs are used.
5910 If you compile the following code with @samp{-fprefix-function-name}
5912 extern void bar (int);
5922 GNU CC will compile the code as if it was written:
5924 extern void prefix_bar (int);
5928 return prefix_bar (a + 5);
5931 This option is designed to be used with @samp{-fcheck-memory-usage}.
5933 @item -finstrument-functions
5934 Generate instrumentation calls for entry and exit to functions. Just
5935 after function entry and just before function exit, the following
5936 profiling functions will be called with the address of the current
5937 function and its call site. (On some platforms,
5938 @code{__builtin_return_address} does not work beyond the current
5939 function, so the call site information may not be available to the
5940 profiling functions otherwise.)
5943 void __cyg_profile_func_enter (void *this_fn, void *call_site);
5944 void __cyg_profile_func_exit (void *this_fn, void *call_site);
5947 The first argument is the address of the start of the current function,
5948 which may be looked up exactly in the symbol table.
5950 This instrumentation is also done for functions expanded inline in other
5951 functions. The profiling calls will indicate where, conceptually, the
5952 inline function is entered and exited. This means that addressable
5953 versions of such functions must be available. If all your uses of a
5954 function are expanded inline, this may mean an additional expansion of
5955 code size. If you use @samp{extern inline} in your C code, an
5956 addressable version of such functions must be provided. (This is
5957 normally the case anyways, but if you get lucky and the optimizer always
5958 expands the functions inline, you might have gotten away without
5959 providing static copies.)
5961 A function may be given the attribute @code{no_instrument_function}, in
5962 which case this instrumentation will not be done. This can be used, for
5963 example, for the profiling functions listed above, high-priority
5964 interrupt routines, and any functions from which the profiling functions
5965 cannot safely be called (perhaps signal handlers, if the profiling
5966 routines generate output or allocate memory).
5969 Generate code to verify that you do not go beyond the boundary of the
5970 stack. You should specify this flag if you are running in an
5971 environment with multiple threads, but only rarely need to specify it in
5972 a single-threaded environment since stack overflow is automatically
5973 detected on nearly all systems if there is only one stack.
5975 @cindex aliasing of parameters
5976 @cindex parameters, aliased
5977 @item -fargument-alias
5978 @itemx -fargument-noalias
5979 @itemx -fargument-noalias-global
5980 Specify the possible relationships among parameters and between
5981 parameters and global data.
5983 @samp{-fargument-alias} specifies that arguments (parameters) may
5984 alias each other and may alias global storage.
5985 @samp{-fargument-noalias} specifies that arguments do not alias
5986 each other, but may alias global storage.
5987 @samp{-fargument-noalias-global} specifies that arguments do not
5988 alias each other and do not alias global storage.
5990 Each language will automatically use whatever option is required by
5991 the language standard. You should not need to use these options yourself.
5994 @node Environment Variables
5995 @section Environment Variables Affecting GNU CC
5996 @cindex environment variables
5998 This section describes several environment variables that affect how GNU
5999 CC operates. Some of them work by specifying directories or prefixes to use
6000 when searching for various kinds of files. Some are used to specify other
6001 aspects of the compilation environment.
6004 Note that you can also specify places to search using options such as
6005 @samp{-B}, @samp{-I} and @samp{-L} (@pxref{Directory Options}). These
6006 take precedence over places specified using environment variables, which
6007 in turn take precedence over those specified by the configuration of GNU
6011 Note that you can also specify places to search using options such as
6012 @samp{-B}, @samp{-I} and @samp{-L} (@pxref{Directory Options}). These
6013 take precedence over places specified using environment variables, which
6014 in turn take precedence over those specified by the configuration of GNU
6021 If @code{TMPDIR} is set, it specifies the directory to use for temporary
6022 files. GNU CC uses temporary files to hold the output of one stage of
6023 compilation which is to be used as input to the next stage: for example,
6024 the output of the preprocessor, which is the input to the compiler
6027 @item GCC_EXEC_PREFIX
6028 @findex GCC_EXEC_PREFIX
6029 If @code{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
6030 names of the subprograms executed by the compiler. No slash is added
6031 when this prefix is combined with the name of a subprogram, but you can
6032 specify a prefix that ends with a slash if you wish.
6034 If GNU CC cannot find the subprogram using the specified prefix, it
6035 tries looking in the usual places for the subprogram.
6037 The default value of @code{GCC_EXEC_PREFIX} is
6038 @file{@var{prefix}/lib/gcc-lib/} where @var{prefix} is the value
6039 of @code{prefix} when you ran the @file{configure} script.
