1 @c Copyright (C) 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000, 2001 Free Software Foundation, Inc.
2 @c This is part of the GCC manual.
3 @c For copying conditions, see the file gcc.texi.
7 Copyright @copyright{} 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997,
8 1998, 1999, 2000, 2001 Free Software Foundation, Inc.
10 Permission is granted to make and distribute verbatim copies of this
11 manual provided the copyright notice and this permission notice are
12 preserved on all copies.
14 Permission is granted to copy and distribute modified versions of this
15 manual under the conditions for verbatim copying, provided also that the
16 entire resulting derived work is distributed under the terms of a
17 permission notice identical to this one.
19 Permission is granted to copy and distribute translations of this manual
20 into another language, under the above conditions for modified versions,
21 except that this permission notice may be included in translations
22 approved by the Free Software Foundation instead of in the original
25 @c Set file name and title for the man page.
27 @settitle GNU project C and C++ compiler
29 gcc [@samp{-c}|@samp{-S}|@samp{-E}] [@samp{-std=}@var{standard}]
30 [@samp{-g}] [@samp{-pg}] [@samp{-O}@var{level}]
31 [@samp{-W}@var{warn}...] [@samp{-pedantic}]
32 [@samp{-I}@var{dir}...] [@samp{-L}@var{dir}...]
33 [@samp{-D}@var{macro}[=@var{defn}]...] [@samp{-U}@var{macro}]
34 [@samp{-f}@var{option}...] [@samp{-m}@var{machine-option}...]
35 [@samp{-o} @var{outfile}] @var{infile}...
37 Only the most useful options are listed here; see below for the
38 remainder. @samp{g++} accepts mostly the same options as @samp{gcc}.
41 cpp(1), gcov(1), g77(1), as(1), ld(1), gdb(1), adb(1), dbx(1), sdb(1)
42 and the Info entries for @file{gcc}, @file{cpp}, @file{g77}, @file{as},
43 @file{ld}, @file{binutils} and @file{gdb}.
46 For instructions on reporting bugs, see
47 @w{@uref{http://gcc.gnu.org/bugs.html}}. Use of the @command{gccbug}
48 script to report bugs is recommended.
51 See the Info entry for @file{gcc}, or
52 @w{@uref{http://gcc.gnu.org/thanks.html}}, for contributors to GCC.
57 @chapter GCC Command Options
58 @cindex GCC command options
59 @cindex command options
60 @cindex options, GCC command
62 @c man begin DESCRIPTION
64 When you invoke GCC, it normally does preprocessing, compilation,
65 assembly and linking. The ``overall options'' allow you to stop this
66 process at an intermediate stage. For example, the @samp{-c} option
67 says not to run the linker. Then the output consists of object files
68 output by the assembler.
70 Other options are passed on to one stage of processing. Some options
71 control the preprocessor and others the compiler itself. Yet other
72 options control the assembler and linker; most of these are not
73 documented here, since you rarely need to use any of them.
75 @cindex C compilation options
76 Most of the command line options that you can use with GCC are useful
77 for C programs; when an option is only useful with another language
78 (usually C++), the explanation says so explicitly. If the description
79 for a particular option does not mention a source language, you can use
80 that option with all supported languages.
82 @cindex C++ compilation options
83 @xref{Invoking G++,,Compiling C++ Programs}, for a summary of special
84 options for compiling C++ programs.
86 @cindex grouping options
87 @cindex options, grouping
88 The @command{gcc} program accepts options and file names as operands. Many
89 options have multi-letter names; therefore multiple single-letter options
90 may @emph{not} be grouped: @samp{-dr} is very different from @w{@samp{-d
93 @cindex order of options
94 @cindex options, order
95 You can mix options and other arguments. For the most part, the order
96 you use doesn't matter. Order does matter when you use several options
97 of the same kind; for example, if you specify @samp{-L} more than once,
98 the directories are searched in the order specified.
100 Many options have long names starting with @samp{-f} or with
101 @samp{-W}---for example, @samp{-fforce-mem},
102 @samp{-fstrength-reduce}, @samp{-Wformat} and so on. Most of
103 these have both positive and negative forms; the negative form of
104 @samp{-ffoo} would be @samp{-fno-foo}. This manual documents
105 only one of these two forms, whichever one is not the default.
110 * Option Summary:: Brief list of all options, without explanations.
111 * Overall Options:: Controlling the kind of output:
112 an executable, object files, assembler files,
113 or preprocessed source.
114 * Invoking G++:: Compiling C++ programs.
115 * C Dialect Options:: Controlling the variant of C language compiled.
116 * C++ Dialect Options:: Variations on C++.
117 * Language Independent Options:: Controlling how diagnostics should be
119 * Warning Options:: How picky should the compiler be?
120 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
121 * Optimize Options:: How much optimization?
122 * Preprocessor Options:: Controlling header files and macro definitions.
123 Also, getting dependency information for Make.
124 * Assembler Options:: Passing options to the assembler.
125 * Link Options:: Specifying libraries and so on.
126 * Directory Options:: Where to find header files and libraries.
127 Where to find the compiler executable files.
128 * Spec Files:: How to pass switches to sub-processes.
129 * Target Options:: Running a cross-compiler, or an old version of GCC.
130 * Submodel Options:: Specifying minor hardware or convention variations,
131 such as 68010 vs 68020.
132 * Code Gen Options:: Specifying conventions for function calls, data layout
134 * Environment Variables:: Env vars that affect GCC.
135 * Running Protoize:: Automatically adding or removing function prototypes.
139 @section Option Summary
141 Here is a summary of all the options, grouped by type. Explanations are
142 in the following sections.
145 @item Overall Options
146 @xref{Overall Options,,Options Controlling the Kind of Output}.
148 -c -S -E -o @var{file} -pipe -pass-exit-codes -x @var{language}
149 -v --target-help --help
152 @item C Language Options
153 @xref{C Dialect Options,,Options Controlling C Dialect}.
155 -ansi -std=@var{standard} -fno-asm -fno-builtin
156 -fhosted -ffreestanding
157 -trigraphs -traditional -traditional-cpp
158 -fallow-single-precision -fcond-mismatch
159 -fsigned-bitfields -fsigned-char
160 -funsigned-bitfields -funsigned-char
161 -fwritable-strings -fshort-wchar
164 @item C++ Language Options
165 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}.
167 -fno-access-control -fcheck-new -fconserve-space
168 -fdollars-in-identifiers -fno-elide-constructors
169 -fno-enforce-eh-specs -fexternal-templates
170 -falt-external-templates
171 -ffor-scope -fno-for-scope -fno-gnu-keywords -fhonor-std
172 -fhuge-objects -fno-implicit-templates
173 -fno-implicit-inline-templates
174 -fno-implement-inlines -fms-extensions
175 -fname-mangling-version-@var{n} -fno-operator-names
176 -fno-optional-diags -fpermissive
177 -frepo -fno-rtti -fsquangle -ftemplate-depth-@var{n}
178 -fuse-cxa-atexit -fvtable-thunks -nostdinc++
179 -fno-default-inline -Wctor-dtor-privacy
180 -Wnon-virtual-dtor -Wreorder
181 -Weffc++ -Wno-deprecated
182 -Wno-non-template-friend -Wold-style-cast
183 -Woverloaded-virtual -Wno-pmf-conversions
187 @item Language Independent Options
188 @xref{Language Independent Options,,Options to Control Diagnostic Messages Formatting}.
190 -fmessage-length=@var{n}
191 -fdiagnostics-show-location=@r{[}once@r{|}every-line@r{]}
194 @item Warning Options
195 @xref{Warning Options,,Options to Request or Suppress Warnings}.
197 -fsyntax-only -pedantic -pedantic-errors
198 -w -W -Wall -Waggregate-return
199 -Wcast-align -Wcast-qual -Wchar-subscripts -Wcomment
200 -Wconversion -Wdisabled-optimization -Werror
201 -Wfloat-equal -Wformat -Wformat=2
202 -Wformat-nonliteral -Wformat-security
203 -Wid-clash-@var{len} -Wimplicit -Wimplicit-int
204 -Wimplicit-function-declaration
205 -Werror-implicit-function-declaration
207 -Wlarger-than-@var{len} -Wlong-long
208 -Wmain -Wmissing-declarations
209 -Wmissing-format-attribute -Wmissing-noreturn
210 -Wmultichar -Wno-format-extra-args -Wno-format-y2k
211 -Wno-import -Wpacked -Wpadded
212 -Wparentheses -Wpointer-arith -Wredundant-decls
213 -Wreturn-type -Wsequence-point -Wshadow
214 -Wsign-compare -Wswitch -Wsystem-headers
215 -Wtrigraphs -Wundef -Wuninitialized
216 -Wunknown-pragmas -Wunreachable-code
217 -Wunused -Wunused-function -Wunused-label -Wunused-parameter
218 -Wunused-value -Wunused-variable -Wwrite-strings
221 @item C-only Warning Options
223 -Wbad-function-cast -Wmissing-prototypes -Wnested-externs
224 -Wstrict-prototypes -Wtraditional
227 @item Debugging Options
228 @xref{Debugging Options,,Options for Debugging Your Program or GCC}.
230 -a -ax -d@var{letters} -fdump-unnumbered -fdump-translation-unit-@var{file}
231 -fpretend-float -fprofile-arcs -ftest-coverage
232 -g -g@var{level} -gcoff -gdwarf -gdwarf-1 -gdwarf-1+ -gdwarf-2
233 -ggdb -gstabs -gstabs+ -gxcoff -gxcoff+
234 -p -pg -print-file-name=@var{library} -print-libgcc-file-name
235 -print-prog-name=@var{program} -print-search-dirs -Q
239 @item Optimization Options
240 @xref{Optimize Options,,Options that Control Optimization}.
242 -falign-functions=@var{n} -falign-jumps=@var{n}
243 -falign-labels=@var{n} -falign-loops=@var{n}
244 -fbranch-probabilities -fcaller-saves
245 -fcse-follow-jumps -fcse-skip-blocks -fdata-sections -fdce
246 -fdelayed-branch -fdelete-null-pointer-checks
247 -fexpensive-optimizations -ffast-math -ffloat-store
248 -fforce-addr -fforce-mem -ffunction-sections -fgcse
249 -finline-functions -finline-limit=@var{n} -fkeep-inline-functions
250 -fkeep-static-consts -fmove-all-movables
251 -fno-default-inline -fno-defer-pop
252 -fno-function-cse -fno-inline -fno-math-errno -fno-peephole
253 -fomit-frame-pointer -foptimize-register-move
254 -foptimize-sibling-calls -freduce-all-givs
255 -fregmove -frename-registers
256 -frerun-cse-after-loop -frerun-loop-opt
257 -fschedule-insns -fschedule-insns2
258 -fsingle-precision-constant -fssa
259 -fstrength-reduce -fstrict-aliasing -fthread-jumps -ftrapv
260 -funroll-all-loops -funroll-loops
261 -O -O0 -O1 -O2 -O3 -Os
264 @item Preprocessor Options
265 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
267 -$ -A@var{question}=@var{answer} -A-@var{question}[=@var{answer}]
269 -D@var{macro}@r{[}=@var{defn}@r{]} -E -H
271 -include @var{file} -imacros @var{file}
272 -iprefix @var{file} -iwithprefix @var{dir}
273 -iwithprefixbefore @var{dir} -isystem @var{dir} -isystem-c++ @var{dir}
274 -M -MM -MF -MG -MP -MQ -MT -nostdinc -P -remap
275 -trigraphs -undef -U@var{macro} -Wp,@var{option}
278 @item Assembler Option
279 @xref{Assembler Options,,Passing Options to the Assembler}.
285 @xref{Link Options,,Options for Linking}.
287 @var{object-file-name} -l@var{library}
288 -nostartfiles -nodefaultlibs -nostdlib
289 -s -static -static-libgcc -shared -shared-libgcc -symbolic
290 -Wl,@var{option} -Xlinker @var{option}
294 @item Directory Options
295 @xref{Directory Options,,Options for Directory Search}.
297 -B@var{prefix} -I@var{dir} -I- -L@var{dir} -specs=@var{file}
301 @c I wrote this xref this way to avoid overfull hbox. -- rms
302 @xref{Target Options}.
304 -b @var{machine} -V @var{version}
307 @item Machine Dependent Options
308 @xref{Submodel Options,,Hardware Models and Configurations}.
310 @emph{M680x0 Options}
311 -m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040
312 -m68060 -mcpu32 -m5200 -m68881 -mbitfield -mc68000 -mc68020
313 -mfpa -mnobitfield -mrtd -mshort -msoft-float -mpcrel
314 -malign-int -mstrict-align
316 @emph{M68hc1x Options}
317 -m6811 -m6812 -m68hc11 -m68hc12
318 -mauto-incdec -mshort -msoft-reg-count=@var{count}
325 -mtune=@var{cpu type}
326 -mcmodel=@var{code model}
328 -mapp-regs -mbroken-saverestore -mcypress
329 -mepilogue -mfaster-structs -mflat
330 -mfpu -mhard-float -mhard-quad-float
331 -mimpure-text -mlive-g0 -mno-app-regs
332 -mno-epilogue -mno-faster-structs -mno-flat -mno-fpu
333 -mno-impure-text -mno-stack-bias -mno-unaligned-doubles
334 -msoft-float -msoft-quad-float -msparclite -mstack-bias
335 -msupersparc -munaligned-doubles -mv8
337 @emph{Convex Options}
338 -mc1 -mc2 -mc32 -mc34 -mc38
339 -margcount -mnoargcount
341 -mvolatile-cache -mvolatile-nocache
343 @emph{AMD29K Options}
344 -m29000 -m29050 -mbw -mnbw -mdw -mndw
345 -mlarge -mnormal -msmall
346 -mkernel-registers -mno-reuse-arg-regs
347 -mno-stack-check -mno-storem-bug
348 -mreuse-arg-regs -msoft-float -mstack-check
349 -mstorem-bug -muser-registers
352 -mapcs-frame -mno-apcs-frame
354 -mapcs-stack-check -mno-apcs-stack-check
355 -mapcs-float -mno-apcs-float
356 -mapcs-reentrant -mno-apcs-reentrant
357 -msched-prolog -mno-sched-prolog
358 -mlittle-endian -mbig-endian -mwords-little-endian
359 -malignment-traps -mno-alignment-traps
360 -msoft-float -mhard-float -mfpe
361 -mthumb-interwork -mno-thumb-interwork
362 -mcpu= -march= -mfpe=
363 -mstructure-size-boundary=
364 -mbsd -mxopen -mno-symrename
366 -mlong-calls -mno-long-calls
367 -mnop-fun-dllimport -mno-nop-fun-dllimport
368 -msingle-pic-base -mno-single-pic-base
372 -mtpcs-frame -mno-tpcs-frame
373 -mtpcs-leaf-frame -mno-tpcs-leaf-frame
374 -mlittle-endian -mbig-endian
375 -mthumb-interwork -mno-thumb-interwork
376 -mstructure-size-boundary=
377 -mnop-fun-dllimport -mno-nop-fun-dllimport
378 -mcallee-super-interworking -mno-callee-super-interworking
379 -mcaller-super-interworking -mno-caller-super-interworking
380 -msingle-pic-base -mno-single-pic-base
383 @emph{MN10200 Options}
386 @emph{MN10300 Options}
393 @emph{M32R/D Options}
394 -mcode-model=@var{model type} -msdata=@var{sdata type}
398 -m88000 -m88100 -m88110 -mbig-pic
399 -mcheck-zero-division -mhandle-large-shift
400 -midentify-revision -mno-check-zero-division
401 -mno-ocs-debug-info -mno-ocs-frame-position
402 -mno-optimize-arg-area -mno-serialize-volatile
403 -mno-underscores -mocs-debug-info
404 -mocs-frame-position -moptimize-arg-area
405 -mserialize-volatile -mshort-data-@var{num} -msvr3
406 -msvr4 -mtrap-large-shift -muse-div-instruction
407 -mversion-03.00 -mwarn-passed-structs
409 @emph{RS/6000 and PowerPC Options}
411 -mtune=@var{cpu type}
412 -mpower -mno-power -mpower2 -mno-power2
413 -mpowerpc -mpowerpc64 -mno-powerpc
414 -mpowerpc-gpopt -mno-powerpc-gpopt
415 -mpowerpc-gfxopt -mno-powerpc-gfxopt
416 -mnew-mnemonics -mold-mnemonics
417 -mfull-toc -mminimal-toc -mno-fop-in-toc -mno-sum-in-toc
418 -m64 -m32 -mxl-call -mno-xl-call -mthreads -mpe
419 -msoft-float -mhard-float -mmultiple -mno-multiple
420 -mstring -mno-string -mupdate -mno-update
421 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align
422 -mstrict-align -mno-strict-align -mrelocatable
423 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib
424 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian
425 -mcall-aix -mcall-sysv -mprototype -mno-prototype
426 -msim -mmvme -mads -myellowknife -memb -msdata
427 -msdata=@var{opt} -mvxworks -G @var{num}
430 -mcall-lib-mul -mfp-arg-in-fpregs -mfp-arg-in-gregs
431 -mfull-fp-blocks -mhc-struct-return -min-line-mul
432 -mminimum-fp-blocks -mnohc-struct-return
435 -mabicalls -mcpu=@var{cpu type}
436 -membedded-data -muninit-const-in-rodata
437 -membedded-pic -mfp32 -mfp64 -mgas -mgp32 -mgp64
438 -mgpopt -mhalf-pic -mhard-float -mint64 -mips1
439 -mips2 -mips3 -mips4 -mlong64 -mlong32 -mlong-calls -mmemcpy
440 -mmips-as -mmips-tfile -mno-abicalls
441 -mno-embedded-data -mno-uninit-const-in-rodata -mno-embedded-pic
442 -mno-gpopt -mno-long-calls
443 -mno-memcpy -mno-mips-tfile -mno-rnames -mno-stats
444 -mrnames -msoft-float
445 -m4650 -msingle-float -mmad
446 -mstats -EL -EB -G @var{num} -nocpp
447 -mabi=32 -mabi=n32 -mabi=64 -mabi=eabi
451 -mcpu=@var{cpu type} -march=@var{cpu type}
452 -mintel-syntax -mieee-fp -mno-fancy-math-387
453 -mno-fp-ret-in-387 -msoft-float -msvr3-shlib
454 -mno-wide-multiply -mrtd -malign-double
455 -mreg-alloc=@var{list} -mregparm=@var{num}
456 -malign-jumps=@var{num} -malign-loops=@var{num}
457 -malign-functions=@var{num} -mpreferred-stack-boundary=@var{num}
458 -mthreads -mno-align-stringops -minline-all-stringops
459 -mpush-args -maccumulate-outgoing-args -m128bit-long-double
463 -march=@var{architecture type}
464 -mbig-switch -mdisable-fpregs -mdisable-indexing
465 -mfast-indirect-calls -mgas -mjump-in-delay
466 -mlong-load-store -mno-big-switch -mno-disable-fpregs
467 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas
468 -mno-jump-in-delay -mno-long-load-store
469 -mno-portable-runtime -mno-soft-float
470 -mno-space-regs -msoft-float -mpa-risc-1-0
471 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime
472 -mschedule=@var{cpu type} -mspace-regs
474 @emph{Intel 960 Options}
475 -m@var{cpu type} -masm-compat -mclean-linkage
476 -mcode-align -mcomplex-addr -mleaf-procedures
477 -mic-compat -mic2.0-compat -mic3.0-compat
478 -mintel-asm -mno-clean-linkage -mno-code-align
479 -mno-complex-addr -mno-leaf-procedures
480 -mno-old-align -mno-strict-align -mno-tail-call
481 -mnumerics -mold-align -msoft-float -mstrict-align
484 @emph{DEC Alpha Options}
485 -mfp-regs -mno-fp-regs -mno-soft-float -msoft-float
487 -mieee -mieee-with-inexact -mieee-conformant
488 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode}
489 -mtrap-precision=@var{mode} -mbuild-constants
491 -mbwx -mno-bwx -mcix -mno-cix -mmax -mno-max
492 -mmemory-latency=@var{time}
494 @emph{Clipper Options}
497 @emph{H8/300 Options}
498 -mrelax -mh -ms -mint32 -malign-300
502 -m4-nofpu -m4-single-only -m4-single -m4
503 -mb -ml -mdalign -mrelax
504 -mbigtable -mfmovd -mhitachi -mnomacsave
505 -misize -mpadstruct -mspace
509 @emph{System V Options}
510 -Qy -Qn -YP,@var{paths} -Ym,@var{dir}
514 -mmangle-cpu -mcpu=@var{cpu} -mtext=@var{text section}
515 -mdata=@var{data section} -mrodata=@var{readonly data section}
517 @emph{TMS320C3x/C4x Options}
518 -mcpu=@var{cpu} -mbig -msmall -mregparm -mmemparm
519 -mfast-fix -mmpyi -mbk -mti -mdp-isr-reload
520 -mrpts=@var{count} -mrptb -mdb -mloop-unsigned
521 -mparallel-insns -mparallel-mpy -mpreserve-float
524 -mlong-calls -mno-long-calls -mep -mno-ep
525 -mprolog-function -mno-prolog-function -mspace
526 -mtda=@var{n} -msda=@var{n} -mzda=@var{n}
530 -m32032 -m32332 -m32532 -m32081 -m32381 -mmult-add -mnomult-add
531 -msoft-float -mrtd -mnortd -mregparam -mnoregparam -msb -mnosb
532 -mbitfield -mnobitfield -mhimem -mnohimem
535 -mmcu=@var{mcu} -msize -minit-stack=@var{n} -mno-interrupts
536 -mcall-prologues -mno-tablejump -mtiny-stack
539 -mhardlit, -mno-hardlit -mdiv -mno-div -mrelax-immediates
540 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields
541 -m4byte-functions -mno-4byte-functions -mcallgraph-data
542 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim
543 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment
546 @item Code Generation Options
547 @xref{Code Gen Options,,Options for Code Generation Conventions}.
549 -fcall-saved-@var{reg} -fcall-used-@var{reg}
550 -fexceptions -funwind-tables -ffixed-@var{reg}
551 -finhibit-size-directive -finstrument-functions
552 -fcheck-memory-usage -fprefix-function-name
553 -fno-common -fno-ident -fno-gnu-linker
554 -fpcc-struct-return -fpic -fPIC
555 -freg-struct-return -fshared-data -fshort-enums
556 -fshort-double -fvolatile -fvolatile-global -fvolatile-static
557 -fverbose-asm -fpack-struct -fstack-check
558 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym}
559 -fargument-alias -fargument-noalias
560 -fargument-noalias-global
566 * Overall Options:: Controlling the kind of output:
567 an executable, object files, assembler files,
568 or preprocessed source.
569 * C Dialect Options:: Controlling the variant of C language compiled.
570 * C++ Dialect Options:: Variations on C++.
571 * Language Independent Options:: Controlling how diagnostics should be
573 * Warning Options:: How picky should the compiler be?
574 * Debugging Options:: Symbol tables, measurements, and debugging dumps.
575 * Optimize Options:: How much optimization?
576 * Preprocessor Options:: Controlling header files and macro definitions.
577 Also, getting dependency information for Make.
578 * Assembler Options:: Passing options to the assembler.
579 * Link Options:: Specifying libraries and so on.
580 * Directory Options:: Where to find header files and libraries.
581 Where to find the compiler executable files.
582 * Spec Files:: How to pass switches to sub-processes.
583 * Target Options:: Running a cross-compiler, or an old version of GCC.
588 @node Overall Options
589 @section Options Controlling the Kind of Output
591 Compilation can involve up to four stages: preprocessing, compilation
592 proper, assembly and linking, always in that order. The first three
593 stages apply to an individual source file, and end by producing an
594 object file; linking combines all the object files (those newly
595 compiled, and those specified as input) into an executable file.
597 @cindex file name suffix
598 For any given input file, the file name suffix determines what kind of
603 C source code which must be preprocessed.
606 C source code which should not be preprocessed.
609 C++ source code which should not be preprocessed.
612 Objective-C source code. Note that you must link with the library
613 @file{libobjc.a} to make an Objective-C program work.
616 Objective-C source code which should not be preprocessed.
619 C header file (not to be compiled or linked).
623 @itemx @var{file}.cxx
624 @itemx @var{file}.cpp
625 @itemx @var{file}.c++
627 C++ source code which must be preprocessed. Note that in @samp{.cxx},
628 the last two letters must both be literally @samp{x}. Likewise,
629 @samp{.C} refers to a literal capital C.
632 @itemx @var{file}.for
633 @itemx @var{file}.FOR
634 Fortran source code which should not be preprocessed.
637 @itemx @var{file}.fpp
638 @itemx @var{file}.FPP
639 Fortran source code which must be preprocessed (with the traditional
643 Fortran source code which must be preprocessed with a RATFOR
644 preprocessor (not included with GCC).
646 @xref{Overall Options,,Options Controlling the Kind of Output, g77,
647 Using and Porting GNU Fortran}, for more details of the handling of
650 @c FIXME: Descriptions of Java file types.
656 @c GCC also knows about some suffixes for languages not yet included:
666 @itemx @var{file}.chi
667 CHILL source code (preprocessed with the traditional preprocessor).
673 Assembler code which must be preprocessed.
676 An object file to be fed straight into linking.
677 Any file name with no recognized suffix is treated this way.
680 You can specify the input language explicitly with the @samp{-x} option:
683 @item -x @var{language}
684 Specify explicitly the @var{language} for the following input files
685 (rather than letting the compiler choose a default based on the file
686 name suffix). This option applies to all following input files until
687 the next @samp{-x} option. Possible values for @var{language} are:
689 c c-header cpp-output
691 objective-c objc-cpp-output
692 assembler assembler-with-cpp
693 f77 f77-cpp-input ratfor
696 @c Also f77-version, for internal use only.
699 Turn off any specification of a language, so that subsequent files are
700 handled according to their file name suffixes (as they are if @samp{-x}
701 has not been used at all).
703 @item -pass-exit-codes
704 Normally the @command{gcc} program will exit with the code of 1 if any
705 phase of the compiler returns a non-success return code. If you specify
706 @samp{-pass-exit-codes}, the @command{gcc} program will instead return with
707 numerically highest error produced by any phase that returned an error
711 If you only want some of the stages of compilation, you can use
712 @samp{-x} (or filename suffixes) to tell @command{gcc} where to start, and
713 one of the options @samp{-c}, @samp{-S}, or @samp{-E} to say where
714 @command{gcc} is to stop. Note that some combinations (for example,
715 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
719 Compile or assemble the source files, but do not link. The linking
720 stage simply is not done. The ultimate output is in the form of an
721 object file for each source file.
723 By default, the object file name for a source file is made by replacing
724 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
726 Unrecognized input files, not requiring compilation or assembly, are
730 Stop after the stage of compilation proper; do not assemble. The output
731 is in the form of an assembler code file for each non-assembler input
734 By default, the assembler file name for a source file is made by
735 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
737 Input files that don't require compilation are ignored.
740 Stop after the preprocessing stage; do not run the compiler proper. The
741 output is in the form of preprocessed source code, which is sent to the
744 Input files which don't require preprocessing are ignored.
746 @cindex output file option
748 Place output in file @var{file}. This applies regardless to whatever
749 sort of output is being produced, whether it be an executable file,
750 an object file, an assembler file or preprocessed C code.
752 Since only one output file can be specified, it does not make sense to
753 use @samp{-o} when compiling more than one input file, unless you are
754 producing an executable file as output.
756 If @samp{-o} is not specified, the default is to put an executable file
757 in @file{a.out}, the object file for @file{@var{source}.@var{suffix}} in
758 @file{@var{source}.o}, its assembler file in @file{@var{source}.s}, and
759 all preprocessed C source on standard output.@refill
762 Print (on standard error output) the commands executed to run the stages
763 of compilation. Also print the version number of the compiler driver
764 program and of the preprocessor and the compiler proper.
767 Use pipes rather than temporary files for communication between the
768 various stages of compilation. This fails to work on some systems where
769 the assembler is unable to read from a pipe; but the GNU assembler has
773 Print (on the standard output) a description of the command line options
774 understood by @command{gcc}. If the @option{-v} option is also specified
775 then @option{--help} will also be passed on to the various processes
776 invoked by @command{gcc}, so that they can display the command line options
777 they accept. If the @option{-W} option is also specified then command
778 line options which have no documentation associated with them will also
782 Print (on the standard output) a description of target specific command
783 line options for each tool.
787 @section Compiling C++ Programs
789 @cindex suffixes for C++ source
790 @cindex C++ source file suffixes
791 C++ source files conventionally use one of the suffixes @samp{.C},
792 @samp{.cc}, @samp{.cpp}, @samp{.c++}, @samp{.cp}, or @samp{.cxx};
793 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
794 files with these names and compiles them as C++ programs even if you
795 call the compiler the same way as for compiling C programs (usually with
796 the name @command{gcc}).
800 However, C++ programs often require class libraries as well as a
801 compiler that understands the C++ language---and under some
802 circumstances, you might want to compile programs from standard input,
803 or otherwise without a suffix that flags them as C++ programs.
804 @command{g++} is a program that calls GCC with the default language
805 set to C++, and automatically specifies linking against the C++
806 library. On many systems, @command{g++} is also
807 installed with the name @command{c++}.
809 @cindex invoking @command{g++}
810 When you compile C++ programs, you may specify many of the same
811 command-line options that you use for compiling programs in any
812 language; or command-line options meaningful for C and related
813 languages; or options that are meaningful only for C++ programs.
814 @xref{C Dialect Options,,Options Controlling C Dialect}, for
815 explanations of options for languages related to C.
816 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
817 explanations of options that are meaningful only for C++ programs.
819 @node C Dialect Options
820 @section Options Controlling C Dialect
821 @cindex dialect options
822 @cindex language dialect options
823 @cindex options, dialect
825 The following options control the dialect of C (or languages derived
826 from C, such as C++ and Objective C) that the compiler accepts:
832 In C mode, support all ISO C89 programs. In C++ mode,
833 remove GNU extensions that conflict with ISO C++.
835 This turns off certain features of GCC that are incompatible with ISO
836 C (when compiling C code), or of standard C++ (when compiling C++ code),
837 such as the @code{asm} and @code{typeof} keywords, and
838 predefined macros such as @code{unix} and @code{vax} that identify the
839 type of system you are using. It also enables the undesirable and
840 rarely used ISO trigraph feature. For the C compiler,
841 it disables recognition of C++ style @samp{//} comments as well as
842 the @code{inline} keyword.
844 The alternate keywords @code{__asm__}, @code{__extension__},
845 @code{__inline__} and @code{__typeof__} continue to work despite
846 @samp{-ansi}. You would not want to use them in an ISO C program, of
847 course, but it is useful to put them in header files that might be included
848 in compilations done with @samp{-ansi}. Alternate predefined macros
849 such as @code{__unix__} and @code{__vax__} are also available, with or
850 without @samp{-ansi}.
852 The @samp{-ansi} option does not cause non-ISO programs to be
853 rejected gratuitously. For that, @samp{-pedantic} is required in
854 addition to @samp{-ansi}. @xref{Warning Options}.
856 The macro @code{__STRICT_ANSI__} is predefined when the @samp{-ansi}
857 option is used. Some header files may notice this macro and refrain
858 from declaring certain functions or defining certain macros that the
859 ISO standard doesn't call for; this is to avoid interfering with any
860 programs that might use these names for other things.
862 Functions which would normally be builtin but do not have semantics
863 defined by ISO C (such as @code{alloca} and @code{ffs}) are not builtin
864 functions with @samp{-ansi} is used. @xref{Other Builtins,,Other
865 built-in functions provided by GNU CC}, for details of the functions
869 Determine the language standard. A value for this option must be provided;
874 Same as @option{-ansi}
877 ISO C as modified in amend. 1
880 ISO C99. Note that this standard is not yet fully supported; see
881 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information.
884 same as @option{-std=iso9899:1990}
887 same as @option{-std=iso9899:1999}
890 default, iso9899:1990 + gnu extensions
893 iso9899:1999 + gnu extensions
896 same as @option{-std=iso9899:1999}, deprecated
899 same as @option{-std=iso9899:1999}, deprecated
902 same as @option{-std=gnu99}, deprecated
906 Even when this option is not specified, you can still use some of the
907 features of newer standards in so far as they do not conflict with
908 previous C standards. For example, you may use @code{__restrict__} even
909 when @option{-std=c99} is not specified.
911 @xref{Standards,,Language Standards Supported by GCC}, for details of
912 these standard versions.
915 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
916 keyword, so that code can use these words as identifiers. You can use
917 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
918 instead. @samp{-ansi} implies @samp{-fno-asm}.
920 In C++, this switch only affects the @code{typeof} keyword, since
921 @code{asm} and @code{inline} are standard keywords. You may want to
922 use the @samp{-fno-gnu-keywords} flag instead, which has the same effect.
925 @cindex builtin functions
926 Don't recognize builtin functions that do not begin with
927 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
928 functions provided by GNU CC}, for details of the functions affected,
929 including those which are not builtin functions when @samp{-ansi} is
930 used because they do not have an ISO standard meaning.
932 GCC normally generates special code to handle certain builtin functions
933 more efficiently; for instance, calls to @code{alloca} may become single
934 instructions that adjust the stack directly, and calls to @code{memcpy}
935 may become inline copy loops. The resulting code is often both smaller
936 and faster, but since the function calls no longer appear as such, you
937 cannot set a breakpoint on those calls, nor can you change the behavior
938 of the functions by linking with a different library.
941 @cindex hosted environment
943 Assert that compilation takes place in a hosted environment. This implies
944 @samp{-fbuiltin}. A hosted environment is one in which the
945 entire standard library is available, and in which @code{main} has a return
946 type of @code{int}. Examples are nearly everything except a kernel.
947 This is equivalent to @samp{-fno-freestanding}.
950 @cindex hosted environment
952 Assert that compilation takes place in a freestanding environment. This
953 implies @samp{-fno-builtin}. A freestanding environment
954 is one in which the standard library may not exist, and program startup may
955 not necessarily be at @code{main}. The most obvious example is an OS kernel.
956 This is equivalent to @samp{-fno-hosted}.
958 @xref{Standards,,Language Standards Supported by GCC}, for details of
959 freestanding and hosted environments.
962 Support ISO C trigraphs. You don't want to know about this
963 brain-damage. The @samp{-ansi} option implies @samp{-trigraphs}.
965 @cindex traditional C language
966 @cindex C language, traditional
968 Attempt to support some aspects of traditional C compilers.
973 All @code{extern} declarations take effect globally even if they
974 are written inside of a function definition. This includes implicit
975 declarations of functions.
978 The newer keywords @code{typeof}, @code{inline}, @code{signed}, @code{const}
979 and @code{volatile} are not recognized. (You can still use the
980 alternative keywords such as @code{__typeof__}, @code{__inline__}, and
984 Comparisons between pointers and integers are always allowed.
987 Integer types @code{unsigned short} and @code{unsigned char} promote
988 to @code{unsigned int}.
991 Out-of-range floating point literals are not an error.
994 Certain constructs which ISO regards as a single invalid preprocessing
995 number, such as @samp{0xe-0xd}, are treated as expressions instead.
998 String ``constants'' are not necessarily constant; they are stored in
999 writable space, and identical looking constants are allocated
1000 separately. (This is the same as the effect of
1001 @samp{-fwritable-strings}.)
1003 @cindex @code{longjmp} and automatic variables
1005 All automatic variables not declared @code{register} are preserved by
1006 @code{longjmp}. Ordinarily, GNU C follows ISO C: automatic variables
1007 not declared @code{volatile} may be clobbered.
1012 @cindex escape sequences, traditional
1013 The character escape sequences @samp{\x} and @samp{\a} evaluate as the
1014 literal characters @samp{x} and @samp{a} respectively. Without
1015 @w{@samp{-traditional}}, @samp{\x} is a prefix for the hexadecimal
1016 representation of a character, and @samp{\a} produces a bell.
1019 You may wish to use @samp{-fno-builtin} as well as @samp{-traditional}
1020 if your program uses names that are normally GNU C builtin functions for
1021 other purposes of its own.
1023 You cannot use @samp{-traditional} if you include any header files that
1024 rely on ISO C features. Some vendors are starting to ship systems with
1025 ISO C header files and you cannot use @samp{-traditional} on such
1026 systems to compile files that include any system headers.
1028 The @samp{-traditional} option also enables @samp{-traditional-cpp},
1029 which is described next.
1031 @item -traditional-cpp
1032 Attempt to support some aspects of traditional C preprocessors.
1037 Comments convert to nothing at all, rather than to a space. This allows
1038 traditional token concatenation.
1041 In a preprocessing directive, the @samp{#} symbol must appear as the first
1042 character of a line.
1045 Macro arguments are recognized within string constants in a macro
1046 definition (and their values are stringified, though without additional
1047 quote marks, when they appear in such a context). The preprocessor
1048 always considers a string constant to end at a newline.
