1 @c Copyright (C) 1988-2016 Free Software Foundation, Inc.
2 @c This is part of the GCC manual.
3 @c For copying conditions, see the file gcc.texi.
10 @c man begin COPYRIGHT
11 Copyright @copyright{} 1988-2016 Free Software Foundation, Inc.
13 Permission is granted to copy, distribute and/or modify this document
14 under the terms of the GNU Free Documentation License, Version 1.3 or
15 any later version published by the Free Software Foundation; with the
16 Invariant Sections being ``GNU General Public License'' and ``Funding
17 Free Software'', the Front-Cover texts being (a) (see below), and with
18 the Back-Cover Texts being (b) (see below). A copy of the license is
19 included in the gfdl(7) man page.
21 (a) The FSF's Front-Cover Text is:
25 (b) The FSF's Back-Cover Text is:
27 You have freedom to copy and modify this GNU Manual, like GNU
28 software. Copies published by the Free Software Foundation raise
29 funds for GNU development.
31 @c Set file name and title for the man page.
33 @settitle GNU project C and C++ compiler
35 gcc [@option{-c}|@option{-S}|@option{-E}] [@option{-std=}@var{standard}]
36 [@option{-g}] [@option{-pg}] [@option{-O}@var{level}]
37 [@option{-W}@var{warn}@dots{}] [@option{-Wpedantic}]
38 [@option{-I}@var{dir}@dots{}] [@option{-L}@var{dir}@dots{}]
39 [@option{-D}@var{macro}[=@var{defn}]@dots{}] [@option{-U}@var{macro}]
40 [@option{-f}@var{option}@dots{}] [@option{-m}@var{machine-option}@dots{}]
41 [@option{-o} @var{outfile}] [@@@var{file}] @var{infile}@dots{}
43 Only the most useful options are listed here; see below for the
44 remainder. @command{g++} accepts mostly the same options as @command{gcc}.
47 gpl(7), gfdl(7), fsf-funding(7),
48 cpp(1), gcov(1), as(1), ld(1), gdb(1), adb(1), dbx(1), sdb(1)
49 and the Info entries for @file{gcc}, @file{cpp}, @file{as},
50 @file{ld}, @file{binutils} and @file{gdb}.
53 For instructions on reporting bugs, see
57 See the Info entry for @command{gcc}, or
58 @w{@uref{http://gcc.gnu.org/onlinedocs/gcc/Contributors.html}},
59 for contributors to GCC@.
64 @chapter GCC Command Options
65 @cindex GCC command options
66 @cindex command options
67 @cindex options, GCC command
69 @c man begin DESCRIPTION
70 When you invoke GCC, it normally does preprocessing, compilation,
71 assembly and linking. The ``overall options'' allow you to stop this
72 process at an intermediate stage. For example, the @option{-c} option
73 says not to run the linker. Then the output consists of object files
74 output by the assembler.
75 @xref{Overall Options,,Options Controlling the Kind of Output}.
77 Other options are passed on to one or more stages of processing. Some options
78 control the preprocessor and others the compiler itself. Yet other
79 options control the assembler and linker; most of these are not
80 documented here, since you rarely need to use any of them.
82 @cindex C compilation options
83 Most of the command-line options that you can use with GCC are useful
84 for C programs; when an option is only useful with another language
85 (usually C++), the explanation says so explicitly. If the description
86 for a particular option does not mention a source language, you can use
87 that option with all supported languages.
89 @cindex cross compiling
90 @cindex specifying machine version
91 @cindex specifying compiler version and target machine
92 @cindex compiler version, specifying
93 @cindex target machine, specifying
94 The usual way to run GCC is to run the executable called @command{gcc}, or
95 @command{@var{machine}-gcc} when cross-compiling, or
96 @command{@var{machine}-gcc-@var{version}} to run a specific version of GCC.
97 When you compile C++ programs, you should invoke GCC as @command{g++}
98 instead. @xref{Invoking G++,,Compiling C++ Programs},
99 for information about the differences in behavior between @command{gcc}
100 and @code{g++} when compiling C++ programs.
102 @cindex grouping options
103 @cindex options, grouping
104 The @command{gcc} program accepts options and file names as operands. Many
105 options have multi-letter names; therefore multiple single-letter options
106 may @emph{not} be grouped: @option{-dv} is very different from @w{@samp{-d
109 @cindex order of options
110 @cindex options, order
111 You can mix options and other arguments. For the most part, the order
112 you use doesn't matter. Order does matter when you use several
113 options of the same kind; for example, if you specify @option{-L} more
114 than once, the directories are searched in the order specified. Also,
115 the placement of the @option{-l} option is significant.
117 Many options have long names starting with @samp{-f} or with
118 @samp{-W}---for example,
119 @option{-fmove-loop-invariants}, @option{-Wformat} and so on. Most of
120 these have both positive and negative forms; the negative form of
121 @option{-ffoo} is @option{-fno-foo}. This manual documents
122 only one of these two forms, whichever one is not the default.
126 @xref{Option Index}, for an index to GCC's options.
129 * Option Summary:: Brief list of all options, without explanations.
130 * Overall Options:: Controlling the kind of output:
131 an executable, object files, assembler files,
132 or preprocessed source.
133 * Invoking G++:: Compiling C++ programs.
134 * C Dialect Options:: Controlling the variant of C language compiled.
135 * C++ Dialect Options:: Variations on C++.
136 * Objective-C and Objective-C++ Dialect Options:: Variations on Objective-C
138 * Diagnostic Message Formatting Options:: Controlling how diagnostics should
140 * Warning Options:: How picky should the compiler be?
141 * Debugging Options:: Producing debuggable code.
142 * Optimize Options:: How much optimization?
143 * Instrumentation Options:: Enabling profiling and extra run-time error checking.
144 * Preprocessor Options:: Controlling header files and macro definitions.
145 Also, getting dependency information for Make.
146 * Assembler Options:: Passing options to the assembler.
147 * Link Options:: Specifying libraries and so on.
148 * Directory Options:: Where to find header files and libraries.
149 Where to find the compiler executable files.
150 * Code Gen Options:: Specifying conventions for function calls, data layout
152 * Developer Options:: Printing GCC configuration info, statistics, and
154 * Submodel Options:: Target-specific options, such as compiling for a
155 specific processor variant.
156 * Spec Files:: How to pass switches to sub-processes.
157 * Environment Variables:: Env vars that affect GCC.
158 * Precompiled Headers:: Compiling a header once, and using it many times.
164 @section Option Summary
166 Here is a summary of all the options, grouped by type. Explanations are
167 in the following sections.
170 @item Overall Options
171 @xref{Overall Options,,Options Controlling the Kind of Output}.
172 @gccoptlist{-c -S -E -o @var{file} -x @var{language} @gol
173 -v -### --help@r{[}=@var{class}@r{[},@dots{}@r{]]} --target-help --version @gol
174 -pass-exit-codes -pipe -specs=@var{file} -wrapper @gol
175 @@@var{file} -fplugin=@var{file} -fplugin-arg-@var{name}=@var{arg} @gol
176 -fdump-ada-spec@r{[}-slim@r{]} -fada-spec-parent=@var{unit} -fdump-go-spec=@var{file}}
178 @item C Language Options
179 @xref{C Dialect Options,,Options Controlling C Dialect}.
180 @gccoptlist{-ansi -std=@var{standard} -fgnu89-inline @gol
181 -aux-info @var{filename} -fallow-parameterless-variadic-functions @gol
182 -fno-asm -fno-builtin -fno-builtin-@var{function} @gol
183 -fhosted -ffreestanding -fopenacc -fopenmp -fopenmp-simd @gol
184 -fms-extensions -fplan9-extensions -fsso-struct=@var{endianness}
185 -fallow-single-precision -fcond-mismatch -flax-vector-conversions @gol
186 -fsigned-bitfields -fsigned-char @gol
187 -funsigned-bitfields -funsigned-char @gol
188 -trigraphs -traditional -traditional-cpp}
190 @item C++ Language Options
191 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}.
192 @gccoptlist{-fabi-version=@var{n} -fno-access-control @gol
193 -faligned-new=@var{n} -fargs-in-order=@var{n} -fcheck-new @gol
194 -fconstexpr-depth=@var{n} -fconstexpr-loop-limit=@var{n} @gol
195 -ffriend-injection @gol
196 -fno-elide-constructors @gol
197 -fno-enforce-eh-specs @gol
198 -ffor-scope -fno-for-scope -fno-gnu-keywords @gol
199 -fno-implicit-templates @gol
200 -fno-implicit-inline-templates @gol
201 -fno-implement-inlines -fms-extensions @gol
202 -fnew-inheriting-ctors @gol
203 -fno-nonansi-builtins -fnothrow-opt -fno-operator-names @gol
204 -fno-optional-diags -fpermissive @gol
205 -fno-pretty-templates @gol
206 -frepo -fno-rtti -fsized-deallocation @gol
207 -ftemplate-backtrace-limit=@var{n} @gol
208 -ftemplate-depth=@var{n} @gol
209 -fno-threadsafe-statics -fuse-cxa-atexit @gol
210 -fno-weak -nostdinc++ @gol
211 -fvisibility-inlines-hidden @gol
212 -fvisibility-ms-compat @gol
213 -fext-numeric-literals @gol
214 -Wabi=@var{n} -Wabi-tag -Wconversion-null -Wctor-dtor-privacy @gol
215 -Wdelete-non-virtual-dtor -Wliteral-suffix -Wmultiple-inheritance @gol
216 -Wnamespaces -Wnarrowing @gol
217 -Wnoexcept -Wnon-virtual-dtor -Wreorder -Wregister @gol
218 -Weffc++ -Wstrict-null-sentinel -Wtemplates @gol
219 -Wno-non-template-friend -Wold-style-cast @gol
220 -Woverloaded-virtual -Wno-pmf-conversions @gol
221 -Wsign-promo -Wvirtual-inheritance}
223 @item Objective-C and Objective-C++ Language Options
224 @xref{Objective-C and Objective-C++ Dialect Options,,Options Controlling
225 Objective-C and Objective-C++ Dialects}.
226 @gccoptlist{-fconstant-string-class=@var{class-name} @gol
227 -fgnu-runtime -fnext-runtime @gol
228 -fno-nil-receivers @gol
229 -fobjc-abi-version=@var{n} @gol
230 -fobjc-call-cxx-cdtors @gol
231 -fobjc-direct-dispatch @gol
232 -fobjc-exceptions @gol
235 -fobjc-std=objc1 @gol
236 -fno-local-ivars @gol
237 -fivar-visibility=@r{[}public@r{|}protected@r{|}private@r{|}package@r{]} @gol
238 -freplace-objc-classes @gol
241 -Wassign-intercept @gol
242 -Wno-protocol -Wselector @gol
243 -Wstrict-selector-match @gol
244 -Wundeclared-selector}
246 @item Diagnostic Message Formatting Options
247 @xref{Diagnostic Message Formatting Options,,Options to Control Diagnostic Messages Formatting}.
248 @gccoptlist{-fmessage-length=@var{n} @gol
249 -fdiagnostics-show-location=@r{[}once@r{|}every-line@r{]} @gol
250 -fdiagnostics-color=@r{[}auto@r{|}never@r{|}always@r{]} @gol
251 -fno-diagnostics-show-option -fno-diagnostics-show-caret @gol
252 -fdiagnostics-parseable-fixits -fdiagnostics-generate-patch}
254 @item Warning Options
255 @xref{Warning Options,,Options to Request or Suppress Warnings}.
256 @gccoptlist{-fsyntax-only -fmax-errors=@var{n} -Wpedantic @gol
257 -pedantic-errors @gol
258 -w -Wextra -Wall -Waddress -Waggregate-return @gol
259 -Walloca -Walloca-larger-than=@var{n} @gol
260 -Wno-aggressive-loop-optimizations -Warray-bounds -Warray-bounds=@var{n} @gol
261 -Wno-attributes -Wbool-compare -Wbool-operation @gol
262 -Wno-builtin-macro-redefined -Wc90-c99-compat -Wc99-c11-compat @gol
263 -Wc++-compat -Wc++11-compat -Wc++14-compat -Wcast-align -Wcast-qual @gol
264 -Wchar-subscripts -Wclobbered -Wcomment -Wconditionally-supported @gol
265 -Wconversion -Wcoverage-mismatch -Wno-cpp -Wdangling-else -Wdate-time @gol
266 -Wdelete-incomplete @gol
267 -Wno-deprecated -Wno-deprecated-declarations -Wno-designated-init @gol
268 -Wdisabled-optimization @gol
269 -Wno-discarded-qualifiers -Wno-discarded-array-qualifiers @gol
270 -Wno-div-by-zero -Wdouble-promotion -Wduplicated-cond @gol
271 -Wempty-body -Wenum-compare -Wno-endif-labels @gol
272 -Werror -Werror=* -Wfatal-errors -Wfloat-equal -Wformat -Wformat=2 @gol
273 -Wno-format-contains-nul -Wno-format-extra-args -Wformat-length=@var{n} @gol
274 -Wformat-nonliteral @gol
275 -Wformat-security -Wformat-signedness -Wformat-y2k -Wframe-address @gol
276 -Wframe-larger-than=@var{len} -Wno-free-nonheap-object -Wjump-misses-init @gol
277 -Wignored-qualifiers -Wignored-attributes -Wincompatible-pointer-types @gol
278 -Wimplicit -Wimplicit-fallthrough -Wimplicit-fallthrough=@var{n} @gol
279 -Wimplicit-function-declaration -Wimplicit-int @gol
280 -Winit-self -Winline -Wno-int-conversion -Wint-in-bool-context @gol
281 -Wno-int-to-pointer-cast -Winvalid-memory-model -Wno-invalid-offsetof @gol
282 -Winvalid-pch -Wlarger-than=@var{len} @gol
283 -Wlogical-op -Wlogical-not-parentheses -Wlong-long @gol
284 -Wmain -Wmaybe-uninitialized -Wmemset-elt-size -Wmemset-transposed-args @gol
285 -Wmisleading-indentation -Wmissing-braces @gol
286 -Wmissing-field-initializers -Wmissing-include-dirs @gol
287 -Wno-multichar -Wnonnull -Wnonnull-compare @gol
288 -Wnormalized=@r{[}none@r{|}id@r{|}nfc@r{|}nfkc@r{]} @gol
289 -Wnull-dereference -Wodr -Wno-overflow -Wopenmp-simd @gol
290 -Woverride-init-side-effects -Woverlength-strings @gol
291 -Wpacked -Wpacked-bitfield-compat -Wpadded @gol
292 -Wparentheses -Wno-pedantic-ms-format @gol
293 -Wplacement-new -Wplacement-new=@var{n} @gol
294 -Wpointer-arith -Wno-pointer-to-int-cast @gol
295 -Wno-pragmas -Wredundant-decls -Wno-return-local-addr @gol
296 -Wreturn-type -Wsequence-point -Wshadow -Wno-shadow-ivar @gol
297 -Wshadow=global, -Wshadow=local, -Wshadow=compatible-local @gol
298 -Wshift-overflow -Wshift-overflow=@var{n} @gol
299 -Wshift-count-negative -Wshift-count-overflow -Wshift-negative-value @gol
300 -Wsign-compare -Wsign-conversion -Wfloat-conversion @gol
301 -Wno-scalar-storage-order @gol
302 -Wsizeof-pointer-memaccess -Wsizeof-array-argument @gol
303 -Wstack-protector -Wstack-usage=@var{len} -Wstrict-aliasing @gol
304 -Wstrict-aliasing=n -Wstrict-overflow -Wstrict-overflow=@var{n} @gol
305 -Wsuggest-attribute=@r{[}pure@r{|}const@r{|}noreturn@r{|}format@r{]} @gol
306 -Wsuggest-final-types @gol -Wsuggest-final-methods -Wsuggest-override @gol
307 -Wmissing-format-attribute -Wsubobject-linkage @gol
308 -Wswitch -Wswitch-bool -Wswitch-default -Wswitch-enum @gol
309 -Wswitch-unreachable -Wsync-nand @gol
310 -Wsystem-headers -Wtautological-compare -Wtrampolines -Wtrigraphs @gol
311 -Wtype-limits -Wundef @gol
312 -Wuninitialized -Wunknown-pragmas -Wunsafe-loop-optimizations @gol
313 -Wunsuffixed-float-constants -Wunused -Wunused-function @gol
314 -Wunused-label -Wunused-local-typedefs -Wunused-parameter @gol
315 -Wno-unused-result -Wunused-value @gol -Wunused-variable @gol
316 -Wunused-const-variable -Wunused-const-variable=@var{n} @gol
317 -Wunused-but-set-parameter -Wunused-but-set-variable @gol
318 -Wuseless-cast -Wvariadic-macros -Wvector-operation-performance @gol
319 -Wvla -Wvla-larger-than=@var{n} -Wvolatile-register-var -Wwrite-strings @gol
320 -Wzero-as-null-pointer-constant -Whsa}
322 @item C and Objective-C-only Warning Options
323 @gccoptlist{-Wbad-function-cast -Wmissing-declarations @gol
324 -Wmissing-parameter-type -Wmissing-prototypes -Wnested-externs @gol
325 -Wold-style-declaration -Wold-style-definition @gol
326 -Wstrict-prototypes -Wtraditional -Wtraditional-conversion @gol
327 -Wdeclaration-after-statement -Wpointer-sign}
329 @item Debugging Options
330 @xref{Debugging Options,,Options for Debugging Your Program}.
331 @gccoptlist{-g -g@var{level} -gcoff -gdwarf -gdwarf-@var{version} @gol
332 -ggdb -grecord-gcc-switches -gno-record-gcc-switches @gol
333 -gstabs -gstabs+ -gstrict-dwarf -gno-strict-dwarf @gol
334 -gvms -gxcoff -gxcoff+ -gz@r{[}=@var{type}@r{]} @gol
335 -fdebug-prefix-map=@var{old}=@var{new} -fdebug-types-section @gol
336 -feliminate-dwarf2-dups -fno-eliminate-unused-debug-types @gol
337 -femit-struct-debug-baseonly -femit-struct-debug-reduced @gol
338 -femit-struct-debug-detailed@r{[}=@var{spec-list}@r{]} @gol
339 -feliminate-unused-debug-symbols -femit-class-debug-always @gol
340 -fno-merge-debug-strings -fno-dwarf2-cfi-asm @gol
341 -fvar-tracking -fvar-tracking-assignments}
343 @item Optimization Options
344 @xref{Optimize Options,,Options that Control Optimization}.
345 @gccoptlist{-faggressive-loop-optimizations -falign-functions[=@var{n}] @gol
346 -falign-jumps[=@var{n}] @gol
347 -falign-labels[=@var{n}] -falign-loops[=@var{n}] @gol
348 -fassociative-math -fauto-profile -fauto-profile[=@var{path}] @gol
349 -fauto-inc-dec -fbranch-probabilities @gol
350 -fbranch-target-load-optimize -fbranch-target-load-optimize2 @gol
351 -fbtr-bb-exclusive -fcaller-saves @gol
352 -fcombine-stack-adjustments -fconserve-stack @gol
353 -fcompare-elim -fcprop-registers -fcrossjumping @gol
354 -fcse-follow-jumps -fcse-skip-blocks -fcx-fortran-rules @gol
355 -fcx-limited-range @gol
356 -fdata-sections -fdce -fdelayed-branch @gol
357 -fdelete-null-pointer-checks -fdevirtualize -fdevirtualize-speculatively @gol
358 -fdevirtualize-at-ltrans -fdse @gol
359 -fearly-inlining -fipa-sra -fexpensive-optimizations -ffat-lto-objects @gol
360 -ffast-math -ffinite-math-only -ffloat-store -fexcess-precision=@var{style} @gol
361 -fforward-propagate -ffp-contract=@var{style} -ffunction-sections @gol
362 -fgcse -fgcse-after-reload -fgcse-las -fgcse-lm -fgraphite-identity @gol
363 -fgcse-sm -fhoist-adjacent-loads -fif-conversion @gol
364 -fif-conversion2 -findirect-inlining @gol
365 -finline-functions -finline-functions-called-once -finline-limit=@var{n} @gol
366 -finline-small-functions -fipa-cp -fipa-cp-clone -fipa-cp-alignment -fipa-bit-cp @gol
367 -fipa-pta -fipa-profile -fipa-pure-const -fipa-reference -fipa-icf @gol
368 -fira-algorithm=@var{algorithm} @gol
369 -fira-region=@var{region} -fira-hoist-pressure @gol
370 -fira-loop-pressure -fno-ira-share-save-slots @gol
371 -fno-ira-share-spill-slots @gol
372 -fisolate-erroneous-paths-dereference -fisolate-erroneous-paths-attribute @gol
373 -fivopts -fkeep-inline-functions -fkeep-static-functions @gol
374 -fkeep-static-consts -flive-range-shrinkage @gol
375 -floop-block -floop-interchange -floop-strip-mine @gol
376 -floop-unroll-and-jam -floop-nest-optimize @gol
377 -floop-parallelize-all -flra-remat -flto -flto-compression-level @gol
378 -flto-partition=@var{alg} -fmerge-all-constants @gol
379 -fmerge-constants -fmodulo-sched -fmodulo-sched-allow-regmoves @gol
380 -fmove-loop-invariants -fno-branch-count-reg @gol
381 -fno-defer-pop -fno-fp-int-builtin-inexact -fno-function-cse @gol
382 -fno-guess-branch-probability -fno-inline -fno-math-errno -fno-peephole @gol
383 -fno-peephole2 -fno-sched-interblock -fno-sched-spec -fno-signed-zeros @gol
384 -fno-toplevel-reorder -fno-trapping-math -fno-zero-initialized-in-bss @gol
385 -fomit-frame-pointer -foptimize-sibling-calls @gol
386 -fpartial-inlining -fpeel-loops -fpredictive-commoning @gol
387 -fprefetch-loop-arrays -fprintf-return-value @gol
388 -fprofile-correction @gol
389 -fprofile-use -fprofile-use=@var{path} -fprofile-values @gol
390 -fprofile-reorder-functions @gol
391 -freciprocal-math -free -frename-registers -freorder-blocks @gol
392 -freorder-blocks-algorithm=@var{algorithm} @gol
393 -freorder-blocks-and-partition -freorder-functions @gol
394 -frerun-cse-after-loop -freschedule-modulo-scheduled-loops @gol
395 -frounding-math -fsched2-use-superblocks -fsched-pressure @gol
396 -fsched-spec-load -fsched-spec-load-dangerous @gol
397 -fsched-stalled-insns-dep[=@var{n}] -fsched-stalled-insns[=@var{n}] @gol
398 -fsched-group-heuristic -fsched-critical-path-heuristic @gol
399 -fsched-spec-insn-heuristic -fsched-rank-heuristic @gol
400 -fsched-last-insn-heuristic -fsched-dep-count-heuristic @gol
401 -fschedule-fusion @gol
402 -fschedule-insns -fschedule-insns2 -fsection-anchors @gol
403 -fselective-scheduling -fselective-scheduling2 @gol
404 -fsel-sched-pipelining -fsel-sched-pipelining-outer-loops @gol
405 -fsemantic-interposition -fshrink-wrap -fshrink-wrap-separate @gol
406 -fsignaling-nans @gol
407 -fsingle-precision-constant -fsplit-ivs-in-unroller -fsplit-loops@gol
409 -fsplit-wide-types -fssa-backprop -fssa-phiopt @gol
410 -fstdarg-opt -fstore-merging -fstrict-aliasing @gol
411 -fstrict-overflow -fthread-jumps -ftracer -ftree-bit-ccp @gol
412 -ftree-builtin-call-dce -ftree-ccp -ftree-ch @gol
413 -ftree-coalesce-vars -ftree-copy-prop -ftree-dce -ftree-dominator-opts @gol
414 -ftree-dse -ftree-forwprop -ftree-fre -fcode-hoisting -ftree-loop-if-convert @gol
415 -ftree-loop-if-convert-stores -ftree-loop-im @gol
416 -ftree-phiprop -ftree-loop-distribution -ftree-loop-distribute-patterns @gol
417 -ftree-loop-ivcanon -ftree-loop-linear -ftree-loop-optimize @gol
418 -ftree-loop-vectorize @gol
419 -ftree-parallelize-loops=@var{n} -ftree-pre -ftree-partial-pre -ftree-pta @gol
420 -ftree-reassoc -ftree-sink -ftree-slsr -ftree-sra @gol
421 -ftree-switch-conversion -ftree-tail-merge @gol
422 -ftree-ter -ftree-vectorize -ftree-vrp -funconstrained-commons @gol
423 -funit-at-a-time -funroll-all-loops -funroll-loops @gol
424 -funsafe-math-optimizations -funswitch-loops @gol
425 -fipa-ra -fvariable-expansion-in-unroller -fvect-cost-model -fvpt @gol
426 -fweb -fwhole-program -fwpa -fuse-linker-plugin @gol
427 --param @var{name}=@var{value}
428 -O -O0 -O1 -O2 -O3 -Os -Ofast -Og}
430 @item Program Instrumentation Options
431 @xref{Instrumentation Options,,Program Instrumentation Options}.
432 @gccoptlist{-p -pg -fprofile-arcs --coverage -ftest-coverage @gol
433 -fprofile-dir=@var{path} -fprofile-generate -fprofile-generate=@var{path} @gol
434 -fsanitize=@var{style} -fsanitize-recover -fsanitize-recover=@var{style} @gol
435 -fasan-shadow-offset=@var{number} -fsanitize-sections=@var{s1},@var{s2},... @gol
436 -fsanitize-undefined-trap-on-error -fbounds-check @gol
437 -fcheck-pointer-bounds -fchkp-check-incomplete-type @gol
438 -fchkp-first-field-has-own-bounds -fchkp-narrow-bounds @gol
439 -fchkp-narrow-to-innermost-array -fchkp-optimize @gol
440 -fchkp-use-fast-string-functions -fchkp-use-nochk-string-functions @gol
441 -fchkp-use-static-bounds -fchkp-use-static-const-bounds @gol
442 -fchkp-treat-zero-dynamic-size-as-infinite -fchkp-check-read @gol
443 -fchkp-check-read -fchkp-check-write -fchkp-store-bounds @gol
444 -fchkp-instrument-calls -fchkp-instrument-marked-only @gol
445 -fchkp-use-wrappers @gol
446 -fstack-protector -fstack-protector-all -fstack-protector-strong @gol
447 -fstack-protector-explicit -fstack-check @gol
448 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
449 -fno-stack-limit -fsplit-stack @gol
450 -fvtable-verify=@r{[}std@r{|}preinit@r{|}none@r{]} @gol
451 -fvtv-counts -fvtv-debug @gol
452 -finstrument-functions @gol
453 -finstrument-functions-exclude-function-list=@var{sym},@var{sym},@dots{} @gol
454 -finstrument-functions-exclude-file-list=@var{file},@var{file},@dots{}}
456 @item Preprocessor Options
457 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
458 @gccoptlist{-A@var{question}=@var{answer} @gol
459 -A-@var{question}@r{[}=@var{answer}@r{]} @gol
460 -C -dD -dI -dM -dN @gol
461 -D@var{macro}@r{[}=@var{defn}@r{]} -E -H @gol
462 -idirafter @var{dir} @gol
463 -include @var{file} -imacros @var{file} @gol
464 -iprefix @var{file} -iwithprefix @var{dir} @gol
465 -iwithprefixbefore @var{dir} -isystem @var{dir} @gol
466 -imultilib @var{dir} -isysroot @var{dir} @gol
467 -M -MM -MF -MG -MP -MQ -MT -nostdinc @gol
468 -P -fdebug-cpp -ftrack-macro-expansion -fworking-directory @gol
469 -remap -trigraphs -undef -U@var{macro} @gol
470 -Wp,@var{option} -Xpreprocessor @var{option} -no-integrated-cpp}
472 @item Assembler Option
473 @xref{Assembler Options,,Passing Options to the Assembler}.
474 @gccoptlist{-Wa,@var{option} -Xassembler @var{option}}
477 @xref{Link Options,,Options for Linking}.
478 @gccoptlist{@var{object-file-name} -fuse-ld=@var{linker} -l@var{library} @gol
479 -nostartfiles -nodefaultlibs -nostdlib -pie -rdynamic @gol
480 -s -static -static-libgcc -static-libstdc++ @gol
481 -static-libasan -static-libtsan -static-liblsan -static-libubsan @gol
482 -static-libmpx -static-libmpxwrappers @gol
483 -shared -shared-libgcc -symbolic @gol
484 -T @var{script} -Wl,@var{option} -Xlinker @var{option} @gol
485 -u @var{symbol} -z @var{keyword}}
487 @item Directory Options
488 @xref{Directory Options,,Options for Directory Search}.
489 @gccoptlist{-B@var{prefix} -I@var{dir} -iplugindir=@var{dir} @gol
490 -iquote@var{dir} -L@var{dir} -no-canonical-prefixes -I- @gol
491 --sysroot=@var{dir} --no-sysroot-suffix}
493 @item Code Generation Options
494 @xref{Code Gen Options,,Options for Code Generation Conventions}.
495 @gccoptlist{-fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
496 -ffixed-@var{reg} -fexceptions @gol
497 -fnon-call-exceptions -fdelete-dead-exceptions -funwind-tables @gol
498 -fasynchronous-unwind-tables @gol
500 -finhibit-size-directive -fno-common -fno-ident @gol
501 -fpcc-struct-return -fpic -fPIC -fpie -fPIE -fno-plt @gol
502 -fno-jump-tables @gol
503 -frecord-gcc-switches @gol
504 -freg-struct-return -fshort-enums -fshort-wchar @gol
505 -fverbose-asm -fpack-struct[=@var{n}] @gol
506 -fleading-underscore -ftls-model=@var{model} @gol
507 -fstack-reuse=@var{reuse_level} @gol
508 -ftrampolines -ftrapv -fwrapv @gol
509 -fvisibility=@r{[}default@r{|}internal@r{|}hidden@r{|}protected@r{]} @gol
510 -fstrict-volatile-bitfields -fsync-libcalls}
512 @item Developer Options
513 @xref{Developer Options,,GCC Developer Options}.
514 @gccoptlist{-d@var{letters} -dumpspecs -dumpmachine -dumpversion @gol
515 -fchecking -fchecking=@var{n} -fdbg-cnt-list @gol
516 -fdbg-cnt=@var{counter-value-list} @gol
517 -fdisable-ipa-@var{pass_name} @gol
518 -fdisable-rtl-@var{pass_name} @gol
519 -fdisable-rtl-@var{pass-name}=@var{range-list} @gol
520 -fdisable-tree-@var{pass_name} @gol
521 -fdisable-tree-@var{pass-name}=@var{range-list} @gol
522 -fdump-noaddr -fdump-unnumbered -fdump-unnumbered-links @gol
523 -fdump-translation-unit@r{[}-@var{n}@r{]} @gol
524 -fdump-class-hierarchy@r{[}-@var{n}@r{]} @gol
525 -fdump-ipa-all -fdump-ipa-cgraph -fdump-ipa-inline @gol
527 -fdump-rtl-@var{pass} -fdump-rtl-@var{pass}=@var{filename} @gol
528 -fdump-statistics @gol
530 -fdump-tree-original@r{[}-@var{n}@r{]} @gol
531 -fdump-tree-optimized@r{[}-@var{n}@r{]} @gol
532 -fdump-tree-cfg -fdump-tree-alias @gol
534 -fdump-tree-ssa@r{[}-@var{n}@r{]} -fdump-tree-pre@r{[}-@var{n}@r{]} @gol
535 -fdump-tree-ccp@r{[}-@var{n}@r{]} -fdump-tree-dce@r{[}-@var{n}@r{]} @gol
536 -fdump-tree-gimple@r{[}-raw@r{]} @gol
537 -fdump-tree-dom@r{[}-@var{n}@r{]} @gol
538 -fdump-tree-dse@r{[}-@var{n}@r{]} @gol
539 -fdump-tree-phiprop@r{[}-@var{n}@r{]} @gol
540 -fdump-tree-phiopt@r{[}-@var{n}@r{]} @gol
541 -fdump-tree-backprop@r{[}-@var{n}@r{]} @gol
542 -fdump-tree-forwprop@r{[}-@var{n}@r{]} @gol
543 -fdump-tree-nrv -fdump-tree-vect @gol
544 -fdump-tree-sink @gol
545 -fdump-tree-sra@r{[}-@var{n}@r{]} @gol
546 -fdump-tree-forwprop@r{[}-@var{n}@r{]} @gol
547 -fdump-tree-fre@r{[}-@var{n}@r{]} @gol
548 -fdump-tree-vtable-verify @gol
549 -fdump-tree-vrp@r{[}-@var{n}@r{]} @gol
550 -fdump-tree-split-paths@r{[}-@var{n}@r{]} @gol
551 -fdump-tree-storeccp@r{[}-@var{n}@r{]} @gol
552 -fdump-final-insns=@var{file} @gol
553 -fcompare-debug@r{[}=@var{opts}@r{]} -fcompare-debug-second @gol
554 -fenable-@var{kind}-@var{pass} @gol
555 -fenable-@var{kind}-@var{pass}=@var{range-list} @gol
556 -fira-verbose=@var{n} @gol
557 -flto-report -flto-report-wpa -fmem-report-wpa @gol
558 -fmem-report -fpre-ipa-mem-report -fpost-ipa-mem-report @gol
559 -fopt-info -fopt-info-@var{options}@r{[}=@var{file}@r{]} @gol
560 -fprofile-report @gol
561 -frandom-seed=@var{string} -fsched-verbose=@var{n} @gol
562 -fsel-sched-verbose -fsel-sched-dump-cfg -fsel-sched-pipelining-verbose @gol
563 -fstats -fstack-usage -ftime-report -ftime-report-details @gol
564 -fvar-tracking-assignments-toggle -gtoggle @gol
565 -print-file-name=@var{library} -print-libgcc-file-name @gol
566 -print-multi-directory -print-multi-lib -print-multi-os-directory @gol
567 -print-prog-name=@var{program} -print-search-dirs -Q @gol
568 -print-sysroot -print-sysroot-headers-suffix @gol
569 -save-temps -save-temps=cwd -save-temps=obj -time@r{[}=@var{file}@r{]}}
571 @item Machine-Dependent Options
572 @xref{Submodel Options,,Machine-Dependent Options}.
573 @c This list is ordered alphanumerically by subsection name.
574 @c Try and put the significant identifier (CPU or system) first,
575 @c so users have a clue at guessing where the ones they want will be.
577 @emph{AArch64 Options}
578 @gccoptlist{-mabi=@var{name} -mbig-endian -mlittle-endian @gol
579 -mgeneral-regs-only @gol
580 -mcmodel=tiny -mcmodel=small -mcmodel=large @gol
582 -momit-leaf-frame-pointer -mno-omit-leaf-frame-pointer @gol
583 -mtls-dialect=desc -mtls-dialect=traditional @gol
584 -mtls-size=@var{size} @gol
585 -mfix-cortex-a53-835769 -mno-fix-cortex-a53-835769 @gol
586 -mfix-cortex-a53-843419 -mno-fix-cortex-a53-843419 @gol
587 -mlow-precision-recip-sqrt -mno-low-precision-recip-sqrt@gol
588 -mlow-precision-sqrt -mno-low-precision-sqrt@gol
589 -mlow-precision-div -mno-low-precision-div @gol
590 -march=@var{name} -mcpu=@var{name} -mtune=@var{name}}
592 @emph{Adapteva Epiphany Options}
593 @gccoptlist{-mhalf-reg-file -mprefer-short-insn-regs @gol
594 -mbranch-cost=@var{num} -mcmove -mnops=@var{num} -msoft-cmpsf @gol
595 -msplit-lohi -mpost-inc -mpost-modify -mstack-offset=@var{num} @gol
596 -mround-nearest -mlong-calls -mshort-calls -msmall16 @gol
597 -mfp-mode=@var{mode} -mvect-double -max-vect-align=@var{num} @gol
598 -msplit-vecmove-early -m1reg-@var{reg}}
601 @gccoptlist{-mbarrel-shifter @gol
602 -mcpu=@var{cpu} -mA6 -mARC600 -mA7 -mARC700 @gol
603 -mdpfp -mdpfp-compact -mdpfp-fast -mno-dpfp-lrsr @gol
604 -mea -mno-mpy -mmul32x16 -mmul64 -matomic @gol
605 -mnorm -mspfp -mspfp-compact -mspfp-fast -msimd -msoft-float -mswap @gol
606 -mcrc -mdsp-packa -mdvbf -mlock -mmac-d16 -mmac-24 -mrtsc -mswape @gol
607 -mtelephony -mxy -misize -mannotate-align -marclinux -marclinux_prof @gol
608 -mlong-calls -mmedium-calls -msdata @gol
609 -mucb-mcount -mvolatile-cache -mtp-regno=@var{regno} @gol
610 -malign-call -mauto-modify-reg -mbbit-peephole -mno-brcc @gol
611 -mcase-vector-pcrel -mcompact-casesi -mno-cond-exec -mearly-cbranchsi @gol
612 -mexpand-adddi -mindexed-loads -mlra -mlra-priority-none @gol
613 -mlra-priority-compact mlra-priority-noncompact -mno-millicode @gol
614 -mmixed-code -mq-class -mRcq -mRcw -msize-level=@var{level} @gol
615 -mtune=@var{cpu} -mmultcost=@var{num} @gol
616 -munalign-prob-threshold=@var{probability} -mmpy-option=@var{multo} @gol
617 -mdiv-rem -mcode-density -mll64 -mfpu=@var{fpu}}
620 @gccoptlist{-mapcs-frame -mno-apcs-frame @gol
621 -mabi=@var{name} @gol
622 -mapcs-stack-check -mno-apcs-stack-check @gol
623 -mapcs-float -mno-apcs-float @gol
624 -mapcs-reentrant -mno-apcs-reentrant @gol
625 -msched-prolog -mno-sched-prolog @gol
626 -mlittle-endian -mbig-endian @gol
627 -mfloat-abi=@var{name} @gol
628 -mfp16-format=@var{name}
629 -mthumb-interwork -mno-thumb-interwork @gol
630 -mcpu=@var{name} -march=@var{name} -mfpu=@var{name} @gol
631 -mtune=@var{name} -mprint-tune-info @gol
632 -mstructure-size-boundary=@var{n} @gol
633 -mabort-on-noreturn @gol
634 -mlong-calls -mno-long-calls @gol
635 -msingle-pic-base -mno-single-pic-base @gol
636 -mpic-register=@var{reg} @gol
637 -mnop-fun-dllimport @gol
638 -mpoke-function-name @gol
640 -mtpcs-frame -mtpcs-leaf-frame @gol
641 -mcaller-super-interworking -mcallee-super-interworking @gol
642 -mtp=@var{name} -mtls-dialect=@var{dialect} @gol
643 -mword-relocations @gol
644 -mfix-cortex-m3-ldrd @gol
645 -munaligned-access @gol
646 -mneon-for-64bits @gol
647 -mslow-flash-data @gol
648 -masm-syntax-unified @gol
653 @gccoptlist{-mmcu=@var{mcu} -maccumulate-args -mbranch-cost=@var{cost} @gol
654 -mcall-prologues -mint8 -mn_flash=@var{size} -mno-interrupts @gol
655 -mrelax -mrmw -mstrict-X -mtiny-stack -mfract-convert-truncate -nodevicelib @gol
656 -Waddr-space-convert -Wmisspelled-isr}
658 @emph{Blackfin Options}
659 @gccoptlist{-mcpu=@var{cpu}@r{[}-@var{sirevision}@r{]} @gol
660 -msim -momit-leaf-frame-pointer -mno-omit-leaf-frame-pointer @gol
661 -mspecld-anomaly -mno-specld-anomaly -mcsync-anomaly -mno-csync-anomaly @gol
662 -mlow-64k -mno-low64k -mstack-check-l1 -mid-shared-library @gol
663 -mno-id-shared-library -mshared-library-id=@var{n} @gol
664 -mleaf-id-shared-library -mno-leaf-id-shared-library @gol
665 -msep-data -mno-sep-data -mlong-calls -mno-long-calls @gol
666 -mfast-fp -minline-plt -mmulticore -mcorea -mcoreb -msdram @gol
670 @gccoptlist{-mbig-endian -mlittle-endian -march=@var{cpu} @gol
671 -msim -msdata=@var{sdata-type}}
674 @gccoptlist{-mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol
675 -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol
676 -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol
677 -mstack-align -mdata-align -mconst-align @gol
678 -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol
679 -melf -maout -melinux -mlinux -sim -sim2 @gol
680 -mmul-bug-workaround -mno-mul-bug-workaround}
683 @gccoptlist{-mmac @gol
684 -mcr16cplus -mcr16c @gol
685 -msim -mint32 -mbit-ops
686 -mdata-model=@var{model}}
688 @emph{Darwin Options}
689 @gccoptlist{-all_load -allowable_client -arch -arch_errors_fatal @gol
690 -arch_only -bind_at_load -bundle -bundle_loader @gol
691 -client_name -compatibility_version -current_version @gol
693 -dependency-file -dylib_file -dylinker_install_name @gol
694 -dynamic -dynamiclib -exported_symbols_list @gol
695 -filelist -flat_namespace -force_cpusubtype_ALL @gol
696 -force_flat_namespace -headerpad_max_install_names @gol
698 -image_base -init -install_name -keep_private_externs @gol
699 -multi_module -multiply_defined -multiply_defined_unused @gol
700 -noall_load -no_dead_strip_inits_and_terms @gol
701 -nofixprebinding -nomultidefs -noprebind -noseglinkedit @gol
702 -pagezero_size -prebind -prebind_all_twolevel_modules @gol
703 -private_bundle -read_only_relocs -sectalign @gol
704 -sectobjectsymbols -whyload -seg1addr @gol
705 -sectcreate -sectobjectsymbols -sectorder @gol
706 -segaddr -segs_read_only_addr -segs_read_write_addr @gol
707 -seg_addr_table -seg_addr_table_filename -seglinkedit @gol
708 -segprot -segs_read_only_addr -segs_read_write_addr @gol
709 -single_module -static -sub_library -sub_umbrella @gol
710 -twolevel_namespace -umbrella -undefined @gol
711 -unexported_symbols_list -weak_reference_mismatches @gol
712 -whatsloaded -F -gused -gfull -mmacosx-version-min=@var{version} @gol
713 -mkernel -mone-byte-bool}
715 @emph{DEC Alpha Options}
716 @gccoptlist{-mno-fp-regs -msoft-float @gol
717 -mieee -mieee-with-inexact -mieee-conformant @gol
718 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol
719 -mtrap-precision=@var{mode} -mbuild-constants @gol
720 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} @gol
721 -mbwx -mmax -mfix -mcix @gol
722 -mfloat-vax -mfloat-ieee @gol
723 -mexplicit-relocs -msmall-data -mlarge-data @gol
724 -msmall-text -mlarge-text @gol
725 -mmemory-latency=@var{time}}
728 @gccoptlist{-msmall-model -mno-lsim}
731 @gccoptlist{-msim -mlra -mnodiv}
734 @gccoptlist{-mgpr-32 -mgpr-64 -mfpr-32 -mfpr-64 @gol
735 -mhard-float -msoft-float @gol
736 -malloc-cc -mfixed-cc -mdword -mno-dword @gol
737 -mdouble -mno-double @gol
738 -mmedia -mno-media -mmuladd -mno-muladd @gol
739 -mfdpic -minline-plt -mgprel-ro -multilib-library-pic @gol
740 -mlinked-fp -mlong-calls -malign-labels @gol
741 -mlibrary-pic -macc-4 -macc-8 @gol
742 -mpack -mno-pack -mno-eflags -mcond-move -mno-cond-move @gol
743 -moptimize-membar -mno-optimize-membar @gol
744 -mscc -mno-scc -mcond-exec -mno-cond-exec @gol
745 -mvliw-branch -mno-vliw-branch @gol
746 -mmulti-cond-exec -mno-multi-cond-exec -mnested-cond-exec @gol
747 -mno-nested-cond-exec -mtomcat-stats @gol
751 @emph{GNU/Linux Options}
752 @gccoptlist{-mglibc -muclibc -mmusl -mbionic -mandroid @gol
753 -tno-android-cc -tno-android-ld}
755 @emph{H8/300 Options}
756 @gccoptlist{-mrelax -mh -ms -mn -mexr -mno-exr -mint32 -malign-300}
759 @gccoptlist{-march=@var{architecture-type} @gol
760 -mdisable-fpregs -mdisable-indexing @gol
761 -mfast-indirect-calls -mgas -mgnu-ld -mhp-ld @gol
762 -mfixed-range=@var{register-range} @gol
763 -mjump-in-delay -mlinker-opt -mlong-calls @gol
764 -mlong-load-store -mno-disable-fpregs @gol
765 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
766 -mno-jump-in-delay -mno-long-load-store @gol
767 -mno-portable-runtime -mno-soft-float @gol
768 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
769 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
770 -mschedule=@var{cpu-type} -mspace-regs -msio -mwsio @gol
771 -munix=@var{unix-std} -nolibdld -static -threads}
774 @gccoptlist{-mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
775 -mvolatile-asm-stop -mregister-names -msdata -mno-sdata @gol
776 -mconstant-gp -mauto-pic -mfused-madd @gol
777 -minline-float-divide-min-latency @gol
778 -minline-float-divide-max-throughput @gol
779 -mno-inline-float-divide @gol
780 -minline-int-divide-min-latency @gol
781 -minline-int-divide-max-throughput @gol
782 -mno-inline-int-divide @gol
783 -minline-sqrt-min-latency -minline-sqrt-max-throughput @gol
784 -mno-inline-sqrt @gol
785 -mdwarf2-asm -mearly-stop-bits @gol
786 -mfixed-range=@var{register-range} -mtls-size=@var{tls-size} @gol
787 -mtune=@var{cpu-type} -milp32 -mlp64 @gol
788 -msched-br-data-spec -msched-ar-data-spec -msched-control-spec @gol
789 -msched-br-in-data-spec -msched-ar-in-data-spec -msched-in-control-spec @gol
790 -msched-spec-ldc -msched-spec-control-ldc @gol
791 -msched-prefer-non-data-spec-insns -msched-prefer-non-control-spec-insns @gol
792 -msched-stop-bits-after-every-cycle -msched-count-spec-in-critical-path @gol
793 -msel-sched-dont-check-control-spec -msched-fp-mem-deps-zero-cost @gol
794 -msched-max-memory-insns-hard-limit -msched-max-memory-insns=@var{max-insns}}
797 @gccoptlist{-mbarrel-shift-enabled -mdivide-enabled -mmultiply-enabled @gol
798 -msign-extend-enabled -muser-enabled}
800 @emph{M32R/D Options}
801 @gccoptlist{-m32r2 -m32rx -m32r @gol
803 -malign-loops -mno-align-loops @gol
804 -missue-rate=@var{number} @gol
805 -mbranch-cost=@var{number} @gol
806 -mmodel=@var{code-size-model-type} @gol
807 -msdata=@var{sdata-type} @gol
808 -mno-flush-func -mflush-func=@var{name} @gol
809 -mno-flush-trap -mflush-trap=@var{number} @gol
813 @gccoptlist{-mcpu=@var{cpu} -msim -memregs=@var{number}}
815 @emph{M680x0 Options}
816 @gccoptlist{-march=@var{arch} -mcpu=@var{cpu} -mtune=@var{tune} @gol
817 -m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
818 -m68060 -mcpu32 -m5200 -m5206e -m528x -m5307 -m5407 @gol
819 -mcfv4e -mbitfield -mno-bitfield -mc68000 -mc68020 @gol
820 -mnobitfield -mrtd -mno-rtd -mdiv -mno-div -mshort @gol
821 -mno-short -mhard-float -m68881 -msoft-float -mpcrel @gol
822 -malign-int -mstrict-align -msep-data -mno-sep-data @gol
823 -mshared-library-id=n -mid-shared-library -mno-id-shared-library @gol
827 @gccoptlist{-mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
828 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
829 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
830 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
831 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
834 @gccoptlist{-mabsdiff -mall-opts -maverage -mbased=@var{n} -mbitops @gol
835 -mc=@var{n} -mclip -mconfig=@var{name} -mcop -mcop32 -mcop64 -mivc2 @gol
836 -mdc -mdiv -meb -mel -mio-volatile -ml -mleadz -mm -mminmax @gol
837 -mmult -mno-opts -mrepeat -ms -msatur -msdram -msim -msimnovec -mtf @gol
840 @emph{MicroBlaze Options}
841 @gccoptlist{-msoft-float -mhard-float -msmall-divides -mcpu=@var{cpu} @gol
842 -mmemcpy -mxl-soft-mul -mxl-soft-div -mxl-barrel-shift @gol
843 -mxl-pattern-compare -mxl-stack-check -mxl-gp-opt -mno-clearbss @gol
844 -mxl-multiply-high -mxl-float-convert -mxl-float-sqrt @gol
845 -mbig-endian -mlittle-endian -mxl-reorder -mxl-mode-@var{app-model}}
848 @gccoptlist{-EL -EB -march=@var{arch} -mtune=@var{arch} @gol
849 -mips1 -mips2 -mips3 -mips4 -mips32 -mips32r2 -mips32r3 -mips32r5 @gol
850 -mips32r6 -mips64 -mips64r2 -mips64r3 -mips64r5 -mips64r6 @gol
851 -mips16 -mno-mips16 -mflip-mips16 @gol
852 -minterlink-compressed -mno-interlink-compressed @gol
853 -minterlink-mips16 -mno-interlink-mips16 @gol
854 -mabi=@var{abi} -mabicalls -mno-abicalls @gol
855 -mshared -mno-shared -mplt -mno-plt -mxgot -mno-xgot @gol
856 -mgp32 -mgp64 -mfp32 -mfpxx -mfp64 -mhard-float -msoft-float @gol
857 -mno-float -msingle-float -mdouble-float @gol
858 -modd-spreg -mno-odd-spreg @gol
859 -mabs=@var{mode} -mnan=@var{encoding} @gol
860 -mdsp -mno-dsp -mdspr2 -mno-dspr2 @gol
863 -mvirt -mno-virt @gol
865 -mmicromips -mno-micromips @gol
867 -mfpu=@var{fpu-type} @gol
868 -msmartmips -mno-smartmips @gol
869 -mpaired-single -mno-paired-single -mdmx -mno-mdmx @gol
870 -mips3d -mno-mips3d -mmt -mno-mt -mllsc -mno-llsc @gol
871 -mlong64 -mlong32 -msym32 -mno-sym32 @gol
872 -G@var{num} -mlocal-sdata -mno-local-sdata @gol
873 -mextern-sdata -mno-extern-sdata -mgpopt -mno-gopt @gol
874 -membedded-data -mno-embedded-data @gol
875 -muninit-const-in-rodata -mno-uninit-const-in-rodata @gol
876 -mcode-readable=@var{setting} @gol
877 -msplit-addresses -mno-split-addresses @gol
878 -mexplicit-relocs -mno-explicit-relocs @gol
879 -mcheck-zero-division -mno-check-zero-division @gol
880 -mdivide-traps -mdivide-breaks @gol
881 -mmemcpy -mno-memcpy -mlong-calls -mno-long-calls @gol
882 -mmad -mno-mad -mimadd -mno-imadd -mfused-madd -mno-fused-madd -nocpp @gol
883 -mfix-24k -mno-fix-24k @gol
884 -mfix-r4000 -mno-fix-r4000 -mfix-r4400 -mno-fix-r4400 @gol
885 -mfix-r10000 -mno-fix-r10000 -mfix-rm7000 -mno-fix-rm7000 @gol
886 -mfix-vr4120 -mno-fix-vr4120 @gol
887 -mfix-vr4130 -mno-fix-vr4130 -mfix-sb1 -mno-fix-sb1 @gol
888 -mflush-func=@var{func} -mno-flush-func @gol
889 -mbranch-cost=@var{num} -mbranch-likely -mno-branch-likely @gol
890 -mcompact-branches=@var{policy} @gol
891 -mfp-exceptions -mno-fp-exceptions @gol
892 -mvr4130-align -mno-vr4130-align -msynci -mno-synci @gol
893 -mrelax-pic-calls -mno-relax-pic-calls -mmcount-ra-address @gol
894 -mframe-header-opt -mno-frame-header-opt}
897 @gccoptlist{-mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
898 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
899 -melf -mbranch-predict -mno-branch-predict -mbase-addresses @gol
900 -mno-base-addresses -msingle-exit -mno-single-exit}
902 @emph{MN10300 Options}
903 @gccoptlist{-mmult-bug -mno-mult-bug @gol
904 -mno-am33 -mam33 -mam33-2 -mam34 @gol
905 -mtune=@var{cpu-type} @gol
906 -mreturn-pointer-on-d0 @gol
907 -mno-crt0 -mrelax -mliw -msetlb}
910 @gccoptlist{-meb -mel -mmul.x -mno-crt0}
912 @emph{MSP430 Options}
913 @gccoptlist{-msim -masm-hex -mmcu= -mcpu= -mlarge -msmall -mrelax @gol
915 -mcode-region= -mdata-region= @gol
916 -msilicon-errata= -msilicon-errata-warn= @gol
920 @gccoptlist{-mbig-endian -mlittle-endian @gol
921 -mreduced-regs -mfull-regs @gol
922 -mcmov -mno-cmov @gol
923 -mperf-ext -mno-perf-ext @gol
924 -mv3push -mno-v3push @gol
925 -m16bit -mno-16bit @gol
926 -misr-vector-size=@var{num} @gol
927 -mcache-block-size=@var{num} @gol
928 -march=@var{arch} @gol
929 -mcmodel=@var{code-model} @gol
932 @emph{Nios II Options}
933 @gccoptlist{-G @var{num} -mgpopt=@var{option} -mgpopt -mno-gpopt @gol
935 -mno-bypass-cache -mbypass-cache @gol
936 -mno-cache-volatile -mcache-volatile @gol
937 -mno-fast-sw-div -mfast-sw-div @gol
938 -mhw-mul -mno-hw-mul -mhw-mulx -mno-hw-mulx -mno-hw-div -mhw-div @gol
939 -mcustom-@var{insn}=@var{N} -mno-custom-@var{insn} @gol
940 -mcustom-fpu-cfg=@var{name} @gol
941 -mhal -msmallc -msys-crt0=@var{name} -msys-lib=@var{name} @gol
942 -march=@var{arch} -mbmx -mno-bmx -mcdx -mno-cdx}
944 @emph{Nvidia PTX Options}
945 @gccoptlist{-m32 -m64 -mmainkernel -moptimize}
947 @emph{PDP-11 Options}
948 @gccoptlist{-mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol
949 -mbcopy -mbcopy-builtin -mint32 -mno-int16 @gol
950 -mint16 -mno-int32 -mfloat32 -mno-float64 @gol
951 -mfloat64 -mno-float32 -mabshi -mno-abshi @gol
952 -mbranch-expensive -mbranch-cheap @gol
953 -munix-asm -mdec-asm}
955 @emph{picoChip Options}
956 @gccoptlist{-mae=@var{ae_type} -mvliw-lookahead=@var{N} @gol
957 -msymbol-as-address -mno-inefficient-warnings}
959 @emph{PowerPC Options}
960 See RS/6000 and PowerPC Options.
963 @gccoptlist{-msim -mmul=none -mmul=g13 -mmul=g14 -mallregs @gol
964 -mcpu=g10 -mcpu=g13 -mcpu=g14 -mg10 -mg13 -mg14 @gol
965 -m64bit-doubles -m32bit-doubles -msave-mduc-in-interrupts}
967 @emph{RS/6000 and PowerPC Options}
968 @gccoptlist{-mcpu=@var{cpu-type} @gol
969 -mtune=@var{cpu-type} @gol
970 -mcmodel=@var{code-model} @gol
972 -maltivec -mno-altivec @gol
973 -mpowerpc-gpopt -mno-powerpc-gpopt @gol
974 -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol
975 -mmfcrf -mno-mfcrf -mpopcntb -mno-popcntb -mpopcntd -mno-popcntd @gol
976 -mfprnd -mno-fprnd @gol
977 -mcmpb -mno-cmpb -mmfpgpr -mno-mfpgpr -mhard-dfp -mno-hard-dfp @gol
978 -mfull-toc -mminimal-toc -mno-fp-in-toc -mno-sum-in-toc @gol
979 -m64 -m32 -mxl-compat -mno-xl-compat -mpe @gol
980 -malign-power -malign-natural @gol
981 -msoft-float -mhard-float -mmultiple -mno-multiple @gol
982 -msingle-float -mdouble-float -msimple-fpu @gol
983 -mstring -mno-string -mupdate -mno-update @gol
984 -mavoid-indexed-addresses -mno-avoid-indexed-addresses @gol
985 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol
986 -mstrict-align -mno-strict-align -mrelocatable @gol
987 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol
988 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol
989 -mdynamic-no-pic -maltivec -mswdiv -msingle-pic-base @gol
990 -mprioritize-restricted-insns=@var{priority} @gol
991 -msched-costly-dep=@var{dependence_type} @gol
992 -minsert-sched-nops=@var{scheme} @gol
993 -mcall-sysv -mcall-netbsd @gol
994 -maix-struct-return -msvr4-struct-return @gol
995 -mabi=@var{abi-type} -msecure-plt -mbss-plt @gol
996 -mblock-move-inline-limit=@var{num} @gol
997 -misel -mno-isel @gol
998 -misel=yes -misel=no @gol
1000 -mspe=yes -mspe=no @gol
1002 -mgen-cell-microcode -mwarn-cell-microcode @gol
1003 -mvrsave -mno-vrsave @gol
1004 -mmulhw -mno-mulhw @gol
1005 -mdlmzb -mno-dlmzb @gol
1006 -mfloat-gprs=yes -mfloat-gprs=no -mfloat-gprs=single -mfloat-gprs=double @gol
1007 -mprototype -mno-prototype @gol
1008 -msim -mmvme -mads -myellowknife -memb -msdata @gol
1009 -msdata=@var{opt} -mvxworks -G @var{num} -pthread @gol
1010 -mrecip -mrecip=@var{opt} -mno-recip -mrecip-precision @gol
1011 -mno-recip-precision @gol
1012 -mveclibabi=@var{type} -mfriz -mno-friz @gol
1013 -mpointers-to-nested-functions -mno-pointers-to-nested-functions @gol
1014 -msave-toc-indirect -mno-save-toc-indirect @gol
1015 -mpower8-fusion -mno-mpower8-fusion -mpower8-vector -mno-power8-vector @gol
1016 -mcrypto -mno-crypto -mhtm -mno-htm -mdirect-move -mno-direct-move @gol
1017 -mquad-memory -mno-quad-memory @gol
1018 -mquad-memory-atomic -mno-quad-memory-atomic @gol
1019 -mcompat-align-parm -mno-compat-align-parm @gol
1020 -mupper-regs-df -mno-upper-regs-df -mupper-regs-sf -mno-upper-regs-sf @gol
1021 -mupper-regs-di -mno-upper-regs-di @gol
1022 -mupper-regs -mno-upper-regs @gol
1023 -mfloat128 -mno-float128 -mfloat128-hardware -mno-float128-hardware @gol
1024 -mgnu-attribute -mno-gnu-attribute @gol
1028 @gccoptlist{-m64bit-doubles -m32bit-doubles -fpu -nofpu@gol
1030 -mbig-endian-data -mlittle-endian-data @gol
1033 -mas100-syntax -mno-as100-syntax@gol
1035 -mmax-constant-size=@gol
1038 -mallow-string-insns -mno-allow-string-insns@gol
1040 -mno-warn-multiple-fast-interrupts@gol
1041 -msave-acc-in-interrupts}
1043 @emph{S/390 and zSeries Options}
1044 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
1045 -mhard-float -msoft-float -mhard-dfp -mno-hard-dfp @gol
1046 -mlong-double-64 -mlong-double-128 @gol
1047 -mbackchain -mno-backchain -mpacked-stack -mno-packed-stack @gol
1048 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
1049 -m64 -m31 -mdebug -mno-debug -mesa -mzarch @gol
1050 -mhtm -mvx -mzvector @gol
1051 -mtpf-trace -mno-tpf-trace -mfused-madd -mno-fused-madd @gol
1052 -mwarn-framesize -mwarn-dynamicstack -mstack-size -mstack-guard @gol
1053 -mhotpatch=@var{halfwords},@var{halfwords}}
1055 @emph{Score Options}
1056 @gccoptlist{-meb -mel @gol
1060 -mscore5 -mscore5u -mscore7 -mscore7d}
1063 @gccoptlist{-m1 -m2 -m2e @gol
1064 -m2a-nofpu -m2a-single-only -m2a-single -m2a @gol
1066 -m4-nofpu -m4-single-only -m4-single -m4 @gol
1067 -m4a-nofpu -m4a-single-only -m4a-single -m4a -m4al @gol
1068 -mb -ml -mdalign -mrelax @gol
1069 -mbigtable -mfmovd -mrenesas -mno-renesas -mnomacsave @gol
1070 -mieee -mno-ieee -mbitops -misize -minline-ic_invalidate -mpadstruct @gol
1071 -mprefergot -musermode -multcost=@var{number} -mdiv=@var{strategy} @gol
1072 -mdivsi3_libfunc=@var{name} -mfixed-range=@var{register-range} @gol
1073 -maccumulate-outgoing-args @gol
1074 -matomic-model=@var{atomic-model} @gol
1075 -mbranch-cost=@var{num} -mzdcbranch -mno-zdcbranch @gol
1076 -mcbranch-force-delay-slot @gol
1077 -mfused-madd -mno-fused-madd -mfsca -mno-fsca -mfsrra -mno-fsrra @gol
1078 -mpretend-cmove -mtas}
1080 @emph{Solaris 2 Options}
1081 @gccoptlist{-mclear-hwcap -mno-clear-hwcap -mimpure-text -mno-impure-text @gol
1084 @emph{SPARC Options}
1085 @gccoptlist{-mcpu=@var{cpu-type} @gol
1086 -mtune=@var{cpu-type} @gol
1087 -mcmodel=@var{code-model} @gol
1088 -mmemory-model=@var{mem-model} @gol
1089 -m32 -m64 -mapp-regs -mno-app-regs @gol
1090 -mfaster-structs -mno-faster-structs -mflat -mno-flat @gol
1091 -mfpu -mno-fpu -mhard-float -msoft-float @gol
1092 -mhard-quad-float -msoft-quad-float @gol
1093 -mstack-bias -mno-stack-bias @gol
1094 -mstd-struct-return -mno-std-struct-return @gol
1095 -munaligned-doubles -mno-unaligned-doubles @gol
1096 -muser-mode -mno-user-mode @gol
1097 -mv8plus -mno-v8plus -mvis -mno-vis @gol
1098 -mvis2 -mno-vis2 -mvis3 -mno-vis3 @gol
1099 -mcbcond -mno-cbcond -mfmaf -mno-fmaf @gol
1100 -mpopc -mno-popc -msubxc -mno-subxc@gol
1101 -mfix-at697f -mfix-ut699}
1104 @gccoptlist{-mwarn-reloc -merror-reloc @gol
1105 -msafe-dma -munsafe-dma @gol
1107 -msmall-mem -mlarge-mem -mstdmain @gol
1108 -mfixed-range=@var{register-range} @gol
1110 -maddress-space-conversion -mno-address-space-conversion @gol
1111 -mcache-size=@var{cache-size} @gol
1112 -matomic-updates -mno-atomic-updates}
1114 @emph{System V Options}
1115 @gccoptlist{-Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
1117 @emph{TILE-Gx Options}
1118 @gccoptlist{-mcpu=CPU -m32 -m64 -mbig-endian -mlittle-endian @gol
1119 -mcmodel=@var{code-model}}
1121 @emph{TILEPro Options}
1122 @gccoptlist{-mcpu=@var{cpu} -m32}
1125 @gccoptlist{-mlong-calls -mno-long-calls -mep -mno-ep @gol
1126 -mprolog-function -mno-prolog-function -mspace @gol
1127 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
1128 -mapp-regs -mno-app-regs @gol
1129 -mdisable-callt -mno-disable-callt @gol
1130 -mv850e2v3 -mv850e2 -mv850e1 -mv850es @gol
1131 -mv850e -mv850 -mv850e3v5 @gol
1142 @gccoptlist{-mg -mgnu -munix}
1144 @emph{Visium Options}
1145 @gccoptlist{-mdebug -msim -mfpu -mno-fpu -mhard-float -msoft-float @gol
1146 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} -msv-mode -muser-mode}
1149 @gccoptlist{-mvms-return-codes -mdebug-main=@var{prefix} -mmalloc64 @gol
1150 -mpointer-size=@var{size}}
1152 @emph{VxWorks Options}
1153 @gccoptlist{-mrtp -non-static -Bstatic -Bdynamic @gol
1154 -Xbind-lazy -Xbind-now}
1157 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
1158 -mtune-ctrl=@var{feature-list} -mdump-tune-features -mno-default @gol
1159 -mfpmath=@var{unit} @gol
1160 -masm=@var{dialect} -mno-fancy-math-387 @gol
1161 -mno-fp-ret-in-387 -m80387 -mhard-float -msoft-float @gol
1162 -mno-wide-multiply -mrtd -malign-double @gol
1163 -mpreferred-stack-boundary=@var{num} @gol
1164 -mincoming-stack-boundary=@var{num} @gol
1165 -mcld -mcx16 -msahf -mmovbe -mcrc32 @gol
1166 -mrecip -mrecip=@var{opt} @gol
1167 -mvzeroupper -mprefer-avx128 @gol
1168 -mmmx -msse -msse2 -msse3 -mssse3 -msse4.1 -msse4.2 -msse4 -mavx @gol
1169 -mavx2 -mavx512f -mavx512pf -mavx512er -mavx512cd -mavx512vl @gol
1170 -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -msha -maes @gol
1171 -mpclmul -mfsgsbase -mrdrnd -mf16c -mfma @gol
1172 -mprefetchwt1 -mclflushopt -mxsavec -mxsaves @gol
1173 -msse4a -m3dnow -mpopcnt -mabm -mbmi -mtbm -mfma4 -mxop -mlzcnt @gol
1174 -mbmi2 -mfxsr -mxsave -mxsaveopt -mrtm -mlwp -mmpx -mmwaitx -mclzero
1175 -mpku -mthreads @gol
1176 -mms-bitfields -mno-align-stringops -minline-all-stringops @gol
1177 -minline-stringops-dynamically -mstringop-strategy=@var{alg} @gol
1178 -mmemcpy-strategy=@var{strategy} -mmemset-strategy=@var{strategy} @gol
1179 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
1180 -m96bit-long-double -mlong-double-64 -mlong-double-80 -mlong-double-128 @gol
1181 -mregparm=@var{num} -msseregparm @gol
1182 -mveclibabi=@var{type} -mvect8-ret-in-mem @gol
1183 -mpc32 -mpc64 -mpc80 -mstackrealign @gol
1184 -momit-leaf-frame-pointer -mno-red-zone -mno-tls-direct-seg-refs @gol
1185 -mcmodel=@var{code-model} -mabi=@var{name} -maddress-mode=@var{mode} @gol
1186 -m32 -m64 -mx32 -m16 -miamcu -mlarge-data-threshold=@var{num} @gol
1187 -msse2avx -mfentry -mrecord-mcount -mnop-mcount -m8bit-idiv @gol
1188 -mavx256-split-unaligned-load -mavx256-split-unaligned-store @gol
1189 -malign-data=@var{type} -mstack-protector-guard=@var{guard} @gol
1190 -mmitigate-rop -mgeneral-regs-only}
1192 @emph{x86 Windows Options}
1193 @gccoptlist{-mconsole -mcygwin -mno-cygwin -mdll @gol
1194 -mnop-fun-dllimport -mthread @gol
1195 -municode -mwin32 -mwindows -fno-set-stack-executable}
1197 @emph{Xstormy16 Options}
1200 @emph{Xtensa Options}
1201 @gccoptlist{-mconst16 -mno-const16 @gol
1202 -mfused-madd -mno-fused-madd @gol
1204 -mserialize-volatile -mno-serialize-volatile @gol
1205 -mtext-section-literals -mno-text-section-literals @gol
1206 -mauto-litpools -mno-auto-litpools @gol
1207 -mtarget-align -mno-target-align @gol
1208 -mlongcalls -mno-longcalls}
1210 @emph{zSeries Options}
1211 See S/390 and zSeries Options.
1215 @node Overall Options
1216 @section Options Controlling the Kind of Output
1218 Compilation can involve up to four stages: preprocessing, compilation
1219 proper, assembly and linking, always in that order. GCC is capable of
1220 preprocessing and compiling several files either into several
1221 assembler input files, or into one assembler input file; then each
1222 assembler input file produces an object file, and linking combines all
1223 the object files (those newly compiled, and those specified as input)
1224 into an executable file.
1226 @cindex file name suffix
1227 For any given input file, the file name suffix determines what kind of
1228 compilation is done:
1232 C source code that must be preprocessed.
1235 C source code that should not be preprocessed.
1238 C++ source code that should not be preprocessed.
1241 Objective-C source code. Note that you must link with the @file{libobjc}
1242 library to make an Objective-C program work.
1245 Objective-C source code that should not be preprocessed.
1249 Objective-C++ source code. Note that you must link with the @file{libobjc}
1250 library to make an Objective-C++ program work. Note that @samp{.M} refers
1251 to a literal capital M@.
1253 @item @var{file}.mii
1254 Objective-C++ source code that should not be preprocessed.
1257 C, C++, Objective-C or Objective-C++ header file to be turned into a
1258 precompiled header (default), or C, C++ header file to be turned into an
1259 Ada spec (via the @option{-fdump-ada-spec} switch).
1262 @itemx @var{file}.cp
1263 @itemx @var{file}.cxx
1264 @itemx @var{file}.cpp
1265 @itemx @var{file}.CPP
1266 @itemx @var{file}.c++
1268 C++ source code that must be preprocessed. Note that in @samp{.cxx},
1269 the last two letters must both be literally @samp{x}. Likewise,
1270 @samp{.C} refers to a literal capital C@.
1274 Objective-C++ source code that must be preprocessed.
1276 @item @var{file}.mii
1277 Objective-C++ source code that should not be preprocessed.
1281 @itemx @var{file}.hp
1282 @itemx @var{file}.hxx
1283 @itemx @var{file}.hpp
1284 @itemx @var{file}.HPP
1285 @itemx @var{file}.h++
1286 @itemx @var{file}.tcc
1287 C++ header file to be turned into a precompiled header or Ada spec.
1290 @itemx @var{file}.for
1291 @itemx @var{file}.ftn
1292 Fixed form Fortran source code that should not be preprocessed.
1295 @itemx @var{file}.FOR
1296 @itemx @var{file}.fpp
1297 @itemx @var{file}.FPP
1298 @itemx @var{file}.FTN
1299 Fixed form Fortran source code that must be preprocessed (with the traditional
1302 @item @var{file}.f90
1303 @itemx @var{file}.f95
1304 @itemx @var{file}.f03
1305 @itemx @var{file}.f08
1306 Free form Fortran source code that should not be preprocessed.
1308 @item @var{file}.F90
1309 @itemx @var{file}.F95
1310 @itemx @var{file}.F03
1311 @itemx @var{file}.F08
1312 Free form Fortran source code that must be preprocessed (with the
1313 traditional preprocessor).
1318 @c FIXME: Descriptions of Java file types.
1324 @item @var{file}.ads
1325 Ada source code file that contains a library unit declaration (a
1326 declaration of a package, subprogram, or generic, or a generic
1327 instantiation), or a library unit renaming declaration (a package,
1328 generic, or subprogram renaming declaration). Such files are also
1331 @item @var{file}.adb
1332 Ada source code file containing a library unit body (a subprogram or
1333 package body). Such files are also called @dfn{bodies}.
1335 @c GCC also knows about some suffixes for languages not yet included:
1346 @itemx @var{file}.sx
1347 Assembler code that must be preprocessed.
1350 An object file to be fed straight into linking.
1351 Any file name with no recognized suffix is treated this way.
1355 You can specify the input language explicitly with the @option{-x} option:
1358 @item -x @var{language}
1359 Specify explicitly the @var{language} for the following input files
1360 (rather than letting the compiler choose a default based on the file
1361 name suffix). This option applies to all following input files until
1362 the next @option{-x} option. Possible values for @var{language} are:
1364 c c-header cpp-output
1365 c++ c++-header c++-cpp-output
1366 objective-c objective-c-header objective-c-cpp-output
1367 objective-c++ objective-c++-header objective-c++-cpp-output
1368 assembler assembler-with-cpp
1370 f77 f77-cpp-input f95 f95-cpp-input
1376 Turn off any specification of a language, so that subsequent files are
1377 handled according to their file name suffixes (as they are if @option{-x}
1378 has not been used at all).
1381 If you only want some of the stages of compilation, you can use
1382 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
1383 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
1384 @command{gcc} is to stop. Note that some combinations (for example,
1385 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
1390 Compile or assemble the source files, but do not link. The linking
1391 stage simply is not done. The ultimate output is in the form of an
1392 object file for each source file.
1394 By default, the object file name for a source file is made by replacing
1395 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
1397 Unrecognized input files, not requiring compilation or assembly, are
1402 Stop after the stage of compilation proper; do not assemble. The output
1403 is in the form of an assembler code file for each non-assembler input
1406 By default, the assembler file name for a source file is made by
1407 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
1409 Input files that don't require compilation are ignored.
1413 Stop after the preprocessing stage; do not run the compiler proper. The
1414 output is in the form of preprocessed source code, which is sent to the
1417 Input files that don't require preprocessing are ignored.
1419 @cindex output file option
1422 Place output in file @var{file}. This applies to whatever
1423 sort of output is being produced, whether it be an executable file,
1424 an object file, an assembler file or preprocessed C code.
1426 If @option{-o} is not specified, the default is to put an executable
1427 file in @file{a.out}, the object file for
1428 @file{@var{source}.@var{suffix}} in @file{@var{source}.o}, its
1429 assembler file in @file{@var{source}.s}, a precompiled header file in
1430 @file{@var{source}.@var{suffix}.gch}, and all preprocessed C source on
1435 Print (on standard error output) the commands executed to run the stages
1436 of compilation. Also print the version number of the compiler driver
1437 program and of the preprocessor and the compiler proper.
1441 Like @option{-v} except the commands are not executed and arguments
1442 are quoted unless they contain only alphanumeric characters or @code{./-_}.
1443 This is useful for shell scripts to capture the driver-generated command lines.
1447 Print (on the standard output) a description of the command-line options
1448 understood by @command{gcc}. If the @option{-v} option is also specified
1449 then @option{--help} is also passed on to the various processes
1450 invoked by @command{gcc}, so that they can display the command-line options
1451 they accept. If the @option{-Wextra} option has also been specified
1452 (prior to the @option{--help} option), then command-line options that
1453 have no documentation associated with them are also displayed.
1456 @opindex target-help
1457 Print (on the standard output) a description of target-specific command-line
1458 options for each tool. For some targets extra target-specific
1459 information may also be printed.
1461 @item --help=@{@var{class}@r{|[}^@r{]}@var{qualifier}@}@r{[},@dots{}@r{]}
1462 Print (on the standard output) a description of the command-line
1463 options understood by the compiler that fit into all specified classes
1464 and qualifiers. These are the supported classes:
1467 @item @samp{optimizers}
1468 Display all of the optimization options supported by the
1471 @item @samp{warnings}
1472 Display all of the options controlling warning messages
1473 produced by the compiler.
1476 Display target-specific options. Unlike the
1477 @option{--target-help} option however, target-specific options of the
1478 linker and assembler are not displayed. This is because those
1479 tools do not currently support the extended @option{--help=} syntax.
1482 Display the values recognized by the @option{--param}
1485 @item @var{language}
1486 Display the options supported for @var{language}, where
1487 @var{language} is the name of one of the languages supported in this
1491 Display the options that are common to all languages.
1494 These are the supported qualifiers:
1497 @item @samp{undocumented}
1498 Display only those options that are undocumented.
1501 Display options taking an argument that appears after an equal
1502 sign in the same continuous piece of text, such as:
1503 @samp{--help=target}.
1505 @item @samp{separate}
1506 Display options taking an argument that appears as a separate word
1507 following the original option, such as: @samp{-o output-file}.
1510 Thus for example to display all the undocumented target-specific
1511 switches supported by the compiler, use:
1514 --help=target,undocumented
1517 The sense of a qualifier can be inverted by prefixing it with the
1518 @samp{^} character, so for example to display all binary warning
1519 options (i.e., ones that are either on or off and that do not take an
1520 argument) that have a description, use:
1523 --help=warnings,^joined,^undocumented
1526 The argument to @option{--help=} should not consist solely of inverted
1529 Combining several classes is possible, although this usually
1530 restricts the output so much that there is nothing to display. One
1531 case where it does work, however, is when one of the classes is
1532 @var{target}. For example, to display all the target-specific
1533 optimization options, use:
1536 --help=target,optimizers
1539 The @option{--help=} option can be repeated on the command line. Each
1540 successive use displays its requested class of options, skipping
1541 those that have already been displayed.
1543 If the @option{-Q} option appears on the command line before the
1544 @option{--help=} option, then the descriptive text displayed by
1545 @option{--help=} is changed. Instead of describing the displayed
1546 options, an indication is given as to whether the option is enabled,
1547 disabled or set to a specific value (assuming that the compiler
1548 knows this at the point where the @option{--help=} option is used).
1550 Here is a truncated example from the ARM port of @command{gcc}:
1553 % gcc -Q -mabi=2 --help=target -c
1554 The following options are target specific:
1556 -mabort-on-noreturn [disabled]
1560 The output is sensitive to the effects of previous command-line
1561 options, so for example it is possible to find out which optimizations
1562 are enabled at @option{-O2} by using:
1565 -Q -O2 --help=optimizers
1568 Alternatively you can discover which binary optimizations are enabled
1569 by @option{-O3} by using:
1572 gcc -c -Q -O3 --help=optimizers > /tmp/O3-opts
1573 gcc -c -Q -O2 --help=optimizers > /tmp/O2-opts
1574 diff /tmp/O2-opts /tmp/O3-opts | grep enabled
1579 Display the version number and copyrights of the invoked GCC@.
1581 @item -pass-exit-codes
1582 @opindex pass-exit-codes
1583 Normally the @command{gcc} program exits with the code of 1 if any
1584 phase of the compiler returns a non-success return code. If you specify
1585 @option{-pass-exit-codes}, the @command{gcc} program instead returns with
1586 the numerically highest error produced by any phase returning an error
1587 indication. The C, C++, and Fortran front ends return 4 if an internal
1588 compiler error is encountered.
1592 Use pipes rather than temporary files for communication between the
1593 various stages of compilation. This fails to work on some systems where
1594 the assembler is unable to read from a pipe; but the GNU assembler has
1597 @item -specs=@var{file}
1599 Process @var{file} after the compiler reads in the standard @file{specs}
1600 file, in order to override the defaults which the @command{gcc} driver
1601 program uses when determining what switches to pass to @command{cc1},
1602 @command{cc1plus}, @command{as}, @command{ld}, etc. More than one
1603 @option{-specs=@var{file}} can be specified on the command line, and they
1604 are processed in order, from left to right. @xref{Spec Files}, for
1605 information about the format of the @var{file}.
1609 Invoke all subcommands under a wrapper program. The name of the
1610 wrapper program and its parameters are passed as a comma separated
1614 gcc -c t.c -wrapper gdb,--args
1618 This invokes all subprograms of @command{gcc} under
1619 @samp{gdb --args}, thus the invocation of @command{cc1} is
1620 @samp{gdb --args cc1 @dots{}}.
1622 @item -fplugin=@var{name}.so
1624 Load the plugin code in file @var{name}.so, assumed to be a
1625 shared object to be dlopen'd by the compiler. The base name of
1626 the shared object file is used to identify the plugin for the
1627 purposes of argument parsing (See
1628 @option{-fplugin-arg-@var{name}-@var{key}=@var{value}} below).
1629 Each plugin should define the callback functions specified in the
1632 @item -fplugin-arg-@var{name}-@var{key}=@var{value}
1633 @opindex fplugin-arg
1634 Define an argument called @var{key} with a value of @var{value}
1635 for the plugin called @var{name}.
1637 @item -fdump-ada-spec@r{[}-slim@r{]}
1638 @opindex fdump-ada-spec
1639 For C and C++ source and include files, generate corresponding Ada specs.
1640 @xref{Generating Ada Bindings for C and C++ headers,,, gnat_ugn,
1641 GNAT User's Guide}, which provides detailed documentation on this feature.
1643 @item -fada-spec-parent=@var{unit}
1644 @opindex fada-spec-parent
1645 In conjunction with @option{-fdump-ada-spec@r{[}-slim@r{]}} above, generate
1646 Ada specs as child units of parent @var{unit}.
1648 @item -fdump-go-spec=@var{file}
1649 @opindex fdump-go-spec
1650 For input files in any language, generate corresponding Go
1651 declarations in @var{file}. This generates Go @code{const},
1652 @code{type}, @code{var}, and @code{func} declarations which may be a
1653 useful way to start writing a Go interface to code written in some
1656 @include @value{srcdir}/../libiberty/at-file.texi
1660 @section Compiling C++ Programs
1662 @cindex suffixes for C++ source
1663 @cindex C++ source file suffixes
1664 C++ source files conventionally use one of the suffixes @samp{.C},
1665 @samp{.cc}, @samp{.cpp}, @samp{.CPP}, @samp{.c++}, @samp{.cp}, or
1666 @samp{.cxx}; C++ header files often use @samp{.hh}, @samp{.hpp},
1667 @samp{.H}, or (for shared template code) @samp{.tcc}; and
1668 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
1669 files with these names and compiles them as C++ programs even if you
1670 call the compiler the same way as for compiling C programs (usually
1671 with the name @command{gcc}).
1675 However, the use of @command{gcc} does not add the C++ library.
1676 @command{g++} is a program that calls GCC and automatically specifies linking
1677 against the C++ library. It treats @samp{.c},
1678 @samp{.h} and @samp{.i} files as C++ source files instead of C source
1679 files unless @option{-x} is used. This program is also useful when
1680 precompiling a C header file with a @samp{.h} extension for use in C++
1681 compilations. On many systems, @command{g++} is also installed with
1682 the name @command{c++}.
1684 @cindex invoking @command{g++}
1685 When you compile C++ programs, you may specify many of the same
1686 command-line options that you use for compiling programs in any
1687 language; or command-line options meaningful for C and related
1688 languages; or options that are meaningful only for C++ programs.
1689 @xref{C Dialect Options,,Options Controlling C Dialect}, for
1690 explanations of options for languages related to C@.
1691 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
1692 explanations of options that are meaningful only for C++ programs.
1694 @node C Dialect Options
1695 @section Options Controlling C Dialect
1696 @cindex dialect options
1697 @cindex language dialect options
1698 @cindex options, dialect
1700 The following options control the dialect of C (or languages derived
1701 from C, such as C++, Objective-C and Objective-C++) that the compiler
1705 @cindex ANSI support
1709 In C mode, this is equivalent to @option{-std=c90}. In C++ mode, it is
1710 equivalent to @option{-std=c++98}.
1712 This turns off certain features of GCC that are incompatible with ISO
1713 C90 (when compiling C code), or of standard C++ (when compiling C++ code),
1714 such as the @code{asm} and @code{typeof} keywords, and
1715 predefined macros such as @code{unix} and @code{vax} that identify the
1716 type of system you are using. It also enables the undesirable and
1717 rarely used ISO trigraph feature. For the C compiler,
1718 it disables recognition of C++ style @samp{//} comments as well as
1719 the @code{inline} keyword.
1721 The alternate keywords @code{__asm__}, @code{__extension__},
1722 @code{__inline__} and @code{__typeof__} continue to work despite
1723 @option{-ansi}. You would not want to use them in an ISO C program, of
1724 course, but it is useful to put them in header files that might be included
1725 in compilations done with @option{-ansi}. Alternate predefined macros
1726 such as @code{__unix__} and @code{__vax__} are also available, with or
1727 without @option{-ansi}.
1729 The @option{-ansi} option does not cause non-ISO programs to be
1730 rejected gratuitously. For that, @option{-Wpedantic} is required in
1731 addition to @option{-ansi}. @xref{Warning Options}.
1733 The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi}
1734 option is used. Some header files may notice this macro and refrain
1735 from declaring certain functions or defining certain macros that the
1736 ISO standard doesn't call for; this is to avoid interfering with any
1737 programs that might use these names for other things.
1739 Functions that are normally built in but do not have semantics
1740 defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in
1741 functions when @option{-ansi} is used. @xref{Other Builtins,,Other
1742 built-in functions provided by GCC}, for details of the functions
1747 Determine the language standard. @xref{Standards,,Language Standards
1748 Supported by GCC}, for details of these standard versions. This option
1749 is currently only supported when compiling C or C++.
1751 The compiler can accept several base standards, such as @samp{c90} or
1752 @samp{c++98}, and GNU dialects of those standards, such as
1753 @samp{gnu90} or @samp{gnu++98}. When a base standard is specified, the
1754 compiler accepts all programs following that standard plus those
1755 using GNU extensions that do not contradict it. For example,
1756 @option{-std=c90} turns off certain features of GCC that are
1757 incompatible with ISO C90, such as the @code{asm} and @code{typeof}
1758 keywords, but not other GNU extensions that do not have a meaning in
1759 ISO C90, such as omitting the middle term of a @code{?:}
1760 expression. On the other hand, when a GNU dialect of a standard is
1761 specified, all features supported by the compiler are enabled, even when
1762 those features change the meaning of the base standard. As a result, some
1763 strict-conforming programs may be rejected. The particular standard
1764 is used by @option{-Wpedantic} to identify which features are GNU
1765 extensions given that version of the standard. For example
1766 @option{-std=gnu90 -Wpedantic} warns about C++ style @samp{//}
1767 comments, while @option{-std=gnu99 -Wpedantic} does not.
1769 A value for this option must be provided; possible values are
1775 Support all ISO C90 programs (certain GNU extensions that conflict
1776 with ISO C90 are disabled). Same as @option{-ansi} for C code.
1778 @item iso9899:199409
1779 ISO C90 as modified in amendment 1.
1785 ISO C99. This standard is substantially completely supported, modulo
1786 bugs and floating-point issues
1787 (mainly but not entirely relating to optional C99 features from
1788 Annexes F and G). See
1789 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The
1790 names @samp{c9x} and @samp{iso9899:199x} are deprecated.
1795 ISO C11, the 2011 revision of the ISO C standard. This standard is
1796 substantially completely supported, modulo bugs, floating-point issues
1797 (mainly but not entirely relating to optional C11 features from
1798 Annexes F and G) and the optional Annexes K (Bounds-checking
1799 interfaces) and L (Analyzability). The name @samp{c1x} is deprecated.
1803 GNU dialect of ISO C90 (including some C99 features).
1807 GNU dialect of ISO C99. The name @samp{gnu9x} is deprecated.
1811 GNU dialect of ISO C11. This is the default for C code.
1812 The name @samp{gnu1x} is deprecated.
1816 The 1998 ISO C++ standard plus the 2003 technical corrigendum and some
1817 additional defect reports. Same as @option{-ansi} for C++ code.
1821 GNU dialect of @option{-std=c++98}.
1825 The 2011 ISO C++ standard plus amendments.
1826 The name @samp{c++0x} is deprecated.
1830 GNU dialect of @option{-std=c++11}.
1831 The name @samp{gnu++0x} is deprecated.
1835 The 2014 ISO C++ standard plus amendments.
1836 The name @samp{c++1y} is deprecated.
1840 GNU dialect of @option{-std=c++14}.
1841 This is the default for C++ code.
1842 The name @samp{gnu++1y} is deprecated.
1845 The next revision of the ISO C++ standard, tentatively planned for
1846 2017. Support is highly experimental, and will almost certainly
1847 change in incompatible ways in future releases.
1850 GNU dialect of @option{-std=c++1z}. Support is highly experimental,
1851 and will almost certainly change in incompatible ways in future
1855 @item -fgnu89-inline
1856 @opindex fgnu89-inline
1857 The option @option{-fgnu89-inline} tells GCC to use the traditional
1858 GNU semantics for @code{inline} functions when in C99 mode.
1859 @xref{Inline,,An Inline Function is As Fast As a Macro}.
1860 Using this option is roughly equivalent to adding the
1861 @code{gnu_inline} function attribute to all inline functions
1862 (@pxref{Function Attributes}).
1864 The option @option{-fno-gnu89-inline} explicitly tells GCC to use the
1865 C99 semantics for @code{inline} when in C99 or gnu99 mode (i.e., it
1866 specifies the default behavior).
1867 This option is not supported in @option{-std=c90} or
1868 @option{-std=gnu90} mode.
1870 The preprocessor macros @code{__GNUC_GNU_INLINE__} and
1871 @code{__GNUC_STDC_INLINE__} may be used to check which semantics are
1872 in effect for @code{inline} functions. @xref{Common Predefined
1873 Macros,,,cpp,The C Preprocessor}.
1875 @item -aux-info @var{filename}
1877 Output to the given filename prototyped declarations for all functions
1878 declared and/or defined in a translation unit, including those in header
1879 files. This option is silently ignored in any language other than C@.
1881 Besides declarations, the file indicates, in comments, the origin of
1882 each declaration (source file and line), whether the declaration was
1883 implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or
1884 @samp{O} for old, respectively, in the first character after the line
1885 number and the colon), and whether it came from a declaration or a
1886 definition (@samp{C} or @samp{F}, respectively, in the following
1887 character). In the case of function definitions, a K&R-style list of
1888 arguments followed by their declarations is also provided, inside
1889 comments, after the declaration.
1891 @item -fallow-parameterless-variadic-functions
1892 @opindex fallow-parameterless-variadic-functions
1893 Accept variadic functions without named parameters.
1895 Although it is possible to define such a function, this is not very
1896 useful as it is not possible to read the arguments. This is only
1897 supported for C as this construct is allowed by C++.
1901 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
1902 keyword, so that code can use these words as identifiers. You can use
1903 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
1904 instead. @option{-ansi} implies @option{-fno-asm}.
1906 In C++, this switch only affects the @code{typeof} keyword, since
1907 @code{asm} and @code{inline} are standard keywords. You may want to
1908 use the @option{-fno-gnu-keywords} flag instead, which has the same
1909 effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this
1910 switch only affects the @code{asm} and @code{typeof} keywords, since
1911 @code{inline} is a standard keyword in ISO C99.
1914 @itemx -fno-builtin-@var{function}
1915 @opindex fno-builtin
1916 @cindex built-in functions
1917 Don't recognize built-in functions that do not begin with
1918 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
1919 functions provided by GCC}, for details of the functions affected,
1920 including those which are not built-in functions when @option{-ansi} or
1921 @option{-std} options for strict ISO C conformance are used because they
1922 do not have an ISO standard meaning.
1924 GCC normally generates special code to handle certain built-in functions
1925 more efficiently; for instance, calls to @code{alloca} may become single
1926 instructions which adjust the stack directly, and calls to @code{memcpy}
1927 may become inline copy loops. The resulting code is often both smaller
1928 and faster, but since the function calls no longer appear as such, you
1929 cannot set a breakpoint on those calls, nor can you change the behavior
1930 of the functions by linking with a different library. In addition,
1931 when a function is recognized as a built-in function, GCC may use
1932 information about that function to warn about problems with calls to
1933 that function, or to generate more efficient code, even if the
1934 resulting code still contains calls to that function. For example,
1935 warnings are given with @option{-Wformat} for bad calls to
1936 @code{printf} when @code{printf} is built in and @code{strlen} is
1937 known not to modify global memory.
1939 With the @option{-fno-builtin-@var{function}} option
1940 only the built-in function @var{function} is
1941 disabled. @var{function} must not begin with @samp{__builtin_}. If a
1942 function is named that is not built-in in this version of GCC, this
1943 option is ignored. There is no corresponding
1944 @option{-fbuiltin-@var{function}} option; if you wish to enable
1945 built-in functions selectively when using @option{-fno-builtin} or
1946 @option{-ffreestanding}, you may define macros such as:
1949 #define abs(n) __builtin_abs ((n))
1950 #define strcpy(d, s) __builtin_strcpy ((d), (s))
1955 @cindex hosted environment
1957 Assert that compilation targets a hosted environment. This implies
1958 @option{-fbuiltin}. A hosted environment is one in which the
1959 entire standard library is available, and in which @code{main} has a return
1960 type of @code{int}. Examples are nearly everything except a kernel.
1961 This is equivalent to @option{-fno-freestanding}.
1963 @item -ffreestanding
1964 @opindex ffreestanding
1965 @cindex hosted environment
1967 Assert that compilation targets a freestanding environment. This
1968 implies @option{-fno-builtin}. A freestanding environment
1969 is one in which the standard library may not exist, and program startup may
1970 not necessarily be at @code{main}. The most obvious example is an OS kernel.
1971 This is equivalent to @option{-fno-hosted}.
1973 @xref{Standards,,Language Standards Supported by GCC}, for details of
1974 freestanding and hosted environments.
1978 @cindex OpenACC accelerator programming
1979 Enable handling of OpenACC directives @code{#pragma acc} in C/C++ and
1980 @code{!$acc} in Fortran. When @option{-fopenacc} is specified, the
1981 compiler generates accelerated code according to the OpenACC Application
1982 Programming Interface v2.0 @w{@uref{http://www.openacc.org/}}. This option
1983 implies @option{-pthread}, and thus is only supported on targets that
1984 have support for @option{-pthread}.
1986 @item -fopenacc-dim=@var{geom}
1987 @opindex fopenacc-dim
1988 @cindex OpenACC accelerator programming
1989 Specify default compute dimensions for parallel offload regions that do
1990 not explicitly specify. The @var{geom} value is a triple of
1991 ':'-separated sizes, in order 'gang', 'worker' and, 'vector'. A size
1992 can be omitted, to use a target-specific default value.
1996 @cindex OpenMP parallel
1997 Enable handling of OpenMP directives @code{#pragma omp} in C/C++ and
1998 @code{!$omp} in Fortran. When @option{-fopenmp} is specified, the
1999 compiler generates parallel code according to the OpenMP Application
2000 Program Interface v4.0 @w{@uref{http://www.openmp.org/}}. This option
2001 implies @option{-pthread}, and thus is only supported on targets that
2002 have support for @option{-pthread}. @option{-fopenmp} implies
2003 @option{-fopenmp-simd}.
2006 @opindex fopenmp-simd
2009 Enable handling of OpenMP's SIMD directives with @code{#pragma omp}
2010 in C/C++ and @code{!$omp} in Fortran. Other OpenMP directives
2015 @cindex Enable Cilk Plus
2016 Enable the usage of Cilk Plus language extension features for C/C++.
2017 When the option @option{-fcilkplus} is specified, enable the usage of
2018 the Cilk Plus Language extension features for C/C++. The present
2019 implementation follows ABI version 1.2. This is an experimental
2020 feature that is only partially complete, and whose interface may
2021 change in future versions of GCC as the official specification
2022 changes. Currently, all features but @code{_Cilk_for} have been
2027 When the option @option{-fgnu-tm} is specified, the compiler
2028 generates code for the Linux variant of Intel's current Transactional
2029 Memory ABI specification document (Revision 1.1, May 6 2009). This is
2030 an experimental feature whose interface may change in future versions
2031 of GCC, as the official specification changes. Please note that not
2032 all architectures are supported for this feature.
2034 For more information on GCC's support for transactional memory,
2035 @xref{Enabling libitm,,The GNU Transactional Memory Library,libitm,GNU
2036 Transactional Memory Library}.
2038 Note that the transactional memory feature is not supported with
2039 non-call exceptions (@option{-fnon-call-exceptions}).
2041 @item -fms-extensions
2042 @opindex fms-extensions
2043 Accept some non-standard constructs used in Microsoft header files.
2045 In C++ code, this allows member names in structures to be similar
2046 to previous types declarations.
2055 Some cases of unnamed fields in structures and unions are only
2056 accepted with this option. @xref{Unnamed Fields,,Unnamed struct/union
2057 fields within structs/unions}, for details.
2059 Note that this option is off for all targets but x86
2060 targets using ms-abi.
2062 @item -fplan9-extensions
2063 @opindex fplan9-extensions
2064 Accept some non-standard constructs used in Plan 9 code.
2066 This enables @option{-fms-extensions}, permits passing pointers to
2067 structures with anonymous fields to functions that expect pointers to
2068 elements of the type of the field, and permits referring to anonymous
2069 fields declared using a typedef. @xref{Unnamed Fields,,Unnamed
2070 struct/union fields within structs/unions}, for details. This is only
2071 supported for C, not C++.
2075 Support ISO C trigraphs. The @option{-ansi} option (and @option{-std}
2076 options for strict ISO C conformance) implies @option{-trigraphs}.
2078 @cindex traditional C language
2079 @cindex C language, traditional
2081 @itemx -traditional-cpp
2082 @opindex traditional-cpp
2083 @opindex traditional
2084 Formerly, these options caused GCC to attempt to emulate a pre-standard
2085 C compiler. They are now only supported with the @option{-E} switch.
2086 The preprocessor continues to support a pre-standard mode. See the GNU
2087 CPP manual for details.
2089 @item -fcond-mismatch
2090 @opindex fcond-mismatch
2091 Allow conditional expressions with mismatched types in the second and
2092 third arguments. The value of such an expression is void. This option
2093 is not supported for C++.
2095 @item -flax-vector-conversions
2096 @opindex flax-vector-conversions
2097 Allow implicit conversions between vectors with differing numbers of
2098 elements and/or incompatible element types. This option should not be
2101 @item -funsigned-char
2102 @opindex funsigned-char
2103 Let the type @code{char} be unsigned, like @code{unsigned char}.
2105 Each kind of machine has a default for what @code{char} should
2106 be. It is either like @code{unsigned char} by default or like
2107 @code{signed char} by default.
2109 Ideally, a portable program should always use @code{signed char} or
2110 @code{unsigned char} when it depends on the signedness of an object.
2111 But many programs have been written to use plain @code{char} and
2112 expect it to be signed, or expect it to be unsigned, depending on the
2113 machines they were written for. This option, and its inverse, let you
2114 make such a program work with the opposite default.
2116 The type @code{char} is always a distinct type from each of
2117 @code{signed char} or @code{unsigned char}, even though its behavior
2118 is always just like one of those two.
2121 @opindex fsigned-char
2122 Let the type @code{char} be signed, like @code{signed char}.
2124 Note that this is equivalent to @option{-fno-unsigned-char}, which is
2125 the negative form of @option{-funsigned-char}. Likewise, the option
2126 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
2128 @item -fsigned-bitfields
2129 @itemx -funsigned-bitfields
2130 @itemx -fno-signed-bitfields
2131 @itemx -fno-unsigned-bitfields
2132 @opindex fsigned-bitfields
2133 @opindex funsigned-bitfields
2134 @opindex fno-signed-bitfields
2135 @opindex fno-unsigned-bitfields
2136 These options control whether a bit-field is signed or unsigned, when the
2137 declaration does not use either @code{signed} or @code{unsigned}. By
2138 default, such a bit-field is signed, because this is consistent: the
2139 basic integer types such as @code{int} are signed types.
2141 @item -fsso-struct=@var{endianness}
2142 @opindex fsso-struct
2143 Set the default scalar storage order of structures and unions to the
2144 specified endianness. The accepted values are @samp{big-endian} and
2145 @samp{little-endian}. If the option is not passed, the compiler uses
2146 the native endianness of the target. This option is not supported for C++.
2148 @strong{Warning:} the @option{-fsso-struct} switch causes GCC to generate
2149 code that is not binary compatible with code generated without it if the
2150 specified endianness is not the native endianness of the target.
2153 @node C++ Dialect Options
2154 @section Options Controlling C++ Dialect
2156 @cindex compiler options, C++
2157 @cindex C++ options, command-line
2158 @cindex options, C++
2159 This section describes the command-line options that are only meaningful
2160 for C++ programs. You can also use most of the GNU compiler options
2161 regardless of what language your program is in. For example, you
2162 might compile a file @file{firstClass.C} like this:
2165 g++ -g -fstrict-enums -O -c firstClass.C
2169 In this example, only @option{-fstrict-enums} is an option meant
2170 only for C++ programs; you can use the other options with any
2171 language supported by GCC@.
2173 Some options for compiling C programs, such as @option{-std}, are also
2174 relevant for C++ programs.
2175 @xref{C Dialect Options,,Options Controlling C Dialect}.
2177 Here is a list of options that are @emph{only} for compiling C++ programs:
2181 @item -fabi-version=@var{n}
2182 @opindex fabi-version
2183 Use version @var{n} of the C++ ABI@. The default is version 0.
2185 Version 0 refers to the version conforming most closely to
2186 the C++ ABI specification. Therefore, the ABI obtained using version 0
2187 will change in different versions of G++ as ABI bugs are fixed.
2189 Version 1 is the version of the C++ ABI that first appeared in G++ 3.2.
2191 Version 2 is the version of the C++ ABI that first appeared in G++
2192 3.4, and was the default through G++ 4.9.
2194 Version 3 corrects an error in mangling a constant address as a
2197 Version 4, which first appeared in G++ 4.5, implements a standard
2198 mangling for vector types.
2200 Version 5, which first appeared in G++ 4.6, corrects the mangling of
2201 attribute const/volatile on function pointer types, decltype of a
2202 plain decl, and use of a function parameter in the declaration of
2205 Version 6, which first appeared in G++ 4.7, corrects the promotion
2206 behavior of C++11 scoped enums and the mangling of template argument
2207 packs, const/static_cast, prefix ++ and --, and a class scope function
2208 used as a template argument.
2210 Version 7, which first appeared in G++ 4.8, that treats nullptr_t as a
2211 builtin type and corrects the mangling of lambdas in default argument
2214 Version 8, which first appeared in G++ 4.9, corrects the substitution
2215 behavior of function types with function-cv-qualifiers.
2217 Version 9, which first appeared in G++ 5.2, corrects the alignment of
2220 Version 10, which first appeared in G++ 6.1, adds mangling of
2221 attributes that affect type identity, such as ia32 calling convention
2222 attributes (e.g. @samp{stdcall}).
2224 Version 11, which first appeared in G++ 7, corrects the mangling of
2225 sizeof... expressions. It also implies
2226 @option{-fnew-inheriting-ctors}.
2228 See also @option{-Wabi}.
2230 @item -fabi-compat-version=@var{n}
2231 @opindex fabi-compat-version
2232 On targets that support strong aliases, G++
2233 works around mangling changes by creating an alias with the correct
2234 mangled name when defining a symbol with an incorrect mangled name.
2235 This switch specifies which ABI version to use for the alias.
2237 With @option{-fabi-version=0} (the default), this defaults to 8 (GCC 5
2238 compatibility). If another ABI version is explicitly selected, this
2239 defaults to 0. For compatibility with GCC versions 3.2 through 4.9,
2240 use @option{-fabi-compat-version=2}.
2242 If this option is not provided but @option{-Wabi=@var{n}} is, that
2243 version is used for compatibility aliases. If this option is provided
2244 along with @option{-Wabi} (without the version), the version from this
2245 option is used for the warning.
2247 @item -fno-access-control
2248 @opindex fno-access-control
2249 Turn off all access checking. This switch is mainly useful for working
2250 around bugs in the access control code.
2253 @opindex faligned-new
2254 Enable support for C++17 @code{new} of types that require more
2255 alignment than @code{void* ::operator new(std::size_t)} provides. A
2256 numeric argument such as @code{-faligned-new=32} can be used to
2257 specify how much alignment (in bytes) is provided by that function,
2258 but few users will need to override the default of
2259 @code{alignof(std::max_align_t)}.
2263 Check that the pointer returned by @code{operator new} is non-null
2264 before attempting to modify the storage allocated. This check is
2265 normally unnecessary because the C++ standard specifies that
2266 @code{operator new} only returns @code{0} if it is declared
2267 @code{throw()}, in which case the compiler always checks the
2268 return value even without this option. In all other cases, when
2269 @code{operator new} has a non-empty exception specification, memory
2270 exhaustion is signalled by throwing @code{std::bad_alloc}. See also
2271 @samp{new (nothrow)}.
2275 Enable support for the C++ Extensions for Concepts Technical
2276 Specification, ISO 19217 (2015), which allows code like
2279 template <class T> concept bool Addable = requires (T t) @{ t + t; @};
2280 template <Addable T> T add (T a, T b) @{ return a + b; @}
2283 @item -fconstexpr-depth=@var{n}
2284 @opindex fconstexpr-depth
2285 Set the maximum nested evaluation depth for C++11 constexpr functions
2286 to @var{n}. A limit is needed to detect endless recursion during
2287 constant expression evaluation. The minimum specified by the standard
2290 @item -fconstexpr-loop-limit=@var{n}
2291 @opindex fconstexpr-loop-limit
2292 Set the maximum number of iterations for a loop in C++14 constexpr functions
2293 to @var{n}. A limit is needed to detect infinite loops during
2294 constant expression evaluation. The default is 262144 (1<<18).
2296 @item -fdeduce-init-list
2297 @opindex fdeduce-init-list
2298 Enable deduction of a template type parameter as
2299 @code{std::initializer_list} from a brace-enclosed initializer list, i.e.@:
2302 template <class T> auto forward(T t) -> decltype (realfn (t))
2309 forward(@{1,2@}); // call forward<std::initializer_list<int>>
2313 This deduction was implemented as a possible extension to the
2314 originally proposed semantics for the C++11 standard, but was not part
2315 of the final standard, so it is disabled by default. This option is
2316 deprecated, and may be removed in a future version of G++.
2318 @item -ffriend-injection
2319 @opindex ffriend-injection
2320 Inject friend functions into the enclosing namespace, so that they are
2321 visible outside the scope of the class in which they are declared.
2322 Friend functions were documented to work this way in the old Annotated
2323 C++ Reference Manual.
2324 However, in ISO C++ a friend function that is not declared
2325 in an enclosing scope can only be found using argument dependent
2326 lookup. GCC defaults to the standard behavior.
2328 This option is for compatibility, and may be removed in a future
2331 @item -fno-elide-constructors
2332 @opindex fno-elide-constructors
2333 The C++ standard allows an implementation to omit creating a temporary
2334 that is only used to initialize another object of the same type.
2335 Specifying this option disables that optimization, and forces G++ to
2336 call the copy constructor in all cases. This option also causes G++
2337 to call trivial member functions which otherwise would be expanded inline.
2339 In C++17, the compiler is required to omit these temporaries, but this
2340 option still affects trivial member functions.
2342 @item -fno-enforce-eh-specs
2343 @opindex fno-enforce-eh-specs
2344 Don't generate code to check for violation of exception specifications
2345 at run time. This option violates the C++ standard, but may be useful
2346 for reducing code size in production builds, much like defining
2347 @code{NDEBUG}. This does not give user code permission to throw
2348 exceptions in violation of the exception specifications; the compiler
2349 still optimizes based on the specifications, so throwing an
2350 unexpected exception results in undefined behavior at run time.
2352 @item -fextern-tls-init
2353 @itemx -fno-extern-tls-init
2354 @opindex fextern-tls-init
2355 @opindex fno-extern-tls-init
2356 The C++11 and OpenMP standards allow @code{thread_local} and
2357 @code{threadprivate} variables to have dynamic (runtime)
2358 initialization. To support this, any use of such a variable goes
2359 through a wrapper function that performs any necessary initialization.
2360 When the use and definition of the variable are in the same
2361 translation unit, this overhead can be optimized away, but when the
2362 use is in a different translation unit there is significant overhead
2363 even if the variable doesn't actually need dynamic initialization. If
2364 the programmer can be sure that no use of the variable in a
2365 non-defining TU needs to trigger dynamic initialization (either
2366 because the variable is statically initialized, or a use of the
2367 variable in the defining TU will be executed before any uses in
2368 another TU), they can avoid this overhead with the
2369 @option{-fno-extern-tls-init} option.
2371 On targets that support symbol aliases, the default is
2372 @option{-fextern-tls-init}. On targets that do not support symbol
2373 aliases, the default is @option{-fno-extern-tls-init}.
2376 @itemx -fno-for-scope
2378 @opindex fno-for-scope
2379 If @option{-ffor-scope} is specified, the scope of variables declared in
2380 a @i{for-init-statement} is limited to the @code{for} loop itself,
2381 as specified by the C++ standard.
2382 If @option{-fno-for-scope} is specified, the scope of variables declared in
2383 a @i{for-init-statement} extends to the end of the enclosing scope,
2384 as was the case in old versions of G++, and other (traditional)
2385 implementations of C++.
2387 If neither flag is given, the default is to follow the standard,
2388 but to allow and give a warning for old-style code that would
2389 otherwise be invalid, or have different behavior.
2391 @item -fno-gnu-keywords
2392 @opindex fno-gnu-keywords
2393 Do not recognize @code{typeof} as a keyword, so that code can use this
2394 word as an identifier. You can use the keyword @code{__typeof__} instead.
2395 This option is implied by the strict ISO C++ dialects: @option{-ansi},
2396 @option{-std=c++98}, @option{-std=c++11}, etc.
2398 @item -fno-implicit-templates
2399 @opindex fno-implicit-templates
2400 Never emit code for non-inline templates that are instantiated
2401 implicitly (i.e.@: by use); only emit code for explicit instantiations.
2402 @xref{Template Instantiation}, for more information.
2404 @item -fno-implicit-inline-templates
2405 @opindex fno-implicit-inline-templates
2406 Don't emit code for implicit instantiations of inline templates, either.
2407 The default is to handle inlines differently so that compiles with and
2408 without optimization need the same set of explicit instantiations.
2410 @item -fno-implement-inlines
2411 @opindex fno-implement-inlines
2412 To save space, do not emit out-of-line copies of inline functions
2413 controlled by @code{#pragma implementation}. This causes linker
2414 errors if these functions are not inlined everywhere they are called.
2416 @item -fms-extensions
2417 @opindex fms-extensions
2418 Disable Wpedantic warnings about constructs used in MFC, such as implicit
2419 int and getting a pointer to member function via non-standard syntax.
2421 @item -fnew-inheriting-ctors
2422 @opindex fnew-inheriting-ctors
2423 Enable the P0136 adjustment to the semantics of C++11 constructor
2424 inheritance. This is part of C++17 but also considered to be a Defect
2425 Report against C++11 and C++14. This flag is enabled by default
2426 unless @option{-fabi-version=10} or lower is specified.
2428 @item -fno-nonansi-builtins
2429 @opindex fno-nonansi-builtins
2430 Disable built-in declarations of functions that are not mandated by
2431 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
2432 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
2435 @opindex fnothrow-opt
2436 Treat a @code{throw()} exception specification as if it were a
2437 @code{noexcept} specification to reduce or eliminate the text size
2438 overhead relative to a function with no exception specification. If
2439 the function has local variables of types with non-trivial
2440 destructors, the exception specification actually makes the
2441 function smaller because the EH cleanups for those variables can be
2442 optimized away. The semantic effect is that an exception thrown out of
2443 a function with such an exception specification results in a call
2444 to @code{terminate} rather than @code{unexpected}.
2446 @item -fno-operator-names
2447 @opindex fno-operator-names
2448 Do not treat the operator name keywords @code{and}, @code{bitand},
2449 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
2450 synonyms as keywords.
2452 @item -fno-optional-diags
2453 @opindex fno-optional-diags
2454 Disable diagnostics that the standard says a compiler does not need to
2455 issue. Currently, the only such diagnostic issued by G++ is the one for
2456 a name having multiple meanings within a class.
2459 @opindex fpermissive
2460 Downgrade some diagnostics about nonconformant code from errors to
2461 warnings. Thus, using @option{-fpermissive} allows some
2462 nonconforming code to compile.
2464 @item -fno-pretty-templates
2465 @opindex fno-pretty-templates
2466 When an error message refers to a specialization of a function
2467 template, the compiler normally prints the signature of the
2468 template followed by the template arguments and any typedefs or
2469 typenames in the signature (e.g. @code{void f(T) [with T = int]}
2470 rather than @code{void f(int)}) so that it's clear which template is
2471 involved. When an error message refers to a specialization of a class
2472 template, the compiler omits any template arguments that match
2473 the default template arguments for that template. If either of these
2474 behaviors make it harder to understand the error message rather than
2475 easier, you can use @option{-fno-pretty-templates} to disable them.
2479 Enable automatic template instantiation at link time. This option also
2480 implies @option{-fno-implicit-templates}. @xref{Template
2481 Instantiation}, for more information.
2485 Disable generation of information about every class with virtual
2486 functions for use by the C++ run-time type identification features
2487 (@code{dynamic_cast} and @code{typeid}). If you don't use those parts
2488 of the language, you can save some space by using this flag. Note that
2489 exception handling uses the same information, but G++ generates it as
2490 needed. The @code{dynamic_cast} operator can still be used for casts that
2491 do not require run-time type information, i.e.@: casts to @code{void *} or to
2492 unambiguous base classes.
2494 @item -fsized-deallocation
2495 @opindex fsized-deallocation
2496 Enable the built-in global declarations
2498 void operator delete (void *, std::size_t) noexcept;
2499 void operator delete[] (void *, std::size_t) noexcept;
2501 as introduced in C++14. This is useful for user-defined replacement
2502 deallocation functions that, for example, use the size of the object
2503 to make deallocation faster. Enabled by default under
2504 @option{-std=c++14} and above. The flag @option{-Wsized-deallocation}
2505 warns about places that might want to add a definition.
2507 @item -fstrict-enums
2508 @opindex fstrict-enums
2509 Allow the compiler to optimize using the assumption that a value of
2510 enumerated type can only be one of the values of the enumeration (as
2511 defined in the C++ standard; basically, a value that can be
2512 represented in the minimum number of bits needed to represent all the
2513 enumerators). This assumption may not be valid if the program uses a
2514 cast to convert an arbitrary integer value to the enumerated type.
2516 @item -fstrong-eval-order
2517 @opindex fstrong-eval-order
2518 Evaluate member access, array subscripting, and shift expressions in
2519 left-to-right order, and evaluate assignment in right-to-left order,
2520 as adopted for C++17. Enabled by default with @option{-std=c++1z}.
2521 @option{-fstrong-eval-order=some} enables just the ordering of member
2522 access and shift expressions, and is the default without
2523 @option{-std=c++1z}.
2525 @item -ftemplate-backtrace-limit=@var{n}
2526 @opindex ftemplate-backtrace-limit
2527 Set the maximum number of template instantiation notes for a single
2528 warning or error to @var{n}. The default value is 10.
2530 @item -ftemplate-depth=@var{n}
2531 @opindex ftemplate-depth
2532 Set the maximum instantiation depth for template classes to @var{n}.
2533 A limit on the template instantiation depth is needed to detect
2534 endless recursions during template class instantiation. ANSI/ISO C++
2535 conforming programs must not rely on a maximum depth greater than 17
2536 (changed to 1024 in C++11). The default value is 900, as the compiler
2537 can run out of stack space before hitting 1024 in some situations.
2539 @item -fno-threadsafe-statics
2540 @opindex fno-threadsafe-statics
2541 Do not emit the extra code to use the routines specified in the C++
2542 ABI for thread-safe initialization of local statics. You can use this
2543 option to reduce code size slightly in code that doesn't need to be
2546 @item -fuse-cxa-atexit
2547 @opindex fuse-cxa-atexit
2548 Register destructors for objects with static storage duration with the
2549 @code{__cxa_atexit} function rather than the @code{atexit} function.
2550 This option is required for fully standards-compliant handling of static
2551 destructors, but only works if your C library supports
2552 @code{__cxa_atexit}.
2554 @item -fno-use-cxa-get-exception-ptr
2555 @opindex fno-use-cxa-get-exception-ptr
2556 Don't use the @code{__cxa_get_exception_ptr} runtime routine. This
2557 causes @code{std::uncaught_exception} to be incorrect, but is necessary
2558 if the runtime routine is not available.
2560 @item -fvisibility-inlines-hidden
2561 @opindex fvisibility-inlines-hidden
2562 This switch declares that the user does not attempt to compare
2563 pointers to inline functions or methods where the addresses of the two functions
2564 are taken in different shared objects.
2566 The effect of this is that GCC may, effectively, mark inline methods with
2567 @code{__attribute__ ((visibility ("hidden")))} so that they do not
2568 appear in the export table of a DSO and do not require a PLT indirection
2569 when used within the DSO@. Enabling this option can have a dramatic effect
2570 on load and link times of a DSO as it massively reduces the size of the
2571 dynamic export table when the library makes heavy use of templates.
2573 The behavior of this switch is not quite the same as marking the
2574 methods as hidden directly, because it does not affect static variables
2575 local to the function or cause the compiler to deduce that
2576 the function is defined in only one shared object.
2578 You may mark a method as having a visibility explicitly to negate the
2579 effect of the switch for that method. For example, if you do want to
2580 compare pointers to a particular inline method, you might mark it as
2581 having default visibility. Marking the enclosing class with explicit
2582 visibility has no effect.
2584 Explicitly instantiated inline methods are unaffected by this option
2585 as their linkage might otherwise cross a shared library boundary.
2586 @xref{Template Instantiation}.
2588 @item -fvisibility-ms-compat
2589 @opindex fvisibility-ms-compat
2590 This flag attempts to use visibility settings to make GCC's C++
2591 linkage model compatible with that of Microsoft Visual Studio.
2593 The flag makes these changes to GCC's linkage model:
2597 It sets the default visibility to @code{hidden}, like
2598 @option{-fvisibility=hidden}.
2601 Types, but not their members, are not hidden by default.
2604 The One Definition Rule is relaxed for types without explicit
2605 visibility specifications that are defined in more than one
2606 shared object: those declarations are permitted if they are
2607 permitted when this option is not used.
2610 In new code it is better to use @option{-fvisibility=hidden} and
2611 export those classes that are intended to be externally visible.
2612 Unfortunately it is possible for code to rely, perhaps accidentally,
2613 on the Visual Studio behavior.
2615 Among the consequences of these changes are that static data members
2616 of the same type with the same name but defined in different shared
2617 objects are different, so changing one does not change the other;
2618 and that pointers to function members defined in different shared
2619 objects may not compare equal. When this flag is given, it is a
2620 violation of the ODR to define types with the same name differently.
2624 Do not use weak symbol support, even if it is provided by the linker.
2625 By default, G++ uses weak symbols if they are available. This
2626 option exists only for testing, and should not be used by end-users;
2627 it results in inferior code and has no benefits. This option may
2628 be removed in a future release of G++.
2632 Do not search for header files in the standard directories specific to
2633 C++, but do still search the other standard directories. (This option
2634 is used when building the C++ library.)
2637 In addition, these optimization, warning, and code generation options
2638 have meanings only for C++ programs:
2641 @item -Wabi @r{(C, Objective-C, C++ and Objective-C++ only)}
2644 Warn when G++ it generates code that is probably not compatible with
2645 the vendor-neutral C++ ABI@. Since G++ now defaults to updating the
2646 ABI with each major release, normally @option{-Wabi} will warn only if
2647 there is a check added later in a release series for an ABI issue
2648 discovered since the initial release. @option{-Wabi} will warn about
2649 more things if an older ABI version is selected (with
2650 @option{-fabi-version=@var{n}}).
2652 @option{-Wabi} can also be used with an explicit version number to
2653 warn about compatibility with a particular @option{-fabi-version}
2654 level, e.g. @option{-Wabi=2} to warn about changes relative to
2655 @option{-fabi-version=2}.
2657 If an explicit version number is provided and
2658 @option{-fabi-compat-version} is not specified, the version number
2659 from this option is used for compatibility aliases. If no explicit
2660 version number is provided with this option, but
2661 @option{-fabi-compat-version} is specified, that version number is
2662 used for ABI warnings.
2664 Although an effort has been made to warn about
2665 all such cases, there are probably some cases that are not warned about,
2666 even though G++ is generating incompatible code. There may also be
2667 cases where warnings are emitted even though the code that is generated
2670 You should rewrite your code to avoid these warnings if you are
2671 concerned about the fact that code generated by G++ may not be binary
2672 compatible with code generated by other compilers.
2674 Known incompatibilities in @option{-fabi-version=2} (which was the
2675 default from GCC 3.4 to 4.9) include:
2680 A template with a non-type template parameter of reference type was
2681 mangled incorrectly:
2684 template <int &> struct S @{@};
2688 This was fixed in @option{-fabi-version=3}.
2691 SIMD vector types declared using @code{__attribute ((vector_size))} were
2692 mangled in a non-standard way that does not allow for overloading of
2693 functions taking vectors of different sizes.
2695 The mangling was changed in @option{-fabi-version=4}.
2698 @code{__attribute ((const))} and @code{noreturn} were mangled as type
2699 qualifiers, and @code{decltype} of a plain declaration was folded away.
2701 These mangling issues were fixed in @option{-fabi-version=5}.
2704 Scoped enumerators passed as arguments to a variadic function are
2705 promoted like unscoped enumerators, causing @code{va_arg} to complain.
2706 On most targets this does not actually affect the parameter passing
2707 ABI, as there is no way to pass an argument smaller than @code{int}.
2709 Also, the ABI changed the mangling of template argument packs,
2710 @code{const_cast}, @code{static_cast}, prefix increment/decrement, and
2711 a class scope function used as a template argument.