6041 Other prefixes specified with @samp{-B} take precedence over this prefix.
6043 This prefix is also used for finding files such as @file{crt0.o} that are
6046 In addition, the prefix is used in an unusual way in finding the
6047 directories to search for header files. For each of the standard
6048 directories whose name normally begins with @samp{/usr/local/lib/gcc-lib}
6049 (more precisely, with the value of @code{GCC_INCLUDE_DIR}), GNU CC tries
6050 replacing that beginning with the specified prefix to produce an
6051 alternate directory name. Thus, with @samp{-Bfoo/}, GNU CC will search
6052 @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}.
6053 These alternate directories are searched first; the standard directories
6057 @findex COMPILER_PATH
6058 The value of @code{COMPILER_PATH} is a colon-separated list of
6059 directories, much like @code{PATH}. GNU CC tries the directories thus
6060 specified when searching for subprograms, if it can't find the
6061 subprograms using @code{GCC_EXEC_PREFIX}.
6064 @findex LIBRARY_PATH
6065 The value of @code{LIBRARY_PATH} is a colon-separated list of
6066 directories, much like @code{PATH}. When configured as a native compiler,
6067 GNU CC tries the directories thus specified when searching for special
6068 linker files, if it can't find them using @code{GCC_EXEC_PREFIX}. Linking
6069 using GNU CC also uses these directories when searching for ordinary
6070 libraries for the @samp{-l} option (but directories specified with
6071 @samp{-L} come first).
6073 @item C_INCLUDE_PATH
6074 @itemx CPLUS_INCLUDE_PATH
6075 @itemx OBJC_INCLUDE_PATH
6076 @findex C_INCLUDE_PATH
6077 @findex CPLUS_INCLUDE_PATH
6078 @findex OBJC_INCLUDE_PATH
6079 @c @itemx OBJCPLUS_INCLUDE_PATH
6080 These environment variables pertain to particular languages. Each
6081 variable's value is a colon-separated list of directories, much like
6082 @code{PATH}. When GNU CC searches for header files, it tries the
6083 directories listed in the variable for the language you are using, after
6084 the directories specified with @samp{-I} but before the standard header
6087 @item DEPENDENCIES_OUTPUT
6088 @findex DEPENDENCIES_OUTPUT
6089 @cindex dependencies for make as output
6090 If this variable is set, its value specifies how to output dependencies
6091 for Make based on the header files processed by the compiler. This
6092 output looks much like the output from the @samp{-M} option
6093 (@pxref{Preprocessor Options}), but it goes to a separate file, and is
6094 in addition to the usual results of compilation.
6096 The value of @code{DEPENDENCIES_OUTPUT} can be just a file name, in
6097 which case the Make rules are written to that file, guessing the target
6098 name from the source file name. Or the value can have the form
6099 @samp{@var{file} @var{target}}, in which case the rules are written to
6100 file @var{file} using @var{target} as the target name.
6104 @cindex locale definition
6105 This variable is used to pass locale information to the compiler. One way in
6106 which this information is used is to determine the character set to be used
6107 when character literals, string literals and comments are parsed in C and C++.
6108 When the compiler is configured to allow multibyte characters,
6109 the following values for @code{LANG} are recognized:
6113 Recognize JIS characters.
6115 Recognize SJIS characters.
6117 Recognize EUCJP characters.
6120 If @code{LANG} is not defined, or if it has some other value, then the
6121 compiler will use mblen and mbtowc as defined by the default locale to
6122 recognize and translate multibyte characters.
6125 @node Running Protoize
6126 @section Running Protoize
6128 The program @code{protoize} is an optional part of GNU C. You can use
6129 it to add prototypes to a program, thus converting the program to ANSI
6130 C in one respect. The companion program @code{unprotoize} does the
6131 reverse: it removes argument types from any prototypes that are found.
6133 When you run these programs, you must specify a set of source files as
6134 command line arguments. The conversion programs start out by compiling
6135 these files to see what functions they define. The information gathered
6136 about a file @var{foo} is saved in a file named @file{@var{foo}.X}.
6138 After scanning comes actual conversion. The specified files are all
6139 eligible to be converted; any files they include (whether sources or
6140 just headers) are eligible as well.
6142 But not all the eligible files are converted. By default,
6143 @code{protoize} and @code{unprotoize} convert only source and header
6144 files in the current directory. You can specify additional directories
6145 whose files should be converted with the @samp{-d @var{directory}}
6146 option. You can also specify particular files to exclude with the
6147 @samp{-x @var{file}} option. A file is converted if it is eligible, its
6148 directory name matches one of the specified directory names, and its
6149 name within the directory has not been excluded.