1051 @cindex detecting @w{@samp{-traditional}}
1052 The predefined macro @code{__STDC__} is not defined when you use
1053 @samp{-traditional}, but @code{__GNUC__} is (since the GNU extensions
1054 which @code{__GNUC__} indicates are not affected by
1055 @samp{-traditional}). If you need to write header files that work
1056 differently depending on whether @samp{-traditional} is in use, by
1057 testing both of these predefined macros you can distinguish four
1058 situations: GNU C, traditional GNU C, other ISO C compilers, and other
1059 old C compilers. The predefined macro @code{__STDC_VERSION__} is also
1060 not defined when you use @samp{-traditional}. @xref{Standard
1061 Predefined,,Standard Predefined Macros,cpp.info,The C Preprocessor},
1062 for more discussion of these and other predefined macros.
1065 @cindex string constants vs newline
1066 @cindex newline vs string constants
1067 The preprocessor considers a string constant to end at a newline (unless
1068 the newline is escaped with @samp{\}). (Without @w{@samp{-traditional}},
1069 string constants can contain the newline character as typed.)
1072 @item -fcond-mismatch
1073 Allow conditional expressions with mismatched types in the second and
1074 third arguments. The value of such an expression is void.
1076 @item -funsigned-char
1077 Let the type @code{char} be unsigned, like @code{unsigned char}.
1079 Each kind of machine has a default for what @code{char} should
1080 be. It is either like @code{unsigned char} by default or like
1081 @code{signed char} by default.
1083 Ideally, a portable program should always use @code{signed char} or
1084 @code{unsigned char} when it depends on the signedness of an object.
1085 But many programs have been written to use plain @code{char} and
1086 expect it to be signed, or expect it to be unsigned, depending on the
1087 machines they were written for. This option, and its inverse, let you
1088 make such a program work with the opposite default.
1090 The type @code{char} is always a distinct type from each of
1091 @code{signed char} or @code{unsigned char}, even though its behavior
1092 is always just like one of those two.
1095 Let the type @code{char} be signed, like @code{signed char}.
1097 Note that this is equivalent to @samp{-fno-unsigned-char}, which is
1098 the negative form of @samp{-funsigned-char}. Likewise, the option
1099 @samp{-fno-signed-char} is equivalent to @samp{-funsigned-char}.
1101 You may wish to use @samp{-fno-builtin} as well as @samp{-traditional}
1102 if your program uses names that are normally GNU C builtin functions for
1103 other purposes of its own.
1105 You cannot use @samp{-traditional} if you include any header files that
1106 rely on ISO C features. Some vendors are starting to ship systems with
1107 ISO C header files and you cannot use @samp{-traditional} on such
1108 systems to compile files that include any system headers.
1110 @item -fsigned-bitfields
1111 @itemx -funsigned-bitfields
1112 @itemx -fno-signed-bitfields
1113 @itemx -fno-unsigned-bitfields
1114 These options control whether a bitfield is signed or unsigned, when the
1115 declaration does not use either @code{signed} or @code{unsigned}. By
1116 default, such a bitfield is signed, because this is consistent: the
1117 basic integer types such as @code{int} are signed types.
1119 However, when @samp{-traditional} is used, bitfields are all unsigned
1122 @item -fwritable-strings
1123 Store string constants in the writable data segment and don't uniquize
1124 them. This is for compatibility with old programs which assume they can
1125 write into string constants. The option @samp{-traditional} also has
1128 Writing into string constants is a very bad idea; ``constants'' should
1131 @item -fallow-single-precision
1132 Do not promote single precision math operations to double precision,
1133 even when compiling with @samp{-traditional}.
1135 Traditional K&R C promotes all floating point operations to double
1136 precision, regardless of the sizes of the operands. On the
1137 architecture for which you are compiling, single precision may be faster
1138 than double precision. If you must use @samp{-traditional}, but want
1139 to use single precision operations when the operands are single
1140 precision, use this option. This option has no effect when compiling
1141 with ISO or GNU C conventions (the default).
1144 Override the underlying type for @samp{wchar_t} to be @samp{short
1145 unsigned int} instead of the default for the target. This option is
1146 useful for building programs to run under WINE.
1149 @node C++ Dialect Options
1150 @section Options Controlling C++ Dialect
1152 @cindex compiler options, C++
1153 @cindex C++ options, command line
1154 @cindex options, C++
1155 This section describes the command-line options that are only meaningful
1156 for C++ programs; but you can also use most of the GNU compiler options
1157 regardless of what language your program is in. For example, you
1158 might compile a file @code{firstClass.C} like this:
1161 g++ -g -frepo -O -c firstClass.C
1165 In this example, only @samp{-frepo} is an option meant
1166 only for C++ programs; you can use the other options with any
1167 language supported by GCC.
1169 Here is a list of options that are @emph{only} for compiling C++ programs:
1172 @item -fno-access-control
1173 Turn off all access checking. This switch is mainly useful for working
1174 around bugs in the access control code.
1177 Check that the pointer returned by @code{operator new} is non-null
1178 before attempting to modify the storage allocated. The current Working
1179 Paper requires that @code{operator new} never return a null pointer, so
1180 this check is normally unnecessary.
1182 An alternative to using this option is to specify that your
1183 @code{operator new} does not throw any exceptions; if you declare it
1184 @samp{throw()}, g++ will check the return value. See also @samp{new
1187 @item -fconserve-space
1188 Put uninitialized or runtime-initialized global variables into the
1189 common segment, as C does. This saves space in the executable at the
1190 cost of not diagnosing duplicate definitions. If you compile with this
1191 flag and your program mysteriously crashes after @code{main()} has
1192 completed, you may have an object that is being destroyed twice because
1193 two definitions were merged.
1195 This option is no longer useful on most targets, now that support has
1196 been added for putting variables into BSS without making them common.
1198 @item -fdollars-in-identifiers
1199 Accept @samp{$} in identifiers. You can also explicitly prohibit use of
1200 @samp{$} with the option @samp{-fno-dollars-in-identifiers}. (GNU C allows
1201 @samp{$} by default on most target systems, but there are a few exceptions.)
1202 Traditional C allowed the character @samp{$} to form part of
1203 identifiers. However, ISO C and C++ forbid @samp{$} in identifiers.
1205 @item -fno-elide-constructors
1206 The C++ standard allows an implementation to omit creating a temporary
1207 which is only used to initialize another object of the same type.
1208 Specifying this option disables that optimization, and forces g++ to
1209 call the copy constructor in all cases.
1211 @item -fno-enforce-eh-specs
1212 Don't check for violation of exception specifications at runtime. This
1213 option violates the C++ standard, but may be useful for reducing code
1214 size in production builds, much like defining @samp{NDEBUG}. The compiler
1215 will still optimize based on the exception specifications.
1217 @item -fexternal-templates
1218 Cause template instantiations to obey @samp{#pragma interface} and
1219 @samp{implementation}; template instances are emitted or not according
1220 to the location of the template definition. @xref{Template
1221 Instantiation}, for more information.
1223 This option is deprecated.
1225 @item -falt-external-templates
1226 Similar to -fexternal-templates, but template instances are emitted or
1227 not according to the place where they are first instantiated.
1228 @xref{Template Instantiation}, for more information.
1230 This option is deprecated.
1233 @itemx -fno-for-scope
1234 If -ffor-scope is specified, the scope of variables declared in
1235 a @i{for-init-statement} is limited to the @samp{for} loop itself,
1236 as specified by the C++ standard.
1237 If -fno-for-scope is specified, the scope of variables declared in
1238 a @i{for-init-statement} extends to the end of the enclosing scope,
1239 as was the case in old versions of gcc, and other (traditional)
1240 implementations of C++.
1242 The default if neither flag is given to follow the standard,
1243 but to allow and give a warning for old-style code that would
1244 otherwise be invalid, or have different behavior.
1246 @item -fno-gnu-keywords
1247 Do not recognize @code{typeof} as a keyword, so that code can use this
1248 word as an identifier. You can use the keyword @code{__typeof__} instead.
1249 @samp{-ansi} implies @samp{-fno-gnu-keywords}.
1252 Treat the @code{namespace std} as a namespace, instead of ignoring
1253 it. For compatibility with earlier versions of g++, the compiler will,
1254 by default, ignore @code{namespace-declarations},
1255 @code{using-declarations}, @code{using-directives}, and
1256 @code{namespace-names}, if they involve @code{std}.
1258 @item -fhuge-objects
1259 Support virtual function calls for objects that exceed the size
1260 representable by a @samp{short int}. Users should not use this flag by
1261 default; if you need to use it, the compiler will tell you so.
1263 This flag is not useful when compiling with -fvtable-thunks.
1265 Like all options that change the ABI, all C++ code, @emph{including
1266 libgcc} must be built with the same setting of this option.
1268 @item -fno-implicit-templates
1269 Never emit code for non-inline templates which are instantiated
1270 implicitly (i.e. by use); only emit code for explicit instantiations.
1271 @xref{Template Instantiation}, for more information.
1273 @item -fno-implicit-inline-templates
1274 Don't emit code for implicit instantiations of inline templates, either.
1275 The default is to handle inlines differently so that compiles with and
1276 without optimization will need the same set of explicit instantiations.
1278 @item -fno-implement-inlines
1279 To save space, do not emit out-of-line copies of inline functions
1280 controlled by @samp{#pragma implementation}. This will cause linker
1281 errors if these functions are not inlined everywhere they are called.
1283 @item -fms-extensions
1284 Disable pedantic warnings about constructs used in MFC, such as implicit
1285 int and getting a pointer to member function via non-standard syntax.
1287 @item -fname-mangling-version-@var{n}
1288 Control the way in which names are mangled. Version 0 is compatible
1289 with versions of g++ before 2.8. Version 1 is the default. Version 1
1290 will allow correct mangling of function templates. For example,
1291 version 0 mangling does not mangle foo<int, double> and foo<int, char>
1292 given this declaration:
1295 template <class T, class U> void foo(T t);
1298 Like all options that change the ABI, all C++ code, @emph{including
1299 libgcc} must be built with the same setting of this option.
1301 @item -fno-operator-names
1302 Do not treat the operator name keywords @code{and}, @code{bitand},
1303 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
1304 synonyms as keywords.
1306 @item -fno-optional-diags
1307 Disable diagnostics that the standard says a compiler does not need to
1308 issue. Currently, the only such diagnostic issued by g++ is the one for
1309 a name having multiple meanings within a class.
1312 Downgrade messages about nonconformant code from errors to warnings. By
1313 default, g++ effectively sets @samp{-pedantic-errors} without
1314 @samp{-pedantic}; this option reverses that. This behavior and this
1315 option are superseded by @samp{-pedantic}, which works as it does for GNU C.
1318 Enable automatic template instantiation. This option also implies
1319 @samp{-fno-implicit-templates}. @xref{Template Instantiation}, for more
1323 Disable generation of information about every class with virtual
1324 functions for use by the C++ runtime type identification features
1325 (@samp{dynamic_cast} and @samp{typeid}). If you don't use those parts
1326 of the language, you can save some space by using this flag. Note that
1327 exception handling uses the same information, but it will generate it as
1331 @itemx -fno-squangle
1332 @samp{-fsquangle} will enable a compressed form of name mangling for
1333 identifiers. In particular, it helps to shorten very long names by recognizing
1334 types and class names which occur more than once, replacing them with special
1335 short ID codes. This option also requires any C++ libraries being used to
1336 be compiled with this option as well. The compiler has this disabled (the
1337 equivalent of @samp{-fno-squangle}) by default.
1339 Like all options that change the ABI, all C++ code, @emph{including
1340 libgcc.a} must be built with the same setting of this option.
1342 @item -ftemplate-depth-@var{n}
1343 Set the maximum instantiation depth for template classes to @var{n}.
1344 A limit on the template instantiation depth is needed to detect
1345 endless recursions during template class instantiation. ANSI/ISO C++
1346 conforming programs must not rely on a maximum depth greater than 17.
1348 @item -fuse-cxa-atexit
1349 Register destructors for objects with static storage duration with the
1350 @code{__cxa_atexit} function rather than the @code{atexit} function.
1351 This option is required for fully standards-compliant handling of static
1352 destructors, but will only work if your C library supports
1353 @code{__cxa_atexit}.
1355 @item -fvtable-thunks
1356 Use @samp{thunks} to implement the virtual function dispatch table
1357 (@samp{vtable}). The traditional (cfront-style) approach to
1358 implementing vtables was to store a pointer to the function and two
1359 offsets for adjusting the @samp{this} pointer at the call site. Newer
1360 implementations store a single pointer to a @samp{thunk} function which
1361 does any necessary adjustment and then calls the target function.
1363 This option also enables a heuristic for controlling emission of
1364 vtables; if a class has any non-inline virtual functions, the vtable
1365 will be emitted in the translation unit containing the first one of
1368 Like all options that change the ABI, all C++ code, @emph{including
1369 libgcc.a} must be built with the same setting of this option.
1372 Do not search for header files in the standard directories specific to
1373 C++, but do still search the other standard directories. (This option
1374 is used when building the C++ library.)
1377 In addition, these optimization, warning, and code generation options
1378 have meanings only for C++ programs:
1381 @item -fno-default-inline
1382 Do not assume @samp{inline} for functions defined inside a class scope.
1383 @xref{Optimize Options,,Options That Control Optimization}. Note that these
1384 functions will have linkage like inline functions; they just won't be
1387 @item -Wctor-dtor-privacy (C++ only)
1388 Warn when a class seems unusable, because all the constructors or
1389 destructors in a class are private and the class has no friends or
1390 public static member functions.
1392 @item -Wnon-virtual-dtor (C++ only)
1393 Warn when a class declares a non-virtual destructor that should probably
1394 be virtual, because it looks like the class will be used polymorphically.
1396 @item -Wreorder (C++ only)
1397 @cindex reordering, warning
1398 @cindex warning for reordering of member initializers
1399 Warn when the order of member initializers given in the code does not
1400 match the order in which they must be executed. For instance:
1406 A(): j (0), i (1) @{ @}
1410 Here the compiler will warn that the member initializers for @samp{i}
1411 and @samp{j} will be rearranged to match the declaration order of the
1415 The following @samp{-W@dots{}} options are not affected by @samp{-Wall}.
1418 @item -Weffc++ (C++ only)
1419 Warn about violations of various style guidelines from Scott Meyers'
1420 @cite{Effective C++} books. If you use this option, you should be aware
1421 that the standard library headers do not obey all of these guidelines;
1422 you can use @samp{grep -v} to filter out those warnings.
1424 @item -Wno-deprecated (C++ only)
1425 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
1427 @item -Wno-non-template-friend (C++ only)
1428 Disable warnings when non-templatized friend functions are declared
1429 within a template. With the advent of explicit template specification
1430 support in g++, if the name of the friend is an unqualified-id (ie,
1431 @samp{friend foo(int)}), the C++ language specification demands that the
1432 friend declare or define an ordinary, nontemplate function. (Section
1433 14.5.3). Before g++ implemented explicit specification, unqualified-ids
1434 could be interpreted as a particular specialization of a templatized
1435 function. Because this non-conforming behavior is no longer the default
1436 behavior for g++, @samp{-Wnon-template-friend} allows the compiler to
1437 check existing code for potential trouble spots, and is on by default.
1438 This new compiler behavior can be turned off with
1439 @samp{-Wno-non-template-friend} which keeps the conformant compiler code
1440 but disables the helpful warning.
1442 @item -Wold-style-cast (C++ only)
1443 Warn if an old-style (C-style) cast is used within a C++ program. The
1444 new-style casts (@samp{static_cast}, @samp{reinterpret_cast}, and
1445 @samp{const_cast}) are less vulnerable to unintended effects.
1447 @item -Woverloaded-virtual (C++ only)
1448 @cindex overloaded virtual fn, warning
1449 @cindex warning for overloaded virtual fn
1450 Warn when a derived class function declaration may be an error in
1451 defining a virtual function. In a derived class, the
1452 definitions of virtual functions must match the type signature of a
1453 virtual function declared in the base class. With this option, the
1454 compiler warns when you define a function with the same name as a
1455 virtual function, but with a type signature that does not match any
1456 declarations from the base class.
1458 @item -Wno-pmf-conversions (C++ only)
1459 Disable the diagnostic for converting a bound pointer to member function
1462 @item -Wsign-promo (C++ only)
1463 Warn when overload resolution chooses a promotion from unsigned or
1464 enumeral type to a signed type over a conversion to an unsigned type of
1465 the same size. Previous versions of g++ would try to preserve
1466 unsignedness, but the standard mandates the current behavior.
1468 @item -Wsynth (C++ only)
1469 @cindex warning for synthesized methods
1470 @cindex synthesized methods, warning
1471 Warn when g++'s synthesis behavior does not match that of cfront. For
1477 A& operator = (int);
1487 In this example, g++ will synthesize a default @samp{A& operator =
1488 (const A&);}, while cfront will use the user-defined @samp{operator =}.
1491 @node Language Independent Options
1492 @section Options to Control Diagnostic Messages Formatting
1493 @cindex options to control diagnostics formatting
1494 @cindex diagnostic messages
1495 @cindex message formatting
1497 Traditionally, diagnostic messages have been formatted irrespective of
1498 the output device's aspect (e.g. its width, ...). The options described
1499 below can be used to control the diagnostic messages formatting
1500 algorithm, e.g. how many characters per line, how often source location
1501 information should be reported. Right now, only the C++ front-end can
1502 honor these options. However it is expected, in the near future, that
1503 the remaining front-ends would be able to digest them correctly.
1506 @item -fmessage-length=@var{n}
1507 Try to format error messages so that they fit on lines of about @var{n}
1508 characters. The default is 72 characters for g++ and 0 for the rest of
1509 the front-ends supported by GCC. If @var{n} is zero, then no
1510 line-wrapping will be done; each error message will appear on a single
1513 @item -fdiagnostics-show-location=once
1514 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
1515 reporter to emit @emph{once} source location information; that is, in
1516 case the message is too long to fit on a single physical line and has to
1517 be wrapped, the source location won't be emitted (as prefix) again,
1518 over and over, in subsequent continuation lines. This is the default
1521 @item -fdiagnostics-show-location=every-line
1522 Only meaningful in line-wrapping mode. Instructs the diagnostic
1523 messages reporter to emit the same source location information (as
1524 prefix) for physical lines that result from the process of breaking a
1525 a message which is too long to fit on a single line.
1529 @node Warning Options
1530 @section Options to Request or Suppress Warnings
1531 @cindex options to control warnings
1532 @cindex warning messages
1533 @cindex messages, warning
1534 @cindex suppressing warnings
1536 Warnings are diagnostic messages that report constructions which
1537 are not inherently erroneous but which are risky or suggest there
1538 may have been an error.
1540 You can request many specific warnings with options beginning @samp{-W},
1541 for example @samp{-Wimplicit} to request warnings on implicit
1542 declarations. Each of these specific warning options also has a
1543 negative form beginning @samp{-Wno-} to turn off warnings;
1544 for example, @samp{-Wno-implicit}. This manual lists only one of the
1545 two forms, whichever is not the default.
1547 These options control the amount and kinds of warnings produced by GCC:
1550 @cindex syntax checking
1552 Check the code for syntax errors, but don't do anything beyond that.
1555 Issue all the warnings demanded by strict ISO C and ISO C++;
1556 reject all programs that use forbidden extensions, and some other
1557 programs that do not follow ISO C and ISO C++. For ISO C, follows the
1558 version of the ISO C standard specified by any @samp{-std} option used.
1560 Valid ISO C and ISO C++ programs should compile properly with or without
1561 this option (though a rare few will require @samp{-ansi}). However,
1562 without this option, certain GNU extensions and traditional C and C++
1563 features are supported as well. With this option, they are rejected.
1565 @samp{-pedantic} does not cause warning messages for use of the
1566 alternate keywords whose names begin and end with @samp{__}. Pedantic
1567 warnings are also disabled in the expression that follows
1568 @code{__extension__}. However, only system header files should use
1569 these escape routes; application programs should avoid them.
1570 @xref{Alternate Keywords}.
1572 Some users try to use @samp{-pedantic} to check programs for strict ISO
1573 C conformance. They soon find that it does not do quite what they want:
1574 it finds some non-ISO practices, but not all---only those for which
1575 ISO C @emph{requires} a diagnostic, and some others for which
1576 diagnostics have been added.
1578 A feature to report any failure to conform to ISO C might be useful in
1579 some instances, but would require considerable additional work and would
1580 be quite different from @samp{-pedantic}. We don't have plans to
1581 support such a feature in the near future.
1583 @item -pedantic-errors
1584 Like @samp{-pedantic}, except that errors are produced rather than
1588 Inhibit all warning messages.
1591 Inhibit warning messages about the use of @samp{#import}.
1593 @item -Wchar-subscripts
1594 Warn if an array subscript has type @code{char}. This is a common cause
1595 of error, as programmers often forget that this type is signed on some
1599 Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*}
1600 comment, or whenever a Backslash-Newline appears in a @samp{//} comment.
1603 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
1604 the arguments supplied have types appropriate to the format string
1605 specified, and that the conversions specified in the format string make
1606 sense. This includes standard functions, and others specified by format
1607 attributes (@pxref{Function Attributes}), in the @code{printf},
1608 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
1609 not in the C standard) families.
1611 The formats are checked against the format features supported by GNU
1612 libc version 2.2. These include all ISO C89 and C99 features, as well
1613 as features from the Single Unix Specification and some BSD and GNU
1614 extensions. Other library implementations may not support all these
1615 features; GCC does not support warning about features that go beyond a
1616 particular library's limitations. However, if @samp{-pedantic} is used
1617 with @samp{-Wformat}, warnings will be given about format features not
1618 in the selected standard version (but not for @code{strfmon} formats,
1619 since those are not in any version of the C standard). @xref{C Dialect
1620 Options,,Options Controlling C Dialect}.
1622 @samp{-Wformat} is included in @samp{-Wall}. For more control over some
1623 aspects of format checking, the options @samp{-Wno-format-y2k},
1624 @samp{-Wno-format-extra-args}, @samp{-Wformat-nonliteral},
1625 @samp{-Wformat-security} and @samp{-Wformat=2} are available, but are
1626 not included in @samp{-Wall}.
1628 @item -Wno-format-y2k
1629 If @samp{-Wformat} is specified, do not warn about @code{strftime}
1630 formats which may yield only a two-digit year.
1632 @item -Wno-format-extra-args
1633 If @samp{-Wformat} is specified, do not warn about excess arguments to a
1634 @code{printf} or @code{scanf} format function. The C standard specifies
1635 that such arguments are ignored.
1637 @item -Wformat-nonliteral
1638 If @samp{-Wformat} is specified, also warn if the format string is not a
1639 string literal and so cannot be checked, unless the format function
1640 takes its format arguments as a @code{va_list}.
1642 @item -Wformat-security
1643 If @samp{-Wformat} is specified, also warn about uses of format
1644 functions that represent possible security problems. At present, this
1645 warns about calls to @code{printf} and @code{scanf} functions where the
1646 format string is not a string literal and there are no format arguments,
1647 as in @code{printf (foo);}. This may be a security hole if the format
1648 string came from untrusted input and contains @samp{%n}. (This is
1649 currently a subset of what @samp{-Wformat-nonliteral} warns about, but
1650 in future warnings may be added to @samp{-Wformat-security} that are not
1651 included in @samp{-Wformat-nonliteral}.)
1654 Enable @samp{-Wformat} plus format checks not included in
1655 @samp{-Wformat}. Currently equivalent to @samp{-Wformat
1656 -Wformat-nonliteral -Wformat-security}.
1658 @item -Wimplicit-int
1659 Warn when a declaration does not specify a type.
1661 @item -Wimplicit-function-declaration
1662 @itemx -Werror-implicit-function-declaration
1663 Give a warning (or error) whenever a function is used before being
1667 Same as @samp{-Wimplicit-int} and @samp{-Wimplicit-function-}@*
1671 Warn if the type of @samp{main} is suspicious. @samp{main} should be a
1672 function with external linkage, returning int, taking either zero
1673 arguments, two, or three arguments of appropriate types.
1676 Warn if a multicharacter constant (@samp{'FOOF'}) is used. Usually they
1677 indicate a typo in the user's code, as they have implementation-defined
1678 values, and should not be used in portable code.
1681 Warn if parentheses are omitted in certain contexts, such
1682 as when there is an assignment in a context where a truth value
1683 is expected, or when operators are nested whose precedence people
1684 often get confused about.
1686 Also warn about constructions where there may be confusion to which
1687 @code{if} statement an @code{else} branch belongs. Here is an example of
1700 In C, every @code{else} branch belongs to the innermost possible @code{if}
1701 statement, which in this example is @code{if (b)}. This is often not
1702 what the programmer expected, as illustrated in the above example by
1703 indentation the programmer chose. When there is the potential for this
1704 confusion, GNU C will issue a warning when this flag is specified.
1705 To eliminate the warning, add explicit braces around the innermost
1706 @code{if} statement so there is no way the @code{else} could belong to
1707 the enclosing @code{if}. The resulting code would look like this:
1721 @item -Wsequence-point
1722 Warn about code that may have undefined semantics because of violations
1723 of sequence point rules in the C standard.
1725 The C standard defines the order in which expressions in a C program are
1726 evaluated in terms of @dfn{sequence points}, which represent a partial
1727 ordering between the execution of parts of the program: those executed
1728 before the sequence point, and those executed after it. These occur
1729 after the evaluation of a full expression (one which is not part of a
1730 larger expression), after the evaluation of the first operand of a
1731 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
1732 function is called (but after the evaluation of its arguments and the
1733 expression denoting the called function), and in certain other places.
1734 Other than as expressed by the sequence point rules, the order of
1735 evaluation of subexpressions of an expression is not specified. All
1736 these rules describe only a partial order rather than a total order,
1737 since, for example, if two functions are called within one expression
1738 with no sequence point between them, the order in which the functions
1739 are called is not specified. However, the standards committee have
1740 ruled that function calls do not overlap.
1742 It is not specified when between sequence points modifications to the
1743 values of objects take effect. Programs whose behavior depends on this
1744 have undefined behavior; the C standard specifies that ``Between the
1745 previous and next sequence point an object shall have its stored value
1746 modified at most once by the evaluation of an expression. Furthermore,
1747 the prior value shall be read only to determine the value to be
1748 stored.''. If a program breaks these rules, the results on any
1749 particular implementation are entirely unpredictable.
1751 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
1752 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
1753 diagnosed by this option, and it may give an occasional false positive
1754 result, but in general it has been found fairly effective at detecting
1755 this sort of problem in programs.
1757 The present implementation of this option only works for C programs. A
1758 future implementation may also work for C++ programs.
1760 There is some controversy over the precise meaning of the sequence point
1761 rules in subtle cases. Alternative formal definitions may be found in
1762 Clive Feather's ``Annex S''
1763 @w{@uref{http://wwwold.dkuug.dk/JTC1/SC22/WG14/www/docs/n925.htm}} and in
1764 Michael Norrish's thesis
1765 @w{@uref{http://www.cl.cam.ac.uk/users/mn200/PhD/thesis-report.ps.gz}}.
1766 Other discussions are by Raymond Mak
1767 @w{@uref{http://wwwold.dkuug.dk/JTC1/SC22/WG14/www/docs/n926.htm}} and
1769 @w{@uref{http://wwwold.dkuug.dk/JTC1/SC22/WG14/www/docs/n927.htm}}.
1772 Warn whenever a function is defined with a return-type that defaults to
1773 @code{int}. Also warn about any @code{return} statement with no
1774 return-value in a function whose return-type is not @code{void}.
1776 For C++, a function without return type always produces a diagnostic
1777 message, even when @samp{-Wno-return-type} is specified. The only
1778 exceptions are @samp{main} and functions defined in system headers.
1781 Warn whenever a @code{switch} statement has an index of enumeral type
1782 and lacks a @code{case} for one or more of the named codes of that
1783 enumeration. (The presence of a @code{default} label prevents this
1784 warning.) @code{case} labels outside the enumeration range also
1785 provoke warnings when this option is used.
1788 Warn if any trigraphs are encountered that might change the meaning of
1789 the program (trigraphs within comments are not warned about).
1791 @item -Wunused-function
1792 Warn whenever a static function is declared but not defined or a
1793 non\-inline static function is unused.
1795 @item -Wunused-label
1796 Warn whenever a label is declared but not used.
1798 To suppress this warning use the @samp{unused} attribute
1799 (@pxref{Variable Attributes}).
1801 @item -Wunused-parameter
1802 Warn whenever a function parameter is unused aside from its declaration.
1804 To suppress this warning use the @samp{unused} attribute
1805 (@pxref{Variable Attributes}).
1807 @item -Wunused-variable
1808 Warn whenever a local variable or non-constant static variable is unused
1809 aside from its declaration
1811 To suppress this warning use the @samp{unused} attribute
1812 (@pxref{Variable Attributes}).
1814 @item -Wunused-value
1815 Warn whenever a statement computes a result that is explicitly not used.
1817 To suppress this warning cast the expression to @samp{void}.
1820 All all the above @samp{-Wunused} options combined.
1822 In order to get a warning about an unused function parameter, you must
1823 either specify @samp{-W -Wunused} or separately specify
1824 @samp{-Wunused-parameter}.
1826 @item -Wuninitialized
1827 Warn if an automatic variable is used without first being initialized or
1828 if a variable may be clobbered by a @code{setjmp} call.
1830 These warnings are possible only in optimizing compilation,
1831 because they require data flow information that is computed only
1832 when optimizing. If you don't specify @samp{-O}, you simply won't
1835 These warnings occur only for variables that are candidates for
1836 register allocation. Therefore, they do not occur for a variable that
1837 is declared @code{volatile}, or whose address is taken, or whose size
1838 is other than 1, 2, 4 or 8 bytes. Also, they do not occur for
1839 structures, unions or arrays, even when they are in registers.
1841 Note that there may be no warning about a variable that is used only
1842 to compute a value that itself is never used, because such
1843 computations may be deleted by data flow analysis before the warnings
1846 These warnings are made optional because GCC is not smart
1847 enough to see all the reasons why the code might be correct
1848 despite appearing to have an error. Here is one example of how
1867 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
1868 always initialized, but GCC doesn't know this. Here is
1869 another common case:
1874 if (change_y) save_y = y, y = new_y;
1876 if (change_y) y = save_y;
1881 This has no bug because @code{save_y} is used only if it is set.
1883 @cindex @code{longjmp} warnings
1884 This option also warns when a non-volatile automatic variable might be
1885 changed by a call to @code{longjmp}. These warnings as well are possible
1886 only in optimizing compilation.
1888 The compiler sees only the calls to @code{setjmp}. It cannot know
1889 where @code{longjmp} will be called; in fact, a signal handler could
1890 call it at any point in the code. As a result, you may get a warning
1891 even when there is in fact no problem because @code{longjmp} cannot
1892 in fact be called at the place which would cause a problem.
1894 Some spurious warnings can be avoided if you declare all the functions
1895 you use that never return as @code{noreturn}. @xref{Function
1898 @item -Wreorder (C++ only)
1899 @cindex reordering, warning
1900 @cindex warning for reordering of member initializers
1901 Warn when the order of member initializers given in the code does not
1902 match the order in which they must be executed. For instance:
1904 @item -Wunknown-pragmas
1905 @cindex warning for unknown pragmas
1906 @cindex unknown pragmas, warning
1907 @cindex pragmas, warning of unknown
1908 Warn when a #pragma directive is encountered which is not understood by
1909 GCC. If this command line option is used, warnings will even be issued
1910 for unknown pragmas in system header files. This is not the case if
1911 the warnings were only enabled by the @samp{-Wall} command line option.
1914 All of the above @samp{-W} options combined. This enables all the
1915 warnings about constructions that some users consider questionable, and
1916 that are easy to avoid (or modify to prevent the warning), even in
1917 conjunction with macros.
1919 @item -Wsystem-headers
1920 @cindex warnings from system headers
1921 @cindex system headers, warnings from
1922 Print warning messages for constructs found in system header files.
1923 Warnings from system headers are normally suppressed, on the assumption
1924 that they usually do not indicate real problems and would only make the
1925 compiler output harder to read. Using this command line option tells
1926 GCC to emit warnings from system headers as if they occurred in user
1927 code. However, note that using @samp{-Wall} in conjunction with this
1928 option will @emph{not} warn about unknown pragmas in system
1929 headers---for that, @samp{-Wunknown-pragmas} must also be used.
1932 The following @samp{-W@dots{}} options are not implied by @samp{-Wall}.
1933 Some of them warn about constructions that users generally do not
1934 consider questionable, but which occasionally you might wish to check
1935 for; others warn about constructions that are necessary or hard to avoid
1936 in some cases, and there is no simple way to modify the code to suppress
1941 Print extra warning messages for these events:
1945 A function can return either with or without a value. (Falling
1946 off the end of the function body is considered returning without
1947 a value.) For example, this function would evoke such a
1961 An expression-statement or the left-hand side of a comma expression
1962 contains no side effects.
1963 To suppress the warning, cast the unused expression to void.
1964 For example, an expression such as @samp{x[i,j]} will cause a warning,
1965 but @samp{x[(void)i,j]} will not.
1968 An unsigned value is compared against zero with @samp{<} or @samp{<=}.
1971 A comparison like @samp{x<=y<=z} appears; this is equivalent to
1972 @samp{(x<=y ? 1 : 0) <= z}, which is a different interpretation from
1973 that of ordinary mathematical notation.
1976 Storage-class specifiers like @code{static} are not the first things in
1977 a declaration. According to the C Standard, this usage is obsolescent.
1980 The return type of a function has a type qualifier such as @code{const}.
1981 Such a type qualifier has no effect, since the value returned by a
1982 function is not an lvalue. (But don't warn about the GNU extension of
1983 @code{volatile void} return types. That extension will be warned about
1984 if @samp{-pedantic} is specified.)
1987 If @samp{-Wall} or @samp{-Wunused} is also specified, warn about unused
1991 A comparison between signed and unsigned values could produce an
1992 incorrect result when the signed value is converted to unsigned.
1993 (But don't warn if @samp{-Wno-sign-compare} is also specified.)
1996 An aggregate has a partly bracketed initializer.
1997 For example, the following code would evoke such a warning,
1998 because braces are missing around the initializer for @code{x.h}:
2001 struct s @{ int f, g; @};
2002 struct t @{ struct s h; int i; @};
2003 struct t x = @{ 1, 2, 3 @};
2007 An aggregate has an initializer which does not initialize all members.
2008 For example, the following code would cause such a warning, because
2009 @code{x.h} would be implicitly initialized to zero:
2012 struct s @{ int f, g, h; @};
2013 struct s x = @{ 3, 4 @};
2018 Warn if floating point values are used in equality comparisons.
2020 The idea behind this is that sometimes it is convenient (for the
2021 programmer) to consider floating-point values as approximations to
2022 infinitely precise real numbers. If you are doing this, then you need
2023 to compute (by analysing the code, or in some other way) the maximum or
2024 likely maximum error that the computation introduces, and allow for it
2025 when performing comparisons (and when producing output, but that's a
2026 different problem). In particular, instead of testing for equality, you
2027 would check to see whether the two values have ranges that overlap; and
2028 this is done with the relational operators, so equality comparisons are
2031 @item -Wtraditional (C only)
2032 Warn about certain constructs that behave differently in traditional and
2037 Macro arguments occurring within string constants in the macro body.
2038 These would substitute the argument in traditional C, but are part of
2039 the constant in ISO C.
2042 A function declared external in one block and then used after the end of
2046 A @code{switch} statement has an operand of type @code{long}.
2049 A non-@code{static} function declaration follows a @code{static} one.
2050 This construct is not accepted by some traditional C compilers.
2053 The ISO type of an integer constant has a different width or
2054 signedness from its traditional type. This warning is only issued if
2055 the base of the constant is ten. I.e. hexadecimal or octal values, which
2056 typically represent bit patterns, are not warned about.
2059 Usage of ISO string concatenation is detected.
2062 A function macro appears without arguments.
2065 The unary plus operator.
2068 Initialization of automatic aggregates.
2071 Identifier conflicts with labels. Traditional C lacks a separate
2072 namespace for labels.
2075 Initialization of unions. If the initializer is zero, the warning is
2076 omitted. This is done under the assumption that the zero initializer in
2077 user code appears conditioned on e.g. @code{__STDC__} to avoid missing
2078 initializer warnings and relies on default initialization to zero in the
2082 The `U' integer constant suffix, or the `F' or `L' floating point
2083 constant suffixes. (Traditonal C does support the `L' suffix on integer
2084 constants.) Note, these suffixes appear in macros defined in the system
2085 headers of most modern systems, e.g. the _MIN/_MAX macros in limits.h.
2086 Use of these macros can lead to spurious warnings as they do not
2087 necessarily reflect whether the code in question is any less portable to
2088 traditional C given that suitable backup definitions are provided.
2092 Warn if an undefined identifier is evaluated in an @samp{#if} directive.
2095 Warn whenever a local variable shadows another local variable.
2097 @item -Wid-clash-@var{len}
2098 Warn whenever two distinct identifiers match in the first @var{len}
2099 characters. This may help you prepare a program that will compile
2100 with certain obsolete, brain-damaged compilers.
