1 @c Copyright (C) 1988-2017 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{} 1988-2017 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), dbx(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 -fpermitted-flt-eval-methods=@var{standard} @gol
182 -aux-info @var{filename} -fallow-parameterless-variadic-functions @gol
183 -fno-asm -fno-builtin -fno-builtin-@var{function} -fgimple@gol
184 -fhosted -ffreestanding -fopenacc -fopenmp -fopenmp-simd @gol
185 -fms-extensions -fplan9-extensions -fsso-struct=@var{endianness} @gol
186 -fallow-single-precision -fcond-mismatch -flax-vector-conversions @gol
187 -fsigned-bitfields -fsigned-char @gol
188 -funsigned-bitfields -funsigned-char}
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 -fnew-ttp-matching @gol
204 -fno-nonansi-builtins -fnothrow-opt -fno-operator-names @gol
205 -fno-optional-diags -fpermissive @gol
206 -fno-pretty-templates @gol
207 -frepo -fno-rtti -fsized-deallocation @gol
208 -ftemplate-backtrace-limit=@var{n} @gol
209 -ftemplate-depth=@var{n} @gol
210 -fno-threadsafe-statics -fuse-cxa-atexit @gol
211 -fno-weak -nostdinc++ @gol
212 -fvisibility-inlines-hidden @gol
213 -fvisibility-ms-compat @gol
214 -fext-numeric-literals @gol
215 -Wabi=@var{n} -Wabi-tag -Wconversion-null -Wctor-dtor-privacy @gol
216 -Wdelete-non-virtual-dtor -Wliteral-suffix -Wmultiple-inheritance @gol
217 -Wnamespaces -Wnarrowing @gol
218 -Wnoexcept -Wnoexcept-type -Wclass-memaccess @gol
219 -Wnon-virtual-dtor -Wreorder -Wregister @gol
220 -Weffc++ -Wstrict-null-sentinel -Wtemplates @gol
221 -Wno-non-template-friend -Wold-style-cast @gol
222 -Woverloaded-virtual -Wno-pmf-conversions @gol
223 -Wsign-promo -Wvirtual-inheritance}
225 @item Objective-C and Objective-C++ Language Options
226 @xref{Objective-C and Objective-C++ Dialect Options,,Options Controlling
227 Objective-C and Objective-C++ Dialects}.
228 @gccoptlist{-fconstant-string-class=@var{class-name} @gol
229 -fgnu-runtime -fnext-runtime @gol
230 -fno-nil-receivers @gol
231 -fobjc-abi-version=@var{n} @gol
232 -fobjc-call-cxx-cdtors @gol
233 -fobjc-direct-dispatch @gol
234 -fobjc-exceptions @gol
237 -fobjc-std=objc1 @gol
238 -fno-local-ivars @gol
239 -fivar-visibility=@r{[}public@r{|}protected@r{|}private@r{|}package@r{]} @gol
240 -freplace-objc-classes @gol
243 -Wassign-intercept @gol
244 -Wno-protocol -Wselector @gol
245 -Wstrict-selector-match @gol
246 -Wundeclared-selector}
248 @item Diagnostic Message Formatting Options
249 @xref{Diagnostic Message Formatting Options,,Options to Control Diagnostic Messages Formatting}.
250 @gccoptlist{-fmessage-length=@var{n} @gol
251 -fdiagnostics-show-location=@r{[}once@r{|}every-line@r{]} @gol
252 -fdiagnostics-color=@r{[}auto@r{|}never@r{|}always@r{]} @gol
253 -fno-diagnostics-show-option -fno-diagnostics-show-caret @gol
254 -fdiagnostics-parseable-fixits -fdiagnostics-generate-patch @gol
255 -fdiagnostics-show-template-tree -fno-elide-type @gol
258 @item Warning Options
259 @xref{Warning Options,,Options to Request or Suppress Warnings}.
260 @gccoptlist{-fsyntax-only -fmax-errors=@var{n} -Wpedantic @gol
261 -pedantic-errors @gol
262 -w -Wextra -Wall -Waddress -Waggregate-return @gol
263 -Walloc-zero -Walloc-size-larger-than=@var{n}
264 -Walloca -Walloca-larger-than=@var{n} @gol
265 -Wno-aggressive-loop-optimizations -Warray-bounds -Warray-bounds=@var{n} @gol
266 -Wno-attributes -Wbool-compare -Wbool-operation @gol
267 -Wno-builtin-declaration-mismatch @gol
268 -Wno-builtin-macro-redefined -Wc90-c99-compat -Wc99-c11-compat @gol
269 -Wc++-compat -Wc++11-compat -Wc++14-compat @gol
270 -Wcast-align -Wcast-align=strict -Wcast-qual @gol
271 -Wchar-subscripts -Wchkp -Wcatch-value -Wcatch-value=@var{n} @gol
272 -Wclobbered -Wcomment -Wconditionally-supported @gol
273 -Wconversion -Wcoverage-mismatch -Wno-cpp -Wdangling-else -Wdate-time @gol
274 -Wdelete-incomplete @gol
275 -Wno-deprecated -Wno-deprecated-declarations -Wno-designated-init @gol
276 -Wdisabled-optimization @gol
277 -Wno-discarded-qualifiers -Wno-discarded-array-qualifiers @gol
278 -Wno-div-by-zero -Wdouble-promotion @gol
279 -Wduplicated-branches -Wduplicated-cond @gol
280 -Wempty-body -Wenum-compare -Wno-endif-labels -Wexpansion-to-defined @gol
281 -Werror -Werror=* -Wextra-semi -Wfatal-errors @gol
282 -Wfloat-equal -Wformat -Wformat=2 @gol
283 -Wno-format-contains-nul -Wno-format-extra-args @gol
284 -Wformat-nonliteral -Wformat-overflow=@var{n} @gol
285 -Wformat-security -Wformat-signedness -Wformat-truncation=@var{n} @gol
286 -Wformat-y2k -Wframe-address @gol
287 -Wframe-larger-than=@var{len} -Wno-free-nonheap-object -Wjump-misses-init @gol
288 -Wif-not-aligned @gol
289 -Wignored-qualifiers -Wignored-attributes -Wincompatible-pointer-types @gol
290 -Wimplicit -Wimplicit-fallthrough -Wimplicit-fallthrough=@var{n} @gol
291 -Wimplicit-function-declaration -Wimplicit-int @gol
292 -Winit-self -Winline -Wno-int-conversion -Wint-in-bool-context @gol
293 -Wno-int-to-pointer-cast -Winvalid-memory-model -Wno-invalid-offsetof @gol
294 -Winvalid-pch -Wlarger-than=@var{len} @gol
295 -Wlogical-op -Wlogical-not-parentheses -Wlong-long @gol
296 -Wmain -Wmaybe-uninitialized -Wmemset-elt-size -Wmemset-transposed-args @gol
297 -Wmisleading-indentation -Wmissing-braces @gol
298 -Wmissing-field-initializers -Wmissing-include-dirs @gol
299 -Wno-multichar -Wmultistatement-macros -Wnonnull -Wnonnull-compare @gol
300 -Wnormalized=@r{[}none@r{|}id@r{|}nfc@r{|}nfkc@r{]} @gol
301 -Wnull-dereference -Wodr -Wno-overflow -Wopenmp-simd @gol
302 -Woverride-init-side-effects -Woverlength-strings @gol
303 -Wpacked -Wpacked-bitfield-compat -Wpacked-not-aligned -Wpadded @gol
304 -Wparentheses -Wno-pedantic-ms-format @gol
305 -Wplacement-new -Wplacement-new=@var{n} @gol
306 -Wpointer-arith -Wpointer-compare -Wno-pointer-to-int-cast @gol
307 -Wno-pragmas -Wredundant-decls -Wrestrict -Wno-return-local-addr @gol
308 -Wreturn-type -Wsequence-point -Wshadow -Wno-shadow-ivar @gol
309 -Wshadow=global, -Wshadow=local, -Wshadow=compatible-local @gol
310 -Wshift-overflow -Wshift-overflow=@var{n} @gol
311 -Wshift-count-negative -Wshift-count-overflow -Wshift-negative-value @gol
312 -Wsign-compare -Wsign-conversion -Wfloat-conversion @gol
313 -Wno-scalar-storage-order -Wsizeof-pointer-div @gol
314 -Wsizeof-pointer-memaccess -Wsizeof-array-argument @gol
315 -Wstack-protector -Wstack-usage=@var{len} -Wstrict-aliasing @gol
316 -Wstrict-aliasing=n -Wstrict-overflow -Wstrict-overflow=@var{n} @gol
317 -Wstringop-overflow=@var{n} -Wstringop-truncation @gol
318 -Wsuggest-attribute=@r{[}pure@r{|}const@r{|}noreturn@r{|}format@r{|}malloc@r{]} @gol
319 -Wsuggest-final-types @gol -Wsuggest-final-methods -Wsuggest-override @gol
320 -Wmissing-format-attribute -Wsubobject-linkage @gol
321 -Wswitch -Wswitch-bool -Wswitch-default -Wswitch-enum @gol
322 -Wswitch-unreachable -Wsync-nand @gol
323 -Wsystem-headers -Wtautological-compare -Wtrampolines -Wtrigraphs @gol
324 -Wtype-limits -Wundef @gol
325 -Wuninitialized -Wunknown-pragmas -Wunsafe-loop-optimizations @gol
326 -Wunsuffixed-float-constants -Wunused -Wunused-function @gol
327 -Wunused-label -Wunused-local-typedefs -Wunused-macros @gol
328 -Wunused-parameter -Wno-unused-result @gol
329 -Wunused-value -Wunused-variable @gol
330 -Wunused-const-variable -Wunused-const-variable=@var{n} @gol
331 -Wunused-but-set-parameter -Wunused-but-set-variable @gol
332 -Wuseless-cast -Wvariadic-macros -Wvector-operation-performance @gol
333 -Wvla -Wvla-larger-than=@var{n} -Wvolatile-register-var -Wwrite-strings @gol
334 -Wzero-as-null-pointer-constant -Whsa}
336 @item C and Objective-C-only Warning Options
337 @gccoptlist{-Wbad-function-cast -Wmissing-declarations @gol
338 -Wmissing-parameter-type -Wmissing-prototypes -Wnested-externs @gol
339 -Wold-style-declaration -Wold-style-definition @gol
340 -Wstrict-prototypes -Wtraditional -Wtraditional-conversion @gol
341 -Wdeclaration-after-statement -Wpointer-sign}
343 @item Debugging Options
344 @xref{Debugging Options,,Options for Debugging Your Program}.
345 @gccoptlist{-g -g@var{level} -gdwarf -gdwarf-@var{version} @gol
346 -ggdb -grecord-gcc-switches -gno-record-gcc-switches @gol
347 -gstabs -gstabs+ -gstrict-dwarf -gno-strict-dwarf @gol
348 -gcolumn-info -gno-column-info @gol
349 -gvms -gxcoff -gxcoff+ -gz@r{[}=@var{type}@r{]} @gol
350 -fdebug-prefix-map=@var{old}=@var{new} -fdebug-types-section @gol
351 -fno-eliminate-unused-debug-types @gol
352 -femit-struct-debug-baseonly -femit-struct-debug-reduced @gol
353 -femit-struct-debug-detailed@r{[}=@var{spec-list}@r{]} @gol
354 -feliminate-unused-debug-symbols -femit-class-debug-always @gol
355 -fno-merge-debug-strings -fno-dwarf2-cfi-asm @gol
356 -fvar-tracking -fvar-tracking-assignments}
358 @item Optimization Options
359 @xref{Optimize Options,,Options that Control Optimization}.
360 @gccoptlist{-faggressive-loop-optimizations -falign-functions[=@var{n}] @gol
361 -falign-jumps[=@var{n}] @gol
362 -falign-labels[=@var{n}] -falign-loops[=@var{n}] @gol
363 -fassociative-math -fauto-profile -fauto-profile[=@var{path}] @gol
364 -fauto-inc-dec -fbranch-probabilities @gol
365 -fbranch-target-load-optimize -fbranch-target-load-optimize2 @gol
366 -fbtr-bb-exclusive -fcaller-saves @gol
367 -fcombine-stack-adjustments -fconserve-stack @gol
368 -fcompare-elim -fcprop-registers -fcrossjumping @gol
369 -fcse-follow-jumps -fcse-skip-blocks -fcx-fortran-rules @gol
370 -fcx-limited-range @gol
371 -fdata-sections -fdce -fdelayed-branch @gol
372 -fdelete-null-pointer-checks -fdevirtualize -fdevirtualize-speculatively @gol
373 -fdevirtualize-at-ltrans -fdse @gol
374 -fearly-inlining -fipa-sra -fexpensive-optimizations -ffat-lto-objects @gol
375 -ffast-math -ffinite-math-only -ffloat-store -fexcess-precision=@var{style} @gol
376 -fforward-propagate -ffp-contract=@var{style} -ffunction-sections @gol
377 -fgcse -fgcse-after-reload -fgcse-las -fgcse-lm -fgraphite-identity @gol
378 -fgcse-sm -fhoist-adjacent-loads -fif-conversion @gol
379 -fif-conversion2 -findirect-inlining @gol
380 -finline-functions -finline-functions-called-once -finline-limit=@var{n} @gol
381 -finline-small-functions -fipa-cp -fipa-cp-clone @gol
382 -fipa-bit-cp -fipa-vrp @gol
383 -fipa-pta -fipa-profile -fipa-pure-const -fipa-reference -fipa-icf @gol
384 -fira-algorithm=@var{algorithm} @gol
385 -fira-region=@var{region} -fira-hoist-pressure @gol
386 -fira-loop-pressure -fno-ira-share-save-slots @gol
387 -fno-ira-share-spill-slots @gol
388 -fisolate-erroneous-paths-dereference -fisolate-erroneous-paths-attribute @gol
389 -fivopts -fkeep-inline-functions -fkeep-static-functions @gol
390 -fkeep-static-consts -flimit-function-alignment -flive-range-shrinkage @gol
391 -floop-block -floop-interchange -floop-strip-mine @gol
392 -floop-unroll-and-jam -floop-nest-optimize @gol
393 -floop-parallelize-all -flra-remat -flto -flto-compression-level @gol
394 -flto-partition=@var{alg} -fmerge-all-constants @gol
395 -fmerge-constants -fmodulo-sched -fmodulo-sched-allow-regmoves @gol
396 -fmove-loop-invariants -fno-branch-count-reg @gol
397 -fno-defer-pop -fno-fp-int-builtin-inexact -fno-function-cse @gol
398 -fno-guess-branch-probability -fno-inline -fno-math-errno -fno-peephole @gol
399 -fno-peephole2 -fno-printf-return-value -fno-sched-interblock @gol
400 -fno-sched-spec -fno-signed-zeros @gol
401 -fno-toplevel-reorder -fno-trapping-math -fno-zero-initialized-in-bss @gol
402 -fomit-frame-pointer -foptimize-sibling-calls @gol
403 -fpartial-inlining -fpeel-loops -fpredictive-commoning @gol
404 -fprefetch-loop-arrays @gol
405 -fprofile-correction @gol
406 -fprofile-use -fprofile-use=@var{path} -fprofile-values @gol
407 -fprofile-reorder-functions @gol
408 -freciprocal-math -free -frename-registers -freorder-blocks @gol
409 -freorder-blocks-algorithm=@var{algorithm} @gol
410 -freorder-blocks-and-partition -freorder-functions @gol
411 -frerun-cse-after-loop -freschedule-modulo-scheduled-loops @gol
412 -frounding-math -fsched2-use-superblocks -fsched-pressure @gol
413 -fsched-spec-load -fsched-spec-load-dangerous @gol
414 -fsched-stalled-insns-dep[=@var{n}] -fsched-stalled-insns[=@var{n}] @gol
415 -fsched-group-heuristic -fsched-critical-path-heuristic @gol
416 -fsched-spec-insn-heuristic -fsched-rank-heuristic @gol
417 -fsched-last-insn-heuristic -fsched-dep-count-heuristic @gol
418 -fschedule-fusion @gol
419 -fschedule-insns -fschedule-insns2 -fsection-anchors @gol
420 -fselective-scheduling -fselective-scheduling2 @gol
421 -fsel-sched-pipelining -fsel-sched-pipelining-outer-loops @gol
422 -fsemantic-interposition -fshrink-wrap -fshrink-wrap-separate @gol
423 -fsignaling-nans @gol
424 -fsingle-precision-constant -fsplit-ivs-in-unroller -fsplit-loops@gol
426 -fsplit-wide-types -fssa-backprop -fssa-phiopt @gol
427 -fstdarg-opt -fstore-merging -fstrict-aliasing @gol
428 -fthread-jumps -ftracer -ftree-bit-ccp @gol
429 -ftree-builtin-call-dce -ftree-ccp -ftree-ch @gol
430 -ftree-coalesce-vars -ftree-copy-prop -ftree-dce -ftree-dominator-opts @gol
431 -ftree-dse -ftree-forwprop -ftree-fre -fcode-hoisting @gol
432 -ftree-loop-if-convert -ftree-loop-im @gol
433 -ftree-phiprop -ftree-loop-distribution -ftree-loop-distribute-patterns @gol
434 -ftree-loop-ivcanon -ftree-loop-linear -ftree-loop-optimize @gol
435 -ftree-loop-vectorize @gol
436 -ftree-parallelize-loops=@var{n} -ftree-pre -ftree-partial-pre -ftree-pta @gol
437 -ftree-reassoc -ftree-sink -ftree-slsr -ftree-sra @gol
438 -ftree-switch-conversion -ftree-tail-merge @gol
439 -ftree-ter -ftree-vectorize -ftree-vrp -funconstrained-commons @gol
440 -funit-at-a-time -funroll-all-loops -funroll-loops @gol
441 -funsafe-math-optimizations -funswitch-loops @gol
442 -fipa-ra -fvariable-expansion-in-unroller -fvect-cost-model -fvpt @gol
443 -fweb -fwhole-program -fwpa -fuse-linker-plugin @gol
444 --param @var{name}=@var{value}
445 -O -O0 -O1 -O2 -O3 -Os -Ofast -Og}
447 @item Program Instrumentation Options
448 @xref{Instrumentation Options,,Program Instrumentation Options}.
449 @gccoptlist{-p -pg -fprofile-arcs --coverage -ftest-coverage @gol
450 -fprofile-abs-path @gol
451 -fprofile-dir=@var{path} -fprofile-generate -fprofile-generate=@var{path} @gol
452 -fsanitize=@var{style} -fsanitize-recover -fsanitize-recover=@var{style} @gol
453 -fasan-shadow-offset=@var{number} -fsanitize-sections=@var{s1},@var{s2},... @gol
454 -fsanitize-undefined-trap-on-error -fbounds-check @gol
455 -fcheck-pointer-bounds -fchkp-check-incomplete-type @gol
456 -fchkp-first-field-has-own-bounds -fchkp-narrow-bounds @gol
457 -fchkp-narrow-to-innermost-array -fchkp-optimize @gol
458 -fchkp-use-fast-string-functions -fchkp-use-nochk-string-functions @gol
459 -fchkp-use-static-bounds -fchkp-use-static-const-bounds @gol
460 -fchkp-treat-zero-dynamic-size-as-infinite -fchkp-check-read @gol
461 -fchkp-check-read -fchkp-check-write -fchkp-store-bounds @gol
462 -fchkp-instrument-calls -fchkp-instrument-marked-only @gol
463 -fchkp-use-wrappers -fchkp-flexible-struct-trailing-arrays@gol
464 -fcf-protection==@r{[}full@r{|}branch@r{|}return@r{|}none@r{]} @gol
465 -fstack-protector -fstack-protector-all -fstack-protector-strong @gol
466 -fstack-protector-explicit -fstack-check @gol
467 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
468 -fno-stack-limit -fsplit-stack @gol
469 -fvtable-verify=@r{[}std@r{|}preinit@r{|}none@r{]} @gol
470 -fvtv-counts -fvtv-debug @gol
471 -finstrument-functions @gol
472 -finstrument-functions-exclude-function-list=@var{sym},@var{sym},@dots{} @gol
473 -finstrument-functions-exclude-file-list=@var{file},@var{file},@dots{}}
475 @item Preprocessor Options
476 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
477 @gccoptlist{-A@var{question}=@var{answer} @gol
478 -A-@var{question}@r{[}=@var{answer}@r{]} @gol
479 -C -CC -D@var{macro}@r{[}=@var{defn}@r{]} @gol
480 -dD -dI -dM -dN -dU @gol
481 -fdebug-cpp -fdirectives-only -fdollars-in-identifiers @gol
482 -fexec-charset=@var{charset} -fextended-identifiers @gol
483 -finput-charset=@var{charset} -fno-canonical-system-headers @gol
484 -fpch-deps -fpch-preprocess -fpreprocessed @gol
485 -ftabstop=@var{width} -ftrack-macro-expansion @gol
486 -fwide-exec-charset=@var{charset} -fworking-directory @gol
487 -H -imacros @var{file} -include @var{file} @gol
488 -M -MD -MF -MG -MM -MMD -MP -MQ -MT @gol
489 -no-integrated-cpp -P -pthread -remap @gol
490 -traditional -traditional-cpp -trigraphs @gol
491 -U@var{macro} -undef @gol
492 -Wp,@var{option} -Xpreprocessor @var{option}}
494 @item Assembler Options
495 @xref{Assembler Options,,Passing Options to the Assembler}.
496 @gccoptlist{-Wa,@var{option} -Xassembler @var{option}}
499 @xref{Link Options,,Options for Linking}.
500 @gccoptlist{@var{object-file-name} -fuse-ld=@var{linker} -l@var{library} @gol
501 -nostartfiles -nodefaultlibs -nostdlib -pie -pthread -rdynamic @gol
502 -s -static -static-pie -static-libgcc -static-libstdc++ @gol
503 -static-libasan -static-libtsan -static-liblsan -static-libubsan @gol
504 -static-libmpx -static-libmpxwrappers @gol
505 -shared -shared-libgcc -symbolic @gol
506 -T @var{script} -Wl,@var{option} -Xlinker @var{option} @gol
507 -u @var{symbol} -z @var{keyword}}
509 @item Directory Options
510 @xref{Directory Options,,Options for Directory Search}.
511 @gccoptlist{-B@var{prefix} -I@var{dir} -I- @gol
512 -idirafter @var{dir} @gol
513 -imacros @var{file} -imultilib @var{dir} @gol
514 -iplugindir=@var{dir} -iprefix @var{file} @gol
515 -iquote @var{dir} -isysroot @var{dir} -isystem @var{dir} @gol
516 -iwithprefix @var{dir} -iwithprefixbefore @var{dir} @gol
517 -L@var{dir} -no-canonical-prefixes --no-sysroot-suffix @gol
518 -nostdinc -nostdinc++ --sysroot=@var{dir}}
520 @item Code Generation Options
521 @xref{Code Gen Options,,Options for Code Generation Conventions}.
522 @gccoptlist{-fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
523 -ffixed-@var{reg} -fexceptions @gol
524 -fnon-call-exceptions -fdelete-dead-exceptions -funwind-tables @gol
525 -fasynchronous-unwind-tables @gol
527 -finhibit-size-directive -fno-common -fno-ident @gol
528 -fpcc-struct-return -fpic -fPIC -fpie -fPIE -fno-plt @gol
529 -fno-jump-tables @gol
530 -frecord-gcc-switches @gol
531 -freg-struct-return -fshort-enums -fshort-wchar @gol
532 -fverbose-asm -fpack-struct[=@var{n}] @gol
533 -fleading-underscore -ftls-model=@var{model} @gol
534 -fstack-reuse=@var{reuse_level} @gol
535 -ftrampolines -ftrapv -fwrapv @gol
536 -fvisibility=@r{[}default@r{|}internal@r{|}hidden@r{|}protected@r{]} @gol
537 -fstrict-volatile-bitfields -fsync-libcalls}
539 @item Developer Options
540 @xref{Developer Options,,GCC Developer Options}.
541 @gccoptlist{-d@var{letters} -dumpspecs -dumpmachine -dumpversion @gol
542 -dumpfullversion -fchecking -fchecking=@var{n} -fdbg-cnt-list @gol
543 -fdbg-cnt=@var{counter-value-list} @gol
544 -fdisable-ipa-@var{pass_name} @gol
545 -fdisable-rtl-@var{pass_name} @gol
546 -fdisable-rtl-@var{pass-name}=@var{range-list} @gol
547 -fdisable-tree-@var{pass_name} @gol
548 -fdisable-tree-@var{pass-name}=@var{range-list} @gol
549 -fdump-noaddr -fdump-unnumbered -fdump-unnumbered-links @gol
550 -fdump-class-hierarchy@r{[}-@var{n}@r{]} @gol
551 -fdump-final-insns@r{[}=@var{file}@r{]}
552 -fdump-ipa-all -fdump-ipa-cgraph -fdump-ipa-inline @gol
554 -fdump-lang-@var{switch} @gol
555 -fdump-lang-@var{switch}-@var{options} @gol
556 -fdump-lang-@var{switch}-@var{options}=@var{filename} @gol
558 -fdump-rtl-@var{pass} -fdump-rtl-@var{pass}=@var{filename} @gol
559 -fdump-statistics @gol
561 -fdump-tree-@var{switch} @gol
562 -fdump-tree-@var{switch}-@var{options} @gol
563 -fdump-tree-@var{switch}-@var{options}=@var{filename} @gol
564 -fcompare-debug@r{[}=@var{opts}@r{]} -fcompare-debug-second @gol
565 -fenable-@var{kind}-@var{pass} @gol
566 -fenable-@var{kind}-@var{pass}=@var{range-list} @gol
567 -fira-verbose=@var{n} @gol
568 -flto-report -flto-report-wpa -fmem-report-wpa @gol
569 -fmem-report -fpre-ipa-mem-report -fpost-ipa-mem-report @gol
570 -fopt-info -fopt-info-@var{options}@r{[}=@var{file}@r{]} @gol
571 -fprofile-report @gol
572 -frandom-seed=@var{string} -fsched-verbose=@var{n} @gol
573 -fsel-sched-verbose -fsel-sched-dump-cfg -fsel-sched-pipelining-verbose @gol
574 -fstats -fstack-usage -ftime-report -ftime-report-details @gol
575 -fvar-tracking-assignments-toggle -gtoggle @gol
576 -print-file-name=@var{library} -print-libgcc-file-name @gol
577 -print-multi-directory -print-multi-lib -print-multi-os-directory @gol
578 -print-prog-name=@var{program} -print-search-dirs -Q @gol
579 -print-sysroot -print-sysroot-headers-suffix @gol
580 -save-temps -save-temps=cwd -save-temps=obj -time@r{[}=@var{file}@r{]}}
582 @item Machine-Dependent Options
583 @xref{Submodel Options,,Machine-Dependent Options}.
584 @c This list is ordered alphanumerically by subsection name.
585 @c Try and put the significant identifier (CPU or system) first,
586 @c so users have a clue at guessing where the ones they want will be.
588 @emph{AArch64 Options}
589 @gccoptlist{-mabi=@var{name} -mbig-endian -mlittle-endian @gol
590 -mgeneral-regs-only @gol
591 -mcmodel=tiny -mcmodel=small -mcmodel=large @gol
593 -momit-leaf-frame-pointer @gol
594 -mtls-dialect=desc -mtls-dialect=traditional @gol
595 -mtls-size=@var{size} @gol
596 -mfix-cortex-a53-835769 -mfix-cortex-a53-843419 @gol
597 -mlow-precision-recip-sqrt -mlow-precision-sqrt -mlow-precision-div @gol
598 -mpc-relative-literal-loads @gol
599 -msign-return-address=@var{scope} @gol
600 -march=@var{name} -mcpu=@var{name} -mtune=@var{name} -moverride=@var{string}}
602 @emph{Adapteva Epiphany Options}
603 @gccoptlist{-mhalf-reg-file -mprefer-short-insn-regs @gol
604 -mbranch-cost=@var{num} -mcmove -mnops=@var{num} -msoft-cmpsf @gol
605 -msplit-lohi -mpost-inc -mpost-modify -mstack-offset=@var{num} @gol
606 -mround-nearest -mlong-calls -mshort-calls -msmall16 @gol
607 -mfp-mode=@var{mode} -mvect-double -max-vect-align=@var{num} @gol
608 -msplit-vecmove-early -m1reg-@var{reg}}
611 @gccoptlist{-mbarrel-shifter @gol
612 -mcpu=@var{cpu} -mA6 -mARC600 -mA7 -mARC700 @gol
613 -mdpfp -mdpfp-compact -mdpfp-fast -mno-dpfp-lrsr @gol
614 -mea -mno-mpy -mmul32x16 -mmul64 -matomic @gol
615 -mnorm -mspfp -mspfp-compact -mspfp-fast -msimd -msoft-float -mswap @gol
616 -mcrc -mdsp-packa -mdvbf -mlock -mmac-d16 -mmac-24 -mrtsc -mswape @gol
617 -mtelephony -mxy -misize -mannotate-align -marclinux -marclinux_prof @gol
618 -mlong-calls -mmedium-calls -msdata -mirq-ctrl-saved @gol
619 -mrgf-banked-regs -mlpc-width=@var{width} -G @var{num} @gol
620 -mvolatile-cache -mtp-regno=@var{regno} @gol
621 -malign-call -mauto-modify-reg -mbbit-peephole -mno-brcc @gol
622 -mcase-vector-pcrel -mcompact-casesi -mno-cond-exec -mearly-cbranchsi @gol
623 -mexpand-adddi -mindexed-loads -mlra -mlra-priority-none @gol
624 -mlra-priority-compact mlra-priority-noncompact -mno-millicode @gol
625 -mmixed-code -mq-class -mRcq -mRcw -msize-level=@var{level} @gol
626 -mtune=@var{cpu} -mmultcost=@var{num} @gol
627 -munalign-prob-threshold=@var{probability} -mmpy-option=@var{multo} @gol
628 -mdiv-rem -mcode-density -mll64 -mfpu=@var{fpu}}
631 @gccoptlist{-mapcs-frame -mno-apcs-frame @gol
632 -mabi=@var{name} @gol
633 -mapcs-stack-check -mno-apcs-stack-check @gol
634 -mapcs-reentrant -mno-apcs-reentrant @gol
635 -msched-prolog -mno-sched-prolog @gol
636 -mlittle-endian -mbig-endian @gol
638 -mfloat-abi=@var{name} @gol
639 -mfp16-format=@var{name}
640 -mthumb-interwork -mno-thumb-interwork @gol
641 -mcpu=@var{name} -march=@var{name} -mfpu=@var{name} @gol
642 -mtune=@var{name} -mprint-tune-info @gol
643 -mstructure-size-boundary=@var{n} @gol
644 -mabort-on-noreturn @gol
645 -mlong-calls -mno-long-calls @gol
646 -msingle-pic-base -mno-single-pic-base @gol
647 -mpic-register=@var{reg} @gol
648 -mnop-fun-dllimport @gol
649 -mpoke-function-name @gol
651 -mtpcs-frame -mtpcs-leaf-frame @gol
652 -mcaller-super-interworking -mcallee-super-interworking @gol
653 -mtp=@var{name} -mtls-dialect=@var{dialect} @gol
654 -mword-relocations @gol
655 -mfix-cortex-m3-ldrd @gol
656 -munaligned-access @gol
657 -mneon-for-64bits @gol
658 -mslow-flash-data @gol
659 -masm-syntax-unified @gol
665 @gccoptlist{-mmcu=@var{mcu} -mabsdata -maccumulate-args @gol
666 -mbranch-cost=@var{cost} @gol
667 -mcall-prologues -mgas-isr-prologues -mint8 @gol
668 -mn_flash=@var{size} -mno-interrupts @gol
669 -mrelax -mrmw -mstrict-X -mtiny-stack -mfract-convert-truncate @gol
670 -mshort-calls -nodevicelib @gol
671 -Waddr-space-convert -Wmisspelled-isr}
673 @emph{Blackfin Options}
674 @gccoptlist{-mcpu=@var{cpu}@r{[}-@var{sirevision}@r{]} @gol
675 -msim -momit-leaf-frame-pointer -mno-omit-leaf-frame-pointer @gol
676 -mspecld-anomaly -mno-specld-anomaly -mcsync-anomaly -mno-csync-anomaly @gol
677 -mlow-64k -mno-low64k -mstack-check-l1 -mid-shared-library @gol
678 -mno-id-shared-library -mshared-library-id=@var{n} @gol
679 -mleaf-id-shared-library -mno-leaf-id-shared-library @gol
680 -msep-data -mno-sep-data -mlong-calls -mno-long-calls @gol
681 -mfast-fp -minline-plt -mmulticore -mcorea -mcoreb -msdram @gol
685 @gccoptlist{-mbig-endian -mlittle-endian -march=@var{cpu} @gol
686 -msim -msdata=@var{sdata-type}}
689 @gccoptlist{-mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol
690 -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol
691 -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol
692 -mstack-align -mdata-align -mconst-align @gol
693 -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol
694 -melf -maout -melinux -mlinux -sim -sim2 @gol
695 -mmul-bug-workaround -mno-mul-bug-workaround}
698 @gccoptlist{-mmac @gol
699 -mcr16cplus -mcr16c @gol
700 -msim -mint32 -mbit-ops
701 -mdata-model=@var{model}}
703 @emph{Darwin Options}
704 @gccoptlist{-all_load -allowable_client -arch -arch_errors_fatal @gol
705 -arch_only -bind_at_load -bundle -bundle_loader @gol
706 -client_name -compatibility_version -current_version @gol
708 -dependency-file -dylib_file -dylinker_install_name @gol
709 -dynamic -dynamiclib -exported_symbols_list @gol
710 -filelist -flat_namespace -force_cpusubtype_ALL @gol
711 -force_flat_namespace -headerpad_max_install_names @gol
713 -image_base -init -install_name -keep_private_externs @gol
714 -multi_module -multiply_defined -multiply_defined_unused @gol
715 -noall_load -no_dead_strip_inits_and_terms @gol
716 -nofixprebinding -nomultidefs -noprebind -noseglinkedit @gol
717 -pagezero_size -prebind -prebind_all_twolevel_modules @gol
718 -private_bundle -read_only_relocs -sectalign @gol
719 -sectobjectsymbols -whyload -seg1addr @gol
720 -sectcreate -sectobjectsymbols -sectorder @gol
721 -segaddr -segs_read_only_addr -segs_read_write_addr @gol
722 -seg_addr_table -seg_addr_table_filename -seglinkedit @gol
723 -segprot -segs_read_only_addr -segs_read_write_addr @gol
724 -single_module -static -sub_library -sub_umbrella @gol
725 -twolevel_namespace -umbrella -undefined @gol
726 -unexported_symbols_list -weak_reference_mismatches @gol
727 -whatsloaded -F -gused -gfull -mmacosx-version-min=@var{version} @gol
728 -mkernel -mone-byte-bool}
730 @emph{DEC Alpha Options}
731 @gccoptlist{-mno-fp-regs -msoft-float @gol
732 -mieee -mieee-with-inexact -mieee-conformant @gol
733 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol
734 -mtrap-precision=@var{mode} -mbuild-constants @gol
735 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} @gol
736 -mbwx -mmax -mfix -mcix @gol
737 -mfloat-vax -mfloat-ieee @gol
738 -mexplicit-relocs -msmall-data -mlarge-data @gol
739 -msmall-text -mlarge-text @gol
740 -mmemory-latency=@var{time}}
743 @gccoptlist{-msmall-model -mno-lsim}
746 @gccoptlist{-msim -mlra -mnodiv -mft32b -mcompress -mnopm}
749 @gccoptlist{-mgpr-32 -mgpr-64 -mfpr-32 -mfpr-64 @gol
750 -mhard-float -msoft-float @gol
751 -malloc-cc -mfixed-cc -mdword -mno-dword @gol
752 -mdouble -mno-double @gol
753 -mmedia -mno-media -mmuladd -mno-muladd @gol
754 -mfdpic -minline-plt -mgprel-ro -multilib-library-pic @gol
755 -mlinked-fp -mlong-calls -malign-labels @gol
756 -mlibrary-pic -macc-4 -macc-8 @gol
757 -mpack -mno-pack -mno-eflags -mcond-move -mno-cond-move @gol
758 -moptimize-membar -mno-optimize-membar @gol
759 -mscc -mno-scc -mcond-exec -mno-cond-exec @gol
760 -mvliw-branch -mno-vliw-branch @gol
761 -mmulti-cond-exec -mno-multi-cond-exec -mnested-cond-exec @gol
762 -mno-nested-cond-exec -mtomcat-stats @gol
766 @emph{GNU/Linux Options}
767 @gccoptlist{-mglibc -muclibc -mmusl -mbionic -mandroid @gol
768 -tno-android-cc -tno-android-ld}
770 @emph{H8/300 Options}
771 @gccoptlist{-mrelax -mh -ms -mn -mexr -mno-exr -mint32 -malign-300}
774 @gccoptlist{-march=@var{architecture-type} @gol
775 -mcaller-copies -mdisable-fpregs -mdisable-indexing @gol
776 -mfast-indirect-calls -mgas -mgnu-ld -mhp-ld @gol
777 -mfixed-range=@var{register-range} @gol
778 -mjump-in-delay -mlinker-opt -mlong-calls @gol
779 -mlong-load-store -mno-disable-fpregs @gol
780 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
781 -mno-jump-in-delay -mno-long-load-store @gol
782 -mno-portable-runtime -mno-soft-float @gol
783 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
784 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
785 -mschedule=@var{cpu-type} -mspace-regs -msio -mwsio @gol
786 -munix=@var{unix-std} -nolibdld -static -threads}
789 @gccoptlist{-mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
790 -mvolatile-asm-stop -mregister-names -msdata -mno-sdata @gol
791 -mconstant-gp -mauto-pic -mfused-madd @gol
792 -minline-float-divide-min-latency @gol
793 -minline-float-divide-max-throughput @gol
794 -mno-inline-float-divide @gol
795 -minline-int-divide-min-latency @gol
796 -minline-int-divide-max-throughput @gol
797 -mno-inline-int-divide @gol
798 -minline-sqrt-min-latency -minline-sqrt-max-throughput @gol
799 -mno-inline-sqrt @gol
800 -mdwarf2-asm -mearly-stop-bits @gol
801 -mfixed-range=@var{register-range} -mtls-size=@var{tls-size} @gol
802 -mtune=@var{cpu-type} -milp32 -mlp64 @gol
803 -msched-br-data-spec -msched-ar-data-spec -msched-control-spec @gol
804 -msched-br-in-data-spec -msched-ar-in-data-spec -msched-in-control-spec @gol
805 -msched-spec-ldc -msched-spec-control-ldc @gol
806 -msched-prefer-non-data-spec-insns -msched-prefer-non-control-spec-insns @gol
807 -msched-stop-bits-after-every-cycle -msched-count-spec-in-critical-path @gol
808 -msel-sched-dont-check-control-spec -msched-fp-mem-deps-zero-cost @gol
809 -msched-max-memory-insns-hard-limit -msched-max-memory-insns=@var{max-insns}}
812 @gccoptlist{-mbarrel-shift-enabled -mdivide-enabled -mmultiply-enabled @gol
813 -msign-extend-enabled -muser-enabled}
815 @emph{M32R/D Options}
816 @gccoptlist{-m32r2 -m32rx -m32r @gol
818 -malign-loops -mno-align-loops @gol
819 -missue-rate=@var{number} @gol
820 -mbranch-cost=@var{number} @gol
821 -mmodel=@var{code-size-model-type} @gol
822 -msdata=@var{sdata-type} @gol
823 -mno-flush-func -mflush-func=@var{name} @gol
824 -mno-flush-trap -mflush-trap=@var{number} @gol
828 @gccoptlist{-mcpu=@var{cpu} -msim -memregs=@var{number}}
830 @emph{M680x0 Options}
831 @gccoptlist{-march=@var{arch} -mcpu=@var{cpu} -mtune=@var{tune} @gol
832 -m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
833 -m68060 -mcpu32 -m5200 -m5206e -m528x -m5307 -m5407 @gol
834 -mcfv4e -mbitfield -mno-bitfield -mc68000 -mc68020 @gol
835 -mnobitfield -mrtd -mno-rtd -mdiv -mno-div -mshort @gol
836 -mno-short -mhard-float -m68881 -msoft-float -mpcrel @gol
837 -malign-int -mstrict-align -msep-data -mno-sep-data @gol
838 -mshared-library-id=n -mid-shared-library -mno-id-shared-library @gol
839 -mxgot -mno-xgot -mlong-jump-table-offsets}
842 @gccoptlist{-mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
843 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
844 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
845 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
846 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
849 @gccoptlist{-mabsdiff -mall-opts -maverage -mbased=@var{n} -mbitops @gol
850 -mc=@var{n} -mclip -mconfig=@var{name} -mcop -mcop32 -mcop64 -mivc2 @gol
851 -mdc -mdiv -meb -mel -mio-volatile -ml -mleadz -mm -mminmax @gol
852 -mmult -mno-opts -mrepeat -ms -msatur -msdram -msim -msimnovec -mtf @gol
855 @emph{MicroBlaze Options}
856 @gccoptlist{-msoft-float -mhard-float -msmall-divides -mcpu=@var{cpu} @gol
857 -mmemcpy -mxl-soft-mul -mxl-soft-div -mxl-barrel-shift @gol
858 -mxl-pattern-compare -mxl-stack-check -mxl-gp-opt -mno-clearbss @gol
859 -mxl-multiply-high -mxl-float-convert -mxl-float-sqrt @gol
860 -mbig-endian -mlittle-endian -mxl-reorder -mxl-mode-@var{app-model}}
863 @gccoptlist{-EL -EB -march=@var{arch} -mtune=@var{arch} @gol
864 -mips1 -mips2 -mips3 -mips4 -mips32 -mips32r2 -mips32r3 -mips32r5 @gol
865 -mips32r6 -mips64 -mips64r2 -mips64r3 -mips64r5 -mips64r6 @gol
866 -mips16 -mno-mips16 -mflip-mips16 @gol
867 -minterlink-compressed -mno-interlink-compressed @gol
868 -minterlink-mips16 -mno-interlink-mips16 @gol
869 -mabi=@var{abi} -mabicalls -mno-abicalls @gol
870 -mshared -mno-shared -mplt -mno-plt -mxgot -mno-xgot @gol
871 -mgp32 -mgp64 -mfp32 -mfpxx -mfp64 -mhard-float -msoft-float @gol
872 -mno-float -msingle-float -mdouble-float @gol
873 -modd-spreg -mno-odd-spreg @gol
874 -mabs=@var{mode} -mnan=@var{encoding} @gol
875 -mdsp -mno-dsp -mdspr2 -mno-dspr2 @gol
878 -mvirt -mno-virt @gol
880 -mmicromips -mno-micromips @gol
882 -mfpu=@var{fpu-type} @gol
883 -msmartmips -mno-smartmips @gol
884 -mpaired-single -mno-paired-single -mdmx -mno-mdmx @gol
885 -mips3d -mno-mips3d -mmt -mno-mt -mllsc -mno-llsc @gol
886 -mlong64 -mlong32 -msym32 -mno-sym32 @gol
887 -G@var{num} -mlocal-sdata -mno-local-sdata @gol
888 -mextern-sdata -mno-extern-sdata -mgpopt -mno-gopt @gol
889 -membedded-data -mno-embedded-data @gol
890 -muninit-const-in-rodata -mno-uninit-const-in-rodata @gol
891 -mcode-readable=@var{setting} @gol
892 -msplit-addresses -mno-split-addresses @gol
893 -mexplicit-relocs -mno-explicit-relocs @gol
894 -mcheck-zero-division -mno-check-zero-division @gol
895 -mdivide-traps -mdivide-breaks @gol
896 -mload-store-pairs -mno-load-store-pairs @gol
897 -mmemcpy -mno-memcpy -mlong-calls -mno-long-calls @gol
898 -mmad -mno-mad -mimadd -mno-imadd -mfused-madd -mno-fused-madd -nocpp @gol
899 -mfix-24k -mno-fix-24k @gol
900 -mfix-r4000 -mno-fix-r4000 -mfix-r4400 -mno-fix-r4400 @gol
901 -mfix-r10000 -mno-fix-r10000 -mfix-rm7000 -mno-fix-rm7000 @gol
902 -mfix-vr4120 -mno-fix-vr4120 @gol
903 -mfix-vr4130 -mno-fix-vr4130 -mfix-sb1 -mno-fix-sb1 @gol
904 -mflush-func=@var{func} -mno-flush-func @gol
905 -mbranch-cost=@var{num} -mbranch-likely -mno-branch-likely @gol
906 -mcompact-branches=@var{policy} @gol
907 -mfp-exceptions -mno-fp-exceptions @gol
908 -mvr4130-align -mno-vr4130-align -msynci -mno-synci @gol
909 -mlxc1-sxc1 -mno-lxc1-sxc1 -mmadd4 -mno-madd4 @gol
910 -mrelax-pic-calls -mno-relax-pic-calls -mmcount-ra-address @gol
911 -mframe-header-opt -mno-frame-header-opt}
914 @gccoptlist{-mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
915 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
916 -melf -mbranch-predict -mno-branch-predict -mbase-addresses @gol
917 -mno-base-addresses -msingle-exit -mno-single-exit}
919 @emph{MN10300 Options}
920 @gccoptlist{-mmult-bug -mno-mult-bug @gol
921 -mno-am33 -mam33 -mam33-2 -mam34 @gol
922 -mtune=@var{cpu-type} @gol
923 -mreturn-pointer-on-d0 @gol
924 -mno-crt0 -mrelax -mliw -msetlb}
927 @gccoptlist{-meb -mel -mmul.x -mno-crt0}
929 @emph{MSP430 Options}
930 @gccoptlist{-msim -masm-hex -mmcu= -mcpu= -mlarge -msmall -mrelax @gol
932 -mcode-region= -mdata-region= @gol
933 -msilicon-errata= -msilicon-errata-warn= @gol
937 @gccoptlist{-mbig-endian -mlittle-endian @gol
938 -mreduced-regs -mfull-regs @gol
939 -mcmov -mno-cmov @gol
940 -mperf-ext -mno-perf-ext @gol
941 -mv3push -mno-v3push @gol
942 -m16bit -mno-16bit @gol
943 -misr-vector-size=@var{num} @gol
944 -mcache-block-size=@var{num} @gol
945 -march=@var{arch} @gol
946 -mcmodel=@var{code-model} @gol
949 @emph{Nios II Options}
950 @gccoptlist{-G @var{num} -mgpopt=@var{option} -mgpopt -mno-gpopt @gol
951 -mgprel-sec=@var{regexp} -mr0rel-sec=@var{regexp} @gol
953 -mno-bypass-cache -mbypass-cache @gol
954 -mno-cache-volatile -mcache-volatile @gol
955 -mno-fast-sw-div -mfast-sw-div @gol
956 -mhw-mul -mno-hw-mul -mhw-mulx -mno-hw-mulx -mno-hw-div -mhw-div @gol
957 -mcustom-@var{insn}=@var{N} -mno-custom-@var{insn} @gol
958 -mcustom-fpu-cfg=@var{name} @gol
959 -mhal -msmallc -msys-crt0=@var{name} -msys-lib=@var{name} @gol
960 -march=@var{arch} -mbmx -mno-bmx -mcdx -mno-cdx}
962 @emph{Nvidia PTX Options}
963 @gccoptlist{-m32 -m64 -mmainkernel -moptimize}
965 @emph{PDP-11 Options}
966 @gccoptlist{-mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol
967 -mbcopy -mbcopy-builtin -mint32 -mno-int16 @gol
968 -mint16 -mno-int32 -mfloat32 -mno-float64 @gol
969 -mfloat64 -mno-float32 -mabshi -mno-abshi @gol
970 -mbranch-expensive -mbranch-cheap @gol
971 -munix-asm -mdec-asm}
973 @emph{picoChip Options}
974 @gccoptlist{-mae=@var{ae_type} -mvliw-lookahead=@var{N} @gol
975 -msymbol-as-address -mno-inefficient-warnings}
977 @emph{PowerPC Options}
978 See RS/6000 and PowerPC Options.
980 @emph{RISC-V Options}
981 @gccoptlist{-mbranch-cost=@var{N-instruction} @gol
982 -mmemcpy -mno-memcpy @gol
984 -mabi=@var{ABI-string} @gol
985 -mfdiv -mno-fdiv @gol
987 -march=@var{ISA-string} @gol
988 -mtune=@var{processor-string} @gol
989 -msmall-data-limit=@var{N-bytes} @gol
990 -msave-restore -mno-save-restore @gol
991 -mstrict-align -mno-strict-align @gol
992 -mcmodel=medlow -mcmodel=medany @gol
993 -mexplicit-relocs -mno-explicit-relocs @gol}
996 @gccoptlist{-msim -mmul=none -mmul=g13 -mmul=g14 -mallregs @gol
997 -mcpu=g10 -mcpu=g13 -mcpu=g14 -mg10 -mg13 -mg14 @gol
998 -m64bit-doubles -m32bit-doubles -msave-mduc-in-interrupts}
1000 @emph{RS/6000 and PowerPC Options}
1001 @gccoptlist{-mcpu=@var{cpu-type} @gol
1002 -mtune=@var{cpu-type} @gol
1003 -mcmodel=@var{code-model} @gol
1005 -maltivec -mno-altivec @gol
1006 -mpowerpc-gpopt -mno-powerpc-gpopt @gol
1007 -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol
1008 -mmfcrf -mno-mfcrf -mpopcntb -mno-popcntb -mpopcntd -mno-popcntd @gol
1009 -mfprnd -mno-fprnd @gol
1010 -mcmpb -mno-cmpb -mmfpgpr -mno-mfpgpr -mhard-dfp -mno-hard-dfp @gol
1011 -mfull-toc -mminimal-toc -mno-fp-in-toc -mno-sum-in-toc @gol
1012 -m64 -m32 -mxl-compat -mno-xl-compat -mpe @gol
1013 -malign-power -malign-natural @gol
1014 -msoft-float -mhard-float -mmultiple -mno-multiple @gol
1015 -msingle-float -mdouble-float -msimple-fpu @gol
1016 -mstring -mno-string -mupdate -mno-update @gol
1017 -mavoid-indexed-addresses -mno-avoid-indexed-addresses @gol
1018 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol
1019 -mstrict-align -mno-strict-align -mrelocatable @gol
1020 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol
1021 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol
1022 -mdynamic-no-pic -maltivec -mswdiv -msingle-pic-base @gol
1023 -mprioritize-restricted-insns=@var{priority} @gol
1024 -msched-costly-dep=@var{dependence_type} @gol
1025 -minsert-sched-nops=@var{scheme} @gol
1026 -mcall-sysv -mcall-netbsd @gol
1027 -maix-struct-return -msvr4-struct-return @gol
1028 -mabi=@var{abi-type} -msecure-plt -mbss-plt @gol
1029 -mblock-move-inline-limit=@var{num} @gol
1030 -misel -mno-isel @gol
1031 -misel=yes -misel=no @gol
1033 -mspe=yes -mspe=no @gol
1035 -mvrsave -mno-vrsave @gol
1036 -mmulhw -mno-mulhw @gol
1037 -mdlmzb -mno-dlmzb @gol
1038 -mfloat-gprs=yes -mfloat-gprs=no -mfloat-gprs=single -mfloat-gprs=double @gol
1039 -mprototype -mno-prototype @gol
1040 -msim -mmvme -mads -myellowknife -memb -msdata @gol
1041 -msdata=@var{opt} -mvxworks -G @var{num} @gol
1042 -mrecip -mrecip=@var{opt} -mno-recip -mrecip-precision @gol
1043 -mno-recip-precision @gol
1044 -mveclibabi=@var{type} -mfriz -mno-friz @gol
1045 -mpointers-to-nested-functions -mno-pointers-to-nested-functions @gol
1046 -msave-toc-indirect -mno-save-toc-indirect @gol
1047 -mpower8-fusion -mno-mpower8-fusion -mpower8-vector -mno-power8-vector @gol
1048 -mcrypto -mno-crypto -mhtm -mno-htm -mdirect-move -mno-direct-move @gol
1049 -mquad-memory -mno-quad-memory @gol
1050 -mquad-memory-atomic -mno-quad-memory-atomic @gol
1051 -mcompat-align-parm -mno-compat-align-parm @gol
1052 -mfloat128 -mno-float128 -mfloat128-hardware -mno-float128-hardware @gol
1053 -mgnu-attribute -mno-gnu-attribute @gol
1054 -mstack-protector-guard=@var{guard} -mstack-protector-guard-reg=@var{reg} @gol
1055 -mstack-protector-guard-offset=@var{offset}}
1058 @gccoptlist{-m64bit-doubles -m32bit-doubles -fpu -nofpu@gol
1060 -mbig-endian-data -mlittle-endian-data @gol
1063 -mas100-syntax -mno-as100-syntax@gol
1065 -mmax-constant-size=@gol
1068 -mallow-string-insns -mno-allow-string-insns@gol
1070 -mno-warn-multiple-fast-interrupts@gol
1071 -msave-acc-in-interrupts}
1073 @emph{S/390 and zSeries Options}
1074 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
1075 -mhard-float -msoft-float -mhard-dfp -mno-hard-dfp @gol
1076 -mlong-double-64 -mlong-double-128 @gol
1077 -mbackchain -mno-backchain -mpacked-stack -mno-packed-stack @gol
1078 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
1079 -m64 -m31 -mdebug -mno-debug -mesa -mzarch @gol
1080 -mhtm -mvx -mzvector @gol
1081 -mtpf-trace -mno-tpf-trace -mfused-madd -mno-fused-madd @gol
1082 -mwarn-framesize -mwarn-dynamicstack -mstack-size -mstack-guard @gol
1083 -mhotpatch=@var{halfwords},@var{halfwords}}
1085 @emph{Score Options}
1086 @gccoptlist{-meb -mel @gol
1090 -mscore5 -mscore5u -mscore7 -mscore7d}
1093 @gccoptlist{-m1 -m2 -m2e @gol
1094 -m2a-nofpu -m2a-single-only -m2a-single -m2a @gol
1096 -m4-nofpu -m4-single-only -m4-single -m4 @gol
1097 -m4a-nofpu -m4a-single-only -m4a-single -m4a -m4al @gol
1098 -mb -ml -mdalign -mrelax @gol
1099 -mbigtable -mfmovd -mrenesas -mno-renesas -mnomacsave @gol
1100 -mieee -mno-ieee -mbitops -misize -minline-ic_invalidate -mpadstruct @gol
1101 -mprefergot -musermode -multcost=@var{number} -mdiv=@var{strategy} @gol
1102 -mdivsi3_libfunc=@var{name} -mfixed-range=@var{register-range} @gol
1103 -maccumulate-outgoing-args @gol
1104 -matomic-model=@var{atomic-model} @gol
1105 -mbranch-cost=@var{num} -mzdcbranch -mno-zdcbranch @gol
1106 -mcbranch-force-delay-slot @gol
1107 -mfused-madd -mno-fused-madd -mfsca -mno-fsca -mfsrra -mno-fsrra @gol
1108 -mpretend-cmove -mtas}
1110 @emph{Solaris 2 Options}
1111 @gccoptlist{-mclear-hwcap -mno-clear-hwcap -mimpure-text -mno-impure-text @gol
1114 @emph{SPARC Options}
1115 @gccoptlist{-mcpu=@var{cpu-type} @gol
1116 -mtune=@var{cpu-type} @gol
1117 -mcmodel=@var{code-model} @gol
1118 -mmemory-model=@var{mem-model} @gol
1119 -m32 -m64 -mapp-regs -mno-app-regs @gol
1120 -mfaster-structs -mno-faster-structs -mflat -mno-flat @gol
1121 -mfpu -mno-fpu -mhard-float -msoft-float @gol
1122 -mhard-quad-float -msoft-quad-float @gol
1123 -mstack-bias -mno-stack-bias @gol
1124 -mstd-struct-return -mno-std-struct-return @gol
1125 -munaligned-doubles -mno-unaligned-doubles @gol
1126 -muser-mode -mno-user-mode @gol
1127 -mv8plus -mno-v8plus -mvis -mno-vis @gol
1128 -mvis2 -mno-vis2 -mvis3 -mno-vis3 @gol
1129 -mvis4 -mno-vis4 -mvis4b -mno-vis4b @gol
1130 -mcbcond -mno-cbcond -mfmaf -mno-fmaf -mfsmuld -mno-fsmuld @gol
1131 -mpopc -mno-popc -msubxc -mno-subxc @gol
1132 -mfix-at697f -mfix-ut699 -mfix-ut700 -mfix-gr712rc @gol
1136 @gccoptlist{-mwarn-reloc -merror-reloc @gol
1137 -msafe-dma -munsafe-dma @gol
1139 -msmall-mem -mlarge-mem -mstdmain @gol
1140 -mfixed-range=@var{register-range} @gol
1142 -maddress-space-conversion -mno-address-space-conversion @gol
1143 -mcache-size=@var{cache-size} @gol
1144 -matomic-updates -mno-atomic-updates}
1146 @emph{System V Options}
1147 @gccoptlist{-Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
1149 @emph{TILE-Gx Options}
1150 @gccoptlist{-mcpu=CPU -m32 -m64 -mbig-endian -mlittle-endian @gol
1151 -mcmodel=@var{code-model}}
1153 @emph{TILEPro Options}
1154 @gccoptlist{-mcpu=@var{cpu} -m32}
1157 @gccoptlist{-mlong-calls -mno-long-calls -mep -mno-ep @gol
1158 -mprolog-function -mno-prolog-function -mspace @gol
1159 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
1160 -mapp-regs -mno-app-regs @gol
1161 -mdisable-callt -mno-disable-callt @gol
1162 -mv850e2v3 -mv850e2 -mv850e1 -mv850es @gol
1163 -mv850e -mv850 -mv850e3v5 @gol
1174 @gccoptlist{-mg -mgnu -munix}
1176 @emph{Visium Options}
1177 @gccoptlist{-mdebug -msim -mfpu -mno-fpu -mhard-float -msoft-float @gol
1178 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} -msv-mode -muser-mode}
1181 @gccoptlist{-mvms-return-codes -mdebug-main=@var{prefix} -mmalloc64 @gol
1182 -mpointer-size=@var{size}}
1184 @emph{VxWorks Options}
1185 @gccoptlist{-mrtp -non-static -Bstatic -Bdynamic @gol
1186 -Xbind-lazy -Xbind-now}
1189 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
1190 -mtune-ctrl=@var{feature-list} -mdump-tune-features -mno-default @gol
1191 -mfpmath=@var{unit} @gol
1192 -masm=@var{dialect} -mno-fancy-math-387 @gol
1193 -mno-fp-ret-in-387 -m80387 -mhard-float -msoft-float @gol
1194 -mno-wide-multiply -mrtd -malign-double @gol
1195 -mpreferred-stack-boundary=@var{num} @gol
1196 -mincoming-stack-boundary=@var{num} @gol
1197 -mcld -mcx16 -msahf -mmovbe -mcrc32 @gol
1198 -mrecip -mrecip=@var{opt} @gol
1199 -mvzeroupper -mprefer-avx128 -mprefer-avx256 @gol
1200 -mmmx -msse -msse2 -msse3 -mssse3 -msse4.1 -msse4.2 -msse4 -mavx @gol
1201 -mavx2 -mavx512f -mavx512pf -mavx512er -mavx512cd -mavx512vl @gol
1202 -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -msha -maes @gol
1203 -mpclmul -mfsgsbase -mrdrnd -mf16c -mfma @gol
1204 -mprefetchwt1 -mclflushopt -mxsavec -mxsaves @gol
1205 -msse4a -m3dnow -m3dnowa -mpopcnt -mabm -mbmi -mtbm -mfma4 -mxop @gol
1206 -mlzcnt -mbmi2 -mfxsr -mxsave -mxsaveopt -mrtm -mlwp -mmpx @gol
1207 -mmwaitx -mclzero -mpku -mthreads -mgfni @gol
1208 -mcet -mibt -mshstk -mforce-indirect-call -mavx512vbmi2 @gol
1209 -mms-bitfields -mno-align-stringops -minline-all-stringops @gol
1210 -minline-stringops-dynamically -mstringop-strategy=@var{alg} @gol
1211 -mmemcpy-strategy=@var{strategy} -mmemset-strategy=@var{strategy} @gol
1212 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
1213 -m96bit-long-double -mlong-double-64 -mlong-double-80 -mlong-double-128 @gol
1214 -mregparm=@var{num} -msseregparm @gol
1215 -mveclibabi=@var{type} -mvect8-ret-in-mem @gol
1216 -mpc32 -mpc64 -mpc80 -mstackrealign @gol
1217 -momit-leaf-frame-pointer -mno-red-zone -mno-tls-direct-seg-refs @gol
1218 -mcmodel=@var{code-model} -mabi=@var{name} -maddress-mode=@var{mode} @gol
1219 -m32 -m64 -mx32 -m16 -miamcu -mlarge-data-threshold=@var{num} @gol
1220 -msse2avx -mfentry -mrecord-mcount -mnop-mcount -m8bit-idiv @gol
1221 -mavx256-split-unaligned-load -mavx256-split-unaligned-store @gol
1222 -malign-data=@var{type} -mstack-protector-guard=@var{guard} @gol
1223 -mstack-protector-guard-reg=@var{reg} @gol
1224 -mstack-protector-guard-offset=@var{offset} @gol
1225 -mstack-protector-guard-symbol=@var{symbol} -mmitigate-rop @gol
1226 -mgeneral-regs-only -mcall-ms2sysv-xlogues}
1228 @emph{x86 Windows Options}
1229 @gccoptlist{-mconsole -mcygwin -mno-cygwin -mdll @gol
1230 -mnop-fun-dllimport -mthread @gol
1231 -municode -mwin32 -mwindows -fno-set-stack-executable}
1233 @emph{Xstormy16 Options}
1236 @emph{Xtensa Options}
1237 @gccoptlist{-mconst16 -mno-const16 @gol
1238 -mfused-madd -mno-fused-madd @gol
1240 -mserialize-volatile -mno-serialize-volatile @gol
1241 -mtext-section-literals -mno-text-section-literals @gol
1242 -mauto-litpools -mno-auto-litpools @gol
1243 -mtarget-align -mno-target-align @gol
1244 -mlongcalls -mno-longcalls}
1246 @emph{zSeries Options}
1247 See S/390 and zSeries Options.
1251 @node Overall Options
1252 @section Options Controlling the Kind of Output
1254 Compilation can involve up to four stages: preprocessing, compilation
1255 proper, assembly and linking, always in that order. GCC is capable of
1256 preprocessing and compiling several files either into several
1257 assembler input files, or into one assembler input file; then each
1258 assembler input file produces an object file, and linking combines all
1259 the object files (those newly compiled, and those specified as input)
1260 into an executable file.
1262 @cindex file name suffix
1263 For any given input file, the file name suffix determines what kind of
1264 compilation is done:
1268 C source code that must be preprocessed.
1271 C source code that should not be preprocessed.
1274 C++ source code that should not be preprocessed.
1277 Objective-C source code. Note that you must link with the @file{libobjc}
1278 library to make an Objective-C program work.
1281 Objective-C source code that should not be preprocessed.
1285 Objective-C++ source code. Note that you must link with the @file{libobjc}
1286 library to make an Objective-C++ program work. Note that @samp{.M} refers
1287 to a literal capital M@.