2713 These issues were corrected in @option{-fabi-version=6}.
2716 Lambdas in default argument scope were mangled incorrectly, and the
2717 ABI changed the mangling of @code{nullptr_t}.
2719 These issues were corrected in @option{-fabi-version=7}.
2722 When mangling a function type with function-cv-qualifiers, the
2723 un-qualified function type was incorrectly treated as a substitution
2726 This was fixed in @option{-fabi-version=8}, the default for GCC 5.1.
2729 @code{decltype(nullptr)} incorrectly had an alignment of 1, leading to
2730 unaligned accesses. Note that this did not affect the ABI of a
2731 function with a @code{nullptr_t} parameter, as parameters have a
2734 This was fixed in @option{-fabi-version=9}, the default for GCC 5.2.
2737 Target-specific attributes that affect the identity of a type, such as
2738 ia32 calling conventions on a function type (stdcall, regparm, etc.),
2739 did not affect the mangled name, leading to name collisions when
2740 function pointers were used as template arguments.
2742 This was fixed in @option{-fabi-version=10}, the default for GCC 6.1.
2746 It also warns about psABI-related changes. The known psABI changes at this
2752 For SysV/x86-64, unions with @code{long double} members are
2753 passed in memory as specified in psABI. For example:
2763 @code{union U} is always passed in memory.
2767 @item -Wabi-tag @r{(C++ and Objective-C++ only)}
2770 Warn when a type with an ABI tag is used in a context that does not
2771 have that ABI tag. See @ref{C++ Attributes} for more information
2774 @item -Wctor-dtor-privacy @r{(C++ and Objective-C++ only)}
2775 @opindex Wctor-dtor-privacy
2776 @opindex Wno-ctor-dtor-privacy
2777 Warn when a class seems unusable because all the constructors or
2778 destructors in that class are private, and it has neither friends nor
2779 public static member functions. Also warn if there are no non-private
2780 methods, and there's at least one private member function that isn't
2781 a constructor or destructor.
2783 @item -Wdelete-non-virtual-dtor @r{(C++ and Objective-C++ only)}
2784 @opindex Wdelete-non-virtual-dtor
2785 @opindex Wno-delete-non-virtual-dtor
2786 Warn when @code{delete} is used to destroy an instance of a class that
2787 has virtual functions and non-virtual destructor. It is unsafe to delete
2788 an instance of a derived class through a pointer to a base class if the
2789 base class does not have a virtual destructor. This warning is enabled
2792 @item -Wliteral-suffix @r{(C++ and Objective-C++ only)}
2793 @opindex Wliteral-suffix
2794 @opindex Wno-literal-suffix
2795 Warn when a string or character literal is followed by a ud-suffix which does
2796 not begin with an underscore. As a conforming extension, GCC treats such
2797 suffixes as separate preprocessing tokens in order to maintain backwards
2798 compatibility with code that uses formatting macros from @code{<inttypes.h>}.
2802 #define __STDC_FORMAT_MACROS
2803 #include <inttypes.h>
2808 printf("My int64: %" PRId64"\n", i64);
2812 In this case, @code{PRId64} is treated as a separate preprocessing token.
2814 This warning is enabled by default.
2816 @item -Wlto-type-mismatch
2817 @opindex Wlto-type-mismatch
2818 @opindex Wno-lto-type-mismatch
2820 During the link-time optimization warn about type mismatches in
2821 global declarations from different compilation units.
2822 Requires @option{-flto} to be enabled. Enabled by default.
2824 @item -Wnarrowing @r{(C++ and Objective-C++ only)}
2826 @opindex Wno-narrowing
2827 With @option{-std=gnu++98} or @option{-std=c++98}, warn when a narrowing
2828 conversion prohibited by C++11 occurs within
2832 int i = @{ 2.2 @}; // error: narrowing from double to int
2835 This flag is included in @option{-Wall} and @option{-Wc++11-compat}.
2837 When a later standard is in effect, e.g. when using @option{-std=c++11},
2838 narrowing conversions are diagnosed by default, as required by the standard.
2839 A narrowing conversion from a constant produces an error,
2840 and a narrowing conversion from a non-constant produces a warning,
2841 but @option{-Wno-narrowing} suppresses the diagnostic.
2842 Note that this does not affect the meaning of well-formed code;
2843 narrowing conversions are still considered ill-formed in SFINAE contexts.
2845 @item -Wnoexcept @r{(C++ and Objective-C++ only)}
2847 @opindex Wno-noexcept
2848 Warn when a noexcept-expression evaluates to false because of a call
2849 to a function that does not have a non-throwing exception
2850 specification (i.e. @code{throw()} or @code{noexcept}) but is known by
2851 the compiler to never throw an exception.
2853 @item -Wnon-virtual-dtor @r{(C++ and Objective-C++ only)}
2854 @opindex Wnon-virtual-dtor
2855 @opindex Wno-non-virtual-dtor
2856 Warn when a class has virtual functions and an accessible non-virtual
2857 destructor itself or in an accessible polymorphic base class, in which
2858 case it is possible but unsafe to delete an instance of a derived
2859 class through a pointer to the class itself or base class. This
2860 warning is automatically enabled if @option{-Weffc++} is specified.
2862 @item -Wregister @r{(C++ and Objective-C++ only)}
2864 @opindex Wno-register
2865 Warn on uses of the @code{register} storage class specifier, except
2866 when it is part of the GNU @ref{Explicit Register Variables} extension.
2867 The use of the @code{register} keyword as storage class specifier has
2868 been deprecated in C++11 and removed in C++17.
2869 Enabled by default with @option{-std=c++1z}.
2871 @item -Wreorder @r{(C++ and Objective-C++ only)}
2873 @opindex Wno-reorder
2874 @cindex reordering, warning
2875 @cindex warning for reordering of member initializers
2876 Warn when the order of member initializers given in the code does not
2877 match the order in which they must be executed. For instance:
2883 A(): j (0), i (1) @{ @}
2888 The compiler rearranges the member initializers for @code{i}
2889 and @code{j} to match the declaration order of the members, emitting
2890 a warning to that effect. This warning is enabled by @option{-Wall}.
2892 @item -fext-numeric-literals @r{(C++ and Objective-C++ only)}
2893 @opindex fext-numeric-literals
2894 @opindex fno-ext-numeric-literals
2895 Accept imaginary, fixed-point, or machine-defined
2896 literal number suffixes as GNU extensions.
2897 When this option is turned off these suffixes are treated
2898 as C++11 user-defined literal numeric suffixes.
2899 This is on by default for all pre-C++11 dialects and all GNU dialects:
2900 @option{-std=c++98}, @option{-std=gnu++98}, @option{-std=gnu++11},
2901 @option{-std=gnu++14}.
2902 This option is off by default
2903 for ISO C++11 onwards (@option{-std=c++11}, ...).
2906 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
2909 @item -Weffc++ @r{(C++ and Objective-C++ only)}
2912 Warn about violations of the following style guidelines from Scott Meyers'
2913 @cite{Effective C++} series of books:
2917 Define a copy constructor and an assignment operator for classes
2918 with dynamically-allocated memory.
2921 Prefer initialization to assignment in constructors.
2924 Have @code{operator=} return a reference to @code{*this}.
2927 Don't try to return a reference when you must return an object.
2930 Distinguish between prefix and postfix forms of increment and
2931 decrement operators.
2934 Never overload @code{&&}, @code{||}, or @code{,}.
2938 This option also enables @option{-Wnon-virtual-dtor}, which is also
2939 one of the effective C++ recommendations. However, the check is
2940 extended to warn about the lack of virtual destructor in accessible
2941 non-polymorphic bases classes too.
2943 When selecting this option, be aware that the standard library
2944 headers do not obey all of these guidelines; use @samp{grep -v}
2945 to filter out those warnings.
2947 @item -Wstrict-null-sentinel @r{(C++ and Objective-C++ only)}
2948 @opindex Wstrict-null-sentinel
2949 @opindex Wno-strict-null-sentinel
2950 Warn about the use of an uncasted @code{NULL} as sentinel. When
2951 compiling only with GCC this is a valid sentinel, as @code{NULL} is defined
2952 to @code{__null}. Although it is a null pointer constant rather than a
2953 null pointer, it is guaranteed to be of the same size as a pointer.
2954 But this use is not portable across different compilers.
2956 @item -Wno-non-template-friend @r{(C++ and Objective-C++ only)}
2957 @opindex Wno-non-template-friend
2958 @opindex Wnon-template-friend
2959 Disable warnings when non-templatized friend functions are declared
2960 within a template. Since the advent of explicit template specification
2961 support in G++, if the name of the friend is an unqualified-id (i.e.,
2962 @samp{friend foo(int)}), the C++ language specification demands that the
2963 friend declare or define an ordinary, nontemplate function. (Section
2964 14.5.3). Before G++ implemented explicit specification, unqualified-ids
2965 could be interpreted as a particular specialization of a templatized
2966 function. Because this non-conforming behavior is no longer the default
2967 behavior for G++, @option{-Wnon-template-friend} allows the compiler to
2968 check existing code for potential trouble spots and is on by default.
2969 This new compiler behavior can be turned off with
2970 @option{-Wno-non-template-friend}, which keeps the conformant compiler code
2971 but disables the helpful warning.
2973 @item -Wold-style-cast @r{(C++ and Objective-C++ only)}
2974 @opindex Wold-style-cast
2975 @opindex Wno-old-style-cast
2976 Warn if an old-style (C-style) cast to a non-void type is used within
2977 a C++ program. The new-style casts (@code{dynamic_cast},
2978 @code{static_cast}, @code{reinterpret_cast}, and @code{const_cast}) are
2979 less vulnerable to unintended effects and much easier to search for.
2981 @item -Woverloaded-virtual @r{(C++ and Objective-C++ only)}
2982 @opindex Woverloaded-virtual
2983 @opindex Wno-overloaded-virtual
2984 @cindex overloaded virtual function, warning
2985 @cindex warning for overloaded virtual function
2986 Warn when a function declaration hides virtual functions from a
2987 base class. For example, in:
2994 struct B: public A @{
2999 the @code{A} class version of @code{f} is hidden in @code{B}, and code
3010 @item -Wno-pmf-conversions @r{(C++ and Objective-C++ only)}
3011 @opindex Wno-pmf-conversions
3012 @opindex Wpmf-conversions
3013 Disable the diagnostic for converting a bound pointer to member function
3016 @item -Wsign-promo @r{(C++ and Objective-C++ only)}
3017 @opindex Wsign-promo
3018 @opindex Wno-sign-promo
3019 Warn when overload resolution chooses a promotion from unsigned or
3020 enumerated type to a signed type, over a conversion to an unsigned type of
3021 the same size. Previous versions of G++ tried to preserve
3022 unsignedness, but the standard mandates the current behavior.
3024 @item -Wtemplates @r{(C++ and Objective-C++ only)}
3026 Warn when a primary template declaration is encountered. Some coding
3027 rules disallow templates, and this may be used to enforce that rule.
3028 The warning is inactive inside a system header file, such as the STL, so
3029 one can still use the STL. One may also instantiate or specialize
3032 @item -Wmultiple-inheritance @r{(C++ and Objective-C++ only)}
3033 @opindex Wmultiple-inheritance
3034 Warn when a class is defined with multiple direct base classes. Some
3035 coding rules disallow multiple inheritance, and this may be used to
3036 enforce that rule. The warning is inactive inside a system header file,
3037 such as the STL, so one can still use the STL. One may also define
3038 classes that indirectly use multiple inheritance.
3040 @item -Wvirtual-inheritance
3041 @opindex Wvirtual-inheritance
3042 Warn when a class is defined with a virtual direct base classe. Some
3043 coding rules disallow multiple inheritance, and this may be used to
3044 enforce that rule. The warning is inactive inside a system header file,
3045 such as the STL, so one can still use the STL. One may also define
3046 classes that indirectly use virtual inheritance.
3049 @opindex Wnamespaces
3050 Warn when a namespace definition is opened. Some coding rules disallow
3051 namespaces, and this may be used to enforce that rule. The warning is
3052 inactive inside a system header file, such as the STL, so one can still
3053 use the STL. One may also use using directives and qualified names.
3055 @item -Wno-terminate @r{(C++ and Objective-C++ only)}
3057 @opindex Wno-terminate
3058 Disable the warning about a throw-expression that will immediately
3059 result in a call to @code{terminate}.
3062 @node Objective-C and Objective-C++ Dialect Options
3063 @section Options Controlling Objective-C and Objective-C++ Dialects
3065 @cindex compiler options, Objective-C and Objective-C++
3066 @cindex Objective-C and Objective-C++ options, command-line
3067 @cindex options, Objective-C and Objective-C++
3068 (NOTE: This manual does not describe the Objective-C and Objective-C++
3069 languages themselves. @xref{Standards,,Language Standards
3070 Supported by GCC}, for references.)
3072 This section describes the command-line options that are only meaningful
3073 for Objective-C and Objective-C++ programs. You can also use most of
3074 the language-independent GNU compiler options.
3075 For example, you might compile a file @file{some_class.m} like this:
3078 gcc -g -fgnu-runtime -O -c some_class.m
3082 In this example, @option{-fgnu-runtime} is an option meant only for
3083 Objective-C and Objective-C++ programs; you can use the other options with
3084 any language supported by GCC@.
3086 Note that since Objective-C is an extension of the C language, Objective-C
3087 compilations may also use options specific to the C front-end (e.g.,
3088 @option{-Wtraditional}). Similarly, Objective-C++ compilations may use
3089 C++-specific options (e.g., @option{-Wabi}).
3091 Here is a list of options that are @emph{only} for compiling Objective-C
3092 and Objective-C++ programs:
3095 @item -fconstant-string-class=@var{class-name}
3096 @opindex fconstant-string-class
3097 Use @var{class-name} as the name of the class to instantiate for each
3098 literal string specified with the syntax @code{@@"@dots{}"}. The default
3099 class name is @code{NXConstantString} if the GNU runtime is being used, and
3100 @code{NSConstantString} if the NeXT runtime is being used (see below). The
3101 @option{-fconstant-cfstrings} option, if also present, overrides the
3102 @option{-fconstant-string-class} setting and cause @code{@@"@dots{}"} literals
3103 to be laid out as constant CoreFoundation strings.
3106 @opindex fgnu-runtime
3107 Generate object code compatible with the standard GNU Objective-C
3108 runtime. This is the default for most types of systems.
3110 @item -fnext-runtime
3111 @opindex fnext-runtime
3112 Generate output compatible with the NeXT runtime. This is the default
3113 for NeXT-based systems, including Darwin and Mac OS X@. The macro
3114 @code{__NEXT_RUNTIME__} is predefined if (and only if) this option is
3117 @item -fno-nil-receivers
3118 @opindex fno-nil-receivers
3119 Assume that all Objective-C message dispatches (@code{[receiver
3120 message:arg]}) in this translation unit ensure that the receiver is
3121 not @code{nil}. This allows for more efficient entry points in the
3122 runtime to be used. This option is only available in conjunction with
3123 the NeXT runtime and ABI version 0 or 1.
3125 @item -fobjc-abi-version=@var{n}
3126 @opindex fobjc-abi-version
3127 Use version @var{n} of the Objective-C ABI for the selected runtime.
3128 This option is currently supported only for the NeXT runtime. In that
3129 case, Version 0 is the traditional (32-bit) ABI without support for
3130 properties and other Objective-C 2.0 additions. Version 1 is the
3131 traditional (32-bit) ABI with support for properties and other
3132 Objective-C 2.0 additions. Version 2 is the modern (64-bit) ABI. If
3133 nothing is specified, the default is Version 0 on 32-bit target
3134 machines, and Version 2 on 64-bit target machines.
3136 @item -fobjc-call-cxx-cdtors
3137 @opindex fobjc-call-cxx-cdtors
3138 For each Objective-C class, check if any of its instance variables is a
3139 C++ object with a non-trivial default constructor. If so, synthesize a
3140 special @code{- (id) .cxx_construct} instance method which runs
3141 non-trivial default constructors on any such instance variables, in order,
3142 and then return @code{self}. Similarly, check if any instance variable
3143 is a C++ object with a non-trivial destructor, and if so, synthesize a
3144 special @code{- (void) .cxx_destruct} method which runs
3145 all such default destructors, in reverse order.
3147 The @code{- (id) .cxx_construct} and @code{- (void) .cxx_destruct}
3148 methods thusly generated only operate on instance variables
3149 declared in the current Objective-C class, and not those inherited
3150 from superclasses. It is the responsibility of the Objective-C
3151 runtime to invoke all such methods in an object's inheritance
3152 hierarchy. The @code{- (id) .cxx_construct} methods are invoked
3153 by the runtime immediately after a new object instance is allocated;
3154 the @code{- (void) .cxx_destruct} methods are invoked immediately
3155 before the runtime deallocates an object instance.
3157 As of this writing, only the NeXT runtime on Mac OS X 10.4 and later has
3158 support for invoking the @code{- (id) .cxx_construct} and
3159 @code{- (void) .cxx_destruct} methods.
3161 @item -fobjc-direct-dispatch
3162 @opindex fobjc-direct-dispatch
3163 Allow fast jumps to the message dispatcher. On Darwin this is
3164 accomplished via the comm page.
3166 @item -fobjc-exceptions
3167 @opindex fobjc-exceptions
3168 Enable syntactic support for structured exception handling in
3169 Objective-C, similar to what is offered by C++ and Java. This option
3170 is required to use the Objective-C keywords @code{@@try},
3171 @code{@@throw}, @code{@@catch}, @code{@@finally} and
3172 @code{@@synchronized}. This option is available with both the GNU
3173 runtime and the NeXT runtime (but not available in conjunction with
3174 the NeXT runtime on Mac OS X 10.2 and earlier).
3178 Enable garbage collection (GC) in Objective-C and Objective-C++
3179 programs. This option is only available with the NeXT runtime; the
3180 GNU runtime has a different garbage collection implementation that
3181 does not require special compiler flags.
3183 @item -fobjc-nilcheck
3184 @opindex fobjc-nilcheck
3185 For the NeXT runtime with version 2 of the ABI, check for a nil
3186 receiver in method invocations before doing the actual method call.
3187 This is the default and can be disabled using
3188 @option{-fno-objc-nilcheck}. Class methods and super calls are never
3189 checked for nil in this way no matter what this flag is set to.
3190 Currently this flag does nothing when the GNU runtime, or an older
3191 version of the NeXT runtime ABI, is used.
3193 @item -fobjc-std=objc1
3195 Conform to the language syntax of Objective-C 1.0, the language
3196 recognized by GCC 4.0. This only affects the Objective-C additions to
3197 the C/C++ language; it does not affect conformance to C/C++ standards,
3198 which is controlled by the separate C/C++ dialect option flags. When
3199 this option is used with the Objective-C or Objective-C++ compiler,
3200 any Objective-C syntax that is not recognized by GCC 4.0 is rejected.
3201 This is useful if you need to make sure that your Objective-C code can
3202 be compiled with older versions of GCC@.
3204 @item -freplace-objc-classes
3205 @opindex freplace-objc-classes
3206 Emit a special marker instructing @command{ld(1)} not to statically link in
3207 the resulting object file, and allow @command{dyld(1)} to load it in at
3208 run time instead. This is used in conjunction with the Fix-and-Continue
3209 debugging mode, where the object file in question may be recompiled and
3210 dynamically reloaded in the course of program execution, without the need
3211 to restart the program itself. Currently, Fix-and-Continue functionality
3212 is only available in conjunction with the NeXT runtime on Mac OS X 10.3
3217 When compiling for the NeXT runtime, the compiler ordinarily replaces calls
3218 to @code{objc_getClass("@dots{}")} (when the name of the class is known at
3219 compile time) with static class references that get initialized at load time,
3220 which improves run-time performance. Specifying the @option{-fzero-link} flag
3221 suppresses this behavior and causes calls to @code{objc_getClass("@dots{}")}
3222 to be retained. This is useful in Zero-Link debugging mode, since it allows
3223 for individual class implementations to be modified during program execution.
3224 The GNU runtime currently always retains calls to @code{objc_get_class("@dots{}")}
3225 regardless of command-line options.
3227 @item -fno-local-ivars
3228 @opindex fno-local-ivars
3229 @opindex flocal-ivars
3230 By default instance variables in Objective-C can be accessed as if
3231 they were local variables from within the methods of the class they're
3232 declared in. This can lead to shadowing between instance variables
3233 and other variables declared either locally inside a class method or
3234 globally with the same name. Specifying the @option{-fno-local-ivars}
3235 flag disables this behavior thus avoiding variable shadowing issues.
3237 @item -fivar-visibility=@r{[}public@r{|}protected@r{|}private@r{|}package@r{]}
3238 @opindex fivar-visibility
3239 Set the default instance variable visibility to the specified option
3240 so that instance variables declared outside the scope of any access
3241 modifier directives default to the specified visibility.
3245 Dump interface declarations for all classes seen in the source file to a
3246 file named @file{@var{sourcename}.decl}.
3248 @item -Wassign-intercept @r{(Objective-C and Objective-C++ only)}
3249 @opindex Wassign-intercept
3250 @opindex Wno-assign-intercept
3251 Warn whenever an Objective-C assignment is being intercepted by the
3254 @item -Wno-protocol @r{(Objective-C and Objective-C++ only)}
3255 @opindex Wno-protocol
3257 If a class is declared to implement a protocol, a warning is issued for
3258 every method in the protocol that is not implemented by the class. The
3259 default behavior is to issue a warning for every method not explicitly
3260 implemented in the class, even if a method implementation is inherited
3261 from the superclass. If you use the @option{-Wno-protocol} option, then
3262 methods inherited from the superclass are considered to be implemented,
3263 and no warning is issued for them.
3265 @item -Wselector @r{(Objective-C and Objective-C++ only)}
3267 @opindex Wno-selector
3268 Warn if multiple methods of different types for the same selector are
3269 found during compilation. The check is performed on the list of methods
3270 in the final stage of compilation. Additionally, a check is performed
3271 for each selector appearing in a @code{@@selector(@dots{})}
3272 expression, and a corresponding method for that selector has been found
3273 during compilation. Because these checks scan the method table only at
3274 the end of compilation, these warnings are not produced if the final
3275 stage of compilation is not reached, for example because an error is
3276 found during compilation, or because the @option{-fsyntax-only} option is
3279 @item -Wstrict-selector-match @r{(Objective-C and Objective-C++ only)}
3280 @opindex Wstrict-selector-match
3281 @opindex Wno-strict-selector-match
3282 Warn if multiple methods with differing argument and/or return types are
3283 found for a given selector when attempting to send a message using this
3284 selector to a receiver of type @code{id} or @code{Class}. When this flag
3285 is off (which is the default behavior), the compiler omits such warnings
3286 if any differences found are confined to types that share the same size
3289 @item -Wundeclared-selector @r{(Objective-C and Objective-C++ only)}
3290 @opindex Wundeclared-selector
3291 @opindex Wno-undeclared-selector
3292 Warn if a @code{@@selector(@dots{})} expression referring to an
3293 undeclared selector is found. A selector is considered undeclared if no
3294 method with that name has been declared before the
3295 @code{@@selector(@dots{})} expression, either explicitly in an
3296 @code{@@interface} or @code{@@protocol} declaration, or implicitly in
3297 an @code{@@implementation} section. This option always performs its
3298 checks as soon as a @code{@@selector(@dots{})} expression is found,
3299 while @option{-Wselector} only performs its checks in the final stage of
3300 compilation. This also enforces the coding style convention
3301 that methods and selectors must be declared before being used.
3303 @item -print-objc-runtime-info
3304 @opindex print-objc-runtime-info
3305 Generate C header describing the largest structure that is passed by
3310 @node Diagnostic Message Formatting Options
3311 @section Options to Control Diagnostic Messages Formatting
3312 @cindex options to control diagnostics formatting
3313 @cindex diagnostic messages
3314 @cindex message formatting
3316 Traditionally, diagnostic messages have been formatted irrespective of
3317 the output device's aspect (e.g.@: its width, @dots{}). You can use the
3318 options described below
3319 to control the formatting algorithm for diagnostic messages,
3320 e.g.@: how many characters per line, how often source location
3321 information should be reported. Note that some language front ends may not
3322 honor these options.
3325 @item -fmessage-length=@var{n}
3326 @opindex fmessage-length
3327 Try to format error messages so that they fit on lines of about
3328 @var{n} characters. If @var{n} is zero, then no line-wrapping is
3329 done; each error message appears on a single line. This is the
3330 default for all front ends.
3332 @item -fdiagnostics-show-location=once
3333 @opindex fdiagnostics-show-location
3334 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
3335 reporter to emit source location information @emph{once}; that is, in
3336 case the message is too long to fit on a single physical line and has to
3337 be wrapped, the source location won't be emitted (as prefix) again,
3338 over and over, in subsequent continuation lines. This is the default
3341 @item -fdiagnostics-show-location=every-line
3342 Only meaningful in line-wrapping mode. Instructs the diagnostic
3343 messages reporter to emit the same source location information (as
3344 prefix) for physical lines that result from the process of breaking
3345 a message which is too long to fit on a single line.
3347 @item -fdiagnostics-color[=@var{WHEN}]
3348 @itemx -fno-diagnostics-color
3349 @opindex fdiagnostics-color
3350 @cindex highlight, color
3351 @vindex GCC_COLORS @r{environment variable}
3352 Use color in diagnostics. @var{WHEN} is @samp{never}, @samp{always},
3353 or @samp{auto}. The default depends on how the compiler has been configured,
3354 it can be any of the above @var{WHEN} options or also @samp{never}
3355 if @env{GCC_COLORS} environment variable isn't present in the environment,
3356 and @samp{auto} otherwise.
3357 @samp{auto} means to use color only when the standard error is a terminal.
3358 The forms @option{-fdiagnostics-color} and @option{-fno-diagnostics-color} are
3359 aliases for @option{-fdiagnostics-color=always} and
3360 @option{-fdiagnostics-color=never}, respectively.
3362 The colors are defined by the environment variable @env{GCC_COLORS}.
3363 Its value is a colon-separated list of capabilities and Select Graphic
3364 Rendition (SGR) substrings. SGR commands are interpreted by the
3365 terminal or terminal emulator. (See the section in the documentation
3366 of your text terminal for permitted values and their meanings as
3367 character attributes.) These substring values are integers in decimal
3368 representation and can be concatenated with semicolons.
3369 Common values to concatenate include
3371 @samp{4} for underline,
3373 @samp{7} for inverse,
3374 @samp{39} for default foreground color,
3375 @samp{30} to @samp{37} for foreground colors,
3376 @samp{90} to @samp{97} for 16-color mode foreground colors,
3377 @samp{38;5;0} to @samp{38;5;255}
3378 for 88-color and 256-color modes foreground colors,
3379 @samp{49} for default background color,
3380 @samp{40} to @samp{47} for background colors,
3381 @samp{100} to @samp{107} for 16-color mode background colors,
3382 and @samp{48;5;0} to @samp{48;5;255}
3383 for 88-color and 256-color modes background colors.
3385 The default @env{GCC_COLORS} is
3387 error=01;31:warning=01;35:note=01;36:range1=32:range2=34:locus=01:quote=01:\
3388 fixit-insert=32:fixit-delete=31:\
3389 diff-filename=01:diff-hunk=32:diff-delete=31:diff-insert=32
3392 where @samp{01;31} is bold red, @samp{01;35} is bold magenta,
3393 @samp{01;36} is bold cyan, @samp{32} is green, @samp{34} is blue,
3394 @samp{01} is bold, and @samp{31} is red.
3395 Setting @env{GCC_COLORS} to the empty string disables colors.
3396 Supported capabilities are as follows.
3400 @vindex error GCC_COLORS @r{capability}
3401 SGR substring for error: markers.
3404 @vindex warning GCC_COLORS @r{capability}
3405 SGR substring for warning: markers.
3408 @vindex note GCC_COLORS @r{capability}
3409 SGR substring for note: markers.
3412 @vindex range1 GCC_COLORS @r{capability}
3413 SGR substring for first additional range.
3416 @vindex range2 GCC_COLORS @r{capability}
3417 SGR substring for second additional range.
3420 @vindex locus GCC_COLORS @r{capability}
3421 SGR substring for location information, @samp{file:line} or
3422 @samp{file:line:column} etc.
3425 @vindex quote GCC_COLORS @r{capability}
3426 SGR substring for information printed within quotes.
3429 @vindex fixit-insert GCC_COLORS @r{capability}
3430 SGR substring for fix-it hints suggesting text to
3431 be inserted or replaced.
3434 @vindex fixit-delete GCC_COLORS @r{capability}
3435 SGR substring for fix-it hints suggesting text to
3438 @item diff-filename=
3439 @vindex diff-filename GCC_COLORS @r{capability}
3440 SGR substring for filename headers within generated patches.
3443 @vindex diff-hunk GCC_COLORS @r{capability}
3444 SGR substring for the starts of hunks within generated patches.
3447 @vindex diff-delete GCC_COLORS @r{capability}
3448 SGR substring for deleted lines within generated patches.
3451 @vindex diff-insert GCC_COLORS @r{capability}
3452 SGR substring for inserted lines within generated patches.
3455 @item -fno-diagnostics-show-option
3456 @opindex fno-diagnostics-show-option
3457 @opindex fdiagnostics-show-option
3458 By default, each diagnostic emitted includes text indicating the
3459 command-line option that directly controls the diagnostic (if such an
3460 option is known to the diagnostic machinery). Specifying the
3461 @option{-fno-diagnostics-show-option} flag suppresses that behavior.
3463 @item -fno-diagnostics-show-caret
3464 @opindex fno-diagnostics-show-caret
3465 @opindex fdiagnostics-show-caret
3466 By default, each diagnostic emitted includes the original source line
3467 and a caret @samp{^} indicating the column. This option suppresses this
3468 information. The source line is truncated to @var{n} characters, if
3469 the @option{-fmessage-length=n} option is given. When the output is done
3470 to the terminal, the width is limited to the width given by the
3471 @env{COLUMNS} environment variable or, if not set, to the terminal width.
3473 @item -fdiagnostics-parseable-fixits
3474 @opindex fdiagnostics-parseable-fixits
3475 Emit fix-it hints in a machine-parseable format, suitable for consumption
3476 by IDEs. For each fix-it, a line will be printed after the relevant
3477 diagnostic, starting with the string ``fix-it:''. For example:
3480 fix-it:"test.c":@{45:3-45:21@}:"gtk_widget_show_all"
3483 The location is expressed as a half-open range, expressed as a count of
3484 bytes, starting at byte 1 for the initial column. In the above example,
3485 bytes 3 through 20 of line 45 of ``test.c'' are to be replaced with the
3489 00000000011111111112222222222
3490 12345678901234567890123456789
3491 gtk_widget_showall (dlg);
3496 The filename and replacement string escape backslash as ``\\", tab as ``\t'',
3497 newline as ``\n'', double quotes as ``\"'', non-printable characters as octal
3498 (e.g. vertical tab as ``\013'').
3500 An empty replacement string indicates that the given range is to be removed.
3501 An empty range (e.g. ``45:3-45:3'') indicates that the string is to
3502 be inserted at the given position.
3504 @item -fdiagnostics-generate-patch
3505 @opindex fdiagnostics-generate-patch
3506 Print fix-it hints to stderr in unified diff format, after any diagnostics
3507 are printed. For example:
3514 void show_cb(GtkDialog *dlg)
3516 - gtk_widget_showall(dlg);
3517 + gtk_widget_show_all(dlg);
3522 The diff may or may not be colorized, following the same rules
3523 as for diagnostics (see @option{-fdiagnostics-color}).
3527 @node Warning Options
3528 @section Options to Request or Suppress Warnings
3529 @cindex options to control warnings
3530 @cindex warning messages
3531 @cindex messages, warning
3532 @cindex suppressing warnings
3534 Warnings are diagnostic messages that report constructions that
3535 are not inherently erroneous but that are risky or suggest there
3536 may have been an error.
3538 The following language-independent options do not enable specific
3539 warnings but control the kinds of diagnostics produced by GCC@.
3542 @cindex syntax checking
3544 @opindex fsyntax-only
3545 Check the code for syntax errors, but don't do anything beyond that.
3547 @item -fmax-errors=@var{n}
3548 @opindex fmax-errors
3549 Limits the maximum number of error messages to @var{n}, at which point
3550 GCC bails out rather than attempting to continue processing the source
3551 code. If @var{n} is 0 (the default), there is no limit on the number
3552 of error messages produced. If @option{-Wfatal-errors} is also
3553 specified, then @option{-Wfatal-errors} takes precedence over this
3558 Inhibit all warning messages.
3563 Make all warnings into errors.
3568 Make the specified warning into an error. The specifier for a warning
3569 is appended; for example @option{-Werror=switch} turns the warnings
3570 controlled by @option{-Wswitch} into errors. This switch takes a
3571 negative form, to be used to negate @option{-Werror} for specific
3572 warnings; for example @option{-Wno-error=switch} makes
3573 @option{-Wswitch} warnings not be errors, even when @option{-Werror}
3576 The warning message for each controllable warning includes the
3577 option that controls the warning. That option can then be used with
3578 @option{-Werror=} and @option{-Wno-error=} as described above.
3579 (Printing of the option in the warning message can be disabled using the
3580 @option{-fno-diagnostics-show-option} flag.)
3582 Note that specifying @option{-Werror=}@var{foo} automatically implies
3583 @option{-W}@var{foo}. However, @option{-Wno-error=}@var{foo} does not
3586 @item -Wfatal-errors
3587 @opindex Wfatal-errors
3588 @opindex Wno-fatal-errors
3589 This option causes the compiler to abort compilation on the first error
3590 occurred rather than trying to keep going and printing further error
3595 You can request many specific warnings with options beginning with
3596 @samp{-W}, for example @option{-Wimplicit} to request warnings on
3597 implicit declarations. Each of these specific warning options also
3598 has a negative form beginning @samp{-Wno-} to turn off warnings; for
3599 example, @option{-Wno-implicit}. This manual lists only one of the
3600 two forms, whichever is not the default. For further
3601 language-specific options also refer to @ref{C++ Dialect Options} and
3602 @ref{Objective-C and Objective-C++ Dialect Options}.
3604 Some options, such as @option{-Wall} and @option{-Wextra}, turn on other
3605 options, such as @option{-Wunused}, which may turn on further options,
3606 such as @option{-Wunused-value}. The combined effect of positive and
3607 negative forms is that more specific options have priority over less
3608 specific ones, independently of their position in the command-line. For
3609 options of the same specificity, the last one takes effect. Options
3610 enabled or disabled via pragmas (@pxref{Diagnostic Pragmas}) take effect
3611 as if they appeared at the end of the command-line.
3613 When an unrecognized warning option is requested (e.g.,
3614 @option{-Wunknown-warning}), GCC emits a diagnostic stating
3615 that the option is not recognized. However, if the @option{-Wno-} form
3616 is used, the behavior is slightly different: no diagnostic is
3617 produced for @option{-Wno-unknown-warning} unless other diagnostics
3618 are being produced. This allows the use of new @option{-Wno-} options
3619 with old compilers, but if something goes wrong, the compiler
3620 warns that an unrecognized option is present.
3627 Issue all the warnings demanded by strict ISO C and ISO C++;
3628 reject all programs that use forbidden extensions, and some other
3629 programs that do not follow ISO C and ISO C++. For ISO C, follows the
3630 version of the ISO C standard specified by any @option{-std} option used.
3632 Valid ISO C and ISO C++ programs should compile properly with or without
3633 this option (though a rare few require @option{-ansi} or a
3634 @option{-std} option specifying the required version of ISO C)@. However,
3635 without this option, certain GNU extensions and traditional C and C++
3636 features are supported as well. With this option, they are rejected.
3638 @option{-Wpedantic} does not cause warning messages for use of the
3639 alternate keywords whose names begin and end with @samp{__}. Pedantic
3640 warnings are also disabled in the expression that follows
3641 @code{__extension__}. However, only system header files should use
3642 these escape routes; application programs should avoid them.
3643 @xref{Alternate Keywords}.
3645 Some users try to use @option{-Wpedantic} to check programs for strict ISO
3646 C conformance. They soon find that it does not do quite what they want:
3647 it finds some non-ISO practices, but not all---only those for which
3648 ISO C @emph{requires} a diagnostic, and some others for which
3649 diagnostics have been added.
3651 A feature to report any failure to conform to ISO C might be useful in
3652 some instances, but would require considerable additional work and would
3653 be quite different from @option{-Wpedantic}. We don't have plans to
3654 support such a feature in the near future.
3656 Where the standard specified with @option{-std} represents a GNU
3657 extended dialect of C, such as @samp{gnu90} or @samp{gnu99}, there is a
3658 corresponding @dfn{base standard}, the version of ISO C on which the GNU
3659 extended dialect is based. Warnings from @option{-Wpedantic} are given
3660 where they are required by the base standard. (It does not make sense
3661 for such warnings to be given only for features not in the specified GNU
3662 C dialect, since by definition the GNU dialects of C include all
3663 features the compiler supports with the given option, and there would be
3664 nothing to warn about.)
3666 @item -pedantic-errors
3667 @opindex pedantic-errors
3668 Give an error whenever the @dfn{base standard} (see @option{-Wpedantic})
3669 requires a diagnostic, in some cases where there is undefined behavior
3670 at compile-time and in some other cases that do not prevent compilation
3671 of programs that are valid according to the standard. This is not
3672 equivalent to @option{-Werror=pedantic}, since there are errors enabled
3673 by this option and not enabled by the latter and vice versa.
3678 This enables all the warnings about constructions that some users
3679 consider questionable, and that are easy to avoid (or modify to
3680 prevent the warning), even in conjunction with macros. This also
3681 enables some language-specific warnings described in @ref{C++ Dialect
3682 Options} and @ref{Objective-C and Objective-C++ Dialect Options}.
3684 @option{-Wall} turns on the following warning flags:
3686 @gccoptlist{-Waddress @gol
3687 -Warray-bounds=1 @r{(only with} @option{-O2}@r{)} @gol
3689 -Wbool-operation @gol
3690 -Wc++11-compat -Wc++14-compat@gol
3691 -Wchar-subscripts @gol
3693 -Wduplicate-decl-specifier @r{(C and Objective-C only)} @gol
3694 -Wenum-compare @r{(in C/ObjC; this is on by default in C++)} @gol
3696 -Wint-in-bool-context @gol
3697 -Wimplicit @r{(C and Objective-C only)} @gol
3698 -Wimplicit-int @r{(C and Objective-C only)} @gol
3699 -Wimplicit-function-declaration @r{(C and Objective-C only)} @gol
3700 -Winit-self @r{(only for C++)} @gol
3701 -Wlogical-not-parentheses
3702 -Wmain @r{(only for C/ObjC and unless} @option{-ffreestanding}@r{)} @gol
3703 -Wmaybe-uninitialized @gol
3704 -Wmemset-elt-size @gol
3705 -Wmemset-transposed-args @gol
3706 -Wmisleading-indentation @r{(only for C/C++)} @gol
3707 -Wmissing-braces @r{(only for C/ObjC)} @gol
3708 -Wnarrowing @r{(only for C++)} @gol
3710 -Wnonnull-compare @gol
3716 -Wsequence-point @gol
3717 -Wsign-compare @r{(only in C++)} @gol
3718 -Wsizeof-pointer-memaccess @gol
3719 -Wstrict-aliasing @gol
3720 -Wstrict-overflow=1 @gol
3722 -Wtautological-compare @gol
3724 -Wuninitialized @gol
3725 -Wunknown-pragmas @gol
3726 -Wunused-function @gol
3729 -Wunused-variable @gol
3730 -Wvolatile-register-var @gol
3733 Note that some warning flags are not implied by @option{-Wall}. Some of
3734 them warn about constructions that users generally do not consider
3735 questionable, but which occasionally you might wish to check for;
3736 others warn about constructions that are necessary or hard to avoid in
3737 some cases, and there is no simple way to modify the code to suppress
3738 the warning. Some of them are enabled by @option{-Wextra} but many of
3739 them must be enabled individually.
3745 This enables some extra warning flags that are not enabled by
3746 @option{-Wall}. (This option used to be called @option{-W}. The older
3747 name is still supported, but the newer name is more descriptive.)
3749 @gccoptlist{-Wclobbered @gol
3751 -Wignored-qualifiers @gol
3752 -Wimplicit-fallthrough=3 @gol
3753 -Wmissing-field-initializers @gol
3754 -Wmissing-parameter-type @r{(C only)} @gol
3755 -Wold-style-declaration @r{(C only)} @gol
3756 -Woverride-init @gol
3757 -Wsign-compare @r{(C only)} @gol
3759 -Wuninitialized @gol
3760 -Wshift-negative-value @r{(in C++03 and in C99 and newer)} @gol
3761 -Wunused-parameter @r{(only with} @option{-Wunused} @r{or} @option{-Wall}@r{)} @gol
3762 -Wunused-but-set-parameter @r{(only with} @option{-Wunused} @r{or} @option{-Wall}@r{)} @gol
3765 The option @option{-Wextra} also prints warning messages for the
3771 A pointer is compared against integer zero with @code{<}, @code{<=},
3772 @code{>}, or @code{>=}.
3775 (C++ only) An enumerator and a non-enumerator both appear in a
3776 conditional expression.
3779 (C++ only) Ambiguous virtual bases.
3782 (C++ only) Subscripting an array that has been declared @code{register}.
3785 (C++ only) Taking the address of a variable that has been declared
3789 (C++ only) A base class is not initialized in the copy constructor
3794 @item -Wchar-subscripts
3795 @opindex Wchar-subscripts
3796 @opindex Wno-char-subscripts
3797 Warn if an array subscript has type @code{char}. This is a common cause
3798 of error, as programmers often forget that this type is signed on some
3800 This warning is enabled by @option{-Wall}.
3804 @opindex Wno-comment
3805 Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*}
3806 comment, or whenever a Backslash-Newline appears in a @samp{//} comment.
3807 This warning is enabled by @option{-Wall}.
3809 @item -Wno-coverage-mismatch
3810 @opindex Wno-coverage-mismatch
3811 Warn if feedback profiles do not match when using the
3812 @option{-fprofile-use} option.
3813 If a source file is changed between compiling with @option{-fprofile-gen} and
3814 with @option{-fprofile-use}, the files with the profile feedback can fail
3815 to match the source file and GCC cannot use the profile feedback
3816 information. By default, this warning is enabled and is treated as an
3817 error. @option{-Wno-coverage-mismatch} can be used to disable the
3818 warning or @option{-Wno-error=coverage-mismatch} can be used to
3819 disable the error. Disabling the error for this warning can result in
3820 poorly optimized code and is useful only in the
3821 case of very minor changes such as bug fixes to an existing code-base.
3822 Completely disabling the warning is not recommended.
3825 @r{(C, Objective-C, C++, Objective-C++ and Fortran only)}
3827 Suppress warning messages emitted by @code{#warning} directives.
3829 @item -Wdouble-promotion @r{(C, C++, Objective-C and Objective-C++ only)}
3830 @opindex Wdouble-promotion
3831 @opindex Wno-double-promotion
3832 Give a warning when a value of type @code{float} is implicitly
3833 promoted to @code{double}. CPUs with a 32-bit ``single-precision''
3834 floating-point unit implement @code{float} in hardware, but emulate
3835 @code{double} in software. On such a machine, doing computations
3836 using @code{double} values is much more expensive because of the
3837 overhead required for software emulation.
3839 It is easy to accidentally do computations with @code{double} because
3840 floating-point literals are implicitly of type @code{double}. For
3844 float area(float radius)
3846 return 3.14159 * radius * radius;
3850 the compiler performs the entire computation with @code{double}
3851 because the floating-point literal is a @code{double}.
3853 @item -Wduplicate-decl-specifier @r{(C and Objective-C only)}
3854 @opindex Wduplicate-decl-specifier
3855 @opindex Wno-duplicate-decl-specifier
3856 Warn if a declaration has duplicate @code{const}, @code{volatile},
3857 @code{restrict} or @code{_Atomic} specifier. This warning is enabled by
3861 @itemx -Wformat=@var{n}
3864 @opindex ffreestanding
3865 @opindex fno-builtin
3867 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
3868 the arguments supplied have types appropriate to the format string
3869 specified, and that the conversions specified in the format string make
3870 sense. This includes standard functions, and others specified by format
3871 attributes (@pxref{Function Attributes}), in the @code{printf},
3872 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
3873 not in the C standard) families (or other target-specific families).
3874 Which functions are checked without format attributes having been
3875 specified depends on the standard version selected, and such checks of
3876 functions without the attribute specified are disabled by
3877 @option{-ffreestanding} or @option{-fno-builtin}.
3879 The formats are checked against the format features supported by GNU
3880 libc version 2.2. These include all ISO C90 and C99 features, as well
3881 as features from the Single Unix Specification and some BSD and GNU
3882 extensions. Other library implementations may not support all these
3883 features; GCC does not support warning about features that go beyond a
3884 particular library's limitations. However, if @option{-Wpedantic} is used
3885 with @option{-Wformat}, warnings are given about format features not
3886 in the selected standard version (but not for @code{strfmon} formats,
3887 since those are not in any version of the C standard). @xref{C Dialect
3888 Options,,Options Controlling C Dialect}.
3895 Option @option{-Wformat} is equivalent to @option{-Wformat=1}, and
3896 @option{-Wno-format} is equivalent to @option{-Wformat=0}. Since
3897 @option{-Wformat} also checks for null format arguments for several
3898 functions, @option{-Wformat} also implies @option{-Wnonnull}. Some
3899 aspects of this level of format checking can be disabled by the
3900 options: @option{-Wno-format-contains-nul},
3901 @option{-Wno-format-extra-args}, and @option{-Wno-format-zero-length}.
3902 @option{-Wformat} is enabled by @option{-Wall}.
3904 @item -Wno-format-contains-nul
3905 @opindex Wno-format-contains-nul
3906 @opindex Wformat-contains-nul
3907 If @option{-Wformat} is specified, do not warn about format strings that
3910 @item -Wno-format-extra-args
3911 @opindex Wno-format-extra-args
3912 @opindex Wformat-extra-args
3913 If @option{-Wformat} is specified, do not warn about excess arguments to a
3914 @code{printf} or @code{scanf} format function. The C standard specifies
3915 that such arguments are ignored.
3917 Where the unused arguments lie between used arguments that are
3918 specified with @samp{$} operand number specifications, normally
3919 warnings are still given, since the implementation could not know what
3920 type to pass to @code{va_arg} to skip the unused arguments. However,
3921 in the case of @code{scanf} formats, this option suppresses the
3922 warning if the unused arguments are all pointers, since the Single
3923 Unix Specification says that such unused arguments are allowed.
3925 @item -Wformat-length
3926 @itemx -Wformat-length=@var{level}
3927 @opindex Wformat-length
3928 @opindex Wno-format-length
3929 Warn about calls to formatted input/output functions such as @code{sprintf}
3930 that might overflow the destination buffer, or about bounded functions such
3931 as @code{snprintf} that might result in output truncation. When the exact
3932 number of bytes written by a format directive cannot be determined at
3933 compile-time it is estimated based on heuristics that depend on the
3934 @var{level} argument and on optimization. While enabling optimization
3935 will in most cases improve the accuracy of the warning, it may also
3936 result in false positives.
3939 @item -Wformat-length
3940 @item -Wformat-length=1
3941 @opindex Wformat-length
3942 @opindex Wno-format-length
3943 Level @var{1} of @option{-Wformat-length} enabled by @option{-Wformat}
3944 employs a conservative approach that warns only about calls that most
3945 likely overflow the buffer or result in output truncation. At this
3946 level, numeric arguments to format directives with unknown values are
3947 assumed to have the value of one, and strings of unknown length to be
3948 empty. Numeric arguments that are known to be bounded to a subrange
3949 of their type, or string arguments whose output is bounded either by
3950 their directive's precision or by a finite set of string literals, are
3951 assumed to take on the value within the range that results in the most
3952 bytes on output. For example, the call to @code{sprintf} below is
3953 diagnosed because even with both @var{a} and @var{b} equal to zero,
3954 the terminating NUL character (@code{'\0'}) appended by the function
3955 to the destination buffer will be written past its end. Increasing
3956 the size of the buffer by a single byte is sufficient to avoid the
3957 warning, though it may not be sufficient to avoid the overflow.
3960 void f (int a, int b)
3963 sprintf (buf, "a = %i, b = %i\n", a, b);
3967 @item -Wformat-length=2
3968 Level @var{2} warns also about calls that might overflow the destination
3969 buffer or result in truncation given an argument of sufficient length
3970 or magnitude. At level @var{2}, unknown numeric arguments are assumed
3971 to have the minimum representable value for signed types with a precision
3972 greater than 1, and the maximum representable value otherwise. Unknown
3973 string arguments whose length cannot be assumed to be bounded either by
3974 the directive's precision, or by a finite set of string literals they
3975 may evaluate to, or the character array they may point to, are assumed
3976 to be 1 character long.
3978 At level @var{2}, the call in the example above is again diagnosed, but
3979 this time because with @var{a} equal to a 32-bit @code{INT_MIN} the first
3980 @code{%i} directive will write some of its digits beyond the end of
3981 the destination buffer. To make the call safe regardless of the values
3982 of the two variables, the size of the destination buffer must be increased
3983 to at least 34 bytes. GCC includes the minimum size of the buffer in
3984 an informational note following the warning.
3986 An alternative to increasing the size of the destination buffer is to
3987 constrain the range of formatted values. The maximum length of string
3988 arguments can be bounded by specifying the precision in the format
3989 directive. When numeric arguments of format directives can be assumed
3990 to be bounded by less than the precision of their type, choosing
3991 an appropriate length modifier to the format specifier will reduce
3992 the required buffer size. For example, if @var{a} and @var{b} in the
3993 example above can be assumed to be within the precision of
3994 the @code{short int} type then using either the @code{%hi} format
3995 directive or casting the argument to @code{short} reduces the maximum
3996 required size of the buffer to 24 bytes.
3999 void f (int a, int b)
4002 sprintf (buf, "a = %hi, b = %i\n", a, (short)b);
4007 @item -Wno-format-zero-length
4008 @opindex Wno-format-zero-length
4009 @opindex Wformat-zero-length
4010 If @option{-Wformat} is specified, do not warn about zero-length formats.
4011 The C standard specifies that zero-length formats are allowed.
4016 Enable @option{-Wformat} plus additional format checks. Currently
4017 equivalent to @option{-Wformat -Wformat-nonliteral -Wformat-security
4020 @item -Wformat-nonliteral
4021 @opindex Wformat-nonliteral
4022 @opindex Wno-format-nonliteral
4023 If @option{-Wformat} is specified, also warn if the format string is not a
4024 string literal and so cannot be checked, unless the format function
4025 takes its format arguments as a @code{va_list}.
4027 @item -Wformat-security
4028 @opindex Wformat-security
4029 @opindex Wno-format-security
4030 If @option{-Wformat} is specified, also warn about uses of format
4031 functions that represent possible security problems. At present, this
4032 warns about calls to @code{printf} and @code{scanf} functions where the
4033 format string is not a string literal and there are no format arguments,
4034 as in @code{printf (foo);}. This may be a security hole if the format
4035 string came from untrusted input and contains @samp{%n}. (This is
4036 currently a subset of what @option{-Wformat-nonliteral} warns about, but
4037 in future warnings may be added to @option{-Wformat-security} that are not
4038 included in @option{-Wformat-nonliteral}.)
4040 @item -Wformat-signedness
4041 @opindex Wformat-signedness
4042 @opindex Wno-format-signedness
4043 If @option{-Wformat} is specified, also warn if the format string
4044 requires an unsigned argument and the argument is signed and vice versa.
4047 @opindex Wformat-y2k
4048 @opindex Wno-format-y2k
4049 If @option{-Wformat} is specified, also warn about @code{strftime}
4050 formats that may yield only a two-digit year.
4055 @opindex Wno-nonnull
4056 Warn about passing a null pointer for arguments marked as
4057 requiring a non-null value by the @code{nonnull} function attribute.
4059 @option{-Wnonnull} is included in @option{-Wall} and @option{-Wformat}. It
4060 can be disabled with the @option{-Wno-nonnull} option.
4062 @item -Wnonnull-compare
4063 @opindex Wnonnull-compare
4064 @opindex Wno-nonnull-compare
4065 Warn when comparing an argument marked with the @code{nonnull}
4066 function attribute against null inside the function.
4068 @option{-Wnonnull-compare} is included in @option{-Wall}. It
4069 can be disabled with the @option{-Wno-nonnull-compare} option.
4071 @item -Wnull-dereference
4072 @opindex Wnull-dereference
4073 @opindex Wno-null-dereference
4074 Warn if the compiler detects paths that trigger erroneous or
4075 undefined behavior due to dereferencing a null pointer. This option
4076 is only active when @option{-fdelete-null-pointer-checks} is active,
4077 which is enabled by optimizations in most targets. The precision of
4078 the warnings depends on the optimization options used.
4080 @item -Winit-self @r{(C, C++, Objective-C and Objective-C++ only)}
4082 @opindex Wno-init-self
4083 Warn about uninitialized variables that are initialized with themselves.
4084 Note this option can only be used with the @option{-Wuninitialized} option.
4086 For example, GCC warns about @code{i} being uninitialized in the
4087 following snippet only when @option{-Winit-self} has been specified:
4098 This warning is enabled by @option{-Wall} in C++.
4100 @item -Wimplicit-int @r{(C and Objective-C only)}
4101 @opindex Wimplicit-int
4102 @opindex Wno-implicit-int
4103 Warn when a declaration does not specify a type.
4104 This warning is enabled by @option{-Wall}.
4106 @item -Wimplicit-function-declaration @r{(C and Objective-C only)}
4107 @opindex Wimplicit-function-declaration
4108 @opindex Wno-implicit-function-declaration
4109 Give a warning whenever a function is used before being declared. In
4110 C99 mode (@option{-std=c99} or @option{-std=gnu99}), this warning is
4111 enabled by default and it is made into an error by
4112 @option{-pedantic-errors}. This warning is also enabled by
4115 @item -Wimplicit @r{(C and Objective-C only)}
4117 @opindex Wno-implicit
4118 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
4119 This warning is enabled by @option{-Wall}.
4121 @item -Wimplicit-fallthrough
4122 @opindex Wimplicit-fallthrough
4123 @opindex Wno-implicit-fallthrough
4124 @option{-Wimplicit-fallthrough} is the same as @option{-Wimplicit-fallthrough=3}
4125 and @option{-Wno-implicit-fallthrough} is the same as
4126 @option{-Wimplicit-fallthrough=0}.
4128 @item -Wimplicit-fallthrough=@var{n}
4129 @opindex Wimplicit-fallthrough=
4130 Warn when a switch case falls through. For example:
4148 This warning does not warn when the last statement of a case cannot
4149 fall through, e.g. when there is a return statement or a call to function
4150 declared with the noreturn attribute. @option{-Wimplicit-fallthrough=}
4151 also takes into account control flow statements, such as ifs, and only
4152 warns when appropriate. E.g.@:
4162 @} else if (i < 1) @{
4172 Since there are occasions where a switch case fall through is desirable,
4173 GCC provides an attribute, @code{__attribute__ ((fallthrough))}, that is
4174 to be used along with a null statement to suppress this warning that
4175 would normally occur:
4183 __attribute__ ((fallthrough));
4190 C++17 provides a standard way to suppress the @option{-Wimplicit-fallthrough}
4191 warning using @code{[[fallthrough]];} instead of the GNU attribute. In C++11
4192 or C++14 users can use @code{[[gnu::fallthrough]];}, which is a GNU extension.
4193 Instead of the these attributes, it is also possible to add a fallthrough
4194 comment to silence the warning. The whole body of the C or C++ style comment
4195 should match the given regular expressions listed below. The option argument
4196 @var{n} specifies what kind of comments are accepted:
4200 @item @option{-Wimplicit-fallthrough=0} disables the warning altogether.
4202 @item @option{-Wimplicit-fallthrough=1} matches @code{.*} regular
4203 expression, any comment is used as fallthrough comment.
4205 @item @option{-Wimplicit-fallthrough=2} case insensitively matches
4206 @code{.*falls?[ \t-]*thr(ough|u).*} regular expression.
4208 @item @option{-Wimplicit-fallthrough=3} case sensitively matches one of the
4209 following regular expressions:
4213 @item @code{-fallthrough}
4215 @item @code{@@fallthrough@@}
4217 @item @code{lint -fallthrough[ \t]*}
4219 @item @code{[ \t.!]*(ELSE,? |INTENTIONAL(LY)? )?@*FALL(S | |-)?THR(OUGH|U)[ \t.!]*(-[^\n\r]*)?}
4221 @item @code{[ \t.!]*(Else,? |Intentional(ly)? )?@*Fall((s | |-)[Tt]|t)hr(ough|u)[ \t.!]*(-[^\n\r]*)?}
4223 @item @code{[ \t.!]*([Ee]lse,? |[Ii]ntentional(ly)? )?@*fall(s | |-)?thr(ough|u)[ \t.!]*(-[^\n\r]*)?}
4227 @item @option{-Wimplicit-fallthrough=4} case sensitively matches one of the
4228 following regular expressions:
4232 @item @code{-fallthrough}
4234 @item @code{@@fallthrough@@}
4236 @item @code{lint -fallthrough[ \t]*}
4238 @item @code{[ \t]*FALLTHR(OUGH|U)[ \t]*}
4242 @item @option{-Wimplicit-fallthrough=5} doesn't recognize any comments as
4243 fallthrough comments, only attributes disable the warning.
4247 The comment needs to be followed after optional whitespace and other comments
4248 by @code{case} or @code{default} keywords or by a user label that preceeds some
4249 @code{case} or @code{default} label.
4264 The @option{-Wimplicit-fallthrough=3} warning is enabled by @option{-Wextra}.
4266 @item -Wignored-qualifiers @r{(C and C++ only)}
4267 @opindex Wignored-qualifiers
4268 @opindex Wno-ignored-qualifiers
4269 Warn if the return type of a function has a type qualifier
4270 such as @code{const}. For ISO C such a type qualifier has no effect,
4271 since the value returned by a function is not an lvalue.
4272 For C++, the warning is only emitted for scalar types or @code{void}.
4273 ISO C prohibits qualified @code{void} return types on function
4274 definitions, so such return types always receive a warning
4275 even without this option.
4277 This warning is also enabled by @option{-Wextra}.
4279 @item -Wignored-attributes @r{(C and C++ only)}
4280 @opindex Wignored-attributes
4281 @opindex Wno-ignored-attributes
4282 Warn when an attribute is ignored. This is different from the
4283 @option{-Wattributes} option in that it warns whenever the compiler decides
4284 to drop an attribute, not that the attribute is either unknown, used in a
4285 wrong place, etc. This warning is enabled by default.
4290 Warn if the type of @code{main} is suspicious. @code{main} should be
4291 a function with external linkage, returning int, taking either zero
4292 arguments, two, or three arguments of appropriate types. This warning
4293 is enabled by default in C++ and is enabled by either @option{-Wall}
4294 or @option{-Wpedantic}.
4296 @item -Wmisleading-indentation @r{(C and C++ only)}
4297 @opindex Wmisleading-indentation
4298 @opindex Wno-misleading-indentation
4299 Warn when the indentation of the code does not reflect the block structure.
4300 Specifically, a warning is issued for @code{if}, @code{else}, @code{while}, and
4301 @code{for} clauses with a guarded statement that does not use braces,
4302 followed by an unguarded statement with the same indentation.
4304 In the following example, the call to ``bar'' is misleadingly indented as
4305 if it were guarded by the ``if'' conditional.
4308 if (some_condition ())
4310 bar (); /* Gotcha: this is not guarded by the "if". */
4313 In the case of mixed tabs and spaces, the warning uses the
4314 @option{-ftabstop=} option to determine if the statements line up
4317 The warning is not issued for code involving multiline preprocessor logic
4318 such as the following example.
4323 #if SOME_CONDITION_THAT_DOES_NOT_HOLD
4329 The warning is not issued after a @code{#line} directive, since this
4330 typically indicates autogenerated code, and no assumptions can be made
4331 about the layout of the file that the directive references.
4333 This warning is enabled by @option{-Wall} in C and C++.
4335 @item -Wmissing-braces
4336 @opindex Wmissing-braces
4337 @opindex Wno-missing-braces
4338 Warn if an aggregate or union initializer is not fully bracketed. In
4339 the following example, the initializer for @code{a} is not fully
4340 bracketed, but that for @code{b} is fully bracketed. This warning is
4341 enabled by @option{-Wall} in C.
4344 int a[2][2] = @{ 0, 1, 2, 3 @};
4345 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
4348 This warning is enabled by @option{-Wall}.
4350 @item -Wmissing-include-dirs @r{(C, C++, Objective-C and Objective-C++ only)}
4351 @opindex Wmissing-include-dirs
4352 @opindex Wno-missing-include-dirs
4353 Warn if a user-supplied include directory does not exist.
4356 @opindex Wparentheses
4357 @opindex Wno-parentheses
4358 Warn if parentheses are omitted in certain contexts, such
4359 as when there is an assignment in a context where a truth value
4360 is expected, or when operators are nested whose precedence people
4361 often get confused about.
4363 Also warn if a comparison like @code{x<=y<=z} appears; this is
4364 equivalent to @code{(x<=y ? 1 : 0) <= z}, which is a different
4365 interpretation from that of ordinary mathematical notation.
4367 Also warn for dangerous uses of the GNU extension to
4368 @code{?:} with omitted middle operand. When the condition
4369 in the @code{?}: operator is a boolean expression, the omitted value is
4370 always 1. Often programmers expect it to be a value computed
4371 inside the conditional expression instead.
4373 This warning is enabled by @option{-Wall}.
4375 @item -Wsequence-point
4376 @opindex Wsequence-point
4377 @opindex Wno-sequence-point
4378 Warn about code that may have undefined semantics because of violations
4379 of sequence point rules in the C and C++ standards.
4381 The C and C++ standards define the order in which expressions in a C/C++
4382 program are evaluated in terms of @dfn{sequence points}, which represent
4383 a partial ordering between the execution of parts of the program: those
4384 executed before the sequence point, and those executed after it. These
4385 occur after the evaluation of a full expression (one which is not part
4386 of a larger expression), after the evaluation of the first operand of a
4387 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
4388 function is called (but after the evaluation of its arguments and the
4389 expression denoting the called function), and in certain other places.
4390 Other than as expressed by the sequence point rules, the order of
4391 evaluation of subexpressions of an expression is not specified. All
4392 these rules describe only a partial order rather than a total order,
4393 since, for example, if two functions are called within one expression
4394 with no sequence point between them, the order in which the functions
4395 are called is not specified. However, the standards committee have
4396 ruled that function calls do not overlap.
4398 It is not specified when between sequence points modifications to the
4399 values of objects take effect. Programs whose behavior depends on this
4400 have undefined behavior; the C and C++ standards specify that ``Between
4401 the previous and next sequence point an object shall have its stored
4402 value modified at most once by the evaluation of an expression.
4403 Furthermore, the prior value shall be read only to determine the value
4404 to be stored.''. If a program breaks these rules, the results on any
4405 particular implementation are entirely unpredictable.
4407 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
4408 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
4409 diagnosed by this option, and it may give an occasional false positive
4410 result, but in general it has been found fairly effective at detecting
4411 this sort of problem in programs.
4413 The C++17 standard will define the order of evaluation of operands in
4414 more cases: in particular it requires that the right-hand side of an
4415 assignment be evaluated before the left-hand side, so the above
4416 examples are no longer undefined. But this warning will still warn
4417 about them, to help people avoid writing code that is undefined in C
4418 and earlier revisions of C++.
4420 The standard is worded confusingly, therefore there is some debate
4421 over the precise meaning of the sequence point rules in subtle cases.
4422 Links to discussions of the problem, including proposed formal
4423 definitions, may be found on the GCC readings page, at
4424 @uref{http://gcc.gnu.org/@/readings.html}.
4426 This warning is enabled by @option{-Wall} for C and C++.
4428 @item -Wno-return-local-addr
4429 @opindex Wno-return-local-addr
4430 @opindex Wreturn-local-addr
4431 Do not warn about returning a pointer (or in C++, a reference) to a
4432 variable that goes out of scope after the function returns.
4435 @opindex Wreturn-type
4436 @opindex Wno-return-type
4437 Warn whenever a function is defined with a return type that defaults
4438 to @code{int}. Also warn about any @code{return} statement with no
4439 return value in a function whose return type is not @code{void}
4440 (falling off the end of the function body is considered returning
4443 For C only, warn about a @code{return} statement with an expression in a
4444 function whose return type is @code{void}, unless the expression type is
4445 also @code{void}. As a GNU extension, the latter case is accepted
4446 without a warning unless @option{-Wpedantic} is used.
4448 For C++, a function without return type always produces a diagnostic
4449 message, even when @option{-Wno-return-type} is specified. The only
4450 exceptions are @code{main} and functions defined in system headers.
4452 This warning is enabled by @option{-Wall}.
4454 @item -Wshift-count-negative
4455 @opindex Wshift-count-negative
4456 @opindex Wno-shift-count-negative
4457 Warn if shift count is negative. This warning is enabled by default.
4459 @item -Wshift-count-overflow
4460 @opindex Wshift-count-overflow
4461 @opindex Wno-shift-count-overflow
4462 Warn if shift count >= width of type. This warning is enabled by default.
4464 @item -Wshift-negative-value
4465 @opindex Wshift-negative-value
4466 @opindex Wno-shift-negative-value
4467 Warn if left shifting a negative value. This warning is enabled by
4468 @option{-Wextra} in C99 and C++11 modes (and newer).
4470 @item -Wshift-overflow
4471 @itemx -Wshift-overflow=@var{n}
4472 @opindex Wshift-overflow
4473 @opindex Wno-shift-overflow
4474 Warn about left shift overflows. This warning is enabled by
4475 default in C99 and C++11 modes (and newer).
4478 @item -Wshift-overflow=1
4479 This is the warning level of @option{-Wshift-overflow} and is enabled
4480 by default in C99 and C++11 modes (and newer). This warning level does
4481 not warn about left-shifting 1 into the sign bit. (However, in C, such
4482 an overflow is still rejected in contexts where an integer constant expression
4485 @item -Wshift-overflow=2
4486 This warning level also warns about left-shifting 1 into the sign bit,
4487 unless C++14 mode is active.
4493 Warn whenever a @code{switch} statement has an index of enumerated type
4494 and lacks a @code{case} for one or more of the named codes of that
4495 enumeration. (The presence of a @code{default} label prevents this
4496 warning.) @code{case} labels outside the enumeration range also
4497 provoke warnings when this option is used (even if there is a
4498 @code{default} label).
4499 This warning is enabled by @option{-Wall}.
4501 @item -Wswitch-default
4502 @opindex Wswitch-default
4503 @opindex Wno-switch-default
4504 Warn whenever a @code{switch} statement does not have a @code{default}
4508 @opindex Wswitch-enum
4509 @opindex Wno-switch-enum
4510 Warn whenever a @code{switch} statement has an index of enumerated type
4511 and lacks a @code{case} for one or more of the named codes of that
4512 enumeration. @code{case} labels outside the enumeration range also
4513 provoke warnings when this option is used. The only difference
4514 between @option{-Wswitch} and this option is that this option gives a
4515 warning about an omitted enumeration code even if there is a
4516 @code{default} label.
4519 @opindex Wswitch-bool
4520 @opindex Wno-switch-bool
4521 Warn whenever a @code{switch} statement has an index of boolean type
4522 and the case values are outside the range of a boolean type.
4523 It is possible to suppress this warning by casting the controlling
4524 expression to a type other than @code{bool}. For example:
4527 switch ((int) (a == 4))
4533 This warning is enabled by default for C and C++ programs.
4535 @item -Wswitch-unreachable
4536 @opindex Wswitch-unreachable
4537 @opindex Wno-switch-unreachable
4538 Warn whenever a @code{switch} statement contains statements between the
4539 controlling expression and the first case label, which will never be
4540 executed. For example:
4552 @option{-Wswitch-unreachable} does not warn if the statement between the
4553 controlling expression and the first case label is just a declaration:
4566 This warning is enabled by default for C and C++ programs.
4568 @item -Wsync-nand @r{(C and C++ only)}
4570 @opindex Wno-sync-nand
4571 Warn when @code{__sync_fetch_and_nand} and @code{__sync_nand_and_fetch}
4572 built-in functions are used. These functions changed semantics in GCC 4.4.
4576 @opindex Wno-trigraphs
4577 Warn if any trigraphs are encountered that might change the meaning of
4578 the program (trigraphs within comments are not warned about).
4579 This warning is enabled by @option{-Wall}.
4581 @item -Wunused-but-set-parameter
4582 @opindex Wunused-but-set-parameter
4583 @opindex Wno-unused-but-set-parameter
4584 Warn whenever a function parameter is assigned to, but otherwise unused
4585 (aside from its declaration).
4587 To suppress this warning use the @code{unused} attribute
4588 (@pxref{Variable Attributes}).
4590 This warning is also enabled by @option{-Wunused} together with
4593 @item -Wunused-but-set-variable
4594 @opindex Wunused-but-set-variable
4595 @opindex Wno-unused-but-set-variable
4596 Warn whenever a local variable is assigned to, but otherwise unused
4597 (aside from its declaration).
4598 This warning is enabled by @option{-Wall}.
4600 To suppress this warning use the @code{unused} attribute
4601 (@pxref{Variable Attributes}).
4603 This warning is also enabled by @option{-Wunused}, which is enabled
4606 @item -Wunused-function
4607 @opindex Wunused-function
4608 @opindex Wno-unused-function
4609 Warn whenever a static function is declared but not defined or a
4610 non-inline static function is unused.
4611 This warning is enabled by @option{-Wall}.
4613 @item -Wunused-label
4614 @opindex Wunused-label
4615 @opindex Wno-unused-label
4616 Warn whenever a label is declared but not used.
4617 This warning is enabled by @option{-Wall}.
4619 To suppress this warning use the @code{unused} attribute
4620 (@pxref{Variable Attributes}).
4622 @item -Wunused-local-typedefs @r{(C, Objective-C, C++ and Objective-C++ only)}
4623 @opindex Wunused-local-typedefs
4624 Warn when a typedef locally defined in a function is not used.
4625 This warning is enabled by @option{-Wall}.
4627 @item -Wunused-parameter
4628 @opindex Wunused-parameter
4629 @opindex Wno-unused-parameter
4630 Warn whenever a function parameter is unused aside from its declaration.
4632 To suppress this warning use the @code{unused} attribute
4633 (@pxref{Variable Attributes}).
4635 @item -Wno-unused-result
4636 @opindex Wunused-result
4637 @opindex Wno-unused-result
4638 Do not warn if a caller of a function marked with attribute
4639 @code{warn_unused_result} (@pxref{Function Attributes}) does not use
4640 its return value. The default is @option{-Wunused-result}.
4642 @item -Wunused-variable
4643 @opindex Wunused-variable
4644 @opindex Wno-unused-variable
4645 Warn whenever a local or static variable is unused aside from its
4646 declaration. This option implies @option{-Wunused-const-variable=1} for C,
4647 but not for C++. This warning is enabled by @option{-Wall}.
4649 To suppress this warning use the @code{unused} attribute
4650 (@pxref{Variable Attributes}).
4652 @item -Wunused-const-variable
4653 @itemx -Wunused-const-variable=@var{n}
4654 @opindex Wunused-const-variable
4655 @opindex Wno-unused-const-variable
4656 Warn whenever a constant static variable is unused aside from its declaration.
4657 @option{-Wunused-const-variable=1} is enabled by @option{-Wunused-variable}
4658 for C, but not for C++. In C this declares variable storage, but in C++ this
4659 is not an error since const variables take the place of @code{#define}s.
4661 To suppress this warning use the @code{unused} attribute
4662 (@pxref{Variable Attributes}).
4665 @item -Wunused-const-variable=1
4666 This is the warning level that is enabled by @option{-Wunused-variable} for
4667 C. It warns only about unused static const variables defined in the main
4668 compilation unit, but not about static const variables declared in any
4671 @item -Wunused-const-variable=2
4672 This warning level also warns for unused constant static variables in
4673 headers (excluding system headers). This is the warning level of
4674 @option{-Wunused-const-variable} and must be explicitly requested since
4675 in C++ this isn't an error and in C it might be harder to clean up all
4679 @item -Wunused-value
4680 @opindex Wunused-value
4681 @opindex Wno-unused-value
4682 Warn whenever a statement computes a result that is explicitly not
4683 used. To suppress this warning cast the unused expression to
4684 @code{void}. This includes an expression-statement or the left-hand
4685 side of a comma expression that contains no side effects. For example,
4686 an expression such as @code{x[i,j]} causes a warning, while
4687 @code{x[(void)i,j]} does not.
4689 This warning is enabled by @option{-Wall}.
4694 All the above @option{-Wunused} options combined.
4696 In order to get a warning about an unused function parameter, you must
4697 either specify @option{-Wextra -Wunused} (note that @option{-Wall} implies
4698 @option{-Wunused}), or separately specify @option{-Wunused-parameter}.
4700 @item -Wuninitialized
4701 @opindex Wuninitialized
4702 @opindex Wno-uninitialized
4703 Warn if an automatic variable is used without first being initialized
4704 or if a variable may be clobbered by a @code{setjmp} call. In C++,
4705 warn if a non-static reference or non-static @code{const} member
4706 appears in a class without constructors.
4708 If you want to warn about code that uses the uninitialized value of the
4709 variable in its own initializer, use the @option{-Winit-self} option.
4711 These warnings occur for individual uninitialized or clobbered
4712 elements of structure, union or array variables as well as for
4713 variables that are uninitialized or clobbered as a whole. They do
4714 not occur for variables or elements declared @code{volatile}. Because
4715 these warnings depend on optimization, the exact variables or elements
4716 for which there are warnings depends on the precise optimization
4717 options and version of GCC used.
4719 Note that there may be no warning about a variable that is used only
4720 to compute a value that itself is never used, because such
4721 computations may be deleted by data flow analysis before the warnings
4724 @item -Winvalid-memory-model
4725 @opindex Winvalid-memory-model
4726 @opindex Wno-invalid-memory-model
4727 Warn for invocations of @ref{__atomic Builtins}, @ref{__sync Builtins},
4728 and the C11 atomic generic functions with a memory consistency argument
4729 that is either invalid for the operation or outside the range of values
4730 of the @code{memory_order} enumeration. For example, since the
4731 @code{__atomic_store} and @code{__atomic_store_n} built-ins are only
4732 defined for the relaxed, release, and sequentially consistent memory
4733 orders the following code is diagnosed:
4738 __atomic_store_n (i, 0, memory_order_consume);
4742 @option{-Winvalid-memory-model} is enabled by default.
4744 @item -Wmaybe-uninitialized
4745 @opindex Wmaybe-uninitialized
4746 @opindex Wno-maybe-uninitialized
4747 For an automatic variable, if there exists a path from the function
4748 entry to a use of the variable that is initialized, but there exist
4749 some other paths for which the variable is not initialized, the compiler
4750 emits a warning if it cannot prove the uninitialized paths are not
4751 executed at run time. These warnings are made optional because GCC is
4752 not smart enough to see all the reasons why the code might be correct
4753 in spite of appearing to have an error. Here is one example of how
4774 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
4775 always initialized, but GCC doesn't know this. To suppress the
4776 warning, you need to provide a default case with assert(0) or
4779 @cindex @code{longjmp} warnings
4780 This option also warns when a non-volatile automatic variable might be
4781 changed by a call to @code{longjmp}. These warnings as well are possible
4782 only in optimizing compilation.
4784 The compiler sees only the calls to @code{setjmp}. It cannot know
4785 where @code{longjmp} will be called; in fact, a signal handler could
4786 call it at any point in the code. As a result, you may get a warning
4787 even when there is in fact no problem because @code{longjmp} cannot
4788 in fact be called at the place that would cause a problem.
4790 Some spurious warnings can be avoided if you declare all the functions
4791 you use that never return as @code{noreturn}. @xref{Function
4794 This warning is enabled by @option{-Wall} or @option{-Wextra}.
4796 @item -Wunknown-pragmas
4797 @opindex Wunknown-pragmas
4798 @opindex Wno-unknown-pragmas
4799 @cindex warning for unknown pragmas
4800 @cindex unknown pragmas, warning
4801 @cindex pragmas, warning of unknown
4802 Warn when a @code{#pragma} directive is encountered that is not understood by
4803 GCC@. If this command-line option is used, warnings are even issued
4804 for unknown pragmas in system header files. This is not the case if
4805 the warnings are only enabled by the @option{-Wall} command-line option.
4808 @opindex Wno-pragmas
4810 Do not warn about misuses of pragmas, such as incorrect parameters,
4811 invalid syntax, or conflicts between pragmas. See also
4812 @option{-Wunknown-pragmas}.
4814 @item -Wstrict-aliasing
4815 @opindex Wstrict-aliasing
4816 @opindex Wno-strict-aliasing
4817 This option is only active when @option{-fstrict-aliasing} is active.
4818 It warns about code that might break the strict aliasing rules that the
4819 compiler is using for optimization. The warning does not catch all
4820 cases, but does attempt to catch the more common pitfalls. It is
4821 included in @option{-Wall}.
4822 It is equivalent to @option{-Wstrict-aliasing=3}
4824 @item -Wstrict-aliasing=n
4825 @opindex Wstrict-aliasing=n
4826 This option is only active when @option{-fstrict-aliasing} is active.
4827 It warns about code that might break the strict aliasing rules that the
4828 compiler is using for optimization.
4829 Higher levels correspond to higher accuracy (fewer false positives).
4830 Higher levels also correspond to more effort, similar to the way @option{-O}
4832 @option{-Wstrict-aliasing} is equivalent to @option{-Wstrict-aliasing=3}.
4834 Level 1: Most aggressive, quick, least accurate.
4835 Possibly useful when higher levels
4836 do not warn but @option{-fstrict-aliasing} still breaks the code, as it has very few
4837 false negatives. However, it has many false positives.
4838 Warns for all pointer conversions between possibly incompatible types,
4839 even if never dereferenced. Runs in the front end only.
4841 Level 2: Aggressive, quick, not too precise.
4842 May still have many false positives (not as many as level 1 though),
4843 and few false negatives (but possibly more than level 1).
4844 Unlike level 1, it only warns when an address is taken. Warns about
4845 incomplete types. Runs in the front end only.
4847 Level 3 (default for @option{-Wstrict-aliasing}):
4848 Should have very few false positives and few false
4849 negatives. Slightly slower than levels 1 or 2 when optimization is enabled.
4850 Takes care of the common pun+dereference pattern in the front end:
4851 @code{*(int*)&some_float}.
4852 If optimization is enabled, it also runs in the back end, where it deals
4853 with multiple statement cases using flow-sensitive points-to information.
4854 Only warns when the converted pointer is dereferenced.
4855 Does not warn about incomplete types.
4857 @item -Wstrict-overflow
4858 @itemx -Wstrict-overflow=@var{n}
4859 @opindex Wstrict-overflow
4860 @opindex Wno-strict-overflow
4861 This option is only active when @option{-fstrict-overflow} is active.
4862 It warns about cases where the compiler optimizes based on the
4863 assumption that signed overflow does not occur. Note that it does not
4864 warn about all cases where the code might overflow: it only warns
4865 about cases where the compiler implements some optimization. Thus
4866 this warning depends on the optimization level.
4868 An optimization that assumes that signed overflow does not occur is
4869 perfectly safe if the values of the variables involved are such that
4870 overflow never does, in fact, occur. Therefore this warning can
4871 easily give a false positive: a warning about code that is not
4872 actually a problem. To help focus on important issues, several
4873 warning levels are defined. No warnings are issued for the use of
4874 undefined signed overflow when estimating how many iterations a loop
4875 requires, in particular when determining whether a loop will be
4879 @item -Wstrict-overflow=1
4880 Warn about cases that are both questionable and easy to avoid. For
4881 example, with @option{-fstrict-overflow}, the compiler simplifies
4882 @code{x + 1 > x} to @code{1}. This level of
4883 @option{-Wstrict-overflow} is enabled by @option{-Wall}; higher levels
4884 are not, and must be explicitly requested.
4886 @item -Wstrict-overflow=2
4887 Also warn about other cases where a comparison is simplified to a
4888 constant. For example: @code{abs (x) >= 0}. This can only be
4889 simplified when @option{-fstrict-overflow} is in effect, because
4890 @code{abs (INT_MIN)} overflows to @code{INT_MIN}, which is less than
4891 zero. @option{-Wstrict-overflow} (with no level) is the same as
4892 @option{-Wstrict-overflow=2}.
4894 @item -Wstrict-overflow=3
4895 Also warn about other cases where a comparison is simplified. For
4896 example: @code{x + 1 > 1} is simplified to @code{x > 0}.
4898 @item -Wstrict-overflow=4
4899 Also warn about other simplifications not covered by the above cases.
4900 For example: @code{(x * 10) / 5} is simplified to @code{x * 2}.
4902 @item -Wstrict-overflow=5
4903 Also warn about cases where the compiler reduces the magnitude of a
4904 constant involved in a comparison. For example: @code{x + 2 > y} is
4905 simplified to @code{x + 1 >= y}. This is reported only at the
4906 highest warning level because this simplification applies to many
4907 comparisons, so this warning level gives a very large number of
4911 @item -Wsuggest-attribute=@r{[}pure@r{|}const@r{|}noreturn@r{|}format@r{]}
4912 @opindex Wsuggest-attribute=
4913 @opindex Wno-suggest-attribute=
4914 Warn for cases where adding an attribute may be beneficial. The
4915 attributes currently supported are listed below.
4918 @item -Wsuggest-attribute=pure
4919 @itemx -Wsuggest-attribute=const
4920 @itemx -Wsuggest-attribute=noreturn
4921 @opindex Wsuggest-attribute=pure
4922 @opindex Wno-suggest-attribute=pure
4923 @opindex Wsuggest-attribute=const
4924 @opindex Wno-suggest-attribute=const
4925 @opindex Wsuggest-attribute=noreturn
4926 @opindex Wno-suggest-attribute=noreturn
4928 Warn about functions that might be candidates for attributes
4929 @code{pure}, @code{const} or @code{noreturn}. The compiler only warns for
4930 functions visible in other compilation units or (in the case of @code{pure} and
4931 @code{const}) if it cannot prove that the function returns normally. A function
4932 returns normally if it doesn't contain an infinite loop or return abnormally
4933 by throwing, calling @code{abort} or trapping. This analysis requires option
4934 @option{-fipa-pure-const}, which is enabled by default at @option{-O} and
4935 higher. Higher optimization levels improve the accuracy of the analysis.
4937 @item -Wsuggest-attribute=format
4938 @itemx -Wmissing-format-attribute
4939 @opindex Wsuggest-attribute=format
4940 @opindex Wmissing-format-attribute
4941 @opindex Wno-suggest-attribute=format
4942 @opindex Wno-missing-format-attribute
4946 Warn about function pointers that might be candidates for @code{format}
4947 attributes. Note these are only possible candidates, not absolute ones.
4948 GCC guesses that function pointers with @code{format} attributes that
4949 are used in assignment, initialization, parameter passing or return
4950 statements should have a corresponding @code{format} attribute in the
4951 resulting type. I.e.@: the left-hand side of the assignment or
4952 initialization, the type of the parameter variable, or the return type
4953 of the containing function respectively should also have a @code{format}
4954 attribute to avoid the warning.
4956 GCC also warns about function definitions that might be
4957 candidates for @code{format} attributes. Again, these are only
4958 possible candidates. GCC guesses that @code{format} attributes
4959 might be appropriate for any function that calls a function like
4960 @code{vprintf} or @code{vscanf}, but this might not always be the
4961 case, and some functions for which @code{format} attributes are
4962 appropriate may not be detected.
4965 @item -Wsuggest-final-types
4966 @opindex Wno-suggest-final-types
4967 @opindex Wsuggest-final-types
4968 Warn about types with virtual methods where code quality would be improved
4969 if the type were declared with the C++11 @code{final} specifier,
4971 declared in an anonymous namespace. This allows GCC to more aggressively
4972 devirtualize the polymorphic calls. This warning is more effective with link
4973 time optimization, where the information about the class hierarchy graph is
4976 @item -Wsuggest-final-methods
4977 @opindex Wno-suggest-final-methods
4978 @opindex Wsuggest-final-methods
4979 Warn about virtual methods where code quality would be improved if the method
4980 were declared with the C++11 @code{final} specifier,
4981 or, if possible, its type were
4982 declared in an anonymous namespace or with the @code{final} specifier.
4984 more effective with link time optimization, where the information about the
4985 class hierarchy graph is more complete. It is recommended to first consider
4986 suggestions of @option{-Wsuggest-final-types} and then rebuild with new
4989 @item -Wsuggest-override
4990 Warn about overriding virtual functions that are not marked with the override
4996 This option warns on all uses of @code{alloca} in the source.
4998 @item -Walloca-larger-than=@var{n}
4999 This option warns on calls to @code{alloca} that are not bounded by a
5000 controlling predicate limiting its argument of integer type to at most
5001 @var{n} bytes, or calls to @code{alloca} where the bound is unknown.
5002 Arguments of non-integer types are considered unbounded even if they
5003 appear to be constrained to the expected range.
5005 For example, a bounded case of @code{alloca} could be:
5008 void func (size_t n)
5019 In the above example, passing @code{-Walloca-larger-than=1000} would not
5020 issue a warning because the call to @code{alloca} is known to be at most
5021 1000 bytes. However, if @code{-Walloca-larger-than=500} were passed,
5022 the compiler would emit a warning.
5024 Unbounded uses, on the other hand, are uses of @code{alloca} with no
5025 controlling predicate constraining its integer argument. For example:
5030 void *p = alloca (n);
5035 If @code{-Walloca-larger-than=500} were passed, the above would trigger
5036 a warning, but this time because of the lack of bounds checking.
5038 Note, that even seemingly correct code involving signed integers could
5042 void func (signed int n)
5052 In the above example, @var{n} could be negative, causing a larger than
5053 expected argument to be implicitly cast into the @code{alloca} call.
5055 This option also warns when @code{alloca} is used in a loop.
5057 This warning is not enabled by @option{-Wall}, and is only active when
5058 @option{-ftree-vrp} is active (default for @option{-O2} and above).
5060 See also @option{-Wvla-larger-than=@var{n}}.
5062 @item -Warray-bounds
5063 @itemx -Warray-bounds=@var{n}
5064 @opindex Wno-array-bounds
5065 @opindex Warray-bounds
5066 This option is only active when @option{-ftree-vrp} is active
5067 (default for @option{-O2} and above). It warns about subscripts to arrays
5068 that are always out of bounds. This warning is enabled by @option{-Wall}.
5071 @item -Warray-bounds=1
5072 This is the warning level of @option{-Warray-bounds} and is enabled
5073 by @option{-Wall}; higher levels are not, and must be explicitly requested.
5075 @item -Warray-bounds=2
5076 This warning level also warns about out of bounds access for
5077 arrays at the end of a struct and for arrays accessed through
5078 pointers. This warning level may give a larger number of
5079 false positives and is deactivated by default.
5082 @item -Wbool-compare
5083 @opindex Wno-bool-compare
5084 @opindex Wbool-compare
5085 Warn about boolean expression compared with an integer value different from
5086 @code{true}/@code{false}. For instance, the following comparison is
5091 if ((n > 1) == 2) @{ @dots{} @}
5093 This warning is enabled by @option{-Wall}.
5095 @item -Wbool-operation
5096 @opindex Wno-bool-operation
5097 @opindex Wbool-operation
5098 Warn about suspicious operations on expressions of a boolean type. For
5099 instance, bitwise negation of a boolean is very likely a bug in the program.
5100 For C, this warning also warns about incrementing or decrementing a boolean,
5101 which rarely makes sense. (In C++, decrementing a boolean is always invalid.
5102 Incrementing a boolean is invalid in C++1z, and deprecated otherwise.)
5104 This warning is enabled by @option{-Wall}.
5106 @item -Wduplicated-cond
5107 @opindex Wno-duplicated-cond
5108 @opindex Wduplicated-cond
5109 Warn about duplicated conditions in an if-else-if chain. For instance,
5110 warn for the following code:
5112 if (p->q != NULL) @{ @dots{} @}
5113 else if (p->q != NULL) @{ @dots{} @}
5116 @item -Wframe-address
5117 @opindex Wno-frame-address
5118 @opindex Wframe-address
5119 Warn when the @samp{__builtin_frame_address} or @samp{__builtin_return_address}
5120 is called with an argument greater than 0. Such calls may return indeterminate
5121 values or crash the program. The warning is included in @option{-Wall}.
5123 @item -Wno-discarded-qualifiers @r{(C and Objective-C only)}
5124 @opindex Wno-discarded-qualifiers
5125 @opindex Wdiscarded-qualifiers
5126 Do not warn if type qualifiers on pointers are being discarded.
5127 Typically, the compiler warns if a @code{const char *} variable is
5128 passed to a function that takes a @code{char *} parameter. This option
5129 can be used to suppress such a warning.
5131 @item -Wno-discarded-array-qualifiers @r{(C and Objective-C only)}
5132 @opindex Wno-discarded-array-qualifiers
5133 @opindex Wdiscarded-array-qualifiers
5134 Do not warn if type qualifiers on arrays which are pointer targets
5135 are being discarded. Typically, the compiler warns if a
5136 @code{const int (*)[]} variable is passed to a function that
5137 takes a @code{int (*)[]} parameter. This option can be used to
5138 suppress such a warning.
5140 @item -Wno-incompatible-pointer-types @r{(C and Objective-C only)}
5141 @opindex Wno-incompatible-pointer-types
5142 @opindex Wincompatible-pointer-types
5143 Do not warn when there is a conversion between pointers that have incompatible
5144 types. This warning is for cases not covered by @option{-Wno-pointer-sign},
5145 which warns for pointer argument passing or assignment with different
5148 @item -Wno-int-conversion @r{(C and Objective-C only)}
5149 @opindex Wno-int-conversion
5150 @opindex Wint-conversion
5151 Do not warn about incompatible integer to pointer and pointer to integer
5152 conversions. This warning is about implicit conversions; for explicit
5153 conversions the warnings @option{-Wno-int-to-pointer-cast} and
5154 @option{-Wno-pointer-to-int-cast} may be used.
5156 @item -Wno-div-by-zero
5157 @opindex Wno-div-by-zero
5158 @opindex Wdiv-by-zero
5159 Do not warn about compile-time integer division by zero. Floating-point
5160 division by zero is not warned about, as it can be a legitimate way of
5161 obtaining infinities and NaNs.
5163 @item -Wsystem-headers
5164 @opindex Wsystem-headers
5165 @opindex Wno-system-headers
5166 @cindex warnings from system headers
5167 @cindex system headers, warnings from
5168 Print warning messages for constructs found in system header files.
5169 Warnings from system headers are normally suppressed, on the assumption
5170 that they usually do not indicate real problems and would only make the
5171 compiler output harder to read. Using this command-line option tells
5172 GCC to emit warnings from system headers as if they occurred in user
5173 code. However, note that using @option{-Wall} in conjunction with this
5174 option does @emph{not} warn about unknown pragmas in system
5175 headers---for that, @option{-Wunknown-pragmas} must also be used.
5177 @item -Wtautological-compare
5178 @opindex Wtautological-compare
5179 @opindex Wno-tautological-compare
5180 Warn if a self-comparison always evaluates to true or false. This
5181 warning detects various mistakes such as:
5185 if (i > i) @{ @dots{} @}
5187 This warning is enabled by @option{-Wall}.
5190 @opindex Wtrampolines
5191 @opindex Wno-trampolines
5192 Warn about trampolines generated for pointers to nested functions.
5193 A trampoline is a small piece of data or code that is created at run
5194 time on the stack when the address of a nested function is taken, and is
5195 used to call the nested function indirectly. For some targets, it is
5196 made up of data only and thus requires no special treatment. But, for
5197 most targets, it is made up of code and thus requires the stack to be
5198 made executable in order for the program to work properly.
5201 @opindex Wfloat-equal
5202 @opindex Wno-float-equal
5203 Warn if floating-point values are used in equality comparisons.
5205 The idea behind this is that sometimes it is convenient (for the
5206 programmer) to consider floating-point values as approximations to
5207 infinitely precise real numbers. If you are doing this, then you need
5208 to compute (by analyzing the code, or in some other way) the maximum or
5209 likely maximum error that the computation introduces, and allow for it
5210 when performing comparisons (and when producing output, but that's a
5211 different problem). In particular, instead of testing for equality, you
5212 should check to see whether the two values have ranges that overlap; and
5213 this is done with the relational operators, so equality comparisons are
5216 @item -Wtraditional @r{(C and Objective-C only)}
5217 @opindex Wtraditional
5218 @opindex Wno-traditional
5219 Warn about certain constructs that behave differently in traditional and
5220 ISO C@. Also warn about ISO C constructs that have no traditional C
5221 equivalent, and/or problematic constructs that should be avoided.
5225 Macro parameters that appear within string literals in the macro body.
5226 In traditional C macro replacement takes place within string literals,
5227 but in ISO C it does not.
5230 In traditional C, some preprocessor directives did not exist.
5231 Traditional preprocessors only considered a line to be a directive
5232 if the @samp{#} appeared in column 1 on the line. Therefore
5233 @option{-Wtraditional} warns about directives that traditional C
5234 understands but ignores because the @samp{#} does not appear as the
5235 first character on the line. It also suggests you hide directives like
5236 @code{#pragma} not understood by traditional C by indenting them. Some
5237 traditional implementations do not recognize @code{#elif}, so this option
5238 suggests avoiding it altogether.
5241 A function-like macro that appears without arguments.
5244 The unary plus operator.
5247 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating-point
5248 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
5249 constants.) Note, these suffixes appear in macros defined in the system
5250 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
5251 Use of these macros in user code might normally lead to spurious
5252 warnings, however GCC's integrated preprocessor has enough context to
5253 avoid warning in these cases.
5256 A function declared external in one block and then used after the end of
5260 A @code{switch} statement has an operand of type @code{long}.
5263 A non-@code{static} function declaration follows a @code{static} one.
5264 This construct is not accepted by some traditional C compilers.
5267 The ISO type of an integer constant has a different width or
5268 signedness from its traditional type. This warning is only issued if
5269 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
5270 typically represent bit patterns, are not warned about.
5273 Usage of ISO string concatenation is detected.
5276 Initialization of automatic aggregates.
5279 Identifier conflicts with labels. Traditional C lacks a separate
5280 namespace for labels.
5283 Initialization of unions. If the initializer is zero, the warning is
5284 omitted. This is done under the assumption that the zero initializer in
5285 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
5286 initializer warnings and relies on default initialization to zero in the
5290 Conversions by prototypes between fixed/floating-point values and vice
5291 versa. The absence of these prototypes when compiling with traditional
5292 C causes serious problems. This is a subset of the possible
5293 conversion warnings; for the full set use @option{-Wtraditional-conversion}.
5296 Use of ISO C style function definitions. This warning intentionally is
5297 @emph{not} issued for prototype declarations or variadic functions
5298 because these ISO C features appear in your code when using
5299 libiberty's traditional C compatibility macros, @code{PARAMS} and
5300 @code{VPARAMS}. This warning is also bypassed for nested functions
5301 because that feature is already a GCC extension and thus not relevant to
5302 traditional C compatibility.
5305 @item -Wtraditional-conversion @r{(C and Objective-C only)}
5306 @opindex Wtraditional-conversion
5307 @opindex Wno-traditional-conversion
5308 Warn if a prototype causes a type conversion that is different from what
5309 would happen to the same argument in the absence of a prototype. This
5310 includes conversions of fixed point to floating and vice versa, and
5311 conversions changing the width or signedness of a fixed-point argument
5312 except when the same as the default promotion.
5314 @item -Wdeclaration-after-statement @r{(C and Objective-C only)}
5315 @opindex Wdeclaration-after-statement
5316 @opindex Wno-declaration-after-statement
5317 Warn when a declaration is found after a statement in a block. This
5318 construct, known from C++, was introduced with ISO C99 and is by default
5319 allowed in GCC@. It is not supported by ISO C90. @xref{Mixed Declarations}.
5324 Warn if an undefined identifier is evaluated in an @code{#if} directive.
5326 @item -Wno-endif-labels
5327 @opindex Wno-endif-labels
5328 @opindex Wendif-labels
5329 Do not warn whenever an @code{#else} or an @code{#endif} are followed by text.
5334 Warn whenever a local variable or type declaration shadows another
5335 variable, parameter, type, class member (in C++), or instance variable
5336 (in Objective-C) or whenever a built-in function is shadowed. Note
5337 that in C++, the compiler warns if a local variable shadows an
5338 explicit typedef, but not if it shadows a struct/class/enum.
5339 Same as @option{-Wshadow=global}.
5341 @item -Wno-shadow-ivar @r{(Objective-C only)}
5342 @opindex Wno-shadow-ivar
5343 @opindex Wshadow-ivar
5344 Do not warn whenever a local variable shadows an instance variable in an
5347 @item -Wshadow=global
5348 @opindex Wshadow=local
5349 The default for @option{-Wshadow}. Warns for any (global) shadowing.
5351 @item -Wshadow=local
5352 @opindex Wshadow=local
5353 Warn when a local variable shadows another local variable or parameter.
5354 This warning is enabled by @option{-Wshadow=global}.
5356 @item -Wshadow=compatible-local
5357 @opindex Wshadow=compatible-local
5358 Warn when a local variable shadows another local variable or parameter
5359 whose type is compatible with that of the shadowing variable. In C++,
5360 type compatibility here means the type of the shadowing variable can be
5361 converted to that of the shadowed variable. The creation of this flag
5362 (in addition to @option{-Wshadow=local}) is based on the idea that when
5363 a local variable shadows another one of incompatible type, it is most
5364 likely intentional, not a bug or typo, as shown in the following example:
5368 for (SomeIterator i = SomeObj.begin(); i != SomeObj.end(); ++i)
5370 for (int i = 0; i < N; ++i)
5379 Since the two variable @code{i} in the example above have incompatible types,
5380 enabling only @option{-Wshadow=compatible-local} will not emit a warning.
5381 Because their types are incompatible, if a programmer accidentally uses one
5382 in place of the other, type checking will catch that and emit an error or
5383 warning. So not warning (about shadowing) in this case will not lead to
5384 undetected bugs. Use of this flag instead of @option{-Wshadow=local} can
5385 possibly reduce the number of warnings triggered by intentional shadowing.
5387 This warning is enabled by @option{-Wshadow=local}.
5389 @item -Wlarger-than=@var{len}
5390 @opindex Wlarger-than=@var{len}
5391 @opindex Wlarger-than-@var{len}
5392 Warn whenever an object of larger than @var{len} bytes is defined.
5394 @item -Wframe-larger-than=@var{len}
5395 @opindex Wframe-larger-than
5396 Warn if the size of a function frame is larger than @var{len} bytes.
5397 The computation done to determine the stack frame size is approximate
5398 and not conservative.
5399 The actual requirements may be somewhat greater than @var{len}
5400 even if you do not get a warning. In addition, any space allocated
5401 via @code{alloca}, variable-length arrays, or related constructs
5402 is not included by the compiler when determining
5403 whether or not to issue a warning.
5405 @item -Wno-free-nonheap-object
5406 @opindex Wno-free-nonheap-object
5407 @opindex Wfree-nonheap-object
5408 Do not warn when attempting to free an object that was not allocated
5411 @item -Wstack-usage=@var{len}
5412 @opindex Wstack-usage
5413 Warn if the stack usage of a function might be larger than @var{len} bytes.
5414 The computation done to determine the stack usage is conservative.
5415 Any space allocated via @code{alloca}, variable-length arrays, or related
5416 constructs is included by the compiler when determining whether or not to
5419 The message is in keeping with the output of @option{-fstack-usage}.
5423 If the stack usage is fully static but exceeds the specified amount, it's:
5426 warning: stack usage is 1120 bytes
5429 If the stack usage is (partly) dynamic but bounded, it's:
5432 warning: stack usage might be 1648 bytes
5435 If the stack usage is (partly) dynamic and not bounded, it's:
5438 warning: stack usage might be unbounded
5442 @item -Wunsafe-loop-optimizations
5443 @opindex Wunsafe-loop-optimizations
5444 @opindex Wno-unsafe-loop-optimizations
5445 Warn if the loop cannot be optimized because the compiler cannot
5446 assume anything on the bounds of the loop indices. With
5447 @option{-funsafe-loop-optimizations} warn if the compiler makes
5450 @item -Wno-pedantic-ms-format @r{(MinGW targets only)}
5451 @opindex Wno-pedantic-ms-format
5452 @opindex Wpedantic-ms-format
5453 When used in combination with @option{-Wformat}
5454 and @option{-pedantic} without GNU extensions, this option
5455 disables the warnings about non-ISO @code{printf} / @code{scanf} format
5456 width specifiers @code{I32}, @code{I64}, and @code{I} used on Windows targets,
5457 which depend on the MS runtime.
5460 @opindex Waligned-new
5461 @opindex Wno-aligned-new
5462 Warn about a new-expression of a type that requires greater alignment
5463 than the @code{alignof(std::max_align_t)} but uses an allocation
5464 function without an explicit alignment parameter. This option is
5465 enabled by @option{-Wall}.
5467 Normally this only warns about global allocation functions, but
5468 @option{-Waligned-new=all} also warns about class member allocation
5471 @item -Wplacement-new
5472 @itemx -Wplacement-new=@var{n}
5473 @opindex Wplacement-new
5474 @opindex Wno-placement-new
5475 Warn about placement new expressions with undefined behavior, such as
5476 constructing an object in a buffer that is smaller than the type of
5477 the object. For example, the placement new expression below is diagnosed
5478 because it attempts to construct an array of 64 integers in a buffer only
5484 This warning is enabled by default.
5487 @item -Wplacement-new=1
5488 This is the default warning level of @option{-Wplacement-new}. At this
5489 level the warning is not issued for some strictly undefined constructs that
5490 GCC allows as extensions for compatibility with legacy code. For example,
5491 the following @code{new} expression is not diagnosed at this level even
5492 though it has undefined behavior according to the C++ standard because
5493 it writes past the end of the one-element array.
5495 struct S @{ int n, a[1]; @};
5496 S *s = (S *)malloc (sizeof *s + 31 * sizeof s->a[0]);
5497 new (s->a)int [32]();
5500 @item -Wplacement-new=2
5501 At this level, in addition to diagnosing all the same constructs as at level
5502 1, a diagnostic is also issued for placement new expressions that construct
5503 an object in the last member of structure whose type is an array of a single
5504 element and whose size is less than the size of the object being constructed.
5505 While the previous example would be diagnosed, the following construct makes
5506 use of the flexible member array extension to avoid the warning at level 2.
5508 struct S @{ int n, a[]; @};
5509 S *s = (S *)malloc (sizeof *s + 32 * sizeof s->a[0]);
5510 new (s->a)int [32]();
5515 @item -Wpointer-arith
5516 @opindex Wpointer-arith
5517 @opindex Wno-pointer-arith
5518 Warn about anything that depends on the ``size of'' a function type or
5519 of @code{void}. GNU C assigns these types a size of 1, for
5520 convenience in calculations with @code{void *} pointers and pointers
5521 to functions. In C++, warn also when an arithmetic operation involves
5522 @code{NULL}. This warning is also enabled by @option{-Wpedantic}.
5525 @opindex Wtype-limits
5526 @opindex Wno-type-limits
5527 Warn if a comparison is always true or always false due to the limited
5528 range of the data type, but do not warn for constant expressions. For
5529 example, warn if an unsigned variable is compared against zero with
5530 @code{<} or @code{>=}. This warning is also enabled by
5533 @item -Wbad-function-cast @r{(C and Objective-C only)}
5534 @opindex Wbad-function-cast
5535 @opindex Wno-bad-function-cast
5536 Warn when a function call is cast to a non-matching type.
5537 For example, warn if a call to a function returning an integer type
5538 is cast to a pointer type.
5540 @item -Wc90-c99-compat @r{(C and Objective-C only)}
5541 @opindex Wc90-c99-compat
5542 @opindex Wno-c90-c99-compat
5543 Warn about features not present in ISO C90, but present in ISO C99.
5544 For instance, warn about use of variable length arrays, @code{long long}
5545 type, @code{bool} type, compound literals, designated initializers, and so
5546 on. This option is independent of the standards mode. Warnings are disabled
5547 in the expression that follows @code{__extension__}.
5549 @item -Wc99-c11-compat @r{(C and Objective-C only)}
5550 @opindex Wc99-c11-compat
5551 @opindex Wno-c99-c11-compat
5552 Warn about features not present in ISO C99, but present in ISO C11.
5553 For instance, warn about use of anonymous structures and unions,
5554 @code{_Atomic} type qualifier, @code{_Thread_local} storage-class specifier,
5555 @code{_Alignas} specifier, @code{Alignof} operator, @code{_Generic} keyword,
5556 and so on. This option is independent of the standards mode. Warnings are
5557 disabled in the expression that follows @code{__extension__}.
5559 @item -Wc++-compat @r{(C and Objective-C only)}
5560 @opindex Wc++-compat
5561 Warn about ISO C constructs that are outside of the common subset of
5562 ISO C and ISO C++, e.g.@: request for implicit conversion from
5563 @code{void *} to a pointer to non-@code{void} type.
5565 @item -Wc++11-compat @r{(C++ and Objective-C++ only)}
5566 @opindex Wc++11-compat
5567 Warn about C++ constructs whose meaning differs between ISO C++ 1998
5568 and ISO C++ 2011, e.g., identifiers in ISO C++ 1998 that are keywords
5569 in ISO C++ 2011. This warning turns on @option{-Wnarrowing} and is
5570 enabled by @option{-Wall}.
5572 @item -Wc++14-compat @r{(C++ and Objective-C++ only)}
5573 @opindex Wc++14-compat
5574 Warn about C++ constructs whose meaning differs between ISO C++ 2011
5575 and ISO C++ 2014. This warning is enabled by @option{-Wall}.
5577 @item -Wc++1z-compat @r{(C++ and Objective-C++ only)}
5578 @opindex Wc++1z-compat
5579 Warn about C++ constructs whose meaning differs between ISO C++ 2014
5580 and the forthoming ISO C++ 2017(?). This warning is enabled by @option{-Wall}.
5584 @opindex Wno-cast-qual
5585 Warn whenever a pointer is cast so as to remove a type qualifier from
5586 the target type. For example, warn if a @code{const char *} is cast
5587 to an ordinary @code{char *}.
5589 Also warn when making a cast that introduces a type qualifier in an
5590 unsafe way. For example, casting @code{char **} to @code{const char **}
5591 is unsafe, as in this example:
5594 /* p is char ** value. */
5595 const char **q = (const char **) p;
5596 /* Assignment of readonly string to const char * is OK. */
5598 /* Now char** pointer points to read-only memory. */
5603 @opindex Wcast-align
5604 @opindex Wno-cast-align
5605 Warn whenever a pointer is cast such that the required alignment of the
5606 target is increased. For example, warn if a @code{char *} is cast to
5607 an @code{int *} on machines where integers can only be accessed at
5608 two- or four-byte boundaries.
5610 @item -Wwrite-strings
5611 @opindex Wwrite-strings
5612 @opindex Wno-write-strings
5613 When compiling C, give string constants the type @code{const
5614 char[@var{length}]} so that copying the address of one into a
5615 non-@code{const} @code{char *} pointer produces a warning. These
5616 warnings help you find at compile time code that can try to write
5617 into a string constant, but only if you have been very careful about
5618 using @code{const} in declarations and prototypes. Otherwise, it is
5619 just a nuisance. This is why we did not make @option{-Wall} request
5622 When compiling C++, warn about the deprecated conversion from string
5623 literals to @code{char *}. This warning is enabled by default for C++
5628 @opindex Wno-clobbered
5629 Warn for variables that might be changed by @code{longjmp} or
5630 @code{vfork}. This warning is also enabled by @option{-Wextra}.
5632 @item -Wconditionally-supported @r{(C++ and Objective-C++ only)}
5633 @opindex Wconditionally-supported
5634 @opindex Wno-conditionally-supported
5635 Warn for conditionally-supported (C++11 [intro.defs]) constructs.
5638 @opindex Wconversion
5639 @opindex Wno-conversion
5640 Warn for implicit conversions that may alter a value. This includes
5641 conversions between real and integer, like @code{abs (x)} when
5642 @code{x} is @code{double}; conversions between signed and unsigned,
5643 like @code{unsigned ui = -1}; and conversions to smaller types, like
5644 @code{sqrtf (M_PI)}. Do not warn for explicit casts like @code{abs
5645 ((int) x)} and @code{ui = (unsigned) -1}, or if the value is not
5646 changed by the conversion like in @code{abs (2.0)}. Warnings about
5647 conversions between signed and unsigned integers can be disabled by
5648 using @option{-Wno-sign-conversion}.
5650 For C++, also warn for confusing overload resolution for user-defined
5651 conversions; and conversions that never use a type conversion
5652 operator: conversions to @code{void}, the same type, a base class or a
5653 reference to them. Warnings about conversions between signed and
5654 unsigned integers are disabled by default in C++ unless
5655 @option{-Wsign-conversion} is explicitly enabled.
5657 @item -Wno-conversion-null @r{(C++ and Objective-C++ only)}
5658 @opindex Wconversion-null
5659 @opindex Wno-conversion-null
5660 Do not warn for conversions between @code{NULL} and non-pointer
5661 types. @option{-Wconversion-null} is enabled by default.
5663 @item -Wzero-as-null-pointer-constant @r{(C++ and Objective-C++ only)}
5664 @opindex Wzero-as-null-pointer-constant
5665 @opindex Wno-zero-as-null-pointer-constant
5666 Warn when a literal @samp{0} is used as null pointer constant. This can
5667 be useful to facilitate the conversion to @code{nullptr} in C++11.
5669 @item -Wsubobject-linkage @r{(C++ and Objective-C++ only)}
5670 @opindex Wsubobject-linkage
5671 @opindex Wno-subobject-linkage
5672 Warn if a class type has a base or a field whose type uses the anonymous
5673 namespace or depends on a type with no linkage. If a type A depends on
5674 a type B with no or internal linkage, defining it in multiple
5675 translation units would be an ODR violation because the meaning of B
5676 is different in each translation unit. If A only appears in a single
5677 translation unit, the best way to silence the warning is to give it
5678 internal linkage by putting it in an anonymous namespace as well. The
5679 compiler doesn't give this warning for types defined in the main .C
5680 file, as those are unlikely to have multiple definitions.
5681 @option{-Wsubobject-linkage} is enabled by default.
5683 @item -Wdangling-else
5684 @opindex Wdangling-else
5685 @opindex Wno-dangling-else
5686 Warn about constructions where there may be confusion to which
5687 @code{if} statement an @code{else} branch belongs. Here is an example of
5702 In C/C++, every @code{else} branch belongs to the innermost possible
5703 @code{if} statement, which in this example is @code{if (b)}. This is
5704 often not what the programmer expected, as illustrated in the above
5705 example by indentation the programmer chose. When there is the
5706 potential for this confusion, GCC issues a warning when this flag
5707 is specified. To eliminate the warning, add explicit braces around
5708 the innermost @code{if} statement so there is no way the @code{else}
5709 can belong to the enclosing @code{if}. The resulting code
5726 This warning is enabled by @option{-Wparentheses}.
5730 @opindex Wno-date-time
5731 Warn when macros @code{__TIME__}, @code{__DATE__} or @code{__TIMESTAMP__}
5732 are encountered as they might prevent bit-wise-identical reproducible
5735 @item -Wdelete-incomplete @r{(C++ and Objective-C++ only)}
5736 @opindex Wdelete-incomplete
5737 @opindex Wno-delete-incomplete
5738 Warn when deleting a pointer to incomplete type, which may cause
5739 undefined behavior at runtime. This warning is enabled by default.
5741 @item -Wuseless-cast @r{(C++ and Objective-C++ only)}
5742 @opindex Wuseless-cast
5743 @opindex Wno-useless-cast
5744 Warn when an expression is casted to its own type.
5747 @opindex Wempty-body
5748 @opindex Wno-empty-body
5749 Warn if an empty body occurs in an @code{if}, @code{else} or @code{do
5750 while} statement. This warning is also enabled by @option{-Wextra}.
5752 @item -Wenum-compare
5753 @opindex Wenum-compare
5754 @opindex Wno-enum-compare
5755 Warn about a comparison between values of different enumerated types.
5756 In C++ enumeral mismatches in conditional expressions are also
5757 diagnosed and the warning is enabled by default. In C this warning is
5758 enabled by @option{-Wall}.
5760 @item -Wjump-misses-init @r{(C, Objective-C only)}
5761 @opindex Wjump-misses-init
5762 @opindex Wno-jump-misses-init
5763 Warn if a @code{goto} statement or a @code{switch} statement jumps
5764 forward across the initialization of a variable, or jumps backward to a
5765 label after the variable has been initialized. This only warns about
5766 variables that are initialized when they are declared. This warning is
5767 only supported for C and Objective-C; in C++ this sort of branch is an
5770 @option{-Wjump-misses-init} is included in @option{-Wc++-compat}. It
5771 can be disabled with the @option{-Wno-jump-misses-init} option.
5773 @item -Wsign-compare
5774 @opindex Wsign-compare
5775 @opindex Wno-sign-compare
5776 @cindex warning for comparison of signed and unsigned values
5777 @cindex comparison of signed and unsigned values, warning
5778 @cindex signed and unsigned values, comparison warning
5779 Warn when a comparison between signed and unsigned values could produce
5780 an incorrect result when the signed value is converted to unsigned.
5781 In C++, this warning is also enabled by @option{-Wall}. In C, it is
5782 also enabled by @option{-Wextra}.
5784 @item -Wsign-conversion
5785 @opindex Wsign-conversion
5786 @opindex Wno-sign-conversion
5787 Warn for implicit conversions that may change the sign of an integer
5788 value, like assigning a signed integer expression to an unsigned
5789 integer variable. An explicit cast silences the warning. In C, this
5790 option is enabled also by @option{-Wconversion}.
5792 @item -Wfloat-conversion
5793 @opindex Wfloat-conversion
5794 @opindex Wno-float-conversion
5795 Warn for implicit conversions that reduce the precision of a real value.
5796 This includes conversions from real to integer, and from higher precision
5797 real to lower precision real values. This option is also enabled by
5798 @option{-Wconversion}.
5800 @item -Wno-scalar-storage-order
5801 @opindex -Wno-scalar-storage-order
5802 @opindex -Wscalar-storage-order
5803 Do not warn on suspicious constructs involving reverse scalar storage order.
5805 @item -Wsized-deallocation @r{(C++ and Objective-C++ only)}
5806 @opindex Wsized-deallocation
5807 @opindex Wno-sized-deallocation
5808 Warn about a definition of an unsized deallocation function
5810 void operator delete (void *) noexcept;
5811 void operator delete[] (void *) noexcept;
5813 without a definition of the corresponding sized deallocation function
5815 void operator delete (void *, std::size_t) noexcept;
5816 void operator delete[] (void *, std::size_t) noexcept;
5818 or vice versa. Enabled by @option{-Wextra} along with
5819 @option{-fsized-deallocation}.
5821 @item -Wsizeof-pointer-memaccess
5822 @opindex Wsizeof-pointer-memaccess
5823 @opindex Wno-sizeof-pointer-memaccess
5824 Warn for suspicious length parameters to certain string and memory built-in
5825 functions if the argument uses @code{sizeof}. This warning warns e.g.@:
5826 about @code{memset (ptr, 0, sizeof (ptr));} if @code{ptr} is not an array,
5827 but a pointer, and suggests a possible fix, or about
5828 @code{memcpy (&foo, ptr, sizeof (&foo));}. This warning is enabled by
5831 @item -Wsizeof-array-argument
5832 @opindex Wsizeof-array-argument
5833 @opindex Wno-sizeof-array-argument
5834 Warn when the @code{sizeof} operator is applied to a parameter that is
5835 declared as an array in a function definition. This warning is enabled by
5836 default for C and C++ programs.
5838 @item -Wmemset-elt-size
5839 @opindex Wmemset-elt-size
5840 @opindex Wno-memset-elt-size
5841 Warn for suspicious calls to the @code{memset} built-in function, if the
5842 first argument references an array, and the third argument is a number
5843 equal to the number of elements, but not equal to the size of the array
5844 in memory. This indicates that the user has omitted a multiplication by
5845 the element size. This warning is enabled by @option{-Wall}.
5847 @item -Wmemset-transposed-args
5848 @opindex Wmemset-transposed-args
5849 @opindex Wno-memset-transposed-args
5850 Warn for suspicious calls to the @code{memset} built-in function, if the
5851 second argument is not zero and the third argument is zero. This warns e.g.@
5852 about @code{memset (buf, sizeof buf, 0)} where most probably
5853 @code{memset (buf, 0, sizeof buf)} was meant instead. The diagnostics
5854 is only emitted if the third argument is literal zero. If it is some
5855 expression that is folded to zero, a cast of zero to some type, etc.,
5856 it is far less likely that the user has mistakenly exchanged the arguments
5857 and no warning is emitted. This warning is enabled by @option{-Wall}.
5861 @opindex Wno-address
5862 Warn about suspicious uses of memory addresses. These include using
5863 the address of a function in a conditional expression, such as
5864 @code{void func(void); if (func)}, and comparisons against the memory
5865 address of a string literal, such as @code{if (x == "abc")}. Such
5866 uses typically indicate a programmer error: the address of a function
5867 always evaluates to true, so their use in a conditional usually
5868 indicate that the programmer forgot the parentheses in a function
5869 call; and comparisons against string literals result in unspecified
5870 behavior and are not portable in C, so they usually indicate that the
5871 programmer intended to use @code{strcmp}. This warning is enabled by
5875 @opindex Wlogical-op
5876 @opindex Wno-logical-op
5877 Warn about suspicious uses of logical operators in expressions.
5878 This includes using logical operators in contexts where a
5879 bit-wise operator is likely to be expected. Also warns when
5880 the operands of a logical operator are the same:
5883 if (a < 0 && a < 0) @{ @dots{} @}
5886 @item -Wlogical-not-parentheses
5887 @opindex Wlogical-not-parentheses
5888 @opindex Wno-logical-not-parentheses
5889 Warn about logical not used on the left hand side operand of a comparison.
5890 This option does not warn if the right operand is considered to be a boolean
5891 expression. Its purpose is to detect suspicious code like the following:
5895 if (!a > 1) @{ @dots{} @}
5898 It is possible to suppress the warning by wrapping the LHS into
5901 if ((!a) > 1) @{ @dots{} @}
5904 This warning is enabled by @option{-Wall}.
5906 @item -Waggregate-return
5907 @opindex Waggregate-return
5908 @opindex Wno-aggregate-return
5909 Warn if any functions that return structures or unions are defined or
5910 called. (In languages where you can return an array, this also elicits
5913 @item -Wno-aggressive-loop-optimizations
5914 @opindex Wno-aggressive-loop-optimizations
5915 @opindex Waggressive-loop-optimizations
5916 Warn if in a loop with constant number of iterations the compiler detects
5917 undefined behavior in some statement during one or more of the iterations.
5919 @item -Wno-attributes
5920 @opindex Wno-attributes
5921 @opindex Wattributes
5922 Do not warn if an unexpected @code{__attribute__} is used, such as
5923 unrecognized attributes, function attributes applied to variables,
5924 etc. This does not stop errors for incorrect use of supported
5927 @item -Wno-builtin-macro-redefined
5928 @opindex Wno-builtin-macro-redefined
5929 @opindex Wbuiltin-macro-redefined
5930 Do not warn if certain built-in macros are redefined. This suppresses
5931 warnings for redefinition of @code{__TIMESTAMP__}, @code{__TIME__},
5932 @code{__DATE__}, @code{__FILE__}, and @code{__BASE_FILE__}.
5934 @item -Wstrict-prototypes @r{(C and Objective-C only)}
5935 @opindex Wstrict-prototypes
5936 @opindex Wno-strict-prototypes
5937 Warn if a function is declared or defined without specifying the
5938 argument types. (An old-style function definition is permitted without
5939 a warning if preceded by a declaration that specifies the argument
5942 @item -Wold-style-declaration @r{(C and Objective-C only)}
5943 @opindex Wold-style-declaration
5944 @opindex Wno-old-style-declaration
5945 Warn for obsolescent usages, according to the C Standard, in a
5946 declaration. For example, warn if storage-class specifiers like
5947 @code{static} are not the first things in a declaration. This warning
5948 is also enabled by @option{-Wextra}.
5950 @item -Wold-style-definition @r{(C and Objective-C only)}
5951 @opindex Wold-style-definition
5952 @opindex Wno-old-style-definition
5953 Warn if an old-style function definition is used. A warning is given
5954 even if there is a previous prototype.
5956 @item -Wmissing-parameter-type @r{(C and Objective-C only)}
5957 @opindex Wmissing-parameter-type
5958 @opindex Wno-missing-parameter-type
5959 A function parameter is declared without a type specifier in K&R-style
5966 This warning is also enabled by @option{-Wextra}.
5968 @item -Wmissing-prototypes @r{(C and Objective-C only)}
5969 @opindex Wmissing-prototypes
5970 @opindex Wno-missing-prototypes
5971 Warn if a global function is defined without a previous prototype
5972 declaration. This warning is issued even if the definition itself
5973 provides a prototype. Use this option to detect global functions
5974 that do not have a matching prototype declaration in a header file.
5975 This option is not valid for C++ because all function declarations
5976 provide prototypes and a non-matching declaration declares an
5977 overload rather than conflict with an earlier declaration.