2102 @item -Wlarger-than-@var{len}
2103 Warn whenever an object of larger than @var{len} bytes is defined.
2105 @item -Wpointer-arith
2106 Warn about anything that depends on the ``size of'' a function type or
2107 of @code{void}. GNU C assigns these types a size of 1, for
2108 convenience in calculations with @code{void *} pointers and pointers
2111 @item -Wbad-function-cast (C only)
2112 Warn whenever a function call is cast to a non-matching type.
2113 For example, warn if @code{int malloc()} is cast to @code{anything *}.
2116 Warn whenever a pointer is cast so as to remove a type qualifier from
2117 the target type. For example, warn if a @code{const char *} is cast
2118 to an ordinary @code{char *}.
2121 Warn whenever a pointer is cast such that the required alignment of the
2122 target is increased. For example, warn if a @code{char *} is cast to
2123 an @code{int *} on machines where integers can only be accessed at
2124 two- or four-byte boundaries.
2126 @item -Wwrite-strings
2127 Give string constants the type @code{const char[@var{length}]} so that
2128 copying the address of one into a non-@code{const} @code{char *}
2129 pointer will get a warning. These warnings will help you find at
2130 compile time code that can try to write into a string constant, but
2131 only if you have been very careful about using @code{const} in
2132 declarations and prototypes. Otherwise, it will just be a nuisance;
2133 this is why we did not make @samp{-Wall} request these warnings.
2136 Warn if a prototype causes a type conversion that is different from what
2137 would happen to the same argument in the absence of a prototype. This
2138 includes conversions of fixed point to floating and vice versa, and
2139 conversions changing the width or signedness of a fixed point argument
2140 except when the same as the default promotion.
2142 Also, warn if a negative integer constant expression is implicitly
2143 converted to an unsigned type. For example, warn about the assignment
2144 @code{x = -1} if @code{x} is unsigned. But do not warn about explicit
2145 casts like @code{(unsigned) -1}.
2147 @item -Wsign-compare
2148 @cindex warning for comparison of signed and unsigned values
2149 @cindex comparison of signed and unsigned values, warning
2150 @cindex signed and unsigned values, comparison warning
2151 Warn when a comparison between signed and unsigned values could produce
2152 an incorrect result when the signed value is converted to unsigned.
2153 This warning is also enabled by @samp{-W}; to get the other warnings
2154 of @samp{-W} without this warning, use @samp{-W -Wno-sign-compare}.
2156 @item -Waggregate-return
2157 Warn if any functions that return structures or unions are defined or
2158 called. (In languages where you can return an array, this also elicits
2161 @item -Wstrict-prototypes (C only)
2162 Warn if a function is declared or defined without specifying the
2163 argument types. (An old-style function definition is permitted without
2164 a warning if preceded by a declaration which specifies the argument
2167 @item -Wmissing-prototypes (C only)
2168 Warn if a global function is defined without a previous prototype
2169 declaration. This warning is issued even if the definition itself
2170 provides a prototype. The aim is to detect global functions that fail
2171 to be declared in header files.
2173 @item -Wmissing-declarations
2174 Warn if a global function is defined without a previous declaration.
2175 Do so even if the definition itself provides a prototype.
2176 Use this option to detect global functions that are not declared in
2179 @item -Wmissing-noreturn
2180 Warn about functions which might be candidates for attribute @code{noreturn}.
2181 Note these are only possible candidates, not absolute ones. Care should
2182 be taken to manually verify functions actually do not ever return before
2183 adding the @code{noreturn} attribute, otherwise subtle code generation
2184 bugs could be introduced. You will not get a warning for @code{main} in
2185 hosted C environments.
2187 @item -Wmissing-format-attribute
2188 If @samp{-Wformat} is enabled, also warn about functions which might be
2189 candidates for @code{format} attributes. Note these are only possible
2190 candidates, not absolute ones. GCC will guess that @code{format}
2191 attributes might be appropriate for any function that calls a function
2192 like @code{vprintf} or @code{vscanf}, but this might not always be the
2193 case, and some functions for which @code{format} attributes are
2194 appropriate may not be detected. This option has no effect unless
2195 @samp{-Wformat} is enabled (possibly by @samp{-Wall}).
2198 Warn if a structure is given the packed attribute, but the packed
2199 attribute has no effect on the layout or size of the structure.
2200 Such structures may be mis-aligned for little benefit. For
2201 instance, in this code, the variable @code{f.x} in @code{struct bar}
2202 will be misaligned even though @code{struct bar} does not itself
2203 have the packed attribute:
2210 @} __attribute__((packed));
2219 Warn if padding is included in a structure, either to align an element
2220 of the structure or to align the whole structure. Sometimes when this
2221 happens it is possible to rearrange the fields of the structure to
2222 reduce the padding and so make the structure smaller.
2224 @item -Wredundant-decls
2225 Warn if anything is declared more than once in the same scope, even in
2226 cases where multiple declaration is valid and changes nothing.
2228 @item -Wnested-externs (C only)
2229 Warn if an @code{extern} declaration is encountered within a function.
2231 @item -Wunreachable-code
2232 Warn if the compiler detects that code will never be executed.
2234 This option is intended to warn when the compiler detects that at
2235 least a whole line of source code will never be executed, because
2236 some condition is never satisfied or because it is after a
2237 procedure that never returns.
2239 It is possible for this option to produce a warning even though there
2240 are circumstances under which part of the affected line can be executed,
2241 so care should be taken when removing apparently-unreachable code.
2243 For instance, when a function is inlined, a warning may mean that the
2244 line is unreachable in only one inlined copy of the function.
2246 This option is not made part of @samp{-Wall} because in a debugging
2247 version of a program there is often substantial code which checks
2248 correct functioning of the program and is, hopefully, unreachable
2249 because the program does work. Another common use of unreachable
2250 code is to provide behaviour which is selectable at compile-time.
2253 Warn if a function can not be inlined and it was declared as inline.
2256 Warn if @samp{long long} type is used. This is default. To inhibit
2257 the warning messages, use @samp{-Wno-long-long}. Flags
2258 @samp{-Wlong-long} and @samp{-Wno-long-long} are taken into account
2259 only when @samp{-pedantic} flag is used.
2261 @item -Wdisabled-optimization
2262 Warn if a requested optimization pass is disabled. This warning does
2263 not generally indicate that there is anything wrong with your code; it
2264 merely indicates that GCC's optimizers were unable to handle the code
2265 effectively. Often, the problem is that your code is too big or too
2266 complex; GCC will refuse to optimize programs when the optimization
2267 itself is likely to take inordinate amounts of time.
2270 Make all warnings into errors.
2273 @node Debugging Options
2274 @section Options for Debugging Your Program or GCC
2275 @cindex options, debugging
2276 @cindex debugging information options
2278 GCC has various special options that are used for debugging
2279 either your program or GCC:
2283 Produce debugging information in the operating system's native format
2284 (stabs, COFF, XCOFF, or DWARF). GDB can work with this debugging
2287 On most systems that use stabs format, @samp{-g} enables use of extra
2288 debugging information that only GDB can use; this extra information
2289 makes debugging work better in GDB but will probably make other debuggers
2291 refuse to read the program. If you want to control for certain whether
2292 to generate the extra information, use @samp{-gstabs+}, @samp{-gstabs},
2293 @samp{-gxcoff+}, @samp{-gxcoff}, @samp{-gdwarf-1+}, or @samp{-gdwarf-1}
2296 Unlike most other C compilers, GCC allows you to use @samp{-g} with
2297 @samp{-O}. The shortcuts taken by optimized code may occasionally
2298 produce surprising results: some variables you declared may not exist
2299 at all; flow of control may briefly move where you did not expect it;
2300 some statements may not be executed because they compute constant
2301 results or their values were already at hand; some statements may
2302 execute in different places because they were moved out of loops.
2304 Nevertheless it proves possible to debug optimized output. This makes
2305 it reasonable to use the optimizer for programs that might have bugs.
2307 The following options are useful when GCC is generated with the
2308 capability for more than one debugging format.
2311 Produce debugging information for use by GDB. This means to use the
2312 most expressive format available (DWARF 2, stabs, or the native format
2313 if neither of those are supported), including GDB extensions if at all
2317 Produce debugging information in stabs format (if that is supported),
2318 without GDB extensions. This is the format used by DBX on most BSD
2319 systems. On MIPS, Alpha and System V Release 4 systems this option
2320 produces stabs debugging output which is not understood by DBX or SDB.
2321 On System V Release 4 systems this option requires the GNU assembler.
2324 Produce debugging information in stabs format (if that is supported),
2325 using GNU extensions understood only by the GNU debugger (GDB). The
2326 use of these extensions is likely to make other debuggers crash or
2327 refuse to read the program.
2330 Produce debugging information in COFF format (if that is supported).
2331 This is the format used by SDB on most System V systems prior to
2335 Produce debugging information in XCOFF format (if that is supported).
2336 This is the format used by the DBX debugger on IBM RS/6000 systems.
2339 Produce debugging information in XCOFF format (if that is supported),
2340 using GNU extensions understood only by the GNU debugger (GDB). The
2341 use of these extensions is likely to make other debuggers crash or
2342 refuse to read the program, and may cause assemblers other than the GNU
2343 assembler (GAS) to fail with an error.
2346 Produce debugging information in DWARF version 1 format (if that is
2347 supported). This is the format used by SDB on most System V Release 4
2351 Produce debugging information in DWARF version 1 format (if that is
2352 supported), using GNU extensions understood only by the GNU debugger
2353 (GDB). The use of these extensions is likely to make other debuggers
2354 crash or refuse to read the program.
2357 Produce debugging information in DWARF version 2 format (if that is
2358 supported). This is the format used by DBX on IRIX 6.
2361 @itemx -ggdb@var{level}
2362 @itemx -gstabs@var{level}
2363 @itemx -gcoff@var{level}
2364 @itemx -gxcoff@var{level}
2365 @itemx -gdwarf@var{level}
2366 @itemx -gdwarf-2@var{level}
2367 Request debugging information and also use @var{level} to specify how
2368 much information. The default level is 2.
2370 Level 1 produces minimal information, enough for making backtraces in
2371 parts of the program that you don't plan to debug. This includes
2372 descriptions of functions and external variables, but no information
2373 about local variables and no line numbers.
2375 Level 3 includes extra information, such as all the macro definitions
2376 present in the program. Some debuggers support macro expansion when
2381 Generate extra code to write profile information suitable for the
2382 analysis program @code{prof}. You must use this option when compiling
2383 the source files you want data about, and you must also use it when
2386 @cindex @code{gprof}
2388 Generate extra code to write profile information suitable for the
2389 analysis program @code{gprof}. You must use this option when compiling
2390 the source files you want data about, and you must also use it when
2395 Generate extra code to write profile information for basic blocks, which will
2396 record the number of times each basic block is executed, the basic block start
2397 address, and the function name containing the basic block. If @samp{-g} is
2398 used, the line number and filename of the start of the basic block will also be
2399 recorded. If not overridden by the machine description, the default action is
2400 to append to the text file @file{bb.out}.
2402 This data could be analyzed by a program like @code{tcov}. Note,
2403 however, that the format of the data is not what @code{tcov} expects.
2404 Eventually GNU @code{gprof} should be extended to process this data.
2407 Makes the compiler print out each function name as it is compiled, and
2408 print some statistics about each pass when it finishes.
2411 Generate extra code to profile basic blocks. Your executable will
2412 produce output that is a superset of that produced when @samp{-a} is
2413 used. Additional output is the source and target address of the basic
2414 blocks where a jump takes place, the number of times a jump is executed,
2415 and (optionally) the complete sequence of basic blocks being executed.
2416 The output is appended to file @file{bb.out}.
2418 You can examine different profiling aspects without recompilation. Your
2419 executable will read a list of function names from file @file{bb.in}.
2420 Profiling starts when a function on the list is entered and stops when
2421 that invocation is exited. To exclude a function from profiling, prefix
2422 its name with `-'. If a function name is not unique, you can
2423 disambiguate it by writing it in the form
2424 @samp{/path/filename.d:functionname}. Your executable will write the
2425 available paths and filenames in file @file{bb.out}.
2427 Several function names have a special meaning:
2430 Write source, target and frequency of jumps to file @file{bb.out}.
2431 @item __bb_hidecall__
2432 Exclude function calls from frequency count.
2433 @item __bb_showret__
2434 Include function returns in frequency count.
2436 Write the sequence of basic blocks executed to file @file{bbtrace.gz}.
2437 The file will be compressed using the program @samp{gzip}, which must
2438 exist in your @env{PATH}. On systems without the @samp{popen}
2439 function, the file will be named @file{bbtrace} and will not be
2440 compressed. @strong{Profiling for even a few seconds on these systems
2441 will produce a very large file.} Note: @code{__bb_hidecall__} and
2442 @code{__bb_showret__} will not affect the sequence written to
2446 Here's a short example using different profiling parameters
2447 in file @file{bb.in}. Assume function @code{foo} consists of basic blocks
2448 1 and 2 and is called twice from block 3 of function @code{main}. After
2449 the calls, block 3 transfers control to block 4 of @code{main}.
2451 With @code{__bb_trace__} and @code{main} contained in file @file{bb.in},
2452 the following sequence of blocks is written to file @file{bbtrace.gz}:
2453 0 3 1 2 1 2 4. The return from block 2 to block 3 is not shown, because
2454 the return is to a point inside the block and not to the top. The
2455 block address 0 always indicates, that control is transferred
2456 to the trace from somewhere outside the observed functions. With
2457 @samp{-foo} added to @file{bb.in}, the blocks of function
2458 @code{foo} are removed from the trace, so only 0 3 4 remains.
2460 With @code{__bb_jumps__} and @code{main} contained in file @file{bb.in},
2461 jump frequencies will be written to file @file{bb.out}. The
2462 frequencies are obtained by constructing a trace of blocks
2463 and incrementing a counter for every neighbouring pair of blocks
2464 in the trace. The trace 0 3 1 2 1 2 4 displays the following
2468 Jump from block 0x0 to block 0x3 executed 1 time(s)
2469 Jump from block 0x3 to block 0x1 executed 1 time(s)
2470 Jump from block 0x1 to block 0x2 executed 2 time(s)
2471 Jump from block 0x2 to block 0x1 executed 1 time(s)
2472 Jump from block 0x2 to block 0x4 executed 1 time(s)
2475 With @code{__bb_hidecall__}, control transfer due to call instructions
2476 is removed from the trace, that is the trace is cut into three parts: 0
2477 3 4, 0 1 2 and 0 1 2. With @code{__bb_showret__}, control transfer due
2478 to return instructions is added to the trace. The trace becomes: 0 3 1
2479 2 3 1 2 3 4. Note, that this trace is not the same, as the sequence
2480 written to @file{bbtrace.gz}. It is solely used for counting jump
2483 @item -fprofile-arcs
2484 Instrument @dfn{arcs} during compilation. For each function of your
2485 program, GCC creates a program flow graph, then finds a spanning tree
2486 for the graph. Only arcs that are not on the spanning tree have to be
2487 instrumented: the compiler adds code to count the number of times that these
2488 arcs are executed. When an arc is the only exit or only entrance to a
2489 block, the instrumentation code can be added to the block; otherwise, a
2490 new basic block must be created to hold the instrumentation code.
2492 Since not every arc in the program must be instrumented, programs
2493 compiled with this option run faster than programs compiled with
2494 @samp{-a}, which adds instrumentation code to every basic block in the
2495 program. The tradeoff: since @code{gcov} does not have
2496 execution counts for all branches, it must start with the execution
2497 counts for the instrumented branches, and then iterate over the program
2498 flow graph until the entire graph has been solved. Hence, @code{gcov}
2499 runs a little more slowly than a program which uses information from
2502 @samp{-fprofile-arcs} also makes it possible to estimate branch
2503 probabilities, and to calculate basic block execution counts. In
2504 general, basic block execution counts do not give enough information to
2505 estimate all branch probabilities. When the compiled program exits, it
2506 saves the arc execution counts to a file called
2507 @file{@var{sourcename}.da}. Use the compiler option
2508 @samp{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
2509 Control Optimization}) when recompiling, to optimize using estimated
2510 branch probabilities.
2513 @item -ftest-coverage
2514 Create data files for the @code{gcov} code-coverage utility
2515 (@pxref{Gcov,, @code{gcov}: a GCC Test Coverage Program}).
2516 The data file names begin with the name of your source file:
2519 @item @var{sourcename}.bb
2520 A mapping from basic blocks to line numbers, which @code{gcov} uses to
2521 associate basic block execution counts with line numbers.
2523 @item @var{sourcename}.bbg
2524 A list of all arcs in the program flow graph. This allows @code{gcov}
2525 to reconstruct the program flow graph, so that it can compute all basic
2526 block and arc execution counts from the information in the
2527 @code{@var{sourcename}.da} file (this last file is the output from
2528 @samp{-fprofile-arcs}).
2531 @item -d@var{letters}
2532 Says to make debugging dumps during compilation at times specified by
2533 @var{letters}. This is used for debugging the compiler. The file names
2534 for most of the dumps are made by appending a pass number and a word to
2535 the source file name (e.g. @file{foo.c.00.rtl} or @file{foo.c.01.sibling}).
2536 Here are the possible letters for use in @var{letters}, and their meanings:
2540 Annotate the assembler output with miscellaneous debugging information.
2542 Dump after computing branch probabilities, to @file{@var{file}.11.bp}.
2544 Dump after block reordering, to @file{@var{file}.26.bbro}.
2546 Dump after instruction combination, to the file @file{@var{file}.14.combine}.
2548 Dump after the first if conversion, to the file @file{@var{file}.15.ce}.
2550 Dump after delayed branch scheduling, to @file{@var{file}.29.dbr}.
2552 Dump all macro definitions, at the end of preprocessing, in addition to
2555 Dump after SSA optimizations, to @file{@var{file}.05.ssa} and
2556 @file{@var{file}.06.ussa}.
2558 Dump after the second if conversion, to @file{@var{file}.24.ce2}.
2560 Dump after life analysis, to @file{@var{file}.13.life}.
2562 Dump after purging @code{ADDRESSOF} codes, to @file{@var{file}.04.addressof}.
2564 Dump after global register allocation, to @file{@var{file}.19.greg}.
2566 Dump after post-reload CSE and other optimizations, to @file{@var{file}.20.postreload}.
2568 Dump after GCSE, to @file{@var{file}.08.gcse}.
2570 Dump after sibling call optimizations, to @file{@var{file}.01.sibling}.
2572 Dump after the first jump optimization, to @file{@var{file}.02.jump}.
2574 Dump after the last jump optimization, to @file{@var{file}.27.jump2}.
2576 Dump after conversion from registers to stack, to @file{@var{file}.29.stack}.
2578 Dump after local register allocation, to @file{@var{file}.18.lreg}.
2580 Dump after loop optimization, to @file{@var{file}.09.loop}.
2582 Dump after performing the machine dependent reorganisation pass, to
2583 @file{@var{file}.28.mach}.
2585 Dump after register renumbering, to @file{@var{file}.23.rnreg}.
2587 Dump after the register move pass, to @file{@var{file}.16.regmove}.
2589 Dump after RTL generation, to @file{@var{file}.00.rtl}.
2591 Dump after the second instruction scheduling pass, to
2592 @file{@var{file}.25.sched2}.
2594 Dump after CSE (including the jump optimization that sometimes follows
2595 CSE), to @file{@var{file}.03.cse}.
2597 Dump after the first instruction scheduling pass, to
2598 @file{@var{file}.17.sched}.
2600 Dump after the second CSE pass (including the jump optimization that
2601 sometimes follows CSE), to @file{@var{file}.10.cse2}.
2603 Dump after the second flow pass, to @file{@var{file}.21.flow2}.
2605 Dump after dead code elimination, to @file{@var{file}.06.dce}.
2607 Dump after the peephole pass, to @file{@var{file}.22.peephole2}.
2609 Produce all the dumps listed above.
2611 Print statistics on memory usage, at the end of the run, to
2614 Annotate the assembler output with a comment indicating which
2615 pattern and alternative was used. The length of each instruction is
2618 Dump the RTL in the assembler output as a comment before each instruction.
2619 Also turns on @samp{-dp} annotation.
2621 For each of the other indicated dump files (except for
2622 @file{@var{file}.00.rtl}), dump a representation of the control flow graph
2623 suitable for viewing with VCG to @file{@var{file}.@var{pass}.vcg}.
2625 Just generate RTL for a function instead of compiling it. Usually used
2628 Dump debugging information during parsing, to standard error.
2631 @item -fdump-unnumbered
2632 When doing debugging dumps (see -d option above), suppress instruction
2633 numbers and line number note output. This makes it more feasible to
2634 use diff on debugging dumps for compiler invocations with different
2635 options, in particular with and without -g.
2637 @item -fdump-translation-unit-@var{file} (C and C++ only)
2638 Dump a representation of the tree structure for the entire translation
2641 @item -fpretend-float
2642 When running a cross-compiler, pretend that the target machine uses the
2643 same floating point format as the host machine. This causes incorrect
2644 output of the actual floating constants, but the actual instruction
2645 sequence will probably be the same as GCC would make when running on
2649 Store the usual ``temporary'' intermediate files permanently; place them
2650 in the current directory and name them based on the source file. Thus,
2651 compiling @file{foo.c} with @samp{-c -save-temps} would produce files
2652 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
2653 preprocessed @file{foo.i} output file even though the compiler now
2654 normally uses an integrated preprocessor.
2657 Report the CPU time taken by each subprocess in the compilation
2658 sequence. For C source files, this is the compiler proper and assembler
2659 (plus the linker if linking is done). The output looks like this:
2666 The first number on each line is the ``user time,'' that is time spent
2667 executing the program itself. The second number is ``system time,''
2668 time spent executing operating system routines on behalf of the program.
2669 Both numbers are in seconds.
2671 @item -print-file-name=@var{library}
2672 Print the full absolute name of the library file @var{library} that
2673 would be used when linking---and don't do anything else. With this
2674 option, GCC does not compile or link anything; it just prints the
2677 @item -print-prog-name=@var{program}
2678 Like @samp{-print-file-name}, but searches for a program such as @samp{cpp}.
2680 @item -print-libgcc-file-name
2681 Same as @samp{-print-file-name=libgcc.a}.
2683 This is useful when you use @samp{-nostdlib} or @samp{-nodefaultlibs}
2684 but you do want to link with @file{libgcc.a}. You can do
2687 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
2690 @item -print-search-dirs
2691 Print the name of the configured installation directory and a list of
2692 program and library directories gcc will search---and don't do anything else.
2694 This is useful when gcc prints the error message
2695 @samp{installation problem, cannot exec cpp0: No such file or directory}.
2696 To resolve this you either need to put @file{cpp0} and the other compiler
2697 components where gcc expects to find them, or you can set the environment
2698 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
2699 Don't forget the trailing '/'.
2700 @xref{Environment Variables}.
2703 @node Optimize Options
2704 @section Options That Control Optimization
2705 @cindex optimize options
2706 @cindex options, optimization
2708 These options control various sorts of optimizations:
2713 Optimize. Optimizing compilation takes somewhat more time, and a lot
2714 more memory for a large function.
2716 Without @samp{-O}, the compiler's goal is to reduce the cost of
2717 compilation and to make debugging produce the expected results.
2718 Statements are independent: if you stop the program with a breakpoint
2719 between statements, you can then assign a new value to any variable or
2720 change the program counter to any other statement in the function and
2721 get exactly the results you would expect from the source code.
2723 Without @samp{-O}, the compiler only allocates variables declared
2724 @code{register} in registers. The resulting compiled code is a little
2725 worse than produced by PCC without @samp{-O}.
2727 With @samp{-O}, the compiler tries to reduce code size and execution
2730 When you specify @samp{-O}, the compiler turns on @samp{-fthread-jumps}
2731 and @samp{-fdefer-pop} on all machines. The compiler turns on
2732 @samp{-fdelayed-branch} on machines that have delay slots, and
2733 @samp{-fomit-frame-pointer} on machines that can support debugging even
2734 without a frame pointer. On some machines the compiler also turns
2735 on other flags.@refill
2738 Optimize even more. GCC performs nearly all supported optimizations
2739 that do not involve a space-speed tradeoff. The compiler does not
2740 perform loop unrolling or function inlining when you specify @samp{-O2}.
2741 As compared to @samp{-O}, this option increases both compilation time
2742 and the performance of the generated code.
2744 @samp{-O2} turns on all optional optimizations except for loop unrolling,
2745 function inlining, and register renaming. It also turns on the
2746 @samp{-fforce-mem} option on all machines and frame pointer elimination
2747 on machines where doing so does not interfere with debugging.
2750 Optimize yet more. @samp{-O3} turns on all optimizations specified by
2751 @samp{-O2} and also turns on the @samp{-finline-functions} and
2752 @samp{-frename-registers} options.
2758 Optimize for size. @samp{-Os} enables all @samp{-O2} optimizations that
2759 do not typically increase code size. It also performs further
2760 optimizations designed to reduce code size.
2762 If you use multiple @samp{-O} options, with or without level numbers,
2763 the last such option is the one that is effective.
2766 Options of the form @samp{-f@var{flag}} specify machine-independent
2767 flags. Most flags have both positive and negative forms; the negative
2768 form of @samp{-ffoo} would be @samp{-fno-foo}. In the table below,
2769 only one of the forms is listed---the one which is not the default.
2770 You can figure out the other form by either removing @samp{no-} or
2775 Do not store floating point variables in registers, and inhibit other
2776 options that might change whether a floating point value is taken from a
2779 @cindex floating point precision
2780 This option prevents undesirable excess precision on machines such as
2781 the 68000 where the floating registers (of the 68881) keep more
2782 precision than a @code{double} is supposed to have. Similarly for the
2783 x86 architecture. For most programs, the excess precision does only
2784 good, but a few programs rely on the precise definition of IEEE floating
2785 point. Use @samp{-ffloat-store} for such programs, after modifying
2786 them to store all pertinent intermediate computations into variables.
2788 @item -fno-default-inline
2789 Do not make member functions inline by default merely because they are
2790 defined inside the class scope (C++ only). Otherwise, when you specify
2791 @w{@samp{-O}}, member functions defined inside class scope are compiled
2792 inline by default; i.e., you don't need to add @samp{inline} in front of
2793 the member function name.
2795 @item -fno-defer-pop
2796 Always pop the arguments to each function call as soon as that function
2797 returns. For machines which must pop arguments after a function call,
2798 the compiler normally lets arguments accumulate on the stack for several
2799 function calls and pops them all at once.
2802 Force memory operands to be copied into registers before doing
2803 arithmetic on them. This produces better code by making all memory
2804 references potential common subexpressions. When they are not common
2805 subexpressions, instruction combination should eliminate the separate
2806 register-load. The @samp{-O2} option turns on this option.
2809 Force memory address constants to be copied into registers before
2810 doing arithmetic on them. This may produce better code just as
2811 @samp{-fforce-mem} may.
2813 @item -fomit-frame-pointer
2814 Don't keep the frame pointer in a register for functions that
2815 don't need one. This avoids the instructions to save, set up and
2816 restore frame pointers; it also makes an extra register available
2817 in many functions. @strong{It also makes debugging impossible on
2821 On some machines, such as the Vax, this flag has no effect, because
2822 the standard calling sequence automatically handles the frame pointer
2823 and nothing is saved by pretending it doesn't exist. The
2824 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
2825 whether a target machine supports this flag. @xref{Registers}.@refill
2828 On some machines, such as the Vax, this flag has no effect, because
2829 the standard calling sequence automatically handles the frame pointer
2830 and nothing is saved by pretending it doesn't exist. The
2831 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
2832 whether a target machine supports this flag. @xref{Registers,,Register
2833 Usage, gcc.info, Using and Porting GCC}.@refill
2836 @item -foptimize-sibling-calls
2837 Optimize sibling and tail recursive calls.
2840 This option generates traps for signed overflow on addition, subtraction,
2841 multiplication operations.
2844 Don't pay attention to the @code{inline} keyword. Normally this option
2845 is used to keep the compiler from expanding any functions inline.
2846 Note that if you are not optimizing, no functions can be expanded inline.
2848 @item -finline-functions
2849 Integrate all simple functions into their callers. The compiler
2850 heuristically decides which functions are simple enough to be worth
2851 integrating in this way.
2853 If all calls to a given function are integrated, and the function is
2854 declared @code{static}, then the function is normally not output as
2855 assembler code in its own right.
2857 @item -finline-limit=@var{n}
2858 By default, gcc limits the size of functions that can be inlined. This flag
2859 allows the control of this limit for functions that are explicitly marked as
2860 inline (ie marked with the inline keyword or defined within the class
2861 definition in c++). @var{n} is the size of functions that can be inlined in
2862 number of pseudo instructions (not counting parameter handling). The default
2863 value of n is 10000. Increasing this value can result in more inlined code at
2864 the cost of compilation time and memory consumption. Decreasing usually makes
2865 the compilation faster and less code will be inlined (which presumably
2866 means slower programs). This option is particularly useful for programs that
2867 use inlining heavily such as those based on recursive templates with c++.
2869 @emph{Note:} pseudo instruction represents, in this particular context, an
2870 abstract measurement of function's size. In no way, it represents a count
2871 of assembly instructions and as such its exact meaning might change from one
2872 release to an another.
2874 @item -fkeep-inline-functions
2875 Even if all calls to a given function are integrated, and the function
2876 is declared @code{static}, nevertheless output a separate run-time
2877 callable version of the function. This switch does not affect
2878 @code{extern inline} functions.
2880 @item -fkeep-static-consts
2881 Emit variables declared @code{static const} when optimization isn't turned
2882 on, even if the variables aren't referenced.
2884 GCC enables this option by default. If you want to force the compiler to
2885 check if the variable was referenced, regardless of whether or not
2886 optimization is turned on, use the @samp{-fno-keep-static-consts} option.
2888 @item -fno-function-cse
2889 Do not put function addresses in registers; make each instruction that
2890 calls a constant function contain the function's address explicitly.
2892 This option results in less efficient code, but some strange hacks
2893 that alter the assembler output may be confused by the optimizations
2894 performed when this option is not used.
2897 This option allows GCC to violate some ISO or IEEE rules and/or
2898 specifications in the interest of optimizing code for speed. For
2899 example, it allows the compiler to assume arguments to the @code{sqrt}
2900 function are non-negative numbers and that no floating-point values
2903 This option should never be turned on by any @samp{-O} option since
2904 it can result in incorrect output for programs which depend on
2905 an exact implementation of IEEE or ISO rules/specifications for
2908 @item -fno-math-errno
2909 Do not set ERRNO after calling math functions that are executed
2910 with a single instruction, e.g., sqrt. A program that relies on
2911 IEEE exceptions for math error handling may want to use this flag
2912 for speed while maintaining IEEE arithmetic compatibility.
2914 The default is @samp{-fmath-errno}. The @samp{-ffast-math} option
2915 sets @samp{-fno-math-errno}.
2918 @c following causes underfulls.. they don't look great, but we deal.
2920 The following options control specific optimizations. The @samp{-O2}
2921 option turns on all of these optimizations except @samp{-funroll-loops}
2922 and @samp{-funroll-all-loops}. On most machines, the @samp{-O} option
2923 turns on the @samp{-fthread-jumps} and @samp{-fdelayed-branch} options,
2924 but specific machines may handle it differently.
2926 You can use the following flags in the rare cases when ``fine-tuning''
2927 of optimizations to be performed is desired.
2930 @item -fstrength-reduce
2931 Perform the optimizations of loop strength reduction and
2932 elimination of iteration variables.
2934 @item -fthread-jumps
2935 Perform optimizations where we check to see if a jump branches to a
2936 location where another comparison subsumed by the first is found. If
2937 so, the first branch is redirected to either the destination of the
2938 second branch or a point immediately following it, depending on whether
2939 the condition is known to be true or false.
2941 @item -fcse-follow-jumps
2942 In common subexpression elimination, scan through jump instructions
2943 when the target of the jump is not reached by any other path. For
2944 example, when CSE encounters an @code{if} statement with an
2945 @code{else} clause, CSE will follow the jump when the condition
2948 @item -fcse-skip-blocks
2949 This is similar to @samp{-fcse-follow-jumps}, but causes CSE to
2950 follow jumps which conditionally skip over blocks. When CSE
2951 encounters a simple @code{if} statement with no else clause,
2952 @samp{-fcse-skip-blocks} causes CSE to follow the jump around the
2953 body of the @code{if}.
2955 @item -frerun-cse-after-loop
2956 Re-run common subexpression elimination after loop optimizations has been
2959 @item -frerun-loop-opt
2960 Run the loop optimizer twice.
2963 Perform a global common subexpression elimination pass.
2964 This pass also performs global constant and copy propagation.
2966 @item -fdelete-null-pointer-checks
2967 Use global dataflow analysis to identify and eliminate useless null
2968 pointer checks. Programs which rely on NULL pointer dereferences @emph{not}
2969 halting the program may not work properly with this option. Use
2970 -fno-delete-null-pointer-checks to disable this optimizing for programs
2971 which depend on that behavior.
2973 @item -fexpensive-optimizations
2974 Perform a number of minor optimizations that are relatively expensive.
2976 @item -foptimize-register-move
2978 Attempt to reassign register numbers in move instructions and as
2979 operands of other simple instructions in order to maximize the amount of
2980 register tying. This is especially helpful on machines with two-operand
2981 instructions. GCC enables this optimization by default with @samp{-O2}
2984 Note @option{-fregmove} and @option{-foptimize-register-move} are the same
2987 @item -fdelayed-branch
2988 If supported for the target machine, attempt to reorder instructions
2989 to exploit instruction slots available after delayed branch
2992 @item -fschedule-insns
2993 If supported for the target machine, attempt to reorder instructions to
2994 eliminate execution stalls due to required data being unavailable. This
2995 helps machines that have slow floating point or memory load instructions
2996 by allowing other instructions to be issued until the result of the load
2997 or floating point instruction is required.
2999 @item -fschedule-insns2
3000 Similar to @samp{-fschedule-insns}, but requests an additional pass of
3001 instruction scheduling after register allocation has been done. This is
3002 especially useful on machines with a relatively small number of
3003 registers and where memory load instructions take more than one cycle.
3005 @item -ffunction-sections
3006 @itemx -fdata-sections
3007 Place each function or data item into its own section in the output
3008 file if the target supports arbitrary sections. The name of the
3009 function or the name of the data item determines the section's name
3012 Use these options on systems where the linker can perform optimizations
3013 to improve locality of reference in the instruction space. HPPA
3014 processors running HP-UX and Sparc processors running Solaris 2 have
3015 linkers with such optimizations. Other systems using the ELF object format
3016 as well as AIX may have these optimizations in the future.
3018 Only use these options when there are significant benefits from doing
3019 so. When you specify these options, the assembler and linker will
3020 create larger object and executable files and will also be slower.
3021 You will not be able to use @code{gprof} on all systems if you
3022 specify this option and you may have problems with debugging if
3023 you specify both this option and @samp{-g}.
3025 @item -fcaller-saves
3026 Enable values to be allocated in registers that will be clobbered by
3027 function calls, by emitting extra instructions to save and restore the
3028 registers around such calls. Such allocation is done only when it
3029 seems to result in better code than would otherwise be produced.
3031 This option is always enabled by default on certain machines, usually
3032 those which have no call-preserved registers to use instead.
3034 For all machines, optimization level 2 and higher enables this flag by
3037 @item -funroll-loops
3038 Perform the optimization of loop unrolling. This is only done for loops
3039 whose number of iterations can be determined at compile time or run time.
3040 @samp{-funroll-loops} implies both @samp{-fstrength-reduce} and
3041 @samp{-frerun-cse-after-loop}.
3043 @item -funroll-all-loops
3044 Perform the optimization of loop unrolling. This is done for all loops
3045 and usually makes programs run more slowly. @samp{-funroll-all-loops}
3046 implies @samp{-fstrength-reduce} as well as @samp{-frerun-cse-after-loop}.
3048 @item -fmove-all-movables
3049 Forces all invariant computations in loops to be moved
3052 @item -freduce-all-givs
3053 Forces all general-induction variables in loops to be
3056 @emph{Note:} When compiling programs written in Fortran,
3057 @samp{-fmove-all-movables} and @samp{-freduce-all-givs} are enabled
3058 by default when you use the optimizer.
3060 These options may generate better or worse code; results are highly
3061 dependent on the structure of loops within the source code.
3063 These two options are intended to be removed someday, once
3064 they have helped determine the efficacy of various
3065 approaches to improving loop optimizations.
3067 Please let us (@w{@email{gcc@@gcc.gnu.org}} and @w{@email{fortran@@gnu.org}})
3068 know how use of these options affects
3069 the performance of your production code.
3070 We're very interested in code that runs @emph{slower}
3071 when these options are @emph{enabled}.
3074 Disable any machine-specific peephole optimizations.
3076 @item -fbranch-probabilities
3077 After running a program compiled with @samp{-fprofile-arcs}
3078 (@pxref{Debugging Options,, Options for Debugging Your Program or
3079 @command{gcc}}), you can compile it a second time using
3080 @samp{-fbranch-probabilities}, to improve optimizations based on
3081 guessing the path a branch might take.