1289 @item @var{file}.mii
1290 Objective-C++ source code that should not be preprocessed.
1293 C, C++, Objective-C or Objective-C++ header file to be turned into a
1294 precompiled header (default), or C, C++ header file to be turned into an
1295 Ada spec (via the @option{-fdump-ada-spec} switch).
1298 @itemx @var{file}.cp
1299 @itemx @var{file}.cxx
1300 @itemx @var{file}.cpp
1301 @itemx @var{file}.CPP
1302 @itemx @var{file}.c++
1304 C++ source code that must be preprocessed. Note that in @samp{.cxx},
1305 the last two letters must both be literally @samp{x}. Likewise,
1306 @samp{.C} refers to a literal capital C@.
1310 Objective-C++ source code that must be preprocessed.
1312 @item @var{file}.mii
1313 Objective-C++ source code that should not be preprocessed.
1317 @itemx @var{file}.hp
1318 @itemx @var{file}.hxx
1319 @itemx @var{file}.hpp
1320 @itemx @var{file}.HPP
1321 @itemx @var{file}.h++
1322 @itemx @var{file}.tcc
1323 C++ header file to be turned into a precompiled header or Ada spec.
1326 @itemx @var{file}.for
1327 @itemx @var{file}.ftn
1328 Fixed form Fortran source code that should not be preprocessed.
1331 @itemx @var{file}.FOR
1332 @itemx @var{file}.fpp
1333 @itemx @var{file}.FPP
1334 @itemx @var{file}.FTN
1335 Fixed form Fortran source code that must be preprocessed (with the traditional
1338 @item @var{file}.f90
1339 @itemx @var{file}.f95
1340 @itemx @var{file}.f03
1341 @itemx @var{file}.f08
1342 Free form Fortran source code that should not be preprocessed.
1344 @item @var{file}.F90
1345 @itemx @var{file}.F95
1346 @itemx @var{file}.F03
1347 @itemx @var{file}.F08
1348 Free form Fortran source code that must be preprocessed (with the
1349 traditional preprocessor).
1354 @item @var{file}.brig
1355 BRIG files (binary representation of HSAIL).
1357 @item @var{file}.ads
1358 Ada source code file that contains a library unit declaration (a
1359 declaration of a package, subprogram, or generic, or a generic
1360 instantiation), or a library unit renaming declaration (a package,
1361 generic, or subprogram renaming declaration). Such files are also
1364 @item @var{file}.adb
1365 Ada source code file containing a library unit body (a subprogram or
1366 package body). Such files are also called @dfn{bodies}.
1368 @c GCC also knows about some suffixes for languages not yet included:
1379 @itemx @var{file}.sx
1380 Assembler code that must be preprocessed.
1383 An object file to be fed straight into linking.
1384 Any file name with no recognized suffix is treated this way.
1388 You can specify the input language explicitly with the @option{-x} option:
1391 @item -x @var{language}
1392 Specify explicitly the @var{language} for the following input files
1393 (rather than letting the compiler choose a default based on the file
1394 name suffix). This option applies to all following input files until
1395 the next @option{-x} option. Possible values for @var{language} are:
1397 c c-header cpp-output
1398 c++ c++-header c++-cpp-output
1399 objective-c objective-c-header objective-c-cpp-output
1400 objective-c++ objective-c++-header objective-c++-cpp-output
1401 assembler assembler-with-cpp
1403 f77 f77-cpp-input f95 f95-cpp-input
1409 Turn off any specification of a language, so that subsequent files are
1410 handled according to their file name suffixes (as they are if @option{-x}
1411 has not been used at all).
1414 If you only want some of the stages of compilation, you can use
1415 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
1416 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
1417 @command{gcc} is to stop. Note that some combinations (for example,
1418 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
1423 Compile or assemble the source files, but do not link. The linking
1424 stage simply is not done. The ultimate output is in the form of an
1425 object file for each source file.
1427 By default, the object file name for a source file is made by replacing
1428 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
1430 Unrecognized input files, not requiring compilation or assembly, are
1435 Stop after the stage of compilation proper; do not assemble. The output
1436 is in the form of an assembler code file for each non-assembler input
1439 By default, the assembler file name for a source file is made by
1440 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
1442 Input files that don't require compilation are ignored.
1446 Stop after the preprocessing stage; do not run the compiler proper. The
1447 output is in the form of preprocessed source code, which is sent to the
1450 Input files that don't require preprocessing are ignored.
1452 @cindex output file option
1455 Place output in file @var{file}. This applies to whatever
1456 sort of output is being produced, whether it be an executable file,
1457 an object file, an assembler file or preprocessed C code.
1459 If @option{-o} is not specified, the default is to put an executable
1460 file in @file{a.out}, the object file for
1461 @file{@var{source}.@var{suffix}} in @file{@var{source}.o}, its
1462 assembler file in @file{@var{source}.s}, a precompiled header file in
1463 @file{@var{source}.@var{suffix}.gch}, and all preprocessed C source on
1468 Print (on standard error output) the commands executed to run the stages
1469 of compilation. Also print the version number of the compiler driver
1470 program and of the preprocessor and the compiler proper.
1474 Like @option{-v} except the commands are not executed and arguments
1475 are quoted unless they contain only alphanumeric characters or @code{./-_}.
1476 This is useful for shell scripts to capture the driver-generated command lines.
1480 Print (on the standard output) a description of the command-line options
1481 understood by @command{gcc}. If the @option{-v} option is also specified
1482 then @option{--help} is also passed on to the various processes
1483 invoked by @command{gcc}, so that they can display the command-line options
1484 they accept. If the @option{-Wextra} option has also been specified
1485 (prior to the @option{--help} option), then command-line options that
1486 have no documentation associated with them are also displayed.
1489 @opindex target-help
1490 Print (on the standard output) a description of target-specific command-line
1491 options for each tool. For some targets extra target-specific
1492 information may also be printed.
1494 @item --help=@{@var{class}@r{|[}^@r{]}@var{qualifier}@}@r{[},@dots{}@r{]}
1495 Print (on the standard output) a description of the command-line
1496 options understood by the compiler that fit into all specified classes
1497 and qualifiers. These are the supported classes:
1500 @item @samp{optimizers}
1501 Display all of the optimization options supported by the
1504 @item @samp{warnings}
1505 Display all of the options controlling warning messages
1506 produced by the compiler.
1509 Display target-specific options. Unlike the
1510 @option{--target-help} option however, target-specific options of the
1511 linker and assembler are not displayed. This is because those
1512 tools do not currently support the extended @option{--help=} syntax.
1515 Display the values recognized by the @option{--param}
1518 @item @var{language}
1519 Display the options supported for @var{language}, where
1520 @var{language} is the name of one of the languages supported in this
1524 Display the options that are common to all languages.
1527 These are the supported qualifiers:
1530 @item @samp{undocumented}
1531 Display only those options that are undocumented.
1534 Display options taking an argument that appears after an equal
1535 sign in the same continuous piece of text, such as:
1536 @samp{--help=target}.
1538 @item @samp{separate}
1539 Display options taking an argument that appears as a separate word
1540 following the original option, such as: @samp{-o output-file}.
1543 Thus for example to display all the undocumented target-specific
1544 switches supported by the compiler, use:
1547 --help=target,undocumented
1550 The sense of a qualifier can be inverted by prefixing it with the
1551 @samp{^} character, so for example to display all binary warning
1552 options (i.e., ones that are either on or off and that do not take an
1553 argument) that have a description, use:
1556 --help=warnings,^joined,^undocumented
1559 The argument to @option{--help=} should not consist solely of inverted
1562 Combining several classes is possible, although this usually
1563 restricts the output so much that there is nothing to display. One
1564 case where it does work, however, is when one of the classes is
1565 @var{target}. For example, to display all the target-specific
1566 optimization options, use:
1569 --help=target,optimizers
1572 The @option{--help=} option can be repeated on the command line. Each
1573 successive use displays its requested class of options, skipping
1574 those that have already been displayed.
1576 If the @option{-Q} option appears on the command line before the
1577 @option{--help=} option, then the descriptive text displayed by
1578 @option{--help=} is changed. Instead of describing the displayed
1579 options, an indication is given as to whether the option is enabled,
1580 disabled or set to a specific value (assuming that the compiler
1581 knows this at the point where the @option{--help=} option is used).
1583 Here is a truncated example from the ARM port of @command{gcc}:
1586 % gcc -Q -mabi=2 --help=target -c
1587 The following options are target specific:
1589 -mabort-on-noreturn [disabled]
1593 The output is sensitive to the effects of previous command-line
1594 options, so for example it is possible to find out which optimizations
1595 are enabled at @option{-O2} by using:
1598 -Q -O2 --help=optimizers
1601 Alternatively you can discover which binary optimizations are enabled
1602 by @option{-O3} by using:
1605 gcc -c -Q -O3 --help=optimizers > /tmp/O3-opts
1606 gcc -c -Q -O2 --help=optimizers > /tmp/O2-opts
1607 diff /tmp/O2-opts /tmp/O3-opts | grep enabled
1612 Display the version number and copyrights of the invoked GCC@.
1614 @item -pass-exit-codes
1615 @opindex pass-exit-codes
1616 Normally the @command{gcc} program exits with the code of 1 if any
1617 phase of the compiler returns a non-success return code. If you specify
1618 @option{-pass-exit-codes}, the @command{gcc} program instead returns with
1619 the numerically highest error produced by any phase returning an error
1620 indication. The C, C++, and Fortran front ends return 4 if an internal
1621 compiler error is encountered.
1625 Use pipes rather than temporary files for communication between the
1626 various stages of compilation. This fails to work on some systems where
1627 the assembler is unable to read from a pipe; but the GNU assembler has
1630 @item -specs=@var{file}
1632 Process @var{file} after the compiler reads in the standard @file{specs}
1633 file, in order to override the defaults which the @command{gcc} driver
1634 program uses when determining what switches to pass to @command{cc1},
1635 @command{cc1plus}, @command{as}, @command{ld}, etc. More than one
1636 @option{-specs=@var{file}} can be specified on the command line, and they
1637 are processed in order, from left to right. @xref{Spec Files}, for
1638 information about the format of the @var{file}.
1642 Invoke all subcommands under a wrapper program. The name of the
1643 wrapper program and its parameters are passed as a comma separated
1647 gcc -c t.c -wrapper gdb,--args
1651 This invokes all subprograms of @command{gcc} under
1652 @samp{gdb --args}, thus the invocation of @command{cc1} is
1653 @samp{gdb --args cc1 @dots{}}.
1655 @item -fplugin=@var{name}.so
1657 Load the plugin code in file @var{name}.so, assumed to be a
1658 shared object to be dlopen'd by the compiler. The base name of
1659 the shared object file is used to identify the plugin for the
1660 purposes of argument parsing (See
1661 @option{-fplugin-arg-@var{name}-@var{key}=@var{value}} below).
1662 Each plugin should define the callback functions specified in the
1665 @item -fplugin-arg-@var{name}-@var{key}=@var{value}
1666 @opindex fplugin-arg
1667 Define an argument called @var{key} with a value of @var{value}
1668 for the plugin called @var{name}.
1670 @item -fdump-ada-spec@r{[}-slim@r{]}
1671 @opindex fdump-ada-spec
1672 For C and C++ source and include files, generate corresponding Ada specs.
1673 @xref{Generating Ada Bindings for C and C++ headers,,, gnat_ugn,
1674 GNAT User's Guide}, which provides detailed documentation on this feature.
1676 @item -fada-spec-parent=@var{unit}
1677 @opindex fada-spec-parent
1678 In conjunction with @option{-fdump-ada-spec@r{[}-slim@r{]}} above, generate
1679 Ada specs as child units of parent @var{unit}.
1681 @item -fdump-go-spec=@var{file}
1682 @opindex fdump-go-spec
1683 For input files in any language, generate corresponding Go
1684 declarations in @var{file}. This generates Go @code{const},
1685 @code{type}, @code{var}, and @code{func} declarations which may be a
1686 useful way to start writing a Go interface to code written in some
1689 @include @value{srcdir}/../libiberty/at-file.texi
1693 @section Compiling C++ Programs
1695 @cindex suffixes for C++ source
1696 @cindex C++ source file suffixes
1697 C++ source files conventionally use one of the suffixes @samp{.C},
1698 @samp{.cc}, @samp{.cpp}, @samp{.CPP}, @samp{.c++}, @samp{.cp}, or
1699 @samp{.cxx}; C++ header files often use @samp{.hh}, @samp{.hpp},
1700 @samp{.H}, or (for shared template code) @samp{.tcc}; and
1701 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
1702 files with these names and compiles them as C++ programs even if you
1703 call the compiler the same way as for compiling C programs (usually
1704 with the name @command{gcc}).
1708 However, the use of @command{gcc} does not add the C++ library.
1709 @command{g++} is a program that calls GCC and automatically specifies linking
1710 against the C++ library. It treats @samp{.c},
1711 @samp{.h} and @samp{.i} files as C++ source files instead of C source
1712 files unless @option{-x} is used. This program is also useful when
1713 precompiling a C header file with a @samp{.h} extension for use in C++
1714 compilations. On many systems, @command{g++} is also installed with
1715 the name @command{c++}.
1717 @cindex invoking @command{g++}
1718 When you compile C++ programs, you may specify many of the same
1719 command-line options that you use for compiling programs in any
1720 language; or command-line options meaningful for C and related
1721 languages; or options that are meaningful only for C++ programs.
1722 @xref{C Dialect Options,,Options Controlling C Dialect}, for
1723 explanations of options for languages related to C@.
1724 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
1725 explanations of options that are meaningful only for C++ programs.
1727 @node C Dialect Options
1728 @section Options Controlling C Dialect
1729 @cindex dialect options
1730 @cindex language dialect options
1731 @cindex options, dialect
1733 The following options control the dialect of C (or languages derived
1734 from C, such as C++, Objective-C and Objective-C++) that the compiler
1738 @cindex ANSI support
1742 In C mode, this is equivalent to @option{-std=c90}. In C++ mode, it is
1743 equivalent to @option{-std=c++98}.
1745 This turns off certain features of GCC that are incompatible with ISO
1746 C90 (when compiling C code), or of standard C++ (when compiling C++ code),
1747 such as the @code{asm} and @code{typeof} keywords, and
1748 predefined macros such as @code{unix} and @code{vax} that identify the
1749 type of system you are using. It also enables the undesirable and
1750 rarely used ISO trigraph feature. For the C compiler,
1751 it disables recognition of C++ style @samp{//} comments as well as
1752 the @code{inline} keyword.
1754 The alternate keywords @code{__asm__}, @code{__extension__},
1755 @code{__inline__} and @code{__typeof__} continue to work despite
1756 @option{-ansi}. You would not want to use them in an ISO C program, of
1757 course, but it is useful to put them in header files that might be included
1758 in compilations done with @option{-ansi}. Alternate predefined macros
1759 such as @code{__unix__} and @code{__vax__} are also available, with or
1760 without @option{-ansi}.
1762 The @option{-ansi} option does not cause non-ISO programs to be
1763 rejected gratuitously. For that, @option{-Wpedantic} is required in
1764 addition to @option{-ansi}. @xref{Warning Options}.
1766 The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi}
1767 option is used. Some header files may notice this macro and refrain
1768 from declaring certain functions or defining certain macros that the
1769 ISO standard doesn't call for; this is to avoid interfering with any
1770 programs that might use these names for other things.
1772 Functions that are normally built in but do not have semantics
1773 defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in
1774 functions when @option{-ansi} is used. @xref{Other Builtins,,Other
1775 built-in functions provided by GCC}, for details of the functions
1780 Determine the language standard. @xref{Standards,,Language Standards
1781 Supported by GCC}, for details of these standard versions. This option
1782 is currently only supported when compiling C or C++.
1784 The compiler can accept several base standards, such as @samp{c90} or
1785 @samp{c++98}, and GNU dialects of those standards, such as
1786 @samp{gnu90} or @samp{gnu++98}. When a base standard is specified, the
1787 compiler accepts all programs following that standard plus those
1788 using GNU extensions that do not contradict it. For example,
1789 @option{-std=c90} turns off certain features of GCC that are
1790 incompatible with ISO C90, such as the @code{asm} and @code{typeof}
1791 keywords, but not other GNU extensions that do not have a meaning in
1792 ISO C90, such as omitting the middle term of a @code{?:}
1793 expression. On the other hand, when a GNU dialect of a standard is
1794 specified, all features supported by the compiler are enabled, even when
1795 those features change the meaning of the base standard. As a result, some
1796 strict-conforming programs may be rejected. The particular standard
1797 is used by @option{-Wpedantic} to identify which features are GNU
1798 extensions given that version of the standard. For example
1799 @option{-std=gnu90 -Wpedantic} warns about C++ style @samp{//}
1800 comments, while @option{-std=gnu99 -Wpedantic} does not.
1802 A value for this option must be provided; possible values are
1808 Support all ISO C90 programs (certain GNU extensions that conflict
1809 with ISO C90 are disabled). Same as @option{-ansi} for C code.
1811 @item iso9899:199409
1812 ISO C90 as modified in amendment 1.
1818 ISO C99. This standard is substantially completely supported, modulo
1819 bugs and floating-point issues
1820 (mainly but not entirely relating to optional C99 features from
1821 Annexes F and G). See
1822 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The
1823 names @samp{c9x} and @samp{iso9899:199x} are deprecated.
1828 ISO C11, the 2011 revision of the ISO C standard. This standard is
1829 substantially completely supported, modulo bugs, floating-point issues
1830 (mainly but not entirely relating to optional C11 features from
1831 Annexes F and G) and the optional Annexes K (Bounds-checking
1832 interfaces) and L (Analyzability). The name @samp{c1x} is deprecated.
1838 ISO C17, the 2017 revision of the ISO C standard (expected to be
1839 published in 2018). This standard is
1840 same as C11 except for corrections of defects (all of which are also
1841 applied with @option{-std=c11}) and a new value of
1842 @code{__STDC_VERSION__}, and so is supported to the same extent as C11.
1846 GNU dialect of ISO C90 (including some C99 features).
1850 GNU dialect of ISO C99. The name @samp{gnu9x} is deprecated.
1854 GNU dialect of ISO C11.
1855 The name @samp{gnu1x} is deprecated.
1859 GNU dialect of ISO C17. This is the default for C code.
1863 The 1998 ISO C++ standard plus the 2003 technical corrigendum and some
1864 additional defect reports. Same as @option{-ansi} for C++ code.
1868 GNU dialect of @option{-std=c++98}.
1872 The 2011 ISO C++ standard plus amendments.
1873 The name @samp{c++0x} is deprecated.
1877 GNU dialect of @option{-std=c++11}.
1878 The name @samp{gnu++0x} is deprecated.
1882 The 2014 ISO C++ standard plus amendments.
1883 The name @samp{c++1y} is deprecated.
1887 GNU dialect of @option{-std=c++14}.
1888 This is the default for C++ code.
1889 The name @samp{gnu++1y} is deprecated.
1893 The 2017 ISO C++ standard plus amendments.
1894 The name @samp{c++1z} is deprecated.
1898 GNU dialect of @option{-std=c++17}.
1899 The name @samp{gnu++1z} is deprecated.
1902 The next revision of the ISO C++ standard, tentatively planned for
1903 2020. Support is highly experimental, and will almost certainly
1904 change in incompatible ways in future releases.
1907 GNU dialect of @option{-std=c++2a}. Support is highly experimental,
1908 and will almost certainly change in incompatible ways in future
1912 @item -fgnu89-inline
1913 @opindex fgnu89-inline
1914 The option @option{-fgnu89-inline} tells GCC to use the traditional
1915 GNU semantics for @code{inline} functions when in C99 mode.
1916 @xref{Inline,,An Inline Function is As Fast As a Macro}.
1917 Using this option is roughly equivalent to adding the
1918 @code{gnu_inline} function attribute to all inline functions
1919 (@pxref{Function Attributes}).
1921 The option @option{-fno-gnu89-inline} explicitly tells GCC to use the
1922 C99 semantics for @code{inline} when in C99 or gnu99 mode (i.e., it
1923 specifies the default behavior).
1924 This option is not supported in @option{-std=c90} or
1925 @option{-std=gnu90} mode.
1927 The preprocessor macros @code{__GNUC_GNU_INLINE__} and
1928 @code{__GNUC_STDC_INLINE__} may be used to check which semantics are
1929 in effect for @code{inline} functions. @xref{Common Predefined
1930 Macros,,,cpp,The C Preprocessor}.
1932 @item -fpermitted-flt-eval-methods=@var{style}
1933 @opindex fpermitted-flt-eval-methods
1934 @opindex fpermitted-flt-eval-methods=c11
1935 @opindex fpermitted-flt-eval-methods=ts-18661-3
1936 ISO/IEC TS 18661-3 defines new permissible values for
1937 @code{FLT_EVAL_METHOD} that indicate that operations and constants with
1938 a semantic type that is an interchange or extended format should be
1939 evaluated to the precision and range of that type. These new values are
1940 a superset of those permitted under C99/C11, which does not specify the
1941 meaning of other positive values of @code{FLT_EVAL_METHOD}. As such, code
1942 conforming to C11 may not have been written expecting the possibility of
1945 @option{-fpermitted-flt-eval-methods} specifies whether the compiler
1946 should allow only the values of @code{FLT_EVAL_METHOD} specified in C99/C11,
1947 or the extended set of values specified in ISO/IEC TS 18661-3.
1949 @var{style} is either @code{c11} or @code{ts-18661-3} as appropriate.
1951 The default when in a standards compliant mode (@option{-std=c11} or similar)
1952 is @option{-fpermitted-flt-eval-methods=c11}. The default when in a GNU
1953 dialect (@option{-std=gnu11} or similar) is
1954 @option{-fpermitted-flt-eval-methods=ts-18661-3}.
1956 @item -aux-info @var{filename}
1958 Output to the given filename prototyped declarations for all functions
1959 declared and/or defined in a translation unit, including those in header
1960 files. This option is silently ignored in any language other than C@.
1962 Besides declarations, the file indicates, in comments, the origin of
1963 each declaration (source file and line), whether the declaration was
1964 implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or
1965 @samp{O} for old, respectively, in the first character after the line
1966 number and the colon), and whether it came from a declaration or a
1967 definition (@samp{C} or @samp{F}, respectively, in the following
1968 character). In the case of function definitions, a K&R-style list of
1969 arguments followed by their declarations is also provided, inside
1970 comments, after the declaration.
1972 @item -fallow-parameterless-variadic-functions
1973 @opindex fallow-parameterless-variadic-functions
1974 Accept variadic functions without named parameters.
1976 Although it is possible to define such a function, this is not very
1977 useful as it is not possible to read the arguments. This is only
1978 supported for C as this construct is allowed by C++.
1982 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
1983 keyword, so that code can use these words as identifiers. You can use
1984 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
1985 instead. @option{-ansi} implies @option{-fno-asm}.
1987 In C++, this switch only affects the @code{typeof} keyword, since
1988 @code{asm} and @code{inline} are standard keywords. You may want to
1989 use the @option{-fno-gnu-keywords} flag instead, which has the same
1990 effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this
1991 switch only affects the @code{asm} and @code{typeof} keywords, since
1992 @code{inline} is a standard keyword in ISO C99.
1995 @itemx -fno-builtin-@var{function}
1996 @opindex fno-builtin
1997 @cindex built-in functions
1998 Don't recognize built-in functions that do not begin with
1999 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
2000 functions provided by GCC}, for details of the functions affected,
2001 including those which are not built-in functions when @option{-ansi} or
2002 @option{-std} options for strict ISO C conformance are used because they
2003 do not have an ISO standard meaning.
2005 GCC normally generates special code to handle certain built-in functions
2006 more efficiently; for instance, calls to @code{alloca} may become single
2007 instructions which adjust the stack directly, and calls to @code{memcpy}
2008 may become inline copy loops. The resulting code is often both smaller
2009 and faster, but since the function calls no longer appear as such, you
2010 cannot set a breakpoint on those calls, nor can you change the behavior
2011 of the functions by linking with a different library. In addition,
2012 when a function is recognized as a built-in function, GCC may use
2013 information about that function to warn about problems with calls to
2014 that function, or to generate more efficient code, even if the
2015 resulting code still contains calls to that function. For example,
2016 warnings are given with @option{-Wformat} for bad calls to
2017 @code{printf} when @code{printf} is built in and @code{strlen} is
2018 known not to modify global memory.
2020 With the @option{-fno-builtin-@var{function}} option
2021 only the built-in function @var{function} is
2022 disabled. @var{function} must not begin with @samp{__builtin_}. If a
2023 function is named that is not built-in in this version of GCC, this
2024 option is ignored. There is no corresponding
2025 @option{-fbuiltin-@var{function}} option; if you wish to enable
2026 built-in functions selectively when using @option{-fno-builtin} or
2027 @option{-ffreestanding}, you may define macros such as:
2030 #define abs(n) __builtin_abs ((n))
2031 #define strcpy(d, s) __builtin_strcpy ((d), (s))
2037 Enable parsing of function definitions marked with @code{__GIMPLE}.
2038 This is an experimental feature that allows unit testing of GIMPLE
2043 @cindex hosted environment
2045 Assert that compilation targets a hosted environment. This implies
2046 @option{-fbuiltin}. A hosted environment is one in which the
2047 entire standard library is available, and in which @code{main} has a return
2048 type of @code{int}. Examples are nearly everything except a kernel.
2049 This is equivalent to @option{-fno-freestanding}.
2051 @item -ffreestanding
2052 @opindex ffreestanding
2053 @cindex hosted environment
2055 Assert that compilation targets a freestanding environment. This
2056 implies @option{-fno-builtin}. A freestanding environment
2057 is one in which the standard library may not exist, and program startup may
2058 not necessarily be at @code{main}. The most obvious example is an OS kernel.
2059 This is equivalent to @option{-fno-hosted}.
2061 @xref{Standards,,Language Standards Supported by GCC}, for details of
2062 freestanding and hosted environments.
2066 @cindex OpenACC accelerator programming
2067 Enable handling of OpenACC directives @code{#pragma acc} in C/C++ and
2068 @code{!$acc} in Fortran. When @option{-fopenacc} is specified, the
2069 compiler generates accelerated code according to the OpenACC Application
2070 Programming Interface v2.0 @w{@uref{https://www.openacc.org}}. This option
2071 implies @option{-pthread}, and thus is only supported on targets that
2072 have support for @option{-pthread}.
2074 @item -fopenacc-dim=@var{geom}
2075 @opindex fopenacc-dim
2076 @cindex OpenACC accelerator programming
2077 Specify default compute dimensions for parallel offload regions that do
2078 not explicitly specify. The @var{geom} value is a triple of
2079 ':'-separated sizes, in order 'gang', 'worker' and, 'vector'. A size
2080 can be omitted, to use a target-specific default value.
2084 @cindex OpenMP parallel
2085 Enable handling of OpenMP directives @code{#pragma omp} in C/C++ and
2086 @code{!$omp} in Fortran. When @option{-fopenmp} is specified, the
2087 compiler generates parallel code according to the OpenMP Application
2088 Program Interface v4.5 @w{@uref{http://www.openmp.org/}}. This option
2089 implies @option{-pthread}, and thus is only supported on targets that
2090 have support for @option{-pthread}. @option{-fopenmp} implies
2091 @option{-fopenmp-simd}.
2094 @opindex fopenmp-simd
2097 Enable handling of OpenMP's SIMD directives with @code{#pragma omp}
2098 in C/C++ and @code{!$omp} in Fortran. Other OpenMP directives
2103 @cindex Enable Cilk Plus
2104 Enable the usage of Cilk Plus language extension features for C/C++.
2105 When the option @option{-fcilkplus} is specified, enable the usage of
2106 the Cilk Plus Language extension features for C/C++. The present
2107 implementation follows ABI version 1.2. This is an experimental
2108 feature that is only partially complete, and whose interface may
2109 change in future versions of GCC as the official specification
2110 changes. Currently, all features but @code{_Cilk_for} have been
2115 When the option @option{-fgnu-tm} is specified, the compiler
2116 generates code for the Linux variant of Intel's current Transactional
2117 Memory ABI specification document (Revision 1.1, May 6 2009). This is
2118 an experimental feature whose interface may change in future versions
2119 of GCC, as the official specification changes. Please note that not
2120 all architectures are supported for this feature.
2122 For more information on GCC's support for transactional memory,
2123 @xref{Enabling libitm,,The GNU Transactional Memory Library,libitm,GNU
2124 Transactional Memory Library}.
2126 Note that the transactional memory feature is not supported with
2127 non-call exceptions (@option{-fnon-call-exceptions}).
2129 @item -fms-extensions
2130 @opindex fms-extensions
2131 Accept some non-standard constructs used in Microsoft header files.
2133 In C++ code, this allows member names in structures to be similar
2134 to previous types declarations.
2143 Some cases of unnamed fields in structures and unions are only
2144 accepted with this option. @xref{Unnamed Fields,,Unnamed struct/union
2145 fields within structs/unions}, for details.
2147 Note that this option is off for all targets but x86
2148 targets using ms-abi.
2150 @item -fplan9-extensions
2151 @opindex fplan9-extensions
2152 Accept some non-standard constructs used in Plan 9 code.
2154 This enables @option{-fms-extensions}, permits passing pointers to
2155 structures with anonymous fields to functions that expect pointers to
2156 elements of the type of the field, and permits referring to anonymous
2157 fields declared using a typedef. @xref{Unnamed Fields,,Unnamed
2158 struct/union fields within structs/unions}, for details. This is only
2159 supported for C, not C++.
2161 @item -fcond-mismatch
2162 @opindex fcond-mismatch
2163 Allow conditional expressions with mismatched types in the second and
2164 third arguments. The value of such an expression is void. This option
2165 is not supported for C++.
2167 @item -flax-vector-conversions
2168 @opindex flax-vector-conversions
2169 Allow implicit conversions between vectors with differing numbers of
2170 elements and/or incompatible element types. This option should not be
2173 @item -funsigned-char
2174 @opindex funsigned-char
2175 Let the type @code{char} be unsigned, like @code{unsigned char}.
2177 Each kind of machine has a default for what @code{char} should
2178 be. It is either like @code{unsigned char} by default or like
2179 @code{signed char} by default.
2181 Ideally, a portable program should always use @code{signed char} or
2182 @code{unsigned char} when it depends on the signedness of an object.
2183 But many programs have been written to use plain @code{char} and
2184 expect it to be signed, or expect it to be unsigned, depending on the
2185 machines they were written for. This option, and its inverse, let you
2186 make such a program work with the opposite default.
2188 The type @code{char} is always a distinct type from each of
2189 @code{signed char} or @code{unsigned char}, even though its behavior
2190 is always just like one of those two.
2193 @opindex fsigned-char
2194 Let the type @code{char} be signed, like @code{signed char}.
2196 Note that this is equivalent to @option{-fno-unsigned-char}, which is
2197 the negative form of @option{-funsigned-char}. Likewise, the option
2198 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
2200 @item -fsigned-bitfields
2201 @itemx -funsigned-bitfields
2202 @itemx -fno-signed-bitfields
2203 @itemx -fno-unsigned-bitfields
2204 @opindex fsigned-bitfields
2205 @opindex funsigned-bitfields
2206 @opindex fno-signed-bitfields
2207 @opindex fno-unsigned-bitfields
2208 These options control whether a bit-field is signed or unsigned, when the
2209 declaration does not use either @code{signed} or @code{unsigned}. By
2210 default, such a bit-field is signed, because this is consistent: the
2211 basic integer types such as @code{int} are signed types.
2213 @item -fsso-struct=@var{endianness}
2214 @opindex fsso-struct
2215 Set the default scalar storage order of structures and unions to the
2216 specified endianness. The accepted values are @samp{big-endian},
2217 @samp{little-endian} and @samp{native} for the native endianness of
2218 the target (the default). This option is not supported for C++.
2220 @strong{Warning:} the @option{-fsso-struct} switch causes GCC to generate
2221 code that is not binary compatible with code generated without it if the
2222 specified endianness is not the native endianness of the target.
2225 @node C++ Dialect Options
2226 @section Options Controlling C++ Dialect
2228 @cindex compiler options, C++
2229 @cindex C++ options, command-line
2230 @cindex options, C++
2231 This section describes the command-line options that are only meaningful
2232 for C++ programs. You can also use most of the GNU compiler options
2233 regardless of what language your program is in. For example, you
2234 might compile a file @file{firstClass.C} like this:
2237 g++ -g -fstrict-enums -O -c firstClass.C
2241 In this example, only @option{-fstrict-enums} is an option meant
2242 only for C++ programs; you can use the other options with any
2243 language supported by GCC@.
2245 Some options for compiling C programs, such as @option{-std}, are also
2246 relevant for C++ programs.
2247 @xref{C Dialect Options,,Options Controlling C Dialect}.
2249 Here is a list of options that are @emph{only} for compiling C++ programs:
2253 @item -fabi-version=@var{n}
2254 @opindex fabi-version
2255 Use version @var{n} of the C++ ABI@. The default is version 0.
2257 Version 0 refers to the version conforming most closely to
2258 the C++ ABI specification. Therefore, the ABI obtained using version 0
2259 will change in different versions of G++ as ABI bugs are fixed.
2261 Version 1 is the version of the C++ ABI that first appeared in G++ 3.2.
2263 Version 2 is the version of the C++ ABI that first appeared in G++
2264 3.4, and was the default through G++ 4.9.
2266 Version 3 corrects an error in mangling a constant address as a
2269 Version 4, which first appeared in G++ 4.5, implements a standard
2270 mangling for vector types.
2272 Version 5, which first appeared in G++ 4.6, corrects the mangling of
2273 attribute const/volatile on function pointer types, decltype of a
2274 plain decl, and use of a function parameter in the declaration of
2277 Version 6, which first appeared in G++ 4.7, corrects the promotion
2278 behavior of C++11 scoped enums and the mangling of template argument
2279 packs, const/static_cast, prefix ++ and --, and a class scope function
2280 used as a template argument.
2282 Version 7, which first appeared in G++ 4.8, that treats nullptr_t as a
2283 builtin type and corrects the mangling of lambdas in default argument
2286 Version 8, which first appeared in G++ 4.9, corrects the substitution
2287 behavior of function types with function-cv-qualifiers.
2289 Version 9, which first appeared in G++ 5.2, corrects the alignment of
2292 Version 10, which first appeared in G++ 6.1, adds mangling of
2293 attributes that affect type identity, such as ia32 calling convention
2294 attributes (e.g. @samp{stdcall}).
2296 Version 11, which first appeared in G++ 7, corrects the mangling of
2297 sizeof... expressions and operator names. For multiple entities with
2298 the same name within a function, that are declared in different scopes,
2299 the mangling now changes starting with the twelfth occurrence. It also
2300 implies @option{-fnew-inheriting-ctors}.
2302 See also @option{-Wabi}.
2304 @item -fabi-compat-version=@var{n}
2305 @opindex fabi-compat-version
2306 On targets that support strong aliases, G++
2307 works around mangling changes by creating an alias with the correct
2308 mangled name when defining a symbol with an incorrect mangled name.
2309 This switch specifies which ABI version to use for the alias.
2311 With @option{-fabi-version=0} (the default), this defaults to 8 (GCC 5
2312 compatibility). If another ABI version is explicitly selected, this
2313 defaults to 0. For compatibility with GCC versions 3.2 through 4.9,
2314 use @option{-fabi-compat-version=2}.
2316 If this option is not provided but @option{-Wabi=@var{n}} is, that
2317 version is used for compatibility aliases. If this option is provided
2318 along with @option{-Wabi} (without the version), the version from this
2319 option is used for the warning.
2321 @item -fno-access-control
2322 @opindex fno-access-control
2323 Turn off all access checking. This switch is mainly useful for working
2324 around bugs in the access control code.
2327 @opindex faligned-new
2328 Enable support for C++17 @code{new} of types that require more
2329 alignment than @code{void* ::operator new(std::size_t)} provides. A
2330 numeric argument such as @code{-faligned-new=32} can be used to
2331 specify how much alignment (in bytes) is provided by that function,
2332 but few users will need to override the default of
2333 @code{alignof(std::max_align_t)}.
2335 This flag is enabled by default for @option{-std=c++17}.
2339 Check that the pointer returned by @code{operator new} is non-null
2340 before attempting to modify the storage allocated. This check is
2341 normally unnecessary because the C++ standard specifies that
2342 @code{operator new} only returns @code{0} if it is declared
2343 @code{throw()}, in which case the compiler always checks the
2344 return value even without this option. In all other cases, when
2345 @code{operator new} has a non-empty exception specification, memory
2346 exhaustion is signalled by throwing @code{std::bad_alloc}. See also
2347 @samp{new (nothrow)}.
2351 Enable support for the C++ Extensions for Concepts Technical
2352 Specification, ISO 19217 (2015), which allows code like
2355 template <class T> concept bool Addable = requires (T t) @{ t + t; @};
2356 template <Addable T> T add (T a, T b) @{ return a + b; @}
2359 @item -fconstexpr-depth=@var{n}
2360 @opindex fconstexpr-depth
2361 Set the maximum nested evaluation depth for C++11 constexpr functions
2362 to @var{n}. A limit is needed to detect endless recursion during
2363 constant expression evaluation. The minimum specified by the standard
2366 @item -fconstexpr-loop-limit=@var{n}
2367 @opindex fconstexpr-loop-limit
2368 Set the maximum number of iterations for a loop in C++14 constexpr functions
2369 to @var{n}. A limit is needed to detect infinite loops during
2370 constant expression evaluation. The default is 262144 (1<<18).
2372 @item -fdeduce-init-list
2373 @opindex fdeduce-init-list
2374 Enable deduction of a template type parameter as
2375 @code{std::initializer_list} from a brace-enclosed initializer list, i.e.@:
2378 template <class T> auto forward(T t) -> decltype (realfn (t))
2385 forward(@{1,2@}); // call forward<std::initializer_list<int>>
2389 This deduction was implemented as a possible extension to the
2390 originally proposed semantics for the C++11 standard, but was not part
2391 of the final standard, so it is disabled by default. This option is
2392 deprecated, and may be removed in a future version of G++.
2394 @item -ffriend-injection
2395 @opindex ffriend-injection
2396 Inject friend functions into the enclosing namespace, so that they are
2397 visible outside the scope of the class in which they are declared.
2398 Friend functions were documented to work this way in the old Annotated
2399 C++ Reference Manual.
2400 However, in ISO C++ a friend function that is not declared
2401 in an enclosing scope can only be found using argument dependent
2402 lookup. GCC defaults to the standard behavior.
2404 This option is for compatibility, and may be removed in a future
2407 @item -fno-elide-constructors
2408 @opindex fno-elide-constructors
2409 The C++ standard allows an implementation to omit creating a temporary
2410 that is only used to initialize another object of the same type.
2411 Specifying this option disables that optimization, and forces G++ to
2412 call the copy constructor in all cases. This option also causes G++
2413 to call trivial member functions which otherwise would be expanded inline.
2415 In C++17, the compiler is required to omit these temporaries, but this
2416 option still affects trivial member functions.
2418 @item -fno-enforce-eh-specs
2419 @opindex fno-enforce-eh-specs
2420 Don't generate code to check for violation of exception specifications
2421 at run time. This option violates the C++ standard, but may be useful
2422 for reducing code size in production builds, much like defining
2423 @code{NDEBUG}. This does not give user code permission to throw
2424 exceptions in violation of the exception specifications; the compiler
2425 still optimizes based on the specifications, so throwing an
2426 unexpected exception results in undefined behavior at run time.
2428 @item -fextern-tls-init
2429 @itemx -fno-extern-tls-init
2430 @opindex fextern-tls-init
2431 @opindex fno-extern-tls-init
2432 The C++11 and OpenMP standards allow @code{thread_local} and
2433 @code{threadprivate} variables to have dynamic (runtime)
2434 initialization. To support this, any use of such a variable goes
2435 through a wrapper function that performs any necessary initialization.
2436 When the use and definition of the variable are in the same
2437 translation unit, this overhead can be optimized away, but when the
2438 use is in a different translation unit there is significant overhead
2439 even if the variable doesn't actually need dynamic initialization. If
2440 the programmer can be sure that no use of the variable in a
2441 non-defining TU needs to trigger dynamic initialization (either
2442 because the variable is statically initialized, or a use of the
2443 variable in the defining TU will be executed before any uses in
2444 another TU), they can avoid this overhead with the
2445 @option{-fno-extern-tls-init} option.
2447 On targets that support symbol aliases, the default is
2448 @option{-fextern-tls-init}. On targets that do not support symbol
2449 aliases, the default is @option{-fno-extern-tls-init}.
2452 @itemx -fno-for-scope
2454 @opindex fno-for-scope
2455 If @option{-ffor-scope} is specified, the scope of variables declared in
2456 a @i{for-init-statement} is limited to the @code{for} loop itself,
2457 as specified by the C++ standard.
2458 If @option{-fno-for-scope} is specified, the scope of variables declared in
2459 a @i{for-init-statement} extends to the end of the enclosing scope,
2460 as was the case in old versions of G++, and other (traditional)
2461 implementations of C++.
2463 If neither flag is given, the default is to follow the standard,
2464 but to allow and give a warning for old-style code that would
2465 otherwise be invalid, or have different behavior.
2467 @item -fno-gnu-keywords
2468 @opindex fno-gnu-keywords
2469 Do not recognize @code{typeof} as a keyword, so that code can use this
2470 word as an identifier. You can use the keyword @code{__typeof__} instead.
2471 This option is implied by the strict ISO C++ dialects: @option{-ansi},
2472 @option{-std=c++98}, @option{-std=c++11}, etc.
2474 @item -fno-implicit-templates
2475 @opindex fno-implicit-templates
2476 Never emit code for non-inline templates that are instantiated
2477 implicitly (i.e.@: by use); only emit code for explicit instantiations.
2478 @xref{Template Instantiation}, for more information.
2480 @item -fno-implicit-inline-templates
2481 @opindex fno-implicit-inline-templates
2482 Don't emit code for implicit instantiations of inline templates, either.
2483 The default is to handle inlines differently so that compiles with and
2484 without optimization need the same set of explicit instantiations.
2486 @item -fno-implement-inlines
2487 @opindex fno-implement-inlines
2488 To save space, do not emit out-of-line copies of inline functions
2489 controlled by @code{#pragma implementation}. This causes linker
2490 errors if these functions are not inlined everywhere they are called.
2492 @item -fms-extensions
2493 @opindex fms-extensions
2494 Disable Wpedantic warnings about constructs used in MFC, such as implicit
2495 int and getting a pointer to member function via non-standard syntax.
2497 @item -fnew-inheriting-ctors
2498 @opindex fnew-inheriting-ctors
2499 Enable the P0136 adjustment to the semantics of C++11 constructor
2500 inheritance. This is part of C++17 but also considered to be a Defect
2501 Report against C++11 and C++14. This flag is enabled by default
2502 unless @option{-fabi-version=10} or lower is specified.
2504 @item -fnew-ttp-matching
2505 @opindex fnew-ttp-matching
2506 Enable the P0522 resolution to Core issue 150, template template
2507 parameters and default arguments: this allows a template with default
2508 template arguments as an argument for a template template parameter
2509 with fewer template parameters. This flag is enabled by default for
2510 @option{-std=c++17}.
2512 @item -fno-nonansi-builtins
2513 @opindex fno-nonansi-builtins
2514 Disable built-in declarations of functions that are not mandated by
2515 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
2516 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
2519 @opindex fnothrow-opt
2520 Treat a @code{throw()} exception specification as if it were a
2521 @code{noexcept} specification to reduce or eliminate the text size
2522 overhead relative to a function with no exception specification. If
2523 the function has local variables of types with non-trivial
2524 destructors, the exception specification actually makes the
2525 function smaller because the EH cleanups for those variables can be
2526 optimized away. The semantic effect is that an exception thrown out of
2527 a function with such an exception specification results in a call
2528 to @code{terminate} rather than @code{unexpected}.
2530 @item -fno-operator-names
2531 @opindex fno-operator-names
2532 Do not treat the operator name keywords @code{and}, @code{bitand},
2533 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
2534 synonyms as keywords.
2536 @item -fno-optional-diags
2537 @opindex fno-optional-diags
2538 Disable diagnostics that the standard says a compiler does not need to
2539 issue. Currently, the only such diagnostic issued by G++ is the one for
2540 a name having multiple meanings within a class.
2543 @opindex fpermissive
2544 Downgrade some diagnostics about nonconformant code from errors to
2545 warnings. Thus, using @option{-fpermissive} allows some
2546 nonconforming code to compile.
2548 @item -fno-pretty-templates
2549 @opindex fno-pretty-templates
2550 When an error message refers to a specialization of a function
2551 template, the compiler normally prints the signature of the
2552 template followed by the template arguments and any typedefs or
2553 typenames in the signature (e.g. @code{void f(T) [with T = int]}
2554 rather than @code{void f(int)}) so that it's clear which template is
2555 involved. When an error message refers to a specialization of a class
2556 template, the compiler omits any template arguments that match
2557 the default template arguments for that template. If either of these
2558 behaviors make it harder to understand the error message rather than
2559 easier, you can use @option{-fno-pretty-templates} to disable them.
2563 Enable automatic template instantiation at link time. This option also
2564 implies @option{-fno-implicit-templates}. @xref{Template
2565 Instantiation}, for more information.
2569 Disable generation of information about every class with virtual
2570 functions for use by the C++ run-time type identification features
2571 (@code{dynamic_cast} and @code{typeid}). If you don't use those parts
2572 of the language, you can save some space by using this flag. Note that
2573 exception handling uses the same information, but G++ generates it as
2574 needed. The @code{dynamic_cast} operator can still be used for casts that
2575 do not require run-time type information, i.e.@: casts to @code{void *} or to
2576 unambiguous base classes.
2578 @item -fsized-deallocation
2579 @opindex fsized-deallocation
2580 Enable the built-in global declarations
2582 void operator delete (void *, std::size_t) noexcept;
2583 void operator delete[] (void *, std::size_t) noexcept;
2585 as introduced in C++14. This is useful for user-defined replacement
2586 deallocation functions that, for example, use the size of the object
2587 to make deallocation faster. Enabled by default under
2588 @option{-std=c++14} and above. The flag @option{-Wsized-deallocation}
2589 warns about places that might want to add a definition.
2591 @item -fstrict-enums
2592 @opindex fstrict-enums
2593 Allow the compiler to optimize using the assumption that a value of
2594 enumerated type can only be one of the values of the enumeration (as
2595 defined in the C++ standard; basically, a value that can be
2596 represented in the minimum number of bits needed to represent all the
2597 enumerators). This assumption may not be valid if the program uses a
2598 cast to convert an arbitrary integer value to the enumerated type.
2600 @item -fstrong-eval-order
2601 @opindex fstrong-eval-order
2602 Evaluate member access, array subscripting, and shift expressions in
2603 left-to-right order, and evaluate assignment in right-to-left order,
2604 as adopted for C++17. Enabled by default with @option{-std=c++17}.
2605 @option{-fstrong-eval-order=some} enables just the ordering of member
2606 access and shift expressions, and is the default without
2607 @option{-std=c++17}.
2609 @item -ftemplate-backtrace-limit=@var{n}
2610 @opindex ftemplate-backtrace-limit
2611 Set the maximum number of template instantiation notes for a single
2612 warning or error to @var{n}. The default value is 10.
2614 @item -ftemplate-depth=@var{n}
2615 @opindex ftemplate-depth
2616 Set the maximum instantiation depth for template classes to @var{n}.
2617 A limit on the template instantiation depth is needed to detect
2618 endless recursions during template class instantiation. ANSI/ISO C++
2619 conforming programs must not rely on a maximum depth greater than 17
2620 (changed to 1024 in C++11). The default value is 900, as the compiler
2621 can run out of stack space before hitting 1024 in some situations.
2623 @item -fno-threadsafe-statics
2624 @opindex fno-threadsafe-statics
2625 Do not emit the extra code to use the routines specified in the C++
2626 ABI for thread-safe initialization of local statics. You can use this
2627 option to reduce code size slightly in code that doesn't need to be
2630 @item -fuse-cxa-atexit
2631 @opindex fuse-cxa-atexit
2632 Register destructors for objects with static storage duration with the
2633 @code{__cxa_atexit} function rather than the @code{atexit} function.
2634 This option is required for fully standards-compliant handling of static
2635 destructors, but only works if your C library supports
2636 @code{__cxa_atexit}.
2638 @item -fno-use-cxa-get-exception-ptr
2639 @opindex fno-use-cxa-get-exception-ptr
2640 Don't use the @code{__cxa_get_exception_ptr} runtime routine. This
2641 causes @code{std::uncaught_exception} to be incorrect, but is necessary
2642 if the runtime routine is not available.
2644 @item -fvisibility-inlines-hidden
2645 @opindex fvisibility-inlines-hidden
2646 This switch declares that the user does not attempt to compare
2647 pointers to inline functions or methods where the addresses of the two functions
2648 are taken in different shared objects.
2650 The effect of this is that GCC may, effectively, mark inline methods with
2651 @code{__attribute__ ((visibility ("hidden")))} so that they do not
2652 appear in the export table of a DSO and do not require a PLT indirection
2653 when used within the DSO@. Enabling this option can have a dramatic effect
2654 on load and link times of a DSO as it massively reduces the size of the
2655 dynamic export table when the library makes heavy use of templates.
2657 The behavior of this switch is not quite the same as marking the
2658 methods as hidden directly, because it does not affect static variables
2659 local to the function or cause the compiler to deduce that
2660 the function is defined in only one shared object.
2662 You may mark a method as having a visibility explicitly to negate the
2663 effect of the switch for that method. For example, if you do want to
2664 compare pointers to a particular inline method, you might mark it as
2665 having default visibility. Marking the enclosing class with explicit
2666 visibility has no effect.
2668 Explicitly instantiated inline methods are unaffected by this option
2669 as their linkage might otherwise cross a shared library boundary.
2670 @xref{Template Instantiation}.
2672 @item -fvisibility-ms-compat
2673 @opindex fvisibility-ms-compat
2674 This flag attempts to use visibility settings to make GCC's C++
2675 linkage model compatible with that of Microsoft Visual Studio.
2677 The flag makes these changes to GCC's linkage model:
2681 It sets the default visibility to @code{hidden}, like
2682 @option{-fvisibility=hidden}.
2685 Types, but not their members, are not hidden by default.
2688 The One Definition Rule is relaxed for types without explicit
2689 visibility specifications that are defined in more than one
2690 shared object: those declarations are permitted if they are
2691 permitted when this option is not used.
2694 In new code it is better to use @option{-fvisibility=hidden} and
2695 export those classes that are intended to be externally visible.
2696 Unfortunately it is possible for code to rely, perhaps accidentally,
2697 on the Visual Studio behavior.
2699 Among the consequences of these changes are that static data members
2700 of the same type with the same name but defined in different shared
2701 objects are different, so changing one does not change the other;
2702 and that pointers to function members defined in different shared
2703 objects may not compare equal. When this flag is given, it is a
2704 violation of the ODR to define types with the same name differently.
2708 Do not use weak symbol support, even if it is provided by the linker.
2709 By default, G++ uses weak symbols if they are available. This
2710 option exists only for testing, and should not be used by end-users;
2711 it results in inferior code and has no benefits. This option may
2712 be removed in a future release of G++.
2716 Do not search for header files in the standard directories specific to
2717 C++, but do still search the other standard directories. (This option
2718 is used when building the C++ library.)
2721 In addition, these optimization, warning, and code generation options
2722 have meanings only for C++ programs:
2725 @item -Wabi @r{(C, Objective-C, C++ and Objective-C++ only)}
2728 Warn when G++ it generates code that is probably not compatible with
2729 the vendor-neutral C++ ABI@. Since G++ now defaults to updating the
2730 ABI with each major release, normally @option{-Wabi} will warn only if
2731 there is a check added later in a release series for an ABI issue
2732 discovered since the initial release. @option{-Wabi} will warn about
2733 more things if an older ABI version is selected (with
2734 @option{-fabi-version=@var{n}}).
2736 @option{-Wabi} can also be used with an explicit version number to
2737 warn about compatibility with a particular @option{-fabi-version}
2738 level, e.g. @option{-Wabi=2} to warn about changes relative to
2739 @option{-fabi-version=2}.
2741 If an explicit version number is provided and
2742 @option{-fabi-compat-version} is not specified, the version number
2743 from this option is used for compatibility aliases. If no explicit
2744 version number is provided with this option, but
2745 @option{-fabi-compat-version} is specified, that version number is
2746 used for ABI warnings.
2748 Although an effort has been made to warn about
2749 all such cases, there are probably some cases that are not warned about,
2750 even though G++ is generating incompatible code. There may also be
2751 cases where warnings are emitted even though the code that is generated
2754 You should rewrite your code to avoid these warnings if you are
2755 concerned about the fact that code generated by G++ may not be binary
2756 compatible with code generated by other compilers.
2758 Known incompatibilities in @option{-fabi-version=2} (which was the
2759 default from GCC 3.4 to 4.9) include:
2764 A template with a non-type template parameter of reference type was
2765 mangled incorrectly:
2768 template <int &> struct S @{@};
2772 This was fixed in @option{-fabi-version=3}.
2775 SIMD vector types declared using @code{__attribute ((vector_size))} were
2776 mangled in a non-standard way that does not allow for overloading of
2777 functions taking vectors of different sizes.
2779 The mangling was changed in @option{-fabi-version=4}.
2782 @code{__attribute ((const))} and @code{noreturn} were mangled as type
2783 qualifiers, and @code{decltype} of a plain declaration was folded away.
2785 These mangling issues were fixed in @option{-fabi-version=5}.
2788 Scoped enumerators passed as arguments to a variadic function are
2789 promoted like unscoped enumerators, causing @code{va_arg} to complain.
2790 On most targets this does not actually affect the parameter passing
2791 ABI, as there is no way to pass an argument smaller than @code{int}.
2793 Also, the ABI changed the mangling of template argument packs,
2794 @code{const_cast}, @code{static_cast}, prefix increment/decrement, and
2795 a class scope function used as a template argument.
2797 These issues were corrected in @option{-fabi-version=6}.
2800 Lambdas in default argument scope were mangled incorrectly, and the
2801 ABI changed the mangling of @code{nullptr_t}.
2803 These issues were corrected in @option{-fabi-version=7}.
2806 When mangling a function type with function-cv-qualifiers, the
2807 un-qualified function type was incorrectly treated as a substitution
2810 This was fixed in @option{-fabi-version=8}, the default for GCC 5.1.
2813 @code{decltype(nullptr)} incorrectly had an alignment of 1, leading to
2814 unaligned accesses. Note that this did not affect the ABI of a
2815 function with a @code{nullptr_t} parameter, as parameters have a
2818 This was fixed in @option{-fabi-version=9}, the default for GCC 5.2.
2821 Target-specific attributes that affect the identity of a type, such as
2822 ia32 calling conventions on a function type (stdcall, regparm, etc.),
2823 did not affect the mangled name, leading to name collisions when
2824 function pointers were used as template arguments.
2826 This was fixed in @option{-fabi-version=10}, the default for GCC 6.1.
2830 It also warns about psABI-related changes. The known psABI changes at this
2836 For SysV/x86-64, unions with @code{long double} members are
2837 passed in memory as specified in psABI. For example:
2847 @code{union U} is always passed in memory.
2851 @item -Wabi-tag @r{(C++ and Objective-C++ only)}
2854 Warn when a type with an ABI tag is used in a context that does not
2855 have that ABI tag. See @ref{C++ Attributes} for more information
2858 @item -Wctor-dtor-privacy @r{(C++ and Objective-C++ only)}
2859 @opindex Wctor-dtor-privacy
2860 @opindex Wno-ctor-dtor-privacy
2861 Warn when a class seems unusable because all the constructors or
2862 destructors in that class are private, and it has neither friends nor
2863 public static member functions. Also warn if there are no non-private
2864 methods, and there's at least one private member function that isn't
2865 a constructor or destructor.
2867 @item -Wdelete-non-virtual-dtor @r{(C++ and Objective-C++ only)}
2868 @opindex Wdelete-non-virtual-dtor
2869 @opindex Wno-delete-non-virtual-dtor
2870 Warn when @code{delete} is used to destroy an instance of a class that
2871 has virtual functions and non-virtual destructor. It is unsafe to delete
2872 an instance of a derived class through a pointer to a base class if the
2873 base class does not have a virtual destructor. This warning is enabled
2876 @item -Wliteral-suffix @r{(C++ and Objective-C++ only)}
2877 @opindex Wliteral-suffix
2878 @opindex Wno-literal-suffix
2879 Warn when a string or character literal is followed by a ud-suffix which does
2880 not begin with an underscore. As a conforming extension, GCC treats such
2881 suffixes as separate preprocessing tokens in order to maintain backwards
2882 compatibility with code that uses formatting macros from @code{<inttypes.h>}.
2886 #define __STDC_FORMAT_MACROS
2887 #include <inttypes.h>
2892 printf("My int64: %" PRId64"\n", i64);
2896 In this case, @code{PRId64} is treated as a separate preprocessing token.
2898 Additionally, warn when a user-defined literal operator is declared with
2899 a literal suffix identifier that doesn't begin with an underscore. Literal
2900 suffix identifiers that don't begin with an underscore are reserved for
2901 future standardization.
2903 This warning is enabled by default.
2905 @item -Wlto-type-mismatch
2906 @opindex Wlto-type-mismatch
2907 @opindex Wno-lto-type-mismatch
2909 During the link-time optimization warn about type mismatches in
2910 global declarations from different compilation units.
2911 Requires @option{-flto} to be enabled. Enabled by default.
2913 @item -Wno-narrowing @r{(C++ and Objective-C++ only)}
2915 @opindex Wno-narrowing
2916 For C++11 and later standards, narrowing conversions are diagnosed by default,
2917 as required by the standard. A narrowing conversion from a constant produces
2918 an error, and a narrowing conversion from a non-constant produces a warning,
2919 but @option{-Wno-narrowing} suppresses the diagnostic.
2920 Note that this does not affect the meaning of well-formed code;
2921 narrowing conversions are still considered ill-formed in SFINAE contexts.
2923 With @option{-Wnarrowing} in C++98, warn when a narrowing
2924 conversion prohibited by C++11 occurs within
2928 int i = @{ 2.2 @}; // error: narrowing from double to int
2931 This flag is included in @option{-Wall} and @option{-Wc++11-compat}.
2933 @item -Wnoexcept @r{(C++ and Objective-C++ only)}
2935 @opindex Wno-noexcept
2936 Warn when a noexcept-expression evaluates to false because of a call
2937 to a function that does not have a non-throwing exception
2938 specification (i.e. @code{throw()} or @code{noexcept}) but is known by
2939 the compiler to never throw an exception.
2941 @item -Wnoexcept-type @r{(C++ and Objective-C++ only)}
2942 @opindex Wnoexcept-type
2943 @opindex Wno-noexcept-type
2944 Warn if the C++17 feature making @code{noexcept} part of a function
2945 type changes the mangled name of a symbol relative to C++14. Enabled
2946 by @option{-Wabi} and @option{-Wc++17-compat}.
2949 template <class T> void f(T t) @{ t(); @};
2951 void h() @{ f(g); @} // in C++14 calls f<void(*)()>, in C++17 calls f<void(*)()noexcept>
2954 @item -Wclass-memaccess @r{(C++ and Objective-C++ only)}
2955 @opindex Wclass-memaccess
2956 Warn when the destination of a call to a raw memory function such as
2957 @code{memset} or @code{memcpy} is an object of class type writing into which
2958 might bypass the class non-trivial or deleted constructor or copy assignment,
2959 violate const-correctness or encapsulation, or corrupt the virtual table.
2960 Modifying the representation of such objects may violate invariants maintained
2961 by member functions of the class. For example, the call to @code{memset}
2962 below is undefined becase it modifies a non-trivial class object and is,
2963 therefore, diagnosed. The safe way to either initialize or clear the storage
2964 of objects of such types is by using the appropriate constructor or assignment
2965 operator, if one is available.
2967 std::string str = "abc";
2968 memset (&str, 0, 3);
2970 The @option{-Wclass-memaccess} option is enabled by @option{-Wall}.
2972 @item -Wnon-virtual-dtor @r{(C++ and Objective-C++ only)}
2973 @opindex Wnon-virtual-dtor
2974 @opindex Wno-non-virtual-dtor
2975 Warn when a class has virtual functions and an accessible non-virtual
2976 destructor itself or in an accessible polymorphic base class, in which
2977 case it is possible but unsafe to delete an instance of a derived
2978 class through a pointer to the class itself or base class. This
2979 warning is automatically enabled if @option{-Weffc++} is specified.
2981 @item -Wregister @r{(C++ and Objective-C++ only)}
2983 @opindex Wno-register
2984 Warn on uses of the @code{register} storage class specifier, except
2985 when it is part of the GNU @ref{Explicit Register Variables} extension.
2986 The use of the @code{register} keyword as storage class specifier has
2987 been deprecated in C++11 and removed in C++17.
2988 Enabled by default with @option{-std=c++17}.
2990 @item -Wreorder @r{(C++ and Objective-C++ only)}
2992 @opindex Wno-reorder
2993 @cindex reordering, warning
2994 @cindex warning for reordering of member initializers
2995 Warn when the order of member initializers given in the code does not
2996 match the order in which they must be executed. For instance:
3002 A(): j (0), i (1) @{ @}
3007 The compiler rearranges the member initializers for @code{i}
3008 and @code{j} to match the declaration order of the members, emitting
3009 a warning to that effect. This warning is enabled by @option{-Wall}.
3011 @item -fext-numeric-literals @r{(C++ and Objective-C++ only)}
3012 @opindex fext-numeric-literals
3013 @opindex fno-ext-numeric-literals
3014 Accept imaginary, fixed-point, or machine-defined
3015 literal number suffixes as GNU extensions.
3016 When this option is turned off these suffixes are treated
3017 as C++11 user-defined literal numeric suffixes.
3018 This is on by default for all pre-C++11 dialects and all GNU dialects:
3019 @option{-std=c++98}, @option{-std=gnu++98}, @option{-std=gnu++11},
3020 @option{-std=gnu++14}.
3021 This option is off by default
3022 for ISO C++11 onwards (@option{-std=c++11}, ...).
3025 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
3028 @item -Weffc++ @r{(C++ and Objective-C++ only)}
3031 Warn about violations of the following style guidelines from Scott Meyers'
3032 @cite{Effective C++} series of books:
3036 Define a copy constructor and an assignment operator for classes
3037 with dynamically-allocated memory.
3040 Prefer initialization to assignment in constructors.
3043 Have @code{operator=} return a reference to @code{*this}.
3046 Don't try to return a reference when you must return an object.
3049 Distinguish between prefix and postfix forms of increment and
3050 decrement operators.
3053 Never overload @code{&&}, @code{||}, or @code{,}.
3057 This option also enables @option{-Wnon-virtual-dtor}, which is also
3058 one of the effective C++ recommendations. However, the check is
3059 extended to warn about the lack of virtual destructor in accessible
3060 non-polymorphic bases classes too.
3062 When selecting this option, be aware that the standard library
3063 headers do not obey all of these guidelines; use @samp{grep -v}
3064 to filter out those warnings.