5978 Use @option{-Wmissing-declarations} to detect missing declarations in C++.
5980 @item -Wmissing-declarations
5981 @opindex Wmissing-declarations
5982 @opindex Wno-missing-declarations
5983 Warn if a global function is defined without a previous declaration.
5984 Do so even if the definition itself provides a prototype.
5985 Use this option to detect global functions that are not declared in
5986 header files. In C, no warnings are issued for functions with previous
5987 non-prototype declarations; use @option{-Wmissing-prototypes} to detect
5988 missing prototypes. In C++, no warnings are issued for function templates,
5989 or for inline functions, or for functions in anonymous namespaces.
5991 @item -Wmissing-field-initializers
5992 @opindex Wmissing-field-initializers
5993 @opindex Wno-missing-field-initializers
5997 Warn if a structure's initializer has some fields missing. For
5998 example, the following code causes such a warning, because
5999 @code{x.h} is implicitly zero:
6002 struct s @{ int f, g, h; @};
6003 struct s x = @{ 3, 4 @};
6006 This option does not warn about designated initializers, so the following
6007 modification does not trigger a warning:
6010 struct s @{ int f, g, h; @};
6011 struct s x = @{ .f = 3, .g = 4 @};
6014 In C++ this option does not warn either about the empty @{ @}
6015 initializer, for example:
6018 struct s @{ int f, g, h; @};
6022 This warning is included in @option{-Wextra}. To get other @option{-Wextra}
6023 warnings without this one, use @option{-Wextra -Wno-missing-field-initializers}.
6025 @item -Wno-multichar
6026 @opindex Wno-multichar
6028 Do not warn if a multicharacter constant (@samp{'FOOF'}) is used.
6029 Usually they indicate a typo in the user's code, as they have
6030 implementation-defined values, and should not be used in portable code.
6032 @item -Wnormalized@r{[}=@r{<}none@r{|}id@r{|}nfc@r{|}nfkc@r{>]}
6033 @opindex Wnormalized=
6034 @opindex Wnormalized
6035 @opindex Wno-normalized
6038 @cindex character set, input normalization
6039 In ISO C and ISO C++, two identifiers are different if they are
6040 different sequences of characters. However, sometimes when characters
6041 outside the basic ASCII character set are used, you can have two
6042 different character sequences that look the same. To avoid confusion,
6043 the ISO 10646 standard sets out some @dfn{normalization rules} which
6044 when applied ensure that two sequences that look the same are turned into
6045 the same sequence. GCC can warn you if you are using identifiers that
6046 have not been normalized; this option controls that warning.
6048 There are four levels of warning supported by GCC@. The default is
6049 @option{-Wnormalized=nfc}, which warns about any identifier that is
6050 not in the ISO 10646 ``C'' normalized form, @dfn{NFC}. NFC is the
6051 recommended form for most uses. It is equivalent to
6052 @option{-Wnormalized}.
6054 Unfortunately, there are some characters allowed in identifiers by
6055 ISO C and ISO C++ that, when turned into NFC, are not allowed in
6056 identifiers. That is, there's no way to use these symbols in portable
6057 ISO C or C++ and have all your identifiers in NFC@.
6058 @option{-Wnormalized=id} suppresses the warning for these characters.
6059 It is hoped that future versions of the standards involved will correct
6060 this, which is why this option is not the default.
6062 You can switch the warning off for all characters by writing
6063 @option{-Wnormalized=none} or @option{-Wno-normalized}. You should
6064 only do this if you are using some other normalization scheme (like
6065 ``D''), because otherwise you can easily create bugs that are
6066 literally impossible to see.
6068 Some characters in ISO 10646 have distinct meanings but look identical
6069 in some fonts or display methodologies, especially once formatting has
6070 been applied. For instance @code{\u207F}, ``SUPERSCRIPT LATIN SMALL
6071 LETTER N'', displays just like a regular @code{n} that has been
6072 placed in a superscript. ISO 10646 defines the @dfn{NFKC}
6073 normalization scheme to convert all these into a standard form as
6074 well, and GCC warns if your code is not in NFKC if you use
6075 @option{-Wnormalized=nfkc}. This warning is comparable to warning
6076 about every identifier that contains the letter O because it might be
6077 confused with the digit 0, and so is not the default, but may be
6078 useful as a local coding convention if the programming environment
6079 cannot be fixed to display these characters distinctly.
6081 @item -Wno-deprecated
6082 @opindex Wno-deprecated
6083 @opindex Wdeprecated
6084 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
6086 @item -Wno-deprecated-declarations
6087 @opindex Wno-deprecated-declarations
6088 @opindex Wdeprecated-declarations
6089 Do not warn about uses of functions (@pxref{Function Attributes}),
6090 variables (@pxref{Variable Attributes}), and types (@pxref{Type
6091 Attributes}) marked as deprecated by using the @code{deprecated}
6095 @opindex Wno-overflow
6097 Do not warn about compile-time overflow in constant expressions.
6102 Warn about One Definition Rule violations during link-time optimization.
6103 Requires @option{-flto-odr-type-merging} to be enabled. Enabled by default.
6106 @opindex Wopenm-simd
6107 Warn if the vectorizer cost model overrides the OpenMP or the Cilk Plus
6108 simd directive set by user. The @option{-fsimd-cost-model=unlimited}
6109 option can be used to relax the cost model.
6111 @item -Woverride-init @r{(C and Objective-C only)}
6112 @opindex Woverride-init
6113 @opindex Wno-override-init
6117 Warn if an initialized field without side effects is overridden when
6118 using designated initializers (@pxref{Designated Inits, , Designated
6121 This warning is included in @option{-Wextra}. To get other
6122 @option{-Wextra} warnings without this one, use @option{-Wextra
6123 -Wno-override-init}.
6125 @item -Woverride-init-side-effects @r{(C and Objective-C only)}
6126 @opindex Woverride-init-side-effects
6127 @opindex Wno-override-init-side-effects
6128 Warn if an initialized field with side effects is overridden when
6129 using designated initializers (@pxref{Designated Inits, , Designated
6130 Initializers}). This warning is enabled by default.
6135 Warn if a structure is given the packed attribute, but the packed
6136 attribute has no effect on the layout or size of the structure.
6137 Such structures may be mis-aligned for little benefit. For
6138 instance, in this code, the variable @code{f.x} in @code{struct bar}
6139 is misaligned even though @code{struct bar} does not itself
6140 have the packed attribute:
6147 @} __attribute__((packed));
6155 @item -Wpacked-bitfield-compat
6156 @opindex Wpacked-bitfield-compat
6157 @opindex Wno-packed-bitfield-compat
6158 The 4.1, 4.2 and 4.3 series of GCC ignore the @code{packed} attribute
6159 on bit-fields of type @code{char}. This has been fixed in GCC 4.4 but
6160 the change can lead to differences in the structure layout. GCC
6161 informs you when the offset of such a field has changed in GCC 4.4.
6162 For example there is no longer a 4-bit padding between field @code{a}
6163 and @code{b} in this structure:
6170 @} __attribute__ ((packed));
6173 This warning is enabled by default. Use
6174 @option{-Wno-packed-bitfield-compat} to disable this warning.
6179 Warn if padding is included in a structure, either to align an element
6180 of the structure or to align the whole structure. Sometimes when this
6181 happens it is possible to rearrange the fields of the structure to
6182 reduce the padding and so make the structure smaller.
6184 @item -Wredundant-decls
6185 @opindex Wredundant-decls
6186 @opindex Wno-redundant-decls
6187 Warn if anything is declared more than once in the same scope, even in
6188 cases where multiple declaration is valid and changes nothing.
6190 @item -Wnested-externs @r{(C and Objective-C only)}
6191 @opindex Wnested-externs
6192 @opindex Wno-nested-externs
6193 Warn if an @code{extern} declaration is encountered within a function.
6195 @item -Wno-inherited-variadic-ctor
6196 @opindex Winherited-variadic-ctor
6197 @opindex Wno-inherited-variadic-ctor
6198 Suppress warnings about use of C++11 inheriting constructors when the
6199 base class inherited from has a C variadic constructor; the warning is
6200 on by default because the ellipsis is not inherited.
6205 Warn if a function that is declared as inline cannot be inlined.
6206 Even with this option, the compiler does not warn about failures to
6207 inline functions declared in system headers.
6209 The compiler uses a variety of heuristics to determine whether or not
6210 to inline a function. For example, the compiler takes into account
6211 the size of the function being inlined and the amount of inlining
6212 that has already been done in the current function. Therefore,
6213 seemingly insignificant changes in the source program can cause the
6214 warnings produced by @option{-Winline} to appear or disappear.
6216 @item -Wno-invalid-offsetof @r{(C++ and Objective-C++ only)}
6217 @opindex Wno-invalid-offsetof
6218 @opindex Winvalid-offsetof
6219 Suppress warnings from applying the @code{offsetof} macro to a non-POD
6220 type. According to the 2014 ISO C++ standard, applying @code{offsetof}
6221 to a non-standard-layout type is undefined. In existing C++ implementations,
6222 however, @code{offsetof} typically gives meaningful results.
6223 This flag is for users who are aware that they are
6224 writing nonportable code and who have deliberately chosen to ignore the
6227 The restrictions on @code{offsetof} may be relaxed in a future version
6228 of the C++ standard.
6230 @item -Wint-in-bool-context
6231 @opindex Wint-in-bool-context
6232 @opindex Wno-int-in-bool-context
6233 Warn for suspicious use of integer values where boolean values are expected,
6234 such as conditional expressions (?:) using non-boolean integer constants in
6235 boolean context, like @code{if (a <= b ? 2 : 3)}. Or left shifting of signed
6236 integers in boolean context, like @code{for (a = 0; 1 << a; a++);}. Likewise
6237 for all kinds of multiplications regardless of the data type.
6238 This warning is enabled by @option{-Wall}.
6240 @item -Wno-int-to-pointer-cast
6241 @opindex Wno-int-to-pointer-cast
6242 @opindex Wint-to-pointer-cast
6243 Suppress warnings from casts to pointer type of an integer of a
6244 different size. In C++, casting to a pointer type of smaller size is
6245 an error. @option{Wint-to-pointer-cast} is enabled by default.
6248 @item -Wno-pointer-to-int-cast @r{(C and Objective-C only)}
6249 @opindex Wno-pointer-to-int-cast
6250 @opindex Wpointer-to-int-cast
6251 Suppress warnings from casts from a pointer to an integer type of a
6255 @opindex Winvalid-pch
6256 @opindex Wno-invalid-pch
6257 Warn if a precompiled header (@pxref{Precompiled Headers}) is found in
6258 the search path but can't be used.
6262 @opindex Wno-long-long
6263 Warn if @code{long long} type is used. This is enabled by either
6264 @option{-Wpedantic} or @option{-Wtraditional} in ISO C90 and C++98
6265 modes. To inhibit the warning messages, use @option{-Wno-long-long}.
6267 @item -Wvariadic-macros
6268 @opindex Wvariadic-macros
6269 @opindex Wno-variadic-macros
6270 Warn if variadic macros are used in ISO C90 mode, or if the GNU
6271 alternate syntax is used in ISO C99 mode. This is enabled by either
6272 @option{-Wpedantic} or @option{-Wtraditional}. To inhibit the warning
6273 messages, use @option{-Wno-variadic-macros}.
6277 @opindex Wno-varargs
6278 Warn upon questionable usage of the macros used to handle variable
6279 arguments like @code{va_start}. This is default. To inhibit the
6280 warning messages, use @option{-Wno-varargs}.
6282 @item -Wvector-operation-performance
6283 @opindex Wvector-operation-performance
6284 @opindex Wno-vector-operation-performance
6285 Warn if vector operation is not implemented via SIMD capabilities of the
6286 architecture. Mainly useful for the performance tuning.
6287 Vector operation can be implemented @code{piecewise}, which means that the
6288 scalar operation is performed on every vector element;
6289 @code{in parallel}, which means that the vector operation is implemented
6290 using scalars of wider type, which normally is more performance efficient;
6291 and @code{as a single scalar}, which means that vector fits into a
6294 @item -Wno-virtual-move-assign
6295 @opindex Wvirtual-move-assign
6296 @opindex Wno-virtual-move-assign
6297 Suppress warnings about inheriting from a virtual base with a
6298 non-trivial C++11 move assignment operator. This is dangerous because
6299 if the virtual base is reachable along more than one path, it is
6300 moved multiple times, which can mean both objects end up in the
6301 moved-from state. If the move assignment operator is written to avoid
6302 moving from a moved-from object, this warning can be disabled.
6307 Warn if a variable-length array is used in the code.
6308 @option{-Wno-vla} prevents the @option{-Wpedantic} warning of
6309 the variable-length array.
6311 @item -Wvla-larger-than=@var{n}
6312 If this option is used, the compiler will warn on uses of
6313 variable-length arrays where the size is either unbounded, or bounded
6314 by an argument that can be larger than @var{n} bytes. This is similar
6315 to how @option{-Walloca-larger-than=@var{n}} works, but with
6316 variable-length arrays.
6318 Note that GCC may optimize small variable-length arrays of a known
6319 value into plain arrays, so this warning may not get triggered for
6322 This warning is not enabled by @option{-Wall}, and is only active when
6323 @option{-ftree-vrp} is active (default for @option{-O2} and above).
6325 See also @option{-Walloca-larger-than=@var{n}}.
6327 @item -Wvolatile-register-var
6328 @opindex Wvolatile-register-var
6329 @opindex Wno-volatile-register-var
6330 Warn if a register variable is declared volatile. The volatile
6331 modifier does not inhibit all optimizations that may eliminate reads
6332 and/or writes to register variables. This warning is enabled by
6335 @item -Wdisabled-optimization
6336 @opindex Wdisabled-optimization
6337 @opindex Wno-disabled-optimization
6338 Warn if a requested optimization pass is disabled. This warning does
6339 not generally indicate that there is anything wrong with your code; it
6340 merely indicates that GCC's optimizers are unable to handle the code
6341 effectively. Often, the problem is that your code is too big or too
6342 complex; GCC refuses to optimize programs when the optimization
6343 itself is likely to take inordinate amounts of time.
6345 @item -Wpointer-sign @r{(C and Objective-C only)}
6346 @opindex Wpointer-sign
6347 @opindex Wno-pointer-sign
6348 Warn for pointer argument passing or assignment with different signedness.
6349 This option is only supported for C and Objective-C@. It is implied by
6350 @option{-Wall} and by @option{-Wpedantic}, which can be disabled with
6351 @option{-Wno-pointer-sign}.
6353 @item -Wstack-protector
6354 @opindex Wstack-protector
6355 @opindex Wno-stack-protector
6356 This option is only active when @option{-fstack-protector} is active. It
6357 warns about functions that are not protected against stack smashing.
6359 @item -Woverlength-strings
6360 @opindex Woverlength-strings
6361 @opindex Wno-overlength-strings
6362 Warn about string constants that are longer than the ``minimum
6363 maximum'' length specified in the C standard. Modern compilers
6364 generally allow string constants that are much longer than the
6365 standard's minimum limit, but very portable programs should avoid
6366 using longer strings.
6368 The limit applies @emph{after} string constant concatenation, and does
6369 not count the trailing NUL@. In C90, the limit was 509 characters; in
6370 C99, it was raised to 4095. C++98 does not specify a normative
6371 minimum maximum, so we do not diagnose overlength strings in C++@.
6373 This option is implied by @option{-Wpedantic}, and can be disabled with
6374 @option{-Wno-overlength-strings}.
6376 @item -Wunsuffixed-float-constants @r{(C and Objective-C only)}
6377 @opindex Wunsuffixed-float-constants
6379 Issue a warning for any floating constant that does not have
6380 a suffix. When used together with @option{-Wsystem-headers} it
6381 warns about such constants in system header files. This can be useful
6382 when preparing code to use with the @code{FLOAT_CONST_DECIMAL64} pragma
6383 from the decimal floating-point extension to C99.
6385 @item -Wno-designated-init @r{(C and Objective-C only)}
6386 Suppress warnings when a positional initializer is used to initialize
6387 a structure that has been marked with the @code{designated_init}
6391 Issue a warning when HSAIL cannot be emitted for the compiled function or
6396 @node Debugging Options
6397 @section Options for Debugging Your Program
6398 @cindex options, debugging
6399 @cindex debugging information options
6401 To tell GCC to emit extra information for use by a debugger, in almost
6402 all cases you need only to add @option{-g} to your other options.
6404 GCC allows you to use @option{-g} with
6405 @option{-O}. The shortcuts taken by optimized code may occasionally
6406 be surprising: some variables you declared may not exist
6407 at all; flow of control may briefly move where you did not expect it;
6408 some statements may not be executed because they compute constant
6409 results or their values are already at hand; some statements may
6410 execute in different places because they have been moved out of loops.
6411 Nevertheless it is possible to debug optimized output. This makes
6412 it reasonable to use the optimizer for programs that might have bugs.
6414 If you are not using some other optimization option, consider
6415 using @option{-Og} (@pxref{Optimize Options}) with @option{-g}.
6416 With no @option{-O} option at all, some compiler passes that collect
6417 information useful for debugging do not run at all, so that
6418 @option{-Og} may result in a better debugging experience.
6423 Produce debugging information in the operating system's native format
6424 (stabs, COFF, XCOFF, or DWARF)@. GDB can work with this debugging
6427 On most systems that use stabs format, @option{-g} enables use of extra
6428 debugging information that only GDB can use; this extra information
6429 makes debugging work better in GDB but probably makes other debuggers
6431 refuse to read the program. If you want to control for certain whether
6432 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
6433 @option{-gxcoff+}, @option{-gxcoff}, or @option{-gvms} (see below).
6437 Produce debugging information for use by GDB@. This means to use the
6438 most expressive format available (DWARF, stabs, or the native format
6439 if neither of those are supported), including GDB extensions if at all
6443 @itemx -gdwarf-@var{version}
6445 Produce debugging information in DWARF format (if that is supported).
6446 The value of @var{version} may be either 2, 3, 4 or 5; the default version
6447 for most targets is 4. DWARF Version 5 is only experimental.
6449 Note that with DWARF Version 2, some ports require and always
6450 use some non-conflicting DWARF 3 extensions in the unwind tables.
6452 Version 4 may require GDB 7.0 and @option{-fvar-tracking-assignments}
6453 for maximum benefit.
6455 GCC no longer supports DWARF Version 1, which is substantially
6456 different than Version 2 and later. For historical reasons, some
6457 other DWARF-related options (including @option{-feliminate-dwarf2-dups}
6458 and @option{-fno-dwarf2-cfi-asm}) retain a reference to DWARF Version 2
6459 in their names, but apply to all currently-supported versions of DWARF.
6463 Produce debugging information in stabs format (if that is supported),
6464 without GDB extensions. This is the format used by DBX on most BSD
6465 systems. On MIPS, Alpha and System V Release 4 systems this option
6466 produces stabs debugging output that is not understood by DBX or SDB@.
6467 On System V Release 4 systems this option requires the GNU assembler.
6471 Produce debugging information in stabs format (if that is supported),
6472 using GNU extensions understood only by the GNU debugger (GDB)@. The
6473 use of these extensions is likely to make other debuggers crash or
6474 refuse to read the program.
6478 Produce debugging information in COFF format (if that is supported).
6479 This is the format used by SDB on most System V systems prior to
6484 Produce debugging information in XCOFF format (if that is supported).
6485 This is the format used by the DBX debugger on IBM RS/6000 systems.
6489 Produce debugging information in XCOFF format (if that is supported),
6490 using GNU extensions understood only by the GNU debugger (GDB)@. The
6491 use of these extensions is likely to make other debuggers crash or
6492 refuse to read the program, and may cause assemblers other than the GNU
6493 assembler (GAS) to fail with an error.
6497 Produce debugging information in Alpha/VMS debug format (if that is
6498 supported). This is the format used by DEBUG on Alpha/VMS systems.
6501 @itemx -ggdb@var{level}
6502 @itemx -gstabs@var{level}
6503 @itemx -gcoff@var{level}
6504 @itemx -gxcoff@var{level}
6505 @itemx -gvms@var{level}
6506 Request debugging information and also use @var{level} to specify how
6507 much information. The default level is 2.
6509 Level 0 produces no debug information at all. Thus, @option{-g0} negates
6512 Level 1 produces minimal information, enough for making backtraces in
6513 parts of the program that you don't plan to debug. This includes
6514 descriptions of functions and external variables, and line number
6515 tables, but no information about local variables.
6517 Level 3 includes extra information, such as all the macro definitions
6518 present in the program. Some debuggers support macro expansion when
6519 you use @option{-g3}.
6521 @option{-gdwarf} does not accept a concatenated debug level, to avoid
6522 confusion with @option{-gdwarf-@var{level}}.
6523 Instead use an additional @option{-g@var{level}} option to change the
6524 debug level for DWARF.
6526 @item -feliminate-unused-debug-symbols
6527 @opindex feliminate-unused-debug-symbols
6528 Produce debugging information in stabs format (if that is supported),
6529 for only symbols that are actually used.
6531 @item -femit-class-debug-always
6532 @opindex femit-class-debug-always
6533 Instead of emitting debugging information for a C++ class in only one
6534 object file, emit it in all object files using the class. This option
6535 should be used only with debuggers that are unable to handle the way GCC
6536 normally emits debugging information for classes because using this
6537 option increases the size of debugging information by as much as a
6540 @item -fno-merge-debug-strings
6541 @opindex fmerge-debug-strings
6542 @opindex fno-merge-debug-strings
6543 Direct the linker to not merge together strings in the debugging
6544 information that are identical in different object files. Merging is
6545 not supported by all assemblers or linkers. Merging decreases the size
6546 of the debug information in the output file at the cost of increasing
6547 link processing time. Merging is enabled by default.
6549 @item -fdebug-prefix-map=@var{old}=@var{new}
6550 @opindex fdebug-prefix-map
6551 When compiling files in directory @file{@var{old}}, record debugging
6552 information describing them as in @file{@var{new}} instead.
6554 @item -fvar-tracking
6555 @opindex fvar-tracking
6556 Run variable tracking pass. It computes where variables are stored at each
6557 position in code. Better debugging information is then generated
6558 (if the debugging information format supports this information).
6560 It is enabled by default when compiling with optimization (@option{-Os},
6561 @option{-O}, @option{-O2}, @dots{}), debugging information (@option{-g}) and
6562 the debug info format supports it.
6564 @item -fvar-tracking-assignments
6565 @opindex fvar-tracking-assignments
6566 @opindex fno-var-tracking-assignments
6567 Annotate assignments to user variables early in the compilation and
6568 attempt to carry the annotations over throughout the compilation all the
6569 way to the end, in an attempt to improve debug information while
6570 optimizing. Use of @option{-gdwarf-4} is recommended along with it.
6572 It can be enabled even if var-tracking is disabled, in which case
6573 annotations are created and maintained, but discarded at the end.
6574 By default, this flag is enabled together with @option{-fvar-tracking},
6575 except when selective scheduling is enabled.
6578 @opindex gsplit-dwarf
6579 Separate as much DWARF debugging information as possible into a
6580 separate output file with the extension @file{.dwo}. This option allows
6581 the build system to avoid linking files with debug information. To
6582 be useful, this option requires a debugger capable of reading @file{.dwo}
6587 Generate DWARF @code{.debug_pubnames} and @code{.debug_pubtypes} sections.
6589 @item -ggnu-pubnames
6590 @opindex ggnu-pubnames
6591 Generate @code{.debug_pubnames} and @code{.debug_pubtypes} sections in a format
6592 suitable for conversion into a GDB@ index. This option is only useful
6593 with a linker that can produce GDB@ index version 7.
6595 @item -fdebug-types-section
6596 @opindex fdebug-types-section
6597 @opindex fno-debug-types-section
6598 When using DWARF Version 4 or higher, type DIEs can be put into
6599 their own @code{.debug_types} section instead of making them part of the
6600 @code{.debug_info} section. It is more efficient to put them in a separate
6601 comdat sections since the linker can then remove duplicates.
6602 But not all DWARF consumers support @code{.debug_types} sections yet
6603 and on some objects @code{.debug_types} produces larger instead of smaller
6604 debugging information.
6606 @item -grecord-gcc-switches
6607 @item -gno-record-gcc-switches
6608 @opindex grecord-gcc-switches
6609 @opindex gno-record-gcc-switches
6610 This switch causes the command-line options used to invoke the
6611 compiler that may affect code generation to be appended to the
6612 DW_AT_producer attribute in DWARF debugging information. The options
6613 are concatenated with spaces separating them from each other and from
6614 the compiler version.
6615 It is enabled by default.
6616 See also @option{-frecord-gcc-switches} for another
6617 way of storing compiler options into the object file.
6619 @item -gstrict-dwarf
6620 @opindex gstrict-dwarf
6621 Disallow using extensions of later DWARF standard version than selected
6622 with @option{-gdwarf-@var{version}}. On most targets using non-conflicting
6623 DWARF extensions from later standard versions is allowed.
6625 @item -gno-strict-dwarf
6626 @opindex gno-strict-dwarf
6627 Allow using extensions of later DWARF standard version than selected with
6628 @option{-gdwarf-@var{version}}.
6630 @item -gz@r{[}=@var{type}@r{]}
6632 Produce compressed debug sections in DWARF format, if that is supported.
6633 If @var{type} is not given, the default type depends on the capabilities
6634 of the assembler and linker used. @var{type} may be one of
6635 @samp{none} (don't compress debug sections), @samp{zlib} (use zlib
6636 compression in ELF gABI format), or @samp{zlib-gnu} (use zlib
6637 compression in traditional GNU format). If the linker doesn't support
6638 writing compressed debug sections, the option is rejected. Otherwise,
6639 if the assembler does not support them, @option{-gz} is silently ignored
6640 when producing object files.
6642 @item -feliminate-dwarf2-dups
6643 @opindex feliminate-dwarf2-dups
6644 Compress DWARF debugging information by eliminating duplicated
6645 information about each symbol. This option only makes sense when
6646 generating DWARF debugging information.
6648 @item -femit-struct-debug-baseonly
6649 @opindex femit-struct-debug-baseonly
6650 Emit debug information for struct-like types
6651 only when the base name of the compilation source file
6652 matches the base name of file in which the struct is defined.
6654 This option substantially reduces the size of debugging information,
6655 but at significant potential loss in type information to the debugger.
6656 See @option{-femit-struct-debug-reduced} for a less aggressive option.
6657 See @option{-femit-struct-debug-detailed} for more detailed control.
6659 This option works only with DWARF debug output.
6661 @item -femit-struct-debug-reduced
6662 @opindex femit-struct-debug-reduced
6663 Emit debug information for struct-like types
6664 only when the base name of the compilation source file
6665 matches the base name of file in which the type is defined,
6666 unless the struct is a template or defined in a system header.
6668 This option significantly reduces the size of debugging information,
6669 with some potential loss in type information to the debugger.
6670 See @option{-femit-struct-debug-baseonly} for a more aggressive option.
6671 See @option{-femit-struct-debug-detailed} for more detailed control.
6673 This option works only with DWARF debug output.
6675 @item -femit-struct-debug-detailed@r{[}=@var{spec-list}@r{]}
6676 @opindex femit-struct-debug-detailed
6677 Specify the struct-like types
6678 for which the compiler generates debug information.
6679 The intent is to reduce duplicate struct debug information
6680 between different object files within the same program.
6682 This option is a detailed version of
6683 @option{-femit-struct-debug-reduced} and @option{-femit-struct-debug-baseonly},
6684 which serves for most needs.
6686 A specification has the syntax@*
6687 [@samp{dir:}|@samp{ind:}][@samp{ord:}|@samp{gen:}](@samp{any}|@samp{sys}|@samp{base}|@samp{none})
6689 The optional first word limits the specification to
6690 structs that are used directly (@samp{dir:}) or used indirectly (@samp{ind:}).
6691 A struct type is used directly when it is the type of a variable, member.
6692 Indirect uses arise through pointers to structs.
6693 That is, when use of an incomplete struct is valid, the use is indirect.
6695 @samp{struct one direct; struct two * indirect;}.
6697 The optional second word limits the specification to
6698 ordinary structs (@samp{ord:}) or generic structs (@samp{gen:}).
6699 Generic structs are a bit complicated to explain.
6700 For C++, these are non-explicit specializations of template classes,
6701 or non-template classes within the above.
6702 Other programming languages have generics,
6703 but @option{-femit-struct-debug-detailed} does not yet implement them.
6705 The third word specifies the source files for those
6706 structs for which the compiler should emit debug information.
6707 The values @samp{none} and @samp{any} have the normal meaning.
6708 The value @samp{base} means that
6709 the base of name of the file in which the type declaration appears
6710 must match the base of the name of the main compilation file.
6711 In practice, this means that when compiling @file{foo.c}, debug information
6712 is generated for types declared in that file and @file{foo.h},
6713 but not other header files.
6714 The value @samp{sys} means those types satisfying @samp{base}
6715 or declared in system or compiler headers.
6717 You may need to experiment to determine the best settings for your application.
6719 The default is @option{-femit-struct-debug-detailed=all}.
6721 This option works only with DWARF debug output.
6723 @item -fno-dwarf2-cfi-asm
6724 @opindex fdwarf2-cfi-asm
6725 @opindex fno-dwarf2-cfi-asm
6726 Emit DWARF unwind info as compiler generated @code{.eh_frame} section
6727 instead of using GAS @code{.cfi_*} directives.
6729 @item -fno-eliminate-unused-debug-types
6730 @opindex feliminate-unused-debug-types
6731 @opindex fno-eliminate-unused-debug-types
6732 Normally, when producing DWARF output, GCC avoids producing debug symbol
6733 output for types that are nowhere used in the source file being compiled.
6734 Sometimes it is useful to have GCC emit debugging
6735 information for all types declared in a compilation
6736 unit, regardless of whether or not they are actually used
6737 in that compilation unit, for example
6738 if, in the debugger, you want to cast a value to a type that is
6739 not actually used in your program (but is declared). More often,
6740 however, this results in a significant amount of wasted space.
6743 @node Optimize Options
6744 @section Options That Control Optimization
6745 @cindex optimize options
6746 @cindex options, optimization
6748 These options control various sorts of optimizations.
6750 Without any optimization option, the compiler's goal is to reduce the
6751 cost of compilation and to make debugging produce the expected
6752 results. Statements are independent: if you stop the program with a
6753 breakpoint between statements, you can then assign a new value to any
6754 variable or change the program counter to any other statement in the
6755 function and get exactly the results you expect from the source
6758 Turning on optimization flags makes the compiler attempt to improve
6759 the performance and/or code size at the expense of compilation time
6760 and possibly the ability to debug the program.
6762 The compiler performs optimization based on the knowledge it has of the
6763 program. Compiling multiple files at once to a single output file mode allows
6764 the compiler to use information gained from all of the files when compiling
6767 Not all optimizations are controlled directly by a flag. Only
6768 optimizations that have a flag are listed in this section.
6770 Most optimizations are only enabled if an @option{-O} level is set on
6771 the command line. Otherwise they are disabled, even if individual
6772 optimization flags are specified.
6774 Depending on the target and how GCC was configured, a slightly different
6775 set of optimizations may be enabled at each @option{-O} level than
6776 those listed here. You can invoke GCC with @option{-Q --help=optimizers}
6777 to find out the exact set of optimizations that are enabled at each level.
6778 @xref{Overall Options}, for examples.
6785 Optimize. Optimizing compilation takes somewhat more time, and a lot
6786 more memory for a large function.
6788 With @option{-O}, the compiler tries to reduce code size and execution
6789 time, without performing any optimizations that take a great deal of
6792 @option{-O} turns on the following optimization flags:
6795 -fbranch-count-reg @gol
6796 -fcombine-stack-adjustments @gol
6798 -fcprop-registers @gol
6801 -fdelayed-branch @gol
6803 -fforward-propagate @gol
6804 -fguess-branch-probability @gol
6805 -fif-conversion2 @gol
6806 -fif-conversion @gol
6807 -finline-functions-called-once @gol
6808 -fipa-pure-const @gol
6810 -fipa-reference @gol
6811 -fmerge-constants @gol
6812 -fmove-loop-invariants @gol
6813 -freorder-blocks @gol
6815 -fshrink-wrap-separate @gol
6816 -fsplit-wide-types @gol
6819 -fstore-merging @gol
6823 -ftree-coalesce-vars @gol
6824 -ftree-copy-prop @gol
6826 -ftree-dominator-opts @gol
6828 -ftree-forwprop @gol
6838 @option{-O} also turns on @option{-fomit-frame-pointer} on machines
6839 where doing so does not interfere with debugging.
6843 Optimize even more. GCC performs nearly all supported optimizations
6844 that do not involve a space-speed tradeoff.
6845 As compared to @option{-O}, this option increases both compilation time
6846 and the performance of the generated code.
6848 @option{-O2} turns on all optimization flags specified by @option{-O}. It
6849 also turns on the following optimization flags:
6850 @gccoptlist{-fthread-jumps @gol
6851 -falign-functions -falign-jumps @gol
6852 -falign-loops -falign-labels @gol
6855 -fcse-follow-jumps -fcse-skip-blocks @gol
6856 -fdelete-null-pointer-checks @gol
6857 -fdevirtualize -fdevirtualize-speculatively @gol
6858 -fexpensive-optimizations @gol
6859 -fgcse -fgcse-lm @gol
6860 -fhoist-adjacent-loads @gol
6861 -finline-small-functions @gol
6862 -findirect-inlining @gol
6864 -fipa-cp-alignment @gol
6868 -fisolate-erroneous-paths-dereference @gol
6870 -foptimize-sibling-calls @gol
6871 -foptimize-strlen @gol
6872 -fpartial-inlining @gol
6874 -freorder-blocks-algorithm=stc @gol
6875 -freorder-blocks-and-partition -freorder-functions @gol
6876 -frerun-cse-after-loop @gol
6877 -fsched-interblock -fsched-spec @gol
6878 -fschedule-insns -fschedule-insns2 @gol
6879 -fstrict-aliasing -fstrict-overflow @gol
6880 -ftree-builtin-call-dce @gol
6881 -ftree-switch-conversion -ftree-tail-merge @gol
6882 -fcode-hoisting @gol
6887 Please note the warning under @option{-fgcse} about
6888 invoking @option{-O2} on programs that use computed gotos.
6892 Optimize yet more. @option{-O3} turns on all optimizations specified
6893 by @option{-O2} and also turns on the @option{-finline-functions},
6894 @option{-funswitch-loops}, @option{-fpredictive-commoning},
6895 @option{-fgcse-after-reload}, @option{-ftree-loop-vectorize},
6896 @option{-ftree-loop-distribute-patterns}, @option{-fsplit-paths}
6897 @option{-ftree-slp-vectorize}, @option{-fvect-cost-model},
6898 @option{-ftree-partial-pre}, @option{-fpeel-loops}
6899 and @option{-fipa-cp-clone} options.
6903 Reduce compilation time and make debugging produce the expected
6904 results. This is the default.
6908 Optimize for size. @option{-Os} enables all @option{-O2} optimizations that
6909 do not typically increase code size. It also performs further
6910 optimizations designed to reduce code size.
6912 @option{-Os} disables the following optimization flags:
6913 @gccoptlist{-falign-functions -falign-jumps -falign-loops @gol
6914 -falign-labels -freorder-blocks -freorder-blocks-algorithm=stc @gol
6915 -freorder-blocks-and-partition -fprefetch-loop-arrays}
6919 Disregard strict standards compliance. @option{-Ofast} enables all
6920 @option{-O3} optimizations. It also enables optimizations that are not
6921 valid for all standard-compliant programs.
6922 It turns on @option{-ffast-math} and the Fortran-specific
6923 @option{-fno-protect-parens} and @option{-fstack-arrays}.
6927 Optimize debugging experience. @option{-Og} enables optimizations
6928 that do not interfere with debugging. It should be the optimization
6929 level of choice for the standard edit-compile-debug cycle, offering
6930 a reasonable level of optimization while maintaining fast compilation
6931 and a good debugging experience.
6934 If you use multiple @option{-O} options, with or without level numbers,
6935 the last such option is the one that is effective.
6937 Options of the form @option{-f@var{flag}} specify machine-independent
6938 flags. Most flags have both positive and negative forms; the negative
6939 form of @option{-ffoo} is @option{-fno-foo}. In the table
6940 below, only one of the forms is listed---the one you typically
6941 use. You can figure out the other form by either removing @samp{no-}
6944 The following options control specific optimizations. They are either
6945 activated by @option{-O} options or are related to ones that are. You
6946 can use the following flags in the rare cases when ``fine-tuning'' of
6947 optimizations to be performed is desired.
6950 @item -fno-defer-pop
6951 @opindex fno-defer-pop
6952 Always pop the arguments to each function call as soon as that function
6953 returns. For machines that must pop arguments after a function call,
6954 the compiler normally lets arguments accumulate on the stack for several
6955 function calls and pops them all at once.
6957 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
6959 @item -fforward-propagate
6960 @opindex fforward-propagate
6961 Perform a forward propagation pass on RTL@. The pass tries to combine two
6962 instructions and checks if the result can be simplified. If loop unrolling
6963 is active, two passes are performed and the second is scheduled after
6966 This option is enabled by default at optimization levels @option{-O},
6967 @option{-O2}, @option{-O3}, @option{-Os}.
6969 @item -ffp-contract=@var{style}
6970 @opindex ffp-contract
6971 @option{-ffp-contract=off} disables floating-point expression contraction.
6972 @option{-ffp-contract=fast} enables floating-point expression contraction
6973 such as forming of fused multiply-add operations if the target has
6974 native support for them.
6975 @option{-ffp-contract=on} enables floating-point expression contraction
6976 if allowed by the language standard. This is currently not implemented
6977 and treated equal to @option{-ffp-contract=off}.
6979 The default is @option{-ffp-contract=fast}.
6981 @item -fomit-frame-pointer
6982 @opindex fomit-frame-pointer
6983 Don't keep the frame pointer in a register for functions that
6984 don't need one. This avoids the instructions to save, set up and
6985 restore frame pointers; it also makes an extra register available
6986 in many functions. @strong{It also makes debugging impossible on
6989 On some machines, such as the VAX, this flag has no effect, because
6990 the standard calling sequence automatically handles the frame pointer
6991 and nothing is saved by pretending it doesn't exist. The
6992 machine-description macro @code{FRAME_POINTER_REQUIRED} controls
6993 whether a target machine supports this flag. @xref{Registers,,Register
6994 Usage, gccint, GNU Compiler Collection (GCC) Internals}.
6996 The default setting (when not optimizing for
6997 size) for 32-bit GNU/Linux x86 and 32-bit Darwin x86 targets is
6998 @option{-fomit-frame-pointer}. You can configure GCC with the
6999 @option{--enable-frame-pointer} configure option to change the default.
7001 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
7003 @item -foptimize-sibling-calls
7004 @opindex foptimize-sibling-calls
7005 Optimize sibling and tail recursive calls.
7007 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
7009 @item -foptimize-strlen
7010 @opindex foptimize-strlen
7011 Optimize various standard C string functions (e.g. @code{strlen},
7012 @code{strchr} or @code{strcpy}) and
7013 their @code{_FORTIFY_SOURCE} counterparts into faster alternatives.
7015 Enabled at levels @option{-O2}, @option{-O3}.
7019 Do not expand any functions inline apart from those marked with
7020 the @code{always_inline} attribute. This is the default when not
7023 Single functions can be exempted from inlining by marking them
7024 with the @code{noinline} attribute.
7026 @item -finline-small-functions
7027 @opindex finline-small-functions
7028 Integrate functions into their callers when their body is smaller than expected
7029 function call code (so overall size of program gets smaller). The compiler
7030 heuristically decides which functions are simple enough to be worth integrating
7031 in this way. This inlining applies to all functions, even those not declared
7034 Enabled at level @option{-O2}.
7036 @item -findirect-inlining
7037 @opindex findirect-inlining
7038 Inline also indirect calls that are discovered to be known at compile
7039 time thanks to previous inlining. This option has any effect only
7040 when inlining itself is turned on by the @option{-finline-functions}
7041 or @option{-finline-small-functions} options.
7043 Enabled at level @option{-O2}.
7045 @item -finline-functions
7046 @opindex finline-functions
7047 Consider all functions for inlining, even if they are not declared inline.
7048 The compiler heuristically decides which functions are worth integrating
7051 If all calls to a given function are integrated, and the function is
7052 declared @code{static}, then the function is normally not output as
7053 assembler code in its own right.
7055 Enabled at level @option{-O3}.
7057 @item -finline-functions-called-once
7058 @opindex finline-functions-called-once
7059 Consider all @code{static} functions called once for inlining into their
7060 caller even if they are not marked @code{inline}. If a call to a given
7061 function is integrated, then the function is not output as assembler code
7064 Enabled at levels @option{-O1}, @option{-O2}, @option{-O3} and @option{-Os}.
7066 @item -fearly-inlining
7067 @opindex fearly-inlining
7068 Inline functions marked by @code{always_inline} and functions whose body seems
7069 smaller than the function call overhead early before doing
7070 @option{-fprofile-generate} instrumentation and real inlining pass. Doing so
7071 makes profiling significantly cheaper and usually inlining faster on programs
7072 having large chains of nested wrapper functions.
7078 Perform interprocedural scalar replacement of aggregates, removal of
7079 unused parameters and replacement of parameters passed by reference
7080 by parameters passed by value.
7082 Enabled at levels @option{-O2}, @option{-O3} and @option{-Os}.
7084 @item -finline-limit=@var{n}
7085 @opindex finline-limit
7086 By default, GCC limits the size of functions that can be inlined. This flag
7087 allows coarse control of this limit. @var{n} is the size of functions that
7088 can be inlined in number of pseudo instructions.
7090 Inlining is actually controlled by a number of parameters, which may be
7091 specified individually by using @option{--param @var{name}=@var{value}}.
7092 The @option{-finline-limit=@var{n}} option sets some of these parameters
7096 @item max-inline-insns-single
7097 is set to @var{n}/2.
7098 @item max-inline-insns-auto
7099 is set to @var{n}/2.
7102 See below for a documentation of the individual
7103 parameters controlling inlining and for the defaults of these parameters.
7105 @emph{Note:} there may be no value to @option{-finline-limit} that results
7106 in default behavior.
7108 @emph{Note:} pseudo instruction represents, in this particular context, an
7109 abstract measurement of function's size. In no way does it represent a count
7110 of assembly instructions and as such its exact meaning might change from one
7111 release to an another.
7113 @item -fno-keep-inline-dllexport
7114 @opindex fno-keep-inline-dllexport
7115 This is a more fine-grained version of @option{-fkeep-inline-functions},
7116 which applies only to functions that are declared using the @code{dllexport}
7117 attribute or declspec (@xref{Function Attributes,,Declaring Attributes of
7120 @item -fkeep-inline-functions
7121 @opindex fkeep-inline-functions
7122 In C, emit @code{static} functions that are declared @code{inline}
7123 into the object file, even if the function has been inlined into all
7124 of its callers. This switch does not affect functions using the
7125 @code{extern inline} extension in GNU C90@. In C++, emit any and all
7126 inline functions into the object file.
7128 @item -fkeep-static-functions
7129 @opindex fkeep-static-functions
7130 Emit @code{static} functions into the object file, even if the function
7133 @item -fkeep-static-consts
7134 @opindex fkeep-static-consts
7135 Emit variables declared @code{static const} when optimization isn't turned
7136 on, even if the variables aren't referenced.
7138 GCC enables this option by default. If you want to force the compiler to
7139 check if a variable is referenced, regardless of whether or not
7140 optimization is turned on, use the @option{-fno-keep-static-consts} option.
7142 @item -fmerge-constants
7143 @opindex fmerge-constants
7144 Attempt to merge identical constants (string constants and floating-point
7145 constants) across compilation units.
7147 This option is the default for optimized compilation if the assembler and
7148 linker support it. Use @option{-fno-merge-constants} to inhibit this
7151 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
7153 @item -fmerge-all-constants
7154 @opindex fmerge-all-constants
7155 Attempt to merge identical constants and identical variables.
7157 This option implies @option{-fmerge-constants}. In addition to
7158 @option{-fmerge-constants} this considers e.g.@: even constant initialized
7159 arrays or initialized constant variables with integral or floating-point
7160 types. Languages like C or C++ require each variable, including multiple
7161 instances of the same variable in recursive calls, to have distinct locations,
7162 so using this option results in non-conforming
7165 @item -fmodulo-sched
7166 @opindex fmodulo-sched
7167 Perform swing modulo scheduling immediately before the first scheduling
7168 pass. This pass looks at innermost loops and reorders their
7169 instructions by overlapping different iterations.
7171 @item -fmodulo-sched-allow-regmoves
7172 @opindex fmodulo-sched-allow-regmoves
7173 Perform more aggressive SMS-based modulo scheduling with register moves
7174 allowed. By setting this flag certain anti-dependences edges are
7175 deleted, which triggers the generation of reg-moves based on the
7176 life-range analysis. This option is effective only with
7177 @option{-fmodulo-sched} enabled.
7179 @item -fno-branch-count-reg
7180 @opindex fno-branch-count-reg
7181 Avoid running a pass scanning for opportunities to use ``decrement and
7182 branch'' instructions on a count register instead of generating sequences
7183 of instructions that decrement a register, compare it against zero, and
7184 then branch based upon the result. This option is only meaningful on
7185 architectures that support such instructions, which include x86, PowerPC,
7186 IA-64 and S/390. Note that the @option{-fno-branch-count-reg} option
7187 doesn't remove the decrement and branch instructions from the generated
7188 instruction stream introduced by other optimization passes.
7190 Enabled by default at @option{-O1} and higher.
7192 The default is @option{-fbranch-count-reg}.
7194 @item -fno-function-cse
7195 @opindex fno-function-cse
7196 Do not put function addresses in registers; make each instruction that
7197 calls a constant function contain the function's address explicitly.
7199 This option results in less efficient code, but some strange hacks
7200 that alter the assembler output may be confused by the optimizations
7201 performed when this option is not used.
7203 The default is @option{-ffunction-cse}
7205 @item -fno-zero-initialized-in-bss
7206 @opindex fno-zero-initialized-in-bss
7207 If the target supports a BSS section, GCC by default puts variables that
7208 are initialized to zero into BSS@. This can save space in the resulting
7211 This option turns off this behavior because some programs explicitly
7212 rely on variables going to the data section---e.g., so that the
7213 resulting executable can find the beginning of that section and/or make
7214 assumptions based on that.
7216 The default is @option{-fzero-initialized-in-bss}.
7218 @item -fthread-jumps
7219 @opindex fthread-jumps
7220 Perform optimizations that check to see if a jump branches to a
7221 location where another comparison subsumed by the first is found. If
7222 so, the first branch is redirected to either the destination of the
7223 second branch or a point immediately following it, depending on whether
7224 the condition is known to be true or false.
7226 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
7228 @item -fsplit-wide-types
7229 @opindex fsplit-wide-types
7230 When using a type that occupies multiple registers, such as @code{long
7231 long} on a 32-bit system, split the registers apart and allocate them
7232 independently. This normally generates better code for those types,
7233 but may make debugging more difficult.
7235 Enabled at levels @option{-O}, @option{-O2}, @option{-O3},
7238 @item -fcse-follow-jumps
7239 @opindex fcse-follow-jumps
7240 In common subexpression elimination (CSE), scan through jump instructions
7241 when the target of the jump is not reached by any other path. For
7242 example, when CSE encounters an @code{if} statement with an
7243 @code{else} clause, CSE follows the jump when the condition
7246 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
7248 @item -fcse-skip-blocks
7249 @opindex fcse-skip-blocks
7250 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
7251 follow jumps that conditionally skip over blocks. When CSE
7252 encounters a simple @code{if} statement with no else clause,
7253 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
7254 body of the @code{if}.
7256 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
7258 @item -frerun-cse-after-loop
7259 @opindex frerun-cse-after-loop
7260 Re-run common subexpression elimination after loop optimizations are
7263 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
7267 Perform a global common subexpression elimination pass.
7268 This pass also performs global constant and copy propagation.
7270 @emph{Note:} When compiling a program using computed gotos, a GCC
7271 extension, you may get better run-time performance if you disable
7272 the global common subexpression elimination pass by adding
7273 @option{-fno-gcse} to the command line.
7275 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
7279 When @option{-fgcse-lm} is enabled, global common subexpression elimination
7280 attempts to move loads that are only killed by stores into themselves. This
7281 allows a loop containing a load/store sequence to be changed to a load outside
7282 the loop, and a copy/store within the loop.
7284 Enabled by default when @option{-fgcse} is enabled.
7288 When @option{-fgcse-sm} is enabled, a store motion pass is run after
7289 global common subexpression elimination. This pass attempts to move
7290 stores out of loops. When used in conjunction with @option{-fgcse-lm},
7291 loops containing a load/store sequence can be changed to a load before
7292 the loop and a store after the loop.
7294 Not enabled at any optimization level.
7298 When @option{-fgcse-las} is enabled, the global common subexpression
7299 elimination pass eliminates redundant loads that come after stores to the
7300 same memory location (both partial and full redundancies).
7302 Not enabled at any optimization level.
7304 @item -fgcse-after-reload
7305 @opindex fgcse-after-reload
7306 When @option{-fgcse-after-reload} is enabled, a redundant load elimination
7307 pass is performed after reload. The purpose of this pass is to clean up
7310 @item -faggressive-loop-optimizations
7311 @opindex faggressive-loop-optimizations
7312 This option tells the loop optimizer to use language constraints to
7313 derive bounds for the number of iterations of a loop. This assumes that
7314 loop code does not invoke undefined behavior by for example causing signed
7315 integer overflows or out-of-bound array accesses. The bounds for the
7316 number of iterations of a loop are used to guide loop unrolling and peeling
7317 and loop exit test optimizations.
7318 This option is enabled by default.
7320 @item -funconstrained-commons
7321 @opindex funconstrained-commons
7322 This option tells the compiler that variables declared in common blocks
7323 (e.g. Fortran) may later be overridden with longer trailing arrays. This
7324 prevents certain optimizations that depend on knowing the array bounds.
7326 @item -fcrossjumping
7327 @opindex fcrossjumping
7328 Perform cross-jumping transformation.
7329 This transformation unifies equivalent code and saves code size. The
7330 resulting code may or may not perform better than without cross-jumping.
7332 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
7334 @item -fauto-inc-dec
7335 @opindex fauto-inc-dec
7336 Combine increments or decrements of addresses with memory accesses.
7337 This pass is always skipped on architectures that do not have
7338 instructions to support this. Enabled by default at @option{-O} and
7339 higher on architectures that support this.
7343 Perform dead code elimination (DCE) on RTL@.
7344 Enabled by default at @option{-O} and higher.
7348 Perform dead store elimination (DSE) on RTL@.
7349 Enabled by default at @option{-O} and higher.
7351 @item -fif-conversion
7352 @opindex fif-conversion
7353 Attempt to transform conditional jumps into branch-less equivalents. This
7354 includes use of conditional moves, min, max, set flags and abs instructions, and
7355 some tricks doable by standard arithmetics. The use of conditional execution
7356 on chips where it is available is controlled by @option{-fif-conversion2}.
7358 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
7360 @item -fif-conversion2
7361 @opindex fif-conversion2
7362 Use conditional execution (where available) to transform conditional jumps into
7363 branch-less equivalents.
7365 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
7367 @item -fdeclone-ctor-dtor
7368 @opindex fdeclone-ctor-dtor
7369 The C++ ABI requires multiple entry points for constructors and
7370 destructors: one for a base subobject, one for a complete object, and
7371 one for a virtual destructor that calls operator delete afterwards.
7372 For a hierarchy with virtual bases, the base and complete variants are
7373 clones, which means two copies of the function. With this option, the
7374 base and complete variants are changed to be thunks that call a common
7377 Enabled by @option{-Os}.
7379 @item -fdelete-null-pointer-checks
7380 @opindex fdelete-null-pointer-checks
7381 Assume that programs cannot safely dereference null pointers, and that
7382 no code or data element resides at address zero.
7383 This option enables simple constant
7384 folding optimizations at all optimization levels. In addition, other
7385 optimization passes in GCC use this flag to control global dataflow
7386 analyses that eliminate useless checks for null pointers; these assume
7387 that a memory access to address zero always results in a trap, so
7388 that if a pointer is checked after it has already been dereferenced,
7391 Note however that in some environments this assumption is not true.
7392 Use @option{-fno-delete-null-pointer-checks} to disable this optimization
7393 for programs that depend on that behavior.
7395 This option is enabled by default on most targets. On Nios II ELF, it
7396 defaults to off. On AVR and CR16, this option is completely disabled.
7398 Passes that use the dataflow information
7399 are enabled independently at different optimization levels.
7401 @item -fdevirtualize
7402 @opindex fdevirtualize
7403 Attempt to convert calls to virtual functions to direct calls. This
7404 is done both within a procedure and interprocedurally as part of
7405 indirect inlining (@option{-findirect-inlining}) and interprocedural constant
7406 propagation (@option{-fipa-cp}).
7407 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
7409 @item -fdevirtualize-speculatively
7410 @opindex fdevirtualize-speculatively
7411 Attempt to convert calls to virtual functions to speculative direct calls.
7412 Based on the analysis of the type inheritance graph, determine for a given call
7413 the set of likely targets. If the set is small, preferably of size 1, change
7414 the call into a conditional deciding between direct and indirect calls. The
7415 speculative calls enable more optimizations, such as inlining. When they seem
7416 useless after further optimization, they are converted back into original form.
7418 @item -fdevirtualize-at-ltrans
7419 @opindex fdevirtualize-at-ltrans
7420 Stream extra information needed for aggressive devirtualization when running
7421 the link-time optimizer in local transformation mode.
7422 This option enables more devirtualization but
7423 significantly increases the size of streamed data. For this reason it is
7424 disabled by default.
7426 @item -fexpensive-optimizations
7427 @opindex fexpensive-optimizations
7428 Perform a number of minor optimizations that are relatively expensive.
7430 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
7434 Attempt to remove redundant extension instructions. This is especially
7435 helpful for the x86-64 architecture, which implicitly zero-extends in 64-bit
7436 registers after writing to their lower 32-bit half.
7438 Enabled for Alpha, AArch64 and x86 at levels @option{-O2},
7439 @option{-O3}, @option{-Os}.
7441 @item -fno-lifetime-dse
7442 @opindex fno-lifetime-dse
7443 In C++ the value of an object is only affected by changes within its
7444 lifetime: when the constructor begins, the object has an indeterminate
7445 value, and any changes during the lifetime of the object are dead when
7446 the object is destroyed. Normally dead store elimination will take
7447 advantage of this; if your code relies on the value of the object
7448 storage persisting beyond the lifetime of the object, you can use this
7449 flag to disable this optimization. To preserve stores before the
7450 constructor starts (e.g. because your operator new clears the object
7451 storage) but still treat the object as dead after the destructor you,
7452 can use @option{-flifetime-dse=1}. The default behavior can be
7453 explicitly selected with @option{-flifetime-dse=2}.
7454 @option{-flifetime-dse=0} is equivalent to @option{-fno-lifetime-dse}.
7456 @item -flive-range-shrinkage
7457 @opindex flive-range-shrinkage
7458 Attempt to decrease register pressure through register live range
7459 shrinkage. This is helpful for fast processors with small or moderate
7462 @item -fira-algorithm=@var{algorithm}
7463 @opindex fira-algorithm
7464 Use the specified coloring algorithm for the integrated register
7465 allocator. The @var{algorithm} argument can be @samp{priority}, which
7466 specifies Chow's priority coloring, or @samp{CB}, which specifies
7467 Chaitin-Briggs coloring. Chaitin-Briggs coloring is not implemented
7468 for all architectures, but for those targets that do support it, it is
7469 the default because it generates better code.
7471 @item -fira-region=@var{region}
7472 @opindex fira-region
7473 Use specified regions for the integrated register allocator. The
7474 @var{region} argument should be one of the following:
7479 Use all loops as register allocation regions.
7480 This can give the best results for machines with a small and/or
7481 irregular register set.
7484 Use all loops except for loops with small register pressure
7485 as the regions. This value usually gives
7486 the best results in most cases and for most architectures,
7487 and is enabled by default when compiling with optimization for speed
7488 (@option{-O}, @option{-O2}, @dots{}).
7491 Use all functions as a single region.
7492 This typically results in the smallest code size, and is enabled by default for
7493 @option{-Os} or @option{-O0}.
7497 @item -fira-hoist-pressure
7498 @opindex fira-hoist-pressure
7499 Use IRA to evaluate register pressure in the code hoisting pass for
7500 decisions to hoist expressions. This option usually results in smaller
7501 code, but it can slow the compiler down.
7503 This option is enabled at level @option{-Os} for all targets.
7505 @item -fira-loop-pressure
7506 @opindex fira-loop-pressure
7507 Use IRA to evaluate register pressure in loops for decisions to move
7508 loop invariants. This option usually results in generation
7509 of faster and smaller code on machines with large register files (>= 32
7510 registers), but it can slow the compiler down.
7512 This option is enabled at level @option{-O3} for some targets.
7514 @item -fno-ira-share-save-slots
7515 @opindex fno-ira-share-save-slots
7516 Disable sharing of stack slots used for saving call-used hard
7517 registers living through a call. Each hard register gets a
7518 separate stack slot, and as a result function stack frames are
7521 @item -fno-ira-share-spill-slots
7522 @opindex fno-ira-share-spill-slots
7523 Disable sharing of stack slots allocated for pseudo-registers. Each
7524 pseudo-register that does not get a hard register gets a separate
7525 stack slot, and as a result function stack frames are larger.
7529 Enable CFG-sensitive rematerialization in LRA. Instead of loading
7530 values of spilled pseudos, LRA tries to rematerialize (recalculate)
7531 values if it is profitable.
7533 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
7535 @item -fdelayed-branch
7536 @opindex fdelayed-branch
7537 If supported for the target machine, attempt to reorder instructions
7538 to exploit instruction slots available after delayed branch
7541 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
7543 @item -fschedule-insns
7544 @opindex fschedule-insns
7545 If supported for the target machine, attempt to reorder instructions to
7546 eliminate execution stalls due to required data being unavailable. This
7547 helps machines that have slow floating point or memory load instructions
7548 by allowing other instructions to be issued until the result of the load
7549 or floating-point instruction is required.
7551 Enabled at levels @option{-O2}, @option{-O3}.
7553 @item -fschedule-insns2
7554 @opindex fschedule-insns2
7555 Similar to @option{-fschedule-insns}, but requests an additional pass of
7556 instruction scheduling after register allocation has been done. This is
7557 especially useful on machines with a relatively small number of
7558 registers and where memory load instructions take more than one cycle.
7560 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
7562 @item -fno-sched-interblock
7563 @opindex fno-sched-interblock
7564 Don't schedule instructions across basic blocks. This is normally
7565 enabled by default when scheduling before register allocation, i.e.@:
7566 with @option{-fschedule-insns} or at @option{-O2} or higher.
7568 @item -fno-sched-spec
7569 @opindex fno-sched-spec
7570 Don't allow speculative motion of non-load instructions. This is normally
7571 enabled by default when scheduling before register allocation, i.e.@:
7572 with @option{-fschedule-insns} or at @option{-O2} or higher.
7574 @item -fsched-pressure
7575 @opindex fsched-pressure
7576 Enable register pressure sensitive insn scheduling before register
7577 allocation. This only makes sense when scheduling before register
7578 allocation is enabled, i.e.@: with @option{-fschedule-insns} or at
7579 @option{-O2} or higher. Usage of this option can improve the
7580 generated code and decrease its size by preventing register pressure
7581 increase above the number of available hard registers and subsequent
7582 spills in register allocation.
7584 @item -fsched-spec-load
7585 @opindex fsched-spec-load
7586 Allow speculative motion of some load instructions. This only makes
7587 sense when scheduling before register allocation, i.e.@: with
7588 @option{-fschedule-insns} or at @option{-O2} or higher.
7590 @item -fsched-spec-load-dangerous
7591 @opindex fsched-spec-load-dangerous
7592 Allow speculative motion of more load instructions. This only makes
7593 sense when scheduling before register allocation, i.e.@: with
7594 @option{-fschedule-insns} or at @option{-O2} or higher.
7596 @item -fsched-stalled-insns
7597 @itemx -fsched-stalled-insns=@var{n}
7598 @opindex fsched-stalled-insns
7599 Define how many insns (if any) can be moved prematurely from the queue
7600 of stalled insns into the ready list during the second scheduling pass.
7601 @option{-fno-sched-stalled-insns} means that no insns are moved
7602 prematurely, @option{-fsched-stalled-insns=0} means there is no limit
7603 on how many queued insns can be moved prematurely.
7604 @option{-fsched-stalled-insns} without a value is equivalent to
7605 @option{-fsched-stalled-insns=1}.
7607 @item -fsched-stalled-insns-dep
7608 @itemx -fsched-stalled-insns-dep=@var{n}
7609 @opindex fsched-stalled-insns-dep
7610 Define how many insn groups (cycles) are examined for a dependency
7611 on a stalled insn that is a candidate for premature removal from the queue
7612 of stalled insns. This has an effect only during the second scheduling pass,
7613 and only if @option{-fsched-stalled-insns} is used.
7614 @option{-fno-sched-stalled-insns-dep} is equivalent to
7615 @option{-fsched-stalled-insns-dep=0}.
7616 @option{-fsched-stalled-insns-dep} without a value is equivalent to
7617 @option{-fsched-stalled-insns-dep=1}.
7619 @item -fsched2-use-superblocks
7620 @opindex fsched2-use-superblocks
7621 When scheduling after register allocation, use superblock scheduling.
7622 This allows motion across basic block boundaries,
7623 resulting in faster schedules. This option is experimental, as not all machine
7624 descriptions used by GCC model the CPU closely enough to avoid unreliable
7625 results from the algorithm.
7627 This only makes sense when scheduling after register allocation, i.e.@: with
7628 @option{-fschedule-insns2} or at @option{-O2} or higher.
7630 @item -fsched-group-heuristic
7631 @opindex fsched-group-heuristic
7632 Enable the group heuristic in the scheduler. This heuristic favors
7633 the instruction that belongs to a schedule group. This is enabled
7634 by default when scheduling is enabled, i.e.@: with @option{-fschedule-insns}
7635 or @option{-fschedule-insns2} or at @option{-O2} or higher.
7637 @item -fsched-critical-path-heuristic
7638 @opindex fsched-critical-path-heuristic
7639 Enable the critical-path heuristic in the scheduler. This heuristic favors
7640 instructions on the critical path. This is enabled by default when
7641 scheduling is enabled, i.e.@: with @option{-fschedule-insns}
7642 or @option{-fschedule-insns2} or at @option{-O2} or higher.
7644 @item -fsched-spec-insn-heuristic
7645 @opindex fsched-spec-insn-heuristic
7646 Enable the speculative instruction heuristic in the scheduler. This
7647 heuristic favors speculative instructions with greater dependency weakness.
7648 This is enabled by default when scheduling is enabled, i.e.@:
7649 with @option{-fschedule-insns} or @option{-fschedule-insns2}
7650 or at @option{-O2} or higher.
7652 @item -fsched-rank-heuristic
7653 @opindex fsched-rank-heuristic
7654 Enable the rank heuristic in the scheduler. This heuristic favors
7655 the instruction belonging to a basic block with greater size or frequency.
7656 This is enabled by default when scheduling is enabled, i.e.@:
7657 with @option{-fschedule-insns} or @option{-fschedule-insns2} or
7658 at @option{-O2} or higher.
7660 @item -fsched-last-insn-heuristic
7661 @opindex fsched-last-insn-heuristic
7662 Enable the last-instruction heuristic in the scheduler. This heuristic
7663 favors the instruction that is less dependent on the last instruction
7664 scheduled. This is enabled by default when scheduling is enabled,
7665 i.e.@: with @option{-fschedule-insns} or @option{-fschedule-insns2} or
7666 at @option{-O2} or higher.
7668 @item -fsched-dep-count-heuristic
7669 @opindex fsched-dep-count-heuristic
7670 Enable the dependent-count heuristic in the scheduler. This heuristic
7671 favors the instruction that has more instructions depending on it.
7672 This is enabled by default when scheduling is enabled, i.e.@:
7673 with @option{-fschedule-insns} or @option{-fschedule-insns2} or
7674 at @option{-O2} or higher.
7676 @item -freschedule-modulo-scheduled-loops
7677 @opindex freschedule-modulo-scheduled-loops
7678 Modulo scheduling is performed before traditional scheduling. If a loop
7679 is modulo scheduled, later scheduling passes may change its schedule.
7680 Use this option to control that behavior.
7682 @item -fselective-scheduling
7683 @opindex fselective-scheduling
7684 Schedule instructions using selective scheduling algorithm. Selective
7685 scheduling runs instead of the first scheduler pass.
7687 @item -fselective-scheduling2
7688 @opindex fselective-scheduling2
7689 Schedule instructions using selective scheduling algorithm. Selective
7690 scheduling runs instead of the second scheduler pass.
7692 @item -fsel-sched-pipelining
7693 @opindex fsel-sched-pipelining
7694 Enable software pipelining of innermost loops during selective scheduling.
7695 This option has no effect unless one of @option{-fselective-scheduling} or
7696 @option{-fselective-scheduling2} is turned on.
7698 @item -fsel-sched-pipelining-outer-loops
7699 @opindex fsel-sched-pipelining-outer-loops
7700 When pipelining loops during selective scheduling, also pipeline outer loops.
7701 This option has no effect unless @option{-fsel-sched-pipelining} is turned on.
7703 @item -fsemantic-interposition
7704 @opindex fsemantic-interposition
7705 Some object formats, like ELF, allow interposing of symbols by the
7707 This means that for symbols exported from the DSO, the compiler cannot perform
7708 interprocedural propagation, inlining and other optimizations in anticipation
7709 that the function or variable in question may change. While this feature is
7710 useful, for example, to rewrite memory allocation functions by a debugging
7711 implementation, it is expensive in the terms of code quality.
7712 With @option{-fno-semantic-interposition} the compiler assumes that
7713 if interposition happens for functions the overwriting function will have
7714 precisely the same semantics (and side effects).
7715 Similarly if interposition happens
7716 for variables, the constructor of the variable will be the same. The flag
7717 has no effect for functions explicitly declared inline
7718 (where it is never allowed for interposition to change semantics)
7719 and for symbols explicitly declared weak.
7722 @opindex fshrink-wrap
7723 Emit function prologues only before parts of the function that need it,
7724 rather than at the top of the function. This flag is enabled by default at
7725 @option{-O} and higher.
7727 @item -fshrink-wrap-separate
7728 @opindex fshrink-wrap-separate
7729 Shrink-wrap separate parts of the prologue and epilogue separately, so that
7730 those parts are only executed when needed.
7731 This option is on by default, but has no effect unless @option{-fshrink-wrap}
7732 is also turned on and the target supports this.
7734 @item -fcaller-saves
7735 @opindex fcaller-saves
7736 Enable allocation of values to registers that are clobbered by
7737 function calls, by emitting extra instructions to save and restore the
7738 registers around such calls. Such allocation is done only when it
7739 seems to result in better code.
7741 This option is always enabled by default on certain machines, usually
7742 those which have no call-preserved registers to use instead.
7744 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
7746 @item -fcombine-stack-adjustments
7747 @opindex fcombine-stack-adjustments
7748 Tracks stack adjustments (pushes and pops) and stack memory references
7749 and then tries to find ways to combine them.
7751 Enabled by default at @option{-O1} and higher.
7755 Use caller save registers for allocation if those registers are not used by
7756 any called function. In that case it is not necessary to save and restore
7757 them around calls. This is only possible if called functions are part of
7758 same compilation unit as current function and they are compiled before it.
7760 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}, however the option
7761 is disabled if generated code will be instrumented for profiling
7762 (@option{-p}, or @option{-pg}) or if callee's register usage cannot be known
7763 exactly (this happens on targets that do not expose prologues
7764 and epilogues in RTL).
7766 @item -fconserve-stack
7767 @opindex fconserve-stack
7768 Attempt to minimize stack usage. The compiler attempts to use less
7769 stack space, even if that makes the program slower. This option
7770 implies setting the @option{large-stack-frame} parameter to 100
7771 and the @option{large-stack-frame-growth} parameter to 400.
7773 @item -ftree-reassoc
7774 @opindex ftree-reassoc
7775 Perform reassociation on trees. This flag is enabled by default
7776 at @option{-O} and higher.
7778 @item -fcode-hoisting
7779 @opindex fcode-hoisting
7780 Perform code hoisting. Code hoisting tries to move the
7781 evaluation of expressions executed on all paths to the function exit
7782 as early as possible. This is especially useful as a code size
7783 optimization, but it often helps for code speed as well.
7784 This flag is enabled by default at @option{-O2} and higher.
7788 Perform partial redundancy elimination (PRE) on trees. This flag is
7789 enabled by default at @option{-O2} and @option{-O3}.
7791 @item -ftree-partial-pre
7792 @opindex ftree-partial-pre
7793 Make partial redundancy elimination (PRE) more aggressive. This flag is
7794 enabled by default at @option{-O3}.
7796 @item -ftree-forwprop
7797 @opindex ftree-forwprop
7798 Perform forward propagation on trees. This flag is enabled by default
7799 at @option{-O} and higher.
7803 Perform full redundancy elimination (FRE) on trees. The difference
7804 between FRE and PRE is that FRE only considers expressions
7805 that are computed on all paths leading to the redundant computation.
7806 This analysis is faster than PRE, though it exposes fewer redundancies.
7807 This flag is enabled by default at @option{-O} and higher.
7809 @item -ftree-phiprop
7810 @opindex ftree-phiprop
7811 Perform hoisting of loads from conditional pointers on trees. This
7812 pass is enabled by default at @option{-O} and higher.
7814 @item -fhoist-adjacent-loads
7815 @opindex fhoist-adjacent-loads
7816 Speculatively hoist loads from both branches of an if-then-else if the
7817 loads are from adjacent locations in the same structure and the target
7818 architecture has a conditional move instruction. This flag is enabled
7819 by default at @option{-O2} and higher.
7821 @item -ftree-copy-prop
7822 @opindex ftree-copy-prop
7823 Perform copy propagation on trees. This pass eliminates unnecessary
7824 copy operations. This flag is enabled by default at @option{-O} and
7827 @item -fipa-pure-const
7828 @opindex fipa-pure-const
7829 Discover which functions are pure or constant.
7830 Enabled by default at @option{-O} and higher.
7832 @item -fipa-reference
7833 @opindex fipa-reference
7834 Discover which static variables do not escape the
7836 Enabled by default at @option{-O} and higher.
7840 Perform interprocedural pointer analysis and interprocedural modification
7841 and reference analysis. This option can cause excessive memory and
7842 compile-time usage on large compilation units. It is not enabled by
7843 default at any optimization level.
7846 @opindex fipa-profile
7847 Perform interprocedural profile propagation. The functions called only from
7848 cold functions are marked as cold. Also functions executed once (such as
7849 @code{cold}, @code{noreturn}, static constructors or destructors) are identified. Cold
7850 functions and loop less parts of functions executed once are then optimized for
7852 Enabled by default at @option{-O} and higher.
7856 Perform interprocedural constant propagation.
7857 This optimization analyzes the program to determine when values passed
7858 to functions are constants and then optimizes accordingly.
7859 This optimization can substantially increase performance
7860 if the application has constants passed to functions.
7861 This flag is enabled by default at @option{-O2}, @option{-Os} and @option{-O3}.
7863 @item -fipa-cp-clone
7864 @opindex fipa-cp-clone
7865 Perform function cloning to make interprocedural constant propagation stronger.
7866 When enabled, interprocedural constant propagation performs function cloning
7867 when externally visible function can be called with constant arguments.
7868 Because this optimization can create multiple copies of functions,
7869 it may significantly increase code size
7870 (see @option{--param ipcp-unit-growth=@var{value}}).
7871 This flag is enabled by default at @option{-O3}.
7873 @item -fipa-cp-alignment
7874 @opindex -fipa-cp-alignment
7875 When enabled, this optimization propagates alignment of function
7876 parameters to support better vectorization and string operations.
7878 This flag is enabled by default at @option{-O2} and @option{-Os}. It
7879 requires that @option{-fipa-cp} is enabled.
7880 @option{-fipa-cp-alignment} is obsolete, use @option{-fipa-bit-cp} instead.
7883 @opindex -fipa-bit-cp
7884 When enabled, perform ipa bitwise constant propagation. This flag is
7885 enabled by default at @option{-O2}. It requires that @option{-fipa-cp}
7890 Perform Identical Code Folding for functions and read-only variables.
7891 The optimization reduces code size and may disturb unwind stacks by replacing
7892 a function by equivalent one with a different name. The optimization works
7893 more effectively with link time optimization enabled.
7895 Nevertheless the behavior is similar to Gold Linker ICF optimization, GCC ICF
7896 works on different levels and thus the optimizations are not same - there are
7897 equivalences that are found only by GCC and equivalences found only by Gold.
7899 This flag is enabled by default at @option{-O2} and @option{-Os}.
7901 @item -fisolate-erroneous-paths-dereference
7902 @opindex fisolate-erroneous-paths-dereference
7903 Detect paths that trigger erroneous or undefined behavior due to
7904 dereferencing a null pointer. Isolate those paths from the main control
7905 flow and turn the statement with erroneous or undefined behavior into a trap.
7906 This flag is enabled by default at @option{-O2} and higher and depends on
7907 @option{-fdelete-null-pointer-checks} also being enabled.
7909 @item -fisolate-erroneous-paths-attribute
7910 @opindex fisolate-erroneous-paths-attribute
7911 Detect paths that trigger erroneous or undefined behavior due a null value
7912 being used in a way forbidden by a @code{returns_nonnull} or @code{nonnull}
7913 attribute. Isolate those paths from the main control flow and turn the
7914 statement with erroneous or undefined behavior into a trap. This is not
7915 currently enabled, but may be enabled by @option{-O2} in the future.
7919 Perform forward store motion on trees. This flag is
7920 enabled by default at @option{-O} and higher.
7922 @item -ftree-bit-ccp
7923 @opindex ftree-bit-ccp
7924 Perform sparse conditional bit constant propagation on trees and propagate
7925 pointer alignment information.
7926 This pass only operates on local scalar variables and is enabled by default
7927 at @option{-O} and higher. It requires that @option{-ftree-ccp} is enabled.
7931 Perform sparse conditional constant propagation (CCP) on trees. This
7932 pass only operates on local scalar variables and is enabled by default
7933 at @option{-O} and higher.
7935 @item -fssa-backprop
7936 @opindex fssa-backprop
7937 Propagate information about uses of a value up the definition chain
7938 in order to simplify the definitions. For example, this pass strips
7939 sign operations if the sign of a value never matters. The flag is
7940 enabled by default at @option{-O} and higher.
7943 @opindex fssa-phiopt
7944 Perform pattern matching on SSA PHI nodes to optimize conditional
7945 code. This pass is enabled by default at @option{-O} and higher.
7947 @item -ftree-switch-conversion
7948 @opindex ftree-switch-conversion
7949 Perform conversion of simple initializations in a switch to
7950 initializations from a scalar array. This flag is enabled by default
7951 at @option{-O2} and higher.
7953 @item -ftree-tail-merge
7954 @opindex ftree-tail-merge
7955 Look for identical code sequences. When found, replace one with a jump to the
7956 other. This optimization is known as tail merging or cross jumping. This flag
7957 is enabled by default at @option{-O2} and higher. The compilation time
7959 be limited using @option{max-tail-merge-comparisons} parameter and
7960 @option{max-tail-merge-iterations} parameter.
7964 Perform dead code elimination (DCE) on trees. This flag is enabled by
7965 default at @option{-O} and higher.
7967 @item -ftree-builtin-call-dce
7968 @opindex ftree-builtin-call-dce
7969 Perform conditional dead code elimination (DCE) for calls to built-in functions
7970 that may set @code{errno} but are otherwise side-effect free. This flag is
7971 enabled by default at @option{-O2} and higher if @option{-Os} is not also
7974 @item -ftree-dominator-opts
7975 @opindex ftree-dominator-opts
7976 Perform a variety of simple scalar cleanups (constant/copy
7977 propagation, redundancy elimination, range propagation and expression
7978 simplification) based on a dominator tree traversal. This also
7979 performs jump threading (to reduce jumps to jumps). This flag is
7980 enabled by default at @option{-O} and higher.
7984 Perform dead store elimination (DSE) on trees. A dead store is a store into
7985 a memory location that is later overwritten by another store without
7986 any intervening loads. In this case the earlier store can be deleted. This
7987 flag is enabled by default at @option{-O} and higher.
7991 Perform loop header copying on trees. This is beneficial since it increases
7992 effectiveness of code motion optimizations. It also saves one jump. This flag
7993 is enabled by default at @option{-O} and higher. It is not enabled
7994 for @option{-Os}, since it usually increases code size.
7996 @item -ftree-loop-optimize
7997 @opindex ftree-loop-optimize
7998 Perform loop optimizations on trees. This flag is enabled by default
7999 at @option{-O} and higher.
8001 @item -ftree-loop-linear
8002 @itemx -floop-interchange
8003 @itemx -floop-strip-mine
8005 @itemx -floop-unroll-and-jam
8006 @opindex ftree-loop-linear
8007 @opindex floop-interchange
8008 @opindex floop-strip-mine
8009 @opindex floop-block
8010 @opindex floop-unroll-and-jam
8011 Perform loop nest optimizations. Same as
8012 @option{-floop-nest-optimize}. To use this code transformation, GCC has
8013 to be configured with @option{--with-isl} to enable the Graphite loop
8014 transformation infrastructure.
8016 @item -fgraphite-identity
8017 @opindex fgraphite-identity
8018 Enable the identity transformation for graphite. For every SCoP we generate
8019 the polyhedral representation and transform it back to gimple. Using
8020 @option{-fgraphite-identity} we can check the costs or benefits of the
8021 GIMPLE -> GRAPHITE -> GIMPLE transformation. Some minimal optimizations
8022 are also performed by the code generator isl, like index splitting and
8023 dead code elimination in loops.
8025 @item -floop-nest-optimize
8026 @opindex floop-nest-optimize
8027 Enable the isl based loop nest optimizer. This is a generic loop nest
8028 optimizer based on the Pluto optimization algorithms. It calculates a loop
8029 structure optimized for data-locality and parallelism. This option
8032 @item -floop-parallelize-all
8033 @opindex floop-parallelize-all
8034 Use the Graphite data dependence analysis to identify loops that can
8035 be parallelized. Parallelize all the loops that can be analyzed to
8036 not contain loop carried dependences without checking that it is
8037 profitable to parallelize the loops.
8039 @item -ftree-coalesce-vars
8040 @opindex ftree-coalesce-vars
8041 While transforming the program out of the SSA representation, attempt to
8042 reduce copying by coalescing versions of different user-defined
8043 variables, instead of just compiler temporaries. This may severely
8044 limit the ability to debug an optimized program compiled with
8045 @option{-fno-var-tracking-assignments}. In the negated form, this flag
8046 prevents SSA coalescing of user variables. This option is enabled by
8047 default if optimization is enabled, and it does very little otherwise.
8049 @item -ftree-loop-if-convert
8050 @opindex ftree-loop-if-convert
8051 Attempt to transform conditional jumps in the innermost loops to
8052 branch-less equivalents. The intent is to remove control-flow from
8053 the innermost loops in order to improve the ability of the
8054 vectorization pass to handle these loops. This is enabled by default
8055 if vectorization is enabled.
8057 @item -ftree-loop-if-convert-stores
8058 @opindex ftree-loop-if-convert-stores
8059 Attempt to also if-convert conditional jumps containing memory writes.
8060 This transformation can be unsafe for multi-threaded programs as it
8061 transforms conditional memory writes into unconditional memory writes.
8064 for (i = 0; i < N; i++)
8070 for (i = 0; i < N; i++)
8071 A[i] = cond ? expr : A[i];
8073 potentially producing data races.
8075 @item -ftree-loop-distribution
8076 @opindex ftree-loop-distribution
8077 Perform loop distribution. This flag can improve cache performance on
8078 big loop bodies and allow further loop optimizations, like
8079 parallelization or vectorization, to take place. For example, the loop
8096 @item -ftree-loop-distribute-patterns
8097 @opindex ftree-loop-distribute-patterns
8098 Perform loop distribution of patterns that can be code generated with
8099 calls to a library. This flag is enabled by default at @option{-O3}.
8101 This pass distributes the initialization loops and generates a call to
8102 memset zero. For example, the loop
8118 and the initialization loop is transformed into a call to memset zero.
8120 @item -ftree-loop-im
8121 @opindex ftree-loop-im
8122 Perform loop invariant motion on trees. This pass moves only invariants that
8123 are hard to handle at RTL level (function calls, operations that expand to
8124 nontrivial sequences of insns). With @option{-funswitch-loops} it also moves
8125 operands of conditions that are invariant out of the loop, so that we can use
8126 just trivial invariantness analysis in loop unswitching. The pass also includes
8129 @item -ftree-loop-ivcanon
8130 @opindex ftree-loop-ivcanon
8131 Create a canonical counter for number of iterations in loops for which
8132 determining number of iterations requires complicated analysis. Later
8133 optimizations then may determine the number easily. Useful especially
8134 in connection with unrolling.
8138 Perform induction variable optimizations (strength reduction, induction
8139 variable merging and induction variable elimination) on trees.
8141 @item -ftree-parallelize-loops=n
8142 @opindex ftree-parallelize-loops
8143 Parallelize loops, i.e., split their iteration space to run in n threads.
8144 This is only possible for loops whose iterations are independent
8145 and can be arbitrarily reordered. The optimization is only
8146 profitable on multiprocessor machines, for loops that are CPU-intensive,
8147 rather than constrained e.g.@: by memory bandwidth. This option
8148 implies @option{-pthread}, and thus is only supported on targets
8149 that have support for @option{-pthread}.
8153 Perform function-local points-to analysis on trees. This flag is
8154 enabled by default at @option{-O} and higher.
8158 Perform scalar replacement of aggregates. This pass replaces structure
8159 references with scalars to prevent committing structures to memory too
8160 early. This flag is enabled by default at @option{-O} and higher.
8162 @item -fstore-merging
8163 @opindex fstore-merging
8164 Perform merging of narrow stores to consecutive memory addresses. This pass
8165 merges contiguous stores of immediate values narrower than a word into fewer
8166 wider stores to reduce the number of instructions. This is enabled by default
8167 at @option{-O} and higher.
8171 Perform temporary expression replacement during the SSA->normal phase. Single
8172 use/single def temporaries are replaced at their use location with their
8173 defining expression. This results in non-GIMPLE code, but gives the expanders
8174 much more complex trees to work on resulting in better RTL generation. This is
8175 enabled by default at @option{-O} and higher.
8179 Perform straight-line strength reduction on trees. This recognizes related
8180 expressions involving multiplications and replaces them by less expensive
8181 calculations when possible. This is enabled by default at @option{-O} and
8184 @item -ftree-vectorize
8185 @opindex ftree-vectorize
8186 Perform vectorization on trees. This flag enables @option{-ftree-loop-vectorize}
8187 and @option{-ftree-slp-vectorize} if not explicitly specified.
8189 @item -ftree-loop-vectorize
8190 @opindex ftree-loop-vectorize
8191 Perform loop vectorization on trees. This flag is enabled by default at
8192 @option{-O3} and when @option{-ftree-vectorize} is enabled.
8194 @item -ftree-slp-vectorize
8195 @opindex ftree-slp-vectorize
8196 Perform basic block vectorization on trees. This flag is enabled by default at
8197 @option{-O3} and when @option{-ftree-vectorize} is enabled.
8199 @item -fvect-cost-model=@var{model}
8200 @opindex fvect-cost-model
8201 Alter the cost model used for vectorization. The @var{model} argument
8202 should be one of @samp{unlimited}, @samp{dynamic} or @samp{cheap}.
8203 With the @samp{unlimited} model the vectorized code-path is assumed
8204 to be profitable while with the @samp{dynamic} model a runtime check
8205 guards the vectorized code-path to enable it only for iteration
8206 counts that will likely execute faster than when executing the original
8207 scalar loop. The @samp{cheap} model disables vectorization of
8208 loops where doing so would be cost prohibitive for example due to
8209 required runtime checks for data dependence or alignment but otherwise
8210 is equal to the @samp{dynamic} model.
8211 The default cost model depends on other optimization flags and is
8212 either @samp{dynamic} or @samp{cheap}.
8214 @item -fsimd-cost-model=@var{model}
8215 @opindex fsimd-cost-model
8216 Alter the cost model used for vectorization of loops marked with the OpenMP
8217 or Cilk Plus simd directive. The @var{model} argument should be one of
8218 @samp{unlimited}, @samp{dynamic}, @samp{cheap}. All values of @var{model}
8219 have the same meaning as described in @option{-fvect-cost-model} and by
8220 default a cost model defined with @option{-fvect-cost-model} is used.
8224 Perform Value Range Propagation on trees. This is similar to the
8225 constant propagation pass, but instead of values, ranges of values are
8226 propagated. This allows the optimizers to remove unnecessary range
8227 checks like array bound checks and null pointer checks. This is
8228 enabled by default at @option{-O2} and higher. Null pointer check
8229 elimination is only done if @option{-fdelete-null-pointer-checks} is
8233 @opindex fsplit-paths
8234 Split paths leading to loop backedges. This can improve dead code
8235 elimination and common subexpression elimination. This is enabled by
8236 default at @option{-O2} and above.
8238 @item -fsplit-ivs-in-unroller
8239 @opindex fsplit-ivs-in-unroller
8240 Enables expression of values of induction variables in later iterations
8241 of the unrolled loop using the value in the first iteration. This breaks
8242 long dependency chains, thus improving efficiency of the scheduling passes.
8244 A combination of @option{-fweb} and CSE is often sufficient to obtain the
8245 same effect. However, that is not reliable in cases where the loop body
8246 is more complicated than a single basic block. It also does not work at all
8247 on some architectures due to restrictions in the CSE pass.
8249 This optimization is enabled by default.
8251 @item -fvariable-expansion-in-unroller
8252 @opindex fvariable-expansion-in-unroller
8253 With this option, the compiler creates multiple copies of some
8254 local variables when unrolling a loop, which can result in superior code.
8256 @item -fpartial-inlining
8257 @opindex fpartial-inlining
8258 Inline parts of functions. This option has any effect only
8259 when inlining itself is turned on by the @option{-finline-functions}
8260 or @option{-finline-small-functions} options.
8262 Enabled at level @option{-O2}.
8264 @item -fpredictive-commoning
8265 @opindex fpredictive-commoning
8266 Perform predictive commoning optimization, i.e., reusing computations
8267 (especially memory loads and stores) performed in previous
8268 iterations of loops.
8270 This option is enabled at level @option{-O3}.
8272 @item -fprefetch-loop-arrays
8273 @opindex fprefetch-loop-arrays
8274 If supported by the target machine, generate instructions to prefetch
8275 memory to improve the performance of loops that access large arrays.
8277 This option may generate better or worse code; results are highly
8278 dependent on the structure of loops within the source code.
8280 Disabled at level @option{-Os}.
8282 @item -fprintf-return-value
8283 @opindex fprintf-return-value
8284 Substitute constants for known return value of formatted output functions
8285 such as @code{sprintf}, @code{snprintf}, @code{vsprintf}, and @code{vsnprintf}
8286 (but not @code{printf} of @code{fprintf}). This transformation allows GCC
8287 to optimize or even eliminate branches based on the known return value of
8288 these functions called with arguments that are either constant, or whose
8289 values are known to be in a range that makes determining the exact return
8290 value possible. For example, both the branch and the body of the @code{if}
8291 statement (but not the call to @code{snprint}) can be optimized away when
8292 @code{i} is a 32-bit or smaller integer because the return value is guaranteed
8297 if (snprintf (buf, "%08x", i) >= sizeof buf)
8301 The @option{-fprintf-return-value} option relies on other optimizations
8302 and yields best results with @option{-O2}. It works in tandem with the
8303 @option{-Wformat-length} option. The @option{-fprintf-return-value}
8304 option is disabled by default.
8307 @itemx -fno-peephole2
8308 @opindex fno-peephole
8309 @opindex fno-peephole2
8310 Disable any machine-specific peephole optimizations. The difference
8311 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
8312 are implemented in the compiler; some targets use one, some use the
8313 other, a few use both.
8315 @option{-fpeephole} is enabled by default.
8316 @option{-fpeephole2} enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
8318 @item -fno-guess-branch-probability
8319 @opindex fno-guess-branch-probability
8320 Do not guess branch probabilities using heuristics.
8322 GCC uses heuristics to guess branch probabilities if they are
8323 not provided by profiling feedback (@option{-fprofile-arcs}). These
8324 heuristics are based on the control flow graph. If some branch probabilities
8325 are specified by @code{__builtin_expect}, then the heuristics are
8326 used to guess branch probabilities for the rest of the control flow graph,
8327 taking the @code{__builtin_expect} info into account. The interactions
8328 between the heuristics and @code{__builtin_expect} can be complex, and in
8329 some cases, it may be useful to disable the heuristics so that the effects
8330 of @code{__builtin_expect} are easier to understand.
8332 The default is @option{-fguess-branch-probability} at levels
8333 @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
8335 @item -freorder-blocks
8336 @opindex freorder-blocks
8337 Reorder basic blocks in the compiled function in order to reduce number of
8338 taken branches and improve code locality.
8340 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
8342 @item -freorder-blocks-algorithm=@var{algorithm}
8343 @opindex freorder-blocks-algorithm
8344 Use the specified algorithm for basic block reordering. The
8345 @var{algorithm} argument can be @samp{simple}, which does not increase
8346 code size (except sometimes due to secondary effects like alignment),
8347 or @samp{stc}, the ``software trace cache'' algorithm, which tries to
8348 put all often executed code together, minimizing the number of branches
8349 executed by making extra copies of code.
8351 The default is @samp{simple} at levels @option{-O}, @option{-Os}, and
8352 @samp{stc} at levels @option{-O2}, @option{-O3}.
8354 @item -freorder-blocks-and-partition
8355 @opindex freorder-blocks-and-partition
8356 In addition to reordering basic blocks in the compiled function, in order
8357 to reduce number of taken branches, partitions hot and cold basic blocks
8358 into separate sections of the assembly and @file{.o} files, to improve
8359 paging and cache locality performance.
8361 This optimization is automatically turned off in the presence of
8362 exception handling, for linkonce sections, for functions with a user-defined
8363 section attribute and on any architecture that does not support named
8366 Enabled for x86 at levels @option{-O2}, @option{-O3}.
8368 @item -freorder-functions
8369 @opindex freorder-functions
8370 Reorder functions in the object file in order to
8371 improve code locality. This is implemented by using special
8372 subsections @code{.text.hot} for most frequently executed functions and
8373 @code{.text.unlikely} for unlikely executed functions. Reordering is done by
8374 the linker so object file format must support named sections and linker must
8375 place them in a reasonable way.
8377 Also profile feedback must be available to make this option effective. See
8378 @option{-fprofile-arcs} for details.
8380 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
8382 @item -fstrict-aliasing
8383 @opindex fstrict-aliasing
8384 Allow the compiler to assume the strictest aliasing rules applicable to
8385 the language being compiled. For C (and C++), this activates
8386 optimizations based on the type of expressions. In particular, an
8387 object of one type is assumed never to reside at the same address as an
8388 object of a different type, unless the types are almost the same. For
8389 example, an @code{unsigned int} can alias an @code{int}, but not a
8390 @code{void*} or a @code{double}. A character type may alias any other
8393 @anchor{Type-punning}Pay special attention to code like this:
8406 The practice of reading from a different union member than the one most
8407 recently written to (called ``type-punning'') is common. Even with
8408 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
8409 is accessed through the union type. So, the code above works as
8410 expected. @xref{Structures unions enumerations and bit-fields
8411 implementation}. However, this code might not:
8422 Similarly, access by taking the address, casting the resulting pointer
8423 and dereferencing the result has undefined behavior, even if the cast
8424 uses a union type, e.g.:
8428 return ((union a_union *) &d)->i;
8432 The @option{-fstrict-aliasing} option is enabled at levels
8433 @option{-O2}, @option{-O3}, @option{-Os}.
8435 @item -fstrict-overflow
8436 @opindex fstrict-overflow
8437 Allow the compiler to assume strict signed overflow rules, depending
8438 on the language being compiled. For C (and C++) this means that
8439 overflow when doing arithmetic with signed numbers is undefined, which
8440 means that the compiler may assume that it does not happen. This
8441 permits various optimizations. For example, the compiler assumes
8442 that an expression like @code{i + 10 > i} is always true for
8443 signed @code{i}. This assumption is only valid if signed overflow is
8444 undefined, as the expression is false if @code{i + 10} overflows when
8445 using twos complement arithmetic. When this option is in effect any
8446 attempt to determine whether an operation on signed numbers
8447 overflows must be written carefully to not actually involve overflow.
8449 This option also allows the compiler to assume strict pointer
8450 semantics: given a pointer to an object, if adding an offset to that
8451 pointer does not produce a pointer to the same object, the addition is
8452 undefined. This permits the compiler to conclude that @code{p + u >
8453 p} is always true for a pointer @code{p} and unsigned integer
8454 @code{u}. This assumption is only valid because pointer wraparound is
8455 undefined, as the expression is false if @code{p + u} overflows using
8456 twos complement arithmetic.
8458 See also the @option{-fwrapv} option. Using @option{-fwrapv} means
8459 that integer signed overflow is fully defined: it wraps. When
8460 @option{-fwrapv} is used, there is no difference between
8461 @option{-fstrict-overflow} and @option{-fno-strict-overflow} for
8462 integers. With @option{-fwrapv} certain types of overflow are
8463 permitted. For example, if the compiler gets an overflow when doing
8464 arithmetic on constants, the overflowed value can still be used with
8465 @option{-fwrapv}, but not otherwise.
8467 The @option{-fstrict-overflow} option is enabled at levels
8468 @option{-O2}, @option{-O3}, @option{-Os}.
8470 @item -falign-functions
8471 @itemx -falign-functions=@var{n}
8472 @opindex falign-functions
8473 Align the start of functions to the next power-of-two greater than
8474 @var{n}, skipping up to @var{n} bytes. For instance,
8475 @option{-falign-functions=32} aligns functions to the next 32-byte
8476 boundary, but @option{-falign-functions=24} aligns to the next
8477 32-byte boundary only if this can be done by skipping 23 bytes or less.
8479 @option{-fno-align-functions} and @option{-falign-functions=1} are
8480 equivalent and mean that functions are not aligned.
8482 Some assemblers only support this flag when @var{n} is a power of two;
8483 in that case, it is rounded up.
8485 If @var{n} is not specified or is zero, use a machine-dependent default.
8487 Enabled at levels @option{-O2}, @option{-O3}.
8489 @item -falign-labels
8490 @itemx -falign-labels=@var{n}
8491 @opindex falign-labels
8492 Align all branch targets to a power-of-two boundary, skipping up to
8493 @var{n} bytes like @option{-falign-functions}. This option can easily
8494 make code slower, because it must insert dummy operations for when the
8495 branch target is reached in the usual flow of the code.
8497 @option{-fno-align-labels} and @option{-falign-labels=1} are
8498 equivalent and mean that labels are not aligned.
8500 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
8501 are greater than this value, then their values are used instead.
8503 If @var{n} is not specified or is zero, use a machine-dependent default
8504 which is very likely to be @samp{1}, meaning no alignment.
8506 Enabled at levels @option{-O2}, @option{-O3}.
8509 @itemx -falign-loops=@var{n}
8510 @opindex falign-loops
8511 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
8512 like @option{-falign-functions}. If the loops are
8513 executed many times, this makes up for any execution of the dummy
8516 @option{-fno-align-loops} and @option{-falign-loops=1} are
8517 equivalent and mean that loops are not aligned.
8519 If @var{n} is not specified or is zero, use a machine-dependent default.
8521 Enabled at levels @option{-O2}, @option{-O3}.
8524 @itemx -falign-jumps=@var{n}
8525 @opindex falign-jumps
8526 Align branch targets to a power-of-two boundary, for branch targets
8527 where the targets can only be reached by jumping, skipping up to @var{n}
8528 bytes like @option{-falign-functions}. In this case, no dummy operations
8531 @option{-fno-align-jumps} and @option{-falign-jumps=1} are
8532 equivalent and mean that loops are not aligned.
8534 If @var{n} is not specified or is zero, use a machine-dependent default.
8536 Enabled at levels @option{-O2}, @option{-O3}.
8538 @item -funit-at-a-time
8539 @opindex funit-at-a-time
8540 This option is left for compatibility reasons. @option{-funit-at-a-time}
8541 has no effect, while @option{-fno-unit-at-a-time} implies
8542 @option{-fno-toplevel-reorder} and @option{-fno-section-anchors}.
8546 @item -fno-toplevel-reorder
8547 @opindex fno-toplevel-reorder
8548 Do not reorder top-level functions, variables, and @code{asm}
8549 statements. Output them in the same order that they appear in the
8550 input file. When this option is used, unreferenced static variables
8551 are not removed. This option is intended to support existing code
8552 that relies on a particular ordering. For new code, it is better to
8553 use attributes when possible.
8555 Enabled at level @option{-O0}. When disabled explicitly, it also implies
8556 @option{-fno-section-anchors}, which is otherwise enabled at @option{-O0} on some
8561 Constructs webs as commonly used for register allocation purposes and assign
8562 each web individual pseudo register. This allows the register allocation pass
8563 to operate on pseudos directly, but also strengthens several other optimization
8564 passes, such as CSE, loop optimizer and trivial dead code remover. It can,
8565 however, make debugging impossible, since variables no longer stay in a
8568 Enabled by default with @option{-funroll-loops}.
8570 @item -fwhole-program
8571 @opindex fwhole-program
8572 Assume that the current compilation unit represents the whole program being
8573 compiled. All public functions and variables with the exception of @code{main}
8574 and those merged by attribute @code{externally_visible} become static functions
8575 and in effect are optimized more aggressively by interprocedural optimizers.
8577 This option should not be used in combination with @option{-flto}.
8578 Instead relying on a linker plugin should provide safer and more precise
8581 @item -flto[=@var{n}]
8583 This option runs the standard link-time optimizer. When invoked
8584 with source code, it generates GIMPLE (one of GCC's internal
8585 representations) and writes it to special ELF sections in the object
8586 file. When the object files are linked together, all the function
8587 bodies are read from these ELF sections and instantiated as if they
8588 had been part of the same translation unit.
8590 To use the link-time optimizer, @option{-flto} and optimization
8591 options should be specified at compile time and during the final link.
8592 It is recommended that you compile all the files participating in the
8593 same link with the same options and also specify those options at
8598 gcc -c -O2 -flto foo.c
8599 gcc -c -O2 -flto bar.c
8600 gcc -o myprog -flto -O2 foo.o bar.o
8603 The first two invocations to GCC save a bytecode representation
8604 of GIMPLE into special ELF sections inside @file{foo.o} and
8605 @file{bar.o}. The final invocation reads the GIMPLE bytecode from
8606 @file{foo.o} and @file{bar.o}, merges the two files into a single
8607 internal image, and compiles the result as usual. Since both
8608 @file{foo.o} and @file{bar.o} are merged into a single image, this
8609 causes all the interprocedural analyses and optimizations in GCC to
8610 work across the two files as if they were a single one. This means,
8611 for example, that the inliner is able to inline functions in
8612 @file{bar.o} into functions in @file{foo.o} and vice-versa.
8614 Another (simpler) way to enable link-time optimization is:
8617 gcc -o myprog -flto -O2 foo.c bar.c
8620 The above generates bytecode for @file{foo.c} and @file{bar.c},
8621 merges them together into a single GIMPLE representation and optimizes
8622 them as usual to produce @file{myprog}.
8624 The only important thing to keep in mind is that to enable link-time
8625 optimizations you need to use the GCC driver to perform the link step.
8626 GCC then automatically performs link-time optimization if any of the
8627 objects involved were compiled with the @option{-flto} command-line option.
8629 should specify the optimization options to be used for link-time
8630 optimization though GCC tries to be clever at guessing an
8631 optimization level to use from the options used at compile time
8632 if you fail to specify one at link time. You can always override
8633 the automatic decision to do link-time optimization at link time
8634 by passing @option{-fno-lto} to the link command.
8636 To make whole program optimization effective, it is necessary to make
8637 certain whole program assumptions. The compiler needs to know
8638 what functions and variables can be accessed by libraries and runtime
8639 outside of the link-time optimized unit. When supported by the linker,
8640 the linker plugin (see @option{-fuse-linker-plugin}) passes information
8641 to the compiler about used and externally visible symbols. When
8642 the linker plugin is not available, @option{-fwhole-program} should be
8643 used to allow the compiler to make these assumptions, which leads
8644 to more aggressive optimization decisions.
8646 When @option{-fuse-linker-plugin} is not enabled, when a file is
8647 compiled with @option{-flto}, the generated object file is larger than
8648 a regular object file because it contains GIMPLE bytecodes and the usual
8649 final code (see @option{-ffat-lto-objects}. This means that
8650 object files with LTO information can be linked as normal object
8651 files; if @option{-fno-lto} is passed to the linker, no
8652 interprocedural optimizations are applied. Note that when
8653 @option{-fno-fat-lto-objects} is enabled the compile stage is faster
8654 but you cannot perform a regular, non-LTO link on them.
8656 Additionally, the optimization flags used to compile individual files
8657 are not necessarily related to those used at link time. For instance,
8660 gcc -c -O0 -ffat-lto-objects -flto foo.c
8661 gcc -c -O0 -ffat-lto-objects -flto bar.c
8662 gcc -o myprog -O3 foo.o bar.o
8665 This produces individual object files with unoptimized assembler
8666 code, but the resulting binary @file{myprog} is optimized at
8667 @option{-O3}. If, instead, the final binary is generated with
8668 @option{-fno-lto}, then @file{myprog} is not optimized.
8670 When producing the final binary, GCC only
8671 applies link-time optimizations to those files that contain bytecode.
8672 Therefore, you can mix and match object files and libraries with
8673 GIMPLE bytecodes and final object code. GCC automatically selects
8674 which files to optimize in LTO mode and which files to link without
8677 There are some code generation flags preserved by GCC when
8678 generating bytecodes, as they need to be used during the final link
8679 stage. Generally options specified at link time override those
8680 specified at compile time.
8682 If you do not specify an optimization level option @option{-O} at
8683 link time, then GCC uses the highest optimization level
8684 used when compiling the object files.
8686 Currently, the following options and their settings are taken from
8687 the first object file that explicitly specifies them:
8688 @option{-fPIC}, @option{-fpic}, @option{-fpie}, @option{-fcommon},
8689 @option{-fexceptions}, @option{-fnon-call-exceptions}, @option{-fgnu-tm}
8690 and all the @option{-m} target flags.
8692 Certain ABI-changing flags are required to match in all compilation units,
8693 and trying to override this at link time with a conflicting value
8694 is ignored. This includes options such as @option{-freg-struct-return}
8695 and @option{-fpcc-struct-return}.
8697 Other options such as @option{-ffp-contract}, @option{-fno-strict-overflow},
8698 @option{-fwrapv}, @option{-fno-trapv} or @option{-fno-strict-aliasing}
8699 are passed through to the link stage and merged conservatively for
8700 conflicting translation units. Specifically
8701 @option{-fno-strict-overflow}, @option{-fwrapv} and @option{-fno-trapv} take
8702 precedence; and for example @option{-ffp-contract=off} takes precedence
8703 over @option{-ffp-contract=fast}. You can override them at link time.
8705 If LTO encounters objects with C linkage declared with incompatible
8706 types in separate translation units to be linked together (undefined
8707 behavior according to ISO C99 6.2.7), a non-fatal diagnostic may be
8708 issued. The behavior is still undefined at run time. Similar
8709 diagnostics may be raised for other languages.
8711 Another feature of LTO is that it is possible to apply interprocedural
8712 optimizations on files written in different languages:
8717 gfortran -c -flto baz.f90
8718 g++ -o myprog -flto -O3 foo.o bar.o baz.o -lgfortran
8721 Notice that the final link is done with @command{g++} to get the C++
8722 runtime libraries and @option{-lgfortran} is added to get the Fortran
8723 runtime libraries. In general, when mixing languages in LTO mode, you
8724 should use the same link command options as when mixing languages in a
8725 regular (non-LTO) compilation.
8727 If object files containing GIMPLE bytecode are stored in a library archive, say
8728 @file{libfoo.a}, it is possible to extract and use them in an LTO link if you
8729 are using a linker with plugin support. To create static libraries suitable
8730 for LTO, use @command{gcc-ar} and @command{gcc-ranlib} instead of @command{ar}
8731 and @command{ranlib};
8732 to show the symbols of object files with GIMPLE bytecode, use
8733 @command{gcc-nm}. Those commands require that @command{ar}, @command{ranlib}
8734 and @command{nm} have been compiled with plugin support. At link time, use the the
8735 flag @option{-fuse-linker-plugin} to ensure that the library participates in
8736 the LTO optimization process:
8739 gcc -o myprog -O2 -flto -fuse-linker-plugin a.o b.o -lfoo
8742 With the linker plugin enabled, the linker extracts the needed
8743 GIMPLE files from @file{libfoo.a} and passes them on to the running GCC
8744 to make them part of the aggregated GIMPLE image to be optimized.
8746 If you are not using a linker with plugin support and/or do not
8747 enable the linker plugin, then the objects inside @file{libfoo.a}
8748 are extracted and linked as usual, but they do not participate
8749 in the LTO optimization process. In order to make a static library suitable
8750 for both LTO optimization and usual linkage, compile its object files with
8751 @option{-flto} @option{-ffat-lto-objects}.
8753 Link-time optimizations do not require the presence of the whole program to
8754 operate. If the program does not require any symbols to be exported, it is
8755 possible to combine @option{-flto} and @option{-fwhole-program} to allow
8756 the interprocedural optimizers to use more aggressive assumptions which may
8757 lead to improved optimization opportunities.
8758 Use of @option{-fwhole-program} is not needed when linker plugin is
8759 active (see @option{-fuse-linker-plugin}).
8761 The current implementation of LTO makes no
8762 attempt to generate bytecode that is portable between different
8763 types of hosts. The bytecode files are versioned and there is a
8764 strict version check, so bytecode files generated in one version of
8765 GCC do not work with an older or newer version of GCC.
8767 Link-time optimization does not work well with generation of debugging
8768 information. Combining @option{-flto} with
8769 @option{-g} is currently experimental and expected to produce unexpected
8772 If you specify the optional @var{n}, the optimization and code
8773 generation done at link time is executed in parallel using @var{n}
8774 parallel jobs by utilizing an installed @command{make} program. The
8775 environment variable @env{MAKE} may be used to override the program
8776 used. The default value for @var{n} is 1.
8778 You can also specify @option{-flto=jobserver} to use GNU make's
8779 job server mode to determine the number of parallel jobs. This
8780 is useful when the Makefile calling GCC is already executing in parallel.
8781 You must prepend a @samp{+} to the command recipe in the parent Makefile
8782 for this to work. This option likely only works if @env{MAKE} is
8785 @item -flto-partition=@var{alg}
8786 @opindex flto-partition
8787 Specify the partitioning algorithm used by the link-time optimizer.
8788 The value is either @samp{1to1} to specify a partitioning mirroring
8789 the original source files or @samp{balanced} to specify partitioning
8790 into equally sized chunks (whenever possible) or @samp{max} to create
8791 new partition for every symbol where possible. Specifying @samp{none}
8792 as an algorithm disables partitioning and streaming completely.
8793 The default value is @samp{balanced}. While @samp{1to1} can be used
8794 as an workaround for various code ordering issues, the @samp{max}
8795 partitioning is intended for internal testing only.
8796 The value @samp{one} specifies that exactly one partition should be
8797 used while the value @samp{none} bypasses partitioning and executes
8798 the link-time optimization step directly from the WPA phase.
8800 @item -flto-odr-type-merging
8801 @opindex flto-odr-type-merging
8802 Enable streaming of mangled types names of C++ types and their unification
8803 at link time. This increases size of LTO object files, but enables
8804 diagnostics about One Definition Rule violations.
8806 @item -flto-compression-level=@var{n}
8807 @opindex flto-compression-level
8808 This option specifies the level of compression used for intermediate
8809 language written to LTO object files, and is only meaningful in
8810 conjunction with LTO mode (@option{-flto}). Valid
8811 values are 0 (no compression) to 9 (maximum compression). Values
8812 outside this range are clamped to either 0 or 9. If the option is not
8813 given, a default balanced compression setting is used.
8815 @item -fuse-linker-plugin
8816 @opindex fuse-linker-plugin
8817 Enables the use of a linker plugin during link-time optimization. This
8818 option relies on plugin support in the linker, which is available in gold
8819 or in GNU ld 2.21 or newer.
8821 This option enables the extraction of object files with GIMPLE bytecode out
8822 of library archives. This improves the quality of optimization by exposing
8823 more code to the link-time optimizer. This information specifies what
8824 symbols can be accessed externally (by non-LTO object or during dynamic
8825 linking). Resulting code quality improvements on binaries (and shared
8826 libraries that use hidden visibility) are similar to @option{-fwhole-program}.
8827 See @option{-flto} for a description of the effect of this flag and how to
8830 This option is enabled by default when LTO support in GCC is enabled
8831 and GCC was configured for use with
8832 a linker supporting plugins (GNU ld 2.21 or newer or gold).
8834 @item -ffat-lto-objects
8835 @opindex ffat-lto-objects
8836 Fat LTO objects are object files that contain both the intermediate language
8837 and the object code. This makes them usable for both LTO linking and normal
8838 linking. This option is effective only when compiling with @option{-flto}
8839 and is ignored at link time.
8841 @option{-fno-fat-lto-objects} improves compilation time over plain LTO, but
8842 requires the complete toolchain to be aware of LTO. It requires a linker with
8843 linker plugin support for basic functionality. Additionally,
8844 @command{nm}, @command{ar} and @command{ranlib}
8845 need to support linker plugins to allow a full-featured build environment
8846 (capable of building static libraries etc). GCC provides the @command{gcc-ar},
8847 @command{gcc-nm}, @command{gcc-ranlib} wrappers to pass the right options
8848 to these tools. With non fat LTO makefiles need to be modified to use them.
8850 The default is @option{-fno-fat-lto-objects} on targets with linker plugin
8853 @item -fcompare-elim
8854 @opindex fcompare-elim
8855 After register allocation and post-register allocation instruction splitting,
8856 identify arithmetic instructions that compute processor flags similar to a
8857 comparison operation based on that arithmetic. If possible, eliminate the
8858 explicit comparison operation.
8860 This pass only applies to certain targets that cannot explicitly represent
8861 the comparison operation before register allocation is complete.
8863 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
8865 @item -fcprop-registers
8866 @opindex fcprop-registers
8867 After register allocation and post-register allocation instruction splitting,
8868 perform a copy-propagation pass to try to reduce scheduling dependencies
8869 and occasionally eliminate the copy.
8871 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
8873 @item -fprofile-correction
8874 @opindex fprofile-correction
8875 Profiles collected using an instrumented binary for multi-threaded programs may
8876 be inconsistent due to missed counter updates. When this option is specified,
8877 GCC uses heuristics to correct or smooth out such inconsistencies. By
8878 default, GCC emits an error message when an inconsistent profile is detected.
8881 @itemx -fprofile-use=@var{path}
8882 @opindex fprofile-use
8883 Enable profile feedback-directed optimizations,
8884 and the following optimizations
8885 which are generally profitable only with profile feedback available:
8886 @option{-fbranch-probabilities}, @option{-fvpt},
8887 @option{-funroll-loops}, @option{-fpeel-loops}, @option{-ftracer},
8888 @option{-ftree-vectorize}, and @option{ftree-loop-distribute-patterns}.
8890 Before you can use this option, you must first generate profiling information.
8891 @xref{Optimize Options}, for information about the @option{-fprofile-generate}
8894 By default, GCC emits an error message if the feedback profiles do not
8895 match the source code. This error can be turned into a warning by using
8896 @option{-Wcoverage-mismatch}. Note this may result in poorly optimized
8899 If @var{path} is specified, GCC looks at the @var{path} to find
8900 the profile feedback data files. See @option{-fprofile-dir}.
8902 @item -fauto-profile
8903 @itemx -fauto-profile=@var{path}
8904 @opindex fauto-profile
8905 Enable sampling-based feedback-directed optimizations,
8906 and the following optimizations
8907 which are generally profitable only with profile feedback available:
8908 @option{-fbranch-probabilities}, @option{-fvpt},
8909 @option{-funroll-loops}, @option{-fpeel-loops}, @option{-ftracer},
8910 @option{-ftree-vectorize},
8911 @option{-finline-functions}, @option{-fipa-cp}, @option{-fipa-cp-clone},
8912 @option{-fpredictive-commoning}, @option{-funswitch-loops},
8913 @option{-fgcse-after-reload}, and @option{-ftree-loop-distribute-patterns}.
8915 @var{path} is the name of a file containing AutoFDO profile information.
8916 If omitted, it defaults to @file{fbdata.afdo} in the current directory.
8918 Producing an AutoFDO profile data file requires running your program
8919 with the @command{perf} utility on a supported GNU/Linux target system.
8920 For more information, see @uref{https://perf.wiki.kernel.org/}.
8924 perf record -e br_inst_retired:near_taken -b -o perf.data \
8928 Then use the @command{create_gcov} tool to convert the raw profile data
8929 to a format that can be used by GCC.@ You must also supply the
8930 unstripped binary for your program to this tool.
8931 See @uref{https://github.com/google/autofdo}.
8935 create_gcov --binary=your_program.unstripped --profile=perf.data \
8940 The following options control compiler behavior regarding floating-point
8941 arithmetic. These options trade off between speed and
8942 correctness. All must be specifically enabled.
8946 @opindex ffloat-store
8947 Do not store floating-point variables in registers, and inhibit other
8948 options that might change whether a floating-point value is taken from a
8951 @cindex floating-point precision
8952 This option prevents undesirable excess precision on machines such as
8953 the 68000 where the floating registers (of the 68881) keep more
8954 precision than a @code{double} is supposed to have. Similarly for the
8955 x86 architecture. For most programs, the excess precision does only
8956 good, but a few programs rely on the precise definition of IEEE floating
8957 point. Use @option{-ffloat-store} for such programs, after modifying
8958 them to store all pertinent intermediate computations into variables.
8960 @item -fexcess-precision=@var{style}
8961 @opindex fexcess-precision
8962 This option allows further control over excess precision on machines
8963 where floating-point registers have more precision than the IEEE
8964 @code{float} and @code{double} types and the processor does not
8965 support operations rounding to those types. By default,
8966 @option{-fexcess-precision=fast} is in effect; this means that
8967 operations are carried out in the precision of the registers and that
8968 it is unpredictable when rounding to the types specified in the source
8969 code takes place. When compiling C, if
8970 @option{-fexcess-precision=standard} is specified then excess
8971 precision follows the rules specified in ISO C99; in particular,
8972 both casts and assignments cause values to be rounded to their
8973 semantic types (whereas @option{-ffloat-store} only affects
8974 assignments). This option is enabled by default for C if a strict
8975 conformance option such as @option{-std=c99} is used.
8978 @option{-fexcess-precision=standard} is not implemented for languages
8979 other than C, and has no effect if
8980 @option{-funsafe-math-optimizations} or @option{-ffast-math} is
8981 specified. On the x86, it also has no effect if @option{-mfpmath=sse}
8982 or @option{-mfpmath=sse+387} is specified; in the former case, IEEE
8983 semantics apply without excess precision, and in the latter, rounding
8988 Sets the options @option{-fno-math-errno}, @option{-funsafe-math-optimizations},
8989 @option{-ffinite-math-only}, @option{-fno-rounding-math},
8990 @option{-fno-signaling-nans} and @option{-fcx-limited-range}.
8992 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
8994 This option is not turned on by any @option{-O} option besides
8995 @option{-Ofast} since it can result in incorrect output for programs
8996 that depend on an exact implementation of IEEE or ISO rules/specifications
8997 for math functions. It may, however, yield faster code for programs
8998 that do not require the guarantees of these specifications.
9000 @item -fno-math-errno
9001 @opindex fno-math-errno
9002 Do not set @code{errno} after calling math functions that are executed
9003 with a single instruction, e.g., @code{sqrt}. A program that relies on
9004 IEEE exceptions for math error handling may want to use this flag
9005 for speed while maintaining IEEE arithmetic compatibility.
9007 This option is not turned on by any @option{-O} option since
9008 it can result in incorrect output for programs that depend on
9009 an exact implementation of IEEE or ISO rules/specifications for
9010 math functions. It may, however, yield faster code for programs
9011 that do not require the guarantees of these specifications.
9013 The default is @option{-fmath-errno}.
9015 On Darwin systems, the math library never sets @code{errno}. There is
9016 therefore no reason for the compiler to consider the possibility that
9017 it might, and @option{-fno-math-errno} is the default.
9019 @item -funsafe-math-optimizations
9020 @opindex funsafe-math-optimizations
9022 Allow optimizations for floating-point arithmetic that (a) assume
9023 that arguments and results are valid and (b) may violate IEEE or
9024 ANSI standards. When used at link time, it may include libraries
9025 or startup files that change the default FPU control word or other
9026 similar optimizations.
9028 This option is not turned on by any @option{-O} option since
9029 it can result in incorrect output for programs that depend on
9030 an exact implementation of IEEE or ISO rules/specifications for
9031 math functions. It may, however, yield faster code for programs
9032 that do not require the guarantees of these specifications.
9033 Enables @option{-fno-signed-zeros}, @option{-fno-trapping-math},
9034 @option{-fassociative-math} and @option{-freciprocal-math}.
9036 The default is @option{-fno-unsafe-math-optimizations}.
9038 @item -fassociative-math
9039 @opindex fassociative-math
9041 Allow re-association of operands in series of floating-point operations.
9042 This violates the ISO C and C++ language standard by possibly changing
9043 computation result. NOTE: re-ordering may change the sign of zero as
9044 well as ignore NaNs and inhibit or create underflow or overflow (and
9045 thus cannot be used on code that relies on rounding behavior like
9046 @code{(x + 2**52) - 2**52}. May also reorder floating-point comparisons
9047 and thus may not be used when ordered comparisons are required.
9048 This option requires that both @option{-fno-signed-zeros} and
9049 @option{-fno-trapping-math} be in effect. Moreover, it doesn't make
9050 much sense with @option{-frounding-math}. For Fortran the option
9051 is automatically enabled when both @option{-fno-signed-zeros} and
9052 @option{-fno-trapping-math} are in effect.
9054 The default is @option{-fno-associative-math}.
9056 @item -freciprocal-math
9057 @opindex freciprocal-math
9059 Allow the reciprocal of a value to be used instead of dividing by
9060 the value if this enables optimizations. For example @code{x / y}
9061 can be replaced with @code{x * (1/y)}, which is useful if @code{(1/y)}
9062 is subject to common subexpression elimination. Note that this loses
9063 precision and increases the number of flops operating on the value.
9065 The default is @option{-fno-reciprocal-math}.
9067 @item -ffinite-math-only
9068 @opindex ffinite-math-only
9069 Allow optimizations for floating-point arithmetic that assume
9070 that arguments and results are not NaNs or +-Infs.
9072 This option is not turned on by any @option{-O} option since
9073 it can result in incorrect output for programs that depend on
9074 an exact implementation of IEEE or ISO rules/specifications for
9075 math functions. It may, however, yield faster code for programs
9076 that do not require the guarantees of these specifications.
9078 The default is @option{-fno-finite-math-only}.
9080 @item -fno-signed-zeros
9081 @opindex fno-signed-zeros
9082 Allow optimizations for floating-point arithmetic that ignore the
9083 signedness of zero. IEEE arithmetic specifies the behavior of
9084 distinct +0.0 and @minus{}0.0 values, which then prohibits simplification
9085 of expressions such as x+0.0 or 0.0*x (even with @option{-ffinite-math-only}).
9086 This option implies that the sign of a zero result isn't significant.
9088 The default is @option{-fsigned-zeros}.
9090 @item -fno-trapping-math
9091 @opindex fno-trapping-math
9092 Compile code assuming that floating-point operations cannot generate
9093 user-visible traps. These traps include division by zero, overflow,
9094 underflow, inexact result and invalid operation. This option requires
9095 that @option{-fno-signaling-nans} be in effect. Setting this option may
9096 allow faster code if one relies on ``non-stop'' IEEE arithmetic, for example.
9098 This option should never be turned on by any @option{-O} option since
9099 it can result in incorrect output for programs that depend on
9100 an exact implementation of IEEE or ISO rules/specifications for
9103 The default is @option{-ftrapping-math}.
9105 @item -frounding-math
9106 @opindex frounding-math
9107 Disable transformations and optimizations that assume default floating-point
9108 rounding behavior. This is round-to-zero for all floating point
9109 to integer conversions, and round-to-nearest for all other arithmetic
9110 truncations. This option should be specified for programs that change
9111 the FP rounding mode dynamically, or that may be executed with a
9112 non-default rounding mode. This option disables constant folding of
9113 floating-point expressions at compile time (which may be affected by
9114 rounding mode) and arithmetic transformations that are unsafe in the
9115 presence of sign-dependent rounding modes.
9117 The default is @option{-fno-rounding-math}.
9119 This option is experimental and does not currently guarantee to
9120 disable all GCC optimizations that are affected by rounding mode.
9121 Future versions of GCC may provide finer control of this setting
9122 using C99's @code{FENV_ACCESS} pragma. This command-line option
9123 will be used to specify the default state for @code{FENV_ACCESS}.