3084 With @samp{-fbranch-probabilities}, GCC puts a @samp{REG_EXEC_COUNT}
3085 note on the first instruction of each basic block, and a
3086 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
3087 These can be used to improve optimization. Currently, they are only
3088 used in one place: in @file{reorg.c}, instead of guessing which path a
3089 branch is mostly to take, the @samp{REG_BR_PROB} values are used to
3090 exactly determine which path is taken more often.
3093 @item -fstrict-aliasing
3094 Allows the compiler to assume the strictest aliasing rules applicable to
3095 the language being compiled. For C (and C++), this activates
3096 optimizations based on the type of expressions. In particular, an
3097 object of one type is assumed never to reside at the same address as an
3098 object of a different type, unless the types are almost the same. For
3099 example, an @code{unsigned int} can alias an @code{int}, but not a
3100 @code{void*} or a @code{double}. A character type may alias any other
3103 Pay special attention to code like this:
3116 The practice of reading from a different union member than the one most
3117 recently written to (called ``type-punning'') is common. Even with
3118 @samp{-fstrict-aliasing}, type-punning is allowed, provided the memory
3119 is accessed through the union type. So, the code above will work as
3120 expected. However, this code might not:
3132 Every language that wishes to perform language-specific alias analysis
3133 should define a function that computes, given an @code{tree}
3134 node, an alias set for the node. Nodes in different alias sets are not
3135 allowed to alias. For an example, see the C front-end function
3136 @code{c_get_alias_set}.
3139 @item -falign-functions
3140 @itemx -falign-functions=@var{n}
3141 Align the start of functions to the next power-of-two greater than
3142 @var{n}, skipping up to @var{n} bytes. For instance,
3143 @samp{-falign-functions=32} aligns functions to the next 32-byte
3144 boundary, but @samp{-falign-functions=24} would align to the next
3145 32-byte boundary only if this can be done by skipping 23 bytes or less.
3147 @samp{-fno-align-functions} and @samp{-falign-functions=1} are
3148 equivalent and mean that functions will not be aligned.
3150 Some assemblers only support this flag when @var{n} is a power of two;
3151 in that case, it is rounded up.
3153 If @var{n} is not specified, use a machine-dependent default.
3155 @item -falign-labels
3156 @itemx -falign-labels=@var{n}
3157 Align all branch targets to a power-of-two boundary, skipping up to
3158 @var{n} bytes like @samp{-falign-functions}. This option can easily
3159 make code slower, because it must insert dummy operations for when the
3160 branch target is reached in the usual flow of the code.
3162 If @samp{-falign-loops} or @samp{-falign-jumps} are applicable and
3163 are greater than this value, then their values are used instead.
3165 If @var{n} is not specified, use a machine-dependent default which is
3166 very likely to be @samp{1}, meaning no alignment.
3169 @itemx -falign-loops=@var{n}
3170 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
3171 like @samp{-falign-functions}. The hope is that the loop will be
3172 executed many times, which will make up for any execution of the dummy
3175 If @var{n} is not specified, use a machine-dependent default.
3178 @itemx -falign-jumps=@var{n}
3179 Align branch targets to a power-of-two boundary, for branch targets
3180 where the targets can only be reached by jumping, skipping up to @var{n}
3181 bytes like @samp{-falign-functions}. In this case, no dummy operations
3184 If @var{n} is not specified, use a machine-dependent default.
3187 Perform optimizations in static single assignment form. Each function's
3188 flow graph is translated into SSA form, optimizations are performed, and
3189 the flow graph is translated back from SSA form. User's should not
3190 specify this option, since it is not yet ready for production use.
3193 Perform dead-code elimination in SSA form. Requires @samp{-fssa}. Like
3194 @samp{-fssa}, this is an experimental feature.
3196 @item -fsingle-precision-constant
3197 Treat floating point constant as single precision constant instead of
3198 implicitly converting it to double precision constant.
3200 @item -frename-registers
3201 Attempt to avoid false dependancies in scheduled code by making use
3202 of registers left over after register allocation. This optimization
3203 will most benefit processors with lots of registers. It can, however,
3204 make debugging impossible, since variables will no longer stay in
3205 a ``home register''.
3208 @node Preprocessor Options
3209 @section Options Controlling the Preprocessor
3210 @cindex preprocessor options
3211 @cindex options, preprocessor
3213 These options control the C preprocessor, which is run on each C source
3214 file before actual compilation.
3216 If you use the @samp{-E} option, nothing is done except preprocessing.
3217 Some of these options make sense only together with @samp{-E} because
3218 they cause the preprocessor output to be unsuitable for actual
3222 @item -include @var{file}
3223 Process @var{file} as input before processing the regular input file.
3224 In effect, the contents of @var{file} are compiled first. Any @samp{-D}
3225 and @samp{-U} options on the command line are always processed before
3226 @samp{-include @var{file}}, regardless of the order in which they are
3227 written. All the @samp{-include} and @samp{-imacros} options are
3228 processed in the order in which they are written.
3230 @item -imacros @var{file}
3231 Process @var{file} as input, discarding the resulting output, before
3232 processing the regular input file. Because the output generated from
3233 @var{file} is discarded, the only effect of @samp{-imacros @var{file}}
3234 is to make the macros defined in @var{file} available for use in the
3235 main input. All the @samp{-include} and @samp{-imacros} options are
3236 processed in the order in which they are written.
3238 @item -idirafter @var{dir}
3239 @cindex second include path
3240 Add the directory @var{dir} to the second include path. The directories
3241 on the second include path are searched when a header file is not found
3242 in any of the directories in the main include path (the one that
3245 @item -iprefix @var{prefix}
3246 Specify @var{prefix} as the prefix for subsequent @samp{-iwithprefix}
3249 @item -iwithprefix @var{dir}
3250 Add a directory to the second include path. The directory's name is
3251 made by concatenating @var{prefix} and @var{dir}, where @var{prefix} was
3252 specified previously with @samp{-iprefix}. If you have not specified a
3253 prefix yet, the directory containing the installed passes of the
3254 compiler is used as the default.
3256 @item -iwithprefixbefore @var{dir}
3257 Add a directory to the main include path. The directory's name is made
3258 by concatenating @var{prefix} and @var{dir}, as in the case of
3259 @samp{-iwithprefix}.
3261 @item -isystem @var{dir}
3262 Add a directory to the beginning of the second include path, marking it
3263 as a system directory, so that it gets the same special treatment as
3264 is applied to the standard system directories.
3267 Do not search the standard system directories for header files. Only
3268 the directories you have specified with @samp{-I} options (and the
3269 current directory, if appropriate) are searched. @xref{Directory
3270 Options}, for information on @samp{-I}.
3272 By using both @samp{-nostdinc} and @samp{-I-}, you can limit the include-file
3273 search path to only those directories you specify explicitly.
3277 When searching for a header file in a directory, remap file names if a
3278 file named @file{header.gcc} exists in that directory. This can be used
3279 to work around limitations of file systems with file name restrictions.
3280 The @file{header.gcc} file should contain a series of lines with two
3281 tokens on each line: the first token is the name to map, and the second
3282 token is the actual name to use.
3285 Do not predefine any nonstandard macros. (Including architecture flags).
3288 Run only the C preprocessor. Preprocess all the C source files
3289 specified and output the results to standard output or to the
3290 specified output file.
3293 Tell the preprocessor not to discard comments. Used with the
3297 Tell the preprocessor not to generate @samp{#line} directives.
3298 Used with the @samp{-E} option.
3301 @cindex dependencies, make
3304 Instead of outputting the result of preprocessing, output a rule
3305 suitable for @code{make} describing the dependencies of the main source
3306 file. The preprocessor outputs one @code{make} rule containing the
3307 object file name for that source file, a colon, and the names of all the
3308 included files. If there are many included files then the rule is split
3309 into several lines using @samp{\}-newline.
3311 @samp{-M} implies @samp{-E}.
3315 Like @samp{-M}, but mention only the files included with @samp{#include
3316 "@var{file}"}. System header files included with @samp{#include
3317 <@var{file}>} are omitted.
3319 @item -MF @var{file}
3321 When used with @samp{-M} or @samp{-MM}, specifies a file to write the
3322 dependencies to. This allows the preprocessor to write the preprocessed
3323 file to stdout normally. If no @samp{-MF} switch is given, CPP sends
3324 the rules to stdout and suppresses normal preprocessed output.
3326 Another way to specify output of a @code{make} rule is by setting
3327 the environment variable @env{DEPENDENCIES_OUTPUT} (@pxref{Environment
3332 When used with @samp{-M} or @samp{-MM}, @samp{-MG} says to treat missing
3333 header files as generated files and assume they live in the same
3334 directory as the source file. It suppresses preprocessed output, as a
3335 missing header file is ordinarily an error.
3337 This feature is used in automatic updating of makefiles.
3341 This option instructs CPP to add a phony target for each dependency
3342 other than the main file, causing each to depend on nothing. These
3343 dummy rules work around errors @code{make} gives if you remove header
3344 files without updating the @code{Makefile} to match.
3346 This is typical output:-
3349 /tmp/test.o: /tmp/test.c /tmp/test.h
3354 @item -MQ @var{target}
3355 @item -MT @var{target}
3358 By default CPP uses the main file name, including any path, and appends
3359 the object suffix, normally ``.o'', to it to obtain the name of the
3360 target for dependency generation. With @samp{-MT} you can specify a
3361 target yourself, overriding the default one.
3363 If you want multiple targets, you can specify them as a single argument
3364 to @samp{-MT}, or use multiple @samp{-MT} options.
3366 The targets you specify are output in the order they appear on the
3367 command line. @samp{-MQ} is identical to @samp{-MT}, except that the
3368 target name is quoted for Make, but with @samp{-MT} it isn't. For
3369 example, -MT '$(objpfx)foo.o' gives
3372 $(objpfx)foo.o: /tmp/foo.c
3375 but -MQ '$(objpfx)foo.o' gives
3378 $$(objpfx)foo.o: /tmp/foo.c
3381 The default target is automatically quoted, as if it were given with
3385 Print the name of each header file used, in addition to other normal
3388 @item -A@var{question}(@var{answer})
3389 Assert the answer @var{answer} for @var{question}, in case it is tested
3390 with a preprocessing conditional such as @samp{#if
3391 #@var{question}(@var{answer})}. @samp{-A-} disables the standard
3392 assertions that normally describe the target machine.
3395 Define macro @var{macro} with the string @samp{1} as its definition.
3397 @item -D@var{macro}=@var{defn}
3398 Define macro @var{macro} as @var{defn}. All instances of @samp{-D} on
3399 the command line are processed before any @samp{-U} options.
3401 Any @samp{-D} and @samp{-U} options on the command line are processed in
3402 order, and always before @samp{-imacros @var{file}}, regardless of the
3403 order in which they are written.
3406 Undefine macro @var{macro}. @samp{-U} options are evaluated after all
3407 @samp{-D} options, but before any @samp{-include} and @samp{-imacros}
3410 Any @samp{-D} and @samp{-U} options on the command line are processed in
3411 order, and always before @samp{-imacros @var{file}}, regardless of the
3412 order in which they are written.
3415 Tell the preprocessor to output only a list of the macro definitions
3416 that are in effect at the end of preprocessing. Used with the @samp{-E}
3420 Tell the preprocessing to pass all macro definitions into the output, in
3421 their proper sequence in the rest of the output.
3424 Like @samp{-dD} except that the macro arguments and contents are omitted.
3425 Only @samp{#define @var{name}} is included in the output.
3429 Output @samp{#include} directives in addition to the result of
3434 Process ISO standard trigraph sequences. These are three-character
3435 sequences, all starting with @samp{??}, that are defined by ISO C to
3436 stand for single characters. For example, @samp{??/} stands for
3437 @samp{\}, so @samp{'??/n'} is a character constant for a newline. By
3438 default, GCC ignores trigraphs, but in standard-conforming modes it
3439 converts them. See the @samp{-std} and @samp{-ansi} options.
3441 The nine trigraph sequences are
3472 Trigraph support is not popular, so many compilers do not implement it
3473 properly. Portable code should not rely on trigraphs being either
3474 converted or ignored.
3476 @item -Wp,@var{option}
3477 Pass @var{option} as an option to the preprocessor. If @var{option}
3478 contains commas, it is split into multiple options at the commas.
3481 @node Assembler Options
3482 @section Passing Options to the Assembler
3484 @c prevent bad page break with this line
3485 You can pass options to the assembler.
3488 @item -Wa,@var{option}
3489 Pass @var{option} as an option to the assembler. If @var{option}
3490 contains commas, it is split into multiple options at the commas.
3494 @section Options for Linking
3495 @cindex link options
3496 @cindex options, linking
3498 These options come into play when the compiler links object files into
3499 an executable output file. They are meaningless if the compiler is
3500 not doing a link step.
3504 @item @var{object-file-name}
3505 A file name that does not end in a special recognized suffix is
3506 considered to name an object file or library. (Object files are
3507 distinguished from libraries by the linker according to the file
3508 contents.) If linking is done, these object files are used as input
3514 If any of these options is used, then the linker is not run, and
3515 object file names should not be used as arguments. @xref{Overall
3519 @item -l@var{library}
3520 Search the library named @var{library} when linking.
3522 It makes a difference where in the command you write this option; the
3523 linker searches processes libraries and object files in the order they
3524 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
3525 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
3526 to functions in @samp{z}, those functions may not be loaded.
3528 The linker searches a standard list of directories for the library,
3529 which is actually a file named @file{lib@var{library}.a}. The linker
3530 then uses this file as if it had been specified precisely by name.
3532 The directories searched include several standard system directories
3533 plus any that you specify with @samp{-L}.
3535 Normally the files found this way are library files---archive files
3536 whose members are object files. The linker handles an archive file by
3537 scanning through it for members which define symbols that have so far
3538 been referenced but not defined. But if the file that is found is an
3539 ordinary object file, it is linked in the usual fashion. The only
3540 difference between using an @samp{-l} option and specifying a file name
3541 is that @samp{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
3542 and searches several directories.
3545 You need this special case of the @samp{-l} option in order to
3546 link an Objective C program.
3549 Do not use the standard system startup files when linking.
3550 The standard system libraries are used normally, unless @option{-nostdlib}
3551 or @option{-nodefaultlibs} is used.
3553 @item -nodefaultlibs
3554 Do not use the standard system libraries when linking.
3555 Only the libraries you specify will be passed to the linker.
3556 The standard startup files are used normally, unless @option{-nostartfiles}
3557 is used. The compiler may generate calls to memcmp, memset, and memcpy
3558 for System V (and ISO C) environments or to bcopy and bzero for
3559 BSD environments. These entries are usually resolved by entries in
3560 libc. These entry points should be supplied through some other
3561 mechanism when this option is specified.
3564 Do not use the standard system startup files or libraries when linking.
3565 No startup files and only the libraries you specify will be passed to
3566 the linker. The compiler may generate calls to memcmp, memset, and memcpy
3567 for System V (and ISO C) environments or to bcopy and bzero for
3568 BSD environments. These entries are usually resolved by entries in
3569 libc. These entry points should be supplied through some other
3570 mechanism when this option is specified.
3572 @cindex @code{-lgcc}, use with @code{-nostdlib}
3573 @cindex @code{-nostdlib} and unresolved references
3574 @cindex unresolved references and @code{-nostdlib}
3575 @cindex @code{-lgcc}, use with @code{-nodefaultlibs}
3576 @cindex @code{-nodefaultlibs} and unresolved references
3577 @cindex unresolved references and @code{-nodefaultlibs}
3578 One of the standard libraries bypassed by @samp{-nostdlib} and
3579 @samp{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
3580 that GCC uses to overcome shortcomings of particular machines, or special
3581 needs for some languages.
3583 (@xref{Interface,,Interfacing to GCC Output}, for more discussion of
3587 (@xref{Interface,,Interfacing to GCC Output,gcc.info,Porting GCC},
3588 for more discussion of @file{libgcc.a}.)
3590 In most cases, you need @file{libgcc.a} even when you want to avoid
3591 other standard libraries. In other words, when you specify @samp{-nostdlib}
3592 or @samp{-nodefaultlibs} you should usually specify @samp{-lgcc} as well.
3593 This ensures that you have no unresolved references to internal GCC
3594 library subroutines. (For example, @samp{__main}, used to ensure C++
3595 constructors will be called; @pxref{Collect2,,@command{collect2}}.)
3598 Remove all symbol table and relocation information from the executable.
3601 On systems that support dynamic linking, this prevents linking with the shared
3602 libraries. On other systems, this option has no effect.
3605 Produce a shared object which can then be linked with other objects to
3606 form an executable. Not all systems support this option. For predictable
3607 results, you must also specify the same set of options that were used to
3608 generate code (@samp{-fpic}, @samp{-fPIC}, or model suboptions)
3609 when you specify this option.@footnote{On some systems, @samp{gcc -shared}
3610 needs to build supplementary stub code for constructors to work. On
3611 multi-libbed systems, @samp{gcc -shared} must select the correct support
3612 libraries to link against. Failing to supply the correct flags may lead
3613 to subtle defects. Supplying them in cases where they are not necessary
3616 @item -shared-libgcc
3617 @itemx -static-libgcc
3618 On systems that provide @file{libgcc} as a shared library, these options
3619 force the use of either the shared or static version respectively.
3620 If no shared version of @file{libgcc} was built when the compiler was
3621 configured, these options have no effect.
3623 There are several situations in which an application should use the
3624 shared @file{libgcc} instead of the static version. The most common
3625 of these is when the application wishes to throw and catch exceptions
3626 across different shared libraries. In that case, each of the libraries
3627 as well as the application itself should use the shared @file{libgcc}.
3629 At present the GCC driver makes no attempt to recognize the situations
3630 in which the shared @file{libgcc} should be used, and defaults to using
3631 the static @file{libgcc} always. This will likely change in the future,
3632 at which time @samp{-static-libgcc} becomes useful as a means for
3633 overriding GCC's choice.
3636 Bind references to global symbols when building a shared object. Warn
3637 about any unresolved references (unless overridden by the link editor
3638 option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support
3641 @item -Xlinker @var{option}
3642 Pass @var{option} as an option to the linker. You can use this to
3643 supply system-specific linker options which GCC does not know how to
3646 If you want to pass an option that takes an argument, you must use
3647 @samp{-Xlinker} twice, once for the option and once for the argument.
3648 For example, to pass @samp{-assert definitions}, you must write
3649 @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write
3650 @samp{-Xlinker "-assert definitions"}, because this passes the entire
3651 string as a single argument, which is not what the linker expects.
3653 @item -Wl,@var{option}
3654 Pass @var{option} as an option to the linker. If @var{option} contains
3655 commas, it is split into multiple options at the commas.
3657 @item -u @var{symbol}
3658 Pretend the symbol @var{symbol} is undefined, to force linking of
3659 library modules to define it. You can use @samp{-u} multiple times with
3660 different symbols to force loading of additional library modules.
3663 @node Directory Options
3664 @section Options for Directory Search
3665 @cindex directory options
3666 @cindex options, directory search
3669 These options specify directories to search for header files, for
3670 libraries and for parts of the compiler:
3674 Add the directory @var{dir} to the head of the list of directories to be
3675 searched for header files. This can be used to override a system header
3676 file, substituting your own version, since these directories are
3677 searched before the system header file directories. If you use more
3678 than one @samp{-I} option, the directories are scanned in left-to-right
3679 order; the standard system directories come after.
3682 Any directories you specify with @samp{-I} options before the @samp{-I-}
3683 option are searched only for the case of @samp{#include "@var{file}"};
3684 they are not searched for @samp{#include <@var{file}>}.
3686 If additional directories are specified with @samp{-I} options after
3687 the @samp{-I-}, these directories are searched for all @samp{#include}
3688 directives. (Ordinarily @emph{all} @samp{-I} directories are used
3691 In addition, the @samp{-I-} option inhibits the use of the current
3692 directory (where the current input file came from) as the first search
3693 directory for @samp{#include "@var{file}"}. There is no way to
3694 override this effect of @samp{-I-}. With @samp{-I.} you can specify
3695 searching the directory which was current when the compiler was
3696 invoked. That is not exactly the same as what the preprocessor does
3697 by default, but it is often satisfactory.
3699 @samp{-I-} does not inhibit the use of the standard system directories
3700 for header files. Thus, @samp{-I-} and @samp{-nostdinc} are
3704 Add directory @var{dir} to the list of directories to be searched
3707 @item -B@var{prefix}
3708 This option specifies where to find the executables, libraries,
3709 include files, and data files of the compiler itself.
3711 The compiler driver program runs one or more of the subprograms
3712 @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries
3713 @var{prefix} as a prefix for each program it tries to run, both with and
3714 without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}).
3716 For each subprogram to be run, the compiler driver first tries the
3717 @samp{-B} prefix, if any. If that name is not found, or if @samp{-B}
3718 was not specified, the driver tries two standard prefixes, which are
3719 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc-lib/}. If neither of
3720 those results in a file name that is found, the unmodified program
3721 name is searched for using the directories specified in your
3722 @env{PATH} environment variable.
3724 @samp{-B} prefixes that effectively specify directory names also apply
3725 to libraries in the linker, because the compiler translates these
3726 options into @samp{-L} options for the linker. They also apply to
3727 includes files in the preprocessor, because the compiler translates these
3728 options into @samp{-isystem} options for the preprocessor. In this case,
3729 the compiler appends @samp{include} to the prefix.
3731 The run-time support file @file{libgcc.a} can also be searched for using
3732 the @samp{-B} prefix, if needed. If it is not found there, the two
3733 standard prefixes above are tried, and that is all. The file is left
3734 out of the link if it is not found by those means.
3736 Another way to specify a prefix much like the @samp{-B} prefix is to use
3737 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
3740 @item -specs=@var{file}
3741 Process @var{file} after the compiler reads in the standard @file{specs}
3742 file, in order to override the defaults that the @file{gcc} driver
3743 program uses when determining what switches to pass to @file{cc1},
3744 @file{cc1plus}, @file{as}, @file{ld}, etc. More than one
3745 @samp{-specs=}@var{file} can be specified on the command line, and they
3746 are processed in order, from left to right.
3752 @section Specifying subprocesses and the switches to pass to them
3754 @command{gcc} is a driver program. It performs its job by invoking a
3755 sequence of other programs to do the work of compiling, assembling and
3756 linking. GCC interprets its command-line parameters and uses these to
3757 deduce which programs it should invoke, and which command-line options
3758 it ought to place on their command lines. This behaviour is controlled
3759 by @dfn{spec strings}. In most cases there is one spec string for each
3760 program that GCC can invoke, but a few programs have multiple spec
3761 strings to control their behaviour. The spec strings built into GCC can
3762 be overridden by using the @samp{-specs=} command-line switch to specify
3765 @dfn{Spec files} are plaintext files that are used to construct spec
3766 strings. They consist of a sequence of directives separated by blank
3767 lines. The type of directive is determined by the first non-whitespace
3768 character on the line and it can be one of the following:
3771 @item %@var{command}
3772 Issues a @var{command} to the spec file processor. The commands that can
3776 @item %include <@var{file}>
3778 Search for @var{file} and insert its text at the current point in the
3781 @item %include_noerr <@var{file}>
3782 @cindex %include_noerr
3783 Just like @samp{%include}, but do not generate an error message if the include
3784 file cannot be found.
3786 @item %rename @var{old_name} @var{new_name}
3788 Rename the spec string @var{old_name} to @var{new_name}.
3792 @item *[@var{spec_name}]:
3793 This tells the compiler to create, override or delete the named spec
3794 string. All lines after this directive up to the next directive or
3795 blank line are considered to be the text for the spec string. If this
3796 results in an empty string then the spec will be deleted. (Or, if the
3797 spec did not exist, then nothing will happened.) Otherwise, if the spec
3798 does not currently exist a new spec will be created. If the spec does
3799 exist then its contents will be overridden by the text of this
3800 directive, unless the first character of that text is the @samp{+}
3801 character, in which case the text will be appended to the spec.
3803 @item [@var{suffix}]:
3804 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
3805 and up to the next directive or blank line are considered to make up the
3806 spec string for the indicated suffix. When the compiler encounters an
3807 input file with the named suffix, it will processes the spec string in
3808 order to work out how to compile that file. For example:
3815 This says that any input file whose name ends in @samp{.ZZ} should be
3816 passed to the program @samp{z-compile}, which should be invoked with the
3817 command-line switch @samp{-input} and with the result of performing the
3818 @samp{%i} substitution. (See below.)
3820 As an alternative to providing a spec string, the text that follows a
3821 suffix directive can be one of the following:
3824 @item @@@var{language}
3825 This says that the suffix is an alias for a known @var{language}. This is
3826 similar to using the @option{-x} command-line switch to GCC to specify a
3827 language explicitly. For example:
3834 Says that .ZZ files are, in fact, C++ source files.
3837 This causes an error messages saying:
3840 @var{name} compiler not installed on this system.
3844 GCC already has an extensive list of suffixes built into it.
3845 This directive will add an entry to the end of the list of suffixes, but
3846 since the list is searched from the end backwards, it is effectively
3847 possible to override earlier entries using this technique.
3851 GCC has the following spec strings built into it. Spec files can
3852 override these strings or create their own. Note that individual
3853 targets can also add their own spec strings to this list.
3856 asm Options to pass to the assembler
3857 asm_final Options to pass to the assembler post-processor
3858 cpp Options to pass to the C preprocessor
3859 cc1 Options to pass to the C compiler
3860 cc1plus Options to pass to the C++ compiler
3861 endfile Object files to include at the end of the link
3862 link Options to pass to the linker
3863 lib Libraries to include on the command line to the linker
3864 libgcc Decides which GCC support library to pass to the linker
3865 linker Sets the name of the linker
3866 predefines Defines to be passed to the C preprocessor
3867 signed_char Defines to pass to CPP to say whether @code{char} is signed by default
3868 startfile Object files to include at the start of the link
3871 Here is a small example of a spec file:
3877 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
3880 This example renames the spec called @samp{lib} to @samp{old_lib} and
3881 then overrides the previous definition of @samp{lib} with a new one.
3882 The new definition adds in some extra command-line options before
3883 including the text of the old definition.
3885 @dfn{Spec strings} are a list of command-line options to be passed to their
3886 corresponding program. In addition, the spec strings can contain
3887 @samp{%}-prefixed sequences to substitute variable text or to
3888 conditionally insert text into the command line. Using these constructs
3889 it is possible to generate quite complex command lines.
3891 Here is a table of all defined @samp{%}-sequences for spec
3892 strings. Note that spaces are not generated automatically around the
3893 results of expanding these sequences. Therefore you can concatenate them
3894 together or combine them with constant text in a single argument.
3898 Substitute one @samp{%} into the program name or argument.
3901 Substitute the name of the input file being processed.
3904 Substitute the basename of the input file being processed.
3905 This is the substring up to (and not including) the last period
3906 and not including the directory.
3909 Marks the argument containing or following the @samp{%d} as a
3910 temporary file name, so that that file will be deleted if GCC exits
3911 successfully. Unlike @samp{%g}, this contributes no text to the
3914 @item %g@var{suffix}
3915 Substitute a file name that has suffix @var{suffix} and is chosen
3916 once per compilation, and mark the argument in the same way as
3917 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
3918 name is now chosen in a way that is hard to predict even when previously
3919 chosen file names are known. For example, @samp{%g.s ... %g.o ... %g.s}
3920 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
3921 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
3922 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
3923 was simply substituted with a file name chosen once per compilation,
3924 without regard to any appended suffix (which was therefore treated
3925 just like ordinary text), making such attacks more likely to succeed.
3927 @item %u@var{suffix}
3928 Like @samp{%g}, but generates a new temporary file name even if
3929 @samp{%u@var{suffix}} was already seen.
3931 @item %U@var{suffix}
3932 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
3933 new one if there is no such last file name. In the absence of any
3934 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
3935 the same suffix @emph{space}, so @samp{%g.s ... %U.s ... %g.s ... %U.s}
3936 would involve the generation of two distinct file names, one
3937 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
3938 simply substituted with a file name chosen for the previous @samp{%u},
3939 without regard to any appended suffix.
3942 Marks the argument containing or following the @samp{%w} as the
3943 designated output file of this compilation. This puts the argument
3944 into the sequence of arguments that @samp{%o} will substitute later.
3947 Substitutes the names of all the output files, with spaces
3948 automatically placed around them. You should write spaces
3949 around the @samp{%o} as well or the results are undefined.
3950 @samp{%o} is for use in the specs for running the linker.
3951 Input files whose names have no recognized suffix are not compiled
3952 at all, but they are included among the output files, so they will
3956 Substitutes the suffix for object files. Note that this is
3957 handled specially when it immediately follows @samp{%g, %u, or %U},
3958 because of the need for those to form complete file names. The
3959 handling is such that @samp{%O} is treated exactly as if it had already
3960 been substituted, except that @samp{%g, %u, and %U} do not currently
3961 support additional @var{suffix} characters following @samp{%O} as they would
3962 following, for example, @samp{.o}.
3965 Substitutes the standard macro predefinitions for the
3966 current target machine. Use this when running @code{cpp}.
3969 Like @samp{%p}, but puts @samp{__} before and after the name of each
3970 predefined macro, except for macros that start with @samp{__} or with
3971 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
3975 Substitute a @samp{-iprefix} option made from GCC_EXEC_PREFIX.
3978 Current argument is the name of a library or startup file of some sort.
3979 Search for that file in a standard list of directories and substitute
3980 the full name found.
3983 Print @var{str} as an error message. @var{str} is terminated by a newline.
3984 Use this when inconsistent options are detected.
3987 Output @samp{-} if the input for the current command is coming from a pipe.
3990 Substitute the contents of spec string @var{name} at this point.
3993 Like @samp{%(...)} but put @samp{__} around @samp{-D} arguments.
3995 @item %x@{@var{option}@}
3996 Accumulate an option for @samp{%X}.
3999 Output the accumulated linker options specified by @samp{-Wl} or a @samp{%x}
4003 Output the accumulated assembler options specified by @samp{-Wa}.
4006 Output the accumulated preprocessor options specified by @samp{-Wp}.
4009 Substitute the major version number of GCC.
4010 (For version 2.9.5, this is 2.)
4013 Substitute the minor version number of GCC.
4014 (For version 2.9.5, this is 9.)
4017 Process the @code{asm} spec. This is used to compute the
4018 switches to be passed to the assembler.
4021 Process the @code{asm_final} spec. This is a spec string for
4022 passing switches to an assembler post-processor, if such a program is
4026 Process the @code{link} spec. This is the spec for computing the
4027 command line passed to the linker. Typically it will make use of the
4028 @samp{%L %G %S %D and %E} sequences.
4031 Dump out a @samp{-L} option for each directory that GCC believes might
4032 contain startup files. If the target supports multilibs then the
4033 current multilib directory will be prepended to each of these paths.
4036 Process the @code{lib} spec. This is a spec string for deciding which
4037 libraries should be included on the command line to the linker.
4040 Process the @code{libgcc} spec. This is a spec string for deciding
4041 which GCC support library should be included on the command line to the linker.
4044 Process the @code{startfile} spec. This is a spec for deciding which
4045 object files should be the first ones passed to the linker. Typically
4046 this might be a file named @file{crt0.o}.
4049 Process the @code{endfile} spec. This is a spec string that specifies
4050 the last object files that will be passed to the linker.
4053 Process the @code{cpp} spec. This is used to construct the arguments
4054 to be passed to the C preprocessor.
4057 Process the @code{signed_char} spec. This is intended to be used
4058 to tell cpp whether a char is signed. It typically has the definition:
4060 %@{funsigned-char:-D__CHAR_UNSIGNED__@}
4064 Process the @code{cc1} spec. This is used to construct the options to be
4065 passed to the actual C compiler (@samp{cc1}).
4068 Process the @code{cc1plus} spec. This is used to construct the options to be
4069 passed to the actual C++ compiler (@samp{cc1plus}).
4072 Substitute the variable part of a matched option. See below.
4073 Note that each comma in the substituted string is replaced by
4077 Substitutes the @code{-S} switch, if that switch was given to GCC.
4078 If that switch was not specified, this substitutes nothing. Note that
4079 the leading dash is omitted when specifying this option, and it is
4080 automatically inserted if the substitution is performed. Thus the spec
4081 string @samp{%@{foo@}} would match the command-line option @samp{-foo}
4082 and would output the command line option @samp{-foo}.
4084 @item %W@{@code{S}@}
4085 Like %@{@code{S}@} but mark last argument supplied within as a file to be
4088 @item %@{@code{S}*@}
4089 Substitutes all the switches specified to GCC whose names start
4090 with @code{-S}, but which also take an argument. This is used for
4091 switches like @samp{-o, -D, -I}, etc. GCC considers @samp{-o foo} as being
4092 one switch whose names starts with @samp{o}. %@{o*@} would substitute this
4093 text, including the space. Thus two arguments would be generated.
4095 @item %@{^@code{S}*@}
4096 Like %@{@code{S}*@}, but don't put a blank between a switch and its
4097 argument. Thus %@{^o*@} would only generate one argument, not two.
4099 @item %@{<@code{S}@}
4100 Remove all occurrences of @code{-S} from the command line. Note - this
4101 command is position dependent. @samp{%} commands in the spec string
4102 before this option will see @code{-S}, @samp{%} commands in the spec
4103 string after this option will not.
4105 @item %@{@code{S}*:@code{X}@}
4106 Substitutes @code{X} if one or more switches whose names start with
4107 @code{-S} are specified to GCC. Note that the tail part of the
4108 @code{-S} option (i.e. the part matched by the @samp{*}) will be substituted
4109 for each occurrence of @samp{%*} within @code{X}.
4111 @item %@{@code{S}:@code{X}@}
4112 Substitutes @code{X}, but only if the @samp{-S} switch was given to GCC.
4114 @item %@{!@code{S}:@code{X}@}
4115 Substitutes @code{X}, but only if the @samp{-S} switch was @emph{not} given to GCC.
4117 @item %@{|@code{S}:@code{X}@}
4118 Like %@{@code{S}:@code{X}@}, but if no @code{S} switch, substitute @samp{-}.
4120 @item %@{|!@code{S}:@code{X}@}
4121 Like %@{!@code{S}:@code{X}@}, but if there is an @code{S} switch, substitute @samp{-}.
4123 @item %@{.@code{S}:@code{X}@}
4124 Substitutes @code{X}, but only if processing a file with suffix @code{S}.
4126 @item %@{!.@code{S}:@code{X}@}
4127 Substitutes @code{X}, but only if @emph{not} processing a file with suffix @code{S}.
4129 @item %@{@code{S}|@code{P}:@code{X}@}
4130 Substitutes @code{X} if either @code{-S} or @code{-P} was given to GCC. This may be
4131 combined with @samp{!} and @samp{.} sequences as well, although they
4132 have a stronger binding than the @samp{|}. For example a spec string
4136 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
4139 will output the following command-line options from the following input
4140 command-line options:
4145 -d fred.c -foo -baz -boggle
4146 -d jim.d -bar -baz -boggle
4151 The conditional text @code{X} in a %@{@code{S}:@code{X}@} or
4152 %@{!@code{S}:@code{X}@} construct may contain other nested @samp{%} constructs
4153 or spaces, or even newlines. They are processed as usual, as described
4156 The @samp{-O, -f, -m, and -W} switches are handled specifically in these
4157 constructs. If another value of @samp{-O} or the negated form of a @samp{-f, -m, or
4158 -W} switch is found later in the command line, the earlier switch
4159 value is ignored, except with @{@code{S}*@} where @code{S} is just one
4160 letter, which passes all matching options.
4162 The character @samp{|} at the beginning of the predicate text is used to indicate
4163 that a command should be piped to the following command, but only if @samp{-pipe}
4166 It is built into GCC which switches take arguments and which do not.
4167 (You might think it would be useful to generalize this to allow each
4168 compiler's spec to say which switches take arguments. But this cannot
4169 be done in a consistent fashion. GCC cannot even decide which input
4170 files have been specified without knowing which switches take arguments,
4171 and it must know which input files to compile in order to tell which
4174 GCC also knows implicitly that arguments starting in @samp{-l} are to be
4175 treated as compiler output files, and passed to the linker in their
4176 proper position among the other output files.
4178 @c man begin OPTIONS
4180 @node Target Options
4181 @section Specifying Target Machine and Compiler Version
4182 @cindex target options
4183 @cindex cross compiling
4184 @cindex specifying machine version
4185 @cindex specifying compiler version and target machine
4186 @cindex compiler version, specifying
4187 @cindex target machine, specifying
4189 By default, GCC compiles code for the same type of machine that you
4190 are using. However, it can also be installed as a cross-compiler, to
4191 compile for some other type of machine. In fact, several different
4192 configurations of GCC, for different target machines, can be
4193 installed side by side. Then you specify which one to use with the
4196 In addition, older and newer versions of GCC can be installed side
4197 by side. One of them (probably the newest) will be the default, but
4198 you may sometimes wish to use another.
4201 @item -b @var{machine}
4202 The argument @var{machine} specifies the target machine for compilation.
4203 This is useful when you have installed GCC as a cross-compiler.
4205 The value to use for @var{machine} is the same as was specified as the
4206 machine type when configuring GCC as a cross-compiler. For
4207 example, if a cross-compiler was configured with @samp{configure
4208 i386v}, meaning to compile for an 80386 running System V, then you
4209 would specify @samp{-b i386v} to run that cross compiler.