3066 @item -Wstrict-null-sentinel @r{(C++ and Objective-C++ only)}
3067 @opindex Wstrict-null-sentinel
3068 @opindex Wno-strict-null-sentinel
3069 Warn about the use of an uncasted @code{NULL} as sentinel. When
3070 compiling only with GCC this is a valid sentinel, as @code{NULL} is defined
3071 to @code{__null}. Although it is a null pointer constant rather than a
3072 null pointer, it is guaranteed to be of the same size as a pointer.
3073 But this use is not portable across different compilers.
3075 @item -Wno-non-template-friend @r{(C++ and Objective-C++ only)}
3076 @opindex Wno-non-template-friend
3077 @opindex Wnon-template-friend
3078 Disable warnings when non-template friend functions are declared
3079 within a template. In very old versions of GCC that predate implementation
3080 of the ISO standard, declarations such as
3081 @samp{friend int foo(int)}, where the name of the friend is an unqualified-id,
3082 could be interpreted as a particular specialization of a template
3083 function; the warning exists to diagnose compatibility problems,
3084 and is enabled by default.
3086 @item -Wold-style-cast @r{(C++ and Objective-C++ only)}
3087 @opindex Wold-style-cast
3088 @opindex Wno-old-style-cast
3089 Warn if an old-style (C-style) cast to a non-void type is used within
3090 a C++ program. The new-style casts (@code{dynamic_cast},
3091 @code{static_cast}, @code{reinterpret_cast}, and @code{const_cast}) are
3092 less vulnerable to unintended effects and much easier to search for.
3094 @item -Woverloaded-virtual @r{(C++ and Objective-C++ only)}
3095 @opindex Woverloaded-virtual
3096 @opindex Wno-overloaded-virtual
3097 @cindex overloaded virtual function, warning
3098 @cindex warning for overloaded virtual function
3099 Warn when a function declaration hides virtual functions from a
3100 base class. For example, in:
3107 struct B: public A @{
3112 the @code{A} class version of @code{f} is hidden in @code{B}, and code
3123 @item -Wno-pmf-conversions @r{(C++ and Objective-C++ only)}
3124 @opindex Wno-pmf-conversions
3125 @opindex Wpmf-conversions
3126 Disable the diagnostic for converting a bound pointer to member function
3129 @item -Wsign-promo @r{(C++ and Objective-C++ only)}
3130 @opindex Wsign-promo
3131 @opindex Wno-sign-promo
3132 Warn when overload resolution chooses a promotion from unsigned or
3133 enumerated type to a signed type, over a conversion to an unsigned type of
3134 the same size. Previous versions of G++ tried to preserve
3135 unsignedness, but the standard mandates the current behavior.
3137 @item -Wtemplates @r{(C++ and Objective-C++ only)}
3139 Warn when a primary template declaration is encountered. Some coding
3140 rules disallow templates, and this may be used to enforce that rule.
3141 The warning is inactive inside a system header file, such as the STL, so
3142 one can still use the STL. One may also instantiate or specialize
3145 @item -Wmultiple-inheritance @r{(C++ and Objective-C++ only)}
3146 @opindex Wmultiple-inheritance
3147 Warn when a class is defined with multiple direct base classes. Some
3148 coding rules disallow multiple inheritance, and this may be used to
3149 enforce that rule. The warning is inactive inside a system header file,
3150 such as the STL, so one can still use the STL. One may also define
3151 classes that indirectly use multiple inheritance.
3153 @item -Wvirtual-inheritance
3154 @opindex Wvirtual-inheritance
3155 Warn when a class is defined with a virtual direct base class. Some
3156 coding rules disallow multiple inheritance, and this may be used to
3157 enforce that rule. The warning is inactive inside a system header file,
3158 such as the STL, so one can still use the STL. One may also define
3159 classes that indirectly use virtual inheritance.
3162 @opindex Wnamespaces
3163 Warn when a namespace definition is opened. Some coding rules disallow
3164 namespaces, and this may be used to enforce that rule. The warning is
3165 inactive inside a system header file, such as the STL, so one can still
3166 use the STL. One may also use using directives and qualified names.
3168 @item -Wno-terminate @r{(C++ and Objective-C++ only)}
3170 @opindex Wno-terminate
3171 Disable the warning about a throw-expression that will immediately
3172 result in a call to @code{terminate}.
3175 @node Objective-C and Objective-C++ Dialect Options
3176 @section Options Controlling Objective-C and Objective-C++ Dialects
3178 @cindex compiler options, Objective-C and Objective-C++
3179 @cindex Objective-C and Objective-C++ options, command-line
3180 @cindex options, Objective-C and Objective-C++
3181 (NOTE: This manual does not describe the Objective-C and Objective-C++
3182 languages themselves. @xref{Standards,,Language Standards
3183 Supported by GCC}, for references.)
3185 This section describes the command-line options that are only meaningful
3186 for Objective-C and Objective-C++ programs. You can also use most of
3187 the language-independent GNU compiler options.
3188 For example, you might compile a file @file{some_class.m} like this:
3191 gcc -g -fgnu-runtime -O -c some_class.m
3195 In this example, @option{-fgnu-runtime} is an option meant only for
3196 Objective-C and Objective-C++ programs; you can use the other options with
3197 any language supported by GCC@.
3199 Note that since Objective-C is an extension of the C language, Objective-C
3200 compilations may also use options specific to the C front-end (e.g.,
3201 @option{-Wtraditional}). Similarly, Objective-C++ compilations may use
3202 C++-specific options (e.g., @option{-Wabi}).
3204 Here is a list of options that are @emph{only} for compiling Objective-C
3205 and Objective-C++ programs:
3208 @item -fconstant-string-class=@var{class-name}
3209 @opindex fconstant-string-class
3210 Use @var{class-name} as the name of the class to instantiate for each
3211 literal string specified with the syntax @code{@@"@dots{}"}. The default
3212 class name is @code{NXConstantString} if the GNU runtime is being used, and
3213 @code{NSConstantString} if the NeXT runtime is being used (see below). The
3214 @option{-fconstant-cfstrings} option, if also present, overrides the
3215 @option{-fconstant-string-class} setting and cause @code{@@"@dots{}"} literals
3216 to be laid out as constant CoreFoundation strings.
3219 @opindex fgnu-runtime
3220 Generate object code compatible with the standard GNU Objective-C
3221 runtime. This is the default for most types of systems.
3223 @item -fnext-runtime
3224 @opindex fnext-runtime
3225 Generate output compatible with the NeXT runtime. This is the default
3226 for NeXT-based systems, including Darwin and Mac OS X@. The macro
3227 @code{__NEXT_RUNTIME__} is predefined if (and only if) this option is
3230 @item -fno-nil-receivers
3231 @opindex fno-nil-receivers
3232 Assume that all Objective-C message dispatches (@code{[receiver
3233 message:arg]}) in this translation unit ensure that the receiver is
3234 not @code{nil}. This allows for more efficient entry points in the
3235 runtime to be used. This option is only available in conjunction with
3236 the NeXT runtime and ABI version 0 or 1.
3238 @item -fobjc-abi-version=@var{n}
3239 @opindex fobjc-abi-version
3240 Use version @var{n} of the Objective-C ABI for the selected runtime.
3241 This option is currently supported only for the NeXT runtime. In that
3242 case, Version 0 is the traditional (32-bit) ABI without support for
3243 properties and other Objective-C 2.0 additions. Version 1 is the
3244 traditional (32-bit) ABI with support for properties and other
3245 Objective-C 2.0 additions. Version 2 is the modern (64-bit) ABI. If
3246 nothing is specified, the default is Version 0 on 32-bit target
3247 machines, and Version 2 on 64-bit target machines.
3249 @item -fobjc-call-cxx-cdtors
3250 @opindex fobjc-call-cxx-cdtors
3251 For each Objective-C class, check if any of its instance variables is a
3252 C++ object with a non-trivial default constructor. If so, synthesize a
3253 special @code{- (id) .cxx_construct} instance method which runs
3254 non-trivial default constructors on any such instance variables, in order,
3255 and then return @code{self}. Similarly, check if any instance variable
3256 is a C++ object with a non-trivial destructor, and if so, synthesize a
3257 special @code{- (void) .cxx_destruct} method which runs
3258 all such default destructors, in reverse order.
3260 The @code{- (id) .cxx_construct} and @code{- (void) .cxx_destruct}
3261 methods thusly generated only operate on instance variables
3262 declared in the current Objective-C class, and not those inherited
3263 from superclasses. It is the responsibility of the Objective-C
3264 runtime to invoke all such methods in an object's inheritance
3265 hierarchy. The @code{- (id) .cxx_construct} methods are invoked
3266 by the runtime immediately after a new object instance is allocated;
3267 the @code{- (void) .cxx_destruct} methods are invoked immediately
3268 before the runtime deallocates an object instance.
3270 As of this writing, only the NeXT runtime on Mac OS X 10.4 and later has
3271 support for invoking the @code{- (id) .cxx_construct} and
3272 @code{- (void) .cxx_destruct} methods.
3274 @item -fobjc-direct-dispatch
3275 @opindex fobjc-direct-dispatch
3276 Allow fast jumps to the message dispatcher. On Darwin this is
3277 accomplished via the comm page.
3279 @item -fobjc-exceptions
3280 @opindex fobjc-exceptions
3281 Enable syntactic support for structured exception handling in
3282 Objective-C, similar to what is offered by C++. This option
3283 is required to use the Objective-C keywords @code{@@try},
3284 @code{@@throw}, @code{@@catch}, @code{@@finally} and
3285 @code{@@synchronized}. This option is available with both the GNU
3286 runtime and the NeXT runtime (but not available in conjunction with
3287 the NeXT runtime on Mac OS X 10.2 and earlier).
3291 Enable garbage collection (GC) in Objective-C and Objective-C++
3292 programs. This option is only available with the NeXT runtime; the
3293 GNU runtime has a different garbage collection implementation that
3294 does not require special compiler flags.
3296 @item -fobjc-nilcheck
3297 @opindex fobjc-nilcheck
3298 For the NeXT runtime with version 2 of the ABI, check for a nil
3299 receiver in method invocations before doing the actual method call.
3300 This is the default and can be disabled using
3301 @option{-fno-objc-nilcheck}. Class methods and super calls are never
3302 checked for nil in this way no matter what this flag is set to.
3303 Currently this flag does nothing when the GNU runtime, or an older
3304 version of the NeXT runtime ABI, is used.
3306 @item -fobjc-std=objc1
3308 Conform to the language syntax of Objective-C 1.0, the language
3309 recognized by GCC 4.0. This only affects the Objective-C additions to
3310 the C/C++ language; it does not affect conformance to C/C++ standards,
3311 which is controlled by the separate C/C++ dialect option flags. When
3312 this option is used with the Objective-C or Objective-C++ compiler,
3313 any Objective-C syntax that is not recognized by GCC 4.0 is rejected.
3314 This is useful if you need to make sure that your Objective-C code can
3315 be compiled with older versions of GCC@.
3317 @item -freplace-objc-classes
3318 @opindex freplace-objc-classes
3319 Emit a special marker instructing @command{ld(1)} not to statically link in
3320 the resulting object file, and allow @command{dyld(1)} to load it in at
3321 run time instead. This is used in conjunction with the Fix-and-Continue
3322 debugging mode, where the object file in question may be recompiled and
3323 dynamically reloaded in the course of program execution, without the need
3324 to restart the program itself. Currently, Fix-and-Continue functionality
3325 is only available in conjunction with the NeXT runtime on Mac OS X 10.3
3330 When compiling for the NeXT runtime, the compiler ordinarily replaces calls
3331 to @code{objc_getClass("@dots{}")} (when the name of the class is known at
3332 compile time) with static class references that get initialized at load time,
3333 which improves run-time performance. Specifying the @option{-fzero-link} flag
3334 suppresses this behavior and causes calls to @code{objc_getClass("@dots{}")}
3335 to be retained. This is useful in Zero-Link debugging mode, since it allows
3336 for individual class implementations to be modified during program execution.
3337 The GNU runtime currently always retains calls to @code{objc_get_class("@dots{}")}
3338 regardless of command-line options.
3340 @item -fno-local-ivars
3341 @opindex fno-local-ivars
3342 @opindex flocal-ivars
3343 By default instance variables in Objective-C can be accessed as if
3344 they were local variables from within the methods of the class they're
3345 declared in. This can lead to shadowing between instance variables
3346 and other variables declared either locally inside a class method or
3347 globally with the same name. Specifying the @option{-fno-local-ivars}
3348 flag disables this behavior thus avoiding variable shadowing issues.
3350 @item -fivar-visibility=@r{[}public@r{|}protected@r{|}private@r{|}package@r{]}
3351 @opindex fivar-visibility
3352 Set the default instance variable visibility to the specified option
3353 so that instance variables declared outside the scope of any access
3354 modifier directives default to the specified visibility.
3358 Dump interface declarations for all classes seen in the source file to a
3359 file named @file{@var{sourcename}.decl}.
3361 @item -Wassign-intercept @r{(Objective-C and Objective-C++ only)}
3362 @opindex Wassign-intercept
3363 @opindex Wno-assign-intercept
3364 Warn whenever an Objective-C assignment is being intercepted by the
3367 @item -Wno-protocol @r{(Objective-C and Objective-C++ only)}
3368 @opindex Wno-protocol
3370 If a class is declared to implement a protocol, a warning is issued for
3371 every method in the protocol that is not implemented by the class. The
3372 default behavior is to issue a warning for every method not explicitly
3373 implemented in the class, even if a method implementation is inherited
3374 from the superclass. If you use the @option{-Wno-protocol} option, then
3375 methods inherited from the superclass are considered to be implemented,
3376 and no warning is issued for them.
3378 @item -Wselector @r{(Objective-C and Objective-C++ only)}
3380 @opindex Wno-selector
3381 Warn if multiple methods of different types for the same selector are
3382 found during compilation. The check is performed on the list of methods
3383 in the final stage of compilation. Additionally, a check is performed
3384 for each selector appearing in a @code{@@selector(@dots{})}
3385 expression, and a corresponding method for that selector has been found
3386 during compilation. Because these checks scan the method table only at
3387 the end of compilation, these warnings are not produced if the final
3388 stage of compilation is not reached, for example because an error is
3389 found during compilation, or because the @option{-fsyntax-only} option is
3392 @item -Wstrict-selector-match @r{(Objective-C and Objective-C++ only)}
3393 @opindex Wstrict-selector-match
3394 @opindex Wno-strict-selector-match
3395 Warn if multiple methods with differing argument and/or return types are
3396 found for a given selector when attempting to send a message using this
3397 selector to a receiver of type @code{id} or @code{Class}. When this flag
3398 is off (which is the default behavior), the compiler omits such warnings
3399 if any differences found are confined to types that share the same size
3402 @item -Wundeclared-selector @r{(Objective-C and Objective-C++ only)}
3403 @opindex Wundeclared-selector
3404 @opindex Wno-undeclared-selector
3405 Warn if a @code{@@selector(@dots{})} expression referring to an
3406 undeclared selector is found. A selector is considered undeclared if no
3407 method with that name has been declared before the
3408 @code{@@selector(@dots{})} expression, either explicitly in an
3409 @code{@@interface} or @code{@@protocol} declaration, or implicitly in
3410 an @code{@@implementation} section. This option always performs its
3411 checks as soon as a @code{@@selector(@dots{})} expression is found,
3412 while @option{-Wselector} only performs its checks in the final stage of
3413 compilation. This also enforces the coding style convention
3414 that methods and selectors must be declared before being used.
3416 @item -print-objc-runtime-info
3417 @opindex print-objc-runtime-info
3418 Generate C header describing the largest structure that is passed by
3423 @node Diagnostic Message Formatting Options
3424 @section Options to Control Diagnostic Messages Formatting
3425 @cindex options to control diagnostics formatting
3426 @cindex diagnostic messages
3427 @cindex message formatting
3429 Traditionally, diagnostic messages have been formatted irrespective of
3430 the output device's aspect (e.g.@: its width, @dots{}). You can use the
3431 options described below
3432 to control the formatting algorithm for diagnostic messages,
3433 e.g.@: how many characters per line, how often source location
3434 information should be reported. Note that some language front ends may not
3435 honor these options.
3438 @item -fmessage-length=@var{n}
3439 @opindex fmessage-length
3440 Try to format error messages so that they fit on lines of about
3441 @var{n} characters. If @var{n} is zero, then no line-wrapping is
3442 done; each error message appears on a single line. This is the
3443 default for all front ends.
3445 @item -fdiagnostics-show-location=once
3446 @opindex fdiagnostics-show-location
3447 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
3448 reporter to emit source location information @emph{once}; that is, in
3449 case the message is too long to fit on a single physical line and has to
3450 be wrapped, the source location won't be emitted (as prefix) again,
3451 over and over, in subsequent continuation lines. This is the default
3454 @item -fdiagnostics-show-location=every-line
3455 Only meaningful in line-wrapping mode. Instructs the diagnostic
3456 messages reporter to emit the same source location information (as
3457 prefix) for physical lines that result from the process of breaking
3458 a message which is too long to fit on a single line.
3460 @item -fdiagnostics-color[=@var{WHEN}]
3461 @itemx -fno-diagnostics-color
3462 @opindex fdiagnostics-color
3463 @cindex highlight, color
3464 @vindex GCC_COLORS @r{environment variable}
3465 Use color in diagnostics. @var{WHEN} is @samp{never}, @samp{always},
3466 or @samp{auto}. The default depends on how the compiler has been configured,
3467 it can be any of the above @var{WHEN} options or also @samp{never}
3468 if @env{GCC_COLORS} environment variable isn't present in the environment,
3469 and @samp{auto} otherwise.
3470 @samp{auto} means to use color only when the standard error is a terminal.
3471 The forms @option{-fdiagnostics-color} and @option{-fno-diagnostics-color} are
3472 aliases for @option{-fdiagnostics-color=always} and
3473 @option{-fdiagnostics-color=never}, respectively.
3475 The colors are defined by the environment variable @env{GCC_COLORS}.
3476 Its value is a colon-separated list of capabilities and Select Graphic
3477 Rendition (SGR) substrings. SGR commands are interpreted by the
3478 terminal or terminal emulator. (See the section in the documentation
3479 of your text terminal for permitted values and their meanings as
3480 character attributes.) These substring values are integers in decimal
3481 representation and can be concatenated with semicolons.
3482 Common values to concatenate include
3484 @samp{4} for underline,
3486 @samp{7} for inverse,
3487 @samp{39} for default foreground color,
3488 @samp{30} to @samp{37} for foreground colors,
3489 @samp{90} to @samp{97} for 16-color mode foreground colors,
3490 @samp{38;5;0} to @samp{38;5;255}
3491 for 88-color and 256-color modes foreground colors,
3492 @samp{49} for default background color,
3493 @samp{40} to @samp{47} for background colors,
3494 @samp{100} to @samp{107} for 16-color mode background colors,
3495 and @samp{48;5;0} to @samp{48;5;255}
3496 for 88-color and 256-color modes background colors.
3498 The default @env{GCC_COLORS} is
3500 error=01;31:warning=01;35:note=01;36:range1=32:range2=34:locus=01:\
3501 quote=01:fixit-insert=32:fixit-delete=31:\
3502 diff-filename=01:diff-hunk=32:diff-delete=31:diff-insert=32:\
3506 where @samp{01;31} is bold red, @samp{01;35} is bold magenta,
3507 @samp{01;36} is bold cyan, @samp{32} is green, @samp{34} is blue,
3508 @samp{01} is bold, and @samp{31} is red.
3509 Setting @env{GCC_COLORS} to the empty string disables colors.
3510 Supported capabilities are as follows.
3514 @vindex error GCC_COLORS @r{capability}
3515 SGR substring for error: markers.
3518 @vindex warning GCC_COLORS @r{capability}
3519 SGR substring for warning: markers.
3522 @vindex note GCC_COLORS @r{capability}
3523 SGR substring for note: markers.
3526 @vindex range1 GCC_COLORS @r{capability}
3527 SGR substring for first additional range.
3530 @vindex range2 GCC_COLORS @r{capability}
3531 SGR substring for second additional range.
3534 @vindex locus GCC_COLORS @r{capability}
3535 SGR substring for location information, @samp{file:line} or
3536 @samp{file:line:column} etc.
3539 @vindex quote GCC_COLORS @r{capability}
3540 SGR substring for information printed within quotes.
3543 @vindex fixit-insert GCC_COLORS @r{capability}
3544 SGR substring for fix-it hints suggesting text to
3545 be inserted or replaced.
3548 @vindex fixit-delete GCC_COLORS @r{capability}
3549 SGR substring for fix-it hints suggesting text to
3552 @item diff-filename=
3553 @vindex diff-filename GCC_COLORS @r{capability}
3554 SGR substring for filename headers within generated patches.
3557 @vindex diff-hunk GCC_COLORS @r{capability}
3558 SGR substring for the starts of hunks within generated patches.
3561 @vindex diff-delete GCC_COLORS @r{capability}
3562 SGR substring for deleted lines within generated patches.
3565 @vindex diff-insert GCC_COLORS @r{capability}
3566 SGR substring for inserted lines within generated patches.
3569 @vindex type-diff GCC_COLORS @r{capability}
3570 SGR substring for highlighting mismatching types within template
3571 arguments in the C++ frontend.
3574 @item -fno-diagnostics-show-option
3575 @opindex fno-diagnostics-show-option
3576 @opindex fdiagnostics-show-option
3577 By default, each diagnostic emitted includes text indicating the
3578 command-line option that directly controls the diagnostic (if such an
3579 option is known to the diagnostic machinery). Specifying the
3580 @option{-fno-diagnostics-show-option} flag suppresses that behavior.
3582 @item -fno-diagnostics-show-caret
3583 @opindex fno-diagnostics-show-caret
3584 @opindex fdiagnostics-show-caret
3585 By default, each diagnostic emitted includes the original source line
3586 and a caret @samp{^} indicating the column. This option suppresses this
3587 information. The source line is truncated to @var{n} characters, if
3588 the @option{-fmessage-length=n} option is given. When the output is done
3589 to the terminal, the width is limited to the width given by the
3590 @env{COLUMNS} environment variable or, if not set, to the terminal width.
3592 @item -fdiagnostics-parseable-fixits
3593 @opindex fdiagnostics-parseable-fixits
3594 Emit fix-it hints in a machine-parseable format, suitable for consumption
3595 by IDEs. For each fix-it, a line will be printed after the relevant
3596 diagnostic, starting with the string ``fix-it:''. For example:
3599 fix-it:"test.c":@{45:3-45:21@}:"gtk_widget_show_all"
3602 The location is expressed as a half-open range, expressed as a count of
3603 bytes, starting at byte 1 for the initial column. In the above example,
3604 bytes 3 through 20 of line 45 of ``test.c'' are to be replaced with the
3608 00000000011111111112222222222
3609 12345678901234567890123456789
3610 gtk_widget_showall (dlg);
3615 The filename and replacement string escape backslash as ``\\", tab as ``\t'',
3616 newline as ``\n'', double quotes as ``\"'', non-printable characters as octal
3617 (e.g. vertical tab as ``\013'').
3619 An empty replacement string indicates that the given range is to be removed.
3620 An empty range (e.g. ``45:3-45:3'') indicates that the string is to
3621 be inserted at the given position.
3623 @item -fdiagnostics-generate-patch
3624 @opindex fdiagnostics-generate-patch
3625 Print fix-it hints to stderr in unified diff format, after any diagnostics
3626 are printed. For example:
3633 void show_cb(GtkDialog *dlg)
3635 - gtk_widget_showall(dlg);
3636 + gtk_widget_show_all(dlg);
3641 The diff may or may not be colorized, following the same rules
3642 as for diagnostics (see @option{-fdiagnostics-color}).
3644 @item -fdiagnostics-show-template-tree
3645 @opindex fdiagnostics-show-template-tree
3647 In the C++ frontend, when printing diagnostics showing mismatching
3648 template types, such as:
3651 could not convert 'std::map<int, std::vector<double> >()'
3652 from 'map<[...],vector<double>>' to 'map<[...],vector<float>>
3655 the @option{-fdiagnostics-show-template-tree} flag enables printing a
3656 tree-like structure showing the common and differing parts of the types,
3666 The parts that differ are highlighted with color (``double'' and
3667 ``float'' in this case).
3669 @item -fno-elide-type
3670 @opindex fno-elide-type
3671 @opindex felide-type
3672 By default when the C++ frontend prints diagnostics showing mismatching
3673 template types, common parts of the types are printed as ``[...]'' to
3674 simplify the error message. For example:
3677 could not convert 'std::map<int, std::vector<double> >()'
3678 from 'map<[...],vector<double>>' to 'map<[...],vector<float>>
3681 Specifying the @option{-fno-elide-type} flag suppresses that behavior.
3682 This flag also affects the output of the
3683 @option{-fdiagnostics-show-template-tree} flag.
3685 @item -fno-show-column
3686 @opindex fno-show-column
3687 Do not print column numbers in diagnostics. This may be necessary if
3688 diagnostics are being scanned by a program that does not understand the
3689 column numbers, such as @command{dejagnu}.
3693 @node Warning Options
3694 @section Options to Request or Suppress Warnings
3695 @cindex options to control warnings
3696 @cindex warning messages
3697 @cindex messages, warning
3698 @cindex suppressing warnings
3700 Warnings are diagnostic messages that report constructions that
3701 are not inherently erroneous but that are risky or suggest there
3702 may have been an error.
3704 The following language-independent options do not enable specific
3705 warnings but control the kinds of diagnostics produced by GCC@.
3708 @cindex syntax checking
3710 @opindex fsyntax-only
3711 Check the code for syntax errors, but don't do anything beyond that.
3713 @item -fmax-errors=@var{n}
3714 @opindex fmax-errors
3715 Limits the maximum number of error messages to @var{n}, at which point
3716 GCC bails out rather than attempting to continue processing the source
3717 code. If @var{n} is 0 (the default), there is no limit on the number
3718 of error messages produced. If @option{-Wfatal-errors} is also
3719 specified, then @option{-Wfatal-errors} takes precedence over this
3724 Inhibit all warning messages.
3729 Make all warnings into errors.
3734 Make the specified warning into an error. The specifier for a warning
3735 is appended; for example @option{-Werror=switch} turns the warnings
3736 controlled by @option{-Wswitch} into errors. This switch takes a
3737 negative form, to be used to negate @option{-Werror} for specific
3738 warnings; for example @option{-Wno-error=switch} makes
3739 @option{-Wswitch} warnings not be errors, even when @option{-Werror}
3742 The warning message for each controllable warning includes the
3743 option that controls the warning. That option can then be used with
3744 @option{-Werror=} and @option{-Wno-error=} as described above.
3745 (Printing of the option in the warning message can be disabled using the
3746 @option{-fno-diagnostics-show-option} flag.)
3748 Note that specifying @option{-Werror=}@var{foo} automatically implies
3749 @option{-W}@var{foo}. However, @option{-Wno-error=}@var{foo} does not
3752 @item -Wfatal-errors
3753 @opindex Wfatal-errors
3754 @opindex Wno-fatal-errors
3755 This option causes the compiler to abort compilation on the first error
3756 occurred rather than trying to keep going and printing further error
3761 You can request many specific warnings with options beginning with
3762 @samp{-W}, for example @option{-Wimplicit} to request warnings on
3763 implicit declarations. Each of these specific warning options also
3764 has a negative form beginning @samp{-Wno-} to turn off warnings; for
3765 example, @option{-Wno-implicit}. This manual lists only one of the
3766 two forms, whichever is not the default. For further
3767 language-specific options also refer to @ref{C++ Dialect Options} and
3768 @ref{Objective-C and Objective-C++ Dialect Options}.
3770 Some options, such as @option{-Wall} and @option{-Wextra}, turn on other
3771 options, such as @option{-Wunused}, which may turn on further options,
3772 such as @option{-Wunused-value}. The combined effect of positive and
3773 negative forms is that more specific options have priority over less
3774 specific ones, independently of their position in the command-line. For
3775 options of the same specificity, the last one takes effect. Options
3776 enabled or disabled via pragmas (@pxref{Diagnostic Pragmas}) take effect
3777 as if they appeared at the end of the command-line.
3779 When an unrecognized warning option is requested (e.g.,
3780 @option{-Wunknown-warning}), GCC emits a diagnostic stating
3781 that the option is not recognized. However, if the @option{-Wno-} form
3782 is used, the behavior is slightly different: no diagnostic is
3783 produced for @option{-Wno-unknown-warning} unless other diagnostics
3784 are being produced. This allows the use of new @option{-Wno-} options
3785 with old compilers, but if something goes wrong, the compiler
3786 warns that an unrecognized option is present.
3793 Issue all the warnings demanded by strict ISO C and ISO C++;
3794 reject all programs that use forbidden extensions, and some other
3795 programs that do not follow ISO C and ISO C++. For ISO C, follows the
3796 version of the ISO C standard specified by any @option{-std} option used.
3798 Valid ISO C and ISO C++ programs should compile properly with or without
3799 this option (though a rare few require @option{-ansi} or a
3800 @option{-std} option specifying the required version of ISO C)@. However,
3801 without this option, certain GNU extensions and traditional C and C++
3802 features are supported as well. With this option, they are rejected.
3804 @option{-Wpedantic} does not cause warning messages for use of the
3805 alternate keywords whose names begin and end with @samp{__}. Pedantic
3806 warnings are also disabled in the expression that follows
3807 @code{__extension__}. However, only system header files should use
3808 these escape routes; application programs should avoid them.
3809 @xref{Alternate Keywords}.
3811 Some users try to use @option{-Wpedantic} to check programs for strict ISO
3812 C conformance. They soon find that it does not do quite what they want:
3813 it finds some non-ISO practices, but not all---only those for which
3814 ISO C @emph{requires} a diagnostic, and some others for which
3815 diagnostics have been added.
3817 A feature to report any failure to conform to ISO C might be useful in
3818 some instances, but would require considerable additional work and would
3819 be quite different from @option{-Wpedantic}. We don't have plans to
3820 support such a feature in the near future.
3822 Where the standard specified with @option{-std} represents a GNU
3823 extended dialect of C, such as @samp{gnu90} or @samp{gnu99}, there is a
3824 corresponding @dfn{base standard}, the version of ISO C on which the GNU
3825 extended dialect is based. Warnings from @option{-Wpedantic} are given
3826 where they are required by the base standard. (It does not make sense
3827 for such warnings to be given only for features not in the specified GNU
3828 C dialect, since by definition the GNU dialects of C include all
3829 features the compiler supports with the given option, and there would be
3830 nothing to warn about.)
3832 @item -pedantic-errors
3833 @opindex pedantic-errors
3834 Give an error whenever the @dfn{base standard} (see @option{-Wpedantic})
3835 requires a diagnostic, in some cases where there is undefined behavior
3836 at compile-time and in some other cases that do not prevent compilation
3837 of programs that are valid according to the standard. This is not
3838 equivalent to @option{-Werror=pedantic}, since there are errors enabled
3839 by this option and not enabled by the latter and vice versa.
3844 This enables all the warnings about constructions that some users
3845 consider questionable, and that are easy to avoid (or modify to
3846 prevent the warning), even in conjunction with macros. This also
3847 enables some language-specific warnings described in @ref{C++ Dialect
3848 Options} and @ref{Objective-C and Objective-C++ Dialect Options}.
3850 @option{-Wall} turns on the following warning flags:
3852 @gccoptlist{-Waddress @gol
3853 -Warray-bounds=1 @r{(only with} @option{-O2}@r{)} @gol
3855 -Wbool-operation @gol
3856 -Wc++11-compat -Wc++14-compat @gol
3857 -Wcatch-value @r{(C++ and Objective-C++ only)} @gol
3858 -Wchar-subscripts @gol
3860 -Wduplicate-decl-specifier @r{(C and Objective-C only)} @gol
3861 -Wenum-compare @r{(in C/ObjC; this is on by default in C++)} @gol
3863 -Wint-in-bool-context @gol
3864 -Wimplicit @r{(C and Objective-C only)} @gol
3865 -Wimplicit-int @r{(C and Objective-C only)} @gol
3866 -Wimplicit-function-declaration @r{(C and Objective-C only)} @gol
3867 -Winit-self @r{(only for C++)} @gol
3868 -Wlogical-not-parentheses @gol
3869 -Wmain @r{(only for C/ObjC and unless} @option{-ffreestanding}@r{)} @gol
3870 -Wmaybe-uninitialized @gol
3871 -Wmemset-elt-size @gol
3872 -Wmemset-transposed-args @gol
3873 -Wmisleading-indentation @r{(only for C/C++)} @gol
3874 -Wmissing-braces @r{(only for C/ObjC)} @gol
3875 -Wmultistatement-macros @gol
3876 -Wnarrowing @r{(only for C++)} @gol
3878 -Wnonnull-compare @gol
3884 -Wsequence-point @gol
3885 -Wsign-compare @r{(only in C++)} @gol
3886 -Wsizeof-pointer-div @gol
3887 -Wsizeof-pointer-memaccess @gol
3888 -Wstrict-aliasing @gol
3889 -Wstrict-overflow=1 @gol
3891 -Wtautological-compare @gol
3893 -Wuninitialized @gol
3894 -Wunknown-pragmas @gol
3895 -Wunused-function @gol
3898 -Wunused-variable @gol
3899 -Wvolatile-register-var @gol
3902 Note that some warning flags are not implied by @option{-Wall}. Some of
3903 them warn about constructions that users generally do not consider
3904 questionable, but which occasionally you might wish to check for;
3905 others warn about constructions that are necessary or hard to avoid in
3906 some cases, and there is no simple way to modify the code to suppress
3907 the warning. Some of them are enabled by @option{-Wextra} but many of
3908 them must be enabled individually.
3914 This enables some extra warning flags that are not enabled by
3915 @option{-Wall}. (This option used to be called @option{-W}. The older
3916 name is still supported, but the newer name is more descriptive.)
3918 @gccoptlist{-Wclobbered @gol
3920 -Wignored-qualifiers @gol
3921 -Wimplicit-fallthrough=3 @gol
3922 -Wmissing-field-initializers @gol
3923 -Wmissing-parameter-type @r{(C only)} @gol
3924 -Wold-style-declaration @r{(C only)} @gol
3925 -Woverride-init @gol
3926 -Wsign-compare @r{(C only)} @gol
3928 -Wuninitialized @gol
3929 -Wshift-negative-value @r{(in C++03 and in C99 and newer)} @gol
3930 -Wunused-parameter @r{(only with} @option{-Wunused} @r{or} @option{-Wall}@r{)} @gol
3931 -Wunused-but-set-parameter @r{(only with} @option{-Wunused} @r{or} @option{-Wall}@r{)} @gol
3934 The option @option{-Wextra} also prints warning messages for the
3940 A pointer is compared against integer zero with @code{<}, @code{<=},
3941 @code{>}, or @code{>=}.
3944 (C++ only) An enumerator and a non-enumerator both appear in a
3945 conditional expression.
3948 (C++ only) Ambiguous virtual bases.
3951 (C++ only) Subscripting an array that has been declared @code{register}.
3954 (C++ only) Taking the address of a variable that has been declared
3958 (C++ only) A base class is not initialized in the copy constructor
3963 @item -Wchar-subscripts
3964 @opindex Wchar-subscripts
3965 @opindex Wno-char-subscripts
3966 Warn if an array subscript has type @code{char}. This is a common cause
3967 of error, as programmers often forget that this type is signed on some
3969 This warning is enabled by @option{-Wall}.
3973 Warn about an invalid memory access that is found by Pointer Bounds Checker
3974 (@option{-fcheck-pointer-bounds}).
3976 @item -Wno-coverage-mismatch
3977 @opindex Wno-coverage-mismatch
3978 Warn if feedback profiles do not match when using the
3979 @option{-fprofile-use} option.
3980 If a source file is changed between compiling with @option{-fprofile-gen} and
3981 with @option{-fprofile-use}, the files with the profile feedback can fail
3982 to match the source file and GCC cannot use the profile feedback
3983 information. By default, this warning is enabled and is treated as an
3984 error. @option{-Wno-coverage-mismatch} can be used to disable the
3985 warning or @option{-Wno-error=coverage-mismatch} can be used to
3986 disable the error. Disabling the error for this warning can result in
3987 poorly optimized code and is useful only in the
3988 case of very minor changes such as bug fixes to an existing code-base.
3989 Completely disabling the warning is not recommended.
3992 @r{(C, Objective-C, C++, Objective-C++ and Fortran only)}
3994 Suppress warning messages emitted by @code{#warning} directives.
3996 @item -Wdouble-promotion @r{(C, C++, Objective-C and Objective-C++ only)}
3997 @opindex Wdouble-promotion
3998 @opindex Wno-double-promotion
3999 Give a warning when a value of type @code{float} is implicitly
4000 promoted to @code{double}. CPUs with a 32-bit ``single-precision''
4001 floating-point unit implement @code{float} in hardware, but emulate
4002 @code{double} in software. On such a machine, doing computations
4003 using @code{double} values is much more expensive because of the
4004 overhead required for software emulation.
4006 It is easy to accidentally do computations with @code{double} because
4007 floating-point literals are implicitly of type @code{double}. For
4011 float area(float radius)
4013 return 3.14159 * radius * radius;
4017 the compiler performs the entire computation with @code{double}
4018 because the floating-point literal is a @code{double}.
4020 @item -Wduplicate-decl-specifier @r{(C and Objective-C only)}
4021 @opindex Wduplicate-decl-specifier
4022 @opindex Wno-duplicate-decl-specifier
4023 Warn if a declaration has duplicate @code{const}, @code{volatile},
4024 @code{restrict} or @code{_Atomic} specifier. This warning is enabled by
4028 @itemx -Wformat=@var{n}
4031 @opindex ffreestanding
4032 @opindex fno-builtin
4034 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
4035 the arguments supplied have types appropriate to the format string
4036 specified, and that the conversions specified in the format string make
4037 sense. This includes standard functions, and others specified by format
4038 attributes (@pxref{Function Attributes}), in the @code{printf},
4039 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
4040 not in the C standard) families (or other target-specific families).
4041 Which functions are checked without format attributes having been
4042 specified depends on the standard version selected, and such checks of
4043 functions without the attribute specified are disabled by
4044 @option{-ffreestanding} or @option{-fno-builtin}.
4046 The formats are checked against the format features supported by GNU
4047 libc version 2.2. These include all ISO C90 and C99 features, as well
4048 as features from the Single Unix Specification and some BSD and GNU
4049 extensions. Other library implementations may not support all these
4050 features; GCC does not support warning about features that go beyond a
4051 particular library's limitations. However, if @option{-Wpedantic} is used
4052 with @option{-Wformat}, warnings are given about format features not
4053 in the selected standard version (but not for @code{strfmon} formats,
4054 since those are not in any version of the C standard). @xref{C Dialect
4055 Options,,Options Controlling C Dialect}.
4062 Option @option{-Wformat} is equivalent to @option{-Wformat=1}, and
4063 @option{-Wno-format} is equivalent to @option{-Wformat=0}. Since
4064 @option{-Wformat} also checks for null format arguments for several
4065 functions, @option{-Wformat} also implies @option{-Wnonnull}. Some
4066 aspects of this level of format checking can be disabled by the
4067 options: @option{-Wno-format-contains-nul},
4068 @option{-Wno-format-extra-args}, and @option{-Wno-format-zero-length}.
4069 @option{-Wformat} is enabled by @option{-Wall}.
4071 @item -Wno-format-contains-nul
4072 @opindex Wno-format-contains-nul
4073 @opindex Wformat-contains-nul
4074 If @option{-Wformat} is specified, do not warn about format strings that
4077 @item -Wno-format-extra-args
4078 @opindex Wno-format-extra-args
4079 @opindex Wformat-extra-args
4080 If @option{-Wformat} is specified, do not warn about excess arguments to a
4081 @code{printf} or @code{scanf} format function. The C standard specifies
4082 that such arguments are ignored.
4084 Where the unused arguments lie between used arguments that are
4085 specified with @samp{$} operand number specifications, normally
4086 warnings are still given, since the implementation could not know what
4087 type to pass to @code{va_arg} to skip the unused arguments. However,
4088 in the case of @code{scanf} formats, this option suppresses the
4089 warning if the unused arguments are all pointers, since the Single
4090 Unix Specification says that such unused arguments are allowed.
4092 @item -Wformat-overflow
4093 @itemx -Wformat-overflow=@var{level}
4094 @opindex Wformat-overflow
4095 @opindex Wno-format-overflow
4096 Warn about calls to formatted input/output functions such as @code{sprintf}
4097 and @code{vsprintf} that might overflow the destination buffer. When the
4098 exact number of bytes written by a format directive cannot be determined
4099 at compile-time it is estimated based on heuristics that depend on the
4100 @var{level} argument and on optimization. While enabling optimization
4101 will in most cases improve the accuracy of the warning, it may also
4102 result in false positives.
4105 @item -Wformat-overflow
4106 @item -Wformat-overflow=1
4107 @opindex Wformat-overflow
4108 @opindex Wno-format-overflow
4109 Level @var{1} of @option{-Wformat-overflow} enabled by @option{-Wformat}
4110 employs a conservative approach that warns only about calls that most
4111 likely overflow the buffer. At this level, numeric arguments to format
4112 directives with unknown values are assumed to have the value of one, and
4113 strings of unknown length to be empty. Numeric arguments that are known
4114 to be bounded to a subrange of their type, or string arguments whose output
4115 is bounded either by their directive's precision or by a finite set of
4116 string literals, are assumed to take on the value within the range that
4117 results in the most bytes on output. For example, the call to @code{sprintf}
4118 below is diagnosed because even with both @var{a} and @var{b} equal to zero,
4119 the terminating NUL character (@code{'\0'}) appended by the function
4120 to the destination buffer will be written past its end. Increasing
4121 the size of the buffer by a single byte is sufficient to avoid the
4122 warning, though it may not be sufficient to avoid the overflow.
4125 void f (int a, int b)
4128 sprintf (buf, "a = %i, b = %i\n", a, b);
4132 @item -Wformat-overflow=2
4133 Level @var{2} warns also about calls that might overflow the destination
4134 buffer given an argument of sufficient length or magnitude. At level
4135 @var{2}, unknown numeric arguments are assumed to have the minimum
4136 representable value for signed types with a precision greater than 1, and
4137 the maximum representable value otherwise. Unknown string arguments whose
4138 length cannot be assumed to be bounded either by the directive's precision,
4139 or by a finite set of string literals they may evaluate to, or the character
4140 array they may point to, are assumed to be 1 character long.
4142 At level @var{2}, the call in the example above is again diagnosed, but
4143 this time because with @var{a} equal to a 32-bit @code{INT_MIN} the first
4144 @code{%i} directive will write some of its digits beyond the end of
4145 the destination buffer. To make the call safe regardless of the values
4146 of the two variables, the size of the destination buffer must be increased
4147 to at least 34 bytes. GCC includes the minimum size of the buffer in
4148 an informational note following the warning.
4150 An alternative to increasing the size of the destination buffer is to
4151 constrain the range of formatted values. The maximum length of string
4152 arguments can be bounded by specifying the precision in the format
4153 directive. When numeric arguments of format directives can be assumed
4154 to be bounded by less than the precision of their type, choosing
4155 an appropriate length modifier to the format specifier will reduce
4156 the required buffer size. For example, if @var{a} and @var{b} in the
4157 example above can be assumed to be within the precision of
4158 the @code{short int} type then using either the @code{%hi} format
4159 directive or casting the argument to @code{short} reduces the maximum
4160 required size of the buffer to 24 bytes.
4163 void f (int a, int b)
4166 sprintf (buf, "a = %hi, b = %i\n", a, (short)b);
4171 @item -Wno-format-zero-length
4172 @opindex Wno-format-zero-length
4173 @opindex Wformat-zero-length
4174 If @option{-Wformat} is specified, do not warn about zero-length formats.
4175 The C standard specifies that zero-length formats are allowed.
4180 Enable @option{-Wformat} plus additional format checks. Currently
4181 equivalent to @option{-Wformat -Wformat-nonliteral -Wformat-security
4184 @item -Wformat-nonliteral
4185 @opindex Wformat-nonliteral
4186 @opindex Wno-format-nonliteral
4187 If @option{-Wformat} is specified, also warn if the format string is not a
4188 string literal and so cannot be checked, unless the format function
4189 takes its format arguments as a @code{va_list}.
4191 @item -Wformat-security
4192 @opindex Wformat-security
4193 @opindex Wno-format-security
4194 If @option{-Wformat} is specified, also warn about uses of format
4195 functions that represent possible security problems. At present, this
4196 warns about calls to @code{printf} and @code{scanf} functions where the
4197 format string is not a string literal and there are no format arguments,
4198 as in @code{printf (foo);}. This may be a security hole if the format
4199 string came from untrusted input and contains @samp{%n}. (This is
4200 currently a subset of what @option{-Wformat-nonliteral} warns about, but
4201 in future warnings may be added to @option{-Wformat-security} that are not
4202 included in @option{-Wformat-nonliteral}.)
4204 @item -Wformat-signedness
4205 @opindex Wformat-signedness
4206 @opindex Wno-format-signedness
4207 If @option{-Wformat} is specified, also warn if the format string
4208 requires an unsigned argument and the argument is signed and vice versa.
4210 @item -Wformat-truncation
4211 @itemx -Wformat-truncation=@var{level}
4212 @opindex Wformat-truncation
4213 @opindex Wno-format-truncation
4214 Warn about calls to formatted input/output functions such as @code{snprintf}
4215 and @code{vsnprintf} that might result in output truncation. When the exact
4216 number of bytes written by a format directive cannot be determined at
4217 compile-time it is estimated based on heuristics that depend on
4218 the @var{level} argument and on optimization. While enabling optimization
4219 will in most cases improve the accuracy of the warning, it may also result
4220 in false positives. Except as noted otherwise, the option uses the same
4221 logic @option{-Wformat-overflow}.
4224 @item -Wformat-truncation
4225 @item -Wformat-truncation=1
4226 @opindex Wformat-truncation
4227 @opindex Wno-format-overflow
4228 Level @var{1} of @option{-Wformat-truncation} enabled by @option{-Wformat}
4229 employs a conservative approach that warns only about calls to bounded
4230 functions whose return value is unused and that will most likely result
4231 in output truncation.
4233 @item -Wformat-truncation=2
4234 Level @var{2} warns also about calls to bounded functions whose return
4235 value is used and that might result in truncation given an argument of
4236 sufficient length or magnitude.
4240 @opindex Wformat-y2k
4241 @opindex Wno-format-y2k
4242 If @option{-Wformat} is specified, also warn about @code{strftime}
4243 formats that may yield only a two-digit year.
4248 @opindex Wno-nonnull
4249 Warn about passing a null pointer for arguments marked as
4250 requiring a non-null value by the @code{nonnull} function attribute.
4252 @option{-Wnonnull} is included in @option{-Wall} and @option{-Wformat}. It
4253 can be disabled with the @option{-Wno-nonnull} option.
4255 @item -Wnonnull-compare
4256 @opindex Wnonnull-compare
4257 @opindex Wno-nonnull-compare
4258 Warn when comparing an argument marked with the @code{nonnull}
4259 function attribute against null inside the function.
4261 @option{-Wnonnull-compare} is included in @option{-Wall}. It
4262 can be disabled with the @option{-Wno-nonnull-compare} option.
4264 @item -Wnull-dereference
4265 @opindex Wnull-dereference
4266 @opindex Wno-null-dereference
4267 Warn if the compiler detects paths that trigger erroneous or
4268 undefined behavior due to dereferencing a null pointer. This option
4269 is only active when @option{-fdelete-null-pointer-checks} is active,
4270 which is enabled by optimizations in most targets. The precision of
4271 the warnings depends on the optimization options used.
4273 @item -Winit-self @r{(C, C++, Objective-C and Objective-C++ only)}
4275 @opindex Wno-init-self
4276 Warn about uninitialized variables that are initialized with themselves.
4277 Note this option can only be used with the @option{-Wuninitialized} option.
4279 For example, GCC warns about @code{i} being uninitialized in the
4280 following snippet only when @option{-Winit-self} has been specified:
4291 This warning is enabled by @option{-Wall} in C++.
4293 @item -Wimplicit-int @r{(C and Objective-C only)}
4294 @opindex Wimplicit-int
4295 @opindex Wno-implicit-int
4296 Warn when a declaration does not specify a type.
4297 This warning is enabled by @option{-Wall}.
4299 @item -Wimplicit-function-declaration @r{(C and Objective-C only)}
4300 @opindex Wimplicit-function-declaration
4301 @opindex Wno-implicit-function-declaration
4302 Give a warning whenever a function is used before being declared. In
4303 C99 mode (@option{-std=c99} or @option{-std=gnu99}), this warning is
4304 enabled by default and it is made into an error by
4305 @option{-pedantic-errors}. This warning is also enabled by
4308 @item -Wimplicit @r{(C and Objective-C only)}
4310 @opindex Wno-implicit
4311 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
4312 This warning is enabled by @option{-Wall}.
4314 @item -Wimplicit-fallthrough
4315 @opindex Wimplicit-fallthrough
4316 @opindex Wno-implicit-fallthrough
4317 @option{-Wimplicit-fallthrough} is the same as @option{-Wimplicit-fallthrough=3}
4318 and @option{-Wno-implicit-fallthrough} is the same as
4319 @option{-Wimplicit-fallthrough=0}.
4321 @item -Wimplicit-fallthrough=@var{n}
4322 @opindex Wimplicit-fallthrough=
4323 Warn when a switch case falls through. For example:
4341 This warning does not warn when the last statement of a case cannot
4342 fall through, e.g. when there is a return statement or a call to function
4343 declared with the noreturn attribute. @option{-Wimplicit-fallthrough=}
4344 also takes into account control flow statements, such as ifs, and only
4345 warns when appropriate. E.g.@:
4355 @} else if (i < 1) @{
4365 Since there are occasions where a switch case fall through is desirable,
4366 GCC provides an attribute, @code{__attribute__ ((fallthrough))}, that is
4367 to be used along with a null statement to suppress this warning that
4368 would normally occur:
4376 __attribute__ ((fallthrough));
4383 C++17 provides a standard way to suppress the @option{-Wimplicit-fallthrough}
4384 warning using @code{[[fallthrough]];} instead of the GNU attribute. In C++11
4385 or C++14 users can use @code{[[gnu::fallthrough]];}, which is a GNU extension.
4386 Instead of these attributes, it is also possible to add a fallthrough comment
4387 to silence the warning. The whole body of the C or C++ style comment should
4388 match the given regular expressions listed below. The option argument @var{n}
4389 specifies what kind of comments are accepted:
4393 @item @option{-Wimplicit-fallthrough=0} disables the warning altogether.
4395 @item @option{-Wimplicit-fallthrough=1} matches @code{.*} regular
4396 expression, any comment is used as fallthrough comment.
4398 @item @option{-Wimplicit-fallthrough=2} case insensitively matches
4399 @code{.*falls?[ \t-]*thr(ough|u).*} regular expression.
4401 @item @option{-Wimplicit-fallthrough=3} case sensitively matches one of the
4402 following regular expressions:
4406 @item @code{-fallthrough}
4408 @item @code{@@fallthrough@@}
4410 @item @code{lint -fallthrough[ \t]*}
4412 @item @code{[ \t.!]*(ELSE,? |INTENTIONAL(LY)? )?@*FALL(S | |-)?THR(OUGH|U)[ \t.!]*(-[^\n\r]*)?}
4414 @item @code{[ \t.!]*(Else,? |Intentional(ly)? )?@*Fall((s | |-)[Tt]|t)hr(ough|u)[ \t.!]*(-[^\n\r]*)?}
4416 @item @code{[ \t.!]*([Ee]lse,? |[Ii]ntentional(ly)? )?@*fall(s | |-)?thr(ough|u)[ \t.!]*(-[^\n\r]*)?}
4420 @item @option{-Wimplicit-fallthrough=4} case sensitively matches one of the
4421 following regular expressions:
4425 @item @code{-fallthrough}
4427 @item @code{@@fallthrough@@}
4429 @item @code{lint -fallthrough[ \t]*}
4431 @item @code{[ \t]*FALLTHR(OUGH|U)[ \t]*}
4435 @item @option{-Wimplicit-fallthrough=5} doesn't recognize any comments as
4436 fallthrough comments, only attributes disable the warning.
4440 The comment needs to be followed after optional whitespace and other comments
4441 by @code{case} or @code{default} keywords or by a user label that precedes some
4442 @code{case} or @code{default} label.
4457 The @option{-Wimplicit-fallthrough=3} warning is enabled by @option{-Wextra}.
4459 @item -Wif-not-aligned @r{(C, C++, Objective-C and Objective-C++ only)}
4460 @opindex Wif-not-aligned
4461 @opindex Wno-if-not-aligned
4462 Control if warning triggered by the @code{warn_if_not_aligned} attribute
4463 should be issued. This is is enabled by default.
4464 Use @option{-Wno-if-not-aligned} to disable it.
4466 @item -Wignored-qualifiers @r{(C and C++ only)}
4467 @opindex Wignored-qualifiers
4468 @opindex Wno-ignored-qualifiers
4469 Warn if the return type of a function has a type qualifier
4470 such as @code{const}. For ISO C such a type qualifier has no effect,
4471 since the value returned by a function is not an lvalue.
4472 For C++, the warning is only emitted for scalar types or @code{void}.
4473 ISO C prohibits qualified @code{void} return types on function
4474 definitions, so such return types always receive a warning
4475 even without this option.
4477 This warning is also enabled by @option{-Wextra}.
4479 @item -Wignored-attributes @r{(C and C++ only)}
4480 @opindex Wignored-attributes
4481 @opindex Wno-ignored-attributes
4482 Warn when an attribute is ignored. This is different from the
4483 @option{-Wattributes} option in that it warns whenever the compiler decides
4484 to drop an attribute, not that the attribute is either unknown, used in a
4485 wrong place, etc. This warning is enabled by default.
4490 Warn if the type of @code{main} is suspicious. @code{main} should be
4491 a function with external linkage, returning int, taking either zero
4492 arguments, two, or three arguments of appropriate types. This warning
4493 is enabled by default in C++ and is enabled by either @option{-Wall}
4494 or @option{-Wpedantic}.
4496 @item -Wmisleading-indentation @r{(C and C++ only)}
4497 @opindex Wmisleading-indentation
4498 @opindex Wno-misleading-indentation
4499 Warn when the indentation of the code does not reflect the block structure.
4500 Specifically, a warning is issued for @code{if}, @code{else}, @code{while}, and
4501 @code{for} clauses with a guarded statement that does not use braces,
4502 followed by an unguarded statement with the same indentation.
4504 In the following example, the call to ``bar'' is misleadingly indented as
4505 if it were guarded by the ``if'' conditional.
4508 if (some_condition ())
4510 bar (); /* Gotcha: this is not guarded by the "if". */
4513 In the case of mixed tabs and spaces, the warning uses the
4514 @option{-ftabstop=} option to determine if the statements line up
4517 The warning is not issued for code involving multiline preprocessor logic
4518 such as the following example.
4523 #if SOME_CONDITION_THAT_DOES_NOT_HOLD
4529 The warning is not issued after a @code{#line} directive, since this
4530 typically indicates autogenerated code, and no assumptions can be made
4531 about the layout of the file that the directive references.
4533 This warning is enabled by @option{-Wall} in C and C++.
4535 @item -Wmissing-braces
4536 @opindex Wmissing-braces
4537 @opindex Wno-missing-braces
4538 Warn if an aggregate or union initializer is not fully bracketed. In
4539 the following example, the initializer for @code{a} is not fully
4540 bracketed, but that for @code{b} is fully bracketed. This warning is
4541 enabled by @option{-Wall} in C.
4544 int a[2][2] = @{ 0, 1, 2, 3 @};
4545 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
4548 This warning is enabled by @option{-Wall}.
4550 @item -Wmissing-include-dirs @r{(C, C++, Objective-C and Objective-C++ only)}
4551 @opindex Wmissing-include-dirs
4552 @opindex Wno-missing-include-dirs
4553 Warn if a user-supplied include directory does not exist.
4555 @item -Wmultistatement-macros
4556 @opindex Wmultistatement-macros
4557 @opindex Wno-multistatement-macros
4558 Warn about unsafe multiple statement macros that appear to be guarded
4559 by a clause such as @code{if}, @code{else}, @code{for}, @code{switch}, or
4560 @code{while}, in which only the first statement is actually guarded after
4561 the macro is expanded.
4566 #define DOIT x++; y++
4571 will increment @code{y} unconditionally, not just when @code{c} holds.
4572 The can usually be fixed by wrapping the macro in a do-while loop:
4574 #define DOIT do @{ x++; y++; @} while (0)
4579 This warning is enabled by @option{-Wall} in C and C++.
4582 @opindex Wparentheses
4583 @opindex Wno-parentheses
4584 Warn if parentheses are omitted in certain contexts, such
4585 as when there is an assignment in a context where a truth value
4586 is expected, or when operators are nested whose precedence people
4587 often get confused about.
4589 Also warn if a comparison like @code{x<=y<=z} appears; this is
4590 equivalent to @code{(x<=y ? 1 : 0) <= z}, which is a different
4591 interpretation from that of ordinary mathematical notation.
4593 Also warn for dangerous uses of the GNU extension to
4594 @code{?:} with omitted middle operand. When the condition
4595 in the @code{?}: operator is a boolean expression, the omitted value is
4596 always 1. Often programmers expect it to be a value computed
4597 inside the conditional expression instead.
4599 For C++ this also warns for some cases of unnecessary parentheses in
4600 declarations, which can indicate an attempt at a function call instead
4604 // Declares a local variable called mymutex.
4605 std::unique_lock<std::mutex> (mymutex);
4606 // User meant std::unique_lock<std::mutex> lock (mymutex);
4610 This warning is enabled by @option{-Wall}.
4612 @item -Wsequence-point
4613 @opindex Wsequence-point
4614 @opindex Wno-sequence-point
4615 Warn about code that may have undefined semantics because of violations
4616 of sequence point rules in the C and C++ standards.
4618 The C and C++ standards define the order in which expressions in a C/C++
4619 program are evaluated in terms of @dfn{sequence points}, which represent
4620 a partial ordering between the execution of parts of the program: those
4621 executed before the sequence point, and those executed after it. These
4622 occur after the evaluation of a full expression (one which is not part
4623 of a larger expression), after the evaluation of the first operand of a
4624 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
4625 function is called (but after the evaluation of its arguments and the
4626 expression denoting the called function), and in certain other places.
4627 Other than as expressed by the sequence point rules, the order of
4628 evaluation of subexpressions of an expression is not specified. All
4629 these rules describe only a partial order rather than a total order,
4630 since, for example, if two functions are called within one expression
4631 with no sequence point between them, the order in which the functions
4632 are called is not specified. However, the standards committee have
4633 ruled that function calls do not overlap.
4635 It is not specified when between sequence points modifications to the
4636 values of objects take effect. Programs whose behavior depends on this
4637 have undefined behavior; the C and C++ standards specify that ``Between
4638 the previous and next sequence point an object shall have its stored
4639 value modified at most once by the evaluation of an expression.
4640 Furthermore, the prior value shall be read only to determine the value
4641 to be stored.''. If a program breaks these rules, the results on any
4642 particular implementation are entirely unpredictable.
4644 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
4645 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
4646 diagnosed by this option, and it may give an occasional false positive
4647 result, but in general it has been found fairly effective at detecting
4648 this sort of problem in programs.
4650 The C++17 standard will define the order of evaluation of operands in
4651 more cases: in particular it requires that the right-hand side of an
4652 assignment be evaluated before the left-hand side, so the above
4653 examples are no longer undefined. But this warning will still warn
4654 about them, to help people avoid writing code that is undefined in C
4655 and earlier revisions of C++.
4657 The standard is worded confusingly, therefore there is some debate
4658 over the precise meaning of the sequence point rules in subtle cases.
4659 Links to discussions of the problem, including proposed formal
4660 definitions, may be found on the GCC readings page, at
4661 @uref{http://gcc.gnu.org/@/readings.html}.
4663 This warning is enabled by @option{-Wall} for C and C++.
4665 @item -Wno-return-local-addr
4666 @opindex Wno-return-local-addr
4667 @opindex Wreturn-local-addr
4668 Do not warn about returning a pointer (or in C++, a reference) to a
4669 variable that goes out of scope after the function returns.
4672 @opindex Wreturn-type
4673 @opindex Wno-return-type
4674 Warn whenever a function is defined with a return type that defaults
4675 to @code{int}. Also warn about any @code{return} statement with no
4676 return value in a function whose return type is not @code{void}
4677 (falling off the end of the function body is considered returning
4680 For C only, warn about a @code{return} statement with an expression in a
4681 function whose return type is @code{void}, unless the expression type is
4682 also @code{void}. As a GNU extension, the latter case is accepted
4683 without a warning unless @option{-Wpedantic} is used.
4685 For C++, a function without return type always produces a diagnostic
4686 message, even when @option{-Wno-return-type} is specified. The only
4687 exceptions are @code{main} and functions defined in system headers.
4689 This warning is enabled by @option{-Wall}.
4691 @item -Wshift-count-negative
4692 @opindex Wshift-count-negative
4693 @opindex Wno-shift-count-negative
4694 Warn if shift count is negative. This warning is enabled by default.
4696 @item -Wshift-count-overflow
4697 @opindex Wshift-count-overflow
4698 @opindex Wno-shift-count-overflow
4699 Warn if shift count >= width of type. This warning is enabled by default.
4701 @item -Wshift-negative-value
4702 @opindex Wshift-negative-value
4703 @opindex Wno-shift-negative-value
4704 Warn if left shifting a negative value. This warning is enabled by
4705 @option{-Wextra} in C99 and C++11 modes (and newer).
4707 @item -Wshift-overflow
4708 @itemx -Wshift-overflow=@var{n}
4709 @opindex Wshift-overflow
4710 @opindex Wno-shift-overflow
4711 Warn about left shift overflows. This warning is enabled by
4712 default in C99 and C++11 modes (and newer).
4715 @item -Wshift-overflow=1
4716 This is the warning level of @option{-Wshift-overflow} and is enabled
4717 by default in C99 and C++11 modes (and newer). This warning level does
4718 not warn about left-shifting 1 into the sign bit. (However, in C, such
4719 an overflow is still rejected in contexts where an integer constant expression
4722 @item -Wshift-overflow=2
4723 This warning level also warns about left-shifting 1 into the sign bit,
4724 unless C++14 mode is active.
4730 Warn whenever a @code{switch} statement has an index of enumerated type
4731 and lacks a @code{case} for one or more of the named codes of that
4732 enumeration. (The presence of a @code{default} label prevents this
4733 warning.) @code{case} labels outside the enumeration range also
4734 provoke warnings when this option is used (even if there is a
4735 @code{default} label).
4736 This warning is enabled by @option{-Wall}.
4738 @item -Wswitch-default
4739 @opindex Wswitch-default
4740 @opindex Wno-switch-default
4741 Warn whenever a @code{switch} statement does not have a @code{default}
4745 @opindex Wswitch-enum
4746 @opindex Wno-switch-enum
4747 Warn whenever a @code{switch} statement has an index of enumerated type
4748 and lacks a @code{case} for one or more of the named codes of that
4749 enumeration. @code{case} labels outside the enumeration range also
4750 provoke warnings when this option is used. The only difference
4751 between @option{-Wswitch} and this option is that this option gives a
4752 warning about an omitted enumeration code even if there is a
4753 @code{default} label.
4756 @opindex Wswitch-bool
4757 @opindex Wno-switch-bool
4758 Warn whenever a @code{switch} statement has an index of boolean type
4759 and the case values are outside the range of a boolean type.