9125 @item -fsignaling-nans
9126 @opindex fsignaling-nans
9127 Compile code assuming that IEEE signaling NaNs may generate user-visible
9128 traps during floating-point operations. Setting this option disables
9129 optimizations that may change the number of exceptions visible with
9130 signaling NaNs. This option implies @option{-ftrapping-math}.
9132 This option causes the preprocessor macro @code{__SUPPORT_SNAN__} to
9135 The default is @option{-fno-signaling-nans}.
9137 This option is experimental and does not currently guarantee to
9138 disable all GCC optimizations that affect signaling NaN behavior.
9140 @item -fno-fp-int-builtin-inexact
9141 @opindex fno-fp-int-builtin-inexact
9142 Do not allow the built-in functions @code{ceil}, @code{floor},
9143 @code{round} and @code{trunc}, and their @code{float} and @code{long
9144 double} variants, to generate code that raises the ``inexact''
9145 floating-point exception for noninteger arguments. ISO C99 and C11
9146 allow these functions to raise the ``inexact'' exception, but ISO/IEC
9147 TS 18661-1:2014, the C bindings to IEEE 754-2008, does not allow these
9150 The default is @option{-ffp-int-builtin-inexact}, allowing the
9151 exception to be raised. This option does nothing unless
9152 @option{-ftrapping-math} is in effect.
9154 Even if @option{-fno-fp-int-builtin-inexact} is used, if the functions
9155 generate a call to a library function then the ``inexact'' exception
9156 may be raised if the library implementation does not follow TS 18661.
9158 @item -fsingle-precision-constant
9159 @opindex fsingle-precision-constant
9160 Treat floating-point constants as single precision instead of
9161 implicitly converting them to double-precision constants.
9163 @item -fcx-limited-range
9164 @opindex fcx-limited-range
9165 When enabled, this option states that a range reduction step is not
9166 needed when performing complex division. Also, there is no checking
9167 whether the result of a complex multiplication or division is @code{NaN
9168 + I*NaN}, with an attempt to rescue the situation in that case. The
9169 default is @option{-fno-cx-limited-range}, but is enabled by
9170 @option{-ffast-math}.
9172 This option controls the default setting of the ISO C99
9173 @code{CX_LIMITED_RANGE} pragma. Nevertheless, the option applies to
9176 @item -fcx-fortran-rules
9177 @opindex fcx-fortran-rules
9178 Complex multiplication and division follow Fortran rules. Range
9179 reduction is done as part of complex division, but there is no checking
9180 whether the result of a complex multiplication or division is @code{NaN
9181 + I*NaN}, with an attempt to rescue the situation in that case.
9183 The default is @option{-fno-cx-fortran-rules}.
9187 The following options control optimizations that may improve
9188 performance, but are not enabled by any @option{-O} options. This
9189 section includes experimental options that may produce broken code.
9192 @item -fbranch-probabilities
9193 @opindex fbranch-probabilities
9194 After running a program compiled with @option{-fprofile-arcs}
9195 (@pxref{Instrumentation Options}),
9196 you can compile it a second time using
9197 @option{-fbranch-probabilities}, to improve optimizations based on
9198 the number of times each branch was taken. When a program
9199 compiled with @option{-fprofile-arcs} exits, it saves arc execution
9200 counts to a file called @file{@var{sourcename}.gcda} for each source
9201 file. The information in this data file is very dependent on the
9202 structure of the generated code, so you must use the same source code
9203 and the same optimization options for both compilations.
9205 With @option{-fbranch-probabilities}, GCC puts a
9206 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
9207 These can be used to improve optimization. Currently, they are only
9208 used in one place: in @file{reorg.c}, instead of guessing which path a
9209 branch is most likely to take, the @samp{REG_BR_PROB} values are used to
9210 exactly determine which path is taken more often.
9212 @item -fprofile-values
9213 @opindex fprofile-values
9214 If combined with @option{-fprofile-arcs}, it adds code so that some
9215 data about values of expressions in the program is gathered.
9217 With @option{-fbranch-probabilities}, it reads back the data gathered
9218 from profiling values of expressions for usage in optimizations.
9220 Enabled with @option{-fprofile-generate} and @option{-fprofile-use}.
9222 @item -fprofile-reorder-functions
9223 @opindex fprofile-reorder-functions
9224 Function reordering based on profile instrumentation collects
9225 first time of execution of a function and orders these functions
9228 Enabled with @option{-fprofile-use}.
9232 If combined with @option{-fprofile-arcs}, this option instructs the compiler
9233 to add code to gather information about values of expressions.
9235 With @option{-fbranch-probabilities}, it reads back the data gathered
9236 and actually performs the optimizations based on them.
9237 Currently the optimizations include specialization of division operations
9238 using the knowledge about the value of the denominator.
9240 @item -frename-registers
9241 @opindex frename-registers
9242 Attempt to avoid false dependencies in scheduled code by making use
9243 of registers left over after register allocation. This optimization
9244 most benefits processors with lots of registers. Depending on the
9245 debug information format adopted by the target, however, it can
9246 make debugging impossible, since variables no longer stay in
9247 a ``home register''.
9249 Enabled by default with @option{-funroll-loops}.
9251 @item -fschedule-fusion
9252 @opindex fschedule-fusion
9253 Performs a target dependent pass over the instruction stream to schedule
9254 instructions of same type together because target machine can execute them
9255 more efficiently if they are adjacent to each other in the instruction flow.
9257 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
9261 Perform tail duplication to enlarge superblock size. This transformation
9262 simplifies the control flow of the function allowing other optimizations to do
9265 Enabled with @option{-fprofile-use}.
9267 @item -funroll-loops
9268 @opindex funroll-loops
9269 Unroll loops whose number of iterations can be determined at compile time or
9270 upon entry to the loop. @option{-funroll-loops} implies
9271 @option{-frerun-cse-after-loop}, @option{-fweb} and @option{-frename-registers}.
9272 It also turns on complete loop peeling (i.e.@: complete removal of loops with
9273 a small constant number of iterations). This option makes code larger, and may
9274 or may not make it run faster.
9276 Enabled with @option{-fprofile-use}.
9278 @item -funroll-all-loops
9279 @opindex funroll-all-loops
9280 Unroll all loops, even if their number of iterations is uncertain when
9281 the loop is entered. This usually makes programs run more slowly.
9282 @option{-funroll-all-loops} implies the same options as
9283 @option{-funroll-loops}.
9286 @opindex fpeel-loops
9287 Peels loops for which there is enough information that they do not
9288 roll much (from profile feedback or static analysis). It also turns on
9289 complete loop peeling (i.e.@: complete removal of loops with small constant
9290 number of iterations).
9292 Enabled with @option{-O3} and/or @option{-fprofile-use}.
9294 @item -fmove-loop-invariants
9295 @opindex fmove-loop-invariants
9296 Enables the loop invariant motion pass in the RTL loop optimizer. Enabled
9297 at level @option{-O1}
9300 @opindex fsplit-loops
9301 Split a loop into two if it contains a condition that's always true
9302 for one side of the iteration space and false for the other.
9304 @item -funswitch-loops
9305 @opindex funswitch-loops
9306 Move branches with loop invariant conditions out of the loop, with duplicates
9307 of the loop on both branches (modified according to result of the condition).
9309 @item -ffunction-sections
9310 @itemx -fdata-sections
9311 @opindex ffunction-sections
9312 @opindex fdata-sections
9313 Place each function or data item into its own section in the output
9314 file if the target supports arbitrary sections. The name of the
9315 function or the name of the data item determines the section's name
9318 Use these options on systems where the linker can perform optimizations
9319 to improve locality of reference in the instruction space. Most systems
9320 using the ELF object format and SPARC processors running Solaris 2 have
9321 linkers with such optimizations. AIX may have these optimizations in
9324 Only use these options when there are significant benefits from doing
9325 so. When you specify these options, the assembler and linker
9326 create larger object and executable files and are also slower.
9327 You cannot use @command{gprof} on all systems if you
9328 specify this option, and you may have problems with debugging if
9329 you specify both this option and @option{-g}.
9331 @item -fbranch-target-load-optimize
9332 @opindex fbranch-target-load-optimize
9333 Perform branch target register load optimization before prologue / epilogue
9335 The use of target registers can typically be exposed only during reload,
9336 thus hoisting loads out of loops and doing inter-block scheduling needs
9337 a separate optimization pass.
9339 @item -fbranch-target-load-optimize2
9340 @opindex fbranch-target-load-optimize2
9341 Perform branch target register load optimization after prologue / epilogue
9344 @item -fbtr-bb-exclusive
9345 @opindex fbtr-bb-exclusive
9346 When performing branch target register load optimization, don't reuse
9347 branch target registers within any basic block.
9350 @opindex fstdarg-opt
9351 Optimize the prologue of variadic argument functions with respect to usage of
9354 @item -fsection-anchors
9355 @opindex fsection-anchors
9356 Try to reduce the number of symbolic address calculations by using
9357 shared ``anchor'' symbols to address nearby objects. This transformation
9358 can help to reduce the number of GOT entries and GOT accesses on some
9361 For example, the implementation of the following function @code{foo}:
9365 int foo (void) @{ return a + b + c; @}
9369 usually calculates the addresses of all three variables, but if you
9370 compile it with @option{-fsection-anchors}, it accesses the variables
9371 from a common anchor point instead. The effect is similar to the
9372 following pseudocode (which isn't valid C):
9377 register int *xr = &x;
9378 return xr[&a - &x] + xr[&b - &x] + xr[&c - &x];
9382 Not all targets support this option.
9384 @item --param @var{name}=@var{value}
9386 In some places, GCC uses various constants to control the amount of
9387 optimization that is done. For example, GCC does not inline functions
9388 that contain more than a certain number of instructions. You can
9389 control some of these constants on the command line using the
9390 @option{--param} option.
9392 The names of specific parameters, and the meaning of the values, are
9393 tied to the internals of the compiler, and are subject to change
9394 without notice in future releases.
9396 In each case, the @var{value} is an integer. The allowable choices for
9400 @item predictable-branch-outcome
9401 When branch is predicted to be taken with probability lower than this threshold
9402 (in percent), then it is considered well predictable. The default is 10.
9404 @item max-rtl-if-conversion-insns
9405 RTL if-conversion tries to remove conditional branches around a block and
9406 replace them with conditionally executed instructions. This parameter
9407 gives the maximum number of instructions in a block which should be
9408 considered for if-conversion. The default is 10, though the compiler will
9409 also use other heuristics to decide whether if-conversion is likely to be
9412 @item max-rtl-if-conversion-predictable-cost
9413 @item max-rtl-if-conversion-unpredictable-cost
9414 RTL if-conversion will try to remove conditional branches around a block
9415 and replace them with conditionally executed instructions. These parameters
9416 give the maximum permissible cost for the sequence that would be generated
9417 by if-conversion depending on whether the branch is statically determined
9418 to be predictable or not. The units for this parameter are the same as
9419 those for the GCC internal seq_cost metric. The compiler will try to
9420 provide a reasonable default for this parameter using the BRANCH_COST
9423 @item max-crossjump-edges
9424 The maximum number of incoming edges to consider for cross-jumping.
9425 The algorithm used by @option{-fcrossjumping} is @math{O(N^2)} in
9426 the number of edges incoming to each block. Increasing values mean
9427 more aggressive optimization, making the compilation time increase with
9428 probably small improvement in executable size.
9430 @item min-crossjump-insns
9431 The minimum number of instructions that must be matched at the end
9432 of two blocks before cross-jumping is performed on them. This
9433 value is ignored in the case where all instructions in the block being
9434 cross-jumped from are matched. The default value is 5.
9436 @item max-grow-copy-bb-insns
9437 The maximum code size expansion factor when copying basic blocks
9438 instead of jumping. The expansion is relative to a jump instruction.
9439 The default value is 8.
9441 @item max-goto-duplication-insns
9442 The maximum number of instructions to duplicate to a block that jumps
9443 to a computed goto. To avoid @math{O(N^2)} behavior in a number of
9444 passes, GCC factors computed gotos early in the compilation process,
9445 and unfactors them as late as possible. Only computed jumps at the
9446 end of a basic blocks with no more than max-goto-duplication-insns are
9447 unfactored. The default value is 8.
9449 @item max-delay-slot-insn-search
9450 The maximum number of instructions to consider when looking for an
9451 instruction to fill a delay slot. If more than this arbitrary number of
9452 instructions are searched, the time savings from filling the delay slot
9453 are minimal, so stop searching. Increasing values mean more
9454 aggressive optimization, making the compilation time increase with probably
9455 small improvement in execution time.
9457 @item max-delay-slot-live-search
9458 When trying to fill delay slots, the maximum number of instructions to
9459 consider when searching for a block with valid live register
9460 information. Increasing this arbitrarily chosen value means more
9461 aggressive optimization, increasing the compilation time. This parameter
9462 should be removed when the delay slot code is rewritten to maintain the
9465 @item max-gcse-memory
9466 The approximate maximum amount of memory that can be allocated in
9467 order to perform the global common subexpression elimination
9468 optimization. If more memory than specified is required, the
9469 optimization is not done.
9471 @item max-gcse-insertion-ratio
9472 If the ratio of expression insertions to deletions is larger than this value
9473 for any expression, then RTL PRE inserts or removes the expression and thus
9474 leaves partially redundant computations in the instruction stream. The default value is 20.
9476 @item max-pending-list-length
9477 The maximum number of pending dependencies scheduling allows
9478 before flushing the current state and starting over. Large functions
9479 with few branches or calls can create excessively large lists which
9480 needlessly consume memory and resources.
9482 @item max-modulo-backtrack-attempts
9483 The maximum number of backtrack attempts the scheduler should make
9484 when modulo scheduling a loop. Larger values can exponentially increase
9487 @item max-inline-insns-single
9488 Several parameters control the tree inliner used in GCC@.
9489 This number sets the maximum number of instructions (counted in GCC's
9490 internal representation) in a single function that the tree inliner
9491 considers for inlining. This only affects functions declared
9492 inline and methods implemented in a class declaration (C++).
9493 The default value is 400.
9495 @item max-inline-insns-auto
9496 When you use @option{-finline-functions} (included in @option{-O3}),
9497 a lot of functions that would otherwise not be considered for inlining
9498 by the compiler are investigated. To those functions, a different
9499 (more restrictive) limit compared to functions declared inline can
9501 The default value is 40.
9503 @item inline-min-speedup
9504 When estimated performance improvement of caller + callee runtime exceeds this
9505 threshold (in precent), the function can be inlined regardless the limit on
9506 @option{--param max-inline-insns-single} and @option{--param
9507 max-inline-insns-auto}.
9509 @item large-function-insns
9510 The limit specifying really large functions. For functions larger than this
9511 limit after inlining, inlining is constrained by
9512 @option{--param large-function-growth}. This parameter is useful primarily
9513 to avoid extreme compilation time caused by non-linear algorithms used by the
9515 The default value is 2700.
9517 @item large-function-growth
9518 Specifies maximal growth of large function caused by inlining in percents.
9519 The default value is 100 which limits large function growth to 2.0 times
9522 @item large-unit-insns
9523 The limit specifying large translation unit. Growth caused by inlining of
9524 units larger than this limit is limited by @option{--param inline-unit-growth}.
9525 For small units this might be too tight.
9526 For example, consider a unit consisting of function A
9527 that is inline and B that just calls A three times. If B is small relative to
9528 A, the growth of unit is 300\% and yet such inlining is very sane. For very
9529 large units consisting of small inlineable functions, however, the overall unit
9530 growth limit is needed to avoid exponential explosion of code size. Thus for
9531 smaller units, the size is increased to @option{--param large-unit-insns}
9532 before applying @option{--param inline-unit-growth}. The default is 10000.
9534 @item inline-unit-growth
9535 Specifies maximal overall growth of the compilation unit caused by inlining.
9536 The default value is 20 which limits unit growth to 1.2 times the original
9537 size. Cold functions (either marked cold via an attribute or by profile
9538 feedback) are not accounted into the unit size.
9540 @item ipcp-unit-growth
9541 Specifies maximal overall growth of the compilation unit caused by
9542 interprocedural constant propagation. The default value is 10 which limits
9543 unit growth to 1.1 times the original size.
9545 @item large-stack-frame
9546 The limit specifying large stack frames. While inlining the algorithm is trying
9547 to not grow past this limit too much. The default value is 256 bytes.
9549 @item large-stack-frame-growth
9550 Specifies maximal growth of large stack frames caused by inlining in percents.
9551 The default value is 1000 which limits large stack frame growth to 11 times
9554 @item max-inline-insns-recursive
9555 @itemx max-inline-insns-recursive-auto
9556 Specifies the maximum number of instructions an out-of-line copy of a
9557 self-recursive inline
9558 function can grow into by performing recursive inlining.
9560 @option{--param max-inline-insns-recursive} applies to functions
9562 For functions not declared inline, recursive inlining
9563 happens only when @option{-finline-functions} (included in @option{-O3}) is
9564 enabled; @option{--param max-inline-insns-recursive-auto} applies instead. The
9565 default value is 450.
9567 @item max-inline-recursive-depth
9568 @itemx max-inline-recursive-depth-auto
9569 Specifies the maximum recursion depth used for recursive inlining.
9571 @option{--param max-inline-recursive-depth} applies to functions
9572 declared inline. For functions not declared inline, recursive inlining
9573 happens only when @option{-finline-functions} (included in @option{-O3}) is
9574 enabled; @option{--param max-inline-recursive-depth-auto} applies instead. The
9577 @item min-inline-recursive-probability
9578 Recursive inlining is profitable only for function having deep recursion
9579 in average and can hurt for function having little recursion depth by
9580 increasing the prologue size or complexity of function body to other
9583 When profile feedback is available (see @option{-fprofile-generate}) the actual
9584 recursion depth can be guessed from the probability that function recurses
9585 via a given call expression. This parameter limits inlining only to call
9586 expressions whose probability exceeds the given threshold (in percents).
9587 The default value is 10.
9589 @item early-inlining-insns
9590 Specify growth that the early inliner can make. In effect it increases
9591 the amount of inlining for code having a large abstraction penalty.
9592 The default value is 14.
9594 @item max-early-inliner-iterations
9595 Limit of iterations of the early inliner. This basically bounds
9596 the number of nested indirect calls the early inliner can resolve.
9597 Deeper chains are still handled by late inlining.
9599 @item comdat-sharing-probability
9600 Probability (in percent) that C++ inline function with comdat visibility
9601 are shared across multiple compilation units. The default value is 20.
9603 @item profile-func-internal-id
9604 A parameter to control whether to use function internal id in profile
9605 database lookup. If the value is 0, the compiler uses an id that
9606 is based on function assembler name and filename, which makes old profile
9607 data more tolerant to source changes such as function reordering etc.
9608 The default value is 0.
9610 @item min-vect-loop-bound
9611 The minimum number of iterations under which loops are not vectorized
9612 when @option{-ftree-vectorize} is used. The number of iterations after
9613 vectorization needs to be greater than the value specified by this option
9614 to allow vectorization. The default value is 0.
9616 @item gcse-cost-distance-ratio
9617 Scaling factor in calculation of maximum distance an expression
9618 can be moved by GCSE optimizations. This is currently supported only in the
9619 code hoisting pass. The bigger the ratio, the more aggressive code hoisting
9620 is with simple expressions, i.e., the expressions that have cost
9621 less than @option{gcse-unrestricted-cost}. Specifying 0 disables
9622 hoisting of simple expressions. The default value is 10.
9624 @item gcse-unrestricted-cost
9625 Cost, roughly measured as the cost of a single typical machine
9626 instruction, at which GCSE optimizations do not constrain
9627 the distance an expression can travel. This is currently
9628 supported only in the code hoisting pass. The lesser the cost,
9629 the more aggressive code hoisting is. Specifying 0
9630 allows all expressions to travel unrestricted distances.
9631 The default value is 3.
9633 @item max-hoist-depth
9634 The depth of search in the dominator tree for expressions to hoist.
9635 This is used to avoid quadratic behavior in hoisting algorithm.
9636 The value of 0 does not limit on the search, but may slow down compilation
9637 of huge functions. The default value is 30.
9639 @item max-tail-merge-comparisons
9640 The maximum amount of similar bbs to compare a bb with. This is used to
9641 avoid quadratic behavior in tree tail merging. The default value is 10.
9643 @item max-tail-merge-iterations
9644 The maximum amount of iterations of the pass over the function. This is used to
9645 limit compilation time in tree tail merging. The default value is 2.
9647 @item store-merging-allow-unaligned
9648 Allow the store merging pass to introduce unaligned stores if it is legal to
9649 do so. The default value is 1.
9651 @item max-stores-to-merge
9652 The maximum number of stores to attempt to merge into wider stores in the store
9653 merging pass. The minimum value is 2 and the default is 64.
9655 @item max-unrolled-insns
9656 The maximum number of instructions that a loop may have to be unrolled.
9657 If a loop is unrolled, this parameter also determines how many times
9658 the loop code is unrolled.
9660 @item max-average-unrolled-insns
9661 The maximum number of instructions biased by probabilities of their execution
9662 that a loop may have to be unrolled. If a loop is unrolled,
9663 this parameter also determines how many times the loop code is unrolled.
9665 @item max-unroll-times
9666 The maximum number of unrollings of a single loop.
9668 @item max-peeled-insns
9669 The maximum number of instructions that a loop may have to be peeled.
9670 If a loop is peeled, this parameter also determines how many times
9671 the loop code is peeled.
9673 @item max-peel-times
9674 The maximum number of peelings of a single loop.
9676 @item max-peel-branches
9677 The maximum number of branches on the hot path through the peeled sequence.
9679 @item max-completely-peeled-insns
9680 The maximum number of insns of a completely peeled loop.
9682 @item max-completely-peel-times
9683 The maximum number of iterations of a loop to be suitable for complete peeling.
9685 @item max-completely-peel-loop-nest-depth
9686 The maximum depth of a loop nest suitable for complete peeling.
9688 @item max-unswitch-insns
9689 The maximum number of insns of an unswitched loop.
9691 @item max-unswitch-level
9692 The maximum number of branches unswitched in a single loop.
9694 @item max-loop-headers-insns
9695 The maximum number of insns in loop header duplicated by he copy loop headers
9699 The minimum cost of an expensive expression in the loop invariant motion.
9701 @item iv-consider-all-candidates-bound
9702 Bound on number of candidates for induction variables, below which
9703 all candidates are considered for each use in induction variable
9704 optimizations. If there are more candidates than this,
9705 only the most relevant ones are considered to avoid quadratic time complexity.
9707 @item iv-max-considered-uses
9708 The induction variable optimizations give up on loops that contain more
9709 induction variable uses.
9711 @item iv-always-prune-cand-set-bound
9712 If the number of candidates in the set is smaller than this value,
9713 always try to remove unnecessary ivs from the set
9714 when adding a new one.
9716 @item avg-loop-niter
9717 Average number of iterations of a loop.
9719 @item scev-max-expr-size
9720 Bound on size of expressions used in the scalar evolutions analyzer.
9721 Large expressions slow the analyzer.
9723 @item scev-max-expr-complexity
9724 Bound on the complexity of the expressions in the scalar evolutions analyzer.
9725 Complex expressions slow the analyzer.
9727 @item max-tree-if-conversion-phi-args
9728 Maximum number of arguments in a PHI supported by TREE if conversion
9729 unless the loop is marked with simd pragma.
9731 @item vect-max-version-for-alignment-checks
9732 The maximum number of run-time checks that can be performed when
9733 doing loop versioning for alignment in the vectorizer.
9735 @item vect-max-version-for-alias-checks
9736 The maximum number of run-time checks that can be performed when
9737 doing loop versioning for alias in the vectorizer.
9739 @item vect-max-peeling-for-alignment
9740 The maximum number of loop peels to enhance access alignment
9741 for vectorizer. Value -1 means no limit.
9743 @item max-iterations-to-track
9744 The maximum number of iterations of a loop the brute-force algorithm
9745 for analysis of the number of iterations of the loop tries to evaluate.
9747 @item hot-bb-count-ws-permille
9748 A basic block profile count is considered hot if it contributes to
9749 the given permillage (i.e. 0...1000) of the entire profiled execution.
9751 @item hot-bb-frequency-fraction
9752 Select fraction of the entry block frequency of executions of basic block in
9753 function given basic block needs to have to be considered hot.
9755 @item max-predicted-iterations
9756 The maximum number of loop iterations we predict statically. This is useful
9757 in cases where a function contains a single loop with known bound and
9758 another loop with unknown bound.
9759 The known number of iterations is predicted correctly, while
9760 the unknown number of iterations average to roughly 10. This means that the
9761 loop without bounds appears artificially cold relative to the other one.
9763 @item builtin-expect-probability
9764 Control the probability of the expression having the specified value. This
9765 parameter takes a percentage (i.e. 0 ... 100) as input.
9766 The default probability of 90 is obtained empirically.
9768 @item align-threshold
9770 Select fraction of the maximal frequency of executions of a basic block in
9771 a function to align the basic block.
9773 @item align-loop-iterations
9775 A loop expected to iterate at least the selected number of iterations is
9778 @item tracer-dynamic-coverage
9779 @itemx tracer-dynamic-coverage-feedback
9781 This value is used to limit superblock formation once the given percentage of
9782 executed instructions is covered. This limits unnecessary code size
9785 The @option{tracer-dynamic-coverage-feedback} parameter
9786 is used only when profile
9787 feedback is available. The real profiles (as opposed to statically estimated
9788 ones) are much less balanced allowing the threshold to be larger value.
9790 @item tracer-max-code-growth
9791 Stop tail duplication once code growth has reached given percentage. This is
9792 a rather artificial limit, as most of the duplicates are eliminated later in
9793 cross jumping, so it may be set to much higher values than is the desired code
9796 @item tracer-min-branch-ratio
9798 Stop reverse growth when the reverse probability of best edge is less than this
9799 threshold (in percent).
9801 @item tracer-min-branch-probability
9802 @itemx tracer-min-branch-probability-feedback
9804 Stop forward growth if the best edge has probability lower than this
9807 Similarly to @option{tracer-dynamic-coverage} two parameters are
9808 provided. @option{tracer-min-branch-probability-feedback} is used for
9809 compilation with profile feedback and @option{tracer-min-branch-probability}
9810 compilation without. The value for compilation with profile feedback
9811 needs to be more conservative (higher) in order to make tracer
9814 @item max-cse-path-length
9816 The maximum number of basic blocks on path that CSE considers.
9820 The maximum number of instructions CSE processes before flushing.
9821 The default is 1000.
9823 @item ggc-min-expand
9825 GCC uses a garbage collector to manage its own memory allocation. This
9826 parameter specifies the minimum percentage by which the garbage
9827 collector's heap should be allowed to expand between collections.
9828 Tuning this may improve compilation speed; it has no effect on code
9831 The default is 30% + 70% * (RAM/1GB) with an upper bound of 100% when
9832 RAM >= 1GB@. If @code{getrlimit} is available, the notion of ``RAM'' is
9833 the smallest of actual RAM and @code{RLIMIT_DATA} or @code{RLIMIT_AS}. If
9834 GCC is not able to calculate RAM on a particular platform, the lower
9835 bound of 30% is used. Setting this parameter and
9836 @option{ggc-min-heapsize} to zero causes a full collection to occur at
9837 every opportunity. This is extremely slow, but can be useful for
9840 @item ggc-min-heapsize
9842 Minimum size of the garbage collector's heap before it begins bothering
9843 to collect garbage. The first collection occurs after the heap expands
9844 by @option{ggc-min-expand}% beyond @option{ggc-min-heapsize}. Again,
9845 tuning this may improve compilation speed, and has no effect on code
9848 The default is the smaller of RAM/8, RLIMIT_RSS, or a limit that
9849 tries to ensure that RLIMIT_DATA or RLIMIT_AS are not exceeded, but
9850 with a lower bound of 4096 (four megabytes) and an upper bound of
9851 131072 (128 megabytes). If GCC is not able to calculate RAM on a
9852 particular platform, the lower bound is used. Setting this parameter
9853 very large effectively disables garbage collection. Setting this
9854 parameter and @option{ggc-min-expand} to zero causes a full collection
9855 to occur at every opportunity.
9857 @item max-reload-search-insns
9858 The maximum number of instruction reload should look backward for equivalent
9859 register. Increasing values mean more aggressive optimization, making the
9860 compilation time increase with probably slightly better performance.
9861 The default value is 100.
9863 @item max-cselib-memory-locations
9864 The maximum number of memory locations cselib should take into account.
9865 Increasing values mean more aggressive optimization, making the compilation time
9866 increase with probably slightly better performance. The default value is 500.
9868 @item max-sched-ready-insns
9869 The maximum number of instructions ready to be issued the scheduler should
9870 consider at any given time during the first scheduling pass. Increasing
9871 values mean more thorough searches, making the compilation time increase
9872 with probably little benefit. The default value is 100.
9874 @item max-sched-region-blocks
9875 The maximum number of blocks in a region to be considered for
9876 interblock scheduling. The default value is 10.
9878 @item max-pipeline-region-blocks
9879 The maximum number of blocks in a region to be considered for
9880 pipelining in the selective scheduler. The default value is 15.
9882 @item max-sched-region-insns
9883 The maximum number of insns in a region to be considered for
9884 interblock scheduling. The default value is 100.
9886 @item max-pipeline-region-insns
9887 The maximum number of insns in a region to be considered for
9888 pipelining in the selective scheduler. The default value is 200.
9891 The minimum probability (in percents) of reaching a source block
9892 for interblock speculative scheduling. The default value is 40.
9894 @item max-sched-extend-regions-iters
9895 The maximum number of iterations through CFG to extend regions.
9896 A value of 0 (the default) disables region extensions.
9898 @item max-sched-insn-conflict-delay
9899 The maximum conflict delay for an insn to be considered for speculative motion.
9900 The default value is 3.
9902 @item sched-spec-prob-cutoff
9903 The minimal probability of speculation success (in percents), so that
9904 speculative insns are scheduled.
9905 The default value is 40.
9907 @item sched-state-edge-prob-cutoff
9908 The minimum probability an edge must have for the scheduler to save its
9910 The default value is 10.
9912 @item sched-mem-true-dep-cost
9913 Minimal distance (in CPU cycles) between store and load targeting same
9914 memory locations. The default value is 1.
9916 @item selsched-max-lookahead
9917 The maximum size of the lookahead window of selective scheduling. It is a
9918 depth of search for available instructions.
9919 The default value is 50.
9921 @item selsched-max-sched-times
9922 The maximum number of times that an instruction is scheduled during
9923 selective scheduling. This is the limit on the number of iterations
9924 through which the instruction may be pipelined. The default value is 2.
9926 @item selsched-insns-to-rename
9927 The maximum number of best instructions in the ready list that are considered
9928 for renaming in the selective scheduler. The default value is 2.
9931 The minimum value of stage count that swing modulo scheduler
9932 generates. The default value is 2.
9934 @item max-last-value-rtl
9935 The maximum size measured as number of RTLs that can be recorded in an expression
9936 in combiner for a pseudo register as last known value of that register. The default
9939 @item max-combine-insns
9940 The maximum number of instructions the RTL combiner tries to combine.
9941 The default value is 2 at @option{-Og} and 4 otherwise.
9943 @item integer-share-limit
9944 Small integer constants can use a shared data structure, reducing the
9945 compiler's memory usage and increasing its speed. This sets the maximum
9946 value of a shared integer constant. The default value is 256.
9948 @item ssp-buffer-size
9949 The minimum size of buffers (i.e.@: arrays) that receive stack smashing
9950 protection when @option{-fstack-protection} is used.
9952 @item min-size-for-stack-sharing
9953 The minimum size of variables taking part in stack slot sharing when not
9954 optimizing. The default value is 32.
9956 @item max-jump-thread-duplication-stmts
9957 Maximum number of statements allowed in a block that needs to be
9958 duplicated when threading jumps.
9960 @item max-fields-for-field-sensitive
9961 Maximum number of fields in a structure treated in
9962 a field sensitive manner during pointer analysis. The default is zero
9963 for @option{-O0} and @option{-O1},
9964 and 100 for @option{-Os}, @option{-O2}, and @option{-O3}.
9966 @item prefetch-latency
9967 Estimate on average number of instructions that are executed before
9968 prefetch finishes. The distance prefetched ahead is proportional
9969 to this constant. Increasing this number may also lead to less
9970 streams being prefetched (see @option{simultaneous-prefetches}).
9972 @item simultaneous-prefetches
9973 Maximum number of prefetches that can run at the same time.
9975 @item l1-cache-line-size
9976 The size of cache line in L1 cache, in bytes.
9979 The size of L1 cache, in kilobytes.
9982 The size of L2 cache, in kilobytes.
9984 @item min-insn-to-prefetch-ratio
9985 The minimum ratio between the number of instructions and the
9986 number of prefetches to enable prefetching in a loop.
9988 @item prefetch-min-insn-to-mem-ratio
9989 The minimum ratio between the number of instructions and the
9990 number of memory references to enable prefetching in a loop.
9992 @item use-canonical-types
9993 Whether the compiler should use the ``canonical'' type system. By
9994 default, this should always be 1, which uses a more efficient internal
9995 mechanism for comparing types in C++ and Objective-C++. However, if
9996 bugs in the canonical type system are causing compilation failures,
9997 set this value to 0 to disable canonical types.
9999 @item switch-conversion-max-branch-ratio
10000 Switch initialization conversion refuses to create arrays that are
10001 bigger than @option{switch-conversion-max-branch-ratio} times the number of
10002 branches in the switch.
10004 @item max-partial-antic-length
10005 Maximum length of the partial antic set computed during the tree
10006 partial redundancy elimination optimization (@option{-ftree-pre}) when
10007 optimizing at @option{-O3} and above. For some sorts of source code
10008 the enhanced partial redundancy elimination optimization can run away,
10009 consuming all of the memory available on the host machine. This
10010 parameter sets a limit on the length of the sets that are computed,
10011 which prevents the runaway behavior. Setting a value of 0 for
10012 this parameter allows an unlimited set length.
10014 @item sccvn-max-scc-size
10015 Maximum size of a strongly connected component (SCC) during SCCVN
10016 processing. If this limit is hit, SCCVN processing for the whole
10017 function is not done and optimizations depending on it are
10018 disabled. The default maximum SCC size is 10000.
10020 @item sccvn-max-alias-queries-per-access
10021 Maximum number of alias-oracle queries we perform when looking for
10022 redundancies for loads and stores. If this limit is hit the search
10023 is aborted and the load or store is not considered redundant. The
10024 number of queries is algorithmically limited to the number of
10025 stores on all paths from the load to the function entry.
10026 The default maximum number of queries is 1000.
10028 @item ira-max-loops-num
10029 IRA uses regional register allocation by default. If a function
10030 contains more loops than the number given by this parameter, only at most
10031 the given number of the most frequently-executed loops form regions
10032 for regional register allocation. The default value of the
10035 @item ira-max-conflict-table-size
10036 Although IRA uses a sophisticated algorithm to compress the conflict
10037 table, the table can still require excessive amounts of memory for
10038 huge functions. If the conflict table for a function could be more
10039 than the size in MB given by this parameter, the register allocator
10040 instead uses a faster, simpler, and lower-quality
10041 algorithm that does not require building a pseudo-register conflict table.
10042 The default value of the parameter is 2000.
10044 @item ira-loop-reserved-regs
10045 IRA can be used to evaluate more accurate register pressure in loops
10046 for decisions to move loop invariants (see @option{-O3}). The number
10047 of available registers reserved for some other purposes is given
10048 by this parameter. The default value of the parameter is 2, which is
10049 the minimal number of registers needed by typical instructions.
10050 This value is the best found from numerous experiments.
10052 @item lra-inheritance-ebb-probability-cutoff
10053 LRA tries to reuse values reloaded in registers in subsequent insns.
10054 This optimization is called inheritance. EBB is used as a region to
10055 do this optimization. The parameter defines a minimal fall-through
10056 edge probability in percentage used to add BB to inheritance EBB in
10057 LRA. The default value of the parameter is 40. The value was chosen
10058 from numerous runs of SPEC2000 on x86-64.
10060 @item loop-invariant-max-bbs-in-loop
10061 Loop invariant motion can be very expensive, both in compilation time and
10062 in amount of needed compile-time memory, with very large loops. Loops
10063 with more basic blocks than this parameter won't have loop invariant
10064 motion optimization performed on them. The default value of the
10065 parameter is 1000 for @option{-O1} and 10000 for @option{-O2} and above.
10067 @item loop-max-datarefs-for-datadeps
10068 Building data dependencies is expensive for very large loops. This
10069 parameter limits the number of data references in loops that are
10070 considered for data dependence analysis. These large loops are no
10071 handled by the optimizations using loop data dependencies.
10072 The default value is 1000.
10074 @item max-vartrack-size
10075 Sets a maximum number of hash table slots to use during variable
10076 tracking dataflow analysis of any function. If this limit is exceeded
10077 with variable tracking at assignments enabled, analysis for that
10078 function is retried without it, after removing all debug insns from
10079 the function. If the limit is exceeded even without debug insns, var
10080 tracking analysis is completely disabled for the function. Setting
10081 the parameter to zero makes it unlimited.
10083 @item max-vartrack-expr-depth
10084 Sets a maximum number of recursion levels when attempting to map
10085 variable names or debug temporaries to value expressions. This trades
10086 compilation time for more complete debug information. If this is set too
10087 low, value expressions that are available and could be represented in
10088 debug information may end up not being used; setting this higher may
10089 enable the compiler to find more complex debug expressions, but compile
10090 time and memory use may grow. The default is 12.
10092 @item min-nondebug-insn-uid
10093 Use uids starting at this parameter for nondebug insns. The range below
10094 the parameter is reserved exclusively for debug insns created by
10095 @option{-fvar-tracking-assignments}, but debug insns may get
10096 (non-overlapping) uids above it if the reserved range is exhausted.
10098 @item ipa-sra-ptr-growth-factor
10099 IPA-SRA replaces a pointer to an aggregate with one or more new
10100 parameters only when their cumulative size is less or equal to
10101 @option{ipa-sra-ptr-growth-factor} times the size of the original
10104 @item sra-max-scalarization-size-Ospeed
10105 @item sra-max-scalarization-size-Osize
10106 The two Scalar Reduction of Aggregates passes (SRA and IPA-SRA) aim to
10107 replace scalar parts of aggregates with uses of independent scalar
10108 variables. These parameters control the maximum size, in storage units,
10109 of aggregate which is considered for replacement when compiling for
10111 (@option{sra-max-scalarization-size-Ospeed}) or size
10112 (@option{sra-max-scalarization-size-Osize}) respectively.
10114 @item tm-max-aggregate-size
10115 When making copies of thread-local variables in a transaction, this
10116 parameter specifies the size in bytes after which variables are
10117 saved with the logging functions as opposed to save/restore code
10118 sequence pairs. This option only applies when using
10121 @item graphite-max-nb-scop-params
10122 To avoid exponential effects in the Graphite loop transforms, the
10123 number of parameters in a Static Control Part (SCoP) is bounded. The
10124 default value is 10 parameters. A variable whose value is unknown at
10125 compilation time and defined outside a SCoP is a parameter of the SCoP.
10127 @item graphite-max-bbs-per-function
10128 To avoid exponential effects in the detection of SCoPs, the size of
10129 the functions analyzed by Graphite is bounded. The default value is
10132 @item loop-block-tile-size
10133 Loop blocking or strip mining transforms, enabled with
10134 @option{-floop-block} or @option{-floop-strip-mine}, strip mine each
10135 loop in the loop nest by a given number of iterations. The strip
10136 length can be changed using the @option{loop-block-tile-size}
10137 parameter. The default value is 51 iterations.
10139 @item loop-unroll-jam-size
10140 Specify the unroll factor for the @option{-floop-unroll-and-jam} option. The
10141 default value is 4.
10143 @item loop-unroll-jam-depth
10144 Specify the dimension to be unrolled (counting from the most inner loop)
10145 for the @option{-floop-unroll-and-jam}. The default value is 2.
10147 @item ipa-cp-value-list-size
10148 IPA-CP attempts to track all possible values and types passed to a function's
10149 parameter in order to propagate them and perform devirtualization.
10150 @option{ipa-cp-value-list-size} is the maximum number of values and types it
10151 stores per one formal parameter of a function.
10153 @item ipa-cp-eval-threshold
10154 IPA-CP calculates its own score of cloning profitability heuristics
10155 and performs those cloning opportunities with scores that exceed
10156 @option{ipa-cp-eval-threshold}.
10158 @item ipa-cp-recursion-penalty
10159 Percentage penalty the recursive functions will receive when they
10160 are evaluated for cloning.
10162 @item ipa-cp-single-call-penalty
10163 Percentage penalty functions containg a single call to another
10164 function will receive when they are evaluated for cloning.
10167 @item ipa-max-agg-items
10168 IPA-CP is also capable to propagate a number of scalar values passed
10169 in an aggregate. @option{ipa-max-agg-items} controls the maximum
10170 number of such values per one parameter.
10172 @item ipa-cp-loop-hint-bonus
10173 When IPA-CP determines that a cloning candidate would make the number
10174 of iterations of a loop known, it adds a bonus of
10175 @option{ipa-cp-loop-hint-bonus} to the profitability score of
10178 @item ipa-cp-array-index-hint-bonus
10179 When IPA-CP determines that a cloning candidate would make the index of
10180 an array access known, it adds a bonus of
10181 @option{ipa-cp-array-index-hint-bonus} to the profitability
10182 score of the candidate.
10184 @item ipa-max-aa-steps
10185 During its analysis of function bodies, IPA-CP employs alias analysis
10186 in order to track values pointed to by function parameters. In order
10187 not spend too much time analyzing huge functions, it gives up and
10188 consider all memory clobbered after examining
10189 @option{ipa-max-aa-steps} statements modifying memory.
10191 @item lto-partitions
10192 Specify desired number of partitions produced during WHOPR compilation.
10193 The number of partitions should exceed the number of CPUs used for compilation.
10194 The default value is 32.
10196 @item lto-min-partition
10197 Size of minimal partition for WHOPR (in estimated instructions).
10198 This prevents expenses of splitting very small programs into too many
10201 @item lto-max-partition
10202 Size of max partition for WHOPR (in estimated instructions).
10203 to provide an upper bound for individual size of partition.
10204 Meant to be used only with balanced partitioning.
10206 @item cxx-max-namespaces-for-diagnostic-help
10207 The maximum number of namespaces to consult for suggestions when C++
10208 name lookup fails for an identifier. The default is 1000.
10210 @item sink-frequency-threshold
10211 The maximum relative execution frequency (in percents) of the target block
10212 relative to a statement's original block to allow statement sinking of a
10213 statement. Larger numbers result in more aggressive statement sinking.
10214 The default value is 75. A small positive adjustment is applied for
10215 statements with memory operands as those are even more profitable so sink.
10217 @item max-stores-to-sink
10218 The maximum number of conditional store pairs that can be sunk. Set to 0
10219 if either vectorization (@option{-ftree-vectorize}) or if-conversion
10220 (@option{-ftree-loop-if-convert}) is disabled. The default is 2.
10222 @item allow-store-data-races
10223 Allow optimizers to introduce new data races on stores.
10224 Set to 1 to allow, otherwise to 0. This option is enabled by default
10225 at optimization level @option{-Ofast}.
10227 @item case-values-threshold
10228 The smallest number of different values for which it is best to use a
10229 jump-table instead of a tree of conditional branches. If the value is
10230 0, use the default for the machine. The default is 0.
10232 @item tree-reassoc-width
10233 Set the maximum number of instructions executed in parallel in
10234 reassociated tree. This parameter overrides target dependent
10235 heuristics used by default if has non zero value.
10237 @item sched-pressure-algorithm
10238 Choose between the two available implementations of
10239 @option{-fsched-pressure}. Algorithm 1 is the original implementation
10240 and is the more likely to prevent instructions from being reordered.
10241 Algorithm 2 was designed to be a compromise between the relatively
10242 conservative approach taken by algorithm 1 and the rather aggressive
10243 approach taken by the default scheduler. It relies more heavily on
10244 having a regular register file and accurate register pressure classes.
10245 See @file{haifa-sched.c} in the GCC sources for more details.
10247 The default choice depends on the target.
10249 @item max-slsr-cand-scan
10250 Set the maximum number of existing candidates that are considered when
10251 seeking a basis for a new straight-line strength reduction candidate.
10254 Enable buffer overflow detection for global objects. This kind
10255 of protection is enabled by default if you are using
10256 @option{-fsanitize=address} option.
10257 To disable global objects protection use @option{--param asan-globals=0}.
10260 Enable buffer overflow detection for stack objects. This kind of
10261 protection is enabled by default when using @option{-fsanitize=address}.
10262 To disable stack protection use @option{--param asan-stack=0} option.
10264 @item asan-instrument-reads
10265 Enable buffer overflow detection for memory reads. This kind of
10266 protection is enabled by default when using @option{-fsanitize=address}.
10267 To disable memory reads protection use
10268 @option{--param asan-instrument-reads=0}.
10270 @item asan-instrument-writes
10271 Enable buffer overflow detection for memory writes. This kind of
10272 protection is enabled by default when using @option{-fsanitize=address}.
10273 To disable memory writes protection use
10274 @option{--param asan-instrument-writes=0} option.
10276 @item asan-memintrin
10277 Enable detection for built-in functions. This kind of protection
10278 is enabled by default when using @option{-fsanitize=address}.
10279 To disable built-in functions protection use
10280 @option{--param asan-memintrin=0}.
10282 @item asan-use-after-return
10283 Enable detection of use-after-return. This kind of protection
10284 is enabled by default when using @option{-fsanitize=address} option.
10285 To disable use-after-return detection use
10286 @option{--param asan-use-after-return=0}.
10288 Note: The check is disabled by default at runtime. To enable the check,
10289 you should set environment variable @env{ASAN_OPTIONS} to
10290 @code{detect_stack_use_after_return=1}.
10292 @item asan-instrumentation-with-call-threshold
10293 If number of memory accesses in function being instrumented
10294 is greater or equal to this number, use callbacks instead of inline checks.
10295 E.g. to disable inline code use
10296 @option{--param asan-instrumentation-with-call-threshold=0}.
10298 @item use-after-scope-direct-emission-threshold
10299 If size of a local variables in bytes is smaller of equal to this number,
10300 direct instruction emission is utilized to poison and unpoison local variables.
10302 @item chkp-max-ctor-size
10303 Static constructors generated by Pointer Bounds Checker may become very
10304 large and significantly increase compile time at optimization level
10305 @option{-O1} and higher. This parameter is a maximum nubmer of statements
10306 in a single generated constructor. Default value is 5000.
10308 @item max-fsm-thread-path-insns
10309 Maximum number of instructions to copy when duplicating blocks on a
10310 finite state automaton jump thread path. The default is 100.
10312 @item max-fsm-thread-length
10313 Maximum number of basic blocks on a finite state automaton jump thread
10314 path. The default is 10.
10316 @item max-fsm-thread-paths
10317 Maximum number of new jump thread paths to create for a finite state
10318 automaton. The default is 50.
10320 @item parloops-chunk-size
10321 Chunk size of omp schedule for loops parallelized by parloops. The default
10324 @item parloops-schedule
10325 Schedule type of omp schedule for loops parallelized by parloops (static,
10326 dynamic, guided, auto, runtime). The default is static.
10328 @item max-ssa-name-query-depth
10329 Maximum depth of recursion when querying properties of SSA names in things
10330 like fold routines. One level of recursion corresponds to following a
10333 @item hsa-gen-debug-stores
10334 Enable emission of special debug stores within HSA kernels which are
10335 then read and reported by libgomp plugin. Generation of these stores
10336 is disabled by default, use @option{--param hsa-gen-debug-stores=1} to
10339 @item max-speculative-devirt-maydefs
10340 The maximum number of may-defs we analyze when looking for a must-def
10341 specifying the dynamic type of an object that invokes a virtual call
10342 we may be able to devirtualize speculatively.
10344 @item max-vrp-switch-assertions
10345 The maximum number of assertions to add along the default edge of a switch
10346 statement during VRP. The default is 10.
10350 @node Instrumentation Options
10351 @section Program Instrumentation Options
10352 @cindex instrumentation options
10353 @cindex program instrumentation options
10354 @cindex run-time error checking options
10355 @cindex profiling options
10356 @cindex options, program instrumentation
10357 @cindex options, run-time error checking
10358 @cindex options, profiling
10360 GCC supports a number of command-line options that control adding
10361 run-time instrumentation to the code it normally generates.
10362 For example, one purpose of instrumentation is collect profiling
10363 statistics for use in finding program hot spots, code coverage
10364 analysis, or profile-guided optimizations.
10365 Another class of program instrumentation is adding run-time checking
10366 to detect programming errors like invalid pointer
10367 dereferences or out-of-bounds array accesses, as well as deliberately
10368 hostile attacks such as stack smashing or C++ vtable hijacking.
10369 There is also a general hook which can be used to implement other
10370 forms of tracing or function-level instrumentation for debug or
10371 program analysis purposes.
10374 @cindex @command{prof}
10377 Generate extra code to write profile information suitable for the
10378 analysis program @command{prof}. You must use this option when compiling
10379 the source files you want data about, and you must also use it when
10382 @cindex @command{gprof}
10385 Generate extra code to write profile information suitable for the
10386 analysis program @command{gprof}. You must use this option when compiling
10387 the source files you want data about, and you must also use it when
10390 @item -fprofile-arcs
10391 @opindex fprofile-arcs
10392 Add code so that program flow @dfn{arcs} are instrumented. During
10393 execution the program records how many times each branch and call is
10394 executed and how many times it is taken or returns. When the compiled
10395 program exits it saves this data to a file called
10396 @file{@var{auxname}.gcda} for each source file. The data may be used for
10397 profile-directed optimizations (@option{-fbranch-probabilities}), or for
10398 test coverage analysis (@option{-ftest-coverage}). Each object file's
10399 @var{auxname} is generated from the name of the output file, if
10400 explicitly specified and it is not the final executable, otherwise it is
10401 the basename of the source file. In both cases any suffix is removed
10402 (e.g.@: @file{foo.gcda} for input file @file{dir/foo.c}, or
10403 @file{dir/foo.gcda} for output file specified as @option{-o dir/foo.o}).
10404 @xref{Cross-profiling}.
10406 @cindex @command{gcov}
10410 This option is used to compile and link code instrumented for coverage
10411 analysis. The option is a synonym for @option{-fprofile-arcs}
10412 @option{-ftest-coverage} (when compiling) and @option{-lgcov} (when
10413 linking). See the documentation for those options for more details.
10418 Compile the source files with @option{-fprofile-arcs} plus optimization
10419 and code generation options. For test coverage analysis, use the
10420 additional @option{-ftest-coverage} option. You do not need to profile
10421 every source file in a program.
10424 Link your object files with @option{-lgcov} or @option{-fprofile-arcs}
10425 (the latter implies the former).
10428 Run the program on a representative workload to generate the arc profile
10429 information. This may be repeated any number of times. You can run
10430 concurrent instances of your program, and provided that the file system
10431 supports locking, the data files will be correctly updated. Also
10432 @code{fork} calls are detected and correctly handled (double counting
10436 For profile-directed optimizations, compile the source files again with
10437 the same optimization and code generation options plus
10438 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
10439 Control Optimization}).
10442 For test coverage analysis, use @command{gcov} to produce human readable
10443 information from the @file{.gcno} and @file{.gcda} files. Refer to the
10444 @command{gcov} documentation for further information.
10448 With @option{-fprofile-arcs}, for each function of your program GCC
10449 creates a program flow graph, then finds a spanning tree for the graph.
10450 Only arcs that are not on the spanning tree have to be instrumented: the
10451 compiler adds code to count the number of times that these arcs are
10452 executed. When an arc is the only exit or only entrance to a block, the
10453 instrumentation code can be added to the block; otherwise, a new basic
10454 block must be created to hold the instrumentation code.
10457 @item -ftest-coverage
10458 @opindex ftest-coverage
10459 Produce a notes file that the @command{gcov} code-coverage utility
10460 (@pxref{Gcov,, @command{gcov}---a Test Coverage Program}) can use to
10461 show program coverage. Each source file's note file is called
10462 @file{@var{auxname}.gcno}. Refer to the @option{-fprofile-arcs} option
10463 above for a description of @var{auxname} and instructions on how to
10464 generate test coverage data. Coverage data matches the source files
10465 more closely if you do not optimize.
10467 @item -fprofile-dir=@var{path}
10468 @opindex fprofile-dir
10470 Set the directory to search for the profile data files in to @var{path}.
10471 This option affects only the profile data generated by
10472 @option{-fprofile-generate}, @option{-ftest-coverage}, @option{-fprofile-arcs}
10473 and used by @option{-fprofile-use} and @option{-fbranch-probabilities}
10474 and its related options. Both absolute and relative paths can be used.
10475 By default, GCC uses the current directory as @var{path}, thus the
10476 profile data file appears in the same directory as the object file.
10478 @item -fprofile-generate
10479 @itemx -fprofile-generate=@var{path}
10480 @opindex fprofile-generate
10482 Enable options usually used for instrumenting application to produce
10483 profile useful for later recompilation with profile feedback based
10484 optimization. You must use @option{-fprofile-generate} both when
10485 compiling and when linking your program.
10487 The following options are enabled: @option{-fprofile-arcs}, @option{-fprofile-values}, @option{-fvpt}.
10489 If @var{path} is specified, GCC looks at the @var{path} to find
10490 the profile feedback data files. See @option{-fprofile-dir}.
10492 To optimize the program based on the collected profile information, use
10493 @option{-fprofile-use}. @xref{Optimize Options}, for more information.
10495 @item -fprofile-update=@var{method}
10496 @opindex fprofile-update
10498 Alter the update method for an application instrumented for profile
10499 feedback based optimization. The @var{method} argument should be one of
10500 @samp{single} or @samp{atomic}. The first one is useful for single-threaded
10501 applications, while the second one prevents profile corruption by emitting
10504 @strong{Warning:} When an application does not properly join all threads
10505 (or creates an detached thread), a profile file can be still corrupted.
10507 @item -fsanitize=address
10508 @opindex fsanitize=address
10509 Enable AddressSanitizer, a fast memory error detector.
10510 Memory access instructions are instrumented to detect
10511 out-of-bounds and use-after-free bugs.
10512 The option enables @option{-fsanitize-address-use-after-scope}.
10513 See @uref{https://github.com/google/sanitizers/wiki/AddressSanitizer} for
10514 more details. The run-time behavior can be influenced using the
10515 @env{ASAN_OPTIONS} environment variable. When set to @code{help=1},
10516 the available options are shown at startup of the instrumented program. See
10517 @url{https://github.com/google/sanitizers/wiki/AddressSanitizerFlags#run-time-flags}
10518 for a list of supported options.
10519 The option can't be combined with @option{-fsanitize=thread}.
10521 @item -fsanitize=kernel-address
10522 @opindex fsanitize=kernel-address
10523 Enable AddressSanitizer for Linux kernel.
10524 The option enables @option{-fsanitize-address-use-after-scope}.
10525 See @uref{https://github.com/google/kasan/wiki} for more details.
10527 @item -fsanitize=thread
10528 @opindex fsanitize=thread
10529 Enable ThreadSanitizer, a fast data race detector.
10530 Memory access instructions are instrumented to detect
10531 data race bugs. See @uref{https://github.com/google/sanitizers/wiki#threadsanitizer} for more
10532 details. The run-time behavior can be influenced using the @env{TSAN_OPTIONS}
10533 environment variable; see
10534 @url{https://github.com/google/sanitizers/wiki/ThreadSanitizerFlags} for a list of
10536 The option can't be combined with @option{-fsanitize=address}
10537 and/or @option{-fsanitize=leak}.
10539 @item -fsanitize=leak
10540 @opindex fsanitize=leak
10541 Enable LeakSanitizer, a memory leak detector.
10542 This option only matters for linking of executables and
10543 the executable is linked against a library that overrides @code{malloc}
10544 and other allocator functions. See
10545 @uref{https://github.com/google/sanitizers/wiki/AddressSanitizerLeakSanitizer} for more
10546 details. The run-time behavior can be influenced using the
10547 @env{LSAN_OPTIONS} environment variable.
10548 The option can't be combined with @option{-fsanitize=thread}.
10550 @item -fsanitize=undefined
10551 @opindex fsanitize=undefined
10552 Enable UndefinedBehaviorSanitizer, a fast undefined behavior detector.
10553 Various computations are instrumented to detect undefined behavior
10554 at runtime. Current suboptions are:
10558 @item -fsanitize=shift
10559 @opindex fsanitize=shift
10560 This option enables checking that the result of a shift operation is
10561 not undefined. Note that what exactly is considered undefined differs
10562 slightly between C and C++, as well as between ISO C90 and C99, etc.
10563 This option has two suboptions, @option{-fsanitize=shift-base} and
10564 @option{-fsanitize=shift-exponent}.
10566 @item -fsanitize=shift-exponent
10567 @opindex fsanitize=shift-exponent
10568 This option enables checking that the second argument of a shift operation
10569 is not negative and is smaller than the precision of the promoted first
10572 @item -fsanitize=shift-base
10573 @opindex fsanitize=shift-base
10574 If the second argument of a shift operation is within range, check that the
10575 result of a shift operation is not undefined. Note that what exactly is
10576 considered undefined differs slightly between C and C++, as well as between
10577 ISO C90 and C99, etc.
10579 @item -fsanitize=integer-divide-by-zero
10580 @opindex fsanitize=integer-divide-by-zero
10581 Detect integer division by zero as well as @code{INT_MIN / -1} division.
10583 @item -fsanitize=unreachable
10584 @opindex fsanitize=unreachable
10585 With this option, the compiler turns the @code{__builtin_unreachable}
10586 call into a diagnostics message call instead. When reaching the
10587 @code{__builtin_unreachable} call, the behavior is undefined.
10589 @item -fsanitize=vla-bound
10590 @opindex fsanitize=vla-bound
10591 This option instructs the compiler to check that the size of a variable
10592 length array is positive.
10594 @item -fsanitize=null
10595 @opindex fsanitize=null
10596 This option enables pointer checking. Particularly, the application
10597 built with this option turned on will issue an error message when it
10598 tries to dereference a NULL pointer, or if a reference (possibly an
10599 rvalue reference) is bound to a NULL pointer, or if a method is invoked
10600 on an object pointed by a NULL pointer.
10602 @item -fsanitize=return
10603 @opindex fsanitize=return
10604 This option enables return statement checking. Programs
10605 built with this option turned on will issue an error message
10606 when the end of a non-void function is reached without actually
10607 returning a value. This option works in C++ only.
10609 @item -fsanitize=signed-integer-overflow
10610 @opindex fsanitize=signed-integer-overflow
10611 This option enables signed integer overflow checking. We check that
10612 the result of @code{+}, @code{*}, and both unary and binary @code{-}
10613 does not overflow in the signed arithmetics. Note, integer promotion
10614 rules must be taken into account. That is, the following is not an
10617 signed char a = SCHAR_MAX;
10621 @item -fsanitize=bounds
10622 @opindex fsanitize=bounds
10623 This option enables instrumentation of array bounds. Various out of bounds
10624 accesses are detected. Flexible array members, flexible array member-like
10625 arrays, and initializers of variables with static storage are not instrumented.
10627 @item -fsanitize=bounds-strict
10628 @opindex fsanitize=bounds-strict
10629 This option enables strict instrumentation of array bounds. Most out of bounds
10630 accesses are detected, including flexible array members and flexible array
10631 member-like arrays. Initializers of variables with static storage are not
10634 @item -fsanitize=alignment
10635 @opindex fsanitize=alignment
10637 This option enables checking of alignment of pointers when they are
10638 dereferenced, or when a reference is bound to insufficiently aligned target,
10639 or when a method or constructor is invoked on insufficiently aligned object.
10641 @item -fsanitize=object-size
10642 @opindex fsanitize=object-size
10643 This option enables instrumentation of memory references using the
10644 @code{__builtin_object_size} function. Various out of bounds pointer
10645 accesses are detected.
10647 @item -fsanitize=float-divide-by-zero
10648 @opindex fsanitize=float-divide-by-zero
10649 Detect floating-point division by zero. Unlike other similar options,
10650 @option{-fsanitize=float-divide-by-zero} is not enabled by
10651 @option{-fsanitize=undefined}, since floating-point division by zero can
10652 be a legitimate way of obtaining infinities and NaNs.
10654 @item -fsanitize=float-cast-overflow
10655 @opindex fsanitize=float-cast-overflow
10656 This option enables floating-point type to integer conversion checking.
10657 We check that the result of the conversion does not overflow.
10658 Unlike other similar options, @option{-fsanitize=float-cast-overflow} is
10659 not enabled by @option{-fsanitize=undefined}.
10660 This option does not work well with @code{FE_INVALID} exceptions enabled.
10662 @item -fsanitize=nonnull-attribute
10663 @opindex fsanitize=nonnull-attribute
10665 This option enables instrumentation of calls, checking whether null values
10666 are not passed to arguments marked as requiring a non-null value by the
10667 @code{nonnull} function attribute.
10669 @item -fsanitize=returns-nonnull-attribute
10670 @opindex fsanitize=returns-nonnull-attribute
10672 This option enables instrumentation of return statements in functions
10673 marked with @code{returns_nonnull} function attribute, to detect returning
10674 of null values from such functions.
10676 @item -fsanitize=bool
10677 @opindex fsanitize=bool
10679 This option enables instrumentation of loads from bool. If a value other
10680 than 0/1 is loaded, a run-time error is issued.
10682 @item -fsanitize=enum
10683 @opindex fsanitize=enum
10685 This option enables instrumentation of loads from an enum type. If
10686 a value outside the range of values for the enum type is loaded,
10687 a run-time error is issued.
10689 @item -fsanitize=vptr
10690 @opindex fsanitize=vptr
10692 This option enables instrumentation of C++ member function calls, member
10693 accesses and some conversions between pointers to base and derived classes,
10694 to verify the referenced object has the correct dynamic type.
10698 While @option{-ftrapv} causes traps for signed overflows to be emitted,
10699 @option{-fsanitize=undefined} gives a diagnostic message.
10700 This currently works only for the C family of languages.
10702 @item -fno-sanitize=all
10703 @opindex fno-sanitize=all
10705 This option disables all previously enabled sanitizers.
10706 @option{-fsanitize=all} is not allowed, as some sanitizers cannot be used
10709 @item -fasan-shadow-offset=@var{number}
10710 @opindex fasan-shadow-offset
10711 This option forces GCC to use custom shadow offset in AddressSanitizer checks.
10712 It is useful for experimenting with different shadow memory layouts in
10713 Kernel AddressSanitizer.
10715 @item -fsanitize-sections=@var{s1},@var{s2},...
10716 @opindex fsanitize-sections
10717 Sanitize global variables in selected user-defined sections. @var{si} may
10720 @item -fsanitize-recover@r{[}=@var{opts}@r{]}
10721 @opindex fsanitize-recover
10722 @opindex fno-sanitize-recover
10723 @option{-fsanitize-recover=} controls error recovery mode for sanitizers
10724 mentioned in comma-separated list of @var{opts}. Enabling this option
10725 for a sanitizer component causes it to attempt to continue
10726 running the program as if no error happened. This means multiple
10727 runtime errors can be reported in a single program run, and the exit
10728 code of the program may indicate success even when errors
10729 have been reported. The @option{-fno-sanitize-recover=} option
10730 can be used to alter
10731 this behavior: only the first detected error is reported
10732 and program then exits with a non-zero exit code.
10734 Currently this feature only works for @option{-fsanitize=undefined} (and its suboptions
10735 except for @option{-fsanitize=unreachable} and @option{-fsanitize=return}),
10736 @option{-fsanitize=float-cast-overflow}, @option{-fsanitize=float-divide-by-zero},
10737 @option{-fsanitize=bounds-strict},
10738 @option{-fsanitize=kernel-address} and @option{-fsanitize=address}.
10739 For these sanitizers error recovery is turned on by default,
10740 except @option{-fsanitize=address}, for which this feature is experimental.
10741 @option{-fsanitize-recover=all} and @option{-fno-sanitize-recover=all} is also
10742 accepted, the former enables recovery for all sanitizers that support it,
10743 the latter disables recovery for all sanitizers that support it.
10745 Even if a recovery mode is turned on the compiler side, it needs to be also
10746 enabled on the runtime library side, otherwise the failures are still fatal.
10747 The runtime library defaults to @code{halt_on_error=0} for
10748 ThreadSanitizer and UndefinedBehaviorSanitizer, while default value for
10749 AddressSanitizer is @code{halt_on_error=1}. This can be overridden through
10750 setting the @code{halt_on_error} flag in the corresponding environment variable.
10752 Syntax without explicit @var{opts} parameter is deprecated. It is equivalent to
10754 -fsanitize-recover=undefined,float-cast-overflow,float-divide-by-zero,bounds-strict
10757 Similarly @option{-fno-sanitize-recover} is equivalent to
10759 -fno-sanitize-recover=undefined,float-cast-overflow,float-divide-by-zero,bounds-strict
10762 @item -fsanitize-address-use-after-scope
10763 @opindex fsanitize-address-use-after-scope
10764 Enable sanitization of local variables to detect use-after-scope bugs.
10765 The option sets @option{-fstack-reuse} to @samp{none}.
10767 @item -fsanitize-undefined-trap-on-error
10768 @opindex fsanitize-undefined-trap-on-error
10769 The @option{-fsanitize-undefined-trap-on-error} option instructs the compiler to
10770 report undefined behavior using @code{__builtin_trap} rather than
10771 a @code{libubsan} library routine. The advantage of this is that the
10772 @code{libubsan} library is not needed and is not linked in, so this
10773 is usable even in freestanding environments.
10775 @item -fsanitize-coverage=trace-pc
10776 @opindex fsanitize-coverage=trace-pc
10777 Enable coverage-guided fuzzing code instrumentation.
10778 Inserts a call to @code{__sanitizer_cov_trace_pc} into every basic block.
10780 @item -fbounds-check
10781 @opindex fbounds-check
10782 For front ends that support it, generate additional code to check that
10783 indices used to access arrays are within the declared range. This is
10784 currently only supported by the Java and Fortran front ends, where
10785 this option defaults to true and false respectively.
10787 @item -fcheck-pointer-bounds
10788 @opindex fcheck-pointer-bounds
10789 @opindex fno-check-pointer-bounds
10790 @cindex Pointer Bounds Checker options
10791 Enable Pointer Bounds Checker instrumentation. Each memory reference
10792 is instrumented with checks of the pointer used for memory access against
10793 bounds associated with that pointer.
10796 is only an implementation for Intel MPX available, thus x86 GNU/Linux target
10797 and @option{-mmpx} are required to enable this feature.
10798 MPX-based instrumentation requires
10799 a runtime library to enable MPX in hardware and handle bounds
10800 violation signals. By default when @option{-fcheck-pointer-bounds}
10801 and @option{-mmpx} options are used to link a program, the GCC driver
10802 links against the @file{libmpx} and @file{libmpxwrappers} libraries.
10803 Bounds checking on calls to dynamic libraries requires a linker
10804 with @option{-z bndplt} support; if GCC was configured with a linker
10805 without support for this option (including the Gold linker and older
10806 versions of ld), a warning is given if you link with @option{-mmpx}
10807 without also specifying @option{-static}, since the overall effectiveness
10808 of the bounds checking protection is reduced.
10809 See also @option{-static-libmpxwrappers}.
10811 MPX-based instrumentation
10812 may be used for debugging and also may be included in production code
10813 to increase program security. Depending on usage, you may
10814 have different requirements for the runtime library. The current version
10815 of the MPX runtime library is more oriented for use as a debugging
10816 tool. MPX runtime library usage implies @option{-lpthread}. See
10817 also @option{-static-libmpx}. The runtime library behavior can be
10818 influenced using various @env{CHKP_RT_*} environment variables. See
10819 @uref{https://gcc.gnu.org/wiki/Intel%20MPX%20support%20in%20the%20GCC%20compiler}
10822 Generated instrumentation may be controlled by various
10823 @option{-fchkp-*} options and by the @code{bnd_variable_size}
10824 structure field attribute (@pxref{Type Attributes}) and
10825 @code{bnd_legacy}, and @code{bnd_instrument} function attributes
10826 (@pxref{Function Attributes}). GCC also provides a number of built-in
10827 functions for controlling the Pointer Bounds Checker. @xref{Pointer
10828 Bounds Checker builtins}, for more information.
10830 @item -fchkp-check-incomplete-type
10831 @opindex fchkp-check-incomplete-type
10832 @opindex fno-chkp-check-incomplete-type
10833 Generate pointer bounds checks for variables with incomplete type.
10834 Enabled by default.
10836 @item -fchkp-narrow-bounds
10837 @opindex fchkp-narrow-bounds
10838 @opindex fno-chkp-narrow-bounds
10839 Controls bounds used by Pointer Bounds Checker for pointers to object
10840 fields. If narrowing is enabled then field bounds are used. Otherwise
10841 object bounds are used. See also @option{-fchkp-narrow-to-innermost-array}
10842 and @option{-fchkp-first-field-has-own-bounds}. Enabled by default.
10844 @item -fchkp-first-field-has-own-bounds
10845 @opindex fchkp-first-field-has-own-bounds
10846 @opindex fno-chkp-first-field-has-own-bounds
10847 Forces Pointer Bounds Checker to use narrowed bounds for the address of the
10848 first field in the structure. By default a pointer to the first field has
10849 the same bounds as a pointer to the whole structure.
10851 @item -fchkp-narrow-to-innermost-array
10852 @opindex fchkp-narrow-to-innermost-array
10853 @opindex fno-chkp-narrow-to-innermost-array
10854 Forces Pointer Bounds Checker to use bounds of the innermost arrays in
10855 case of nested static array access. By default this option is disabled and
10856 bounds of the outermost array are used.
10858 @item -fchkp-optimize
10859 @opindex fchkp-optimize
10860 @opindex fno-chkp-optimize
10861 Enables Pointer Bounds Checker optimizations. Enabled by default at
10862 optimization levels @option{-O}, @option{-O2}, @option{-O3}.
10864 @item -fchkp-use-fast-string-functions
10865 @opindex fchkp-use-fast-string-functions
10866 @opindex fno-chkp-use-fast-string-functions
10867 Enables use of @code{*_nobnd} versions of string functions (not copying bounds)
10868 by Pointer Bounds Checker. Disabled by default.
10870 @item -fchkp-use-nochk-string-functions
10871 @opindex fchkp-use-nochk-string-functions
10872 @opindex fno-chkp-use-nochk-string-functions
10873 Enables use of @code{*_nochk} versions of string functions (not checking bounds)
10874 by Pointer Bounds Checker. Disabled by default.
10876 @item -fchkp-use-static-bounds
10877 @opindex fchkp-use-static-bounds
10878 @opindex fno-chkp-use-static-bounds
10879 Allow Pointer Bounds Checker to generate static bounds holding
10880 bounds of static variables. Enabled by default.
10882 @item -fchkp-use-static-const-bounds
10883 @opindex fchkp-use-static-const-bounds
10884 @opindex fno-chkp-use-static-const-bounds
10885 Use statically-initialized bounds for constant bounds instead of
10886 generating them each time they are required. By default enabled when
10887 @option{-fchkp-use-static-bounds} is enabled.
10889 @item -fchkp-treat-zero-dynamic-size-as-infinite
10890 @opindex fchkp-treat-zero-dynamic-size-as-infinite
10891 @opindex fno-chkp-treat-zero-dynamic-size-as-infinite
10892 With this option, objects with incomplete type whose
10893 dynamically-obtained size is zero are treated as having infinite size
10894 instead by Pointer Bounds
10895 Checker. This option may be helpful if a program is linked with a library
10896 missing size information for some symbols. Disabled by default.
10898 @item -fchkp-check-read
10899 @opindex fchkp-check-read
10900 @opindex fno-chkp-check-read
10901 Instructs Pointer Bounds Checker to generate checks for all read
10902 accesses to memory. Enabled by default.
10904 @item -fchkp-check-write
10905 @opindex fchkp-check-write
10906 @opindex fno-chkp-check-write
10907 Instructs Pointer Bounds Checker to generate checks for all write
10908 accesses to memory. Enabled by default.
10910 @item -fchkp-store-bounds
10911 @opindex fchkp-store-bounds
10912 @opindex fno-chkp-store-bounds
10913 Instructs Pointer Bounds Checker to generate bounds stores for
10914 pointer writes. Enabled by default.
10916 @item -fchkp-instrument-calls
10917 @opindex fchkp-instrument-calls
10918 @opindex fno-chkp-instrument-calls
10919 Instructs Pointer Bounds Checker to pass pointer bounds to calls.
10920 Enabled by default.
10922 @item -fchkp-instrument-marked-only
10923 @opindex fchkp-instrument-marked-only
10924 @opindex fno-chkp-instrument-marked-only
10925 Instructs Pointer Bounds Checker to instrument only functions
10926 marked with the @code{bnd_instrument} attribute
10927 (@pxref{Function Attributes}). Disabled by default.
10929 @item -fchkp-use-wrappers
10930 @opindex fchkp-use-wrappers
10931 @opindex fno-chkp-use-wrappers
10932 Allows Pointer Bounds Checker to replace calls to built-in functions
10933 with calls to wrapper functions. When @option{-fchkp-use-wrappers}
10934 is used to link a program, the GCC driver automatically links
10935 against @file{libmpxwrappers}. See also @option{-static-libmpxwrappers}.
10936 Enabled by default.
10938 @item -fstack-protector
10939 @opindex fstack-protector
10940 Emit extra code to check for buffer overflows, such as stack smashing
10941 attacks. This is done by adding a guard variable to functions with
10942 vulnerable objects. This includes functions that call @code{alloca}, and
10943 functions with buffers larger than 8 bytes. The guards are initialized
10944 when a function is entered and then checked when the function exits.
10945 If a guard check fails, an error message is printed and the program exits.
10947 @item -fstack-protector-all
10948 @opindex fstack-protector-all
10949 Like @option{-fstack-protector} except that all functions are protected.
10951 @item -fstack-protector-strong
10952 @opindex fstack-protector-strong
10953 Like @option{-fstack-protector} but includes additional functions to
10954 be protected --- those that have local array definitions, or have
10955 references to local frame addresses.
10957 @item -fstack-protector-explicit
10958 @opindex fstack-protector-explicit
10959 Like @option{-fstack-protector} but only protects those functions which
10960 have the @code{stack_protect} attribute.
10962 @item -fstack-check
10963 @opindex fstack-check
10964 Generate code to verify that you do not go beyond the boundary of the
10965 stack. You should specify this flag if you are running in an
10966 environment with multiple threads, but you only rarely need to specify it in
10967 a single-threaded environment since stack overflow is automatically
10968 detected on nearly all systems if there is only one stack.
10970 Note that this switch does not actually cause checking to be done; the
10971 operating system or the language runtime must do that. The switch causes
10972 generation of code to ensure that they see the stack being extended.
10974 You can additionally specify a string parameter: @samp{no} means no
10975 checking, @samp{generic} means force the use of old-style checking,
10976 @samp{specific} means use the best checking method and is equivalent
10977 to bare @option{-fstack-check}.
10979 Old-style checking is a generic mechanism that requires no specific
10980 target support in the compiler but comes with the following drawbacks:
10984 Modified allocation strategy for large objects: they are always
10985 allocated dynamically if their size exceeds a fixed threshold.
10988 Fixed limit on the size of the static frame of functions: when it is
10989 topped by a particular function, stack checking is not reliable and
10990 a warning is issued by the compiler.
10993 Inefficiency: because of both the modified allocation strategy and the
10994 generic implementation, code performance is hampered.
10997 Note that old-style stack checking is also the fallback method for
10998 @samp{specific} if no target support has been added in the compiler.
11000 @item -fstack-limit-register=@var{reg}
11001 @itemx -fstack-limit-symbol=@var{sym}
11002 @itemx -fno-stack-limit
11003 @opindex fstack-limit-register
11004 @opindex fstack-limit-symbol
11005 @opindex fno-stack-limit
11006 Generate code to ensure that the stack does not grow beyond a certain value,
11007 either the value of a register or the address of a symbol. If a larger
11008 stack is required, a signal is raised at run time. For most targets,
11009 the signal is raised before the stack overruns the boundary, so
11010 it is possible to catch the signal without taking special precautions.
11012 For instance, if the stack starts at absolute address @samp{0x80000000}
11013 and grows downwards, you can use the flags
11014 @option{-fstack-limit-symbol=__stack_limit} and
11015 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
11016 of 128KB@. Note that this may only work with the GNU linker.
11018 You can locally override stack limit checking by using the
11019 @code{no_stack_limit} function attribute (@pxref{Function Attributes}).
11021 @item -fsplit-stack
11022 @opindex fsplit-stack
11023 Generate code to automatically split the stack before it overflows.
11024 The resulting program has a discontiguous stack which can only
11025 overflow if the program is unable to allocate any more memory. This
11026 is most useful when running threaded programs, as it is no longer
11027 necessary to calculate a good stack size to use for each thread. This
11028 is currently only implemented for the x86 targets running
11031 When code compiled with @option{-fsplit-stack} calls code compiled
11032 without @option{-fsplit-stack}, there may not be much stack space
11033 available for the latter code to run. If compiling all code,
11034 including library code, with @option{-fsplit-stack} is not an option,
11035 then the linker can fix up these calls so that the code compiled
11036 without @option{-fsplit-stack} always has a large stack. Support for
11037 this is implemented in the gold linker in GNU binutils release 2.21
11040 @item -fvtable-verify=@r{[}std@r{|}preinit@r{|}none@r{]}
11041 @opindex fvtable-verify
11042 This option is only available when compiling C++ code.
11043 It turns on (or off, if using @option{-fvtable-verify=none}) the security
11044 feature that verifies at run time, for every virtual call, that
11045 the vtable pointer through which the call is made is valid for the type of
11046 the object, and has not been corrupted or overwritten. If an invalid vtable
11047 pointer is detected at run time, an error is reported and execution of the
11048 program is immediately halted.
11050 This option causes run-time data structures to be built at program startup,
11051 which are used for verifying the vtable pointers.
11052 The options @samp{std} and @samp{preinit}
11053 control the timing of when these data structures are built. In both cases the
11054 data structures are built before execution reaches @code{main}. Using
11055 @option{-fvtable-verify=std} causes the data structures to be built after
11056 shared libraries have been loaded and initialized.
11057 @option{-fvtable-verify=preinit} causes them to be built before shared
11058 libraries have been loaded and initialized.
11060 If this option appears multiple times in the command line with different
11061 values specified, @samp{none} takes highest priority over both @samp{std} and
11062 @samp{preinit}; @samp{preinit} takes priority over @samp{std}.
11065 @opindex fvtv-debug
11066 When used in conjunction with @option{-fvtable-verify=std} or
11067 @option{-fvtable-verify=preinit}, causes debug versions of the
11068 runtime functions for the vtable verification feature to be called.
11069 This flag also causes the compiler to log information about which
11070 vtable pointers it finds for each class.
11071 This information is written to a file named @file{vtv_set_ptr_data.log}
11072 in the directory named by the environment variable @env{VTV_LOGS_DIR}
11073 if that is defined or the current working directory otherwise.
11075 Note: This feature @emph{appends} data to the log file. If you want a fresh log
11076 file, be sure to delete any existing one.
11079 @opindex fvtv-counts
11080 This is a debugging flag. When used in conjunction with
11081 @option{-fvtable-verify=std} or @option{-fvtable-verify=preinit}, this
11082 causes the compiler to keep track of the total number of virtual calls
11083 it encounters and the number of verifications it inserts. It also
11084 counts the number of calls to certain run-time library functions
11085 that it inserts and logs this information for each compilation unit.
11086 The compiler writes this information to a file named
11087 @file{vtv_count_data.log} in the directory named by the environment
11088 variable @env{VTV_LOGS_DIR} if that is defined or the current working
11089 directory otherwise. It also counts the size of the vtable pointer sets
11090 for each class, and writes this information to @file{vtv_class_set_sizes.log}
11091 in the same directory.
11093 Note: This feature @emph{appends} data to the log files. To get fresh log
11094 files, be sure to delete any existing ones.
11096 @item -finstrument-functions
11097 @opindex finstrument-functions
11098 Generate instrumentation calls for entry and exit to functions. Just
11099 after function entry and just before function exit, the following
11100 profiling functions are called with the address of the current
11101 function and its call site. (On some platforms,
11102 @code{__builtin_return_address} does not work beyond the current
11103 function, so the call site information may not be available to the
11104 profiling functions otherwise.)
11107 void __cyg_profile_func_enter (void *this_fn,
11109 void __cyg_profile_func_exit (void *this_fn,
11113 The first argument is the address of the start of the current function,
11114 which may be looked up exactly in the symbol table.
11116 This instrumentation is also done for functions expanded inline in other
11117 functions. The profiling calls indicate where, conceptually, the
11118 inline function is entered and exited. This means that addressable
11119 versions of such functions must be available. If all your uses of a
11120 function are expanded inline, this may mean an additional expansion of
11121 code size. If you use @code{extern inline} in your C code, an
11122 addressable version of such functions must be provided. (This is
11123 normally the case anyway, but if you get lucky and the optimizer always
11124 expands the functions inline, you might have gotten away without
11125 providing static copies.)
11127 A function may be given the attribute @code{no_instrument_function}, in
11128 which case this instrumentation is not done. This can be used, for
11129 example, for the profiling functions listed above, high-priority
11130 interrupt routines, and any functions from which the profiling functions
11131 cannot safely be called (perhaps signal handlers, if the profiling
11132 routines generate output or allocate memory).
11134 @item -finstrument-functions-exclude-file-list=@var{file},@var{file},@dots{}
11135 @opindex finstrument-functions-exclude-file-list
11137 Set the list of functions that are excluded from instrumentation (see
11138 the description of @option{-finstrument-functions}). If the file that
11139 contains a function definition matches with one of @var{file}, then
11140 that function is not instrumented. The match is done on substrings:
11141 if the @var{file} parameter is a substring of the file name, it is
11142 considered to be a match.
11147 -finstrument-functions-exclude-file-list=/bits/stl,include/sys
11151 excludes any inline function defined in files whose pathnames
11152 contain @file{/bits/stl} or @file{include/sys}.
11154 If, for some reason, you want to include letter @samp{,} in one of
11155 @var{sym}, write @samp{\,}. For example,
11156 @option{-finstrument-functions-exclude-file-list='\,\,tmp'}
11157 (note the single quote surrounding the option).
11159 @item -finstrument-functions-exclude-function-list=@var{sym},@var{sym},@dots{}
11160 @opindex finstrument-functions-exclude-function-list
11162 This is similar to @option{-finstrument-functions-exclude-file-list},
11163 but this option sets the list of function names to be excluded from
11164 instrumentation. The function name to be matched is its user-visible
11165 name, such as @code{vector<int> blah(const vector<int> &)}, not the
11166 internal mangled name (e.g., @code{_Z4blahRSt6vectorIiSaIiEE}). The
11167 match is done on substrings: if the @var{sym} parameter is a substring
11168 of the function name, it is considered to be a match. For C99 and C++
11169 extended identifiers, the function name must be given in UTF-8, not
11170 using universal character names.
11175 @node Preprocessor Options
11176 @section Options Controlling the Preprocessor
11177 @cindex preprocessor options
11178 @cindex options, preprocessor
11180 These options control the C preprocessor, which is run on each C source
11181 file before actual compilation.
11183 If you use the @option{-E} option, nothing is done except preprocessing.
11184 Some of these options make sense only together with @option{-E} because
11185 they cause the preprocessor output to be unsuitable for actual
11189 @item -Wp,@var{option}
11191 You can use @option{-Wp,@var{option}} to bypass the compiler driver
11192 and pass @var{option} directly through to the preprocessor. If
11193 @var{option} contains commas, it is split into multiple options at the
11194 commas. However, many options are modified, translated or interpreted
11195 by the compiler driver before being passed to the preprocessor, and
11196 @option{-Wp} forcibly bypasses this phase. The preprocessor's direct
11197 interface is undocumented and subject to change, so whenever possible
11198 you should avoid using @option{-Wp} and let the driver handle the
11201 @item -Xpreprocessor @var{option}
11202 @opindex Xpreprocessor
11203 Pass @var{option} as an option to the preprocessor. You can use this to
11204 supply system-specific preprocessor options that GCC does not
11207 If you want to pass an option that takes an argument, you must use
11208 @option{-Xpreprocessor} twice, once for the option and once for the argument.
11210 @item -no-integrated-cpp
11211 @opindex no-integrated-cpp
11212 Perform preprocessing as a separate pass before compilation.
11213 By default, GCC performs preprocessing as an integrated part of
11214 input tokenization and parsing.
11215 If this option is provided, the appropriate language front end
11216 (@command{cc1}, @command{cc1plus}, or @command{cc1obj} for C, C++,
11217 and Objective-C, respectively) is instead invoked twice,
11218 once for preprocessing only and once for actual compilation
11219 of the preprocessed input.
11220 This option may be useful in conjunction with the @option{-B} or
11221 @option{-wrapper} options to specify an alternate preprocessor or
11222 perform additional processing of the program source between
11223 normal preprocessing and compilation.
11226 @include cppopts.texi
11228 @node Assembler Options
11229 @section Passing Options to the Assembler
11231 @c prevent bad page break with this line
11232 You can pass options to the assembler.
11235 @item -Wa,@var{option}
11237 Pass @var{option} as an option to the assembler. If @var{option}
11238 contains commas, it is split into multiple options at the commas.
11240 @item -Xassembler @var{option}
11241 @opindex Xassembler
11242 Pass @var{option} as an option to the assembler. You can use this to
11243 supply system-specific assembler options that GCC does not
11246 If you want to pass an option that takes an argument, you must use
11247 @option{-Xassembler} twice, once for the option and once for the argument.
11252 @section Options for Linking
11253 @cindex link options
11254 @cindex options, linking
11256 These options come into play when the compiler links object files into
11257 an executable output file. They are meaningless if the compiler is
11258 not doing a link step.
11262 @item @var{object-file-name}
11263 A file name that does not end in a special recognized suffix is
11264 considered to name an object file or library. (Object files are
11265 distinguished from libraries by the linker according to the file
11266 contents.) If linking is done, these object files are used as input
11275 If any of these options is used, then the linker is not run, and
11276 object file names should not be used as arguments. @xref{Overall
11280 @opindex fuse-ld=bfd
11281 Use the @command{bfd} linker instead of the default linker.
11283 @item -fuse-ld=gold
11284 @opindex fuse-ld=gold
11285 Use the @command{gold} linker instead of the default linker.
11288 @item -l@var{library}
11289 @itemx -l @var{library}
11291 Search the library named @var{library} when linking. (The second
11292 alternative with the library as a separate argument is only for
11293 POSIX compliance and is not recommended.)
11295 It makes a difference where in the command you write this option; the
11296 linker searches and processes libraries and object files in the order they
11297 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
11298 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
11299 to functions in @samp{z}, those functions may not be loaded.
11301 The linker searches a standard list of directories for the library,
11302 which is actually a file named @file{lib@var{library}.a}. The linker
11303 then uses this file as if it had been specified precisely by name.
11305 The directories searched include several standard system directories
11306 plus any that you specify with @option{-L}.
11308 Normally the files found this way are library files---archive files
11309 whose members are object files. The linker handles an archive file by
11310 scanning through it for members which define symbols that have so far
11311 been referenced but not defined. But if the file that is found is an
11312 ordinary object file, it is linked in the usual fashion. The only
11313 difference between using an @option{-l} option and specifying a file name
11314 is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
11315 and searches several directories.
11319 You need this special case of the @option{-l} option in order to
11320 link an Objective-C or Objective-C++ program.
11322 @item -nostartfiles
11323 @opindex nostartfiles
11324 Do not use the standard system startup files when linking.
11325 The standard system libraries are used normally, unless @option{-nostdlib}
11326 or @option{-nodefaultlibs} is used.
11328 @item -nodefaultlibs
11329 @opindex nodefaultlibs
11330 Do not use the standard system libraries when linking.
11331 Only the libraries you specify are passed to the linker, and options
11332 specifying linkage of the system libraries, such as @option{-static-libgcc}
11333 or @option{-shared-libgcc}, are ignored.
11334 The standard startup files are used normally, unless @option{-nostartfiles}
11337 The compiler may generate calls to @code{memcmp},
11338 @code{memset}, @code{memcpy} and @code{memmove}.
11339 These entries are usually resolved by entries in
11340 libc. These entry points should be supplied through some other
11341 mechanism when this option is specified.
11345 Do not use the standard system startup files or libraries when linking.
11346 No startup files and only the libraries you specify are passed to
11347 the linker, and options specifying linkage of the system libraries, such as
11348 @option{-static-libgcc} or @option{-shared-libgcc}, are ignored.
11350 The compiler may generate calls to @code{memcmp}, @code{memset},
11351 @code{memcpy} and @code{memmove}.
11352 These entries are usually resolved by entries in
11353 libc. These entry points should be supplied through some other
11354 mechanism when this option is specified.
11356 @cindex @option{-lgcc}, use with @option{-nostdlib}
11357 @cindex @option{-nostdlib} and unresolved references
11358 @cindex unresolved references and @option{-nostdlib}
11359 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
11360 @cindex @option{-nodefaultlibs} and unresolved references
11361 @cindex unresolved references and @option{-nodefaultlibs}
11362 One of the standard libraries bypassed by @option{-nostdlib} and
11363 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
11364 which GCC uses to overcome shortcomings of particular machines, or special
11365 needs for some languages.
11366 (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler
11367 Collection (GCC) Internals},
11368 for more discussion of @file{libgcc.a}.)
11369 In most cases, you need @file{libgcc.a} even when you want to avoid
11370 other standard libraries. In other words, when you specify @option{-nostdlib}
11371 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
11372 This ensures that you have no unresolved references to internal GCC
11373 library subroutines.
11374 (An example of such an internal subroutine is @code{__main}, used to ensure C++
11375 constructors are called; @pxref{Collect2,,@code{collect2}, gccint,
11376 GNU Compiler Collection (GCC) Internals}.)
11380 Produce a position independent executable on targets that support it.
11381 For predictable results, you must also specify the same set of options
11382 used for compilation (@option{-fpie}, @option{-fPIE},
11383 or model suboptions) when you specify this linker option.
11387 Don't produce a position independent executable.
11391 Pass the flag @option{-export-dynamic} to the ELF linker, on targets
11392 that support it. This instructs the linker to add all symbols, not
11393 only used ones, to the dynamic symbol table. This option is needed
11394 for some uses of @code{dlopen} or to allow obtaining backtraces
11395 from within a program.
11399 Remove all symbol table and relocation information from the executable.
11403 On systems that support dynamic linking, this prevents linking with the shared
11404 libraries. On other systems, this option has no effect.
11408 Produce a shared object which can then be linked with other objects to
11409 form an executable. Not all systems support this option. For predictable
11410 results, you must also specify the same set of options used for compilation
11411 (@option{-fpic}, @option{-fPIC}, or model suboptions) when
11412 you specify this linker option.@footnote{On some systems, @samp{gcc -shared}
11413 needs to build supplementary stub code for constructors to work. On
11414 multi-libbed systems, @samp{gcc -shared} must select the correct support
11415 libraries to link against. Failing to supply the correct flags may lead
11416 to subtle defects. Supplying them in cases where they are not necessary
11419 @item -shared-libgcc
11420 @itemx -static-libgcc
11421 @opindex shared-libgcc
11422 @opindex static-libgcc
11423 On systems that provide @file{libgcc} as a shared library, these options
11424 force the use of either the shared or static version, respectively.
11425 If no shared version of @file{libgcc} was built when the compiler was
11426 configured, these options have no effect.
11428 There are several situations in which an application should use the
11429 shared @file{libgcc} instead of the static version. The most common
11430 of these is when the application wishes to throw and catch exceptions
11431 across different shared libraries. In that case, each of the libraries
11432 as well as the application itself should use the shared @file{libgcc}.
11434 Therefore, the G++ and GCJ drivers automatically add
11435 @option{-shared-libgcc} whenever you build a shared library or a main
11436 executable, because C++ and Java programs typically use exceptions, so
11437 this is the right thing to do.
11439 If, instead, you use the GCC driver to create shared libraries, you may
11440 find that they are not always linked with the shared @file{libgcc}.
11441 If GCC finds, at its configuration time, that you have a non-GNU linker
11442 or a GNU linker that does not support option @option{--eh-frame-hdr},
11443 it links the shared version of @file{libgcc} into shared libraries
11444 by default. Otherwise, it takes advantage of the linker and optimizes
11445 away the linking with the shared version of @file{libgcc}, linking with
11446 the static version of libgcc by default. This allows exceptions to
11447 propagate through such shared libraries, without incurring relocation
11448 costs at library load time.
11450 However, if a library or main executable is supposed to throw or catch
11451 exceptions, you must link it using the G++ or GCJ driver, as appropriate
11452 for the languages used in the program, or using the option
11453 @option{-shared-libgcc}, such that it is linked with the shared
11456 @item -static-libasan
11457 @opindex static-libasan
11458 When the @option{-fsanitize=address} option is used to link a program,
11459 the GCC driver automatically links against @option{libasan}. If
11460 @file{libasan} is available as a shared library, and the @option{-static}
11461 option is not used, then this links against the shared version of
11462 @file{libasan}. The @option{-static-libasan} option directs the GCC
11463 driver to link @file{libasan} statically, without necessarily linking
11464 other libraries statically.
11466 @item -static-libtsan
11467 @opindex static-libtsan
11468 When the @option{-fsanitize=thread} option is used to link a program,
11469 the GCC driver automatically links against @option{libtsan}. If
11470 @file{libtsan} is available as a shared library, and the @option{-static}
11471 option is not used, then this links against the shared version of
11472 @file{libtsan}. The @option{-static-libtsan} option directs the GCC
11473 driver to link @file{libtsan} statically, without necessarily linking
11474 other libraries statically.
11476 @item -static-liblsan
11477 @opindex static-liblsan
11478 When the @option{-fsanitize=leak} option is used to link a program,
11479 the GCC driver automatically links against @option{liblsan}. If
11480 @file{liblsan} is available as a shared library, and the @option{-static}
11481 option is not used, then this links against the shared version of
11482 @file{liblsan}. The @option{-static-liblsan} option directs the GCC
11483 driver to link @file{liblsan} statically, without necessarily linking
11484 other libraries statically.
11486 @item -static-libubsan
11487 @opindex static-libubsan
11488 When the @option{-fsanitize=undefined} option is used to link a program,
11489 the GCC driver automatically links against @option{libubsan}. If
11490 @file{libubsan} is available as a shared library, and the @option{-static}
11491 option is not used, then this links against the shared version of
11492 @file{libubsan}. The @option{-static-libubsan} option directs the GCC
11493 driver to link @file{libubsan} statically, without necessarily linking
11494 other libraries statically.
11496 @item -static-libmpx
11497 @opindex static-libmpx
11498 When the @option{-fcheck-pointer bounds} and @option{-mmpx} options are
11499 used to link a program, the GCC driver automatically links against
11500 @file{libmpx}. If @file{libmpx} is available as a shared library,
11501 and the @option{-static} option is not used, then this links against
11502 the shared version of @file{libmpx}. The @option{-static-libmpx}
11503 option directs the GCC driver to link @file{libmpx} statically,
11504 without necessarily linking other libraries statically.
11506 @item -static-libmpxwrappers
11507 @opindex static-libmpxwrappers
11508 When the @option{-fcheck-pointer bounds} and @option{-mmpx} options are used
11509 to link a program without also using @option{-fno-chkp-use-wrappers}, the
11510 GCC driver automatically links against @file{libmpxwrappers}. If
11511 @file{libmpxwrappers} is available as a shared library, and the
11512 @option{-static} option is not used, then this links against the shared
11513 version of @file{libmpxwrappers}. The @option{-static-libmpxwrappers}
11514 option directs the GCC driver to link @file{libmpxwrappers} statically,
11515 without necessarily linking other libraries statically.
11517 @item -static-libstdc++
11518 @opindex static-libstdc++
11519 When the @command{g++} program is used to link a C++ program, it
11520 normally automatically links against @option{libstdc++}. If
11521 @file{libstdc++} is available as a shared library, and the
11522 @option{-static} option is not used, then this links against the
11523 shared version of @file{libstdc++}. That is normally fine. However, it
11524 is sometimes useful to freeze the version of @file{libstdc++} used by
11525 the program without going all the way to a fully static link. The
11526 @option{-static-libstdc++} option directs the @command{g++} driver to
11527 link @file{libstdc++} statically, without necessarily linking other
11528 libraries statically.
11532 Bind references to global symbols when building a shared object. Warn
11533 about any unresolved references (unless overridden by the link editor
11534 option @option{-Xlinker -z -Xlinker defs}). Only a few systems support
11537 @item -T @var{script}
11539 @cindex linker script
11540 Use @var{script} as the linker script. This option is supported by most
11541 systems using the GNU linker. On some targets, such as bare-board
11542 targets without an operating system, the @option{-T} option may be required
11543 when linking to avoid references to undefined symbols.
11545 @item -Xlinker @var{option}
11547 Pass @var{option} as an option to the linker. You can use this to
11548 supply system-specific linker options that GCC does not recognize.
11550 If you want to pass an option that takes a separate argument, you must use
11551 @option{-Xlinker} twice, once for the option and once for the argument.
11552 For example, to pass @option{-assert definitions}, you must write
11553 @option{-Xlinker -assert -Xlinker definitions}. It does not work to write
11554 @option{-Xlinker "-assert definitions"}, because this passes the entire
11555 string as a single argument, which is not what the linker expects.
11557 When using the GNU linker, it is usually more convenient to pass
11558 arguments to linker options using the @option{@var{option}=@var{value}}
11559 syntax than as separate arguments. For example, you can specify
11560 @option{-Xlinker -Map=output.map} rather than
11561 @option{-Xlinker -Map -Xlinker output.map}. Other linkers may not support
11562 this syntax for command-line options.
11564 @item -Wl,@var{option}
11566 Pass @var{option} as an option to the linker. If @var{option} contains
11567 commas, it is split into multiple options at the commas. You can use this
11568 syntax to pass an argument to the option.
11569 For example, @option{-Wl,-Map,output.map} passes @option{-Map output.map} to the
11570 linker. When using the GNU linker, you can also get the same effect with
11571 @option{-Wl,-Map=output.map}.
11573 @item -u @var{symbol}
11575 Pretend the symbol @var{symbol} is undefined, to force linking of
11576 library modules to define it. You can use @option{-u} multiple times with
11577 different symbols to force loading of additional library modules.
11579 @item -z @var{keyword}
11581 @option{-z} is passed directly on to the linker along with the keyword
11582 @var{keyword}. See the section in the documentation of your linker for
11583 permitted values and their meanings.
11586 @node Directory Options
11587 @section Options for Directory Search
11588 @cindex directory options
11589 @cindex options, directory search
11590 @cindex search path
11592 These options specify directories to search for header files, for
11593 libraries and for parts of the compiler:
11598 Add the directory @var{dir} to the head of the list of directories to be
11599 searched for header files. This can be used to override a system header
11600 file, substituting your own version, since these directories are
11601 searched before the system header file directories. However, you should
11602 not use this option to add directories that contain vendor-supplied
11603 system header files (use @option{-isystem} for that). If you use more than
11604 one @option{-I} option, the directories are scanned in left-to-right
11605 order; the standard system directories come after.
11607 If a standard system include directory, or a directory specified with
11608 @option{-isystem}, is also specified with @option{-I}, the @option{-I}
11609 option is ignored. The directory is still searched but as a
11610 system directory at its normal position in the system include chain.
11611 This is to ensure that GCC's procedure to fix buggy system headers and
11612 the ordering for the @code{include_next} directive are not inadvertently changed.
11613 If you really need to change the search order for system directories,
11614 use the @option{-nostdinc} and/or @option{-isystem} options.
11616 @item -iplugindir=@var{dir}
11617 @opindex iplugindir=
11618 Set the directory to search for plugins that are passed
11619 by @option{-fplugin=@var{name}} instead of
11620 @option{-fplugin=@var{path}/@var{name}.so}. This option is not meant
11621 to be used by the user, but only passed by the driver.
11623 @item -iquote@var{dir}
11625 Add the directory @var{dir} to the head of the list of directories to
11626 be searched for header files only for the case of @code{#include
11627 "@var{file}"}; they are not searched for @code{#include <@var{file}>},
11628 otherwise just like @option{-I}.
11632 Add directory @var{dir} to the list of directories to be searched
11635 @item -B@var{prefix}
11637 This option specifies where to find the executables, libraries,
11638 include files, and data files of the compiler itself.
11640 The compiler driver program runs one or more of the subprograms
11641 @command{cpp}, @command{cc1}, @command{as} and @command{ld}. It tries
11642 @var{prefix} as a prefix for each program it tries to run, both with and
11643 without @samp{@var{machine}/@var{version}/} for the corresponding target
11644 machine and compiler version.
11646 For each subprogram to be run, the compiler driver first tries the
11647 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
11648 is not specified, the driver tries two standard prefixes,
11649 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc/}. If neither of
11650 those results in a file name that is found, the unmodified program
11651 name is searched for using the directories specified in your
11652 @env{PATH} environment variable.
11654 The compiler checks to see if the path provided by @option{-B}
11655 refers to a directory, and if necessary it adds a directory
11656 separator character at the end of the path.
11658 @option{-B} prefixes that effectively specify directory names also apply
11659 to libraries in the linker, because the compiler translates these
11660 options into @option{-L} options for the linker. They also apply to
11661 include files in the preprocessor, because the compiler translates these
11662 options into @option{-isystem} options for the preprocessor. In this case,
11663 the compiler appends @samp{include} to the prefix.
11665 The runtime support file @file{libgcc.a} can also be searched for using
11666 the @option{-B} prefix, if needed. If it is not found there, the two
11667 standard prefixes above are tried, and that is all. The file is left
11668 out of the link if it is not found by those means.
11670 Another way to specify a prefix much like the @option{-B} prefix is to use
11671 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
11674 As a special kludge, if the path provided by @option{-B} is
11675 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
11676 9, then it is replaced by @file{[dir/]include}. This is to help
11677 with boot-strapping the compiler.
11679 @item -no-canonical-prefixes
11680 @opindex no-canonical-prefixes
11681 Do not expand any symbolic links, resolve references to @samp{/../}
11682 or @samp{/./}, or make the path absolute when generating a relative
11685 @item --sysroot=@var{dir}
11687 Use @var{dir} as the logical root directory for headers and libraries.
11688 For example, if the compiler normally searches for headers in
11689 @file{/usr/include} and libraries in @file{/usr/lib}, it instead
11690 searches @file{@var{dir}/usr/include} and @file{@var{dir}/usr/lib}.
11692 If you use both this option and the @option{-isysroot} option, then
11693 the @option{--sysroot} option applies to libraries, but the
11694 @option{-isysroot} option applies to header files.
11696 The GNU linker (beginning with version 2.16) has the necessary support
11697 for this option. If your linker does not support this option, the
11698 header file aspect of @option{--sysroot} still works, but the
11699 library aspect does not.
11701 @item --no-sysroot-suffix
11702 @opindex no-sysroot-suffix
11703 For some targets, a suffix is added to the root directory specified
11704 with @option{--sysroot}, depending on the other options used, so that
11705 headers may for example be found in
11706 @file{@var{dir}/@var{suffix}/usr/include} instead of
11707 @file{@var{dir}/usr/include}. This option disables the addition of
11712 This option has been deprecated. Please use @option{-iquote} instead for
11713 @option{-I} directories before the @option{-I-} and remove the @option{-I-}
11715 Any directories you specify with @option{-I} options before the @option{-I-}
11716 option are searched only for the case of @code{#include "@var{file}"};
11717 they are not searched for @code{#include <@var{file}>}.
11719 If additional directories are specified with @option{-I} options after
11720 the @option{-I-} option, these directories are searched for all @code{#include}
11721 directives. (Ordinarily @emph{all} @option{-I} directories are used
11724 In addition, the @option{-I-} option inhibits the use of the current
11725 directory (where the current input file came from) as the first search
11726 directory for @code{#include "@var{file}"}. There is no way to
11727 override this effect of @option{-I-}. With @option{-I.} you can specify
11728 searching the directory that is current when the compiler is
11729 invoked. That is not exactly the same as what the preprocessor does
11730 by default, but it is often satisfactory.
11732 @option{-I-} does not inhibit the use of the standard system directories
11733 for header files. Thus, @option{-I-} and @option{-nostdinc} are
11737 @node Code Gen Options
11738 @section Options for Code Generation Conventions
11739 @cindex code generation conventions
11740 @cindex options, code generation
11741 @cindex run-time options
11743 These machine-independent options control the interface conventions
11744 used in code generation.
11746 Most of them have both positive and negative forms; the negative form
11747 of @option{-ffoo} is @option{-fno-foo}. In the table below, only
11748 one of the forms is listed---the one that is not the default. You
11749 can figure out the other form by either removing @samp{no-} or adding
11753 @item -fstack-reuse=@var{reuse-level}
11754 @opindex fstack_reuse
11755 This option controls stack space reuse for user declared local/auto variables
11756 and compiler generated temporaries. @var{reuse_level} can be @samp{all},
11757 @samp{named_vars}, or @samp{none}. @samp{all} enables stack reuse for all
11758 local variables and temporaries, @samp{named_vars} enables the reuse only for
11759 user defined local variables with names, and @samp{none} disables stack reuse
11760 completely. The default value is @samp{all}. The option is needed when the
11761 program extends the lifetime of a scoped local variable or a compiler generated
11762 temporary beyond the end point defined by the language. When a lifetime of
11763 a variable ends, and if the variable lives in memory, the optimizing compiler
11764 has the freedom to reuse its stack space with other temporaries or scoped
11765 local variables whose live range does not overlap with it. Legacy code extending
11766 local lifetime is likely to break with the stack reuse optimization.
11785 if (*p == 10) // out of scope use of local1
11796 A(int k) : i(k), j(k) @{ @}
11803 void foo(const A& ar)
11810 foo(A(10)); // temp object's lifetime ends when foo returns
11816 ap->i+= 10; // ap references out of scope temp whose space
11817 // is reused with a. What is the value of ap->i?
11822 The lifetime of a compiler generated temporary is well defined by the C++
11823 standard. When a lifetime of a temporary ends, and if the temporary lives
11824 in memory, the optimizing compiler has the freedom to reuse its stack
11825 space with other temporaries or scoped local variables whose live range
11826 does not overlap with it. However some of the legacy code relies on
11827 the behavior of older compilers in which temporaries' stack space is
11828 not reused, the aggressive stack reuse can lead to runtime errors. This
11829 option is used to control the temporary stack reuse optimization.
11833 This option generates traps for signed overflow on addition, subtraction,
11834 multiplication operations.
11835 The options @option{-ftrapv} and @option{-fwrapv} override each other, so using
11836 @option{-ftrapv} @option{-fwrapv} on the command-line results in
11837 @option{-fwrapv} being effective. Note that only active options override, so
11838 using @option{-ftrapv} @option{-fwrapv} @option{-fno-wrapv} on the command-line
11839 results in @option{-ftrapv} being effective.
11843 This option instructs the compiler to assume that signed arithmetic
11844 overflow of addition, subtraction and multiplication wraps around
11845 using twos-complement representation. This flag enables some optimizations
11846 and disables others. This option is enabled by default for the Java
11847 front end, as required by the Java language specification.
11848 The options @option{-ftrapv} and @option{-fwrapv} override each other, so using
11849 @option{-ftrapv} @option{-fwrapv} on the command-line results in
11850 @option{-fwrapv} being effective. Note that only active options override, so
11851 using @option{-ftrapv} @option{-fwrapv} @option{-fno-wrapv} on the command-line
11852 results in @option{-ftrapv} being effective.
11855 @opindex fexceptions
11856 Enable exception handling. Generates extra code needed to propagate
11857 exceptions. For some targets, this implies GCC generates frame
11858 unwind information for all functions, which can produce significant data
11859 size overhead, although it does not affect execution. If you do not
11860 specify this option, GCC enables it by default for languages like
11861 C++ that normally require exception handling, and disables it for
11862 languages like C that do not normally require it. However, you may need
11863 to enable this option when compiling C code that needs to interoperate
11864 properly with exception handlers written in C++. You may also wish to
11865 disable this option if you are compiling older C++ programs that don't
11866 use exception handling.
11868 @item -fnon-call-exceptions
11869 @opindex fnon-call-exceptions
11870 Generate code that allows trapping instructions to throw exceptions.
11871 Note that this requires platform-specific runtime support that does
11872 not exist everywhere. Moreover, it only allows @emph{trapping}
11873 instructions to throw exceptions, i.e.@: memory references or floating-point
11874 instructions. It does not allow exceptions to be thrown from
11875 arbitrary signal handlers such as @code{SIGALRM}.
11877 @item -fdelete-dead-exceptions
11878 @opindex fdelete-dead-exceptions
11879 Consider that instructions that may throw exceptions but don't otherwise
11880 contribute to the execution of the program can be optimized away.
11881 This option is enabled by default for the Ada front end, as permitted by
11882 the Ada language specification.
11883 Optimization passes that cause dead exceptions to be removed are enabled independently at different optimization levels.
11885 @item -funwind-tables
11886 @opindex funwind-tables
11887 Similar to @option{-fexceptions}, except that it just generates any needed
11888 static data, but does not affect the generated code in any other way.
11889 You normally do not need to enable this option; instead, a language processor
11890 that needs this handling enables it on your behalf.
11892 @item -fasynchronous-unwind-tables
11893 @opindex fasynchronous-unwind-tables
11894 Generate unwind table in DWARF format, if supported by target machine. The
11895 table is exact at each instruction boundary, so it can be used for stack
11896 unwinding from asynchronous events (such as debugger or garbage collector).
11898 @item -fno-gnu-unique
11899 @opindex fno-gnu-unique
11900 On systems with recent GNU assembler and C library, the C++ compiler
11901 uses the @code{STB_GNU_UNIQUE} binding to make sure that definitions
11902 of template static data members and static local variables in inline
11903 functions are unique even in the presence of @code{RTLD_LOCAL}; this
11904 is necessary to avoid problems with a library used by two different
11905 @code{RTLD_LOCAL} plugins depending on a definition in one of them and
11906 therefore disagreeing with the other one about the binding of the
11907 symbol. But this causes @code{dlclose} to be ignored for affected
11908 DSOs; if your program relies on reinitialization of a DSO via
11909 @code{dlclose} and @code{dlopen}, you can use
11910 @option{-fno-gnu-unique}.
11912 @item -fpcc-struct-return
11913 @opindex fpcc-struct-return
11914 Return ``short'' @code{struct} and @code{union} values in memory like
11915 longer ones, rather than in registers. This convention is less
11916 efficient, but it has the advantage of allowing intercallability between
11917 GCC-compiled files and files compiled with other compilers, particularly
11918 the Portable C Compiler (pcc).
11920 The precise convention for returning structures in memory depends
11921 on the target configuration macros.
11923 Short structures and unions are those whose size and alignment match
11924 that of some integer type.
11926 @strong{Warning:} code compiled with the @option{-fpcc-struct-return}
11927 switch is not binary compatible with code compiled with the
11928 @option{-freg-struct-return} switch.
11929 Use it to conform to a non-default application binary interface.
11931 @item -freg-struct-return
11932 @opindex freg-struct-return
11933 Return @code{struct} and @code{union} values in registers when possible.
11934 This is more efficient for small structures than
11935 @option{-fpcc-struct-return}.
11937 If you specify neither @option{-fpcc-struct-return} nor
11938 @option{-freg-struct-return}, GCC defaults to whichever convention is
11939 standard for the target. If there is no standard convention, GCC
11940 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
11941 the principal compiler. In those cases, we can choose the standard, and
11942 we chose the more efficient register return alternative.
11944 @strong{Warning:} code compiled with the @option{-freg-struct-return}
11945 switch is not binary compatible with code compiled with the
11946 @option{-fpcc-struct-return} switch.
11947 Use it to conform to a non-default application binary interface.
11949 @item -fshort-enums
11950 @opindex fshort-enums
11951 Allocate to an @code{enum} type only as many bytes as it needs for the
11952 declared range of possible values. Specifically, the @code{enum} type
11953 is equivalent to the smallest integer type that has enough room.
11955 @strong{Warning:} the @option{-fshort-enums} switch causes GCC to generate
11956 code that is not binary compatible with code generated without that switch.
11957 Use it to conform to a non-default application binary interface.
11959 @item -fshort-wchar
11960 @opindex fshort-wchar
11961 Override the underlying type for @code{wchar_t} to be @code{short
11962 unsigned int} instead of the default for the target. This option is
11963 useful for building programs to run under WINE@.
11965 @strong{Warning:} the @option{-fshort-wchar} switch causes GCC to generate
11966 code that is not binary compatible with code generated without that switch.
11967 Use it to conform to a non-default application binary interface.
11970 @opindex fno-common
11971 @cindex tentative definitions
11972 In C code, this option controls the placement of global variables
11973 defined without an initializer, known as @dfn{tentative definitions}
11974 in the C standard. Tentative definitions are distinct from declarations
11975 of a variable with the @code{extern} keyword, which do not allocate storage.
11977 Unix C compilers have traditionally allocated storage for
11978 uninitialized global variables in a common block. This allows the
11979 linker to resolve all tentative definitions of the same variable
11980 in different compilation units to the same object, or to a non-tentative
11982 This is the behavior specified by @option{-fcommon}, and is the default for
11983 GCC on most targets.
11984 On the other hand, this behavior is not required by ISO
11985 C, and on some targets may carry a speed or code size penalty on
11986 variable references.
11988 The @option{-fno-common} option specifies that the compiler should instead
11989 place uninitialized global variables in the data section of the object file.
11990 This inhibits the merging of tentative definitions by the linker so
11991 you get a multiple-definition error if the same
11992 variable is defined in more than one compilation unit.
11993 Compiling with @option{-fno-common} is useful on targets for which
11994 it provides better performance, or if you wish to verify that the
11995 program will work on other systems that always treat uninitialized
11996 variable definitions this way.
12000 Ignore the @code{#ident} directive.
12002 @item -finhibit-size-directive
12003 @opindex finhibit-size-directive
12004 Don't output a @code{.size} assembler directive, or anything else that
12005 would cause trouble if the function is split in the middle, and the
12006 two halves are placed at locations far apart in memory. This option is
12007 used when compiling @file{crtstuff.c}; you should not need to use it
12010 @item -fverbose-asm
12011 @opindex fverbose-asm
12012 Put extra commentary information in the generated assembly code to
12013 make it more readable. This option is generally only of use to those
12014 who actually need to read the generated assembly code (perhaps while
12015 debugging the compiler itself).
12017 @option{-fno-verbose-asm}, the default, causes the
12018 extra information to be omitted and is useful when comparing two assembler
12021 The added comments include:
12026 information on the compiler version and command-line options,
12029 the source code lines associated with the assembly instructions,
12030 in the form FILENAME:LINENUMBER:CONTENT OF LINE,
12033 hints on which high-level expressions correspond to
12034 the various assembly instruction operands.
12038 For example, given this C source file:
12046 for (i = 0; i < n; i++)
12053 compiling to (x86_64) assembly via @option{-S} and emitting the result
12054 direct to stdout via @option{-o} @option{-}
12057 gcc -S test.c -fverbose-asm -Os -o -
12060 gives output similar to this:
12064 # GNU C11 (GCC) version 7.0.0 20160809 (experimental) (x86_64-pc-linux-gnu)
12071 .type test, @@function
12075 # test.c:4: int total = 0;
12076 xorl %eax, %eax # <retval>
12077 # test.c:6: for (i = 0; i < n; i++)
12078 xorl %edx, %edx # i
12080 # test.c:6: for (i = 0; i < n; i++)
12081 cmpl %edi, %edx # n, i
12083 # test.c:7: total += i * i;
12084 movl %edx, %ecx # i, tmp92
12085 imull %edx, %ecx # i, tmp92
12086 # test.c:6: for (i = 0; i < n; i++)
12088 # test.c:7: total += i * i;
12089 addl %ecx, %eax # tmp92, <retval>
12097 .ident "GCC: (GNU) 7.0.0 20160809 (experimental)"
12098 .section .note.GNU-stack,"",@@progbits
12101 The comments are intended for humans rather than machines and hence the
12102 precise format of the comments is subject to change.
12104 @item -frecord-gcc-switches
12105 @opindex frecord-gcc-switches
12106 This switch causes the command line used to invoke the
12107 compiler to be recorded into the object file that is being created.
12108 This switch is only implemented on some targets and the exact format
12109 of the recording is target and binary file format dependent, but it
12110 usually takes the form of a section containing ASCII text. This
12111 switch is related to the @option{-fverbose-asm} switch, but that
12112 switch only records information in the assembler output file as
12113 comments, so it never reaches the object file.
12114 See also @option{-grecord-gcc-switches} for another
12115 way of storing compiler options into the object file.
12119 @cindex global offset table
12121 Generate position-independent code (PIC) suitable for use in a shared
12122 library, if supported for the target machine. Such code accesses all
12123 constant addresses through a global offset table (GOT)@. The dynamic
12124 loader resolves the GOT entries when the program starts (the dynamic
12125 loader is not part of GCC; it is part of the operating system). If
12126 the GOT size for the linked executable exceeds a machine-specific
12127 maximum size, you get an error message from the linker indicating that
12128 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
12129 instead. (These maximums are 8k on the SPARC, 28k on AArch64 and 32k
12130 on the m68k and RS/6000. The x86 has no such limit.)
12132 Position-independent code requires special support, and therefore works
12133 only on certain machines. For the x86, GCC supports PIC for System V
12134 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
12135 position-independent.
12137 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
12142 If supported for the target machine, emit position-independent code,
12143 suitable for dynamic linking and avoiding any limit on the size of the
12144 global offset table. This option makes a difference on AArch64, m68k,
12145 PowerPC and SPARC@.
12147 Position-independent code requires special support, and therefore works
12148 only on certain machines.
12150 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
12157 These options are similar to @option{-fpic} and @option{-fPIC}, but
12158 generated position independent code can be only linked into executables.
12159 Usually these options are used when @option{-pie} GCC option is
12160 used during linking.
12162 @option{-fpie} and @option{-fPIE} both define the macros
12163 @code{__pie__} and @code{__PIE__}. The macros have the value 1
12164 for @option{-fpie} and 2 for @option{-fPIE}.
12168 Do not use the PLT for external function calls in position-independent code.
12169 Instead, load the callee address at call sites from the GOT and branch to it.
12170 This leads to more efficient code by eliminating PLT stubs and exposing
12171 GOT loads to optimizations. On architectures such as 32-bit x86 where
12172 PLT stubs expect the GOT pointer in a specific register, this gives more
12173 register allocation freedom to the compiler.
12174 Lazy binding requires use of the PLT;
12175 with @option{-fno-plt} all external symbols are resolved at load time.
12177 Alternatively, the function attribute @code{noplt} can be used to avoid calls
12178 through the PLT for specific external functions.
12180 In position-dependent code, a few targets also convert calls to
12181 functions that are marked to not use the PLT to use the GOT instead.
12183 @item -fno-jump-tables
12184 @opindex fno-jump-tables
12185 Do not use jump tables for switch statements even where it would be
12186 more efficient than other code generation strategies. This option is
12187 of use in conjunction with @option{-fpic} or @option{-fPIC} for
12188 building code that forms part of a dynamic linker and cannot
12189 reference the address of a jump table. On some targets, jump tables
12190 do not require a GOT and this option is not needed.
12192 @item -ffixed-@var{reg}
12194 Treat the register named @var{reg} as a fixed register; generated code
12195 should never refer to it (except perhaps as a stack pointer, frame
12196 pointer or in some other fixed role).
12198 @var{reg} must be the name of a register. The register names accepted
12199 are machine-specific and are defined in the @code{REGISTER_NAMES}
12200 macro in the machine description macro file.
12202 This flag does not have a negative form, because it specifies a
12205 @item -fcall-used-@var{reg}
12206 @opindex fcall-used
12207 Treat the register named @var{reg} as an allocable register that is
12208 clobbered by function calls. It may be allocated for temporaries or
12209 variables that do not live across a call. Functions compiled this way
12210 do not save and restore the register @var{reg}.
12212 It is an error to use this flag with the frame pointer or stack pointer.
12213 Use of this flag for other registers that have fixed pervasive roles in
12214 the machine's execution model produces disastrous results.
12216 This flag does not have a negative form, because it specifies a
12219 @item -fcall-saved-@var{reg}
12220 @opindex fcall-saved
12221 Treat the register named @var{reg} as an allocable register saved by
12222 functions. It may be allocated even for temporaries or variables that
12223 live across a call. Functions compiled this way save and restore
12224 the register @var{reg} if they use it.
12226 It is an error to use this flag with the frame pointer or stack pointer.
12227 Use of this flag for other registers that have fixed pervasive roles in
12228 the machine's execution model produces disastrous results.
12230 A different sort of disaster results from the use of this flag for
12231 a register in which function values may be returned.
12233 This flag does not have a negative form, because it specifies a
12236 @item -fpack-struct[=@var{n}]
12237 @opindex fpack-struct
12238 Without a value specified, pack all structure members together without
12239 holes. When a value is specified (which must be a small power of two), pack
12240 structure members according to this value, representing the maximum
12241 alignment (that is, objects with default alignment requirements larger than
12242 this are output potentially unaligned at the next fitting location.