4211 When you do not specify @samp{-b}, it normally means to compile for
4212 the same type of machine that you are using.
4214 @item -V @var{version}
4215 The argument @var{version} specifies which version of GCC to run.
4216 This is useful when multiple versions are installed. For example,
4217 @var{version} might be @samp{2.0}, meaning to run GCC version 2.0.
4219 The default version, when you do not specify @samp{-V}, is the last
4220 version of GCC that you installed.
4223 The @samp{-b} and @samp{-V} options actually work by controlling part of
4224 the file name used for the executable files and libraries used for
4225 compilation. A given version of GCC, for a given target machine, is
4226 normally kept in the directory @file{/usr/local/lib/gcc-lib/@var{machine}/@var{version}}.@refill
4228 Thus, sites can customize the effect of @samp{-b} or @samp{-V} either by
4229 changing the names of these directories or adding alternate names (or
4230 symbolic links). If in directory @file{/usr/local/lib/gcc-lib/} the
4231 file @file{80386} is a link to the file @file{i386v}, then @samp{-b
4232 80386} becomes an alias for @samp{-b i386v}.
4234 In one respect, the @samp{-b} or @samp{-V} do not completely change
4235 to a different compiler: the top-level driver program @command{gcc}
4236 that you originally invoked continues to run and invoke the other
4237 executables (preprocessor, compiler per se, assembler and linker)
4238 that do the real work. However, since no real work is done in the
4239 driver program, it usually does not matter that the driver program
4240 in use is not the one for the specified target. It is common for the
4241 interface to the other executables to change incompatibly between
4242 compiler versions, so unless the version specified is very close to that
4243 of the driver (for example, @samp{-V 3.0} with a driver program from GCC
4244 version 3.0.1), use of @samp{-V} may not work; for example, using
4245 @samp{-V 2.95.2} will not work with a driver program from GCC 3.0.
4247 The only way that the driver program depends on the target machine is
4248 in the parsing and handling of special machine-specific options.
4249 However, this is controlled by a file which is found, along with the
4250 other executables, in the directory for the specified version and
4251 target machine. As a result, a single installed driver program adapts
4252 to any specified target machine, and sufficiently similar compiler
4255 The driver program executable does control one significant thing,
4256 however: the default version and target machine. Therefore, you can
4257 install different instances of the driver program, compiled for
4258 different targets or versions, under different names.
4260 For example, if the driver for version 2.0 is installed as @command{ogcc}
4261 and that for version 2.1 is installed as @command{gcc}, then the command
4262 @command{gcc} will use version 2.1 by default, while @command{ogcc} will use
4263 2.0 by default. However, you can choose either version with either
4264 command with the @samp{-V} option.
4266 @node Submodel Options
4267 @section Hardware Models and Configurations
4268 @cindex submodel options
4269 @cindex specifying hardware config
4270 @cindex hardware models and configurations, specifying
4271 @cindex machine dependent options
4273 Earlier we discussed the standard option @samp{-b} which chooses among
4274 different installed compilers for completely different target
4275 machines, such as Vax vs. 68000 vs. 80386.
4277 In addition, each of these target machine types can have its own
4278 special options, starting with @samp{-m}, to choose among various
4279 hardware models or configurations---for example, 68010 vs 68020,
4280 floating coprocessor or none. A single installed version of the
4281 compiler can compile for any model or configuration, according to the
4284 Some configurations of the compiler also support additional special
4285 options, usually for compatibility with other compilers on the same
4289 These options are defined by the macro @code{TARGET_SWITCHES} in the
4290 machine description. The default for the options is also defined by
4291 that macro, which enables you to change the defaults.
4307 * RS/6000 and PowerPC Options::
4312 * Intel 960 Options::
4313 * DEC Alpha Options::
4317 * System V Options::
4318 * TMS320C3x/C4x Options::
4327 @node M680x0 Options
4328 @subsection M680x0 Options
4329 @cindex M680x0 options
4331 These are the @samp{-m} options defined for the 68000 series. The default
4332 values for these options depends on which style of 68000 was selected when
4333 the compiler was configured; the defaults for the most common choices are
4339 Generate output for a 68000. This is the default
4340 when the compiler is configured for 68000-based systems.
4342 Use this option for microcontrollers with a 68000 or EC000 core,
4343 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
4347 Generate output for a 68020. This is the default
4348 when the compiler is configured for 68020-based systems.
4351 Generate output containing 68881 instructions for floating point.
4352 This is the default for most 68020 systems unless @samp{-nfp} was
4353 specified when the compiler was configured.
4356 Generate output for a 68030. This is the default when the compiler is
4357 configured for 68030-based systems.
4360 Generate output for a 68040. This is the default when the compiler is
4361 configured for 68040-based systems.
4363 This option inhibits the use of 68881/68882 instructions that have to be
4364 emulated by software on the 68040. Use this option if your 68040 does not
4365 have code to emulate those instructions.
4368 Generate output for a 68060. This is the default when the compiler is
4369 configured for 68060-based systems.
4371 This option inhibits the use of 68020 and 68881/68882 instructions that
4372 have to be emulated by software on the 68060. Use this option if your 68060
4373 does not have code to emulate those instructions.
4376 Generate output for a CPU32. This is the default
4377 when the compiler is configured for CPU32-based systems.
4379 Use this option for microcontrollers with a
4380 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
4381 68336, 68340, 68341, 68349 and 68360.
4384 Generate output for a 520X "coldfire" family cpu. This is the default
4385 when the compiler is configured for 520X-based systems.
4387 Use this option for microcontroller with a 5200 core, including
4388 the MCF5202, MCF5203, MCF5204 and MCF5202.
4392 Generate output for a 68040, without using any of the new instructions.
4393 This results in code which can run relatively efficiently on either a
4394 68020/68881 or a 68030 or a 68040. The generated code does use the
4395 68881 instructions that are emulated on the 68040.
4398 Generate output for a 68060, without using any of the new instructions.
4399 This results in code which can run relatively efficiently on either a
4400 68020/68881 or a 68030 or a 68040. The generated code does use the
4401 68881 instructions that are emulated on the 68060.
4404 Generate output containing Sun FPA instructions for floating point.
4407 Generate output containing library calls for floating point.
4408 @strong{Warning:} the requisite libraries are not available for all m68k
4409 targets. Normally the facilities of the machine's usual C compiler are
4410 used, but this can't be done directly in cross-compilation. You must
4411 make your own arrangements to provide suitable library functions for
4412 cross-compilation. The embedded targets @samp{m68k-*-aout} and
4413 @samp{m68k-*-coff} do provide software floating point support.
4416 Consider type @code{int} to be 16 bits wide, like @code{short int}.
4419 Do not use the bit-field instructions. The @samp{-m68000}, @samp{-mcpu32}
4420 and @samp{-m5200} options imply @w{@samp{-mnobitfield}}.
4423 Do use the bit-field instructions. The @samp{-m68020} option implies
4424 @samp{-mbitfield}. This is the default if you use a configuration
4425 designed for a 68020.
4428 Use a different function-calling convention, in which functions
4429 that take a fixed number of arguments return with the @code{rtd}
4430 instruction, which pops their arguments while returning. This
4431 saves one instruction in the caller since there is no need to pop
4432 the arguments there.
4434 This calling convention is incompatible with the one normally
4435 used on Unix, so you cannot use it if you need to call libraries
4436 compiled with the Unix compiler.
4438 Also, you must provide function prototypes for all functions that
4439 take variable numbers of arguments (including @code{printf});
4440 otherwise incorrect code will be generated for calls to those
4443 In addition, seriously incorrect code will result if you call a
4444 function with too many arguments. (Normally, extra arguments are
4445 harmlessly ignored.)
4447 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
4448 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
4451 @itemx -mno-align-int
4452 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
4453 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
4454 boundary (@samp{-malign-int}) or a 16-bit boundary (@samp{-mno-align-int}).
4455 Aligning variables on 32-bit boundaries produces code that runs somewhat
4456 faster on processors with 32-bit busses at the expense of more memory.
4458 @strong{Warning:} if you use the @samp{-malign-int} switch, GCC will
4459 align structures containing the above types differently than
4460 most published application binary interface specifications for the m68k.
4463 Use the pc-relative addressing mode of the 68000 directly, instead of
4464 using a global offset table. At present, this option implies -fpic,
4465 allowing at most a 16-bit offset for pc-relative addressing. -fPIC is
4466 not presently supported with -mpcrel, though this could be supported for
4467 68020 and higher processors.
4469 @item -mno-strict-align
4470 @itemx -mstrict-align
4471 @kindex -mstrict-align
4472 Do not (do) assume that unaligned memory references will be handled by
4477 @node M68hc1x Options
4478 @subsection M68hc1x Options
4479 @cindex M68hc1x options
4481 These are the @samp{-m} options defined for the 68hc11 and 68hc12
4482 microcontrollers. The default values for these options depends on
4483 which style of microcontroller was selected when the compiler was configured;
4484 the defaults for the most common choices are given below.
4489 Generate output for a 68HC11. This is the default
4490 when the compiler is configured for 68HC11-based systems.
4494 Generate output for a 68HC12. This is the default
4495 when the compiler is configured for 68HC12-based systems.
4498 Enable the use of 68HC12 pre and post auto-increment and auto-decrement
4502 Consider type @code{int} to be 16 bits wide, like @code{short int}.
4504 @item -msoft-reg-count=@var{count}
4505 Specify the number of pseudo-soft registers which are used for the
4506 code generation. The maximum number is 32. Using more pseudo-soft
4507 register may or may not result in better code depending on the program.
4508 The default is 4 for 68HC11 and 2 for 68HC12.
4513 @subsection VAX Options
4516 These @samp{-m} options are defined for the Vax:
4520 Do not output certain jump instructions (@code{aobleq} and so on)
4521 that the Unix assembler for the Vax cannot handle across long
4525 Do output those jump instructions, on the assumption that you
4526 will assemble with the GNU assembler.
4529 Output code for g-format floating point numbers instead of d-format.
4533 @subsection SPARC Options
4534 @cindex SPARC options
4536 These @samp{-m} switches are supported on the SPARC:
4541 Specify @samp{-mapp-regs} to generate output using the global registers
4542 2 through 4, which the SPARC SVR4 ABI reserves for applications. This
4545 To be fully SVR4 ABI compliant at the cost of some performance loss,
4546 specify @samp{-mno-app-regs}. You should compile libraries and system
4547 software with this option.
4551 Generate output containing floating point instructions. This is the
4556 Generate output containing library calls for floating point.
4557 @strong{Warning:} the requisite libraries are not available for all SPARC
4558 targets. Normally the facilities of the machine's usual C compiler are
4559 used, but this cannot be done directly in cross-compilation. You must make
4560 your own arrangements to provide suitable library functions for
4561 cross-compilation. The embedded targets @samp{sparc-*-aout} and
4562 @samp{sparclite-*-*} do provide software floating point support.
4564 @samp{-msoft-float} changes the calling convention in the output file;
4565 therefore, it is only useful if you compile @emph{all} of a program with
4566 this option. In particular, you need to compile @file{libgcc.a}, the
4567 library that comes with GCC, with @samp{-msoft-float} in order for
4570 @item -mhard-quad-float
4571 Generate output containing quad-word (long double) floating point
4574 @item -msoft-quad-float
4575 Generate output containing library calls for quad-word (long double)
4576 floating point instructions. The functions called are those specified
4577 in the SPARC ABI. This is the default.
4579 As of this writing, there are no sparc implementations that have hardware
4580 support for the quad-word floating point instructions. They all invoke
4581 a trap handler for one of these instructions, and then the trap handler
4582 emulates the effect of the instruction. Because of the trap handler overhead,
4583 this is much slower than calling the ABI library routines. Thus the
4584 @samp{-msoft-quad-float} option is the default.
4588 With @samp{-mepilogue} (the default), the compiler always emits code for
4589 function exit at the end of each function. Any function exit in
4590 the middle of the function (such as a return statement in C) will
4591 generate a jump to the exit code at the end of the function.
4593 With @samp{-mno-epilogue}, the compiler tries to emit exit code inline
4594 at every function exit.
4598 With @samp{-mflat}, the compiler does not generate save/restore instructions
4599 and will use a "flat" or single register window calling convention.
4600 This model uses %i7 as the frame pointer and is compatible with the normal
4601 register window model. Code from either may be intermixed.
4602 The local registers and the input registers (0-5) are still treated as
4603 "call saved" registers and will be saved on the stack as necessary.
4605 With @samp{-mno-flat} (the default), the compiler emits save/restore
4606 instructions (except for leaf functions) and is the normal mode of operation.
4608 @item -mno-unaligned-doubles
4609 @itemx -munaligned-doubles
4610 Assume that doubles have 8 byte alignment. This is the default.
4612 With @samp{-munaligned-doubles}, GCC assumes that doubles have 8 byte
4613 alignment only if they are contained in another type, or if they have an
4614 absolute address. Otherwise, it assumes they have 4 byte alignment.
4615 Specifying this option avoids some rare compatibility problems with code
4616 generated by other compilers. It is not the default because it results
4617 in a performance loss, especially for floating point code.
4619 @item -mno-faster-structs
4620 @itemx -mfaster-structs
4621 With @samp{-mfaster-structs}, the compiler assumes that structures
4622 should have 8 byte alignment. This enables the use of pairs of
4623 @code{ldd} and @code{std} instructions for copies in structure
4624 assignment, in place of twice as many @code{ld} and @code{st} pairs.
4625 However, the use of this changed alignment directly violates the Sparc
4626 ABI. Thus, it's intended only for use on targets where the developer
4627 acknowledges that their resulting code will not be directly in line with
4628 the rules of the ABI.
4632 These two options select variations on the SPARC architecture.
4634 By default (unless specifically configured for the Fujitsu SPARClite),
4635 GCC generates code for the v7 variant of the SPARC architecture.
4637 @samp{-mv8} will give you SPARC v8 code. The only difference from v7
4638 code is that the compiler emits the integer multiply and integer
4639 divide instructions which exist in SPARC v8 but not in SPARC v7.
4641 @samp{-msparclite} will give you SPARClite code. This adds the integer
4642 multiply, integer divide step and scan (@code{ffs}) instructions which
4643 exist in SPARClite but not in SPARC v7.
4645 These options are deprecated and will be deleted in a future GCC release.
4646 They have been replaced with @samp{-mcpu=xxx}.
4650 These two options select the processor for which the code is optimised.
4652 With @samp{-mcypress} (the default), the compiler optimizes code for the
4653 Cypress CY7C602 chip, as used in the SparcStation/SparcServer 3xx series.
4654 This is also appropriate for the older SparcStation 1, 2, IPX etc.
4656 With @samp{-msupersparc} the compiler optimizes code for the SuperSparc cpu, as
4657 used in the SparcStation 10, 1000 and 2000 series. This flag also enables use
4658 of the full SPARC v8 instruction set.
4660 These options are deprecated and will be deleted in a future GCC release.
4661 They have been replaced with @samp{-mcpu=xxx}.
4663 @item -mcpu=@var{cpu_type}
4664 Set the instruction set, register set, and instruction scheduling parameters
4665 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
4666 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite},
4667 @samp{hypersparc}, @samp{sparclite86x}, @samp{f930}, @samp{f934},
4668 @samp{sparclet}, @samp{tsc701}, @samp{v9}, and @samp{ultrasparc}.
4670 Default instruction scheduling parameters are used for values that select
4671 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
4672 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
4674 Here is a list of each supported architecture and their supported
4679 v8: supersparc, hypersparc
4680 sparclite: f930, f934, sparclite86x
4685 @item -mtune=@var{cpu_type}
4686 Set the instruction scheduling parameters for machine type
4687 @var{cpu_type}, but do not set the instruction set or register set that the
4688 option @samp{-mcpu=}@var{cpu_type} would.
4690 The same values for @samp{-mcpu=}@var{cpu_type} are used for
4691 @samp{-mtune=}@*@var{cpu_type}, though the only useful values are those that
4692 select a particular cpu implementation: @samp{cypress}, @samp{supersparc},
4693 @samp{hypersparc}, @samp{f930}, @samp{f934}, @samp{sparclite86x},
4694 @samp{tsc701}, @samp{ultrasparc}.
4698 These @samp{-m} switches are supported in addition to the above
4699 on the SPARCLET processor.
4702 @item -mlittle-endian
4703 Generate code for a processor running in little-endian mode.
4706 Treat register @code{%g0} as a normal register.
4707 GCC will continue to clobber it as necessary but will not assume
4708 it always reads as 0.
4710 @item -mbroken-saverestore
4711 Generate code that does not use non-trivial forms of the @code{save} and
4712 @code{restore} instructions. Early versions of the SPARCLET processor do
4713 not correctly handle @code{save} and @code{restore} instructions used with
4714 arguments. They correctly handle them used without arguments. A @code{save}
4715 instruction used without arguments increments the current window pointer
4716 but does not allocate a new stack frame. It is assumed that the window
4717 overflow trap handler will properly handle this case as will interrupt
4721 These @samp{-m} switches are supported in addition to the above
4722 on SPARC V9 processors in 64 bit environments.
4725 @item -mlittle-endian
4726 Generate code for a processor running in little-endian mode.
4730 Generate code for a 32 bit or 64 bit environment.
4731 The 32 bit environment sets int, long and pointer to 32 bits.
4732 The 64 bit environment sets int to 32 bits and long and pointer
4735 @item -mcmodel=medlow
4736 Generate code for the Medium/Low code model: the program must be linked
4737 in the low 32 bits of the address space. Pointers are 64 bits.
4738 Programs can be statically or dynamically linked.
4740 @item -mcmodel=medmid
4741 Generate code for the Medium/Middle code model: the program must be linked
4742 in the low 44 bits of the address space, the text segment must be less than
4743 2G bytes, and data segment must be within 2G of the text segment.
4744 Pointers are 64 bits.
4746 @item -mcmodel=medany
4747 Generate code for the Medium/Anywhere code model: the program may be linked
4748 anywhere in the address space, the text segment must be less than
4749 2G bytes, and data segment must be within 2G of the text segment.
4750 Pointers are 64 bits.
4752 @item -mcmodel=embmedany
4753 Generate code for the Medium/Anywhere code model for embedded systems:
4754 assume a 32 bit text and a 32 bit data segment, both starting anywhere
4755 (determined at link time). Register %g4 points to the base of the
4756 data segment. Pointers still 64 bits.
4757 Programs are statically linked, PIC is not supported.
4760 @itemx -mno-stack-bias
4761 With @samp{-mstack-bias}, GCC assumes that the stack pointer, and
4762 frame pointer if present, are offset by -2047 which must be added back
4763 when making stack frame references.
4764 Otherwise, assume no such offset is present.
4767 @node Convex Options
4768 @subsection Convex Options
4769 @cindex Convex options
4771 These @samp{-m} options are defined for Convex:
4775 Generate output for C1. The code will run on any Convex machine.
4776 The preprocessor symbol @code{__convex__c1__} is defined.
4779 Generate output for C2. Uses instructions not available on C1.
4780 Scheduling and other optimizations are chosen for max performance on C2.
4781 The preprocessor symbol @code{__convex_c2__} is defined.
4784 Generate output for C32xx. Uses instructions not available on C1.
4785 Scheduling and other optimizations are chosen for max performance on C32.
4786 The preprocessor symbol @code{__convex_c32__} is defined.
4789 Generate output for C34xx. Uses instructions not available on C1.
4790 Scheduling and other optimizations are chosen for max performance on C34.
4791 The preprocessor symbol @code{__convex_c34__} is defined.
4794 Generate output for C38xx. Uses instructions not available on C1.
4795 Scheduling and other optimizations are chosen for max performance on C38.
4796 The preprocessor symbol @code{__convex_c38__} is defined.
4799 Generate code which puts an argument count in the word preceding each
4800 argument list. This is compatible with regular CC, and a few programs
4801 may need the argument count word. GDB and other source-level debuggers
4802 do not need it; this info is in the symbol table.
4805 Omit the argument count word. This is the default.
4807 @item -mvolatile-cache
4808 Allow volatile references to be cached. This is the default.
4810 @item -mvolatile-nocache
4811 Volatile references bypass the data cache, going all the way to memory.
4812 This is only needed for multi-processor code that does not use standard
4813 synchronization instructions. Making non-volatile references to volatile
4814 locations will not necessarily work.
4817 Type long is 32 bits, the same as type int. This is the default.
4820 Type long is 64 bits, the same as type long long. This option is useless,
4821 because no library support exists for it.
4824 @node AMD29K Options
4825 @subsection AMD29K Options
4826 @cindex AMD29K options
4828 These @samp{-m} options are defined for the AMD Am29000:
4833 @cindex DW bit (29k)
4834 Generate code that assumes the @code{DW} bit is set, i.e., that byte and
4835 halfword operations are directly supported by the hardware. This is the
4840 Generate code that assumes the @code{DW} bit is not set.
4844 @cindex byte writes (29k)
4845 Generate code that assumes the system supports byte and halfword write
4846 operations. This is the default.
4850 Generate code that assumes the systems does not support byte and
4851 halfword write operations. @samp{-mnbw} implies @samp{-mndw}.
4855 @cindex memory model (29k)
4856 Use a small memory model that assumes that all function addresses are
4857 either within a single 256 KB segment or at an absolute address of less
4858 than 256k. This allows the @code{call} instruction to be used instead
4859 of a @code{const}, @code{consth}, @code{calli} sequence.
4863 Use the normal memory model: Generate @code{call} instructions only when
4864 calling functions in the same file and @code{calli} instructions
4865 otherwise. This works if each file occupies less than 256 KB but allows
4866 the entire executable to be larger than 256 KB. This is the default.
4869 Always use @code{calli} instructions. Specify this option if you expect
4870 a single file to compile into more than 256 KB of code.
4874 @cindex processor selection (29k)
4875 Generate code for the Am29050.
4879 Generate code for the Am29000. This is the default.
4881 @item -mkernel-registers
4882 @kindex -mkernel-registers
4883 @cindex kernel and user registers (29k)
4884 Generate references to registers @code{gr64-gr95} instead of to
4885 registers @code{gr96-gr127}. This option can be used when compiling
4886 kernel code that wants a set of global registers disjoint from that used
4889 Note that when this option is used, register names in @samp{-f} flags
4890 must use the normal, user-mode, names.
4892 @item -muser-registers
4893 @kindex -muser-registers
4894 Use the normal set of global registers, @code{gr96-gr127}. This is the
4898 @itemx -mno-stack-check
4899 @kindex -mstack-check
4900 @cindex stack checks (29k)
4901 Insert (or do not insert) a call to @code{__msp_check} after each stack
4902 adjustment. This is often used for kernel code.
4905 @itemx -mno-storem-bug
4906 @kindex -mstorem-bug
4907 @cindex storem bug (29k)
4908 @samp{-mstorem-bug} handles 29k processors which cannot handle the
4909 separation of a mtsrim insn and a storem instruction (most 29000 chips
4910 to date, but not the 29050).
4912 @item -mno-reuse-arg-regs
4913 @itemx -mreuse-arg-regs
4914 @kindex -mreuse-arg-regs
4915 @samp{-mno-reuse-arg-regs} tells the compiler to only use incoming argument
4916 registers for copying out arguments. This helps detect calling a function
4917 with fewer arguments than it was declared with.
4919 @item -mno-impure-text
4920 @itemx -mimpure-text
4921 @kindex -mimpure-text
4922 @samp{-mimpure-text}, used in addition to @samp{-shared}, tells the compiler to
4923 not pass @samp{-assert pure-text} to the linker when linking a shared object.
4926 @kindex -msoft-float
4927 Generate output containing library calls for floating point.
4928 @strong{Warning:} the requisite libraries are not part of GCC.
4929 Normally the facilities of the machine's usual C compiler are used, but
4930 this can't be done directly in cross-compilation. You must make your
4931 own arrangements to provide suitable library functions for
4936 Do not generate multm or multmu instructions. This is useful for some embedded
4937 systems which do not have trap handlers for these instructions.
4941 @subsection ARM Options
4944 These @samp{-m} options are defined for Advanced RISC Machines (ARM)
4949 @kindex -mapcs-frame
4950 Generate a stack frame that is compliant with the ARM Procedure Call
4951 Standard for all functions, even if this is not strictly necessary for
4952 correct execution of the code. Specifying @samp{-fomit-frame-pointer}
4953 with this option will cause the stack frames not to be generated for
4954 leaf functions. The default is @samp{-mno-apcs-frame}.
4958 This is a synonym for @samp{-mapcs-frame}.
4962 Generate code for a processor running with a 26-bit program counter,
4963 and conforming to the function calling standards for the APCS 26-bit
4964 option. This option replaces the @samp{-m2} and @samp{-m3} options
4965 of previous releases of the compiler.
4969 Generate code for a processor running with a 32-bit program counter,
4970 and conforming to the function calling standards for the APCS 32-bit
4971 option. This option replaces the @samp{-m6} option of previous releases
4974 @item -mapcs-stack-check
4975 @kindex -mapcs-stack-check
4976 @kindex -mno-apcs-stack-check
4977 Generate code to check the amount of stack space available upon entry to
4978 every function (that actually uses some stack space). If there is
4979 insufficient space available then either the function
4980 @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be
4981 called, depending upon the amount of stack space required. The run time
4982 system is required to provide these functions. The default is
4983 @samp{-mno-apcs-stack-check}, since this produces smaller code.
4986 @kindex -mapcs-float
4987 @kindex -mno-apcs-float
4988 Pass floating point arguments using the float point registers. This is
4989 one of the variants of the APCS. This option is recommended if the
4990 target hardware has a floating point unit or if a lot of floating point
4991 arithmetic is going to be performed by the code. The default is
4992 @samp{-mno-apcs-float}, since integer only code is slightly increased in
4993 size if @samp{-mapcs-float} is used.
4995 @item -mapcs-reentrant
4996 @kindex -mapcs-reentrant
4997 @kindex -mno-apcs-reentrant
4998 Generate reentrant, position independent code. This is the equivalent
4999 to specifying the @samp{-fpic} option. The default is
5000 @samp{-mno-apcs-reentrant}.
5002 @item -mthumb-interwork
5003 @kindex -mthumb-interwork
5004 @kindex -mno-thumb-interwork
5005 Generate code which supports calling between the ARM and THUMB
5006 instruction sets. Without this option the two instruction sets cannot
5007 be reliably used inside one program. The default is
5008 @samp{-mno-thumb-interwork}, since slightly larger code is generated
5009 when @samp{-mthumb-interwork} is specified.
5011 @item -mno-sched-prolog
5012 @kindex -mno-sched-prolog
5013 @kindex -msched-prolog
5014 Prevent the reordering of instructions in the function prolog, or the
5015 merging of those instruction with the instructions in the function's
5016 body. This means that all functions will start with a recognizable set
5017 of instructions (or in fact one of a choice from a small set of
5018 different function prologues), and this information can be used to
5019 locate the start if functions inside an executable piece of code. The
5020 default is @samp{-msched-prolog}.
5023 Generate output containing floating point instructions. This is the
5027 Generate output containing library calls for floating point.
5028 @strong{Warning:} the requisite libraries are not available for all ARM
5029 targets. Normally the facilities of the machine's usual C compiler are
5030 used, but this cannot be done directly in cross-compilation. You must make
5031 your own arrangements to provide suitable library functions for
5034 @samp{-msoft-float} changes the calling convention in the output file;
5035 therefore, it is only useful if you compile @emph{all} of a program with
5036 this option. In particular, you need to compile @file{libgcc.a}, the
5037 library that comes with GCC, with @samp{-msoft-float} in order for
5040 @item -mlittle-endian
5041 Generate code for a processor running in little-endian mode. This is
5042 the default for all standard configurations.
5045 Generate code for a processor running in big-endian mode; the default is
5046 to compile code for a little-endian processor.
5048 @item -mwords-little-endian
5049 This option only applies when generating code for big-endian processors.
5050 Generate code for a little-endian word order but a big-endian byte
5051 order. That is, a byte order of the form @samp{32107654}. Note: this
5052 option should only be used if you require compatibility with code for
5053 big-endian ARM processors generated by versions of the compiler prior to
5056 @item -malignment-traps
5057 @kindex -malignment-traps
5058 Generate code that will not trap if the MMU has alignment traps enabled.
5059 On ARM architectures prior to ARMv4, there were no instructions to
5060 access half-word objects stored in memory. However, when reading from
5061 memory a feature of the ARM architecture allows a word load to be used,
5062 even if the address is unaligned, and the processor core will rotate the
5063 data as it is being loaded. This option tells the compiler that such
5064 misaligned accesses will cause a MMU trap and that it should instead
5065 synthesise the access as a series of byte accesses. The compiler can
5066 still use word accesses to load half-word data if it knows that the
5067 address is aligned to a word boundary.
5069 This option is ignored when compiling for ARM architecture 4 or later,
5070 since these processors have instructions to directly access half-word
5073 @item -mno-alignment-traps
5074 @kindex -mno-alignment-traps
5075 Generate code that assumes that the MMU will not trap unaligned
5076 accesses. This produces better code when the target instruction set
5077 does not have half-word memory operations (implementations prior to
5080 Note that you cannot use this option to access unaligned word objects,
5081 since the processor will only fetch one 32-bit aligned object from
5084 The default setting for most targets is -mno-alignment-traps, since
5085 this produces better code when there are no half-word memory
5086 instructions available.
5088 @item -mshort-load-bytes
5089 @kindex -mshort-load-bytes
5090 This is a deprecated alias for @samp{-malignment-traps}.
5092 @item -mno-short-load-bytes
5093 @kindex -mno-short-load-bytes
5094 This is a deprecated alias for @samp{-mno-alignment-traps}.
5096 @item -mshort-load-words
5097 @kindex -mshort-load-words
5098 This is a deprecated alias for @samp{-mno-alignment-traps}.
5100 @item -mno-short-load-words
5101 @kindex -mno-short-load-words
5102 This is a deprecated alias for @samp{-malignment-traps}.
5106 This option only applies to RISC iX. Emulate the native BSD-mode
5107 compiler. This is the default if @samp{-ansi} is not specified.
5111 This option only applies to RISC iX. Emulate the native X/Open-mode
5114 @item -mno-symrename
5115 @kindex -mno-symrename
5116 This option only applies to RISC iX. Do not run the assembler
5117 post-processor, @samp{symrename}, after code has been assembled.
5118 Normally it is necessary to modify some of the standard symbols in
5119 preparation for linking with the RISC iX C library; this option
5120 suppresses this pass. The post-processor is never run when the
5121 compiler is built for cross-compilation.
5125 This specifies the name of the target ARM processor. GCC uses this name
5126 to determine what kind of instructions it can use when generating
5127 assembly code. Permissible names are: arm2, arm250, arm3, arm6, arm60,
5128 arm600, arm610, arm620, arm7, arm7m, arm7d, arm7dm, arm7di, arm7dmi,
5129 arm70, arm700, arm700i, arm710, arm710c, arm7100, arm7500, arm7500fe,
5130 arm7tdmi, arm8, strongarm, strongarm110, strongarm1100, arm8, arm810,
5131 arm9, arm920, arm920t, arm9tdmi.
5133 @itemx -mtune=<name>
5135 This option is very similar to the @samp{-mcpu=} option, except that
5136 instead of specifying the actual target processor type, and hence
5137 restricting which instructions can be used, it specifies that GCC should
5138 tune the performance of the code as if the target were of the type
5139 specified in this option, but still choosing the instructions that it
5140 will generate based on the cpu specified by a @samp{-mcpu=} option.
5141 For some arm implementations better performance can be obtained by using
5146 This specifies the name of the target ARM architecture. GCC uses this
5147 name to determine what kind of instructions it can use when generating
5148 assembly code. This option can be used in conjunction with or instead
5149 of the @samp{-mcpu=} option. Permissible names are: armv2, armv2a,
5150 armv3, armv3m, armv4, armv4t, armv5.
5152 @item -mfpe=<number>
5153 @itemx -mfp=<number>
5156 This specifies the version of the floating point emulation available on
5157 the target. Permissible values are 2 and 3. @samp{-mfp=} is a synonym
5158 for @samp{-mfpe=} to support older versions of GCC.
5160 @item -mstructure-size-boundary=<n>
5161 @kindex -mstructure-size-boundary
5162 The size of all structures and unions will be rounded up to a multiple
5163 of the number of bits set by this option. Permissible values are 8 and
5164 32. The default value varies for different toolchains. For the COFF
5165 targeted toolchain the default value is 8. Specifying the larger number
5166 can produce faster, more efficient code, but can also increase the size
5167 of the program. The two values are potentially incompatible. Code
5168 compiled with one value cannot necessarily expect to work with code or
5169 libraries compiled with the other value, if they exchange information
5170 using structures or unions. Programmers are encouraged to use the 32
5171 value as future versions of the toolchain may default to this value.
5173 @item -mabort-on-noreturn
5174 @kindex -mabort-on-noreturn
5175 @kindex -mnoabort-on-noreturn
5176 Generate a call to the function abort at the end of a noreturn function.
5177 It will be executed if the function tries to return.
5180 @itemx -mno-long-calls
5181 Tells the compiler to perform function calls by first loading the
5182 address of the function into a register and then performing a subroutine
5183 call on this register. This switch is needed if the target function
5184 will lie outside of the 64 megabyte addressing range of the offset based
5185 version of subroutine call instruction.
5187 Even if this switch is enabled, not all function calls will be turned
5188 into long calls. The heuristic is that static functions, functions
5189 which have the @samp{short-call} attribute, functions that are inside
5190 the scope of a @samp{#pragma no_long_calls} directive and functions whose
5191 definitions have already been compiled within the current compilation
5192 unit, will not be turned into long calls. The exception to this rule is
5193 that weak function definitions, functions with the @samp{long-call}
5194 attribute or the @samp{section} attribute, and functions that are within
5195 the scope of a @samp{#pragma long_calls} directive, will always be
5196 turned into long calls.
5198 This feature is not enabled by default. Specifying
5199 @samp{--no-long-calls} will restore the default behaviour, as will
5200 placing the function calls within the scope of a @samp{#pragma
5201 long_calls_off} directive. Note these switches have no effect on how
5202 the compiler generates code to handle function calls via function
5205 @item -mnop-fun-dllimport
5206 @kindex -mnop-fun-dllimport
5207 Disable the support for the @emph{dllimport} attribute.
5209 @item -msingle-pic-base
5210 @kindex -msingle-pic-base
5211 Treat the register used for PIC addressing as read-only, rather than
5212 loading it in the prologue for each function. The run-time system is
5213 responsible for initialising this register with an appropriate value
5214 before execution begins.
5216 @item -mpic-register=<reg>
5217 @kindex -mpic-register=
5218 Specify the register to be used for PIC addressing. The default is R10
5219 unless stack-checking is enabled, when R9 is used.
5224 @subsection Thumb Options
5225 @cindex Thumb Options
5229 @item -mthumb-interwork
5230 @kindex -mthumb-interwork
5231 @kindex -mno-thumb-interwork
5232 Generate code which supports calling between the THUMB and ARM
5233 instruction sets. Without this option the two instruction sets cannot
5234 be reliably used inside one program. The default is
5235 @samp{-mno-thumb-interwork}, since slightly smaller code is generated
5239 @kindex -mtpcs-frame
5240 @kindex -mno-tpcs-frame
5241 Generate a stack frame that is compliant with the Thumb Procedure Call
5242 Standard for all non-leaf functions. (A leaf function is one that does
5243 not call any other functions). The default is @samp{-mno-apcs-frame}.
5245 @item -mtpcs-leaf-frame
5246 @kindex -mtpcs-leaf-frame
5247 @kindex -mno-tpcs-leaf-frame
5248 Generate a stack frame that is compliant with the Thumb Procedure Call
5249 Standard for all leaf functions. (A leaf function is one that does
5250 not call any other functions). The default is @samp{-mno-apcs-leaf-frame}.
5252 @item -mlittle-endian
5253 @kindex -mlittle-endian
5254 Generate code for a processor running in little-endian mode. This is
5255 the default for all standard configurations.
5258 @kindex -mbig-endian
5259 Generate code for a processor running in big-endian mode.
5261 @item -mstructure-size-boundary=<n>
5262 @kindex -mstructure-size-boundary
5263 The size of all structures and unions will be rounded up to a multiple
5264 of the number of bits set by this option. Permissible values are 8 and
5265 32. The default value varies for different toolchains. For the COFF
5266 targeted toolchain the default value is 8. Specifying the larger number
5267 can produced faster, more efficient code, but can also increase the size
5268 of the program. The two values are potentially incompatible. Code
5269 compiled with one value cannot necessarily expect to work with code or
5270 libraries compiled with the other value, if they exchange information
5271 using structures or unions. Programmers are encouraged to use the 32
5272 value as future versions of the toolchain may default to this value.
5274 @item -mnop-fun-dllimport
5275 @kindex -mnop-fun-dllimport
5276 Disable the support for the @emph{dllimport} attribute.
5278 @item -mcallee-super-interworking
5279 @kindex -mcallee-super-interworking
5280 Gives all externally visible functions in the file being compiled an ARM
5281 instruction set header which switches to Thumb mode before executing the
5282 rest of the function. This allows these functions to be called from
5283 non-interworking code.