4760 It is possible to suppress this warning by casting the controlling
4761 expression to a type other than @code{bool}. For example:
4764 switch ((int) (a == 4))
4770 This warning is enabled by default for C and C++ programs.
4772 @item -Wswitch-unreachable
4773 @opindex Wswitch-unreachable
4774 @opindex Wno-switch-unreachable
4775 Warn whenever a @code{switch} statement contains statements between the
4776 controlling expression and the first case label, which will never be
4777 executed. For example:
4789 @option{-Wswitch-unreachable} does not warn if the statement between the
4790 controlling expression and the first case label is just a declaration:
4803 This warning is enabled by default for C and C++ programs.
4805 @item -Wsync-nand @r{(C and C++ only)}
4807 @opindex Wno-sync-nand
4808 Warn when @code{__sync_fetch_and_nand} and @code{__sync_nand_and_fetch}
4809 built-in functions are used. These functions changed semantics in GCC 4.4.
4811 @item -Wunused-but-set-parameter
4812 @opindex Wunused-but-set-parameter
4813 @opindex Wno-unused-but-set-parameter
4814 Warn whenever a function parameter is assigned to, but otherwise unused
4815 (aside from its declaration).
4817 To suppress this warning use the @code{unused} attribute
4818 (@pxref{Variable Attributes}).
4820 This warning is also enabled by @option{-Wunused} together with
4823 @item -Wunused-but-set-variable
4824 @opindex Wunused-but-set-variable
4825 @opindex Wno-unused-but-set-variable
4826 Warn whenever a local variable is assigned to, but otherwise unused
4827 (aside from its declaration).
4828 This warning is enabled by @option{-Wall}.
4830 To suppress this warning use the @code{unused} attribute
4831 (@pxref{Variable Attributes}).
4833 This warning is also enabled by @option{-Wunused}, which is enabled
4836 @item -Wunused-function
4837 @opindex Wunused-function
4838 @opindex Wno-unused-function
4839 Warn whenever a static function is declared but not defined or a
4840 non-inline static function is unused.
4841 This warning is enabled by @option{-Wall}.
4843 @item -Wunused-label
4844 @opindex Wunused-label
4845 @opindex Wno-unused-label
4846 Warn whenever a label is declared but not used.
4847 This warning is enabled by @option{-Wall}.
4849 To suppress this warning use the @code{unused} attribute
4850 (@pxref{Variable Attributes}).
4852 @item -Wunused-local-typedefs @r{(C, Objective-C, C++ and Objective-C++ only)}
4853 @opindex Wunused-local-typedefs
4854 Warn when a typedef locally defined in a function is not used.
4855 This warning is enabled by @option{-Wall}.
4857 @item -Wunused-parameter
4858 @opindex Wunused-parameter
4859 @opindex Wno-unused-parameter
4860 Warn whenever a function parameter is unused aside from its declaration.
4862 To suppress this warning use the @code{unused} attribute
4863 (@pxref{Variable Attributes}).
4865 @item -Wno-unused-result
4866 @opindex Wunused-result
4867 @opindex Wno-unused-result
4868 Do not warn if a caller of a function marked with attribute
4869 @code{warn_unused_result} (@pxref{Function Attributes}) does not use
4870 its return value. The default is @option{-Wunused-result}.
4872 @item -Wunused-variable
4873 @opindex Wunused-variable
4874 @opindex Wno-unused-variable
4875 Warn whenever a local or static variable is unused aside from its
4876 declaration. This option implies @option{-Wunused-const-variable=1} for C,
4877 but not for C++. This warning is enabled by @option{-Wall}.
4879 To suppress this warning use the @code{unused} attribute
4880 (@pxref{Variable Attributes}).
4882 @item -Wunused-const-variable
4883 @itemx -Wunused-const-variable=@var{n}
4884 @opindex Wunused-const-variable
4885 @opindex Wno-unused-const-variable
4886 Warn whenever a constant static variable is unused aside from its declaration.
4887 @option{-Wunused-const-variable=1} is enabled by @option{-Wunused-variable}
4888 for C, but not for C++. In C this declares variable storage, but in C++ this
4889 is not an error since const variables take the place of @code{#define}s.
4891 To suppress this warning use the @code{unused} attribute
4892 (@pxref{Variable Attributes}).
4895 @item -Wunused-const-variable=1
4896 This is the warning level that is enabled by @option{-Wunused-variable} for
4897 C. It warns only about unused static const variables defined in the main
4898 compilation unit, but not about static const variables declared in any
4901 @item -Wunused-const-variable=2
4902 This warning level also warns for unused constant static variables in
4903 headers (excluding system headers). This is the warning level of
4904 @option{-Wunused-const-variable} and must be explicitly requested since
4905 in C++ this isn't an error and in C it might be harder to clean up all
4909 @item -Wunused-value
4910 @opindex Wunused-value
4911 @opindex Wno-unused-value
4912 Warn whenever a statement computes a result that is explicitly not
4913 used. To suppress this warning cast the unused expression to
4914 @code{void}. This includes an expression-statement or the left-hand
4915 side of a comma expression that contains no side effects. For example,
4916 an expression such as @code{x[i,j]} causes a warning, while
4917 @code{x[(void)i,j]} does not.
4919 This warning is enabled by @option{-Wall}.
4924 All the above @option{-Wunused} options combined.
4926 In order to get a warning about an unused function parameter, you must
4927 either specify @option{-Wextra -Wunused} (note that @option{-Wall} implies
4928 @option{-Wunused}), or separately specify @option{-Wunused-parameter}.
4930 @item -Wuninitialized
4931 @opindex Wuninitialized
4932 @opindex Wno-uninitialized
4933 Warn if an automatic variable is used without first being initialized
4934 or if a variable may be clobbered by a @code{setjmp} call. In C++,
4935 warn if a non-static reference or non-static @code{const} member
4936 appears in a class without constructors.
4938 If you want to warn about code that uses the uninitialized value of the
4939 variable in its own initializer, use the @option{-Winit-self} option.
4941 These warnings occur for individual uninitialized or clobbered
4942 elements of structure, union or array variables as well as for
4943 variables that are uninitialized or clobbered as a whole. They do
4944 not occur for variables or elements declared @code{volatile}. Because
4945 these warnings depend on optimization, the exact variables or elements
4946 for which there are warnings depends on the precise optimization
4947 options and version of GCC used.
4949 Note that there may be no warning about a variable that is used only
4950 to compute a value that itself is never used, because such
4951 computations may be deleted by data flow analysis before the warnings
4954 @item -Winvalid-memory-model
4955 @opindex Winvalid-memory-model
4956 @opindex Wno-invalid-memory-model
4957 Warn for invocations of @ref{__atomic Builtins}, @ref{__sync Builtins},
4958 and the C11 atomic generic functions with a memory consistency argument
4959 that is either invalid for the operation or outside the range of values
4960 of the @code{memory_order} enumeration. For example, since the
4961 @code{__atomic_store} and @code{__atomic_store_n} built-ins are only
4962 defined for the relaxed, release, and sequentially consistent memory
4963 orders the following code is diagnosed:
4968 __atomic_store_n (i, 0, memory_order_consume);
4972 @option{-Winvalid-memory-model} is enabled by default.
4974 @item -Wmaybe-uninitialized
4975 @opindex Wmaybe-uninitialized
4976 @opindex Wno-maybe-uninitialized
4977 For an automatic variable, if there exists a path from the function
4978 entry to a use of the variable that is initialized, but there exist
4979 some other paths for which the variable is not initialized, the compiler
4980 emits a warning if it cannot prove the uninitialized paths are not
4981 executed at run time. These warnings are made optional because GCC is
4982 not smart enough to see all the reasons why the code might be correct
4983 in spite of appearing to have an error. Here is one example of how
5004 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
5005 always initialized, but GCC doesn't know this. To suppress the
5006 warning, you need to provide a default case with assert(0) or
5009 @cindex @code{longjmp} warnings
5010 This option also warns when a non-volatile automatic variable might be
5011 changed by a call to @code{longjmp}. These warnings as well are possible
5012 only in optimizing compilation.
5014 The compiler sees only the calls to @code{setjmp}. It cannot know
5015 where @code{longjmp} will be called; in fact, a signal handler could
5016 call it at any point in the code. As a result, you may get a warning
5017 even when there is in fact no problem because @code{longjmp} cannot
5018 in fact be called at the place that would cause a problem.
5020 Some spurious warnings can be avoided if you declare all the functions
5021 you use that never return as @code{noreturn}. @xref{Function
5024 This warning is enabled by @option{-Wall} or @option{-Wextra}.
5026 @item -Wunknown-pragmas
5027 @opindex Wunknown-pragmas
5028 @opindex Wno-unknown-pragmas
5029 @cindex warning for unknown pragmas
5030 @cindex unknown pragmas, warning
5031 @cindex pragmas, warning of unknown
5032 Warn when a @code{#pragma} directive is encountered that is not understood by
5033 GCC@. If this command-line option is used, warnings are even issued
5034 for unknown pragmas in system header files. This is not the case if
5035 the warnings are only enabled by the @option{-Wall} command-line option.
5038 @opindex Wno-pragmas
5040 Do not warn about misuses of pragmas, such as incorrect parameters,
5041 invalid syntax, or conflicts between pragmas. See also
5042 @option{-Wunknown-pragmas}.
5044 @item -Wstrict-aliasing
5045 @opindex Wstrict-aliasing
5046 @opindex Wno-strict-aliasing
5047 This option is only active when @option{-fstrict-aliasing} is active.
5048 It warns about code that might break the strict aliasing rules that the
5049 compiler is using for optimization. The warning does not catch all
5050 cases, but does attempt to catch the more common pitfalls. It is
5051 included in @option{-Wall}.
5052 It is equivalent to @option{-Wstrict-aliasing=3}
5054 @item -Wstrict-aliasing=n
5055 @opindex Wstrict-aliasing=n
5056 This option is only active when @option{-fstrict-aliasing} is active.
5057 It warns about code that might break the strict aliasing rules that the
5058 compiler is using for optimization.
5059 Higher levels correspond to higher accuracy (fewer false positives).
5060 Higher levels also correspond to more effort, similar to the way @option{-O}
5062 @option{-Wstrict-aliasing} is equivalent to @option{-Wstrict-aliasing=3}.
5064 Level 1: Most aggressive, quick, least accurate.
5065 Possibly useful when higher levels
5066 do not warn but @option{-fstrict-aliasing} still breaks the code, as it has very few
5067 false negatives. However, it has many false positives.
5068 Warns for all pointer conversions between possibly incompatible types,
5069 even if never dereferenced. Runs in the front end only.
5071 Level 2: Aggressive, quick, not too precise.
5072 May still have many false positives (not as many as level 1 though),
5073 and few false negatives (but possibly more than level 1).
5074 Unlike level 1, it only warns when an address is taken. Warns about
5075 incomplete types. Runs in the front end only.
5077 Level 3 (default for @option{-Wstrict-aliasing}):
5078 Should have very few false positives and few false
5079 negatives. Slightly slower than levels 1 or 2 when optimization is enabled.
5080 Takes care of the common pun+dereference pattern in the front end:
5081 @code{*(int*)&some_float}.
5082 If optimization is enabled, it also runs in the back end, where it deals
5083 with multiple statement cases using flow-sensitive points-to information.
5084 Only warns when the converted pointer is dereferenced.
5085 Does not warn about incomplete types.
5087 @item -Wstrict-overflow
5088 @itemx -Wstrict-overflow=@var{n}
5089 @opindex Wstrict-overflow
5090 @opindex Wno-strict-overflow
5091 This option is only active when signed overflow is undefined.
5092 It warns about cases where the compiler optimizes based on the
5093 assumption that signed overflow does not occur. Note that it does not
5094 warn about all cases where the code might overflow: it only warns
5095 about cases where the compiler implements some optimization. Thus
5096 this warning depends on the optimization level.
5098 An optimization that assumes that signed overflow does not occur is
5099 perfectly safe if the values of the variables involved are such that
5100 overflow never does, in fact, occur. Therefore this warning can
5101 easily give a false positive: a warning about code that is not
5102 actually a problem. To help focus on important issues, several
5103 warning levels are defined. No warnings are issued for the use of
5104 undefined signed overflow when estimating how many iterations a loop
5105 requires, in particular when determining whether a loop will be
5109 @item -Wstrict-overflow=1
5110 Warn about cases that are both questionable and easy to avoid. For
5111 example the compiler simplifies
5112 @code{x + 1 > x} to @code{1}. This level of
5113 @option{-Wstrict-overflow} is enabled by @option{-Wall}; higher levels
5114 are not, and must be explicitly requested.
5116 @item -Wstrict-overflow=2
5117 Also warn about other cases where a comparison is simplified to a
5118 constant. For example: @code{abs (x) >= 0}. This can only be
5119 simplified when signed integer overflow is undefined, because
5120 @code{abs (INT_MIN)} overflows to @code{INT_MIN}, which is less than
5121 zero. @option{-Wstrict-overflow} (with no level) is the same as
5122 @option{-Wstrict-overflow=2}.
5124 @item -Wstrict-overflow=3
5125 Also warn about other cases where a comparison is simplified. For
5126 example: @code{x + 1 > 1} is simplified to @code{x > 0}.
5128 @item -Wstrict-overflow=4
5129 Also warn about other simplifications not covered by the above cases.
5130 For example: @code{(x * 10) / 5} is simplified to @code{x * 2}.
5132 @item -Wstrict-overflow=5
5133 Also warn about cases where the compiler reduces the magnitude of a
5134 constant involved in a comparison. For example: @code{x + 2 > y} is
5135 simplified to @code{x + 1 >= y}. This is reported only at the
5136 highest warning level because this simplification applies to many
5137 comparisons, so this warning level gives a very large number of
5141 @item -Wstringop-overflow
5142 @itemx -Wstringop-overflow=@var{type}
5143 @opindex Wstringop-overflow
5144 @opindex Wno-stringop-overflow
5145 Warn for calls to string manipulation functions such as @code{memcpy} and
5146 @code{strcpy} that are determined to overflow the destination buffer. The
5147 optional argument is one greater than the type of Object Size Checking to
5148 perform to determine the size of the destination. @xref{Object Size Checking}.
5149 The argument is meaningful only for functions that operate on character arrays
5150 but not for raw memory functions like @code{memcpy} which always make use
5151 of Object Size type-0. The option also warns for calls that specify a size
5152 in excess of the largest possible object or at most @code{SIZE_MAX / 2} bytes.
5153 The option produces the best results with optimization enabled but can detect
5154 a small subset of simple buffer overflows even without optimization in
5155 calls to the GCC built-in functions like @code{__builtin_memcpy} that
5156 correspond to the standard functions. In any case, the option warns about
5157 just a subset of buffer overflows detected by the corresponding overflow
5158 checking built-ins. For example, the option will issue a warning for
5159 the @code{strcpy} call below because it copies at least 5 characters
5160 (the string @code{"blue"} including the terminating NUL) into the buffer
5164 enum Color @{ blue, purple, yellow @};
5165 const char* f (enum Color clr)
5167 static char buf [4];
5171 case blue: str = "blue"; break;
5172 case purple: str = "purple"; break;
5173 case yellow: str = "yellow"; break;
5176 return strcpy (buf, str); // warning here
5180 Option @option{-Wstringop-overflow=2} is enabled by default.
5183 @item -Wstringop-overflow
5184 @item -Wstringop-overflow=1
5185 @opindex Wstringop-overflow
5186 @opindex Wno-stringop-overflow
5187 The @option{-Wstringop-overflow=1} option uses type-zero Object Size Checking
5188 to determine the sizes of destination objects. This is the default setting
5189 of the option. At this setting the option will not warn for writes past
5190 the end of subobjects of larger objects accessed by pointers unless the
5191 size of the largest surrounding object is known. When the destination may
5192 be one of several objects it is assumed to be the largest one of them. On
5193 Linux systems, when optimization is enabled at this setting the option warns
5194 for the same code as when the @code{_FORTIFY_SOURCE} macro is defined to
5197 @item -Wstringop-overflow=2
5198 The @option{-Wstringop-overflow=2} option uses type-one Object Size Checking
5199 to determine the sizes of destination objects. At this setting the option
5200 will warn about overflows when writing to members of the largest complete
5201 objects whose exact size is known. It will, however, not warn for excessive
5202 writes to the same members of unknown objects referenced by pointers since
5203 they may point to arrays containing unknown numbers of elements.
5205 @item -Wstringop-overflow=3
5206 The @option{-Wstringop-overflow=3} option uses type-two Object Size Checking
5207 to determine the sizes of destination objects. At this setting the option
5208 warns about overflowing the smallest object or data member. This is the
5209 most restrictive setting of the option that may result in warnings for safe
5212 @item -Wstringop-overflow=4
5213 The @option{-Wstringop-overflow=4} option uses type-three Object Size Checking
5214 to determine the sizes of destination objects. At this setting the option
5215 will warn about overflowing any data members, and when the destination is
5216 one of several objects it uses the size of the largest of them to decide
5217 whether to issue a warning. Similarly to @option{-Wstringop-overflow=3} this
5218 setting of the option may result in warnings for benign code.
5221 @item -Wstringop-truncation
5222 @opindex Wstringop-truncation
5223 @opindex Wno-stringop-truncation
5224 Warn for calls to bounded string manipulation functions such as @code{strncat},
5225 @code{strncpy}, and @code{stpncpy} that may either truncate the copied string
5226 or leave the destination unchanged.
5228 In the following example, the call to @code{strncat} specifies a bound that
5229 is less than the length of the source string. As a result, the copy of
5230 the source will be truncated and so the call is diagnosed. To avoid the
5231 warning use @code{bufsize - strlen (buf) - 1)} as the bound.
5234 void append (char *buf, size_t bufsize)
5236 strncat (buf, ".txt", 3);
5240 As another example, the following call to @code{strncpy} results in copying
5241 to @code{d} just the characters preceding the terminating NUL, without
5242 appending the NUL to the end. Assuming the result of @code{strncpy} is
5243 necessarily a NUL-terminated string is a common mistake, and so the call
5244 is diagnosed. To avoid the warning when the result is not expected to be
5245 NUL-terminated, call @code{memcpy} instead.
5248 void copy (char *d, const char *s)
5250 strncpy (d, s, strlen (s));
5254 In the following example, the call to @code{strncpy} specifies the size
5255 of the destination buffer as the bound. If the length of the source
5256 string is equal to or greater than this size the result of the copy will
5257 not be NUL-terminated. Therefore, the call is also diagnosed. To avoid
5258 the warning, specify @code{sizeof buf - 1} as the bound and set the last
5259 element of the buffer to @code{NUL}.
5262 void copy (const char *s)
5265 strncpy (buf, s, sizeof buf);
5270 @item -Wsuggest-attribute=@r{[}pure@r{|}const@r{|}noreturn@r{|}format@r{|}cold@r{|}malloc@r{]}
5271 @opindex Wsuggest-attribute=
5272 @opindex Wno-suggest-attribute=
5273 Warn for cases where adding an attribute may be beneficial. The
5274 attributes currently supported are listed below.
5277 @item -Wsuggest-attribute=pure
5278 @itemx -Wsuggest-attribute=const
5279 @itemx -Wsuggest-attribute=noreturn
5280 @itemx -Wsuggest-attribute=malloc
5281 @opindex Wsuggest-attribute=pure
5282 @opindex Wno-suggest-attribute=pure
5283 @opindex Wsuggest-attribute=const
5284 @opindex Wno-suggest-attribute=const
5285 @opindex Wsuggest-attribute=noreturn
5286 @opindex Wno-suggest-attribute=noreturn
5287 @opindex Wsuggest-attribute=malloc
5288 @opindex Wno-suggest-attribute=malloc
5290 Warn about functions that might be candidates for attributes
5291 @code{pure}, @code{const} or @code{noreturn} or @code{malloc}. The compiler
5292 only warns for functions visible in other compilation units or (in the case of
5293 @code{pure} and @code{const}) if it cannot prove that the function returns
5294 normally. A function returns normally if it doesn't contain an infinite loop or
5295 return abnormally by throwing, calling @code{abort} or trapping. This analysis
5296 requires option @option{-fipa-pure-const}, which is enabled by default at
5297 @option{-O} and higher. Higher optimization levels improve the accuracy
5300 @item -Wsuggest-attribute=format
5301 @itemx -Wmissing-format-attribute
5302 @opindex Wsuggest-attribute=format
5303 @opindex Wmissing-format-attribute
5304 @opindex Wno-suggest-attribute=format
5305 @opindex Wno-missing-format-attribute
5309 Warn about function pointers that might be candidates for @code{format}
5310 attributes. Note these are only possible candidates, not absolute ones.
5311 GCC guesses that function pointers with @code{format} attributes that
5312 are used in assignment, initialization, parameter passing or return
5313 statements should have a corresponding @code{format} attribute in the
5314 resulting type. I.e.@: the left-hand side of the assignment or
5315 initialization, the type of the parameter variable, or the return type
5316 of the containing function respectively should also have a @code{format}
5317 attribute to avoid the warning.
5319 GCC also warns about function definitions that might be
5320 candidates for @code{format} attributes. Again, these are only
5321 possible candidates. GCC guesses that @code{format} attributes
5322 might be appropriate for any function that calls a function like
5323 @code{vprintf} or @code{vscanf}, but this might not always be the
5324 case, and some functions for which @code{format} attributes are
5325 appropriate may not be detected.
5327 @item -Wsuggest-attribute=cold
5328 @opindex Wsuggest-attribute=cold
5329 @opindex Wno-suggest-attribute=cold
5331 Warn about functions that might be candidates for @code{cold} attribute. This
5332 is based on static detection and generally will only warn about functions which
5333 always leads to a call to another @code{cold} function such as wrappers of
5334 C++ @code{throw} or fatal error reporting functions leading to @code{abort}.
5337 @item -Wsuggest-final-types
5338 @opindex Wno-suggest-final-types
5339 @opindex Wsuggest-final-types
5340 Warn about types with virtual methods where code quality would be improved
5341 if the type were declared with the C++11 @code{final} specifier,
5343 declared in an anonymous namespace. This allows GCC to more aggressively
5344 devirtualize the polymorphic calls. This warning is more effective with link
5345 time optimization, where the information about the class hierarchy graph is
5348 @item -Wsuggest-final-methods
5349 @opindex Wno-suggest-final-methods
5350 @opindex Wsuggest-final-methods
5351 Warn about virtual methods where code quality would be improved if the method
5352 were declared with the C++11 @code{final} specifier,
5353 or, if possible, its type were
5354 declared in an anonymous namespace or with the @code{final} specifier.
5356 more effective with link-time optimization, where the information about the
5357 class hierarchy graph is more complete. It is recommended to first consider
5358 suggestions of @option{-Wsuggest-final-types} and then rebuild with new
5361 @item -Wsuggest-override
5362 Warn about overriding virtual functions that are not marked with the override
5366 @opindex Wno-alloc-zero
5367 @opindex Walloc-zero
5368 Warn about calls to allocation functions decorated with attribute
5369 @code{alloc_size} that specify zero bytes, including those to the built-in
5370 forms of the functions @code{aligned_alloc}, @code{alloca}, @code{calloc},
5371 @code{malloc}, and @code{realloc}. Because the behavior of these functions
5372 when called with a zero size differs among implementations (and in the case
5373 of @code{realloc} has been deprecated) relying on it may result in subtle
5374 portability bugs and should be avoided.
5376 @item -Walloc-size-larger-than=@var{n}
5377 Warn about calls to functions decorated with attribute @code{alloc_size}
5378 that attempt to allocate objects larger than the specified number of bytes,
5379 or where the result of the size computation in an integer type with infinite
5380 precision would exceed @code{SIZE_MAX / 2}. The option argument @var{n}
5381 may end in one of the standard suffixes designating a multiple of bytes
5382 such as @code{kB} and @code{KiB} for kilobyte and kibibyte, respectively,
5383 @code{MB} and @code{MiB} for megabyte and mebibyte, and so on.
5384 @xref{Function Attributes}.
5389 This option warns on all uses of @code{alloca} in the source.
5391 @item -Walloca-larger-than=@var{n}
5392 This option warns on calls to @code{alloca} that are not bounded by a
5393 controlling predicate limiting its argument of integer type to at most
5394 @var{n} bytes, or calls to @code{alloca} where the bound is unknown.
5395 Arguments of non-integer types are considered unbounded even if they
5396 appear to be constrained to the expected range.
5398 For example, a bounded case of @code{alloca} could be:
5401 void func (size_t n)
5412 In the above example, passing @code{-Walloca-larger-than=1000} would not
5413 issue a warning because the call to @code{alloca} is known to be at most
5414 1000 bytes. However, if @code{-Walloca-larger-than=500} were passed,
5415 the compiler would emit a warning.
5417 Unbounded uses, on the other hand, are uses of @code{alloca} with no
5418 controlling predicate constraining its integer argument. For example:
5423 void *p = alloca (n);
5428 If @code{-Walloca-larger-than=500} were passed, the above would trigger
5429 a warning, but this time because of the lack of bounds checking.
5431 Note, that even seemingly correct code involving signed integers could
5435 void func (signed int n)
5445 In the above example, @var{n} could be negative, causing a larger than
5446 expected argument to be implicitly cast into the @code{alloca} call.
5448 This option also warns when @code{alloca} is used in a loop.
5450 This warning is not enabled by @option{-Wall}, and is only active when
5451 @option{-ftree-vrp} is active (default for @option{-O2} and above).
5453 See also @option{-Wvla-larger-than=@var{n}}.
5455 @item -Warray-bounds
5456 @itemx -Warray-bounds=@var{n}
5457 @opindex Wno-array-bounds
5458 @opindex Warray-bounds
5459 This option is only active when @option{-ftree-vrp} is active
5460 (default for @option{-O2} and above). It warns about subscripts to arrays
5461 that are always out of bounds. This warning is enabled by @option{-Wall}.
5464 @item -Warray-bounds=1
5465 This is the warning level of @option{-Warray-bounds} and is enabled
5466 by @option{-Wall}; higher levels are not, and must be explicitly requested.
5468 @item -Warray-bounds=2
5469 This warning level also warns about out of bounds access for
5470 arrays at the end of a struct and for arrays accessed through
5471 pointers. This warning level may give a larger number of
5472 false positives and is deactivated by default.
5475 @item -Wattribute-alias
5476 Warn about declarations using the @code{alias} and similar attributes whose
5477 target is incompatible with the type of the alias. @xref{Function Attributes,
5478 ,Declaring Attributes of Functions}.
5480 @item -Wbool-compare
5481 @opindex Wno-bool-compare
5482 @opindex Wbool-compare
5483 Warn about boolean expression compared with an integer value different from
5484 @code{true}/@code{false}. For instance, the following comparison is
5489 if ((n > 1) == 2) @{ @dots{} @}
5491 This warning is enabled by @option{-Wall}.
5493 @item -Wbool-operation
5494 @opindex Wno-bool-operation
5495 @opindex Wbool-operation
5496 Warn about suspicious operations on expressions of a boolean type. For
5497 instance, bitwise negation of a boolean is very likely a bug in the program.
5498 For C, this warning also warns about incrementing or decrementing a boolean,
5499 which rarely makes sense. (In C++, decrementing a boolean is always invalid.
5500 Incrementing a boolean is invalid in C++17, and deprecated otherwise.)
5502 This warning is enabled by @option{-Wall}.
5504 @item -Wduplicated-branches
5505 @opindex Wno-duplicated-branches
5506 @opindex Wduplicated-branches
5507 Warn when an if-else has identical branches. This warning detects cases like
5514 It doesn't warn when both branches contain just a null statement. This warning
5515 also warn for conditional operators:
5517 int i = x ? *p : *p;
5520 @item -Wduplicated-cond
5521 @opindex Wno-duplicated-cond
5522 @opindex Wduplicated-cond
5523 Warn about duplicated conditions in an if-else-if chain. For instance,
5524 warn for the following code:
5526 if (p->q != NULL) @{ @dots{} @}
5527 else if (p->q != NULL) @{ @dots{} @}
5530 @item -Wframe-address
5531 @opindex Wno-frame-address
5532 @opindex Wframe-address
5533 Warn when the @samp{__builtin_frame_address} or @samp{__builtin_return_address}
5534 is called with an argument greater than 0. Such calls may return indeterminate
5535 values or crash the program. The warning is included in @option{-Wall}.
5537 @item -Wno-discarded-qualifiers @r{(C and Objective-C only)}
5538 @opindex Wno-discarded-qualifiers
5539 @opindex Wdiscarded-qualifiers
5540 Do not warn if type qualifiers on pointers are being discarded.
5541 Typically, the compiler warns if a @code{const char *} variable is
5542 passed to a function that takes a @code{char *} parameter. This option
5543 can be used to suppress such a warning.
5545 @item -Wno-discarded-array-qualifiers @r{(C and Objective-C only)}
5546 @opindex Wno-discarded-array-qualifiers
5547 @opindex Wdiscarded-array-qualifiers
5548 Do not warn if type qualifiers on arrays which are pointer targets
5549 are being discarded. Typically, the compiler warns if a
5550 @code{const int (*)[]} variable is passed to a function that
5551 takes a @code{int (*)[]} parameter. This option can be used to
5552 suppress such a warning.
5554 @item -Wno-incompatible-pointer-types @r{(C and Objective-C only)}
5555 @opindex Wno-incompatible-pointer-types
5556 @opindex Wincompatible-pointer-types
5557 Do not warn when there is a conversion between pointers that have incompatible
5558 types. This warning is for cases not covered by @option{-Wno-pointer-sign},
5559 which warns for pointer argument passing or assignment with different
5562 @item -Wno-int-conversion @r{(C and Objective-C only)}
5563 @opindex Wno-int-conversion
5564 @opindex Wint-conversion
5565 Do not warn about incompatible integer to pointer and pointer to integer
5566 conversions. This warning is about implicit conversions; for explicit
5567 conversions the warnings @option{-Wno-int-to-pointer-cast} and
5568 @option{-Wno-pointer-to-int-cast} may be used.
5570 @item -Wno-div-by-zero
5571 @opindex Wno-div-by-zero
5572 @opindex Wdiv-by-zero
5573 Do not warn about compile-time integer division by zero. Floating-point
5574 division by zero is not warned about, as it can be a legitimate way of
5575 obtaining infinities and NaNs.
5577 @item -Wsystem-headers
5578 @opindex Wsystem-headers
5579 @opindex Wno-system-headers
5580 @cindex warnings from system headers
5581 @cindex system headers, warnings from
5582 Print warning messages for constructs found in system header files.
5583 Warnings from system headers are normally suppressed, on the assumption
5584 that they usually do not indicate real problems and would only make the
5585 compiler output harder to read. Using this command-line option tells
5586 GCC to emit warnings from system headers as if they occurred in user
5587 code. However, note that using @option{-Wall} in conjunction with this
5588 option does @emph{not} warn about unknown pragmas in system
5589 headers---for that, @option{-Wunknown-pragmas} must also be used.
5591 @item -Wtautological-compare
5592 @opindex Wtautological-compare
5593 @opindex Wno-tautological-compare
5594 Warn if a self-comparison always evaluates to true or false. This
5595 warning detects various mistakes such as:
5599 if (i > i) @{ @dots{} @}
5602 This warning also warns about bitwise comparisons that always evaluate
5603 to true or false, for instance:
5605 if ((a & 16) == 10) @{ @dots{} @}
5607 will always be false.
5609 This warning is enabled by @option{-Wall}.
5612 @opindex Wtrampolines
5613 @opindex Wno-trampolines
5614 Warn about trampolines generated for pointers to nested functions.
5615 A trampoline is a small piece of data or code that is created at run
5616 time on the stack when the address of a nested function is taken, and is
5617 used to call the nested function indirectly. For some targets, it is
5618 made up of data only and thus requires no special treatment. But, for
5619 most targets, it is made up of code and thus requires the stack to be
5620 made executable in order for the program to work properly.
5623 @opindex Wfloat-equal
5624 @opindex Wno-float-equal
5625 Warn if floating-point values are used in equality comparisons.
5627 The idea behind this is that sometimes it is convenient (for the
5628 programmer) to consider floating-point values as approximations to
5629 infinitely precise real numbers. If you are doing this, then you need
5630 to compute (by analyzing the code, or in some other way) the maximum or
5631 likely maximum error that the computation introduces, and allow for it
5632 when performing comparisons (and when producing output, but that's a
5633 different problem). In particular, instead of testing for equality, you
5634 should check to see whether the two values have ranges that overlap; and
5635 this is done with the relational operators, so equality comparisons are
5638 @item -Wtraditional @r{(C and Objective-C only)}
5639 @opindex Wtraditional
5640 @opindex Wno-traditional
5641 Warn about certain constructs that behave differently in traditional and
5642 ISO C@. Also warn about ISO C constructs that have no traditional C
5643 equivalent, and/or problematic constructs that should be avoided.
5647 Macro parameters that appear within string literals in the macro body.
5648 In traditional C macro replacement takes place within string literals,
5649 but in ISO C it does not.
5652 In traditional C, some preprocessor directives did not exist.
5653 Traditional preprocessors only considered a line to be a directive
5654 if the @samp{#} appeared in column 1 on the line. Therefore
5655 @option{-Wtraditional} warns about directives that traditional C
5656 understands but ignores because the @samp{#} does not appear as the
5657 first character on the line. It also suggests you hide directives like
5658 @code{#pragma} not understood by traditional C by indenting them. Some
5659 traditional implementations do not recognize @code{#elif}, so this option
5660 suggests avoiding it altogether.
5663 A function-like macro that appears without arguments.
5666 The unary plus operator.
5669 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating-point
5670 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
5671 constants.) Note, these suffixes appear in macros defined in the system
5672 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
5673 Use of these macros in user code might normally lead to spurious
5674 warnings, however GCC's integrated preprocessor has enough context to
5675 avoid warning in these cases.
5678 A function declared external in one block and then used after the end of
5682 A @code{switch} statement has an operand of type @code{long}.
5685 A non-@code{static} function declaration follows a @code{static} one.
5686 This construct is not accepted by some traditional C compilers.
5689 The ISO type of an integer constant has a different width or
5690 signedness from its traditional type. This warning is only issued if
5691 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
5692 typically represent bit patterns, are not warned about.
5695 Usage of ISO string concatenation is detected.
5698 Initialization of automatic aggregates.
5701 Identifier conflicts with labels. Traditional C lacks a separate
5702 namespace for labels.
5705 Initialization of unions. If the initializer is zero, the warning is
5706 omitted. This is done under the assumption that the zero initializer in
5707 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
5708 initializer warnings and relies on default initialization to zero in the
5712 Conversions by prototypes between fixed/floating-point values and vice
5713 versa. The absence of these prototypes when compiling with traditional
5714 C causes serious problems. This is a subset of the possible
5715 conversion warnings; for the full set use @option{-Wtraditional-conversion}.
5718 Use of ISO C style function definitions. This warning intentionally is
5719 @emph{not} issued for prototype declarations or variadic functions
5720 because these ISO C features appear in your code when using
5721 libiberty's traditional C compatibility macros, @code{PARAMS} and
5722 @code{VPARAMS}. This warning is also bypassed for nested functions
5723 because that feature is already a GCC extension and thus not relevant to
5724 traditional C compatibility.
5727 @item -Wtraditional-conversion @r{(C and Objective-C only)}
5728 @opindex Wtraditional-conversion
5729 @opindex Wno-traditional-conversion
5730 Warn if a prototype causes a type conversion that is different from what
5731 would happen to the same argument in the absence of a prototype. This
5732 includes conversions of fixed point to floating and vice versa, and
5733 conversions changing the width or signedness of a fixed-point argument
5734 except when the same as the default promotion.
5736 @item -Wdeclaration-after-statement @r{(C and Objective-C only)}
5737 @opindex Wdeclaration-after-statement
5738 @opindex Wno-declaration-after-statement
5739 Warn when a declaration is found after a statement in a block. This
5740 construct, known from C++, was introduced with ISO C99 and is by default
5741 allowed in GCC@. It is not supported by ISO C90. @xref{Mixed Declarations}.
5746 Warn whenever a local variable or type declaration shadows another
5747 variable, parameter, type, class member (in C++), or instance variable
5748 (in Objective-C) or whenever a built-in function is shadowed. Note
5749 that in C++, the compiler warns if a local variable shadows an
5750 explicit typedef, but not if it shadows a struct/class/enum.
5751 Same as @option{-Wshadow=global}.
5753 @item -Wno-shadow-ivar @r{(Objective-C only)}
5754 @opindex Wno-shadow-ivar
5755 @opindex Wshadow-ivar
5756 Do not warn whenever a local variable shadows an instance variable in an
5759 @item -Wshadow=global
5760 @opindex Wshadow=local
5761 The default for @option{-Wshadow}. Warns for any (global) shadowing.
5763 @item -Wshadow=local
5764 @opindex Wshadow=local
5765 Warn when a local variable shadows another local variable or parameter.
5766 This warning is enabled by @option{-Wshadow=global}.
5768 @item -Wshadow=compatible-local
5769 @opindex Wshadow=compatible-local
5770 Warn when a local variable shadows another local variable or parameter
5771 whose type is compatible with that of the shadowing variable. In C++,
5772 type compatibility here means the type of the shadowing variable can be
5773 converted to that of the shadowed variable. The creation of this flag
5774 (in addition to @option{-Wshadow=local}) is based on the idea that when
5775 a local variable shadows another one of incompatible type, it is most
5776 likely intentional, not a bug or typo, as shown in the following example:
5780 for (SomeIterator i = SomeObj.begin(); i != SomeObj.end(); ++i)
5782 for (int i = 0; i < N; ++i)
5791 Since the two variable @code{i} in the example above have incompatible types,
5792 enabling only @option{-Wshadow=compatible-local} will not emit a warning.
5793 Because their types are incompatible, if a programmer accidentally uses one
5794 in place of the other, type checking will catch that and emit an error or
5795 warning. So not warning (about shadowing) in this case will not lead to
5796 undetected bugs. Use of this flag instead of @option{-Wshadow=local} can
5797 possibly reduce the number of warnings triggered by intentional shadowing.
5799 This warning is enabled by @option{-Wshadow=local}.
5801 @item -Wlarger-than=@var{len}
5802 @opindex Wlarger-than=@var{len}
5803 @opindex Wlarger-than-@var{len}
5804 Warn whenever an object of larger than @var{len} bytes is defined.
5806 @item -Wframe-larger-than=@var{len}
5807 @opindex Wframe-larger-than
5808 Warn if the size of a function frame is larger than @var{len} bytes.
5809 The computation done to determine the stack frame size is approximate
5810 and not conservative.
5811 The actual requirements may be somewhat greater than @var{len}
5812 even if you do not get a warning. In addition, any space allocated
5813 via @code{alloca}, variable-length arrays, or related constructs
5814 is not included by the compiler when determining
5815 whether or not to issue a warning.
5817 @item -Wno-free-nonheap-object
5818 @opindex Wno-free-nonheap-object
5819 @opindex Wfree-nonheap-object
5820 Do not warn when attempting to free an object that was not allocated
5823 @item -Wstack-usage=@var{len}
5824 @opindex Wstack-usage
5825 Warn if the stack usage of a function might be larger than @var{len} bytes.
5826 The computation done to determine the stack usage is conservative.
5827 Any space allocated via @code{alloca}, variable-length arrays, or related
5828 constructs is included by the compiler when determining whether or not to
5831 The message is in keeping with the output of @option{-fstack-usage}.
5835 If the stack usage is fully static but exceeds the specified amount, it's:
5838 warning: stack usage is 1120 bytes
5841 If the stack usage is (partly) dynamic but bounded, it's:
5844 warning: stack usage might be 1648 bytes
5847 If the stack usage is (partly) dynamic and not bounded, it's:
5850 warning: stack usage might be unbounded
5854 @item -Wunsafe-loop-optimizations
5855 @opindex Wunsafe-loop-optimizations
5856 @opindex Wno-unsafe-loop-optimizations
5857 Warn if the loop cannot be optimized because the compiler cannot
5858 assume anything on the bounds of the loop indices. With
5859 @option{-funsafe-loop-optimizations} warn if the compiler makes
5862 @item -Wno-pedantic-ms-format @r{(MinGW targets only)}
5863 @opindex Wno-pedantic-ms-format
5864 @opindex Wpedantic-ms-format
5865 When used in combination with @option{-Wformat}
5866 and @option{-pedantic} without GNU extensions, this option
5867 disables the warnings about non-ISO @code{printf} / @code{scanf} format
5868 width specifiers @code{I32}, @code{I64}, and @code{I} used on Windows targets,
5869 which depend on the MS runtime.
5872 @opindex Waligned-new
5873 @opindex Wno-aligned-new
5874 Warn about a new-expression of a type that requires greater alignment
5875 than the @code{alignof(std::max_align_t)} but uses an allocation
5876 function without an explicit alignment parameter. This option is
5877 enabled by @option{-Wall}.
5879 Normally this only warns about global allocation functions, but
5880 @option{-Waligned-new=all} also warns about class member allocation
5883 @item -Wplacement-new
5884 @itemx -Wplacement-new=@var{n}
5885 @opindex Wplacement-new
5886 @opindex Wno-placement-new
5887 Warn about placement new expressions with undefined behavior, such as
5888 constructing an object in a buffer that is smaller than the type of
5889 the object. For example, the placement new expression below is diagnosed
5890 because it attempts to construct an array of 64 integers in a buffer only
5896 This warning is enabled by default.
5899 @item -Wplacement-new=1
5900 This is the default warning level of @option{-Wplacement-new}. At this
5901 level the warning is not issued for some strictly undefined constructs that
5902 GCC allows as extensions for compatibility with legacy code. For example,
5903 the following @code{new} expression is not diagnosed at this level even
5904 though it has undefined behavior according to the C++ standard because
5905 it writes past the end of the one-element array.
5907 struct S @{ int n, a[1]; @};
5908 S *s = (S *)malloc (sizeof *s + 31 * sizeof s->a[0]);
5909 new (s->a)int [32]();
5912 @item -Wplacement-new=2
5913 At this level, in addition to diagnosing all the same constructs as at level
5914 1, a diagnostic is also issued for placement new expressions that construct
5915 an object in the last member of structure whose type is an array of a single
5916 element and whose size is less than the size of the object being constructed.
5917 While the previous example would be diagnosed, the following construct makes
5918 use of the flexible member array extension to avoid the warning at level 2.
5920 struct S @{ int n, a[]; @};
5921 S *s = (S *)malloc (sizeof *s + 32 * sizeof s->a[0]);
5922 new (s->a)int [32]();
5927 @item -Wpointer-arith
5928 @opindex Wpointer-arith
5929 @opindex Wno-pointer-arith
5930 Warn about anything that depends on the ``size of'' a function type or
5931 of @code{void}. GNU C assigns these types a size of 1, for
5932 convenience in calculations with @code{void *} pointers and pointers
5933 to functions. In C++, warn also when an arithmetic operation involves
5934 @code{NULL}. This warning is also enabled by @option{-Wpedantic}.
5936 @item -Wpointer-compare
5937 @opindex Wpointer-compare
5938 @opindex Wno-pointer-compare
5939 Warn if a pointer is compared with a zero character constant. This usually
5940 means that the pointer was meant to be dereferenced. For example:
5943 const char *p = foo ();
5948 Note that the code above is invalid in C++11.
5950 This warning is enabled by default.
5953 @opindex Wtype-limits
5954 @opindex Wno-type-limits
5955 Warn if a comparison is always true or always false due to the limited
5956 range of the data type, but do not warn for constant expressions. For
5957 example, warn if an unsigned variable is compared against zero with
5958 @code{<} or @code{>=}. This warning is also enabled by
5961 @include cppwarnopts.texi
5963 @item -Wbad-function-cast @r{(C and Objective-C only)}
5964 @opindex Wbad-function-cast
5965 @opindex Wno-bad-function-cast
5966 Warn when a function call is cast to a non-matching type.
5967 For example, warn if a call to a function returning an integer type
5968 is cast to a pointer type.
5970 @item -Wc90-c99-compat @r{(C and Objective-C only)}
5971 @opindex Wc90-c99-compat
5972 @opindex Wno-c90-c99-compat
5973 Warn about features not present in ISO C90, but present in ISO C99.
5974 For instance, warn about use of variable length arrays, @code{long long}
5975 type, @code{bool} type, compound literals, designated initializers, and so
5976 on. This option is independent of the standards mode. Warnings are disabled
5977 in the expression that follows @code{__extension__}.
5979 @item -Wc99-c11-compat @r{(C and Objective-C only)}
5980 @opindex Wc99-c11-compat
5981 @opindex Wno-c99-c11-compat
5982 Warn about features not present in ISO C99, but present in ISO C11.
5983 For instance, warn about use of anonymous structures and unions,
5984 @code{_Atomic} type qualifier, @code{_Thread_local} storage-class specifier,
5985 @code{_Alignas} specifier, @code{Alignof} operator, @code{_Generic} keyword,
5986 and so on. This option is independent of the standards mode. Warnings are
5987 disabled in the expression that follows @code{__extension__}.
5989 @item -Wc++-compat @r{(C and Objective-C only)}
5990 @opindex Wc++-compat
5991 Warn about ISO C constructs that are outside of the common subset of
5992 ISO C and ISO C++, e.g.@: request for implicit conversion from
5993 @code{void *} to a pointer to non-@code{void} type.
5995 @item -Wc++11-compat @r{(C++ and Objective-C++ only)}
5996 @opindex Wc++11-compat
5997 Warn about C++ constructs whose meaning differs between ISO C++ 1998
5998 and ISO C++ 2011, e.g., identifiers in ISO C++ 1998 that are keywords
5999 in ISO C++ 2011. This warning turns on @option{-Wnarrowing} and is
6000 enabled by @option{-Wall}.
6002 @item -Wc++14-compat @r{(C++ and Objective-C++ only)}
6003 @opindex Wc++14-compat
6004 Warn about C++ constructs whose meaning differs between ISO C++ 2011
6005 and ISO C++ 2014. This warning is enabled by @option{-Wall}.
6007 @item -Wc++17-compat @r{(C++ and Objective-C++ only)}
6008 @opindex Wc++17-compat
6009 Warn about C++ constructs whose meaning differs between ISO C++ 2014
6010 and ISO C++ 2017. This warning is enabled by @option{-Wall}.
6014 @opindex Wno-cast-qual
6015 Warn whenever a pointer is cast so as to remove a type qualifier from
6016 the target type. For example, warn if a @code{const char *} is cast
6017 to an ordinary @code{char *}.
6019 Also warn when making a cast that introduces a type qualifier in an
6020 unsafe way. For example, casting @code{char **} to @code{const char **}
6021 is unsafe, as in this example:
6024 /* p is char ** value. */
6025 const char **q = (const char **) p;
6026 /* Assignment of readonly string to const char * is OK. */
6028 /* Now char** pointer points to read-only memory. */
6033 @opindex Wcast-align
6034 @opindex Wno-cast-align
6035 Warn whenever a pointer is cast such that the required alignment of the
6036 target is increased. For example, warn if a @code{char *} is cast to
6037 an @code{int *} on machines where integers can only be accessed at
6038 two- or four-byte boundaries.
6040 @item -Wcast-align=strict
6041 @opindex Wcast-align=strict
6042 Warn whenever a pointer is cast such that the required alignment of the
6043 target is increased. For example, warn if a @code{char *} is cast to
6044 an @code{int *} regardless of the target machine.
6046 @item -Wwrite-strings
6047 @opindex Wwrite-strings
6048 @opindex Wno-write-strings
6049 When compiling C, give string constants the type @code{const
6050 char[@var{length}]} so that copying the address of one into a
6051 non-@code{const} @code{char *} pointer produces a warning. These
6052 warnings help you find at compile time code that can try to write
6053 into a string constant, but only if you have been very careful about
6054 using @code{const} in declarations and prototypes. Otherwise, it is
6055 just a nuisance. This is why we did not make @option{-Wall} request
6058 When compiling C++, warn about the deprecated conversion from string
6059 literals to @code{char *}. This warning is enabled by default for C++
6063 @itemx -Wcatch-value=@var{n} @r{(C++ and Objective-C++ only)}
6064 @opindex Wcatch-value
6065 @opindex Wno-catch-value
6066 Warn about catch handlers that do not catch via reference.
6067 With @option{-Wcatch-value=1} (or @option{-Wcatch-value} for short)
6068 warn about polymorphic class types that are caught by value.
6069 With @option{-Wcatch-value=2} warn about all class types that are caught
6070 by value. With @option{-Wcatch-value=3} warn about all types that are
6071 not caught by reference. @option{-Wcatch-value} is enabled by @option{-Wall}.
6075 @opindex Wno-clobbered
6076 Warn for variables that might be changed by @code{longjmp} or
6077 @code{vfork}. This warning is also enabled by @option{-Wextra}.
6079 @item -Wconditionally-supported @r{(C++ and Objective-C++ only)}
6080 @opindex Wconditionally-supported
6081 @opindex Wno-conditionally-supported
6082 Warn for conditionally-supported (C++11 [intro.defs]) constructs.
6085 @opindex Wconversion
6086 @opindex Wno-conversion
6087 Warn for implicit conversions that may alter a value. This includes
6088 conversions between real and integer, like @code{abs (x)} when
6089 @code{x} is @code{double}; conversions between signed and unsigned,
6090 like @code{unsigned ui = -1}; and conversions to smaller types, like
6091 @code{sqrtf (M_PI)}. Do not warn for explicit casts like @code{abs
6092 ((int) x)} and @code{ui = (unsigned) -1}, or if the value is not
6093 changed by the conversion like in @code{abs (2.0)}. Warnings about
6094 conversions between signed and unsigned integers can be disabled by
6095 using @option{-Wno-sign-conversion}.
6097 For C++, also warn for confusing overload resolution for user-defined
6098 conversions; and conversions that never use a type conversion
6099 operator: conversions to @code{void}, the same type, a base class or a
6100 reference to them. Warnings about conversions between signed and
6101 unsigned integers are disabled by default in C++ unless
6102 @option{-Wsign-conversion} is explicitly enabled.
6104 @item -Wno-conversion-null @r{(C++ and Objective-C++ only)}
6105 @opindex Wconversion-null
6106 @opindex Wno-conversion-null
6107 Do not warn for conversions between @code{NULL} and non-pointer
6108 types. @option{-Wconversion-null} is enabled by default.
6110 @item -Wzero-as-null-pointer-constant @r{(C++ and Objective-C++ only)}
6111 @opindex Wzero-as-null-pointer-constant
6112 @opindex Wno-zero-as-null-pointer-constant
6113 Warn when a literal @samp{0} is used as null pointer constant. This can
6114 be useful to facilitate the conversion to @code{nullptr} in C++11.
6116 @item -Wsubobject-linkage @r{(C++ and Objective-C++ only)}
6117 @opindex Wsubobject-linkage
6118 @opindex Wno-subobject-linkage
6119 Warn if a class type has a base or a field whose type uses the anonymous
6120 namespace or depends on a type with no linkage. If a type A depends on
6121 a type B with no or internal linkage, defining it in multiple
6122 translation units would be an ODR violation because the meaning of B
6123 is different in each translation unit. If A only appears in a single
6124 translation unit, the best way to silence the warning is to give it
6125 internal linkage by putting it in an anonymous namespace as well. The
6126 compiler doesn't give this warning for types defined in the main .C
6127 file, as those are unlikely to have multiple definitions.
6128 @option{-Wsubobject-linkage} is enabled by default.
6130 @item -Wdangling-else
6131 @opindex Wdangling-else
6132 @opindex Wno-dangling-else
6133 Warn about constructions where there may be confusion to which
6134 @code{if} statement an @code{else} branch belongs. Here is an example of
6149 In C/C++, every @code{else} branch belongs to the innermost possible
6150 @code{if} statement, which in this example is @code{if (b)}. This is
6151 often not what the programmer expected, as illustrated in the above
6152 example by indentation the programmer chose. When there is the
6153 potential for this confusion, GCC issues a warning when this flag
6154 is specified. To eliminate the warning, add explicit braces around
6155 the innermost @code{if} statement so there is no way the @code{else}
6156 can belong to the enclosing @code{if}. The resulting code
6173 This warning is enabled by @option{-Wparentheses}.
6177 @opindex Wno-date-time
6178 Warn when macros @code{__TIME__}, @code{__DATE__} or @code{__TIMESTAMP__}
6179 are encountered as they might prevent bit-wise-identical reproducible
6182 @item -Wdelete-incomplete @r{(C++ and Objective-C++ only)}
6183 @opindex Wdelete-incomplete
6184 @opindex Wno-delete-incomplete
6185 Warn when deleting a pointer to incomplete type, which may cause
6186 undefined behavior at runtime. This warning is enabled by default.
6188 @item -Wuseless-cast @r{(C++ and Objective-C++ only)}
6189 @opindex Wuseless-cast
6190 @opindex Wno-useless-cast
6191 Warn when an expression is casted to its own type.
6194 @opindex Wempty-body
6195 @opindex Wno-empty-body
6196 Warn if an empty body occurs in an @code{if}, @code{else} or @code{do
6197 while} statement. This warning is also enabled by @option{-Wextra}.
6199 @item -Wenum-compare
6200 @opindex Wenum-compare
6201 @opindex Wno-enum-compare
6202 Warn about a comparison between values of different enumerated types.
6203 In C++ enumerated type mismatches in conditional expressions are also
6204 diagnosed and the warning is enabled by default. In C this warning is
6205 enabled by @option{-Wall}.
6207 @item -Wextra-semi @r{(C++, Objective-C++ only)}
6208 @opindex Wextra-semi
6209 @opindex Wno-extra-semi
6210 Warn about redundant semicolon after in-class function definition.
6212 @item -Wjump-misses-init @r{(C, Objective-C only)}
6213 @opindex Wjump-misses-init
6214 @opindex Wno-jump-misses-init
6215 Warn if a @code{goto} statement or a @code{switch} statement jumps
6216 forward across the initialization of a variable, or jumps backward to a
6217 label after the variable has been initialized. This only warns about
6218 variables that are initialized when they are declared. This warning is
6219 only supported for C and Objective-C; in C++ this sort of branch is an
6222 @option{-Wjump-misses-init} is included in @option{-Wc++-compat}. It
6223 can be disabled with the @option{-Wno-jump-misses-init} option.
6225 @item -Wsign-compare
6226 @opindex Wsign-compare
6227 @opindex Wno-sign-compare
6228 @cindex warning for comparison of signed and unsigned values
6229 @cindex comparison of signed and unsigned values, warning
6230 @cindex signed and unsigned values, comparison warning
6231 Warn when a comparison between signed and unsigned values could produce
6232 an incorrect result when the signed value is converted to unsigned.
6233 In C++, this warning is also enabled by @option{-Wall}. In C, it is
6234 also enabled by @option{-Wextra}.
6236 @item -Wsign-conversion
6237 @opindex Wsign-conversion
6238 @opindex Wno-sign-conversion
6239 Warn for implicit conversions that may change the sign of an integer
6240 value, like assigning a signed integer expression to an unsigned
6241 integer variable. An explicit cast silences the warning. In C, this
6242 option is enabled also by @option{-Wconversion}.
6244 @item -Wfloat-conversion
6245 @opindex Wfloat-conversion
6246 @opindex Wno-float-conversion
6247 Warn for implicit conversions that reduce the precision of a real value.
6248 This includes conversions from real to integer, and from higher precision
6249 real to lower precision real values. This option is also enabled by
6250 @option{-Wconversion}.
6252 @item -Wno-scalar-storage-order
6253 @opindex -Wno-scalar-storage-order
6254 @opindex -Wscalar-storage-order
6255 Do not warn on suspicious constructs involving reverse scalar storage order.
6257 @item -Wsized-deallocation @r{(C++ and Objective-C++ only)}
6258 @opindex Wsized-deallocation
6259 @opindex Wno-sized-deallocation
6260 Warn about a definition of an unsized deallocation function
6262 void operator delete (void *) noexcept;
6263 void operator delete[] (void *) noexcept;
6265 without a definition of the corresponding sized deallocation function
6267 void operator delete (void *, std::size_t) noexcept;
6268 void operator delete[] (void *, std::size_t) noexcept;
6270 or vice versa. Enabled by @option{-Wextra} along with
6271 @option{-fsized-deallocation}.
6273 @item -Wsizeof-pointer-div
6274 @opindex Wsizeof-pointer-div
6275 @opindex Wno-sizeof-pointer-div
6276 Warn for suspicious divisions of two sizeof expressions that divide
6277 the pointer size by the element size, which is the usual way to compute
6278 the array size but won't work out correctly with pointers. This warning
6279 warns e.g.@: about @code{sizeof (ptr) / sizeof (ptr[0])} if @code{ptr} is
6280 not an array, but a pointer. This warning is enabled by @option{-Wall}.
6282 @item -Wsizeof-pointer-memaccess
6283 @opindex Wsizeof-pointer-memaccess
6284 @opindex Wno-sizeof-pointer-memaccess
6285 Warn for suspicious length parameters to certain string and memory built-in
6286 functions if the argument uses @code{sizeof}. This warning triggers for
6287 example for @code{memset (ptr, 0, sizeof (ptr));} if @code{ptr} is not
6288 an array, but a pointer, and suggests a possible fix, or about
6289 @code{memcpy (&foo, ptr, sizeof (&foo));}. @option{-Wsizeof-pointer-memaccess}
6290 also warns about calls to bounded string copy functions like @code{strncat}
6291 or @code{strncpy} that specify as the bound a @code{sizeof} expression of
6292 the source array. For example, in the following function the call to
6293 @code{strncat} specifies the size of the source string as the bound. That
6294 is almost certainly a mistake and so the call is diagnosed.
6296 void make_file (const char *name)
6298 char path[PATH_MAX];
6299 strncpy (path, name, sizeof path - 1);
6300 strncat (path, ".text", sizeof ".text");
6305 The @option{-Wsizeof-pointer-memaccess} option is enabled by @option{-Wall}.
6307 @item -Wsizeof-array-argument
6308 @opindex Wsizeof-array-argument
6309 @opindex Wno-sizeof-array-argument
6310 Warn when the @code{sizeof} operator is applied to a parameter that is
6311 declared as an array in a function definition. This warning is enabled by
6312 default for C and C++ programs.
6314 @item -Wmemset-elt-size
6315 @opindex Wmemset-elt-size
6316 @opindex Wno-memset-elt-size
6317 Warn for suspicious calls to the @code{memset} built-in function, if the
6318 first argument references an array, and the third argument is a number
6319 equal to the number of elements, but not equal to the size of the array
6320 in memory. This indicates that the user has omitted a multiplication by
6321 the element size. This warning is enabled by @option{-Wall}.
6323 @item -Wmemset-transposed-args
6324 @opindex Wmemset-transposed-args
6325 @opindex Wno-memset-transposed-args
6326 Warn for suspicious calls to the @code{memset} built-in function, if the
6327 second argument is not zero and the third argument is zero. This warns e.g.@
6328 about @code{memset (buf, sizeof buf, 0)} where most probably
6329 @code{memset (buf, 0, sizeof buf)} was meant instead. The diagnostics
6330 is only emitted if the third argument is literal zero. If it is some
6331 expression that is folded to zero, a cast of zero to some type, etc.,
6332 it is far less likely that the user has mistakenly exchanged the arguments
6333 and no warning is emitted. This warning is enabled by @option{-Wall}.
6337 @opindex Wno-address
6338 Warn about suspicious uses of memory addresses. These include using
6339 the address of a function in a conditional expression, such as
6340 @code{void func(void); if (func)}, and comparisons against the memory
6341 address of a string literal, such as @code{if (x == "abc")}. Such
6342 uses typically indicate a programmer error: the address of a function
6343 always evaluates to true, so their use in a conditional usually
6344 indicate that the programmer forgot the parentheses in a function
6345 call; and comparisons against string literals result in unspecified
6346 behavior and are not portable in C, so they usually indicate that the
6347 programmer intended to use @code{strcmp}. This warning is enabled by
6351 @opindex Wlogical-op
6352 @opindex Wno-logical-op
6353 Warn about suspicious uses of logical operators in expressions.
6354 This includes using logical operators in contexts where a
6355 bit-wise operator is likely to be expected. Also warns when
6356 the operands of a logical operator are the same:
6359 if (a < 0 && a < 0) @{ @dots{} @}
6362 @item -Wlogical-not-parentheses
6363 @opindex Wlogical-not-parentheses
6364 @opindex Wno-logical-not-parentheses
6365 Warn about logical not used on the left hand side operand of a comparison.
6366 This option does not warn if the right operand is considered to be a boolean
6367 expression. Its purpose is to detect suspicious code like the following:
6371 if (!a > 1) @{ @dots{} @}
6374 It is possible to suppress the warning by wrapping the LHS into
6377 if ((!a) > 1) @{ @dots{} @}
6380 This warning is enabled by @option{-Wall}.
6382 @item -Waggregate-return
6383 @opindex Waggregate-return
6384 @opindex Wno-aggregate-return
6385 Warn if any functions that return structures or unions are defined or
6386 called. (In languages where you can return an array, this also elicits
6389 @item -Wno-aggressive-loop-optimizations
6390 @opindex Wno-aggressive-loop-optimizations
6391 @opindex Waggressive-loop-optimizations
6392 Warn if in a loop with constant number of iterations the compiler detects
6393 undefined behavior in some statement during one or more of the iterations.
6395 @item -Wno-attributes
6396 @opindex Wno-attributes
6397 @opindex Wattributes
6398 Do not warn if an unexpected @code{__attribute__} is used, such as
6399 unrecognized attributes, function attributes applied to variables,
6400 etc. This does not stop errors for incorrect use of supported
6403 @item -Wno-builtin-declaration-mismatch
6404 @opindex Wno-builtin-declaration-mismatch
6405 @opindex Wbuiltin-declaration-mismatch
6406 Warn if a built-in function is declared with the wrong signature or
6408 This warning is enabled by default.
6410 @item -Wno-builtin-macro-redefined
6411 @opindex Wno-builtin-macro-redefined
6412 @opindex Wbuiltin-macro-redefined
6413 Do not warn if certain built-in macros are redefined. This suppresses
6414 warnings for redefinition of @code{__TIMESTAMP__}, @code{__TIME__},
6415 @code{__DATE__}, @code{__FILE__}, and @code{__BASE_FILE__}.
6417 @item -Wstrict-prototypes @r{(C and Objective-C only)}
6418 @opindex Wstrict-prototypes
6419 @opindex Wno-strict-prototypes
6420 Warn if a function is declared or defined without specifying the
6421 argument types. (An old-style function definition is permitted without
6422 a warning if preceded by a declaration that specifies the argument
6425 @item -Wold-style-declaration @r{(C and Objective-C only)}
6426 @opindex Wold-style-declaration
6427 @opindex Wno-old-style-declaration
6428 Warn for obsolescent usages, according to the C Standard, in a
6429 declaration. For example, warn if storage-class specifiers like
6430 @code{static} are not the first things in a declaration. This warning
6431 is also enabled by @option{-Wextra}.
6433 @item -Wold-style-definition @r{(C and Objective-C only)}
6434 @opindex Wold-style-definition
6435 @opindex Wno-old-style-definition
6436 Warn if an old-style function definition is used. A warning is given
6437 even if there is a previous prototype.