12244 @strong{Warning:} the @option{-fpack-struct} switch causes GCC to generate
12245 code that is not binary compatible with code generated without that switch.
12246 Additionally, it makes the code suboptimal.
12247 Use it to conform to a non-default application binary interface.
12249 @item -fleading-underscore
12250 @opindex fleading-underscore
12251 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
12252 change the way C symbols are represented in the object file. One use
12253 is to help link with legacy assembly code.
12255 @strong{Warning:} the @option{-fleading-underscore} switch causes GCC to
12256 generate code that is not binary compatible with code generated without that
12257 switch. Use it to conform to a non-default application binary interface.
12258 Not all targets provide complete support for this switch.
12260 @item -ftls-model=@var{model}
12261 @opindex ftls-model
12262 Alter the thread-local storage model to be used (@pxref{Thread-Local}).
12263 The @var{model} argument should be one of @samp{global-dynamic},
12264 @samp{local-dynamic}, @samp{initial-exec} or @samp{local-exec}.
12265 Note that the choice is subject to optimization: the compiler may use
12266 a more efficient model for symbols not visible outside of the translation
12267 unit, or if @option{-fpic} is not given on the command line.
12269 The default without @option{-fpic} is @samp{initial-exec}; with
12270 @option{-fpic} the default is @samp{global-dynamic}.
12272 @item -ftrampolines
12273 @opindex ftrampolines
12274 For targets that normally need trampolines for nested functions, always
12275 generate them instead of using descriptors. Otherwise, for targets that
12276 do not need them, like for example HP-PA or IA-64, do nothing.
12278 A trampoline is a small piece of code that is created at run time on the
12279 stack when the address of a nested function is taken, and is used to call
12280 the nested function indirectly. Therefore, it requires the stack to be
12281 made executable in order for the program to work properly.
12283 @option{-fno-trampolines} is enabled by default on a language by language
12284 basis to let the compiler avoid generating them, if it computes that this
12285 is safe, and replace them with descriptors. Descriptors are made up of data
12286 only, but the generated code must be prepared to deal with them. As of this
12287 writing, @option{-fno-trampolines} is enabled by default only for Ada.
12289 Moreover, code compiled with @option{-ftrampolines} and code compiled with
12290 @option{-fno-trampolines} are not binary compatible if nested functions are
12291 present. This option must therefore be used on a program-wide basis and be
12292 manipulated with extreme care.
12294 @item -fvisibility=@r{[}default@r{|}internal@r{|}hidden@r{|}protected@r{]}
12295 @opindex fvisibility
12296 Set the default ELF image symbol visibility to the specified option---all
12297 symbols are marked with this unless overridden within the code.
12298 Using this feature can very substantially improve linking and
12299 load times of shared object libraries, produce more optimized
12300 code, provide near-perfect API export and prevent symbol clashes.
12301 It is @strong{strongly} recommended that you use this in any shared objects
12304 Despite the nomenclature, @samp{default} always means public; i.e.,
12305 available to be linked against from outside the shared object.
12306 @samp{protected} and @samp{internal} are pretty useless in real-world
12307 usage so the only other commonly used option is @samp{hidden}.
12308 The default if @option{-fvisibility} isn't specified is
12309 @samp{default}, i.e., make every symbol public.
12311 A good explanation of the benefits offered by ensuring ELF
12312 symbols have the correct visibility is given by ``How To Write
12313 Shared Libraries'' by Ulrich Drepper (which can be found at
12314 @w{@uref{http://www.akkadia.org/drepper/}})---however a superior
12315 solution made possible by this option to marking things hidden when
12316 the default is public is to make the default hidden and mark things
12317 public. This is the norm with DLLs on Windows and with @option{-fvisibility=hidden}
12318 and @code{__attribute__ ((visibility("default")))} instead of
12319 @code{__declspec(dllexport)} you get almost identical semantics with
12320 identical syntax. This is a great boon to those working with
12321 cross-platform projects.
12323 For those adding visibility support to existing code, you may find
12324 @code{#pragma GCC visibility} of use. This works by you enclosing
12325 the declarations you wish to set visibility for with (for example)
12326 @code{#pragma GCC visibility push(hidden)} and
12327 @code{#pragma GCC visibility pop}.
12328 Bear in mind that symbol visibility should be viewed @strong{as
12329 part of the API interface contract} and thus all new code should
12330 always specify visibility when it is not the default; i.e., declarations
12331 only for use within the local DSO should @strong{always} be marked explicitly
12332 as hidden as so to avoid PLT indirection overheads---making this
12333 abundantly clear also aids readability and self-documentation of the code.
12334 Note that due to ISO C++ specification requirements, @code{operator new} and
12335 @code{operator delete} must always be of default visibility.
12337 Be aware that headers from outside your project, in particular system
12338 headers and headers from any other library you use, may not be
12339 expecting to be compiled with visibility other than the default. You
12340 may need to explicitly say @code{#pragma GCC visibility push(default)}
12341 before including any such headers.
12343 @code{extern} declarations are not affected by @option{-fvisibility}, so
12344 a lot of code can be recompiled with @option{-fvisibility=hidden} with
12345 no modifications. However, this means that calls to @code{extern}
12346 functions with no explicit visibility use the PLT, so it is more
12347 effective to use @code{__attribute ((visibility))} and/or
12348 @code{#pragma GCC visibility} to tell the compiler which @code{extern}
12349 declarations should be treated as hidden.
12351 Note that @option{-fvisibility} does affect C++ vague linkage
12352 entities. This means that, for instance, an exception class that is
12353 be thrown between DSOs must be explicitly marked with default
12354 visibility so that the @samp{type_info} nodes are unified between
12357 An overview of these techniques, their benefits and how to use them
12358 is at @uref{http://gcc.gnu.org/@/wiki/@/Visibility}.
12360 @item -fstrict-volatile-bitfields
12361 @opindex fstrict-volatile-bitfields
12362 This option should be used if accesses to volatile bit-fields (or other
12363 structure fields, although the compiler usually honors those types
12364 anyway) should use a single access of the width of the
12365 field's type, aligned to a natural alignment if possible. For
12366 example, targets with memory-mapped peripheral registers might require
12367 all such accesses to be 16 bits wide; with this flag you can
12368 declare all peripheral bit-fields as @code{unsigned short} (assuming short
12369 is 16 bits on these targets) to force GCC to use 16-bit accesses
12370 instead of, perhaps, a more efficient 32-bit access.
12372 If this option is disabled, the compiler uses the most efficient
12373 instruction. In the previous example, that might be a 32-bit load
12374 instruction, even though that accesses bytes that do not contain
12375 any portion of the bit-field, or memory-mapped registers unrelated to
12376 the one being updated.
12378 In some cases, such as when the @code{packed} attribute is applied to a
12379 structure field, it may not be possible to access the field with a single
12380 read or write that is correctly aligned for the target machine. In this
12381 case GCC falls back to generating multiple accesses rather than code that
12382 will fault or truncate the result at run time.
12384 Note: Due to restrictions of the C/C++11 memory model, write accesses are
12385 not allowed to touch non bit-field members. It is therefore recommended
12386 to define all bits of the field's type as bit-field members.
12388 The default value of this option is determined by the application binary
12389 interface for the target processor.
12391 @item -fsync-libcalls
12392 @opindex fsync-libcalls
12393 This option controls whether any out-of-line instance of the @code{__sync}
12394 family of functions may be used to implement the C++11 @code{__atomic}
12395 family of functions.
12397 The default value of this option is enabled, thus the only useful form
12398 of the option is @option{-fno-sync-libcalls}. This option is used in
12399 the implementation of the @file{libatomic} runtime library.
12403 @node Developer Options
12404 @section GCC Developer Options
12405 @cindex developer options
12406 @cindex debugging GCC
12407 @cindex debug dump options
12408 @cindex dump options
12409 @cindex compilation statistics
12411 This section describes command-line options that are primarily of
12412 interest to GCC developers, including options to support compiler
12413 testing and investigation of compiler bugs and compile-time
12414 performance problems. This includes options that produce debug dumps
12415 at various points in the compilation; that print statistics such as
12416 memory use and execution time; and that print information about GCC's
12417 configuration, such as where it searches for libraries. You should
12418 rarely need to use any of these options for ordinary compilation and
12423 @item -d@var{letters}
12424 @itemx -fdump-rtl-@var{pass}
12425 @itemx -fdump-rtl-@var{pass}=@var{filename}
12427 @opindex fdump-rtl-@var{pass}
12428 Says to make debugging dumps during compilation at times specified by
12429 @var{letters}. This is used for debugging the RTL-based passes of the
12430 compiler. The file names for most of the dumps are made by appending
12431 a pass number and a word to the @var{dumpname}, and the files are
12432 created in the directory of the output file. In case of
12433 @option{=@var{filename}} option, the dump is output on the given file
12434 instead of the pass numbered dump files. Note that the pass number is
12435 assigned as passes are registered into the pass manager. Most passes
12436 are registered in the order that they will execute and for these passes
12437 the number corresponds to the pass execution order. However, passes
12438 registered by plugins, passes specific to compilation targets, or
12439 passes that are otherwise registered after all the other passes are
12440 numbered higher than a pass named "final", even if they are executed
12441 earlier. @var{dumpname} is generated from the name of the output
12442 file if explicitly specified and not an executable, otherwise it is
12443 the basename of the source file. These switches may have different
12444 effects when @option{-E} is used for preprocessing.
12446 Debug dumps can be enabled with a @option{-fdump-rtl} switch or some
12447 @option{-d} option @var{letters}. Here are the possible
12448 letters for use in @var{pass} and @var{letters}, and their meanings:
12452 @item -fdump-rtl-alignments
12453 @opindex fdump-rtl-alignments
12454 Dump after branch alignments have been computed.
12456 @item -fdump-rtl-asmcons
12457 @opindex fdump-rtl-asmcons
12458 Dump after fixing rtl statements that have unsatisfied in/out constraints.
12460 @item -fdump-rtl-auto_inc_dec
12461 @opindex fdump-rtl-auto_inc_dec
12462 Dump after auto-inc-dec discovery. This pass is only run on
12463 architectures that have auto inc or auto dec instructions.
12465 @item -fdump-rtl-barriers
12466 @opindex fdump-rtl-barriers
12467 Dump after cleaning up the barrier instructions.
12469 @item -fdump-rtl-bbpart
12470 @opindex fdump-rtl-bbpart
12471 Dump after partitioning hot and cold basic blocks.
12473 @item -fdump-rtl-bbro
12474 @opindex fdump-rtl-bbro
12475 Dump after block reordering.
12477 @item -fdump-rtl-btl1
12478 @itemx -fdump-rtl-btl2
12479 @opindex fdump-rtl-btl2
12480 @opindex fdump-rtl-btl2
12481 @option{-fdump-rtl-btl1} and @option{-fdump-rtl-btl2} enable dumping
12482 after the two branch
12483 target load optimization passes.
12485 @item -fdump-rtl-bypass
12486 @opindex fdump-rtl-bypass
12487 Dump after jump bypassing and control flow optimizations.
12489 @item -fdump-rtl-combine
12490 @opindex fdump-rtl-combine
12491 Dump after the RTL instruction combination pass.
12493 @item -fdump-rtl-compgotos
12494 @opindex fdump-rtl-compgotos
12495 Dump after duplicating the computed gotos.
12497 @item -fdump-rtl-ce1
12498 @itemx -fdump-rtl-ce2
12499 @itemx -fdump-rtl-ce3
12500 @opindex fdump-rtl-ce1
12501 @opindex fdump-rtl-ce2
12502 @opindex fdump-rtl-ce3
12503 @option{-fdump-rtl-ce1}, @option{-fdump-rtl-ce2}, and
12504 @option{-fdump-rtl-ce3} enable dumping after the three
12505 if conversion passes.
12507 @item -fdump-rtl-cprop_hardreg
12508 @opindex fdump-rtl-cprop_hardreg
12509 Dump after hard register copy propagation.
12511 @item -fdump-rtl-csa
12512 @opindex fdump-rtl-csa
12513 Dump after combining stack adjustments.
12515 @item -fdump-rtl-cse1
12516 @itemx -fdump-rtl-cse2
12517 @opindex fdump-rtl-cse1
12518 @opindex fdump-rtl-cse2
12519 @option{-fdump-rtl-cse1} and @option{-fdump-rtl-cse2} enable dumping after
12520 the two common subexpression elimination passes.
12522 @item -fdump-rtl-dce
12523 @opindex fdump-rtl-dce
12524 Dump after the standalone dead code elimination passes.
12526 @item -fdump-rtl-dbr
12527 @opindex fdump-rtl-dbr
12528 Dump after delayed branch scheduling.
12530 @item -fdump-rtl-dce1
12531 @itemx -fdump-rtl-dce2
12532 @opindex fdump-rtl-dce1
12533 @opindex fdump-rtl-dce2
12534 @option{-fdump-rtl-dce1} and @option{-fdump-rtl-dce2} enable dumping after
12535 the two dead store elimination passes.
12537 @item -fdump-rtl-eh
12538 @opindex fdump-rtl-eh
12539 Dump after finalization of EH handling code.
12541 @item -fdump-rtl-eh_ranges
12542 @opindex fdump-rtl-eh_ranges
12543 Dump after conversion of EH handling range regions.
12545 @item -fdump-rtl-expand
12546 @opindex fdump-rtl-expand
12547 Dump after RTL generation.
12549 @item -fdump-rtl-fwprop1
12550 @itemx -fdump-rtl-fwprop2
12551 @opindex fdump-rtl-fwprop1
12552 @opindex fdump-rtl-fwprop2
12553 @option{-fdump-rtl-fwprop1} and @option{-fdump-rtl-fwprop2} enable
12554 dumping after the two forward propagation passes.
12556 @item -fdump-rtl-gcse1
12557 @itemx -fdump-rtl-gcse2
12558 @opindex fdump-rtl-gcse1
12559 @opindex fdump-rtl-gcse2
12560 @option{-fdump-rtl-gcse1} and @option{-fdump-rtl-gcse2} enable dumping
12561 after global common subexpression elimination.
12563 @item -fdump-rtl-init-regs
12564 @opindex fdump-rtl-init-regs
12565 Dump after the initialization of the registers.
12567 @item -fdump-rtl-initvals
12568 @opindex fdump-rtl-initvals
12569 Dump after the computation of the initial value sets.
12571 @item -fdump-rtl-into_cfglayout
12572 @opindex fdump-rtl-into_cfglayout
12573 Dump after converting to cfglayout mode.
12575 @item -fdump-rtl-ira
12576 @opindex fdump-rtl-ira
12577 Dump after iterated register allocation.
12579 @item -fdump-rtl-jump
12580 @opindex fdump-rtl-jump
12581 Dump after the second jump optimization.
12583 @item -fdump-rtl-loop2
12584 @opindex fdump-rtl-loop2
12585 @option{-fdump-rtl-loop2} enables dumping after the rtl
12586 loop optimization passes.
12588 @item -fdump-rtl-mach
12589 @opindex fdump-rtl-mach
12590 Dump after performing the machine dependent reorganization pass, if that
12593 @item -fdump-rtl-mode_sw
12594 @opindex fdump-rtl-mode_sw
12595 Dump after removing redundant mode switches.
12597 @item -fdump-rtl-rnreg
12598 @opindex fdump-rtl-rnreg
12599 Dump after register renumbering.
12601 @item -fdump-rtl-outof_cfglayout
12602 @opindex fdump-rtl-outof_cfglayout
12603 Dump after converting from cfglayout mode.
12605 @item -fdump-rtl-peephole2
12606 @opindex fdump-rtl-peephole2
12607 Dump after the peephole pass.
12609 @item -fdump-rtl-postreload
12610 @opindex fdump-rtl-postreload
12611 Dump after post-reload optimizations.
12613 @item -fdump-rtl-pro_and_epilogue
12614 @opindex fdump-rtl-pro_and_epilogue
12615 Dump after generating the function prologues and epilogues.
12617 @item -fdump-rtl-sched1
12618 @itemx -fdump-rtl-sched2
12619 @opindex fdump-rtl-sched1
12620 @opindex fdump-rtl-sched2
12621 @option{-fdump-rtl-sched1} and @option{-fdump-rtl-sched2} enable dumping
12622 after the basic block scheduling passes.
12624 @item -fdump-rtl-ree
12625 @opindex fdump-rtl-ree
12626 Dump after sign/zero extension elimination.
12628 @item -fdump-rtl-seqabstr
12629 @opindex fdump-rtl-seqabstr
12630 Dump after common sequence discovery.
12632 @item -fdump-rtl-shorten
12633 @opindex fdump-rtl-shorten
12634 Dump after shortening branches.
12636 @item -fdump-rtl-sibling
12637 @opindex fdump-rtl-sibling
12638 Dump after sibling call optimizations.
12640 @item -fdump-rtl-split1
12641 @itemx -fdump-rtl-split2
12642 @itemx -fdump-rtl-split3
12643 @itemx -fdump-rtl-split4
12644 @itemx -fdump-rtl-split5
12645 @opindex fdump-rtl-split1
12646 @opindex fdump-rtl-split2
12647 @opindex fdump-rtl-split3
12648 @opindex fdump-rtl-split4
12649 @opindex fdump-rtl-split5
12650 These options enable dumping after five rounds of
12651 instruction splitting.
12653 @item -fdump-rtl-sms
12654 @opindex fdump-rtl-sms
12655 Dump after modulo scheduling. This pass is only run on some
12658 @item -fdump-rtl-stack
12659 @opindex fdump-rtl-stack
12660 Dump after conversion from GCC's ``flat register file'' registers to the
12661 x87's stack-like registers. This pass is only run on x86 variants.
12663 @item -fdump-rtl-subreg1
12664 @itemx -fdump-rtl-subreg2
12665 @opindex fdump-rtl-subreg1
12666 @opindex fdump-rtl-subreg2
12667 @option{-fdump-rtl-subreg1} and @option{-fdump-rtl-subreg2} enable dumping after
12668 the two subreg expansion passes.
12670 @item -fdump-rtl-unshare
12671 @opindex fdump-rtl-unshare
12672 Dump after all rtl has been unshared.
12674 @item -fdump-rtl-vartrack
12675 @opindex fdump-rtl-vartrack
12676 Dump after variable tracking.
12678 @item -fdump-rtl-vregs
12679 @opindex fdump-rtl-vregs
12680 Dump after converting virtual registers to hard registers.
12682 @item -fdump-rtl-web
12683 @opindex fdump-rtl-web
12684 Dump after live range splitting.
12686 @item -fdump-rtl-regclass
12687 @itemx -fdump-rtl-subregs_of_mode_init
12688 @itemx -fdump-rtl-subregs_of_mode_finish
12689 @itemx -fdump-rtl-dfinit
12690 @itemx -fdump-rtl-dfinish
12691 @opindex fdump-rtl-regclass
12692 @opindex fdump-rtl-subregs_of_mode_init
12693 @opindex fdump-rtl-subregs_of_mode_finish
12694 @opindex fdump-rtl-dfinit
12695 @opindex fdump-rtl-dfinish
12696 These dumps are defined but always produce empty files.
12699 @itemx -fdump-rtl-all
12701 @opindex fdump-rtl-all
12702 Produce all the dumps listed above.
12706 Annotate the assembler output with miscellaneous debugging information.
12710 Dump all macro definitions, at the end of preprocessing, in addition to
12715 Produce a core dump whenever an error occurs.
12719 Annotate the assembler output with a comment indicating which
12720 pattern and alternative is used. The length of each instruction is
12725 Dump the RTL in the assembler output as a comment before each instruction.
12726 Also turns on @option{-dp} annotation.
12730 Just generate RTL for a function instead of compiling it. Usually used
12731 with @option{-fdump-rtl-expand}.
12734 @item -fdump-noaddr
12735 @opindex fdump-noaddr
12736 When doing debugging dumps, suppress address output. This makes it more
12737 feasible to use diff on debugging dumps for compiler invocations with
12738 different compiler binaries and/or different
12739 text / bss / data / heap / stack / dso start locations.
12742 @opindex freport-bug
12743 Collect and dump debug information into a temporary file if an
12744 internal compiler error (ICE) occurs.
12746 @item -fdump-unnumbered
12747 @opindex fdump-unnumbered
12748 When doing debugging dumps, suppress instruction numbers and address output.
12749 This makes it more feasible to use diff on debugging dumps for compiler
12750 invocations with different options, in particular with and without
12753 @item -fdump-unnumbered-links
12754 @opindex fdump-unnumbered-links
12755 When doing debugging dumps (see @option{-d} option above), suppress
12756 instruction numbers for the links to the previous and next instructions
12759 @item -fdump-translation-unit @r{(C++ only)}
12760 @itemx -fdump-translation-unit-@var{options} @r{(C++ only)}
12761 @opindex fdump-translation-unit
12762 Dump a representation of the tree structure for the entire translation
12763 unit to a file. The file name is made by appending @file{.tu} to the
12764 source file name, and the file is created in the same directory as the
12765 output file. If the @samp{-@var{options}} form is used, @var{options}
12766 controls the details of the dump as described for the
12767 @option{-fdump-tree} options.
12769 @item -fdump-class-hierarchy @r{(C++ only)}
12770 @itemx -fdump-class-hierarchy-@var{options} @r{(C++ only)}
12771 @opindex fdump-class-hierarchy
12772 Dump a representation of each class's hierarchy and virtual function
12773 table layout to a file. The file name is made by appending
12774 @file{.class} to the source file name, and the file is created in the
12775 same directory as the output file. If the @samp{-@var{options}} form
12776 is used, @var{options} controls the details of the dump as described
12777 for the @option{-fdump-tree} options.
12779 @item -fdump-ipa-@var{switch}
12781 Control the dumping at various stages of inter-procedural analysis
12782 language tree to a file. The file name is generated by appending a
12783 switch specific suffix to the source file name, and the file is created
12784 in the same directory as the output file. The following dumps are
12789 Enables all inter-procedural analysis dumps.
12792 Dumps information about call-graph optimization, unused function removal,
12793 and inlining decisions.
12796 Dump after function inlining.
12800 @item -fdump-passes
12801 @opindex fdump-passes
12802 Dump the list of optimization passes that are turned on and off by
12803 the current command-line options.
12805 @item -fdump-statistics-@var{option}
12806 @opindex fdump-statistics
12807 Enable and control dumping of pass statistics in a separate file. The
12808 file name is generated by appending a suffix ending in
12809 @samp{.statistics} to the source file name, and the file is created in
12810 the same directory as the output file. If the @samp{-@var{option}}
12811 form is used, @samp{-stats} causes counters to be summed over the
12812 whole compilation unit while @samp{-details} dumps every event as
12813 the passes generate them. The default with no option is to sum
12814 counters for each function compiled.
12816 @item -fdump-tree-@var{switch}
12817 @itemx -fdump-tree-@var{switch}-@var{options}
12818 @itemx -fdump-tree-@var{switch}-@var{options}=@var{filename}
12819 @opindex fdump-tree
12820 Control the dumping at various stages of processing the intermediate
12821 language tree to a file. The file name is generated by appending a
12822 switch-specific suffix to the source file name, and the file is
12823 created in the same directory as the output file. In case of
12824 @option{=@var{filename}} option, the dump is output on the given file
12825 instead of the auto named dump files. If the @samp{-@var{options}}
12826 form is used, @var{options} is a list of @samp{-} separated options
12827 which control the details of the dump. Not all options are applicable
12828 to all dumps; those that are not meaningful are ignored. The
12829 following options are available
12833 Print the address of each node. Usually this is not meaningful as it
12834 changes according to the environment and source file. Its primary use
12835 is for tying up a dump file with a debug environment.
12837 If @code{DECL_ASSEMBLER_NAME} has been set for a given decl, use that
12838 in the dump instead of @code{DECL_NAME}. Its primary use is ease of
12839 use working backward from mangled names in the assembly file.
12841 When dumping front-end intermediate representations, inhibit dumping
12842 of members of a scope or body of a function merely because that scope
12843 has been reached. Only dump such items when they are directly reachable
12844 by some other path.
12846 When dumping pretty-printed trees, this option inhibits dumping the
12847 bodies of control structures.
12849 When dumping RTL, print the RTL in slim (condensed) form instead of
12850 the default LISP-like representation.
12852 Print a raw representation of the tree. By default, trees are
12853 pretty-printed into a C-like representation.
12855 Enable more detailed dumps (not honored by every dump option). Also
12856 include information from the optimization passes.
12858 Enable dumping various statistics about the pass (not honored by every dump
12861 Enable showing basic block boundaries (disabled in raw dumps).
12863 For each of the other indicated dump files (@option{-fdump-rtl-@var{pass}}),
12864 dump a representation of the control flow graph suitable for viewing with
12865 GraphViz to @file{@var{file}.@var{passid}.@var{pass}.dot}. Each function in
12866 the file is pretty-printed as a subgraph, so that GraphViz can render them
12867 all in a single plot.
12869 This option currently only works for RTL dumps, and the RTL is always
12870 dumped in slim form.
12872 Enable showing virtual operands for every statement.
12874 Enable showing line numbers for statements.
12876 Enable showing the unique ID (@code{DECL_UID}) for each variable.
12878 Enable showing the tree dump for each statement.
12880 Enable showing the EH region number holding each statement.
12882 Enable showing scalar evolution analysis details.
12884 Enable showing optimization information (only available in certain
12887 Enable showing missed optimization information (only available in certain
12890 Enable other detailed optimization information (only available in
12892 @item =@var{filename}
12893 Instead of an auto named dump file, output into the given file
12894 name. The file names @file{stdout} and @file{stderr} are treated
12895 specially and are considered already open standard streams. For
12899 gcc -O2 -ftree-vectorize -fdump-tree-vect-blocks=foo.dump
12900 -fdump-tree-pre=stderr file.c
12903 outputs vectorizer dump into @file{foo.dump}, while the PRE dump is
12904 output on to @file{stderr}. If two conflicting dump filenames are
12905 given for the same pass, then the latter option overrides the earlier
12909 @opindex fdump-tree-split-paths
12910 Dump each function after splitting paths to loop backedges. The file
12911 name is made by appending @file{.split-paths} to the source file name.
12914 Turn on all options, except @option{raw}, @option{slim}, @option{verbose}
12915 and @option{lineno}.
12918 Turn on all optimization options, i.e., @option{optimized},
12919 @option{missed}, and @option{note}.
12922 The following tree dumps are possible:
12926 @opindex fdump-tree-original
12927 Dump before any tree based optimization, to @file{@var{file}.original}.
12930 @opindex fdump-tree-optimized
12931 Dump after all tree based optimization, to @file{@var{file}.optimized}.
12934 @opindex fdump-tree-gimple
12935 Dump each function before and after the gimplification pass to a file. The
12936 file name is made by appending @file{.gimple} to the source file name.
12939 @opindex fdump-tree-cfg
12940 Dump the control flow graph of each function to a file. The file name is
12941 made by appending @file{.cfg} to the source file name.
12944 @opindex fdump-tree-ch
12945 Dump each function after copying loop headers. The file name is made by
12946 appending @file{.ch} to the source file name.
12949 @opindex fdump-tree-ssa
12950 Dump SSA related information to a file. The file name is made by appending
12951 @file{.ssa} to the source file name.
12954 @opindex fdump-tree-alias
12955 Dump aliasing information for each function. The file name is made by
12956 appending @file{.alias} to the source file name.
12959 @opindex fdump-tree-ccp
12960 Dump each function after CCP@. The file name is made by appending
12961 @file{.ccp} to the source file name.
12964 @opindex fdump-tree-storeccp
12965 Dump each function after STORE-CCP@. The file name is made by appending
12966 @file{.storeccp} to the source file name.
12969 @opindex fdump-tree-pre
12970 Dump trees after partial redundancy elimination and/or code hoisting.
12971 The file name is made by appending @file{.pre} to the source file name.
12974 @opindex fdump-tree-fre
12975 Dump trees after full redundancy elimination. The file name is made
12976 by appending @file{.fre} to the source file name.
12979 @opindex fdump-tree-copyprop
12980 Dump trees after copy propagation. The file name is made
12981 by appending @file{.copyprop} to the source file name.
12983 @item store_copyprop
12984 @opindex fdump-tree-store_copyprop
12985 Dump trees after store copy-propagation. The file name is made
12986 by appending @file{.store_copyprop} to the source file name.
12989 @opindex fdump-tree-dce
12990 Dump each function after dead code elimination. The file name is made by
12991 appending @file{.dce} to the source file name.
12994 @opindex fdump-tree-sra
12995 Dump each function after performing scalar replacement of aggregates. The
12996 file name is made by appending @file{.sra} to the source file name.
12999 @opindex fdump-tree-sink
13000 Dump each function after performing code sinking. The file name is made
13001 by appending @file{.sink} to the source file name.
13004 @opindex fdump-tree-dom
13005 Dump each function after applying dominator tree optimizations. The file
13006 name is made by appending @file{.dom} to the source file name.
13009 @opindex fdump-tree-dse
13010 Dump each function after applying dead store elimination. The file
13011 name is made by appending @file{.dse} to the source file name.
13014 @opindex fdump-tree-phiopt
13015 Dump each function after optimizing PHI nodes into straightline code. The file
13016 name is made by appending @file{.phiopt} to the source file name.
13019 @opindex fdump-tree-backprop
13020 Dump each function after back-propagating use information up the definition
13021 chain. The file name is made by appending @file{.backprop} to the
13025 @opindex fdump-tree-forwprop
13026 Dump each function after forward propagating single use variables. The file
13027 name is made by appending @file{.forwprop} to the source file name.
13030 @opindex fdump-tree-nrv
13031 Dump each function after applying the named return value optimization on
13032 generic trees. The file name is made by appending @file{.nrv} to the source
13036 @opindex fdump-tree-vect
13037 Dump each function after applying vectorization of loops. The file name is
13038 made by appending @file{.vect} to the source file name.
13041 @opindex fdump-tree-slp
13042 Dump each function after applying vectorization of basic blocks. The file name
13043 is made by appending @file{.slp} to the source file name.
13046 @opindex fdump-tree-vrp
13047 Dump each function after Value Range Propagation (VRP). The file name
13048 is made by appending @file{.vrp} to the source file name.
13051 @opindex fdump-tree-evrp
13052 Dump each function after Early Value Range Propagation (EVRP). The file name
13053 is made by appending @file{.evrp} to the source file name.
13056 @opindex fdump-tree-oaccdevlow
13057 Dump each function after applying device-specific OpenACC transformations.
13058 The file name is made by appending @file{.oaccdevlow} to the source file name.
13061 @opindex fdump-tree-all
13062 Enable all the available tree dumps with the flags provided in this option.
13066 @itemx -fopt-info-@var{options}
13067 @itemx -fopt-info-@var{options}=@var{filename}
13069 Controls optimization dumps from various optimization passes. If the
13070 @samp{-@var{options}} form is used, @var{options} is a list of
13071 @samp{-} separated option keywords to select the dump details and
13074 The @var{options} can be divided into two groups: options describing the
13075 verbosity of the dump, and options describing which optimizations
13076 should be included. The options from both the groups can be freely
13077 mixed as they are non-overlapping. However, in case of any conflicts,
13078 the later options override the earlier options on the command
13081 The following options control the dump verbosity:
13085 Print information when an optimization is successfully applied. It is
13086 up to a pass to decide which information is relevant. For example, the
13087 vectorizer passes print the source location of loops which are
13088 successfully vectorized.
13090 Print information about missed optimizations. Individual passes
13091 control which information to include in the output.
13093 Print verbose information about optimizations, such as certain
13094 transformations, more detailed messages about decisions etc.
13096 Print detailed optimization information. This includes
13097 @samp{optimized}, @samp{missed}, and @samp{note}.
13100 One or more of the following option keywords can be used to describe a
13101 group of optimizations:
13105 Enable dumps from all interprocedural optimizations.
13107 Enable dumps from all loop optimizations.
13109 Enable dumps from all inlining optimizations.
13111 Enable dumps from all vectorization optimizations.
13113 Enable dumps from all optimizations. This is a superset of
13114 the optimization groups listed above.
13117 If @var{options} is
13118 omitted, it defaults to @samp{optimized-optall}, which means to dump all
13119 info about successful optimizations from all the passes.
13121 If the @var{filename} is provided, then the dumps from all the
13122 applicable optimizations are concatenated into the @var{filename}.
13123 Otherwise the dump is output onto @file{stderr}. Though multiple
13124 @option{-fopt-info} options are accepted, only one of them can include
13125 a @var{filename}. If other filenames are provided then all but the
13126 first such option are ignored.
13128 Note that the output @var{filename} is overwritten
13129 in case of multiple translation units. If a combined output from
13130 multiple translation units is desired, @file{stderr} should be used
13133 In the following example, the optimization info is output to
13142 gcc -O3 -fopt-info-missed=missed.all
13146 outputs missed optimization report from all the passes into
13147 @file{missed.all}, and this one:
13150 gcc -O2 -ftree-vectorize -fopt-info-vec-missed
13154 prints information about missed optimization opportunities from
13155 vectorization passes on @file{stderr}.
13156 Note that @option{-fopt-info-vec-missed} is equivalent to
13157 @option{-fopt-info-missed-vec}.
13159 As another example,
13161 gcc -O3 -fopt-info-inline-optimized-missed=inline.txt
13165 outputs information about missed optimizations as well as
13166 optimized locations from all the inlining passes into
13172 gcc -fopt-info-vec-missed=vec.miss -fopt-info-loop-optimized=loop.opt
13176 Here the two output filenames @file{vec.miss} and @file{loop.opt} are
13177 in conflict since only one output file is allowed. In this case, only
13178 the first option takes effect and the subsequent options are
13179 ignored. Thus only @file{vec.miss} is produced which contains
13180 dumps from the vectorizer about missed opportunities.
13182 @item -fsched-verbose=@var{n}
13183 @opindex fsched-verbose
13184 On targets that use instruction scheduling, this option controls the
13185 amount of debugging output the scheduler prints to the dump files.
13187 For @var{n} greater than zero, @option{-fsched-verbose} outputs the
13188 same information as @option{-fdump-rtl-sched1} and @option{-fdump-rtl-sched2}.
13189 For @var{n} greater than one, it also output basic block probabilities,
13190 detailed ready list information and unit/insn info. For @var{n} greater
13191 than two, it includes RTL at abort point, control-flow and regions info.
13192 And for @var{n} over four, @option{-fsched-verbose} also includes
13197 @item -fenable-@var{kind}-@var{pass}
13198 @itemx -fdisable-@var{kind}-@var{pass}=@var{range-list}
13202 This is a set of options that are used to explicitly disable/enable
13203 optimization passes. These options are intended for use for debugging GCC.
13204 Compiler users should use regular options for enabling/disabling
13209 @item -fdisable-ipa-@var{pass}
13210 Disable IPA pass @var{pass}. @var{pass} is the pass name. If the same pass is
13211 statically invoked in the compiler multiple times, the pass name should be
13212 appended with a sequential number starting from 1.
13214 @item -fdisable-rtl-@var{pass}
13215 @itemx -fdisable-rtl-@var{pass}=@var{range-list}
13216 Disable RTL pass @var{pass}. @var{pass} is the pass name. If the same pass is
13217 statically invoked in the compiler multiple times, the pass name should be
13218 appended with a sequential number starting from 1. @var{range-list} is a
13219 comma-separated list of function ranges or assembler names. Each range is a number
13220 pair separated by a colon. The range is inclusive in both ends. If the range
13221 is trivial, the number pair can be simplified as a single number. If the
13222 function's call graph node's @var{uid} falls within one of the specified ranges,
13223 the @var{pass} is disabled for that function. The @var{uid} is shown in the
13224 function header of a dump file, and the pass names can be dumped by using
13225 option @option{-fdump-passes}.
13227 @item -fdisable-tree-@var{pass}
13228 @itemx -fdisable-tree-@var{pass}=@var{range-list}
13229 Disable tree pass @var{pass}. See @option{-fdisable-rtl} for the description of
13232 @item -fenable-ipa-@var{pass}
13233 Enable IPA pass @var{pass}. @var{pass} is the pass name. If the same pass is
13234 statically invoked in the compiler multiple times, the pass name should be
13235 appended with a sequential number starting from 1.
13237 @item -fenable-rtl-@var{pass}
13238 @itemx -fenable-rtl-@var{pass}=@var{range-list}
13239 Enable RTL pass @var{pass}. See @option{-fdisable-rtl} for option argument
13240 description and examples.
13242 @item -fenable-tree-@var{pass}
13243 @itemx -fenable-tree-@var{pass}=@var{range-list}
13244 Enable tree pass @var{pass}. See @option{-fdisable-rtl} for the description
13245 of option arguments.
13249 Here are some examples showing uses of these options.
13253 # disable ccp1 for all functions
13254 -fdisable-tree-ccp1
13255 # disable complete unroll for function whose cgraph node uid is 1
13256 -fenable-tree-cunroll=1
13257 # disable gcse2 for functions at the following ranges [1,1],
13258 # [300,400], and [400,1000]
13259 # disable gcse2 for functions foo and foo2
13260 -fdisable-rtl-gcse2=foo,foo2
13261 # disable early inlining
13262 -fdisable-tree-einline
13263 # disable ipa inlining
13264 -fdisable-ipa-inline
13265 # enable tree full unroll
13266 -fenable-tree-unroll
13271 @itemx -fchecking=@var{n}
13273 @opindex fno-checking
13274 Enable internal consistency checking. The default depends on
13275 the compiler configuration. @option{-fchecking=2} enables further
13276 internal consistency checking that might affect code generation.
13278 @item -frandom-seed=@var{string}
13279 @opindex frandom-seed
13280 This option provides a seed that GCC uses in place of
13281 random numbers in generating certain symbol names
13282 that have to be different in every compiled file. It is also used to
13283 place unique stamps in coverage data files and the object files that
13284 produce them. You can use the @option{-frandom-seed} option to produce
13285 reproducibly identical object files.
13287 The @var{string} can either be a number (decimal, octal or hex) or an
13288 arbitrary string (in which case it's converted to a number by
13291 The @var{string} should be different for every file you compile.
13294 @itemx -save-temps=cwd
13295 @opindex save-temps
13296 Store the usual ``temporary'' intermediate files permanently; place them
13297 in the current directory and name them based on the source file. Thus,
13298 compiling @file{foo.c} with @option{-c -save-temps} produces files
13299 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
13300 preprocessed @file{foo.i} output file even though the compiler now
13301 normally uses an integrated preprocessor.
13303 When used in combination with the @option{-x} command-line option,
13304 @option{-save-temps} is sensible enough to avoid over writing an
13305 input source file with the same extension as an intermediate file.
13306 The corresponding intermediate file may be obtained by renaming the
13307 source file before using @option{-save-temps}.
13309 If you invoke GCC in parallel, compiling several different source
13310 files that share a common base name in different subdirectories or the
13311 same source file compiled for multiple output destinations, it is
13312 likely that the different parallel compilers will interfere with each
13313 other, and overwrite the temporary files. For instance:
13316 gcc -save-temps -o outdir1/foo.o indir1/foo.c&
13317 gcc -save-temps -o outdir2/foo.o indir2/foo.c&
13320 may result in @file{foo.i} and @file{foo.o} being written to
13321 simultaneously by both compilers.
13323 @item -save-temps=obj
13324 @opindex save-temps=obj
13325 Store the usual ``temporary'' intermediate files permanently. If the
13326 @option{-o} option is used, the temporary files are based on the
13327 object file. If the @option{-o} option is not used, the
13328 @option{-save-temps=obj} switch behaves like @option{-save-temps}.
13333 gcc -save-temps=obj -c foo.c
13334 gcc -save-temps=obj -c bar.c -o dir/xbar.o
13335 gcc -save-temps=obj foobar.c -o dir2/yfoobar
13339 creates @file{foo.i}, @file{foo.s}, @file{dir/xbar.i},
13340 @file{dir/xbar.s}, @file{dir2/yfoobar.i}, @file{dir2/yfoobar.s}, and
13341 @file{dir2/yfoobar.o}.
13343 @item -time@r{[}=@var{file}@r{]}
13345 Report the CPU time taken by each subprocess in the compilation
13346 sequence. For C source files, this is the compiler proper and assembler
13347 (plus the linker if linking is done).
13349 Without the specification of an output file, the output looks like this:
13356 The first number on each line is the ``user time'', that is time spent
13357 executing the program itself. The second number is ``system time'',
13358 time spent executing operating system routines on behalf of the program.
13359 Both numbers are in seconds.
13361 With the specification of an output file, the output is appended to the
13362 named file, and it looks like this:
13365 0.12 0.01 cc1 @var{options}
13366 0.00 0.01 as @var{options}
13369 The ``user time'' and the ``system time'' are moved before the program
13370 name, and the options passed to the program are displayed, so that one
13371 can later tell what file was being compiled, and with which options.
13373 @item -fdump-final-insns@r{[}=@var{file}@r{]}
13374 @opindex fdump-final-insns
13375 Dump the final internal representation (RTL) to @var{file}. If the
13376 optional argument is omitted (or if @var{file} is @code{.}), the name
13377 of the dump file is determined by appending @code{.gkd} to the
13378 compilation output file name.
13380 @item -fcompare-debug@r{[}=@var{opts}@r{]}
13381 @opindex fcompare-debug
13382 @opindex fno-compare-debug
13383 If no error occurs during compilation, run the compiler a second time,
13384 adding @var{opts} and @option{-fcompare-debug-second} to the arguments
13385 passed to the second compilation. Dump the final internal
13386 representation in both compilations, and print an error if they differ.
13388 If the equal sign is omitted, the default @option{-gtoggle} is used.
13390 The environment variable @env{GCC_COMPARE_DEBUG}, if defined, non-empty
13391 and nonzero, implicitly enables @option{-fcompare-debug}. If
13392 @env{GCC_COMPARE_DEBUG} is defined to a string starting with a dash,
13393 then it is used for @var{opts}, otherwise the default @option{-gtoggle}
13396 @option{-fcompare-debug=}, with the equal sign but without @var{opts},
13397 is equivalent to @option{-fno-compare-debug}, which disables the dumping
13398 of the final representation and the second compilation, preventing even
13399 @env{GCC_COMPARE_DEBUG} from taking effect.
13401 To verify full coverage during @option{-fcompare-debug} testing, set
13402 @env{GCC_COMPARE_DEBUG} to say @option{-fcompare-debug-not-overridden},
13403 which GCC rejects as an invalid option in any actual compilation
13404 (rather than preprocessing, assembly or linking). To get just a
13405 warning, setting @env{GCC_COMPARE_DEBUG} to @samp{-w%n-fcompare-debug
13406 not overridden} will do.
13408 @item -fcompare-debug-second
13409 @opindex fcompare-debug-second
13410 This option is implicitly passed to the compiler for the second
13411 compilation requested by @option{-fcompare-debug}, along with options to
13412 silence warnings, and omitting other options that would cause
13413 side-effect compiler outputs to files or to the standard output. Dump
13414 files and preserved temporary files are renamed so as to contain the
13415 @code{.gk} additional extension during the second compilation, to avoid
13416 overwriting those generated by the first.
13418 When this option is passed to the compiler driver, it causes the
13419 @emph{first} compilation to be skipped, which makes it useful for little
13420 other than debugging the compiler proper.
13424 Turn off generation of debug info, if leaving out this option
13425 generates it, or turn it on at level 2 otherwise. The position of this
13426 argument in the command line does not matter; it takes effect after all
13427 other options are processed, and it does so only once, no matter how
13428 many times it is given. This is mainly intended to be used with
13429 @option{-fcompare-debug}.
13431 @item -fvar-tracking-assignments-toggle
13432 @opindex fvar-tracking-assignments-toggle
13433 @opindex fno-var-tracking-assignments-toggle
13434 Toggle @option{-fvar-tracking-assignments}, in the same way that
13435 @option{-gtoggle} toggles @option{-g}.
13439 Makes the compiler print out each function name as it is compiled, and
13440 print some statistics about each pass when it finishes.
13442 @item -ftime-report
13443 @opindex ftime-report
13444 Makes the compiler print some statistics about the time consumed by each
13445 pass when it finishes.
13447 @item -ftime-report-details
13448 @opindex ftime-report-details
13449 Record the time consumed by infrastructure parts separately for each pass.
13451 @item -fira-verbose=@var{n}
13452 @opindex fira-verbose
13453 Control the verbosity of the dump file for the integrated register allocator.
13454 The default value is 5. If the value @var{n} is greater or equal to 10,
13455 the dump output is sent to stderr using the same format as @var{n} minus 10.
13458 @opindex flto-report
13459 Prints a report with internal details on the workings of the link-time
13460 optimizer. The contents of this report vary from version to version.
13461 It is meant to be useful to GCC developers when processing object
13462 files in LTO mode (via @option{-flto}).
13464 Disabled by default.
13466 @item -flto-report-wpa
13467 @opindex flto-report-wpa
13468 Like @option{-flto-report}, but only print for the WPA phase of Link
13472 @opindex fmem-report
13473 Makes the compiler print some statistics about permanent memory
13474 allocation when it finishes.
13476 @item -fmem-report-wpa
13477 @opindex fmem-report-wpa
13478 Makes the compiler print some statistics about permanent memory
13479 allocation for the WPA phase only.
13481 @item -fpre-ipa-mem-report
13482 @opindex fpre-ipa-mem-report
13483 @item -fpost-ipa-mem-report
13484 @opindex fpost-ipa-mem-report
13485 Makes the compiler print some statistics about permanent memory
13486 allocation before or after interprocedural optimization.
13488 @item -fprofile-report
13489 @opindex fprofile-report
13490 Makes the compiler print some statistics about consistency of the
13491 (estimated) profile and effect of individual passes.
13493 @item -fstack-usage
13494 @opindex fstack-usage
13495 Makes the compiler output stack usage information for the program, on a
13496 per-function basis. The filename for the dump is made by appending
13497 @file{.su} to the @var{auxname}. @var{auxname} is generated from the name of
13498 the output file, if explicitly specified and it is not an executable,
13499 otherwise it is the basename of the source file. An entry is made up
13504 The name of the function.
13508 One or more qualifiers: @code{static}, @code{dynamic}, @code{bounded}.
13511 The qualifier @code{static} means that the function manipulates the stack
13512 statically: a fixed number of bytes are allocated for the frame on function
13513 entry and released on function exit; no stack adjustments are otherwise made
13514 in the function. The second field is this fixed number of bytes.
13516 The qualifier @code{dynamic} means that the function manipulates the stack
13517 dynamically: in addition to the static allocation described above, stack
13518 adjustments are made in the body of the function, for example to push/pop
13519 arguments around function calls. If the qualifier @code{bounded} is also
13520 present, the amount of these adjustments is bounded at compile time and
13521 the second field is an upper bound of the total amount of stack used by
13522 the function. If it is not present, the amount of these adjustments is
13523 not bounded at compile time and the second field only represents the
13528 Emit statistics about front-end processing at the end of the compilation.
13529 This option is supported only by the C++ front end, and
13530 the information is generally only useful to the G++ development team.
13532 @item -fdbg-cnt-list
13533 @opindex fdbg-cnt-list
13534 Print the name and the counter upper bound for all debug counters.
13537 @item -fdbg-cnt=@var{counter-value-list}
13539 Set the internal debug counter upper bound. @var{counter-value-list}
13540 is a comma-separated list of @var{name}:@var{value} pairs
13541 which sets the upper bound of each debug counter @var{name} to @var{value}.
13542 All debug counters have the initial upper bound of @code{UINT_MAX};
13543 thus @code{dbg_cnt} returns true always unless the upper bound
13544 is set by this option.
13545 For example, with @option{-fdbg-cnt=dce:10,tail_call:0},
13546 @code{dbg_cnt(dce)} returns true only for first 10 invocations.
13548 @item -print-file-name=@var{library}
13549 @opindex print-file-name
13550 Print the full absolute name of the library file @var{library} that
13551 would be used when linking---and don't do anything else. With this
13552 option, GCC does not compile or link anything; it just prints the
13555 @item -print-multi-directory
13556 @opindex print-multi-directory
13557 Print the directory name corresponding to the multilib selected by any
13558 other switches present in the command line. This directory is supposed
13559 to exist in @env{GCC_EXEC_PREFIX}.
13561 @item -print-multi-lib
13562 @opindex print-multi-lib
13563 Print the mapping from multilib directory names to compiler switches
13564 that enable them. The directory name is separated from the switches by
13565 @samp{;}, and each switch starts with an @samp{@@} instead of the
13566 @samp{-}, without spaces between multiple switches. This is supposed to
13567 ease shell processing.
13569 @item -print-multi-os-directory
13570 @opindex print-multi-os-directory
13571 Print the path to OS libraries for the selected
13572 multilib, relative to some @file{lib} subdirectory. If OS libraries are
13573 present in the @file{lib} subdirectory and no multilibs are used, this is
13574 usually just @file{.}, if OS libraries are present in @file{lib@var{suffix}}
13575 sibling directories this prints e.g.@: @file{../lib64}, @file{../lib} or
13576 @file{../lib32}, or if OS libraries are present in @file{lib/@var{subdir}}
13577 subdirectories it prints e.g.@: @file{amd64}, @file{sparcv9} or @file{ev6}.
13579 @item -print-multiarch
13580 @opindex print-multiarch
13581 Print the path to OS libraries for the selected multiarch,
13582 relative to some @file{lib} subdirectory.
13584 @item -print-prog-name=@var{program}
13585 @opindex print-prog-name
13586 Like @option{-print-file-name}, but searches for a program such as @command{cpp}.
13588 @item -print-libgcc-file-name
13589 @opindex print-libgcc-file-name
13590 Same as @option{-print-file-name=libgcc.a}.
13592 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
13593 but you do want to link with @file{libgcc.a}. You can do:
13596 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
13599 @item -print-search-dirs
13600 @opindex print-search-dirs
13601 Print the name of the configured installation directory and a list of
13602 program and library directories @command{gcc} searches---and don't do anything else.
13604 This is useful when @command{gcc} prints the error message
13605 @samp{installation problem, cannot exec cpp0: No such file or directory}.
13606 To resolve this you either need to put @file{cpp0} and the other compiler
13607 components where @command{gcc} expects to find them, or you can set the environment
13608 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
13609 Don't forget the trailing @samp{/}.
13610 @xref{Environment Variables}.
13612 @item -print-sysroot
13613 @opindex print-sysroot
13614 Print the target sysroot directory that is used during
13615 compilation. This is the target sysroot specified either at configure
13616 time or using the @option{--sysroot} option, possibly with an extra
13617 suffix that depends on compilation options. If no target sysroot is
13618 specified, the option prints nothing.
13620 @item -print-sysroot-headers-suffix
13621 @opindex print-sysroot-headers-suffix
13622 Print the suffix added to the target sysroot when searching for
13623 headers, or give an error if the compiler is not configured with such
13624 a suffix---and don't do anything else.
13627 @opindex dumpmachine
13628 Print the compiler's target machine (for example,
13629 @samp{i686-pc-linux-gnu})---and don't do anything else.
13632 @opindex dumpversion
13633 Print the compiler version (for example, @code{3.0})---and don't do
13638 Print the compiler's built-in specs---and don't do anything else. (This
13639 is used when GCC itself is being built.) @xref{Spec Files}.
13642 @node Submodel Options
13643 @section Machine-Dependent Options
13644 @cindex submodel options
13645 @cindex specifying hardware config
13646 @cindex hardware models and configurations, specifying
13647 @cindex target-dependent options
13648 @cindex machine-dependent options
13650 Each target machine supported by GCC can have its own options---for
13651 example, to allow you to compile for a particular processor variant or
13652 ABI, or to control optimizations specific to that machine. By
13653 convention, the names of machine-specific options start with
13656 Some configurations of the compiler also support additional target-specific
13657 options, usually for compatibility with other compilers on the same
13660 @c This list is ordered alphanumerically by subsection name.
13661 @c It should be the same order and spelling as these options are listed
13662 @c in Machine Dependent Options
13665 * AArch64 Options::
13666 * Adapteva Epiphany Options::
13670 * Blackfin Options::
13675 * DEC Alpha Options::
13679 * GNU/Linux Options::
13689 * MicroBlaze Options::
13692 * MN10300 Options::
13696 * Nios II Options::
13697 * Nvidia PTX Options::
13699 * picoChip Options::
13700 * PowerPC Options::
13702 * RS/6000 and PowerPC Options::
13704 * S/390 and zSeries Options::
13707 * Solaris 2 Options::
13710 * System V Options::
13711 * TILE-Gx Options::
13712 * TILEPro Options::
13717 * VxWorks Options::
13719 * x86 Windows Options::
13720 * Xstormy16 Options::
13722 * zSeries Options::
13725 @node AArch64 Options
13726 @subsection AArch64 Options
13727 @cindex AArch64 Options
13729 These options are defined for AArch64 implementations:
13733 @item -mabi=@var{name}
13735 Generate code for the specified data model. Permissible values
13736 are @samp{ilp32} for SysV-like data model where int, long int and pointers
13737 are 32 bits, and @samp{lp64} for SysV-like data model where int is 32 bits,
13738 but long int and pointers are 64 bits.
13740 The default depends on the specific target configuration. Note that
13741 the LP64 and ILP32 ABIs are not link-compatible; you must compile your
13742 entire program with the same ABI, and link with a compatible set of libraries.
13745 @opindex mbig-endian
13746 Generate big-endian code. This is the default when GCC is configured for an
13747 @samp{aarch64_be-*-*} target.
13749 @item -mgeneral-regs-only
13750 @opindex mgeneral-regs-only
13751 Generate code which uses only the general-purpose registers. This will prevent
13752 the compiler from using floating-point and Advanced SIMD registers but will not
13753 impose any restrictions on the assembler.
13755 @item -mlittle-endian
13756 @opindex mlittle-endian
13757 Generate little-endian code. This is the default when GCC is configured for an
13758 @samp{aarch64-*-*} but not an @samp{aarch64_be-*-*} target.
13760 @item -mcmodel=tiny
13761 @opindex mcmodel=tiny
13762 Generate code for the tiny code model. The program and its statically defined
13763 symbols must be within 1MB of each other. Programs can be statically or
13764 dynamically linked.
13766 @item -mcmodel=small
13767 @opindex mcmodel=small
13768 Generate code for the small code model. The program and its statically defined
13769 symbols must be within 4GB of each other. Programs can be statically or
13770 dynamically linked. This is the default code model.
13772 @item -mcmodel=large
13773 @opindex mcmodel=large
13774 Generate code for the large code model. This makes no assumptions about
13775 addresses and sizes of sections. Programs can be statically linked only.
13777 @item -mstrict-align
13778 @opindex mstrict-align
13779 Avoid generating memory accesses that may not be aligned on a natural object
13780 boundary as described in the architecture specification.
13782 @item -momit-leaf-frame-pointer
13783 @itemx -mno-omit-leaf-frame-pointer
13784 @opindex momit-leaf-frame-pointer
13785 @opindex mno-omit-leaf-frame-pointer
13786 Omit or keep the frame pointer in leaf functions. The former behavior is the
13789 @item -mtls-dialect=desc
13790 @opindex mtls-dialect=desc
13791 Use TLS descriptors as the thread-local storage mechanism for dynamic accesses
13792 of TLS variables. This is the default.
13794 @item -mtls-dialect=traditional
13795 @opindex mtls-dialect=traditional
13796 Use traditional TLS as the thread-local storage mechanism for dynamic accesses
13799 @item -mtls-size=@var{size}
13801 Specify bit size of immediate TLS offsets. Valid values are 12, 24, 32, 48.
13802 This option requires binutils 2.26 or newer.
13804 @item -mfix-cortex-a53-835769
13805 @itemx -mno-fix-cortex-a53-835769
13806 @opindex mfix-cortex-a53-835769
13807 @opindex mno-fix-cortex-a53-835769
13808 Enable or disable the workaround for the ARM Cortex-A53 erratum number 835769.
13809 This involves inserting a NOP instruction between memory instructions and
13810 64-bit integer multiply-accumulate instructions.
13812 @item -mfix-cortex-a53-843419
13813 @itemx -mno-fix-cortex-a53-843419
13814 @opindex mfix-cortex-a53-843419
13815 @opindex mno-fix-cortex-a53-843419
13816 Enable or disable the workaround for the ARM Cortex-A53 erratum number 843419.
13817 This erratum workaround is made at link time and this will only pass the
13818 corresponding flag to the linker.
13820 @item -mlow-precision-recip-sqrt
13821 @item -mno-low-precision-recip-sqrt
13822 @opindex mlow-precision-recip-sqrt
13823 @opindex mno-low-precision-recip-sqrt
13824 Enable or disable the reciprocal square root approximation.
13825 This option only has an effect if @option{-ffast-math} or
13826 @option{-funsafe-math-optimizations} is used as well. Enabling this reduces
13827 precision of reciprocal square root results to about 16 bits for
13828 single precision and to 32 bits for double precision.
13830 @item -mlow-precision-sqrt
13831 @item -mno-low-precision-sqrt
13832 @opindex -mlow-precision-sqrt
13833 @opindex -mno-low-precision-sqrt
13834 Enable or disable the square root approximation.
13835 This option only has an effect if @option{-ffast-math} or
13836 @option{-funsafe-math-optimizations} is used as well. Enabling this reduces
13837 precision of square root results to about 16 bits for
13838 single precision and to 32 bits for double precision.
13839 If enabled, it implies @option{-mlow-precision-recip-sqrt}.
13841 @item -mlow-precision-div
13842 @item -mno-low-precision-div
13843 @opindex -mlow-precision-div
13844 @opindex -mno-low-precision-div
13845 Enable or disable the division approximation.
13846 This option only has an effect if @option{-ffast-math} or
13847 @option{-funsafe-math-optimizations} is used as well. Enabling this reduces
13848 precision of division results to about 16 bits for
13849 single precision and to 32 bits for double precision.
13851 @item -march=@var{name}
13853 Specify the name of the target architecture and, optionally, one or
13854 more feature modifiers. This option has the form
13855 @option{-march=@var{arch}@r{@{}+@r{[}no@r{]}@var{feature}@r{@}*}}.
13857 The permissible values for @var{arch} are @samp{armv8-a},
13858 @samp{armv8.1-a}, @samp{armv8.2-a} or @var{native}.
13860 The value @samp{armv8.2-a} implies @samp{armv8.1-a} and enables compiler
13861 support for the ARMv8.2-A architecture extensions.
13863 The value @samp{armv8.1-a} implies @samp{armv8-a} and enables compiler
13864 support for the ARMv8.1-A architecture extension. In particular, it
13865 enables the @samp{+crc} and @samp{+lse} features.
13867 The value @samp{native} is available on native AArch64 GNU/Linux and
13868 causes the compiler to pick the architecture of the host system. This
13869 option has no effect if the compiler is unable to recognize the
13870 architecture of the host system,
13872 The permissible values for @var{feature} are listed in the sub-section
13873 on @ref{aarch64-feature-modifiers,,@option{-march} and @option{-mcpu}
13874 Feature Modifiers}. Where conflicting feature modifiers are
13875 specified, the right-most feature is used.
13877 GCC uses @var{name} to determine what kind of instructions it can emit
13878 when generating assembly code. If @option{-march} is specified
13879 without either of @option{-mtune} or @option{-mcpu} also being
13880 specified, the code is tuned to perform well across a range of target
13881 processors implementing the target architecture.
13883 @item -mtune=@var{name}
13885 Specify the name of the target processor for which GCC should tune the
13886 performance of the code. Permissible values for this option are:
13887 @samp{generic}, @samp{cortex-a35}, @samp{cortex-a53}, @samp{cortex-a57},
13888 @samp{cortex-a72}, @samp{cortex-a73}, @samp{exynos-m1}, @samp{falkor},
13889 @samp{qdf24xx}, @samp{thunderx}, @samp{xgene1}, @samp{vulcan},
13890 @samp{cortex-a57.cortex-a53}, @samp{cortex-a72.cortex-a53},
13891 @samp{cortex-a73.cortex-a35}, @samp{cortex-a73.cortex-a53}, @samp{native}.
13893 The values @samp{cortex-a57.cortex-a53}, @samp{cortex-a72.cortex-a53},
13894 @samp{cortex-a73.cortex-a35}, @samp{cortex-a73.cortex-a53}
13895 specify that GCC should tune for a big.LITTLE system.
13897 Additionally on native AArch64 GNU/Linux systems the value
13898 @samp{native} tunes performance to the host system. This option has no effect
13899 if the compiler is unable to recognize the processor of the host system.
13901 Where none of @option{-mtune=}, @option{-mcpu=} or @option{-march=}
13902 are specified, the code is tuned to perform well across a range
13903 of target processors.
13905 This option cannot be suffixed by feature modifiers.
13907 @item -mcpu=@var{name}
13909 Specify the name of the target processor, optionally suffixed by one
13910 or more feature modifiers. This option has the form
13911 @option{-mcpu=@var{cpu}@r{@{}+@r{[}no@r{]}@var{feature}@r{@}*}}, where
13912 the permissible values for @var{cpu} are the same as those available
13913 for @option{-mtune}. The permissible values for @var{feature} are
13914 documented in the sub-section on
13915 @ref{aarch64-feature-modifiers,,@option{-march} and @option{-mcpu}
13916 Feature Modifiers}. Where conflicting feature modifiers are
13917 specified, the right-most feature is used.
13919 GCC uses @var{name} to determine what kind of instructions it can emit when
13920 generating assembly code (as if by @option{-march}) and to determine
13921 the target processor for which to tune for performance (as if
13922 by @option{-mtune}). Where this option is used in conjunction
13923 with @option{-march} or @option{-mtune}, those options take precedence
13924 over the appropriate part of this option.
13926 @item -moverride=@var{string}
13928 Override tuning decisions made by the back-end in response to a
13929 @option{-mtune=} switch. The syntax, semantics, and accepted values
13930 for @var{string} in this option are not guaranteed to be consistent
13933 This option is only intended to be useful when developing GCC.
13935 @item -mpc-relative-literal-loads
13936 @opindex mpc-relative-literal-loads
13937 Enable PC-relative literal loads. With this option literal pools are
13938 accessed using a single instruction and emitted after each function. This
13939 limits the maximum size of functions to 1MB. This is enabled by default for
13940 @option{-mcmodel=tiny}.
13944 @subsubsection @option{-march} and @option{-mcpu} Feature Modifiers
13945 @anchor{aarch64-feature-modifiers}
13946 @cindex @option{-march} feature modifiers
13947 @cindex @option{-mcpu} feature modifiers
13948 Feature modifiers used with @option{-march} and @option{-mcpu} can be any of
13949 the following and their inverses @option{no@var{feature}}:
13953 Enable CRC extension. This is on by default for
13954 @option{-march=armv8.1-a}.
13956 Enable Crypto extension. This also enables Advanced SIMD and floating-point
13959 Enable floating-point instructions. This is on by default for all possible
13960 values for options @option{-march} and @option{-mcpu}.
13962 Enable Advanced SIMD instructions. This also enables floating-point
13963 instructions. This is on by default for all possible values for options
13964 @option{-march} and @option{-mcpu}.
13966 Enable Large System Extension instructions. This is on by default for
13967 @option{-march=armv8.1-a}.
13969 Enable FP16 extension. This also enables floating-point instructions.
13973 Feature @option{crypto} implies @option{simd}, which implies @option{fp}.
13974 Conversely, @option{nofp} implies @option{nosimd}, which implies
13977 @node Adapteva Epiphany Options
13978 @subsection Adapteva Epiphany Options
13980 These @samp{-m} options are defined for Adapteva Epiphany:
13983 @item -mhalf-reg-file
13984 @opindex mhalf-reg-file
13985 Don't allocate any register in the range @code{r32}@dots{}@code{r63}.
13986 That allows code to run on hardware variants that lack these registers.
13988 @item -mprefer-short-insn-regs
13989 @opindex mprefer-short-insn-regs
13990 Preferentially allocate registers that allow short instruction generation.
13991 This can result in increased instruction count, so this may either reduce or
13992 increase overall code size.
13994 @item -mbranch-cost=@var{num}
13995 @opindex mbranch-cost
13996 Set the cost of branches to roughly @var{num} ``simple'' instructions.
13997 This cost is only a heuristic and is not guaranteed to produce
13998 consistent results across releases.
14002 Enable the generation of conditional moves.
14004 @item -mnops=@var{num}
14006 Emit @var{num} NOPs before every other generated instruction.
14008 @item -mno-soft-cmpsf
14009 @opindex mno-soft-cmpsf
14010 For single-precision floating-point comparisons, emit an @code{fsub} instruction
14011 and test the flags. This is faster than a software comparison, but can
14012 get incorrect results in the presence of NaNs, or when two different small
14013 numbers are compared such that their difference is calculated as zero.
14014 The default is @option{-msoft-cmpsf}, which uses slower, but IEEE-compliant,
14015 software comparisons.
14017 @item -mstack-offset=@var{num}
14018 @opindex mstack-offset
14019 Set the offset between the top of the stack and the stack pointer.
14020 E.g., a value of 8 means that the eight bytes in the range @code{sp+0@dots{}sp+7}
14021 can be used by leaf functions without stack allocation.
14022 Values other than @samp{8} or @samp{16} are untested and unlikely to work.
14023 Note also that this option changes the ABI; compiling a program with a
14024 different stack offset than the libraries have been compiled with
14025 generally does not work.
14026 This option can be useful if you want to evaluate if a different stack
14027 offset would give you better code, but to actually use a different stack
14028 offset to build working programs, it is recommended to configure the
14029 toolchain with the appropriate @option{--with-stack-offset=@var{num}} option.
14031 @item -mno-round-nearest
14032 @opindex mno-round-nearest
14033 Make the scheduler assume that the rounding mode has been set to
14034 truncating. The default is @option{-mround-nearest}.
14037 @opindex mlong-calls
14038 If not otherwise specified by an attribute, assume all calls might be beyond
14039 the offset range of the @code{b} / @code{bl} instructions, and therefore load the
14040 function address into a register before performing a (otherwise direct) call.
14041 This is the default.
14043 @item -mshort-calls
14044 @opindex short-calls
14045 If not otherwise specified by an attribute, assume all direct calls are
14046 in the range of the @code{b} / @code{bl} instructions, so use these instructions
14047 for direct calls. The default is @option{-mlong-calls}.
14051 Assume addresses can be loaded as 16-bit unsigned values. This does not
14052 apply to function addresses for which @option{-mlong-calls} semantics
14055 @item -mfp-mode=@var{mode}
14057 Set the prevailing mode of the floating-point unit.
14058 This determines the floating-point mode that is provided and expected
14059 at function call and return time. Making this mode match the mode you
14060 predominantly need at function start can make your programs smaller and
14061 faster by avoiding unnecessary mode switches.
14063 @var{mode} can be set to one the following values:
14067 Any mode at function entry is valid, and retained or restored when
14068 the function returns, and when it calls other functions.
14069 This mode is useful for compiling libraries or other compilation units
14070 you might want to incorporate into different programs with different
14071 prevailing FPU modes, and the convenience of being able to use a single
14072 object file outweighs the size and speed overhead for any extra
14073 mode switching that might be needed, compared with what would be needed
14074 with a more specific choice of prevailing FPU mode.
14077 This is the mode used for floating-point calculations with
14078 truncating (i.e.@: round towards zero) rounding mode. That includes
14079 conversion from floating point to integer.
14081 @item round-nearest
14082 This is the mode used for floating-point calculations with
14083 round-to-nearest-or-even rounding mode.
14086 This is the mode used to perform integer calculations in the FPU, e.g.@:
14087 integer multiply, or integer multiply-and-accumulate.
14090 The default is @option{-mfp-mode=caller}
14092 @item -mnosplit-lohi
14093 @itemx -mno-postinc
14094 @itemx -mno-postmodify
14095 @opindex mnosplit-lohi
14096 @opindex mno-postinc
14097 @opindex mno-postmodify
14098 Code generation tweaks that disable, respectively, splitting of 32-bit
14099 loads, generation of post-increment addresses, and generation of
14100 post-modify addresses. The defaults are @option{msplit-lohi},
14101 @option{-mpost-inc}, and @option{-mpost-modify}.
14103 @item -mnovect-double
14104 @opindex mno-vect-double
14105 Change the preferred SIMD mode to SImode. The default is
14106 @option{-mvect-double}, which uses DImode as preferred SIMD mode.
14108 @item -max-vect-align=@var{num}
14109 @opindex max-vect-align
14110 The maximum alignment for SIMD vector mode types.
14111 @var{num} may be 4 or 8. The default is 8.
14112 Note that this is an ABI change, even though many library function
14113 interfaces are unaffected if they don't use SIMD vector modes
14114 in places that affect size and/or alignment of relevant types.
14116 @item -msplit-vecmove-early
14117 @opindex msplit-vecmove-early
14118 Split vector moves into single word moves before reload. In theory this
14119 can give better register allocation, but so far the reverse seems to be
14120 generally the case.
14122 @item -m1reg-@var{reg}
14124 Specify a register to hold the constant @minus{}1, which makes loading small negative
14125 constants and certain bitmasks faster.
14126 Allowable values for @var{reg} are @samp{r43} and @samp{r63},
14127 which specify use of that register as a fixed register,
14128 and @samp{none}, which means that no register is used for this
14129 purpose. The default is @option{-m1reg-none}.
14134 @subsection ARC Options
14135 @cindex ARC options
14137 The following options control the architecture variant for which code
14140 @c architecture variants
14143 @item -mbarrel-shifter
14144 @opindex mbarrel-shifter
14145 Generate instructions supported by barrel shifter. This is the default
14146 unless @option{-mcpu=ARC601} or @samp{-mcpu=ARCEM} is in effect.
14148 @item -mcpu=@var{cpu}
14150 Set architecture type, register usage, and instruction scheduling
14151 parameters for @var{cpu}. There are also shortcut alias options
14152 available for backward compatibility and convenience. Supported
14153 values for @var{cpu} are
14160 Compile for ARC600. Aliases: @option{-mA6}, @option{-mARC600}.
14165 Compile for ARC601. Alias: @option{-mARC601}.
14171 Compile for ARC700. Aliases: @option{-mA7}, @option{-mARC700}.
14172 This is the default when configured with @option{--with-cpu=arc700}@.
14176 Compile for ARC EM.
14180 Compile for ARC HS.
14185 @itemx -mdpfp-compact
14186 @opindex mdpfp-compact
14187 FPX: Generate Double Precision FPX instructions, tuned for the compact
14191 @opindex mdpfp-fast
14192 FPX: Generate Double Precision FPX instructions, tuned for the fast
14195 @item -mno-dpfp-lrsr
14196 @opindex mno-dpfp-lrsr
14197 Disable LR and SR instructions from using FPX extension aux registers.
14201 Generate Extended arithmetic instructions. Currently only
14202 @code{divaw}, @code{adds}, @code{subs}, and @code{sat16} are
14203 supported. This is always enabled for @option{-mcpu=ARC700}.
14207 Do not generate mpy instructions for ARC700.
14211 Generate 32x16 bit multiply and mac instructions.
14215 Generate mul64 and mulu64 instructions. Only valid for @option{-mcpu=ARC600}.
14219 Generate norm instruction. This is the default if @option{-mcpu=ARC700}
14224 @itemx -mspfp-compact
14225 @opindex mspfp-compact
14226 FPX: Generate Single Precision FPX instructions, tuned for the compact
14230 @opindex mspfp-fast
14231 FPX: Generate Single Precision FPX instructions, tuned for the fast
14236 Enable generation of ARC SIMD instructions via target-specific
14237 builtins. Only valid for @option{-mcpu=ARC700}.
14240 @opindex msoft-float
14241 This option ignored; it is provided for compatibility purposes only.
14242 Software floating point code is emitted by default, and this default
14243 can overridden by FPX options; @samp{mspfp}, @samp{mspfp-compact}, or
14244 @samp{mspfp-fast} for single precision, and @samp{mdpfp},
14245 @samp{mdpfp-compact}, or @samp{mdpfp-fast} for double precision.
14249 Generate swap instructions.
14253 This enables Locked Load/Store Conditional extension to implement
14254 atomic memopry built-in functions. Not available for ARC 6xx or ARC
14259 Enable DIV/REM instructions for ARCv2 cores.
14261 @item -mcode-density
14262 @opindex mcode-density
14263 Enable code density instructions for ARC EM, default on for ARC HS.
14267 Enable double load/store operations for ARC HS cores.
14269 @item -mtp-regno=@var{regno}
14271 Specify thread pointer register number.
14273 @item -mmpy-option=@var{multo}
14274 @opindex mmpy-option
14275 Compile ARCv2 code with a multiplier design option. @samp{wlh1} is
14276 the default value. The recognized values for @var{multo} are:
14280 No multiplier available.
14284 The multiply option is set to w: 16x16 multiplier, fully pipelined.
14285 The following instructions are enabled: MPYW, and MPYUW.
14289 The multiply option is set to wlh1: 32x32 multiplier, fully
14290 pipelined (1 stage). The following instructions are additionally
14291 enabled: MPY, MPYU, MPYM, MPYMU, and MPY_S.
14295 The multiply option is set to wlh2: 32x32 multiplier, fully pipelined
14296 (2 stages). The following instructions are additionally enabled: MPY,
14297 MPYU, MPYM, MPYMU, and MPY_S.
14301 The multiply option is set to wlh3: Two 16x16 multiplier, blocking,
14302 sequential. The following instructions are additionally enabled: MPY,
14303 MPYU, MPYM, MPYMU, and MPY_S.
14307 The multiply option is set to wlh4: One 16x16 multiplier, blocking,
14308 sequential. The following instructions are additionally enabled: MPY,
14309 MPYU, MPYM, MPYMU, and MPY_S.
14313 The multiply option is set to wlh5: One 32x4 multiplier, blocking,
14314 sequential. The following instructions are additionally enabled: MPY,
14315 MPYU, MPYM, MPYMU, and MPY_S.
14319 This option is only available for ARCv2 cores@.
14321 @item -mfpu=@var{fpu}
14323 Enables specific floating-point hardware extension for ARCv2
14324 core. Supported values for @var{fpu} are:
14330 Enables support for single precision floating point hardware
14335 Enables support for double precision floating point hardware
14336 extensions. The single precision floating point extension is also
14337 enabled. Not available for ARC EM@.
14341 Enables support for double precision floating point hardware
14342 extensions using double precision assist instructions. The single
14343 precision floating point extension is also enabled. This option is
14344 only available for ARC EM@.
14348 Enables support for double precision floating point hardware
14349 extensions using double precision assist instructions, and simple
14350 precision square-root and divide hardware extensions. The single
14351 precision floating point extension is also enabled. This option is
14352 only available for ARC EM@.
14356 Enables support for double precision floating point hardware
14357 extensions using double precision assist instructions, and simple
14358 precision fused multiple and add hardware extension. The single
14359 precision floating point extension is also enabled. This option is
14360 only available for ARC EM@.
14364 Enables support for double precision floating point hardware
14365 extensions using double precision assist instructions, and all simple
14366 precision hardware extensions. The single precision floating point
14367 extension is also enabled. This option is only available for ARC EM@.
14371 Enables support for single precision floating point, and single
14372 precision square-root and divide hardware extensions@.
14376 Enables support for double precision floating point, and double
14377 precision square-root and divide hardware extensions. This option
14378 includes option @samp{fpus_div}. Not available for ARC EM@.
14382 Enables support for single precision floating point, and single
14383 precision fused multiple and add hardware extensions@.
14387 Enables support for double precision floating point, and double
14388 precision fused multiple and add hardware extensions. This option
14389 includes option @samp{fpus_fma}. Not available for ARC EM@.
14393 Enables support for all single precision floating point hardware
14398 Enables support for all single and double precision floating point
14399 hardware extensions. Not available for ARC EM@.
14405 The following options are passed through to the assembler, and also
14406 define preprocessor macro symbols.
14408 @c Flags used by the assembler, but for which we define preprocessor
14409 @c macro symbols as well.
14412 @opindex mdsp-packa
14413 Passed down to the assembler to enable the DSP Pack A extensions.
14414 Also sets the preprocessor symbol @code{__Xdsp_packa}.
14418 Passed down to the assembler to enable the dual viterbi butterfly
14419 extension. Also sets the preprocessor symbol @code{__Xdvbf}.
14421 @c ARC700 4.10 extension instruction
14424 Passed down to the assembler to enable the Locked Load/Store
14425 Conditional extension. Also sets the preprocessor symbol
14430 Passed down to the assembler. Also sets the preprocessor symbol
14431 @code{__Xxmac_d16}.
14435 Passed down to the assembler. Also sets the preprocessor symbol
14438 @c ARC700 4.10 extension instruction
14441 Passed down to the assembler to enable the 64-bit Time-Stamp Counter
14442 extension instruction. Also sets the preprocessor symbol
14445 @c ARC700 4.10 extension instruction
14448 Passed down to the assembler to enable the swap byte ordering
14449 extension instruction. Also sets the preprocessor symbol
14453 @opindex mtelephony
14454 Passed down to the assembler to enable dual and single operand
14455 instructions for telephony. Also sets the preprocessor symbol
14456 @code{__Xtelephony}.
14460 Passed down to the assembler to enable the XY Memory extension. Also
14461 sets the preprocessor symbol @code{__Xxy}.
14465 The following options control how the assembly code is annotated:
14467 @c Assembly annotation options
14471 Annotate assembler instructions with estimated addresses.
14473 @item -mannotate-align
14474 @opindex mannotate-align
14475 Explain what alignment considerations lead to the decision to make an
14476 instruction short or long.
14480 The following options are passed through to the linker:
14482 @c options passed through to the linker
14486 Passed through to the linker, to specify use of the @code{arclinux} emulation.
14487 This option is enabled by default in tool chains built for
14488 @w{@code{arc-linux-uclibc}} and @w{@code{arceb-linux-uclibc}} targets
14489 when profiling is not requested.
14491 @item -marclinux_prof
14492 @opindex marclinux_prof
14493 Passed through to the linker, to specify use of the
14494 @code{arclinux_prof} emulation. This option is enabled by default in
14495 tool chains built for @w{@code{arc-linux-uclibc}} and
14496 @w{@code{arceb-linux-uclibc}} targets when profiling is requested.
14500 The following options control the semantics of generated code:
14502 @c semantically relevant code generation options
14505 @opindex mlong-calls
14506 Generate call insns as register indirect calls, thus providing access
14507 to the full 32-bit address range.
14509 @item -mmedium-calls
14510 @opindex mmedium-calls
14511 Don't use less than 25 bit addressing range for calls, which is the
14512 offset available for an unconditional branch-and-link
14513 instruction. Conditional execution of function calls is suppressed, to
14514 allow use of the 25-bit range, rather than the 21-bit range with
14515 conditional branch-and-link. This is the default for tool chains built
14516 for @w{@code{arc-linux-uclibc}} and @w{@code{arceb-linux-uclibc}} targets.
14520 Do not generate sdata references. This is the default for tool chains
14521 built for @w{@code{arc-linux-uclibc}} and @w{@code{arceb-linux-uclibc}}
14525 @opindex mucb-mcount
14526 Instrument with mcount calls as used in UCB code. I.e. do the
14527 counting in the callee, not the caller. By default ARC instrumentation
14528 counts in the caller.
14530 @item -mvolatile-cache
14531 @opindex mvolatile-cache
14532 Use ordinarily cached memory accesses for volatile references. This is the
14535 @item -mno-volatile-cache
14536 @opindex mno-volatile-cache
14537 Enable cache bypass for volatile references.
14541 The following options fine tune code generation:
14542 @c code generation tuning options
14545 @opindex malign-call
14546 Do alignment optimizations for call instructions.
14548 @item -mauto-modify-reg
14549 @opindex mauto-modify-reg
14550 Enable the use of pre/post modify with register displacement.
14552 @item -mbbit-peephole
14553 @opindex mbbit-peephole
14554 Enable bbit peephole2.
14558 This option disables a target-specific pass in @file{arc_reorg} to
14559 generate @code{BRcc} instructions. It has no effect on @code{BRcc}
14560 generation driven by the combiner pass.
14562 @item -mcase-vector-pcrel
14563 @opindex mcase-vector-pcrel
14564 Use pc-relative switch case tables - this enables case table shortening.
14565 This is the default for @option{-Os}.
14567 @item -mcompact-casesi
14568 @opindex mcompact-casesi
14569 Enable compact casesi pattern. This is the default for @option{-Os},
14570 and only available for ARCv1 cores.
14572 @item -mno-cond-exec
14573 @opindex mno-cond-exec
14574 Disable ARCompact specific pass to generate conditional execution instructions.
14575 Due to delay slot scheduling and interactions between operand numbers,
14576 literal sizes, instruction lengths, and the support for conditional execution,
14577 the target-independent pass to generate conditional execution is often lacking,
14578 so the ARC port has kept a special pass around that tries to find more
14579 conditional execution generating opportunities after register allocation,
14580 branch shortening, and delay slot scheduling have been done. This pass
14581 generally, but not always, improves performance and code size, at the cost of
14582 extra compilation time, which is why there is an option to switch it off.
14583 If you have a problem with call instructions exceeding their allowable
14584 offset range because they are conditionalized, you should consider using
14585 @option{-mmedium-calls} instead.
14587 @item -mearly-cbranchsi
14588 @opindex mearly-cbranchsi
14589 Enable pre-reload use of the cbranchsi pattern.
14591 @item -mexpand-adddi
14592 @opindex mexpand-adddi
14593 Expand @code{adddi3} and @code{subdi3} at rtl generation time into
14594 @code{add.f}, @code{adc} etc.
14596 @item -mindexed-loads
14597 @opindex mindexed-loads
14598 Enable the use of indexed loads. This can be problematic because some
14599 optimizers then assume that indexed stores exist, which is not
14603 Enable Local Register Allocation. This is still experimental for ARC,
14604 so by default the compiler uses standard reload
14605 (i.e. @option{-mno-lra}).
14607 @item -mlra-priority-none
14608 @opindex mlra-priority-none
14609 Don't indicate any priority for target registers.
14611 @item -mlra-priority-compact
14612 @opindex mlra-priority-compact
14613 Indicate target register priority for r0..r3 / r12..r15.
14615 @item -mlra-priority-noncompact
14616 @opindex mlra-priority-noncompact
14617 Reduce target register priority for r0..r3 / r12..r15.
14619 @item -mno-millicode
14620 @opindex mno-millicode
14621 When optimizing for size (using @option{-Os}), prologues and epilogues
14622 that have to save or restore a large number of registers are often
14623 shortened by using call to a special function in libgcc; this is
14624 referred to as a @emph{millicode} call. As these calls can pose
14625 performance issues, and/or cause linking issues when linking in a
14626 nonstandard way, this option is provided to turn off millicode call
14630 @opindex mmixed-code
14631 Tweak register allocation to help 16-bit instruction generation.
14632 This generally has the effect of decreasing the average instruction size
14633 while increasing the instruction count.
14637 Enable 'q' instruction alternatives.
14638 This is the default for @option{-Os}.
14642 Enable Rcq constraint handling - most short code generation depends on this.
14643 This is the default.
14647 Enable Rcw constraint handling - ccfsm condexec mostly depends on this.
14648 This is the default.
14650 @item -msize-level=@var{level}
14651 @opindex msize-level
14652 Fine-tune size optimization with regards to instruction lengths and alignment.
14653 The recognized values for @var{level} are:
14656 No size optimization. This level is deprecated and treated like @samp{1}.
14659 Short instructions are used opportunistically.
14662 In addition, alignment of loops and of code after barriers are dropped.
14665 In addition, optional data alignment is dropped, and the option @option{Os} is enabled.
14669 This defaults to @samp{3} when @option{-Os} is in effect. Otherwise,
14670 the behavior when this is not set is equivalent to level @samp{1}.
14672 @item -mtune=@var{cpu}
14674 Set instruction scheduling parameters for @var{cpu}, overriding any implied
14675 by @option{-mcpu=}.
14677 Supported values for @var{cpu} are
14681 Tune for ARC600 cpu.
14684 Tune for ARC601 cpu.
14687 Tune for ARC700 cpu with standard multiplier block.
14690 Tune for ARC700 cpu with XMAC block.
14693 Tune for ARC725D cpu.
14696 Tune for ARC750D cpu.
14700 @item -mmultcost=@var{num}
14702 Cost to assume for a multiply instruction, with @samp{4} being equal to a
14703 normal instruction.
14705 @item -munalign-prob-threshold=@var{probability}
14706 @opindex munalign-prob-threshold
14707 Set probability threshold for unaligning branches.
14708 When tuning for @samp{ARC700} and optimizing for speed, branches without
14709 filled delay slot are preferably emitted unaligned and long, unless
14710 profiling indicates that the probability for the branch to be taken
14711 is below @var{probability}. @xref{Cross-profiling}.
14712 The default is (REG_BR_PROB_BASE/2), i.e.@: 5000.
14716 The following options are maintained for backward compatibility, but
14717 are now deprecated and will be removed in a future release:
14719 @c Deprecated options
14727 @opindex mbig-endian
14730 Compile code for big endian targets. Use of these options is now
14731 deprecated. Users wanting big-endian code, should use the
14732 @w{@code{arceb-elf32}} and @w{@code{arceb-linux-uclibc}} targets when
14733 building the tool chain, for which big-endian is the default.
14735 @item -mlittle-endian
14736 @opindex mlittle-endian
14739 Compile code for little endian targets. Use of these options is now
14740 deprecated. Users wanting little-endian code should use the
14741 @w{@code{arc-elf32}} and @w{@code{arc-linux-uclibc}} targets when
14742 building the tool chain, for which little-endian is the default.
14744 @item -mbarrel_shifter
14745 @opindex mbarrel_shifter
14746 Replaced by @option{-mbarrel-shifter}.
14748 @item -mdpfp_compact
14749 @opindex mdpfp_compact
14750 Replaced by @option{-mdpfp-compact}.
14753 @opindex mdpfp_fast
14754 Replaced by @option{-mdpfp-fast}.
14757 @opindex mdsp_packa
14758 Replaced by @option{-mdsp-packa}.
14762 Replaced by @option{-mea}.
14766 Replaced by @option{-mmac-24}.
14770 Replaced by @option{-mmac-d16}.
14772 @item -mspfp_compact
14773 @opindex mspfp_compact
14774 Replaced by @option{-mspfp-compact}.
14777 @opindex mspfp_fast
14778 Replaced by @option{-mspfp-fast}.
14780 @item -mtune=@var{cpu}
14782 Values @samp{arc600}, @samp{arc601}, @samp{arc700} and
14783 @samp{arc700-xmac} for @var{cpu} are replaced by @samp{ARC600},
14784 @samp{ARC601}, @samp{ARC700} and @samp{ARC700-xmac} respectively
14786 @item -multcost=@var{num}
14788 Replaced by @option{-mmultcost}.
14793 @subsection ARM Options
14794 @cindex ARM options
14796 These @samp{-m} options are defined for the ARM port:
14799 @item -mabi=@var{name}
14801 Generate code for the specified ABI@. Permissible values are: @samp{apcs-gnu},
14802 @samp{atpcs}, @samp{aapcs}, @samp{aapcs-linux} and @samp{iwmmxt}.
14805 @opindex mapcs-frame
14806 Generate a stack frame that is compliant with the ARM Procedure Call
14807 Standard for all functions, even if this is not strictly necessary for
14808 correct execution of the code. Specifying @option{-fomit-frame-pointer}
14809 with this option causes the stack frames not to be generated for
14810 leaf functions. The default is @option{-mno-apcs-frame}.
14811 This option is deprecated.
14815 This is a synonym for @option{-mapcs-frame} and is deprecated.
14818 @c not currently implemented
14819 @item -mapcs-stack-check
14820 @opindex mapcs-stack-check
14821 Generate code to check the amount of stack space available upon entry to
14822 every function (that actually uses some stack space). If there is
14823 insufficient space available then either the function
14824 @code{__rt_stkovf_split_small} or @code{__rt_stkovf_split_big} is
14825 called, depending upon the amount of stack space required. The runtime
14826 system is required to provide these functions. The default is
14827 @option{-mno-apcs-stack-check}, since this produces smaller code.
14829 @c not currently implemented
14831 @opindex mapcs-float
14832 Pass floating-point arguments using the floating-point registers. This is
14833 one of the variants of the APCS@. This option is recommended if the
14834 target hardware has a floating-point unit or if a lot of floating-point
14835 arithmetic is going to be performed by the code. The default is
14836 @option{-mno-apcs-float}, since the size of integer-only code is
14837 slightly increased if @option{-mapcs-float} is used.
14839 @c not currently implemented
14840 @item -mapcs-reentrant
14841 @opindex mapcs-reentrant
14842 Generate reentrant, position-independent code. The default is
14843 @option{-mno-apcs-reentrant}.
14846 @item -mthumb-interwork
14847 @opindex mthumb-interwork
14848 Generate code that supports calling between the ARM and Thumb
14849 instruction sets. Without this option, on pre-v5 architectures, the
14850 two instruction sets cannot be reliably used inside one program. The
14851 default is @option{-mno-thumb-interwork}, since slightly larger code
14852 is generated when @option{-mthumb-interwork} is specified. In AAPCS
14853 configurations this option is meaningless.
14855 @item -mno-sched-prolog
14856 @opindex mno-sched-prolog
14857 Prevent the reordering of instructions in the function prologue, or the
14858 merging of those instruction with the instructions in the function's
14859 body. This means that all functions start with a recognizable set
14860 of instructions (or in fact one of a choice from a small set of
14861 different function prologues), and this information can be used to
14862 locate the start of functions inside an executable piece of code. The
14863 default is @option{-msched-prolog}.
14865 @item -mfloat-abi=@var{name}
14866 @opindex mfloat-abi
14867 Specifies which floating-point ABI to use. Permissible values
14868 are: @samp{soft}, @samp{softfp} and @samp{hard}.
14870 Specifying @samp{soft} causes GCC to generate output containing
14871 library calls for floating-point operations.
14872 @samp{softfp} allows the generation of code using hardware floating-point
14873 instructions, but still uses the soft-float calling conventions.
14874 @samp{hard} allows generation of floating-point instructions
14875 and uses FPU-specific calling conventions.
14877 The default depends on the specific target configuration. Note that
14878 the hard-float and soft-float ABIs are not link-compatible; you must
14879 compile your entire program with the same ABI, and link with a
14880 compatible set of libraries.
14882 @item -mlittle-endian
14883 @opindex mlittle-endian
14884 Generate code for a processor running in little-endian mode. This is
14885 the default for all standard configurations.
14888 @opindex mbig-endian
14889 Generate code for a processor running in big-endian mode; the default is
14890 to compile code for a little-endian processor.
14892 @item -march=@var{name}
14894 This specifies the name of the target ARM architecture. GCC uses this
14895 name to determine what kind of instructions it can emit when generating
14896 assembly code. This option can be used in conjunction with or instead
14897 of the @option{-mcpu=} option. Permissible names are: @samp{armv2},
14898 @samp{armv2a}, @samp{armv3}, @samp{armv3m}, @samp{armv4}, @samp{armv4t},
14899 @samp{armv5}, @samp{armv5e}, @samp{armv5t}, @samp{armv5te},
14900 @samp{armv6}, @samp{armv6-m}, @samp{armv6j}, @samp{armv6k},
14901 @samp{armv6kz}, @samp{armv6s-m},
14902 @samp{armv6t2}, @samp{armv6z}, @samp{armv6zk},
14903 @samp{armv7}, @samp{armv7-a}, @samp{armv7-m}, @samp{armv7-r}, @samp{armv7e-m},
14904 @samp{armv7ve}, @samp{armv8-a}, @samp{armv8-a+crc}, @samp{armv8.1-a},
14905 @samp{armv8.1-a+crc}, @samp{armv8-m.base}, @samp{armv8-m.main},
14906 @samp{armv8-m.main+dsp}, @samp{iwmmxt}, @samp{iwmmxt2}.
14908 Architecture revisions older than @samp{armv4t} are deprecated.
14910 @option{-march=armv6s-m} is the @samp{armv6-m} architecture with support for
14911 the (now mandatory) SVC instruction.
14913 @option{-march=armv6zk} is an alias for @samp{armv6kz}, existing for backwards
14916 @option{-march=armv7ve} is the @samp{armv7-a} architecture with virtualization
14919 @option{-march=armv8-a+crc} enables code generation for the ARMv8-A
14920 architecture together with the optional CRC32 extensions.
14922 @option{-march=armv8.1-a} enables compiler support for the ARMv8.1-A
14923 architecture. This also enables the features provided by
14924 @option{-march=armv8-a+crc}.
14926 @option{-march=armv8.2-a} enables compiler support for the ARMv8.2-A
14927 architecture. This also enables the features provided by
14928 @option{-march=armv8.1-a}.
14930 @option{-march=armv8.2-a+fp16} enables compiler support for the
14931 ARMv8.2-A architecture with the optional FP16 instructions extension.
14932 This also enables the features provided by @option{-march=armv8.1-a}
14933 and implies @option{-mfp16-format=ieee}.
14935 @option{-march=native} causes the compiler to auto-detect the architecture
14936 of the build computer. At present, this feature is only supported on
14937 GNU/Linux, and not all architectures are recognized. If the auto-detect
14938 is unsuccessful the option has no effect.
14940 @item -mtune=@var{name}
14942 This option specifies the name of the target ARM processor for
14943 which GCC should tune the performance of the code.
14944 For some ARM implementations better performance can be obtained by using
14946 Permissible names are: @samp{arm2}, @samp{arm250},
14947 @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610},
14948 @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm},
14949 @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700},
14950 @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100},
14952 @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm7tdmi-s},
14953 @samp{arm710t}, @samp{arm720t}, @samp{arm740t},
14954 @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100},
14955 @samp{strongarm1110},
14956 @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920},
14957 @samp{arm920t}, @samp{arm922t}, @samp{arm946e-s}, @samp{arm966e-s},
14958 @samp{arm968e-s}, @samp{arm926ej-s}, @samp{arm940t}, @samp{arm9tdmi},
14959 @samp{arm10tdmi}, @samp{arm1020t}, @samp{arm1026ej-s},
14960 @samp{arm10e}, @samp{arm1020e}, @samp{arm1022e},
14961 @samp{arm1136j-s}, @samp{arm1136jf-s}, @samp{mpcore}, @samp{mpcorenovfp},
14962 @samp{arm1156t2-s}, @samp{arm1156t2f-s}, @samp{arm1176jz-s}, @samp{arm1176jzf-s},
14963 @samp{generic-armv7-a}, @samp{cortex-a5}, @samp{cortex-a7}, @samp{cortex-a8},
14964 @samp{cortex-a9}, @samp{cortex-a12}, @samp{cortex-a15}, @samp{cortex-a17},
14965 @samp{cortex-a32}, @samp{cortex-a35}, @samp{cortex-a53}, @samp{cortex-a57},
14966 @samp{cortex-a72}, @samp{cortex-a73}, @samp{cortex-r4},
14967 @samp{cortex-r4f}, @samp{cortex-r5}, @samp{cortex-r7}, @samp{cortex-r8},
14975 @samp{cortex-m0plus},
14976 @samp{cortex-m1.small-multiply},
14977 @samp{cortex-m0.small-multiply},
14978 @samp{cortex-m0plus.small-multiply},
14982 @samp{marvell-pj4},
14983 @samp{xscale}, @samp{iwmmxt}, @samp{iwmmxt2}, @samp{ep9312},
14984 @samp{fa526}, @samp{fa626},
14985 @samp{fa606te}, @samp{fa626te}, @samp{fmp626}, @samp{fa726te},
14988 Additionally, this option can specify that GCC should tune the performance
14989 of the code for a big.LITTLE system. Permissible names are:
14990 @samp{cortex-a15.cortex-a7}, @samp{cortex-a17.cortex-a7},
14991 @samp{cortex-a57.cortex-a53}, @samp{cortex-a72.cortex-a53},
14992 @samp{cortex-a72.cortex-a35}, @samp{cortex-a73.cortex-a53}.
14994 @option{-mtune=generic-@var{arch}} specifies that GCC should tune the
14995 performance for a blend of processors within architecture @var{arch}.
14996 The aim is to generate code that run well on the current most popular
14997 processors, balancing between optimizations that benefit some CPUs in the
14998 range, and avoiding performance pitfalls of other CPUs. The effects of
14999 this option may change in future GCC versions as CPU models come and go.
15001 @option{-mtune=native} causes the compiler to auto-detect the CPU
15002 of the build computer. At present, this feature is only supported on
15003 GNU/Linux, and not all architectures are recognized. If the auto-detect is
15004 unsuccessful the option has no effect.
15006 @item -mcpu=@var{name}
15008 This specifies the name of the target ARM processor. GCC uses this name
15009 to derive the name of the target ARM architecture (as if specified
15010 by @option{-march}) and the ARM processor type for which to tune for
15011 performance (as if specified by @option{-mtune}). Where this option
15012 is used in conjunction with @option{-march} or @option{-mtune},
15013 those options take precedence over the appropriate part of this option.
15015 Permissible names for this option are the same as those for
15018 @option{-mcpu=generic-@var{arch}} is also permissible, and is
15019 equivalent to @option{-march=@var{arch} -mtune=generic-@var{arch}}.
15020 See @option{-mtune} for more information.
15022 @option{-mcpu=native} causes the compiler to auto-detect the CPU
15023 of the build computer. At present, this feature is only supported on
15024 GNU/Linux, and not all architectures are recognized. If the auto-detect
15025 is unsuccessful the option has no effect.
15027 @item -mfpu=@var{name}
15029 This specifies what floating-point hardware (or hardware emulation) is
15030 available on the target. Permissible names are: @samp{vfp}, @samp{vfpv3},
15031 @samp{vfpv3-fp16}, @samp{vfpv3-d16}, @samp{vfpv3-d16-fp16}, @samp{vfpv3xd},
15032 @samp{vfpv3xd-fp16}, @samp{neon}, @samp{neon-fp16}, @samp{vfpv4},
15033 @samp{vfpv4-d16}, @samp{fpv4-sp-d16}, @samp{neon-vfpv4},
15034 @samp{fpv5-d16}, @samp{fpv5-sp-d16},
15035 @samp{fp-armv8}, @samp{neon-fp-armv8} and @samp{crypto-neon-fp-armv8}.
15037 If @option{-msoft-float} is specified this specifies the format of
15038 floating-point values.
15040 If the selected floating-point hardware includes the NEON extension
15041 (e.g. @option{-mfpu}=@samp{neon}), note that floating-point
15042 operations are not generated by GCC's auto-vectorization pass unless
15043 @option{-funsafe-math-optimizations} is also specified. This is
15044 because NEON hardware does not fully implement the IEEE 754 standard for
15045 floating-point arithmetic (in particular denormal values are treated as
15046 zero), so the use of NEON instructions may lead to a loss of precision.
15048 You can also set the fpu name at function level by using the @code{target("fpu=")} function attributes (@pxref{ARM Function Attributes}) or pragmas (@pxref{Function Specific Option Pragmas}).
15050 @item -mfp16-format=@var{name}
15051 @opindex mfp16-format
15052 Specify the format of the @code{__fp16} half-precision floating-point type.
15053 Permissible names are @samp{none}, @samp{ieee}, and @samp{alternative};
15054 the default is @samp{none}, in which case the @code{__fp16} type is not
15055 defined. @xref{Half-Precision}, for more information.
15057 @item -mstructure-size-boundary=@var{n}
15058 @opindex mstructure-size-boundary
15059 The sizes of all structures and unions are rounded up to a multiple
15060 of the number of bits set by this option. Permissible values are 8, 32
15061 and 64. The default value varies for different toolchains. For the COFF
15062 targeted toolchain the default value is 8. A value of 64 is only allowed
15063 if the underlying ABI supports it.
15065 Specifying a larger number can produce faster, more efficient code, but
15066 can also increase the size of the program. Different values are potentially
15067 incompatible. Code compiled with one value cannot necessarily expect to
15068 work with code or libraries compiled with another value, if they exchange
15069 information using structures or unions.
15071 @item -mabort-on-noreturn
15072 @opindex mabort-on-noreturn
15073 Generate a call to the function @code{abort} at the end of a
15074 @code{noreturn} function. It is executed if the function tries to
15078 @itemx -mno-long-calls
15079 @opindex mlong-calls
15080 @opindex mno-long-calls
15081 Tells the compiler to perform function calls by first loading the
15082 address of the function into a register and then performing a subroutine
15083 call on this register. This switch is needed if the target function
15084 lies outside of the 64-megabyte addressing range of the offset-based
15085 version of subroutine call instruction.
15087 Even if this switch is enabled, not all function calls are turned
15088 into long calls. The heuristic is that static functions, functions
15089 that have the @code{short_call} attribute, functions that are inside
15090 the scope of a @code{#pragma no_long_calls} directive, and functions whose
15091 definitions have already been compiled within the current compilation
15092 unit are not turned into long calls. The exceptions to this rule are
15093 that weak function definitions, functions with the @code{long_call}
15094 attribute or the @code{section} attribute, and functions that are within
15095 the scope of a @code{#pragma long_calls} directive are always
15096 turned into long calls.
15098 This feature is not enabled by default. Specifying
15099 @option{-mno-long-calls} restores the default behavior, as does
15100 placing the function calls within the scope of a @code{#pragma
15101 long_calls_off} directive. Note these switches have no effect on how
15102 the compiler generates code to handle function calls via function
15105 @item -msingle-pic-base
15106 @opindex msingle-pic-base
15107 Treat the register used for PIC addressing as read-only, rather than
15108 loading it in the prologue for each function. The runtime system is
15109 responsible for initializing this register with an appropriate value
15110 before execution begins.
15112 @item -mpic-register=@var{reg}
15113 @opindex mpic-register
15114 Specify the register to be used for PIC addressing.
15115 For standard PIC base case, the default is any suitable register
15116 determined by compiler. For single PIC base case, the default is
15117 @samp{R9} if target is EABI based or stack-checking is enabled,
15118 otherwise the default is @samp{R10}.
15120 @item -mpic-data-is-text-relative
15121 @opindex mpic-data-is-text-relative
15122 Assume that the displacement between the text and data segments is fixed
15123 at static link time. This permits using PC-relative addressing
15124 operations to access data known to be in the data segment. For
15125 non-VxWorks RTP targets, this option is enabled by default. When
15126 disabled on such targets, it will enable @option{-msingle-pic-base} by
15129 @item -mpoke-function-name
15130 @opindex mpoke-function-name
15131 Write the name of each function into the text section, directly
15132 preceding the function prologue. The generated code is similar to this:
15136 .ascii "arm_poke_function_name", 0
15139 .word 0xff000000 + (t1 - t0)
15140 arm_poke_function_name
15142 stmfd sp!, @{fp, ip, lr, pc@}
15146 When performing a stack backtrace, code can inspect the value of
15147 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
15148 location @code{pc - 12} and the top 8 bits are set, then we know that
15149 there is a function name embedded immediately preceding this location
15150 and has length @code{((pc[-3]) & 0xff000000)}.
15157 Select between generating code that executes in ARM and Thumb
15158 states. The default for most configurations is to generate code
15159 that executes in ARM state, but the default can be changed by
15160 configuring GCC with the @option{--with-mode=}@var{state}
15163 You can also override the ARM and Thumb mode for each function
15164 by using the @code{target("thumb")} and @code{target("arm")} function attributes
15165 (@pxref{ARM Function Attributes}) or pragmas (@pxref{Function Specific Option Pragmas}).
15168 @opindex mtpcs-frame
15169 Generate a stack frame that is compliant with the Thumb Procedure Call
15170 Standard for all non-leaf functions. (A leaf function is one that does
15171 not call any other functions.) The default is @option{-mno-tpcs-frame}.
15173 @item -mtpcs-leaf-frame
15174 @opindex mtpcs-leaf-frame
15175 Generate a stack frame that is compliant with the Thumb Procedure Call
15176 Standard for all leaf functions. (A leaf function is one that does
15177 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
15179 @item -mcallee-super-interworking
15180 @opindex mcallee-super-interworking
15181 Gives all externally visible functions in the file being compiled an ARM
15182 instruction set header which switches to Thumb mode before executing the
15183 rest of the function. This allows these functions to be called from
15184 non-interworking code. This option is not valid in AAPCS configurations
15185 because interworking is enabled by default.
15187 @item -mcaller-super-interworking
15188 @opindex mcaller-super-interworking
15189 Allows calls via function pointers (including virtual functions) to
15190 execute correctly regardless of whether the target code has been
15191 compiled for interworking or not. There is a small overhead in the cost
15192 of executing a function pointer if this option is enabled. This option
15193 is not valid in AAPCS configurations because interworking is enabled
15196 @item -mtp=@var{name}
15198 Specify the access model for the thread local storage pointer. The valid
15199 models are @samp{soft}, which generates calls to @code{__aeabi_read_tp},
15200 @samp{cp15}, which fetches the thread pointer from @code{cp15} directly
15201 (supported in the arm6k architecture), and @samp{auto}, which uses the
15202 best available method for the selected processor. The default setting is
15205 @item -mtls-dialect=@var{dialect}
15206 @opindex mtls-dialect
15207 Specify the dialect to use for accessing thread local storage. Two
15208 @var{dialect}s are supported---@samp{gnu} and @samp{gnu2}. The
15209 @samp{gnu} dialect selects the original GNU scheme for supporting
15210 local and global dynamic TLS models. The @samp{gnu2} dialect
15211 selects the GNU descriptor scheme, which provides better performance
15212 for shared libraries. The GNU descriptor scheme is compatible with
15213 the original scheme, but does require new assembler, linker and
15214 library support. Initial and local exec TLS models are unaffected by
15215 this option and always use the original scheme.
15217 @item -mword-relocations
15218 @opindex mword-relocations
15219 Only generate absolute relocations on word-sized values (i.e. R_ARM_ABS32).
15220 This is enabled by default on targets (uClinux, SymbianOS) where the runtime
15221 loader imposes this restriction, and when @option{-fpic} or @option{-fPIC}
15224 @item -mfix-cortex-m3-ldrd
15225 @opindex mfix-cortex-m3-ldrd
15226 Some Cortex-M3 cores can cause data corruption when @code{ldrd} instructions
15227 with overlapping destination and base registers are used. This option avoids
15228 generating these instructions. This option is enabled by default when
15229 @option{-mcpu=cortex-m3} is specified.
15231 @item -munaligned-access
15232 @itemx -mno-unaligned-access
15233 @opindex munaligned-access
15234 @opindex mno-unaligned-access
15235 Enables (or disables) reading and writing of 16- and 32- bit values
15236 from addresses that are not 16- or 32- bit aligned. By default
15237 unaligned access is disabled for all pre-ARMv6, all ARMv6-M and for
15238 ARMv8-M Baseline architectures, and enabled for all other
15239 architectures. If unaligned access is not enabled then words in packed
15240 data structures are accessed a byte at a time.
15242 The ARM attribute @code{Tag_CPU_unaligned_access} is set in the
15243 generated object file to either true or false, depending upon the
15244 setting of this option. If unaligned access is enabled then the
15245 preprocessor symbol @code{__ARM_FEATURE_UNALIGNED} is also
15248 @item -mneon-for-64bits
15249 @opindex mneon-for-64bits
15250 Enables using Neon to handle scalar 64-bits operations. This is
15251 disabled by default since the cost of moving data from core registers
15254 @item -mslow-flash-data
15255 @opindex mslow-flash-data
15256 Assume loading data from flash is slower than fetching instruction.
15257 Therefore literal load is minimized for better performance.
15258 This option is only supported when compiling for ARMv7 M-profile and
15261 @item -masm-syntax-unified
15262 @opindex masm-syntax-unified
15263 Assume inline assembler is using unified asm syntax. The default is
15264 currently off which implies divided syntax. This option has no impact
15265 on Thumb2. However, this may change in future releases of GCC.
15266 Divided syntax should be considered deprecated.
15268 @item -mrestrict-it
15269 @opindex mrestrict-it
15270 Restricts generation of IT blocks to conform to the rules of ARMv8.
15271 IT blocks can only contain a single 16-bit instruction from a select
15272 set of instructions. This option is on by default for ARMv8 Thumb mode.
15274 @item -mprint-tune-info
15275 @opindex mprint-tune-info
15276 Print CPU tuning information as comment in assembler file. This is
15277 an option used only for regression testing of the compiler and not
15278 intended for ordinary use in compiling code. This option is disabled
15282 @opindex mpure-code
15283 Do not allow constant data to be placed in code sections.
15284 Additionally, when compiling for ELF object format give all text sections the
15285 ELF processor-specific section attribute @code{SHF_ARM_PURECODE}. This option
15286 is only available when generating non-pic code for ARMv7-M targets.
15291 @subsection AVR Options
15292 @cindex AVR Options
15294 These options are defined for AVR implementations:
15297 @item -mmcu=@var{mcu}
15299 Specify Atmel AVR instruction set architectures (ISA) or MCU type.
15301 The default for this option is@tie{}@samp{avr2}.
15303 GCC supports the following AVR devices and ISAs:
15305 @include avr-mmcu.texi
15307 @item -maccumulate-args
15308 @opindex maccumulate-args
15309 Accumulate outgoing function arguments and acquire/release the needed
15310 stack space for outgoing function arguments once in function
15311 prologue/epilogue. Without this option, outgoing arguments are pushed
15312 before calling a function and popped afterwards.
15314 Popping the arguments after the function call can be expensive on
15315 AVR so that accumulating the stack space might lead to smaller
15316 executables because arguments need not to be removed from the
15317 stack after such a function call.
15319 This option can lead to reduced code size for functions that perform
15320 several calls to functions that get their arguments on the stack like
15321 calls to printf-like functions.
15323 @item -mbranch-cost=@var{cost}
15324 @opindex mbranch-cost
15325 Set the branch costs for conditional branch instructions to
15326 @var{cost}. Reasonable values for @var{cost} are small, non-negative
15327 integers. The default branch cost is 0.
15329 @item -mcall-prologues
15330 @opindex mcall-prologues
15331 Functions prologues/epilogues are expanded as calls to appropriate
15332 subroutines. Code size is smaller.
15336 Assume @code{int} to be 8-bit integer. This affects the sizes of all types: a
15337 @code{char} is 1 byte, an @code{int} is 1 byte, a @code{long} is 2 bytes,
15338 and @code{long long} is 4 bytes. Please note that this option does not
15339 conform to the C standards, but it results in smaller code
15342 @item -mn-flash=@var{num}
15344 Assume that the flash memory has a size of
15345 @var{num} times 64@tie{}KiB.
15347 @item -mno-interrupts
15348 @opindex mno-interrupts
15349 Generated code is not compatible with hardware interrupts.
15350 Code size is smaller.
15354 Try to replace @code{CALL} resp.@: @code{JMP} instruction by the shorter
15355 @code{RCALL} resp.@: @code{RJMP} instruction if applicable.
15356 Setting @option{-mrelax} just adds the @option{--mlink-relax} option to
15357 the assembler's command line and the @option{--relax} option to the
15358 linker's command line.
15360 Jump relaxing is performed by the linker because jump offsets are not
15361 known before code is located. Therefore, the assembler code generated by the
15362 compiler is the same, but the instructions in the executable may
15363 differ from instructions in the assembler code.
15365 Relaxing must be turned on if linker stubs are needed, see the
15366 section on @code{EIND} and linker stubs below.
15370 Assume that the device supports the Read-Modify-Write
15371 instructions @code{XCH}, @code{LAC}, @code{LAS} and @code{LAT}.
15375 Treat the stack pointer register as an 8-bit register,
15376 i.e.@: assume the high byte of the stack pointer is zero.
15377 In general, you don't need to set this option by hand.
15379 This option is used internally by the compiler to select and
15380 build multilibs for architectures @code{avr2} and @code{avr25}.
15381 These architectures mix devices with and without @code{SPH}.
15382 For any setting other than @option{-mmcu=avr2} or @option{-mmcu=avr25}
15383 the compiler driver adds or removes this option from the compiler
15384 proper's command line, because the compiler then knows if the device
15385 or architecture has an 8-bit stack pointer and thus no @code{SPH}
15390 Use address register @code{X} in a way proposed by the hardware. This means
15391 that @code{X} is only used in indirect, post-increment or
15392 pre-decrement addressing.
15394 Without this option, the @code{X} register may be used in the same way
15395 as @code{Y} or @code{Z} which then is emulated by additional
15397 For example, loading a value with @code{X+const} addressing with a
15398 small non-negative @code{const < 64} to a register @var{Rn} is
15402 adiw r26, const ; X += const
15403 ld @var{Rn}, X ; @var{Rn} = *X
15404 sbiw r26, const ; X -= const
15408 @opindex mtiny-stack
15409 Only change the lower 8@tie{}bits of the stack pointer.
15411 @item -mfract-convert-truncate
15412 @opindex mfract-convert-truncate
15413 Allow to use truncation instead of rounding towards zero for fractional fixed-point types.
15416 @opindex nodevicelib
15417 Don't link against AVR-LibC's device specific library @code{lib<mcu>.a}.
15419 @item -Waddr-space-convert
15420 @opindex Waddr-space-convert
15421 Warn about conversions between address spaces in the case where the
15422 resulting address space is not contained in the incoming address space.
15424 @item -Wmisspelled-isr
15425 @opindex Wmisspelled-isr
15426 Warn if the ISR is misspelled, i.e. without __vector prefix.
15427 Enabled by default.
15430 @subsubsection @code{EIND} and Devices with More Than 128 Ki Bytes of Flash
15431 @cindex @code{EIND}
15432 Pointers in the implementation are 16@tie{}bits wide.
15433 The address of a function or label is represented as word address so
15434 that indirect jumps and calls can target any code address in the
15435 range of 64@tie{}Ki words.
15437 In order to facilitate indirect jump on devices with more than 128@tie{}Ki
15438 bytes of program memory space, there is a special function register called
15439 @code{EIND} that serves as most significant part of the target address
15440 when @code{EICALL} or @code{EIJMP} instructions are used.
15442 Indirect jumps and calls on these devices are handled as follows by
15443 the compiler and are subject to some limitations:
15448 The compiler never sets @code{EIND}.
15451 The compiler uses @code{EIND} implicitly in @code{EICALL}/@code{EIJMP}
15452 instructions or might read @code{EIND} directly in order to emulate an
15453 indirect call/jump by means of a @code{RET} instruction.
15456 The compiler assumes that @code{EIND} never changes during the startup
15457 code or during the application. In particular, @code{EIND} is not
15458 saved/restored in function or interrupt service routine
15462 For indirect calls to functions and computed goto, the linker
15463 generates @emph{stubs}. Stubs are jump pads sometimes also called
15464 @emph{trampolines}. Thus, the indirect call/jump jumps to such a stub.
15465 The stub contains a direct jump to the desired address.
15468 Linker relaxation must be turned on so that the linker generates
15469 the stubs correctly in all situations. See the compiler option
15470 @option{-mrelax} and the linker option @option{--relax}.
15471 There are corner cases where the linker is supposed to generate stubs
15472 but aborts without relaxation and without a helpful error message.
15475 The default linker script is arranged for code with @code{EIND = 0}.
15476 If code is supposed to work for a setup with @code{EIND != 0}, a custom
15477 linker script has to be used in order to place the sections whose
15478 name start with @code{.trampolines} into the segment where @code{EIND}
15482 The startup code from libgcc never sets @code{EIND}.
15483 Notice that startup code is a blend of code from libgcc and AVR-LibC.
15484 For the impact of AVR-LibC on @code{EIND}, see the
15485 @w{@uref{http://nongnu.org/avr-libc/user-manual/,AVR-LibC user manual}}.
15488 It is legitimate for user-specific startup code to set up @code{EIND}
15489 early, for example by means of initialization code located in
15490 section @code{.init3}. Such code runs prior to general startup code
15491 that initializes RAM and calls constructors, but after the bit
15492 of startup code from AVR-LibC that sets @code{EIND} to the segment
15493 where the vector table is located.
15495 #include <avr/io.h>
15498 __attribute__((section(".init3"),naked,used,no_instrument_function))
15499 init3_set_eind (void)
15501 __asm volatile ("ldi r24,pm_hh8(__trampolines_start)\n\t"
15502 "out %i0,r24" :: "n" (&EIND) : "r24","memory");
15507 The @code{__trampolines_start} symbol is defined in the linker script.
15510 Stubs are generated automatically by the linker if
15511 the following two conditions are met:
15514 @item The address of a label is taken by means of the @code{gs} modifier
15515 (short for @emph{generate stubs}) like so:
15517 LDI r24, lo8(gs(@var{func}))
15518 LDI r25, hi8(gs(@var{func}))
15520 @item The final location of that label is in a code segment
15521 @emph{outside} the segment where the stubs are located.
15525 The compiler emits such @code{gs} modifiers for code labels in the
15526 following situations:
15528 @item Taking address of a function or code label.
15529 @item Computed goto.
15530 @item If prologue-save function is used, see @option{-mcall-prologues}
15531 command-line option.
15532 @item Switch/case dispatch tables. If you do not want such dispatch
15533 tables you can specify the @option{-fno-jump-tables} command-line option.
15534 @item C and C++ constructors/destructors called during startup/shutdown.
15535 @item If the tools hit a @code{gs()} modifier explained above.
15539 Jumping to non-symbolic addresses like so is @emph{not} supported:
15544 /* Call function at word address 0x2 */
15545 return ((int(*)(void)) 0x2)();
15549 Instead, a stub has to be set up, i.e.@: the function has to be called
15550 through a symbol (@code{func_4} in the example):
15555 extern int func_4 (void);
15557 /* Call function at byte address 0x4 */
15562 and the application be linked with @option{-Wl,--defsym,func_4=0x4}.
15563 Alternatively, @code{func_4} can be defined in the linker script.
15566 @subsubsection Handling of the @code{RAMPD}, @code{RAMPX}, @code{RAMPY} and @code{RAMPZ} Special Function Registers
15567 @cindex @code{RAMPD}
15568 @cindex @code{RAMPX}
15569 @cindex @code{RAMPY}
15570 @cindex @code{RAMPZ}
15571 Some AVR devices support memories larger than the 64@tie{}KiB range
15572 that can be accessed with 16-bit pointers. To access memory locations
15573 outside this 64@tie{}KiB range, the contentent of a @code{RAMP}
15574 register is used as high part of the address:
15575 The @code{X}, @code{Y}, @code{Z} address register is concatenated
15576 with the @code{RAMPX}, @code{RAMPY}, @code{RAMPZ} special function
15577 register, respectively, to get a wide address. Similarly,
15578 @code{RAMPD} is used together with direct addressing.
15582 The startup code initializes the @code{RAMP} special function
15583 registers with zero.
15586 If a @ref{AVR Named Address Spaces,named address space} other than
15587 generic or @code{__flash} is used, then @code{RAMPZ} is set
15588 as needed before the operation.
15591 If the device supports RAM larger than 64@tie{}KiB and the compiler
15592 needs to change @code{RAMPZ} to accomplish an operation, @code{RAMPZ}
15593 is reset to zero after the operation.
15596 If the device comes with a specific @code{RAMP} register, the ISR
15597 prologue/epilogue saves/restores that SFR and initializes it with
15598 zero in case the ISR code might (implicitly) use it.
15601 RAM larger than 64@tie{}KiB is not supported by GCC for AVR targets.
15602 If you use inline assembler to read from locations outside the
15603 16-bit address range and change one of the @code{RAMP} registers,
15604 you must reset it to zero after the access.
15608 @subsubsection AVR Built-in Macros
15610 GCC defines several built-in macros so that the user code can test
15611 for the presence or absence of features. Almost any of the following
15612 built-in macros are deduced from device capabilities and thus
15613 triggered by the @option{-mmcu=} command-line option.
15615 For even more AVR-specific built-in macros see
15616 @ref{AVR Named Address Spaces} and @ref{AVR Built-in Functions}.
15621 Build-in macro that resolves to a decimal number that identifies the
15622 architecture and depends on the @option{-mmcu=@var{mcu}} option.
15623 Possible values are:
15625 @code{2}, @code{25}, @code{3}, @code{31}, @code{35},
15626 @code{4}, @code{5}, @code{51}, @code{6}
15628 for @var{mcu}=@code{avr2}, @code{avr25}, @code{avr3}, @code{avr31},
15629 @code{avr35}, @code{avr4}, @code{avr5}, @code{avr51}, @code{avr6},
15633 @code{100}, @code{102}, @code{104},
15634 @code{105}, @code{106}, @code{107}
15636 for @var{mcu}=@code{avrtiny}, @code{avrxmega2}, @code{avrxmega4},
15637 @code{avrxmega5}, @code{avrxmega6}, @code{avrxmega7}, respectively.
15638 If @var{mcu} specifies a device, this built-in macro is set
15639 accordingly. For example, with @option{-mmcu=atmega8} the macro is
15640 defined to @code{4}.
15642 @item __AVR_@var{Device}__
15643 Setting @option{-mmcu=@var{device}} defines this built-in macro which reflects
15644 the device's name. For example, @option{-mmcu=atmega8} defines the
15645 built-in macro @code{__AVR_ATmega8__}, @option{-mmcu=attiny261a} defines
15646 @code{__AVR_ATtiny261A__}, etc.
15648 The built-in macros' names follow
15649 the scheme @code{__AVR_@var{Device}__} where @var{Device} is
15650 the device name as from the AVR user manual. The difference between
15651 @var{Device} in the built-in macro and @var{device} in
15652 @option{-mmcu=@var{device}} is that the latter is always lowercase.
15654 If @var{device} is not a device but only a core architecture like
15655 @samp{avr51}, this macro is not defined.
15657 @item __AVR_DEVICE_NAME__
15658 Setting @option{-mmcu=@var{device}} defines this built-in macro to
15659 the device's name. For example, with @option{-mmcu=atmega8} the macro
15660 is defined to @code{atmega8}.