6151 Basic conversion with @code{protoize} consists of rewriting most
6152 function definitions and function declarations to specify the types of
6153 the arguments. The only ones not rewritten are those for varargs
6156 @code{protoize} optionally inserts prototype declarations at the
6157 beginning of the source file, to make them available for any calls that
6158 precede the function's definition. Or it can insert prototype
6159 declarations with block scope in the blocks where undeclared functions
6162 Basic conversion with @code{unprotoize} consists of rewriting most
6163 function declarations to remove any argument types, and rewriting
6164 function definitions to the old-style pre-ANSI form.
6166 Both conversion programs print a warning for any function declaration or
6167 definition that they can't convert. You can suppress these warnings
6170 The output from @code{protoize} or @code{unprotoize} replaces the
6171 original source file. The original file is renamed to a name ending
6172 with @samp{.save}. If the @samp{.save} file already exists, then
6173 the source file is simply discarded.
6175 @code{protoize} and @code{unprotoize} both depend on GNU CC itself to
6176 scan the program and collect information about the functions it uses.
6177 So neither of these programs will work until GNU CC is installed.
6179 Here is a table of the options you can use with @code{protoize} and
6180 @code{unprotoize}. Each option works with both programs unless
6184 @item -B @var{directory}
6185 Look for the file @file{SYSCALLS.c.X} in @var{directory}, instead of the
6186 usual directory (normally @file{/usr/local/lib}). This file contains
6187 prototype information about standard system functions. This option
6188 applies only to @code{protoize}.
6190 @item -c @var{compilation-options}
6191 Use @var{compilation-options} as the options when running @code{gcc} to
6192 produce the @samp{.X} files. The special option @samp{-aux-info} is
6193 always passed in addition, to tell @code{gcc} to write a @samp{.X} file.
6195 Note that the compilation options must be given as a single argument to
6196 @code{protoize} or @code{unprotoize}. If you want to specify several
6197 @code{gcc} options, you must quote the entire set of compilation options
6198 to make them a single word in the shell.
6200 There are certain @code{gcc} arguments that you cannot use, because they
6201 would produce the wrong kind of output. These include @samp{-g},
6202 @samp{-O}, @samp{-c}, @samp{-S}, and @samp{-o} If you include these in
6203 the @var{compilation-options}, they are ignored.
6206 Rename files to end in @samp{.C} instead of @samp{.c}.
6207 This is convenient if you are converting a C program to C++.
6208 This option applies only to @code{protoize}.
6211 Add explicit global declarations. This means inserting explicit
6212 declarations at the beginning of each source file for each function
6213 that is called in the file and was not declared. These declarations
6214 precede the first function definition that contains a call to an
6215 undeclared function. This option applies only to @code{protoize}.
6217 @item -i @var{string}
6218 Indent old-style parameter declarations with the string @var{string}.
6219 This option applies only to @code{protoize}.
6221 @code{unprotoize} converts prototyped function definitions to old-style
6222 function definitions, where the arguments are declared between the
6223 argument list and the initial @samp{@{}. By default, @code{unprotoize}
6224 uses five spaces as the indentation. If you want to indent with just
6225 one space instead, use @samp{-i " "}.
6228 Keep the @samp{.X} files. Normally, they are deleted after conversion
6232 Add explicit local declarations. @code{protoize} with @samp{-l} inserts
6233 a prototype declaration for each function in each block which calls the
6234 function without any declaration. This option applies only to
6238 Make no real changes. This mode just prints information about the conversions
6239 that would have been done without @samp{-n}.
6242 Make no @samp{.save} files. The original files are simply deleted.
6243 Use this option with caution.
6245 @item -p @var{program}
6246 Use the program @var{program} as the compiler. Normally, the name
6250 Work quietly. Most warnings are suppressed.
6253 Print the version number, just like @samp{-v} for @code{gcc}.
6256 If you need special compiler options to compile one of your program's
6257 source files, then you should generate that file's @samp{.X} file
6258 specially, by running @code{gcc} on that source file with the
6259 appropriate options and the option @samp{-aux-info}. Then run
6260 @code{protoize} on the entire set of files. @code{protoize} will use
6261 the existing @samp{.X} file because it is newer than the source file.
6265 gcc -Dfoo=bar file1.c -aux-info
6270 You need to include the special files along with the rest in the
6271 @code{protoize} command, even though their @samp{.X} files already
6272 exist, because otherwise they won't get converted.
6274 @xref{Protoize Caveats}, for more information on how to use
6275 @code{protoize} successfully.