6439 @item -Wmissing-parameter-type @r{(C and Objective-C only)}
6440 @opindex Wmissing-parameter-type
6441 @opindex Wno-missing-parameter-type
6442 A function parameter is declared without a type specifier in K&R-style
6449 This warning is also enabled by @option{-Wextra}.
6451 @item -Wmissing-prototypes @r{(C and Objective-C only)}
6452 @opindex Wmissing-prototypes
6453 @opindex Wno-missing-prototypes
6454 Warn if a global function is defined without a previous prototype
6455 declaration. This warning is issued even if the definition itself
6456 provides a prototype. Use this option to detect global functions
6457 that do not have a matching prototype declaration in a header file.
6458 This option is not valid for C++ because all function declarations
6459 provide prototypes and a non-matching declaration declares an
6460 overload rather than conflict with an earlier declaration.
6461 Use @option{-Wmissing-declarations} to detect missing declarations in C++.
6463 @item -Wmissing-declarations
6464 @opindex Wmissing-declarations
6465 @opindex Wno-missing-declarations
6466 Warn if a global function is defined without a previous declaration.
6467 Do so even if the definition itself provides a prototype.
6468 Use this option to detect global functions that are not declared in
6469 header files. In C, no warnings are issued for functions with previous
6470 non-prototype declarations; use @option{-Wmissing-prototypes} to detect
6471 missing prototypes. In C++, no warnings are issued for function templates,
6472 or for inline functions, or for functions in anonymous namespaces.
6474 @item -Wmissing-field-initializers
6475 @opindex Wmissing-field-initializers
6476 @opindex Wno-missing-field-initializers
6480 Warn if a structure's initializer has some fields missing. For
6481 example, the following code causes such a warning, because
6482 @code{x.h} is implicitly zero:
6485 struct s @{ int f, g, h; @};
6486 struct s x = @{ 3, 4 @};
6489 This option does not warn about designated initializers, so the following
6490 modification does not trigger a warning:
6493 struct s @{ int f, g, h; @};
6494 struct s x = @{ .f = 3, .g = 4 @};
6497 In C this option does not warn about the universal zero initializer
6501 struct s @{ int f, g, h; @};
6502 struct s x = @{ 0 @};
6505 Likewise, in C++ this option does not warn about the empty @{ @}
6506 initializer, for example:
6509 struct s @{ int f, g, h; @};
6513 This warning is included in @option{-Wextra}. To get other @option{-Wextra}
6514 warnings without this one, use @option{-Wextra -Wno-missing-field-initializers}.
6516 @item -Wno-multichar
6517 @opindex Wno-multichar
6519 Do not warn if a multicharacter constant (@samp{'FOOF'}) is used.
6520 Usually they indicate a typo in the user's code, as they have
6521 implementation-defined values, and should not be used in portable code.
6523 @item -Wnormalized=@r{[}none@r{|}id@r{|}nfc@r{|}nfkc@r{]}
6524 @opindex Wnormalized=
6525 @opindex Wnormalized
6526 @opindex Wno-normalized
6529 @cindex character set, input normalization
6530 In ISO C and ISO C++, two identifiers are different if they are
6531 different sequences of characters. However, sometimes when characters
6532 outside the basic ASCII character set are used, you can have two
6533 different character sequences that look the same. To avoid confusion,
6534 the ISO 10646 standard sets out some @dfn{normalization rules} which
6535 when applied ensure that two sequences that look the same are turned into
6536 the same sequence. GCC can warn you if you are using identifiers that
6537 have not been normalized; this option controls that warning.
6539 There are four levels of warning supported by GCC@. The default is
6540 @option{-Wnormalized=nfc}, which warns about any identifier that is
6541 not in the ISO 10646 ``C'' normalized form, @dfn{NFC}. NFC is the
6542 recommended form for most uses. It is equivalent to
6543 @option{-Wnormalized}.
6545 Unfortunately, there are some characters allowed in identifiers by
6546 ISO C and ISO C++ that, when turned into NFC, are not allowed in
6547 identifiers. That is, there's no way to use these symbols in portable
6548 ISO C or C++ and have all your identifiers in NFC@.
6549 @option{-Wnormalized=id} suppresses the warning for these characters.
6550 It is hoped that future versions of the standards involved will correct
6551 this, which is why this option is not the default.
6553 You can switch the warning off for all characters by writing
6554 @option{-Wnormalized=none} or @option{-Wno-normalized}. You should
6555 only do this if you are using some other normalization scheme (like
6556 ``D''), because otherwise you can easily create bugs that are
6557 literally impossible to see.
6559 Some characters in ISO 10646 have distinct meanings but look identical
6560 in some fonts or display methodologies, especially once formatting has
6561 been applied. For instance @code{\u207F}, ``SUPERSCRIPT LATIN SMALL
6562 LETTER N'', displays just like a regular @code{n} that has been
6563 placed in a superscript. ISO 10646 defines the @dfn{NFKC}
6564 normalization scheme to convert all these into a standard form as
6565 well, and GCC warns if your code is not in NFKC if you use
6566 @option{-Wnormalized=nfkc}. This warning is comparable to warning
6567 about every identifier that contains the letter O because it might be
6568 confused with the digit 0, and so is not the default, but may be
6569 useful as a local coding convention if the programming environment
6570 cannot be fixed to display these characters distinctly.
6572 @item -Wno-deprecated
6573 @opindex Wno-deprecated
6574 @opindex Wdeprecated
6575 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
6577 @item -Wno-deprecated-declarations
6578 @opindex Wno-deprecated-declarations
6579 @opindex Wdeprecated-declarations
6580 Do not warn about uses of functions (@pxref{Function Attributes}),
6581 variables (@pxref{Variable Attributes}), and types (@pxref{Type
6582 Attributes}) marked as deprecated by using the @code{deprecated}
6586 @opindex Wno-overflow
6588 Do not warn about compile-time overflow in constant expressions.
6593 Warn about One Definition Rule violations during link-time optimization.
6594 Requires @option{-flto-odr-type-merging} to be enabled. Enabled by default.
6597 @opindex Wopenm-simd
6598 Warn if the vectorizer cost model overrides the OpenMP or the Cilk Plus
6599 simd directive set by user. The @option{-fsimd-cost-model=unlimited}
6600 option can be used to relax the cost model.
6602 @item -Woverride-init @r{(C and Objective-C only)}
6603 @opindex Woverride-init
6604 @opindex Wno-override-init
6608 Warn if an initialized field without side effects is overridden when
6609 using designated initializers (@pxref{Designated Inits, , Designated
6612 This warning is included in @option{-Wextra}. To get other
6613 @option{-Wextra} warnings without this one, use @option{-Wextra
6614 -Wno-override-init}.
6616 @item -Woverride-init-side-effects @r{(C and Objective-C only)}
6617 @opindex Woverride-init-side-effects
6618 @opindex Wno-override-init-side-effects
6619 Warn if an initialized field with side effects is overridden when
6620 using designated initializers (@pxref{Designated Inits, , Designated
6621 Initializers}). This warning is enabled by default.
6626 Warn if a structure is given the packed attribute, but the packed
6627 attribute has no effect on the layout or size of the structure.
6628 Such structures may be mis-aligned for little benefit. For
6629 instance, in this code, the variable @code{f.x} in @code{struct bar}
6630 is misaligned even though @code{struct bar} does not itself
6631 have the packed attribute:
6638 @} __attribute__((packed));
6646 @item -Wpacked-bitfield-compat
6647 @opindex Wpacked-bitfield-compat
6648 @opindex Wno-packed-bitfield-compat
6649 The 4.1, 4.2 and 4.3 series of GCC ignore the @code{packed} attribute
6650 on bit-fields of type @code{char}. This has been fixed in GCC 4.4 but
6651 the change can lead to differences in the structure layout. GCC
6652 informs you when the offset of such a field has changed in GCC 4.4.
6653 For example there is no longer a 4-bit padding between field @code{a}
6654 and @code{b} in this structure:
6661 @} __attribute__ ((packed));
6664 This warning is enabled by default. Use
6665 @option{-Wno-packed-bitfield-compat} to disable this warning.
6667 @item -Wpacked-not-aligned @r{(C, C++, Objective-C and Objective-C++ only)}
6668 @opindex Wpacked-not-aligned
6669 @opindex Wno-packed-not-aligned
6670 Warn if a structure field with explicitly specified alignment in a
6671 packed struct or union is misaligned. For example, a warning will
6672 be issued on @code{struct S}, like, @code{warning: alignment 1 of
6673 'struct S' is less than 8}, in this code:
6677 struct __attribute__ ((aligned (8))) S8 @{ char a[8]; @};
6678 struct __attribute__ ((packed)) S @{
6684 This warning is enabled by @option{-Wall}.
6689 Warn if padding is included in a structure, either to align an element
6690 of the structure or to align the whole structure. Sometimes when this
6691 happens it is possible to rearrange the fields of the structure to
6692 reduce the padding and so make the structure smaller.
6694 @item -Wredundant-decls
6695 @opindex Wredundant-decls
6696 @opindex Wno-redundant-decls
6697 Warn if anything is declared more than once in the same scope, even in
6698 cases where multiple declaration is valid and changes nothing.
6702 @opindex Wno-restrict
6703 Warn when an argument passed to a restrict-qualified parameter
6704 aliases with another argument.
6706 @item -Wnested-externs @r{(C and Objective-C only)}
6707 @opindex Wnested-externs
6708 @opindex Wno-nested-externs
6709 Warn if an @code{extern} declaration is encountered within a function.
6711 @item -Wno-inherited-variadic-ctor
6712 @opindex Winherited-variadic-ctor
6713 @opindex Wno-inherited-variadic-ctor
6714 Suppress warnings about use of C++11 inheriting constructors when the
6715 base class inherited from has a C variadic constructor; the warning is
6716 on by default because the ellipsis is not inherited.
6721 Warn if a function that is declared as inline cannot be inlined.
6722 Even with this option, the compiler does not warn about failures to
6723 inline functions declared in system headers.
6725 The compiler uses a variety of heuristics to determine whether or not
6726 to inline a function. For example, the compiler takes into account
6727 the size of the function being inlined and the amount of inlining
6728 that has already been done in the current function. Therefore,
6729 seemingly insignificant changes in the source program can cause the
6730 warnings produced by @option{-Winline} to appear or disappear.
6732 @item -Wno-invalid-offsetof @r{(C++ and Objective-C++ only)}
6733 @opindex Wno-invalid-offsetof
6734 @opindex Winvalid-offsetof
6735 Suppress warnings from applying the @code{offsetof} macro to a non-POD
6736 type. According to the 2014 ISO C++ standard, applying @code{offsetof}
6737 to a non-standard-layout type is undefined. In existing C++ implementations,
6738 however, @code{offsetof} typically gives meaningful results.
6739 This flag is for users who are aware that they are
6740 writing nonportable code and who have deliberately chosen to ignore the
6743 The restrictions on @code{offsetof} may be relaxed in a future version
6744 of the C++ standard.
6746 @item -Wint-in-bool-context
6747 @opindex Wint-in-bool-context
6748 @opindex Wno-int-in-bool-context
6749 Warn for suspicious use of integer values where boolean values are expected,
6750 such as conditional expressions (?:) using non-boolean integer constants in
6751 boolean context, like @code{if (a <= b ? 2 : 3)}. Or left shifting of signed
6752 integers in boolean context, like @code{for (a = 0; 1 << a; a++);}. Likewise
6753 for all kinds of multiplications regardless of the data type.
6754 This warning is enabled by @option{-Wall}.
6756 @item -Wno-int-to-pointer-cast
6757 @opindex Wno-int-to-pointer-cast
6758 @opindex Wint-to-pointer-cast
6759 Suppress warnings from casts to pointer type of an integer of a
6760 different size. In C++, casting to a pointer type of smaller size is
6761 an error. @option{Wint-to-pointer-cast} is enabled by default.
6764 @item -Wno-pointer-to-int-cast @r{(C and Objective-C only)}
6765 @opindex Wno-pointer-to-int-cast
6766 @opindex Wpointer-to-int-cast
6767 Suppress warnings from casts from a pointer to an integer type of a
6771 @opindex Winvalid-pch
6772 @opindex Wno-invalid-pch
6773 Warn if a precompiled header (@pxref{Precompiled Headers}) is found in
6774 the search path but cannot be used.
6778 @opindex Wno-long-long
6779 Warn if @code{long long} type is used. This is enabled by either
6780 @option{-Wpedantic} or @option{-Wtraditional} in ISO C90 and C++98
6781 modes. To inhibit the warning messages, use @option{-Wno-long-long}.
6783 @item -Wvariadic-macros
6784 @opindex Wvariadic-macros
6785 @opindex Wno-variadic-macros
6786 Warn if variadic macros are used in ISO C90 mode, or if the GNU
6787 alternate syntax is used in ISO C99 mode. This is enabled by either
6788 @option{-Wpedantic} or @option{-Wtraditional}. To inhibit the warning
6789 messages, use @option{-Wno-variadic-macros}.
6793 @opindex Wno-varargs
6794 Warn upon questionable usage of the macros used to handle variable
6795 arguments like @code{va_start}. This is default. To inhibit the
6796 warning messages, use @option{-Wno-varargs}.
6798 @item -Wvector-operation-performance
6799 @opindex Wvector-operation-performance
6800 @opindex Wno-vector-operation-performance
6801 Warn if vector operation is not implemented via SIMD capabilities of the
6802 architecture. Mainly useful for the performance tuning.
6803 Vector operation can be implemented @code{piecewise}, which means that the
6804 scalar operation is performed on every vector element;
6805 @code{in parallel}, which means that the vector operation is implemented
6806 using scalars of wider type, which normally is more performance efficient;
6807 and @code{as a single scalar}, which means that vector fits into a
6810 @item -Wno-virtual-move-assign
6811 @opindex Wvirtual-move-assign
6812 @opindex Wno-virtual-move-assign
6813 Suppress warnings about inheriting from a virtual base with a
6814 non-trivial C++11 move assignment operator. This is dangerous because
6815 if the virtual base is reachable along more than one path, it is
6816 moved multiple times, which can mean both objects end up in the
6817 moved-from state. If the move assignment operator is written to avoid
6818 moving from a moved-from object, this warning can be disabled.
6823 Warn if a variable-length array is used in the code.
6824 @option{-Wno-vla} prevents the @option{-Wpedantic} warning of
6825 the variable-length array.
6827 @item -Wvla-larger-than=@var{n}
6828 If this option is used, the compiler will warn on uses of
6829 variable-length arrays where the size is either unbounded, or bounded
6830 by an argument that can be larger than @var{n} bytes. This is similar
6831 to how @option{-Walloca-larger-than=@var{n}} works, but with
6832 variable-length arrays.
6834 Note that GCC may optimize small variable-length arrays of a known
6835 value into plain arrays, so this warning may not get triggered for
6838 This warning is not enabled by @option{-Wall}, and is only active when
6839 @option{-ftree-vrp} is active (default for @option{-O2} and above).
6841 See also @option{-Walloca-larger-than=@var{n}}.
6843 @item -Wvolatile-register-var
6844 @opindex Wvolatile-register-var
6845 @opindex Wno-volatile-register-var
6846 Warn if a register variable is declared volatile. The volatile
6847 modifier does not inhibit all optimizations that may eliminate reads
6848 and/or writes to register variables. This warning is enabled by
6851 @item -Wdisabled-optimization
6852 @opindex Wdisabled-optimization
6853 @opindex Wno-disabled-optimization
6854 Warn if a requested optimization pass is disabled. This warning does
6855 not generally indicate that there is anything wrong with your code; it
6856 merely indicates that GCC's optimizers are unable to handle the code
6857 effectively. Often, the problem is that your code is too big or too
6858 complex; GCC refuses to optimize programs when the optimization
6859 itself is likely to take inordinate amounts of time.
6861 @item -Wpointer-sign @r{(C and Objective-C only)}
6862 @opindex Wpointer-sign
6863 @opindex Wno-pointer-sign
6864 Warn for pointer argument passing or assignment with different signedness.
6865 This option is only supported for C and Objective-C@. It is implied by
6866 @option{-Wall} and by @option{-Wpedantic}, which can be disabled with
6867 @option{-Wno-pointer-sign}.
6869 @item -Wstack-protector
6870 @opindex Wstack-protector
6871 @opindex Wno-stack-protector
6872 This option is only active when @option{-fstack-protector} is active. It
6873 warns about functions that are not protected against stack smashing.
6875 @item -Woverlength-strings
6876 @opindex Woverlength-strings
6877 @opindex Wno-overlength-strings
6878 Warn about string constants that are longer than the ``minimum
6879 maximum'' length specified in the C standard. Modern compilers
6880 generally allow string constants that are much longer than the
6881 standard's minimum limit, but very portable programs should avoid
6882 using longer strings.
6884 The limit applies @emph{after} string constant concatenation, and does
6885 not count the trailing NUL@. In C90, the limit was 509 characters; in
6886 C99, it was raised to 4095. C++98 does not specify a normative
6887 minimum maximum, so we do not diagnose overlength strings in C++@.
6889 This option is implied by @option{-Wpedantic}, and can be disabled with
6890 @option{-Wno-overlength-strings}.
6892 @item -Wunsuffixed-float-constants @r{(C and Objective-C only)}
6893 @opindex Wunsuffixed-float-constants
6895 Issue a warning for any floating constant that does not have
6896 a suffix. When used together with @option{-Wsystem-headers} it
6897 warns about such constants in system header files. This can be useful
6898 when preparing code to use with the @code{FLOAT_CONST_DECIMAL64} pragma
6899 from the decimal floating-point extension to C99.
6901 @item -Wno-designated-init @r{(C and Objective-C only)}
6902 Suppress warnings when a positional initializer is used to initialize
6903 a structure that has been marked with the @code{designated_init}
6907 Issue a warning when HSAIL cannot be emitted for the compiled function or
6912 @node Debugging Options
6913 @section Options for Debugging Your Program
6914 @cindex options, debugging
6915 @cindex debugging information options
6917 To tell GCC to emit extra information for use by a debugger, in almost
6918 all cases you need only to add @option{-g} to your other options.
6920 GCC allows you to use @option{-g} with
6921 @option{-O}. The shortcuts taken by optimized code may occasionally
6922 be surprising: some variables you declared may not exist
6923 at all; flow of control may briefly move where you did not expect it;
6924 some statements may not be executed because they compute constant
6925 results or their values are already at hand; some statements may
6926 execute in different places because they have been moved out of loops.
6927 Nevertheless it is possible to debug optimized output. This makes
6928 it reasonable to use the optimizer for programs that might have bugs.
6930 If you are not using some other optimization option, consider
6931 using @option{-Og} (@pxref{Optimize Options}) with @option{-g}.
6932 With no @option{-O} option at all, some compiler passes that collect
6933 information useful for debugging do not run at all, so that
6934 @option{-Og} may result in a better debugging experience.
6939 Produce debugging information in the operating system's native format
6940 (stabs, COFF, XCOFF, or DWARF)@. GDB can work with this debugging
6943 On most systems that use stabs format, @option{-g} enables use of extra
6944 debugging information that only GDB can use; this extra information
6945 makes debugging work better in GDB but probably makes other debuggers
6947 refuse to read the program. If you want to control for certain whether
6948 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
6949 @option{-gxcoff+}, @option{-gxcoff}, or @option{-gvms} (see below).
6953 Produce debugging information for use by GDB@. This means to use the
6954 most expressive format available (DWARF, stabs, or the native format
6955 if neither of those are supported), including GDB extensions if at all
6959 @itemx -gdwarf-@var{version}
6961 Produce debugging information in DWARF format (if that is supported).
6962 The value of @var{version} may be either 2, 3, 4 or 5; the default version
6963 for most targets is 4. DWARF Version 5 is only experimental.
6965 Note that with DWARF Version 2, some ports require and always
6966 use some non-conflicting DWARF 3 extensions in the unwind tables.
6968 Version 4 may require GDB 7.0 and @option{-fvar-tracking-assignments}
6969 for maximum benefit.
6971 GCC no longer supports DWARF Version 1, which is substantially
6972 different than Version 2 and later. For historical reasons, some
6973 other DWARF-related options such as
6974 @option{-fno-dwarf2-cfi-asm}) retain a reference to DWARF Version 2
6975 in their names, but apply to all currently-supported versions of DWARF.
6979 Produce debugging information in stabs format (if that is supported),
6980 without GDB extensions. This is the format used by DBX on most BSD
6981 systems. On MIPS, Alpha and System V Release 4 systems this option
6982 produces stabs debugging output that is not understood by DBX@.
6983 On System V Release 4 systems this option requires the GNU assembler.
6987 Produce debugging information in stabs format (if that is supported),
6988 using GNU extensions understood only by the GNU debugger (GDB)@. The
6989 use of these extensions is likely to make other debuggers crash or
6990 refuse to read the program.
6994 Produce debugging information in XCOFF format (if that is supported).
6995 This is the format used by the DBX debugger on IBM RS/6000 systems.
6999 Produce debugging information in XCOFF format (if that is supported),
7000 using GNU extensions understood only by the GNU debugger (GDB)@. The
7001 use of these extensions is likely to make other debuggers crash or
7002 refuse to read the program, and may cause assemblers other than the GNU
7003 assembler (GAS) to fail with an error.
7007 Produce debugging information in Alpha/VMS debug format (if that is
7008 supported). This is the format used by DEBUG on Alpha/VMS systems.
7011 @itemx -ggdb@var{level}
7012 @itemx -gstabs@var{level}
7013 @itemx -gxcoff@var{level}
7014 @itemx -gvms@var{level}
7015 Request debugging information and also use @var{level} to specify how
7016 much information. The default level is 2.
7018 Level 0 produces no debug information at all. Thus, @option{-g0} negates
7021 Level 1 produces minimal information, enough for making backtraces in
7022 parts of the program that you don't plan to debug. This includes
7023 descriptions of functions and external variables, and line number
7024 tables, but no information about local variables.
7026 Level 3 includes extra information, such as all the macro definitions
7027 present in the program. Some debuggers support macro expansion when
7028 you use @option{-g3}.
7030 @option{-gdwarf} does not accept a concatenated debug level, to avoid
7031 confusion with @option{-gdwarf-@var{level}}.
7032 Instead use an additional @option{-g@var{level}} option to change the
7033 debug level for DWARF.
7035 @item -feliminate-unused-debug-symbols
7036 @opindex feliminate-unused-debug-symbols
7037 Produce debugging information in stabs format (if that is supported),
7038 for only symbols that are actually used.
7040 @item -femit-class-debug-always
7041 @opindex femit-class-debug-always
7042 Instead of emitting debugging information for a C++ class in only one
7043 object file, emit it in all object files using the class. This option
7044 should be used only with debuggers that are unable to handle the way GCC
7045 normally emits debugging information for classes because using this
7046 option increases the size of debugging information by as much as a
7049 @item -fno-merge-debug-strings
7050 @opindex fmerge-debug-strings
7051 @opindex fno-merge-debug-strings
7052 Direct the linker to not merge together strings in the debugging
7053 information that are identical in different object files. Merging is
7054 not supported by all assemblers or linkers. Merging decreases the size
7055 of the debug information in the output file at the cost of increasing
7056 link processing time. Merging is enabled by default.
7058 @item -fdebug-prefix-map=@var{old}=@var{new}
7059 @opindex fdebug-prefix-map
7060 When compiling files in directory @file{@var{old}}, record debugging
7061 information describing them as in @file{@var{new}} instead. This can be
7062 used to replace a build-time path with an install-time path in the debug info.
7063 It can also be used to change an absolute path to a relative path by using
7064 @file{.} for @var{new}. This can give more reproducible builds, which are
7065 location independent, but may require an extra command to tell GDB where to
7066 find the source files.
7068 @item -fvar-tracking
7069 @opindex fvar-tracking
7070 Run variable tracking pass. It computes where variables are stored at each
7071 position in code. Better debugging information is then generated
7072 (if the debugging information format supports this information).
7074 It is enabled by default when compiling with optimization (@option{-Os},
7075 @option{-O}, @option{-O2}, @dots{}), debugging information (@option{-g}) and
7076 the debug info format supports it.
7078 @item -fvar-tracking-assignments
7079 @opindex fvar-tracking-assignments
7080 @opindex fno-var-tracking-assignments
7081 Annotate assignments to user variables early in the compilation and
7082 attempt to carry the annotations over throughout the compilation all the
7083 way to the end, in an attempt to improve debug information while
7084 optimizing. Use of @option{-gdwarf-4} is recommended along with it.
7086 It can be enabled even if var-tracking is disabled, in which case
7087 annotations are created and maintained, but discarded at the end.
7088 By default, this flag is enabled together with @option{-fvar-tracking},
7089 except when selective scheduling is enabled.
7092 @opindex gsplit-dwarf
7093 Separate as much DWARF debugging information as possible into a
7094 separate output file with the extension @file{.dwo}. This option allows
7095 the build system to avoid linking files with debug information. To
7096 be useful, this option requires a debugger capable of reading @file{.dwo}
7101 Generate DWARF @code{.debug_pubnames} and @code{.debug_pubtypes} sections.
7103 @item -ggnu-pubnames
7104 @opindex ggnu-pubnames
7105 Generate @code{.debug_pubnames} and @code{.debug_pubtypes} sections in a format
7106 suitable for conversion into a GDB@ index. This option is only useful
7107 with a linker that can produce GDB@ index version 7.
7109 @item -fdebug-types-section
7110 @opindex fdebug-types-section
7111 @opindex fno-debug-types-section
7112 When using DWARF Version 4 or higher, type DIEs can be put into
7113 their own @code{.debug_types} section instead of making them part of the
7114 @code{.debug_info} section. It is more efficient to put them in a separate
7115 comdat sections since the linker can then remove duplicates.
7116 But not all DWARF consumers support @code{.debug_types} sections yet
7117 and on some objects @code{.debug_types} produces larger instead of smaller
7118 debugging information.
7120 @item -grecord-gcc-switches
7121 @item -gno-record-gcc-switches
7122 @opindex grecord-gcc-switches
7123 @opindex gno-record-gcc-switches
7124 This switch causes the command-line options used to invoke the
7125 compiler that may affect code generation to be appended to the
7126 DW_AT_producer attribute in DWARF debugging information. The options
7127 are concatenated with spaces separating them from each other and from
7128 the compiler version.
7129 It is enabled by default.
7130 See also @option{-frecord-gcc-switches} for another
7131 way of storing compiler options into the object file.
7133 @item -gstrict-dwarf
7134 @opindex gstrict-dwarf
7135 Disallow using extensions of later DWARF standard version than selected
7136 with @option{-gdwarf-@var{version}}. On most targets using non-conflicting
7137 DWARF extensions from later standard versions is allowed.
7139 @item -gno-strict-dwarf
7140 @opindex gno-strict-dwarf
7141 Allow using extensions of later DWARF standard version than selected with
7142 @option{-gdwarf-@var{version}}.
7145 @item -gno-column-info
7146 @opindex gcolumn-info
7147 @opindex gno-column-info
7148 Emit location column information into DWARF debugging information, rather
7149 than just file and line.
7150 This option is enabled by default.
7152 @item -gz@r{[}=@var{type}@r{]}
7154 Produce compressed debug sections in DWARF format, if that is supported.
7155 If @var{type} is not given, the default type depends on the capabilities
7156 of the assembler and linker used. @var{type} may be one of
7157 @samp{none} (don't compress debug sections), @samp{zlib} (use zlib
7158 compression in ELF gABI format), or @samp{zlib-gnu} (use zlib
7159 compression in traditional GNU format). If the linker doesn't support
7160 writing compressed debug sections, the option is rejected. Otherwise,
7161 if the assembler does not support them, @option{-gz} is silently ignored
7162 when producing object files.
7164 @item -femit-struct-debug-baseonly
7165 @opindex femit-struct-debug-baseonly
7166 Emit debug information for struct-like types
7167 only when the base name of the compilation source file
7168 matches the base name of file in which the struct is defined.
7170 This option substantially reduces the size of debugging information,
7171 but at significant potential loss in type information to the debugger.
7172 See @option{-femit-struct-debug-reduced} for a less aggressive option.
7173 See @option{-femit-struct-debug-detailed} for more detailed control.
7175 This option works only with DWARF debug output.
7177 @item -femit-struct-debug-reduced
7178 @opindex femit-struct-debug-reduced
7179 Emit debug information for struct-like types
7180 only when the base name of the compilation source file
7181 matches the base name of file in which the type is defined,
7182 unless the struct is a template or defined in a system header.
7184 This option significantly reduces the size of debugging information,
7185 with some potential loss in type information to the debugger.
7186 See @option{-femit-struct-debug-baseonly} for a more aggressive option.
7187 See @option{-femit-struct-debug-detailed} for more detailed control.
7189 This option works only with DWARF debug output.
7191 @item -femit-struct-debug-detailed@r{[}=@var{spec-list}@r{]}
7192 @opindex femit-struct-debug-detailed
7193 Specify the struct-like types
7194 for which the compiler generates debug information.
7195 The intent is to reduce duplicate struct debug information
7196 between different object files within the same program.
7198 This option is a detailed version of
7199 @option{-femit-struct-debug-reduced} and @option{-femit-struct-debug-baseonly},
7200 which serves for most needs.
7202 A specification has the syntax@*
7203 [@samp{dir:}|@samp{ind:}][@samp{ord:}|@samp{gen:}](@samp{any}|@samp{sys}|@samp{base}|@samp{none})
7205 The optional first word limits the specification to
7206 structs that are used directly (@samp{dir:}) or used indirectly (@samp{ind:}).
7207 A struct type is used directly when it is the type of a variable, member.
7208 Indirect uses arise through pointers to structs.
7209 That is, when use of an incomplete struct is valid, the use is indirect.
7211 @samp{struct one direct; struct two * indirect;}.
7213 The optional second word limits the specification to
7214 ordinary structs (@samp{ord:}) or generic structs (@samp{gen:}).
7215 Generic structs are a bit complicated to explain.
7216 For C++, these are non-explicit specializations of template classes,
7217 or non-template classes within the above.
7218 Other programming languages have generics,
7219 but @option{-femit-struct-debug-detailed} does not yet implement them.
7221 The third word specifies the source files for those
7222 structs for which the compiler should emit debug information.
7223 The values @samp{none} and @samp{any} have the normal meaning.
7224 The value @samp{base} means that
7225 the base of name of the file in which the type declaration appears
7226 must match the base of the name of the main compilation file.
7227 In practice, this means that when compiling @file{foo.c}, debug information
7228 is generated for types declared in that file and @file{foo.h},
7229 but not other header files.
7230 The value @samp{sys} means those types satisfying @samp{base}
7231 or declared in system or compiler headers.
7233 You may need to experiment to determine the best settings for your application.
7235 The default is @option{-femit-struct-debug-detailed=all}.
7237 This option works only with DWARF debug output.
7239 @item -fno-dwarf2-cfi-asm
7240 @opindex fdwarf2-cfi-asm
7241 @opindex fno-dwarf2-cfi-asm
7242 Emit DWARF unwind info as compiler generated @code{.eh_frame} section
7243 instead of using GAS @code{.cfi_*} directives.
7245 @item -fno-eliminate-unused-debug-types
7246 @opindex feliminate-unused-debug-types
7247 @opindex fno-eliminate-unused-debug-types
7248 Normally, when producing DWARF output, GCC avoids producing debug symbol
7249 output for types that are nowhere used in the source file being compiled.
7250 Sometimes it is useful to have GCC emit debugging
7251 information for all types declared in a compilation
7252 unit, regardless of whether or not they are actually used
7253 in that compilation unit, for example
7254 if, in the debugger, you want to cast a value to a type that is
7255 not actually used in your program (but is declared). More often,
7256 however, this results in a significant amount of wasted space.
7259 @node Optimize Options
7260 @section Options That Control Optimization
7261 @cindex optimize options
7262 @cindex options, optimization
7264 These options control various sorts of optimizations.
7266 Without any optimization option, the compiler's goal is to reduce the
7267 cost of compilation and to make debugging produce the expected
7268 results. Statements are independent: if you stop the program with a
7269 breakpoint between statements, you can then assign a new value to any
7270 variable or change the program counter to any other statement in the
7271 function and get exactly the results you expect from the source
7274 Turning on optimization flags makes the compiler attempt to improve
7275 the performance and/or code size at the expense of compilation time
7276 and possibly the ability to debug the program.
7278 The compiler performs optimization based on the knowledge it has of the
7279 program. Compiling multiple files at once to a single output file mode allows
7280 the compiler to use information gained from all of the files when compiling
7283 Not all optimizations are controlled directly by a flag. Only
7284 optimizations that have a flag are listed in this section.
7286 Most optimizations are only enabled if an @option{-O} level is set on
7287 the command line. Otherwise they are disabled, even if individual
7288 optimization flags are specified.
7290 Depending on the target and how GCC was configured, a slightly different
7291 set of optimizations may be enabled at each @option{-O} level than
7292 those listed here. You can invoke GCC with @option{-Q --help=optimizers}
7293 to find out the exact set of optimizations that are enabled at each level.
7294 @xref{Overall Options}, for examples.
7301 Optimize. Optimizing compilation takes somewhat more time, and a lot
7302 more memory for a large function.
7304 With @option{-O}, the compiler tries to reduce code size and execution
7305 time, without performing any optimizations that take a great deal of
7308 @option{-O} turns on the following optimization flags:
7311 -fbranch-count-reg @gol
7312 -fcombine-stack-adjustments @gol
7314 -fcprop-registers @gol
7317 -fdelayed-branch @gol
7319 -fforward-propagate @gol
7320 -fguess-branch-probability @gol
7321 -fif-conversion2 @gol
7322 -fif-conversion @gol
7323 -finline-functions-called-once @gol
7324 -fipa-pure-const @gol
7326 -fipa-reference @gol
7327 -fmerge-constants @gol
7328 -fmove-loop-invariants @gol
7329 -fomit-frame-pointer @gol
7330 -freorder-blocks @gol
7332 -fshrink-wrap-separate @gol
7333 -fsplit-wide-types @gol
7339 -ftree-coalesce-vars @gol
7340 -ftree-copy-prop @gol
7342 -ftree-dominator-opts @gol
7344 -ftree-forwprop @gol
7356 Optimize even more. GCC performs nearly all supported optimizations
7357 that do not involve a space-speed tradeoff.
7358 As compared to @option{-O}, this option increases both compilation time
7359 and the performance of the generated code.
7361 @option{-O2} turns on all optimization flags specified by @option{-O}. It
7362 also turns on the following optimization flags:
7363 @gccoptlist{-fthread-jumps @gol
7364 -falign-functions -falign-jumps @gol
7365 -falign-loops -falign-labels @gol
7368 -fcse-follow-jumps -fcse-skip-blocks @gol
7369 -fdelete-null-pointer-checks @gol
7370 -fdevirtualize -fdevirtualize-speculatively @gol
7371 -fexpensive-optimizations @gol
7372 -fgcse -fgcse-lm @gol
7373 -fhoist-adjacent-loads @gol
7374 -finline-small-functions @gol
7375 -findirect-inlining @gol
7381 -fisolate-erroneous-paths-dereference @gol
7383 -foptimize-sibling-calls @gol
7384 -foptimize-strlen @gol
7385 -fpartial-inlining @gol
7387 -freorder-blocks-algorithm=stc @gol
7388 -freorder-blocks-and-partition -freorder-functions @gol
7389 -frerun-cse-after-loop @gol
7390 -fsched-interblock -fsched-spec @gol
7391 -fschedule-insns -fschedule-insns2 @gol
7392 -fstore-merging @gol
7393 -fstrict-aliasing @gol
7394 -ftree-builtin-call-dce @gol
7395 -ftree-switch-conversion -ftree-tail-merge @gol
7396 -fcode-hoisting @gol
7401 Please note the warning under @option{-fgcse} about
7402 invoking @option{-O2} on programs that use computed gotos.
7406 Optimize yet more. @option{-O3} turns on all optimizations specified
7407 by @option{-O2} and also turns on the following optimization flags:
7408 @gccoptlist{-finline-functions @gol
7409 -funswitch-loops @gol
7410 -fpredictive-commoning @gol
7411 -fgcse-after-reload @gol
7412 -ftree-loop-vectorize @gol
7413 -ftree-loop-distribution @gol
7414 -ftree-loop-distribute-patterns @gol
7416 -ftree-slp-vectorize @gol
7417 -fvect-cost-model @gol
7418 -ftree-partial-pre @gol
7424 Reduce compilation time and make debugging produce the expected
7425 results. This is the default.
7429 Optimize for size. @option{-Os} enables all @option{-O2} optimizations that
7430 do not typically increase code size. It also performs further
7431 optimizations designed to reduce code size.
7433 @option{-Os} disables the following optimization flags:
7434 @gccoptlist{-falign-functions -falign-jumps -falign-loops @gol
7435 -falign-labels -freorder-blocks -freorder-blocks-algorithm=stc @gol
7436 -freorder-blocks-and-partition -fprefetch-loop-arrays}
7440 Disregard strict standards compliance. @option{-Ofast} enables all
7441 @option{-O3} optimizations. It also enables optimizations that are not
7442 valid for all standard-compliant programs.
7443 It turns on @option{-ffast-math} and the Fortran-specific
7444 @option{-fstack-arrays}, unless @option{-fmax-stack-var-size} is
7445 specified, and @option{-fno-protect-parens}.
7449 Optimize debugging experience. @option{-Og} enables optimizations
7450 that do not interfere with debugging. It should be the optimization
7451 level of choice for the standard edit-compile-debug cycle, offering
7452 a reasonable level of optimization while maintaining fast compilation
7453 and a good debugging experience.
7456 If you use multiple @option{-O} options, with or without level numbers,
7457 the last such option is the one that is effective.
7459 Options of the form @option{-f@var{flag}} specify machine-independent
7460 flags. Most flags have both positive and negative forms; the negative
7461 form of @option{-ffoo} is @option{-fno-foo}. In the table
7462 below, only one of the forms is listed---the one you typically
7463 use. You can figure out the other form by either removing @samp{no-}
7466 The following options control specific optimizations. They are either
7467 activated by @option{-O} options or are related to ones that are. You
7468 can use the following flags in the rare cases when ``fine-tuning'' of
7469 optimizations to be performed is desired.
7472 @item -fno-defer-pop
7473 @opindex fno-defer-pop
7474 Always pop the arguments to each function call as soon as that function
7475 returns. For machines that must pop arguments after a function call,
7476 the compiler normally lets arguments accumulate on the stack for several
7477 function calls and pops them all at once.
7479 Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
7481 @item -fforward-propagate
7482 @opindex fforward-propagate
7483 Perform a forward propagation pass on RTL@. The pass tries to combine two
7484 instructions and checks if the result can be simplified. If loop unrolling
7485 is active, two passes are performed and the second is scheduled after
7488 This option is enabled by default at optimization levels @option{-O},
7489 @option{-O2}, @option{-O3}, @option{-Os}.
7491 @item -ffp-contract=@var{style}
7492 @opindex ffp-contract
7493 @option{-ffp-contract=off} disables floating-point expression contraction.
7494 @option{-ffp-contract=fast} enables floating-point expression contraction
7495 such as forming of fused multiply-add operations if the target has
7496 native support for them.
7497 @option{-ffp-contract=on} enables floating-point expression contraction
7498 if allowed by the language standard. This is currently not implemented
7499 and treated equal to @option{-ffp-contract=off}.
7501 The default is @option{-ffp-contract=fast}.
7503 @item -fomit-frame-pointer
7504 @opindex fomit-frame-pointer
7505 Omit the frame pointer in functions that don't need one. This avoids the
7506 instructions to save, set up and restore the frame pointer; on many targets
7507 it also makes an extra register available.
7509 On some targets this flag has no effect because the standard calling sequence
7510 always uses a frame pointer, so it cannot be omitted.
7512 Note that @option{-fno-omit-frame-pointer} doesn't guarantee the frame pointer
7513 is used in all functions. Several targets always omit the frame pointer in
7516 Enabled by default at @option{-O} and higher.
7518 @item -foptimize-sibling-calls
7519 @opindex foptimize-sibling-calls
7520 Optimize sibling and tail recursive calls.
7522 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
7524 @item -foptimize-strlen
7525 @opindex foptimize-strlen
7526 Optimize various standard C string functions (e.g. @code{strlen},
7527 @code{strchr} or @code{strcpy}) and
7528 their @code{_FORTIFY_SOURCE} counterparts into faster alternatives.
7530 Enabled at levels @option{-O2}, @option{-O3}.
7534 Do not expand any functions inline apart from those marked with
7535 the @code{always_inline} attribute. This is the default when not
7538 Single functions can be exempted from inlining by marking them
7539 with the @code{noinline} attribute.
7541 @item -finline-small-functions
7542 @opindex finline-small-functions
7543 Integrate functions into their callers when their body is smaller than expected
7544 function call code (so overall size of program gets smaller). The compiler
7545 heuristically decides which functions are simple enough to be worth integrating
7546 in this way. This inlining applies to all functions, even those not declared
7549 Enabled at level @option{-O2}.
7551 @item -findirect-inlining
7552 @opindex findirect-inlining
7553 Inline also indirect calls that are discovered to be known at compile
7554 time thanks to previous inlining. This option has any effect only
7555 when inlining itself is turned on by the @option{-finline-functions}
7556 or @option{-finline-small-functions} options.
7558 Enabled at level @option{-O2}.
7560 @item -finline-functions
7561 @opindex finline-functions
7562 Consider all functions for inlining, even if they are not declared inline.
7563 The compiler heuristically decides which functions are worth integrating
7566 If all calls to a given function are integrated, and the function is
7567 declared @code{static}, then the function is normally not output as
7568 assembler code in its own right.
7570 Enabled at level @option{-O3}.
7572 @item -finline-functions-called-once
7573 @opindex finline-functions-called-once
7574 Consider all @code{static} functions called once for inlining into their
7575 caller even if they are not marked @code{inline}. If a call to a given
7576 function is integrated, then the function is not output as assembler code
7579 Enabled at levels @option{-O1}, @option{-O2}, @option{-O3} and @option{-Os}.
7581 @item -fearly-inlining
7582 @opindex fearly-inlining
7583 Inline functions marked by @code{always_inline} and functions whose body seems
7584 smaller than the function call overhead early before doing
7585 @option{-fprofile-generate} instrumentation and real inlining pass. Doing so
7586 makes profiling significantly cheaper and usually inlining faster on programs
7587 having large chains of nested wrapper functions.
7593 Perform interprocedural scalar replacement of aggregates, removal of
7594 unused parameters and replacement of parameters passed by reference
7595 by parameters passed by value.
7597 Enabled at levels @option{-O2}, @option{-O3} and @option{-Os}.
7599 @item -finline-limit=@var{n}
7600 @opindex finline-limit
7601 By default, GCC limits the size of functions that can be inlined. This flag
7602 allows coarse control of this limit. @var{n} is the size of functions that
7603 can be inlined in number of pseudo instructions.
7605 Inlining is actually controlled by a number of parameters, which may be
7606 specified individually by using @option{--param @var{name}=@var{value}}.
7607 The @option{-finline-limit=@var{n}} option sets some of these parameters
7611 @item max-inline-insns-single
7612 is set to @var{n}/2.
7613 @item max-inline-insns-auto
7614 is set to @var{n}/2.
7617 See below for a documentation of the individual
7618 parameters controlling inlining and for the defaults of these parameters.
7620 @emph{Note:} there may be no value to @option{-finline-limit} that results
7621 in default behavior.
7623 @emph{Note:} pseudo instruction represents, in this particular context, an
7624 abstract measurement of function's size. In no way does it represent a count
7625 of assembly instructions and as such its exact meaning might change from one
7626 release to an another.
7628 @item -fno-keep-inline-dllexport
7629 @opindex fno-keep-inline-dllexport
7630 This is a more fine-grained version of @option{-fkeep-inline-functions},
7631 which applies only to functions that are declared using the @code{dllexport}
7632 attribute or declspec. @xref{Function Attributes,,Declaring Attributes of
7635 @item -fkeep-inline-functions
7636 @opindex fkeep-inline-functions
7637 In C, emit @code{static} functions that are declared @code{inline}
7638 into the object file, even if the function has been inlined into all
7639 of its callers. This switch does not affect functions using the
7640 @code{extern inline} extension in GNU C90@. In C++, emit any and all
7641 inline functions into the object file.
7643 @item -fkeep-static-functions
7644 @opindex fkeep-static-functions
7645 Emit @code{static} functions into the object file, even if the function
7648 @item -fkeep-static-consts
7649 @opindex fkeep-static-consts
7650 Emit variables declared @code{static const} when optimization isn't turned
7651 on, even if the variables aren't referenced.
7653 GCC enables this option by default. If you want to force the compiler to
7654 check if a variable is referenced, regardless of whether or not
7655 optimization is turned on, use the @option{-fno-keep-static-consts} option.
7657 @item -fmerge-constants
7658 @opindex fmerge-constants
7659 Attempt to merge identical constants (string constants and floating-point
7660 constants) across compilation units.
7662 This option is the default for optimized compilation if the assembler and
7663 linker support it. Use @option{-fno-merge-constants} to inhibit this
7666 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
7668 @item -fmerge-all-constants
7669 @opindex fmerge-all-constants
7670 Attempt to merge identical constants and identical variables.
7672 This option implies @option{-fmerge-constants}. In addition to
7673 @option{-fmerge-constants} this considers e.g.@: even constant initialized
7674 arrays or initialized constant variables with integral or floating-point
7675 types. Languages like C or C++ require each variable, including multiple
7676 instances of the same variable in recursive calls, to have distinct locations,
7677 so using this option results in non-conforming
7680 @item -fmodulo-sched
7681 @opindex fmodulo-sched
7682 Perform swing modulo scheduling immediately before the first scheduling
7683 pass. This pass looks at innermost loops and reorders their
7684 instructions by overlapping different iterations.
7686 @item -fmodulo-sched-allow-regmoves
7687 @opindex fmodulo-sched-allow-regmoves
7688 Perform more aggressive SMS-based modulo scheduling with register moves
7689 allowed. By setting this flag certain anti-dependences edges are
7690 deleted, which triggers the generation of reg-moves based on the
7691 life-range analysis. This option is effective only with
7692 @option{-fmodulo-sched} enabled.
7694 @item -fno-branch-count-reg
7695 @opindex fno-branch-count-reg
7696 Avoid running a pass scanning for opportunities to use ``decrement and
7697 branch'' instructions on a count register instead of generating sequences
7698 of instructions that decrement a register, compare it against zero, and
7699 then branch based upon the result. This option is only meaningful on
7700 architectures that support such instructions, which include x86, PowerPC,
7701 IA-64 and S/390. Note that the @option{-fno-branch-count-reg} option
7702 doesn't remove the decrement and branch instructions from the generated
7703 instruction stream introduced by other optimization passes.
7705 Enabled by default at @option{-O1} and higher.
7707 The default is @option{-fbranch-count-reg}.
7709 @item -fno-function-cse
7710 @opindex fno-function-cse
7711 Do not put function addresses in registers; make each instruction that
7712 calls a constant function contain the function's address explicitly.
7714 This option results in less efficient code, but some strange hacks
7715 that alter the assembler output may be confused by the optimizations
7716 performed when this option is not used.
7718 The default is @option{-ffunction-cse}
7720 @item -fno-zero-initialized-in-bss
7721 @opindex fno-zero-initialized-in-bss
7722 If the target supports a BSS section, GCC by default puts variables that
7723 are initialized to zero into BSS@. This can save space in the resulting
7726 This option turns off this behavior because some programs explicitly
7727 rely on variables going to the data section---e.g., so that the
7728 resulting executable can find the beginning of that section and/or make
7729 assumptions based on that.
7731 The default is @option{-fzero-initialized-in-bss}.
7733 @item -fthread-jumps
7734 @opindex fthread-jumps
7735 Perform optimizations that check to see if a jump branches to a
7736 location where another comparison subsumed by the first is found. If
7737 so, the first branch is redirected to either the destination of the
7738 second branch or a point immediately following it, depending on whether
7739 the condition is known to be true or false.
7741 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
7743 @item -fsplit-wide-types
7744 @opindex fsplit-wide-types
7745 When using a type that occupies multiple registers, such as @code{long
7746 long} on a 32-bit system, split the registers apart and allocate them
7747 independently. This normally generates better code for those types,
7748 but may make debugging more difficult.
7750 Enabled at levels @option{-O}, @option{-O2}, @option{-O3},
7753 @item -fcse-follow-jumps
7754 @opindex fcse-follow-jumps
7755 In common subexpression elimination (CSE), scan through jump instructions
7756 when the target of the jump is not reached by any other path. For
7757 example, when CSE encounters an @code{if} statement with an
7758 @code{else} clause, CSE follows the jump when the condition
7761 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
7763 @item -fcse-skip-blocks
7764 @opindex fcse-skip-blocks
7765 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
7766 follow jumps that conditionally skip over blocks. When CSE
7767 encounters a simple @code{if} statement with no else clause,
7768 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
7769 body of the @code{if}.
7771 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
7773 @item -frerun-cse-after-loop
7774 @opindex frerun-cse-after-loop
7775 Re-run common subexpression elimination after loop optimizations are
7778 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
7782 Perform a global common subexpression elimination pass.
7783 This pass also performs global constant and copy propagation.
7785 @emph{Note:} When compiling a program using computed gotos, a GCC
7786 extension, you may get better run-time performance if you disable
7787 the global common subexpression elimination pass by adding
7788 @option{-fno-gcse} to the command line.
7790 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
7794 When @option{-fgcse-lm} is enabled, global common subexpression elimination
7795 attempts to move loads that are only killed by stores into themselves. This
7796 allows a loop containing a load/store sequence to be changed to a load outside
7797 the loop, and a copy/store within the loop.
7799 Enabled by default when @option{-fgcse} is enabled.
7803 When @option{-fgcse-sm} is enabled, a store motion pass is run after
7804 global common subexpression elimination. This pass attempts to move
7805 stores out of loops. When used in conjunction with @option{-fgcse-lm},
7806 loops containing a load/store sequence can be changed to a load before
7807 the loop and a store after the loop.
7809 Not enabled at any optimization level.
7813 When @option{-fgcse-las} is enabled, the global common subexpression
7814 elimination pass eliminates redundant loads that come after stores to the
7815 same memory location (both partial and full redundancies).
7817 Not enabled at any optimization level.
7819 @item -fgcse-after-reload
7820 @opindex fgcse-after-reload
7821 When @option{-fgcse-after-reload} is enabled, a redundant load elimination
7822 pass is performed after reload. The purpose of this pass is to clean up
7825 @item -faggressive-loop-optimizations
7826 @opindex faggressive-loop-optimizations
7827 This option tells the loop optimizer to use language constraints to
7828 derive bounds for the number of iterations of a loop. This assumes that
7829 loop code does not invoke undefined behavior by for example causing signed
7830 integer overflows or out-of-bound array accesses. The bounds for the
7831 number of iterations of a loop are used to guide loop unrolling and peeling
7832 and loop exit test optimizations.
7833 This option is enabled by default.
7835 @item -funconstrained-commons
7836 @opindex funconstrained-commons
7837 This option tells the compiler that variables declared in common blocks
7838 (e.g. Fortran) may later be overridden with longer trailing arrays. This
7839 prevents certain optimizations that depend on knowing the array bounds.
7841 @item -fcrossjumping
7842 @opindex fcrossjumping
7843 Perform cross-jumping transformation.
7844 This transformation unifies equivalent code and saves code size. The
7845 resulting code may or may not perform better than without cross-jumping.
7847 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
7849 @item -fauto-inc-dec
7850 @opindex fauto-inc-dec
7851 Combine increments or decrements of addresses with memory accesses.
7852 This pass is always skipped on architectures that do not have
7853 instructions to support this. Enabled by default at @option{-O} and
7854 higher on architectures that support this.
7858 Perform dead code elimination (DCE) on RTL@.
7859 Enabled by default at @option{-O} and higher.
7863 Perform dead store elimination (DSE) on RTL@.
7864 Enabled by default at @option{-O} and higher.
7866 @item -fif-conversion
7867 @opindex fif-conversion
7868 Attempt to transform conditional jumps into branch-less equivalents. This
7869 includes use of conditional moves, min, max, set flags and abs instructions, and
7870 some tricks doable by standard arithmetics. The use of conditional execution
7871 on chips where it is available is controlled by @option{-fif-conversion2}.
7873 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
7875 @item -fif-conversion2
7876 @opindex fif-conversion2
7877 Use conditional execution (where available) to transform conditional jumps into
7878 branch-less equivalents.
7880 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
7882 @item -fdeclone-ctor-dtor
7883 @opindex fdeclone-ctor-dtor
7884 The C++ ABI requires multiple entry points for constructors and
7885 destructors: one for a base subobject, one for a complete object, and
7886 one for a virtual destructor that calls operator delete afterwards.
7887 For a hierarchy with virtual bases, the base and complete variants are
7888 clones, which means two copies of the function. With this option, the
7889 base and complete variants are changed to be thunks that call a common
7892 Enabled by @option{-Os}.
7894 @item -fdelete-null-pointer-checks
7895 @opindex fdelete-null-pointer-checks
7896 Assume that programs cannot safely dereference null pointers, and that
7897 no code or data element resides at address zero.
7898 This option enables simple constant
7899 folding optimizations at all optimization levels. In addition, other
7900 optimization passes in GCC use this flag to control global dataflow
7901 analyses that eliminate useless checks for null pointers; these assume
7902 that a memory access to address zero always results in a trap, so
7903 that if a pointer is checked after it has already been dereferenced,
7906 Note however that in some environments this assumption is not true.
7907 Use @option{-fno-delete-null-pointer-checks} to disable this optimization
7908 for programs that depend on that behavior.
7910 This option is enabled by default on most targets. On Nios II ELF, it
7911 defaults to off. On AVR, CR16, and MSP430, this option is completely disabled.
7913 Passes that use the dataflow information
7914 are enabled independently at different optimization levels.
7916 @item -fdevirtualize
7917 @opindex fdevirtualize
7918 Attempt to convert calls to virtual functions to direct calls. This
7919 is done both within a procedure and interprocedurally as part of
7920 indirect inlining (@option{-findirect-inlining}) and interprocedural constant
7921 propagation (@option{-fipa-cp}).
7922 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
7924 @item -fdevirtualize-speculatively
7925 @opindex fdevirtualize-speculatively
7926 Attempt to convert calls to virtual functions to speculative direct calls.
7927 Based on the analysis of the type inheritance graph, determine for a given call
7928 the set of likely targets. If the set is small, preferably of size 1, change
7929 the call into a conditional deciding between direct and indirect calls. The
7930 speculative calls enable more optimizations, such as inlining. When they seem
7931 useless after further optimization, they are converted back into original form.
7933 @item -fdevirtualize-at-ltrans
7934 @opindex fdevirtualize-at-ltrans
7935 Stream extra information needed for aggressive devirtualization when running
7936 the link-time optimizer in local transformation mode.
7937 This option enables more devirtualization but
7938 significantly increases the size of streamed data. For this reason it is
7939 disabled by default.
7941 @item -fexpensive-optimizations
7942 @opindex fexpensive-optimizations
7943 Perform a number of minor optimizations that are relatively expensive.
7945 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
7949 Attempt to remove redundant extension instructions. This is especially
7950 helpful for the x86-64 architecture, which implicitly zero-extends in 64-bit
7951 registers after writing to their lower 32-bit half.
7953 Enabled for Alpha, AArch64 and x86 at levels @option{-O2},
7954 @option{-O3}, @option{-Os}.
7956 @item -fno-lifetime-dse
7957 @opindex fno-lifetime-dse
7958 In C++ the value of an object is only affected by changes within its
7959 lifetime: when the constructor begins, the object has an indeterminate
7960 value, and any changes during the lifetime of the object are dead when
7961 the object is destroyed. Normally dead store elimination will take
7962 advantage of this; if your code relies on the value of the object
7963 storage persisting beyond the lifetime of the object, you can use this
7964 flag to disable this optimization. To preserve stores before the
7965 constructor starts (e.g. because your operator new clears the object
7966 storage) but still treat the object as dead after the destructor you,
7967 can use @option{-flifetime-dse=1}. The default behavior can be
7968 explicitly selected with @option{-flifetime-dse=2}.
7969 @option{-flifetime-dse=0} is equivalent to @option{-fno-lifetime-dse}.
7971 @item -flive-range-shrinkage
7972 @opindex flive-range-shrinkage
7973 Attempt to decrease register pressure through register live range
7974 shrinkage. This is helpful for fast processors with small or moderate
7977 @item -fira-algorithm=@var{algorithm}
7978 @opindex fira-algorithm
7979 Use the specified coloring algorithm for the integrated register
7980 allocator. The @var{algorithm} argument can be @samp{priority}, which
7981 specifies Chow's priority coloring, or @samp{CB}, which specifies
7982 Chaitin-Briggs coloring. Chaitin-Briggs coloring is not implemented
7983 for all architectures, but for those targets that do support it, it is
7984 the default because it generates better code.
7986 @item -fira-region=@var{region}
7987 @opindex fira-region
7988 Use specified regions for the integrated register allocator. The
7989 @var{region} argument should be one of the following:
7994 Use all loops as register allocation regions.
7995 This can give the best results for machines with a small and/or
7996 irregular register set.
7999 Use all loops except for loops with small register pressure
8000 as the regions. This value usually gives
8001 the best results in most cases and for most architectures,
8002 and is enabled by default when compiling with optimization for speed
8003 (@option{-O}, @option{-O2}, @dots{}).
8006 Use all functions as a single region.
8007 This typically results in the smallest code size, and is enabled by default for
8008 @option{-Os} or @option{-O0}.
8012 @item -fira-hoist-pressure
8013 @opindex fira-hoist-pressure
8014 Use IRA to evaluate register pressure in the code hoisting pass for
8015 decisions to hoist expressions. This option usually results in smaller
8016 code, but it can slow the compiler down.
8018 This option is enabled at level @option{-Os} for all targets.
8020 @item -fira-loop-pressure
8021 @opindex fira-loop-pressure
8022 Use IRA to evaluate register pressure in loops for decisions to move
8023 loop invariants. This option usually results in generation
8024 of faster and smaller code on machines with large register files (>= 32
8025 registers), but it can slow the compiler down.
8027 This option is enabled at level @option{-O3} for some targets.
8029 @item -fno-ira-share-save-slots
8030 @opindex fno-ira-share-save-slots
8031 Disable sharing of stack slots used for saving call-used hard
8032 registers living through a call. Each hard register gets a
8033 separate stack slot, and as a result function stack frames are
8036 @item -fno-ira-share-spill-slots
8037 @opindex fno-ira-share-spill-slots
8038 Disable sharing of stack slots allocated for pseudo-registers. Each
8039 pseudo-register that does not get a hard register gets a separate
8040 stack slot, and as a result function stack frames are larger.
8044 Enable CFG-sensitive rematerialization in LRA. Instead of loading
8045 values of spilled pseudos, LRA tries to rematerialize (recalculate)
8046 values if it is profitable.
8048 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
8050 @item -fdelayed-branch
8051 @opindex fdelayed-branch
8052 If supported for the target machine, attempt to reorder instructions
8053 to exploit instruction slots available after delayed branch
8056 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
8058 @item -fschedule-insns
8059 @opindex fschedule-insns
8060 If supported for the target machine, attempt to reorder instructions to
8061 eliminate execution stalls due to required data being unavailable. This
8062 helps machines that have slow floating point or memory load instructions
8063 by allowing other instructions to be issued until the result of the load
8064 or floating-point instruction is required.
8066 Enabled at levels @option{-O2}, @option{-O3}.
8068 @item -fschedule-insns2
8069 @opindex fschedule-insns2
8070 Similar to @option{-fschedule-insns}, but requests an additional pass of
8071 instruction scheduling after register allocation has been done. This is
8072 especially useful on machines with a relatively small number of
8073 registers and where memory load instructions take more than one cycle.
8075 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
8077 @item -fno-sched-interblock
8078 @opindex fno-sched-interblock
8079 Don't schedule instructions across basic blocks. This is normally
8080 enabled by default when scheduling before register allocation, i.e.@:
8081 with @option{-fschedule-insns} or at @option{-O2} or higher.
8083 @item -fno-sched-spec
8084 @opindex fno-sched-spec
8085 Don't allow speculative motion of non-load instructions. This is normally
8086 enabled by default when scheduling before register allocation, i.e.@:
8087 with @option{-fschedule-insns} or at @option{-O2} or higher.
8089 @item -fsched-pressure
8090 @opindex fsched-pressure
8091 Enable register pressure sensitive insn scheduling before register
8092 allocation. This only makes sense when scheduling before register
8093 allocation is enabled, i.e.@: with @option{-fschedule-insns} or at
8094 @option{-O2} or higher. Usage of this option can improve the
8095 generated code and decrease its size by preventing register pressure
8096 increase above the number of available hard registers and subsequent
8097 spills in register allocation.
8099 @item -fsched-spec-load
8100 @opindex fsched-spec-load
8101 Allow speculative motion of some load instructions. This only makes
8102 sense when scheduling before register allocation, i.e.@: with
8103 @option{-fschedule-insns} or at @option{-O2} or higher.
8105 @item -fsched-spec-load-dangerous
8106 @opindex fsched-spec-load-dangerous
8107 Allow speculative motion of more load instructions. This only makes
8108 sense when scheduling before register allocation, i.e.@: with
8109 @option{-fschedule-insns} or at @option{-O2} or higher.
8111 @item -fsched-stalled-insns
8112 @itemx -fsched-stalled-insns=@var{n}
8113 @opindex fsched-stalled-insns
8114 Define how many insns (if any) can be moved prematurely from the queue
8115 of stalled insns into the ready list during the second scheduling pass.
8116 @option{-fno-sched-stalled-insns} means that no insns are moved
8117 prematurely, @option{-fsched-stalled-insns=0} means there is no limit
8118 on how many queued insns can be moved prematurely.
8119 @option{-fsched-stalled-insns} without a value is equivalent to
8120 @option{-fsched-stalled-insns=1}.
8122 @item -fsched-stalled-insns-dep
8123 @itemx -fsched-stalled-insns-dep=@var{n}
8124 @opindex fsched-stalled-insns-dep
8125 Define how many insn groups (cycles) are examined for a dependency
8126 on a stalled insn that is a candidate for premature removal from the queue
8127 of stalled insns. This has an effect only during the second scheduling pass,
8128 and only if @option{-fsched-stalled-insns} is used.
8129 @option{-fno-sched-stalled-insns-dep} is equivalent to
8130 @option{-fsched-stalled-insns-dep=0}.
8131 @option{-fsched-stalled-insns-dep} without a value is equivalent to
8132 @option{-fsched-stalled-insns-dep=1}.
8134 @item -fsched2-use-superblocks
8135 @opindex fsched2-use-superblocks
8136 When scheduling after register allocation, use superblock scheduling.
8137 This allows motion across basic block boundaries,
8138 resulting in faster schedules. This option is experimental, as not all machine
8139 descriptions used by GCC model the CPU closely enough to avoid unreliable
8140 results from the algorithm.
8142 This only makes sense when scheduling after register allocation, i.e.@: with
8143 @option{-fschedule-insns2} or at @option{-O2} or higher.
8145 @item -fsched-group-heuristic
8146 @opindex fsched-group-heuristic
8147 Enable the group heuristic in the scheduler. This heuristic favors
8148 the instruction that belongs to a schedule group. This is enabled
8149 by default when scheduling is enabled, i.e.@: with @option{-fschedule-insns}
8150 or @option{-fschedule-insns2} or at @option{-O2} or higher.