15662 If @var{device} is not a device but only a core architecture like
15663 @samp{avr51}, this macro is not defined.
15665 @item __AVR_XMEGA__
15666 The device / architecture belongs to the XMEGA family of devices.
15668 @item __AVR_HAVE_ELPM__
15669 The device has the @code{ELPM} instruction.
15671 @item __AVR_HAVE_ELPMX__
15672 The device has the @code{ELPM R@var{n},Z} and @code{ELPM
15673 R@var{n},Z+} instructions.
15675 @item __AVR_HAVE_MOVW__
15676 The device has the @code{MOVW} instruction to perform 16-bit
15677 register-register moves.
15679 @item __AVR_HAVE_LPMX__
15680 The device has the @code{LPM R@var{n},Z} and
15681 @code{LPM R@var{n},Z+} instructions.
15683 @item __AVR_HAVE_MUL__
15684 The device has a hardware multiplier.
15686 @item __AVR_HAVE_JMP_CALL__
15687 The device has the @code{JMP} and @code{CALL} instructions.
15688 This is the case for devices with at least 16@tie{}KiB of program
15691 @item __AVR_HAVE_EIJMP_EICALL__
15692 @itemx __AVR_3_BYTE_PC__
15693 The device has the @code{EIJMP} and @code{EICALL} instructions.
15694 This is the case for devices with more than 128@tie{}KiB of program memory.
15695 This also means that the program counter
15696 (PC) is 3@tie{}bytes wide.
15698 @item __AVR_2_BYTE_PC__
15699 The program counter (PC) is 2@tie{}bytes wide. This is the case for devices
15700 with up to 128@tie{}KiB of program memory.
15702 @item __AVR_HAVE_8BIT_SP__
15703 @itemx __AVR_HAVE_16BIT_SP__
15704 The stack pointer (SP) register is treated as 8-bit respectively
15705 16-bit register by the compiler.
15706 The definition of these macros is affected by @option{-mtiny-stack}.
15708 @item __AVR_HAVE_SPH__
15710 The device has the SPH (high part of stack pointer) special function
15711 register or has an 8-bit stack pointer, respectively.
15712 The definition of these macros is affected by @option{-mmcu=} and
15713 in the cases of @option{-mmcu=avr2} and @option{-mmcu=avr25} also
15716 @item __AVR_HAVE_RAMPD__
15717 @itemx __AVR_HAVE_RAMPX__
15718 @itemx __AVR_HAVE_RAMPY__
15719 @itemx __AVR_HAVE_RAMPZ__
15720 The device has the @code{RAMPD}, @code{RAMPX}, @code{RAMPY},
15721 @code{RAMPZ} special function register, respectively.
15723 @item __NO_INTERRUPTS__
15724 This macro reflects the @option{-mno-interrupts} command-line option.
15726 @item __AVR_ERRATA_SKIP__
15727 @itemx __AVR_ERRATA_SKIP_JMP_CALL__
15728 Some AVR devices (AT90S8515, ATmega103) must not skip 32-bit
15729 instructions because of a hardware erratum. Skip instructions are
15730 @code{SBRS}, @code{SBRC}, @code{SBIS}, @code{SBIC} and @code{CPSE}.
15731 The second macro is only defined if @code{__AVR_HAVE_JMP_CALL__} is also
15734 @item __AVR_ISA_RMW__
15735 The device has Read-Modify-Write instructions (XCH, LAC, LAS and LAT).
15737 @item __AVR_SFR_OFFSET__=@var{offset}
15738 Instructions that can address I/O special function registers directly
15739 like @code{IN}, @code{OUT}, @code{SBI}, etc.@: may use a different
15740 address as if addressed by an instruction to access RAM like @code{LD}
15741 or @code{STS}. This offset depends on the device architecture and has
15742 to be subtracted from the RAM address in order to get the
15743 respective I/O@tie{}address.
15745 @item __WITH_AVRLIBC__
15746 The compiler is configured to be used together with AVR-Libc.
15747 See the @option{--with-avrlibc} configure option.
15751 @node Blackfin Options
15752 @subsection Blackfin Options
15753 @cindex Blackfin Options
15756 @item -mcpu=@var{cpu}@r{[}-@var{sirevision}@r{]}
15758 Specifies the name of the target Blackfin processor. Currently, @var{cpu}
15759 can be one of @samp{bf512}, @samp{bf514}, @samp{bf516}, @samp{bf518},
15760 @samp{bf522}, @samp{bf523}, @samp{bf524}, @samp{bf525}, @samp{bf526},
15761 @samp{bf527}, @samp{bf531}, @samp{bf532}, @samp{bf533},
15762 @samp{bf534}, @samp{bf536}, @samp{bf537}, @samp{bf538}, @samp{bf539},
15763 @samp{bf542}, @samp{bf544}, @samp{bf547}, @samp{bf548}, @samp{bf549},
15764 @samp{bf542m}, @samp{bf544m}, @samp{bf547m}, @samp{bf548m}, @samp{bf549m},
15765 @samp{bf561}, @samp{bf592}.
15767 The optional @var{sirevision} specifies the silicon revision of the target
15768 Blackfin processor. Any workarounds available for the targeted silicon revision
15769 are enabled. If @var{sirevision} is @samp{none}, no workarounds are enabled.
15770 If @var{sirevision} is @samp{any}, all workarounds for the targeted processor
15771 are enabled. The @code{__SILICON_REVISION__} macro is defined to two
15772 hexadecimal digits representing the major and minor numbers in the silicon
15773 revision. If @var{sirevision} is @samp{none}, the @code{__SILICON_REVISION__}
15774 is not defined. If @var{sirevision} is @samp{any}, the
15775 @code{__SILICON_REVISION__} is defined to be @code{0xffff}.
15776 If this optional @var{sirevision} is not used, GCC assumes the latest known
15777 silicon revision of the targeted Blackfin processor.
15779 GCC defines a preprocessor macro for the specified @var{cpu}.
15780 For the @samp{bfin-elf} toolchain, this option causes the hardware BSP
15781 provided by libgloss to be linked in if @option{-msim} is not given.
15783 Without this option, @samp{bf532} is used as the processor by default.
15785 Note that support for @samp{bf561} is incomplete. For @samp{bf561},
15786 only the preprocessor macro is defined.
15790 Specifies that the program will be run on the simulator. This causes
15791 the simulator BSP provided by libgloss to be linked in. This option
15792 has effect only for @samp{bfin-elf} toolchain.
15793 Certain other options, such as @option{-mid-shared-library} and
15794 @option{-mfdpic}, imply @option{-msim}.
15796 @item -momit-leaf-frame-pointer
15797 @opindex momit-leaf-frame-pointer
15798 Don't keep the frame pointer in a register for leaf functions. This
15799 avoids the instructions to save, set up and restore frame pointers and
15800 makes an extra register available in leaf functions. The option
15801 @option{-fomit-frame-pointer} removes the frame pointer for all functions,
15802 which might make debugging harder.
15804 @item -mspecld-anomaly
15805 @opindex mspecld-anomaly
15806 When enabled, the compiler ensures that the generated code does not
15807 contain speculative loads after jump instructions. If this option is used,
15808 @code{__WORKAROUND_SPECULATIVE_LOADS} is defined.
15810 @item -mno-specld-anomaly
15811 @opindex mno-specld-anomaly
15812 Don't generate extra code to prevent speculative loads from occurring.
15814 @item -mcsync-anomaly
15815 @opindex mcsync-anomaly
15816 When enabled, the compiler ensures that the generated code does not
15817 contain CSYNC or SSYNC instructions too soon after conditional branches.
15818 If this option is used, @code{__WORKAROUND_SPECULATIVE_SYNCS} is defined.
15820 @item -mno-csync-anomaly
15821 @opindex mno-csync-anomaly
15822 Don't generate extra code to prevent CSYNC or SSYNC instructions from
15823 occurring too soon after a conditional branch.
15827 When enabled, the compiler is free to take advantage of the knowledge that
15828 the entire program fits into the low 64k of memory.
15831 @opindex mno-low-64k
15832 Assume that the program is arbitrarily large. This is the default.
15834 @item -mstack-check-l1
15835 @opindex mstack-check-l1
15836 Do stack checking using information placed into L1 scratchpad memory by the
15839 @item -mid-shared-library
15840 @opindex mid-shared-library
15841 Generate code that supports shared libraries via the library ID method.
15842 This allows for execute in place and shared libraries in an environment
15843 without virtual memory management. This option implies @option{-fPIC}.
15844 With a @samp{bfin-elf} target, this option implies @option{-msim}.
15846 @item -mno-id-shared-library
15847 @opindex mno-id-shared-library
15848 Generate code that doesn't assume ID-based shared libraries are being used.
15849 This is the default.
15851 @item -mleaf-id-shared-library
15852 @opindex mleaf-id-shared-library
15853 Generate code that supports shared libraries via the library ID method,
15854 but assumes that this library or executable won't link against any other
15855 ID shared libraries. That allows the compiler to use faster code for jumps
15858 @item -mno-leaf-id-shared-library
15859 @opindex mno-leaf-id-shared-library
15860 Do not assume that the code being compiled won't link against any ID shared
15861 libraries. Slower code is generated for jump and call insns.
15863 @item -mshared-library-id=n
15864 @opindex mshared-library-id
15865 Specifies the identification number of the ID-based shared library being
15866 compiled. Specifying a value of 0 generates more compact code; specifying
15867 other values forces the allocation of that number to the current
15868 library but is no more space- or time-efficient than omitting this option.
15872 Generate code that allows the data segment to be located in a different
15873 area of memory from the text segment. This allows for execute in place in
15874 an environment without virtual memory management by eliminating relocations
15875 against the text section.
15877 @item -mno-sep-data
15878 @opindex mno-sep-data
15879 Generate code that assumes that the data segment follows the text segment.
15880 This is the default.
15883 @itemx -mno-long-calls
15884 @opindex mlong-calls
15885 @opindex mno-long-calls
15886 Tells the compiler to perform function calls by first loading the
15887 address of the function into a register and then performing a subroutine
15888 call on this register. This switch is needed if the target function
15889 lies outside of the 24-bit addressing range of the offset-based
15890 version of subroutine call instruction.
15892 This feature is not enabled by default. Specifying
15893 @option{-mno-long-calls} restores the default behavior. Note these
15894 switches have no effect on how the compiler generates code to handle
15895 function calls via function pointers.
15899 Link with the fast floating-point library. This library relaxes some of
15900 the IEEE floating-point standard's rules for checking inputs against
15901 Not-a-Number (NAN), in the interest of performance.
15904 @opindex minline-plt
15905 Enable inlining of PLT entries in function calls to functions that are
15906 not known to bind locally. It has no effect without @option{-mfdpic}.
15909 @opindex mmulticore
15910 Build a standalone application for multicore Blackfin processors.
15911 This option causes proper start files and link scripts supporting
15912 multicore to be used, and defines the macro @code{__BFIN_MULTICORE}.
15913 It can only be used with @option{-mcpu=bf561@r{[}-@var{sirevision}@r{]}}.
15915 This option can be used with @option{-mcorea} or @option{-mcoreb}, which
15916 selects the one-application-per-core programming model. Without
15917 @option{-mcorea} or @option{-mcoreb}, the single-application/dual-core
15918 programming model is used. In this model, the main function of Core B
15919 should be named as @code{coreb_main}.
15921 If this option is not used, the single-core application programming
15926 Build a standalone application for Core A of BF561 when using
15927 the one-application-per-core programming model. Proper start files
15928 and link scripts are used to support Core A, and the macro
15929 @code{__BFIN_COREA} is defined.
15930 This option can only be used in conjunction with @option{-mmulticore}.
15934 Build a standalone application for Core B of BF561 when using
15935 the one-application-per-core programming model. Proper start files
15936 and link scripts are used to support Core B, and the macro
15937 @code{__BFIN_COREB} is defined. When this option is used, @code{coreb_main}
15938 should be used instead of @code{main}.
15939 This option can only be used in conjunction with @option{-mmulticore}.
15943 Build a standalone application for SDRAM. Proper start files and
15944 link scripts are used to put the application into SDRAM, and the macro
15945 @code{__BFIN_SDRAM} is defined.
15946 The loader should initialize SDRAM before loading the application.
15950 Assume that ICPLBs are enabled at run time. This has an effect on certain
15951 anomaly workarounds. For Linux targets, the default is to assume ICPLBs
15952 are enabled; for standalone applications the default is off.
15956 @subsection C6X Options
15957 @cindex C6X Options
15960 @item -march=@var{name}
15962 This specifies the name of the target architecture. GCC uses this
15963 name to determine what kind of instructions it can emit when generating
15964 assembly code. Permissible names are: @samp{c62x},
15965 @samp{c64x}, @samp{c64x+}, @samp{c67x}, @samp{c67x+}, @samp{c674x}.
15968 @opindex mbig-endian
15969 Generate code for a big-endian target.
15971 @item -mlittle-endian
15972 @opindex mlittle-endian
15973 Generate code for a little-endian target. This is the default.
15977 Choose startup files and linker script suitable for the simulator.
15979 @item -msdata=default
15980 @opindex msdata=default
15981 Put small global and static data in the @code{.neardata} section,
15982 which is pointed to by register @code{B14}. Put small uninitialized
15983 global and static data in the @code{.bss} section, which is adjacent
15984 to the @code{.neardata} section. Put small read-only data into the
15985 @code{.rodata} section. The corresponding sections used for large
15986 pieces of data are @code{.fardata}, @code{.far} and @code{.const}.
15989 @opindex msdata=all
15990 Put all data, not just small objects, into the sections reserved for
15991 small data, and use addressing relative to the @code{B14} register to
15995 @opindex msdata=none
15996 Make no use of the sections reserved for small data, and use absolute
15997 addresses to access all data. Put all initialized global and static
15998 data in the @code{.fardata} section, and all uninitialized data in the
15999 @code{.far} section. Put all constant data into the @code{.const}
16004 @subsection CRIS Options
16005 @cindex CRIS Options
16007 These options are defined specifically for the CRIS ports.
16010 @item -march=@var{architecture-type}
16011 @itemx -mcpu=@var{architecture-type}
16014 Generate code for the specified architecture. The choices for
16015 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
16016 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX@.
16017 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
16020 @item -mtune=@var{architecture-type}
16022 Tune to @var{architecture-type} everything applicable about the generated
16023 code, except for the ABI and the set of available instructions. The
16024 choices for @var{architecture-type} are the same as for
16025 @option{-march=@var{architecture-type}}.
16027 @item -mmax-stack-frame=@var{n}
16028 @opindex mmax-stack-frame
16029 Warn when the stack frame of a function exceeds @var{n} bytes.
16035 The options @option{-metrax4} and @option{-metrax100} are synonyms for
16036 @option{-march=v3} and @option{-march=v8} respectively.
16038 @item -mmul-bug-workaround
16039 @itemx -mno-mul-bug-workaround
16040 @opindex mmul-bug-workaround
16041 @opindex mno-mul-bug-workaround
16042 Work around a bug in the @code{muls} and @code{mulu} instructions for CPU
16043 models where it applies. This option is active by default.
16047 Enable CRIS-specific verbose debug-related information in the assembly
16048 code. This option also has the effect of turning off the @samp{#NO_APP}
16049 formatted-code indicator to the assembler at the beginning of the
16054 Do not use condition-code results from previous instruction; always emit
16055 compare and test instructions before use of condition codes.
16057 @item -mno-side-effects
16058 @opindex mno-side-effects
16059 Do not emit instructions with side effects in addressing modes other than
16062 @item -mstack-align
16063 @itemx -mno-stack-align
16064 @itemx -mdata-align
16065 @itemx -mno-data-align
16066 @itemx -mconst-align
16067 @itemx -mno-const-align
16068 @opindex mstack-align
16069 @opindex mno-stack-align
16070 @opindex mdata-align
16071 @opindex mno-data-align
16072 @opindex mconst-align
16073 @opindex mno-const-align
16074 These options (@samp{no-} options) arrange (eliminate arrangements) for the
16075 stack frame, individual data and constants to be aligned for the maximum
16076 single data access size for the chosen CPU model. The default is to
16077 arrange for 32-bit alignment. ABI details such as structure layout are
16078 not affected by these options.
16086 Similar to the stack- data- and const-align options above, these options
16087 arrange for stack frame, writable data and constants to all be 32-bit,
16088 16-bit or 8-bit aligned. The default is 32-bit alignment.
16090 @item -mno-prologue-epilogue
16091 @itemx -mprologue-epilogue
16092 @opindex mno-prologue-epilogue
16093 @opindex mprologue-epilogue
16094 With @option{-mno-prologue-epilogue}, the normal function prologue and
16095 epilogue which set up the stack frame are omitted and no return
16096 instructions or return sequences are generated in the code. Use this
16097 option only together with visual inspection of the compiled code: no
16098 warnings or errors are generated when call-saved registers must be saved,
16099 or storage for local variables needs to be allocated.
16103 @opindex mno-gotplt
16105 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
16106 instruction sequences that load addresses for functions from the PLT part
16107 of the GOT rather than (traditional on other architectures) calls to the
16108 PLT@. The default is @option{-mgotplt}.
16112 Legacy no-op option only recognized with the cris-axis-elf and
16113 cris-axis-linux-gnu targets.
16117 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
16121 This option, recognized for the cris-axis-elf, arranges
16122 to link with input-output functions from a simulator library. Code,
16123 initialized data and zero-initialized data are allocated consecutively.
16127 Like @option{-sim}, but pass linker options to locate initialized data at
16128 0x40000000 and zero-initialized data at 0x80000000.
16132 @subsection CR16 Options
16133 @cindex CR16 Options
16135 These options are defined specifically for the CR16 ports.
16141 Enable the use of multiply-accumulate instructions. Disabled by default.
16145 @opindex mcr16cplus
16147 Generate code for CR16C or CR16C+ architecture. CR16C+ architecture
16152 Links the library libsim.a which is in compatible with simulator. Applicable
16153 to ELF compiler only.
16157 Choose integer type as 32-bit wide.
16161 Generates @code{sbit}/@code{cbit} instructions for bit manipulations.
16163 @item -mdata-model=@var{model}
16164 @opindex mdata-model
16165 Choose a data model. The choices for @var{model} are @samp{near},
16166 @samp{far} or @samp{medium}. @samp{medium} is default.
16167 However, @samp{far} is not valid with @option{-mcr16c}, as the
16168 CR16C architecture does not support the far data model.
16171 @node Darwin Options
16172 @subsection Darwin Options
16173 @cindex Darwin options
16175 These options are defined for all architectures running the Darwin operating
16178 FSF GCC on Darwin does not create ``fat'' object files; it creates
16179 an object file for the single architecture that GCC was built to
16180 target. Apple's GCC on Darwin does create ``fat'' files if multiple
16181 @option{-arch} options are used; it does so by running the compiler or
16182 linker multiple times and joining the results together with
16185 The subtype of the file created (like @samp{ppc7400} or @samp{ppc970} or
16186 @samp{i686}) is determined by the flags that specify the ISA
16187 that GCC is targeting, like @option{-mcpu} or @option{-march}. The
16188 @option{-force_cpusubtype_ALL} option can be used to override this.
16190 The Darwin tools vary in their behavior when presented with an ISA
16191 mismatch. The assembler, @file{as}, only permits instructions to
16192 be used that are valid for the subtype of the file it is generating,
16193 so you cannot put 64-bit instructions in a @samp{ppc750} object file.
16194 The linker for shared libraries, @file{/usr/bin/libtool}, fails
16195 and prints an error if asked to create a shared library with a less
16196 restrictive subtype than its input files (for instance, trying to put
16197 a @samp{ppc970} object file in a @samp{ppc7400} library). The linker
16198 for executables, @command{ld}, quietly gives the executable the most
16199 restrictive subtype of any of its input files.
16204 Add the framework directory @var{dir} to the head of the list of
16205 directories to be searched for header files. These directories are
16206 interleaved with those specified by @option{-I} options and are
16207 scanned in a left-to-right order.
16209 A framework directory is a directory with frameworks in it. A
16210 framework is a directory with a @file{Headers} and/or
16211 @file{PrivateHeaders} directory contained directly in it that ends
16212 in @file{.framework}. The name of a framework is the name of this
16213 directory excluding the @file{.framework}. Headers associated with
16214 the framework are found in one of those two directories, with
16215 @file{Headers} being searched first. A subframework is a framework
16216 directory that is in a framework's @file{Frameworks} directory.
16217 Includes of subframework headers can only appear in a header of a
16218 framework that contains the subframework, or in a sibling subframework
16219 header. Two subframeworks are siblings if they occur in the same
16220 framework. A subframework should not have the same name as a
16221 framework; a warning is issued if this is violated. Currently a
16222 subframework cannot have subframeworks; in the future, the mechanism
16223 may be extended to support this. The standard frameworks can be found
16224 in @file{/System/Library/Frameworks} and
16225 @file{/Library/Frameworks}. An example include looks like
16226 @code{#include <Framework/header.h>}, where @file{Framework} denotes
16227 the name of the framework and @file{header.h} is found in the
16228 @file{PrivateHeaders} or @file{Headers} directory.
16230 @item -iframework@var{dir}
16231 @opindex iframework
16232 Like @option{-F} except the directory is a treated as a system
16233 directory. The main difference between this @option{-iframework} and
16234 @option{-F} is that with @option{-iframework} the compiler does not
16235 warn about constructs contained within header files found via
16236 @var{dir}. This option is valid only for the C family of languages.
16240 Emit debugging information for symbols that are used. For stabs
16241 debugging format, this enables @option{-feliminate-unused-debug-symbols}.
16242 This is by default ON@.
16246 Emit debugging information for all symbols and types.
16248 @item -mmacosx-version-min=@var{version}
16249 The earliest version of MacOS X that this executable will run on
16250 is @var{version}. Typical values of @var{version} include @code{10.1},
16251 @code{10.2}, and @code{10.3.9}.
16253 If the compiler was built to use the system's headers by default,
16254 then the default for this option is the system version on which the
16255 compiler is running, otherwise the default is to make choices that
16256 are compatible with as many systems and code bases as possible.
16260 Enable kernel development mode. The @option{-mkernel} option sets
16261 @option{-static}, @option{-fno-common}, @option{-fno-use-cxa-atexit},
16262 @option{-fno-exceptions}, @option{-fno-non-call-exceptions},
16263 @option{-fapple-kext}, @option{-fno-weak} and @option{-fno-rtti} where
16264 applicable. This mode also sets @option{-mno-altivec},
16265 @option{-msoft-float}, @option{-fno-builtin} and
16266 @option{-mlong-branch} for PowerPC targets.
16268 @item -mone-byte-bool
16269 @opindex mone-byte-bool
16270 Override the defaults for @code{bool} so that @code{sizeof(bool)==1}.
16271 By default @code{sizeof(bool)} is @code{4} when compiling for
16272 Darwin/PowerPC and @code{1} when compiling for Darwin/x86, so this
16273 option has no effect on x86.
16275 @strong{Warning:} The @option{-mone-byte-bool} switch causes GCC
16276 to generate code that is not binary compatible with code generated
16277 without that switch. Using this switch may require recompiling all
16278 other modules in a program, including system libraries. Use this
16279 switch to conform to a non-default data model.
16281 @item -mfix-and-continue
16282 @itemx -ffix-and-continue
16283 @itemx -findirect-data
16284 @opindex mfix-and-continue
16285 @opindex ffix-and-continue
16286 @opindex findirect-data
16287 Generate code suitable for fast turnaround development, such as to
16288 allow GDB to dynamically load @file{.o} files into already-running
16289 programs. @option{-findirect-data} and @option{-ffix-and-continue}
16290 are provided for backwards compatibility.
16294 Loads all members of static archive libraries.
16295 See man ld(1) for more information.
16297 @item -arch_errors_fatal
16298 @opindex arch_errors_fatal
16299 Cause the errors having to do with files that have the wrong architecture
16302 @item -bind_at_load
16303 @opindex bind_at_load
16304 Causes the output file to be marked such that the dynamic linker will
16305 bind all undefined references when the file is loaded or launched.
16309 Produce a Mach-o bundle format file.
16310 See man ld(1) for more information.
16312 @item -bundle_loader @var{executable}
16313 @opindex bundle_loader
16314 This option specifies the @var{executable} that will load the build
16315 output file being linked. See man ld(1) for more information.
16318 @opindex dynamiclib
16319 When passed this option, GCC produces a dynamic library instead of
16320 an executable when linking, using the Darwin @file{libtool} command.
16322 @item -force_cpusubtype_ALL
16323 @opindex force_cpusubtype_ALL
16324 This causes GCC's output file to have the @samp{ALL} subtype, instead of
16325 one controlled by the @option{-mcpu} or @option{-march} option.
16327 @item -allowable_client @var{client_name}
16328 @itemx -client_name
16329 @itemx -compatibility_version
16330 @itemx -current_version
16332 @itemx -dependency-file
16334 @itemx -dylinker_install_name
16336 @itemx -exported_symbols_list
16339 @itemx -flat_namespace
16340 @itemx -force_flat_namespace
16341 @itemx -headerpad_max_install_names
16344 @itemx -install_name
16345 @itemx -keep_private_externs
16346 @itemx -multi_module
16347 @itemx -multiply_defined
16348 @itemx -multiply_defined_unused
16351 @itemx -no_dead_strip_inits_and_terms
16352 @itemx -nofixprebinding
16353 @itemx -nomultidefs
16355 @itemx -noseglinkedit
16356 @itemx -pagezero_size
16358 @itemx -prebind_all_twolevel_modules
16359 @itemx -private_bundle
16361 @itemx -read_only_relocs
16363 @itemx -sectobjectsymbols
16367 @itemx -sectobjectsymbols
16370 @itemx -segs_read_only_addr
16372 @itemx -segs_read_write_addr
16373 @itemx -seg_addr_table
16374 @itemx -seg_addr_table_filename
16375 @itemx -seglinkedit
16377 @itemx -segs_read_only_addr
16378 @itemx -segs_read_write_addr
16379 @itemx -single_module
16381 @itemx -sub_library
16383 @itemx -sub_umbrella
16384 @itemx -twolevel_namespace
16387 @itemx -unexported_symbols_list
16388 @itemx -weak_reference_mismatches
16389 @itemx -whatsloaded
16390 @opindex allowable_client
16391 @opindex client_name
16392 @opindex compatibility_version
16393 @opindex current_version
16394 @opindex dead_strip
16395 @opindex dependency-file
16396 @opindex dylib_file
16397 @opindex dylinker_install_name
16399 @opindex exported_symbols_list
16401 @opindex flat_namespace
16402 @opindex force_flat_namespace
16403 @opindex headerpad_max_install_names
16404 @opindex image_base
16406 @opindex install_name
16407 @opindex keep_private_externs
16408 @opindex multi_module
16409 @opindex multiply_defined
16410 @opindex multiply_defined_unused
16411 @opindex noall_load
16412 @opindex no_dead_strip_inits_and_terms
16413 @opindex nofixprebinding
16414 @opindex nomultidefs
16416 @opindex noseglinkedit
16417 @opindex pagezero_size
16419 @opindex prebind_all_twolevel_modules
16420 @opindex private_bundle
16421 @opindex read_only_relocs
16423 @opindex sectobjectsymbols
16426 @opindex sectcreate
16427 @opindex sectobjectsymbols
16430 @opindex segs_read_only_addr
16431 @opindex segs_read_write_addr
16432 @opindex seg_addr_table
16433 @opindex seg_addr_table_filename
16434 @opindex seglinkedit
16436 @opindex segs_read_only_addr
16437 @opindex segs_read_write_addr
16438 @opindex single_module
16440 @opindex sub_library
16441 @opindex sub_umbrella
16442 @opindex twolevel_namespace
16445 @opindex unexported_symbols_list
16446 @opindex weak_reference_mismatches
16447 @opindex whatsloaded
16448 These options are passed to the Darwin linker. The Darwin linker man page
16449 describes them in detail.
16452 @node DEC Alpha Options
16453 @subsection DEC Alpha Options
16455 These @samp{-m} options are defined for the DEC Alpha implementations:
16458 @item -mno-soft-float
16459 @itemx -msoft-float
16460 @opindex mno-soft-float
16461 @opindex msoft-float
16462 Use (do not use) the hardware floating-point instructions for
16463 floating-point operations. When @option{-msoft-float} is specified,
16464 functions in @file{libgcc.a} are used to perform floating-point
16465 operations. Unless they are replaced by routines that emulate the
16466 floating-point operations, or compiled in such a way as to call such
16467 emulations routines, these routines issue floating-point
16468 operations. If you are compiling for an Alpha without floating-point
16469 operations, you must ensure that the library is built so as not to call
16472 Note that Alpha implementations without floating-point operations are
16473 required to have floating-point registers.
16476 @itemx -mno-fp-regs
16478 @opindex mno-fp-regs
16479 Generate code that uses (does not use) the floating-point register set.
16480 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
16481 register set is not used, floating-point operands are passed in integer
16482 registers as if they were integers and floating-point results are passed
16483 in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence,
16484 so any function with a floating-point argument or return value called by code
16485 compiled with @option{-mno-fp-regs} must also be compiled with that
16488 A typical use of this option is building a kernel that does not use,
16489 and hence need not save and restore, any floating-point registers.
16493 The Alpha architecture implements floating-point hardware optimized for
16494 maximum performance. It is mostly compliant with the IEEE floating-point
16495 standard. However, for full compliance, software assistance is
16496 required. This option generates code fully IEEE-compliant code
16497 @emph{except} that the @var{inexact-flag} is not maintained (see below).
16498 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
16499 defined during compilation. The resulting code is less efficient but is
16500 able to correctly support denormalized numbers and exceptional IEEE
16501 values such as not-a-number and plus/minus infinity. Other Alpha
16502 compilers call this option @option{-ieee_with_no_inexact}.
16504 @item -mieee-with-inexact
16505 @opindex mieee-with-inexact
16506 This is like @option{-mieee} except the generated code also maintains
16507 the IEEE @var{inexact-flag}. Turning on this option causes the
16508 generated code to implement fully-compliant IEEE math. In addition to
16509 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
16510 macro. On some Alpha implementations the resulting code may execute
16511 significantly slower than the code generated by default. Since there is
16512 very little code that depends on the @var{inexact-flag}, you should
16513 normally not specify this option. Other Alpha compilers call this
16514 option @option{-ieee_with_inexact}.
16516 @item -mfp-trap-mode=@var{trap-mode}
16517 @opindex mfp-trap-mode
16518 This option controls what floating-point related traps are enabled.
16519 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
16520 The trap mode can be set to one of four values:
16524 This is the default (normal) setting. The only traps that are enabled
16525 are the ones that cannot be disabled in software (e.g., division by zero
16529 In addition to the traps enabled by @samp{n}, underflow traps are enabled
16533 Like @samp{u}, but the instructions are marked to be safe for software
16534 completion (see Alpha architecture manual for details).
16537 Like @samp{su}, but inexact traps are enabled as well.
16540 @item -mfp-rounding-mode=@var{rounding-mode}
16541 @opindex mfp-rounding-mode
16542 Selects the IEEE rounding mode. Other Alpha compilers call this option
16543 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
16548 Normal IEEE rounding mode. Floating-point numbers are rounded towards
16549 the nearest machine number or towards the even machine number in case
16553 Round towards minus infinity.
16556 Chopped rounding mode. Floating-point numbers are rounded towards zero.
16559 Dynamic rounding mode. A field in the floating-point control register
16560 (@var{fpcr}, see Alpha architecture reference manual) controls the
16561 rounding mode in effect. The C library initializes this register for
16562 rounding towards plus infinity. Thus, unless your program modifies the
16563 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
16566 @item -mtrap-precision=@var{trap-precision}
16567 @opindex mtrap-precision
16568 In the Alpha architecture, floating-point traps are imprecise. This
16569 means without software assistance it is impossible to recover from a
16570 floating trap and program execution normally needs to be terminated.
16571 GCC can generate code that can assist operating system trap handlers
16572 in determining the exact location that caused a floating-point trap.
16573 Depending on the requirements of an application, different levels of
16574 precisions can be selected:
16578 Program precision. This option is the default and means a trap handler
16579 can only identify which program caused a floating-point exception.
16582 Function precision. The trap handler can determine the function that
16583 caused a floating-point exception.
16586 Instruction precision. The trap handler can determine the exact
16587 instruction that caused a floating-point exception.
16590 Other Alpha compilers provide the equivalent options called
16591 @option{-scope_safe} and @option{-resumption_safe}.
16593 @item -mieee-conformant
16594 @opindex mieee-conformant
16595 This option marks the generated code as IEEE conformant. You must not
16596 use this option unless you also specify @option{-mtrap-precision=i} and either
16597 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
16598 is to emit the line @samp{.eflag 48} in the function prologue of the
16599 generated assembly file.
16601 @item -mbuild-constants
16602 @opindex mbuild-constants
16603 Normally GCC examines a 32- or 64-bit integer constant to
16604 see if it can construct it from smaller constants in two or three
16605 instructions. If it cannot, it outputs the constant as a literal and
16606 generates code to load it from the data segment at run time.
16608 Use this option to require GCC to construct @emph{all} integer constants
16609 using code, even if it takes more instructions (the maximum is six).
16611 You typically use this option to build a shared library dynamic
16612 loader. Itself a shared library, it must relocate itself in memory
16613 before it can find the variables and constants in its own data segment.
16631 Indicate whether GCC should generate code to use the optional BWX,
16632 CIX, FIX and MAX instruction sets. The default is to use the instruction
16633 sets supported by the CPU type specified via @option{-mcpu=} option or that
16634 of the CPU on which GCC was built if none is specified.
16637 @itemx -mfloat-ieee
16638 @opindex mfloat-vax
16639 @opindex mfloat-ieee
16640 Generate code that uses (does not use) VAX F and G floating-point
16641 arithmetic instead of IEEE single and double precision.
16643 @item -mexplicit-relocs
16644 @itemx -mno-explicit-relocs
16645 @opindex mexplicit-relocs
16646 @opindex mno-explicit-relocs
16647 Older Alpha assemblers provided no way to generate symbol relocations
16648 except via assembler macros. Use of these macros does not allow
16649 optimal instruction scheduling. GNU binutils as of version 2.12
16650 supports a new syntax that allows the compiler to explicitly mark
16651 which relocations should apply to which instructions. This option
16652 is mostly useful for debugging, as GCC detects the capabilities of
16653 the assembler when it is built and sets the default accordingly.
16656 @itemx -mlarge-data
16657 @opindex msmall-data
16658 @opindex mlarge-data
16659 When @option{-mexplicit-relocs} is in effect, static data is
16660 accessed via @dfn{gp-relative} relocations. When @option{-msmall-data}
16661 is used, objects 8 bytes long or smaller are placed in a @dfn{small data area}
16662 (the @code{.sdata} and @code{.sbss} sections) and are accessed via
16663 16-bit relocations off of the @code{$gp} register. This limits the
16664 size of the small data area to 64KB, but allows the variables to be
16665 directly accessed via a single instruction.
16667 The default is @option{-mlarge-data}. With this option the data area
16668 is limited to just below 2GB@. Programs that require more than 2GB of
16669 data must use @code{malloc} or @code{mmap} to allocate the data in the
16670 heap instead of in the program's data segment.
16672 When generating code for shared libraries, @option{-fpic} implies
16673 @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}.
16676 @itemx -mlarge-text
16677 @opindex msmall-text
16678 @opindex mlarge-text
16679 When @option{-msmall-text} is used, the compiler assumes that the
16680 code of the entire program (or shared library) fits in 4MB, and is
16681 thus reachable with a branch instruction. When @option{-msmall-data}
16682 is used, the compiler can assume that all local symbols share the
16683 same @code{$gp} value, and thus reduce the number of instructions
16684 required for a function call from 4 to 1.
16686 The default is @option{-mlarge-text}.
16688 @item -mcpu=@var{cpu_type}
16690 Set the instruction set and instruction scheduling parameters for
16691 machine type @var{cpu_type}. You can specify either the @samp{EV}
16692 style name or the corresponding chip number. GCC supports scheduling
16693 parameters for the EV4, EV5 and EV6 family of processors and
16694 chooses the default values for the instruction set from the processor
16695 you specify. If you do not specify a processor type, GCC defaults
16696 to the processor on which the compiler was built.
16698 Supported values for @var{cpu_type} are
16704 Schedules as an EV4 and has no instruction set extensions.
16708 Schedules as an EV5 and has no instruction set extensions.
16712 Schedules as an EV5 and supports the BWX extension.
16717 Schedules as an EV5 and supports the BWX and MAX extensions.
16721 Schedules as an EV6 and supports the BWX, FIX, and MAX extensions.
16725 Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions.
16728 Native toolchains also support the value @samp{native},
16729 which selects the best architecture option for the host processor.
16730 @option{-mcpu=native} has no effect if GCC does not recognize
16733 @item -mtune=@var{cpu_type}
16735 Set only the instruction scheduling parameters for machine type
16736 @var{cpu_type}. The instruction set is not changed.
16738 Native toolchains also support the value @samp{native},
16739 which selects the best architecture option for the host processor.
16740 @option{-mtune=native} has no effect if GCC does not recognize
16743 @item -mmemory-latency=@var{time}
16744 @opindex mmemory-latency
16745 Sets the latency the scheduler should assume for typical memory
16746 references as seen by the application. This number is highly
16747 dependent on the memory access patterns used by the application
16748 and the size of the external cache on the machine.
16750 Valid options for @var{time} are
16754 A decimal number representing clock cycles.
16760 The compiler contains estimates of the number of clock cycles for
16761 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
16762 (also called Dcache, Scache, and Bcache), as well as to main memory.
16763 Note that L3 is only valid for EV5.
16769 @subsection FR30 Options
16770 @cindex FR30 Options
16772 These options are defined specifically for the FR30 port.
16776 @item -msmall-model
16777 @opindex msmall-model
16778 Use the small address space model. This can produce smaller code, but
16779 it does assume that all symbolic values and addresses fit into a
16784 Assume that runtime support has been provided and so there is no need
16785 to include the simulator library (@file{libsim.a}) on the linker
16791 @subsection FT32 Options
16792 @cindex FT32 Options
16794 These options are defined specifically for the FT32 port.
16800 Specifies that the program will be run on the simulator. This causes
16801 an alternate runtime startup and library to be linked.
16802 You must not use this option when generating programs that will run on
16803 real hardware; you must provide your own runtime library for whatever
16804 I/O functions are needed.
16808 Enable Local Register Allocation. This is still experimental for FT32,
16809 so by default the compiler uses standard reload.
16813 Do not use div and mod instructions.
16818 @subsection FRV Options
16819 @cindex FRV Options
16825 Only use the first 32 general-purpose registers.
16830 Use all 64 general-purpose registers.
16835 Use only the first 32 floating-point registers.
16840 Use all 64 floating-point registers.
16843 @opindex mhard-float
16845 Use hardware instructions for floating-point operations.
16848 @opindex msoft-float
16850 Use library routines for floating-point operations.
16855 Dynamically allocate condition code registers.
16860 Do not try to dynamically allocate condition code registers, only
16861 use @code{icc0} and @code{fcc0}.
16866 Change ABI to use double word insns.
16871 Do not use double word instructions.
16876 Use floating-point double instructions.
16879 @opindex mno-double
16881 Do not use floating-point double instructions.
16886 Use media instructions.
16891 Do not use media instructions.
16896 Use multiply and add/subtract instructions.
16899 @opindex mno-muladd
16901 Do not use multiply and add/subtract instructions.
16906 Select the FDPIC ABI, which uses function descriptors to represent
16907 pointers to functions. Without any PIC/PIE-related options, it
16908 implies @option{-fPIE}. With @option{-fpic} or @option{-fpie}, it
16909 assumes GOT entries and small data are within a 12-bit range from the
16910 GOT base address; with @option{-fPIC} or @option{-fPIE}, GOT offsets
16911 are computed with 32 bits.
16912 With a @samp{bfin-elf} target, this option implies @option{-msim}.
16915 @opindex minline-plt
16917 Enable inlining of PLT entries in function calls to functions that are
16918 not known to bind locally. It has no effect without @option{-mfdpic}.
16919 It's enabled by default if optimizing for speed and compiling for
16920 shared libraries (i.e., @option{-fPIC} or @option{-fpic}), or when an
16921 optimization option such as @option{-O3} or above is present in the
16927 Assume a large TLS segment when generating thread-local code.
16932 Do not assume a large TLS segment when generating thread-local code.
16937 Enable the use of @code{GPREL} relocations in the FDPIC ABI for data
16938 that is known to be in read-only sections. It's enabled by default,
16939 except for @option{-fpic} or @option{-fpie}: even though it may help
16940 make the global offset table smaller, it trades 1 instruction for 4.
16941 With @option{-fPIC} or @option{-fPIE}, it trades 3 instructions for 4,
16942 one of which may be shared by multiple symbols, and it avoids the need
16943 for a GOT entry for the referenced symbol, so it's more likely to be a
16944 win. If it is not, @option{-mno-gprel-ro} can be used to disable it.
16946 @item -multilib-library-pic
16947 @opindex multilib-library-pic
16949 Link with the (library, not FD) pic libraries. It's implied by
16950 @option{-mlibrary-pic}, as well as by @option{-fPIC} and
16951 @option{-fpic} without @option{-mfdpic}. You should never have to use
16955 @opindex mlinked-fp
16957 Follow the EABI requirement of always creating a frame pointer whenever
16958 a stack frame is allocated. This option is enabled by default and can
16959 be disabled with @option{-mno-linked-fp}.
16962 @opindex mlong-calls
16964 Use indirect addressing to call functions outside the current
16965 compilation unit. This allows the functions to be placed anywhere
16966 within the 32-bit address space.
16968 @item -malign-labels
16969 @opindex malign-labels
16971 Try to align labels to an 8-byte boundary by inserting NOPs into the
16972 previous packet. This option only has an effect when VLIW packing
16973 is enabled. It doesn't create new packets; it merely adds NOPs to
16976 @item -mlibrary-pic
16977 @opindex mlibrary-pic
16979 Generate position-independent EABI code.
16984 Use only the first four media accumulator registers.
16989 Use all eight media accumulator registers.
16994 Pack VLIW instructions.
16999 Do not pack VLIW instructions.
17002 @opindex mno-eflags
17004 Do not mark ABI switches in e_flags.
17007 @opindex mcond-move
17009 Enable the use of conditional-move instructions (default).
17011 This switch is mainly for debugging the compiler and will likely be removed
17012 in a future version.
17014 @item -mno-cond-move
17015 @opindex mno-cond-move
17017 Disable the use of conditional-move instructions.
17019 This switch is mainly for debugging the compiler and will likely be removed
17020 in a future version.
17025 Enable the use of conditional set instructions (default).
17027 This switch is mainly for debugging the compiler and will likely be removed
17028 in a future version.
17033 Disable the use of conditional set instructions.
17035 This switch is mainly for debugging the compiler and will likely be removed
17036 in a future version.
17039 @opindex mcond-exec
17041 Enable the use of conditional execution (default).
17043 This switch is mainly for debugging the compiler and will likely be removed
17044 in a future version.
17046 @item -mno-cond-exec
17047 @opindex mno-cond-exec
17049 Disable the use of conditional execution.
17051 This switch is mainly for debugging the compiler and will likely be removed
17052 in a future version.
17054 @item -mvliw-branch
17055 @opindex mvliw-branch
17057 Run a pass to pack branches into VLIW instructions (default).
17059 This switch is mainly for debugging the compiler and will likely be removed
17060 in a future version.
17062 @item -mno-vliw-branch
17063 @opindex mno-vliw-branch
17065 Do not run a pass to pack branches into VLIW instructions.
17067 This switch is mainly for debugging the compiler and will likely be removed
17068 in a future version.
17070 @item -mmulti-cond-exec
17071 @opindex mmulti-cond-exec
17073 Enable optimization of @code{&&} and @code{||} in conditional execution
17076 This switch is mainly for debugging the compiler and will likely be removed
17077 in a future version.
17079 @item -mno-multi-cond-exec
17080 @opindex mno-multi-cond-exec
17082 Disable optimization of @code{&&} and @code{||} in conditional execution.
17084 This switch is mainly for debugging the compiler and will likely be removed
17085 in a future version.
17087 @item -mnested-cond-exec
17088 @opindex mnested-cond-exec
17090 Enable nested conditional execution optimizations (default).
17092 This switch is mainly for debugging the compiler and will likely be removed
17093 in a future version.
17095 @item -mno-nested-cond-exec
17096 @opindex mno-nested-cond-exec
17098 Disable nested conditional execution optimizations.
17100 This switch is mainly for debugging the compiler and will likely be removed
17101 in a future version.
17103 @item -moptimize-membar
17104 @opindex moptimize-membar
17106 This switch removes redundant @code{membar} instructions from the
17107 compiler-generated code. It is enabled by default.
17109 @item -mno-optimize-membar
17110 @opindex mno-optimize-membar
17112 This switch disables the automatic removal of redundant @code{membar}
17113 instructions from the generated code.
17115 @item -mtomcat-stats
17116 @opindex mtomcat-stats
17118 Cause gas to print out tomcat statistics.
17120 @item -mcpu=@var{cpu}
17123 Select the processor type for which to generate code. Possible values are
17124 @samp{frv}, @samp{fr550}, @samp{tomcat}, @samp{fr500}, @samp{fr450},
17125 @samp{fr405}, @samp{fr400}, @samp{fr300} and @samp{simple}.
17129 @node GNU/Linux Options
17130 @subsection GNU/Linux Options
17132 These @samp{-m} options are defined for GNU/Linux targets:
17137 Use the GNU C library. This is the default except
17138 on @samp{*-*-linux-*uclibc*}, @samp{*-*-linux-*musl*} and
17139 @samp{*-*-linux-*android*} targets.
17143 Use uClibc C library. This is the default on
17144 @samp{*-*-linux-*uclibc*} targets.
17148 Use the musl C library. This is the default on
17149 @samp{*-*-linux-*musl*} targets.
17153 Use Bionic C library. This is the default on
17154 @samp{*-*-linux-*android*} targets.
17158 Compile code compatible with Android platform. This is the default on
17159 @samp{*-*-linux-*android*} targets.
17161 When compiling, this option enables @option{-mbionic}, @option{-fPIC},
17162 @option{-fno-exceptions} and @option{-fno-rtti} by default. When linking,
17163 this option makes the GCC driver pass Android-specific options to the linker.
17164 Finally, this option causes the preprocessor macro @code{__ANDROID__}
17167 @item -tno-android-cc
17168 @opindex tno-android-cc
17169 Disable compilation effects of @option{-mandroid}, i.e., do not enable
17170 @option{-mbionic}, @option{-fPIC}, @option{-fno-exceptions} and
17171 @option{-fno-rtti} by default.
17173 @item -tno-android-ld
17174 @opindex tno-android-ld
17175 Disable linking effects of @option{-mandroid}, i.e., pass standard Linux
17176 linking options to the linker.
17180 @node H8/300 Options
17181 @subsection H8/300 Options
17183 These @samp{-m} options are defined for the H8/300 implementations:
17188 Shorten some address references at link time, when possible; uses the
17189 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
17190 ld, Using ld}, for a fuller description.
17194 Generate code for the H8/300H@.
17198 Generate code for the H8S@.
17202 Generate code for the H8S and H8/300H in the normal mode. This switch
17203 must be used either with @option{-mh} or @option{-ms}.
17207 Generate code for the H8S/2600. This switch must be used with @option{-ms}.
17211 Extended registers are stored on stack before execution of function
17212 with monitor attribute. Default option is @option{-mexr}.
17213 This option is valid only for H8S targets.
17217 Extended registers are not stored on stack before execution of function
17218 with monitor attribute. Default option is @option{-mno-exr}.
17219 This option is valid only for H8S targets.
17223 Make @code{int} data 32 bits by default.
17226 @opindex malign-300
17227 On the H8/300H and H8S, use the same alignment rules as for the H8/300.
17228 The default for the H8/300H and H8S is to align longs and floats on
17230 @option{-malign-300} causes them to be aligned on 2-byte boundaries.
17231 This option has no effect on the H8/300.
17235 @subsection HPPA Options
17236 @cindex HPPA Options
17238 These @samp{-m} options are defined for the HPPA family of computers:
17241 @item -march=@var{architecture-type}
17243 Generate code for the specified architecture. The choices for
17244 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
17245 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
17246 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
17247 architecture option for your machine. Code compiled for lower numbered
17248 architectures runs on higher numbered architectures, but not the
17251 @item -mpa-risc-1-0
17252 @itemx -mpa-risc-1-1
17253 @itemx -mpa-risc-2-0
17254 @opindex mpa-risc-1-0
17255 @opindex mpa-risc-1-1
17256 @opindex mpa-risc-2-0
17257 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
17259 @item -mjump-in-delay
17260 @opindex mjump-in-delay
17261 This option is ignored and provided for compatibility purposes only.
17263 @item -mdisable-fpregs
17264 @opindex mdisable-fpregs
17265 Prevent floating-point registers from being used in any manner. This is
17266 necessary for compiling kernels that perform lazy context switching of
17267 floating-point registers. If you use this option and attempt to perform
17268 floating-point operations, the compiler aborts.
17270 @item -mdisable-indexing
17271 @opindex mdisable-indexing
17272 Prevent the compiler from using indexing address modes. This avoids some
17273 rather obscure problems when compiling MIG generated code under MACH@.
17275 @item -mno-space-regs
17276 @opindex mno-space-regs
17277 Generate code that assumes the target has no space registers. This allows
17278 GCC to generate faster indirect calls and use unscaled index address modes.
17280 Such code is suitable for level 0 PA systems and kernels.
17282 @item -mfast-indirect-calls
17283 @opindex mfast-indirect-calls
17284 Generate code that assumes calls never cross space boundaries. This
17285 allows GCC to emit code that performs faster indirect calls.
17287 This option does not work in the presence of shared libraries or nested
17290 @item -mfixed-range=@var{register-range}
17291 @opindex mfixed-range
17292 Generate code treating the given register range as fixed registers.
17293 A fixed register is one that the register allocator cannot use. This is
17294 useful when compiling kernel code. A register range is specified as
17295 two registers separated by a dash. Multiple register ranges can be
17296 specified separated by a comma.
17298 @item -mlong-load-store
17299 @opindex mlong-load-store
17300 Generate 3-instruction load and store sequences as sometimes required by
17301 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
17304 @item -mportable-runtime
17305 @opindex mportable-runtime
17306 Use the portable calling conventions proposed by HP for ELF systems.
17310 Enable the use of assembler directives only GAS understands.
17312 @item -mschedule=@var{cpu-type}
17314 Schedule code according to the constraints for the machine type
17315 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
17316 @samp{7100}, @samp{7100LC}, @samp{7200}, @samp{7300} and @samp{8000}. Refer
17317 to @file{/usr/lib/sched.models} on an HP-UX system to determine the
17318 proper scheduling option for your machine. The default scheduling is
17322 @opindex mlinker-opt
17323 Enable the optimization pass in the HP-UX linker. Note this makes symbolic
17324 debugging impossible. It also triggers a bug in the HP-UX 8 and HP-UX 9
17325 linkers in which they give bogus error messages when linking some programs.
17328 @opindex msoft-float
17329 Generate output containing library calls for floating point.
17330 @strong{Warning:} the requisite libraries are not available for all HPPA
17331 targets. Normally the facilities of the machine's usual C compiler are
17332 used, but this cannot be done directly in cross-compilation. You must make
17333 your own arrangements to provide suitable library functions for
17336 @option{-msoft-float} changes the calling convention in the output file;
17337 therefore, it is only useful if you compile @emph{all} of a program with
17338 this option. In particular, you need to compile @file{libgcc.a}, the
17339 library that comes with GCC, with @option{-msoft-float} in order for
17344 Generate the predefine, @code{_SIO}, for server IO@. The default is
17345 @option{-mwsio}. This generates the predefines, @code{__hp9000s700},
17346 @code{__hp9000s700__} and @code{_WSIO}, for workstation IO@. These
17347 options are available under HP-UX and HI-UX@.
17351 Use options specific to GNU @command{ld}.
17352 This passes @option{-shared} to @command{ld} when
17353 building a shared library. It is the default when GCC is configured,
17354 explicitly or implicitly, with the GNU linker. This option does not
17355 affect which @command{ld} is called; it only changes what parameters
17356 are passed to that @command{ld}.
17357 The @command{ld} that is called is determined by the
17358 @option{--with-ld} configure option, GCC's program search path, and
17359 finally by the user's @env{PATH}. The linker used by GCC can be printed
17360 using @samp{which `gcc -print-prog-name=ld`}. This option is only available
17361 on the 64-bit HP-UX GCC, i.e.@: configured with @samp{hppa*64*-*-hpux*}.
17365 Use options specific to HP @command{ld}.
17366 This passes @option{-b} to @command{ld} when building
17367 a shared library and passes @option{+Accept TypeMismatch} to @command{ld} on all
17368 links. It is the default when GCC is configured, explicitly or
17369 implicitly, with the HP linker. This option does not affect
17370 which @command{ld} is called; it only changes what parameters are passed to that
17372 The @command{ld} that is called is determined by the @option{--with-ld}
17373 configure option, GCC's program search path, and finally by the user's
17374 @env{PATH}. The linker used by GCC can be printed using @samp{which
17375 `gcc -print-prog-name=ld`}. This option is only available on the 64-bit
17376 HP-UX GCC, i.e.@: configured with @samp{hppa*64*-*-hpux*}.
17379 @opindex mno-long-calls
17380 Generate code that uses long call sequences. This ensures that a call
17381 is always able to reach linker generated stubs. The default is to generate
17382 long calls only when the distance from the call site to the beginning
17383 of the function or translation unit, as the case may be, exceeds a
17384 predefined limit set by the branch type being used. The limits for
17385 normal calls are 7,600,000 and 240,000 bytes, respectively for the
17386 PA 2.0 and PA 1.X architectures. Sibcalls are always limited at
17389 Distances are measured from the beginning of functions when using the
17390 @option{-ffunction-sections} option, or when using the @option{-mgas}
17391 and @option{-mno-portable-runtime} options together under HP-UX with
17394 It is normally not desirable to use this option as it degrades
17395 performance. However, it may be useful in large applications,
17396 particularly when partial linking is used to build the application.
17398 The types of long calls used depends on the capabilities of the
17399 assembler and linker, and the type of code being generated. The
17400 impact on systems that support long absolute calls, and long pic
17401 symbol-difference or pc-relative calls should be relatively small.
17402 However, an indirect call is used on 32-bit ELF systems in pic code
17403 and it is quite long.
17405 @item -munix=@var{unix-std}
17407 Generate compiler predefines and select a startfile for the specified
17408 UNIX standard. The choices for @var{unix-std} are @samp{93}, @samp{95}
17409 and @samp{98}. @samp{93} is supported on all HP-UX versions. @samp{95}
17410 is available on HP-UX 10.10 and later. @samp{98} is available on HP-UX
17411 11.11 and later. The default values are @samp{93} for HP-UX 10.00,
17412 @samp{95} for HP-UX 10.10 though to 11.00, and @samp{98} for HP-UX 11.11
17415 @option{-munix=93} provides the same predefines as GCC 3.3 and 3.4.
17416 @option{-munix=95} provides additional predefines for @code{XOPEN_UNIX}
17417 and @code{_XOPEN_SOURCE_EXTENDED}, and the startfile @file{unix95.o}.
17418 @option{-munix=98} provides additional predefines for @code{_XOPEN_UNIX},
17419 @code{_XOPEN_SOURCE_EXTENDED}, @code{_INCLUDE__STDC_A1_SOURCE} and
17420 @code{_INCLUDE_XOPEN_SOURCE_500}, and the startfile @file{unix98.o}.
17422 It is @emph{important} to note that this option changes the interfaces
17423 for various library routines. It also affects the operational behavior
17424 of the C library. Thus, @emph{extreme} care is needed in using this
17427 Library code that is intended to operate with more than one UNIX
17428 standard must test, set and restore the variable @code{__xpg4_extended_mask}
17429 as appropriate. Most GNU software doesn't provide this capability.
17433 Suppress the generation of link options to search libdld.sl when the
17434 @option{-static} option is specified on HP-UX 10 and later.
17438 The HP-UX implementation of setlocale in libc has a dependency on
17439 libdld.sl. There isn't an archive version of libdld.sl. Thus,
17440 when the @option{-static} option is specified, special link options
17441 are needed to resolve this dependency.
17443 On HP-UX 10 and later, the GCC driver adds the necessary options to
17444 link with libdld.sl when the @option{-static} option is specified.
17445 This causes the resulting binary to be dynamic. On the 64-bit port,
17446 the linkers generate dynamic binaries by default in any case. The
17447 @option{-nolibdld} option can be used to prevent the GCC driver from
17448 adding these link options.
17452 Add support for multithreading with the @dfn{dce thread} library
17453 under HP-UX@. This option sets flags for both the preprocessor and
17457 @node IA-64 Options
17458 @subsection IA-64 Options
17459 @cindex IA-64 Options
17461 These are the @samp{-m} options defined for the Intel IA-64 architecture.
17465 @opindex mbig-endian
17466 Generate code for a big-endian target. This is the default for HP-UX@.
17468 @item -mlittle-endian
17469 @opindex mlittle-endian
17470 Generate code for a little-endian target. This is the default for AIX5
17476 @opindex mno-gnu-as
17477 Generate (or don't) code for the GNU assembler. This is the default.
17478 @c Also, this is the default if the configure option @option{--with-gnu-as}
17484 @opindex mno-gnu-ld
17485 Generate (or don't) code for the GNU linker. This is the default.
17486 @c Also, this is the default if the configure option @option{--with-gnu-ld}
17491 Generate code that does not use a global pointer register. The result
17492 is not position independent code, and violates the IA-64 ABI@.
17494 @item -mvolatile-asm-stop
17495 @itemx -mno-volatile-asm-stop
17496 @opindex mvolatile-asm-stop
17497 @opindex mno-volatile-asm-stop
17498 Generate (or don't) a stop bit immediately before and after volatile asm
17501 @item -mregister-names
17502 @itemx -mno-register-names
17503 @opindex mregister-names
17504 @opindex mno-register-names
17505 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
17506 the stacked registers. This may make assembler output more readable.
17512 Disable (or enable) optimizations that use the small data section. This may
17513 be useful for working around optimizer bugs.
17515 @item -mconstant-gp
17516 @opindex mconstant-gp
17517 Generate code that uses a single constant global pointer value. This is
17518 useful when compiling kernel code.
17522 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
17523 This is useful when compiling firmware code.
17525 @item -minline-float-divide-min-latency
17526 @opindex minline-float-divide-min-latency
17527 Generate code for inline divides of floating-point values
17528 using the minimum latency algorithm.
17530 @item -minline-float-divide-max-throughput
17531 @opindex minline-float-divide-max-throughput
17532 Generate code for inline divides of floating-point values
17533 using the maximum throughput algorithm.
17535 @item -mno-inline-float-divide
17536 @opindex mno-inline-float-divide
17537 Do not generate inline code for divides of floating-point values.
17539 @item -minline-int-divide-min-latency
17540 @opindex minline-int-divide-min-latency
17541 Generate code for inline divides of integer values
17542 using the minimum latency algorithm.
17544 @item -minline-int-divide-max-throughput
17545 @opindex minline-int-divide-max-throughput
17546 Generate code for inline divides of integer values
17547 using the maximum throughput algorithm.
17549 @item -mno-inline-int-divide
17550 @opindex mno-inline-int-divide
17551 Do not generate inline code for divides of integer values.
17553 @item -minline-sqrt-min-latency
17554 @opindex minline-sqrt-min-latency
17555 Generate code for inline square roots
17556 using the minimum latency algorithm.
17558 @item -minline-sqrt-max-throughput
17559 @opindex minline-sqrt-max-throughput
17560 Generate code for inline square roots
17561 using the maximum throughput algorithm.
17563 @item -mno-inline-sqrt
17564 @opindex mno-inline-sqrt
17565 Do not generate inline code for @code{sqrt}.
17568 @itemx -mno-fused-madd
17569 @opindex mfused-madd
17570 @opindex mno-fused-madd
17571 Do (don't) generate code that uses the fused multiply/add or multiply/subtract
17572 instructions. The default is to use these instructions.
17574 @item -mno-dwarf2-asm
17575 @itemx -mdwarf2-asm
17576 @opindex mno-dwarf2-asm
17577 @opindex mdwarf2-asm
17578 Don't (or do) generate assembler code for the DWARF line number debugging
17579 info. This may be useful when not using the GNU assembler.
17581 @item -mearly-stop-bits
17582 @itemx -mno-early-stop-bits
17583 @opindex mearly-stop-bits
17584 @opindex mno-early-stop-bits
17585 Allow stop bits to be placed earlier than immediately preceding the
17586 instruction that triggered the stop bit. This can improve instruction
17587 scheduling, but does not always do so.
17589 @item -mfixed-range=@var{register-range}
17590 @opindex mfixed-range
17591 Generate code treating the given register range as fixed registers.
17592 A fixed register is one that the register allocator cannot use. This is
17593 useful when compiling kernel code. A register range is specified as
17594 two registers separated by a dash. Multiple register ranges can be
17595 specified separated by a comma.
17597 @item -mtls-size=@var{tls-size}
17599 Specify bit size of immediate TLS offsets. Valid values are 14, 22, and
17602 @item -mtune=@var{cpu-type}
17604 Tune the instruction scheduling for a particular CPU, Valid values are
17605 @samp{itanium}, @samp{itanium1}, @samp{merced}, @samp{itanium2},
17606 and @samp{mckinley}.
17612 Generate code for a 32-bit or 64-bit environment.
17613 The 32-bit environment sets int, long and pointer to 32 bits.
17614 The 64-bit environment sets int to 32 bits and long and pointer
17615 to 64 bits. These are HP-UX specific flags.
17617 @item -mno-sched-br-data-spec
17618 @itemx -msched-br-data-spec
17619 @opindex mno-sched-br-data-spec
17620 @opindex msched-br-data-spec
17621 (Dis/En)able data speculative scheduling before reload.
17622 This results in generation of @code{ld.a} instructions and
17623 the corresponding check instructions (@code{ld.c} / @code{chk.a}).
17624 The default setting is disabled.
17626 @item -msched-ar-data-spec
17627 @itemx -mno-sched-ar-data-spec
17628 @opindex msched-ar-data-spec
17629 @opindex mno-sched-ar-data-spec
17630 (En/Dis)able data speculative scheduling after reload.
17631 This results in generation of @code{ld.a} instructions and
17632 the corresponding check instructions (@code{ld.c} / @code{chk.a}).
17633 The default setting is enabled.
17635 @item -mno-sched-control-spec
17636 @itemx -msched-control-spec
17637 @opindex mno-sched-control-spec
17638 @opindex msched-control-spec
17639 (Dis/En)able control speculative scheduling. This feature is
17640 available only during region scheduling (i.e.@: before reload).
17641 This results in generation of the @code{ld.s} instructions and
17642 the corresponding check instructions @code{chk.s}.
17643 The default setting is disabled.
17645 @item -msched-br-in-data-spec
17646 @itemx -mno-sched-br-in-data-spec
17647 @opindex msched-br-in-data-spec
17648 @opindex mno-sched-br-in-data-spec
17649 (En/Dis)able speculative scheduling of the instructions that
17650 are dependent on the data speculative loads before reload.
17651 This is effective only with @option{-msched-br-data-spec} enabled.
17652 The default setting is enabled.
17654 @item -msched-ar-in-data-spec
17655 @itemx -mno-sched-ar-in-data-spec
17656 @opindex msched-ar-in-data-spec
17657 @opindex mno-sched-ar-in-data-spec
17658 (En/Dis)able speculative scheduling of the instructions that
17659 are dependent on the data speculative loads after reload.
17660 This is effective only with @option{-msched-ar-data-spec} enabled.
17661 The default setting is enabled.
17663 @item -msched-in-control-spec
17664 @itemx -mno-sched-in-control-spec
17665 @opindex msched-in-control-spec
17666 @opindex mno-sched-in-control-spec
17667 (En/Dis)able speculative scheduling of the instructions that
17668 are dependent on the control speculative loads.
17669 This is effective only with @option{-msched-control-spec} enabled.
17670 The default setting is enabled.
17672 @item -mno-sched-prefer-non-data-spec-insns
17673 @itemx -msched-prefer-non-data-spec-insns
17674 @opindex mno-sched-prefer-non-data-spec-insns
17675 @opindex msched-prefer-non-data-spec-insns
17676 If enabled, data-speculative instructions are chosen for schedule
17677 only if there are no other choices at the moment. This makes
17678 the use of the data speculation much more conservative.
17679 The default setting is disabled.
17681 @item -mno-sched-prefer-non-control-spec-insns
17682 @itemx -msched-prefer-non-control-spec-insns
17683 @opindex mno-sched-prefer-non-control-spec-insns
17684 @opindex msched-prefer-non-control-spec-insns
17685 If enabled, control-speculative instructions are chosen for schedule
17686 only if there are no other choices at the moment. This makes
17687 the use of the control speculation much more conservative.
17688 The default setting is disabled.
17690 @item -mno-sched-count-spec-in-critical-path
17691 @itemx -msched-count-spec-in-critical-path
17692 @opindex mno-sched-count-spec-in-critical-path
17693 @opindex msched-count-spec-in-critical-path
17694 If enabled, speculative dependencies are considered during
17695 computation of the instructions priorities. This makes the use of the
17696 speculation a bit more conservative.
17697 The default setting is disabled.
17699 @item -msched-spec-ldc
17700 @opindex msched-spec-ldc
17701 Use a simple data speculation check. This option is on by default.
17703 @item -msched-control-spec-ldc
17704 @opindex msched-spec-ldc
17705 Use a simple check for control speculation. This option is on by default.
17707 @item -msched-stop-bits-after-every-cycle
17708 @opindex msched-stop-bits-after-every-cycle
17709 Place a stop bit after every cycle when scheduling. This option is on
17712 @item -msched-fp-mem-deps-zero-cost
17713 @opindex msched-fp-mem-deps-zero-cost
17714 Assume that floating-point stores and loads are not likely to cause a conflict
17715 when placed into the same instruction group. This option is disabled by
17718 @item -msel-sched-dont-check-control-spec
17719 @opindex msel-sched-dont-check-control-spec
17720 Generate checks for control speculation in selective scheduling.
17721 This flag is disabled by default.
17723 @item -msched-max-memory-insns=@var{max-insns}
17724 @opindex msched-max-memory-insns
17725 Limit on the number of memory insns per instruction group, giving lower
17726 priority to subsequent memory insns attempting to schedule in the same
17727 instruction group. Frequently useful to prevent cache bank conflicts.
17728 The default value is 1.
17730 @item -msched-max-memory-insns-hard-limit
17731 @opindex msched-max-memory-insns-hard-limit
17732 Makes the limit specified by @option{msched-max-memory-insns} a hard limit,
17733 disallowing more than that number in an instruction group.
17734 Otherwise, the limit is ``soft'', meaning that non-memory operations
17735 are preferred when the limit is reached, but memory operations may still
17741 @subsection LM32 Options
17742 @cindex LM32 options
17744 These @option{-m} options are defined for the LatticeMico32 architecture:
17747 @item -mbarrel-shift-enabled
17748 @opindex mbarrel-shift-enabled
17749 Enable barrel-shift instructions.
17751 @item -mdivide-enabled
17752 @opindex mdivide-enabled
17753 Enable divide and modulus instructions.
17755 @item -mmultiply-enabled
17756 @opindex multiply-enabled
17757 Enable multiply instructions.
17759 @item -msign-extend-enabled
17760 @opindex msign-extend-enabled
17761 Enable sign extend instructions.
17763 @item -muser-enabled
17764 @opindex muser-enabled
17765 Enable user-defined instructions.
17770 @subsection M32C Options
17771 @cindex M32C options
17774 @item -mcpu=@var{name}
17776 Select the CPU for which code is generated. @var{name} may be one of
17777 @samp{r8c} for the R8C/Tiny series, @samp{m16c} for the M16C (up to
17778 /60) series, @samp{m32cm} for the M16C/80 series, or @samp{m32c} for
17779 the M32C/80 series.
17783 Specifies that the program will be run on the simulator. This causes
17784 an alternate runtime library to be linked in which supports, for
17785 example, file I/O@. You must not use this option when generating
17786 programs that will run on real hardware; you must provide your own
17787 runtime library for whatever I/O functions are needed.
17789 @item -memregs=@var{number}
17791 Specifies the number of memory-based pseudo-registers GCC uses
17792 during code generation. These pseudo-registers are used like real
17793 registers, so there is a tradeoff between GCC's ability to fit the
17794 code into available registers, and the performance penalty of using
17795 memory instead of registers. Note that all modules in a program must
17796 be compiled with the same value for this option. Because of that, you
17797 must not use this option with GCC's default runtime libraries.
17801 @node M32R/D Options
17802 @subsection M32R/D Options
17803 @cindex M32R/D options
17805 These @option{-m} options are defined for Renesas M32R/D architectures:
17810 Generate code for the M32R/2@.
17814 Generate code for the M32R/X@.
17818 Generate code for the M32R@. This is the default.
17820 @item -mmodel=small
17821 @opindex mmodel=small
17822 Assume all objects live in the lower 16MB of memory (so that their addresses
17823 can be loaded with the @code{ld24} instruction), and assume all subroutines
17824 are reachable with the @code{bl} instruction.
17825 This is the default.
17827 The addressability of a particular object can be set with the
17828 @code{model} attribute.
17830 @item -mmodel=medium
17831 @opindex mmodel=medium
17832 Assume objects may be anywhere in the 32-bit address space (the compiler
17833 generates @code{seth/add3} instructions to load their addresses), and
17834 assume all subroutines are reachable with the @code{bl} instruction.
17836 @item -mmodel=large
17837 @opindex mmodel=large
17838 Assume objects may be anywhere in the 32-bit address space (the compiler
17839 generates @code{seth/add3} instructions to load their addresses), and
17840 assume subroutines may not be reachable with the @code{bl} instruction
17841 (the compiler generates the much slower @code{seth/add3/jl}
17842 instruction sequence).
17845 @opindex msdata=none
17846 Disable use of the small data area. Variables are put into
17847 one of @code{.data}, @code{.bss}, or @code{.rodata} (unless the
17848 @code{section} attribute has been specified).
17849 This is the default.
17851 The small data area consists of sections @code{.sdata} and @code{.sbss}.
17852 Objects may be explicitly put in the small data area with the
17853 @code{section} attribute using one of these sections.
17855 @item -msdata=sdata
17856 @opindex msdata=sdata
17857 Put small global and static data in the small data area, but do not
17858 generate special code to reference them.
17861 @opindex msdata=use
17862 Put small global and static data in the small data area, and generate
17863 special instructions to reference them.
17867 @cindex smaller data references
17868 Put global and static objects less than or equal to @var{num} bytes
17869 into the small data or BSS sections instead of the normal data or BSS
17870 sections. The default value of @var{num} is 8.
17871 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
17872 for this option to have any effect.
17874 All modules should be compiled with the same @option{-G @var{num}} value.
17875 Compiling with different values of @var{num} may or may not work; if it
17876 doesn't the linker gives an error message---incorrect code is not
17881 Makes the M32R-specific code in the compiler display some statistics
17882 that might help in debugging programs.
17884 @item -malign-loops
17885 @opindex malign-loops
17886 Align all loops to a 32-byte boundary.
17888 @item -mno-align-loops
17889 @opindex mno-align-loops
17890 Do not enforce a 32-byte alignment for loops. This is the default.
17892 @item -missue-rate=@var{number}
17893 @opindex missue-rate=@var{number}
17894 Issue @var{number} instructions per cycle. @var{number} can only be 1
17897 @item -mbranch-cost=@var{number}
17898 @opindex mbranch-cost=@var{number}
17899 @var{number} can only be 1 or 2. If it is 1 then branches are
17900 preferred over conditional code, if it is 2, then the opposite applies.
17902 @item -mflush-trap=@var{number}
17903 @opindex mflush-trap=@var{number}
17904 Specifies the trap number to use to flush the cache. The default is
17905 12. Valid numbers are between 0 and 15 inclusive.
17907 @item -mno-flush-trap
17908 @opindex mno-flush-trap
17909 Specifies that the cache cannot be flushed by using a trap.
17911 @item -mflush-func=@var{name}
17912 @opindex mflush-func=@var{name}
17913 Specifies the name of the operating system function to call to flush
17914 the cache. The default is @samp{_flush_cache}, but a function call
17915 is only used if a trap is not available.
17917 @item -mno-flush-func
17918 @opindex mno-flush-func
17919 Indicates that there is no OS function for flushing the cache.
17923 @node M680x0 Options
17924 @subsection M680x0 Options
17925 @cindex M680x0 options
17927 These are the @samp{-m} options defined for M680x0 and ColdFire processors.
17928 The default settings depend on which architecture was selected when
17929 the compiler was configured; the defaults for the most common choices
17933 @item -march=@var{arch}
17935 Generate code for a specific M680x0 or ColdFire instruction set
17936 architecture. Permissible values of @var{arch} for M680x0
17937 architectures are: @samp{68000}, @samp{68010}, @samp{68020},
17938 @samp{68030}, @samp{68040}, @samp{68060} and @samp{cpu32}. ColdFire
17939 architectures are selected according to Freescale's ISA classification
17940 and the permissible values are: @samp{isaa}, @samp{isaaplus},
17941 @samp{isab} and @samp{isac}.
17943 GCC defines a macro @code{__mcf@var{arch}__} whenever it is generating
17944 code for a ColdFire target. The @var{arch} in this macro is one of the
17945 @option{-march} arguments given above.
17947 When used together, @option{-march} and @option{-mtune} select code
17948 that runs on a family of similar processors but that is optimized
17949 for a particular microarchitecture.
17951 @item -mcpu=@var{cpu}
17953 Generate code for a specific M680x0 or ColdFire processor.
17954 The M680x0 @var{cpu}s are: @samp{68000}, @samp{68010}, @samp{68020},
17955 @samp{68030}, @samp{68040}, @samp{68060}, @samp{68302}, @samp{68332}
17956 and @samp{cpu32}. The ColdFire @var{cpu}s are given by the table
17957 below, which also classifies the CPUs into families:
17959 @multitable @columnfractions 0.20 0.80
17960 @item @strong{Family} @tab @strong{@samp{-mcpu} arguments}
17961 @item @samp{51} @tab @samp{51} @samp{51ac} @samp{51ag} @samp{51cn} @samp{51em} @samp{51je} @samp{51jf} @samp{51jg} @samp{51jm} @samp{51mm} @samp{51qe} @samp{51qm}
17962 @item @samp{5206} @tab @samp{5202} @samp{5204} @samp{5206}
17963 @item @samp{5206e} @tab @samp{5206e}
17964 @item @samp{5208} @tab @samp{5207} @samp{5208}
17965 @item @samp{5211a} @tab @samp{5210a} @samp{5211a}
17966 @item @samp{5213} @tab @samp{5211} @samp{5212} @samp{5213}
17967 @item @samp{5216} @tab @samp{5214} @samp{5216}
17968 @item @samp{52235} @tab @samp{52230} @samp{52231} @samp{52232} @samp{52233} @samp{52234} @samp{52235}
17969 @item @samp{5225} @tab @samp{5224} @samp{5225}
17970 @item @samp{52259} @tab @samp{52252} @samp{52254} @samp{52255} @samp{52256} @samp{52258} @samp{52259}
17971 @item @samp{5235} @tab @samp{5232} @samp{5233} @samp{5234} @samp{5235} @samp{523x}
17972 @item @samp{5249} @tab @samp{5249}
17973 @item @samp{5250} @tab @samp{5250}
17974 @item @samp{5271} @tab @samp{5270} @samp{5271}
17975 @item @samp{5272} @tab @samp{5272}
17976 @item @samp{5275} @tab @samp{5274} @samp{5275}
17977 @item @samp{5282} @tab @samp{5280} @samp{5281} @samp{5282} @samp{528x}
17978 @item @samp{53017} @tab @samp{53011} @samp{53012} @samp{53013} @samp{53014} @samp{53015} @samp{53016} @samp{53017}
17979 @item @samp{5307} @tab @samp{5307}
17980 @item @samp{5329} @tab @samp{5327} @samp{5328} @samp{5329} @samp{532x}
17981 @item @samp{5373} @tab @samp{5372} @samp{5373} @samp{537x}
17982 @item @samp{5407} @tab @samp{5407}
17983 @item @samp{5475} @tab @samp{5470} @samp{5471} @samp{5472} @samp{5473} @samp{5474} @samp{5475} @samp{547x} @samp{5480} @samp{5481} @samp{5482} @samp{5483} @samp{5484} @samp{5485}
17986 @option{-mcpu=@var{cpu}} overrides @option{-march=@var{arch}} if
17987 @var{arch} is compatible with @var{cpu}. Other combinations of
17988 @option{-mcpu} and @option{-march} are rejected.
17990 GCC defines the macro @code{__mcf_cpu_@var{cpu}} when ColdFire target
17991 @var{cpu} is selected. It also defines @code{__mcf_family_@var{family}},
17992 where the value of @var{family} is given by the table above.
17994 @item -mtune=@var{tune}
17996 Tune the code for a particular microarchitecture within the
17997 constraints set by @option{-march} and @option{-mcpu}.
17998 The M680x0 microarchitectures are: @samp{68000}, @samp{68010},
17999 @samp{68020}, @samp{68030}, @samp{68040}, @samp{68060}
18000 and @samp{cpu32}. The ColdFire microarchitectures
18001 are: @samp{cfv1}, @samp{cfv2}, @samp{cfv3}, @samp{cfv4} and @samp{cfv4e}.
18003 You can also use @option{-mtune=68020-40} for code that needs
18004 to run relatively well on 68020, 68030 and 68040 targets.
18005 @option{-mtune=68020-60} is similar but includes 68060 targets
18006 as well. These two options select the same tuning decisions as
18007 @option{-m68020-40} and @option{-m68020-60} respectively.
18009 GCC defines the macros @code{__mc@var{arch}} and @code{__mc@var{arch}__}
18010 when tuning for 680x0 architecture @var{arch}. It also defines
18011 @code{mc@var{arch}} unless either @option{-ansi} or a non-GNU @option{-std}
18012 option is used. If GCC is tuning for a range of architectures,
18013 as selected by @option{-mtune=68020-40} or @option{-mtune=68020-60},
18014 it defines the macros for every architecture in the range.
18016 GCC also defines the macro @code{__m@var{uarch}__} when tuning for
18017 ColdFire microarchitecture @var{uarch}, where @var{uarch} is one
18018 of the arguments given above.
18024 Generate output for a 68000. This is the default
18025 when the compiler is configured for 68000-based systems.
18026 It is equivalent to @option{-march=68000}.
18028 Use this option for microcontrollers with a 68000 or EC000 core,
18029 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
18033 Generate output for a 68010. This is the default
18034 when the compiler is configured for 68010-based systems.
18035 It is equivalent to @option{-march=68010}.
18041 Generate output for a 68020. This is the default
18042 when the compiler is configured for 68020-based systems.
18043 It is equivalent to @option{-march=68020}.
18047 Generate output for a 68030. This is the default when the compiler is
18048 configured for 68030-based systems. It is equivalent to
18049 @option{-march=68030}.
18053 Generate output for a 68040. This is the default when the compiler is
18054 configured for 68040-based systems. It is equivalent to
18055 @option{-march=68040}.
18057 This option inhibits the use of 68881/68882 instructions that have to be
18058 emulated by software on the 68040. Use this option if your 68040 does not
18059 have code to emulate those instructions.
18063 Generate output for a 68060. This is the default when the compiler is
18064 configured for 68060-based systems. It is equivalent to
18065 @option{-march=68060}.
18067 This option inhibits the use of 68020 and 68881/68882 instructions that
18068 have to be emulated by software on the 68060. Use this option if your 68060
18069 does not have code to emulate those instructions.
18073 Generate output for a CPU32. This is the default
18074 when the compiler is configured for CPU32-based systems.
18075 It is equivalent to @option{-march=cpu32}.
18077 Use this option for microcontrollers with a
18078 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
18079 68336, 68340, 68341, 68349 and 68360.
18083 Generate output for a 520X ColdFire CPU@. This is the default
18084 when the compiler is configured for 520X-based systems.
18085 It is equivalent to @option{-mcpu=5206}, and is now deprecated
18086 in favor of that option.
18088 Use this option for microcontroller with a 5200 core, including
18089 the MCF5202, MCF5203, MCF5204 and MCF5206.
18093 Generate output for a 5206e ColdFire CPU@. The option is now
18094 deprecated in favor of the equivalent @option{-mcpu=5206e}.
18098 Generate output for a member of the ColdFire 528X family.
18099 The option is now deprecated in favor of the equivalent
18100 @option{-mcpu=528x}.
18104 Generate output for a ColdFire 5307 CPU@. The option is now deprecated
18105 in favor of the equivalent @option{-mcpu=5307}.
18109 Generate output for a ColdFire 5407 CPU@. The option is now deprecated
18110 in favor of the equivalent @option{-mcpu=5407}.
18114 Generate output for a ColdFire V4e family CPU (e.g.@: 547x/548x).
18115 This includes use of hardware floating-point instructions.
18116 The option is equivalent to @option{-mcpu=547x}, and is now
18117 deprecated in favor of that option.
18121 Generate output for a 68040, without using any of the new instructions.
18122 This results in code that can run relatively efficiently on either a
18123 68020/68881 or a 68030 or a 68040. The generated code does use the
18124 68881 instructions that are emulated on the 68040.
18126 The option is equivalent to @option{-march=68020} @option{-mtune=68020-40}.
18130 Generate output for a 68060, without using any of the new instructions.
18131 This results in code that can run relatively efficiently on either a
18132 68020/68881 or a 68030 or a 68040. The generated code does use the
18133 68881 instructions that are emulated on the 68060.
18135 The option is equivalent to @option{-march=68020} @option{-mtune=68020-60}.
18139 @opindex mhard-float
18141 Generate floating-point instructions. This is the default for 68020
18142 and above, and for ColdFire devices that have an FPU@. It defines the
18143 macro @code{__HAVE_68881__} on M680x0 targets and @code{__mcffpu__}
18144 on ColdFire targets.
18147 @opindex msoft-float
18148 Do not generate floating-point instructions; use library calls instead.
18149 This is the default for 68000, 68010, and 68832 targets. It is also
18150 the default for ColdFire devices that have no FPU.
18156 Generate (do not generate) ColdFire hardware divide and remainder
18157 instructions. If @option{-march} is used without @option{-mcpu},
18158 the default is ``on'' for ColdFire architectures and ``off'' for M680x0
18159 architectures. Otherwise, the default is taken from the target CPU
18160 (either the default CPU, or the one specified by @option{-mcpu}). For
18161 example, the default is ``off'' for @option{-mcpu=5206} and ``on'' for
18162 @option{-mcpu=5206e}.
18164 GCC defines the macro @code{__mcfhwdiv__} when this option is enabled.
18168 Consider type @code{int} to be 16 bits wide, like @code{short int}.
18169 Additionally, parameters passed on the stack are also aligned to a
18170 16-bit boundary even on targets whose API mandates promotion to 32-bit.
18174 Do not consider type @code{int} to be 16 bits wide. This is the default.
18177 @itemx -mno-bitfield
18178 @opindex mnobitfield
18179 @opindex mno-bitfield
18180 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
18181 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
18185 Do use the bit-field instructions. The @option{-m68020} option implies
18186 @option{-mbitfield}. This is the default if you use a configuration
18187 designed for a 68020.
18191 Use a different function-calling convention, in which functions
18192 that take a fixed number of arguments return with the @code{rtd}
18193 instruction, which pops their arguments while returning. This
18194 saves one instruction in the caller since there is no need to pop
18195 the arguments there.
18197 This calling convention is incompatible with the one normally
18198 used on Unix, so you cannot use it if you need to call libraries
18199 compiled with the Unix compiler.
18201 Also, you must provide function prototypes for all functions that
18202 take variable numbers of arguments (including @code{printf});
18203 otherwise incorrect code is generated for calls to those
18206 In addition, seriously incorrect code results if you call a
18207 function with too many arguments. (Normally, extra arguments are
18208 harmlessly ignored.)
18210 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
18211 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
18215 Do not use the calling conventions selected by @option{-mrtd}.
18216 This is the default.
18219 @itemx -mno-align-int
18220 @opindex malign-int
18221 @opindex mno-align-int
18222 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
18223 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
18224 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
18225 Aligning variables on 32-bit boundaries produces code that runs somewhat
18226 faster on processors with 32-bit busses at the expense of more memory.
18228 @strong{Warning:} if you use the @option{-malign-int} switch, GCC
18229 aligns structures containing the above types differently than
18230 most published application binary interface specifications for the m68k.
18234 Use the pc-relative addressing mode of the 68000 directly, instead of
18235 using a global offset table. At present, this option implies @option{-fpic},
18236 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
18237 not presently supported with @option{-mpcrel}, though this could be supported for
18238 68020 and higher processors.
18240 @item -mno-strict-align
18241 @itemx -mstrict-align
18242 @opindex mno-strict-align
18243 @opindex mstrict-align
18244 Do not (do) assume that unaligned memory references are handled by
18248 Generate code that allows the data segment to be located in a different
18249 area of memory from the text segment. This allows for execute-in-place in
18250 an environment without virtual memory management. This option implies
18253 @item -mno-sep-data
18254 Generate code that assumes that the data segment follows the text segment.
18255 This is the default.
18257 @item -mid-shared-library
18258 Generate code that supports shared libraries via the library ID method.
18259 This allows for execute-in-place and shared libraries in an environment
18260 without virtual memory management. This option implies @option{-fPIC}.
18262 @item -mno-id-shared-library
18263 Generate code that doesn't assume ID-based shared libraries are being used.
18264 This is the default.
18266 @item -mshared-library-id=n
18267 Specifies the identification number of the ID-based shared library being
18268 compiled. Specifying a value of 0 generates more compact code; specifying
18269 other values forces the allocation of that number to the current
18270 library, but is no more space- or time-efficient than omitting this option.
18276 When generating position-independent code for ColdFire, generate code
18277 that works if the GOT has more than 8192 entries. This code is
18278 larger and slower than code generated without this option. On M680x0
18279 processors, this option is not needed; @option{-fPIC} suffices.
18281 GCC normally uses a single instruction to load values from the GOT@.
18282 While this is relatively efficient, it only works if the GOT
18283 is smaller than about 64k. Anything larger causes the linker
18284 to report an error such as:
18286 @cindex relocation truncated to fit (ColdFire)
18288 relocation truncated to fit: R_68K_GOT16O foobar
18291 If this happens, you should recompile your code with @option{-mxgot}.
18292 It should then work with very large GOTs. However, code generated with
18293 @option{-mxgot} is less efficient, since it takes 4 instructions to fetch
18294 the value of a global symbol.
18296 Note that some linkers, including newer versions of the GNU linker,
18297 can create multiple GOTs and sort GOT entries. If you have such a linker,
18298 you should only need to use @option{-mxgot} when compiling a single
18299 object file that accesses more than 8192 GOT entries. Very few do.
18301 These options have no effect unless GCC is generating
18302 position-independent code.
18306 @node MCore Options
18307 @subsection MCore Options
18308 @cindex MCore options
18310 These are the @samp{-m} options defined for the Motorola M*Core
18316 @itemx -mno-hardlit
18318 @opindex mno-hardlit
18319 Inline constants into the code stream if it can be done in two
18320 instructions or less.
18326 Use the divide instruction. (Enabled by default).
18328 @item -mrelax-immediate
18329 @itemx -mno-relax-immediate
18330 @opindex mrelax-immediate
18331 @opindex mno-relax-immediate
18332 Allow arbitrary-sized immediates in bit operations.
18334 @item -mwide-bitfields
18335 @itemx -mno-wide-bitfields
18336 @opindex mwide-bitfields
18337 @opindex mno-wide-bitfields
18338 Always treat bit-fields as @code{int}-sized.
18340 @item -m4byte-functions
18341 @itemx -mno-4byte-functions
18342 @opindex m4byte-functions
18343 @opindex mno-4byte-functions
18344 Force all functions to be aligned to a 4-byte boundary.
18346 @item -mcallgraph-data
18347 @itemx -mno-callgraph-data
18348 @opindex mcallgraph-data
18349 @opindex mno-callgraph-data
18350 Emit callgraph information.
18353 @itemx -mno-slow-bytes
18354 @opindex mslow-bytes
18355 @opindex mno-slow-bytes
18356 Prefer word access when reading byte quantities.
18358 @item -mlittle-endian
18359 @itemx -mbig-endian
18360 @opindex mlittle-endian
18361 @opindex mbig-endian
18362 Generate code for a little-endian target.
18368 Generate code for the 210 processor.
18372 Assume that runtime support has been provided and so omit the
18373 simulator library (@file{libsim.a)} from the linker command line.
18375 @item -mstack-increment=@var{size}
18376 @opindex mstack-increment
18377 Set the maximum amount for a single stack increment operation. Large
18378 values can increase the speed of programs that contain functions
18379 that need a large amount of stack space, but they can also trigger a
18380 segmentation fault if the stack is extended too much. The default
18386 @subsection MeP Options
18387 @cindex MeP options
18393 Enables the @code{abs} instruction, which is the absolute difference
18394 between two registers.
18398 Enables all the optional instructions---average, multiply, divide, bit
18399 operations, leading zero, absolute difference, min/max, clip, and
18405 Enables the @code{ave} instruction, which computes the average of two
18408 @item -mbased=@var{n}
18410 Variables of size @var{n} bytes or smaller are placed in the
18411 @code{.based} section by default. Based variables use the @code{$tp}
18412 register as a base register, and there is a 128-byte limit to the
18413 @code{.based} section.
18417 Enables the bit operation instructions---bit test (@code{btstm}), set
18418 (@code{bsetm}), clear (@code{bclrm}), invert (@code{bnotm}), and
18419 test-and-set (@code{tas}).
18421 @item -mc=@var{name}
18423 Selects which section constant data is placed in. @var{name} may
18424 be @samp{tiny}, @samp{near}, or @samp{far}.
18428 Enables the @code{clip} instruction. Note that @option{-mclip} is not
18429 useful unless you also provide @option{-mminmax}.
18431 @item -mconfig=@var{name}
18433 Selects one of the built-in core configurations. Each MeP chip has
18434 one or more modules in it; each module has a core CPU and a variety of
18435 coprocessors, optional instructions, and peripherals. The
18436 @code{MeP-Integrator} tool, not part of GCC, provides these
18437 configurations through this option; using this option is the same as
18438 using all the corresponding command-line options. The default
18439 configuration is @samp{default}.
18443 Enables the coprocessor instructions. By default, this is a 32-bit
18444 coprocessor. Note that the coprocessor is normally enabled via the
18445 @option{-mconfig=} option.
18449 Enables the 32-bit coprocessor's instructions.
18453 Enables the 64-bit coprocessor's instructions.
18457 Enables IVC2 scheduling. IVC2 is a 64-bit VLIW coprocessor.
18461 Causes constant variables to be placed in the @code{.near} section.
18465 Enables the @code{div} and @code{divu} instructions.
18469 Generate big-endian code.
18473 Generate little-endian code.
18475 @item -mio-volatile
18476 @opindex mio-volatile
18477 Tells the compiler that any variable marked with the @code{io}
18478 attribute is to be considered volatile.
18482 Causes variables to be assigned to the @code{.far} section by default.
18486 Enables the @code{leadz} (leading zero) instruction.
18490 Causes variables to be assigned to the @code{.near} section by default.
18494 Enables the @code{min} and @code{max} instructions.
18498 Enables the multiplication and multiply-accumulate instructions.
18502 Disables all the optional instructions enabled by @option{-mall-opts}.
18506 Enables the @code{repeat} and @code{erepeat} instructions, used for
18507 low-overhead looping.
18511 Causes all variables to default to the @code{.tiny} section. Note
18512 that there is a 65536-byte limit to this section. Accesses to these
18513 variables use the @code{%gp} base register.
18517 Enables the saturation instructions. Note that the compiler does not
18518 currently generate these itself, but this option is included for
18519 compatibility with other tools, like @code{as}.
18523 Link the SDRAM-based runtime instead of the default ROM-based runtime.
18527 Link the simulator run-time libraries.
18531 Link the simulator runtime libraries, excluding built-in support
18532 for reset and exception vectors and tables.
18536 Causes all functions to default to the @code{.far} section. Without
18537 this option, functions default to the @code{.near} section.
18539 @item -mtiny=@var{n}
18541 Variables that are @var{n} bytes or smaller are allocated to the
18542 @code{.tiny} section. These variables use the @code{$gp} base
18543 register. The default for this option is 4, but note that there's a
18544 65536-byte limit to the @code{.tiny} section.
18548 @node MicroBlaze Options
18549 @subsection MicroBlaze Options
18550 @cindex MicroBlaze Options
18555 @opindex msoft-float
18556 Use software emulation for floating point (default).
18559 @opindex mhard-float
18560 Use hardware floating-point instructions.
18564 Do not optimize block moves, use @code{memcpy}.
18566 @item -mno-clearbss
18567 @opindex mno-clearbss
18568 This option is deprecated. Use @option{-fno-zero-initialized-in-bss} instead.
18570 @item -mcpu=@var{cpu-type}
18572 Use features of, and schedule code for, the given CPU.
18573 Supported values are in the format @samp{v@var{X}.@var{YY}.@var{Z}},
18574 where @var{X} is a major version, @var{YY} is the minor version, and
18575 @var{Z} is compatibility code. Example values are @samp{v3.00.a},
18576 @samp{v4.00.b}, @samp{v5.00.a}, @samp{v5.00.b}, @samp{v5.00.b}, @samp{v6.00.a}.
18578 @item -mxl-soft-mul
18579 @opindex mxl-soft-mul
18580 Use software multiply emulation (default).
18582 @item -mxl-soft-div
18583 @opindex mxl-soft-div
18584 Use software emulation for divides (default).
18586 @item -mxl-barrel-shift
18587 @opindex mxl-barrel-shift
18588 Use the hardware barrel shifter.
18590 @item -mxl-pattern-compare
18591 @opindex mxl-pattern-compare
18592 Use pattern compare instructions.
18594 @item -msmall-divides
18595 @opindex msmall-divides
18596 Use table lookup optimization for small signed integer divisions.
18598 @item -mxl-stack-check
18599 @opindex mxl-stack-check
18600 This option is deprecated. Use @option{-fstack-check} instead.
18603 @opindex mxl-gp-opt
18604 Use GP-relative @code{.sdata}/@code{.sbss} sections.
18606 @item -mxl-multiply-high
18607 @opindex mxl-multiply-high
18608 Use multiply high instructions for high part of 32x32 multiply.
18610 @item -mxl-float-convert
18611 @opindex mxl-float-convert
18612 Use hardware floating-point conversion instructions.
18614 @item -mxl-float-sqrt
18615 @opindex mxl-float-sqrt
18616 Use hardware floating-point square root instruction.
18619 @opindex mbig-endian
18620 Generate code for a big-endian target.
18622 @item -mlittle-endian
18623 @opindex mlittle-endian
18624 Generate code for a little-endian target.
18627 @opindex mxl-reorder
18628 Use reorder instructions (swap and byte reversed load/store).
18630 @item -mxl-mode-@var{app-model}
18631 Select application model @var{app-model}. Valid models are
18634 normal executable (default), uses startup code @file{crt0.o}.
18637 for use with Xilinx Microprocessor Debugger (XMD) based
18638 software intrusive debug agent called xmdstub. This uses startup file
18639 @file{crt1.o} and sets the start address of the program to 0x800.
18642 for applications that are loaded using a bootloader.
18643 This model uses startup file @file{crt2.o} which does not contain a processor
18644 reset vector handler. This is suitable for transferring control on a
18645 processor reset to the bootloader rather than the application.
18648 for applications that do not require any of the
18649 MicroBlaze vectors. This option may be useful for applications running
18650 within a monitoring application. This model uses @file{crt3.o} as a startup file.
18653 Option @option{-xl-mode-@var{app-model}} is a deprecated alias for
18654 @option{-mxl-mode-@var{app-model}}.
18659 @subsection MIPS Options
18660 @cindex MIPS options
18666 Generate big-endian code.
18670 Generate little-endian code. This is the default for @samp{mips*el-*-*}
18673 @item -march=@var{arch}
18675 Generate code that runs on @var{arch}, which can be the name of a
18676 generic MIPS ISA, or the name of a particular processor.
18678 @samp{mips1}, @samp{mips2}, @samp{mips3}, @samp{mips4},
18679 @samp{mips32}, @samp{mips32r2}, @samp{mips32r3}, @samp{mips32r5},
18680 @samp{mips32r6}, @samp{mips64}, @samp{mips64r2}, @samp{mips64r3},
18681 @samp{mips64r5} and @samp{mips64r6}.
18682 The processor names are:
18683 @samp{4kc}, @samp{4km}, @samp{4kp}, @samp{4ksc},
18684 @samp{4kec}, @samp{4kem}, @samp{4kep}, @samp{4ksd},
18685 @samp{5kc}, @samp{5kf},
18687 @samp{24kc}, @samp{24kf2_1}, @samp{24kf1_1},
18688 @samp{24kec}, @samp{24kef2_1}, @samp{24kef1_1},
18689 @samp{34kc}, @samp{34kf2_1}, @samp{34kf1_1}, @samp{34kn},
18690 @samp{74kc}, @samp{74kf2_1}, @samp{74kf1_1}, @samp{74kf3_2},
18691 @samp{1004kc}, @samp{1004kf2_1}, @samp{1004kf1_1},
18694 @samp{loongson2e}, @samp{loongson2f}, @samp{loongson3a},
18696 @samp{m14k}, @samp{m14kc}, @samp{m14ke}, @samp{m14kec},
18697 @samp{m5100}, @samp{m5101},
18698 @samp{octeon}, @samp{octeon+}, @samp{octeon2}, @samp{octeon3},
18701 @samp{r2000}, @samp{r3000}, @samp{r3900}, @samp{r4000}, @samp{r4400},
18702 @samp{r4600}, @samp{r4650}, @samp{r4700}, @samp{r6000}, @samp{r8000},
18703 @samp{rm7000}, @samp{rm9000},
18704 @samp{r10000}, @samp{r12000}, @samp{r14000}, @samp{r16000},
18707 @samp{vr4100}, @samp{vr4111}, @samp{vr4120}, @samp{vr4130}, @samp{vr4300},
18708 @samp{vr5000}, @samp{vr5400}, @samp{vr5500},
18709 @samp{xlr} and @samp{xlp}.
18710 The special value @samp{from-abi} selects the
18711 most compatible architecture for the selected ABI (that is,
18712 @samp{mips1} for 32-bit ABIs and @samp{mips3} for 64-bit ABIs)@.
18714 The native Linux/GNU toolchain also supports the value @samp{native},
18715 which selects the best architecture option for the host processor.
18716 @option{-march=native} has no effect if GCC does not recognize
18719 In processor names, a final @samp{000} can be abbreviated as @samp{k}
18720 (for example, @option{-march=r2k}). Prefixes are optional, and
18721 @samp{vr} may be written @samp{r}.
18723 Names of the form @samp{@var{n}f2_1} refer to processors with
18724 FPUs clocked at half the rate of the core, names of the form
18725 @samp{@var{n}f1_1} refer to processors with FPUs clocked at the same
18726 rate as the core, and names of the form @samp{@var{n}f3_2} refer to
18727 processors with FPUs clocked a ratio of 3:2 with respect to the core.
18728 For compatibility reasons, @samp{@var{n}f} is accepted as a synonym
18729 for @samp{@var{n}f2_1} while @samp{@var{n}x} and @samp{@var{b}fx} are
18730 accepted as synonyms for @samp{@var{n}f1_1}.
18732 GCC defines two macros based on the value of this option. The first
18733 is @code{_MIPS_ARCH}, which gives the name of target architecture, as
18734 a string. The second has the form @code{_MIPS_ARCH_@var{foo}},
18735 where @var{foo} is the capitalized value of @code{_MIPS_ARCH}@.
18736 For example, @option{-march=r2000} sets @code{_MIPS_ARCH}
18737 to @code{"r2000"} and defines the macro @code{_MIPS_ARCH_R2000}.
18739 Note that the @code{_MIPS_ARCH} macro uses the processor names given
18740 above. In other words, it has the full prefix and does not
18741 abbreviate @samp{000} as @samp{k}. In the case of @samp{from-abi},
18742 the macro names the resolved architecture (either @code{"mips1"} or
18743 @code{"mips3"}). It names the default architecture when no
18744 @option{-march} option is given.
18746 @item -mtune=@var{arch}
18748 Optimize for @var{arch}. Among other things, this option controls
18749 the way instructions are scheduled, and the perceived cost of arithmetic
18750 operations. The list of @var{arch} values is the same as for
18753 When this option is not used, GCC optimizes for the processor
18754 specified by @option{-march}. By using @option{-march} and
18755 @option{-mtune} together, it is possible to generate code that
18756 runs on a family of processors, but optimize the code for one
18757 particular member of that family.
18759 @option{-mtune} defines the macros @code{_MIPS_TUNE} and
18760 @code{_MIPS_TUNE_@var{foo}}, which work in the same way as the
18761 @option{-march} ones described above.
18765 Equivalent to @option{-march=mips1}.
18769 Equivalent to @option{-march=mips2}.
18773 Equivalent to @option{-march=mips3}.
18777 Equivalent to @option{-march=mips4}.
18781 Equivalent to @option{-march=mips32}.
18785 Equivalent to @option{-march=mips32r3}.
18789 Equivalent to @option{-march=mips32r5}.
18793 Equivalent to @option{-march=mips32r6}.
18797 Equivalent to @option{-march=mips64}.
18801 Equivalent to @option{-march=mips64r2}.
18805 Equivalent to @option{-march=mips64r3}.
18809 Equivalent to @option{-march=mips64r5}.
18813 Equivalent to @option{-march=mips64r6}.
18818 @opindex mno-mips16
18819 Generate (do not generate) MIPS16 code. If GCC is targeting a
18820 MIPS32 or MIPS64 architecture, it makes use of the MIPS16e ASE@.
18822 MIPS16 code generation can also be controlled on a per-function basis
18823 by means of @code{mips16} and @code{nomips16} attributes.
18824 @xref{Function Attributes}, for more information.
18826 @item -mflip-mips16
18827 @opindex mflip-mips16
18828 Generate MIPS16 code on alternating functions. This option is provided
18829 for regression testing of mixed MIPS16/non-MIPS16 code generation, and is
18830 not intended for ordinary use in compiling user code.
18832 @item -minterlink-compressed
18833 @item -mno-interlink-compressed
18834 @opindex minterlink-compressed
18835 @opindex mno-interlink-compressed
18836 Require (do not require) that code using the standard (uncompressed) MIPS ISA
18837 be link-compatible with MIPS16 and microMIPS code, and vice versa.
18839 For example, code using the standard ISA encoding cannot jump directly
18840 to MIPS16 or microMIPS code; it must either use a call or an indirect jump.
18841 @option{-minterlink-compressed} therefore disables direct jumps unless GCC
18842 knows that the target of the jump is not compressed.
18844 @item -minterlink-mips16
18845 @itemx -mno-interlink-mips16
18846 @opindex minterlink-mips16
18847 @opindex mno-interlink-mips16
18848 Aliases of @option{-minterlink-compressed} and
18849 @option{-mno-interlink-compressed}. These options predate the microMIPS ASE
18850 and are retained for backwards compatibility.
18862 Generate code for the given ABI@.
18864 Note that the EABI has a 32-bit and a 64-bit variant. GCC normally
18865 generates 64-bit code when you select a 64-bit architecture, but you
18866 can use @option{-mgp32} to get 32-bit code instead.
18868 For information about the O64 ABI, see
18869 @uref{http://gcc.gnu.org/@/projects/@/mipso64-abi.html}.
18871 GCC supports a variant of the o32 ABI in which floating-point registers
18872 are 64 rather than 32 bits wide. You can select this combination with
18873 @option{-mabi=32} @option{-mfp64}. This ABI relies on the @code{mthc1}
18874 and @code{mfhc1} instructions and is therefore only supported for
18875 MIPS32R2, MIPS32R3 and MIPS32R5 processors.
18877 The register assignments for arguments and return values remain the
18878 same, but each scalar value is passed in a single 64-bit register
18879 rather than a pair of 32-bit registers. For example, scalar
18880 floating-point values are returned in @samp{$f0} only, not a
18881 @samp{$f0}/@samp{$f1} pair. The set of call-saved registers also
18882 remains the same in that the even-numbered double-precision registers
18885 Two additional variants of the o32 ABI are supported to enable
18886 a transition from 32-bit to 64-bit registers. These are FPXX
18887 (@option{-mfpxx}) and FP64A (@option{-mfp64} @option{-mno-odd-spreg}).
18888 The FPXX extension mandates that all code must execute correctly
18889 when run using 32-bit or 64-bit registers. The code can be interlinked
18890 with either FP32 or FP64, but not both.
18891 The FP64A extension is similar to the FP64 extension but forbids the
18892 use of odd-numbered single-precision registers. This can be used
18893 in conjunction with the @code{FRE} mode of FPUs in MIPS32R5
18894 processors and allows both FP32 and FP64A code to interlink and
18895 run in the same process without changing FPU modes.
18898 @itemx -mno-abicalls
18900 @opindex mno-abicalls
18901 Generate (do not generate) code that is suitable for SVR4-style
18902 dynamic objects. @option{-mabicalls} is the default for SVR4-based
18907 Generate (do not generate) code that is fully position-independent,
18908 and that can therefore be linked into shared libraries. This option
18909 only affects @option{-mabicalls}.
18911 All @option{-mabicalls} code has traditionally been position-independent,
18912 regardless of options like @option{-fPIC} and @option{-fpic}. However,
18913 as an extension, the GNU toolchain allows executables to use absolute
18914 accesses for locally-binding symbols. It can also use shorter GP
18915 initialization sequences and generate direct calls to locally-defined
18916 functions. This mode is selected by @option{-mno-shared}.
18918 @option{-mno-shared} depends on binutils 2.16 or higher and generates
18919 objects that can only be linked by the GNU linker. However, the option
18920 does not affect the ABI of the final executable; it only affects the ABI
18921 of relocatable objects. Using @option{-mno-shared} generally makes
18922 executables both smaller and quicker.
18924 @option{-mshared} is the default.
18930 Assume (do not assume) that the static and dynamic linkers
18931 support PLTs and copy relocations. This option only affects
18932 @option{-mno-shared -mabicalls}. For the n64 ABI, this option
18933 has no effect without @option{-msym32}.
18935 You can make @option{-mplt} the default by configuring
18936 GCC with @option{--with-mips-plt}. The default is
18937 @option{-mno-plt} otherwise.
18943 Lift (do not lift) the usual restrictions on the size of the global
18946 GCC normally uses a single instruction to load values from the GOT@.
18947 While this is relatively efficient, it only works if the GOT
18948 is smaller than about 64k. Anything larger causes the linker
18949 to report an error such as:
18951 @cindex relocation truncated to fit (MIPS)
18953 relocation truncated to fit: R_MIPS_GOT16 foobar
18956 If this happens, you should recompile your code with @option{-mxgot}.
18957 This works with very large GOTs, although the code is also
18958 less efficient, since it takes three instructions to fetch the
18959 value of a global symbol.
18961 Note that some linkers can create multiple GOTs. If you have such a
18962 linker, you should only need to use @option{-mxgot} when a single object
18963 file accesses more than 64k's worth of GOT entries. Very few do.
18965 These options have no effect unless GCC is generating position
18970 Assume that general-purpose registers are 32 bits wide.
18974 Assume that general-purpose registers are 64 bits wide.
18978 Assume that floating-point registers are 32 bits wide.
18982 Assume that floating-point registers are 64 bits wide.
18986 Do not assume the width of floating-point registers.
18989 @opindex mhard-float
18990 Use floating-point coprocessor instructions.
18993 @opindex msoft-float
18994 Do not use floating-point coprocessor instructions. Implement
18995 floating-point calculations using library calls instead.
18999 Equivalent to @option{-msoft-float}, but additionally asserts that the
19000 program being compiled does not perform any floating-point operations.
19001 This option is presently supported only by some bare-metal MIPS
19002 configurations, where it may select a special set of libraries
19003 that lack all floating-point support (including, for example, the
19004 floating-point @code{printf} formats).
19005 If code compiled with @option{-mno-float} accidentally contains
19006 floating-point operations, it is likely to suffer a link-time
19007 or run-time failure.
19009 @item -msingle-float
19010 @opindex msingle-float
19011 Assume that the floating-point coprocessor only supports single-precision
19014 @item -mdouble-float
19015 @opindex mdouble-float
19016 Assume that the floating-point coprocessor supports double-precision
19017 operations. This is the default.
19020 @itemx -mno-odd-spreg
19021 @opindex modd-spreg
19022 @opindex mno-odd-spreg
19023 Enable the use of odd-numbered single-precision floating-point registers
19024 for the o32 ABI. This is the default for processors that are known to
19025 support these registers. When using the o32 FPXX ABI, @option{-mno-odd-spreg}
19029 @itemx -mabs=legacy
19031 @opindex mabs=legacy
19032 These options control the treatment of the special not-a-number (NaN)
19033 IEEE 754 floating-point data with the @code{abs.@i{fmt}} and
19034 @code{neg.@i{fmt}} machine instructions.
19036 By default or when @option{-mabs=legacy} is used the legacy
19037 treatment is selected. In this case these instructions are considered
19038 arithmetic and avoided where correct operation is required and the
19039 input operand might be a NaN. A longer sequence of instructions that
19040 manipulate the sign bit of floating-point datum manually is used
19041 instead unless the @option{-ffinite-math-only} option has also been
19044 The @option{-mabs=2008} option selects the IEEE 754-2008 treatment. In
19045 this case these instructions are considered non-arithmetic and therefore
19046 operating correctly in all cases, including in particular where the
19047 input operand is a NaN. These instructions are therefore always used
19048 for the respective operations.
19051 @itemx -mnan=legacy
19053 @opindex mnan=legacy
19054 These options control the encoding of the special not-a-number (NaN)
19055 IEEE 754 floating-point data.
19057 The @option{-mnan=legacy} option selects the legacy encoding. In this
19058 case quiet NaNs (qNaNs) are denoted by the first bit of their trailing
19059 significand field being 0, whereas signaling NaNs (sNaNs) are denoted
19060 by the first bit of their trailing significand field being 1.
19062 The @option{-mnan=2008} option selects the IEEE 754-2008 encoding. In
19063 this case qNaNs are denoted by the first bit of their trailing
19064 significand field being 1, whereas sNaNs are denoted by the first bit of
19065 their trailing significand field being 0.
19067 The default is @option{-mnan=legacy} unless GCC has been configured with
19068 @option{--with-nan=2008}.
19074 Use (do not use) @samp{ll}, @samp{sc}, and @samp{sync} instructions to
19075 implement atomic memory built-in functions. When neither option is
19076 specified, GCC uses the instructions if the target architecture
19079 @option{-mllsc} is useful if the runtime environment can emulate the
19080 instructions and @option{-mno-llsc} can be useful when compiling for
19081 nonstandard ISAs. You can make either option the default by
19082 configuring GCC with @option{--with-llsc} and @option{--without-llsc}
19083 respectively. @option{--with-llsc} is the default for some
19084 configurations; see the installation documentation for details.
19090 Use (do not use) revision 1 of the MIPS DSP ASE@.
19091 @xref{MIPS DSP Built-in Functions}. This option defines the
19092 preprocessor macro @code{__mips_dsp}. It also defines
19093 @code{__mips_dsp_rev} to 1.
19099 Use (do not use) revision 2 of the MIPS DSP ASE@.
19100 @xref{MIPS DSP Built-in Functions}. This option defines the
19101 preprocessor macros @code{__mips_dsp} and @code{__mips_dspr2}.
19102 It also defines @code{__mips_dsp_rev} to 2.
19105 @itemx -mno-smartmips
19106 @opindex msmartmips
19107 @opindex mno-smartmips
19108 Use (do not use) the MIPS SmartMIPS ASE.
19110 @item -mpaired-single
19111 @itemx -mno-paired-single
19112 @opindex mpaired-single
19113 @opindex mno-paired-single
19114 Use (do not use) paired-single floating-point instructions.
19115 @xref{MIPS Paired-Single Support}. This option requires
19116 hardware floating-point support to be enabled.
19122 Use (do not use) MIPS Digital Media Extension instructions.
19123 This option can only be used when generating 64-bit code and requires
19124 hardware floating-point support to be enabled.
19129 @opindex mno-mips3d
19130 Use (do not use) the MIPS-3D ASE@. @xref{MIPS-3D Built-in Functions}.
19131 The option @option{-mips3d} implies @option{-mpaired-single}.
19134 @itemx -mno-micromips
19135 @opindex mmicromips
19136 @opindex mno-mmicromips
19137 Generate (do not generate) microMIPS code.
19139 MicroMIPS code generation can also be controlled on a per-function basis
19140 by means of @code{micromips} and @code{nomicromips} attributes.
19141 @xref{Function Attributes}, for more information.
19147 Use (do not use) MT Multithreading instructions.
19153 Use (do not use) the MIPS MCU ASE instructions.
19159 Use (do not use) the MIPS Enhanced Virtual Addressing instructions.
19165 Use (do not use) the MIPS Virtualization Application Specific instructions.
19171 Use (do not use) the MIPS eXtended Physical Address (XPA) instructions.
19175 Force @code{long} types to be 64 bits wide. See @option{-mlong32} for
19176 an explanation of the default and the way that the pointer size is
19181 Force @code{long}, @code{int}, and pointer types to be 32 bits wide.
19183 The default size of @code{int}s, @code{long}s and pointers depends on
19184 the ABI@. All the supported ABIs use 32-bit @code{int}s. The n64 ABI
19185 uses 64-bit @code{long}s, as does the 64-bit EABI; the others use
19186 32-bit @code{long}s. Pointers are the same size as @code{long}s,
19187 or the same size as integer registers, whichever is smaller.
19193 Assume (do not assume) that all symbols have 32-bit values, regardless
19194 of the selected ABI@. This option is useful in combination with
19195 @option{-mabi=64} and @option{-mno-abicalls} because it allows GCC
19196 to generate shorter and faster references to symbolic addresses.
19200 Put definitions of externally-visible data in a small data section
19201 if that data is no bigger than @var{num} bytes. GCC can then generate
19202 more efficient accesses to the data; see @option{-mgpopt} for details.
19204 The default @option{-G} option depends on the configuration.
19206 @item -mlocal-sdata
19207 @itemx -mno-local-sdata
19208 @opindex mlocal-sdata
19209 @opindex mno-local-sdata
19210 Extend (do not extend) the @option{-G} behavior to local data too,
19211 such as to static variables in C@. @option{-mlocal-sdata} is the
19212 default for all configurations.
19214 If the linker complains that an application is using too much small data,
19215 you might want to try rebuilding the less performance-critical parts with
19216 @option{-mno-local-sdata}. You might also want to build large
19217 libraries with @option{-mno-local-sdata}, so that the libraries leave
19218 more room for the main program.
19220 @item -mextern-sdata
19221 @itemx -mno-extern-sdata
19222 @opindex mextern-sdata
19223 @opindex mno-extern-sdata
19224 Assume (do not assume) that externally-defined data is in
19225 a small data section if the size of that data is within the @option{-G} limit.
19226 @option{-mextern-sdata} is the default for all configurations.
19228 If you compile a module @var{Mod} with @option{-mextern-sdata} @option{-G
19229 @var{num}} @option{-mgpopt}, and @var{Mod} references a variable @var{Var}
19230 that is no bigger than @var{num} bytes, you must make sure that @var{Var}
19231 is placed in a small data section. If @var{Var} is defined by another
19232 module, you must either compile that module with a high-enough
19233 @option{-G} setting or attach a @code{section} attribute to @var{Var}'s
19234 definition. If @var{Var} is common, you must link the application
19235 with a high-enough @option{-G} setting.
19237 The easiest way of satisfying these restrictions is to compile
19238 and link every module with the same @option{-G} option. However,
19239 you may wish to build a library that supports several different
19240 small data limits. You can do this by compiling the library with
19241 the highest supported @option{-G} setting and additionally using
19242 @option{-mno-extern-sdata} to stop the library from making assumptions
19243 about externally-defined data.
19249 Use (do not use) GP-relative accesses for symbols that are known to be
19250 in a small data section; see @option{-G}, @option{-mlocal-sdata} and
19251 @option{-mextern-sdata}. @option{-mgpopt} is the default for all
19254 @option{-mno-gpopt} is useful for cases where the @code{$gp} register
19255 might not hold the value of @code{_gp}. For example, if the code is
19256 part of a library that might be used in a boot monitor, programs that
19257 call boot monitor routines pass an unknown value in @code{$gp}.
19258 (In such situations, the boot monitor itself is usually compiled
19259 with @option{-G0}.)
19261 @option{-mno-gpopt} implies @option{-mno-local-sdata} and
19262 @option{-mno-extern-sdata}.
19264 @item -membedded-data
19265 @itemx -mno-embedded-data
19266 @opindex membedded-data
19267 @opindex mno-embedded-data
19268 Allocate variables to the read-only data section first if possible, then
19269 next in the small data section if possible, otherwise in data. This gives
19270 slightly slower code than the default, but reduces the amount of RAM required
19271 when executing, and thus may be preferred for some embedded systems.
19273 @item -muninit-const-in-rodata
19274 @itemx -mno-uninit-const-in-rodata
19275 @opindex muninit-const-in-rodata
19276 @opindex mno-uninit-const-in-rodata
19277 Put uninitialized @code{const} variables in the read-only data section.
19278 This option is only meaningful in conjunction with @option{-membedded-data}.
19280 @item -mcode-readable=@var{setting}
19281 @opindex mcode-readable
19282 Specify whether GCC may generate code that reads from executable sections.
19283 There are three possible settings:
19286 @item -mcode-readable=yes
19287 Instructions may freely access executable sections. This is the
19290 @item -mcode-readable=pcrel
19291 MIPS16 PC-relative load instructions can access executable sections,
19292 but other instructions must not do so. This option is useful on 4KSc
19293 and 4KSd processors when the code TLBs have the Read Inhibit bit set.
19294 It is also useful on processors that can be configured to have a dual
19295 instruction/data SRAM interface and that, like the M4K, automatically
19296 redirect PC-relative loads to the instruction RAM.
19298 @item -mcode-readable=no
19299 Instructions must not access executable sections. This option can be
19300 useful on targets that are configured to have a dual instruction/data
19301 SRAM interface but that (unlike the M4K) do not automatically redirect
19302 PC-relative loads to the instruction RAM.
19305 @item -msplit-addresses
19306 @itemx -mno-split-addresses
19307 @opindex msplit-addresses
19308 @opindex mno-split-addresses
19309 Enable (disable) use of the @code{%hi()} and @code{%lo()} assembler
19310 relocation operators. This option has been superseded by
19311 @option{-mexplicit-relocs} but is retained for backwards compatibility.
19313 @item -mexplicit-relocs
19314 @itemx -mno-explicit-relocs
19315 @opindex mexplicit-relocs
19316 @opindex mno-explicit-relocs
19317 Use (do not use) assembler relocation operators when dealing with symbolic
19318 addresses. The alternative, selected by @option{-mno-explicit-relocs},
19319 is to use assembler macros instead.
19321 @option{-mexplicit-relocs} is the default if GCC was configured
19322 to use an assembler that supports relocation operators.
19324 @item -mcheck-zero-division
19325 @itemx -mno-check-zero-division
19326 @opindex mcheck-zero-division
19327 @opindex mno-check-zero-division
19328 Trap (do not trap) on integer division by zero.
19330 The default is @option{-mcheck-zero-division}.
19332 @item -mdivide-traps
19333 @itemx -mdivide-breaks
19334 @opindex mdivide-traps
19335 @opindex mdivide-breaks
19336 MIPS systems check for division by zero by generating either a
19337 conditional trap or a break instruction. Using traps results in
19338 smaller code, but is only supported on MIPS II and later. Also, some
19339 versions of the Linux kernel have a bug that prevents trap from
19340 generating the proper signal (@code{SIGFPE}). Use @option{-mdivide-traps} to
19341 allow conditional traps on architectures that support them and
19342 @option{-mdivide-breaks} to force the use of breaks.
19344 The default is usually @option{-mdivide-traps}, but this can be
19345 overridden at configure time using @option{--with-divide=breaks}.
19346 Divide-by-zero checks can be completely disabled using
19347 @option{-mno-check-zero-division}.
19352 @opindex mno-memcpy
19353 Force (do not force) the use of @code{memcpy} for non-trivial block
19354 moves. The default is @option{-mno-memcpy}, which allows GCC to inline
19355 most constant-sized copies.
19358 @itemx -mno-long-calls
19359 @opindex mlong-calls
19360 @opindex mno-long-calls
19361 Disable (do not disable) use of the @code{jal} instruction. Calling
19362 functions using @code{jal} is more efficient but requires the caller
19363 and callee to be in the same 256 megabyte segment.
19365 This option has no effect on abicalls code. The default is
19366 @option{-mno-long-calls}.
19372 Enable (disable) use of the @code{mad}, @code{madu} and @code{mul}
19373 instructions, as provided by the R4650 ISA@.