5285 @item -mcaller-super-interworking
5286 @kindex -mcaller-super-interworking
5287 Allows calls via function pointers (including virtual functions) to
5288 execute correctly regardless of whether the target code has been
5289 compiled for interworking or not. There is a small overhead in the cost
5290 of executing a function pointer if this option is enabled.
5292 @item -msingle-pic-base
5293 @kindex -msingle-pic-base
5294 Treat the register used for PIC addressing as read-only, rather than
5295 loading it in the prologue for each function. The run-time system is
5296 responsible for initialising this register with an appropriate value
5297 before execution begins.
5299 @item -mpic-register=<reg>
5300 @kindex -mpic-register=
5301 Specify the register to be used for PIC addressing. The default is R10.
5305 @node MN10200 Options
5306 @subsection MN10200 Options
5307 @cindex MN10200 options
5308 These @samp{-m} options are defined for Matsushita MN10200 architectures:
5312 Indicate to the linker that it should perform a relaxation optimization pass
5313 to shorten branches, calls and absolute memory addresses. This option only
5314 has an effect when used on the command line for the final link step.
5316 This option makes symbolic debugging impossible.
5319 @node MN10300 Options
5320 @subsection MN10300 Options
5321 @cindex MN10300 options
5322 These @samp{-m} options are defined for Matsushita MN10300 architectures:
5326 Generate code to avoid bugs in the multiply instructions for the MN10300
5327 processors. This is the default.
5330 Do not generate code to avoid bugs in the multiply instructions for the
5334 Generate code which uses features specific to the AM33 processor.
5337 Do not generate code which uses features specific to the AM33 processor. This
5341 Indicate to the linker that it should perform a relaxation optimization pass
5342 to shorten branches, calls and absolute memory addresses. This option only
5343 has an effect when used on the command line for the final link step.
5345 This option makes symbolic debugging impossible.
5349 @node M32R/D Options
5350 @subsection M32R/D Options
5351 @cindex M32R/D options
5353 These @samp{-m} options are defined for Mitsubishi M32R/D architectures:
5356 @item -mcode-model=small
5357 Assume all objects live in the lower 16MB of memory (so that their addresses
5358 can be loaded with the @code{ld24} instruction), and assume all subroutines
5359 are reachable with the @code{bl} instruction.
5360 This is the default.
5362 The addressability of a particular object can be set with the
5363 @code{model} attribute.
5365 @item -mcode-model=medium
5366 Assume objects may be anywhere in the 32 bit address space (the compiler
5367 will generate @code{seth/add3} instructions to load their addresses), and
5368 assume all subroutines are reachable with the @code{bl} instruction.
5370 @item -mcode-model=large
5371 Assume objects may be anywhere in the 32 bit address space (the compiler
5372 will generate @code{seth/add3} instructions to load their addresses), and
5373 assume subroutines may not be reachable with the @code{bl} instruction
5374 (the compiler will generate the much slower @code{seth/add3/jl}
5375 instruction sequence).
5378 Disable use of the small data area. Variables will be put into
5379 one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the
5380 @code{section} attribute has been specified).
5381 This is the default.
5383 The small data area consists of sections @samp{.sdata} and @samp{.sbss}.
5384 Objects may be explicitly put in the small data area with the
5385 @code{section} attribute using one of these sections.
5388 Put small global and static data in the small data area, but do not
5389 generate special code to reference them.
5392 Put small global and static data in the small data area, and generate
5393 special instructions to reference them.
5396 @cindex smaller data references
5397 Put global and static objects less than or equal to @var{num} bytes
5398 into the small data or bss sections instead of the normal data or bss
5399 sections. The default value of @var{num} is 8.
5400 The @samp{-msdata} option must be set to one of @samp{sdata} or @samp{use}
5401 for this option to have any effect.
5403 All modules should be compiled with the same @samp{-G @var{num}} value.
5404 Compiling with different values of @var{num} may or may not work; if it
5405 doesn't the linker will give an error message - incorrect code will not be
5411 @subsection M88K Options
5412 @cindex M88k options
5414 These @samp{-m} options are defined for Motorola 88k architectures:
5419 Generate code that works well on both the m88100 and the
5424 Generate code that works best for the m88100, but that also
5429 Generate code that works best for the m88110, and may not run
5434 Obsolete option to be removed from the next revision.
5437 @item -midentify-revision
5438 @kindex -midentify-revision
5440 @cindex identifying source, compiler (88k)
5441 Include an @code{ident} directive in the assembler output recording the
5442 source file name, compiler name and version, timestamp, and compilation
5445 @item -mno-underscores
5446 @kindex -mno-underscores
5447 @cindex underscores, avoiding (88k)
5448 In assembler output, emit symbol names without adding an underscore
5449 character at the beginning of each name. The default is to use an
5450 underscore as prefix on each name.
5452 @item -mocs-debug-info
5453 @itemx -mno-ocs-debug-info
5454 @kindex -mocs-debug-info
5455 @kindex -mno-ocs-debug-info
5457 @cindex debugging, 88k OCS
5458 Include (or omit) additional debugging information (about registers used
5459 in each stack frame) as specified in the 88open Object Compatibility
5460 Standard, ``OCS''. This extra information allows debugging of code that
5461 has had the frame pointer eliminated. The default for DG/UX, SVr4, and
5462 Delta 88 SVr3.2 is to include this information; other 88k configurations
5463 omit this information by default.
5465 @item -mocs-frame-position
5466 @kindex -mocs-frame-position
5467 @cindex register positions in frame (88k)
5468 When emitting COFF debugging information for automatic variables and
5469 parameters stored on the stack, use the offset from the canonical frame
5470 address, which is the stack pointer (register 31) on entry to the
5471 function. The DG/UX, SVr4, Delta88 SVr3.2, and BCS configurations use
5472 @samp{-mocs-frame-position}; other 88k configurations have the default
5473 @samp{-mno-ocs-frame-position}.
5475 @item -mno-ocs-frame-position
5476 @kindex -mno-ocs-frame-position
5477 @cindex register positions in frame (88k)
5478 When emitting COFF debugging information for automatic variables and
5479 parameters stored on the stack, use the offset from the frame pointer
5480 register (register 30). When this option is in effect, the frame
5481 pointer is not eliminated when debugging information is selected by the
5484 @item -moptimize-arg-area
5485 @itemx -mno-optimize-arg-area
5486 @kindex -moptimize-arg-area
5487 @kindex -mno-optimize-arg-area
5488 @cindex arguments in frame (88k)
5489 Control how function arguments are stored in stack frames.
5490 @samp{-moptimize-arg-area} saves space by optimizing them, but this
5491 conflicts with the 88open specifications. The opposite alternative,
5492 @samp{-mno-optimize-arg-area}, agrees with 88open standards. By default
5493 GCC does not optimize the argument area.
5495 @item -mshort-data-@var{num}
5496 @kindex -mshort-data-@var{num}
5497 @cindex smaller data references (88k)
5498 @cindex r0-relative references (88k)
5499 Generate smaller data references by making them relative to @code{r0},
5500 which allows loading a value using a single instruction (rather than the
5501 usual two). You control which data references are affected by
5502 specifying @var{num} with this option. For example, if you specify
5503 @samp{-mshort-data-512}, then the data references affected are those
5504 involving displacements of less than 512 bytes.
5505 @samp{-mshort-data-@var{num}} is not effective for @var{num} greater
5508 @item -mserialize-volatile
5509 @kindex -mserialize-volatile
5510 @itemx -mno-serialize-volatile
5511 @kindex -mno-serialize-volatile
5512 @cindex sequential consistency on 88k
5513 Do, or don't, generate code to guarantee sequential consistency
5514 of volatile memory references. By default, consistency is
5517 The order of memory references made by the MC88110 processor does
5518 not always match the order of the instructions requesting those
5519 references. In particular, a load instruction may execute before
5520 a preceding store instruction. Such reordering violates
5521 sequential consistency of volatile memory references, when there
5522 are multiple processors. When consistency must be guaranteed,
5523 GNU C generates special instructions, as needed, to force
5524 execution in the proper order.
5526 The MC88100 processor does not reorder memory references and so
5527 always provides sequential consistency. However, by default, GNU
5528 C generates the special instructions to guarantee consistency
5529 even when you use @samp{-m88100}, so that the code may be run on an
5530 MC88110 processor. If you intend to run your code only on the
5531 MC88100 processor, you may use @samp{-mno-serialize-volatile}.
5533 The extra code generated to guarantee consistency may affect the
5534 performance of your application. If you know that you can safely
5535 forgo this guarantee, you may use @samp{-mno-serialize-volatile}.
5541 @cindex assembler syntax, 88k
5543 Turn on (@samp{-msvr4}) or off (@samp{-msvr3}) compiler extensions
5544 related to System V release 4 (SVr4). This controls the following:
5548 Which variant of the assembler syntax to emit.
5550 @samp{-msvr4} makes the C preprocessor recognize @samp{#pragma weak}
5551 that is used on System V release 4.
5553 @samp{-msvr4} makes GCC issue additional declaration directives used in
5557 @samp{-msvr4} is the default for the m88k-motorola-sysv4 and
5558 m88k-dg-dgux m88k configurations. @samp{-msvr3} is the default for all
5559 other m88k configurations.
5561 @item -mversion-03.00
5562 @kindex -mversion-03.00
5563 This option is obsolete, and is ignored.
5564 @c ??? which asm syntax better for GAS? option there too?
5566 @item -mno-check-zero-division
5567 @itemx -mcheck-zero-division
5568 @kindex -mno-check-zero-division
5569 @kindex -mcheck-zero-division
5570 @cindex zero division on 88k
5571 Do, or don't, generate code to guarantee that integer division by
5572 zero will be detected. By default, detection is guaranteed.
5574 Some models of the MC88100 processor fail to trap upon integer
5575 division by zero under certain conditions. By default, when
5576 compiling code that might be run on such a processor, GNU C
5577 generates code that explicitly checks for zero-valued divisors
5578 and traps with exception number 503 when one is detected. Use of
5579 mno-check-zero-division suppresses such checking for code
5580 generated to run on an MC88100 processor.
5582 GNU C assumes that the MC88110 processor correctly detects all
5583 instances of integer division by zero. When @samp{-m88110} is
5584 specified, both @samp{-mcheck-zero-division} and
5585 @samp{-mno-check-zero-division} are ignored, and no explicit checks for
5586 zero-valued divisors are generated.
5588 @item -muse-div-instruction
5589 @kindex -muse-div-instruction
5590 @cindex divide instruction, 88k
5591 Use the div instruction for signed integer division on the
5592 MC88100 processor. By default, the div instruction is not used.
5594 On the MC88100 processor the signed integer division instruction
5595 div) traps to the operating system on a negative operand. The
5596 operating system transparently completes the operation, but at a
5597 large cost in execution time. By default, when compiling code
5598 that might be run on an MC88100 processor, GNU C emulates signed
5599 integer division using the unsigned integer division instruction
5600 divu), thereby avoiding the large penalty of a trap to the
5601 operating system. Such emulation has its own, smaller, execution
5602 cost in both time and space. To the extent that your code's
5603 important signed integer division operations are performed on two
5604 nonnegative operands, it may be desirable to use the div
5605 instruction directly.
5607 On the MC88110 processor the div instruction (also known as the
5608 divs instruction) processes negative operands without trapping to
5609 the operating system. When @samp{-m88110} is specified,
5610 @samp{-muse-div-instruction} is ignored, and the div instruction is used
5611 for signed integer division.
5613 Note that the result of dividing INT_MIN by -1 is undefined. In
5614 particular, the behavior of such a division with and without
5615 @samp{-muse-div-instruction} may differ.
5617 @item -mtrap-large-shift
5618 @itemx -mhandle-large-shift
5619 @kindex -mtrap-large-shift
5620 @kindex -mhandle-large-shift
5621 @cindex bit shift overflow (88k)
5622 @cindex large bit shifts (88k)
5623 Include code to detect bit-shifts of more than 31 bits; respectively,
5624 trap such shifts or emit code to handle them properly. By default GCC
5625 makes no special provision for large bit shifts.
5627 @item -mwarn-passed-structs
5628 @kindex -mwarn-passed-structs
5629 @cindex structure passing (88k)
5630 Warn when a function passes a struct as an argument or result.
5631 Structure-passing conventions have changed during the evolution of the C
5632 language, and are often the source of portability problems. By default,
5633 GCC issues no such warning.
5636 @node RS/6000 and PowerPC Options
5637 @subsection IBM RS/6000 and PowerPC Options
5638 @cindex RS/6000 and PowerPC Options
5639 @cindex IBM RS/6000 and PowerPC Options
5641 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
5649 @itemx -mpowerpc-gpopt
5650 @itemx -mno-powerpc-gpopt
5651 @itemx -mpowerpc-gfxopt
5652 @itemx -mno-powerpc-gfxopt
5654 @itemx -mno-powerpc64
5658 @kindex -mpowerpc-gpopt
5659 @kindex -mpowerpc-gfxopt
5661 GCC supports two related instruction set architectures for the
5662 RS/6000 and PowerPC. The @dfn{POWER} instruction set are those
5663 instructions supported by the @samp{rios} chip set used in the original
5664 RS/6000 systems and the @dfn{PowerPC} instruction set is the
5665 architecture of the Motorola MPC5xx, MPC6xx, MPC8xx microprocessors, and
5666 the IBM 4xx microprocessors.
5668 Neither architecture is a subset of the other. However there is a
5669 large common subset of instructions supported by both. An MQ
5670 register is included in processors supporting the POWER architecture.
5672 You use these options to specify which instructions are available on the
5673 processor you are using. The default value of these options is
5674 determined when configuring GCC. Specifying the
5675 @samp{-mcpu=@var{cpu_type}} overrides the specification of these
5676 options. We recommend you use the @samp{-mcpu=@var{cpu_type}} option
5677 rather than the options listed above.
5679 The @samp{-mpower} option allows GCC to generate instructions that
5680 are found only in the POWER architecture and to use the MQ register.
5681 Specifying @samp{-mpower2} implies @samp{-power} and also allows GCC
5682 to generate instructions that are present in the POWER2 architecture but
5683 not the original POWER architecture.
5685 The @samp{-mpowerpc} option allows GCC to generate instructions that
5686 are found only in the 32-bit subset of the PowerPC architecture.
5687 Specifying @samp{-mpowerpc-gpopt} implies @samp{-mpowerpc} and also allows
5688 GCC to use the optional PowerPC architecture instructions in the
5689 General Purpose group, including floating-point square root. Specifying
5690 @samp{-mpowerpc-gfxopt} implies @samp{-mpowerpc} and also allows GCC to
5691 use the optional PowerPC architecture instructions in the Graphics
5692 group, including floating-point select.
5694 The @samp{-mpowerpc64} option allows GCC to generate the additional
5695 64-bit instructions that are found in the full PowerPC64 architecture
5696 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
5697 @samp{-mno-powerpc64}.
5699 If you specify both @samp{-mno-power} and @samp{-mno-powerpc}, GCC
5700 will use only the instructions in the common subset of both
5701 architectures plus some special AIX common-mode calls, and will not use
5702 the MQ register. Specifying both @samp{-mpower} and @samp{-mpowerpc}
5703 permits GCC to use any instruction from either architecture and to
5704 allow use of the MQ register; specify this for the Motorola MPC601.
5706 @item -mnew-mnemonics
5707 @itemx -mold-mnemonics
5708 @kindex -mnew-mnemonics
5709 @kindex -mold-mnemonics
5710 Select which mnemonics to use in the generated assembler code.
5711 @samp{-mnew-mnemonics} requests output that uses the assembler mnemonics
5712 defined for the PowerPC architecture, while @samp{-mold-mnemonics}
5713 requests the assembler mnemonics defined for the POWER architecture.
5714 Instructions defined in only one architecture have only one mnemonic;
5715 GCC uses that mnemonic irrespective of which of these options is
5718 GCC defaults to the mnemonics appropriate for the architecture in
5719 use. Specifying @samp{-mcpu=@var{cpu_type}} sometimes overrides the
5720 value of these option. Unless you are building a cross-compiler, you
5721 should normally not specify either @samp{-mnew-mnemonics} or
5722 @samp{-mold-mnemonics}, but should instead accept the default.
5724 @item -mcpu=@var{cpu_type}
5726 Set architecture type, register usage, choice of mnemonics, and
5727 instruction scheduling parameters for machine type @var{cpu_type}.
5728 Supported values for @var{cpu_type} are @samp{rios}, @samp{rios1},
5729 @samp{rsc}, @samp{rios2}, @samp{rs64a}, @samp{601}, @samp{602},
5730 @samp{603}, @samp{603e}, @samp{604}, @samp{604e}, @samp{620},
5731 @samp{630}, @samp{740}, @samp{750}, @samp{power}, @samp{power2},
5732 @samp{powerpc}, @samp{403}, @samp{505}, @samp{801}, @samp{821},
5733 @samp{823}, and @samp{860} and @samp{common}. @samp{-mcpu=power},
5734 @samp{-mcpu=power2}, @samp{-mcpu=powerpc}, and @samp{-mcpu=powerpc64}
5735 specify generic POWER, POWER2, pure 32-bit PowerPC (i.e., not MPC601),
5736 and 64-bit PowerPC architecture machine types, with an appropriate,
5737 generic processor model assumed for scheduling purposes.@refill
5739 Specifying any of the following options:
5740 @samp{-mcpu=rios1}, @samp{-mcpu=rios2}, @samp{-mcpu=rsc},
5741 @samp{-mcpu=power}, or @samp{-mcpu=power2}
5742 enables the @samp{-mpower} option and disables the @samp{-mpowerpc} option;
5743 @samp{-mcpu=601} enables both the @samp{-mpower} and @samp{-mpowerpc} options.
5744 All of @samp{-mcpu=rs64a}, @samp{-mcpu=602}, @samp{-mcpu=603},
5745 @samp{-mcpu=603e}, @samp{-mcpu=604}, @samp{-mcpu=620}, @samp{-mcpu=630},
5746 @samp{-mcpu=740}, and @samp{-mcpu=750}
5747 enable the @samp{-mpowerpc} option and disable the @samp{-mpower} option.
5748 Exactly similarly, all of @samp{-mcpu=403},
5749 @samp{-mcpu=505}, @samp{-mcpu=821}, @samp{-mcpu=860} and @samp{-mcpu=powerpc}
5750 enable the @samp{-mpowerpc} option and disable the @samp{-mpower} option.
5751 @samp{-mcpu=common} disables both the
5752 @samp{-mpower} and @samp{-mpowerpc} options.@refill
5754 AIX versions 4 or greater selects @samp{-mcpu=common} by default, so
5755 that code will operate on all members of the RS/6000 POWER and PowerPC
5756 families. In that case, GCC will use only the instructions in the
5757 common subset of both architectures plus some special AIX common-mode
5758 calls, and will not use the MQ register. GCC assumes a generic
5759 processor model for scheduling purposes.
5761 Specifying any of the options @samp{-mcpu=rios1}, @samp{-mcpu=rios2},
5762 @samp{-mcpu=rsc}, @samp{-mcpu=power}, or @samp{-mcpu=power2} also
5763 disables the @samp{new-mnemonics} option. Specifying @samp{-mcpu=601},
5764 @samp{-mcpu=602}, @samp{-mcpu=603}, @samp{-mcpu=603e}, @samp{-mcpu=604},
5765 @samp{-mcpu=620}, @samp{-mcpu=630}, @samp{-mcpu=403}, @samp{-mcpu=505},
5766 @samp{-mcpu=821}, @samp{-mcpu=860} or @samp{-mcpu=powerpc} also enables
5767 the @samp{new-mnemonics} option.@refill
5769 Specifying @samp{-mcpu=403}, @samp{-mcpu=821}, or @samp{-mcpu=860} also
5770 enables the @samp{-msoft-float} option.
5772 @item -mtune=@var{cpu_type}
5773 Set the instruction scheduling parameters for machine type
5774 @var{cpu_type}, but do not set the architecture type, register usage,
5775 choice of mnemonics like @samp{-mcpu=}@var{cpu_type} would. The same
5776 values for @var{cpu_type} are used for @samp{-mtune=}@var{cpu_type} as
5777 for @samp{-mcpu=}@var{cpu_type}. The @samp{-mtune=}@var{cpu_type}
5778 option overrides the @samp{-mcpu=}@var{cpu_type} option in terms of
5779 instruction scheduling parameters.
5782 @itemx -mno-fp-in-toc
5783 @itemx -mno-sum-in-toc
5784 @itemx -mminimal-toc
5785 @kindex -mminimal-toc
5786 Modify generation of the TOC (Table Of Contents), which is created for
5787 every executable file. The @samp{-mfull-toc} option is selected by
5788 default. In that case, GCC will allocate at least one TOC entry for
5789 each unique non-automatic variable reference in your program. GCC
5790 will also place floating-point constants in the TOC. However, only
5791 16,384 entries are available in the TOC.
5793 If you receive a linker error message that saying you have overflowed
5794 the available TOC space, you can reduce the amount of TOC space used
5795 with the @samp{-mno-fp-in-toc} and @samp{-mno-sum-in-toc} options.
5796 @samp{-mno-fp-in-toc} prevents GCC from putting floating-point
5797 constants in the TOC and @samp{-mno-sum-in-toc} forces GCC to
5798 generate code to calculate the sum of an address and a constant at
5799 run-time instead of putting that sum into the TOC. You may specify one
5800 or both of these options. Each causes GCC to produce very slightly
5801 slower and larger code at the expense of conserving TOC space.
5803 If you still run out of space in the TOC even when you specify both of
5804 these options, specify @samp{-mminimal-toc} instead. This option causes
5805 GCC to make only one TOC entry for every file. When you specify this
5806 option, GCC will produce code that is slower and larger but which
5807 uses extremely little TOC space. You may wish to use this option
5808 only on files that contain less frequently executed code. @refill
5814 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
5815 @code{long} type, and the infrastructure needed to support them.
5816 Specifying @samp{-maix64} implies @samp{-mpowerpc64} and
5817 @samp{-mpowerpc}, while @samp{-maix32} disables the 64-bit ABI and
5818 implies @samp{-mno-powerpc64}. GCC defaults to @samp{-maix32}.
5823 On AIX, pass floating-point arguments to prototyped functions beyond the
5824 register save area (RSA) on the stack in addition to argument FPRs. The
5825 AIX calling convention was extended but not initially documented to
5826 handle an obscure K&R C case of calling a function that takes the
5827 address of its arguments with fewer arguments than declared. AIX XL
5828 compilers access floating point arguments which do not fit in the
5829 RSA from the stack when a subroutine is compiled without
5830 optimization. Because always storing floating-point arguments on the
5831 stack is inefficient and rarely needed, this option is not enabled by
5832 default and only is necessary when calling subroutines compiled by AIX
5833 XL compilers without optimization.
5837 Support @dfn{AIX Threads}. Link an application written to use
5838 @dfn{pthreads} with special libraries and startup code to enable the
5843 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE). Link an
5844 application written to use message passing with special startup code to
5845 enable the application to run. The system must have PE installed in the
5846 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
5847 must be overridden with the @samp{-specs=} option to specify the
5848 appropriate directory location. The Parallel Environment does not
5849 support threads, so the @samp{-mpe} option and the @samp{-mthreads}
5850 option are incompatible.
5854 @kindex -msoft-float
5855 Generate code that does not use (uses) the floating-point register set.
5856 Software floating point emulation is provided if you use the
5857 @samp{-msoft-float} option, and pass the option to GCC when linking.
5860 @itemx -mno-multiple
5861 Generate code that uses (does not use) the load multiple word
5862 instructions and the store multiple word instructions. These
5863 instructions are generated by default on POWER systems, and not
5864 generated on PowerPC systems. Do not use @samp{-mmultiple} on little
5865 endian PowerPC systems, since those instructions do not work when the
5866 processor is in little endian mode. The exceptions are PPC740 and
5867 PPC750 which permit the instructions usage in little endian mode.
5872 Generate code that uses (does not use) the load string instructions
5873 and the store string word instructions to save multiple registers and
5874 do small block moves. These instructions are generated by default on
5875 POWER systems, and not generated on PowerPC systems. Do not use
5876 @samp{-mstring} on little endian PowerPC systems, since those
5877 instructions do not work when the processor is in little endian mode.
5878 The exceptions are PPC740 and PPC750 which permit the instructions
5879 usage in little endian mode.
5884 Generate code that uses (does not use) the load or store instructions
5885 that update the base register to the address of the calculated memory
5886 location. These instructions are generated by default. If you use
5887 @samp{-mno-update}, there is a small window between the time that the
5888 stack pointer is updated and the address of the previous frame is
5889 stored, which means code that walks the stack frame across interrupts or
5890 signals may get corrupted data.
5893 @itemx -mno-fused-madd
5894 @kindex -mfused-madd
5895 Generate code that uses (does not use) the floating point multiply and
5896 accumulate instructions. These instructions are generated by default if
5897 hardware floating is used.
5899 @item -mno-bit-align
5902 On System V.4 and embedded PowerPC systems do not (do) force structures
5903 and unions that contain bit fields to be aligned to the base type of the
5906 For example, by default a structure containing nothing but 8
5907 @code{unsigned} bitfields of length 1 would be aligned to a 4 byte
5908 boundary and have a size of 4 bytes. By using @samp{-mno-bit-align},
5909 the structure would be aligned to a 1 byte boundary and be one byte in
5912 @item -mno-strict-align
5913 @itemx -mstrict-align
5914 @kindex -mstrict-align
5915 On System V.4 and embedded PowerPC systems do not (do) assume that
5916 unaligned memory references will be handled by the system.
5919 @itemx -mno-relocatable
5920 @kindex -mrelocatable
5921 On embedded PowerPC systems generate code that allows (does not allow)
5922 the program to be relocated to a different address at runtime. If you
5923 use @samp{-mrelocatable} on any module, all objects linked together must
5924 be compiled with @samp{-mrelocatable} or @samp{-mrelocatable-lib}.
5926 @item -mrelocatable-lib
5927 @itemx -mno-relocatable-lib
5928 On embedded PowerPC systems generate code that allows (does not allow)
5929 the program to be relocated to a different address at runtime. Modules
5930 compiled with @samp{-mrelocatable-lib} can be linked with either modules
5931 compiled without @samp{-mrelocatable} and @samp{-mrelocatable-lib} or
5932 with modules compiled with the @samp{-mrelocatable} options.
5936 On System V.4 and embedded PowerPC systems do not (do) assume that
5937 register 2 contains a pointer to a global area pointing to the addresses
5938 used in the program.
5941 @itemx -mlittle-endian
5942 On System V.4 and embedded PowerPC systems compile code for the
5943 processor in little endian mode. The @samp{-mlittle-endian} option is
5944 the same as @samp{-mlittle}.
5948 On System V.4 and embedded PowerPC systems compile code for the
5949 processor in big endian mode. The @samp{-mbig-endian} option is
5950 the same as @samp{-mbig}.
5953 On System V.4 and embedded PowerPC systems compile code using calling
5954 conventions that adheres to the March 1995 draft of the System V
5955 Application Binary Interface, PowerPC processor supplement. This is the
5956 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
5958 @item -mcall-sysv-eabi
5959 Specify both @samp{-mcall-sysv} and @samp{-meabi} options.
5961 @item -mcall-sysv-noeabi
5962 Specify both @samp{-mcall-sysv} and @samp{-mno-eabi} options.
5965 On System V.4 and embedded PowerPC systems compile code using calling
5966 conventions that are similar to those used on AIX. This is the
5967 default if you configured GCC using @samp{powerpc-*-eabiaix}.
5969 @item -mcall-solaris
5970 On System V.4 and embedded PowerPC systems compile code for the Solaris
5974 On System V.4 and embedded PowerPC systems compile code for the
5975 Linux-based GNU system.
5978 @itemx -mno-prototype
5979 On System V.4 and embedded PowerPC systems assume that all calls to
5980 variable argument functions are properly prototyped. Otherwise, the
5981 compiler must insert an instruction before every non prototyped call to
5982 set or clear bit 6 of the condition code register (@var{CR}) to
5983 indicate whether floating point values were passed in the floating point
5984 registers in case the function takes a variable arguments. With
5985 @samp{-mprototype}, only calls to prototyped variable argument functions
5986 will set or clear the bit.
5989 On embedded PowerPC systems, assume that the startup module is called
5990 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
5991 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}.
5995 On embedded PowerPC systems, assume that the startup module is called
5996 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
6000 On embedded PowerPC systems, assume that the startup module is called
6001 @file{crt0.o} and the standard C libraries are @file{libads.a} and
6005 On embedded PowerPC systems, assume that the startup module is called
6006 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
6010 On System V.4 and embedded PowerPC systems, specify that you are
6011 compiling for a VxWorks system.
6014 On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags
6015 header to indicate that @samp{eabi} extended relocations are used.
6019 On System V.4 and embedded PowerPC systems do (do not) adhere to the
6020 Embedded Applications Binary Interface (eabi) which is a set of
6021 modifications to the System V.4 specifications. Selecting @option{-meabi}
6022 means that the stack is aligned to an 8 byte boundary, a function
6023 @code{__eabi} is called to from @code{main} to set up the eabi
6024 environment, and the @samp{-msdata} option can use both @code{r2} and
6025 @code{r13} to point to two separate small data areas. Selecting
6026 @option{-mno-eabi} means that the stack is aligned to a 16 byte boundary,
6027 do not call an initialization function from @code{main}, and the
6028 @samp{-msdata} option will only use @code{r13} to point to a single
6029 small data area. The @samp{-meabi} option is on by default if you
6030 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
6033 On System V.4 and embedded PowerPC systems, put small initialized
6034 @code{const} global and static data in the @samp{.sdata2} section, which
6035 is pointed to by register @code{r2}. Put small initialized
6036 non-@code{const} global and static data in the @samp{.sdata} section,
6037 which is pointed to by register @code{r13}. Put small uninitialized
6038 global and static data in the @samp{.sbss} section, which is adjacent to
6039 the @samp{.sdata} section. The @samp{-msdata=eabi} option is
6040 incompatible with the @samp{-mrelocatable} option. The
6041 @samp{-msdata=eabi} option also sets the @samp{-memb} option.
6044 On System V.4 and embedded PowerPC systems, put small global and static
6045 data in the @samp{.sdata} section, which is pointed to by register
6046 @code{r13}. Put small uninitialized global and static data in the
6047 @samp{.sbss} section, which is adjacent to the @samp{.sdata} section.
6048 The @samp{-msdata=sysv} option is incompatible with the
6049 @samp{-mrelocatable} option.
6051 @item -msdata=default
6053 On System V.4 and embedded PowerPC systems, if @samp{-meabi} is used,
6054 compile code the same as @samp{-msdata=eabi}, otherwise compile code the
6055 same as @samp{-msdata=sysv}.
6058 On System V.4 and embedded PowerPC systems, put small global and static
6059 data in the @samp{.sdata} section. Put small uninitialized global and
6060 static data in the @samp{.sbss} section. Do not use register @code{r13}
6061 to address small data however. This is the default behavior unless
6062 other @samp{-msdata} options are used.
6066 On embedded PowerPC systems, put all initialized global and static data
6067 in the @samp{.data} section, and all uninitialized data in the
6068 @samp{.bss} section.
6071 @cindex smaller data references (PowerPC)
6072 @cindex .sdata/.sdata2 references (PowerPC)
6073 On embedded PowerPC systems, put global and static items less than or
6074 equal to @var{num} bytes into the small data or bss sections instead of
6075 the normal data or bss section. By default, @var{num} is 8. The
6076 @samp{-G @var{num}} switch is also passed to the linker.
6077 All modules should be compiled with the same @samp{-G @var{num}} value.
6080 @itemx -mno-regnames
6081 On System V.4 and embedded PowerPC systems do (do not) emit register
6082 names in the assembly language output using symbolic forms.
6087 @subsection IBM RT Options
6089 @cindex IBM RT options
6091 These @samp{-m} options are defined for the IBM RT PC:
6095 Use an in-line code sequence for integer multiplies. This is the
6098 @item -mcall-lib-mul
6099 Call @code{lmul$$} for integer multiples.
6101 @item -mfull-fp-blocks
6102 Generate full-size floating point data blocks, including the minimum
6103 amount of scratch space recommended by IBM. This is the default.
6105 @item -mminimum-fp-blocks
6106 Do not include extra scratch space in floating point data blocks. This
6107 results in smaller code, but slower execution, since scratch space must
6108 be allocated dynamically.
6110 @cindex @file{varargs.h} and RT PC
6111 @cindex @file{stdarg.h} and RT PC
6112 @item -mfp-arg-in-fpregs
6113 Use a calling sequence incompatible with the IBM calling convention in
6114 which floating point arguments are passed in floating point registers.
6115 Note that @code{varargs.h} and @code{stdargs.h} will not work with
6116 floating point operands if this option is specified.
6118 @item -mfp-arg-in-gregs
6119 Use the normal calling convention for floating point arguments. This is
6122 @item -mhc-struct-return
6123 Return structures of more than one word in memory, rather than in a
6124 register. This provides compatibility with the MetaWare HighC (hc)
6125 compiler. Use the option @samp{-fpcc-struct-return} for compatibility
6126 with the Portable C Compiler (pcc).
6128 @item -mnohc-struct-return
6129 Return some structures of more than one word in registers, when
6130 convenient. This is the default. For compatibility with the
6131 IBM-supplied compilers, use the option @samp{-fpcc-struct-return} or the
6132 option @samp{-mhc-struct-return}.
6136 @subsection MIPS Options
6137 @cindex MIPS options
6139 These @samp{-m} options are defined for the MIPS family of computers:
6142 @item -mcpu=@var{cpu type}
6143 Assume the defaults for the machine type @var{cpu type} when scheduling
6144 instructions. The choices for @var{cpu type} are @samp{r2000}, @samp{r3000},
6145 @samp{r3900}, @samp{r4000}, @samp{r4100}, @samp{r4300}, @samp{r4400},
6146 @samp{r4600}, @samp{r4650}, @samp{r5000}, @samp{r6000}, @samp{r8000},
6147 and @samp{orion}. Additionally, the @samp{r2000}, @samp{r3000},
6148 @samp{r4000}, @samp{r5000}, and @samp{r6000} can be abbreviated as
6149 @samp{r2k} (or @samp{r2K}), @samp{r3k}, etc. While picking a specific
6150 @var{cpu type} will schedule things appropriately for that particular
6151 chip, the compiler will not generate any code that does not meet level 1
6152 of the MIPS ISA (instruction set architecture) without a @samp{-mipsX}
6153 or @samp{-mabi} switch being used.
6156 Issue instructions from level 1 of the MIPS ISA. This is the default.
6157 @samp{r3000} is the default @var{cpu type} at this ISA level.
6160 Issue instructions from level 2 of the MIPS ISA (branch likely, square
6161 root instructions). @samp{r6000} is the default @var{cpu type} at this
6165 Issue instructions from level 3 of the MIPS ISA (64 bit instructions).
6166 @samp{r4000} is the default @var{cpu type} at this ISA level.
6169 Issue instructions from level 4 of the MIPS ISA (conditional move,
6170 prefetch, enhanced FPU instructions). @samp{r8000} is the default
6171 @var{cpu type} at this ISA level.
6174 Assume that 32 32-bit floating point registers are available. This is
6178 Assume that 32 64-bit floating point registers are available. This is
6179 the default when the @samp{-mips3} option is used.
6182 Assume that 32 32-bit general purpose registers are available. This is
6186 Assume that 32 64-bit general purpose registers are available. This is
6187 the default when the @samp{-mips3} option is used.
6190 Force int and long types to be 64 bits wide. See @samp{-mlong32} for an
6191 explanation of the default, and the width of pointers.
6194 Force long types to be 64 bits wide. See @samp{-mlong32} for an
6195 explanation of the default, and the width of pointers.
6198 Force long, int, and pointer types to be 32 bits wide.
6200 If none of @samp{-mlong32}, @samp{-mlong64}, or @samp{-mint64} are set,
6201 the size of ints, longs, and pointers depends on the ABI and ISA chosen.
6202 For @samp{-mabi=32}, and @samp{-mabi=n32}, ints and longs are 32 bits
6203 wide. For @samp{-mabi=64}, ints are 32 bits, and longs are 64 bits wide.
6204 For @samp{-mabi=eabi} and either @samp{-mips1} or @samp{-mips2}, ints
6205 and longs are 32 bits wide. For @samp{-mabi=eabi} and higher ISAs, ints
6206 are 32 bits, and longs are 64 bits wide. The width of pointer types is
6207 the smaller of the width of longs or the width of general purpose
6208 registers (which in turn depends on the ISA).