8152 @item -fsched-critical-path-heuristic
8153 @opindex fsched-critical-path-heuristic
8154 Enable the critical-path heuristic in the scheduler. This heuristic favors
8155 instructions on the critical path. This is enabled by default when
8156 scheduling is enabled, i.e.@: with @option{-fschedule-insns}
8157 or @option{-fschedule-insns2} or at @option{-O2} or higher.
8159 @item -fsched-spec-insn-heuristic
8160 @opindex fsched-spec-insn-heuristic
8161 Enable the speculative instruction heuristic in the scheduler. This
8162 heuristic favors speculative instructions with greater dependency weakness.
8163 This is enabled by default when scheduling is enabled, i.e.@:
8164 with @option{-fschedule-insns} or @option{-fschedule-insns2}
8165 or at @option{-O2} or higher.
8167 @item -fsched-rank-heuristic
8168 @opindex fsched-rank-heuristic
8169 Enable the rank heuristic in the scheduler. This heuristic favors
8170 the instruction belonging to a basic block with greater size or frequency.
8171 This is enabled by default when scheduling is enabled, i.e.@:
8172 with @option{-fschedule-insns} or @option{-fschedule-insns2} or
8173 at @option{-O2} or higher.
8175 @item -fsched-last-insn-heuristic
8176 @opindex fsched-last-insn-heuristic
8177 Enable the last-instruction heuristic in the scheduler. This heuristic
8178 favors the instruction that is less dependent on the last instruction
8179 scheduled. This is enabled by default when scheduling is enabled,
8180 i.e.@: with @option{-fschedule-insns} or @option{-fschedule-insns2} or
8181 at @option{-O2} or higher.
8183 @item -fsched-dep-count-heuristic
8184 @opindex fsched-dep-count-heuristic
8185 Enable the dependent-count heuristic in the scheduler. This heuristic
8186 favors the instruction that has more instructions depending on it.
8187 This is enabled by default when scheduling is enabled, i.e.@:
8188 with @option{-fschedule-insns} or @option{-fschedule-insns2} or
8189 at @option{-O2} or higher.
8191 @item -freschedule-modulo-scheduled-loops
8192 @opindex freschedule-modulo-scheduled-loops
8193 Modulo scheduling is performed before traditional scheduling. If a loop
8194 is modulo scheduled, later scheduling passes may change its schedule.
8195 Use this option to control that behavior.
8197 @item -fselective-scheduling
8198 @opindex fselective-scheduling
8199 Schedule instructions using selective scheduling algorithm. Selective
8200 scheduling runs instead of the first scheduler pass.
8202 @item -fselective-scheduling2
8203 @opindex fselective-scheduling2
8204 Schedule instructions using selective scheduling algorithm. Selective
8205 scheduling runs instead of the second scheduler pass.
8207 @item -fsel-sched-pipelining
8208 @opindex fsel-sched-pipelining
8209 Enable software pipelining of innermost loops during selective scheduling.
8210 This option has no effect unless one of @option{-fselective-scheduling} or
8211 @option{-fselective-scheduling2} is turned on.
8213 @item -fsel-sched-pipelining-outer-loops
8214 @opindex fsel-sched-pipelining-outer-loops
8215 When pipelining loops during selective scheduling, also pipeline outer loops.
8216 This option has no effect unless @option{-fsel-sched-pipelining} is turned on.
8218 @item -fsemantic-interposition
8219 @opindex fsemantic-interposition
8220 Some object formats, like ELF, allow interposing of symbols by the
8222 This means that for symbols exported from the DSO, the compiler cannot perform
8223 interprocedural propagation, inlining and other optimizations in anticipation
8224 that the function or variable in question may change. While this feature is
8225 useful, for example, to rewrite memory allocation functions by a debugging
8226 implementation, it is expensive in the terms of code quality.
8227 With @option{-fno-semantic-interposition} the compiler assumes that
8228 if interposition happens for functions the overwriting function will have
8229 precisely the same semantics (and side effects).
8230 Similarly if interposition happens
8231 for variables, the constructor of the variable will be the same. The flag
8232 has no effect for functions explicitly declared inline
8233 (where it is never allowed for interposition to change semantics)
8234 and for symbols explicitly declared weak.
8237 @opindex fshrink-wrap
8238 Emit function prologues only before parts of the function that need it,
8239 rather than at the top of the function. This flag is enabled by default at
8240 @option{-O} and higher.
8242 @item -fshrink-wrap-separate
8243 @opindex fshrink-wrap-separate
8244 Shrink-wrap separate parts of the prologue and epilogue separately, so that
8245 those parts are only executed when needed.
8246 This option is on by default, but has no effect unless @option{-fshrink-wrap}
8247 is also turned on and the target supports this.
8249 @item -fcaller-saves
8250 @opindex fcaller-saves
8251 Enable allocation of values to registers that are clobbered by
8252 function calls, by emitting extra instructions to save and restore the
8253 registers around such calls. Such allocation is done only when it
8254 seems to result in better code.
8256 This option is always enabled by default on certain machines, usually
8257 those which have no call-preserved registers to use instead.
8259 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
8261 @item -fcombine-stack-adjustments
8262 @opindex fcombine-stack-adjustments
8263 Tracks stack adjustments (pushes and pops) and stack memory references
8264 and then tries to find ways to combine them.
8266 Enabled by default at @option{-O1} and higher.
8270 Use caller save registers for allocation if those registers are not used by
8271 any called function. In that case it is not necessary to save and restore
8272 them around calls. This is only possible if called functions are part of
8273 same compilation unit as current function and they are compiled before it.
8275 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}, however the option
8276 is disabled if generated code will be instrumented for profiling
8277 (@option{-p}, or @option{-pg}) or if callee's register usage cannot be known
8278 exactly (this happens on targets that do not expose prologues
8279 and epilogues in RTL).
8281 @item -fconserve-stack
8282 @opindex fconserve-stack
8283 Attempt to minimize stack usage. The compiler attempts to use less
8284 stack space, even if that makes the program slower. This option
8285 implies setting the @option{large-stack-frame} parameter to 100
8286 and the @option{large-stack-frame-growth} parameter to 400.
8288 @item -ftree-reassoc
8289 @opindex ftree-reassoc
8290 Perform reassociation on trees. This flag is enabled by default
8291 at @option{-O} and higher.
8293 @item -fcode-hoisting
8294 @opindex fcode-hoisting
8295 Perform code hoisting. Code hoisting tries to move the
8296 evaluation of expressions executed on all paths to the function exit
8297 as early as possible. This is especially useful as a code size
8298 optimization, but it often helps for code speed as well.
8299 This flag is enabled by default at @option{-O2} and higher.
8303 Perform partial redundancy elimination (PRE) on trees. This flag is
8304 enabled by default at @option{-O2} and @option{-O3}.
8306 @item -ftree-partial-pre
8307 @opindex ftree-partial-pre
8308 Make partial redundancy elimination (PRE) more aggressive. This flag is
8309 enabled by default at @option{-O3}.
8311 @item -ftree-forwprop
8312 @opindex ftree-forwprop
8313 Perform forward propagation on trees. This flag is enabled by default
8314 at @option{-O} and higher.
8318 Perform full redundancy elimination (FRE) on trees. The difference
8319 between FRE and PRE is that FRE only considers expressions
8320 that are computed on all paths leading to the redundant computation.
8321 This analysis is faster than PRE, though it exposes fewer redundancies.
8322 This flag is enabled by default at @option{-O} and higher.
8324 @item -ftree-phiprop
8325 @opindex ftree-phiprop
8326 Perform hoisting of loads from conditional pointers on trees. This
8327 pass is enabled by default at @option{-O} and higher.
8329 @item -fhoist-adjacent-loads
8330 @opindex fhoist-adjacent-loads
8331 Speculatively hoist loads from both branches of an if-then-else if the
8332 loads are from adjacent locations in the same structure and the target
8333 architecture has a conditional move instruction. This flag is enabled
8334 by default at @option{-O2} and higher.
8336 @item -ftree-copy-prop
8337 @opindex ftree-copy-prop
8338 Perform copy propagation on trees. This pass eliminates unnecessary
8339 copy operations. This flag is enabled by default at @option{-O} and
8342 @item -fipa-pure-const
8343 @opindex fipa-pure-const
8344 Discover which functions are pure or constant.
8345 Enabled by default at @option{-O} and higher.
8347 @item -fipa-reference
8348 @opindex fipa-reference
8349 Discover which static variables do not escape the
8351 Enabled by default at @option{-O} and higher.
8355 Perform interprocedural pointer analysis and interprocedural modification
8356 and reference analysis. This option can cause excessive memory and
8357 compile-time usage on large compilation units. It is not enabled by
8358 default at any optimization level.
8361 @opindex fipa-profile
8362 Perform interprocedural profile propagation. The functions called only from
8363 cold functions are marked as cold. Also functions executed once (such as
8364 @code{cold}, @code{noreturn}, static constructors or destructors) are identified. Cold
8365 functions and loop less parts of functions executed once are then optimized for
8367 Enabled by default at @option{-O} and higher.
8371 Perform interprocedural constant propagation.
8372 This optimization analyzes the program to determine when values passed
8373 to functions are constants and then optimizes accordingly.
8374 This optimization can substantially increase performance
8375 if the application has constants passed to functions.
8376 This flag is enabled by default at @option{-O2}, @option{-Os} and @option{-O3}.
8378 @item -fipa-cp-clone
8379 @opindex fipa-cp-clone
8380 Perform function cloning to make interprocedural constant propagation stronger.
8381 When enabled, interprocedural constant propagation performs function cloning
8382 when externally visible function can be called with constant arguments.
8383 Because this optimization can create multiple copies of functions,
8384 it may significantly increase code size
8385 (see @option{--param ipcp-unit-growth=@var{value}}).
8386 This flag is enabled by default at @option{-O3}.
8389 @opindex -fipa-bit-cp
8390 When enabled, perform interprocedural bitwise constant
8391 propagation. This flag is enabled by default at @option{-O2}. It
8392 requires that @option{-fipa-cp} is enabled.
8396 When enabled, perform interprocedural propagation of value
8397 ranges. This flag is enabled by default at @option{-O2}. It requires
8398 that @option{-fipa-cp} is enabled.
8402 Perform Identical Code Folding for functions and read-only variables.
8403 The optimization reduces code size and may disturb unwind stacks by replacing
8404 a function by equivalent one with a different name. The optimization works
8405 more effectively with link-time optimization enabled.
8407 Nevertheless the behavior is similar to Gold Linker ICF optimization, GCC ICF
8408 works on different levels and thus the optimizations are not same - there are
8409 equivalences that are found only by GCC and equivalences found only by Gold.
8411 This flag is enabled by default at @option{-O2} and @option{-Os}.
8413 @item -fisolate-erroneous-paths-dereference
8414 @opindex fisolate-erroneous-paths-dereference
8415 Detect paths that trigger erroneous or undefined behavior due to
8416 dereferencing a null pointer. Isolate those paths from the main control
8417 flow and turn the statement with erroneous or undefined behavior into a trap.
8418 This flag is enabled by default at @option{-O2} and higher and depends on
8419 @option{-fdelete-null-pointer-checks} also being enabled.
8421 @item -fisolate-erroneous-paths-attribute
8422 @opindex fisolate-erroneous-paths-attribute
8423 Detect paths that trigger erroneous or undefined behavior due a null value
8424 being used in a way forbidden by a @code{returns_nonnull} or @code{nonnull}
8425 attribute. Isolate those paths from the main control flow and turn the
8426 statement with erroneous or undefined behavior into a trap. This is not
8427 currently enabled, but may be enabled by @option{-O2} in the future.
8431 Perform forward store motion on trees. This flag is
8432 enabled by default at @option{-O} and higher.
8434 @item -ftree-bit-ccp
8435 @opindex ftree-bit-ccp
8436 Perform sparse conditional bit constant propagation on trees and propagate
8437 pointer alignment information.
8438 This pass only operates on local scalar variables and is enabled by default
8439 at @option{-O} and higher. It requires that @option{-ftree-ccp} is enabled.
8443 Perform sparse conditional constant propagation (CCP) on trees. This
8444 pass only operates on local scalar variables and is enabled by default
8445 at @option{-O} and higher.
8447 @item -fssa-backprop
8448 @opindex fssa-backprop
8449 Propagate information about uses of a value up the definition chain
8450 in order to simplify the definitions. For example, this pass strips
8451 sign operations if the sign of a value never matters. The flag is
8452 enabled by default at @option{-O} and higher.
8455 @opindex fssa-phiopt
8456 Perform pattern matching on SSA PHI nodes to optimize conditional
8457 code. This pass is enabled by default at @option{-O} and higher.
8459 @item -ftree-switch-conversion
8460 @opindex ftree-switch-conversion
8461 Perform conversion of simple initializations in a switch to
8462 initializations from a scalar array. This flag is enabled by default
8463 at @option{-O2} and higher.
8465 @item -ftree-tail-merge
8466 @opindex ftree-tail-merge
8467 Look for identical code sequences. When found, replace one with a jump to the
8468 other. This optimization is known as tail merging or cross jumping. This flag
8469 is enabled by default at @option{-O2} and higher. The compilation time
8471 be limited using @option{max-tail-merge-comparisons} parameter and
8472 @option{max-tail-merge-iterations} parameter.
8476 Perform dead code elimination (DCE) on trees. This flag is enabled by
8477 default at @option{-O} and higher.
8479 @item -ftree-builtin-call-dce
8480 @opindex ftree-builtin-call-dce
8481 Perform conditional dead code elimination (DCE) for calls to built-in functions
8482 that may set @code{errno} but are otherwise side-effect free. This flag is
8483 enabled by default at @option{-O2} and higher if @option{-Os} is not also
8486 @item -ftree-dominator-opts
8487 @opindex ftree-dominator-opts
8488 Perform a variety of simple scalar cleanups (constant/copy
8489 propagation, redundancy elimination, range propagation and expression
8490 simplification) based on a dominator tree traversal. This also
8491 performs jump threading (to reduce jumps to jumps). This flag is
8492 enabled by default at @option{-O} and higher.
8496 Perform dead store elimination (DSE) on trees. A dead store is a store into
8497 a memory location that is later overwritten by another store without
8498 any intervening loads. In this case the earlier store can be deleted. This
8499 flag is enabled by default at @option{-O} and higher.
8503 Perform loop header copying on trees. This is beneficial since it increases
8504 effectiveness of code motion optimizations. It also saves one jump. This flag
8505 is enabled by default at @option{-O} and higher. It is not enabled
8506 for @option{-Os}, since it usually increases code size.
8508 @item -ftree-loop-optimize
8509 @opindex ftree-loop-optimize
8510 Perform loop optimizations on trees. This flag is enabled by default
8511 at @option{-O} and higher.
8513 @item -ftree-loop-linear
8514 @itemx -floop-interchange
8515 @itemx -floop-strip-mine
8517 @itemx -floop-unroll-and-jam
8518 @opindex ftree-loop-linear
8519 @opindex floop-interchange
8520 @opindex floop-strip-mine
8521 @opindex floop-block
8522 @opindex floop-unroll-and-jam
8523 Perform loop nest optimizations. Same as
8524 @option{-floop-nest-optimize}. To use this code transformation, GCC has
8525 to be configured with @option{--with-isl} to enable the Graphite loop
8526 transformation infrastructure.
8528 @item -fgraphite-identity
8529 @opindex fgraphite-identity
8530 Enable the identity transformation for graphite. For every SCoP we generate
8531 the polyhedral representation and transform it back to gimple. Using
8532 @option{-fgraphite-identity} we can check the costs or benefits of the
8533 GIMPLE -> GRAPHITE -> GIMPLE transformation. Some minimal optimizations
8534 are also performed by the code generator isl, like index splitting and
8535 dead code elimination in loops.
8537 @item -floop-nest-optimize
8538 @opindex floop-nest-optimize
8539 Enable the isl based loop nest optimizer. This is a generic loop nest
8540 optimizer based on the Pluto optimization algorithms. It calculates a loop
8541 structure optimized for data-locality and parallelism. This option
8544 @item -floop-parallelize-all
8545 @opindex floop-parallelize-all
8546 Use the Graphite data dependence analysis to identify loops that can
8547 be parallelized. Parallelize all the loops that can be analyzed to
8548 not contain loop carried dependences without checking that it is
8549 profitable to parallelize the loops.
8551 @item -ftree-coalesce-vars
8552 @opindex ftree-coalesce-vars
8553 While transforming the program out of the SSA representation, attempt to
8554 reduce copying by coalescing versions of different user-defined
8555 variables, instead of just compiler temporaries. This may severely
8556 limit the ability to debug an optimized program compiled with
8557 @option{-fno-var-tracking-assignments}. In the negated form, this flag
8558 prevents SSA coalescing of user variables. This option is enabled by
8559 default if optimization is enabled, and it does very little otherwise.
8561 @item -ftree-loop-if-convert
8562 @opindex ftree-loop-if-convert
8563 Attempt to transform conditional jumps in the innermost loops to
8564 branch-less equivalents. The intent is to remove control-flow from
8565 the innermost loops in order to improve the ability of the
8566 vectorization pass to handle these loops. This is enabled by default
8567 if vectorization is enabled.
8569 @item -ftree-loop-distribution
8570 @opindex ftree-loop-distribution
8571 Perform loop distribution. This flag can improve cache performance on
8572 big loop bodies and allow further loop optimizations, like
8573 parallelization or vectorization, to take place. For example, the loop
8590 @item -ftree-loop-distribute-patterns
8591 @opindex ftree-loop-distribute-patterns
8592 Perform loop distribution of patterns that can be code generated with
8593 calls to a library. This flag is enabled by default at @option{-O3}.
8595 This pass distributes the initialization loops and generates a call to
8596 memset zero. For example, the loop
8612 and the initialization loop is transformed into a call to memset zero.
8614 @item -ftree-loop-im
8615 @opindex ftree-loop-im
8616 Perform loop invariant motion on trees. This pass moves only invariants that
8617 are hard to handle at RTL level (function calls, operations that expand to
8618 nontrivial sequences of insns). With @option{-funswitch-loops} it also moves
8619 operands of conditions that are invariant out of the loop, so that we can use
8620 just trivial invariantness analysis in loop unswitching. The pass also includes
8623 @item -ftree-loop-ivcanon
8624 @opindex ftree-loop-ivcanon
8625 Create a canonical counter for number of iterations in loops for which
8626 determining number of iterations requires complicated analysis. Later
8627 optimizations then may determine the number easily. Useful especially
8628 in connection with unrolling.
8632 Perform induction variable optimizations (strength reduction, induction
8633 variable merging and induction variable elimination) on trees.
8635 @item -ftree-parallelize-loops=n
8636 @opindex ftree-parallelize-loops
8637 Parallelize loops, i.e., split their iteration space to run in n threads.
8638 This is only possible for loops whose iterations are independent
8639 and can be arbitrarily reordered. The optimization is only
8640 profitable on multiprocessor machines, for loops that are CPU-intensive,
8641 rather than constrained e.g.@: by memory bandwidth. This option
8642 implies @option{-pthread}, and thus is only supported on targets
8643 that have support for @option{-pthread}.
8647 Perform function-local points-to analysis on trees. This flag is
8648 enabled by default at @option{-O} and higher.
8652 Perform scalar replacement of aggregates. This pass replaces structure
8653 references with scalars to prevent committing structures to memory too
8654 early. This flag is enabled by default at @option{-O} and higher.
8656 @item -fstore-merging
8657 @opindex fstore-merging
8658 Perform merging of narrow stores to consecutive memory addresses. This pass
8659 merges contiguous stores of immediate values narrower than a word into fewer
8660 wider stores to reduce the number of instructions. This is enabled by default
8661 at @option{-O2} and higher as well as @option{-Os}.
8665 Perform temporary expression replacement during the SSA->normal phase. Single
8666 use/single def temporaries are replaced at their use location with their
8667 defining expression. This results in non-GIMPLE code, but gives the expanders
8668 much more complex trees to work on resulting in better RTL generation. This is
8669 enabled by default at @option{-O} and higher.
8673 Perform straight-line strength reduction on trees. This recognizes related
8674 expressions involving multiplications and replaces them by less expensive
8675 calculations when possible. This is enabled by default at @option{-O} and
8678 @item -ftree-vectorize
8679 @opindex ftree-vectorize
8680 Perform vectorization on trees. This flag enables @option{-ftree-loop-vectorize}
8681 and @option{-ftree-slp-vectorize} if not explicitly specified.
8683 @item -ftree-loop-vectorize
8684 @opindex ftree-loop-vectorize
8685 Perform loop vectorization on trees. This flag is enabled by default at
8686 @option{-O3} and when @option{-ftree-vectorize} is enabled.
8688 @item -ftree-slp-vectorize
8689 @opindex ftree-slp-vectorize
8690 Perform basic block vectorization on trees. This flag is enabled by default at
8691 @option{-O3} and when @option{-ftree-vectorize} is enabled.
8693 @item -fvect-cost-model=@var{model}
8694 @opindex fvect-cost-model
8695 Alter the cost model used for vectorization. The @var{model} argument
8696 should be one of @samp{unlimited}, @samp{dynamic} or @samp{cheap}.
8697 With the @samp{unlimited} model the vectorized code-path is assumed
8698 to be profitable while with the @samp{dynamic} model a runtime check
8699 guards the vectorized code-path to enable it only for iteration
8700 counts that will likely execute faster than when executing the original
8701 scalar loop. The @samp{cheap} model disables vectorization of
8702 loops where doing so would be cost prohibitive for example due to
8703 required runtime checks for data dependence or alignment but otherwise
8704 is equal to the @samp{dynamic} model.
8705 The default cost model depends on other optimization flags and is
8706 either @samp{dynamic} or @samp{cheap}.
8708 @item -fsimd-cost-model=@var{model}
8709 @opindex fsimd-cost-model
8710 Alter the cost model used for vectorization of loops marked with the OpenMP
8711 or Cilk Plus simd directive. The @var{model} argument should be one of
8712 @samp{unlimited}, @samp{dynamic}, @samp{cheap}. All values of @var{model}
8713 have the same meaning as described in @option{-fvect-cost-model} and by
8714 default a cost model defined with @option{-fvect-cost-model} is used.
8718 Perform Value Range Propagation on trees. This is similar to the
8719 constant propagation pass, but instead of values, ranges of values are
8720 propagated. This allows the optimizers to remove unnecessary range
8721 checks like array bound checks and null pointer checks. This is
8722 enabled by default at @option{-O2} and higher. Null pointer check
8723 elimination is only done if @option{-fdelete-null-pointer-checks} is
8727 @opindex fsplit-paths
8728 Split paths leading to loop backedges. This can improve dead code
8729 elimination and common subexpression elimination. This is enabled by
8730 default at @option{-O2} and above.
8732 @item -fsplit-ivs-in-unroller
8733 @opindex fsplit-ivs-in-unroller
8734 Enables expression of values of induction variables in later iterations
8735 of the unrolled loop using the value in the first iteration. This breaks
8736 long dependency chains, thus improving efficiency of the scheduling passes.
8738 A combination of @option{-fweb} and CSE is often sufficient to obtain the
8739 same effect. However, that is not reliable in cases where the loop body
8740 is more complicated than a single basic block. It also does not work at all
8741 on some architectures due to restrictions in the CSE pass.
8743 This optimization is enabled by default.
8745 @item -fvariable-expansion-in-unroller
8746 @opindex fvariable-expansion-in-unroller
8747 With this option, the compiler creates multiple copies of some
8748 local variables when unrolling a loop, which can result in superior code.
8750 @item -fpartial-inlining
8751 @opindex fpartial-inlining
8752 Inline parts of functions. This option has any effect only
8753 when inlining itself is turned on by the @option{-finline-functions}
8754 or @option{-finline-small-functions} options.
8756 Enabled at level @option{-O2}.
8758 @item -fpredictive-commoning
8759 @opindex fpredictive-commoning
8760 Perform predictive commoning optimization, i.e., reusing computations
8761 (especially memory loads and stores) performed in previous
8762 iterations of loops.
8764 This option is enabled at level @option{-O3}.
8766 @item -fprefetch-loop-arrays
8767 @opindex fprefetch-loop-arrays
8768 If supported by the target machine, generate instructions to prefetch
8769 memory to improve the performance of loops that access large arrays.
8771 This option may generate better or worse code; results are highly
8772 dependent on the structure of loops within the source code.
8774 Disabled at level @option{-Os}.
8776 @item -fno-printf-return-value
8777 @opindex fno-printf-return-value
8778 Do not substitute constants for known return value of formatted output
8779 functions such as @code{sprintf}, @code{snprintf}, @code{vsprintf}, and
8780 @code{vsnprintf} (but not @code{printf} of @code{fprintf}). This
8781 transformation allows GCC to optimize or even eliminate branches based
8782 on the known return value of these functions called with arguments that
8783 are either constant, or whose values are known to be in a range that
8784 makes determining the exact return value possible. For example, when
8785 @option{-fprintf-return-value} is in effect, both the branch and the
8786 body of the @code{if} statement (but not the call to @code{snprint})
8787 can be optimized away when @code{i} is a 32-bit or smaller integer
8788 because the return value is guaranteed to be at most 8.
8792 if (snprintf (buf, "%08x", i) >= sizeof buf)
8796 The @option{-fprintf-return-value} option relies on other optimizations
8797 and yields best results with @option{-O2}. It works in tandem with the
8798 @option{-Wformat-overflow} and @option{-Wformat-truncation} options.
8799 The @option{-fprintf-return-value} option is enabled by default.
8802 @itemx -fno-peephole2
8803 @opindex fno-peephole
8804 @opindex fno-peephole2
8805 Disable any machine-specific peephole optimizations. The difference
8806 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
8807 are implemented in the compiler; some targets use one, some use the
8808 other, a few use both.
8810 @option{-fpeephole} is enabled by default.
8811 @option{-fpeephole2} enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
8813 @item -fno-guess-branch-probability
8814 @opindex fno-guess-branch-probability
8815 Do not guess branch probabilities using heuristics.
8817 GCC uses heuristics to guess branch probabilities if they are
8818 not provided by profiling feedback (@option{-fprofile-arcs}). These
8819 heuristics are based on the control flow graph. If some branch probabilities
8820 are specified by @code{__builtin_expect}, then the heuristics are
8821 used to guess branch probabilities for the rest of the control flow graph,
8822 taking the @code{__builtin_expect} info into account. The interactions
8823 between the heuristics and @code{__builtin_expect} can be complex, and in
8824 some cases, it may be useful to disable the heuristics so that the effects
8825 of @code{__builtin_expect} are easier to understand.
8827 The default is @option{-fguess-branch-probability} at levels
8828 @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
8830 @item -freorder-blocks
8831 @opindex freorder-blocks
8832 Reorder basic blocks in the compiled function in order to reduce number of
8833 taken branches and improve code locality.
8835 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
8837 @item -freorder-blocks-algorithm=@var{algorithm}
8838 @opindex freorder-blocks-algorithm
8839 Use the specified algorithm for basic block reordering. The
8840 @var{algorithm} argument can be @samp{simple}, which does not increase
8841 code size (except sometimes due to secondary effects like alignment),
8842 or @samp{stc}, the ``software trace cache'' algorithm, which tries to
8843 put all often executed code together, minimizing the number of branches
8844 executed by making extra copies of code.
8846 The default is @samp{simple} at levels @option{-O}, @option{-Os}, and
8847 @samp{stc} at levels @option{-O2}, @option{-O3}.
8849 @item -freorder-blocks-and-partition
8850 @opindex freorder-blocks-and-partition
8851 In addition to reordering basic blocks in the compiled function, in order
8852 to reduce number of taken branches, partitions hot and cold basic blocks
8853 into separate sections of the assembly and @file{.o} files, to improve
8854 paging and cache locality performance.
8856 This optimization is automatically turned off in the presence of
8857 exception handling or unwind tables (on targets using setjump/longjump or target specific scheme), for linkonce sections, for functions with a user-defined
8858 section attribute and on any architecture that does not support named
8859 sections. When @option{-fsplit-stack} is used this option is not
8860 enabled by default (to avoid linker errors), but may be enabled
8861 explicitly (if using a working linker).
8863 Enabled for x86 at levels @option{-O2}, @option{-O3}.
8865 @item -freorder-functions
8866 @opindex freorder-functions
8867 Reorder functions in the object file in order to
8868 improve code locality. This is implemented by using special
8869 subsections @code{.text.hot} for most frequently executed functions and
8870 @code{.text.unlikely} for unlikely executed functions. Reordering is done by
8871 the linker so object file format must support named sections and linker must
8872 place them in a reasonable way.
8874 Also profile feedback must be available to make this option effective. See
8875 @option{-fprofile-arcs} for details.
8877 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
8879 @item -fstrict-aliasing
8880 @opindex fstrict-aliasing
8881 Allow the compiler to assume the strictest aliasing rules applicable to
8882 the language being compiled. For C (and C++), this activates
8883 optimizations based on the type of expressions. In particular, an
8884 object of one type is assumed never to reside at the same address as an
8885 object of a different type, unless the types are almost the same. For
8886 example, an @code{unsigned int} can alias an @code{int}, but not a
8887 @code{void*} or a @code{double}. A character type may alias any other
8890 @anchor{Type-punning}Pay special attention to code like this:
8903 The practice of reading from a different union member than the one most
8904 recently written to (called ``type-punning'') is common. Even with
8905 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
8906 is accessed through the union type. So, the code above works as
8907 expected. @xref{Structures unions enumerations and bit-fields
8908 implementation}. However, this code might not:
8919 Similarly, access by taking the address, casting the resulting pointer
8920 and dereferencing the result has undefined behavior, even if the cast
8921 uses a union type, e.g.:
8925 return ((union a_union *) &d)->i;
8929 The @option{-fstrict-aliasing} option is enabled at levels
8930 @option{-O2}, @option{-O3}, @option{-Os}.
8932 @item -falign-functions
8933 @itemx -falign-functions=@var{n}
8934 @opindex falign-functions
8935 Align the start of functions to the next power-of-two greater than
8936 @var{n}, skipping up to @var{n} bytes. For instance,
8937 @option{-falign-functions=32} aligns functions to the next 32-byte
8938 boundary, but @option{-falign-functions=24} aligns to the next
8939 32-byte boundary only if this can be done by skipping 23 bytes or less.
8941 @option{-fno-align-functions} and @option{-falign-functions=1} are
8942 equivalent and mean that functions are not aligned.
8944 Some assemblers only support this flag when @var{n} is a power of two;
8945 in that case, it is rounded up.
8947 If @var{n} is not specified or is zero, use a machine-dependent default.
8949 Enabled at levels @option{-O2}, @option{-O3}.
8951 @item -flimit-function-alignment
8952 If this option is enabled, the compiler tries to avoid unnecessarily
8953 overaligning functions. It attempts to instruct the assembler to align
8954 by the amount specified by @option{-falign-functions}, but not to
8955 skip more bytes than the size of the function.
8957 @item -falign-labels
8958 @itemx -falign-labels=@var{n}
8959 @opindex falign-labels
8960 Align all branch targets to a power-of-two boundary, skipping up to
8961 @var{n} bytes like @option{-falign-functions}. This option can easily
8962 make code slower, because it must insert dummy operations for when the
8963 branch target is reached in the usual flow of the code.
8965 @option{-fno-align-labels} and @option{-falign-labels=1} are
8966 equivalent and mean that labels are not aligned.
8968 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
8969 are greater than this value, then their values are used instead.
8971 If @var{n} is not specified or is zero, use a machine-dependent default
8972 which is very likely to be @samp{1}, meaning no alignment.
8974 Enabled at levels @option{-O2}, @option{-O3}.
8977 @itemx -falign-loops=@var{n}
8978 @opindex falign-loops
8979 Align loops to a power-of-two boundary, skipping up to @var{n} bytes
8980 like @option{-falign-functions}. If the loops are
8981 executed many times, this makes up for any execution of the dummy
8984 @option{-fno-align-loops} and @option{-falign-loops=1} are
8985 equivalent and mean that loops are not aligned.
8987 If @var{n} is not specified or is zero, use a machine-dependent default.
8989 Enabled at levels @option{-O2}, @option{-O3}.
8992 @itemx -falign-jumps=@var{n}
8993 @opindex falign-jumps
8994 Align branch targets to a power-of-two boundary, for branch targets
8995 where the targets can only be reached by jumping, skipping up to @var{n}
8996 bytes like @option{-falign-functions}. In this case, no dummy operations
8999 @option{-fno-align-jumps} and @option{-falign-jumps=1} are
9000 equivalent and mean that loops are not aligned.
9002 If @var{n} is not specified or is zero, use a machine-dependent default.
9004 Enabled at levels @option{-O2}, @option{-O3}.
9006 @item -funit-at-a-time
9007 @opindex funit-at-a-time
9008 This option is left for compatibility reasons. @option{-funit-at-a-time}
9009 has no effect, while @option{-fno-unit-at-a-time} implies
9010 @option{-fno-toplevel-reorder} and @option{-fno-section-anchors}.
9014 @item -fno-toplevel-reorder
9015 @opindex fno-toplevel-reorder
9016 Do not reorder top-level functions, variables, and @code{asm}
9017 statements. Output them in the same order that they appear in the
9018 input file. When this option is used, unreferenced static variables
9019 are not removed. This option is intended to support existing code
9020 that relies on a particular ordering. For new code, it is better to
9021 use attributes when possible.
9023 Enabled at level @option{-O0}. When disabled explicitly, it also implies
9024 @option{-fno-section-anchors}, which is otherwise enabled at @option{-O0} on some
9029 Constructs webs as commonly used for register allocation purposes and assign
9030 each web individual pseudo register. This allows the register allocation pass
9031 to operate on pseudos directly, but also strengthens several other optimization
9032 passes, such as CSE, loop optimizer and trivial dead code remover. It can,
9033 however, make debugging impossible, since variables no longer stay in a
9036 Enabled by default with @option{-funroll-loops}.
9038 @item -fwhole-program
9039 @opindex fwhole-program
9040 Assume that the current compilation unit represents the whole program being
9041 compiled. All public functions and variables with the exception of @code{main}
9042 and those merged by attribute @code{externally_visible} become static functions
9043 and in effect are optimized more aggressively by interprocedural optimizers.
9045 This option should not be used in combination with @option{-flto}.
9046 Instead relying on a linker plugin should provide safer and more precise
9049 @item -flto[=@var{n}]
9051 This option runs the standard link-time optimizer. When invoked
9052 with source code, it generates GIMPLE (one of GCC's internal
9053 representations) and writes it to special ELF sections in the object
9054 file. When the object files are linked together, all the function
9055 bodies are read from these ELF sections and instantiated as if they
9056 had been part of the same translation unit.
9058 To use the link-time optimizer, @option{-flto} and optimization
9059 options should be specified at compile time and during the final link.
9060 It is recommended that you compile all the files participating in the
9061 same link with the same options and also specify those options at
9066 gcc -c -O2 -flto foo.c
9067 gcc -c -O2 -flto bar.c
9068 gcc -o myprog -flto -O2 foo.o bar.o
9071 The first two invocations to GCC save a bytecode representation
9072 of GIMPLE into special ELF sections inside @file{foo.o} and
9073 @file{bar.o}. The final invocation reads the GIMPLE bytecode from
9074 @file{foo.o} and @file{bar.o}, merges the two files into a single
9075 internal image, and compiles the result as usual. Since both
9076 @file{foo.o} and @file{bar.o} are merged into a single image, this
9077 causes all the interprocedural analyses and optimizations in GCC to
9078 work across the two files as if they were a single one. This means,
9079 for example, that the inliner is able to inline functions in
9080 @file{bar.o} into functions in @file{foo.o} and vice-versa.
9082 Another (simpler) way to enable link-time optimization is:
9085 gcc -o myprog -flto -O2 foo.c bar.c
9088 The above generates bytecode for @file{foo.c} and @file{bar.c},
9089 merges them together into a single GIMPLE representation and optimizes
9090 them as usual to produce @file{myprog}.
9092 The only important thing to keep in mind is that to enable link-time
9093 optimizations you need to use the GCC driver to perform the link step.
9094 GCC then automatically performs link-time optimization if any of the
9095 objects involved were compiled with the @option{-flto} command-line option.
9097 should specify the optimization options to be used for link-time
9098 optimization though GCC tries to be clever at guessing an
9099 optimization level to use from the options used at compile time
9100 if you fail to specify one at link time. You can always override
9101 the automatic decision to do link-time optimization
9102 by passing @option{-fno-lto} to the link command.
9104 To make whole program optimization effective, it is necessary to make
9105 certain whole program assumptions. The compiler needs to know
9106 what functions and variables can be accessed by libraries and runtime
9107 outside of the link-time optimized unit. When supported by the linker,
9108 the linker plugin (see @option{-fuse-linker-plugin}) passes information
9109 to the compiler about used and externally visible symbols. When
9110 the linker plugin is not available, @option{-fwhole-program} should be
9111 used to allow the compiler to make these assumptions, which leads
9112 to more aggressive optimization decisions.
9114 When @option{-fuse-linker-plugin} is not enabled, when a file is
9115 compiled with @option{-flto}, the generated object file is larger than
9116 a regular object file because it contains GIMPLE bytecodes and the usual
9117 final code (see @option{-ffat-lto-objects}. This means that
9118 object files with LTO information can be linked as normal object
9119 files; if @option{-fno-lto} is passed to the linker, no
9120 interprocedural optimizations are applied. Note that when
9121 @option{-fno-fat-lto-objects} is enabled the compile stage is faster
9122 but you cannot perform a regular, non-LTO link on them.
9124 Additionally, the optimization flags used to compile individual files
9125 are not necessarily related to those used at link time. For instance,
9128 gcc -c -O0 -ffat-lto-objects -flto foo.c
9129 gcc -c -O0 -ffat-lto-objects -flto bar.c
9130 gcc -o myprog -O3 foo.o bar.o
9133 This produces individual object files with unoptimized assembler
9134 code, but the resulting binary @file{myprog} is optimized at
9135 @option{-O3}. If, instead, the final binary is generated with
9136 @option{-fno-lto}, then @file{myprog} is not optimized.
9138 When producing the final binary, GCC only
9139 applies link-time optimizations to those files that contain bytecode.
9140 Therefore, you can mix and match object files and libraries with
9141 GIMPLE bytecodes and final object code. GCC automatically selects
9142 which files to optimize in LTO mode and which files to link without
9145 There are some code generation flags preserved by GCC when
9146 generating bytecodes, as they need to be used during the final link
9147 stage. Generally options specified at link time override those
9148 specified at compile time.
9150 If you do not specify an optimization level option @option{-O} at
9151 link time, then GCC uses the highest optimization level
9152 used when compiling the object files.
9154 Currently, the following options and their settings are taken from
9155 the first object file that explicitly specifies them:
9156 @option{-fPIC}, @option{-fpic}, @option{-fpie}, @option{-fcommon},
9157 @option{-fexceptions}, @option{-fnon-call-exceptions}, @option{-fgnu-tm}
9158 and all the @option{-m} target flags.
9160 Certain ABI-changing flags are required to match in all compilation units,
9161 and trying to override this at link time with a conflicting value
9162 is ignored. This includes options such as @option{-freg-struct-return}
9163 and @option{-fpcc-struct-return}.
9165 Other options such as @option{-ffp-contract}, @option{-fno-strict-overflow},
9166 @option{-fwrapv}, @option{-fno-trapv} or @option{-fno-strict-aliasing}
9167 are passed through to the link stage and merged conservatively for
9168 conflicting translation units. Specifically
9169 @option{-fno-strict-overflow}, @option{-fwrapv} and @option{-fno-trapv} take
9170 precedence; and for example @option{-ffp-contract=off} takes precedence
9171 over @option{-ffp-contract=fast}. You can override them at link time.
9173 If LTO encounters objects with C linkage declared with incompatible
9174 types in separate translation units to be linked together (undefined
9175 behavior according to ISO C99 6.2.7), a non-fatal diagnostic may be
9176 issued. The behavior is still undefined at run time. Similar
9177 diagnostics may be raised for other languages.
9179 Another feature of LTO is that it is possible to apply interprocedural
9180 optimizations on files written in different languages:
9185 gfortran -c -flto baz.f90
9186 g++ -o myprog -flto -O3 foo.o bar.o baz.o -lgfortran
9189 Notice that the final link is done with @command{g++} to get the C++
9190 runtime libraries and @option{-lgfortran} is added to get the Fortran
9191 runtime libraries. In general, when mixing languages in LTO mode, you
9192 should use the same link command options as when mixing languages in a
9193 regular (non-LTO) compilation.
9195 If object files containing GIMPLE bytecode are stored in a library archive, say
9196 @file{libfoo.a}, it is possible to extract and use them in an LTO link if you
9197 are using a linker with plugin support. To create static libraries suitable
9198 for LTO, use @command{gcc-ar} and @command{gcc-ranlib} instead of @command{ar}
9199 and @command{ranlib};
9200 to show the symbols of object files with GIMPLE bytecode, use
9201 @command{gcc-nm}. Those commands require that @command{ar}, @command{ranlib}
9202 and @command{nm} have been compiled with plugin support. At link time, use the the
9203 flag @option{-fuse-linker-plugin} to ensure that the library participates in
9204 the LTO optimization process:
9207 gcc -o myprog -O2 -flto -fuse-linker-plugin a.o b.o -lfoo
9210 With the linker plugin enabled, the linker extracts the needed
9211 GIMPLE files from @file{libfoo.a} and passes them on to the running GCC
9212 to make them part of the aggregated GIMPLE image to be optimized.
9214 If you are not using a linker with plugin support and/or do not
9215 enable the linker plugin, then the objects inside @file{libfoo.a}
9216 are extracted and linked as usual, but they do not participate
9217 in the LTO optimization process. In order to make a static library suitable
9218 for both LTO optimization and usual linkage, compile its object files with
9219 @option{-flto} @option{-ffat-lto-objects}.
9221 Link-time optimizations do not require the presence of the whole program to
9222 operate. If the program does not require any symbols to be exported, it is
9223 possible to combine @option{-flto} and @option{-fwhole-program} to allow
9224 the interprocedural optimizers to use more aggressive assumptions which may
9225 lead to improved optimization opportunities.
9226 Use of @option{-fwhole-program} is not needed when linker plugin is
9227 active (see @option{-fuse-linker-plugin}).
9229 The current implementation of LTO makes no
9230 attempt to generate bytecode that is portable between different
9231 types of hosts. The bytecode files are versioned and there is a
9232 strict version check, so bytecode files generated in one version of
9233 GCC do not work with an older or newer version of GCC.
9235 Link-time optimization does not work well with generation of debugging
9236 information. Combining @option{-flto} with
9237 @option{-g} is currently experimental and expected to produce unexpected
9240 If you specify the optional @var{n}, the optimization and code
9241 generation done at link time is executed in parallel using @var{n}
9242 parallel jobs by utilizing an installed @command{make} program. The
9243 environment variable @env{MAKE} may be used to override the program
9244 used. The default value for @var{n} is 1.
9246 You can also specify @option{-flto=jobserver} to use GNU make's
9247 job server mode to determine the number of parallel jobs. This
9248 is useful when the Makefile calling GCC is already executing in parallel.
9249 You must prepend a @samp{+} to the command recipe in the parent Makefile
9250 for this to work. This option likely only works if @env{MAKE} is
9253 @item -flto-partition=@var{alg}
9254 @opindex flto-partition
9255 Specify the partitioning algorithm used by the link-time optimizer.
9256 The value is either @samp{1to1} to specify a partitioning mirroring
9257 the original source files or @samp{balanced} to specify partitioning
9258 into equally sized chunks (whenever possible) or @samp{max} to create
9259 new partition for every symbol where possible. Specifying @samp{none}
9260 as an algorithm disables partitioning and streaming completely.
9261 The default value is @samp{balanced}. While @samp{1to1} can be used
9262 as an workaround for various code ordering issues, the @samp{max}
9263 partitioning is intended for internal testing only.
9264 The value @samp{one} specifies that exactly one partition should be
9265 used while the value @samp{none} bypasses partitioning and executes
9266 the link-time optimization step directly from the WPA phase.
9268 @item -flto-odr-type-merging
9269 @opindex flto-odr-type-merging
9270 Enable streaming of mangled types names of C++ types and their unification
9271 at link time. This increases size of LTO object files, but enables
9272 diagnostics about One Definition Rule violations.
9274 @item -flto-compression-level=@var{n}
9275 @opindex flto-compression-level
9276 This option specifies the level of compression used for intermediate
9277 language written to LTO object files, and is only meaningful in
9278 conjunction with LTO mode (@option{-flto}). Valid
9279 values are 0 (no compression) to 9 (maximum compression). Values
9280 outside this range are clamped to either 0 or 9. If the option is not
9281 given, a default balanced compression setting is used.
9283 @item -fuse-linker-plugin
9284 @opindex fuse-linker-plugin
9285 Enables the use of a linker plugin during link-time optimization. This
9286 option relies on plugin support in the linker, which is available in gold
9287 or in GNU ld 2.21 or newer.
9289 This option enables the extraction of object files with GIMPLE bytecode out
9290 of library archives. This improves the quality of optimization by exposing
9291 more code to the link-time optimizer. This information specifies what
9292 symbols can be accessed externally (by non-LTO object or during dynamic
9293 linking). Resulting code quality improvements on binaries (and shared
9294 libraries that use hidden visibility) are similar to @option{-fwhole-program}.
9295 See @option{-flto} for a description of the effect of this flag and how to
9298 This option is enabled by default when LTO support in GCC is enabled
9299 and GCC was configured for use with
9300 a linker supporting plugins (GNU ld 2.21 or newer or gold).
9302 @item -ffat-lto-objects
9303 @opindex ffat-lto-objects
9304 Fat LTO objects are object files that contain both the intermediate language
9305 and the object code. This makes them usable for both LTO linking and normal
9306 linking. This option is effective only when compiling with @option{-flto}
9307 and is ignored at link time.
9309 @option{-fno-fat-lto-objects} improves compilation time over plain LTO, but
9310 requires the complete toolchain to be aware of LTO. It requires a linker with
9311 linker plugin support for basic functionality. Additionally,
9312 @command{nm}, @command{ar} and @command{ranlib}
9313 need to support linker plugins to allow a full-featured build environment
9314 (capable of building static libraries etc). GCC provides the @command{gcc-ar},
9315 @command{gcc-nm}, @command{gcc-ranlib} wrappers to pass the right options
9316 to these tools. With non fat LTO makefiles need to be modified to use them.
9318 The default is @option{-fno-fat-lto-objects} on targets with linker plugin
9321 @item -fcompare-elim
9322 @opindex fcompare-elim
9323 After register allocation and post-register allocation instruction splitting,
9324 identify arithmetic instructions that compute processor flags similar to a
9325 comparison operation based on that arithmetic. If possible, eliminate the
9326 explicit comparison operation.
9328 This pass only applies to certain targets that cannot explicitly represent
9329 the comparison operation before register allocation is complete.
9331 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
9333 @item -fcprop-registers
9334 @opindex fcprop-registers
9335 After register allocation and post-register allocation instruction splitting,
9336 perform a copy-propagation pass to try to reduce scheduling dependencies
9337 and occasionally eliminate the copy.
9339 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
9341 @item -fprofile-correction
9342 @opindex fprofile-correction
9343 Profiles collected using an instrumented binary for multi-threaded programs may
9344 be inconsistent due to missed counter updates. When this option is specified,
9345 GCC uses heuristics to correct or smooth out such inconsistencies. By
9346 default, GCC emits an error message when an inconsistent profile is detected.
9349 @itemx -fprofile-use=@var{path}
9350 @opindex fprofile-use
9351 Enable profile feedback-directed optimizations,
9352 and the following optimizations
9353 which are generally profitable only with profile feedback available:
9354 @option{-fbranch-probabilities}, @option{-fvpt},
9355 @option{-funroll-loops}, @option{-fpeel-loops}, @option{-ftracer},
9356 @option{-ftree-vectorize}, and @option{ftree-loop-distribute-patterns}.
9358 Before you can use this option, you must first generate profiling information.
9359 @xref{Instrumentation Options}, for information about the
9360 @option{-fprofile-generate} option.
9362 By default, GCC emits an error message if the feedback profiles do not
9363 match the source code. This error can be turned into a warning by using
9364 @option{-Wcoverage-mismatch}. Note this may result in poorly optimized
9367 If @var{path} is specified, GCC looks at the @var{path} to find
9368 the profile feedback data files. See @option{-fprofile-dir}.
9370 @item -fauto-profile
9371 @itemx -fauto-profile=@var{path}
9372 @opindex fauto-profile
9373 Enable sampling-based feedback-directed optimizations,
9374 and the following optimizations
9375 which are generally profitable only with profile feedback available:
9376 @option{-fbranch-probabilities}, @option{-fvpt},
9377 @option{-funroll-loops}, @option{-fpeel-loops}, @option{-ftracer},
9378 @option{-ftree-vectorize},
9379 @option{-finline-functions}, @option{-fipa-cp}, @option{-fipa-cp-clone},
9380 @option{-fpredictive-commoning}, @option{-funswitch-loops},
9381 @option{-fgcse-after-reload}, and @option{-ftree-loop-distribute-patterns}.
9383 @var{path} is the name of a file containing AutoFDO profile information.
9384 If omitted, it defaults to @file{fbdata.afdo} in the current directory.
9386 Producing an AutoFDO profile data file requires running your program
9387 with the @command{perf} utility on a supported GNU/Linux target system.
9388 For more information, see @uref{https://perf.wiki.kernel.org/}.
9392 perf record -e br_inst_retired:near_taken -b -o perf.data \
9396 Then use the @command{create_gcov} tool to convert the raw profile data
9397 to a format that can be used by GCC.@ You must also supply the
9398 unstripped binary for your program to this tool.
9399 See @uref{https://github.com/google/autofdo}.
9403 create_gcov --binary=your_program.unstripped --profile=perf.data \
9408 The following options control compiler behavior regarding floating-point
9409 arithmetic. These options trade off between speed and
9410 correctness. All must be specifically enabled.
9414 @opindex ffloat-store
9415 Do not store floating-point variables in registers, and inhibit other
9416 options that might change whether a floating-point value is taken from a
9419 @cindex floating-point precision
9420 This option prevents undesirable excess precision on machines such as
9421 the 68000 where the floating registers (of the 68881) keep more
9422 precision than a @code{double} is supposed to have. Similarly for the
9423 x86 architecture. For most programs, the excess precision does only
9424 good, but a few programs rely on the precise definition of IEEE floating
9425 point. Use @option{-ffloat-store} for such programs, after modifying
9426 them to store all pertinent intermediate computations into variables.
9428 @item -fexcess-precision=@var{style}
9429 @opindex fexcess-precision
9430 This option allows further control over excess precision on machines
9431 where floating-point operations occur in a format with more precision or
9432 range than the IEEE standard and interchange floating-point types. By
9433 default, @option{-fexcess-precision=fast} is in effect; this means that
9434 operations may be carried out in a wider precision than the types specified
9435 in the source if that would result in faster code, and it is unpredictable
9436 when rounding to the types specified in the source code takes place.
9437 When compiling C, if @option{-fexcess-precision=standard} is specified then
9438 excess precision follows the rules specified in ISO C99; in particular,
9439 both casts and assignments cause values to be rounded to their
9440 semantic types (whereas @option{-ffloat-store} only affects
9441 assignments). This option is enabled by default for C if a strict
9442 conformance option such as @option{-std=c99} is used.
9443 @option{-ffast-math} enables @option{-fexcess-precision=fast} by default
9444 regardless of whether a strict conformance option is used.
9447 @option{-fexcess-precision=standard} is not implemented for languages
9448 other than C. On the x86, it has no effect if @option{-mfpmath=sse}
9449 or @option{-mfpmath=sse+387} is specified; in the former case, IEEE
9450 semantics apply without excess precision, and in the latter, rounding
9455 Sets the options @option{-fno-math-errno}, @option{-funsafe-math-optimizations},
9456 @option{-ffinite-math-only}, @option{-fno-rounding-math},
9457 @option{-fno-signaling-nans}, @option{-fcx-limited-range} and
9458 @option{-fexcess-precision=fast}.
9460 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
9462 This option is not turned on by any @option{-O} option besides
9463 @option{-Ofast} since it can result in incorrect output for programs
9464 that depend on an exact implementation of IEEE or ISO rules/specifications
9465 for math functions. It may, however, yield faster code for programs
9466 that do not require the guarantees of these specifications.
9468 @item -fno-math-errno
9469 @opindex fno-math-errno
9470 Do not set @code{errno} after calling math functions that are executed
9471 with a single instruction, e.g., @code{sqrt}. A program that relies on
9472 IEEE exceptions for math error handling may want to use this flag
9473 for speed while maintaining IEEE arithmetic compatibility.
9475 This option is not turned on by any @option{-O} option since
9476 it can result in incorrect output for programs that depend on
9477 an exact implementation of IEEE or ISO rules/specifications for
9478 math functions. It may, however, yield faster code for programs
9479 that do not require the guarantees of these specifications.
9481 The default is @option{-fmath-errno}.
9483 On Darwin systems, the math library never sets @code{errno}. There is
9484 therefore no reason for the compiler to consider the possibility that
9485 it might, and @option{-fno-math-errno} is the default.
9487 @item -funsafe-math-optimizations
9488 @opindex funsafe-math-optimizations
9490 Allow optimizations for floating-point arithmetic that (a) assume
9491 that arguments and results are valid and (b) may violate IEEE or
9492 ANSI standards. When used at link time, it may include libraries
9493 or startup files that change the default FPU control word or other
9494 similar optimizations.
9496 This option is not turned on by any @option{-O} option since
9497 it can result in incorrect output for programs that depend on
9498 an exact implementation of IEEE or ISO rules/specifications for
9499 math functions. It may, however, yield faster code for programs
9500 that do not require the guarantees of these specifications.
9501 Enables @option{-fno-signed-zeros}, @option{-fno-trapping-math},
9502 @option{-fassociative-math} and @option{-freciprocal-math}.
9504 The default is @option{-fno-unsafe-math-optimizations}.
9506 @item -fassociative-math
9507 @opindex fassociative-math
9509 Allow re-association of operands in series of floating-point operations.
9510 This violates the ISO C and C++ language standard by possibly changing
9511 computation result. NOTE: re-ordering may change the sign of zero as
9512 well as ignore NaNs and inhibit or create underflow or overflow (and
9513 thus cannot be used on code that relies on rounding behavior like
9514 @code{(x + 2**52) - 2**52}. May also reorder floating-point comparisons
9515 and thus may not be used when ordered comparisons are required.
9516 This option requires that both @option{-fno-signed-zeros} and
9517 @option{-fno-trapping-math} be in effect. Moreover, it doesn't make
9518 much sense with @option{-frounding-math}. For Fortran the option
9519 is automatically enabled when both @option{-fno-signed-zeros} and
9520 @option{-fno-trapping-math} are in effect.
9522 The default is @option{-fno-associative-math}.
9524 @item -freciprocal-math
9525 @opindex freciprocal-math
9527 Allow the reciprocal of a value to be used instead of dividing by
9528 the value if this enables optimizations. For example @code{x / y}
9529 can be replaced with @code{x * (1/y)}, which is useful if @code{(1/y)}
9530 is subject to common subexpression elimination. Note that this loses
9531 precision and increases the number of flops operating on the value.
9533 The default is @option{-fno-reciprocal-math}.
9535 @item -ffinite-math-only
9536 @opindex ffinite-math-only
9537 Allow optimizations for floating-point arithmetic that assume
9538 that arguments and results are not NaNs or +-Infs.
9540 This option is not turned on by any @option{-O} option since
9541 it can result in incorrect output for programs that depend on
9542 an exact implementation of IEEE or ISO rules/specifications for
9543 math functions. It may, however, yield faster code for programs
9544 that do not require the guarantees of these specifications.
9546 The default is @option{-fno-finite-math-only}.
9548 @item -fno-signed-zeros
9549 @opindex fno-signed-zeros
9550 Allow optimizations for floating-point arithmetic that ignore the
9551 signedness of zero. IEEE arithmetic specifies the behavior of
9552 distinct +0.0 and @minus{}0.0 values, which then prohibits simplification
9553 of expressions such as x+0.0 or 0.0*x (even with @option{-ffinite-math-only}).
9554 This option implies that the sign of a zero result isn't significant.
9556 The default is @option{-fsigned-zeros}.
9558 @item -fno-trapping-math
9559 @opindex fno-trapping-math
9560 Compile code assuming that floating-point operations cannot generate
9561 user-visible traps. These traps include division by zero, overflow,
9562 underflow, inexact result and invalid operation. This option requires
9563 that @option{-fno-signaling-nans} be in effect. Setting this option may
9564 allow faster code if one relies on ``non-stop'' IEEE arithmetic, for example.
9566 This option should never be turned on by any @option{-O} option since
9567 it can result in incorrect output for programs that depend on
9568 an exact implementation of IEEE or ISO rules/specifications for
9571 The default is @option{-ftrapping-math}.
9573 @item -frounding-math
9574 @opindex frounding-math
9575 Disable transformations and optimizations that assume default floating-point
9576 rounding behavior. This is round-to-zero for all floating point
9577 to integer conversions, and round-to-nearest for all other arithmetic
9578 truncations. This option should be specified for programs that change
9579 the FP rounding mode dynamically, or that may be executed with a
9580 non-default rounding mode. This option disables constant folding of
9581 floating-point expressions at compile time (which may be affected by
9582 rounding mode) and arithmetic transformations that are unsafe in the
9583 presence of sign-dependent rounding modes.
9585 The default is @option{-fno-rounding-math}.
9587 This option is experimental and does not currently guarantee to
9588 disable all GCC optimizations that are affected by rounding mode.
9589 Future versions of GCC may provide finer control of this setting
9590 using C99's @code{FENV_ACCESS} pragma. This command-line option
9591 will be used to specify the default state for @code{FENV_ACCESS}.
9593 @item -fsignaling-nans
9594 @opindex fsignaling-nans
9595 Compile code assuming that IEEE signaling NaNs may generate user-visible
9596 traps during floating-point operations. Setting this option disables
9597 optimizations that may change the number of exceptions visible with
9598 signaling NaNs. This option implies @option{-ftrapping-math}.
9600 This option causes the preprocessor macro @code{__SUPPORT_SNAN__} to
9603 The default is @option{-fno-signaling-nans}.
9605 This option is experimental and does not currently guarantee to
9606 disable all GCC optimizations that affect signaling NaN behavior.
9608 @item -fno-fp-int-builtin-inexact
9609 @opindex fno-fp-int-builtin-inexact
9610 Do not allow the built-in functions @code{ceil}, @code{floor},
9611 @code{round} and @code{trunc}, and their @code{float} and @code{long
9612 double} variants, to generate code that raises the ``inexact''
9613 floating-point exception for noninteger arguments. ISO C99 and C11
9614 allow these functions to raise the ``inexact'' exception, but ISO/IEC
9615 TS 18661-1:2014, the C bindings to IEEE 754-2008, does not allow these
9618 The default is @option{-ffp-int-builtin-inexact}, allowing the
9619 exception to be raised. This option does nothing unless
9620 @option{-ftrapping-math} is in effect.
9622 Even if @option{-fno-fp-int-builtin-inexact} is used, if the functions
9623 generate a call to a library function then the ``inexact'' exception
9624 may be raised if the library implementation does not follow TS 18661.
9626 @item -fsingle-precision-constant
9627 @opindex fsingle-precision-constant
9628 Treat floating-point constants as single precision instead of
9629 implicitly converting them to double-precision constants.
9631 @item -fcx-limited-range
9632 @opindex fcx-limited-range
9633 When enabled, this option states that a range reduction step is not
9634 needed when performing complex division. Also, there is no checking
9635 whether the result of a complex multiplication or division is @code{NaN
9636 + I*NaN}, with an attempt to rescue the situation in that case. The
9637 default is @option{-fno-cx-limited-range}, but is enabled by
9638 @option{-ffast-math}.
9640 This option controls the default setting of the ISO C99
9641 @code{CX_LIMITED_RANGE} pragma. Nevertheless, the option applies to
9644 @item -fcx-fortran-rules
9645 @opindex fcx-fortran-rules
9646 Complex multiplication and division follow Fortran rules. Range
9647 reduction is done as part of complex division, but there is no checking
9648 whether the result of a complex multiplication or division is @code{NaN
9649 + I*NaN}, with an attempt to rescue the situation in that case.
9651 The default is @option{-fno-cx-fortran-rules}.
9655 The following options control optimizations that may improve
9656 performance, but are not enabled by any @option{-O} options. This
9657 section includes experimental options that may produce broken code.
9660 @item -fbranch-probabilities
9661 @opindex fbranch-probabilities
9662 After running a program compiled with @option{-fprofile-arcs}
9663 (@pxref{Instrumentation Options}),
9664 you can compile it a second time using
9665 @option{-fbranch-probabilities}, to improve optimizations based on
9666 the number of times each branch was taken. When a program
9667 compiled with @option{-fprofile-arcs} exits, it saves arc execution
9668 counts to a file called @file{@var{sourcename}.gcda} for each source
9669 file. The information in this data file is very dependent on the
9670 structure of the generated code, so you must use the same source code
9671 and the same optimization options for both compilations.
9673 With @option{-fbranch-probabilities}, GCC puts a
9674 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
9675 These can be used to improve optimization. Currently, they are only
9676 used in one place: in @file{reorg.c}, instead of guessing which path a
9677 branch is most likely to take, the @samp{REG_BR_PROB} values are used to
9678 exactly determine which path is taken more often.
9680 @item -fprofile-values
9681 @opindex fprofile-values
9682 If combined with @option{-fprofile-arcs}, it adds code so that some
9683 data about values of expressions in the program is gathered.
9685 With @option{-fbranch-probabilities}, it reads back the data gathered
9686 from profiling values of expressions for usage in optimizations.
9688 Enabled with @option{-fprofile-generate} and @option{-fprofile-use}.
9690 @item -fprofile-reorder-functions
9691 @opindex fprofile-reorder-functions
9692 Function reordering based on profile instrumentation collects
9693 first time of execution of a function and orders these functions
9696 Enabled with @option{-fprofile-use}.
9700 If combined with @option{-fprofile-arcs}, this option instructs the compiler
9701 to add code to gather information about values of expressions.
9703 With @option{-fbranch-probabilities}, it reads back the data gathered
9704 and actually performs the optimizations based on them.
9705 Currently the optimizations include specialization of division operations
9706 using the knowledge about the value of the denominator.
9708 @item -frename-registers
9709 @opindex frename-registers
9710 Attempt to avoid false dependencies in scheduled code by making use
9711 of registers left over after register allocation. This optimization
9712 most benefits processors with lots of registers. Depending on the
9713 debug information format adopted by the target, however, it can
9714 make debugging impossible, since variables no longer stay in
9715 a ``home register''.
9717 Enabled by default with @option{-funroll-loops}.
9719 @item -fschedule-fusion
9720 @opindex fschedule-fusion
9721 Performs a target dependent pass over the instruction stream to schedule
9722 instructions of same type together because target machine can execute them
9723 more efficiently if they are adjacent to each other in the instruction flow.
9725 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
9729 Perform tail duplication to enlarge superblock size. This transformation
9730 simplifies the control flow of the function allowing other optimizations to do
9733 Enabled with @option{-fprofile-use}.
9735 @item -funroll-loops
9736 @opindex funroll-loops
9737 Unroll loops whose number of iterations can be determined at compile time or
9738 upon entry to the loop. @option{-funroll-loops} implies
9739 @option{-frerun-cse-after-loop}, @option{-fweb} and @option{-frename-registers}.