19379 Enable (disable) use of the @code{madd} and @code{msub} integer
19380 instructions. The default is @option{-mimadd} on architectures
19381 that support @code{madd} and @code{msub} except for the 74k
19382 architecture where it was found to generate slower code.
19385 @itemx -mno-fused-madd
19386 @opindex mfused-madd
19387 @opindex mno-fused-madd
19388 Enable (disable) use of the floating-point multiply-accumulate
19389 instructions, when they are available. The default is
19390 @option{-mfused-madd}.
19392 On the R8000 CPU when multiply-accumulate instructions are used,
19393 the intermediate product is calculated to infinite precision
19394 and is not subject to the FCSR Flush to Zero bit. This may be
19395 undesirable in some circumstances. On other processors the result
19396 is numerically identical to the equivalent computation using
19397 separate multiply, add, subtract and negate instructions.
19401 Tell the MIPS assembler to not run its preprocessor over user
19402 assembler files (with a @samp{.s} suffix) when assembling them.
19407 @opindex mno-fix-24k
19408 Work around the 24K E48 (lost data on stores during refill) errata.
19409 The workarounds are implemented by the assembler rather than by GCC@.
19412 @itemx -mno-fix-r4000
19413 @opindex mfix-r4000
19414 @opindex mno-fix-r4000
19415 Work around certain R4000 CPU errata:
19418 A double-word or a variable shift may give an incorrect result if executed
19419 immediately after starting an integer division.
19421 A double-word or a variable shift may give an incorrect result if executed
19422 while an integer multiplication is in progress.
19424 An integer division may give an incorrect result if started in a delay slot
19425 of a taken branch or a jump.
19429 @itemx -mno-fix-r4400
19430 @opindex mfix-r4400
19431 @opindex mno-fix-r4400
19432 Work around certain R4400 CPU errata:
19435 A double-word or a variable shift may give an incorrect result if executed
19436 immediately after starting an integer division.
19440 @itemx -mno-fix-r10000
19441 @opindex mfix-r10000
19442 @opindex mno-fix-r10000
19443 Work around certain R10000 errata:
19446 @code{ll}/@code{sc} sequences may not behave atomically on revisions
19447 prior to 3.0. They may deadlock on revisions 2.6 and earlier.
19450 This option can only be used if the target architecture supports
19451 branch-likely instructions. @option{-mfix-r10000} is the default when
19452 @option{-march=r10000} is used; @option{-mno-fix-r10000} is the default
19456 @itemx -mno-fix-rm7000
19457 @opindex mfix-rm7000
19458 Work around the RM7000 @code{dmult}/@code{dmultu} errata. The
19459 workarounds are implemented by the assembler rather than by GCC@.
19462 @itemx -mno-fix-vr4120
19463 @opindex mfix-vr4120
19464 Work around certain VR4120 errata:
19467 @code{dmultu} does not always produce the correct result.
19469 @code{div} and @code{ddiv} do not always produce the correct result if one
19470 of the operands is negative.
19472 The workarounds for the division errata rely on special functions in
19473 @file{libgcc.a}. At present, these functions are only provided by
19474 the @code{mips64vr*-elf} configurations.
19476 Other VR4120 errata require a NOP to be inserted between certain pairs of
19477 instructions. These errata are handled by the assembler, not by GCC itself.
19480 @opindex mfix-vr4130
19481 Work around the VR4130 @code{mflo}/@code{mfhi} errata. The
19482 workarounds are implemented by the assembler rather than by GCC,
19483 although GCC avoids using @code{mflo} and @code{mfhi} if the
19484 VR4130 @code{macc}, @code{macchi}, @code{dmacc} and @code{dmacchi}
19485 instructions are available instead.
19488 @itemx -mno-fix-sb1
19490 Work around certain SB-1 CPU core errata.
19491 (This flag currently works around the SB-1 revision 2
19492 ``F1'' and ``F2'' floating-point errata.)
19494 @item -mr10k-cache-barrier=@var{setting}
19495 @opindex mr10k-cache-barrier
19496 Specify whether GCC should insert cache barriers to avoid the
19497 side-effects of speculation on R10K processors.
19499 In common with many processors, the R10K tries to predict the outcome
19500 of a conditional branch and speculatively executes instructions from
19501 the ``taken'' branch. It later aborts these instructions if the
19502 predicted outcome is wrong. However, on the R10K, even aborted
19503 instructions can have side effects.
19505 This problem only affects kernel stores and, depending on the system,
19506 kernel loads. As an example, a speculatively-executed store may load
19507 the target memory into cache and mark the cache line as dirty, even if
19508 the store itself is later aborted. If a DMA operation writes to the
19509 same area of memory before the ``dirty'' line is flushed, the cached
19510 data overwrites the DMA-ed data. See the R10K processor manual
19511 for a full description, including other potential problems.
19513 One workaround is to insert cache barrier instructions before every memory
19514 access that might be speculatively executed and that might have side
19515 effects even if aborted. @option{-mr10k-cache-barrier=@var{setting}}
19516 controls GCC's implementation of this workaround. It assumes that
19517 aborted accesses to any byte in the following regions does not have
19522 the memory occupied by the current function's stack frame;
19525 the memory occupied by an incoming stack argument;
19528 the memory occupied by an object with a link-time-constant address.
19531 It is the kernel's responsibility to ensure that speculative
19532 accesses to these regions are indeed safe.
19534 If the input program contains a function declaration such as:
19540 then the implementation of @code{foo} must allow @code{j foo} and
19541 @code{jal foo} to be executed speculatively. GCC honors this
19542 restriction for functions it compiles itself. It expects non-GCC
19543 functions (such as hand-written assembly code) to do the same.
19545 The option has three forms:
19548 @item -mr10k-cache-barrier=load-store
19549 Insert a cache barrier before a load or store that might be
19550 speculatively executed and that might have side effects even
19553 @item -mr10k-cache-barrier=store
19554 Insert a cache barrier before a store that might be speculatively
19555 executed and that might have side effects even if aborted.
19557 @item -mr10k-cache-barrier=none
19558 Disable the insertion of cache barriers. This is the default setting.
19561 @item -mflush-func=@var{func}
19562 @itemx -mno-flush-func
19563 @opindex mflush-func
19564 Specifies the function to call to flush the I and D caches, or to not
19565 call any such function. If called, the function must take the same
19566 arguments as the common @code{_flush_func}, that is, the address of the
19567 memory range for which the cache is being flushed, the size of the
19568 memory range, and the number 3 (to flush both caches). The default
19569 depends on the target GCC was configured for, but commonly is either
19570 @code{_flush_func} or @code{__cpu_flush}.
19572 @item mbranch-cost=@var{num}
19573 @opindex mbranch-cost
19574 Set the cost of branches to roughly @var{num} ``simple'' instructions.
19575 This cost is only a heuristic and is not guaranteed to produce
19576 consistent results across releases. A zero cost redundantly selects
19577 the default, which is based on the @option{-mtune} setting.
19579 @item -mbranch-likely
19580 @itemx -mno-branch-likely
19581 @opindex mbranch-likely
19582 @opindex mno-branch-likely
19583 Enable or disable use of Branch Likely instructions, regardless of the
19584 default for the selected architecture. By default, Branch Likely
19585 instructions may be generated if they are supported by the selected
19586 architecture. An exception is for the MIPS32 and MIPS64 architectures
19587 and processors that implement those architectures; for those, Branch
19588 Likely instructions are not be generated by default because the MIPS32
19589 and MIPS64 architectures specifically deprecate their use.
19591 @item -mcompact-branches=never
19592 @itemx -mcompact-branches=optimal
19593 @itemx -mcompact-branches=always
19594 @opindex mcompact-branches=never
19595 @opindex mcompact-branches=optimal
19596 @opindex mcompact-branches=always
19597 These options control which form of branches will be generated. The
19598 default is @option{-mcompact-branches=optimal}.
19600 The @option{-mcompact-branches=never} option ensures that compact branch
19601 instructions will never be generated.
19603 The @option{-mcompact-branches=always} option ensures that a compact
19604 branch instruction will be generated if available. If a compact branch
19605 instruction is not available, a delay slot form of the branch will be
19608 This option is supported from MIPS Release 6 onwards.
19610 The @option{-mcompact-branches=optimal} option will cause a delay slot
19611 branch to be used if one is available in the current ISA and the delay
19612 slot is successfully filled. If the delay slot is not filled, a compact
19613 branch will be chosen if one is available.
19615 @item -mfp-exceptions
19616 @itemx -mno-fp-exceptions
19617 @opindex mfp-exceptions
19618 Specifies whether FP exceptions are enabled. This affects how
19619 FP instructions are scheduled for some processors.
19620 The default is that FP exceptions are
19623 For instance, on the SB-1, if FP exceptions are disabled, and we are emitting
19624 64-bit code, then we can use both FP pipes. Otherwise, we can only use one
19627 @item -mvr4130-align
19628 @itemx -mno-vr4130-align
19629 @opindex mvr4130-align
19630 The VR4130 pipeline is two-way superscalar, but can only issue two
19631 instructions together if the first one is 8-byte aligned. When this
19632 option is enabled, GCC aligns pairs of instructions that it
19633 thinks should execute in parallel.
19635 This option only has an effect when optimizing for the VR4130.
19636 It normally makes code faster, but at the expense of making it bigger.
19637 It is enabled by default at optimization level @option{-O3}.
19642 Enable (disable) generation of @code{synci} instructions on
19643 architectures that support it. The @code{synci} instructions (if
19644 enabled) are generated when @code{__builtin___clear_cache} is
19647 This option defaults to @option{-mno-synci}, but the default can be
19648 overridden by configuring GCC with @option{--with-synci}.
19650 When compiling code for single processor systems, it is generally safe
19651 to use @code{synci}. However, on many multi-core (SMP) systems, it
19652 does not invalidate the instruction caches on all cores and may lead
19653 to undefined behavior.
19655 @item -mrelax-pic-calls
19656 @itemx -mno-relax-pic-calls
19657 @opindex mrelax-pic-calls
19658 Try to turn PIC calls that are normally dispatched via register
19659 @code{$25} into direct calls. This is only possible if the linker can
19660 resolve the destination at link time and if the destination is within
19661 range for a direct call.
19663 @option{-mrelax-pic-calls} is the default if GCC was configured to use
19664 an assembler and a linker that support the @code{.reloc} assembly
19665 directive and @option{-mexplicit-relocs} is in effect. With
19666 @option{-mno-explicit-relocs}, this optimization can be performed by the
19667 assembler and the linker alone without help from the compiler.
19669 @item -mmcount-ra-address
19670 @itemx -mno-mcount-ra-address
19671 @opindex mmcount-ra-address
19672 @opindex mno-mcount-ra-address
19673 Emit (do not emit) code that allows @code{_mcount} to modify the
19674 calling function's return address. When enabled, this option extends
19675 the usual @code{_mcount} interface with a new @var{ra-address}
19676 parameter, which has type @code{intptr_t *} and is passed in register
19677 @code{$12}. @code{_mcount} can then modify the return address by
19678 doing both of the following:
19681 Returning the new address in register @code{$31}.
19683 Storing the new address in @code{*@var{ra-address}},
19684 if @var{ra-address} is nonnull.
19687 The default is @option{-mno-mcount-ra-address}.
19689 @item -mframe-header-opt
19690 @itemx -mno-frame-header-opt
19691 @opindex mframe-header-opt
19692 Enable (disable) frame header optimization in the o32 ABI. When using the
19693 o32 ABI, calling functions will allocate 16 bytes on the stack for the called
19694 function to write out register arguments. When enabled, this optimization
19695 will suppress the allocation of the frame header if it can be determined that
19698 This optimization is off by default at all optimization levels.
19703 @subsection MMIX Options
19704 @cindex MMIX Options
19706 These options are defined for the MMIX:
19710 @itemx -mno-libfuncs
19712 @opindex mno-libfuncs
19713 Specify that intrinsic library functions are being compiled, passing all
19714 values in registers, no matter the size.
19717 @itemx -mno-epsilon
19719 @opindex mno-epsilon
19720 Generate floating-point comparison instructions that compare with respect
19721 to the @code{rE} epsilon register.
19723 @item -mabi=mmixware
19725 @opindex mabi=mmixware
19727 Generate code that passes function parameters and return values that (in
19728 the called function) are seen as registers @code{$0} and up, as opposed to
19729 the GNU ABI which uses global registers @code{$231} and up.
19731 @item -mzero-extend
19732 @itemx -mno-zero-extend
19733 @opindex mzero-extend
19734 @opindex mno-zero-extend
19735 When reading data from memory in sizes shorter than 64 bits, use (do not
19736 use) zero-extending load instructions by default, rather than
19737 sign-extending ones.
19740 @itemx -mno-knuthdiv
19742 @opindex mno-knuthdiv
19743 Make the result of a division yielding a remainder have the same sign as
19744 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
19745 remainder follows the sign of the dividend. Both methods are
19746 arithmetically valid, the latter being almost exclusively used.
19748 @item -mtoplevel-symbols
19749 @itemx -mno-toplevel-symbols
19750 @opindex mtoplevel-symbols
19751 @opindex mno-toplevel-symbols
19752 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
19753 code can be used with the @code{PREFIX} assembly directive.
19757 Generate an executable in the ELF format, rather than the default
19758 @samp{mmo} format used by the @command{mmix} simulator.
19760 @item -mbranch-predict
19761 @itemx -mno-branch-predict
19762 @opindex mbranch-predict
19763 @opindex mno-branch-predict
19764 Use (do not use) the probable-branch instructions, when static branch
19765 prediction indicates a probable branch.
19767 @item -mbase-addresses
19768 @itemx -mno-base-addresses
19769 @opindex mbase-addresses
19770 @opindex mno-base-addresses
19771 Generate (do not generate) code that uses @emph{base addresses}. Using a
19772 base address automatically generates a request (handled by the assembler
19773 and the linker) for a constant to be set up in a global register. The
19774 register is used for one or more base address requests within the range 0
19775 to 255 from the value held in the register. The generally leads to short
19776 and fast code, but the number of different data items that can be
19777 addressed is limited. This means that a program that uses lots of static
19778 data may require @option{-mno-base-addresses}.
19780 @item -msingle-exit
19781 @itemx -mno-single-exit
19782 @opindex msingle-exit
19783 @opindex mno-single-exit
19784 Force (do not force) generated code to have a single exit point in each
19788 @node MN10300 Options
19789 @subsection MN10300 Options
19790 @cindex MN10300 options
19792 These @option{-m} options are defined for Matsushita MN10300 architectures:
19797 Generate code to avoid bugs in the multiply instructions for the MN10300
19798 processors. This is the default.
19800 @item -mno-mult-bug
19801 @opindex mno-mult-bug
19802 Do not generate code to avoid bugs in the multiply instructions for the
19803 MN10300 processors.
19807 Generate code using features specific to the AM33 processor.
19811 Do not generate code using features specific to the AM33 processor. This
19816 Generate code using features specific to the AM33/2.0 processor.
19820 Generate code using features specific to the AM34 processor.
19822 @item -mtune=@var{cpu-type}
19824 Use the timing characteristics of the indicated CPU type when
19825 scheduling instructions. This does not change the targeted processor
19826 type. The CPU type must be one of @samp{mn10300}, @samp{am33},
19827 @samp{am33-2} or @samp{am34}.
19829 @item -mreturn-pointer-on-d0
19830 @opindex mreturn-pointer-on-d0
19831 When generating a function that returns a pointer, return the pointer
19832 in both @code{a0} and @code{d0}. Otherwise, the pointer is returned
19833 only in @code{a0}, and attempts to call such functions without a prototype
19834 result in errors. Note that this option is on by default; use
19835 @option{-mno-return-pointer-on-d0} to disable it.
19839 Do not link in the C run-time initialization object file.
19843 Indicate to the linker that it should perform a relaxation optimization pass
19844 to shorten branches, calls and absolute memory addresses. This option only
19845 has an effect when used on the command line for the final link step.
19847 This option makes symbolic debugging impossible.
19851 Allow the compiler to generate @emph{Long Instruction Word}
19852 instructions if the target is the @samp{AM33} or later. This is the
19853 default. This option defines the preprocessor macro @code{__LIW__}.
19857 Do not allow the compiler to generate @emph{Long Instruction Word}
19858 instructions. This option defines the preprocessor macro
19863 Allow the compiler to generate the @emph{SETLB} and @emph{Lcc}
19864 instructions if the target is the @samp{AM33} or later. This is the
19865 default. This option defines the preprocessor macro @code{__SETLB__}.
19869 Do not allow the compiler to generate @emph{SETLB} or @emph{Lcc}
19870 instructions. This option defines the preprocessor macro
19871 @code{__NO_SETLB__}.
19875 @node Moxie Options
19876 @subsection Moxie Options
19877 @cindex Moxie Options
19883 Generate big-endian code. This is the default for @samp{moxie-*-*}
19888 Generate little-endian code.
19892 Generate mul.x and umul.x instructions. This is the default for
19893 @samp{moxiebox-*-*} configurations.
19897 Do not link in the C run-time initialization object file.
19901 @node MSP430 Options
19902 @subsection MSP430 Options
19903 @cindex MSP430 Options
19905 These options are defined for the MSP430:
19911 Force assembly output to always use hex constants. Normally such
19912 constants are signed decimals, but this option is available for
19913 testsuite and/or aesthetic purposes.
19917 Select the MCU to target. This is used to create a C preprocessor
19918 symbol based upon the MCU name, converted to upper case and pre- and
19919 post-fixed with @samp{__}. This in turn is used by the
19920 @file{msp430.h} header file to select an MCU-specific supplementary
19923 The option also sets the ISA to use. If the MCU name is one that is
19924 known to only support the 430 ISA then that is selected, otherwise the
19925 430X ISA is selected. A generic MCU name of @samp{msp430} can also be
19926 used to select the 430 ISA. Similarly the generic @samp{msp430x} MCU
19927 name selects the 430X ISA.
19929 In addition an MCU-specific linker script is added to the linker
19930 command line. The script's name is the name of the MCU with
19931 @file{.ld} appended. Thus specifying @option{-mmcu=xxx} on the @command{gcc}
19932 command line defines the C preprocessor symbol @code{__XXX__} and
19933 cause the linker to search for a script called @file{xxx.ld}.
19935 This option is also passed on to the assembler.
19938 @itemx -mno-warn-mcu
19940 @opindex mno-warn-mcu
19941 This option enables or disables warnings about conflicts between the
19942 MCU name specified by the @option{-mmcu} option and the ISA set by the
19943 @option{-mcpu} option and/or the hardware multiply support set by the
19944 @option{-mhwmult} option. It also toggles warnings about unrecognized
19945 MCU names. This option is on by default.
19949 Specifies the ISA to use. Accepted values are @samp{msp430},
19950 @samp{msp430x} and @samp{msp430xv2}. This option is deprecated. The
19951 @option{-mmcu=} option should be used to select the ISA.
19955 Link to the simulator runtime libraries and linker script. Overrides
19956 any scripts that would be selected by the @option{-mmcu=} option.
19960 Use large-model addressing (20-bit pointers, 32-bit @code{size_t}).
19964 Use small-model addressing (16-bit pointers, 16-bit @code{size_t}).
19968 This option is passed to the assembler and linker, and allows the
19969 linker to perform certain optimizations that cannot be done until
19974 Describes the type of hardware multiply supported by the target.
19975 Accepted values are @samp{none} for no hardware multiply, @samp{16bit}
19976 for the original 16-bit-only multiply supported by early MCUs.
19977 @samp{32bit} for the 16/32-bit multiply supported by later MCUs and
19978 @samp{f5series} for the 16/32-bit multiply supported by F5-series MCUs.
19979 A value of @samp{auto} can also be given. This tells GCC to deduce
19980 the hardware multiply support based upon the MCU name provided by the
19981 @option{-mmcu} option. If no @option{-mmcu} option is specified or if
19982 the MCU name is not recognized then no hardware multiply support is
19983 assumed. @code{auto} is the default setting.
19985 Hardware multiplies are normally performed by calling a library
19986 routine. This saves space in the generated code. When compiling at
19987 @option{-O3} or higher however the hardware multiplier is invoked
19988 inline. This makes for bigger, but faster code.
19990 The hardware multiply routines disable interrupts whilst running and
19991 restore the previous interrupt state when they finish. This makes
19992 them safe to use inside interrupt handlers as well as in normal code.
19996 Enable the use of a minimum runtime environment - no static
19997 initializers or constructors. This is intended for memory-constrained
19998 devices. The compiler includes special symbols in some objects
19999 that tell the linker and runtime which code fragments are required.
20001 @item -mcode-region=
20002 @itemx -mdata-region=
20003 @opindex mcode-region
20004 @opindex mdata-region
20005 These options tell the compiler where to place functions and data that
20006 do not have one of the @code{lower}, @code{upper}, @code{either} or
20007 @code{section} attributes. Possible values are @code{lower},
20008 @code{upper}, @code{either} or @code{any}. The first three behave
20009 like the corresponding attribute. The fourth possible value -
20010 @code{any} - is the default. It leaves placement entirely up to the
20011 linker script and how it assigns the standard sections
20012 (@code{.text}, @code{.data}, etc) to the memory regions.
20014 @item -msilicon-errata=
20015 @opindex msilicon-errata
20016 This option passes on a request to assembler to enable the fixes for
20017 the named silicon errata.
20019 @item -msilicon-errata-warn=
20020 @opindex msilicon-errata-warn
20021 This option passes on a request to the assembler to enable warning
20022 messages when a silicon errata might need to be applied.
20026 @node NDS32 Options
20027 @subsection NDS32 Options
20028 @cindex NDS32 Options
20030 These options are defined for NDS32 implementations:
20035 @opindex mbig-endian
20036 Generate code in big-endian mode.
20038 @item -mlittle-endian
20039 @opindex mlittle-endian
20040 Generate code in little-endian mode.
20042 @item -mreduced-regs
20043 @opindex mreduced-regs
20044 Use reduced-set registers for register allocation.
20047 @opindex mfull-regs
20048 Use full-set registers for register allocation.
20052 Generate conditional move instructions.
20056 Do not generate conditional move instructions.
20060 Generate performance extension instructions.
20062 @item -mno-perf-ext
20063 @opindex mno-perf-ext
20064 Do not generate performance extension instructions.
20068 Generate v3 push25/pop25 instructions.
20071 @opindex mno-v3push
20072 Do not generate v3 push25/pop25 instructions.
20076 Generate 16-bit instructions.
20079 @opindex mno-16-bit
20080 Do not generate 16-bit instructions.
20082 @item -misr-vector-size=@var{num}
20083 @opindex misr-vector-size
20084 Specify the size of each interrupt vector, which must be 4 or 16.
20086 @item -mcache-block-size=@var{num}
20087 @opindex mcache-block-size
20088 Specify the size of each cache block,
20089 which must be a power of 2 between 4 and 512.
20091 @item -march=@var{arch}
20093 Specify the name of the target architecture.
20095 @item -mcmodel=@var{code-model}
20097 Set the code model to one of
20100 All the data and read-only data segments must be within 512KB addressing space.
20101 The text segment must be within 16MB addressing space.
20102 @item @samp{medium}
20103 The data segment must be within 512KB while the read-only data segment can be
20104 within 4GB addressing space. The text segment should be still within 16MB
20107 All the text and data segments can be within 4GB addressing space.
20111 @opindex mctor-dtor
20112 Enable constructor/destructor feature.
20116 Guide linker to relax instructions.
20120 @node Nios II Options
20121 @subsection Nios II Options
20122 @cindex Nios II options
20123 @cindex Altera Nios II options
20125 These are the options defined for the Altera Nios II processor.
20131 @cindex smaller data references
20132 Put global and static objects less than or equal to @var{num} bytes
20133 into the small data or BSS sections instead of the normal data or BSS
20134 sections. The default value of @var{num} is 8.
20136 @item -mgpopt=@var{option}
20141 Generate (do not generate) GP-relative accesses. The following
20142 @var{option} names are recognized:
20147 Do not generate GP-relative accesses.
20150 Generate GP-relative accesses for small data objects that are not
20151 external, weak, or uninitialized common symbols.
20152 Also use GP-relative addressing for objects that
20153 have been explicitly placed in a small data section via a @code{section}
20157 As for @samp{local}, but also generate GP-relative accesses for
20158 small data objects that are external, weak, or common. If you use this option,
20159 you must ensure that all parts of your program (including libraries) are
20160 compiled with the same @option{-G} setting.
20163 Generate GP-relative accesses for all data objects in the program. If you
20164 use this option, the entire data and BSS segments
20165 of your program must fit in 64K of memory and you must use an appropriate
20166 linker script to allocate them within the addressable range of the
20170 Generate GP-relative addresses for function pointers as well as data
20171 pointers. If you use this option, the entire text, data, and BSS segments
20172 of your program must fit in 64K of memory and you must use an appropriate
20173 linker script to allocate them within the addressable range of the
20178 @option{-mgpopt} is equivalent to @option{-mgpopt=local}, and
20179 @option{-mno-gpopt} is equivalent to @option{-mgpopt=none}.
20181 The default is @option{-mgpopt} except when @option{-fpic} or
20182 @option{-fPIC} is specified to generate position-independent code.
20183 Note that the Nios II ABI does not permit GP-relative accesses from
20186 You may need to specify @option{-mno-gpopt} explicitly when building
20187 programs that include large amounts of small data, including large
20188 GOT data sections. In this case, the 16-bit offset for GP-relative
20189 addressing may not be large enough to allow access to the entire
20190 small data section.
20196 Generate little-endian (default) or big-endian (experimental) code,
20199 @item -march=@var{arch}
20201 This specifies the name of the target Nios II architecture. GCC uses this
20202 name to determine what kind of instructions it can emit when generating
20203 assembly code. Permissible names are: @samp{r1}, @samp{r2}.
20205 The preprocessor macro @code{__nios2_arch__} is available to programs,
20206 with value 1 or 2, indicating the targeted ISA level.
20208 @item -mbypass-cache
20209 @itemx -mno-bypass-cache
20210 @opindex mno-bypass-cache
20211 @opindex mbypass-cache
20212 Force all load and store instructions to always bypass cache by
20213 using I/O variants of the instructions. The default is not to
20216 @item -mno-cache-volatile
20217 @itemx -mcache-volatile
20218 @opindex mcache-volatile
20219 @opindex mno-cache-volatile
20220 Volatile memory access bypass the cache using the I/O variants of
20221 the load and store instructions. The default is not to bypass the cache.
20223 @item -mno-fast-sw-div
20224 @itemx -mfast-sw-div
20225 @opindex mno-fast-sw-div
20226 @opindex mfast-sw-div
20227 Do not use table-based fast divide for small numbers. The default
20228 is to use the fast divide at @option{-O3} and above.
20232 @itemx -mno-hw-mulx
20236 @opindex mno-hw-mul
20238 @opindex mno-hw-mulx
20240 @opindex mno-hw-div
20242 Enable or disable emitting @code{mul}, @code{mulx} and @code{div} family of
20243 instructions by the compiler. The default is to emit @code{mul}
20244 and not emit @code{div} and @code{mulx}.
20250 Enable or disable generation of Nios II R2 BMX (bit manipulation) and
20251 CDX (code density) instructions. Enabling these instructions also
20252 requires @option{-march=r2}. Since these instructions are optional
20253 extensions to the R2 architecture, the default is not to emit them.
20255 @item -mcustom-@var{insn}=@var{N}
20256 @itemx -mno-custom-@var{insn}
20257 @opindex mcustom-@var{insn}
20258 @opindex mno-custom-@var{insn}
20259 Each @option{-mcustom-@var{insn}=@var{N}} option enables use of a
20260 custom instruction with encoding @var{N} when generating code that uses
20261 @var{insn}. For example, @option{-mcustom-fadds=253} generates custom
20262 instruction 253 for single-precision floating-point add operations instead
20263 of the default behavior of using a library call.
20265 The following values of @var{insn} are supported. Except as otherwise
20266 noted, floating-point operations are expected to be implemented with
20267 normal IEEE 754 semantics and correspond directly to the C operators or the
20268 equivalent GCC built-in functions (@pxref{Other Builtins}).
20270 Single-precision floating point:
20273 @item @samp{fadds}, @samp{fsubs}, @samp{fdivs}, @samp{fmuls}
20274 Binary arithmetic operations.
20280 Unary absolute value.
20282 @item @samp{fcmpeqs}, @samp{fcmpges}, @samp{fcmpgts}, @samp{fcmples}, @samp{fcmplts}, @samp{fcmpnes}
20283 Comparison operations.
20285 @item @samp{fmins}, @samp{fmaxs}
20286 Floating-point minimum and maximum. These instructions are only
20287 generated if @option{-ffinite-math-only} is specified.
20289 @item @samp{fsqrts}
20290 Unary square root operation.
20292 @item @samp{fcoss}, @samp{fsins}, @samp{ftans}, @samp{fatans}, @samp{fexps}, @samp{flogs}
20293 Floating-point trigonometric and exponential functions. These instructions
20294 are only generated if @option{-funsafe-math-optimizations} is also specified.
20298 Double-precision floating point:
20301 @item @samp{faddd}, @samp{fsubd}, @samp{fdivd}, @samp{fmuld}
20302 Binary arithmetic operations.
20308 Unary absolute value.
20310 @item @samp{fcmpeqd}, @samp{fcmpged}, @samp{fcmpgtd}, @samp{fcmpled}, @samp{fcmpltd}, @samp{fcmpned}
20311 Comparison operations.
20313 @item @samp{fmind}, @samp{fmaxd}
20314 Double-precision minimum and maximum. These instructions are only
20315 generated if @option{-ffinite-math-only} is specified.
20317 @item @samp{fsqrtd}
20318 Unary square root operation.
20320 @item @samp{fcosd}, @samp{fsind}, @samp{ftand}, @samp{fatand}, @samp{fexpd}, @samp{flogd}
20321 Double-precision trigonometric and exponential functions. These instructions
20322 are only generated if @option{-funsafe-math-optimizations} is also specified.
20328 @item @samp{fextsd}
20329 Conversion from single precision to double precision.
20331 @item @samp{ftruncds}
20332 Conversion from double precision to single precision.
20334 @item @samp{fixsi}, @samp{fixsu}, @samp{fixdi}, @samp{fixdu}
20335 Conversion from floating point to signed or unsigned integer types, with
20336 truncation towards zero.
20339 Conversion from single-precision floating point to signed integer,
20340 rounding to the nearest integer and ties away from zero.
20341 This corresponds to the @code{__builtin_lroundf} function when
20342 @option{-fno-math-errno} is used.
20344 @item @samp{floatis}, @samp{floatus}, @samp{floatid}, @samp{floatud}
20345 Conversion from signed or unsigned integer types to floating-point types.
20349 In addition, all of the following transfer instructions for internal
20350 registers X and Y must be provided to use any of the double-precision
20351 floating-point instructions. Custom instructions taking two
20352 double-precision source operands expect the first operand in the
20353 64-bit register X. The other operand (or only operand of a unary
20354 operation) is given to the custom arithmetic instruction with the
20355 least significant half in source register @var{src1} and the most
20356 significant half in @var{src2}. A custom instruction that returns a
20357 double-precision result returns the most significant 32 bits in the
20358 destination register and the other half in 32-bit register Y.
20359 GCC automatically generates the necessary code sequences to write
20360 register X and/or read register Y when double-precision floating-point
20361 instructions are used.
20366 Write @var{src1} into the least significant half of X and @var{src2} into
20367 the most significant half of X.
20370 Write @var{src1} into Y.
20372 @item @samp{frdxhi}, @samp{frdxlo}
20373 Read the most or least (respectively) significant half of X and store it in
20377 Read the value of Y and store it into @var{dest}.
20380 Note that you can gain more local control over generation of Nios II custom
20381 instructions by using the @code{target("custom-@var{insn}=@var{N}")}
20382 and @code{target("no-custom-@var{insn}")} function attributes
20383 (@pxref{Function Attributes})
20384 or pragmas (@pxref{Function Specific Option Pragmas}).
20386 @item -mcustom-fpu-cfg=@var{name}
20387 @opindex mcustom-fpu-cfg
20389 This option enables a predefined, named set of custom instruction encodings
20390 (see @option{-mcustom-@var{insn}} above).
20391 Currently, the following sets are defined:
20393 @option{-mcustom-fpu-cfg=60-1} is equivalent to:
20394 @gccoptlist{-mcustom-fmuls=252 @gol
20395 -mcustom-fadds=253 @gol
20396 -mcustom-fsubs=254 @gol
20397 -fsingle-precision-constant}
20399 @option{-mcustom-fpu-cfg=60-2} is equivalent to:
20400 @gccoptlist{-mcustom-fmuls=252 @gol
20401 -mcustom-fadds=253 @gol
20402 -mcustom-fsubs=254 @gol
20403 -mcustom-fdivs=255 @gol
20404 -fsingle-precision-constant}
20406 @option{-mcustom-fpu-cfg=72-3} is equivalent to:
20407 @gccoptlist{-mcustom-floatus=243 @gol
20408 -mcustom-fixsi=244 @gol
20409 -mcustom-floatis=245 @gol
20410 -mcustom-fcmpgts=246 @gol
20411 -mcustom-fcmples=249 @gol
20412 -mcustom-fcmpeqs=250 @gol
20413 -mcustom-fcmpnes=251 @gol
20414 -mcustom-fmuls=252 @gol
20415 -mcustom-fadds=253 @gol
20416 -mcustom-fsubs=254 @gol
20417 -mcustom-fdivs=255 @gol
20418 -fsingle-precision-constant}
20420 Custom instruction assignments given by individual
20421 @option{-mcustom-@var{insn}=} options override those given by
20422 @option{-mcustom-fpu-cfg=}, regardless of the
20423 order of the options on the command line.
20425 Note that you can gain more local control over selection of a FPU
20426 configuration by using the @code{target("custom-fpu-cfg=@var{name}")}
20427 function attribute (@pxref{Function Attributes})
20428 or pragma (@pxref{Function Specific Option Pragmas}).
20432 These additional @samp{-m} options are available for the Altera Nios II
20433 ELF (bare-metal) target:
20439 Link with HAL BSP. This suppresses linking with the GCC-provided C runtime
20440 startup and termination code, and is typically used in conjunction with
20441 @option{-msys-crt0=} to specify the location of the alternate startup code
20442 provided by the HAL BSP.
20446 Link with a limited version of the C library, @option{-lsmallc}, rather than
20449 @item -msys-crt0=@var{startfile}
20451 @var{startfile} is the file name of the startfile (crt0) to use
20452 when linking. This option is only useful in conjunction with @option{-mhal}.
20454 @item -msys-lib=@var{systemlib}
20456 @var{systemlib} is the library name of the library that provides
20457 low-level system calls required by the C library,
20458 e.g. @code{read} and @code{write}.
20459 This option is typically used to link with a library provided by a HAL BSP.
20463 @node Nvidia PTX Options
20464 @subsection Nvidia PTX Options
20465 @cindex Nvidia PTX options
20466 @cindex nvptx options
20468 These options are defined for Nvidia PTX:
20476 Generate code for 32-bit or 64-bit ABI.
20479 @opindex mmainkernel
20480 Link in code for a __main kernel. This is for stand-alone instead of
20481 offloading execution.
20485 Apply partitioned execution optimizations. This is the default when any
20486 level of optimization is selected.
20490 @node PDP-11 Options
20491 @subsection PDP-11 Options
20492 @cindex PDP-11 Options
20494 These options are defined for the PDP-11:
20499 Use hardware FPP floating point. This is the default. (FIS floating
20500 point on the PDP-11/40 is not supported.)
20503 @opindex msoft-float
20504 Do not use hardware floating point.
20508 Return floating-point results in ac0 (fr0 in Unix assembler syntax).
20512 Return floating-point results in memory. This is the default.
20516 Generate code for a PDP-11/40.
20520 Generate code for a PDP-11/45. This is the default.
20524 Generate code for a PDP-11/10.
20526 @item -mbcopy-builtin
20527 @opindex mbcopy-builtin
20528 Use inline @code{movmemhi} patterns for copying memory. This is the
20533 Do not use inline @code{movmemhi} patterns for copying memory.
20539 Use 16-bit @code{int}. This is the default.
20545 Use 32-bit @code{int}.
20548 @itemx -mno-float32
20550 @opindex mno-float32
20551 Use 64-bit @code{float}. This is the default.
20554 @itemx -mno-float64
20556 @opindex mno-float64
20557 Use 32-bit @code{float}.
20561 Use @code{abshi2} pattern. This is the default.
20565 Do not use @code{abshi2} pattern.
20567 @item -mbranch-expensive
20568 @opindex mbranch-expensive
20569 Pretend that branches are expensive. This is for experimenting with
20570 code generation only.
20572 @item -mbranch-cheap
20573 @opindex mbranch-cheap
20574 Do not pretend that branches are expensive. This is the default.
20578 Use Unix assembler syntax. This is the default when configured for
20579 @samp{pdp11-*-bsd}.
20583 Use DEC assembler syntax. This is the default when configured for any
20584 PDP-11 target other than @samp{pdp11-*-bsd}.
20587 @node picoChip Options
20588 @subsection picoChip Options
20589 @cindex picoChip options
20591 These @samp{-m} options are defined for picoChip implementations:
20595 @item -mae=@var{ae_type}
20597 Set the instruction set, register set, and instruction scheduling
20598 parameters for array element type @var{ae_type}. Supported values
20599 for @var{ae_type} are @samp{ANY}, @samp{MUL}, and @samp{MAC}.
20601 @option{-mae=ANY} selects a completely generic AE type. Code
20602 generated with this option runs on any of the other AE types. The
20603 code is not as efficient as it would be if compiled for a specific
20604 AE type, and some types of operation (e.g., multiplication) do not
20605 work properly on all types of AE.
20607 @option{-mae=MUL} selects a MUL AE type. This is the most useful AE type
20608 for compiled code, and is the default.
20610 @option{-mae=MAC} selects a DSP-style MAC AE. Code compiled with this
20611 option may suffer from poor performance of byte (char) manipulation,
20612 since the DSP AE does not provide hardware support for byte load/stores.
20614 @item -msymbol-as-address
20615 Enable the compiler to directly use a symbol name as an address in a
20616 load/store instruction, without first loading it into a
20617 register. Typically, the use of this option generates larger
20618 programs, which run faster than when the option isn't used. However, the
20619 results vary from program to program, so it is left as a user option,
20620 rather than being permanently enabled.
20622 @item -mno-inefficient-warnings
20623 Disables warnings about the generation of inefficient code. These
20624 warnings can be generated, for example, when compiling code that
20625 performs byte-level memory operations on the MAC AE type. The MAC AE has
20626 no hardware support for byte-level memory operations, so all byte
20627 load/stores must be synthesized from word load/store operations. This is
20628 inefficient and a warning is generated to indicate
20629 that you should rewrite the code to avoid byte operations, or to target
20630 an AE type that has the necessary hardware support. This option disables
20635 @node PowerPC Options
20636 @subsection PowerPC Options
20637 @cindex PowerPC options
20639 These are listed under @xref{RS/6000 and PowerPC Options}.
20642 @subsection RL78 Options
20643 @cindex RL78 Options
20649 Links in additional target libraries to support operation within a
20658 Specifies the type of hardware multiplication and division support to
20659 be used. The simplest is @code{none}, which uses software for both
20660 multiplication and division. This is the default. The @code{g13}
20661 value is for the hardware multiply/divide peripheral found on the
20662 RL78/G13 (S2 core) targets. The @code{g14} value selects the use of
20663 the multiplication and division instructions supported by the RL78/G14
20664 (S3 core) parts. The value @code{rl78} is an alias for @code{g14} and
20665 the value @code{mg10} is an alias for @code{none}.
20667 In addition a C preprocessor macro is defined, based upon the setting
20668 of this option. Possible values are: @code{__RL78_MUL_NONE__},
20669 @code{__RL78_MUL_G13__} or @code{__RL78_MUL_G14__}.
20676 Specifies the RL78 core to target. The default is the G14 core, also
20677 known as an S3 core or just RL78. The G13 or S2 core does not have
20678 multiply or divide instructions, instead it uses a hardware peripheral
20679 for these operations. The G10 or S1 core does not have register
20680 banks, so it uses a different calling convention.
20682 If this option is set it also selects the type of hardware multiply
20683 support to use, unless this is overridden by an explicit
20684 @option{-mmul=none} option on the command line. Thus specifying
20685 @option{-mcpu=g13} enables the use of the G13 hardware multiply
20686 peripheral and specifying @option{-mcpu=g10} disables the use of
20687 hardware multiplications altogether.
20689 Note, although the RL78/G14 core is the default target, specifying
20690 @option{-mcpu=g14} or @option{-mcpu=rl78} on the command line does
20691 change the behavior of the toolchain since it also enables G14
20692 hardware multiply support. If these options are not specified on the
20693 command line then software multiplication routines will be used even
20694 though the code targets the RL78 core. This is for backwards
20695 compatibility with older toolchains which did not have hardware
20696 multiply and divide support.
20698 In addition a C preprocessor macro is defined, based upon the setting
20699 of this option. Possible values are: @code{__RL78_G10__},
20700 @code{__RL78_G13__} or @code{__RL78_G14__}.
20710 These are aliases for the corresponding @option{-mcpu=} option. They
20711 are provided for backwards compatibility.
20715 Allow the compiler to use all of the available registers. By default
20716 registers @code{r24..r31} are reserved for use in interrupt handlers.
20717 With this option enabled these registers can be used in ordinary
20720 @item -m64bit-doubles
20721 @itemx -m32bit-doubles
20722 @opindex m64bit-doubles
20723 @opindex m32bit-doubles
20724 Make the @code{double} data type be 64 bits (@option{-m64bit-doubles})
20725 or 32 bits (@option{-m32bit-doubles}) in size. The default is
20726 @option{-m32bit-doubles}.
20728 @item -msave-mduc-in-interrupts
20729 @item -mno-save-mduc-in-interrupts
20730 @opindex msave-mduc-in-interrupts
20731 @opindex mno-save-mduc-in-interrupts
20732 Specifies that interrupt handler functions should preserve the
20733 MDUC registers. This is only necessary if normal code might use
20734 the MDUC registers, for example because it performs multiplication
20735 and division operations. The default is to ignore the MDUC registers
20736 as this makes the interrupt handlers faster. The target option -mg13
20737 needs to be passed for this to work as this feature is only available
20738 on the G13 target (S2 core). The MDUC registers will only be saved
20739 if the interrupt handler performs a multiplication or division
20740 operation or it calls another function.
20744 @node RS/6000 and PowerPC Options
20745 @subsection IBM RS/6000 and PowerPC Options
20746 @cindex RS/6000 and PowerPC Options
20747 @cindex IBM RS/6000 and PowerPC Options
20749 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
20751 @item -mpowerpc-gpopt
20752 @itemx -mno-powerpc-gpopt
20753 @itemx -mpowerpc-gfxopt
20754 @itemx -mno-powerpc-gfxopt
20757 @itemx -mno-powerpc64
20761 @itemx -mno-popcntb
20763 @itemx -mno-popcntd
20772 @itemx -mno-hard-dfp
20773 @opindex mpowerpc-gpopt
20774 @opindex mno-powerpc-gpopt
20775 @opindex mpowerpc-gfxopt
20776 @opindex mno-powerpc-gfxopt
20777 @opindex mpowerpc64
20778 @opindex mno-powerpc64
20782 @opindex mno-popcntb
20784 @opindex mno-popcntd
20790 @opindex mno-mfpgpr
20792 @opindex mno-hard-dfp
20793 You use these options to specify which instructions are available on the
20794 processor you are using. The default value of these options is
20795 determined when configuring GCC@. Specifying the
20796 @option{-mcpu=@var{cpu_type}} overrides the specification of these
20797 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
20798 rather than the options listed above.
20800 Specifying @option{-mpowerpc-gpopt} allows
20801 GCC to use the optional PowerPC architecture instructions in the
20802 General Purpose group, including floating-point square root. Specifying
20803 @option{-mpowerpc-gfxopt} allows GCC to
20804 use the optional PowerPC architecture instructions in the Graphics
20805 group, including floating-point select.
20807 The @option{-mmfcrf} option allows GCC to generate the move from
20808 condition register field instruction implemented on the POWER4
20809 processor and other processors that support the PowerPC V2.01
20811 The @option{-mpopcntb} option allows GCC to generate the popcount and
20812 double-precision FP reciprocal estimate instruction implemented on the
20813 POWER5 processor and other processors that support the PowerPC V2.02
20815 The @option{-mpopcntd} option allows GCC to generate the popcount
20816 instruction implemented on the POWER7 processor and other processors
20817 that support the PowerPC V2.06 architecture.
20818 The @option{-mfprnd} option allows GCC to generate the FP round to
20819 integer instructions implemented on the POWER5+ processor and other
20820 processors that support the PowerPC V2.03 architecture.
20821 The @option{-mcmpb} option allows GCC to generate the compare bytes
20822 instruction implemented on the POWER6 processor and other processors
20823 that support the PowerPC V2.05 architecture.
20824 The @option{-mmfpgpr} option allows GCC to generate the FP move to/from
20825 general-purpose register instructions implemented on the POWER6X
20826 processor and other processors that support the extended PowerPC V2.05
20828 The @option{-mhard-dfp} option allows GCC to generate the decimal
20829 floating-point instructions implemented on some POWER processors.
20831 The @option{-mpowerpc64} option allows GCC to generate the additional
20832 64-bit instructions that are found in the full PowerPC64 architecture
20833 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
20834 @option{-mno-powerpc64}.
20836 @item -mcpu=@var{cpu_type}
20838 Set architecture type, register usage, and
20839 instruction scheduling parameters for machine type @var{cpu_type}.
20840 Supported values for @var{cpu_type} are @samp{401}, @samp{403},
20841 @samp{405}, @samp{405fp}, @samp{440}, @samp{440fp}, @samp{464}, @samp{464fp},
20842 @samp{476}, @samp{476fp}, @samp{505}, @samp{601}, @samp{602}, @samp{603},
20843 @samp{603e}, @samp{604}, @samp{604e}, @samp{620}, @samp{630}, @samp{740},
20844 @samp{7400}, @samp{7450}, @samp{750}, @samp{801}, @samp{821}, @samp{823},
20845 @samp{860}, @samp{970}, @samp{8540}, @samp{a2}, @samp{e300c2},
20846 @samp{e300c3}, @samp{e500mc}, @samp{e500mc64}, @samp{e5500},
20847 @samp{e6500}, @samp{ec603e}, @samp{G3}, @samp{G4}, @samp{G5},
20848 @samp{titan}, @samp{power3}, @samp{power4}, @samp{power5}, @samp{power5+},
20849 @samp{power6}, @samp{power6x}, @samp{power7}, @samp{power8},
20850 @samp{power9}, @samp{powerpc}, @samp{powerpc64}, @samp{powerpc64le},
20853 @option{-mcpu=powerpc}, @option{-mcpu=powerpc64}, and
20854 @option{-mcpu=powerpc64le} specify pure 32-bit PowerPC (either
20855 endian), 64-bit big endian PowerPC and 64-bit little endian PowerPC
20856 architecture machine types, with an appropriate, generic processor
20857 model assumed for scheduling purposes.
20859 The other options specify a specific processor. Code generated under
20860 those options runs best on that processor, and may not run at all on
20863 The @option{-mcpu} options automatically enable or disable the
20866 @gccoptlist{-maltivec -mfprnd -mhard-float -mmfcrf -mmultiple @gol
20867 -mpopcntb -mpopcntd -mpowerpc64 @gol
20868 -mpowerpc-gpopt -mpowerpc-gfxopt -msingle-float -mdouble-float @gol
20869 -msimple-fpu -mstring -mmulhw -mdlmzb -mmfpgpr -mvsx @gol
20870 -mcrypto -mdirect-move -mhtm -mpower8-fusion -mpower8-vector @gol
20871 -mquad-memory -mquad-memory-atomic -mfloat128 -mfloat128-hardware}
20873 The particular options set for any particular CPU varies between
20874 compiler versions, depending on what setting seems to produce optimal
20875 code for that CPU; it doesn't necessarily reflect the actual hardware's
20876 capabilities. If you wish to set an individual option to a particular
20877 value, you may specify it after the @option{-mcpu} option, like
20878 @option{-mcpu=970 -mno-altivec}.
20880 On AIX, the @option{-maltivec} and @option{-mpowerpc64} options are
20881 not enabled or disabled by the @option{-mcpu} option at present because
20882 AIX does not have full support for these options. You may still
20883 enable or disable them individually if you're sure it'll work in your
20886 @item -mtune=@var{cpu_type}
20888 Set the instruction scheduling parameters for machine type
20889 @var{cpu_type}, but do not set the architecture type or register usage,
20890 as @option{-mcpu=@var{cpu_type}} does. The same
20891 values for @var{cpu_type} are used for @option{-mtune} as for
20892 @option{-mcpu}. If both are specified, the code generated uses the
20893 architecture and registers set by @option{-mcpu}, but the
20894 scheduling parameters set by @option{-mtune}.
20896 @item -mcmodel=small
20897 @opindex mcmodel=small
20898 Generate PowerPC64 code for the small model: The TOC is limited to
20901 @item -mcmodel=medium
20902 @opindex mcmodel=medium
20903 Generate PowerPC64 code for the medium model: The TOC and other static
20904 data may be up to a total of 4G in size.
20906 @item -mcmodel=large
20907 @opindex mcmodel=large
20908 Generate PowerPC64 code for the large model: The TOC may be up to 4G
20909 in size. Other data and code is only limited by the 64-bit address
20913 @itemx -mno-altivec
20915 @opindex mno-altivec
20916 Generate code that uses (does not use) AltiVec instructions, and also
20917 enable the use of built-in functions that allow more direct access to
20918 the AltiVec instruction set. You may also need to set
20919 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
20922 When @option{-maltivec} is used, rather than @option{-maltivec=le} or
20923 @option{-maltivec=be}, the element order for AltiVec intrinsics such
20924 as @code{vec_splat}, @code{vec_extract}, and @code{vec_insert}
20925 match array element order corresponding to the endianness of the
20926 target. That is, element zero identifies the leftmost element in a
20927 vector register when targeting a big-endian platform, and identifies
20928 the rightmost element in a vector register when targeting a
20929 little-endian platform.
20932 @opindex maltivec=be
20933 Generate AltiVec instructions using big-endian element order,
20934 regardless of whether the target is big- or little-endian. This is
20935 the default when targeting a big-endian platform.
20937 The element order is used to interpret element numbers in AltiVec
20938 intrinsics such as @code{vec_splat}, @code{vec_extract}, and
20939 @code{vec_insert}. By default, these match array element order
20940 corresponding to the endianness for the target.
20943 @opindex maltivec=le
20944 Generate AltiVec instructions using little-endian element order,
20945 regardless of whether the target is big- or little-endian. This is
20946 the default when targeting a little-endian platform. This option is
20947 currently ignored when targeting a big-endian platform.
20949 The element order is used to interpret element numbers in AltiVec
20950 intrinsics such as @code{vec_splat}, @code{vec_extract}, and
20951 @code{vec_insert}. By default, these match array element order
20952 corresponding to the endianness for the target.
20957 @opindex mno-vrsave
20958 Generate VRSAVE instructions when generating AltiVec code.
20960 @item -mgen-cell-microcode
20961 @opindex mgen-cell-microcode
20962 Generate Cell microcode instructions.
20964 @item -mwarn-cell-microcode
20965 @opindex mwarn-cell-microcode
20966 Warn when a Cell microcode instruction is emitted. An example
20967 of a Cell microcode instruction is a variable shift.
20970 @opindex msecure-plt
20971 Generate code that allows @command{ld} and @command{ld.so}
20972 to build executables and shared
20973 libraries with non-executable @code{.plt} and @code{.got} sections.
20975 32-bit SYSV ABI option.
20979 Generate code that uses a BSS @code{.plt} section that @command{ld.so}
20981 requires @code{.plt} and @code{.got}
20982 sections that are both writable and executable.
20983 This is a PowerPC 32-bit SYSV ABI option.
20989 This switch enables or disables the generation of ISEL instructions.
20991 @item -misel=@var{yes/no}
20992 This switch has been deprecated. Use @option{-misel} and
20993 @option{-mno-isel} instead.
20997 Enable Local Register Allocation. This is still experimental for PowerPC,
20998 so by default the compiler uses standard reload
20999 (i.e. @option{-mno-lra}).
21005 This switch enables or disables the generation of SPE simd
21011 @opindex mno-paired
21012 This switch enables or disables the generation of PAIRED simd
21015 @item -mspe=@var{yes/no}
21016 This option has been deprecated. Use @option{-mspe} and
21017 @option{-mno-spe} instead.
21023 Generate code that uses (does not use) vector/scalar (VSX)
21024 instructions, and also enable the use of built-in functions that allow
21025 more direct access to the VSX instruction set.
21030 @opindex mno-crypto
21031 Enable the use (disable) of the built-in functions that allow direct
21032 access to the cryptographic instructions that were added in version
21033 2.07 of the PowerPC ISA.
21035 @item -mdirect-move
21036 @itemx -mno-direct-move
21037 @opindex mdirect-move
21038 @opindex mno-direct-move
21039 Generate code that uses (does not use) the instructions to move data
21040 between the general purpose registers and the vector/scalar (VSX)
21041 registers that were added in version 2.07 of the PowerPC ISA.
21047 Enable (disable) the use of the built-in functions that allow direct
21048 access to the Hardware Transactional Memory (HTM) instructions that
21049 were added in version 2.07 of the PowerPC ISA.
21051 @item -mpower8-fusion
21052 @itemx -mno-power8-fusion
21053 @opindex mpower8-fusion
21054 @opindex mno-power8-fusion
21055 Generate code that keeps (does not keeps) some integer operations
21056 adjacent so that the instructions can be fused together on power8 and
21059 @item -mpower8-vector
21060 @itemx -mno-power8-vector
21061 @opindex mpower8-vector
21062 @opindex mno-power8-vector
21063 Generate code that uses (does not use) the vector and scalar
21064 instructions that were added in version 2.07 of the PowerPC ISA. Also
21065 enable the use of built-in functions that allow more direct access to
21066 the vector instructions.
21068 @item -mquad-memory
21069 @itemx -mno-quad-memory
21070 @opindex mquad-memory
21071 @opindex mno-quad-memory
21072 Generate code that uses (does not use) the non-atomic quad word memory
21073 instructions. The @option{-mquad-memory} option requires use of
21076 @item -mquad-memory-atomic
21077 @itemx -mno-quad-memory-atomic
21078 @opindex mquad-memory-atomic
21079 @opindex mno-quad-memory-atomic
21080 Generate code that uses (does not use) the atomic quad word memory
21081 instructions. The @option{-mquad-memory-atomic} option requires use of
21084 @item -mupper-regs-di
21085 @itemx -mno-upper-regs-di
21086 @opindex mupper-regs-di
21087 @opindex mno-upper-regs-di
21088 Generate code that uses (does not use) the scalar instructions that
21089 target all 64 registers in the vector/scalar floating point register
21090 set that were added in version 2.06 of the PowerPC ISA when processing
21091 integers. @option{-mupper-regs-di} is turned on by default if you use
21092 any of the @option{-mcpu=power7}, @option{-mcpu=power8},
21093 @option{-mcpu=power9}, or @option{-mvsx} options.
21095 @item -mupper-regs-df
21096 @itemx -mno-upper-regs-df
21097 @opindex mupper-regs-df
21098 @opindex mno-upper-regs-df
21099 Generate code that uses (does not use) the scalar double precision
21100 instructions that target all 64 registers in the vector/scalar
21101 floating point register set that were added in version 2.06 of the
21102 PowerPC ISA. @option{-mupper-regs-df} is turned on by default if you
21103 use any of the @option{-mcpu=power7}, @option{-mcpu=power8},
21104 @option{-mcpu=power9}, or @option{-mvsx} options.
21106 @item -mupper-regs-sf
21107 @itemx -mno-upper-regs-sf
21108 @opindex mupper-regs-sf
21109 @opindex mno-upper-regs-sf
21110 Generate code that uses (does not use) the scalar single precision
21111 instructions that target all 64 registers in the vector/scalar
21112 floating point register set that were added in version 2.07 of the
21113 PowerPC ISA. @option{-mupper-regs-sf} is turned on by default if you
21114 use either of the @option{-mcpu=power8}, @option{-mpower8-vector}, or
21115 @option{-mcpu=power9} options.
21118 @itemx -mno-upper-regs
21119 @opindex mupper-regs
21120 @opindex mno-upper-regs
21121 Generate code that uses (does not use) the scalar
21122 instructions that target all 64 registers in the vector/scalar
21123 floating point register set, depending on the model of the machine.
21125 If the @option{-mno-upper-regs} option is used, it turns off both
21126 @option{-mupper-regs-sf} and @option{-mupper-regs-df} options.
21129 @itemx -mno-float128
21131 @opindex mno-float128
21132 Enable/disable the @var{__float128} keyword for IEEE 128-bit floating point
21133 and use either software emulation for IEEE 128-bit floating point or
21134 hardware instructions.
21136 The VSX instruction set (@option{-mvsx}, @option{-mcpu=power7}, or
21137 @option{-mcpu=power8}) must be enabled to use the @option{-mfloat128}
21138 option. The @option{-mfloat128} option only works on PowerPC 64-bit
21141 If you use the ISA 3.0 instruction set (@option{-mcpu=power9}), the
21142 @option{-mfloat128} option will also enable the generation of ISA 3.0
21143 IEEE 128-bit floating point instructions. Otherwise, IEEE 128-bit
21144 floating point will be done with software emulation.
21146 @item -mfloat128-hardware
21147 @itemx -mno-float128-hardware
21148 @opindex mfloat128-hardware
21149 @opindex mno-float128-hardware
21150 Enable/disable using ISA 3.0 hardware instructions to support the
21151 @var{__float128} data type.
21153 If you use @option{-mfloat128-hardware}, it will enable the option
21154 @option{-mfloat128} as well.
21156 If you select ISA 3.0 instructions with @option{-mcpu=power9}, but do
21157 not use either @option{-mfloat128} or @option{-mfloat128-hardware},
21158 the IEEE 128-bit floating point support will not be enabled.
21160 @item -mfloat-gprs=@var{yes/single/double/no}
21161 @itemx -mfloat-gprs
21162 @opindex mfloat-gprs
21163 This switch enables or disables the generation of floating-point
21164 operations on the general-purpose registers for architectures that
21167 The argument @samp{yes} or @samp{single} enables the use of
21168 single-precision floating-point operations.
21170 The argument @samp{double} enables the use of single and
21171 double-precision floating-point operations.
21173 The argument @samp{no} disables floating-point operations on the
21174 general-purpose registers.
21176 This option is currently only available on the MPC854x.
21182 Generate code for 32-bit or 64-bit environments of Darwin and SVR4
21183 targets (including GNU/Linux). The 32-bit environment sets int, long
21184 and pointer to 32 bits and generates code that runs on any PowerPC
21185 variant. The 64-bit environment sets int to 32 bits and long and
21186 pointer to 64 bits, and generates code for PowerPC64, as for
21187 @option{-mpowerpc64}.
21190 @itemx -mno-fp-in-toc
21191 @itemx -mno-sum-in-toc
21192 @itemx -mminimal-toc
21194 @opindex mno-fp-in-toc
21195 @opindex mno-sum-in-toc
21196 @opindex mminimal-toc
21197 Modify generation of the TOC (Table Of Contents), which is created for
21198 every executable file. The @option{-mfull-toc} option is selected by
21199 default. In that case, GCC allocates at least one TOC entry for
21200 each unique non-automatic variable reference in your program. GCC
21201 also places floating-point constants in the TOC@. However, only
21202 16,384 entries are available in the TOC@.
21204 If you receive a linker error message that saying you have overflowed
21205 the available TOC space, you can reduce the amount of TOC space used
21206 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
21207 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
21208 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
21209 generate code to calculate the sum of an address and a constant at
21210 run time instead of putting that sum into the TOC@. You may specify one
21211 or both of these options. Each causes GCC to produce very slightly
21212 slower and larger code at the expense of conserving TOC space.
21214 If you still run out of space in the TOC even when you specify both of
21215 these options, specify @option{-mminimal-toc} instead. This option causes
21216 GCC to make only one TOC entry for every file. When you specify this
21217 option, GCC produces code that is slower and larger but which
21218 uses extremely little TOC space. You may wish to use this option
21219 only on files that contain less frequently-executed code.
21225 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
21226 @code{long} type, and the infrastructure needed to support them.
21227 Specifying @option{-maix64} implies @option{-mpowerpc64},
21228 while @option{-maix32} disables the 64-bit ABI and
21229 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
21232 @itemx -mno-xl-compat
21233 @opindex mxl-compat
21234 @opindex mno-xl-compat
21235 Produce code that conforms more closely to IBM XL compiler semantics
21236 when using AIX-compatible ABI@. Pass floating-point arguments to
21237 prototyped functions beyond the register save area (RSA) on the stack
21238 in addition to argument FPRs. Do not assume that most significant
21239 double in 128-bit long double value is properly rounded when comparing
21240 values and converting to double. Use XL symbol names for long double
21243 The AIX calling convention was extended but not initially documented to
21244 handle an obscure K&R C case of calling a function that takes the
21245 address of its arguments with fewer arguments than declared. IBM XL
21246 compilers access floating-point arguments that do not fit in the
21247 RSA from the stack when a subroutine is compiled without
21248 optimization. Because always storing floating-point arguments on the
21249 stack is inefficient and rarely needed, this option is not enabled by
21250 default and only is necessary when calling subroutines compiled by IBM
21251 XL compilers without optimization.
21255 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
21256 application written to use message passing with special startup code to
21257 enable the application to run. The system must have PE installed in the
21258 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
21259 must be overridden with the @option{-specs=} option to specify the
21260 appropriate directory location. The Parallel Environment does not
21261 support threads, so the @option{-mpe} option and the @option{-pthread}
21262 option are incompatible.
21264 @item -malign-natural
21265 @itemx -malign-power
21266 @opindex malign-natural
21267 @opindex malign-power
21268 On AIX, 32-bit Darwin, and 64-bit PowerPC GNU/Linux, the option
21269 @option{-malign-natural} overrides the ABI-defined alignment of larger
21270 types, such as floating-point doubles, on their natural size-based boundary.
21271 The option @option{-malign-power} instructs GCC to follow the ABI-specified
21272 alignment rules. GCC defaults to the standard alignment defined in the ABI@.
21274 On 64-bit Darwin, natural alignment is the default, and @option{-malign-power}
21278 @itemx -mhard-float
21279 @opindex msoft-float
21280 @opindex mhard-float
21281 Generate code that does not use (uses) the floating-point register set.
21282 Software floating-point emulation is provided if you use the
21283 @option{-msoft-float} option, and pass the option to GCC when linking.
21285 @item -msingle-float
21286 @itemx -mdouble-float
21287 @opindex msingle-float
21288 @opindex mdouble-float
21289 Generate code for single- or double-precision floating-point operations.
21290 @option{-mdouble-float} implies @option{-msingle-float}.
21293 @opindex msimple-fpu
21294 Do not generate @code{sqrt} and @code{div} instructions for hardware
21295 floating-point unit.
21297 @item -mfpu=@var{name}
21299 Specify type of floating-point unit. Valid values for @var{name} are
21300 @samp{sp_lite} (equivalent to @option{-msingle-float -msimple-fpu}),
21301 @samp{dp_lite} (equivalent to @option{-mdouble-float -msimple-fpu}),
21302 @samp{sp_full} (equivalent to @option{-msingle-float}),
21303 and @samp{dp_full} (equivalent to @option{-mdouble-float}).
21306 @opindex mxilinx-fpu
21307 Perform optimizations for the floating-point unit on Xilinx PPC 405/440.
21310 @itemx -mno-multiple
21312 @opindex mno-multiple
21313 Generate code that uses (does not use) the load multiple word
21314 instructions and the store multiple word instructions. These
21315 instructions are generated by default on POWER systems, and not
21316 generated on PowerPC systems. Do not use @option{-mmultiple} on little-endian
21317 PowerPC systems, since those instructions do not work when the
21318 processor is in little-endian mode. The exceptions are PPC740 and
21319 PPC750 which permit these instructions in little-endian mode.
21324 @opindex mno-string
21325 Generate code that uses (does not use) the load string instructions
21326 and the store string word instructions to save multiple registers and
21327 do small block moves. These instructions are generated by default on
21328 POWER systems, and not generated on PowerPC systems. Do not use
21329 @option{-mstring} on little-endian PowerPC systems, since those
21330 instructions do not work when the processor is in little-endian mode.
21331 The exceptions are PPC740 and PPC750 which permit these instructions
21332 in little-endian mode.
21337 @opindex mno-update
21338 Generate code that uses (does not use) the load or store instructions
21339 that update the base register to the address of the calculated memory
21340 location. These instructions are generated by default. If you use
21341 @option{-mno-update}, there is a small window between the time that the
21342 stack pointer is updated and the address of the previous frame is
21343 stored, which means code that walks the stack frame across interrupts or
21344 signals may get corrupted data.
21346 @item -mavoid-indexed-addresses
21347 @itemx -mno-avoid-indexed-addresses
21348 @opindex mavoid-indexed-addresses
21349 @opindex mno-avoid-indexed-addresses
21350 Generate code that tries to avoid (not avoid) the use of indexed load
21351 or store instructions. These instructions can incur a performance
21352 penalty on Power6 processors in certain situations, such as when
21353 stepping through large arrays that cross a 16M boundary. This option
21354 is enabled by default when targeting Power6 and disabled otherwise.
21357 @itemx -mno-fused-madd
21358 @opindex mfused-madd
21359 @opindex mno-fused-madd
21360 Generate code that uses (does not use) the floating-point multiply and
21361 accumulate instructions. These instructions are generated by default
21362 if hardware floating point is used. The machine-dependent
21363 @option{-mfused-madd} option is now mapped to the machine-independent
21364 @option{-ffp-contract=fast} option, and @option{-mno-fused-madd} is
21365 mapped to @option{-ffp-contract=off}.
21371 Generate code that uses (does not use) the half-word multiply and
21372 multiply-accumulate instructions on the IBM 405, 440, 464 and 476 processors.
21373 These instructions are generated by default when targeting those
21380 Generate code that uses (does not use) the string-search @samp{dlmzb}
21381 instruction on the IBM 405, 440, 464 and 476 processors. This instruction is
21382 generated by default when targeting those processors.
21384 @item -mno-bit-align
21386 @opindex mno-bit-align
21387 @opindex mbit-align
21388 On System V.4 and embedded PowerPC systems do not (do) force structures
21389 and unions that contain bit-fields to be aligned to the base type of the
21392 For example, by default a structure containing nothing but 8
21393 @code{unsigned} bit-fields of length 1 is aligned to a 4-byte
21394 boundary and has a size of 4 bytes. By using @option{-mno-bit-align},
21395 the structure is aligned to a 1-byte boundary and is 1 byte in
21398 @item -mno-strict-align
21399 @itemx -mstrict-align
21400 @opindex mno-strict-align
21401 @opindex mstrict-align
21402 On System V.4 and embedded PowerPC systems do not (do) assume that
21403 unaligned memory references are handled by the system.
21405 @item -mrelocatable
21406 @itemx -mno-relocatable
21407 @opindex mrelocatable
21408 @opindex mno-relocatable
21409 Generate code that allows (does not allow) a static executable to be
21410 relocated to a different address at run time. A simple embedded
21411 PowerPC system loader should relocate the entire contents of
21412 @code{.got2} and 4-byte locations listed in the @code{.fixup} section,
21413 a table of 32-bit addresses generated by this option. For this to
21414 work, all objects linked together must be compiled with
21415 @option{-mrelocatable} or @option{-mrelocatable-lib}.
21416 @option{-mrelocatable} code aligns the stack to an 8-byte boundary.
21418 @item -mrelocatable-lib
21419 @itemx -mno-relocatable-lib
21420 @opindex mrelocatable-lib
21421 @opindex mno-relocatable-lib
21422 Like @option{-mrelocatable}, @option{-mrelocatable-lib} generates a
21423 @code{.fixup} section to allow static executables to be relocated at
21424 run time, but @option{-mrelocatable-lib} does not use the smaller stack
21425 alignment of @option{-mrelocatable}. Objects compiled with
21426 @option{-mrelocatable-lib} may be linked with objects compiled with
21427 any combination of the @option{-mrelocatable} options.
21433 On System V.4 and embedded PowerPC systems do not (do) assume that
21434 register 2 contains a pointer to a global area pointing to the addresses
21435 used in the program.
21438 @itemx -mlittle-endian
21440 @opindex mlittle-endian
21441 On System V.4 and embedded PowerPC systems compile code for the
21442 processor in little-endian mode. The @option{-mlittle-endian} option is
21443 the same as @option{-mlittle}.
21446 @itemx -mbig-endian
21448 @opindex mbig-endian
21449 On System V.4 and embedded PowerPC systems compile code for the
21450 processor in big-endian mode. The @option{-mbig-endian} option is
21451 the same as @option{-mbig}.
21453 @item -mdynamic-no-pic
21454 @opindex mdynamic-no-pic
21455 On Darwin and Mac OS X systems, compile code so that it is not
21456 relocatable, but that its external references are relocatable. The
21457 resulting code is suitable for applications, but not shared
21460 @item -msingle-pic-base
21461 @opindex msingle-pic-base
21462 Treat the register used for PIC addressing as read-only, rather than
21463 loading it in the prologue for each function. The runtime system is
21464 responsible for initializing this register with an appropriate value
21465 before execution begins.
21467 @item -mprioritize-restricted-insns=@var{priority}
21468 @opindex mprioritize-restricted-insns
21469 This option controls the priority that is assigned to
21470 dispatch-slot restricted instructions during the second scheduling
21471 pass. The argument @var{priority} takes the value @samp{0}, @samp{1},
21472 or @samp{2} to assign no, highest, or second-highest (respectively)
21473 priority to dispatch-slot restricted
21476 @item -msched-costly-dep=@var{dependence_type}
21477 @opindex msched-costly-dep
21478 This option controls which dependences are considered costly
21479 by the target during instruction scheduling. The argument
21480 @var{dependence_type} takes one of the following values:
21484 No dependence is costly.
21487 All dependences are costly.
21489 @item @samp{true_store_to_load}
21490 A true dependence from store to load is costly.
21492 @item @samp{store_to_load}
21493 Any dependence from store to load is costly.
21496 Any dependence for which the latency is greater than or equal to
21497 @var{number} is costly.
21500 @item -minsert-sched-nops=@var{scheme}
21501 @opindex minsert-sched-nops
21502 This option controls which NOP insertion scheme is used during
21503 the second scheduling pass. The argument @var{scheme} takes one of the
21511 Pad with NOPs any dispatch group that has vacant issue slots,
21512 according to the scheduler's grouping.
21514 @item @samp{regroup_exact}
21515 Insert NOPs to force costly dependent insns into
21516 separate groups. Insert exactly as many NOPs as needed to force an insn
21517 to a new group, according to the estimated processor grouping.
21520 Insert NOPs to force costly dependent insns into
21521 separate groups. Insert @var{number} NOPs to force an insn to a new group.
21525 @opindex mcall-sysv
21526 On System V.4 and embedded PowerPC systems compile code using calling
21527 conventions that adhere to the March 1995 draft of the System V
21528 Application Binary Interface, PowerPC processor supplement. This is the
21529 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
21531 @item -mcall-sysv-eabi
21533 @opindex mcall-sysv-eabi
21534 @opindex mcall-eabi
21535 Specify both @option{-mcall-sysv} and @option{-meabi} options.
21537 @item -mcall-sysv-noeabi
21538 @opindex mcall-sysv-noeabi
21539 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
21541 @item -mcall-aixdesc
21543 On System V.4 and embedded PowerPC systems compile code for the AIX
21547 @opindex mcall-linux
21548 On System V.4 and embedded PowerPC systems compile code for the
21549 Linux-based GNU system.
21551 @item -mcall-freebsd
21552 @opindex mcall-freebsd
21553 On System V.4 and embedded PowerPC systems compile code for the
21554 FreeBSD operating system.
21556 @item -mcall-netbsd
21557 @opindex mcall-netbsd
21558 On System V.4 and embedded PowerPC systems compile code for the
21559 NetBSD operating system.
21561 @item -mcall-openbsd
21562 @opindex mcall-netbsd
21563 On System V.4 and embedded PowerPC systems compile code for the
21564 OpenBSD operating system.
21566 @item -maix-struct-return
21567 @opindex maix-struct-return
21568 Return all structures in memory (as specified by the AIX ABI)@.
21570 @item -msvr4-struct-return
21571 @opindex msvr4-struct-return
21572 Return structures smaller than 8 bytes in registers (as specified by the
21575 @item -mabi=@var{abi-type}
21577 Extend the current ABI with a particular extension, or remove such extension.
21578 Valid values are @samp{altivec}, @samp{no-altivec}, @samp{spe},
21579 @samp{no-spe}, @samp{ibmlongdouble}, @samp{ieeelongdouble},
21580 @samp{elfv1}, @samp{elfv2}@.
21584 Extend the current ABI with SPE ABI extensions. This does not change
21585 the default ABI, instead it adds the SPE ABI extensions to the current
21589 @opindex mabi=no-spe
21590 Disable Book-E SPE ABI extensions for the current ABI@.
21592 @item -mabi=ibmlongdouble
21593 @opindex mabi=ibmlongdouble
21594 Change the current ABI to use IBM extended-precision long double.
21595 This is a PowerPC 32-bit SYSV ABI option.
21597 @item -mabi=ieeelongdouble
21598 @opindex mabi=ieeelongdouble
21599 Change the current ABI to use IEEE extended-precision long double.
21600 This is a PowerPC 32-bit Linux ABI option.
21603 @opindex mabi=elfv1
21604 Change the current ABI to use the ELFv1 ABI.
21605 This is the default ABI for big-endian PowerPC 64-bit Linux.
21606 Overriding the default ABI requires special system support and is
21607 likely to fail in spectacular ways.
21610 @opindex mabi=elfv2
21611 Change the current ABI to use the ELFv2 ABI.
21612 This is the default ABI for little-endian PowerPC 64-bit Linux.
21613 Overriding the default ABI requires special system support and is
21614 likely to fail in spectacular ways.
21616 @item -mgnu-attribute
21617 @itemx -mno-gnu-attribute
21618 @opindex mgnu-attribute
21619 @opindex mno-gnu-attribute
21620 Emit .gnu_attribute assembly directives to set tag/value pairs in a
21621 .gnu.attributes section that specify ABI variations in function
21622 parameters or return values.
21625 @itemx -mno-prototype
21626 @opindex mprototype
21627 @opindex mno-prototype
21628 On System V.4 and embedded PowerPC systems assume that all calls to
21629 variable argument functions are properly prototyped. Otherwise, the
21630 compiler must insert an instruction before every non-prototyped call to
21631 set or clear bit 6 of the condition code register (@code{CR}) to
21632 indicate whether floating-point values are passed in the floating-point
21633 registers in case the function takes variable arguments. With
21634 @option{-mprototype}, only calls to prototyped variable argument functions
21635 set or clear the bit.
21639 On embedded PowerPC systems, assume that the startup module is called
21640 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
21641 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}
21646 On embedded PowerPC systems, assume that the startup module is called
21647 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
21652 On embedded PowerPC systems, assume that the startup module is called
21653 @file{crt0.o} and the standard C libraries are @file{libads.a} and
21656 @item -myellowknife
21657 @opindex myellowknife
21658 On embedded PowerPC systems, assume that the startup module is called
21659 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
21664 On System V.4 and embedded PowerPC systems, specify that you are
21665 compiling for a VxWorks system.
21669 On embedded PowerPC systems, set the @code{PPC_EMB} bit in the ELF flags
21670 header to indicate that @samp{eabi} extended relocations are used.
21676 On System V.4 and embedded PowerPC systems do (do not) adhere to the
21677 Embedded Applications Binary Interface (EABI), which is a set of
21678 modifications to the System V.4 specifications. Selecting @option{-meabi}
21679 means that the stack is aligned to an 8-byte boundary, a function
21680 @code{__eabi} is called from @code{main} to set up the EABI
21681 environment, and the @option{-msdata} option can use both @code{r2} and
21682 @code{r13} to point to two separate small data areas. Selecting
21683 @option{-mno-eabi} means that the stack is aligned to a 16-byte boundary,
21684 no EABI initialization function is called from @code{main}, and the
21685 @option{-msdata} option only uses @code{r13} to point to a single
21686 small data area. The @option{-meabi} option is on by default if you
21687 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
21690 @opindex msdata=eabi
21691 On System V.4 and embedded PowerPC systems, put small initialized
21692 @code{const} global and static data in the @code{.sdata2} section, which
21693 is pointed to by register @code{r2}. Put small initialized
21694 non-@code{const} global and static data in the @code{.sdata} section,
21695 which is pointed to by register @code{r13}. Put small uninitialized
21696 global and static data in the @code{.sbss} section, which is adjacent to
21697 the @code{.sdata} section. The @option{-msdata=eabi} option is
21698 incompatible with the @option{-mrelocatable} option. The
21699 @option{-msdata=eabi} option also sets the @option{-memb} option.
21702 @opindex msdata=sysv
21703 On System V.4 and embedded PowerPC systems, put small global and static
21704 data in the @code{.sdata} section, which is pointed to by register
21705 @code{r13}. Put small uninitialized global and static data in the
21706 @code{.sbss} section, which is adjacent to the @code{.sdata} section.
21707 The @option{-msdata=sysv} option is incompatible with the
21708 @option{-mrelocatable} option.
21710 @item -msdata=default
21712 @opindex msdata=default
21714 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
21715 compile code the same as @option{-msdata=eabi}, otherwise compile code the
21716 same as @option{-msdata=sysv}.
21719 @opindex msdata=data
21720 On System V.4 and embedded PowerPC systems, put small global
21721 data in the @code{.sdata} section. Put small uninitialized global
21722 data in the @code{.sbss} section. Do not use register @code{r13}
21723 to address small data however. This is the default behavior unless
21724 other @option{-msdata} options are used.
21728 @opindex msdata=none
21730 On embedded PowerPC systems, put all initialized global and static data
21731 in the @code{.data} section, and all uninitialized data in the
21732 @code{.bss} section.
21734 @item -mblock-move-inline-limit=@var{num}
21735 @opindex mblock-move-inline-limit
21736 Inline all block moves (such as calls to @code{memcpy} or structure
21737 copies) less than or equal to @var{num} bytes. The minimum value for
21738 @var{num} is 32 bytes on 32-bit targets and 64 bytes on 64-bit
21739 targets. The default value is target-specific.
21743 @cindex smaller data references (PowerPC)
21744 @cindex .sdata/.sdata2 references (PowerPC)
21745 On embedded PowerPC systems, put global and static items less than or
21746 equal to @var{num} bytes into the small data or BSS sections instead of
21747 the normal data or BSS section. By default, @var{num} is 8. The
21748 @option{-G @var{num}} switch is also passed to the linker.
21749 All modules should be compiled with the same @option{-G @var{num}} value.
21752 @itemx -mno-regnames
21754 @opindex mno-regnames
21755 On System V.4 and embedded PowerPC systems do (do not) emit register
21756 names in the assembly language output using symbolic forms.
21759 @itemx -mno-longcall
21761 @opindex mno-longcall
21762 By default assume that all calls are far away so that a longer and more
21763 expensive calling sequence is required. This is required for calls
21764 farther than 32 megabytes (33,554,432 bytes) from the current location.
21765 A short call is generated if the compiler knows
21766 the call cannot be that far away. This setting can be overridden by
21767 the @code{shortcall} function attribute, or by @code{#pragma
21770 Some linkers are capable of detecting out-of-range calls and generating
21771 glue code on the fly. On these systems, long calls are unnecessary and
21772 generate slower code. As of this writing, the AIX linker can do this,
21773 as can the GNU linker for PowerPC/64. It is planned to add this feature
21774 to the GNU linker for 32-bit PowerPC systems as well.
21776 On Darwin/PPC systems, @code{#pragma longcall} generates @code{jbsr
21777 callee, L42}, plus a @dfn{branch island} (glue code). The two target
21778 addresses represent the callee and the branch island. The
21779 Darwin/PPC linker prefers the first address and generates a @code{bl
21780 callee} if the PPC @code{bl} instruction reaches the callee directly;
21781 otherwise, the linker generates @code{bl L42} to call the branch
21782 island. The branch island is appended to the body of the
21783 calling function; it computes the full 32-bit address of the callee
21786 On Mach-O (Darwin) systems, this option directs the compiler emit to
21787 the glue for every direct call, and the Darwin linker decides whether
21788 to use or discard it.
21790 In the future, GCC may ignore all longcall specifications
21791 when the linker is known to generate glue.
21793 @item -mtls-markers
21794 @itemx -mno-tls-markers
21795 @opindex mtls-markers
21796 @opindex mno-tls-markers
21797 Mark (do not mark) calls to @code{__tls_get_addr} with a relocation
21798 specifying the function argument. The relocation allows the linker to
21799 reliably associate function call with argument setup instructions for
21800 TLS optimization, which in turn allows GCC to better schedule the
21805 Adds support for multithreading with the @dfn{pthreads} library.
21806 This option sets flags for both the preprocessor and linker.
21811 This option enables use of the reciprocal estimate and
21812 reciprocal square root estimate instructions with additional
21813 Newton-Raphson steps to increase precision instead of doing a divide or
21814 square root and divide for floating-point arguments. You should use
21815 the @option{-ffast-math} option when using @option{-mrecip} (or at
21816 least @option{-funsafe-math-optimizations},
21817 @option{-ffinite-math-only}, @option{-freciprocal-math} and
21818 @option{-fno-trapping-math}). Note that while the throughput of the
21819 sequence is generally higher than the throughput of the non-reciprocal
21820 instruction, the precision of the sequence can be decreased by up to 2
21821 ulp (i.e.@: the inverse of 1.0 equals 0.99999994) for reciprocal square
21824 @item -mrecip=@var{opt}
21825 @opindex mrecip=opt
21826 This option controls which reciprocal estimate instructions
21827 may be used. @var{opt} is a comma-separated list of options, which may
21828 be preceded by a @code{!} to invert the option:
21833 Enable all estimate instructions.
21836 Enable the default instructions, equivalent to @option{-mrecip}.
21839 Disable all estimate instructions, equivalent to @option{-mno-recip}.
21842 Enable the reciprocal approximation instructions for both
21843 single and double precision.
21846 Enable the single-precision reciprocal approximation instructions.
21849 Enable the double-precision reciprocal approximation instructions.
21852 Enable the reciprocal square root approximation instructions for both
21853 single and double precision.
21856 Enable the single-precision reciprocal square root approximation instructions.
21859 Enable the double-precision reciprocal square root approximation instructions.
21863 So, for example, @option{-mrecip=all,!rsqrtd} enables
21864 all of the reciprocal estimate instructions, except for the
21865 @code{FRSQRTE}, @code{XSRSQRTEDP}, and @code{XVRSQRTEDP} instructions
21866 which handle the double-precision reciprocal square root calculations.
21868 @item -mrecip-precision
21869 @itemx -mno-recip-precision
21870 @opindex mrecip-precision
21871 Assume (do not assume) that the reciprocal estimate instructions
21872 provide higher-precision estimates than is mandated by the PowerPC
21873 ABI. Selecting @option{-mcpu=power6}, @option{-mcpu=power7} or
21874 @option{-mcpu=power8} automatically selects @option{-mrecip-precision}.
21875 The double-precision square root estimate instructions are not generated by
21876 default on low-precision machines, since they do not provide an
21877 estimate that converges after three steps.
21879 @item -mveclibabi=@var{type}
21880 @opindex mveclibabi
21881 Specifies the ABI type to use for vectorizing intrinsics using an
21882 external library. The only type supported at present is @samp{mass},
21883 which specifies to use IBM's Mathematical Acceleration Subsystem
21884 (MASS) libraries for vectorizing intrinsics using external libraries.
21885 GCC currently emits calls to @code{acosd2}, @code{acosf4},
21886 @code{acoshd2}, @code{acoshf4}, @code{asind2}, @code{asinf4},
21887 @code{asinhd2}, @code{asinhf4}, @code{atan2d2}, @code{atan2f4},
21888 @code{atand2}, @code{atanf4}, @code{atanhd2}, @code{atanhf4},
21889 @code{cbrtd2}, @code{cbrtf4}, @code{cosd2}, @code{cosf4},
21890 @code{coshd2}, @code{coshf4}, @code{erfcd2}, @code{erfcf4},
21891 @code{erfd2}, @code{erff4}, @code{exp2d2}, @code{exp2f4},
21892 @code{expd2}, @code{expf4}, @code{expm1d2}, @code{expm1f4},
21893 @code{hypotd2}, @code{hypotf4}, @code{lgammad2}, @code{lgammaf4},
21894 @code{log10d2}, @code{log10f4}, @code{log1pd2}, @code{log1pf4},
21895 @code{log2d2}, @code{log2f4}, @code{logd2}, @code{logf4},
21896 @code{powd2}, @code{powf4}, @code{sind2}, @code{sinf4}, @code{sinhd2},
21897 @code{sinhf4}, @code{sqrtd2}, @code{sqrtf4}, @code{tand2},
21898 @code{tanf4}, @code{tanhd2}, and @code{tanhf4} when generating code
21899 for power7. Both @option{-ftree-vectorize} and
21900 @option{-funsafe-math-optimizations} must also be enabled. The MASS
21901 libraries must be specified at link time.
21906 Generate (do not generate) the @code{friz} instruction when the
21907 @option{-funsafe-math-optimizations} option is used to optimize
21908 rounding of floating-point values to 64-bit integer and back to floating
21909 point. The @code{friz} instruction does not return the same value if
21910 the floating-point number is too large to fit in an integer.
21912 @item -mpointers-to-nested-functions
21913 @itemx -mno-pointers-to-nested-functions
21914 @opindex mpointers-to-nested-functions
21915 Generate (do not generate) code to load up the static chain register
21916 (@code{r11}) when calling through a pointer on AIX and 64-bit Linux
21917 systems where a function pointer points to a 3-word descriptor giving
21918 the function address, TOC value to be loaded in register @code{r2}, and
21919 static chain value to be loaded in register @code{r11}. The
21920 @option{-mpointers-to-nested-functions} is on by default. You cannot
21921 call through pointers to nested functions or pointers
21922 to functions compiled in other languages that use the static chain if
21923 you use @option{-mno-pointers-to-nested-functions}.
21925 @item -msave-toc-indirect
21926 @itemx -mno-save-toc-indirect
21927 @opindex msave-toc-indirect
21928 Generate (do not generate) code to save the TOC value in the reserved
21929 stack location in the function prologue if the function calls through
21930 a pointer on AIX and 64-bit Linux systems. If the TOC value is not
21931 saved in the prologue, it is saved just before the call through the
21932 pointer. The @option{-mno-save-toc-indirect} option is the default.
21934 @item -mcompat-align-parm
21935 @itemx -mno-compat-align-parm
21936 @opindex mcompat-align-parm
21937 Generate (do not generate) code to pass structure parameters with a
21938 maximum alignment of 64 bits, for compatibility with older versions
21941 Older versions of GCC (prior to 4.9.0) incorrectly did not align a
21942 structure parameter on a 128-bit boundary when that structure contained
21943 a member requiring 128-bit alignment. This is corrected in more
21944 recent versions of GCC. This option may be used to generate code
21945 that is compatible with functions compiled with older versions of
21948 The @option{-mno-compat-align-parm} option is the default.
21952 @subsection RX Options
21955 These command-line options are defined for RX targets:
21958 @item -m64bit-doubles
21959 @itemx -m32bit-doubles
21960 @opindex m64bit-doubles
21961 @opindex m32bit-doubles
21962 Make the @code{double} data type be 64 bits (@option{-m64bit-doubles})
21963 or 32 bits (@option{-m32bit-doubles}) in size. The default is
21964 @option{-m32bit-doubles}. @emph{Note} RX floating-point hardware only
21965 works on 32-bit values, which is why the default is
21966 @option{-m32bit-doubles}.
21972 Enables (@option{-fpu}) or disables (@option{-nofpu}) the use of RX
21973 floating-point hardware. The default is enabled for the RX600
21974 series and disabled for the RX200 series.
21976 Floating-point instructions are only generated for 32-bit floating-point
21977 values, however, so the FPU hardware is not used for doubles if the
21978 @option{-m64bit-doubles} option is used.
21980 @emph{Note} If the @option{-fpu} option is enabled then
21981 @option{-funsafe-math-optimizations} is also enabled automatically.
21982 This is because the RX FPU instructions are themselves unsafe.
21984 @item -mcpu=@var{name}
21986 Selects the type of RX CPU to be targeted. Currently three types are
21987 supported, the generic @samp{RX600} and @samp{RX200} series hardware and
21988 the specific @samp{RX610} CPU. The default is @samp{RX600}.
21990 The only difference between @samp{RX600} and @samp{RX610} is that the
21991 @samp{RX610} does not support the @code{MVTIPL} instruction.
21993 The @samp{RX200} series does not have a hardware floating-point unit
21994 and so @option{-nofpu} is enabled by default when this type is
21997 @item -mbig-endian-data
21998 @itemx -mlittle-endian-data
21999 @opindex mbig-endian-data
22000 @opindex mlittle-endian-data
22001 Store data (but not code) in the big-endian format. The default is
22002 @option{-mlittle-endian-data}, i.e.@: to store data in the little-endian
22005 @item -msmall-data-limit=@var{N}
22006 @opindex msmall-data-limit
22007 Specifies the maximum size in bytes of global and static variables
22008 which can be placed into the small data area. Using the small data
22009 area can lead to smaller and faster code, but the size of area is
22010 limited and it is up to the programmer to ensure that the area does
22011 not overflow. Also when the small data area is used one of the RX's
22012 registers (usually @code{r13}) is reserved for use pointing to this
22013 area, so it is no longer available for use by the compiler. This
22014 could result in slower and/or larger code if variables are pushed onto
22015 the stack instead of being held in this register.
22017 Note, common variables (variables that have not been initialized) and
22018 constants are not placed into the small data area as they are assigned
22019 to other sections in the output executable.
22021 The default value is zero, which disables this feature. Note, this
22022 feature is not enabled by default with higher optimization levels
22023 (@option{-O2} etc) because of the potentially detrimental effects of
22024 reserving a register. It is up to the programmer to experiment and
22025 discover whether this feature is of benefit to their program. See the
22026 description of the @option{-mpid} option for a description of how the
22027 actual register to hold the small data area pointer is chosen.
22033 Use the simulator runtime. The default is to use the libgloss
22034 board-specific runtime.
22036 @item -mas100-syntax
22037 @itemx -mno-as100-syntax
22038 @opindex mas100-syntax
22039 @opindex mno-as100-syntax
22040 When generating assembler output use a syntax that is compatible with
22041 Renesas's AS100 assembler. This syntax can also be handled by the GAS
22042 assembler, but it has some restrictions so it is not generated by default.
22044 @item -mmax-constant-size=@var{N}
22045 @opindex mmax-constant-size
22046 Specifies the maximum size, in bytes, of a constant that can be used as
22047 an operand in a RX instruction. Although the RX instruction set does
22048 allow constants of up to 4 bytes in length to be used in instructions,
22049 a longer value equates to a longer instruction. Thus in some
22050 circumstances it can be beneficial to restrict the size of constants
22051 that are used in instructions. Constants that are too big are instead
22052 placed into a constant pool and referenced via register indirection.
22054 The value @var{N} can be between 0 and 4. A value of 0 (the default)
22055 or 4 means that constants of any size are allowed.
22059 Enable linker relaxation. Linker relaxation is a process whereby the
22060 linker attempts to reduce the size of a program by finding shorter
22061 versions of various instructions. Disabled by default.
22063 @item -mint-register=@var{N}
22064 @opindex mint-register
22065 Specify the number of registers to reserve for fast interrupt handler
22066 functions. The value @var{N} can be between 0 and 4. A value of 1
22067 means that register @code{r13} is reserved for the exclusive use
22068 of fast interrupt handlers. A value of 2 reserves @code{r13} and
22069 @code{r12}. A value of 3 reserves @code{r13}, @code{r12} and
22070 @code{r11}, and a value of 4 reserves @code{r13} through @code{r10}.
22071 A value of 0, the default, does not reserve any registers.
22073 @item -msave-acc-in-interrupts
22074 @opindex msave-acc-in-interrupts
22075 Specifies that interrupt handler functions should preserve the
22076 accumulator register. This is only necessary if normal code might use
22077 the accumulator register, for example because it performs 64-bit
22078 multiplications. The default is to ignore the accumulator as this
22079 makes the interrupt handlers faster.
22085 Enables the generation of position independent data. When enabled any
22086 access to constant data is done via an offset from a base address
22087 held in a register. This allows the location of constant data to be
22088 determined at run time without requiring the executable to be
22089 relocated, which is a benefit to embedded applications with tight
22090 memory constraints. Data that can be modified is not affected by this
22093 Note, using this feature reserves a register, usually @code{r13}, for
22094 the constant data base address. This can result in slower and/or
22095 larger code, especially in complicated functions.
22097 The actual register chosen to hold the constant data base address
22098 depends upon whether the @option{-msmall-data-limit} and/or the
22099 @option{-mint-register} command-line options are enabled. Starting
22100 with register @code{r13} and proceeding downwards, registers are
22101 allocated first to satisfy the requirements of @option{-mint-register},
22102 then @option{-mpid} and finally @option{-msmall-data-limit}. Thus it
22103 is possible for the small data area register to be @code{r8} if both
22104 @option{-mint-register=4} and @option{-mpid} are specified on the
22107 By default this feature is not enabled. The default can be restored
22108 via the @option{-mno-pid} command-line option.
22110 @item -mno-warn-multiple-fast-interrupts
22111 @itemx -mwarn-multiple-fast-interrupts
22112 @opindex mno-warn-multiple-fast-interrupts
22113 @opindex mwarn-multiple-fast-interrupts
22114 Prevents GCC from issuing a warning message if it finds more than one
22115 fast interrupt handler when it is compiling a file. The default is to
22116 issue a warning for each extra fast interrupt handler found, as the RX
22117 only supports one such interrupt.
22119 @item -mallow-string-insns
22120 @itemx -mno-allow-string-insns
22121 @opindex mallow-string-insns
22122 @opindex mno-allow-string-insns
22123 Enables or disables the use of the string manipulation instructions
22124 @code{SMOVF}, @code{SCMPU}, @code{SMOVB}, @code{SMOVU}, @code{SUNTIL}
22125 @code{SWHILE} and also the @code{RMPA} instruction. These
22126 instructions may prefetch data, which is not safe to do if accessing
22127 an I/O register. (See section 12.2.7 of the RX62N Group User's Manual
22128 for more information).
22130 The default is to allow these instructions, but it is not possible for
22131 GCC to reliably detect all circumstances where a string instruction
22132 might be used to access an I/O register, so their use cannot be
22133 disabled automatically. Instead it is reliant upon the programmer to
22134 use the @option{-mno-allow-string-insns} option if their program
22135 accesses I/O space.
22137 When the instructions are enabled GCC defines the C preprocessor
22138 symbol @code{__RX_ALLOW_STRING_INSNS__}, otherwise it defines the
22139 symbol @code{__RX_DISALLOW_STRING_INSNS__}.
22145 Use only (or not only) @code{JSR} instructions to access functions.
22146 This option can be used when code size exceeds the range of @code{BSR}
22147 instructions. Note that @option{-mno-jsr} does not mean to not use
22148 @code{JSR} but instead means that any type of branch may be used.
22151 @emph{Note:} The generic GCC command-line option @option{-ffixed-@var{reg}}
22152 has special significance to the RX port when used with the
22153 @code{interrupt} function attribute. This attribute indicates a
22154 function intended to process fast interrupts. GCC ensures
22155 that it only uses the registers @code{r10}, @code{r11}, @code{r12}
22156 and/or @code{r13} and only provided that the normal use of the
22157 corresponding registers have been restricted via the
22158 @option{-ffixed-@var{reg}} or @option{-mint-register} command-line
22161 @node S/390 and zSeries Options
22162 @subsection S/390 and zSeries Options
22163 @cindex S/390 and zSeries Options
22165 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
22169 @itemx -msoft-float
22170 @opindex mhard-float
22171 @opindex msoft-float
22172 Use (do not use) the hardware floating-point instructions and registers
22173 for floating-point operations. When @option{-msoft-float} is specified,
22174 functions in @file{libgcc.a} are used to perform floating-point
22175 operations. When @option{-mhard-float} is specified, the compiler
22176 generates IEEE floating-point instructions. This is the default.
22179 @itemx -mno-hard-dfp
22181 @opindex mno-hard-dfp
22182 Use (do not use) the hardware decimal-floating-point instructions for
22183 decimal-floating-point operations. When @option{-mno-hard-dfp} is
22184 specified, functions in @file{libgcc.a} are used to perform
22185 decimal-floating-point operations. When @option{-mhard-dfp} is
22186 specified, the compiler generates decimal-floating-point hardware
22187 instructions. This is the default for @option{-march=z9-ec} or higher.
22189 @item -mlong-double-64
22190 @itemx -mlong-double-128
22191 @opindex mlong-double-64
22192 @opindex mlong-double-128
22193 These switches control the size of @code{long double} type. A size
22194 of 64 bits makes the @code{long double} type equivalent to the @code{double}
22195 type. This is the default.
22198 @itemx -mno-backchain
22199 @opindex mbackchain
22200 @opindex mno-backchain
22201 Store (do not store) the address of the caller's frame as backchain pointer
22202 into the callee's stack frame.
22203 A backchain may be needed to allow debugging using tools that do not understand
22204 DWARF call frame information.
22205 When @option{-mno-packed-stack} is in effect, the backchain pointer is stored
22206 at the bottom of the stack frame; when @option{-mpacked-stack} is in effect,
22207 the backchain is placed into the topmost word of the 96/160 byte register
22210 In general, code compiled with @option{-mbackchain} is call-compatible with
22211 code compiled with @option{-mmo-backchain}; however, use of the backchain
22212 for debugging purposes usually requires that the whole binary is built with
22213 @option{-mbackchain}. Note that the combination of @option{-mbackchain},
22214 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
22215 to build a linux kernel use @option{-msoft-float}.
22217 The default is to not maintain the backchain.
22219 @item -mpacked-stack
22220 @itemx -mno-packed-stack
22221 @opindex mpacked-stack
22222 @opindex mno-packed-stack
22223 Use (do not use) the packed stack layout. When @option{-mno-packed-stack} is
22224 specified, the compiler uses the all fields of the 96/160 byte register save
22225 area only for their default purpose; unused fields still take up stack space.
22226 When @option{-mpacked-stack} is specified, register save slots are densely
22227 packed at the top of the register save area; unused space is reused for other
22228 purposes, allowing for more efficient use of the available stack space.
22229 However, when @option{-mbackchain} is also in effect, the topmost word of
22230 the save area is always used to store the backchain, and the return address
22231 register is always saved two words below the backchain.
22233 As long as the stack frame backchain is not used, code generated with
22234 @option{-mpacked-stack} is call-compatible with code generated with
22235 @option{-mno-packed-stack}. Note that some non-FSF releases of GCC 2.95 for
22236 S/390 or zSeries generated code that uses the stack frame backchain at run
22237 time, not just for debugging purposes. Such code is not call-compatible
22238 with code compiled with @option{-mpacked-stack}. Also, note that the
22239 combination of @option{-mbackchain},
22240 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
22241 to build a linux kernel use @option{-msoft-float}.
22243 The default is to not use the packed stack layout.
22246 @itemx -mno-small-exec
22247 @opindex msmall-exec
22248 @opindex mno-small-exec
22249 Generate (or do not generate) code using the @code{bras} instruction
22250 to do subroutine calls.
22251 This only works reliably if the total executable size does not
22252 exceed 64k. The default is to use the @code{basr} instruction instead,
22253 which does not have this limitation.
22259 When @option{-m31} is specified, generate code compliant to the
22260 GNU/Linux for S/390 ABI@. When @option{-m64} is specified, generate
22261 code compliant to the GNU/Linux for zSeries ABI@. This allows GCC in
22262 particular to generate 64-bit instructions. For the @samp{s390}
22263 targets, the default is @option{-m31}, while the @samp{s390x}
22264 targets default to @option{-m64}.
22270 When @option{-mzarch} is specified, generate code using the
22271 instructions available on z/Architecture.
22272 When @option{-mesa} is specified, generate code using the
22273 instructions available on ESA/390. Note that @option{-mesa} is
22274 not possible with @option{-m64}.
22275 When generating code compliant to the GNU/Linux for S/390 ABI,
22276 the default is @option{-mesa}. When generating code compliant
22277 to the GNU/Linux for zSeries ABI, the default is @option{-mzarch}.
22283 The @option{-mhtm} option enables a set of builtins making use of
22284 instructions available with the transactional execution facility
22285 introduced with the IBM zEnterprise EC12 machine generation
22286 @ref{S/390 System z Built-in Functions}.
22287 @option{-mhtm} is enabled by default when using @option{-march=zEC12}.
22293 When @option{-mvx} is specified, generate code using the instructions
22294 available with the vector extension facility introduced with the IBM
22295 z13 machine generation.
22296 This option changes the ABI for some vector type values with regard to
22297 alignment and calling conventions. In case vector type values are
22298 being used in an ABI-relevant context a GAS @samp{.gnu_attribute}
22299 command will be added to mark the resulting binary with the ABI used.
22300 @option{-mvx} is enabled by default when using @option{-march=z13}.
22303 @itemx -mno-zvector
22305 @opindex mno-zvector
22306 The @option{-mzvector} option enables vector language extensions and
22307 builtins using instructions available with the vector extension
22308 facility introduced with the IBM z13 machine generation.
22309 This option adds support for @samp{vector} to be used as a keyword to
22310 define vector type variables and arguments. @samp{vector} is only
22311 available when GNU extensions are enabled. It will not be expanded
22312 when requesting strict standard compliance e.g. with @option{-std=c99}.
22313 In addition to the GCC low-level builtins @option{-mzvector} enables
22314 a set of builtins added for compatibility with AltiVec-style
22315 implementations like Power and Cell. In order to make use of these
22316 builtins the header file @file{vecintrin.h} needs to be included.
22317 @option{-mzvector} is disabled by default.
22323 Generate (or do not generate) code using the @code{mvcle} instruction
22324 to perform block moves. When @option{-mno-mvcle} is specified,
22325 use a @code{mvc} loop instead. This is the default unless optimizing for
22332 Print (or do not print) additional debug information when compiling.
22333 The default is to not print debug information.
22335 @item -march=@var{cpu-type}
22337 Generate code that runs on @var{cpu-type}, which is the name of a
22338 system representing a certain processor type. Possible values for
22339 @var{cpu-type} are @samp{z900}/@samp{arch5}, @samp{z990}/@samp{arch6},
22340 @samp{z9-109}, @samp{z9-ec}/@samp{arch7}, @samp{z10}/@samp{arch8},
22341 @samp{z196}/@samp{arch9}, @samp{zEC12}, @samp{z13}/@samp{arch11}, and
22344 The default is @option{-march=z900}. @samp{g5}/@samp{arch3} and
22345 @samp{g6} are deprecated and will be removed with future releases.
22347 Specifying @samp{native} as cpu type can be used to select the best
22348 architecture option for the host processor.
22349 @option{-march=native} has no effect if GCC does not recognize the
22352 @item -mtune=@var{cpu-type}
22354 Tune to @var{cpu-type} everything applicable about the generated code,
22355 except for the ABI and the set of available instructions.
22356 The list of @var{cpu-type} values is the same as for @option{-march}.
22357 The default is the value used for @option{-march}.
22360 @itemx -mno-tpf-trace
22361 @opindex mtpf-trace
22362 @opindex mno-tpf-trace
22363 Generate code that adds (does not add) in TPF OS specific branches to trace
22364 routines in the operating system. This option is off by default, even
22365 when compiling for the TPF OS@.
22368 @itemx -mno-fused-madd
22369 @opindex mfused-madd
22370 @opindex mno-fused-madd
22371 Generate code that uses (does not use) the floating-point multiply and
22372 accumulate instructions. These instructions are generated by default if
22373 hardware floating point is used.
22375 @item -mwarn-framesize=@var{framesize}
22376 @opindex mwarn-framesize
22377 Emit a warning if the current function exceeds the given frame size. Because
22378 this is a compile-time check it doesn't need to be a real problem when the program
22379 runs. It is intended to identify functions that most probably cause
22380 a stack overflow. It is useful to be used in an environment with limited stack
22381 size e.g.@: the linux kernel.
22383 @item -mwarn-dynamicstack
22384 @opindex mwarn-dynamicstack
22385 Emit a warning if the function calls @code{alloca} or uses dynamically-sized
22386 arrays. This is generally a bad idea with a limited stack size.
22388 @item -mstack-guard=@var{stack-guard}
22389 @itemx -mstack-size=@var{stack-size}
22390 @opindex mstack-guard
22391 @opindex mstack-size
22392 If these options are provided the S/390 back end emits additional instructions in
22393 the function prologue that trigger a trap if the stack size is @var{stack-guard}
22394 bytes above the @var{stack-size} (remember that the stack on S/390 grows downward).
22395 If the @var{stack-guard} option is omitted the smallest power of 2 larger than
22396 the frame size of the compiled function is chosen.
22397 These options are intended to be used to help debugging stack overflow problems.
22398 The additionally emitted code causes only little overhead and hence can also be
22399 used in production-like systems without greater performance degradation. The given
22400 values have to be exact powers of 2 and @var{stack-size} has to be greater than
22401 @var{stack-guard} without exceeding 64k.
22402 In order to be efficient the extra code makes the assumption that the stack starts
22403 at an address aligned to the value given by @var{stack-size}.
22404 The @var{stack-guard} option can only be used in conjunction with @var{stack-size}.
22406 @item -mhotpatch=@var{pre-halfwords},@var{post-halfwords}
22408 If the hotpatch option is enabled, a ``hot-patching'' function
22409 prologue is generated for all functions in the compilation unit.
22410 The funtion label is prepended with the given number of two-byte
22411 NOP instructions (@var{pre-halfwords}, maximum 1000000). After
22412 the label, 2 * @var{post-halfwords} bytes are appended, using the
22413 largest NOP like instructions the architecture allows (maximum
22416 If both arguments are zero, hotpatching is disabled.
22418 This option can be overridden for individual functions with the
22419 @code{hotpatch} attribute.
22422 @node Score Options
22423 @subsection Score Options
22424 @cindex Score Options
22426 These options are defined for Score implementations:
22431 Compile code for big-endian mode. This is the default.
22435 Compile code for little-endian mode.
22439 Disable generation of @code{bcnz} instructions.
22443 Enable generation of unaligned load and store instructions.
22447 Enable the use of multiply-accumulate instructions. Disabled by default.
22451 Specify the SCORE5 as the target architecture.
22455 Specify the SCORE5U of the target architecture.
22459 Specify the SCORE7 as the target architecture. This is the default.
22463 Specify the SCORE7D as the target architecture.
22467 @subsection SH Options
22469 These @samp{-m} options are defined for the SH implementations:
22474 Generate code for the SH1.
22478 Generate code for the SH2.
22481 Generate code for the SH2e.
22485 Generate code for the SH2a without FPU, or for a SH2a-FPU in such a way
22486 that the floating-point unit is not used.
22488 @item -m2a-single-only
22489 @opindex m2a-single-only
22490 Generate code for the SH2a-FPU, in such a way that no double-precision
22491 floating-point operations are used.
22494 @opindex m2a-single
22495 Generate code for the SH2a-FPU assuming the floating-point unit is in
22496 single-precision mode by default.
22500 Generate code for the SH2a-FPU assuming the floating-point unit is in
22501 double-precision mode by default.
22505 Generate code for the SH3.
22509 Generate code for the SH3e.
22513 Generate code for the SH4 without a floating-point unit.
22515 @item -m4-single-only
22516 @opindex m4-single-only
22517 Generate code for the SH4 with a floating-point unit that only
22518 supports single-precision arithmetic.
22522 Generate code for the SH4 assuming the floating-point unit is in
22523 single-precision mode by default.
22527 Generate code for the SH4.
22531 Generate code for SH4-100.
22533 @item -m4-100-nofpu
22534 @opindex m4-100-nofpu
22535 Generate code for SH4-100 in such a way that the
22536 floating-point unit is not used.
22538 @item -m4-100-single
22539 @opindex m4-100-single
22540 Generate code for SH4-100 assuming the floating-point unit is in
22541 single-precision mode by default.
22543 @item -m4-100-single-only
22544 @opindex m4-100-single-only
22545 Generate code for SH4-100 in such a way that no double-precision
22546 floating-point operations are used.
22550 Generate code for SH4-200.
22552 @item -m4-200-nofpu
22553 @opindex m4-200-nofpu
22554 Generate code for SH4-200 without in such a way that the
22555 floating-point unit is not used.
22557 @item -m4-200-single
22558 @opindex m4-200-single
22559 Generate code for SH4-200 assuming the floating-point unit is in
22560 single-precision mode by default.
22562 @item -m4-200-single-only
22563 @opindex m4-200-single-only
22564 Generate code for SH4-200 in such a way that no double-precision
22565 floating-point operations are used.
22569 Generate code for SH4-300.
22571 @item -m4-300-nofpu
22572 @opindex m4-300-nofpu
22573 Generate code for SH4-300 without in such a way that the
22574 floating-point unit is not used.
22576 @item -m4-300-single
22577 @opindex m4-300-single
22578 Generate code for SH4-300 in such a way that no double-precision
22579 floating-point operations are used.
22581 @item -m4-300-single-only
22582 @opindex m4-300-single-only
22583 Generate code for SH4-300 in such a way that no double-precision
22584 floating-point operations are used.
22588 Generate code for SH4-340 (no MMU, no FPU).
22592 Generate code for SH4-500 (no FPU). Passes @option{-isa=sh4-nofpu} to the
22597 Generate code for the SH4al-dsp, or for a SH4a in such a way that the
22598 floating-point unit is not used.
22600 @item -m4a-single-only
22601 @opindex m4a-single-only
22602 Generate code for the SH4a, in such a way that no double-precision
22603 floating-point operations are used.
22606 @opindex m4a-single
22607 Generate code for the SH4a assuming the floating-point unit is in
22608 single-precision mode by default.
22612 Generate code for the SH4a.
22616 Same as @option{-m4a-nofpu}, except that it implicitly passes
22617 @option{-dsp} to the assembler. GCC doesn't generate any DSP
22618 instructions at the moment.
22622 Compile code for the processor in big-endian mode.
22626 Compile code for the processor in little-endian mode.
22630 Align doubles at 64-bit boundaries. Note that this changes the calling
22631 conventions, and thus some functions from the standard C library do
22632 not work unless you recompile it first with @option{-mdalign}.
22636 Shorten some address references at link time, when possible; uses the
22637 linker option @option{-relax}.
22641 Use 32-bit offsets in @code{switch} tables. The default is to use
22646 Enable the use of bit manipulation instructions on SH2A.
22650 Enable the use of the instruction @code{fmovd}. Check @option{-mdalign} for
22651 alignment constraints.
22655 Comply with the calling conventions defined by Renesas.
22658 @opindex mno-renesas
22659 Comply with the calling conventions defined for GCC before the Renesas
22660 conventions were available. This option is the default for all
22661 targets of the SH toolchain.
22664 @opindex mnomacsave
22665 Mark the @code{MAC} register as call-clobbered, even if
22666 @option{-mrenesas} is given.
22672 Control the IEEE compliance of floating-point comparisons, which affects the
22673 handling of cases where the result of a comparison is unordered. By default
22674 @option{-mieee} is implicitly enabled. If @option{-ffinite-math-only} is
22675 enabled @option{-mno-ieee} is implicitly set, which results in faster
22676 floating-point greater-equal and less-equal comparisons. The implicit settings
22677 can be overridden by specifying either @option{-mieee} or @option{-mno-ieee}.
22679 @item -minline-ic_invalidate
22680 @opindex minline-ic_invalidate
22681 Inline code to invalidate instruction cache entries after setting up
22682 nested function trampolines.
22683 This option has no effect if @option{-musermode} is in effect and the selected
22684 code generation option (e.g. @option{-m4}) does not allow the use of the @code{icbi}
22686 If the selected code generation option does not allow the use of the @code{icbi}
22687 instruction, and @option{-musermode} is not in effect, the inlined code
22688 manipulates the instruction cache address array directly with an associative
22689 write. This not only requires privileged mode at run time, but it also
22690 fails if the cache line had been mapped via the TLB and has become unmapped.
22694 Dump instruction size and location in the assembly code.
22697 @opindex mpadstruct
22698 This option is deprecated. It pads structures to multiple of 4 bytes,
22699 which is incompatible with the SH ABI@.
22701 @item -matomic-model=@var{model}
22702 @opindex matomic-model=@var{model}
22703 Sets the model of atomic operations and additional parameters as a comma
22704 separated list. For details on the atomic built-in functions see
22705 @ref{__atomic Builtins}. The following models and parameters are supported:
22710 Disable compiler generated atomic sequences and emit library calls for atomic
22711 operations. This is the default if the target is not @code{sh*-*-linux*}.
22714 Generate GNU/Linux compatible gUSA software atomic sequences for the atomic
22715 built-in functions. The generated atomic sequences require additional support
22716 from the interrupt/exception handling code of the system and are only suitable
22717 for SH3* and SH4* single-core systems. This option is enabled by default when
22718 the target is @code{sh*-*-linux*} and SH3* or SH4*. When the target is SH4A,
22719 this option also partially utilizes the hardware atomic instructions
22720 @code{movli.l} and @code{movco.l} to create more efficient code, unless
22721 @samp{strict} is specified.
22724 Generate software atomic sequences that use a variable in the thread control
22725 block. This is a variation of the gUSA sequences which can also be used on
22726 SH1* and SH2* targets. The generated atomic sequences require additional
22727 support from the interrupt/exception handling code of the system and are only
22728 suitable for single-core systems. When using this model, the @samp{gbr-offset=}
22729 parameter has to be specified as well.
22732 Generate software atomic sequences that temporarily disable interrupts by
22733 setting @code{SR.IMASK = 1111}. This model works only when the program runs
22734 in privileged mode and is only suitable for single-core systems. Additional
22735 support from the interrupt/exception handling code of the system is not
22736 required. This model is enabled by default when the target is
22737 @code{sh*-*-linux*} and SH1* or SH2*.
22740 Generate hardware atomic sequences using the @code{movli.l} and @code{movco.l}
22741 instructions only. This is only available on SH4A and is suitable for
22742 multi-core systems. Since the hardware instructions support only 32 bit atomic
22743 variables access to 8 or 16 bit variables is emulated with 32 bit accesses.
22744 Code compiled with this option is also compatible with other software
22745 atomic model interrupt/exception handling systems if executed on an SH4A
22746 system. Additional support from the interrupt/exception handling code of the
22747 system is not required for this model.
22750 This parameter specifies the offset in bytes of the variable in the thread
22751 control block structure that should be used by the generated atomic sequences
22752 when the @samp{soft-tcb} model has been selected. For other models this
22753 parameter is ignored. The specified value must be an integer multiple of four
22754 and in the range 0-1020.
22757 This parameter prevents mixed usage of multiple atomic models, even if they
22758 are compatible, and makes the compiler generate atomic sequences of the
22759 specified model only.
22765 Generate the @code{tas.b} opcode for @code{__atomic_test_and_set}.
22766 Notice that depending on the particular hardware and software configuration
22767 this can degrade overall performance due to the operand cache line flushes
22768 that are implied by the @code{tas.b} instruction. On multi-core SH4A
22769 processors the @code{tas.b} instruction must be used with caution since it
22770 can result in data corruption for certain cache configurations.
22773 @opindex mprefergot
22774 When generating position-independent code, emit function calls using
22775 the Global Offset Table instead of the Procedure Linkage Table.
22778 @itemx -mno-usermode
22780 @opindex mno-usermode
22781 Don't allow (allow) the compiler generating privileged mode code. Specifying
22782 @option{-musermode} also implies @option{-mno-inline-ic_invalidate} if the
22783 inlined code would not work in user mode. @option{-musermode} is the default
22784 when the target is @code{sh*-*-linux*}. If the target is SH1* or SH2*
22785 @option{-musermode} has no effect, since there is no user mode.
22787 @item -multcost=@var{number}
22788 @opindex multcost=@var{number}
22789 Set the cost to assume for a multiply insn.
22791 @item -mdiv=@var{strategy}
22792 @opindex mdiv=@var{strategy}
22793 Set the division strategy to be used for integer division operations.
22794 @var{strategy} can be one of:
22799 Calls a library function that uses the single-step division instruction
22800 @code{div1} to perform the operation. Division by zero calculates an
22801 unspecified result and does not trap. This is the default except for SH4,
22802 SH2A and SHcompact.
22805 Calls a library function that performs the operation in double precision
22806 floating point. Division by zero causes a floating-point exception. This is
22807 the default for SHcompact with FPU. Specifying this for targets that do not
22808 have a double precision FPU defaults to @code{call-div1}.
22811 Calls a library function that uses a lookup table for small divisors and
22812 the @code{div1} instruction with case distinction for larger divisors. Division
22813 by zero calculates an unspecified result and does not trap. This is the default
22814 for SH4. Specifying this for targets that do not have dynamic shift
22815 instructions defaults to @code{call-div1}.
22819 When a division strategy has not been specified the default strategy is
22820 selected based on the current target. For SH2A the default strategy is to
22821 use the @code{divs} and @code{divu} instructions instead of library function
22824 @item -maccumulate-outgoing-args
22825 @opindex maccumulate-outgoing-args
22826 Reserve space once for outgoing arguments in the function prologue rather
22827 than around each call. Generally beneficial for performance and size. Also
22828 needed for unwinding to avoid changing the stack frame around conditional code.
22830 @item -mdivsi3_libfunc=@var{name}
22831 @opindex mdivsi3_libfunc=@var{name}
22832 Set the name of the library function used for 32-bit signed division to
22834 This only affects the name used in the @samp{call} division strategies, and
22835 the compiler still expects the same sets of input/output/clobbered registers as
22836 if this option were not present.
22838 @item -mfixed-range=@var{register-range}
22839 @opindex mfixed-range
22840 Generate code treating the given register range as fixed registers.
22841 A fixed register is one that the register allocator can not use. This is
22842 useful when compiling kernel code. A register range is specified as
22843 two registers separated by a dash. Multiple register ranges can be
22844 specified separated by a comma.
22846 @item -mbranch-cost=@var{num}
22847 @opindex mbranch-cost=@var{num}
22848 Assume @var{num} to be the cost for a branch instruction. Higher numbers
22849 make the compiler try to generate more branch-free code if possible.
22850 If not specified the value is selected depending on the processor type that
22851 is being compiled for.
22854 @itemx -mno-zdcbranch
22855 @opindex mzdcbranch
22856 @opindex mno-zdcbranch
22857 Assume (do not assume) that zero displacement conditional branch instructions
22858 @code{bt} and @code{bf} are fast. If @option{-mzdcbranch} is specified, the
22859 compiler prefers zero displacement branch code sequences. This is
22860 enabled by default when generating code for SH4 and SH4A. It can be explicitly
22861 disabled by specifying @option{-mno-zdcbranch}.
22863 @item -mcbranch-force-delay-slot
22864 @opindex mcbranch-force-delay-slot
22865 Force the usage of delay slots for conditional branches, which stuffs the delay
22866 slot with a @code{nop} if a suitable instruction can't be found. By default
22867 this option is disabled. It can be enabled to work around hardware bugs as
22868 found in the original SH7055.
22871 @itemx -mno-fused-madd
22872 @opindex mfused-madd
22873 @opindex mno-fused-madd
22874 Generate code that uses (does not use) the floating-point multiply and
22875 accumulate instructions. These instructions are generated by default
22876 if hardware floating point is used. The machine-dependent
22877 @option{-mfused-madd} option is now mapped to the machine-independent
22878 @option{-ffp-contract=fast} option, and @option{-mno-fused-madd} is
22879 mapped to @option{-ffp-contract=off}.
22885 Allow or disallow the compiler to emit the @code{fsca} instruction for sine
22886 and cosine approximations. The option @option{-mfsca} must be used in
22887 combination with @option{-funsafe-math-optimizations}. It is enabled by default
22888 when generating code for SH4A. Using @option{-mno-fsca} disables sine and cosine
22889 approximations even if @option{-funsafe-math-optimizations} is in effect.
22895 Allow or disallow the compiler to emit the @code{fsrra} instruction for
22896 reciprocal square root approximations. The option @option{-mfsrra} must be used
22897 in combination with @option{-funsafe-math-optimizations} and
22898 @option{-ffinite-math-only}. It is enabled by default when generating code for
22899 SH4A. Using @option{-mno-fsrra} disables reciprocal square root approximations
22900 even if @option{-funsafe-math-optimizations} and @option{-ffinite-math-only} are
22903 @item -mpretend-cmove
22904 @opindex mpretend-cmove
22905 Prefer zero-displacement conditional branches for conditional move instruction
22906 patterns. This can result in faster code on the SH4 processor.
22910 Generate code using the FDPIC ABI.
22914 @node Solaris 2 Options
22915 @subsection Solaris 2 Options
22916 @cindex Solaris 2 options
22918 These @samp{-m} options are supported on Solaris 2:
22921 @item -mclear-hwcap
22922 @opindex mclear-hwcap
22923 @option{-mclear-hwcap} tells the compiler to remove the hardware
22924 capabilities generated by the Solaris assembler. This is only necessary
22925 when object files use ISA extensions not supported by the current
22926 machine, but check at runtime whether or not to use them.