6215 Generate code for the indicated ABI. The default instruction level is
6216 @samp{-mips1} for @samp{32}, @samp{-mips3} for @samp{n32}, and
6217 @samp{-mips4} otherwise. Conversely, with @samp{-mips1} or
6218 @samp{-mips2}, the default ABI is @samp{32}; otherwise, the default ABI
6222 Generate code for the MIPS assembler, and invoke @file{mips-tfile} to
6223 add normal debug information. This is the default for all
6224 platforms except for the OSF/1 reference platform, using the OSF/rose
6225 object format. If the either of the @samp{-gstabs} or @samp{-gstabs+}
6226 switches are used, the @file{mips-tfile} program will encapsulate the
6227 stabs within MIPS ECOFF.
6230 Generate code for the GNU assembler. This is the default on the OSF/1
6231 reference platform, using the OSF/rose object format. Also, this is
6232 the default if the configure option @samp{--with-gnu-as} is used.
6234 @item -msplit-addresses
6235 @itemx -mno-split-addresses
6236 Generate code to load the high and low parts of address constants separately.
6237 This allows @code{gcc} to optimize away redundant loads of the high order
6238 bits of addresses. This optimization requires GNU as and GNU ld.
6239 This optimization is enabled by default for some embedded targets where
6240 GNU as and GNU ld are standard.
6244 The @samp{-mrnames} switch says to output code using the MIPS software
6245 names for the registers, instead of the hardware names (ie, @var{a0}
6246 instead of @var{$4}). The only known assembler that supports this option
6247 is the Algorithmics assembler.
6251 The @samp{-mgpopt} switch says to write all of the data declarations
6252 before the instructions in the text section, this allows the MIPS
6253 assembler to generate one word memory references instead of using two
6254 words for short global or static data items. This is on by default if
6255 optimization is selected.
6259 For each non-inline function processed, the @samp{-mstats} switch
6260 causes the compiler to emit one line to the standard error file to
6261 print statistics about the program (number of registers saved, stack
6266 The @samp{-mmemcpy} switch makes all block moves call the appropriate
6267 string function (@samp{memcpy} or @samp{bcopy}) instead of possibly
6268 generating inline code.
6271 @itemx -mno-mips-tfile
6272 The @samp{-mno-mips-tfile} switch causes the compiler not
6273 postprocess the object file with the @file{mips-tfile} program,
6274 after the MIPS assembler has generated it to add debug support. If
6275 @file{mips-tfile} is not run, then no local variables will be
6276 available to the debugger. In addition, @file{stage2} and
6277 @file{stage3} objects will have the temporary file names passed to the
6278 assembler embedded in the object file, which means the objects will
6279 not compare the same. The @samp{-mno-mips-tfile} switch should only
6280 be used when there are bugs in the @file{mips-tfile} program that
6281 prevents compilation.
6284 Generate output containing library calls for floating point.
6285 @strong{Warning:} the requisite libraries are not part of GCC.
6286 Normally the facilities of the machine's usual C compiler are used, but
6287 this can't be done directly in cross-compilation. You must make your
6288 own arrangements to provide suitable library functions for
6292 Generate output containing floating point instructions. This is the
6293 default if you use the unmodified sources.
6296 @itemx -mno-abicalls
6297 Emit (or do not emit) the pseudo operations @samp{.abicalls},
6298 @samp{.cpload}, and @samp{.cprestore} that some System V.4 ports use for
6299 position independent code.
6302 @itemx -mno-long-calls
6303 Do all calls with the @samp{JALR} instruction, which requires
6304 loading up a function's address into a register before the call.
6305 You need to use this switch, if you call outside of the current
6306 512 megabyte segment to functions that are not through pointers.
6309 @itemx -mno-half-pic
6310 Put pointers to extern references into the data section and load them
6311 up, rather than put the references in the text section.
6313 @item -membedded-pic
6314 @itemx -mno-embedded-pic
6315 Generate PIC code suitable for some embedded systems. All calls are
6316 made using PC relative address, and all data is addressed using the $gp
6317 register. No more than 65536 bytes of global data may be used. This
6318 requires GNU as and GNU ld which do most of the work. This currently
6319 only works on targets which use ECOFF; it does not work with ELF.
6321 @item -membedded-data
6322 @itemx -mno-embedded-data
6323 Allocate variables to the read-only data section first if possible, then
6324 next in the small data section if possible, otherwise in data. This gives
6325 slightly slower code than the default, but reduces the amount of RAM required
6326 when executing, and thus may be preferred for some embedded systems.
6328 @item -muninit-const-in-rodata
6329 @itemx -mno-uninit-const-in-rodata
6330 When used together with -membedded-data, it will always store uninitialized
6331 const variables in the read-only data section.
6333 @item -msingle-float
6334 @itemx -mdouble-float
6335 The @samp{-msingle-float} switch tells gcc to assume that the floating
6336 point coprocessor only supports single precision operations, as on the
6337 @samp{r4650} chip. The @samp{-mdouble-float} switch permits gcc to use
6338 double precision operations. This is the default.
6342 Permit use of the @samp{mad}, @samp{madu} and @samp{mul} instructions,
6343 as on the @samp{r4650} chip.
6346 Turns on @samp{-msingle-float}, @samp{-mmad}, and, at least for now,
6351 Enable 16-bit instructions.
6354 Use the entry and exit pseudo ops. This option can only be used with
6358 Compile code for the processor in little endian mode.
6359 The requisite libraries are assumed to exist.
6362 Compile code for the processor in big endian mode.
6363 The requisite libraries are assumed to exist.
6366 @cindex smaller data references (MIPS)
6367 @cindex gp-relative references (MIPS)
6368 Put global and static items less than or equal to @var{num} bytes into
6369 the small data or bss sections instead of the normal data or bss
6370 section. This allows the assembler to emit one word memory reference
6371 instructions based on the global pointer (@var{gp} or @var{$28}),
6372 instead of the normal two words used. By default, @var{num} is 8 when
6373 the MIPS assembler is used, and 0 when the GNU assembler is used. The
6374 @samp{-G @var{num}} switch is also passed to the assembler and linker.
6375 All modules should be compiled with the same @samp{-G @var{num}}
6379 Tell the MIPS assembler to not run its preprocessor over user
6380 assembler files (with a @samp{.s} suffix) when assembling them.
6383 Pass an option to gas which will cause nops to be inserted if
6384 the read of the destination register of an mfhi or mflo instruction
6385 occurs in the following two instructions.
6388 Do not include the default crt0.
6392 These options are defined by the macro
6393 @code{TARGET_SWITCHES} in the machine description. The default for the
6394 options is also defined by that macro, which enables you to change the
6399 @subsection Intel 386 Options
6400 @cindex i386 Options
6401 @cindex Intel 386 Options
6403 These @samp{-m} options are defined for the i386 family of computers:
6406 @item -mcpu=@var{cpu type}
6407 Assume the defaults for the machine type @var{cpu type} when scheduling
6408 instructions. The choices for @var{cpu type} are @samp{i386},
6409 @samp{i486}, @samp{i586}, @samp{i686}, @samp{pentium},
6410 @samp{pentiumpro}, @samp{k6}, and @samp{athlon}
6412 While picking a specific @var{cpu type} will schedule things appropriately
6413 for that particular chip, the compiler will not generate any code that
6414 does not run on the i386 without the @samp{-march=@var{cpu type}} option
6415 being used. @samp{i586} is equivalent to @samp{pentium} and @samp{i686}
6416 is equivalent to @samp{pentiumpro}. @samp{k6} is the AMD chip as
6417 opposed to the Intel ones.
6419 @item -march=@var{cpu type}
6420 Generate instructions for the machine type @var{cpu type}. The choices
6421 for @var{cpu type} are the same as for @samp{-mcpu}. Moreover,
6422 specifying @samp{-march=@var{cpu type}} implies @samp{-mcpu=@var{cpu type}}.
6428 Synonyms for -mcpu=i386, -mcpu=i486, -mcpu=pentium, and -mcpu=pentiumpro
6429 respectively. These synonyms are deprecated.
6431 @item -mintel-syntax
6432 Emit assembly using Intel syntax opcodes instead of AT&T syntax.
6436 Control whether or not the compiler uses IEEE floating point
6437 comparisons. These handle correctly the case where the result of a
6438 comparison is unordered.
6441 Generate output containing library calls for floating point.
6442 @strong{Warning:} the requisite libraries are not part of GCC.
6443 Normally the facilities of the machine's usual C compiler are used, but
6444 this can't be done directly in cross-compilation. You must make your
6445 own arrangements to provide suitable library functions for
6448 On machines where a function returns floating point results in the 80387
6449 register stack, some floating point opcodes may be emitted even if
6450 @samp{-msoft-float} is used.
6452 @item -mno-fp-ret-in-387
6453 Do not use the FPU registers for return values of functions.
6455 The usual calling convention has functions return values of types
6456 @code{float} and @code{double} in an FPU register, even if there
6457 is no FPU. The idea is that the operating system should emulate
6460 The option @samp{-mno-fp-ret-in-387} causes such values to be returned
6461 in ordinary CPU registers instead.
6463 @item -mno-fancy-math-387
6464 Some 387 emulators do not support the @code{sin}, @code{cos} and
6465 @code{sqrt} instructions for the 387. Specify this option to avoid
6466 generating those instructions. This option is the default on FreeBSD.
6467 As of revision 2.6.1, these instructions are not generated unless you
6468 also use the @samp{-ffast-math} switch.
6470 @item -malign-double
6471 @itemx -mno-align-double
6472 Control whether GCC aligns @code{double}, @code{long double}, and
6473 @code{long long} variables on a two word boundary or a one word
6474 boundary. Aligning @code{double} variables on a two word boundary will
6475 produce code that runs somewhat faster on a @samp{Pentium} at the
6476 expense of more memory.
6478 @item -m128bit-long-double
6479 @itemx -m128bit-long-double
6480 Control the size of @code{long double} type. i386 application binary interface
6481 specify the size to be 12 bytes, while modern architectures (Pentium and newer)
6482 preffer @code{long double} aligned to 8 or 16 byte boundary. This is
6483 impossible to reach with 12 byte long doubles in the array accesses.
6485 @strong{Warning:} if you use the @samp{-m128bit-long-double} switch, the
6486 structures and arrays containing @code{long double} will change their size as
6487 well as function calling convention for function taking @code{long double}
6490 @item -m96bit-long-double
6491 @itemx -m96bit-long-double
6492 Set the size of @code{long double} to 96 bits as required by the i386
6493 application binary interface. This is the default.
6496 @itemx -mno-svr3-shlib
6497 Control whether GCC places uninitialized locals into @code{bss} or
6498 @code{data}. @samp{-msvr3-shlib} places these locals into @code{bss}.
6499 These options are meaningful only on System V Release 3.
6501 @item -mno-wide-multiply
6502 @itemx -mwide-multiply
6503 Control whether GCC uses the @code{mul} and @code{imul} that produce
6504 64 bit results in @code{eax:edx} from 32 bit operands to do @code{long
6505 long} multiplies and 32-bit division by constants.
6508 Use a different function-calling convention, in which functions that
6509 take a fixed number of arguments return with the @code{ret} @var{num}
6510 instruction, which pops their arguments while returning. This saves one
6511 instruction in the caller since there is no need to pop the arguments
6514 You can specify that an individual function is called with this calling
6515 sequence with the function attribute @samp{stdcall}. You can also
6516 override the @samp{-mrtd} option by using the function attribute
6517 @samp{cdecl}. @xref{Function Attributes}.
6519 @strong{Warning:} this calling convention is incompatible with the one
6520 normally used on Unix, so you cannot use it if you need to call
6521 libraries compiled with the Unix compiler.
6523 Also, you must provide function prototypes for all functions that
6524 take variable numbers of arguments (including @code{printf});
6525 otherwise incorrect code will be generated for calls to those
6528 In addition, seriously incorrect code will result if you call a
6529 function with too many arguments. (Normally, extra arguments are
6530 harmlessly ignored.)
6532 @item -mreg-alloc=@var{regs}
6533 Control the default allocation order of integer registers. The
6534 string @var{regs} is a series of letters specifying a register. The
6535 supported letters are: @code{a} allocate EAX; @code{b} allocate EBX;
6536 @code{c} allocate ECX; @code{d} allocate EDX; @code{S} allocate ESI;
6537 @code{D} allocate EDI; @code{B} allocate EBP.
6539 @item -mregparm=@var{num}
6540 Control how many registers are used to pass integer arguments. By
6541 default, no registers are used to pass arguments, and at most 3
6542 registers can be used. You can control this behavior for a specific
6543 function by using the function attribute @samp{regparm}.
6544 @xref{Function Attributes}.
6546 @strong{Warning:} if you use this switch, and
6547 @var{num} is nonzero, then you must build all modules with the same
6548 value, including any libraries. This includes the system libraries and
6551 @item -malign-loops=@var{num}
6552 Align loops to a 2 raised to a @var{num} byte boundary. If
6553 @samp{-malign-loops} is not specified, the default is 2 unless
6554 gas 2.8 (or later) is being used in which case the default is
6555 to align the loop on a 16 byte boundary if it is less than 8
6558 @item -malign-jumps=@var{num}
6559 Align instructions that are only jumped to to a 2 raised to a @var{num}
6560 byte boundary. If @samp{-malign-jumps} is not specified, the default is
6561 2 if optimizing for a 386, and 4 if optimizing for a 486 unless
6562 gas 2.8 (or later) is being used in which case the default is
6563 to align the instruction on a 16 byte boundary if it is less
6566 @item -malign-functions=@var{num}
6567 Align the start of functions to a 2 raised to @var{num} byte boundary.
6568 If @samp{-malign-functions} is not specified, the default is 2 if optimizing
6569 for a 386, and 4 if optimizing for a 486.
6571 @item -mpreferred-stack-boundary=@var{num}
6572 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
6573 byte boundary. If @samp{-mpreferred-stack-boundary} is not specified,
6574 the default is 4 (16 bytes or 128 bits).
6576 The stack is required to be aligned on a 4 byte boundary. On Pentium
6577 and PentiumPro, @code{double} and @code{long double} values should be
6578 aligned to an 8 byte boundary (see @samp{-malign-double}) or suffer
6579 significant run time performance penalties. On Pentium III, the
6580 Streaming SIMD Extension (SSE) data type @code{__m128} suffers similar
6581 penalties if it is not 16 byte aligned.
6583 To ensure proper alignment of this values on the stack, the stack boundary
6584 must be as aligned as that required by any value stored on the stack.
6585 Further, every function must be generated such that it keeps the stack
6586 aligned. Thus calling a function compiled with a higher preferred
6587 stack boundary from a function compiled with a lower preferred stack
6588 boundary will most likely misalign the stack. It is recommended that
6589 libraries that use callbacks always use the default setting.
6591 This extra alignment does consume extra stack space. Code that is sensitive
6592 to stack space usage, such as embedded systems and operating system kernels,
6593 may want to reduce the preferred alignment to
6594 @samp{-mpreferred-stack-boundary=2}.
6598 Use PUSH operations to store outgoing parameters. This method is shorter
6599 and usually equally fast as method using SUB/MOV operations and is enabled
6600 by default. In some cases disabling it may improve performance because of
6601 improved scheduling and reduced dependencies.
6603 @item -maccumulate-outgoing-args
6604 @kindex -maccumulate-outgoing-args
6605 If enabled, the maximum amount of space required for outgoing arguments will be
6606 computed in the function prologue. This in faster on most modern CPUs
6607 because of reduced dependencies, improved scheduling and reduced stack usage
6608 when preferred stack boundary is not equal to 2. The drawback is a notable
6609 increase in code size. This switch implies -mno-push-args.
6613 Support thread-safe exception handling on @samp{Mingw32}. Code that relies
6614 on thread-safe exception handling must compile and link all code with the
6615 @samp{-mthreads} option. When compiling, @samp{-mthreads} defines
6616 @samp{-D_MT}; when linking, it links in a special thread helper library
6617 @samp{-lmingwthrd} which cleans up per thread exception handling data.
6619 @item -mno-align-stringops
6620 @kindex -mno-align-stringops
6621 Do not align destination of inlined string operations. This switch reduces
6622 code size and improves performance in case the destination is already aligned,
6623 but gcc don't know about it.
6625 @item -minline-all-stringops
6626 @kindex -minline-all-stringops
6627 By default GCC inlines string operations only when destination is known to be
6628 aligned at least to 4 byte boundary. This enables more inlining, increase code
6629 size, but may improve performance of code that depends on fast memcpy, strlen
6630 and memset for short lengths.
6634 @subsection HPPA Options
6635 @cindex HPPA Options
6637 These @samp{-m} options are defined for the HPPA family of computers:
6640 @item -march=@var{architecture type}
6641 Generate code for the specified architecture. The choices for
6642 @var{architecture type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
6643 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
6644 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
6645 architecture option for your machine. Code compiled for lower numbered
6646 architectures will run on higher numbered architectures, but not the
6649 PA 2.0 support currently requires gas snapshot 19990413 or later. The
6650 next release of binutils (current is 2.9.1) will probably contain PA 2.0
6654 @itemx -mpa-risc-1-1
6655 @itemx -mpa-risc-2-0
6656 Synonyms for -march=1.0, -march=1.1, and -march=2.0 respectively.
6659 Generate code suitable for big switch tables. Use this option only if
6660 the assembler/linker complain about out of range branches within a switch
6663 @item -mjump-in-delay
6664 Fill delay slots of function calls with unconditional jump instructions
6665 by modifying the return pointer for the function call to be the target
6666 of the conditional jump.
6668 @item -mdisable-fpregs
6669 Prevent floating point registers from being used in any manner. This is
6670 necessary for compiling kernels which perform lazy context switching of
6671 floating point registers. If you use this option and attempt to perform
6672 floating point operations, the compiler will abort.
6674 @item -mdisable-indexing
6675 Prevent the compiler from using indexing address modes. This avoids some
6676 rather obscure problems when compiling MIG generated code under MACH.
6678 @item -mno-space-regs
6679 Generate code that assumes the target has no space registers. This allows
6680 GCC to generate faster indirect calls and use unscaled index address modes.
6682 Such code is suitable for level 0 PA systems and kernels.
6684 @item -mfast-indirect-calls
6685 Generate code that assumes calls never cross space boundaries. This
6686 allows GCC to emit code which performs faster indirect calls.
6688 This option will not work in the presence of shared libraries or nested
6691 @item -mlong-load-store
6692 Generate 3-instruction load and store sequences as sometimes required by
6693 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
6696 @item -mportable-runtime
6697 Use the portable calling conventions proposed by HP for ELF systems.
6700 Enable the use of assembler directives only GAS understands.
6702 @item -mschedule=@var{cpu type}
6703 Schedule code according to the constraints for the machine type
6704 @var{cpu type}. The choices for @var{cpu type} are @samp{700}
6705 @samp{7100}, @samp{7100LC}, @samp{7200}, and @samp{8000}. Refer to
6706 @file{/usr/lib/sched.models} on an HP-UX system to determine the
6707 proper scheduling option for your machine.
6710 Enable the optimization pass in the HPUX linker. Note this makes symbolic
6711 debugging impossible. It also triggers a bug in the HPUX 8 and HPUX 9 linkers
6712 in which they give bogus error messages when linking some programs.
6715 Generate output containing library calls for floating point.
6716 @strong{Warning:} the requisite libraries are not available for all HPPA
6717 targets. Normally the facilities of the machine's usual C compiler are
6718 used, but this cannot be done directly in cross-compilation. You must make
6719 your own arrangements to provide suitable library functions for
6720 cross-compilation. The embedded target @samp{hppa1.1-*-pro}
6721 does provide software floating point support.
6723 @samp{-msoft-float} changes the calling convention in the output file;
6724 therefore, it is only useful if you compile @emph{all} of a program with
6725 this option. In particular, you need to compile @file{libgcc.a}, the
6726 library that comes with GCC, with @samp{-msoft-float} in order for
6730 @node Intel 960 Options
6731 @subsection Intel 960 Options
6733 These @samp{-m} options are defined for the Intel 960 implementations:
6736 @item -m@var{cpu type}
6737 Assume the defaults for the machine type @var{cpu type} for some of
6738 the other options, including instruction scheduling, floating point
6739 support, and addressing modes. The choices for @var{cpu type} are
6740 @samp{ka}, @samp{kb}, @samp{mc}, @samp{ca}, @samp{cf},
6741 @samp{sa}, and @samp{sb}.
6747 The @samp{-mnumerics} option indicates that the processor does support
6748 floating-point instructions. The @samp{-msoft-float} option indicates
6749 that floating-point support should not be assumed.
6751 @item -mleaf-procedures
6752 @itemx -mno-leaf-procedures
6753 Do (or do not) attempt to alter leaf procedures to be callable with the
6754 @code{bal} instruction as well as @code{call}. This will result in more
6755 efficient code for explicit calls when the @code{bal} instruction can be
6756 substituted by the assembler or linker, but less efficient code in other
6757 cases, such as calls via function pointers, or using a linker that doesn't
6758 support this optimization.
6761 @itemx -mno-tail-call
6762 Do (or do not) make additional attempts (beyond those of the
6763 machine-independent portions of the compiler) to optimize tail-recursive
6764 calls into branches. You may not want to do this because the detection of
6765 cases where this is not valid is not totally complete. The default is
6766 @samp{-mno-tail-call}.
6768 @item -mcomplex-addr
6769 @itemx -mno-complex-addr
6770 Assume (or do not assume) that the use of a complex addressing mode is a
6771 win on this implementation of the i960. Complex addressing modes may not
6772 be worthwhile on the K-series, but they definitely are on the C-series.
6773 The default is currently @samp{-mcomplex-addr} for all processors except
6777 @itemx -mno-code-align
6778 Align code to 8-byte boundaries for faster fetching (or don't bother).
6779 Currently turned on by default for C-series implementations only.
6782 @item -mclean-linkage
6783 @itemx -mno-clean-linkage
6784 These options are not fully implemented.
6788 @itemx -mic2.0-compat
6789 @itemx -mic3.0-compat
6790 Enable compatibility with iC960 v2.0 or v3.0.
6794 Enable compatibility with the iC960 assembler.
6796 @item -mstrict-align
6797 @itemx -mno-strict-align
6798 Do not permit (do permit) unaligned accesses.
6801 Enable structure-alignment compatibility with Intel's gcc release version
6802 1.3 (based on gcc 1.37). This option implies @samp{-mstrict-align}.
6804 @item -mlong-double-64
6805 Implement type @samp{long double} as 64-bit floating point numbers.
6806 Without the option @samp{long double} is implemented by 80-bit
6807 floating point numbers. The only reason we have it because there is
6808 no 128-bit @samp{long double} support in @samp{fp-bit.c} yet. So it
6809 is only useful for people using soft-float targets. Otherwise, we
6810 should recommend against use of it.
6814 @node DEC Alpha Options
6815 @subsection DEC Alpha Options
6817 These @samp{-m} options are defined for the DEC Alpha implementations:
6820 @item -mno-soft-float
6822 Use (do not use) the hardware floating-point instructions for
6823 floating-point operations. When @option{-msoft-float} is specified,
6824 functions in @file{libgcc1.c} will be used to perform floating-point
6825 operations. Unless they are replaced by routines that emulate the
6826 floating-point operations, or compiled in such a way as to call such
6827 emulations routines, these routines will issue floating-point
6828 operations. If you are compiling for an Alpha without floating-point
6829 operations, you must ensure that the library is built so as not to call
6832 Note that Alpha implementations without floating-point operations are
6833 required to have floating-point registers.
6837 Generate code that uses (does not use) the floating-point register set.
6838 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
6839 register set is not used, floating point operands are passed in integer
6840 registers as if they were integers and floating-point results are passed
6841 in $0 instead of $f0. This is a non-standard calling sequence, so any
6842 function with a floating-point argument or return value called by code
6843 compiled with @option{-mno-fp-regs} must also be compiled with that
6846 A typical use of this option is building a kernel that does not use,
6847 and hence need not save and restore, any floating-point registers.
6850 The Alpha architecture implements floating-point hardware optimized for
6851 maximum performance. It is mostly compliant with the IEEE floating
6852 point standard. However, for full compliance, software assistance is
6853 required. This option generates code fully IEEE compliant code
6854 @emph{except} that the @var{inexact flag} is not maintained (see below).
6855 If this option is turned on, the CPP macro @code{_IEEE_FP} is defined
6856 during compilation. The option is a shorthand for: @samp{-D_IEEE_FP
6857 -mfp-trap-mode=su -mtrap-precision=i -mieee-conformant}. The resulting
6858 code is less efficient but is able to correctly support denormalized
6859 numbers and exceptional IEEE values such as not-a-number and plus/minus
6860 infinity. Other Alpha compilers call this option
6861 @option{-ieee_with_no_inexact}.
6863 @item -mieee-with-inexact
6864 @c overfull hbox here --bob 22 jul96
6865 @c original text between ignore ... end ignore
6867 This is like @samp{-mieee} except the generated code also maintains the
6868 IEEE @var{inexact flag}. Turning on this option causes the generated
6869 code to implement fully-compliant IEEE math. The option is a shorthand
6870 for @samp{-D_IEEE_FP -D_IEEE_FP_INEXACT} plus @samp{-mieee-conformant},
6871 @samp{-mfp-trap-mode=sui}, and @samp{-mtrap-precision=i}. On some Alpha
6872 implementations the resulting code may execute significantly slower than
6873 the code generated by default. Since there is very little code that
6874 depends on the @var{inexact flag}, you should normally not specify this
6875 option. Other Alpha compilers call this option
6876 @samp{-ieee_with_inexact}.
6878 @c changed paragraph
6879 This is like @samp{-mieee} except the generated code also maintains the
6880 IEEE @var{inexact flag}. Turning on this option causes the generated
6881 code to implement fully-compliant IEEE math. The option is a shorthand
6882 for @samp{-D_IEEE_FP -D_IEEE_FP_INEXACT} plus the three following:
6883 @samp{-mieee-conformant},
6884 @samp{-mfp-trap-mode=sui},
6885 and @samp{-mtrap-precision=i}.
6886 On some Alpha implementations the resulting code may execute
6887 significantly slower than the code generated by default. Since there
6888 is very little code that depends on the @var{inexact flag}, you should
6889 normally not specify this option. Other Alpha compilers call this
6890 option @samp{-ieee_with_inexact}.
6891 @c end changes to prevent overfull hboxes
6893 @item -mfp-trap-mode=@var{trap mode}
6894 This option controls what floating-point related traps are enabled.
6895 Other Alpha compilers call this option @samp{-fptm }@var{trap mode}.
6896 The trap mode can be set to one of four values:
6900 This is the default (normal) setting. The only traps that are enabled
6901 are the ones that cannot be disabled in software (e.g., division by zero
6905 In addition to the traps enabled by @samp{n}, underflow traps are enabled
6909 Like @samp{su}, but the instructions are marked to be safe for software
6910 completion (see Alpha architecture manual for details).
6913 Like @samp{su}, but inexact traps are enabled as well.
6916 @item -mfp-rounding-mode=@var{rounding mode}
6917 Selects the IEEE rounding mode. Other Alpha compilers call this option
6918 @samp{-fprm }@var{rounding mode}. The @var{rounding mode} can be one
6923 Normal IEEE rounding mode. Floating point numbers are rounded towards
6924 the nearest machine number or towards the even machine number in case
6928 Round towards minus infinity.
6931 Chopped rounding mode. Floating point numbers are rounded towards zero.
6934 Dynamic rounding mode. A field in the floating point control register
6935 (@var{fpcr}, see Alpha architecture reference manual) controls the
6936 rounding mode in effect. The C library initializes this register for
6937 rounding towards plus infinity. Thus, unless your program modifies the
6938 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
6941 @item -mtrap-precision=@var{trap precision}
6942 In the Alpha architecture, floating point traps are imprecise. This
6943 means without software assistance it is impossible to recover from a
6944 floating trap and program execution normally needs to be terminated.
6945 GCC can generate code that can assist operating system trap handlers
6946 in determining the exact location that caused a floating point trap.
6947 Depending on the requirements of an application, different levels of
6948 precisions can be selected:
6952 Program precision. This option is the default and means a trap handler
6953 can only identify which program caused a floating point exception.
6956 Function precision. The trap handler can determine the function that
6957 caused a floating point exception.
6960 Instruction precision. The trap handler can determine the exact
6961 instruction that caused a floating point exception.
6964 Other Alpha compilers provide the equivalent options called
6965 @samp{-scope_safe} and @samp{-resumption_safe}.
6967 @item -mieee-conformant
6968 This option marks the generated code as IEEE conformant. You must not
6969 use this option unless you also specify @samp{-mtrap-precision=i} and either
6970 @samp{-mfp-trap-mode=su} or @samp{-mfp-trap-mode=sui}. Its only effect
6971 is to emit the line @samp{.eflag 48} in the function prologue of the
6972 generated assembly file. Under DEC Unix, this has the effect that
6973 IEEE-conformant math library routines will be linked in.
6975 @item -mbuild-constants
6976 Normally GCC examines a 32- or 64-bit integer constant to
6977 see if it can construct it from smaller constants in two or three
6978 instructions. If it cannot, it will output the constant as a literal and
6979 generate code to load it from the data segment at runtime.
6981 Use this option to require GCC to construct @emph{all} integer constants
6982 using code, even if it takes more instructions (the maximum is six).
6984 You would typically use this option to build a shared library dynamic
6985 loader. Itself a shared library, it must relocate itself in memory
6986 before it can find the variables and constants in its own data segment.
6990 Select whether to generate code to be assembled by the vendor-supplied
6991 assembler (@samp{-malpha-as}) or by the GNU assembler @samp{-mgas}.
6999 Indicate whether GCC should generate code to use the optional BWX,
7000 CIX, and MAX instruction sets. The default is to use the instruction sets
7001 supported by the CPU type specified via @samp{-mcpu=} option or that
7002 of the CPU on which GCC was built if none was specified.
7004 @item -mcpu=@var{cpu_type}
7005 Set the instruction set, register set, and instruction scheduling
7006 parameters for machine type @var{cpu_type}. You can specify either the
7007 @samp{EV} style name or the corresponding chip number. GCC
7008 supports scheduling parameters for the EV4 and EV5 family of processors
7009 and will choose the default values for the instruction set from
7010 the processor you specify. If you do not specify a processor type,
7011 GCC will default to the processor on which the compiler was built.
7013 Supported values for @var{cpu_type} are
7018 Schedules as an EV4 and has no instruction set extensions.
7022 Schedules as an EV5 and has no instruction set extensions.
7026 Schedules as an EV5 and supports the BWX extension.
7031 Schedules as an EV5 and supports the BWX and MAX extensions.
7035 Schedules as an EV5 (until Digital releases the scheduling parameters
7036 for the EV6) and supports the BWX, CIX, and MAX extensions.
7039 @item -mmemory-latency=@var{time}
7040 Sets the latency the scheduler should assume for typical memory
7041 references as seen by the application. This number is highly
7042 dependent on the memory access patterns used by the application
7043 and the size of the external cache on the machine.
7045 Valid options for @var{time} are
7049 A decimal number representing clock cycles.
7055 The compiler contains estimates of the number of clock cycles for
7056 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
7057 (also called Dcache, Scache, and Bcache), as well as to main memory.
7058 Note that L3 is only valid for EV5.
7063 @node Clipper Options
7064 @subsection Clipper Options
7066 These @samp{-m} options are defined for the Clipper implementations:
7070 Produce code for a C300 Clipper processor. This is the default.
7073 Produce code for a C400 Clipper processor i.e. use floating point
7077 @node H8/300 Options
7078 @subsection H8/300 Options
7080 These @samp{-m} options are defined for the H8/300 implementations:
7084 Shorten some address references at link time, when possible; uses the
7085 linker option @samp{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
7086 ld.info, Using ld}, for a fuller description.
7089 Generate code for the H8/300H.
7092 Generate code for the H8/S.
7095 Generate code for the H8/S2600. This switch must be used with -ms.
7098 Make @code{int} data 32 bits by default.
7101 On the H8/300H and H8/S, use the same alignment rules as for the H8/300.
7102 The default for the H8/300H and H8/S is to align longs and floats on 4
7104 @samp{-malign-300} causes them to be aligned on 2 byte boundaries.
7105 This option has no effect on the H8/300.
7109 @subsection SH Options
7111 These @samp{-m} options are defined for the SH implementations:
7115 Generate code for the SH1.
7118 Generate code for the SH2.
7121 Generate code for the SH3.
7124 Generate code for the SH3e.
7127 Generate code for the SH4 without a floating-point unit.
7129 @item -m4-single-only
7130 Generate code for the SH4 with a floating-point unit that only
7131 supports single-precision arithmentic.
7134 Generate code for the SH4 assuming the floating-point unit is in
7135 single-precision mode by default.
7138 Generate code for the SH4.
7141 Compile code for the processor in big endian mode.
7144 Compile code for the processor in little endian mode.
7147 Align doubles at 64 bit boundaries. Note that this changes the calling
7148 conventions, and thus some functions from the standard C library will
7149 not work unless you recompile it first with -mdalign.
7152 Shorten some address references at link time, when possible; uses the
7153 linker option @samp{-relax}.
7156 Use 32-bit offsets in @code{switch} tables. The default is to use
7160 Enable the use of the instruction @code{fmovd}.
7163 Comply with the calling conventions defined by Hitachi.
7166 Mark the @code{MAC} register as call-clobbered, even if
7167 @option{-mhitachi} is given.
7170 Dump instruction size and location in the assembly code.
7173 This option is deprecated. It pads structures to multiple of 4 bytes,
7174 which is incompatible with the SH ABI.
7177 Optimize for space instead of speed. Implied by @option{-Os}.
7180 When generating position-independent code, emit function calls using
7181 the Global Offset Table instead of the Procedure Linkage Table.
7184 Generate a library function call to invalidate instruction cache
7185 entries, after fixing up a trampoline. This library function call
7186 doesn't assume it can write to the whole memory address space. This
7187 is the default when the target is @code{sh-*-linux*}.
7190 @node System V Options
7191 @subsection Options for System V
7193 These additional options are available on System V Release 4 for
7194 compatibility with other compilers on those systems:
7198 Create a shared object.
7199 It is recommended that @samp{-symbolic} or @samp{-shared} be used instead.
7202 Identify the versions of each tool used by the compiler, in a
7203 @code{.ident} assembler directive in the output.
7206 Refrain from adding @code{.ident} directives to the output file (this is
7209 @item -YP,@var{dirs}
7210 Search the directories @var{dirs}, and no others, for libraries
7211 specified with @samp{-l}.
7214 Look in the directory @var{dir} to find the M4 preprocessor.
7215 The assembler uses this option.
7216 @c This is supposed to go with a -Yd for predefined M4 macro files, but
7217 @c the generic assembler that comes with Solaris takes just -Ym.
7220 @node TMS320C3x/C4x Options
7221 @subsection TMS320C3x/C4x Options
7222 @cindex TMS320C3x/C4x Options
7224 These @samp{-m} options are defined for TMS320C3x/C4x implementations:
7228 @item -mcpu=@var{cpu_type}
7229 Set the instruction set, register set, and instruction scheduling
7230 parameters for machine type @var{cpu_type}. Supported values for
7231 @var{cpu_type} are @samp{c30}, @samp{c31}, @samp{c32}, @samp{c40}, and
7232 @samp{c44}. The default is @samp{c40} to generate code for the
7237 @itemx -msmall-memory
7239 Generates code for the big or small memory model. The small memory
7240 model assumed that all data fits into one 64K word page. At run-time
7241 the data page (DP) register must be set to point to the 64K page
7242 containing the .bss and .data program sections. The big memory model is
7243 the default and requires reloading of the DP register for every direct
7248 Allow (disallow) allocation of general integer operands into the block
7253 Enable (disable) generation of code using decrement and branch,
7254 DBcond(D), instructions. This is enabled by default for the C4x. To be
7255 on the safe side, this is disabled for the C3x, since the maximum
7256 iteration count on the C3x is 2^23 + 1 (but who iterates loops more than
7257 2^23 times on the C3x?). Note that GCC will try to reverse a loop so
7258 that it can utilise the decrement and branch instruction, but will give
7259 up if there is more than one memory reference in the loop. Thus a loop
7260 where the loop counter is decremented can generate slightly more
7261 efficient code, in cases where the RPTB instruction cannot be utilised.
7263 @item -mdp-isr-reload
7265 Force the DP register to be saved on entry to an interrupt service
7266 routine (ISR), reloaded to point to the data section, and restored on
7267 exit from the ISR. This should not be required unless someone has
7268 violated the small memory model by modifying the DP register, say within
7273 For the C3x use the 24-bit MPYI instruction for integer multiplies
7274 instead of a library call to guarantee 32-bit results. Note that if one
7275 of the operands is a constant, then the multiplication will be performed
7276 using shifts and adds. If the -mmpyi option is not specified for the C3x,
7277 then squaring operations are performed inline instead of a library call.
7280 @itemx -mno-fast-fix
7281 The C3x/C4x FIX instruction to convert a floating point value to an
7282 integer value chooses the nearest integer less than or equal to the
7283 floating point value rather than to the nearest integer. Thus if the
7284 floating point number is negative, the result will be incorrectly
7285 truncated an additional code is necessary to detect and correct this
7286 case. This option can be used to disable generation of the additional
7287 code required to correct the result.