9740 It also turns on complete loop peeling (i.e.@: complete removal of loops with
9741 a small constant number of iterations). This option makes code larger, and may
9742 or may not make it run faster.
9744 Enabled with @option{-fprofile-use}.
9746 @item -funroll-all-loops
9747 @opindex funroll-all-loops
9748 Unroll all loops, even if their number of iterations is uncertain when
9749 the loop is entered. This usually makes programs run more slowly.
9750 @option{-funroll-all-loops} implies the same options as
9751 @option{-funroll-loops}.
9754 @opindex fpeel-loops
9755 Peels loops for which there is enough information that they do not
9756 roll much (from profile feedback or static analysis). It also turns on
9757 complete loop peeling (i.e.@: complete removal of loops with small constant
9758 number of iterations).
9760 Enabled with @option{-O3} and/or @option{-fprofile-use}.
9762 @item -fmove-loop-invariants
9763 @opindex fmove-loop-invariants
9764 Enables the loop invariant motion pass in the RTL loop optimizer. Enabled
9765 at level @option{-O1}
9768 @opindex fsplit-loops
9769 Split a loop into two if it contains a condition that's always true
9770 for one side of the iteration space and false for the other.
9772 @item -funswitch-loops
9773 @opindex funswitch-loops
9774 Move branches with loop invariant conditions out of the loop, with duplicates
9775 of the loop on both branches (modified according to result of the condition).
9777 @item -ffunction-sections
9778 @itemx -fdata-sections
9779 @opindex ffunction-sections
9780 @opindex fdata-sections
9781 Place each function or data item into its own section in the output
9782 file if the target supports arbitrary sections. The name of the
9783 function or the name of the data item determines the section's name
9786 Use these options on systems where the linker can perform optimizations to
9787 improve locality of reference in the instruction space. Most systems using the
9788 ELF object format have linkers with such optimizations. On AIX, the linker
9789 rearranges sections (CSECTs) based on the call graph. The performance impact
9792 Together with a linker garbage collection (linker @option{--gc-sections}
9793 option) these options may lead to smaller statically-linked executables (after
9796 On ELF/DWARF systems these options do not degenerate the quality of the debug
9797 information. There could be issues with other object files/debug info formats.
9799 Only use these options when there are significant benefits from doing so. When
9800 you specify these options, the assembler and linker create larger object and
9801 executable files and are also slower. These options affect code generation.
9802 They prevent optimizations by the compiler and assembler using relative
9803 locations inside a translation unit since the locations are unknown until
9804 link time. An example of such an optimization is relaxing calls to short call
9807 @item -fbranch-target-load-optimize
9808 @opindex fbranch-target-load-optimize
9809 Perform branch target register load optimization before prologue / epilogue
9811 The use of target registers can typically be exposed only during reload,
9812 thus hoisting loads out of loops and doing inter-block scheduling needs
9813 a separate optimization pass.
9815 @item -fbranch-target-load-optimize2
9816 @opindex fbranch-target-load-optimize2
9817 Perform branch target register load optimization after prologue / epilogue
9820 @item -fbtr-bb-exclusive
9821 @opindex fbtr-bb-exclusive
9822 When performing branch target register load optimization, don't reuse
9823 branch target registers within any basic block.
9826 @opindex fstdarg-opt
9827 Optimize the prologue of variadic argument functions with respect to usage of
9830 @item -fsection-anchors
9831 @opindex fsection-anchors
9832 Try to reduce the number of symbolic address calculations by using
9833 shared ``anchor'' symbols to address nearby objects. This transformation
9834 can help to reduce the number of GOT entries and GOT accesses on some
9837 For example, the implementation of the following function @code{foo}:
9841 int foo (void) @{ return a + b + c; @}
9845 usually calculates the addresses of all three variables, but if you
9846 compile it with @option{-fsection-anchors}, it accesses the variables
9847 from a common anchor point instead. The effect is similar to the
9848 following pseudocode (which isn't valid C):
9853 register int *xr = &x;
9854 return xr[&a - &x] + xr[&b - &x] + xr[&c - &x];
9858 Not all targets support this option.
9860 @item --param @var{name}=@var{value}
9862 In some places, GCC uses various constants to control the amount of
9863 optimization that is done. For example, GCC does not inline functions
9864 that contain more than a certain number of instructions. You can
9865 control some of these constants on the command line using the
9866 @option{--param} option.
9868 The names of specific parameters, and the meaning of the values, are
9869 tied to the internals of the compiler, and are subject to change
9870 without notice in future releases.
9872 In each case, the @var{value} is an integer. The allowable choices for
9876 @item predictable-branch-outcome
9877 When branch is predicted to be taken with probability lower than this threshold
9878 (in percent), then it is considered well predictable. The default is 10.
9880 @item max-rtl-if-conversion-insns
9881 RTL if-conversion tries to remove conditional branches around a block and
9882 replace them with conditionally executed instructions. This parameter
9883 gives the maximum number of instructions in a block which should be
9884 considered for if-conversion. The default is 10, though the compiler will
9885 also use other heuristics to decide whether if-conversion is likely to be
9888 @item max-rtl-if-conversion-predictable-cost
9889 @item max-rtl-if-conversion-unpredictable-cost
9890 RTL if-conversion will try to remove conditional branches around a block
9891 and replace them with conditionally executed instructions. These parameters
9892 give the maximum permissible cost for the sequence that would be generated
9893 by if-conversion depending on whether the branch is statically determined
9894 to be predictable or not. The units for this parameter are the same as
9895 those for the GCC internal seq_cost metric. The compiler will try to
9896 provide a reasonable default for this parameter using the BRANCH_COST
9899 @item max-crossjump-edges
9900 The maximum number of incoming edges to consider for cross-jumping.
9901 The algorithm used by @option{-fcrossjumping} is @math{O(N^2)} in
9902 the number of edges incoming to each block. Increasing values mean
9903 more aggressive optimization, making the compilation time increase with
9904 probably small improvement in executable size.
9906 @item min-crossjump-insns
9907 The minimum number of instructions that must be matched at the end
9908 of two blocks before cross-jumping is performed on them. This
9909 value is ignored in the case where all instructions in the block being
9910 cross-jumped from are matched. The default value is 5.
9912 @item max-grow-copy-bb-insns
9913 The maximum code size expansion factor when copying basic blocks
9914 instead of jumping. The expansion is relative to a jump instruction.
9915 The default value is 8.
9917 @item max-goto-duplication-insns
9918 The maximum number of instructions to duplicate to a block that jumps
9919 to a computed goto. To avoid @math{O(N^2)} behavior in a number of
9920 passes, GCC factors computed gotos early in the compilation process,
9921 and unfactors them as late as possible. Only computed jumps at the
9922 end of a basic blocks with no more than max-goto-duplication-insns are
9923 unfactored. The default value is 8.
9925 @item max-delay-slot-insn-search
9926 The maximum number of instructions to consider when looking for an
9927 instruction to fill a delay slot. If more than this arbitrary number of
9928 instructions are searched, the time savings from filling the delay slot
9929 are minimal, so stop searching. Increasing values mean more
9930 aggressive optimization, making the compilation time increase with probably
9931 small improvement in execution time.
9933 @item max-delay-slot-live-search
9934 When trying to fill delay slots, the maximum number of instructions to
9935 consider when searching for a block with valid live register
9936 information. Increasing this arbitrarily chosen value means more
9937 aggressive optimization, increasing the compilation time. This parameter
9938 should be removed when the delay slot code is rewritten to maintain the
9941 @item max-gcse-memory
9942 The approximate maximum amount of memory that can be allocated in
9943 order to perform the global common subexpression elimination
9944 optimization. If more memory than specified is required, the
9945 optimization is not done.
9947 @item max-gcse-insertion-ratio
9948 If the ratio of expression insertions to deletions is larger than this value
9949 for any expression, then RTL PRE inserts or removes the expression and thus
9950 leaves partially redundant computations in the instruction stream. The default value is 20.
9952 @item max-pending-list-length
9953 The maximum number of pending dependencies scheduling allows
9954 before flushing the current state and starting over. Large functions
9955 with few branches or calls can create excessively large lists which
9956 needlessly consume memory and resources.
9958 @item max-modulo-backtrack-attempts
9959 The maximum number of backtrack attempts the scheduler should make
9960 when modulo scheduling a loop. Larger values can exponentially increase
9963 @item max-inline-insns-single
9964 Several parameters control the tree inliner used in GCC@.
9965 This number sets the maximum number of instructions (counted in GCC's
9966 internal representation) in a single function that the tree inliner
9967 considers for inlining. This only affects functions declared
9968 inline and methods implemented in a class declaration (C++).
9969 The default value is 400.
9971 @item max-inline-insns-auto
9972 When you use @option{-finline-functions} (included in @option{-O3}),
9973 a lot of functions that would otherwise not be considered for inlining
9974 by the compiler are investigated. To those functions, a different
9975 (more restrictive) limit compared to functions declared inline can
9977 The default value is 40.
9979 @item inline-min-speedup
9980 When estimated performance improvement of caller + callee runtime exceeds this
9981 threshold (in percent), the function can be inlined regardless of the limit on
9982 @option{--param max-inline-insns-single} and @option{--param
9983 max-inline-insns-auto}.
9985 @item large-function-insns
9986 The limit specifying really large functions. For functions larger than this
9987 limit after inlining, inlining is constrained by
9988 @option{--param large-function-growth}. This parameter is useful primarily
9989 to avoid extreme compilation time caused by non-linear algorithms used by the
9991 The default value is 2700.
9993 @item large-function-growth
9994 Specifies maximal growth of large function caused by inlining in percents.
9995 The default value is 100 which limits large function growth to 2.0 times
9998 @item large-unit-insns
9999 The limit specifying large translation unit. Growth caused by inlining of
10000 units larger than this limit is limited by @option{--param inline-unit-growth}.
10001 For small units this might be too tight.
10002 For example, consider a unit consisting of function A
10003 that is inline and B that just calls A three times. If B is small relative to
10004 A, the growth of unit is 300\% and yet such inlining is very sane. For very
10005 large units consisting of small inlineable functions, however, the overall unit
10006 growth limit is needed to avoid exponential explosion of code size. Thus for
10007 smaller units, the size is increased to @option{--param large-unit-insns}
10008 before applying @option{--param inline-unit-growth}. The default is 10000.
10010 @item inline-unit-growth
10011 Specifies maximal overall growth of the compilation unit caused by inlining.
10012 The default value is 20 which limits unit growth to 1.2 times the original
10013 size. Cold functions (either marked cold via an attribute or by profile
10014 feedback) are not accounted into the unit size.
10016 @item ipcp-unit-growth
10017 Specifies maximal overall growth of the compilation unit caused by
10018 interprocedural constant propagation. The default value is 10 which limits
10019 unit growth to 1.1 times the original size.
10021 @item large-stack-frame
10022 The limit specifying large stack frames. While inlining the algorithm is trying
10023 to not grow past this limit too much. The default value is 256 bytes.
10025 @item large-stack-frame-growth
10026 Specifies maximal growth of large stack frames caused by inlining in percents.
10027 The default value is 1000 which limits large stack frame growth to 11 times
10030 @item max-inline-insns-recursive
10031 @itemx max-inline-insns-recursive-auto
10032 Specifies the maximum number of instructions an out-of-line copy of a
10033 self-recursive inline
10034 function can grow into by performing recursive inlining.
10036 @option{--param max-inline-insns-recursive} applies to functions
10038 For functions not declared inline, recursive inlining
10039 happens only when @option{-finline-functions} (included in @option{-O3}) is
10040 enabled; @option{--param max-inline-insns-recursive-auto} applies instead. The
10041 default value is 450.
10043 @item max-inline-recursive-depth
10044 @itemx max-inline-recursive-depth-auto
10045 Specifies the maximum recursion depth used for recursive inlining.
10047 @option{--param max-inline-recursive-depth} applies to functions
10048 declared inline. For functions not declared inline, recursive inlining
10049 happens only when @option{-finline-functions} (included in @option{-O3}) is
10050 enabled; @option{--param max-inline-recursive-depth-auto} applies instead. The
10051 default value is 8.
10053 @item min-inline-recursive-probability
10054 Recursive inlining is profitable only for function having deep recursion
10055 in average and can hurt for function having little recursion depth by
10056 increasing the prologue size or complexity of function body to other
10059 When profile feedback is available (see @option{-fprofile-generate}) the actual
10060 recursion depth can be guessed from the probability that function recurses
10061 via a given call expression. This parameter limits inlining only to call
10062 expressions whose probability exceeds the given threshold (in percents).
10063 The default value is 10.
10065 @item early-inlining-insns
10066 Specify growth that the early inliner can make. In effect it increases
10067 the amount of inlining for code having a large abstraction penalty.
10068 The default value is 14.
10070 @item max-early-inliner-iterations
10071 Limit of iterations of the early inliner. This basically bounds
10072 the number of nested indirect calls the early inliner can resolve.
10073 Deeper chains are still handled by late inlining.
10075 @item comdat-sharing-probability
10076 Probability (in percent) that C++ inline function with comdat visibility
10077 are shared across multiple compilation units. The default value is 20.
10079 @item profile-func-internal-id
10080 A parameter to control whether to use function internal id in profile
10081 database lookup. If the value is 0, the compiler uses an id that
10082 is based on function assembler name and filename, which makes old profile
10083 data more tolerant to source changes such as function reordering etc.
10084 The default value is 0.
10086 @item min-vect-loop-bound
10087 The minimum number of iterations under which loops are not vectorized
10088 when @option{-ftree-vectorize} is used. The number of iterations after
10089 vectorization needs to be greater than the value specified by this option
10090 to allow vectorization. The default value is 0.
10092 @item gcse-cost-distance-ratio
10093 Scaling factor in calculation of maximum distance an expression
10094 can be moved by GCSE optimizations. This is currently supported only in the
10095 code hoisting pass. The bigger the ratio, the more aggressive code hoisting
10096 is with simple expressions, i.e., the expressions that have cost
10097 less than @option{gcse-unrestricted-cost}. Specifying 0 disables
10098 hoisting of simple expressions. The default value is 10.
10100 @item gcse-unrestricted-cost
10101 Cost, roughly measured as the cost of a single typical machine
10102 instruction, at which GCSE optimizations do not constrain
10103 the distance an expression can travel. This is currently
10104 supported only in the code hoisting pass. The lesser the cost,
10105 the more aggressive code hoisting is. Specifying 0
10106 allows all expressions to travel unrestricted distances.
10107 The default value is 3.
10109 @item max-hoist-depth
10110 The depth of search in the dominator tree for expressions to hoist.
10111 This is used to avoid quadratic behavior in hoisting algorithm.
10112 The value of 0 does not limit on the search, but may slow down compilation
10113 of huge functions. The default value is 30.
10115 @item max-tail-merge-comparisons
10116 The maximum amount of similar bbs to compare a bb with. This is used to
10117 avoid quadratic behavior in tree tail merging. The default value is 10.
10119 @item max-tail-merge-iterations
10120 The maximum amount of iterations of the pass over the function. This is used to
10121 limit compilation time in tree tail merging. The default value is 2.
10123 @item store-merging-allow-unaligned
10124 Allow the store merging pass to introduce unaligned stores if it is legal to
10125 do so. The default value is 1.
10127 @item max-stores-to-merge
10128 The maximum number of stores to attempt to merge into wider stores in the store
10129 merging pass. The minimum value is 2 and the default is 64.
10131 @item max-unrolled-insns
10132 The maximum number of instructions that a loop may have to be unrolled.
10133 If a loop is unrolled, this parameter also determines how many times
10134 the loop code is unrolled.
10136 @item max-average-unrolled-insns
10137 The maximum number of instructions biased by probabilities of their execution
10138 that a loop may have to be unrolled. If a loop is unrolled,
10139 this parameter also determines how many times the loop code is unrolled.
10141 @item max-unroll-times
10142 The maximum number of unrollings of a single loop.
10144 @item max-peeled-insns
10145 The maximum number of instructions that a loop may have to be peeled.
10146 If a loop is peeled, this parameter also determines how many times
10147 the loop code is peeled.
10149 @item max-peel-times
10150 The maximum number of peelings of a single loop.
10152 @item max-peel-branches
10153 The maximum number of branches on the hot path through the peeled sequence.
10155 @item max-completely-peeled-insns
10156 The maximum number of insns of a completely peeled loop.
10158 @item max-completely-peel-times
10159 The maximum number of iterations of a loop to be suitable for complete peeling.
10161 @item max-completely-peel-loop-nest-depth
10162 The maximum depth of a loop nest suitable for complete peeling.
10164 @item max-unswitch-insns
10165 The maximum number of insns of an unswitched loop.
10167 @item max-unswitch-level
10168 The maximum number of branches unswitched in a single loop.
10170 @item max-loop-headers-insns
10171 The maximum number of insns in loop header duplicated by the copy loop headers
10174 @item lim-expensive
10175 The minimum cost of an expensive expression in the loop invariant motion.
10177 @item iv-consider-all-candidates-bound
10178 Bound on number of candidates for induction variables, below which
10179 all candidates are considered for each use in induction variable
10180 optimizations. If there are more candidates than this,
10181 only the most relevant ones are considered to avoid quadratic time complexity.
10183 @item iv-max-considered-uses
10184 The induction variable optimizations give up on loops that contain more
10185 induction variable uses.
10187 @item iv-always-prune-cand-set-bound
10188 If the number of candidates in the set is smaller than this value,
10189 always try to remove unnecessary ivs from the set
10190 when adding a new one.
10192 @item avg-loop-niter
10193 Average number of iterations of a loop.
10195 @item dse-max-object-size
10196 Maximum size (in bytes) of objects tracked bytewise by dead store elimination.
10197 Larger values may result in larger compilation times.
10199 @item scev-max-expr-size
10200 Bound on size of expressions used in the scalar evolutions analyzer.
10201 Large expressions slow the analyzer.
10203 @item scev-max-expr-complexity
10204 Bound on the complexity of the expressions in the scalar evolutions analyzer.
10205 Complex expressions slow the analyzer.
10207 @item max-tree-if-conversion-phi-args
10208 Maximum number of arguments in a PHI supported by TREE if conversion
10209 unless the loop is marked with simd pragma.
10211 @item vect-max-version-for-alignment-checks
10212 The maximum number of run-time checks that can be performed when
10213 doing loop versioning for alignment in the vectorizer.
10215 @item vect-max-version-for-alias-checks
10216 The maximum number of run-time checks that can be performed when
10217 doing loop versioning for alias in the vectorizer.
10219 @item vect-max-peeling-for-alignment
10220 The maximum number of loop peels to enhance access alignment
10221 for vectorizer. Value -1 means no limit.
10223 @item max-iterations-to-track
10224 The maximum number of iterations of a loop the brute-force algorithm
10225 for analysis of the number of iterations of the loop tries to evaluate.
10227 @item hot-bb-count-ws-permille
10228 A basic block profile count is considered hot if it contributes to
10229 the given permillage (i.e. 0...1000) of the entire profiled execution.
10231 @item hot-bb-frequency-fraction
10232 Select fraction of the entry block frequency of executions of basic block in
10233 function given basic block needs to have to be considered hot.
10235 @item max-predicted-iterations
10236 The maximum number of loop iterations we predict statically. This is useful
10237 in cases where a function contains a single loop with known bound and
10238 another loop with unknown bound.
10239 The known number of iterations is predicted correctly, while
10240 the unknown number of iterations average to roughly 10. This means that the
10241 loop without bounds appears artificially cold relative to the other one.
10243 @item builtin-expect-probability
10244 Control the probability of the expression having the specified value. This
10245 parameter takes a percentage (i.e. 0 ... 100) as input.
10246 The default probability of 90 is obtained empirically.
10248 @item align-threshold
10250 Select fraction of the maximal frequency of executions of a basic block in
10251 a function to align the basic block.
10253 @item align-loop-iterations
10255 A loop expected to iterate at least the selected number of iterations is
10258 @item tracer-dynamic-coverage
10259 @itemx tracer-dynamic-coverage-feedback
10261 This value is used to limit superblock formation once the given percentage of
10262 executed instructions is covered. This limits unnecessary code size
10265 The @option{tracer-dynamic-coverage-feedback} parameter
10266 is used only when profile
10267 feedback is available. The real profiles (as opposed to statically estimated
10268 ones) are much less balanced allowing the threshold to be larger value.
10270 @item tracer-max-code-growth
10271 Stop tail duplication once code growth has reached given percentage. This is
10272 a rather artificial limit, as most of the duplicates are eliminated later in
10273 cross jumping, so it may be set to much higher values than is the desired code
10276 @item tracer-min-branch-ratio
10278 Stop reverse growth when the reverse probability of best edge is less than this
10279 threshold (in percent).
10281 @item tracer-min-branch-probability
10282 @itemx tracer-min-branch-probability-feedback
10284 Stop forward growth if the best edge has probability lower than this
10287 Similarly to @option{tracer-dynamic-coverage} two parameters are
10288 provided. @option{tracer-min-branch-probability-feedback} is used for
10289 compilation with profile feedback and @option{tracer-min-branch-probability}
10290 compilation without. The value for compilation with profile feedback
10291 needs to be more conservative (higher) in order to make tracer
10294 @item stack-clash-protection-guard-size
10295 Specify the size of the operating system provided stack guard as
10296 2 raised to @var{num} bytes. The default value is 12 (4096 bytes).
10297 Acceptable values are between 12 and 30. Higher values may reduce the
10298 number of explicit probes, but a value larger than the operating system
10299 provided guard will leave code vulnerable to stack clash style attacks.
10301 @item stack-clash-protection-probe-interval
10302 Stack clash protection involves probing stack space as it is allocated. This
10303 param controls the maximum distance between probes into the stack as 2 raised
10304 to @var{num} bytes. Acceptable values are between 10 and 16 and defaults to
10305 12. Higher values may reduce the number of explicit probes, but a value
10306 larger than the operating system provided guard will leave code vulnerable to
10307 stack clash style attacks.
10309 @item max-cse-path-length
10311 The maximum number of basic blocks on path that CSE considers.
10314 @item max-cse-insns
10315 The maximum number of instructions CSE processes before flushing.
10316 The default is 1000.
10318 @item ggc-min-expand
10320 GCC uses a garbage collector to manage its own memory allocation. This
10321 parameter specifies the minimum percentage by which the garbage
10322 collector's heap should be allowed to expand between collections.
10323 Tuning this may improve compilation speed; it has no effect on code
10326 The default is 30% + 70% * (RAM/1GB) with an upper bound of 100% when
10327 RAM >= 1GB@. If @code{getrlimit} is available, the notion of ``RAM'' is
10328 the smallest of actual RAM and @code{RLIMIT_DATA} or @code{RLIMIT_AS}. If
10329 GCC is not able to calculate RAM on a particular platform, the lower
10330 bound of 30% is used. Setting this parameter and
10331 @option{ggc-min-heapsize} to zero causes a full collection to occur at
10332 every opportunity. This is extremely slow, but can be useful for
10335 @item ggc-min-heapsize
10337 Minimum size of the garbage collector's heap before it begins bothering
10338 to collect garbage. The first collection occurs after the heap expands
10339 by @option{ggc-min-expand}% beyond @option{ggc-min-heapsize}. Again,
10340 tuning this may improve compilation speed, and has no effect on code
10343 The default is the smaller of RAM/8, RLIMIT_RSS, or a limit that
10344 tries to ensure that RLIMIT_DATA or RLIMIT_AS are not exceeded, but
10345 with a lower bound of 4096 (four megabytes) and an upper bound of
10346 131072 (128 megabytes). If GCC is not able to calculate RAM on a
10347 particular platform, the lower bound is used. Setting this parameter
10348 very large effectively disables garbage collection. Setting this
10349 parameter and @option{ggc-min-expand} to zero causes a full collection
10350 to occur at every opportunity.
10352 @item max-reload-search-insns
10353 The maximum number of instruction reload should look backward for equivalent
10354 register. Increasing values mean more aggressive optimization, making the
10355 compilation time increase with probably slightly better performance.
10356 The default value is 100.
10358 @item max-cselib-memory-locations
10359 The maximum number of memory locations cselib should take into account.
10360 Increasing values mean more aggressive optimization, making the compilation time
10361 increase with probably slightly better performance. The default value is 500.
10363 @item max-sched-ready-insns
10364 The maximum number of instructions ready to be issued the scheduler should
10365 consider at any given time during the first scheduling pass. Increasing
10366 values mean more thorough searches, making the compilation time increase
10367 with probably little benefit. The default value is 100.
10369 @item max-sched-region-blocks
10370 The maximum number of blocks in a region to be considered for
10371 interblock scheduling. The default value is 10.
10373 @item max-pipeline-region-blocks
10374 The maximum number of blocks in a region to be considered for
10375 pipelining in the selective scheduler. The default value is 15.
10377 @item max-sched-region-insns
10378 The maximum number of insns in a region to be considered for
10379 interblock scheduling. The default value is 100.
10381 @item max-pipeline-region-insns
10382 The maximum number of insns in a region to be considered for
10383 pipelining in the selective scheduler. The default value is 200.
10385 @item min-spec-prob
10386 The minimum probability (in percents) of reaching a source block
10387 for interblock speculative scheduling. The default value is 40.
10389 @item max-sched-extend-regions-iters
10390 The maximum number of iterations through CFG to extend regions.
10391 A value of 0 (the default) disables region extensions.
10393 @item max-sched-insn-conflict-delay
10394 The maximum conflict delay for an insn to be considered for speculative motion.
10395 The default value is 3.
10397 @item sched-spec-prob-cutoff
10398 The minimal probability of speculation success (in percents), so that
10399 speculative insns are scheduled.
10400 The default value is 40.
10402 @item sched-state-edge-prob-cutoff
10403 The minimum probability an edge must have for the scheduler to save its
10405 The default value is 10.
10407 @item sched-mem-true-dep-cost
10408 Minimal distance (in CPU cycles) between store and load targeting same
10409 memory locations. The default value is 1.
10411 @item selsched-max-lookahead
10412 The maximum size of the lookahead window of selective scheduling. It is a
10413 depth of search for available instructions.
10414 The default value is 50.
10416 @item selsched-max-sched-times
10417 The maximum number of times that an instruction is scheduled during
10418 selective scheduling. This is the limit on the number of iterations
10419 through which the instruction may be pipelined. The default value is 2.
10421 @item selsched-insns-to-rename
10422 The maximum number of best instructions in the ready list that are considered
10423 for renaming in the selective scheduler. The default value is 2.
10426 The minimum value of stage count that swing modulo scheduler
10427 generates. The default value is 2.
10429 @item max-last-value-rtl
10430 The maximum size measured as number of RTLs that can be recorded in an expression
10431 in combiner for a pseudo register as last known value of that register. The default
10434 @item max-combine-insns
10435 The maximum number of instructions the RTL combiner tries to combine.
10436 The default value is 2 at @option{-Og} and 4 otherwise.
10438 @item integer-share-limit
10439 Small integer constants can use a shared data structure, reducing the
10440 compiler's memory usage and increasing its speed. This sets the maximum
10441 value of a shared integer constant. The default value is 256.
10443 @item ssp-buffer-size
10444 The minimum size of buffers (i.e.@: arrays) that receive stack smashing
10445 protection when @option{-fstack-protection} is used.
10447 @item min-size-for-stack-sharing
10448 The minimum size of variables taking part in stack slot sharing when not
10449 optimizing. The default value is 32.
10451 @item max-jump-thread-duplication-stmts
10452 Maximum number of statements allowed in a block that needs to be
10453 duplicated when threading jumps.
10455 @item max-fields-for-field-sensitive
10456 Maximum number of fields in a structure treated in
10457 a field sensitive manner during pointer analysis. The default is zero
10458 for @option{-O0} and @option{-O1},
10459 and 100 for @option{-Os}, @option{-O2}, and @option{-O3}.
10461 @item prefetch-latency
10462 Estimate on average number of instructions that are executed before
10463 prefetch finishes. The distance prefetched ahead is proportional
10464 to this constant. Increasing this number may also lead to less
10465 streams being prefetched (see @option{simultaneous-prefetches}).
10467 @item simultaneous-prefetches
10468 Maximum number of prefetches that can run at the same time.
10470 @item l1-cache-line-size
10471 The size of cache line in L1 cache, in bytes.
10473 @item l1-cache-size
10474 The size of L1 cache, in kilobytes.
10476 @item l2-cache-size
10477 The size of L2 cache, in kilobytes.
10479 @item min-insn-to-prefetch-ratio
10480 The minimum ratio between the number of instructions and the
10481 number of prefetches to enable prefetching in a loop.
10483 @item prefetch-min-insn-to-mem-ratio
10484 The minimum ratio between the number of instructions and the
10485 number of memory references to enable prefetching in a loop.
10487 @item use-canonical-types
10488 Whether the compiler should use the ``canonical'' type system. By
10489 default, this should always be 1, which uses a more efficient internal
10490 mechanism for comparing types in C++ and Objective-C++. However, if
10491 bugs in the canonical type system are causing compilation failures,
10492 set this value to 0 to disable canonical types.
10494 @item switch-conversion-max-branch-ratio
10495 Switch initialization conversion refuses to create arrays that are
10496 bigger than @option{switch-conversion-max-branch-ratio} times the number of
10497 branches in the switch.
10499 @item max-partial-antic-length
10500 Maximum length of the partial antic set computed during the tree
10501 partial redundancy elimination optimization (@option{-ftree-pre}) when
10502 optimizing at @option{-O3} and above. For some sorts of source code
10503 the enhanced partial redundancy elimination optimization can run away,
10504 consuming all of the memory available on the host machine. This
10505 parameter sets a limit on the length of the sets that are computed,
10506 which prevents the runaway behavior. Setting a value of 0 for
10507 this parameter allows an unlimited set length.
10509 @item sccvn-max-scc-size
10510 Maximum size of a strongly connected component (SCC) during SCCVN
10511 processing. If this limit is hit, SCCVN processing for the whole
10512 function is not done and optimizations depending on it are
10513 disabled. The default maximum SCC size is 10000.
10515 @item sccvn-max-alias-queries-per-access
10516 Maximum number of alias-oracle queries we perform when looking for
10517 redundancies for loads and stores. If this limit is hit the search
10518 is aborted and the load or store is not considered redundant. The
10519 number of queries is algorithmically limited to the number of
10520 stores on all paths from the load to the function entry.
10521 The default maximum number of queries is 1000.
10523 @item ira-max-loops-num
10524 IRA uses regional register allocation by default. If a function
10525 contains more loops than the number given by this parameter, only at most
10526 the given number of the most frequently-executed loops form regions
10527 for regional register allocation. The default value of the
10530 @item ira-max-conflict-table-size
10531 Although IRA uses a sophisticated algorithm to compress the conflict
10532 table, the table can still require excessive amounts of memory for
10533 huge functions. If the conflict table for a function could be more
10534 than the size in MB given by this parameter, the register allocator
10535 instead uses a faster, simpler, and lower-quality
10536 algorithm that does not require building a pseudo-register conflict table.
10537 The default value of the parameter is 2000.
10539 @item ira-loop-reserved-regs
10540 IRA can be used to evaluate more accurate register pressure in loops
10541 for decisions to move loop invariants (see @option{-O3}). The number
10542 of available registers reserved for some other purposes is given
10543 by this parameter. The default value of the parameter is 2, which is
10544 the minimal number of registers needed by typical instructions.
10545 This value is the best found from numerous experiments.
10547 @item lra-inheritance-ebb-probability-cutoff
10548 LRA tries to reuse values reloaded in registers in subsequent insns.
10549 This optimization is called inheritance. EBB is used as a region to
10550 do this optimization. The parameter defines a minimal fall-through
10551 edge probability in percentage used to add BB to inheritance EBB in
10552 LRA. The default value of the parameter is 40. The value was chosen
10553 from numerous runs of SPEC2000 on x86-64.
10555 @item loop-invariant-max-bbs-in-loop
10556 Loop invariant motion can be very expensive, both in compilation time and
10557 in amount of needed compile-time memory, with very large loops. Loops
10558 with more basic blocks than this parameter won't have loop invariant
10559 motion optimization performed on them. The default value of the
10560 parameter is 1000 for @option{-O1} and 10000 for @option{-O2} and above.
10562 @item loop-max-datarefs-for-datadeps
10563 Building data dependencies is expensive for very large loops. This
10564 parameter limits the number of data references in loops that are
10565 considered for data dependence analysis. These large loops are no
10566 handled by the optimizations using loop data dependencies.
10567 The default value is 1000.
10569 @item max-vartrack-size
10570 Sets a maximum number of hash table slots to use during variable
10571 tracking dataflow analysis of any function. If this limit is exceeded
10572 with variable tracking at assignments enabled, analysis for that
10573 function is retried without it, after removing all debug insns from
10574 the function. If the limit is exceeded even without debug insns, var
10575 tracking analysis is completely disabled for the function. Setting
10576 the parameter to zero makes it unlimited.
10578 @item max-vartrack-expr-depth
10579 Sets a maximum number of recursion levels when attempting to map
10580 variable names or debug temporaries to value expressions. This trades
10581 compilation time for more complete debug information. If this is set too
10582 low, value expressions that are available and could be represented in
10583 debug information may end up not being used; setting this higher may
10584 enable the compiler to find more complex debug expressions, but compile
10585 time and memory use may grow. The default is 12.
10587 @item min-nondebug-insn-uid
10588 Use uids starting at this parameter for nondebug insns. The range below
10589 the parameter is reserved exclusively for debug insns created by
10590 @option{-fvar-tracking-assignments}, but debug insns may get
10591 (non-overlapping) uids above it if the reserved range is exhausted.
10593 @item ipa-sra-ptr-growth-factor
10594 IPA-SRA replaces a pointer to an aggregate with one or more new
10595 parameters only when their cumulative size is less or equal to
10596 @option{ipa-sra-ptr-growth-factor} times the size of the original
10599 @item sra-max-scalarization-size-Ospeed
10600 @item sra-max-scalarization-size-Osize
10601 The two Scalar Reduction of Aggregates passes (SRA and IPA-SRA) aim to
10602 replace scalar parts of aggregates with uses of independent scalar
10603 variables. These parameters control the maximum size, in storage units,
10604 of aggregate which is considered for replacement when compiling for
10606 (@option{sra-max-scalarization-size-Ospeed}) or size
10607 (@option{sra-max-scalarization-size-Osize}) respectively.
10609 @item tm-max-aggregate-size
10610 When making copies of thread-local variables in a transaction, this
10611 parameter specifies the size in bytes after which variables are
10612 saved with the logging functions as opposed to save/restore code
10613 sequence pairs. This option only applies when using
10616 @item graphite-max-nb-scop-params
10617 To avoid exponential effects in the Graphite loop transforms, the
10618 number of parameters in a Static Control Part (SCoP) is bounded. The
10619 default value is 10 parameters, a value of zero can be used to lift
10620 the bound. A variable whose value is unknown at compilation time and
10621 defined outside a SCoP is a parameter of the SCoP.
10623 @item loop-block-tile-size
10624 Loop blocking or strip mining transforms, enabled with
10625 @option{-floop-block} or @option{-floop-strip-mine}, strip mine each
10626 loop in the loop nest by a given number of iterations. The strip
10627 length can be changed using the @option{loop-block-tile-size}
10628 parameter. The default value is 51 iterations.
10630 @item loop-unroll-jam-size
10631 Specify the unroll factor for the @option{-floop-unroll-and-jam} option. The
10632 default value is 4.
10634 @item loop-unroll-jam-depth
10635 Specify the dimension to be unrolled (counting from the most inner loop)
10636 for the @option{-floop-unroll-and-jam}. The default value is 2.
10638 @item ipa-cp-value-list-size
10639 IPA-CP attempts to track all possible values and types passed to a function's
10640 parameter in order to propagate them and perform devirtualization.
10641 @option{ipa-cp-value-list-size} is the maximum number of values and types it
10642 stores per one formal parameter of a function.
10644 @item ipa-cp-eval-threshold
10645 IPA-CP calculates its own score of cloning profitability heuristics
10646 and performs those cloning opportunities with scores that exceed
10647 @option{ipa-cp-eval-threshold}.
10649 @item ipa-cp-recursion-penalty
10650 Percentage penalty the recursive functions will receive when they
10651 are evaluated for cloning.
10653 @item ipa-cp-single-call-penalty
10654 Percentage penalty functions containing a single call to another
10655 function will receive when they are evaluated for cloning.
10658 @item ipa-max-agg-items
10659 IPA-CP is also capable to propagate a number of scalar values passed
10660 in an aggregate. @option{ipa-max-agg-items} controls the maximum
10661 number of such values per one parameter.
10663 @item ipa-cp-loop-hint-bonus
10664 When IPA-CP determines that a cloning candidate would make the number
10665 of iterations of a loop known, it adds a bonus of
10666 @option{ipa-cp-loop-hint-bonus} to the profitability score of
10669 @item ipa-cp-array-index-hint-bonus
10670 When IPA-CP determines that a cloning candidate would make the index of
10671 an array access known, it adds a bonus of
10672 @option{ipa-cp-array-index-hint-bonus} to the profitability
10673 score of the candidate.
10675 @item ipa-max-aa-steps
10676 During its analysis of function bodies, IPA-CP employs alias analysis
10677 in order to track values pointed to by function parameters. In order
10678 not spend too much time analyzing huge functions, it gives up and
10679 consider all memory clobbered after examining
10680 @option{ipa-max-aa-steps} statements modifying memory.
10682 @item lto-partitions
10683 Specify desired number of partitions produced during WHOPR compilation.
10684 The number of partitions should exceed the number of CPUs used for compilation.
10685 The default value is 32.
10687 @item lto-min-partition
10688 Size of minimal partition for WHOPR (in estimated instructions).
10689 This prevents expenses of splitting very small programs into too many
10692 @item lto-max-partition
10693 Size of max partition for WHOPR (in estimated instructions).
10694 to provide an upper bound for individual size of partition.
10695 Meant to be used only with balanced partitioning.
10697 @item cxx-max-namespaces-for-diagnostic-help
10698 The maximum number of namespaces to consult for suggestions when C++
10699 name lookup fails for an identifier. The default is 1000.
10701 @item sink-frequency-threshold
10702 The maximum relative execution frequency (in percents) of the target block
10703 relative to a statement's original block to allow statement sinking of a
10704 statement. Larger numbers result in more aggressive statement sinking.
10705 The default value is 75. A small positive adjustment is applied for
10706 statements with memory operands as those are even more profitable so sink.
10708 @item max-stores-to-sink
10709 The maximum number of conditional store pairs that can be sunk. Set to 0
10710 if either vectorization (@option{-ftree-vectorize}) or if-conversion
10711 (@option{-ftree-loop-if-convert}) is disabled. The default is 2.
10713 @item allow-store-data-races
10714 Allow optimizers to introduce new data races on stores.
10715 Set to 1 to allow, otherwise to 0. This option is enabled by default
10716 at optimization level @option{-Ofast}.
10718 @item case-values-threshold
10719 The smallest number of different values for which it is best to use a
10720 jump-table instead of a tree of conditional branches. If the value is
10721 0, use the default for the machine. The default is 0.
10723 @item tree-reassoc-width
10724 Set the maximum number of instructions executed in parallel in
10725 reassociated tree. This parameter overrides target dependent
10726 heuristics used by default if has non zero value.
10728 @item sched-pressure-algorithm
10729 Choose between the two available implementations of
10730 @option{-fsched-pressure}. Algorithm 1 is the original implementation
10731 and is the more likely to prevent instructions from being reordered.
10732 Algorithm 2 was designed to be a compromise between the relatively
10733 conservative approach taken by algorithm 1 and the rather aggressive
10734 approach taken by the default scheduler. It relies more heavily on
10735 having a regular register file and accurate register pressure classes.
10736 See @file{haifa-sched.c} in the GCC sources for more details.
10738 The default choice depends on the target.
10740 @item max-slsr-cand-scan
10741 Set the maximum number of existing candidates that are considered when
10742 seeking a basis for a new straight-line strength reduction candidate.
10745 Enable buffer overflow detection for global objects. This kind
10746 of protection is enabled by default if you are using
10747 @option{-fsanitize=address} option.
10748 To disable global objects protection use @option{--param asan-globals=0}.
10751 Enable buffer overflow detection for stack objects. This kind of
10752 protection is enabled by default when using @option{-fsanitize=address}.
10753 To disable stack protection use @option{--param asan-stack=0} option.
10755 @item asan-instrument-reads
10756 Enable buffer overflow detection for memory reads. This kind of
10757 protection is enabled by default when using @option{-fsanitize=address}.
10758 To disable memory reads protection use
10759 @option{--param asan-instrument-reads=0}.
10761 @item asan-instrument-writes
10762 Enable buffer overflow detection for memory writes. This kind of
10763 protection is enabled by default when using @option{-fsanitize=address}.
10764 To disable memory writes protection use
10765 @option{--param asan-instrument-writes=0} option.
10767 @item asan-memintrin
10768 Enable detection for built-in functions. This kind of protection
10769 is enabled by default when using @option{-fsanitize=address}.
10770 To disable built-in functions protection use
10771 @option{--param asan-memintrin=0}.
10773 @item asan-use-after-return
10774 Enable detection of use-after-return. This kind of protection
10775 is enabled by default when using the @option{-fsanitize=address} option.
10776 To disable it use @option{--param asan-use-after-return=0}.
10778 Note: By default the check is disabled at run time. To enable it,
10779 add @code{detect_stack_use_after_return=1} to the environment variable
10780 @env{ASAN_OPTIONS}.
10782 @item asan-instrumentation-with-call-threshold
10783 If number of memory accesses in function being instrumented
10784 is greater or equal to this number, use callbacks instead of inline checks.
10785 E.g. to disable inline code use
10786 @option{--param asan-instrumentation-with-call-threshold=0}.
10788 @item use-after-scope-direct-emission-threshold
10789 If the size of a local variable in bytes is smaller or equal to this
10790 number, directly poison (or unpoison) shadow memory instead of using
10791 run-time callbacks. The default value is 256.
10793 @item chkp-max-ctor-size
10794 Static constructors generated by Pointer Bounds Checker may become very
10795 large and significantly increase compile time at optimization level
10796 @option{-O1} and higher. This parameter is a maximum number of statements
10797 in a single generated constructor. Default value is 5000.
10799 @item max-fsm-thread-path-insns
10800 Maximum number of instructions to copy when duplicating blocks on a
10801 finite state automaton jump thread path. The default is 100.
10803 @item max-fsm-thread-length
10804 Maximum number of basic blocks on a finite state automaton jump thread
10805 path. The default is 10.
10807 @item max-fsm-thread-paths
10808 Maximum number of new jump thread paths to create for a finite state
10809 automaton. The default is 50.
10811 @item parloops-chunk-size
10812 Chunk size of omp schedule for loops parallelized by parloops. The default
10815 @item parloops-schedule
10816 Schedule type of omp schedule for loops parallelized by parloops (static,
10817 dynamic, guided, auto, runtime). The default is static.
10819 @item max-ssa-name-query-depth
10820 Maximum depth of recursion when querying properties of SSA names in things
10821 like fold routines. One level of recursion corresponds to following a
10824 @item hsa-gen-debug-stores
10825 Enable emission of special debug stores within HSA kernels which are
10826 then read and reported by libgomp plugin. Generation of these stores
10827 is disabled by default, use @option{--param hsa-gen-debug-stores=1} to
10830 @item max-speculative-devirt-maydefs
10831 The maximum number of may-defs we analyze when looking for a must-def
10832 specifying the dynamic type of an object that invokes a virtual call
10833 we may be able to devirtualize speculatively.
10835 @item max-vrp-switch-assertions
10836 The maximum number of assertions to add along the default edge of a switch
10837 statement during VRP. The default is 10.
10841 @node Instrumentation Options
10842 @section Program Instrumentation Options
10843 @cindex instrumentation options
10844 @cindex program instrumentation options
10845 @cindex run-time error checking options
10846 @cindex profiling options
10847 @cindex options, program instrumentation
10848 @cindex options, run-time error checking
10849 @cindex options, profiling
10851 GCC supports a number of command-line options that control adding
10852 run-time instrumentation to the code it normally generates.
10853 For example, one purpose of instrumentation is collect profiling
10854 statistics for use in finding program hot spots, code coverage
10855 analysis, or profile-guided optimizations.
10856 Another class of program instrumentation is adding run-time checking
10857 to detect programming errors like invalid pointer
10858 dereferences or out-of-bounds array accesses, as well as deliberately
10859 hostile attacks such as stack smashing or C++ vtable hijacking.
10860 There is also a general hook which can be used to implement other
10861 forms of tracing or function-level instrumentation for debug or
10862 program analysis purposes.
10865 @cindex @command{prof}
10868 Generate extra code to write profile information suitable for the
10869 analysis program @command{prof}. You must use this option when compiling
10870 the source files you want data about, and you must also use it when
10873 @cindex @command{gprof}
10876 Generate extra code to write profile information suitable for the
10877 analysis program @command{gprof}. You must use this option when compiling
10878 the source files you want data about, and you must also use it when
10881 @item -fprofile-arcs
10882 @opindex fprofile-arcs
10883 Add code so that program flow @dfn{arcs} are instrumented. During
10884 execution the program records how many times each branch and call is
10885 executed and how many times it is taken or returns. On targets that support
10886 constructors with priority support, profiling properly handles constructors,
10887 destructors and C++ constructors (and destructors) of classes which are used
10888 as a type of a global variable.
10891 program exits it saves this data to a file called
10892 @file{@var{auxname}.gcda} for each source file. The data may be used for
10893 profile-directed optimizations (@option{-fbranch-probabilities}), or for
10894 test coverage analysis (@option{-ftest-coverage}). Each object file's
10895 @var{auxname} is generated from the name of the output file, if
10896 explicitly specified and it is not the final executable, otherwise it is
10897 the basename of the source file. In both cases any suffix is removed
10898 (e.g.@: @file{foo.gcda} for input file @file{dir/foo.c}, or
10899 @file{dir/foo.gcda} for output file specified as @option{-o dir/foo.o}).
10900 @xref{Cross-profiling}.
10902 @cindex @command{gcov}
10906 This option is used to compile and link code instrumented for coverage
10907 analysis. The option is a synonym for @option{-fprofile-arcs}
10908 @option{-ftest-coverage} (when compiling) and @option{-lgcov} (when
10909 linking). See the documentation for those options for more details.
10914 Compile the source files with @option{-fprofile-arcs} plus optimization
10915 and code generation options. For test coverage analysis, use the
10916 additional @option{-ftest-coverage} option. You do not need to profile
10917 every source file in a program.
10920 Compile the source files additionally with @option{-fprofile-abs-path}
10921 to create absolute path names in the @file{.gcno} files. This allows
10922 @command{gcov} to find the correct sources in projects where compilations
10923 occur with different working directories.
10926 Link your object files with @option{-lgcov} or @option{-fprofile-arcs}
10927 (the latter implies the former).
10930 Run the program on a representative workload to generate the arc profile
10931 information. This may be repeated any number of times. You can run
10932 concurrent instances of your program, and provided that the file system
10933 supports locking, the data files will be correctly updated. Unless
10934 a strict ISO C dialect option is in effect, @code{fork} calls are
10935 detected and correctly handled without double counting.
10938 For profile-directed optimizations, compile the source files again with
10939 the same optimization and code generation options plus
10940 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
10941 Control Optimization}).
10944 For test coverage analysis, use @command{gcov} to produce human readable
10945 information from the @file{.gcno} and @file{.gcda} files. Refer to the
10946 @command{gcov} documentation for further information.
10950 With @option{-fprofile-arcs}, for each function of your program GCC
10951 creates a program flow graph, then finds a spanning tree for the graph.
10952 Only arcs that are not on the spanning tree have to be instrumented: the
10953 compiler adds code to count the number of times that these arcs are
10954 executed. When an arc is the only exit or only entrance to a block, the
10955 instrumentation code can be added to the block; otherwise, a new basic
10956 block must be created to hold the instrumentation code.
10959 @item -ftest-coverage
10960 @opindex ftest-coverage
10961 Produce a notes file that the @command{gcov} code-coverage utility
10962 (@pxref{Gcov,, @command{gcov}---a Test Coverage Program}) can use to
10963 show program coverage. Each source file's note file is called
10964 @file{@var{auxname}.gcno}. Refer to the @option{-fprofile-arcs} option
10965 above for a description of @var{auxname} and instructions on how to
10966 generate test coverage data. Coverage data matches the source files
10967 more closely if you do not optimize.
10969 @item -fprofile-abs-path
10970 @opindex fprofile-abs-path
10971 Automatically convert relative source file names to absolute path names
10972 in the @file{.gcno} files. This allows @command{gcov} to find the correct
10973 sources in projects where compilations occur with different working
10976 @item -fprofile-dir=@var{path}
10977 @opindex fprofile-dir
10979 Set the directory to search for the profile data files in to @var{path}.
10980 This option affects only the profile data generated by
10981 @option{-fprofile-generate}, @option{-ftest-coverage}, @option{-fprofile-arcs}
10982 and used by @option{-fprofile-use} and @option{-fbranch-probabilities}
10983 and its related options. Both absolute and relative paths can be used.
10984 By default, GCC uses the current directory as @var{path}, thus the
10985 profile data file appears in the same directory as the object file.
10987 @item -fprofile-generate
10988 @itemx -fprofile-generate=@var{path}
10989 @opindex fprofile-generate
10991 Enable options usually used for instrumenting application to produce
10992 profile useful for later recompilation with profile feedback based
10993 optimization. You must use @option{-fprofile-generate} both when
10994 compiling and when linking your program.
10996 The following options are enabled: @option{-fprofile-arcs}, @option{-fprofile-values}, @option{-fvpt}.
10998 If @var{path} is specified, GCC looks at the @var{path} to find
10999 the profile feedback data files. See @option{-fprofile-dir}.
11001 To optimize the program based on the collected profile information, use
11002 @option{-fprofile-use}. @xref{Optimize Options}, for more information.
11004 @item -fprofile-update=@var{method}
11005 @opindex fprofile-update
11007 Alter the update method for an application instrumented for profile
11008 feedback based optimization. The @var{method} argument should be one of
11009 @samp{single}, @samp{atomic} or @samp{prefer-atomic}.
11010 The first one is useful for single-threaded applications,
11011 while the second one prevents profile corruption by emitting thread-safe code.
11013 @strong{Warning:} When an application does not properly join all threads
11014 (or creates an detached thread), a profile file can be still corrupted.
11016 Using @samp{prefer-atomic} would be transformed either to @samp{atomic},
11017 when supported by a target, or to @samp{single} otherwise. The GCC driver
11018 automatically selects @samp{prefer-atomic} when @option{-pthread}
11019 is present in the command line.
11021 @item -fsanitize=address
11022 @opindex fsanitize=address
11023 Enable AddressSanitizer, a fast memory error detector.
11024 Memory access instructions are instrumented to detect
11025 out-of-bounds and use-after-free bugs.
11026 The option enables @option{-fsanitize-address-use-after-scope}.
11027 See @uref{https://github.com/google/sanitizers/wiki/AddressSanitizer} for
11028 more details. The run-time behavior can be influenced using the
11029 @env{ASAN_OPTIONS} environment variable. When set to @code{help=1},
11030 the available options are shown at startup of the instrumented program. See
11031 @url{https://github.com/google/sanitizers/wiki/AddressSanitizerFlags#run-time-flags}
11032 for a list of supported options.
11033 The option cannot be combined with @option{-fsanitize=thread}
11034 and/or @option{-fcheck-pointer-bounds}.
11036 @item -fsanitize=kernel-address
11037 @opindex fsanitize=kernel-address
11038 Enable AddressSanitizer for Linux kernel.
11039 See @uref{https://github.com/google/kasan/wiki} for more details.
11040 The option cannot be combined with @option{-fcheck-pointer-bounds}.
11042 @item -fsanitize=thread
11043 @opindex fsanitize=thread
11044 Enable ThreadSanitizer, a fast data race detector.
11045 Memory access instructions are instrumented to detect
11046 data race bugs. See @uref{https://github.com/google/sanitizers/wiki#threadsanitizer} for more
11047 details. The run-time behavior can be influenced using the @env{TSAN_OPTIONS}
11048 environment variable; see
11049 @url{https://github.com/google/sanitizers/wiki/ThreadSanitizerFlags} for a list of
11051 The option cannot be combined with @option{-fsanitize=address},
11052 @option{-fsanitize=leak} and/or @option{-fcheck-pointer-bounds}.
11054 Note that sanitized atomic builtins cannot throw exceptions when
11055 operating on invalid memory addresses with non-call exceptions
11056 (@option{-fnon-call-exceptions}).
11058 @item -fsanitize=leak
11059 @opindex fsanitize=leak
11060 Enable LeakSanitizer, a memory leak detector.
11061 This option only matters for linking of executables and
11062 the executable is linked against a library that overrides @code{malloc}
11063 and other allocator functions. See
11064 @uref{https://github.com/google/sanitizers/wiki/AddressSanitizerLeakSanitizer} for more
11065 details. The run-time behavior can be influenced using the
11066 @env{LSAN_OPTIONS} environment variable.
11067 The option cannot be combined with @option{-fsanitize=thread}.
11069 @item -fsanitize=undefined
11070 @opindex fsanitize=undefined
11071 Enable UndefinedBehaviorSanitizer, a fast undefined behavior detector.
11072 Various computations are instrumented to detect undefined behavior
11073 at runtime. Current suboptions are:
11077 @item -fsanitize=shift
11078 @opindex fsanitize=shift
11079 This option enables checking that the result of a shift operation is
11080 not undefined. Note that what exactly is considered undefined differs
11081 slightly between C and C++, as well as between ISO C90 and C99, etc.
11082 This option has two suboptions, @option{-fsanitize=shift-base} and
11083 @option{-fsanitize=shift-exponent}.
11085 @item -fsanitize=shift-exponent
11086 @opindex fsanitize=shift-exponent
11087 This option enables checking that the second argument of a shift operation
11088 is not negative and is smaller than the precision of the promoted first
11091 @item -fsanitize=shift-base
11092 @opindex fsanitize=shift-base
11093 If the second argument of a shift operation is within range, check that the
11094 result of a shift operation is not undefined. Note that what exactly is
11095 considered undefined differs slightly between C and C++, as well as between
11096 ISO C90 and C99, etc.
11098 @item -fsanitize=integer-divide-by-zero
11099 @opindex fsanitize=integer-divide-by-zero
11100 Detect integer division by zero as well as @code{INT_MIN / -1} division.
11102 @item -fsanitize=unreachable
11103 @opindex fsanitize=unreachable
11104 With this option, the compiler turns the @code{__builtin_unreachable}
11105 call into a diagnostics message call instead. When reaching the
11106 @code{__builtin_unreachable} call, the behavior is undefined.
11108 @item -fsanitize=vla-bound
11109 @opindex fsanitize=vla-bound
11110 This option instructs the compiler to check that the size of a variable
11111 length array is positive.
11113 @item -fsanitize=null
11114 @opindex fsanitize=null
11115 This option enables pointer checking. Particularly, the application
11116 built with this option turned on will issue an error message when it
11117 tries to dereference a NULL pointer, or if a reference (possibly an
11118 rvalue reference) is bound to a NULL pointer, or if a method is invoked
11119 on an object pointed by a NULL pointer.
11121 @item -fsanitize=return
11122 @opindex fsanitize=return
11123 This option enables return statement checking. Programs
11124 built with this option turned on will issue an error message
11125 when the end of a non-void function is reached without actually
11126 returning a value. This option works in C++ only.
11128 @item -fsanitize=signed-integer-overflow
11129 @opindex fsanitize=signed-integer-overflow
11130 This option enables signed integer overflow checking. We check that
11131 the result of @code{+}, @code{*}, and both unary and binary @code{-}
11132 does not overflow in the signed arithmetics. Note, integer promotion
11133 rules must be taken into account. That is, the following is not an
11136 signed char a = SCHAR_MAX;
11140 @item -fsanitize=bounds
11141 @opindex fsanitize=bounds
11142 This option enables instrumentation of array bounds. Various out of bounds
11143 accesses are detected. Flexible array members, flexible array member-like
11144 arrays, and initializers of variables with static storage are not instrumented.
11145 The option cannot be combined with @option{-fcheck-pointer-bounds}.
11147 @item -fsanitize=bounds-strict
11148 @opindex fsanitize=bounds-strict
11149 This option enables strict instrumentation of array bounds. Most out of bounds
11150 accesses are detected, including flexible array members and flexible array
11151 member-like arrays. Initializers of variables with static storage are not
11152 instrumented. The option cannot be combined
11153 with @option{-fcheck-pointer-bounds}.
11155 @item -fsanitize=alignment
11156 @opindex fsanitize=alignment
11158 This option enables checking of alignment of pointers when they are
11159 dereferenced, or when a reference is bound to insufficiently aligned target,
11160 or when a method or constructor is invoked on insufficiently aligned object.
11162 @item -fsanitize=object-size
11163 @opindex fsanitize=object-size
11164 This option enables instrumentation of memory references using the
11165 @code{__builtin_object_size} function. Various out of bounds pointer
11166 accesses are detected.
11168 @item -fsanitize=float-divide-by-zero
11169 @opindex fsanitize=float-divide-by-zero
11170 Detect floating-point division by zero. Unlike other similar options,
11171 @option{-fsanitize=float-divide-by-zero} is not enabled by
11172 @option{-fsanitize=undefined}, since floating-point division by zero can
11173 be a legitimate way of obtaining infinities and NaNs.
11175 @item -fsanitize=float-cast-overflow
11176 @opindex fsanitize=float-cast-overflow
11177 This option enables floating-point type to integer conversion checking.
11178 We check that the result of the conversion does not overflow.
11179 Unlike other similar options, @option{-fsanitize=float-cast-overflow} is
11180 not enabled by @option{-fsanitize=undefined}.
11181 This option does not work well with @code{FE_INVALID} exceptions enabled.
11183 @item -fsanitize=nonnull-attribute
11184 @opindex fsanitize=nonnull-attribute
11186 This option enables instrumentation of calls, checking whether null values
11187 are not passed to arguments marked as requiring a non-null value by the
11188 @code{nonnull} function attribute.
11190 @item -fsanitize=returns-nonnull-attribute
11191 @opindex fsanitize=returns-nonnull-attribute
11193 This option enables instrumentation of return statements in functions
11194 marked with @code{returns_nonnull} function attribute, to detect returning
11195 of null values from such functions.
11197 @item -fsanitize=bool
11198 @opindex fsanitize=bool
11200 This option enables instrumentation of loads from bool. If a value other
11201 than 0/1 is loaded, a run-time error is issued.
11203 @item -fsanitize=enum
11204 @opindex fsanitize=enum
11206 This option enables instrumentation of loads from an enum type. If
11207 a value outside the range of values for the enum type is loaded,
11208 a run-time error is issued.
11210 @item -fsanitize=vptr
11211 @opindex fsanitize=vptr
11213 This option enables instrumentation of C++ member function calls, member
11214 accesses and some conversions between pointers to base and derived classes,
11215 to verify the referenced object has the correct dynamic type.
11217 @item -fsanitize=pointer-overflow
11218 @opindex fsanitize=pointer-overflow
11220 This option enables instrumentation of pointer arithmetics. If the pointer
11221 arithmetics overflows, a run-time error is issued.
11223 @item -fsanitize=builtin
11224 @opindex fsanitize=builtin
11226 This option enables instrumentation of arguments to selected builtin
11227 functions. If an invalid value is passed to such arguments, a run-time
11228 error is issued. E.g.@ passing 0 as the argument to @code{__builtin_ctz}
11229 or @code{__builtin_clz} invokes undefined behavior and is diagnosed
11234 While @option{-ftrapv} causes traps for signed overflows to be emitted,
11235 @option{-fsanitize=undefined} gives a diagnostic message.
11236 This currently works only for the C family of languages.
11238 @item -fno-sanitize=all
11239 @opindex fno-sanitize=all
11241 This option disables all previously enabled sanitizers.
11242 @option{-fsanitize=all} is not allowed, as some sanitizers cannot be used
11245 @item -fasan-shadow-offset=@var{number}
11246 @opindex fasan-shadow-offset
11247 This option forces GCC to use custom shadow offset in AddressSanitizer checks.
11248 It is useful for experimenting with different shadow memory layouts in
11249 Kernel AddressSanitizer.
11251 @item -fsanitize-sections=@var{s1},@var{s2},...
11252 @opindex fsanitize-sections
11253 Sanitize global variables in selected user-defined sections. @var{si} may
11256 @item -fsanitize-recover@r{[}=@var{opts}@r{]}
11257 @opindex fsanitize-recover
11258 @opindex fno-sanitize-recover
11259 @option{-fsanitize-recover=} controls error recovery mode for sanitizers
11260 mentioned in comma-separated list of @var{opts}. Enabling this option
11261 for a sanitizer component causes it to attempt to continue
11262 running the program as if no error happened. This means multiple
11263 runtime errors can be reported in a single program run, and the exit
11264 code of the program may indicate success even when errors
11265 have been reported. The @option{-fno-sanitize-recover=} option
11266 can be used to alter
11267 this behavior: only the first detected error is reported
11268 and program then exits with a non-zero exit code.
11270 Currently this feature only works for @option{-fsanitize=undefined} (and its suboptions
11271 except for @option{-fsanitize=unreachable} and @option{-fsanitize=return}),
11272 @option{-fsanitize=float-cast-overflow}, @option{-fsanitize=float-divide-by-zero},
11273 @option{-fsanitize=bounds-strict},
11274 @option{-fsanitize=kernel-address} and @option{-fsanitize=address}.
11275 For these sanitizers error recovery is turned on by default,
11276 except @option{-fsanitize=address}, for which this feature is experimental.
11277 @option{-fsanitize-recover=all} and @option{-fno-sanitize-recover=all} is also
11278 accepted, the former enables recovery for all sanitizers that support it,
11279 the latter disables recovery for all sanitizers that support it.
11281 Even if a recovery mode is turned on the compiler side, it needs to be also
11282 enabled on the runtime library side, otherwise the failures are still fatal.
11283 The runtime library defaults to @code{halt_on_error=0} for
11284 ThreadSanitizer and UndefinedBehaviorSanitizer, while default value for
11285 AddressSanitizer is @code{halt_on_error=1}. This can be overridden through
11286 setting the @code{halt_on_error} flag in the corresponding environment variable.
11288 Syntax without an explicit @var{opts} parameter is deprecated. It is
11289 equivalent to specifying an @var{opts} list of:
11292 undefined,float-cast-overflow,float-divide-by-zero,bounds-strict
11295 @item -fsanitize-address-use-after-scope
11296 @opindex fsanitize-address-use-after-scope
11297 Enable sanitization of local variables to detect use-after-scope bugs.
11298 The option sets @option{-fstack-reuse} to @samp{none}.
11300 @item -fsanitize-undefined-trap-on-error
11301 @opindex fsanitize-undefined-trap-on-error
11302 The @option{-fsanitize-undefined-trap-on-error} option instructs the compiler to
11303 report undefined behavior using @code{__builtin_trap} rather than
11304 a @code{libubsan} library routine. The advantage of this is that the
11305 @code{libubsan} library is not needed and is not linked in, so this
11306 is usable even in freestanding environments.