22928 @item -mimpure-text
22929 @opindex mimpure-text
22930 @option{-mimpure-text}, used in addition to @option{-shared}, tells
22931 the compiler to not pass @option{-z text} to the linker when linking a
22932 shared object. Using this option, you can link position-dependent
22933 code into a shared object.
22935 @option{-mimpure-text} suppresses the ``relocations remain against
22936 allocatable but non-writable sections'' linker error message.
22937 However, the necessary relocations trigger copy-on-write, and the
22938 shared object is not actually shared across processes. Instead of
22939 using @option{-mimpure-text}, you should compile all source code with
22940 @option{-fpic} or @option{-fPIC}.
22944 These switches are supported in addition to the above on Solaris 2:
22949 Add support for multithreading using the POSIX threads library. This
22950 option sets flags for both the preprocessor and linker. This option does
22951 not affect the thread safety of object code produced by the compiler or
22952 that of libraries supplied with it.
22956 This is a synonym for @option{-pthreads}.
22959 @node SPARC Options
22960 @subsection SPARC Options
22961 @cindex SPARC options
22963 These @samp{-m} options are supported on the SPARC:
22966 @item -mno-app-regs
22968 @opindex mno-app-regs
22970 Specify @option{-mapp-regs} to generate output using the global registers
22971 2 through 4, which the SPARC SVR4 ABI reserves for applications. Like the
22972 global register 1, each global register 2 through 4 is then treated as an
22973 allocable register that is clobbered by function calls. This is the default.
22975 To be fully SVR4 ABI-compliant at the cost of some performance loss,
22976 specify @option{-mno-app-regs}. You should compile libraries and system
22977 software with this option.
22983 With @option{-mflat}, the compiler does not generate save/restore instructions
22984 and uses a ``flat'' or single register window model. This model is compatible
22985 with the regular register window model. The local registers and the input
22986 registers (0--5) are still treated as ``call-saved'' registers and are
22987 saved on the stack as needed.
22989 With @option{-mno-flat} (the default), the compiler generates save/restore
22990 instructions (except for leaf functions). This is the normal operating mode.
22993 @itemx -mhard-float
22995 @opindex mhard-float
22996 Generate output containing floating-point instructions. This is the
23000 @itemx -msoft-float
23002 @opindex msoft-float
23003 Generate output containing library calls for floating point.
23004 @strong{Warning:} the requisite libraries are not available for all SPARC
23005 targets. Normally the facilities of the machine's usual C compiler are
23006 used, but this cannot be done directly in cross-compilation. You must make
23007 your own arrangements to provide suitable library functions for
23008 cross-compilation. The embedded targets @samp{sparc-*-aout} and
23009 @samp{sparclite-*-*} do provide software floating-point support.
23011 @option{-msoft-float} changes the calling convention in the output file;
23012 therefore, it is only useful if you compile @emph{all} of a program with
23013 this option. In particular, you need to compile @file{libgcc.a}, the
23014 library that comes with GCC, with @option{-msoft-float} in order for
23017 @item -mhard-quad-float
23018 @opindex mhard-quad-float
23019 Generate output containing quad-word (long double) floating-point
23022 @item -msoft-quad-float
23023 @opindex msoft-quad-float
23024 Generate output containing library calls for quad-word (long double)
23025 floating-point instructions. The functions called are those specified
23026 in the SPARC ABI@. This is the default.
23028 As of this writing, there are no SPARC implementations that have hardware
23029 support for the quad-word floating-point instructions. They all invoke
23030 a trap handler for one of these instructions, and then the trap handler
23031 emulates the effect of the instruction. Because of the trap handler overhead,
23032 this is much slower than calling the ABI library routines. Thus the
23033 @option{-msoft-quad-float} option is the default.
23035 @item -mno-unaligned-doubles
23036 @itemx -munaligned-doubles
23037 @opindex mno-unaligned-doubles
23038 @opindex munaligned-doubles
23039 Assume that doubles have 8-byte alignment. This is the default.
23041 With @option{-munaligned-doubles}, GCC assumes that doubles have 8-byte
23042 alignment only if they are contained in another type, or if they have an
23043 absolute address. Otherwise, it assumes they have 4-byte alignment.
23044 Specifying this option avoids some rare compatibility problems with code
23045 generated by other compilers. It is not the default because it results
23046 in a performance loss, especially for floating-point code.
23049 @itemx -mno-user-mode
23050 @opindex muser-mode
23051 @opindex mno-user-mode
23052 Do not generate code that can only run in supervisor mode. This is relevant
23053 only for the @code{casa} instruction emitted for the LEON3 processor. This
23056 @item -mfaster-structs
23057 @itemx -mno-faster-structs
23058 @opindex mfaster-structs
23059 @opindex mno-faster-structs
23060 With @option{-mfaster-structs}, the compiler assumes that structures
23061 should have 8-byte alignment. This enables the use of pairs of
23062 @code{ldd} and @code{std} instructions for copies in structure
23063 assignment, in place of twice as many @code{ld} and @code{st} pairs.
23064 However, the use of this changed alignment directly violates the SPARC
23065 ABI@. Thus, it's intended only for use on targets where the developer
23066 acknowledges that their resulting code is not directly in line with
23067 the rules of the ABI@.
23069 @item -mstd-struct-return
23070 @itemx -mno-std-struct-return
23071 @opindex mstd-struct-return
23072 @opindex mno-std-struct-return
23073 With @option{-mstd-struct-return}, the compiler generates checking code
23074 in functions returning structures or unions to detect size mismatches
23075 between the two sides of function calls, as per the 32-bit ABI@.
23077 The default is @option{-mno-std-struct-return}. This option has no effect
23080 @item -mcpu=@var{cpu_type}
23082 Set the instruction set, register set, and instruction scheduling parameters
23083 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
23084 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{hypersparc},
23085 @samp{leon}, @samp{leon3}, @samp{leon3v7}, @samp{sparclite}, @samp{f930},
23086 @samp{f934}, @samp{sparclite86x}, @samp{sparclet}, @samp{tsc701}, @samp{v9},
23087 @samp{ultrasparc}, @samp{ultrasparc3}, @samp{niagara}, @samp{niagara2},
23088 @samp{niagara3}, @samp{niagara4} and @samp{niagara7}.
23090 Native Solaris and GNU/Linux toolchains also support the value @samp{native},
23091 which selects the best architecture option for the host processor.
23092 @option{-mcpu=native} has no effect if GCC does not recognize
23095 Default instruction scheduling parameters are used for values that select
23096 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
23097 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
23099 Here is a list of each supported architecture and their supported
23107 supersparc, hypersparc, leon, leon3
23110 f930, f934, sparclite86x
23116 ultrasparc, ultrasparc3, niagara, niagara2, niagara3, niagara4, niagara7
23119 By default (unless configured otherwise), GCC generates code for the V7
23120 variant of the SPARC architecture. With @option{-mcpu=cypress}, the compiler
23121 additionally optimizes it for the Cypress CY7C602 chip, as used in the
23122 SPARCStation/SPARCServer 3xx series. This is also appropriate for the older
23123 SPARCStation 1, 2, IPX etc.
23125 With @option{-mcpu=v8}, GCC generates code for the V8 variant of the SPARC
23126 architecture. The only difference from V7 code is that the compiler emits
23127 the integer multiply and integer divide instructions which exist in SPARC-V8
23128 but not in SPARC-V7. With @option{-mcpu=supersparc}, the compiler additionally
23129 optimizes it for the SuperSPARC chip, as used in the SPARCStation 10, 1000 and
23132 With @option{-mcpu=sparclite}, GCC generates code for the SPARClite variant of
23133 the SPARC architecture. This adds the integer multiply, integer divide step
23134 and scan (@code{ffs}) instructions which exist in SPARClite but not in SPARC-V7.
23135 With @option{-mcpu=f930}, the compiler additionally optimizes it for the
23136 Fujitsu MB86930 chip, which is the original SPARClite, with no FPU@. With
23137 @option{-mcpu=f934}, the compiler additionally optimizes it for the Fujitsu
23138 MB86934 chip, which is the more recent SPARClite with FPU@.
23140 With @option{-mcpu=sparclet}, GCC generates code for the SPARClet variant of
23141 the SPARC architecture. This adds the integer multiply, multiply/accumulate,
23142 integer divide step and scan (@code{ffs}) instructions which exist in SPARClet
23143 but not in SPARC-V7. With @option{-mcpu=tsc701}, the compiler additionally
23144 optimizes it for the TEMIC SPARClet chip.
23146 With @option{-mcpu=v9}, GCC generates code for the V9 variant of the SPARC
23147 architecture. This adds 64-bit integer and floating-point move instructions,
23148 3 additional floating-point condition code registers and conditional move
23149 instructions. With @option{-mcpu=ultrasparc}, the compiler additionally
23150 optimizes it for the Sun UltraSPARC I/II/IIi chips. With
23151 @option{-mcpu=ultrasparc3}, the compiler additionally optimizes it for the
23152 Sun UltraSPARC III/III+/IIIi/IIIi+/IV/IV+ chips. With
23153 @option{-mcpu=niagara}, the compiler additionally optimizes it for
23154 Sun UltraSPARC T1 chips. With @option{-mcpu=niagara2}, the compiler
23155 additionally optimizes it for Sun UltraSPARC T2 chips. With
23156 @option{-mcpu=niagara3}, the compiler additionally optimizes it for Sun
23157 UltraSPARC T3 chips. With @option{-mcpu=niagara4}, the compiler
23158 additionally optimizes it for Sun UltraSPARC T4 chips. With
23159 @option{-mcpu=niagara7}, the compiler additionally optimizes it for
23160 Oracle SPARC M7 chips.
23162 @item -mtune=@var{cpu_type}
23164 Set the instruction scheduling parameters for machine type
23165 @var{cpu_type}, but do not set the instruction set or register set that the
23166 option @option{-mcpu=@var{cpu_type}} does.
23168 The same values for @option{-mcpu=@var{cpu_type}} can be used for
23169 @option{-mtune=@var{cpu_type}}, but the only useful values are those
23170 that select a particular CPU implementation. Those are
23171 @samp{cypress}, @samp{supersparc}, @samp{hypersparc}, @samp{leon},
23172 @samp{leon3}, @samp{leon3v7}, @samp{f930}, @samp{f934},
23173 @samp{sparclite86x}, @samp{tsc701}, @samp{ultrasparc},
23174 @samp{ultrasparc3}, @samp{niagara}, @samp{niagara2}, @samp{niagara3},
23175 @samp{niagara4} and @samp{niagara7}. With native Solaris and
23176 GNU/Linux toolchains, @samp{native} can also be used.
23181 @opindex mno-v8plus
23182 With @option{-mv8plus}, GCC generates code for the SPARC-V8+ ABI@. The
23183 difference from the V8 ABI is that the global and out registers are
23184 considered 64 bits wide. This is enabled by default on Solaris in 32-bit
23185 mode for all SPARC-V9 processors.
23191 With @option{-mvis}, GCC generates code that takes advantage of the UltraSPARC
23192 Visual Instruction Set extensions. The default is @option{-mno-vis}.
23198 With @option{-mvis2}, GCC generates code that takes advantage of
23199 version 2.0 of the UltraSPARC Visual Instruction Set extensions. The
23200 default is @option{-mvis2} when targeting a cpu that supports such
23201 instructions, such as UltraSPARC-III and later. Setting @option{-mvis2}
23202 also sets @option{-mvis}.
23208 With @option{-mvis3}, GCC generates code that takes advantage of
23209 version 3.0 of the UltraSPARC Visual Instruction Set extensions. The
23210 default is @option{-mvis3} when targeting a cpu that supports such
23211 instructions, such as niagara-3 and later. Setting @option{-mvis3}
23212 also sets @option{-mvis2} and @option{-mvis}.
23218 With @option{-mvis4}, GCC generates code that takes advantage of
23219 version 4.0 of the UltraSPARC Visual Instruction Set extensions. The
23220 default is @option{-mvis4} when targeting a cpu that supports such
23221 instructions, such as niagara-7 and later. Setting @option{-mvis4}
23222 also sets @option{-mvis3}, @option{-mvis2} and @option{-mvis}.
23227 @opindex mno-cbcond
23228 With @option{-mcbcond}, GCC generates code that takes advantage of the UltraSPARC
23229 Compare-and-Branch-on-Condition instructions. The default is @option{-mcbcond}
23230 when targeting a CPU that supports such instructions, such as Niagara-4 and
23237 With @option{-mfmaf}, GCC generates code that takes advantage of the UltraSPARC
23238 Fused Multiply-Add Floating-point instructions. The default is @option{-mfmaf}
23239 when targeting a CPU that supports such instructions, such as Niagara-3 and
23246 With @option{-mpopc}, GCC generates code that takes advantage of the UltraSPARC
23247 Population Count instruction. The default is @option{-mpopc}
23248 when targeting a CPU that supports such an instruction, such as Niagara-2 and
23255 With @option{-msubxc}, GCC generates code that takes advantage of the UltraSPARC
23256 Subtract-Extended-with-Carry instruction. The default is @option{-msubxc}
23257 when targeting a CPU that supports such an instruction, such as Niagara-7 and
23261 @opindex mfix-at697f
23262 Enable the documented workaround for the single erratum of the Atmel AT697F
23263 processor (which corresponds to erratum #13 of the AT697E processor).
23266 @opindex mfix-ut699
23267 Enable the documented workarounds for the floating-point errata and the data
23268 cache nullify errata of the UT699 processor.
23271 These @samp{-m} options are supported in addition to the above
23272 on SPARC-V9 processors in 64-bit environments:
23279 Generate code for a 32-bit or 64-bit environment.
23280 The 32-bit environment sets int, long and pointer to 32 bits.
23281 The 64-bit environment sets int to 32 bits and long and pointer
23284 @item -mcmodel=@var{which}
23286 Set the code model to one of
23290 The Medium/Low code model: 64-bit addresses, programs
23291 must be linked in the low 32 bits of memory. Programs can be statically
23292 or dynamically linked.
23295 The Medium/Middle code model: 64-bit addresses, programs
23296 must be linked in the low 44 bits of memory, the text and data segments must
23297 be less than 2GB in size and the data segment must be located within 2GB of
23301 The Medium/Anywhere code model: 64-bit addresses, programs
23302 may be linked anywhere in memory, the text and data segments must be less
23303 than 2GB in size and the data segment must be located within 2GB of the
23307 The Medium/Anywhere code model for embedded systems:
23308 64-bit addresses, the text and data segments must be less than 2GB in
23309 size, both starting anywhere in memory (determined at link time). The
23310 global register %g4 points to the base of the data segment. Programs
23311 are statically linked and PIC is not supported.
23314 @item -mmemory-model=@var{mem-model}
23315 @opindex mmemory-model
23316 Set the memory model in force on the processor to one of
23320 The default memory model for the processor and operating system.
23323 Relaxed Memory Order
23326 Partial Store Order
23332 Sequential Consistency
23335 These memory models are formally defined in Appendix D of the Sparc V9
23336 architecture manual, as set in the processor's @code{PSTATE.MM} field.
23339 @itemx -mno-stack-bias
23340 @opindex mstack-bias
23341 @opindex mno-stack-bias
23342 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
23343 frame pointer if present, are offset by @minus{}2047 which must be added back
23344 when making stack frame references. This is the default in 64-bit mode.
23345 Otherwise, assume no such offset is present.
23349 @subsection SPU Options
23350 @cindex SPU options
23352 These @samp{-m} options are supported on the SPU:
23356 @itemx -merror-reloc
23357 @opindex mwarn-reloc
23358 @opindex merror-reloc
23360 The loader for SPU does not handle dynamic relocations. By default, GCC
23361 gives an error when it generates code that requires a dynamic
23362 relocation. @option{-mno-error-reloc} disables the error,
23363 @option{-mwarn-reloc} generates a warning instead.
23366 @itemx -munsafe-dma
23368 @opindex munsafe-dma
23370 Instructions that initiate or test completion of DMA must not be
23371 reordered with respect to loads and stores of the memory that is being
23373 With @option{-munsafe-dma} you must use the @code{volatile} keyword to protect
23374 memory accesses, but that can lead to inefficient code in places where the
23375 memory is known to not change. Rather than mark the memory as volatile,
23376 you can use @option{-msafe-dma} to tell the compiler to treat
23377 the DMA instructions as potentially affecting all memory.
23379 @item -mbranch-hints
23380 @opindex mbranch-hints
23382 By default, GCC generates a branch hint instruction to avoid
23383 pipeline stalls for always-taken or probably-taken branches. A hint
23384 is not generated closer than 8 instructions away from its branch.
23385 There is little reason to disable them, except for debugging purposes,
23386 or to make an object a little bit smaller.
23390 @opindex msmall-mem
23391 @opindex mlarge-mem
23393 By default, GCC generates code assuming that addresses are never larger
23394 than 18 bits. With @option{-mlarge-mem} code is generated that assumes
23395 a full 32-bit address.
23400 By default, GCC links against startup code that assumes the SPU-style
23401 main function interface (which has an unconventional parameter list).
23402 With @option{-mstdmain}, GCC links your program against startup
23403 code that assumes a C99-style interface to @code{main}, including a
23404 local copy of @code{argv} strings.
23406 @item -mfixed-range=@var{register-range}
23407 @opindex mfixed-range
23408 Generate code treating the given register range as fixed registers.
23409 A fixed register is one that the register allocator cannot use. This is
23410 useful when compiling kernel code. A register range is specified as
23411 two registers separated by a dash. Multiple register ranges can be
23412 specified separated by a comma.
23418 Compile code assuming that pointers to the PPU address space accessed
23419 via the @code{__ea} named address space qualifier are either 32 or 64
23420 bits wide. The default is 32 bits. As this is an ABI-changing option,
23421 all object code in an executable must be compiled with the same setting.
23423 @item -maddress-space-conversion
23424 @itemx -mno-address-space-conversion
23425 @opindex maddress-space-conversion
23426 @opindex mno-address-space-conversion
23427 Allow/disallow treating the @code{__ea} address space as superset
23428 of the generic address space. This enables explicit type casts
23429 between @code{__ea} and generic pointer as well as implicit
23430 conversions of generic pointers to @code{__ea} pointers. The
23431 default is to allow address space pointer conversions.
23433 @item -mcache-size=@var{cache-size}
23434 @opindex mcache-size
23435 This option controls the version of libgcc that the compiler links to an
23436 executable and selects a software-managed cache for accessing variables
23437 in the @code{__ea} address space with a particular cache size. Possible
23438 options for @var{cache-size} are @samp{8}, @samp{16}, @samp{32}, @samp{64}
23439 and @samp{128}. The default cache size is 64KB.
23441 @item -matomic-updates
23442 @itemx -mno-atomic-updates
23443 @opindex matomic-updates
23444 @opindex mno-atomic-updates
23445 This option controls the version of libgcc that the compiler links to an
23446 executable and selects whether atomic updates to the software-managed
23447 cache of PPU-side variables are used. If you use atomic updates, changes
23448 to a PPU variable from SPU code using the @code{__ea} named address space
23449 qualifier do not interfere with changes to other PPU variables residing
23450 in the same cache line from PPU code. If you do not use atomic updates,
23451 such interference may occur; however, writing back cache lines is
23452 more efficient. The default behavior is to use atomic updates.
23455 @itemx -mdual-nops=@var{n}
23456 @opindex mdual-nops
23457 By default, GCC inserts NOPs to increase dual issue when it expects
23458 it to increase performance. @var{n} can be a value from 0 to 10. A
23459 smaller @var{n} inserts fewer NOPs. 10 is the default, 0 is the
23460 same as @option{-mno-dual-nops}. Disabled with @option{-Os}.
23462 @item -mhint-max-nops=@var{n}
23463 @opindex mhint-max-nops
23464 Maximum number of NOPs to insert for a branch hint. A branch hint must
23465 be at least 8 instructions away from the branch it is affecting. GCC
23466 inserts up to @var{n} NOPs to enforce this, otherwise it does not
23467 generate the branch hint.
23469 @item -mhint-max-distance=@var{n}
23470 @opindex mhint-max-distance
23471 The encoding of the branch hint instruction limits the hint to be within
23472 256 instructions of the branch it is affecting. By default, GCC makes
23473 sure it is within 125.
23476 @opindex msafe-hints
23477 Work around a hardware bug that causes the SPU to stall indefinitely.
23478 By default, GCC inserts the @code{hbrp} instruction to make sure
23479 this stall won't happen.
23483 @node System V Options
23484 @subsection Options for System V
23486 These additional options are available on System V Release 4 for
23487 compatibility with other compilers on those systems:
23492 Create a shared object.
23493 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
23497 Identify the versions of each tool used by the compiler, in a
23498 @code{.ident} assembler directive in the output.
23502 Refrain from adding @code{.ident} directives to the output file (this is
23505 @item -YP,@var{dirs}
23507 Search the directories @var{dirs}, and no others, for libraries
23508 specified with @option{-l}.
23510 @item -Ym,@var{dir}
23512 Look in the directory @var{dir} to find the M4 preprocessor.
23513 The assembler uses this option.
23514 @c This is supposed to go with a -Yd for predefined M4 macro files, but
23515 @c the generic assembler that comes with Solaris takes just -Ym.
23518 @node TILE-Gx Options
23519 @subsection TILE-Gx Options
23520 @cindex TILE-Gx options
23522 These @samp{-m} options are supported on the TILE-Gx:
23525 @item -mcmodel=small
23526 @opindex mcmodel=small
23527 Generate code for the small model. The distance for direct calls is
23528 limited to 500M in either direction. PC-relative addresses are 32
23529 bits. Absolute addresses support the full address range.
23531 @item -mcmodel=large
23532 @opindex mcmodel=large
23533 Generate code for the large model. There is no limitation on call
23534 distance, pc-relative addresses, or absolute addresses.
23536 @item -mcpu=@var{name}
23538 Selects the type of CPU to be targeted. Currently the only supported
23539 type is @samp{tilegx}.
23545 Generate code for a 32-bit or 64-bit environment. The 32-bit
23546 environment sets int, long, and pointer to 32 bits. The 64-bit
23547 environment sets int to 32 bits and long and pointer to 64 bits.
23550 @itemx -mlittle-endian
23551 @opindex mbig-endian
23552 @opindex mlittle-endian
23553 Generate code in big/little endian mode, respectively.
23556 @node TILEPro Options
23557 @subsection TILEPro Options
23558 @cindex TILEPro options
23560 These @samp{-m} options are supported on the TILEPro:
23563 @item -mcpu=@var{name}
23565 Selects the type of CPU to be targeted. Currently the only supported
23566 type is @samp{tilepro}.
23570 Generate code for a 32-bit environment, which sets int, long, and
23571 pointer to 32 bits. This is the only supported behavior so the flag
23572 is essentially ignored.
23576 @subsection V850 Options
23577 @cindex V850 Options
23579 These @samp{-m} options are defined for V850 implementations:
23583 @itemx -mno-long-calls
23584 @opindex mlong-calls
23585 @opindex mno-long-calls
23586 Treat all calls as being far away (near). If calls are assumed to be
23587 far away, the compiler always loads the function's address into a
23588 register, and calls indirect through the pointer.
23594 Do not optimize (do optimize) basic blocks that use the same index
23595 pointer 4 or more times to copy pointer into the @code{ep} register, and
23596 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
23597 option is on by default if you optimize.
23599 @item -mno-prolog-function
23600 @itemx -mprolog-function
23601 @opindex mno-prolog-function
23602 @opindex mprolog-function
23603 Do not use (do use) external functions to save and restore registers
23604 at the prologue and epilogue of a function. The external functions
23605 are slower, but use less code space if more than one function saves
23606 the same number of registers. The @option{-mprolog-function} option
23607 is on by default if you optimize.
23611 Try to make the code as small as possible. At present, this just turns
23612 on the @option{-mep} and @option{-mprolog-function} options.
23614 @item -mtda=@var{n}
23616 Put static or global variables whose size is @var{n} bytes or less into
23617 the tiny data area that register @code{ep} points to. The tiny data
23618 area can hold up to 256 bytes in total (128 bytes for byte references).
23620 @item -msda=@var{n}
23622 Put static or global variables whose size is @var{n} bytes or less into
23623 the small data area that register @code{gp} points to. The small data
23624 area can hold up to 64 kilobytes.
23626 @item -mzda=@var{n}
23628 Put static or global variables whose size is @var{n} bytes or less into
23629 the first 32 kilobytes of memory.
23633 Specify that the target processor is the V850.
23637 Specify that the target processor is the V850E3V5. The preprocessor
23638 constant @code{__v850e3v5__} is defined if this option is used.
23642 Specify that the target processor is the V850E3V5. This is an alias for
23643 the @option{-mv850e3v5} option.
23647 Specify that the target processor is the V850E2V3. The preprocessor
23648 constant @code{__v850e2v3__} is defined if this option is used.
23652 Specify that the target processor is the V850E2. The preprocessor
23653 constant @code{__v850e2__} is defined if this option is used.
23657 Specify that the target processor is the V850E1. The preprocessor
23658 constants @code{__v850e1__} and @code{__v850e__} are defined if
23659 this option is used.
23663 Specify that the target processor is the V850ES. This is an alias for
23664 the @option{-mv850e1} option.
23668 Specify that the target processor is the V850E@. The preprocessor
23669 constant @code{__v850e__} is defined if this option is used.
23671 If neither @option{-mv850} nor @option{-mv850e} nor @option{-mv850e1}
23672 nor @option{-mv850e2} nor @option{-mv850e2v3} nor @option{-mv850e3v5}
23673 are defined then a default target processor is chosen and the
23674 relevant @samp{__v850*__} preprocessor constant is defined.
23676 The preprocessor constants @code{__v850} and @code{__v851__} are always
23677 defined, regardless of which processor variant is the target.
23679 @item -mdisable-callt
23680 @itemx -mno-disable-callt
23681 @opindex mdisable-callt
23682 @opindex mno-disable-callt
23683 This option suppresses generation of the @code{CALLT} instruction for the
23684 v850e, v850e1, v850e2, v850e2v3 and v850e3v5 flavors of the v850
23687 This option is enabled by default when the RH850 ABI is
23688 in use (see @option{-mrh850-abi}), and disabled by default when the
23689 GCC ABI is in use. If @code{CALLT} instructions are being generated
23690 then the C preprocessor symbol @code{__V850_CALLT__} is defined.
23696 Pass on (or do not pass on) the @option{-mrelax} command-line option
23700 @itemx -mno-long-jumps
23701 @opindex mlong-jumps
23702 @opindex mno-long-jumps
23703 Disable (or re-enable) the generation of PC-relative jump instructions.
23706 @itemx -mhard-float
23707 @opindex msoft-float
23708 @opindex mhard-float
23709 Disable (or re-enable) the generation of hardware floating point
23710 instructions. This option is only significant when the target
23711 architecture is @samp{V850E2V3} or higher. If hardware floating point
23712 instructions are being generated then the C preprocessor symbol
23713 @code{__FPU_OK__} is defined, otherwise the symbol
23714 @code{__NO_FPU__} is defined.
23718 Enables the use of the e3v5 LOOP instruction. The use of this
23719 instruction is not enabled by default when the e3v5 architecture is
23720 selected because its use is still experimental.
23724 @opindex mrh850-abi
23726 Enables support for the RH850 version of the V850 ABI. This is the
23727 default. With this version of the ABI the following rules apply:
23731 Integer sized structures and unions are returned via a memory pointer
23732 rather than a register.
23735 Large structures and unions (more than 8 bytes in size) are passed by
23739 Functions are aligned to 16-bit boundaries.
23742 The @option{-m8byte-align} command-line option is supported.
23745 The @option{-mdisable-callt} command-line option is enabled by
23746 default. The @option{-mno-disable-callt} command-line option is not
23750 When this version of the ABI is enabled the C preprocessor symbol
23751 @code{__V850_RH850_ABI__} is defined.
23755 Enables support for the old GCC version of the V850 ABI. With this
23756 version of the ABI the following rules apply:
23760 Integer sized structures and unions are returned in register @code{r10}.
23763 Large structures and unions (more than 8 bytes in size) are passed by
23767 Functions are aligned to 32-bit boundaries, unless optimizing for
23771 The @option{-m8byte-align} command-line option is not supported.
23774 The @option{-mdisable-callt} command-line option is supported but not
23775 enabled by default.
23778 When this version of the ABI is enabled the C preprocessor symbol
23779 @code{__V850_GCC_ABI__} is defined.
23781 @item -m8byte-align
23782 @itemx -mno-8byte-align
23783 @opindex m8byte-align
23784 @opindex mno-8byte-align
23785 Enables support for @code{double} and @code{long long} types to be
23786 aligned on 8-byte boundaries. The default is to restrict the
23787 alignment of all objects to at most 4-bytes. When
23788 @option{-m8byte-align} is in effect the C preprocessor symbol
23789 @code{__V850_8BYTE_ALIGN__} is defined.
23792 @opindex mbig-switch
23793 Generate code suitable for big switch tables. Use this option only if
23794 the assembler/linker complain about out of range branches within a switch
23799 This option causes r2 and r5 to be used in the code generated by
23800 the compiler. This setting is the default.
23802 @item -mno-app-regs
23803 @opindex mno-app-regs
23804 This option causes r2 and r5 to be treated as fixed registers.
23809 @subsection VAX Options
23810 @cindex VAX options
23812 These @samp{-m} options are defined for the VAX:
23817 Do not output certain jump instructions (@code{aobleq} and so on)
23818 that the Unix assembler for the VAX cannot handle across long
23823 Do output those jump instructions, on the assumption that the
23824 GNU assembler is being used.
23828 Output code for G-format floating-point numbers instead of D-format.
23831 @node Visium Options
23832 @subsection Visium Options
23833 @cindex Visium options
23839 A program which performs file I/O and is destined to run on an MCM target
23840 should be linked with this option. It causes the libraries libc.a and
23841 libdebug.a to be linked. The program should be run on the target under
23842 the control of the GDB remote debugging stub.
23846 A program which performs file I/O and is destined to run on the simulator
23847 should be linked with option. This causes libraries libc.a and libsim.a to
23851 @itemx -mhard-float
23853 @opindex mhard-float
23854 Generate code containing floating-point instructions. This is the
23858 @itemx -msoft-float
23860 @opindex msoft-float
23861 Generate code containing library calls for floating-point.
23863 @option{-msoft-float} changes the calling convention in the output file;
23864 therefore, it is only useful if you compile @emph{all} of a program with
23865 this option. In particular, you need to compile @file{libgcc.a}, the
23866 library that comes with GCC, with @option{-msoft-float} in order for
23869 @item -mcpu=@var{cpu_type}
23871 Set the instruction set, register set, and instruction scheduling parameters
23872 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
23873 @samp{mcm}, @samp{gr5} and @samp{gr6}.
23875 @samp{mcm} is a synonym of @samp{gr5} present for backward compatibility.
23877 By default (unless configured otherwise), GCC generates code for the GR5
23878 variant of the Visium architecture.
23880 With @option{-mcpu=gr6}, GCC generates code for the GR6 variant of the Visium
23881 architecture. The only difference from GR5 code is that the compiler will
23882 generate block move instructions.
23884 @item -mtune=@var{cpu_type}
23886 Set the instruction scheduling parameters for machine type @var{cpu_type},
23887 but do not set the instruction set or register set that the option
23888 @option{-mcpu=@var{cpu_type}} would.
23892 Generate code for the supervisor mode, where there are no restrictions on
23893 the access to general registers. This is the default.
23896 @opindex muser-mode
23897 Generate code for the user mode, where the access to some general registers
23898 is forbidden: on the GR5, registers r24 to r31 cannot be accessed in this
23899 mode; on the GR6, only registers r29 to r31 are affected.
23903 @subsection VMS Options
23905 These @samp{-m} options are defined for the VMS implementations:
23908 @item -mvms-return-codes
23909 @opindex mvms-return-codes
23910 Return VMS condition codes from @code{main}. The default is to return POSIX-style
23911 condition (e.g.@ error) codes.
23913 @item -mdebug-main=@var{prefix}
23914 @opindex mdebug-main=@var{prefix}
23915 Flag the first routine whose name starts with @var{prefix} as the main
23916 routine for the debugger.
23920 Default to 64-bit memory allocation routines.
23922 @item -mpointer-size=@var{size}
23923 @opindex mpointer-size=@var{size}
23924 Set the default size of pointers. Possible options for @var{size} are
23925 @samp{32} or @samp{short} for 32 bit pointers, @samp{64} or @samp{long}
23926 for 64 bit pointers, and @samp{no} for supporting only 32 bit pointers.
23927 The later option disables @code{pragma pointer_size}.
23930 @node VxWorks Options
23931 @subsection VxWorks Options
23932 @cindex VxWorks Options
23934 The options in this section are defined for all VxWorks targets.
23935 Options specific to the target hardware are listed with the other
23936 options for that target.
23941 GCC can generate code for both VxWorks kernels and real time processes
23942 (RTPs). This option switches from the former to the latter. It also
23943 defines the preprocessor macro @code{__RTP__}.
23946 @opindex non-static
23947 Link an RTP executable against shared libraries rather than static
23948 libraries. The options @option{-static} and @option{-shared} can
23949 also be used for RTPs (@pxref{Link Options}); @option{-static}
23956 These options are passed down to the linker. They are defined for
23957 compatibility with Diab.
23960 @opindex Xbind-lazy
23961 Enable lazy binding of function calls. This option is equivalent to
23962 @option{-Wl,-z,now} and is defined for compatibility with Diab.
23966 Disable lazy binding of function calls. This option is the default and
23967 is defined for compatibility with Diab.
23971 @subsection x86 Options
23972 @cindex x86 Options
23974 These @samp{-m} options are defined for the x86 family of computers.
23978 @item -march=@var{cpu-type}
23980 Generate instructions for the machine type @var{cpu-type}. In contrast to
23981 @option{-mtune=@var{cpu-type}}, which merely tunes the generated code
23982 for the specified @var{cpu-type}, @option{-march=@var{cpu-type}} allows GCC
23983 to generate code that may not run at all on processors other than the one
23984 indicated. Specifying @option{-march=@var{cpu-type}} implies
23985 @option{-mtune=@var{cpu-type}}.
23987 The choices for @var{cpu-type} are:
23991 This selects the CPU to generate code for at compilation time by determining
23992 the processor type of the compiling machine. Using @option{-march=native}
23993 enables all instruction subsets supported by the local machine (hence
23994 the result might not run on different machines). Using @option{-mtune=native}
23995 produces code optimized for the local machine under the constraints
23996 of the selected instruction set.
23999 Original Intel i386 CPU@.
24002 Intel i486 CPU@. (No scheduling is implemented for this chip.)
24006 Intel Pentium CPU with no MMX support.
24009 Intel Lakemont MCU, based on Intel Pentium CPU.
24012 Intel Pentium MMX CPU, based on Pentium core with MMX instruction set support.
24015 Intel Pentium Pro CPU@.
24018 When used with @option{-march}, the Pentium Pro
24019 instruction set is used, so the code runs on all i686 family chips.
24020 When used with @option{-mtune}, it has the same meaning as @samp{generic}.
24023 Intel Pentium II CPU, based on Pentium Pro core with MMX instruction set
24028 Intel Pentium III CPU, based on Pentium Pro core with MMX and SSE instruction
24032 Intel Pentium M; low-power version of Intel Pentium III CPU
24033 with MMX, SSE and SSE2 instruction set support. Used by Centrino notebooks.
24037 Intel Pentium 4 CPU with MMX, SSE and SSE2 instruction set support.
24040 Improved version of Intel Pentium 4 CPU with MMX, SSE, SSE2 and SSE3 instruction
24044 Improved version of Intel Pentium 4 CPU with 64-bit extensions, MMX, SSE,
24045 SSE2 and SSE3 instruction set support.
24048 Intel Core 2 CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3 and SSSE3
24049 instruction set support.
24052 Intel Nehalem CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3, SSSE3,
24053 SSE4.1, SSE4.2 and POPCNT instruction set support.
24056 Intel Westmere CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3, SSSE3,
24057 SSE4.1, SSE4.2, POPCNT, AES and PCLMUL instruction set support.
24060 Intel Sandy Bridge CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3, SSSE3,
24061 SSE4.1, SSE4.2, POPCNT, AVX, AES and PCLMUL instruction set support.
24064 Intel Ivy Bridge CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3, SSSE3,
24065 SSE4.1, SSE4.2, POPCNT, AVX, AES, PCLMUL, FSGSBASE, RDRND and F16C
24066 instruction set support.
24069 Intel Haswell CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3, SSSE3,
24070 SSE4.1, SSE4.2, POPCNT, AVX, AVX2, AES, PCLMUL, FSGSBASE, RDRND, FMA,
24071 BMI, BMI2 and F16C instruction set support.
24074 Intel Broadwell CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3, SSSE3,
24075 SSE4.1, SSE4.2, POPCNT, AVX, AVX2, AES, PCLMUL, FSGSBASE, RDRND, FMA,
24076 BMI, BMI2, F16C, RDSEED, ADCX and PREFETCHW instruction set support.
24079 Intel Skylake CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3, SSSE3,
24080 SSE4.1, SSE4.2, POPCNT, AVX, AVX2, AES, PCLMUL, FSGSBASE, RDRND, FMA,
24081 BMI, BMI2, F16C, RDSEED, ADCX, PREFETCHW, CLFLUSHOPT, XSAVEC and
24082 XSAVES instruction set support.
24085 Intel Bonnell CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3 and SSSE3
24086 instruction set support.
24089 Intel Silvermont CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3, SSSE3,
24090 SSE4.1, SSE4.2, POPCNT, AES, PCLMUL and RDRND instruction set support.
24093 Intel Knight's Landing CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3,
24094 SSSE3, SSE4.1, SSE4.2, POPCNT, AVX, AVX2, AES, PCLMUL, FSGSBASE, RDRND, FMA,
24095 BMI, BMI2, F16C, RDSEED, ADCX, PREFETCHW, AVX512F, AVX512PF, AVX512ER and
24096 AVX512CD instruction set support.
24098 @item skylake-avx512
24099 Intel Skylake Server CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3,
24100 SSSE3, SSE4.1, SSE4.2, POPCNT, PKU, AVX, AVX2, AES, PCLMUL, FSGSBASE, RDRND, FMA,
24101 BMI, BMI2, F16C, RDSEED, ADCX, PREFETCHW, CLFLUSHOPT, XSAVEC, XSAVES, AVX512F,
24102 AVX512VL, AVX512BW, AVX512DQ and AVX512CD instruction set support.
24105 AMD K6 CPU with MMX instruction set support.
24109 Improved versions of AMD K6 CPU with MMX and 3DNow!@: instruction set support.
24112 @itemx athlon-tbird
24113 AMD Athlon CPU with MMX, 3dNOW!, enhanced 3DNow!@: and SSE prefetch instructions
24119 Improved AMD Athlon CPU with MMX, 3DNow!, enhanced 3DNow!@: and full SSE
24120 instruction set support.
24126 Processors based on the AMD K8 core with x86-64 instruction set support,
24127 including the AMD Opteron, Athlon 64, and Athlon 64 FX processors.
24128 (This supersets MMX, SSE, SSE2, 3DNow!, enhanced 3DNow!@: and 64-bit
24129 instruction set extensions.)
24132 @itemx opteron-sse3
24133 @itemx athlon64-sse3
24134 Improved versions of AMD K8 cores with SSE3 instruction set support.
24138 CPUs based on AMD Family 10h cores with x86-64 instruction set support. (This
24139 supersets MMX, SSE, SSE2, SSE3, SSE4A, 3DNow!, enhanced 3DNow!, ABM and 64-bit
24140 instruction set extensions.)
24143 CPUs based on AMD Family 15h cores with x86-64 instruction set support. (This
24144 supersets FMA4, AVX, XOP, LWP, AES, PCL_MUL, CX16, MMX, SSE, SSE2, SSE3, SSE4A,
24145 SSSE3, SSE4.1, SSE4.2, ABM and 64-bit instruction set extensions.)
24147 AMD Family 15h core based CPUs with x86-64 instruction set support. (This
24148 supersets BMI, TBM, F16C, FMA, FMA4, AVX, XOP, LWP, AES, PCL_MUL, CX16, MMX,
24149 SSE, SSE2, SSE3, SSE4A, SSSE3, SSE4.1, SSE4.2, ABM and 64-bit instruction set
24152 AMD Family 15h core based CPUs with x86-64 instruction set support. (This
24153 supersets BMI, TBM, F16C, FMA, FMA4, FSGSBASE, AVX, XOP, LWP, AES,
24154 PCL_MUL, CX16, MMX, SSE, SSE2, SSE3, SSE4A, SSSE3, SSE4.1, SSE4.2, ABM and
24155 64-bit instruction set extensions.
24157 AMD Family 15h core based CPUs with x86-64 instruction set support. (This
24158 supersets BMI, BMI2, TBM, F16C, FMA, FMA4, FSGSBASE, AVX, AVX2, XOP, LWP,
24159 AES, PCL_MUL, CX16, MOVBE, MMX, SSE, SSE2, SSE3, SSE4A, SSSE3, SSE4.1,
24160 SSE4.2, ABM and 64-bit instruction set extensions.
24163 AMD Family 17h core based CPUs with x86-64 instruction set support. (This
24164 supersets BMI, BMI2, F16C, FMA, FSGSBASE, AVX, AVX2, ADCX, RDSEED, MWAITX,
24165 SHA, CLZERO, AES, PCL_MUL, CX16, MOVBE, MMX, SSE, SSE2, SSE3, SSE4A, SSSE3,
24166 SSE4.1, SSE4.2, ABM, XSAVEC, XSAVES, CLFLUSHOPT, POPCNT, and 64-bit
24167 instruction set extensions.
24170 CPUs based on AMD Family 14h cores with x86-64 instruction set support. (This
24171 supersets MMX, SSE, SSE2, SSE3, SSSE3, SSE4A, CX16, ABM and 64-bit
24172 instruction set extensions.)
24175 CPUs based on AMD Family 16h cores with x86-64 instruction set support. This
24176 includes MOVBE, F16C, BMI, AVX, PCL_MUL, AES, SSE4.2, SSE4.1, CX16, ABM,
24177 SSE4A, SSSE3, SSE3, SSE2, SSE, MMX and 64-bit instruction set extensions.
24180 IDT WinChip C6 CPU, dealt in same way as i486 with additional MMX instruction
24184 IDT WinChip 2 CPU, dealt in same way as i486 with additional MMX and 3DNow!@:
24185 instruction set support.
24188 VIA C3 CPU with MMX and 3DNow!@: instruction set support.
24189 (No scheduling is implemented for this chip.)
24192 VIA C3-2 (Nehemiah/C5XL) CPU with MMX and SSE instruction set support.
24193 (No scheduling is implemented for this chip.)
24196 VIA C7 (Esther) CPU with MMX, SSE, SSE2 and SSE3 instruction set support.
24197 (No scheduling is implemented for this chip.)
24200 VIA Eden Samuel 2 CPU with MMX and 3DNow!@: instruction set support.
24201 (No scheduling is implemented for this chip.)
24204 VIA Eden Nehemiah CPU with MMX and SSE instruction set support.
24205 (No scheduling is implemented for this chip.)
24208 VIA Eden Esther CPU with MMX, SSE, SSE2 and SSE3 instruction set support.
24209 (No scheduling is implemented for this chip.)
24212 VIA Eden X2 CPU with x86-64, MMX, SSE, SSE2 and SSE3 instruction set support.
24213 (No scheduling is implemented for this chip.)
24216 VIA Eden X4 CPU with x86-64, MMX, SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2,
24217 AVX and AVX2 instruction set support.
24218 (No scheduling is implemented for this chip.)
24221 Generic VIA Nano CPU with x86-64, MMX, SSE, SSE2, SSE3 and SSSE3
24222 instruction set support.
24223 (No scheduling is implemented for this chip.)
24226 VIA Nano 1xxx CPU with x86-64, MMX, SSE, SSE2, SSE3 and SSSE3
24227 instruction set support.
24228 (No scheduling is implemented for this chip.)
24231 VIA Nano 2xxx CPU with x86-64, MMX, SSE, SSE2, SSE3 and SSSE3
24232 instruction set support.
24233 (No scheduling is implemented for this chip.)
24236 VIA Nano 3xxx CPU with x86-64, MMX, SSE, SSE2, SSE3, SSSE3 and SSE4.1
24237 instruction set support.
24238 (No scheduling is implemented for this chip.)
24241 VIA Nano Dual Core CPU with x86-64, MMX, SSE, SSE2, SSE3, SSSE3 and SSE4.1
24242 instruction set support.
24243 (No scheduling is implemented for this chip.)
24246 VIA Nano Quad Core CPU with x86-64, MMX, SSE, SSE2, SSE3, SSSE3 and SSE4.1
24247 instruction set support.
24248 (No scheduling is implemented for this chip.)
24251 AMD Geode embedded processor with MMX and 3DNow!@: instruction set support.
24254 @item -mtune=@var{cpu-type}
24256 Tune to @var{cpu-type} everything applicable about the generated code, except
24257 for the ABI and the set of available instructions.
24258 While picking a specific @var{cpu-type} schedules things appropriately
24259 for that particular chip, the compiler does not generate any code that
24260 cannot run on the default machine type unless you use a
24261 @option{-march=@var{cpu-type}} option.
24262 For example, if GCC is configured for i686-pc-linux-gnu
24263 then @option{-mtune=pentium4} generates code that is tuned for Pentium 4
24264 but still runs on i686 machines.
24266 The choices for @var{cpu-type} are the same as for @option{-march}.
24267 In addition, @option{-mtune} supports 2 extra choices for @var{cpu-type}:
24271 Produce code optimized for the most common IA32/@/AMD64/@/EM64T processors.
24272 If you know the CPU on which your code will run, then you should use
24273 the corresponding @option{-mtune} or @option{-march} option instead of
24274 @option{-mtune=generic}. But, if you do not know exactly what CPU users
24275 of your application will have, then you should use this option.
24277 As new processors are deployed in the marketplace, the behavior of this
24278 option will change. Therefore, if you upgrade to a newer version of
24279 GCC, code generation controlled by this option will change to reflect
24281 that are most common at the time that version of GCC is released.
24283 There is no @option{-march=generic} option because @option{-march}
24284 indicates the instruction set the compiler can use, and there is no
24285 generic instruction set applicable to all processors. In contrast,
24286 @option{-mtune} indicates the processor (or, in this case, collection of
24287 processors) for which the code is optimized.
24290 Produce code optimized for the most current Intel processors, which are
24291 Haswell and Silvermont for this version of GCC. If you know the CPU
24292 on which your code will run, then you should use the corresponding
24293 @option{-mtune} or @option{-march} option instead of @option{-mtune=intel}.
24294 But, if you want your application performs better on both Haswell and
24295 Silvermont, then you should use this option.
24297 As new Intel processors are deployed in the marketplace, the behavior of
24298 this option will change. Therefore, if you upgrade to a newer version of
24299 GCC, code generation controlled by this option will change to reflect
24300 the most current Intel processors at the time that version of GCC is
24303 There is no @option{-march=intel} option because @option{-march} indicates
24304 the instruction set the compiler can use, and there is no common
24305 instruction set applicable to all processors. In contrast,
24306 @option{-mtune} indicates the processor (or, in this case, collection of
24307 processors) for which the code is optimized.
24310 @item -mcpu=@var{cpu-type}
24312 A deprecated synonym for @option{-mtune}.
24314 @item -mfpmath=@var{unit}
24316 Generate floating-point arithmetic for selected unit @var{unit}. The choices
24317 for @var{unit} are:
24321 Use the standard 387 floating-point coprocessor present on the majority of chips and
24322 emulated otherwise. Code compiled with this option runs almost everywhere.
24323 The temporary results are computed in 80-bit precision instead of the precision
24324 specified by the type, resulting in slightly different results compared to most
24325 of other chips. See @option{-ffloat-store} for more detailed description.
24327 This is the default choice for x86-32 targets.
24330 Use scalar floating-point instructions present in the SSE instruction set.
24331 This instruction set is supported by Pentium III and newer chips,
24332 and in the AMD line
24333 by Athlon-4, Athlon XP and Athlon MP chips. The earlier version of the SSE
24334 instruction set supports only single-precision arithmetic, thus the double and
24335 extended-precision arithmetic are still done using 387. A later version, present
24336 only in Pentium 4 and AMD x86-64 chips, supports double-precision
24339 For the x86-32 compiler, you must use @option{-march=@var{cpu-type}}, @option{-msse}
24340 or @option{-msse2} switches to enable SSE extensions and make this option
24341 effective. For the x86-64 compiler, these extensions are enabled by default.
24343 The resulting code should be considerably faster in the majority of cases and avoid
24344 the numerical instability problems of 387 code, but may break some existing
24345 code that expects temporaries to be 80 bits.
24347 This is the default choice for the x86-64 compiler.
24352 Attempt to utilize both instruction sets at once. This effectively doubles the
24353 amount of available registers, and on chips with separate execution units for
24354 387 and SSE the execution resources too. Use this option with care, as it is
24355 still experimental, because the GCC register allocator does not model separate
24356 functional units well, resulting in unstable performance.
24359 @item -masm=@var{dialect}
24360 @opindex masm=@var{dialect}
24361 Output assembly instructions using selected @var{dialect}. Also affects
24362 which dialect is used for basic @code{asm} (@pxref{Basic Asm}) and
24363 extended @code{asm} (@pxref{Extended Asm}). Supported choices (in dialect
24364 order) are @samp{att} or @samp{intel}. The default is @samp{att}. Darwin does
24365 not support @samp{intel}.
24368 @itemx -mno-ieee-fp
24370 @opindex mno-ieee-fp
24371 Control whether or not the compiler uses IEEE floating-point
24372 comparisons. These correctly handle the case where the result of a
24373 comparison is unordered.
24378 @opindex mhard-float
24379 Generate output containing 80387 instructions for floating point.
24384 @opindex msoft-float
24385 Generate output containing library calls for floating point.
24387 @strong{Warning:} the requisite libraries are not part of GCC@.
24388 Normally the facilities of the machine's usual C compiler are used, but
24389 this can't be done directly in cross-compilation. You must make your
24390 own arrangements to provide suitable library functions for
24393 On machines where a function returns floating-point results in the 80387
24394 register stack, some floating-point opcodes may be emitted even if
24395 @option{-msoft-float} is used.
24397 @item -mno-fp-ret-in-387
24398 @opindex mno-fp-ret-in-387
24399 Do not use the FPU registers for return values of functions.
24401 The usual calling convention has functions return values of types
24402 @code{float} and @code{double} in an FPU register, even if there
24403 is no FPU@. The idea is that the operating system should emulate
24406 The option @option{-mno-fp-ret-in-387} causes such values to be returned
24407 in ordinary CPU registers instead.
24409 @item -mno-fancy-math-387
24410 @opindex mno-fancy-math-387
24411 Some 387 emulators do not support the @code{sin}, @code{cos} and
24412 @code{sqrt} instructions for the 387. Specify this option to avoid
24413 generating those instructions. This option is the default on
24414 OpenBSD and NetBSD@. This option is overridden when @option{-march}
24415 indicates that the target CPU always has an FPU and so the
24416 instruction does not need emulation. These
24417 instructions are not generated unless you also use the
24418 @option{-funsafe-math-optimizations} switch.
24420 @item -malign-double
24421 @itemx -mno-align-double
24422 @opindex malign-double
24423 @opindex mno-align-double
24424 Control whether GCC aligns @code{double}, @code{long double}, and
24425 @code{long long} variables on a two-word boundary or a one-word
24426 boundary. Aligning @code{double} variables on a two-word boundary
24427 produces code that runs somewhat faster on a Pentium at the
24428 expense of more memory.
24430 On x86-64, @option{-malign-double} is enabled by default.
24432 @strong{Warning:} if you use the @option{-malign-double} switch,
24433 structures containing the above types are aligned differently than
24434 the published application binary interface specifications for the x86-32
24435 and are not binary compatible with structures in code compiled
24436 without that switch.
24438 @item -m96bit-long-double
24439 @itemx -m128bit-long-double
24440 @opindex m96bit-long-double
24441 @opindex m128bit-long-double
24442 These switches control the size of @code{long double} type. The x86-32
24443 application binary interface specifies the size to be 96 bits,
24444 so @option{-m96bit-long-double} is the default in 32-bit mode.
24446 Modern architectures (Pentium and newer) prefer @code{long double}
24447 to be aligned to an 8- or 16-byte boundary. In arrays or structures
24448 conforming to the ABI, this is not possible. So specifying
24449 @option{-m128bit-long-double} aligns @code{long double}
24450 to a 16-byte boundary by padding the @code{long double} with an additional
24453 In the x86-64 compiler, @option{-m128bit-long-double} is the default choice as
24454 its ABI specifies that @code{long double} is aligned on 16-byte boundary.
24456 Notice that neither of these options enable any extra precision over the x87
24457 standard of 80 bits for a @code{long double}.
24459 @strong{Warning:} if you override the default value for your target ABI, this
24460 changes the size of
24461 structures and arrays containing @code{long double} variables,
24462 as well as modifying the function calling convention for functions taking
24463 @code{long double}. Hence they are not binary-compatible
24464 with code compiled without that switch.
24466 @item -mlong-double-64
24467 @itemx -mlong-double-80
24468 @itemx -mlong-double-128
24469 @opindex mlong-double-64
24470 @opindex mlong-double-80
24471 @opindex mlong-double-128
24472 These switches control the size of @code{long double} type. A size
24473 of 64 bits makes the @code{long double} type equivalent to the @code{double}
24474 type. This is the default for 32-bit Bionic C library. A size
24475 of 128 bits makes the @code{long double} type equivalent to the
24476 @code{__float128} type. This is the default for 64-bit Bionic C library.
24478 @strong{Warning:} if you override the default value for your target ABI, this
24479 changes the size of
24480 structures and arrays containing @code{long double} variables,
24481 as well as modifying the function calling convention for functions taking
24482 @code{long double}. Hence they are not binary-compatible
24483 with code compiled without that switch.
24485 @item -malign-data=@var{type}
24486 @opindex malign-data
24487 Control how GCC aligns variables. Supported values for @var{type} are
24488 @samp{compat} uses increased alignment value compatible uses GCC 4.8
24489 and earlier, @samp{abi} uses alignment value as specified by the
24490 psABI, and @samp{cacheline} uses increased alignment value to match
24491 the cache line size. @samp{compat} is the default.
24493 @item -mlarge-data-threshold=@var{threshold}
24494 @opindex mlarge-data-threshold
24495 When @option{-mcmodel=medium} is specified, data objects larger than
24496 @var{threshold} are placed in the large data section. This value must be the
24497 same across all objects linked into the binary, and defaults to 65535.
24501 Use a different function-calling convention, in which functions that
24502 take a fixed number of arguments return with the @code{ret @var{num}}
24503 instruction, which pops their arguments while returning. This saves one
24504 instruction in the caller since there is no need to pop the arguments
24507 You can specify that an individual function is called with this calling
24508 sequence with the function attribute @code{stdcall}. You can also
24509 override the @option{-mrtd} option by using the function attribute
24510 @code{cdecl}. @xref{Function Attributes}.
24512 @strong{Warning:} this calling convention is incompatible with the one
24513 normally used on Unix, so you cannot use it if you need to call
24514 libraries compiled with the Unix compiler.
24516 Also, you must provide function prototypes for all functions that
24517 take variable numbers of arguments (including @code{printf});
24518 otherwise incorrect code is generated for calls to those
24521 In addition, seriously incorrect code results if you call a
24522 function with too many arguments. (Normally, extra arguments are
24523 harmlessly ignored.)
24525 @item -mregparm=@var{num}
24527 Control how many registers are used to pass integer arguments. By
24528 default, no registers are used to pass arguments, and at most 3
24529 registers can be used. You can control this behavior for a specific
24530 function by using the function attribute @code{regparm}.
24531 @xref{Function Attributes}.
24533 @strong{Warning:} if you use this switch, and
24534 @var{num} is nonzero, then you must build all modules with the same
24535 value, including any libraries. This includes the system libraries and
24539 @opindex msseregparm
24540 Use SSE register passing conventions for float and double arguments
24541 and return values. You can control this behavior for a specific
24542 function by using the function attribute @code{sseregparm}.
24543 @xref{Function Attributes}.
24545 @strong{Warning:} if you use this switch then you must build all
24546 modules with the same value, including any libraries. This includes
24547 the system libraries and startup modules.
24549 @item -mvect8-ret-in-mem
24550 @opindex mvect8-ret-in-mem
24551 Return 8-byte vectors in memory instead of MMX registers. This is the
24552 default on Solaris@tie{}8 and 9 and VxWorks to match the ABI of the Sun
24553 Studio compilers until version 12. Later compiler versions (starting
24554 with Studio 12 Update@tie{}1) follow the ABI used by other x86 targets, which
24555 is the default on Solaris@tie{}10 and later. @emph{Only} use this option if
24556 you need to remain compatible with existing code produced by those
24557 previous compiler versions or older versions of GCC@.
24566 Set 80387 floating-point precision to 32, 64 or 80 bits. When @option{-mpc32}
24567 is specified, the significands of results of floating-point operations are
24568 rounded to 24 bits (single precision); @option{-mpc64} rounds the
24569 significands of results of floating-point operations to 53 bits (double
24570 precision) and @option{-mpc80} rounds the significands of results of
24571 floating-point operations to 64 bits (extended double precision), which is
24572 the default. When this option is used, floating-point operations in higher
24573 precisions are not available to the programmer without setting the FPU
24574 control word explicitly.
24576 Setting the rounding of floating-point operations to less than the default
24577 80 bits can speed some programs by 2% or more. Note that some mathematical
24578 libraries assume that extended-precision (80-bit) floating-point operations
24579 are enabled by default; routines in such libraries could suffer significant
24580 loss of accuracy, typically through so-called ``catastrophic cancellation'',
24581 when this option is used to set the precision to less than extended precision.
24583 @item -mstackrealign
24584 @opindex mstackrealign
24585 Realign the stack at entry. On the x86, the @option{-mstackrealign}
24586 option generates an alternate prologue and epilogue that realigns the
24587 run-time stack if necessary. This supports mixing legacy codes that keep
24588 4-byte stack alignment with modern codes that keep 16-byte stack alignment for
24589 SSE compatibility. See also the attribute @code{force_align_arg_pointer},
24590 applicable to individual functions.
24592 @item -mpreferred-stack-boundary=@var{num}
24593 @opindex mpreferred-stack-boundary
24594 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
24595 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
24596 the default is 4 (16 bytes or 128 bits).
24598 @strong{Warning:} When generating code for the x86-64 architecture with
24599 SSE extensions disabled, @option{-mpreferred-stack-boundary=3} can be
24600 used to keep the stack boundary aligned to 8 byte boundary. Since
24601 x86-64 ABI require 16 byte stack alignment, this is ABI incompatible and
24602 intended to be used in controlled environment where stack space is
24603 important limitation. This option leads to wrong code when functions
24604 compiled with 16 byte stack alignment (such as functions from a standard
24605 library) are called with misaligned stack. In this case, SSE
24606 instructions may lead to misaligned memory access traps. In addition,
24607 variable arguments are handled incorrectly for 16 byte aligned
24608 objects (including x87 long double and __int128), leading to wrong
24609 results. You must build all modules with
24610 @option{-mpreferred-stack-boundary=3}, including any libraries. This
24611 includes the system libraries and startup modules.
24613 @item -mincoming-stack-boundary=@var{num}
24614 @opindex mincoming-stack-boundary
24615 Assume the incoming stack is aligned to a 2 raised to @var{num} byte
24616 boundary. If @option{-mincoming-stack-boundary} is not specified,
24617 the one specified by @option{-mpreferred-stack-boundary} is used.
24619 On Pentium and Pentium Pro, @code{double} and @code{long double} values
24620 should be aligned to an 8-byte boundary (see @option{-malign-double}) or
24621 suffer significant run time performance penalties. On Pentium III, the
24622 Streaming SIMD Extension (SSE) data type @code{__m128} may not work
24623 properly if it is not 16-byte aligned.
24625 To ensure proper alignment of this values on the stack, the stack boundary
24626 must be as aligned as that required by any value stored on the stack.
24627 Further, every function must be generated such that it keeps the stack
24628 aligned. Thus calling a function compiled with a higher preferred
24629 stack boundary from a function compiled with a lower preferred stack
24630 boundary most likely misaligns the stack. It is recommended that
24631 libraries that use callbacks always use the default setting.
24633 This extra alignment does consume extra stack space, and generally
24634 increases code size. Code that is sensitive to stack space usage, such
24635 as embedded systems and operating system kernels, may want to reduce the
24636 preferred alignment to @option{-mpreferred-stack-boundary=2}.
24693 @itemx -mavx512ifma
24694 @opindex mavx512ifma
24696 @itemx -mavx512vbmi
24697 @opindex mavx512vbmi
24709 @opindex mclfushopt
24726 @itemx -mprefetchwt1
24727 @opindex mprefetchwt1
24783 These switches enable the use of instructions in the MMX, SSE,
24784 SSE2, SSE3, SSSE3, SSE4.1, AVX, AVX2, AVX512F, AVX512PF, AVX512ER, AVX512CD,
24785 SHA, AES, PCLMUL, FSGSBASE, RDRND, F16C, FMA, SSE4A, FMA4, XOP, LWP, ABM,
24786 AVX512VL, AVX512BW, AVX512DQ, AVX512IFMA AVX512VBMI, BMI, BMI2, FXSR,
24787 XSAVE, XSAVEOPT, LZCNT, RTM, MPX, MWAITX, PKU or 3DNow!@:
24788 extended instruction sets. Each has a corresponding @option{-mno-} option
24789 to disable use of these instructions.
24791 These extensions are also available as built-in functions: see
24792 @ref{x86 Built-in Functions}, for details of the functions enabled and
24793 disabled by these switches.
24795 To generate SSE/SSE2 instructions automatically from floating-point
24796 code (as opposed to 387 instructions), see @option{-mfpmath=sse}.
24798 GCC depresses SSEx instructions when @option{-mavx} is used. Instead, it
24799 generates new AVX instructions or AVX equivalence for all SSEx instructions
24802 These options enable GCC to use these extended instructions in
24803 generated code, even without @option{-mfpmath=sse}. Applications that
24804 perform run-time CPU detection must compile separate files for each
24805 supported architecture, using the appropriate flags. In particular,
24806 the file containing the CPU detection code should be compiled without
24809 @item -mdump-tune-features
24810 @opindex mdump-tune-features
24811 This option instructs GCC to dump the names of the x86 performance
24812 tuning features and default settings. The names can be used in
24813 @option{-mtune-ctrl=@var{feature-list}}.
24815 @item -mtune-ctrl=@var{feature-list}
24816 @opindex mtune-ctrl=@var{feature-list}
24817 This option is used to do fine grain control of x86 code generation features.
24818 @var{feature-list} is a comma separated list of @var{feature} names. See also
24819 @option{-mdump-tune-features}. When specified, the @var{feature} is turned
24820 on if it is not preceded with @samp{^}, otherwise, it is turned off.
24821 @option{-mtune-ctrl=@var{feature-list}} is intended to be used by GCC
24822 developers. Using it may lead to code paths not covered by testing and can
24823 potentially result in compiler ICEs or runtime errors.
24826 @opindex mno-default
24827 This option instructs GCC to turn off all tunable features. See also
24828 @option{-mtune-ctrl=@var{feature-list}} and @option{-mdump-tune-features}.
24832 This option instructs GCC to emit a @code{cld} instruction in the prologue
24833 of functions that use string instructions. String instructions depend on
24834 the DF flag to select between autoincrement or autodecrement mode. While the
24835 ABI specifies the DF flag to be cleared on function entry, some operating
24836 systems violate this specification by not clearing the DF flag in their
24837 exception dispatchers. The exception handler can be invoked with the DF flag
24838 set, which leads to wrong direction mode when string instructions are used.
24839 This option can be enabled by default on 32-bit x86 targets by configuring
24840 GCC with the @option{--enable-cld} configure option. Generation of @code{cld}
24841 instructions can be suppressed with the @option{-mno-cld} compiler option
24845 @opindex mvzeroupper
24846 This option instructs GCC to emit a @code{vzeroupper} instruction
24847 before a transfer of control flow out of the function to minimize
24848 the AVX to SSE transition penalty as well as remove unnecessary @code{zeroupper}
24851 @item -mprefer-avx128
24852 @opindex mprefer-avx128
24853 This option instructs GCC to use 128-bit AVX instructions instead of
24854 256-bit AVX instructions in the auto-vectorizer.
24858 This option enables GCC to generate @code{CMPXCHG16B} instructions.
24859 @code{CMPXCHG16B} allows for atomic operations on 128-bit double quadword
24860 (or oword) data types.
24861 This is useful for high-resolution counters that can be updated
24862 by multiple processors (or cores). This instruction is generated as part of
24863 atomic built-in functions: see @ref{__sync Builtins} or
24864 @ref{__atomic Builtins} for details.
24868 This option enables generation of @code{SAHF} instructions in 64-bit code.
24869 Early Intel Pentium 4 CPUs with Intel 64 support,
24870 prior to the introduction of Pentium 4 G1 step in December 2005,
24871 lacked the @code{LAHF} and @code{SAHF} instructions
24872 which are supported by AMD64.
24873 These are load and store instructions, respectively, for certain status flags.
24874 In 64-bit mode, the @code{SAHF} instruction is used to optimize @code{fmod},
24875 @code{drem}, and @code{remainder} built-in functions;
24876 see @ref{Other Builtins} for details.
24880 This option enables use of the @code{movbe} instruction to implement
24881 @code{__builtin_bswap32} and @code{__builtin_bswap64}.
24885 This option enables built-in functions @code{__builtin_ia32_crc32qi},
24886 @code{__builtin_ia32_crc32hi}, @code{__builtin_ia32_crc32si} and
24887 @code{__builtin_ia32_crc32di} to generate the @code{crc32} machine instruction.
24891 This option enables use of @code{RCPSS} and @code{RSQRTSS} instructions
24892 (and their vectorized variants @code{RCPPS} and @code{RSQRTPS})
24893 with an additional Newton-Raphson step
24894 to increase precision instead of @code{DIVSS} and @code{SQRTSS}
24895 (and their vectorized
24896 variants) for single-precision floating-point arguments. These instructions
24897 are generated only when @option{-funsafe-math-optimizations} is enabled
24898 together with @option{-ffinite-math-only} and @option{-fno-trapping-math}.
24899 Note that while the throughput of the sequence is higher than the throughput
24900 of the non-reciprocal instruction, the precision of the sequence can be
24901 decreased by up to 2 ulp (i.e. the inverse of 1.0 equals 0.99999994).
24903 Note that GCC implements @code{1.0f/sqrtf(@var{x})} in terms of @code{RSQRTSS}
24904 (or @code{RSQRTPS}) already with @option{-ffast-math} (or the above option
24905 combination), and doesn't need @option{-mrecip}.
24907 Also note that GCC emits the above sequence with additional Newton-Raphson step
24908 for vectorized single-float division and vectorized @code{sqrtf(@var{x})}
24909 already with @option{-ffast-math} (or the above option combination), and
24910 doesn't need @option{-mrecip}.
24912 @item -mrecip=@var{opt}
24913 @opindex mrecip=opt
24914 This option controls which reciprocal estimate instructions
24915 may be used. @var{opt} is a comma-separated list of options, which may
24916 be preceded by a @samp{!} to invert the option:
24920 Enable all estimate instructions.
24923 Enable the default instructions, equivalent to @option{-mrecip}.
24926 Disable all estimate instructions, equivalent to @option{-mno-recip}.
24929 Enable the approximation for scalar division.
24932 Enable the approximation for vectorized division.
24935 Enable the approximation for scalar square root.
24938 Enable the approximation for vectorized square root.
24941 So, for example, @option{-mrecip=all,!sqrt} enables
24942 all of the reciprocal approximations, except for square root.
24944 @item -mveclibabi=@var{type}
24945 @opindex mveclibabi
24946 Specifies the ABI type to use for vectorizing intrinsics using an
24947 external library. Supported values for @var{type} are @samp{svml}
24948 for the Intel short
24949 vector math library and @samp{acml} for the AMD math core library.
24950 To use this option, both @option{-ftree-vectorize} and
24951 @option{-funsafe-math-optimizations} have to be enabled, and an SVML or ACML
24952 ABI-compatible library must be specified at link time.
24954 GCC currently emits calls to @code{vmldExp2},
24955 @code{vmldLn2}, @code{vmldLog102}, @code{vmldLog102}, @code{vmldPow2},
24956 @code{vmldTanh2}, @code{vmldTan2}, @code{vmldAtan2}, @code{vmldAtanh2},
24957 @code{vmldCbrt2}, @code{vmldSinh2}, @code{vmldSin2}, @code{vmldAsinh2},
24958 @code{vmldAsin2}, @code{vmldCosh2}, @code{vmldCos2}, @code{vmldAcosh2},
24959 @code{vmldAcos2}, @code{vmlsExp4}, @code{vmlsLn4}, @code{vmlsLog104},
24960 @code{vmlsLog104}, @code{vmlsPow4}, @code{vmlsTanh4}, @code{vmlsTan4},
24961 @code{vmlsAtan4}, @code{vmlsAtanh4}, @code{vmlsCbrt4}, @code{vmlsSinh4},
24962 @code{vmlsSin4}, @code{vmlsAsinh4}, @code{vmlsAsin4}, @code{vmlsCosh4},
24963 @code{vmlsCos4}, @code{vmlsAcosh4} and @code{vmlsAcos4} for corresponding
24964 function type when @option{-mveclibabi=svml} is used, and @code{__vrd2_sin},
24965 @code{__vrd2_cos}, @code{__vrd2_exp}, @code{__vrd2_log}, @code{__vrd2_log2},
24966 @code{__vrd2_log10}, @code{__vrs4_sinf}, @code{__vrs4_cosf},
24967 @code{__vrs4_expf}, @code{__vrs4_logf}, @code{__vrs4_log2f},
24968 @code{__vrs4_log10f} and @code{__vrs4_powf} for the corresponding function type
24969 when @option{-mveclibabi=acml} is used.
24971 @item -mabi=@var{name}
24973 Generate code for the specified calling convention. Permissible values
24974 are @samp{sysv} for the ABI used on GNU/Linux and other systems, and
24975 @samp{ms} for the Microsoft ABI. The default is to use the Microsoft
24976 ABI when targeting Microsoft Windows and the SysV ABI on all other systems.
24977 You can control this behavior for specific functions by
24978 using the function attributes @code{ms_abi} and @code{sysv_abi}.
24979 @xref{Function Attributes}.
24981 @item -mtls-dialect=@var{type}
24982 @opindex mtls-dialect
24983 Generate code to access thread-local storage using the @samp{gnu} or
24984 @samp{gnu2} conventions. @samp{gnu} is the conservative default;
24985 @samp{gnu2} is more efficient, but it may add compile- and run-time
24986 requirements that cannot be satisfied on all systems.
24989 @itemx -mno-push-args
24990 @opindex mpush-args
24991 @opindex mno-push-args
24992 Use PUSH operations to store outgoing parameters. This method is shorter
24993 and usually equally fast as method using SUB/MOV operations and is enabled
24994 by default. In some cases disabling it may improve performance because of
24995 improved scheduling and reduced dependencies.
24997 @item -maccumulate-outgoing-args
24998 @opindex maccumulate-outgoing-args
24999 If enabled, the maximum amount of space required for outgoing arguments is
25000 computed in the function prologue. This is faster on most modern CPUs
25001 because of reduced dependencies, improved scheduling and reduced stack usage
25002 when the preferred stack boundary is not equal to 2. The drawback is a notable
25003 increase in code size. This switch implies @option{-mno-push-args}.
25007 Support thread-safe exception handling on MinGW. Programs that rely
25008 on thread-safe exception handling must compile and link all code with the
25009 @option{-mthreads} option. When compiling, @option{-mthreads} defines
25010 @option{-D_MT}; when linking, it links in a special thread helper library
25011 @option{-lmingwthrd} which cleans up per-thread exception-handling data.
25013 @item -mms-bitfields
25014 @itemx -mno-ms-bitfields
25015 @opindex mms-bitfields
25016 @opindex mno-ms-bitfields
25018 Enable/disable bit-field layout compatible with the native Microsoft
25021 If @code{packed} is used on a structure, or if bit-fields are used,
25022 it may be that the Microsoft ABI lays out the structure differently
25023 than the way GCC normally does. Particularly when moving packed
25024 data between functions compiled with GCC and the native Microsoft compiler
25025 (either via function call or as data in a file), it may be necessary to access
25028 This option is enabled by default for Microsoft Windows
25029 targets. This behavior can also be controlled locally by use of variable
25030 or type attributes. For more information, see @ref{x86 Variable Attributes}
25031 and @ref{x86 Type Attributes}.
25033 The Microsoft structure layout algorithm is fairly simple with the exception
25034 of the bit-field packing.
25035 The padding and alignment of members of structures and whether a bit-field
25036 can straddle a storage-unit boundary are determine by these rules:
25039 @item Structure members are stored sequentially in the order in which they are
25040 declared: the first member has the lowest memory address and the last member
25043 @item Every data object has an alignment requirement. The alignment requirement
25044 for all data except structures, unions, and arrays is either the size of the
25045 object or the current packing size (specified with either the
25046 @code{aligned} attribute or the @code{pack} pragma),
25047 whichever is less. For structures, unions, and arrays,
25048 the alignment requirement is the largest alignment requirement of its members.
25049 Every object is allocated an offset so that:
25052 offset % alignment_requirement == 0
25055 @item Adjacent bit-fields are packed into the same 1-, 2-, or 4-byte allocation
25056 unit if the integral types are the same size and if the next bit-field fits
25057 into the current allocation unit without crossing the boundary imposed by the
25058 common alignment requirements of the bit-fields.
25061 MSVC interprets zero-length bit-fields in the following ways:
25064 @item If a zero-length bit-field is inserted between two bit-fields that
25065 are normally coalesced, the bit-fields are not coalesced.
25072 unsigned long bf_1 : 12;
25074 unsigned long bf_2 : 12;
25079 The size of @code{t1} is 8 bytes with the zero-length bit-field. If the
25080 zero-length bit-field were removed, @code{t1}'s size would be 4 bytes.
25082 @item If a zero-length bit-field is inserted after a bit-field, @code{foo}, and the
25083 alignment of the zero-length bit-field is greater than the member that follows it,
25084 @code{bar}, @code{bar} is aligned as the type of the zero-length bit-field.
25105 For @code{t2}, @code{bar} is placed at offset 2, rather than offset 1.
25106 Accordingly, the size of @code{t2} is 4. For @code{t3}, the zero-length
25107 bit-field does not affect the alignment of @code{bar} or, as a result, the size
25110 Taking this into account, it is important to note the following:
25113 @item If a zero-length bit-field follows a normal bit-field, the type of the
25114 zero-length bit-field may affect the alignment of the structure as whole. For
25115 example, @code{t2} has a size of 4 bytes, since the zero-length bit-field follows a
25116 normal bit-field, and is of type short.
25118 @item Even if a zero-length bit-field is not followed by a normal bit-field, it may
25119 still affect the alignment of the structure:
25130 Here, @code{t4} takes up 4 bytes.
25133 @item Zero-length bit-fields following non-bit-field members are ignored:
25145 Here, @code{t5} takes up 2 bytes.
25149 @item -mno-align-stringops
25150 @opindex mno-align-stringops
25151 Do not align the destination of inlined string operations. This switch reduces
25152 code size and improves performance in case the destination is already aligned,
25153 but GCC doesn't know about it.
25155 @item -minline-all-stringops
25156 @opindex minline-all-stringops
25157 By default GCC inlines string operations only when the destination is
25158 known to be aligned to least a 4-byte boundary.
25159 This enables more inlining and increases code
25160 size, but may improve performance of code that depends on fast
25161 @code{memcpy}, @code{strlen},
25162 and @code{memset} for short lengths.
25164 @item -minline-stringops-dynamically
25165 @opindex minline-stringops-dynamically
25166 For string operations of unknown size, use run-time checks with
25167 inline code for small blocks and a library call for large blocks.
25169 @item -mstringop-strategy=@var{alg}
25170 @opindex mstringop-strategy=@var{alg}
25171 Override the internal decision heuristic for the particular algorithm to use
25172 for inlining string operations. The allowed values for @var{alg} are:
25178 Expand using i386 @code{rep} prefix of the specified size.
25182 @itemx unrolled_loop
25183 Expand into an inline loop.
25186 Always use a library call.
25189 @item -mmemcpy-strategy=@var{strategy}
25190 @opindex mmemcpy-strategy=@var{strategy}
25191 Override the internal decision heuristic to decide if @code{__builtin_memcpy}
25192 should be inlined and what inline algorithm to use when the expected size
25193 of the copy operation is known. @var{strategy}
25194 is a comma-separated list of @var{alg}:@var{max_size}:@var{dest_align} triplets.
25195 @var{alg} is specified in @option{-mstringop-strategy}, @var{max_size} specifies
25196 the max byte size with which inline algorithm @var{alg} is allowed. For the last
25197 triplet, the @var{max_size} must be @code{-1}. The @var{max_size} of the triplets
25198 in the list must be specified in increasing order. The minimal byte size for
25199 @var{alg} is @code{0} for the first triplet and @code{@var{max_size} + 1} of the
25202 @item -mmemset-strategy=@var{strategy}
25203 @opindex mmemset-strategy=@var{strategy}
25204 The option is similar to @option{-mmemcpy-strategy=} except that it is to control
25205 @code{__builtin_memset} expansion.
25207 @item -momit-leaf-frame-pointer
25208 @opindex momit-leaf-frame-pointer
25209 Don't keep the frame pointer in a register for leaf functions. This
25210 avoids the instructions to save, set up, and restore frame pointers and
25211 makes an extra register available in leaf functions. The option
25212 @option{-fomit-leaf-frame-pointer} removes the frame pointer for leaf functions,
25213 which might make debugging harder.
25215 @item -mtls-direct-seg-refs
25216 @itemx -mno-tls-direct-seg-refs
25217 @opindex mtls-direct-seg-refs
25218 Controls whether TLS variables may be accessed with offsets from the
25219 TLS segment register (@code{%gs} for 32-bit, @code{%fs} for 64-bit),
25220 or whether the thread base pointer must be added. Whether or not this
25221 is valid depends on the operating system, and whether it maps the
25222 segment to cover the entire TLS area.
25224 For systems that use the GNU C Library, the default is on.
25227 @itemx -mno-sse2avx
25229 Specify that the assembler should encode SSE instructions with VEX
25230 prefix. The option @option{-mavx} turns this on by default.
25235 If profiling is active (@option{-pg}), put the profiling
25236 counter call before the prologue.
25237 Note: On x86 architectures the attribute @code{ms_hook_prologue}
25238 isn't possible at the moment for @option{-mfentry} and @option{-pg}.
25240 @item -mrecord-mcount
25241 @itemx -mno-record-mcount
25242 @opindex mrecord-mcount
25243 If profiling is active (@option{-pg}), generate a __mcount_loc section
25244 that contains pointers to each profiling call. This is useful for
25245 automatically patching and out calls.
25248 @itemx -mno-nop-mcount
25249 @opindex mnop-mcount
25250 If profiling is active (@option{-pg}), generate the calls to
25251 the profiling functions as NOPs. This is useful when they
25252 should be patched in later dynamically. This is likely only
25253 useful together with @option{-mrecord-mcount}.
25255 @item -mskip-rax-setup
25256 @itemx -mno-skip-rax-setup
25257 @opindex mskip-rax-setup
25258 When generating code for the x86-64 architecture with SSE extensions
25259 disabled, @option{-mskip-rax-setup} can be used to skip setting up RAX
25260 register when there are no variable arguments passed in vector registers.
25262 @strong{Warning:} Since RAX register is used to avoid unnecessarily
25263 saving vector registers on stack when passing variable arguments, the
25264 impacts of this option are callees may waste some stack space,
25265 misbehave or jump to a random location. GCC 4.4 or newer don't have
25266 those issues, regardless the RAX register value.
25269 @itemx -mno-8bit-idiv
25270 @opindex m8bit-idiv
25271 On some processors, like Intel Atom, 8-bit unsigned integer divide is
25272 much faster than 32-bit/64-bit integer divide. This option generates a
25273 run-time check. If both dividend and divisor are within range of 0
25274 to 255, 8-bit unsigned integer divide is used instead of
25275 32-bit/64-bit integer divide.
25277 @item -mavx256-split-unaligned-load
25278 @itemx -mavx256-split-unaligned-store
25279 @opindex mavx256-split-unaligned-load
25280 @opindex mavx256-split-unaligned-store
25281 Split 32-byte AVX unaligned load and store.
25283 @item -mstack-protector-guard=@var{guard}
25284 @opindex mstack-protector-guard=@var{guard}
25285 Generate stack protection code using canary at @var{guard}. Supported
25286 locations are @samp{global} for global canary or @samp{tls} for per-thread
25287 canary in the TLS block (the default). This option has effect only when
25288 @option{-fstack-protector} or @option{-fstack-protector-all} is specified.
25290 @item -mmitigate-rop
25291 @opindex mmitigate-rop
25292 Try to avoid generating code sequences that contain unintended return
25293 opcodes, to mitigate against certain forms of attack. At the moment,
25294 this option is limited in what it can do and should not be relied
25295 on to provide serious protection.
25297 @item -mgeneral-regs-only
25298 @opindex mgeneral-regs-only
25299 Generate code that uses only the general-purpose registers. This
25300 prevents the compiler from using floating-point, vector, mask and bound
25305 These @samp{-m} switches are supported in addition to the above
25306 on x86-64 processors in 64-bit environments.
25319 Generate code for a 16-bit, 32-bit or 64-bit environment.
25320 The @option{-m32} option sets @code{int}, @code{long}, and pointer types
25322 generates code that runs on any i386 system.
25324 The @option{-m64} option sets @code{int} to 32 bits and @code{long} and pointer
25325 types to 64 bits, and generates code for the x86-64 architecture.
25326 For Darwin only the @option{-m64} option also turns off the @option{-fno-pic}
25327 and @option{-mdynamic-no-pic} options.
25329 The @option{-mx32} option sets @code{int}, @code{long}, and pointer types
25331 generates code for the x86-64 architecture.
25333 The @option{-m16} option is the same as @option{-m32}, except for that
25334 it outputs the @code{.code16gcc} assembly directive at the beginning of
25335 the assembly output so that the binary can run in 16-bit mode.
25337 The @option{-miamcu} option generates code which conforms to Intel MCU
25338 psABI. It requires the @option{-m32} option to be turned on.
25340 @item -mno-red-zone
25341 @opindex mno-red-zone
25342 Do not use a so-called ``red zone'' for x86-64 code. The red zone is mandated
25343 by the x86-64 ABI; it is a 128-byte area beyond the location of the
25344 stack pointer that is not modified by signal or interrupt handlers
25345 and therefore can be used for temporary data without adjusting the stack
25346 pointer. The flag @option{-mno-red-zone} disables this red zone.
25348 @item -mcmodel=small
25349 @opindex mcmodel=small
25350 Generate code for the small code model: the program and its symbols must
25351 be linked in the lower 2 GB of the address space. Pointers are 64 bits.
25352 Programs can be statically or dynamically linked. This is the default
25355 @item -mcmodel=kernel
25356 @opindex mcmodel=kernel
25357 Generate code for the kernel code model. The kernel runs in the
25358 negative 2 GB of the address space.
25359 This model has to be used for Linux kernel code.
25361 @item -mcmodel=medium
25362 @opindex mcmodel=medium
25363 Generate code for the medium model: the program is linked in the lower 2
25364 GB of the address space. Small symbols are also placed there. Symbols
25365 with sizes larger than @option{-mlarge-data-threshold} are put into
25366 large data or BSS sections and can be located above 2GB. Programs can
25367 be statically or dynamically linked.
25369 @item -mcmodel=large
25370 @opindex mcmodel=large
25371 Generate code for the large model. This model makes no assumptions
25372 about addresses and sizes of sections.
25374 @item -maddress-mode=long
25375 @opindex maddress-mode=long
25376 Generate code for long address mode. This is only supported for 64-bit
25377 and x32 environments. It is the default address mode for 64-bit
25380 @item -maddress-mode=short
25381 @opindex maddress-mode=short
25382 Generate code for short address mode. This is only supported for 32-bit
25383 and x32 environments. It is the default address mode for 32-bit and
25387 @node x86 Windows Options
25388 @subsection x86 Windows Options
25389 @cindex x86 Windows Options
25390 @cindex Windows Options for x86
25392 These additional options are available for Microsoft Windows targets:
25398 specifies that a console application is to be generated, by
25399 instructing the linker to set the PE header subsystem type
25400 required for console applications.
25401 This option is available for Cygwin and MinGW targets and is
25402 enabled by default on those targets.
25406 This option is available for Cygwin and MinGW targets. It
25407 specifies that a DLL---a dynamic link library---is to be
25408 generated, enabling the selection of the required runtime
25409 startup object and entry point.
25411 @item -mnop-fun-dllimport
25412 @opindex mnop-fun-dllimport
25413 This option is available for Cygwin and MinGW targets. It
25414 specifies that the @code{dllimport} attribute should be ignored.
25418 This option is available for MinGW targets. It specifies
25419 that MinGW-specific thread support is to be used.
25423 This option is available for MinGW-w64 targets. It causes
25424 the @code{UNICODE} preprocessor macro to be predefined, and
25425 chooses Unicode-capable runtime startup code.
25429 This option is available for Cygwin and MinGW targets. It
25430 specifies that the typical Microsoft Windows predefined macros are to
25431 be set in the pre-processor, but does not influence the choice
25432 of runtime library/startup code.
25436 This option is available for Cygwin and MinGW targets. It
25437 specifies that a GUI application is to be generated by
25438 instructing the linker to set the PE header subsystem type
25441 @item -fno-set-stack-executable
25442 @opindex fno-set-stack-executable
25443 This option is available for MinGW targets. It specifies that
25444 the executable flag for the stack used by nested functions isn't
25445 set. This is necessary for binaries running in kernel mode of
25446 Microsoft Windows, as there the User32 API, which is used to set executable
25447 privileges, isn't available.
25449 @item -fwritable-relocated-rdata
25450 @opindex fno-writable-relocated-rdata
25451 This option is available for MinGW and Cygwin targets. It specifies
25452 that relocated-data in read-only section is put into the @code{.data}
25453 section. This is a necessary for older runtimes not supporting
25454 modification of @code{.rdata} sections for pseudo-relocation.
25456 @item -mpe-aligned-commons
25457 @opindex mpe-aligned-commons
25458 This option is available for Cygwin and MinGW targets. It
25459 specifies that the GNU extension to the PE file format that
25460 permits the correct alignment of COMMON variables should be
25461 used when generating code. It is enabled by default if
25462 GCC detects that the target assembler found during configuration
25463 supports the feature.
25466 See also under @ref{x86 Options} for standard options.
25468 @node Xstormy16 Options
25469 @subsection Xstormy16 Options
25470 @cindex Xstormy16 Options
25472 These options are defined for Xstormy16:
25477 Choose startup files and linker script suitable for the simulator.
25480 @node Xtensa Options
25481 @subsection Xtensa Options
25482 @cindex Xtensa Options
25484 These options are supported for Xtensa targets:
25488 @itemx -mno-const16
25490 @opindex mno-const16
25491 Enable or disable use of @code{CONST16} instructions for loading
25492 constant values. The @code{CONST16} instruction is currently not a
25493 standard option from Tensilica. When enabled, @code{CONST16}
25494 instructions are always used in place of the standard @code{L32R}
25495 instructions. The use of @code{CONST16} is enabled by default only if
25496 the @code{L32R} instruction is not available.
25499 @itemx -mno-fused-madd
25500 @opindex mfused-madd
25501 @opindex mno-fused-madd
25502 Enable or disable use of fused multiply/add and multiply/subtract
25503 instructions in the floating-point option. This has no effect if the
25504 floating-point option is not also enabled. Disabling fused multiply/add
25505 and multiply/subtract instructions forces the compiler to use separate
25506 instructions for the multiply and add/subtract operations. This may be
25507 desirable in some cases where strict IEEE 754-compliant results are
25508 required: the fused multiply add/subtract instructions do not round the
25509 intermediate result, thereby producing results with @emph{more} bits of
25510 precision than specified by the IEEE standard. Disabling fused multiply
25511 add/subtract instructions also ensures that the program output is not
25512 sensitive to the compiler's ability to combine multiply and add/subtract
25515 @item -mserialize-volatile
25516 @itemx -mno-serialize-volatile
25517 @opindex mserialize-volatile
25518 @opindex mno-serialize-volatile
25519 When this option is enabled, GCC inserts @code{MEMW} instructions before
25520 @code{volatile} memory references to guarantee sequential consistency.
25521 The default is @option{-mserialize-volatile}. Use
25522 @option{-mno-serialize-volatile} to omit the @code{MEMW} instructions.
25524 @item -mforce-no-pic
25525 @opindex mforce-no-pic
25526 For targets, like GNU/Linux, where all user-mode Xtensa code must be
25527 position-independent code (PIC), this option disables PIC for compiling
25530 @item -mtext-section-literals
25531 @itemx -mno-text-section-literals
25532 @opindex mtext-section-literals
25533 @opindex mno-text-section-literals
25534 These options control the treatment of literal pools. The default is
25535 @option{-mno-text-section-literals}, which places literals in a separate
25536 section in the output file. This allows the literal pool to be placed
25537 in a data RAM/ROM, and it also allows the linker to combine literal
25538 pools from separate object files to remove redundant literals and
25539 improve code size. With @option{-mtext-section-literals}, the literals
25540 are interspersed in the text section in order to keep them as close as
25541 possible to their references. This may be necessary for large assembly
25542 files. Literals for each function are placed right before that function.
25544 @item -mauto-litpools
25545 @itemx -mno-auto-litpools
25546 @opindex mauto-litpools
25547 @opindex mno-auto-litpools
25548 These options control the treatment of literal pools. The default is
25549 @option{-mno-auto-litpools}, which places literals in a separate
25550 section in the output file unless @option{-mtext-section-literals} is
25551 used. With @option{-mauto-litpools} the literals are interspersed in
25552 the text section by the assembler. Compiler does not produce explicit
25553 @code{.literal} directives and loads literals into registers with
25554 @code{MOVI} instructions instead of @code{L32R} to let the assembler
25555 do relaxation and place literals as necessary. This option allows
25556 assembler to create several literal pools per function and assemble
25557 very big functions, which may not be possible with
25558 @option{-mtext-section-literals}.
25560 @item -mtarget-align
25561 @itemx -mno-target-align
25562 @opindex mtarget-align
25563 @opindex mno-target-align
25564 When this option is enabled, GCC instructs the assembler to
25565 automatically align instructions to reduce branch penalties at the
25566 expense of some code density. The assembler attempts to widen density
25567 instructions to align branch targets and the instructions following call
25568 instructions. If there are not enough preceding safe density
25569 instructions to align a target, no widening is performed. The
25570 default is @option{-mtarget-align}. These options do not affect the
25571 treatment of auto-aligned instructions like @code{LOOP}, which the
25572 assembler always aligns, either by widening density instructions or
25573 by inserting NOP instructions.
25576 @itemx -mno-longcalls
25577 @opindex mlongcalls
25578 @opindex mno-longcalls
25579 When this option is enabled, GCC instructs the assembler to translate
25580 direct calls to indirect calls unless it can determine that the target
25581 of a direct call is in the range allowed by the call instruction. This
25582 translation typically occurs for calls to functions in other source
25583 files. Specifically, the assembler translates a direct @code{CALL}
25584 instruction into an @code{L32R} followed by a @code{CALLX} instruction.
25585 The default is @option{-mno-longcalls}. This option should be used in
25586 programs where the call target can potentially be out of range. This
25587 option is implemented in the assembler, not the compiler, so the
25588 assembly code generated by GCC still shows direct call
25589 instructions---look at the disassembled object code to see the actual
25590 instructions. Note that the assembler uses an indirect call for
25591 every cross-file call, not just those that really are out of range.
25594 @node zSeries Options
25595 @subsection zSeries Options
25596 @cindex zSeries options
25598 These are listed under @xref{S/390 and zSeries Options}.
25604 @section Specifying Subprocesses and the Switches to Pass to Them
25607 @command{gcc} is a driver program. It performs its job by invoking a
25608 sequence of other programs to do the work of compiling, assembling and
25609 linking. GCC interprets its command-line parameters and uses these to
25610 deduce which programs it should invoke, and which command-line options
25611 it ought to place on their command lines. This behavior is controlled
25612 by @dfn{spec strings}. In most cases there is one spec string for each
25613 program that GCC can invoke, but a few programs have multiple spec
25614 strings to control their behavior. The spec strings built into GCC can
25615 be overridden by using the @option{-specs=} command-line switch to specify
25618 @dfn{Spec files} are plain-text files that are used to construct spec
25619 strings. They consist of a sequence of directives separated by blank
25620 lines. The type of directive is determined by the first non-whitespace
25621 character on the line, which can be one of the following:
25624 @item %@var{command}
25625 Issues a @var{command} to the spec file processor. The commands that can
25629 @item %include <@var{file}>
25630 @cindex @code{%include}
25631 Search for @var{file} and insert its text at the current point in the
25634 @item %include_noerr <@var{file}>
25635 @cindex @code{%include_noerr}
25636 Just like @samp{%include}, but do not generate an error message if the include
25637 file cannot be found.
25639 @item %rename @var{old_name} @var{new_name}
25640 @cindex @code{%rename}
25641 Rename the spec string @var{old_name} to @var{new_name}.
25645 @item *[@var{spec_name}]:
25646 This tells the compiler to create, override or delete the named spec
25647 string. All lines after this directive up to the next directive or
25648 blank line are considered to be the text for the spec string. If this
25649 results in an empty string then the spec is deleted. (Or, if the
25650 spec did not exist, then nothing happens.) Otherwise, if the spec
25651 does not currently exist a new spec is created. If the spec does
25652 exist then its contents are overridden by the text of this
25653 directive, unless the first character of that text is the @samp{+}
25654 character, in which case the text is appended to the spec.
25656 @item [@var{suffix}]:
25657 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
25658 and up to the next directive or blank line are considered to make up the
25659 spec string for the indicated suffix. When the compiler encounters an
25660 input file with the named suffix, it processes the spec string in
25661 order to work out how to compile that file. For example:
25665 z-compile -input %i
25668 This says that any input file whose name ends in @samp{.ZZ} should be
25669 passed to the program @samp{z-compile}, which should be invoked with the
25670 command-line switch @option{-input} and with the result of performing the
25671 @samp{%i} substitution. (See below.)
25673 As an alternative to providing a spec string, the text following a
25674 suffix directive can be one of the following:
25677 @item @@@var{language}
25678 This says that the suffix is an alias for a known @var{language}. This is
25679 similar to using the @option{-x} command-line switch to GCC to specify a
25680 language explicitly. For example:
25687 Says that .ZZ files are, in fact, C++ source files.
25690 This causes an error messages saying:
25693 @var{name} compiler not installed on this system.
25697 GCC already has an extensive list of suffixes built into it.
25698 This directive adds an entry to the end of the list of suffixes, but
25699 since the list is searched from the end backwards, it is effectively
25700 possible to override earlier entries using this technique.
25704 GCC has the following spec strings built into it. Spec files can
25705 override these strings or create their own. Note that individual
25706 targets can also add their own spec strings to this list.
25709 asm Options to pass to the assembler
25710 asm_final Options to pass to the assembler post-processor
25711 cpp Options to pass to the C preprocessor
25712 cc1 Options to pass to the C compiler
25713 cc1plus Options to pass to the C++ compiler
25714 endfile Object files to include at the end of the link
25715 link Options to pass to the linker
25716 lib Libraries to include on the command line to the linker
25717 libgcc Decides which GCC support library to pass to the linker
25718 linker Sets the name of the linker
25719 predefines Defines to be passed to the C preprocessor
25720 signed_char Defines to pass to CPP to say whether @code{char} is signed
25722 startfile Object files to include at the start of the link
25725 Here is a small example of a spec file:
25728 %rename lib old_lib
25731 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
25734 This example renames the spec called @samp{lib} to @samp{old_lib} and
25735 then overrides the previous definition of @samp{lib} with a new one.
25736 The new definition adds in some extra command-line options before
25737 including the text of the old definition.
25739 @dfn{Spec strings} are a list of command-line options to be passed to their
25740 corresponding program. In addition, the spec strings can contain
25741 @samp{%}-prefixed sequences to substitute variable text or to
25742 conditionally insert text into the command line. Using these constructs
25743 it is possible to generate quite complex command lines.
25745 Here is a table of all defined @samp{%}-sequences for spec
25746 strings. Note that spaces are not generated automatically around the
25747 results of expanding these sequences. Therefore you can concatenate them
25748 together or combine them with constant text in a single argument.
25752 Substitute one @samp{%} into the program name or argument.
25755 Substitute the name of the input file being processed.
25758 Substitute the basename of the input file being processed.
25759 This is the substring up to (and not including) the last period
25760 and not including the directory.
25763 This is the same as @samp{%b}, but include the file suffix (text after
25767 Marks the argument containing or following the @samp{%d} as a
25768 temporary file name, so that that file is deleted if GCC exits
25769 successfully. Unlike @samp{%g}, this contributes no text to the
25772 @item %g@var{suffix}
25773 Substitute a file name that has suffix @var{suffix} and is chosen
25774 once per compilation, and mark the argument in the same way as
25775 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
25776 name is now chosen in a way that is hard to predict even when previously
25777 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
25778 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
25779 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
25780 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
25781 was simply substituted with a file name chosen once per compilation,
25782 without regard to any appended suffix (which was therefore treated
25783 just like ordinary text), making such attacks more likely to succeed.
25785 @item %u@var{suffix}
25786 Like @samp{%g}, but generates a new temporary file name
25787 each time it appears instead of once per compilation.
25789 @item %U@var{suffix}
25790 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
25791 new one if there is no such last file name. In the absence of any
25792 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
25793 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
25794 involves the generation of two distinct file names, one
25795 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
25796 simply substituted with a file name chosen for the previous @samp{%u},
25797 without regard to any appended suffix.
25799 @item %j@var{suffix}
25800 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
25801 writable, and if @option{-save-temps} is not used;
25802 otherwise, substitute the name
25803 of a temporary file, just like @samp{%u}. This temporary file is not
25804 meant for communication between processes, but rather as a junk
25805 disposal mechanism.
25807 @item %|@var{suffix}
25808 @itemx %m@var{suffix}
25809 Like @samp{%g}, except if @option{-pipe} is in effect. In that case
25810 @samp{%|} substitutes a single dash and @samp{%m} substitutes nothing at
25811 all. These are the two most common ways to instruct a program that it
25812 should read from standard input or write to standard output. If you
25813 need something more elaborate you can use an @samp{%@{pipe:@code{X}@}}
25814 construct: see for example @file{f/lang-specs.h}.
25816 @item %.@var{SUFFIX}
25817 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
25818 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
25819 terminated by the next space or %.
25822 Marks the argument containing or following the @samp{%w} as the
25823 designated output file of this compilation. This puts the argument
25824 into the sequence of arguments that @samp{%o} substitutes.
25827 Substitutes the names of all the output files, with spaces
25828 automatically placed around them. You should write spaces
25829 around the @samp{%o} as well or the results are undefined.
25830 @samp{%o} is for use in the specs for running the linker.
25831 Input files whose names have no recognized suffix are not compiled
25832 at all, but they are included among the output files, so they are
25836 Substitutes the suffix for object files. Note that this is
25837 handled specially when it immediately follows @samp{%g, %u, or %U},
25838 because of the need for those to form complete file names. The
25839 handling is such that @samp{%O} is treated exactly as if it had already
25840 been substituted, except that @samp{%g, %u, and %U} do not currently
25841 support additional @var{suffix} characters following @samp{%O} as they do
25842 following, for example, @samp{.o}.
25845 Substitutes the standard macro predefinitions for the
25846 current target machine. Use this when running @command{cpp}.
25849 Like @samp{%p}, but puts @samp{__} before and after the name of each
25850 predefined macro, except for macros that start with @samp{__} or with
25851 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
25855 Substitute any of @option{-iprefix} (made from @env{GCC_EXEC_PREFIX}),
25856 @option{-isysroot} (made from @env{TARGET_SYSTEM_ROOT}),
25857 @option{-isystem} (made from @env{COMPILER_PATH} and @option{-B} options)
25858 and @option{-imultilib} as necessary.
25861 Current argument is the name of a library or startup file of some sort.
25862 Search for that file in a standard list of directories and substitute
25863 the full name found. The current working directory is included in the
25864 list of directories scanned.
25867 Current argument is the name of a linker script. Search for that file
25868 in the current list of directories to scan for libraries. If the file
25869 is located insert a @option{--script} option into the command line
25870 followed by the full path name found. If the file is not found then
25871 generate an error message. Note: the current working directory is not
25875 Print @var{str} as an error message. @var{str} is terminated by a newline.
25876 Use this when inconsistent options are detected.
25878 @item %(@var{name})
25879 Substitute the contents of spec string @var{name} at this point.
25881 @item %x@{@var{option}@}
25882 Accumulate an option for @samp{%X}.
25885 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
25889 Output the accumulated assembler options specified by @option{-Wa}.
25892 Output the accumulated preprocessor options specified by @option{-Wp}.
25895 Process the @code{asm} spec. This is used to compute the
25896 switches to be passed to the assembler.
25899 Process the @code{asm_final} spec. This is a spec string for
25900 passing switches to an assembler post-processor, if such a program is
25904 Process the @code{link} spec. This is the spec for computing the
25905 command line passed to the linker. Typically it makes use of the
25906 @samp{%L %G %S %D and %E} sequences.
25909 Dump out a @option{-L} option for each directory that GCC believes might
25910 contain startup files. If the target supports multilibs then the
25911 current multilib directory is prepended to each of these paths.
25914 Process the @code{lib} spec. This is a spec string for deciding which
25915 libraries are included on the command line to the linker.
25918 Process the @code{libgcc} spec. This is a spec string for deciding
25919 which GCC support library is included on the command line to the linker.
25922 Process the @code{startfile} spec. This is a spec for deciding which
25923 object files are the first ones passed to the linker. Typically
25924 this might be a file named @file{crt0.o}.
25927 Process the @code{endfile} spec. This is a spec string that specifies
25928 the last object files that are passed to the linker.
25931 Process the @code{cpp} spec. This is used to construct the arguments
25932 to be passed to the C preprocessor.
25935 Process the @code{cc1} spec. This is used to construct the options to be
25936 passed to the actual C compiler (@command{cc1}).
25939 Process the @code{cc1plus} spec. This is used to construct the options to be
25940 passed to the actual C++ compiler (@command{cc1plus}).
25943 Substitute the variable part of a matched option. See below.
25944 Note that each comma in the substituted string is replaced by
25948 Remove all occurrences of @code{-S} from the command line. Note---this
25949 command is position dependent. @samp{%} commands in the spec string
25950 before this one see @code{-S}, @samp{%} commands in the spec string
25951 after this one do not.
25953 @item %:@var{function}(@var{args})
25954 Call the named function @var{function}, passing it @var{args}.
25955 @var{args} is first processed as a nested spec string, then split
25956 into an argument vector in the usual fashion. The function returns
25957 a string which is processed as if it had appeared literally as part
25958 of the current spec.
25960 The following built-in spec functions are provided:
25963 @item @code{getenv}
25964 The @code{getenv} spec function takes two arguments: an environment
25965 variable name and a string. If the environment variable is not
25966 defined, a fatal error is issued. Otherwise, the return value is the
25967 value of the environment variable concatenated with the string. For
25968 example, if @env{TOPDIR} is defined as @file{/path/to/top}, then:
25971 %:getenv(TOPDIR /include)
25974 expands to @file{/path/to/top/include}.
25976 @item @code{if-exists}
25977 The @code{if-exists} spec function takes one argument, an absolute
25978 pathname to a file. If the file exists, @code{if-exists} returns the
25979 pathname. Here is a small example of its usage:
25983 crt0%O%s %:if-exists(crti%O%s) crtbegin%O%s
25986 @item @code{if-exists-else}
25987 The @code{if-exists-else} spec function is similar to the @code{if-exists}
25988 spec function, except that it takes two arguments. The first argument is
25989 an absolute pathname to a file. If the file exists, @code{if-exists-else}
25990 returns the pathname. If it does not exist, it returns the second argument.
25991 This way, @code{if-exists-else} can be used to select one file or another,
25992 based on the existence of the first. Here is a small example of its usage:
25996 crt0%O%s %:if-exists(crti%O%s) \
25997 %:if-exists-else(crtbeginT%O%s crtbegin%O%s)
26000 @item @code{replace-outfile}
26001 The @code{replace-outfile} spec function takes two arguments. It looks for the
26002 first argument in the outfiles array and replaces it with the second argument. Here
26003 is a small example of its usage:
26006 %@{fgnu-runtime:%:replace-outfile(-lobjc -lobjc-gnu)@}
26009 @item @code{remove-outfile}
26010 The @code{remove-outfile} spec function takes one argument. It looks for the
26011 first argument in the outfiles array and removes it. Here is a small example
26015 %:remove-outfile(-lm)
26018 @item @code{pass-through-libs}
26019 The @code{pass-through-libs} spec function takes any number of arguments. It
26020 finds any @option{-l} options and any non-options ending in @file{.a} (which it
26021 assumes are the names of linker input library archive files) and returns a
26022 result containing all the found arguments each prepended by
26023 @option{-plugin-opt=-pass-through=} and joined by spaces. This list is
26024 intended to be passed to the LTO linker plugin.
26027 %:pass-through-libs(%G %L %G)
26030 @item @code{print-asm-header}
26031 The @code{print-asm-header} function takes no arguments and simply
26032 prints a banner like:
26038 Use "-Wa,OPTION" to pass "OPTION" to the assembler.
26041 It is used to separate compiler options from assembler options
26042 in the @option{--target-help} output.
26045 @item %@{@code{S}@}
26046 Substitutes the @code{-S} switch, if that switch is given to GCC@.
26047 If that switch is not specified, this substitutes nothing. Note that
26048 the leading dash is omitted when specifying this option, and it is
26049 automatically inserted if the substitution is performed. Thus the spec
26050 string @samp{%@{foo@}} matches the command-line option @option{-foo}
26051 and outputs the command-line option @option{-foo}.
26053 @item %W@{@code{S}@}
26054 Like %@{@code{S}@} but mark last argument supplied within as a file to be
26055 deleted on failure.
26057 @item %@{@code{S}*@}
26058 Substitutes all the switches specified to GCC whose names start
26059 with @code{-S}, but which also take an argument. This is used for
26060 switches like @option{-o}, @option{-D}, @option{-I}, etc.
26061 GCC considers @option{-o foo} as being
26062 one switch whose name starts with @samp{o}. %@{o*@} substitutes this
26063 text, including the space. Thus two arguments are generated.
26065 @item %@{@code{S}*&@code{T}*@}
26066 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
26067 (the order of @code{S} and @code{T} in the spec is not significant).
26068 There can be any number of ampersand-separated variables; for each the
26069 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
26071 @item %@{@code{S}:@code{X}@}
26072 Substitutes @code{X}, if the @option{-S} switch is given to GCC@.
26074 @item %@{!@code{S}:@code{X}@}
26075 Substitutes @code{X}, if the @option{-S} switch is @emph{not} given to GCC@.
26077 @item %@{@code{S}*:@code{X}@}
26078 Substitutes @code{X} if one or more switches whose names start with
26079 @code{-S} are specified to GCC@. Normally @code{X} is substituted only
26080 once, no matter how many such switches appeared. However, if @code{%*}
26081 appears somewhere in @code{X}, then @code{X} is substituted once
26082 for each matching switch, with the @code{%*} replaced by the part of
26083 that switch matching the @code{*}.
26085 If @code{%*} appears as the last part of a spec sequence then a space
26086 is added after the end of the last substitution. If there is more
26087 text in the sequence, however, then a space is not generated. This
26088 allows the @code{%*} substitution to be used as part of a larger
26089 string. For example, a spec string like this:
26092 %@{mcu=*:--script=%*/memory.ld@}
26096 when matching an option like @option{-mcu=newchip} produces:
26099 --script=newchip/memory.ld
26102 @item %@{.@code{S}:@code{X}@}
26103 Substitutes @code{X}, if processing a file with suffix @code{S}.
26105 @item %@{!.@code{S}:@code{X}@}
26106 Substitutes @code{X}, if @emph{not} processing a file with suffix @code{S}.
26108 @item %@{,@code{S}:@code{X}@}
26109 Substitutes @code{X}, if processing a file for language @code{S}.
26111 @item %@{!,@code{S}:@code{X}@}
26112 Substitutes @code{X}, if not processing a file for language @code{S}.
26114 @item %@{@code{S}|@code{P}:@code{X}@}
26115 Substitutes @code{X} if either @code{-S} or @code{-P} is given to
26116 GCC@. This may be combined with @samp{!}, @samp{.}, @samp{,}, and
26117 @code{*} sequences as well, although they have a stronger binding than
26118 the @samp{|}. If @code{%*} appears in @code{X}, all of the
26119 alternatives must be starred, and only the first matching alternative
26122 For example, a spec string like this:
26125 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
26129 outputs the following command-line options from the following input
26130 command-line options:
26135 -d fred.c -foo -baz -boggle
26136 -d jim.d -bar -baz -boggle
26139 @item %@{S:X; T:Y; :D@}
26141 If @code{S} is given to GCC, substitutes @code{X}; else if @code{T} is
26142 given to GCC, substitutes @code{Y}; else substitutes @code{D}. There can
26143 be as many clauses as you need. This may be combined with @code{.},
26144 @code{,}, @code{!}, @code{|}, and @code{*} as needed.
26149 The conditional text @code{X} in a %@{@code{S}:@code{X}@} or similar
26150 construct may contain other nested @samp{%} constructs or spaces, or
26151 even newlines. They are processed as usual, as described above.
26152 Trailing white space in @code{X} is ignored. White space may also
26153 appear anywhere on the left side of the colon in these constructs,
26154 except between @code{.} or @code{*} and the corresponding word.
26156 The @option{-O}, @option{-f}, @option{-m}, and @option{-W} switches are
26157 handled specifically in these constructs. If another value of
26158 @option{-O} or the negated form of a @option{-f}, @option{-m}, or
26159 @option{-W} switch is found later in the command line, the earlier
26160 switch value is ignored, except with @{@code{S}*@} where @code{S} is
26161 just one letter, which passes all matching options.
26163 The character @samp{|} at the beginning of the predicate text is used to
26164 indicate that a command should be piped to the following command, but
26165 only if @option{-pipe} is specified.
26167 It is built into GCC which switches take arguments and which do not.
26168 (You might think it would be useful to generalize this to allow each
26169 compiler's spec to say which switches take arguments. But this cannot
26170 be done in a consistent fashion. GCC cannot even decide which input
26171 files have been specified without knowing which switches take arguments,
26172 and it must know which input files to compile in order to tell which
26175 GCC also knows implicitly that arguments starting in @option{-l} are to be
26176 treated as compiler output files, and passed to the linker in their
26177 proper position among the other output files.
26179 @node Environment Variables
26180 @section Environment Variables Affecting GCC
26181 @cindex environment variables
26183 @c man begin ENVIRONMENT
26184 This section describes several environment variables that affect how GCC
26185 operates. Some of them work by specifying directories or prefixes to use
26186 when searching for various kinds of files. Some are used to specify other
26187 aspects of the compilation environment.
26189 Note that you can also specify places to search using options such as
26190 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
26191 take precedence over places specified using environment variables, which
26192 in turn take precedence over those specified by the configuration of GCC@.
26193 @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint,
26194 GNU Compiler Collection (GCC) Internals}.
26199 @c @itemx LC_COLLATE
26201 @c @itemx LC_MONETARY
26202 @c @itemx LC_NUMERIC
26207 @c @findex LC_COLLATE
26208 @findex LC_MESSAGES
26209 @c @findex LC_MONETARY
26210 @c @findex LC_NUMERIC
26214 These environment variables control the way that GCC uses
26215 localization information which allows GCC to work with different
26216 national conventions. GCC inspects the locale categories
26217 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
26218 so. These locale categories can be set to any value supported by your
26219 installation. A typical value is @samp{en_GB.UTF-8} for English in the United
26220 Kingdom encoded in UTF-8.
26222 The @env{LC_CTYPE} environment variable specifies character
26223 classification. GCC uses it to determine the character boundaries in
26224 a string; this is needed for some multibyte encodings that contain quote
26225 and escape characters that are otherwise interpreted as a string
26228 The @env{LC_MESSAGES} environment variable specifies the language to
26229 use in diagnostic messages.
26231 If the @env{LC_ALL} environment variable is set, it overrides the value
26232 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
26233 and @env{LC_MESSAGES} default to the value of the @env{LANG}
26234 environment variable. If none of these variables are set, GCC
26235 defaults to traditional C English behavior.
26239 If @env{TMPDIR} is set, it specifies the directory to use for temporary
26240 files. GCC uses temporary files to hold the output of one stage of
26241 compilation which is to be used as input to the next stage: for example,
26242 the output of the preprocessor, which is the input to the compiler
26245 @item GCC_COMPARE_DEBUG
26246 @findex GCC_COMPARE_DEBUG
26247 Setting @env{GCC_COMPARE_DEBUG} is nearly equivalent to passing
26248 @option{-fcompare-debug} to the compiler driver. See the documentation
26249 of this option for more details.
26251 @item GCC_EXEC_PREFIX
26252 @findex GCC_EXEC_PREFIX
26253 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
26254 names of the subprograms executed by the compiler. No slash is added
26255 when this prefix is combined with the name of a subprogram, but you can
26256 specify a prefix that ends with a slash if you wish.
26258 If @env{GCC_EXEC_PREFIX} is not set, GCC attempts to figure out
26259 an appropriate prefix to use based on the pathname it is invoked with.
26261 If GCC cannot find the subprogram using the specified prefix, it
26262 tries looking in the usual places for the subprogram.
26264 The default value of @env{GCC_EXEC_PREFIX} is
26265 @file{@var{prefix}/lib/gcc/} where @var{prefix} is the prefix to
26266 the installed compiler. In many cases @var{prefix} is the value
26267 of @code{prefix} when you ran the @file{configure} script.
26269 Other prefixes specified with @option{-B} take precedence over this prefix.
26271 This prefix is also used for finding files such as @file{crt0.o} that are
26274 In addition, the prefix is used in an unusual way in finding the
26275 directories to search for header files. For each of the standard
26276 directories whose name normally begins with @samp{/usr/local/lib/gcc}
26277 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
26278 replacing that beginning with the specified prefix to produce an
26279 alternate directory name. Thus, with @option{-Bfoo/}, GCC searches
26280 @file{foo/bar} just before it searches the standard directory
26281 @file{/usr/local/lib/bar}.
26282 If a standard directory begins with the configured
26283 @var{prefix} then the value of @var{prefix} is replaced by
26284 @env{GCC_EXEC_PREFIX} when looking for header files.
26286 @item COMPILER_PATH
26287 @findex COMPILER_PATH
26288 The value of @env{COMPILER_PATH} is a colon-separated list of
26289 directories, much like @env{PATH}. GCC tries the directories thus
26290 specified when searching for subprograms, if it can't find the
26291 subprograms using @env{GCC_EXEC_PREFIX}.
26294 @findex LIBRARY_PATH
26295 The value of @env{LIBRARY_PATH} is a colon-separated list of
26296 directories, much like @env{PATH}. When configured as a native compiler,
26297 GCC tries the directories thus specified when searching for special
26298 linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking
26299 using GCC also uses these directories when searching for ordinary
26300 libraries for the @option{-l} option (but directories specified with
26301 @option{-L} come first).
26305 @cindex locale definition
26306 This variable is used to pass locale information to the compiler. One way in
26307 which this information is used is to determine the character set to be used
26308 when character literals, string literals and comments are parsed in C and C++.
26309 When the compiler is configured to allow multibyte characters,
26310 the following values for @env{LANG} are recognized:
26314 Recognize JIS characters.
26316 Recognize SJIS characters.
26318 Recognize EUCJP characters.
26321 If @env{LANG} is not defined, or if it has some other value, then the
26322 compiler uses @code{mblen} and @code{mbtowc} as defined by the default locale to
26323 recognize and translate multibyte characters.
26327 Some additional environment variables affect the behavior of the
26330 @include cppenv.texi
26334 @node Precompiled Headers
26335 @section Using Precompiled Headers
26336 @cindex precompiled headers
26337 @cindex speed of compilation
26339 Often large projects have many header files that are included in every
26340 source file. The time the compiler takes to process these header files
26341 over and over again can account for nearly all of the time required to
26342 build the project. To make builds faster, GCC allows you to
26343 @dfn{precompile} a header file.
26345 To create a precompiled header file, simply compile it as you would any
26346 other file, if necessary using the @option{-x} option to make the driver
26347 treat it as a C or C++ header file. You may want to use a
26348 tool like @command{make} to keep the precompiled header up-to-date when
26349 the headers it contains change.
26351 A precompiled header file is searched for when @code{#include} is
26352 seen in the compilation. As it searches for the included file
26353 (@pxref{Search Path,,Search Path,cpp,The C Preprocessor}) the
26354 compiler looks for a precompiled header in each directory just before it
26355 looks for the include file in that directory. The name searched for is
26356 the name specified in the @code{#include} with @samp{.gch} appended. If
26357 the precompiled header file can't be used, it is ignored.
26359 For instance, if you have @code{#include "all.h"}, and you have
26360 @file{all.h.gch} in the same directory as @file{all.h}, then the
26361 precompiled header file is used if possible, and the original
26362 header is used otherwise.
26364 Alternatively, you might decide to put the precompiled header file in a
26365 directory and use @option{-I} to ensure that directory is searched
26366 before (or instead of) the directory containing the original header.
26367 Then, if you want to check that the precompiled header file is always
26368 used, you can put a file of the same name as the original header in this
26369 directory containing an @code{#error} command.
26371 This also works with @option{-include}. So yet another way to use
26372 precompiled headers, good for projects not designed with precompiled
26373 header files in mind, is to simply take most of the header files used by
26374 a project, include them from another header file, precompile that header
26375 file, and @option{-include} the precompiled header. If the header files
26376 have guards against multiple inclusion, they are skipped because
26377 they've already been included (in the precompiled header).
26379 If you need to precompile the same header file for different
26380 languages, targets, or compiler options, you can instead make a
26381 @emph{directory} named like @file{all.h.gch}, and put each precompiled
26382 header in the directory, perhaps using @option{-o}. It doesn't matter
26383 what you call the files in the directory; every precompiled header in
26384 the directory is considered. The first precompiled header
26385 encountered in the directory that is valid for this compilation is
26386 used; they're searched in no particular order.
26388 There are many other possibilities, limited only by your imagination,
26389 good sense, and the constraints of your build system.
26391 A precompiled header file can be used only when these conditions apply:
26395 Only one precompiled header can be used in a particular compilation.
26398 A precompiled header can't be used once the first C token is seen. You
26399 can have preprocessor directives before a precompiled header; you cannot
26400 include a precompiled header from inside another header.
26403 The precompiled header file must be produced for the same language as
26404 the current compilation. You can't use a C precompiled header for a C++
26408 The precompiled header file must have been produced by the same compiler
26409 binary as the current compilation is using.
26412 Any macros defined before the precompiled header is included must
26413 either be defined in the same way as when the precompiled header was
26414 generated, or must not affect the precompiled header, which usually
26415 means that they don't appear in the precompiled header at all.
26417 The @option{-D} option is one way to define a macro before a
26418 precompiled header is included; using a @code{#define} can also do it.
26419 There are also some options that define macros implicitly, like
26420 @option{-O} and @option{-Wdeprecated}; the same rule applies to macros
26423 @item If debugging information is output when using the precompiled
26424 header, using @option{-g} or similar, the same kind of debugging information
26425 must have been output when building the precompiled header. However,
26426 a precompiled header built using @option{-g} can be used in a compilation
26427 when no debugging information is being output.
26429 @item The same @option{-m} options must generally be used when building
26430 and using the precompiled header. @xref{Submodel Options},
26431 for any cases where this rule is relaxed.
26433 @item Each of the following options must be the same when building and using
26434 the precompiled header:
26436 @gccoptlist{-fexceptions}
26439 Some other command-line options starting with @option{-f},
26440 @option{-p}, or @option{-O} must be defined in the same way as when
26441 the precompiled header was generated. At present, it's not clear
26442 which options are safe to change and which are not; the safest choice
26443 is to use exactly the same options when generating and using the
26444 precompiled header. The following are known to be safe:
26446 @gccoptlist{-fmessage-length= -fpreprocessed -fsched-interblock @gol
26447 -fsched-spec -fsched-spec-load -fsched-spec-load-dangerous @gol
26448 -fsched-verbose=@var{number} -fschedule-insns -fvisibility= @gol
26453 For all of these except the last, the compiler automatically
26454 ignores the precompiled header if the conditions aren't met. If you
26455 find an option combination that doesn't work and doesn't cause the
26456 precompiled header to be ignored, please consider filing a bug report,
26459 If you do use differing options when generating and using the
26460 precompiled header, the actual behavior is a mixture of the
26461 behavior for the options. For instance, if you use @option{-g} to
26462 generate the precompiled header but not when using it, you may or may
26463 not get debugging information for routines in the precompiled header.