7291 Enable (disable) generation of repeat block sequences using the RPTB
7292 instruction for zero overhead looping. The RPTB construct is only used
7293 for innermost loops that do not call functions or jump across the loop
7294 boundaries. There is no advantage having nested RPTB loops due to the
7295 overhead required to save and restore the RC, RS, and RE registers.
7296 This is enabled by default with -O2.
7298 @item -mrpts=@var{count}
7300 Enable (disable) the use of the single instruction repeat instruction
7301 RPTS. If a repeat block contains a single instruction, and the loop
7302 count can be guaranteed to be less than the value @var{count}, GCC will
7303 emit a RPTS instruction instead of a RPTB. If no value is specified,
7304 then a RPTS will be emitted even if the loop count cannot be determined
7305 at compile time. Note that the repeated instruction following RPTS does
7306 not have to be reloaded from memory each iteration, thus freeing up the
7307 CPU buses for operands. However, since interrupts are blocked by this
7308 instruction, it is disabled by default.
7310 @item -mloop-unsigned
7311 @itemx -mno-loop-unsigned
7312 The maximum iteration count when using RPTS and RPTB (and DB on the C40)
7313 is 2^31 + 1 since these instructions test if the iteration count is
7314 negative to terminate the loop. If the iteration count is unsigned
7315 there is a possibility than the 2^31 + 1 maximum iteration count may be
7316 exceeded. This switch allows an unsigned iteration count.
7319 Try to emit an assembler syntax that the TI assembler (asm30) is happy
7320 with. This also enforces compatibility with the API employed by the TI
7321 C3x C compiler. For example, long doubles are passed as structures
7322 rather than in floating point registers.
7326 Generate code that uses registers (stack) for passing arguments to functions.
7327 By default, arguments are passed in registers where possible rather
7328 than by pushing arguments on to the stack.
7330 @item -mparallel-insns
7331 @itemx -mno-parallel-insns
7332 Allow the generation of parallel instructions. This is enabled by
7335 @item -mparallel-mpy
7336 @itemx -mno-parallel-mpy
7337 Allow the generation of MPY||ADD and MPY||SUB parallel instructions,
7338 provided -mparallel-insns is also specified. These instructions have
7339 tight register constraints which can pessimize the code generation
7345 @subsection V850 Options
7346 @cindex V850 Options
7348 These @samp{-m} options are defined for V850 implementations:
7352 @itemx -mno-long-calls
7353 Treat all calls as being far away (near). If calls are assumed to be
7354 far away, the compiler will always load the functions address up into a
7355 register, and call indirect through the pointer.
7359 Do not optimize (do optimize) basic blocks that use the same index
7360 pointer 4 or more times to copy pointer into the @code{ep} register, and
7361 use the shorter @code{sld} and @code{sst} instructions. The @samp{-mep}
7362 option is on by default if you optimize.
7364 @item -mno-prolog-function
7365 @itemx -mprolog-function
7366 Do not use (do use) external functions to save and restore registers at
7367 the prolog and epilog of a function. The external functions are slower,
7368 but use less code space if more than one function saves the same number
7369 of registers. The @samp{-mprolog-function} option is on by default if
7373 Try to make the code as small as possible. At present, this just turns
7374 on the @samp{-mep} and @samp{-mprolog-function} options.
7377 Put static or global variables whose size is @var{n} bytes or less into
7378 the tiny data area that register @code{ep} points to. The tiny data
7379 area can hold up to 256 bytes in total (128 bytes for byte references).
7382 Put static or global variables whose size is @var{n} bytes or less into
7383 the small data area that register @code{gp} points to. The small data
7384 area can hold up to 64 kilobytes.
7387 Put static or global variables whose size is @var{n} bytes or less into
7388 the first 32 kilobytes of memory.
7391 Specify that the target processor is the V850.
7394 Generate code suitable for big switch tables. Use this option only if
7395 the assembler/linker complain about out of range branches within a switch
7400 @subsection ARC Options
7403 These options are defined for ARC implementations:
7407 Compile code for little endian mode. This is the default.
7410 Compile code for big endian mode.
7413 Prepend the name of the cpu to all public symbol names.
7414 In multiple-processor systems, there are many ARC variants with different
7415 instruction and register set characteristics. This flag prevents code
7416 compiled for one cpu to be linked with code compiled for another.
7417 No facility exists for handling variants that are "almost identical".
7418 This is an all or nothing option.
7420 @item -mcpu=@var{cpu}
7421 Compile code for ARC variant @var{cpu}.
7422 Which variants are supported depend on the configuration.
7423 All variants support @samp{-mcpu=base}, this is the default.
7425 @item -mtext=@var{text section}
7426 @itemx -mdata=@var{data section}
7427 @itemx -mrodata=@var{readonly data section}
7428 Put functions, data, and readonly data in @var{text section},
7429 @var{data section}, and @var{readonly data section} respectively
7430 by default. This can be overridden with the @code{section} attribute.
7431 @xref{Variable Attributes}.
7436 @subsection NS32K Options
7437 @cindex NS32K options
7439 These are the @samp{-m} options defined for the 32000 series. The default
7440 values for these options depends on which style of 32000 was selected when
7441 the compiler was configured; the defaults for the most common choices are
7447 Generate output for a 32032. This is the default
7448 when the compiler is configured for 32032 and 32016 based systems.
7452 Generate output for a 32332. This is the default
7453 when the compiler is configured for 32332-based systems.
7457 Generate output for a 32532. This is the default
7458 when the compiler is configured for 32532-based systems.
7461 Generate output containing 32081 instructions for floating point.
7462 This is the default for all systems.
7465 Generate output containing 32381 instructions for floating point. This
7466 also implies @samp{-m32081}. The 32381 is only compatible with the 32332
7467 and 32532 cpus. This is the default for the pc532-netbsd configuration.
7470 Try and generate multiply-add floating point instructions @code{polyF}
7471 and @code{dotF}. This option is only available if the @samp{-m32381}
7472 option is in effect. Using these instructions requires changes to to
7473 register allocation which generally has a negative impact on
7474 performance. This option should only be enabled when compiling code
7475 particularly likely to make heavy use of multiply-add instructions.
7478 Do not try and generate multiply-add floating point instructions
7479 @code{polyF} and @code{dotF}. This is the default on all platforms.
7482 Generate output containing library calls for floating point.
7483 @strong{Warning:} the requisite libraries may not be available.
7486 Do not use the bit-field instructions. On some machines it is faster to
7487 use shifting and masking operations. This is the default for the pc532.
7490 Do use the bit-field instructions. This is the default for all platforms
7494 Use a different function-calling convention, in which functions
7495 that take a fixed number of arguments return pop their
7496 arguments on return with the @code{ret} instruction.
7498 This calling convention is incompatible with the one normally
7499 used on Unix, so you cannot use it if you need to call libraries
7500 compiled with the Unix compiler.
7502 Also, you must provide function prototypes for all functions that
7503 take variable numbers of arguments (including @code{printf});
7504 otherwise incorrect code will be generated for calls to those
7507 In addition, seriously incorrect code will result if you call a
7508 function with too many arguments. (Normally, extra arguments are
7509 harmlessly ignored.)
7511 This option takes its name from the 680x0 @code{rtd} instruction.
7515 Use a different function-calling convention where the first two arguments
7516 are passed in registers.
7518 This calling convention is incompatible with the one normally
7519 used on Unix, so you cannot use it if you need to call libraries
7520 compiled with the Unix compiler.
7523 Do not pass any arguments in registers. This is the default for all
7527 It is OK to use the sb as an index register which is always loaded with
7528 zero. This is the default for the pc532-netbsd target.
7531 The sb register is not available for use or has not been initialized to
7532 zero by the run time system. This is the default for all targets except
7533 the pc532-netbsd. It is also implied whenever @samp{-mhimem} or
7534 @samp{-fpic} is set.
7537 Many ns32000 series addressing modes use displacements of up to 512MB.
7538 If an address is above 512MB then displacements from zero can not be used.
7539 This option causes code to be generated which can be loaded above 512MB.
7540 This may be useful for operating systems or ROM code.
7543 Assume code will be loaded in the first 512MB of virtual address space.
7544 This is the default for all platforms.
7550 @subsection AVR Options
7553 These options are defined for AVR implementations:
7556 @item -mmcu=@var{mcu}
7557 Specify ATMEL AVR instruction set or MCU type.
7559 Instruction set avr1 is for the minimal AVR core, not supported by the C
7560 compiler, only for assembler programs (MCU types: at90s1200, attiny10,
7561 attiny11, attiny12, attiny15, attiny28).
7563 Instruction set avr2 (default) is for the classic AVR core with up to
7564 8K program memory space (MCU types: at90s2313, at90s2323, attiny22,
7565 at90s2333, at90s2343, at90s4414, at90s4433, at90s4434, at90s8515,
7566 at90c8534, at90s8535).
7568 Instruction set avr3 is for the classic AVR core with up to 128K program
7569 memory space (MCU types: atmega103, atmega603).
7571 Instruction set avr4 is for the enhanced AVR core with up to 8K program
7572 memory space (MCU types: atmega83, atmega85).
7574 Instruction set avr5 is for the enhanced AVR core with up to 128K program
7575 memory space (MCU types: atmega161, atmega163, atmega32, at94k).
7578 Output instruction sizes to the asm file.
7580 @item -minit-stack=@var{N}
7581 Specify the initial stack address, which may be a symbol or numeric value,
7582 __stack is the default.
7584 @item -mno-interrupts
7585 Generated code is not compatible with hardware interrupts.
7586 Code size will be smaller.
7588 @item -mcall-prologues
7589 Functions prologues/epilogues expanded as call to appropriate
7590 subroutines. Code size will be smaller.
7592 @item -mno-tablejump
7593 Do not generate tablejump insns which sometimes increase code size.
7596 Change only the low 8 bits of the stack pointer.
7600 @subsection MCore Options
7601 @cindex MCore options
7603 These are the @samp{-m} options defined for the Motorola M*Core
7611 Inline constants into the code stream if it can be done in two
7612 instructions or less.
7617 Use the divide instruction. (Enabled by default).
7619 @item -mrelax-immediate
7620 @itemx -mrelax-immediate
7621 @itemx -mno-relax-immediate
7622 Allow arbitrary sized immediates in bit operations.
7624 @item -mwide-bitfields
7625 @itemx -mwide-bitfields
7626 @itemx -mno-wide-bitfields
7627 Always treat bitfields as int-sized.
7629 @item -m4byte-functions
7630 @itemx -m4byte-functions
7631 @itemx -mno-4byte-functions
7632 Force all functions to be aligned to a four byte boundary.
7634 @item -mcallgraph-data
7635 @itemx -mcallgraph-data
7636 @itemx -mno-callgraph-data
7637 Emit callgraph information.
7641 @itemx -mno-slow-bytes
7642 Prefer word access when reading byte quantities.
7644 @item -mlittle-endian
7645 @itemx -mlittle-endian
7647 Generate code for a little endian target.
7652 Generate code for the 210 processor.
7656 @subsection D30V Options
7657 @cindex D30V Options
7659 These @samp{-m} options are defined for D30V implementations:
7663 Link the @samp{.text}, @samp{.data}, @samp{.bss}, @samp{.strings},
7664 @samp{.rodata}, @samp{.rodata1}, @samp{.data1} sections into external
7665 memory, which starts at location @code{0x80000000}.
7668 Same as the @samp{-mextmem} switch.
7671 Link the @samp{.text} section into onchip text memory, which starts at
7672 location @code{0x0}. Also link @samp{.data}, @samp{.bss},
7673 @samp{.strings}, @samp{.rodata}, @samp{.rodata1}, @samp{.data1} sections
7674 into onchip data memory, which starts at location @code{0x20000000}.
7676 @item -mno-asm-optimize
7677 @itemx -masm-optimize
7678 Disable (enable) passing @samp{-O} to the assembler when optimizing.
7679 The assembler uses the @samp{-O} option to automatically parallelize
7680 adjacent short instructions where possible.
7682 @item -mbranch-cost=@var{n}
7683 Increase the internal costs of branches to @var{n}. Higher costs means
7684 that the compiler will issue more instructions to avoid doing a branch.
7687 @item -mcond-exec=@var{n}
7688 Specify the maximum number of conditionally executed instructions that
7689 replace a branch. The default is 4.
7692 @node Code Gen Options
7693 @section Options for Code Generation Conventions
7694 @cindex code generation conventions
7695 @cindex options, code generation
7696 @cindex run-time options
7698 These machine-independent options control the interface conventions
7699 used in code generation.
7701 Most of them have both positive and negative forms; the negative form
7702 of @samp{-ffoo} would be @samp{-fno-foo}. In the table below, only
7703 one of the forms is listed---the one which is not the default. You
7704 can figure out the other form by either removing @samp{no-} or adding
7709 Enable exception handling. Generates extra code needed to propagate
7710 exceptions. For some targets, this implies GNU CC will generate frame
7711 unwind information for all functions, which can produce significant data
7712 size overhead, although it does not affect execution. If you do not
7713 specify this option, GNU CC will enable it by default for languages like
7714 C++ which normally require exception handling, and disable itfor
7715 languages like C that do not normally require it. However, you may need
7716 to enable this option when compiling C code that needs to interoperate
7717 properly with exception handlers written in C++. You may also wish to
7718 disable this option if you are compiling older C++ programs that don't
7719 use exception handling.
7721 @item -funwind-tables
7722 Similar to @option{-fexceptions}, except that it will just generate any needed
7723 static data, but will not affect the generated code in any other way.
7724 You will normally not enable this option; instead, a language processor
7725 that needs this handling would enable it on your behalf.
7727 @item -fpcc-struct-return
7728 Return ``short'' @code{struct} and @code{union} values in memory like
7729 longer ones, rather than in registers. This convention is less
7730 efficient, but it has the advantage of allowing intercallability between
7731 GCC-compiled files and files compiled with other compilers.
7733 The precise convention for returning structures in memory depends
7734 on the target configuration macros.
7736 Short structures and unions are those whose size and alignment match
7737 that of some integer type.
7739 @item -freg-struct-return
7740 Use the convention that @code{struct} and @code{union} values are
7741 returned in registers when possible. This is more efficient for small
7742 structures than @samp{-fpcc-struct-return}.
7744 If you specify neither @samp{-fpcc-struct-return} nor its contrary
7745 @samp{-freg-struct-return}, GCC defaults to whichever convention is
7746 standard for the target. If there is no standard convention, GCC
7747 defaults to @samp{-fpcc-struct-return}, except on targets where GCC
7748 is the principal compiler. In those cases, we can choose the standard,
7749 and we chose the more efficient register return alternative.
7752 Allocate to an @code{enum} type only as many bytes as it needs for the
7753 declared range of possible values. Specifically, the @code{enum} type
7754 will be equivalent to the smallest integer type which has enough room.
7756 @item -fshort-double
7757 Use the same size for @code{double} as for @code{float}.
7760 Requests that the data and non-@code{const} variables of this
7761 compilation be shared data rather than private data. The distinction
7762 makes sense only on certain operating systems, where shared data is
7763 shared between processes running the same program, while private data
7764 exists in one copy per process.
7767 Allocate even uninitialized global variables in the data section of the
7768 object file, rather than generating them as common blocks. This has the
7769 effect that if the same variable is declared (without @code{extern}) in
7770 two different compilations, you will get an error when you link them.
7771 The only reason this might be useful is if you wish to verify that the
7772 program will work on other systems which always work this way.
7775 Ignore the @samp{#ident} directive.
7777 @item -fno-gnu-linker
7778 Do not output global initializations (such as C++ constructors and
7779 destructors) in the form used by the GNU linker (on systems where the GNU
7780 linker is the standard method of handling them). Use this option when
7781 you want to use a non-GNU linker, which also requires using the
7782 @command{collect2} program to make sure the system linker includes
7783 constructors and destructors. (@command{collect2} is included in the GCC
7784 distribution.) For systems which @emph{must} use @command{collect2}, the
7785 compiler driver @command{gcc} is configured to do this automatically.
7787 @item -finhibit-size-directive
7788 Don't output a @code{.size} assembler directive, or anything else that
7789 would cause trouble if the function is split in the middle, and the
7790 two halves are placed at locations far apart in memory. This option is
7791 used when compiling @file{crtstuff.c}; you should not need to use it
7795 Put extra commentary information in the generated assembly code to
7796 make it more readable. This option is generally only of use to those
7797 who actually need to read the generated assembly code (perhaps while
7798 debugging the compiler itself).
7800 @samp{-fno-verbose-asm}, the default, causes the
7801 extra information to be omitted and is useful when comparing two assembler
7805 Consider all memory references through pointers to be volatile.
7807 @item -fvolatile-global
7808 Consider all memory references to extern and global data items to
7809 be volatile. GCC does not consider static data items to be volatile
7810 because of this switch.
7812 @item -fvolatile-static
7813 Consider all memory references to static data to be volatile.
7816 @cindex global offset table
7818 Generate position-independent code (PIC) suitable for use in a shared
7819 library, if supported for the target machine. Such code accesses all
7820 constant addresses through a global offset table (GOT). The dynamic
7821 loader resolves the GOT entries when the program starts (the dynamic
7822 loader is not part of GCC; it is part of the operating system). If
7823 the GOT size for the linked executable exceeds a machine-specific
7824 maximum size, you get an error message from the linker indicating that
7825 @samp{-fpic} does not work; in that case, recompile with @samp{-fPIC}
7826 instead. (These maximums are 16k on the m88k, 8k on the Sparc, and 32k
7827 on the m68k and RS/6000. The 386 has no such limit.)
7829 Position-independent code requires special support, and therefore works
7830 only on certain machines. For the 386, GCC supports PIC for System V
7831 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
7832 position-independent.
7835 If supported for the target machine, emit position-independent code,
7836 suitable for dynamic linking and avoiding any limit on the size of the
7837 global offset table. This option makes a difference on the m68k, m88k,
7840 Position-independent code requires special support, and therefore works
7841 only on certain machines.
7843 @item -ffixed-@var{reg}
7844 Treat the register named @var{reg} as a fixed register; generated code
7845 should never refer to it (except perhaps as a stack pointer, frame
7846 pointer or in some other fixed role).
7848 @var{reg} must be the name of a register. The register names accepted
7849 are machine-specific and are defined in the @code{REGISTER_NAMES}
7850 macro in the machine description macro file.
7852 This flag does not have a negative form, because it specifies a
7855 @item -fcall-used-@var{reg}
7856 Treat the register named @var{reg} as an allocable register that is
7857 clobbered by function calls. It may be allocated for temporaries or
7858 variables that do not live across a call. Functions compiled this way
7859 will not save and restore the register @var{reg}.
7861 It is an error to used this flag with the frame pointer or stack pointer.
7862 Use of this flag for other registers that have fixed pervasive roles in
7863 the machine's execution model will produce disastrous results.
7865 This flag does not have a negative form, because it specifies a
7868 @item -fcall-saved-@var{reg}
7869 Treat the register named @var{reg} as an allocable register saved by
7870 functions. It may be allocated even for temporaries or variables that
7871 live across a call. Functions compiled this way will save and restore
7872 the register @var{reg} if they use it.
7874 It is an error to used this flag with the frame pointer or stack pointer.
7875 Use of this flag for other registers that have fixed pervasive roles in
7876 the machine's execution model will produce disastrous results.
7878 A different sort of disaster will result from the use of this flag for
7879 a register in which function values may be returned.
7881 This flag does not have a negative form, because it specifies a
7885 Pack all structure members together without holes. Usually you would
7886 not want to use this option, since it makes the code suboptimal, and
7887 the offsets of structure members won't agree with system libraries.
7889 @item -fcheck-memory-usage
7890 Generate extra code to check each memory access. GCC will generate
7891 code that is suitable for a detector of bad memory accesses such as
7894 Normally, you should compile all, or none, of your code with this option.
7896 If you do mix code compiled with and without this option,
7897 you must ensure that all code that has side effects
7898 and that is called by code compiled with this option
7899 is, itself, compiled with this option.
7900 If you do not, you might get erroneous messages from the detector.
7902 If you use functions from a library that have side-effects (such as
7903 @code{read}), you might not be able to recompile the library and
7904 specify this option. In that case, you can enable the
7905 @samp{-fprefix-function-name} option, which requests GCC to encapsulate
7906 your code and make other functions look as if they were compiled with
7907 @samp{-fcheck-memory-usage}. This is done by calling ``stubs'',
7908 which are provided by the detector. If you cannot find or build
7909 stubs for every function you call, you might have to specify
7910 @samp{-fcheck-memory-usage} without @samp{-fprefix-function-name}.
7912 If you specify this option, you can not use the @code{asm} or
7913 @code{__asm__} keywords in functions with memory checking enabled. GNU
7914 CC cannot understand what the @code{asm} statement may do, and therefore
7915 cannot generate the appropriate code, so it will reject it. However, if
7916 you specify the function attribute @code{no_check_memory_usage}
7917 (@pxref{Function Attributes}), GNU CC will disable memory checking within a
7918 function; you may use @code{asm} statements inside such functions. You
7919 may have an inline expansion of a non-checked function within a checked
7920 function; in that case GNU CC will not generate checks for the inlined
7921 function's memory accesses.
7923 If you move your @code{asm} statements to non-checked inline functions
7924 and they do access memory, you can add calls to the support code in your
7925 inline function, to indicate any reads, writes, or copies being done.
7926 These calls would be similar to those done in the stubs described above.
7928 @item -fprefix-function-name
7929 Request GCC to add a prefix to the symbols generated for function names.
7930 GCC adds a prefix to the names of functions defined as well as
7931 functions called. Code compiled with this option and code compiled
7932 without the option can't be linked together, unless stubs are used.
7934 If you compile the following code with @samp{-fprefix-function-name}
7936 extern void bar (int);
7945 GCC will compile the code as if it was written:
7947 extern void prefix_bar (int);
7951 return prefix_bar (a + 5);
7954 This option is designed to be used with @samp{-fcheck-memory-usage}.
7956 @item -finstrument-functions
7957 Generate instrumentation calls for entry and exit to functions. Just
7958 after function entry and just before function exit, the following
7959 profiling functions will be called with the address of the current
7960 function and its call site. (On some platforms,
7961 @code{__builtin_return_address} does not work beyond the current
7962 function, so the call site information may not be available to the
7963 profiling functions otherwise.)
7966 void __cyg_profile_func_enter (void *this_fn, void *call_site);
7967 void __cyg_profile_func_exit (void *this_fn, void *call_site);
7970 The first argument is the address of the start of the current function,
7971 which may be looked up exactly in the symbol table.
7973 This instrumentation is also done for functions expanded inline in other
7974 functions. The profiling calls will indicate where, conceptually, the
7975 inline function is entered and exited. This means that addressable
7976 versions of such functions must be available. If all your uses of a
7977 function are expanded inline, this may mean an additional expansion of
7978 code size. If you use @samp{extern inline} in your C code, an
7979 addressable version of such functions must be provided. (This is
7980 normally the case anyways, but if you get lucky and the optimizer always
7981 expands the functions inline, you might have gotten away without
7982 providing static copies.)
7984 A function may be given the attribute @code{no_instrument_function}, in
7985 which case this instrumentation will not be done. This can be used, for
7986 example, for the profiling functions listed above, high-priority
7987 interrupt routines, and any functions from which the profiling functions
7988 cannot safely be called (perhaps signal handlers, if the profiling
7989 routines generate output or allocate memory).
7992 Generate code to verify that you do not go beyond the boundary of the
7993 stack. You should specify this flag if you are running in an
7994 environment with multiple threads, but only rarely need to specify it in
7995 a single-threaded environment since stack overflow is automatically
7996 detected on nearly all systems if there is only one stack.
7998 Note that this switch does not actually cause checking to be done; the
7999 operating system must do that. The switch causes generation of code
8000 to ensure that the operating system sees the stack being extended.
8002 @item -fstack-limit-register=@var{reg}
8003 @itemx -fstack-limit-symbol=@var{sym}
8004 @itemx -fno-stack-limit
8005 Generate code to ensure that the stack does not grow beyond a certain value,
8006 either the value of a register or the address of a symbol. If the stack
8007 would grow beyond the value, a signal is raised. For most targets,
8008 the signal is raised before the stack overruns the boundary, so
8009 it is possible to catch the signal without taking special precautions.
8011 For instance, if the stack starts at address @samp{0x80000000} and grows
8012 downwards you can use the flags
8013 @samp{-fstack-limit-symbol=__stack_limit}
8014 @samp{-Wl,--defsym,__stack_limit=0x7ffe0000} which will enforce a stack
8017 @cindex aliasing of parameters
8018 @cindex parameters, aliased
8019 @item -fargument-alias
8020 @itemx -fargument-noalias
8021 @itemx -fargument-noalias-global
8022 Specify the possible relationships among parameters and between
8023 parameters and global data.
8025 @samp{-fargument-alias} specifies that arguments (parameters) may
8026 alias each other and may alias global storage.
8027 @samp{-fargument-noalias} specifies that arguments do not alias
8028 each other, but may alias global storage.
8029 @samp{-fargument-noalias-global} specifies that arguments do not
8030 alias each other and do not alias global storage.
8032 Each language will automatically use whatever option is required by
8033 the language standard. You should not need to use these options yourself.
8035 @item -fleading-underscore
8036 This option and its counterpart, -fno-leading-underscore, forcibly
8037 change the way C symbols are represented in the object file. One use
8038 is to help link with legacy assembly code.
8040 Be warned that you should know what you are doing when invoking this
8041 option, and that not all targets provide complete support for it.
8046 @node Environment Variables
8047 @section Environment Variables Affecting GCC
8048 @cindex environment variables
8050 @c man begin ENVIRONMENT
8052 This section describes several environment variables that affect how GCC
8053 operates. Some of them work by specifying directories or prefixes to use
8054 when searching for various kinds of files. Some are used to specify other
8055 aspects of the compilation environment.
8058 Note that you can also specify places to search using options such as
8059 @samp{-B}, @samp{-I} and @samp{-L} (@pxref{Directory Options}). These
8060 take precedence over places specified using environment variables, which
8061 in turn take precedence over those specified by the configuration of GCC.
8065 Note that you can also specify places to search using options such as
8066 @samp{-B}, @samp{-I} and @samp{-L} (@pxref{Directory Options}). These
8067 take precedence over places specified using environment variables, which
8068 in turn take precedence over those specified by the configuration of GCC.
8075 @c @itemx LC_COLLATE
8077 @c @itemx LC_MONETARY
8078 @c @itemx LC_NUMERIC
8083 @c @findex LC_COLLATE
8085 @c @findex LC_MONETARY
8086 @c @findex LC_NUMERIC
8090 These environment variables control the way that GCC uses
8091 localization information that allow GCC to work with different
8092 national conventions. GCC inspects the locale categories
8093 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
8094 so. These locale categories can be set to any value supported by your
8095 installation. A typical value is @samp{en_UK} for English in the United
8098 The @env{LC_CTYPE} environment variable specifies character
8099 classification. GCC uses it to determine the character boundaries in
8100 a string; this is needed for some multibyte encodings that contain quote
8101 and escape characters that would otherwise be interpreted as a string
8104 The @env{LC_MESSAGES} environment variable specifies the language to
8105 use in diagnostic messages.
8107 If the @env{LC_ALL} environment variable is set, it overrides the value
8108 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
8109 and @env{LC_MESSAGES} default to the value of the @env{LANG}
8110 environment variable. If none of these variables are set, GCC
8111 defaults to traditional C English behavior.
8115 If @env{TMPDIR} is set, it specifies the directory to use for temporary
8116 files. GCC uses temporary files to hold the output of one stage of
8117 compilation which is to be used as input to the next stage: for example,
8118 the output of the preprocessor, which is the input to the compiler
8121 @item GCC_EXEC_PREFIX
8122 @findex GCC_EXEC_PREFIX
8123 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
8124 names of the subprograms executed by the compiler. No slash is added
8125 when this prefix is combined with the name of a subprogram, but you can
8126 specify a prefix that ends with a slash if you wish.
8128 If @env{GCC_EXEC_PREFIX} is not set, GNU CC will attempt to figure out
8129 an appropriate prefix to use based on the pathname it was invoked with.
8131 If GCC cannot find the subprogram using the specified prefix, it
8132 tries looking in the usual places for the subprogram.
8134 The default value of @env{GCC_EXEC_PREFIX} is
8135 @file{@var{prefix}/lib/gcc-lib/} where @var{prefix} is the value
8136 of @code{prefix} when you ran the @file{configure} script.
8138 Other prefixes specified with @samp{-B} take precedence over this prefix.
8140 This prefix is also used for finding files such as @file{crt0.o} that are
8143 In addition, the prefix is used in an unusual way in finding the
8144 directories to search for header files. For each of the standard
8145 directories whose name normally begins with @samp{/usr/local/lib/gcc-lib}
8146 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
8147 replacing that beginning with the specified prefix to produce an
8148 alternate directory name. Thus, with @samp{-Bfoo/}, GCC will search
8149 @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}.
8150 These alternate directories are searched first; the standard directories
8154 @findex COMPILER_PATH
8155 The value of @env{COMPILER_PATH} is a colon-separated list of
8156 directories, much like @env{PATH}. GCC tries the directories thus
8157 specified when searching for subprograms, if it can't find the
8158 subprograms using @env{GCC_EXEC_PREFIX}.
8161 @findex LIBRARY_PATH
8162 The value of @env{LIBRARY_PATH} is a colon-separated list of
8163 directories, much like @env{PATH}. When configured as a native compiler,
8164 GCC tries the directories thus specified when searching for special
8165 linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking
8166 using GCC also uses these directories when searching for ordinary
8167 libraries for the @samp{-l} option (but directories specified with
8168 @samp{-L} come first).
8170 @item C_INCLUDE_PATH
8171 @itemx CPLUS_INCLUDE_PATH
8172 @itemx OBJC_INCLUDE_PATH
8173 @findex C_INCLUDE_PATH
8174 @findex CPLUS_INCLUDE_PATH
8175 @findex OBJC_INCLUDE_PATH
8176 @c @itemx OBJCPLUS_INCLUDE_PATH
8177 These environment variables pertain to particular languages. Each
8178 variable's value is a colon-separated list of directories, much like
8179 @env{PATH}. When GCC searches for header files, it tries the
8180 directories listed in the variable for the language you are using, after
8181 the directories specified with @samp{-I} but before the standard header
8184 @item DEPENDENCIES_OUTPUT
8185 @findex DEPENDENCIES_OUTPUT
8186 @cindex dependencies for make as output
8187 If this variable is set, its value specifies how to output dependencies
8188 for Make based on the header files processed by the compiler. This
8189 output looks much like the output from the @samp{-M} option
8190 (@pxref{Preprocessor Options}), but it goes to a separate file, and is
8191 in addition to the usual results of compilation.
8193 The value of @env{DEPENDENCIES_OUTPUT} can be just a file name, in
8194 which case the Make rules are written to that file, guessing the target
8195 name from the source file name. Or the value can have the form
8196 @samp{@var{file} @var{target}}, in which case the rules are written to
8197 file @var{file} using @var{target} as the target name.
8201 @cindex locale definition
8202 This variable is used to pass locale information to the compiler. One way in
8203 which this information is used is to determine the character set to be used
8204 when character literals, string literals and comments are parsed in C and C++.
8205 When the compiler is configured to allow multibyte characters,
8206 the following values for @env{LANG} are recognized:
8210 Recognize JIS characters.
8212 Recognize SJIS characters.
8214 Recognize EUCJP characters.
8217 If @env{LANG} is not defined, or if it has some other value, then the
8218 compiler will use mblen and mbtowc as defined by the default locale to
8219 recognize and translate multibyte characters.
8224 @node Running Protoize
8225 @section Running Protoize
8227 The program @code{protoize} is an optional part of GNU C. You can use
8228 it to add prototypes to a program, thus converting the program to ISO
8229 C in one respect. The companion program @code{unprotoize} does the
8230 reverse: it removes argument types from any prototypes that are found.
8232 When you run these programs, you must specify a set of source files as
8233 command line arguments. The conversion programs start out by compiling
8234 these files to see what functions they define. The information gathered
8235 about a file @var{foo} is saved in a file named @file{@var{foo}.X}.
8237 After scanning comes actual conversion. The specified files are all
8238 eligible to be converted; any files they include (whether sources or
8239 just headers) are eligible as well.
8241 But not all the eligible files are converted. By default,
8242 @code{protoize} and @code{unprotoize} convert only source and header
8243 files in the current directory. You can specify additional directories
8244 whose files should be converted with the @samp{-d @var{directory}}
8245 option. You can also specify particular files to exclude with the
8246 @samp{-x @var{file}} option. A file is converted if it is eligible, its
8247 directory name matches one of the specified directory names, and its
8248 name within the directory has not been excluded.
8250 Basic conversion with @code{protoize} consists of rewriting most
8251 function definitions and function declarations to specify the types of
8252 the arguments. The only ones not rewritten are those for varargs
8255 @code{protoize} optionally inserts prototype declarations at the
8256 beginning of the source file, to make them available for any calls that
8257 precede the function's definition. Or it can insert prototype
8258 declarations with block scope in the blocks where undeclared functions
8261 Basic conversion with @code{unprotoize} consists of rewriting most
8262 function declarations to remove any argument types, and rewriting
8263 function definitions to the old-style pre-ISO form.
8265 Both conversion programs print a warning for any function declaration or
8266 definition that they can't convert. You can suppress these warnings
8269 The output from @code{protoize} or @code{unprotoize} replaces the
8270 original source file. The original file is renamed to a name ending
8271 with @samp{.save} (for DOS, the saved filename ends in @samp{.sav}
8272 without the original @samp{.c} suffix). If the @samp{.save} (@samp{.sav}
8273 for DOS) file already exists, then the source file is simply discarded.
8275 @code{protoize} and @code{unprotoize} both depend on GCC itself to
8276 scan the program and collect information about the functions it uses.
8277 So neither of these programs will work until GCC is installed.
8279 Here is a table of the options you can use with @code{protoize} and
8280 @code{unprotoize}. Each option works with both programs unless
8284 @item -B @var{directory}
8285 Look for the file @file{SYSCALLS.c.X} in @var{directory}, instead of the
8286 usual directory (normally @file{/usr/local/lib}). This file contains
8287 prototype information about standard system functions. This option
8288 applies only to @code{protoize}.
8290 @item -c @var{compilation-options}
8291 Use @var{compilation-options} as the options when running @code{gcc} to
8292 produce the @samp{.X} files. The special option @samp{-aux-info} is
8293 always passed in addition, to tell @code{gcc} to write a @samp{.X} file.
8295 Note that the compilation options must be given as a single argument to
8296 @code{protoize} or @code{unprotoize}. If you want to specify several
8297 @code{gcc} options, you must quote the entire set of compilation options
8298 to make them a single word in the shell.
8300 There are certain @code{gcc} arguments that you cannot use, because they
8301 would produce the wrong kind of output. These include @samp{-g},
8302 @samp{-O}, @samp{-c}, @samp{-S}, and @samp{-o} If you include these in
8303 the @var{compilation-options}, they are ignored.
8306 Rename files to end in @samp{.C} (@samp{.cc} for DOS-based file
8307 systems) instead of @samp{.c}. This is convenient if you are converting
8308 a C program to C++. This option applies only to @code{protoize}.
8311 Add explicit global declarations. This means inserting explicit
8312 declarations at the beginning of each source file for each function
8313 that is called in the file and was not declared. These declarations
8314 precede the first function definition that contains a call to an
8315 undeclared function. This option applies only to @code{protoize}.
8317 @item -i @var{string}
8318 Indent old-style parameter declarations with the string @var{string}.
8319 This option applies only to @code{protoize}.
8321 @code{unprotoize} converts prototyped function definitions to old-style
8322 function definitions, where the arguments are declared between the
8323 argument list and the initial @samp{@{}. By default, @code{unprotoize}
8324 uses five spaces as the indentation. If you want to indent with just
8325 one space instead, use @samp{-i " "}.
8328 Keep the @samp{.X} files. Normally, they are deleted after conversion
8332 Add explicit local declarations. @code{protoize} with @samp{-l} inserts
8333 a prototype declaration for each function in each block which calls the
8334 function without any declaration. This option applies only to
8338 Make no real changes. This mode just prints information about the conversions
8339 that would have been done without @samp{-n}.
8342 Make no @samp{.save} files. The original files are simply deleted.
8343 Use this option with caution.
8345 @item -p @var{program}
8346 Use the program @var{program} as the compiler. Normally, the name
8350 Work quietly. Most warnings are suppressed.
8353 Print the version number, just like @samp{-v} for @code{gcc}.
8356 If you need special compiler options to compile one of your program's
8357 source files, then you should generate that file's @samp{.X} file
8358 specially, by running @code{gcc} on that source file with the
8359 appropriate options and the option @samp{-aux-info}. Then run
8360 @code{protoize} on the entire set of files. @code{protoize} will use
8361 the existing @samp{.X} file because it is newer than the source file.
8365 gcc -Dfoo=bar file1.c -aux-info
8370 You need to include the special files along with the rest in the
8371 @code{protoize} command, even though their @samp{.X} files already
8372 exist, because otherwise they won't get converted.
8374 @xref{Protoize Caveats}, for more information on how to use
8375 @code{protoize} successfully.