11308 @item -fsanitize-coverage=trace-pc
11309 @opindex fsanitize-coverage=trace-pc
11310 Enable coverage-guided fuzzing code instrumentation.
11311 Inserts a call to @code{__sanitizer_cov_trace_pc} into every basic block.
11313 @item -fsanitize-coverage=trace-cmp
11314 @opindex fsanitize-coverage=trace-cmp
11315 Enable dataflow guided fuzzing code instrumentation.
11316 Inserts a call to @code{__sanitizer_cov_trace_cmp1},
11317 @code{__sanitizer_cov_trace_cmp2}, @code{__sanitizer_cov_trace_cmp4} or
11318 @code{__sanitizer_cov_trace_cmp8} for integral comparison with both operands
11319 variable or @code{__sanitizer_cov_trace_const_cmp1},
11320 @code{__sanitizer_cov_trace_const_cmp2},
11321 @code{__sanitizer_cov_trace_const_cmp4} or
11322 @code{__sanitizer_cov_trace_const_cmp8} for integral comparison with one
11323 operand constant, @code{__sanitizer_cov_trace_cmpf} or
11324 @code{__sanitizer_cov_trace_cmpd} for float or double comparisons and
11325 @code{__sanitizer_cov_trace_switch} for switch statements.
11327 @item -fbounds-check
11328 @opindex fbounds-check
11329 For front ends that support it, generate additional code to check that
11330 indices used to access arrays are within the declared range. This is
11331 currently only supported by the Fortran front end, where this option
11334 @item -fcheck-pointer-bounds
11335 @opindex fcheck-pointer-bounds
11336 @opindex fno-check-pointer-bounds
11337 @cindex Pointer Bounds Checker options
11338 Enable Pointer Bounds Checker instrumentation. Each memory reference
11339 is instrumented with checks of the pointer used for memory access against
11340 bounds associated with that pointer.
11343 is only an implementation for Intel MPX available, thus x86 GNU/Linux target
11344 and @option{-mmpx} are required to enable this feature.
11345 MPX-based instrumentation requires
11346 a runtime library to enable MPX in hardware and handle bounds
11347 violation signals. By default when @option{-fcheck-pointer-bounds}
11348 and @option{-mmpx} options are used to link a program, the GCC driver
11349 links against the @file{libmpx} and @file{libmpxwrappers} libraries.
11350 Bounds checking on calls to dynamic libraries requires a linker
11351 with @option{-z bndplt} support; if GCC was configured with a linker
11352 without support for this option (including the Gold linker and older
11353 versions of ld), a warning is given if you link with @option{-mmpx}
11354 without also specifying @option{-static}, since the overall effectiveness
11355 of the bounds checking protection is reduced.
11356 See also @option{-static-libmpxwrappers}.
11358 MPX-based instrumentation
11359 may be used for debugging and also may be included in production code
11360 to increase program security. Depending on usage, you may
11361 have different requirements for the runtime library. The current version
11362 of the MPX runtime library is more oriented for use as a debugging
11363 tool. MPX runtime library usage implies @option{-lpthread}. See
11364 also @option{-static-libmpx}. The runtime library behavior can be
11365 influenced using various @env{CHKP_RT_*} environment variables. See
11366 @uref{https://gcc.gnu.org/wiki/Intel%20MPX%20support%20in%20the%20GCC%20compiler}
11369 Generated instrumentation may be controlled by various
11370 @option{-fchkp-*} options and by the @code{bnd_variable_size}
11371 structure field attribute (@pxref{Type Attributes}) and
11372 @code{bnd_legacy}, and @code{bnd_instrument} function attributes
11373 (@pxref{Function Attributes}). GCC also provides a number of built-in
11374 functions for controlling the Pointer Bounds Checker. @xref{Pointer
11375 Bounds Checker builtins}, for more information.
11377 @item -fchkp-check-incomplete-type
11378 @opindex fchkp-check-incomplete-type
11379 @opindex fno-chkp-check-incomplete-type
11380 Generate pointer bounds checks for variables with incomplete type.
11381 Enabled by default.
11383 @item -fchkp-narrow-bounds
11384 @opindex fchkp-narrow-bounds
11385 @opindex fno-chkp-narrow-bounds
11386 Controls bounds used by Pointer Bounds Checker for pointers to object
11387 fields. If narrowing is enabled then field bounds are used. Otherwise
11388 object bounds are used. See also @option{-fchkp-narrow-to-innermost-array}
11389 and @option{-fchkp-first-field-has-own-bounds}. Enabled by default.
11391 @item -fchkp-first-field-has-own-bounds
11392 @opindex fchkp-first-field-has-own-bounds
11393 @opindex fno-chkp-first-field-has-own-bounds
11394 Forces Pointer Bounds Checker to use narrowed bounds for the address of the
11395 first field in the structure. By default a pointer to the first field has
11396 the same bounds as a pointer to the whole structure.
11398 @item -fchkp-flexible-struct-trailing-arrays
11399 @opindex fchkp-flexible-struct-trailing-arrays
11400 @opindex fno-chkp-flexible-struct-trailing-arrays
11401 Forces Pointer Bounds Checker to treat all trailing arrays in structures as
11402 possibly flexible. By default only array fields with zero length or that are
11403 marked with attribute bnd_variable_size are treated as flexible.
11405 @item -fchkp-narrow-to-innermost-array
11406 @opindex fchkp-narrow-to-innermost-array
11407 @opindex fno-chkp-narrow-to-innermost-array
11408 Forces Pointer Bounds Checker to use bounds of the innermost arrays in
11409 case of nested static array access. By default this option is disabled and
11410 bounds of the outermost array are used.
11412 @item -fchkp-optimize
11413 @opindex fchkp-optimize
11414 @opindex fno-chkp-optimize
11415 Enables Pointer Bounds Checker optimizations. Enabled by default at
11416 optimization levels @option{-O}, @option{-O2}, @option{-O3}.
11418 @item -fchkp-use-fast-string-functions
11419 @opindex fchkp-use-fast-string-functions
11420 @opindex fno-chkp-use-fast-string-functions
11421 Enables use of @code{*_nobnd} versions of string functions (not copying bounds)
11422 by Pointer Bounds Checker. Disabled by default.
11424 @item -fchkp-use-nochk-string-functions
11425 @opindex fchkp-use-nochk-string-functions
11426 @opindex fno-chkp-use-nochk-string-functions
11427 Enables use of @code{*_nochk} versions of string functions (not checking bounds)
11428 by Pointer Bounds Checker. Disabled by default.
11430 @item -fchkp-use-static-bounds
11431 @opindex fchkp-use-static-bounds
11432 @opindex fno-chkp-use-static-bounds
11433 Allow Pointer Bounds Checker to generate static bounds holding
11434 bounds of static variables. Enabled by default.
11436 @item -fchkp-use-static-const-bounds
11437 @opindex fchkp-use-static-const-bounds
11438 @opindex fno-chkp-use-static-const-bounds
11439 Use statically-initialized bounds for constant bounds instead of
11440 generating them each time they are required. By default enabled when
11441 @option{-fchkp-use-static-bounds} is enabled.
11443 @item -fchkp-treat-zero-dynamic-size-as-infinite
11444 @opindex fchkp-treat-zero-dynamic-size-as-infinite
11445 @opindex fno-chkp-treat-zero-dynamic-size-as-infinite
11446 With this option, objects with incomplete type whose
11447 dynamically-obtained size is zero are treated as having infinite size
11448 instead by Pointer Bounds
11449 Checker. This option may be helpful if a program is linked with a library
11450 missing size information for some symbols. Disabled by default.
11452 @item -fchkp-check-read
11453 @opindex fchkp-check-read
11454 @opindex fno-chkp-check-read
11455 Instructs Pointer Bounds Checker to generate checks for all read
11456 accesses to memory. Enabled by default.
11458 @item -fchkp-check-write
11459 @opindex fchkp-check-write
11460 @opindex fno-chkp-check-write
11461 Instructs Pointer Bounds Checker to generate checks for all write
11462 accesses to memory. Enabled by default.
11464 @item -fchkp-store-bounds
11465 @opindex fchkp-store-bounds
11466 @opindex fno-chkp-store-bounds
11467 Instructs Pointer Bounds Checker to generate bounds stores for
11468 pointer writes. Enabled by default.
11470 @item -fchkp-instrument-calls
11471 @opindex fchkp-instrument-calls
11472 @opindex fno-chkp-instrument-calls
11473 Instructs Pointer Bounds Checker to pass pointer bounds to calls.
11474 Enabled by default.
11476 @item -fchkp-instrument-marked-only
11477 @opindex fchkp-instrument-marked-only
11478 @opindex fno-chkp-instrument-marked-only
11479 Instructs Pointer Bounds Checker to instrument only functions
11480 marked with the @code{bnd_instrument} attribute
11481 (@pxref{Function Attributes}). Disabled by default.
11483 @item -fchkp-use-wrappers
11484 @opindex fchkp-use-wrappers
11485 @opindex fno-chkp-use-wrappers
11486 Allows Pointer Bounds Checker to replace calls to built-in functions
11487 with calls to wrapper functions. When @option{-fchkp-use-wrappers}
11488 is used to link a program, the GCC driver automatically links
11489 against @file{libmpxwrappers}. See also @option{-static-libmpxwrappers}.
11490 Enabled by default.
11492 @item -fcf-protection==@r{[}full@r{|}branch@r{|}return@r{|}none@r{]}
11493 @opindex fcf-protection
11494 Enable code instrumentation of control-flow transfers to increase
11495 program security by checking that target addresses of control-flow
11496 transfer instructions (such as indirect function call, function return,
11497 indirect jump) are valid. This prevents diverting the flow of control
11498 to an unexpected target. This is intended to protect against such
11499 threats as Return-oriented Programming (ROP), and similarly
11500 call/jmp-oriented programming (COP/JOP).
11502 The value @code{branch} tells the compiler to implement checking of
11503 validity of control-flow transfer at the point of indirect branch
11504 instructions, i.e. call/jmp instructions. The value @code{return}
11505 implements checking of validity at the point of returning from a
11506 function. The value @code{full} is an alias for specifying both
11507 @code{branch} and @code{return}. The value @code{none} turns off
11510 You can also use the @code{nocf_check} attribute to identify
11511 which functions and calls should be skipped from instrumentation
11512 (@pxref{Function Attributes}).
11514 Currently the x86 GNU/Linux target provides an implementation based
11515 on Intel Control-flow Enforcement Technology (CET). Instrumentation
11516 for x86 is controlled by target-specific options @option{-mcet},
11517 @option{-mibt} and @option{-mshstk} (@pxref{x86 Options}).
11519 @item -fstack-protector
11520 @opindex fstack-protector
11521 Emit extra code to check for buffer overflows, such as stack smashing
11522 attacks. This is done by adding a guard variable to functions with
11523 vulnerable objects. This includes functions that call @code{alloca}, and
11524 functions with buffers larger than 8 bytes. The guards are initialized
11525 when a function is entered and then checked when the function exits.
11526 If a guard check fails, an error message is printed and the program exits.
11528 @item -fstack-protector-all
11529 @opindex fstack-protector-all
11530 Like @option{-fstack-protector} except that all functions are protected.
11532 @item -fstack-protector-strong
11533 @opindex fstack-protector-strong
11534 Like @option{-fstack-protector} but includes additional functions to
11535 be protected --- those that have local array definitions, or have
11536 references to local frame addresses.
11538 @item -fstack-protector-explicit
11539 @opindex fstack-protector-explicit
11540 Like @option{-fstack-protector} but only protects those functions which
11541 have the @code{stack_protect} attribute.
11543 @item -fstack-check
11544 @opindex fstack-check
11545 Generate code to verify that you do not go beyond the boundary of the
11546 stack. You should specify this flag if you are running in an
11547 environment with multiple threads, but you only rarely need to specify it in
11548 a single-threaded environment since stack overflow is automatically
11549 detected on nearly all systems if there is only one stack.
11551 Note that this switch does not actually cause checking to be done; the
11552 operating system or the language runtime must do that. The switch causes
11553 generation of code to ensure that they see the stack being extended.
11555 You can additionally specify a string parameter: @samp{no} means no
11556 checking, @samp{generic} means force the use of old-style checking,
11557 @samp{specific} means use the best checking method and is equivalent
11558 to bare @option{-fstack-check}.
11560 Old-style checking is a generic mechanism that requires no specific
11561 target support in the compiler but comes with the following drawbacks:
11565 Modified allocation strategy for large objects: they are always
11566 allocated dynamically if their size exceeds a fixed threshold. Note this
11567 may change the semantics of some code.
11570 Fixed limit on the size of the static frame of functions: when it is
11571 topped by a particular function, stack checking is not reliable and
11572 a warning is issued by the compiler.
11575 Inefficiency: because of both the modified allocation strategy and the
11576 generic implementation, code performance is hampered.
11579 Note that old-style stack checking is also the fallback method for
11580 @samp{specific} if no target support has been added in the compiler.
11582 @samp{-fstack-check=} is designed for Ada's needs to detect infinite recursion
11583 and stack overflows. @samp{specific} is an excellent choice when compiling
11584 Ada code. It is not generally sufficient to protect against stack-clash
11585 attacks. To protect against those you want @samp{-fstack-clash-protection}.
11587 @item -fstack-clash-protection
11588 @opindex fstack-clash-protection
11589 Generate code to prevent stack clash style attacks. When this option is
11590 enabled, the compiler will only allocate one page of stack space at a time
11591 and each page is accessed immediately after allocation. Thus, it prevents
11592 allocations from jumping over any stack guard page provided by the
11595 Most targets do not fully support stack clash protection. However, on
11596 those targets @option{-fstack-clash-protection} will protect dynamic stack
11597 allocations. @option{-fstack-clash-protection} may also provide limited
11598 protection for static stack allocations if the target supports
11599 @option{-fstack-check=specific}.
11601 @item -fstack-limit-register=@var{reg}
11602 @itemx -fstack-limit-symbol=@var{sym}
11603 @itemx -fno-stack-limit
11604 @opindex fstack-limit-register
11605 @opindex fstack-limit-symbol
11606 @opindex fno-stack-limit
11607 Generate code to ensure that the stack does not grow beyond a certain value,
11608 either the value of a register or the address of a symbol. If a larger
11609 stack is required, a signal is raised at run time. For most targets,
11610 the signal is raised before the stack overruns the boundary, so
11611 it is possible to catch the signal without taking special precautions.
11613 For instance, if the stack starts at absolute address @samp{0x80000000}
11614 and grows downwards, you can use the flags
11615 @option{-fstack-limit-symbol=__stack_limit} and
11616 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
11617 of 128KB@. Note that this may only work with the GNU linker.
11619 You can locally override stack limit checking by using the
11620 @code{no_stack_limit} function attribute (@pxref{Function Attributes}).
11622 @item -fsplit-stack
11623 @opindex fsplit-stack
11624 Generate code to automatically split the stack before it overflows.
11625 The resulting program has a discontiguous stack which can only
11626 overflow if the program is unable to allocate any more memory. This
11627 is most useful when running threaded programs, as it is no longer
11628 necessary to calculate a good stack size to use for each thread. This
11629 is currently only implemented for the x86 targets running
11632 When code compiled with @option{-fsplit-stack} calls code compiled
11633 without @option{-fsplit-stack}, there may not be much stack space
11634 available for the latter code to run. If compiling all code,
11635 including library code, with @option{-fsplit-stack} is not an option,
11636 then the linker can fix up these calls so that the code compiled
11637 without @option{-fsplit-stack} always has a large stack. Support for
11638 this is implemented in the gold linker in GNU binutils release 2.21
11641 @item -fvtable-verify=@r{[}std@r{|}preinit@r{|}none@r{]}
11642 @opindex fvtable-verify
11643 This option is only available when compiling C++ code.
11644 It turns on (or off, if using @option{-fvtable-verify=none}) the security
11645 feature that verifies at run time, for every virtual call, that
11646 the vtable pointer through which the call is made is valid for the type of
11647 the object, and has not been corrupted or overwritten. If an invalid vtable
11648 pointer is detected at run time, an error is reported and execution of the
11649 program is immediately halted.
11651 This option causes run-time data structures to be built at program startup,
11652 which are used for verifying the vtable pointers.
11653 The options @samp{std} and @samp{preinit}
11654 control the timing of when these data structures are built. In both cases the
11655 data structures are built before execution reaches @code{main}. Using
11656 @option{-fvtable-verify=std} causes the data structures to be built after
11657 shared libraries have been loaded and initialized.
11658 @option{-fvtable-verify=preinit} causes them to be built before shared
11659 libraries have been loaded and initialized.
11661 If this option appears multiple times in the command line with different
11662 values specified, @samp{none} takes highest priority over both @samp{std} and
11663 @samp{preinit}; @samp{preinit} takes priority over @samp{std}.
11666 @opindex fvtv-debug
11667 When used in conjunction with @option{-fvtable-verify=std} or
11668 @option{-fvtable-verify=preinit}, causes debug versions of the
11669 runtime functions for the vtable verification feature to be called.
11670 This flag also causes the compiler to log information about which
11671 vtable pointers it finds for each class.
11672 This information is written to a file named @file{vtv_set_ptr_data.log}
11673 in the directory named by the environment variable @env{VTV_LOGS_DIR}
11674 if that is defined or the current working directory otherwise.
11676 Note: This feature @emph{appends} data to the log file. If you want a fresh log
11677 file, be sure to delete any existing one.
11680 @opindex fvtv-counts
11681 This is a debugging flag. When used in conjunction with
11682 @option{-fvtable-verify=std} or @option{-fvtable-verify=preinit}, this
11683 causes the compiler to keep track of the total number of virtual calls
11684 it encounters and the number of verifications it inserts. It also
11685 counts the number of calls to certain run-time library functions
11686 that it inserts and logs this information for each compilation unit.
11687 The compiler writes this information to a file named
11688 @file{vtv_count_data.log} in the directory named by the environment
11689 variable @env{VTV_LOGS_DIR} if that is defined or the current working
11690 directory otherwise. It also counts the size of the vtable pointer sets
11691 for each class, and writes this information to @file{vtv_class_set_sizes.log}
11692 in the same directory.
11694 Note: This feature @emph{appends} data to the log files. To get fresh log
11695 files, be sure to delete any existing ones.
11697 @item -finstrument-functions
11698 @opindex finstrument-functions
11699 Generate instrumentation calls for entry and exit to functions. Just
11700 after function entry and just before function exit, the following
11701 profiling functions are called with the address of the current
11702 function and its call site. (On some platforms,
11703 @code{__builtin_return_address} does not work beyond the current
11704 function, so the call site information may not be available to the
11705 profiling functions otherwise.)
11708 void __cyg_profile_func_enter (void *this_fn,
11710 void __cyg_profile_func_exit (void *this_fn,
11714 The first argument is the address of the start of the current function,
11715 which may be looked up exactly in the symbol table.
11717 This instrumentation is also done for functions expanded inline in other
11718 functions. The profiling calls indicate where, conceptually, the
11719 inline function is entered and exited. This means that addressable
11720 versions of such functions must be available. If all your uses of a
11721 function are expanded inline, this may mean an additional expansion of
11722 code size. If you use @code{extern inline} in your C code, an
11723 addressable version of such functions must be provided. (This is
11724 normally the case anyway, but if you get lucky and the optimizer always
11725 expands the functions inline, you might have gotten away without
11726 providing static copies.)
11728 A function may be given the attribute @code{no_instrument_function}, in
11729 which case this instrumentation is not done. This can be used, for
11730 example, for the profiling functions listed above, high-priority
11731 interrupt routines, and any functions from which the profiling functions
11732 cannot safely be called (perhaps signal handlers, if the profiling
11733 routines generate output or allocate memory).
11735 @item -finstrument-functions-exclude-file-list=@var{file},@var{file},@dots{}
11736 @opindex finstrument-functions-exclude-file-list
11738 Set the list of functions that are excluded from instrumentation (see
11739 the description of @option{-finstrument-functions}). If the file that
11740 contains a function definition matches with one of @var{file}, then
11741 that function is not instrumented. The match is done on substrings:
11742 if the @var{file} parameter is a substring of the file name, it is
11743 considered to be a match.
11748 -finstrument-functions-exclude-file-list=/bits/stl,include/sys
11752 excludes any inline function defined in files whose pathnames
11753 contain @file{/bits/stl} or @file{include/sys}.
11755 If, for some reason, you want to include letter @samp{,} in one of
11756 @var{sym}, write @samp{\,}. For example,
11757 @option{-finstrument-functions-exclude-file-list='\,\,tmp'}
11758 (note the single quote surrounding the option).
11760 @item -finstrument-functions-exclude-function-list=@var{sym},@var{sym},@dots{}
11761 @opindex finstrument-functions-exclude-function-list
11763 This is similar to @option{-finstrument-functions-exclude-file-list},
11764 but this option sets the list of function names to be excluded from
11765 instrumentation. The function name to be matched is its user-visible
11766 name, such as @code{vector<int> blah(const vector<int> &)}, not the
11767 internal mangled name (e.g., @code{_Z4blahRSt6vectorIiSaIiEE}). The
11768 match is done on substrings: if the @var{sym} parameter is a substring
11769 of the function name, it is considered to be a match. For C99 and C++
11770 extended identifiers, the function name must be given in UTF-8, not
11771 using universal character names.
11773 @item -fpatchable-function-entry=@var{N}[,@var{M}]
11774 @opindex fpatchable-function-entry
11775 Generate @var{N} NOPs right at the beginning
11776 of each function, with the function entry point before the @var{M}th NOP.
11777 If @var{M} is omitted, it defaults to @code{0} so the
11778 function entry points to the address just at the first NOP.
11779 The NOP instructions reserve extra space which can be used to patch in
11780 any desired instrumentation at run time, provided that the code segment
11781 is writable. The amount of space is controllable indirectly via
11782 the number of NOPs; the NOP instruction used corresponds to the instruction
11783 emitted by the internal GCC back-end interface @code{gen_nop}. This behavior
11784 is target-specific and may also depend on the architecture variant and/or
11785 other compilation options.
11787 For run-time identification, the starting addresses of these areas,
11788 which correspond to their respective function entries minus @var{M},
11789 are additionally collected in the @code{__patchable_function_entries}
11790 section of the resulting binary.
11792 Note that the value of @code{__attribute__ ((patchable_function_entry
11793 (N,M)))} takes precedence over command-line option
11794 @option{-fpatchable-function-entry=N,M}. This can be used to increase
11795 the area size or to remove it completely on a single function.
11796 If @code{N=0}, no pad location is recorded.
11798 The NOP instructions are inserted at---and maybe before, depending on
11799 @var{M}---the function entry address, even before the prologue.
11804 @node Preprocessor Options
11805 @section Options Controlling the Preprocessor
11806 @cindex preprocessor options
11807 @cindex options, preprocessor
11809 These options control the C preprocessor, which is run on each C source
11810 file before actual compilation.
11812 If you use the @option{-E} option, nothing is done except preprocessing.
11813 Some of these options make sense only together with @option{-E} because
11814 they cause the preprocessor output to be unsuitable for actual
11817 In addition to the options listed here, there are a number of options
11818 to control search paths for include files documented in
11819 @ref{Directory Options}.
11820 Options to control preprocessor diagnostics are listed in
11821 @ref{Warning Options}.
11824 @include cppopts.texi
11826 @item -Wp,@var{option}
11828 You can use @option{-Wp,@var{option}} to bypass the compiler driver
11829 and pass @var{option} directly through to the preprocessor. If
11830 @var{option} contains commas, it is split into multiple options at the
11831 commas. However, many options are modified, translated or interpreted
11832 by the compiler driver before being passed to the preprocessor, and
11833 @option{-Wp} forcibly bypasses this phase. The preprocessor's direct
11834 interface is undocumented and subject to change, so whenever possible
11835 you should avoid using @option{-Wp} and let the driver handle the
11838 @item -Xpreprocessor @var{option}
11839 @opindex Xpreprocessor
11840 Pass @var{option} as an option to the preprocessor. You can use this to
11841 supply system-specific preprocessor options that GCC does not
11844 If you want to pass an option that takes an argument, you must use
11845 @option{-Xpreprocessor} twice, once for the option and once for the argument.
11847 @item -no-integrated-cpp
11848 @opindex no-integrated-cpp
11849 Perform preprocessing as a separate pass before compilation.
11850 By default, GCC performs preprocessing as an integrated part of
11851 input tokenization and parsing.
11852 If this option is provided, the appropriate language front end
11853 (@command{cc1}, @command{cc1plus}, or @command{cc1obj} for C, C++,
11854 and Objective-C, respectively) is instead invoked twice,
11855 once for preprocessing only and once for actual compilation
11856 of the preprocessed input.
11857 This option may be useful in conjunction with the @option{-B} or
11858 @option{-wrapper} options to specify an alternate preprocessor or
11859 perform additional processing of the program source between
11860 normal preprocessing and compilation.
11864 @node Assembler Options
11865 @section Passing Options to the Assembler
11867 @c prevent bad page break with this line
11868 You can pass options to the assembler.
11871 @item -Wa,@var{option}
11873 Pass @var{option} as an option to the assembler. If @var{option}
11874 contains commas, it is split into multiple options at the commas.
11876 @item -Xassembler @var{option}
11877 @opindex Xassembler
11878 Pass @var{option} as an option to the assembler. You can use this to
11879 supply system-specific assembler options that GCC does not
11882 If you want to pass an option that takes an argument, you must use
11883 @option{-Xassembler} twice, once for the option and once for the argument.
11888 @section Options for Linking
11889 @cindex link options
11890 @cindex options, linking
11892 These options come into play when the compiler links object files into
11893 an executable output file. They are meaningless if the compiler is
11894 not doing a link step.
11898 @item @var{object-file-name}
11899 A file name that does not end in a special recognized suffix is
11900 considered to name an object file or library. (Object files are
11901 distinguished from libraries by the linker according to the file
11902 contents.) If linking is done, these object files are used as input
11911 If any of these options is used, then the linker is not run, and
11912 object file names should not be used as arguments. @xref{Overall
11916 @opindex fuse-ld=bfd
11917 Use the @command{bfd} linker instead of the default linker.
11919 @item -fuse-ld=gold
11920 @opindex fuse-ld=gold
11921 Use the @command{gold} linker instead of the default linker.
11924 @item -l@var{library}
11925 @itemx -l @var{library}
11927 Search the library named @var{library} when linking. (The second
11928 alternative with the library as a separate argument is only for
11929 POSIX compliance and is not recommended.)
11931 It makes a difference where in the command you write this option; the
11932 linker searches and processes libraries and object files in the order they
11933 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
11934 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
11935 to functions in @samp{z}, those functions may not be loaded.
11937 The linker searches a standard list of directories for the library,
11938 which is actually a file named @file{lib@var{library}.a}. The linker
11939 then uses this file as if it had been specified precisely by name.
11941 The directories searched include several standard system directories
11942 plus any that you specify with @option{-L}.
11944 Normally the files found this way are library files---archive files
11945 whose members are object files. The linker handles an archive file by
11946 scanning through it for members which define symbols that have so far
11947 been referenced but not defined. But if the file that is found is an
11948 ordinary object file, it is linked in the usual fashion. The only
11949 difference between using an @option{-l} option and specifying a file name
11950 is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a}
11951 and searches several directories.
11955 You need this special case of the @option{-l} option in order to
11956 link an Objective-C or Objective-C++ program.
11958 @item -nostartfiles
11959 @opindex nostartfiles
11960 Do not use the standard system startup files when linking.
11961 The standard system libraries are used normally, unless @option{-nostdlib}
11962 or @option{-nodefaultlibs} is used.
11964 @item -nodefaultlibs
11965 @opindex nodefaultlibs
11966 Do not use the standard system libraries when linking.
11967 Only the libraries you specify are passed to the linker, and options
11968 specifying linkage of the system libraries, such as @option{-static-libgcc}
11969 or @option{-shared-libgcc}, are ignored.
11970 The standard startup files are used normally, unless @option{-nostartfiles}
11973 The compiler may generate calls to @code{memcmp},
11974 @code{memset}, @code{memcpy} and @code{memmove}.
11975 These entries are usually resolved by entries in
11976 libc. These entry points should be supplied through some other
11977 mechanism when this option is specified.
11981 Do not use the standard system startup files or libraries when linking.
11982 No startup files and only the libraries you specify are passed to
11983 the linker, and options specifying linkage of the system libraries, such as
11984 @option{-static-libgcc} or @option{-shared-libgcc}, are ignored.
11986 The compiler may generate calls to @code{memcmp}, @code{memset},
11987 @code{memcpy} and @code{memmove}.
11988 These entries are usually resolved by entries in
11989 libc. These entry points should be supplied through some other
11990 mechanism when this option is specified.
11992 @cindex @option{-lgcc}, use with @option{-nostdlib}
11993 @cindex @option{-nostdlib} and unresolved references
11994 @cindex unresolved references and @option{-nostdlib}
11995 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
11996 @cindex @option{-nodefaultlibs} and unresolved references
11997 @cindex unresolved references and @option{-nodefaultlibs}
11998 One of the standard libraries bypassed by @option{-nostdlib} and
11999 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
12000 which GCC uses to overcome shortcomings of particular machines, or special
12001 needs for some languages.
12002 (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler
12003 Collection (GCC) Internals},
12004 for more discussion of @file{libgcc.a}.)
12005 In most cases, you need @file{libgcc.a} even when you want to avoid
12006 other standard libraries. In other words, when you specify @option{-nostdlib}
12007 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
12008 This ensures that you have no unresolved references to internal GCC
12009 library subroutines.
12010 (An example of such an internal subroutine is @code{__main}, used to ensure C++
12011 constructors are called; @pxref{Collect2,,@code{collect2}, gccint,
12012 GNU Compiler Collection (GCC) Internals}.)
12016 Produce a dynamically linked position independent executable on targets
12017 that support it. For predictable results, you must also specify the same
12018 set of options used for compilation (@option{-fpie}, @option{-fPIE},
12019 or model suboptions) when you specify this linker option.
12023 Don't produce a dynamically linked position independent executable.
12026 @opindex static-pie
12027 Produce a static position independent executable on targets that support
12028 it. A static position independent executable is similar to a static
12029 executable, but can be loaded at any address without a dynamic linker.
12030 For predictable results, you must also specify the same set of options
12031 used for compilation (@option{-fpie}, @option{-fPIE}, or model
12032 suboptions) when you specify this linker option.
12036 Link with the POSIX threads library. This option is supported on
12037 GNU/Linux targets, most other Unix derivatives, and also on
12038 x86 Cygwin and MinGW targets. On some targets this option also sets
12039 flags for the preprocessor, so it should be used consistently for both
12040 compilation and linking.
12044 Pass the flag @option{-export-dynamic} to the ELF linker, on targets
12045 that support it. This instructs the linker to add all symbols, not
12046 only used ones, to the dynamic symbol table. This option is needed
12047 for some uses of @code{dlopen} or to allow obtaining backtraces
12048 from within a program.
12052 Remove all symbol table and relocation information from the executable.
12056 On systems that support dynamic linking, this overrides @option{-pie}
12057 and prevents linking with the shared libraries. On other systems, this
12058 option has no effect.
12062 Produce a shared object which can then be linked with other objects to
12063 form an executable. Not all systems support this option. For predictable
12064 results, you must also specify the same set of options used for compilation
12065 (@option{-fpic}, @option{-fPIC}, or model suboptions) when
12066 you specify this linker option.@footnote{On some systems, @samp{gcc -shared}
12067 needs to build supplementary stub code for constructors to work. On
12068 multi-libbed systems, @samp{gcc -shared} must select the correct support
12069 libraries to link against. Failing to supply the correct flags may lead
12070 to subtle defects. Supplying them in cases where they are not necessary
12073 @item -shared-libgcc
12074 @itemx -static-libgcc
12075 @opindex shared-libgcc
12076 @opindex static-libgcc
12077 On systems that provide @file{libgcc} as a shared library, these options
12078 force the use of either the shared or static version, respectively.
12079 If no shared version of @file{libgcc} was built when the compiler was
12080 configured, these options have no effect.
12082 There are several situations in which an application should use the
12083 shared @file{libgcc} instead of the static version. The most common
12084 of these is when the application wishes to throw and catch exceptions
12085 across different shared libraries. In that case, each of the libraries
12086 as well as the application itself should use the shared @file{libgcc}.
12088 Therefore, the G++ and driver automatically adds @option{-shared-libgcc}
12089 whenever you build a shared library or a main executable, because C++
12090 programs typically use exceptions, so this is the right thing to do.
12092 If, instead, you use the GCC driver to create shared libraries, you may
12093 find that they are not always linked with the shared @file{libgcc}.
12094 If GCC finds, at its configuration time, that you have a non-GNU linker
12095 or a GNU linker that does not support option @option{--eh-frame-hdr},
12096 it links the shared version of @file{libgcc} into shared libraries
12097 by default. Otherwise, it takes advantage of the linker and optimizes
12098 away the linking with the shared version of @file{libgcc}, linking with
12099 the static version of libgcc by default. This allows exceptions to
12100 propagate through such shared libraries, without incurring relocation
12101 costs at library load time.
12103 However, if a library or main executable is supposed to throw or catch
12104 exceptions, you must link it using the G++ driver, as appropriate
12105 for the languages used in the program, or using the option
12106 @option{-shared-libgcc}, such that it is linked with the shared
12109 @item -static-libasan
12110 @opindex static-libasan
12111 When the @option{-fsanitize=address} option is used to link a program,
12112 the GCC driver automatically links against @option{libasan}. If
12113 @file{libasan} is available as a shared library, and the @option{-static}
12114 option is not used, then this links against the shared version of
12115 @file{libasan}. The @option{-static-libasan} option directs the GCC
12116 driver to link @file{libasan} statically, without necessarily linking
12117 other libraries statically.
12119 @item -static-libtsan
12120 @opindex static-libtsan
12121 When the @option{-fsanitize=thread} option is used to link a program,
12122 the GCC driver automatically links against @option{libtsan}. If
12123 @file{libtsan} is available as a shared library, and the @option{-static}
12124 option is not used, then this links against the shared version of
12125 @file{libtsan}. The @option{-static-libtsan} option directs the GCC
12126 driver to link @file{libtsan} statically, without necessarily linking
12127 other libraries statically.
12129 @item -static-liblsan
12130 @opindex static-liblsan
12131 When the @option{-fsanitize=leak} option is used to link a program,
12132 the GCC driver automatically links against @option{liblsan}. If
12133 @file{liblsan} is available as a shared library, and the @option{-static}
12134 option is not used, then this links against the shared version of
12135 @file{liblsan}. The @option{-static-liblsan} option directs the GCC
12136 driver to link @file{liblsan} statically, without necessarily linking
12137 other libraries statically.
12139 @item -static-libubsan
12140 @opindex static-libubsan
12141 When the @option{-fsanitize=undefined} option is used to link a program,
12142 the GCC driver automatically links against @option{libubsan}. If
12143 @file{libubsan} is available as a shared library, and the @option{-static}
12144 option is not used, then this links against the shared version of
12145 @file{libubsan}. The @option{-static-libubsan} option directs the GCC
12146 driver to link @file{libubsan} statically, without necessarily linking
12147 other libraries statically.
12149 @item -static-libmpx
12150 @opindex static-libmpx
12151 When the @option{-fcheck-pointer bounds} and @option{-mmpx} options are
12152 used to link a program, the GCC driver automatically links against
12153 @file{libmpx}. If @file{libmpx} is available as a shared library,
12154 and the @option{-static} option is not used, then this links against
12155 the shared version of @file{libmpx}. The @option{-static-libmpx}
12156 option directs the GCC driver to link @file{libmpx} statically,
12157 without necessarily linking other libraries statically.
12159 @item -static-libmpxwrappers
12160 @opindex static-libmpxwrappers
12161 When the @option{-fcheck-pointer bounds} and @option{-mmpx} options are used
12162 to link a program without also using @option{-fno-chkp-use-wrappers}, the
12163 GCC driver automatically links against @file{libmpxwrappers}. If
12164 @file{libmpxwrappers} is available as a shared library, and the
12165 @option{-static} option is not used, then this links against the shared
12166 version of @file{libmpxwrappers}. The @option{-static-libmpxwrappers}
12167 option directs the GCC driver to link @file{libmpxwrappers} statically,
12168 without necessarily linking other libraries statically.
12170 @item -static-libstdc++
12171 @opindex static-libstdc++
12172 When the @command{g++} program is used to link a C++ program, it
12173 normally automatically links against @option{libstdc++}. If
12174 @file{libstdc++} is available as a shared library, and the
12175 @option{-static} option is not used, then this links against the
12176 shared version of @file{libstdc++}. That is normally fine. However, it
12177 is sometimes useful to freeze the version of @file{libstdc++} used by
12178 the program without going all the way to a fully static link. The
12179 @option{-static-libstdc++} option directs the @command{g++} driver to
12180 link @file{libstdc++} statically, without necessarily linking other
12181 libraries statically.
12185 Bind references to global symbols when building a shared object. Warn
12186 about any unresolved references (unless overridden by the link editor
12187 option @option{-Xlinker -z -Xlinker defs}). Only a few systems support
12190 @item -T @var{script}
12192 @cindex linker script
12193 Use @var{script} as the linker script. This option is supported by most
12194 systems using the GNU linker. On some targets, such as bare-board
12195 targets without an operating system, the @option{-T} option may be required
12196 when linking to avoid references to undefined symbols.
12198 @item -Xlinker @var{option}
12200 Pass @var{option} as an option to the linker. You can use this to
12201 supply system-specific linker options that GCC does not recognize.
12203 If you want to pass an option that takes a separate argument, you must use
12204 @option{-Xlinker} twice, once for the option and once for the argument.
12205 For example, to pass @option{-assert definitions}, you must write
12206 @option{-Xlinker -assert -Xlinker definitions}. It does not work to write
12207 @option{-Xlinker "-assert definitions"}, because this passes the entire
12208 string as a single argument, which is not what the linker expects.
12210 When using the GNU linker, it is usually more convenient to pass
12211 arguments to linker options using the @option{@var{option}=@var{value}}
12212 syntax than as separate arguments. For example, you can specify
12213 @option{-Xlinker -Map=output.map} rather than
12214 @option{-Xlinker -Map -Xlinker output.map}. Other linkers may not support
12215 this syntax for command-line options.
12217 @item -Wl,@var{option}
12219 Pass @var{option} as an option to the linker. If @var{option} contains
12220 commas, it is split into multiple options at the commas. You can use this
12221 syntax to pass an argument to the option.
12222 For example, @option{-Wl,-Map,output.map} passes @option{-Map output.map} to the
12223 linker. When using the GNU linker, you can also get the same effect with
12224 @option{-Wl,-Map=output.map}.
12226 @item -u @var{symbol}
12228 Pretend the symbol @var{symbol} is undefined, to force linking of
12229 library modules to define it. You can use @option{-u} multiple times with
12230 different symbols to force loading of additional library modules.
12232 @item -z @var{keyword}
12234 @option{-z} is passed directly on to the linker along with the keyword
12235 @var{keyword}. See the section in the documentation of your linker for
12236 permitted values and their meanings.
12239 @node Directory Options
12240 @section Options for Directory Search
12241 @cindex directory options
12242 @cindex options, directory search
12243 @cindex search path
12245 These options specify directories to search for header files, for
12246 libraries and for parts of the compiler:
12249 @include cppdiropts.texi
12251 @item -iplugindir=@var{dir}
12252 @opindex iplugindir=
12253 Set the directory to search for plugins that are passed
12254 by @option{-fplugin=@var{name}} instead of
12255 @option{-fplugin=@var{path}/@var{name}.so}. This option is not meant
12256 to be used by the user, but only passed by the driver.
12260 Add directory @var{dir} to the list of directories to be searched
12263 @item -B@var{prefix}
12265 This option specifies where to find the executables, libraries,
12266 include files, and data files of the compiler itself.
12268 The compiler driver program runs one or more of the subprograms
12269 @command{cpp}, @command{cc1}, @command{as} and @command{ld}. It tries
12270 @var{prefix} as a prefix for each program it tries to run, both with and
12271 without @samp{@var{machine}/@var{version}/} for the corresponding target
12272 machine and compiler version.
12274 For each subprogram to be run, the compiler driver first tries the
12275 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
12276 is not specified, the driver tries two standard prefixes,
12277 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc/}. If neither of
12278 those results in a file name that is found, the unmodified program
12279 name is searched for using the directories specified in your
12280 @env{PATH} environment variable.
12282 The compiler checks to see if the path provided by @option{-B}
12283 refers to a directory, and if necessary it adds a directory
12284 separator character at the end of the path.
12286 @option{-B} prefixes that effectively specify directory names also apply
12287 to libraries in the linker, because the compiler translates these
12288 options into @option{-L} options for the linker. They also apply to
12289 include files in the preprocessor, because the compiler translates these
12290 options into @option{-isystem} options for the preprocessor. In this case,
12291 the compiler appends @samp{include} to the prefix.
12293 The runtime support file @file{libgcc.a} can also be searched for using
12294 the @option{-B} prefix, if needed. If it is not found there, the two
12295 standard prefixes above are tried, and that is all. The file is left
12296 out of the link if it is not found by those means.
12298 Another way to specify a prefix much like the @option{-B} prefix is to use
12299 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
12302 As a special kludge, if the path provided by @option{-B} is
12303 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
12304 9, then it is replaced by @file{[dir/]include}. This is to help
12305 with boot-strapping the compiler.
12307 @item -no-canonical-prefixes
12308 @opindex no-canonical-prefixes
12309 Do not expand any symbolic links, resolve references to @samp{/../}
12310 or @samp{/./}, or make the path absolute when generating a relative
12313 @item --sysroot=@var{dir}
12315 Use @var{dir} as the logical root directory for headers and libraries.
12316 For example, if the compiler normally searches for headers in
12317 @file{/usr/include} and libraries in @file{/usr/lib}, it instead
12318 searches @file{@var{dir}/usr/include} and @file{@var{dir}/usr/lib}.
12320 If you use both this option and the @option{-isysroot} option, then
12321 the @option{--sysroot} option applies to libraries, but the
12322 @option{-isysroot} option applies to header files.
12324 The GNU linker (beginning with version 2.16) has the necessary support
12325 for this option. If your linker does not support this option, the
12326 header file aspect of @option{--sysroot} still works, but the
12327 library aspect does not.
12329 @item --no-sysroot-suffix
12330 @opindex no-sysroot-suffix
12331 For some targets, a suffix is added to the root directory specified
12332 with @option{--sysroot}, depending on the other options used, so that
12333 headers may for example be found in
12334 @file{@var{dir}/@var{suffix}/usr/include} instead of
12335 @file{@var{dir}/usr/include}. This option disables the addition of
12340 @node Code Gen Options
12341 @section Options for Code Generation Conventions
12342 @cindex code generation conventions
12343 @cindex options, code generation
12344 @cindex run-time options
12346 These machine-independent options control the interface conventions
12347 used in code generation.
12349 Most of them have both positive and negative forms; the negative form
12350 of @option{-ffoo} is @option{-fno-foo}. In the table below, only
12351 one of the forms is listed---the one that is not the default. You
12352 can figure out the other form by either removing @samp{no-} or adding
12356 @item -fstack-reuse=@var{reuse-level}
12357 @opindex fstack_reuse
12358 This option controls stack space reuse for user declared local/auto variables
12359 and compiler generated temporaries. @var{reuse_level} can be @samp{all},
12360 @samp{named_vars}, or @samp{none}. @samp{all} enables stack reuse for all
12361 local variables and temporaries, @samp{named_vars} enables the reuse only for
12362 user defined local variables with names, and @samp{none} disables stack reuse
12363 completely. The default value is @samp{all}. The option is needed when the
12364 program extends the lifetime of a scoped local variable or a compiler generated
12365 temporary beyond the end point defined by the language. When a lifetime of
12366 a variable ends, and if the variable lives in memory, the optimizing compiler
12367 has the freedom to reuse its stack space with other temporaries or scoped
12368 local variables whose live range does not overlap with it. Legacy code extending
12369 local lifetime is likely to break with the stack reuse optimization.
12388 if (*p == 10) // out of scope use of local1
12399 A(int k) : i(k), j(k) @{ @}
12406 void foo(const A& ar)
12413 foo(A(10)); // temp object's lifetime ends when foo returns
12419 ap->i+= 10; // ap references out of scope temp whose space
12420 // is reused with a. What is the value of ap->i?
12425 The lifetime of a compiler generated temporary is well defined by the C++
12426 standard. When a lifetime of a temporary ends, and if the temporary lives
12427 in memory, the optimizing compiler has the freedom to reuse its stack
12428 space with other temporaries or scoped local variables whose live range
12429 does not overlap with it. However some of the legacy code relies on
12430 the behavior of older compilers in which temporaries' stack space is
12431 not reused, the aggressive stack reuse can lead to runtime errors. This
12432 option is used to control the temporary stack reuse optimization.
12436 This option generates traps for signed overflow on addition, subtraction,
12437 multiplication operations.
12438 The options @option{-ftrapv} and @option{-fwrapv} override each other, so using
12439 @option{-ftrapv} @option{-fwrapv} on the command-line results in
12440 @option{-fwrapv} being effective. Note that only active options override, so
12441 using @option{-ftrapv} @option{-fwrapv} @option{-fno-wrapv} on the command-line
12442 results in @option{-ftrapv} being effective.
12446 This option instructs the compiler to assume that signed arithmetic
12447 overflow of addition, subtraction and multiplication wraps around
12448 using twos-complement representation. This flag enables some optimizations
12449 and disables others.
12450 The options @option{-ftrapv} and @option{-fwrapv} override each other, so using
12451 @option{-ftrapv} @option{-fwrapv} on the command-line results in
12452 @option{-fwrapv} being effective. Note that only active options override, so
12453 using @option{-ftrapv} @option{-fwrapv} @option{-fno-wrapv} on the command-line
12454 results in @option{-ftrapv} being effective.
12457 @opindex fexceptions
12458 Enable exception handling. Generates extra code needed to propagate
12459 exceptions. For some targets, this implies GCC generates frame
12460 unwind information for all functions, which can produce significant data
12461 size overhead, although it does not affect execution. If you do not
12462 specify this option, GCC enables it by default for languages like
12463 C++ that normally require exception handling, and disables it for
12464 languages like C that do not normally require it. However, you may need
12465 to enable this option when compiling C code that needs to interoperate
12466 properly with exception handlers written in C++. You may also wish to
12467 disable this option if you are compiling older C++ programs that don't
12468 use exception handling.
12470 @item -fnon-call-exceptions
12471 @opindex fnon-call-exceptions
12472 Generate code that allows trapping instructions to throw exceptions.
12473 Note that this requires platform-specific runtime support that does
12474 not exist everywhere. Moreover, it only allows @emph{trapping}
12475 instructions to throw exceptions, i.e.@: memory references or floating-point
12476 instructions. It does not allow exceptions to be thrown from
12477 arbitrary signal handlers such as @code{SIGALRM}.
12479 @item -fdelete-dead-exceptions
12480 @opindex fdelete-dead-exceptions
12481 Consider that instructions that may throw exceptions but don't otherwise
12482 contribute to the execution of the program can be optimized away.
12483 This option is enabled by default for the Ada front end, as permitted by
12484 the Ada language specification.
12485 Optimization passes that cause dead exceptions to be removed are enabled independently at different optimization levels.
12487 @item -funwind-tables
12488 @opindex funwind-tables
12489 Similar to @option{-fexceptions}, except that it just generates any needed
12490 static data, but does not affect the generated code in any other way.
12491 You normally do not need to enable this option; instead, a language processor
12492 that needs this handling enables it on your behalf.
12494 @item -fasynchronous-unwind-tables
12495 @opindex fasynchronous-unwind-tables
12496 Generate unwind table in DWARF format, if supported by target machine. The
12497 table is exact at each instruction boundary, so it can be used for stack
12498 unwinding from asynchronous events (such as debugger or garbage collector).
12500 @item -fno-gnu-unique
12501 @opindex fno-gnu-unique
12502 On systems with recent GNU assembler and C library, the C++ compiler
12503 uses the @code{STB_GNU_UNIQUE} binding to make sure that definitions
12504 of template static data members and static local variables in inline
12505 functions are unique even in the presence of @code{RTLD_LOCAL}; this
12506 is necessary to avoid problems with a library used by two different
12507 @code{RTLD_LOCAL} plugins depending on a definition in one of them and
12508 therefore disagreeing with the other one about the binding of the
12509 symbol. But this causes @code{dlclose} to be ignored for affected
12510 DSOs; if your program relies on reinitialization of a DSO via
12511 @code{dlclose} and @code{dlopen}, you can use
12512 @option{-fno-gnu-unique}.
12514 @item -fpcc-struct-return
12515 @opindex fpcc-struct-return
12516 Return ``short'' @code{struct} and @code{union} values in memory like
12517 longer ones, rather than in registers. This convention is less
12518 efficient, but it has the advantage of allowing intercallability between
12519 GCC-compiled files and files compiled with other compilers, particularly
12520 the Portable C Compiler (pcc).
12522 The precise convention for returning structures in memory depends
12523 on the target configuration macros.
12525 Short structures and unions are those whose size and alignment match
12526 that of some integer type.
12528 @strong{Warning:} code compiled with the @option{-fpcc-struct-return}
12529 switch is not binary compatible with code compiled with the
12530 @option{-freg-struct-return} switch.
12531 Use it to conform to a non-default application binary interface.
12533 @item -freg-struct-return
12534 @opindex freg-struct-return
12535 Return @code{struct} and @code{union} values in registers when possible.
12536 This is more efficient for small structures than
12537 @option{-fpcc-struct-return}.
12539 If you specify neither @option{-fpcc-struct-return} nor
12540 @option{-freg-struct-return}, GCC defaults to whichever convention is
12541 standard for the target. If there is no standard convention, GCC
12542 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
12543 the principal compiler. In those cases, we can choose the standard, and
12544 we chose the more efficient register return alternative.
12546 @strong{Warning:} code compiled with the @option{-freg-struct-return}
12547 switch is not binary compatible with code compiled with the
12548 @option{-fpcc-struct-return} switch.
12549 Use it to conform to a non-default application binary interface.
12551 @item -fshort-enums
12552 @opindex fshort-enums
12553 Allocate to an @code{enum} type only as many bytes as it needs for the
12554 declared range of possible values. Specifically, the @code{enum} type
12555 is equivalent to the smallest integer type that has enough room.
12557 @strong{Warning:} the @option{-fshort-enums} switch causes GCC to generate
12558 code that is not binary compatible with code generated without that switch.
12559 Use it to conform to a non-default application binary interface.
12561 @item -fshort-wchar
12562 @opindex fshort-wchar
12563 Override the underlying type for @code{wchar_t} to be @code{short
12564 unsigned int} instead of the default for the target. This option is
12565 useful for building programs to run under WINE@.
12567 @strong{Warning:} the @option{-fshort-wchar} switch causes GCC to generate
12568 code that is not binary compatible with code generated without that switch.
12569 Use it to conform to a non-default application binary interface.
12572 @opindex fno-common
12573 @cindex tentative definitions
12574 In C code, this option controls the placement of global variables
12575 defined without an initializer, known as @dfn{tentative definitions}
12576 in the C standard. Tentative definitions are distinct from declarations
12577 of a variable with the @code{extern} keyword, which do not allocate storage.
12579 Unix C compilers have traditionally allocated storage for
12580 uninitialized global variables in a common block. This allows the
12581 linker to resolve all tentative definitions of the same variable
12582 in different compilation units to the same object, or to a non-tentative
12584 This is the behavior specified by @option{-fcommon}, and is the default for
12585 GCC on most targets.
12586 On the other hand, this behavior is not required by ISO
12587 C, and on some targets may carry a speed or code size penalty on
12588 variable references.
12590 The @option{-fno-common} option specifies that the compiler should instead
12591 place uninitialized global variables in the data section of the object file.
12592 This inhibits the merging of tentative definitions by the linker so
12593 you get a multiple-definition error if the same
12594 variable is defined in more than one compilation unit.
12595 Compiling with @option{-fno-common} is useful on targets for which
12596 it provides better performance, or if you wish to verify that the
12597 program will work on other systems that always treat uninitialized
12598 variable definitions this way.
12602 Ignore the @code{#ident} directive.
12604 @item -finhibit-size-directive
12605 @opindex finhibit-size-directive
12606 Don't output a @code{.size} assembler directive, or anything else that
12607 would cause trouble if the function is split in the middle, and the
12608 two halves are placed at locations far apart in memory. This option is
12609 used when compiling @file{crtstuff.c}; you should not need to use it
12612 @item -fverbose-asm
12613 @opindex fverbose-asm
12614 Put extra commentary information in the generated assembly code to
12615 make it more readable. This option is generally only of use to those
12616 who actually need to read the generated assembly code (perhaps while
12617 debugging the compiler itself).
12619 @option{-fno-verbose-asm}, the default, causes the
12620 extra information to be omitted and is useful when comparing two assembler
12623 The added comments include:
12628 information on the compiler version and command-line options,
12631 the source code lines associated with the assembly instructions,
12632 in the form FILENAME:LINENUMBER:CONTENT OF LINE,
12635 hints on which high-level expressions correspond to
12636 the various assembly instruction operands.
12640 For example, given this C source file:
12648 for (i = 0; i < n; i++)
12655 compiling to (x86_64) assembly via @option{-S} and emitting the result
12656 direct to stdout via @option{-o} @option{-}
12659 gcc -S test.c -fverbose-asm -Os -o -
12662 gives output similar to this:
12666 # GNU C11 (GCC) version 7.0.0 20160809 (experimental) (x86_64-pc-linux-gnu)
12673 .type test, @@function
12677 # test.c:4: int total = 0;
12678 xorl %eax, %eax # <retval>
12679 # test.c:6: for (i = 0; i < n; i++)
12680 xorl %edx, %edx # i
12682 # test.c:6: for (i = 0; i < n; i++)
12683 cmpl %edi, %edx # n, i
12685 # test.c:7: total += i * i;
12686 movl %edx, %ecx # i, tmp92
12687 imull %edx, %ecx # i, tmp92
12688 # test.c:6: for (i = 0; i < n; i++)
12690 # test.c:7: total += i * i;
12691 addl %ecx, %eax # tmp92, <retval>
12699 .ident "GCC: (GNU) 7.0.0 20160809 (experimental)"
12700 .section .note.GNU-stack,"",@@progbits
12703 The comments are intended for humans rather than machines and hence the
12704 precise format of the comments is subject to change.
12706 @item -frecord-gcc-switches
12707 @opindex frecord-gcc-switches
12708 This switch causes the command line used to invoke the
12709 compiler to be recorded into the object file that is being created.
12710 This switch is only implemented on some targets and the exact format
12711 of the recording is target and binary file format dependent, but it
12712 usually takes the form of a section containing ASCII text. This
12713 switch is related to the @option{-fverbose-asm} switch, but that
12714 switch only records information in the assembler output file as
12715 comments, so it never reaches the object file.
12716 See also @option{-grecord-gcc-switches} for another
12717 way of storing compiler options into the object file.
12721 @cindex global offset table
12723 Generate position-independent code (PIC) suitable for use in a shared
12724 library, if supported for the target machine. Such code accesses all
12725 constant addresses through a global offset table (GOT)@. The dynamic
12726 loader resolves the GOT entries when the program starts (the dynamic
12727 loader is not part of GCC; it is part of the operating system). If
12728 the GOT size for the linked executable exceeds a machine-specific
12729 maximum size, you get an error message from the linker indicating that
12730 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
12731 instead. (These maximums are 8k on the SPARC, 28k on AArch64 and 32k
12732 on the m68k and RS/6000. The x86 has no such limit.)
12734 Position-independent code requires special support, and therefore works
12735 only on certain machines. For the x86, GCC supports PIC for System V
12736 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
12737 position-independent.
12739 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
12744 If supported for the target machine, emit position-independent code,
12745 suitable for dynamic linking and avoiding any limit on the size of the
12746 global offset table. This option makes a difference on AArch64, m68k,
12747 PowerPC and SPARC@.
12749 Position-independent code requires special support, and therefore works
12750 only on certain machines.
12752 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
12759 These options are similar to @option{-fpic} and @option{-fPIC}, but
12760 generated position independent code can be only linked into executables.
12761 Usually these options are used when @option{-pie} GCC option is
12762 used during linking.
12764 @option{-fpie} and @option{-fPIE} both define the macros
12765 @code{__pie__} and @code{__PIE__}. The macros have the value 1
12766 for @option{-fpie} and 2 for @option{-fPIE}.
12770 Do not use the PLT for external function calls in position-independent code.
12771 Instead, load the callee address at call sites from the GOT and branch to it.
12772 This leads to more efficient code by eliminating PLT stubs and exposing
12773 GOT loads to optimizations. On architectures such as 32-bit x86 where
12774 PLT stubs expect the GOT pointer in a specific register, this gives more
12775 register allocation freedom to the compiler.
12776 Lazy binding requires use of the PLT;
12777 with @option{-fno-plt} all external symbols are resolved at load time.
12779 Alternatively, the function attribute @code{noplt} can be used to avoid calls
12780 through the PLT for specific external functions.
12782 In position-dependent code, a few targets also convert calls to
12783 functions that are marked to not use the PLT to use the GOT instead.
12785 @item -fno-jump-tables
12786 @opindex fno-jump-tables
12787 Do not use jump tables for switch statements even where it would be
12788 more efficient than other code generation strategies. This option is
12789 of use in conjunction with @option{-fpic} or @option{-fPIC} for
12790 building code that forms part of a dynamic linker and cannot
12791 reference the address of a jump table. On some targets, jump tables
12792 do not require a GOT and this option is not needed.
12794 @item -ffixed-@var{reg}
12796 Treat the register named @var{reg} as a fixed register; generated code
12797 should never refer to it (except perhaps as a stack pointer, frame
12798 pointer or in some other fixed role).
12800 @var{reg} must be the name of a register. The register names accepted
12801 are machine-specific and are defined in the @code{REGISTER_NAMES}
12802 macro in the machine description macro file.
12804 This flag does not have a negative form, because it specifies a
12807 @item -fcall-used-@var{reg}
12808 @opindex fcall-used
12809 Treat the register named @var{reg} as an allocable register that is
12810 clobbered by function calls. It may be allocated for temporaries or
12811 variables that do not live across a call. Functions compiled this way
12812 do not save and restore the register @var{reg}.
12814 It is an error to use this flag with the frame pointer or stack pointer.
12815 Use of this flag for other registers that have fixed pervasive roles in
12816 the machine's execution model produces disastrous results.
12818 This flag does not have a negative form, because it specifies a
12821 @item -fcall-saved-@var{reg}
12822 @opindex fcall-saved
12823 Treat the register named @var{reg} as an allocable register saved by
12824 functions. It may be allocated even for temporaries or variables that
12825 live across a call. Functions compiled this way save and restore
12826 the register @var{reg} if they use it.
12828 It is an error to use this flag with the frame pointer or stack pointer.
12829 Use of this flag for other registers that have fixed pervasive roles in
12830 the machine's execution model produces disastrous results.
12832 A different sort of disaster results from the use of this flag for
12833 a register in which function values may be returned.
12835 This flag does not have a negative form, because it specifies a
12838 @item -fpack-struct[=@var{n}]
12839 @opindex fpack-struct
12840 Without a value specified, pack all structure members together without
12841 holes. When a value is specified (which must be a small power of two), pack
12842 structure members according to this value, representing the maximum
12843 alignment (that is, objects with default alignment requirements larger than
12844 this are output potentially unaligned at the next fitting location.
12846 @strong{Warning:} the @option{-fpack-struct} switch causes GCC to generate
12847 code that is not binary compatible with code generated without that switch.
12848 Additionally, it makes the code suboptimal.
12849 Use it to conform to a non-default application binary interface.
12851 @item -fleading-underscore
12852 @opindex fleading-underscore
12853 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
12854 change the way C symbols are represented in the object file. One use
12855 is to help link with legacy assembly code.
12857 @strong{Warning:} the @option{-fleading-underscore} switch causes GCC to
12858 generate code that is not binary compatible with code generated without that
12859 switch. Use it to conform to a non-default application binary interface.
12860 Not all targets provide complete support for this switch.
12862 @item -ftls-model=@var{model}
12863 @opindex ftls-model
12864 Alter the thread-local storage model to be used (@pxref{Thread-Local}).
12865 The @var{model} argument should be one of @samp{global-dynamic},
12866 @samp{local-dynamic}, @samp{initial-exec} or @samp{local-exec}.
12867 Note that the choice is subject to optimization: the compiler may use
12868 a more efficient model for symbols not visible outside of the translation
12869 unit, or if @option{-fpic} is not given on the command line.
12871 The default without @option{-fpic} is @samp{initial-exec}; with
12872 @option{-fpic} the default is @samp{global-dynamic}.
12874 @item -ftrampolines
12875 @opindex ftrampolines
12876 For targets that normally need trampolines for nested functions, always
12877 generate them instead of using descriptors. Otherwise, for targets that
12878 do not need them, like for example HP-PA or IA-64, do nothing.
12880 A trampoline is a small piece of code that is created at run time on the
12881 stack when the address of a nested function is taken, and is used to call
12882 the nested function indirectly. Therefore, it requires the stack to be
12883 made executable in order for the program to work properly.
12885 @option{-fno-trampolines} is enabled by default on a language by language
12886 basis to let the compiler avoid generating them, if it computes that this
12887 is safe, and replace them with descriptors. Descriptors are made up of data
12888 only, but the generated code must be prepared to deal with them. As of this
12889 writing, @option{-fno-trampolines} is enabled by default only for Ada.
12891 Moreover, code compiled with @option{-ftrampolines} and code compiled with
12892 @option{-fno-trampolines} are not binary compatible if nested functions are
12893 present. This option must therefore be used on a program-wide basis and be
12894 manipulated with extreme care.
12896 @item -fvisibility=@r{[}default@r{|}internal@r{|}hidden@r{|}protected@r{]}
12897 @opindex fvisibility
12898 Set the default ELF image symbol visibility to the specified option---all
12899 symbols are marked with this unless overridden within the code.
12900 Using this feature can very substantially improve linking and
12901 load times of shared object libraries, produce more optimized
12902 code, provide near-perfect API export and prevent symbol clashes.
12903 It is @strong{strongly} recommended that you use this in any shared objects
12906 Despite the nomenclature, @samp{default} always means public; i.e.,
12907 available to be linked against from outside the shared object.
12908 @samp{protected} and @samp{internal} are pretty useless in real-world
12909 usage so the only other commonly used option is @samp{hidden}.
12910 The default if @option{-fvisibility} isn't specified is
12911 @samp{default}, i.e., make every symbol public.
12913 A good explanation of the benefits offered by ensuring ELF
12914 symbols have the correct visibility is given by ``How To Write
12915 Shared Libraries'' by Ulrich Drepper (which can be found at
12916 @w{@uref{https://www.akkadia.org/drepper/}})---however a superior
12917 solution made possible by this option to marking things hidden when
12918 the default is public is to make the default hidden and mark things
12919 public. This is the norm with DLLs on Windows and with @option{-fvisibility=hidden}
12920 and @code{__attribute__ ((visibility("default")))} instead of
12921 @code{__declspec(dllexport)} you get almost identical semantics with
12922 identical syntax. This is a great boon to those working with
12923 cross-platform projects.
12925 For those adding visibility support to existing code, you may find
12926 @code{#pragma GCC visibility} of use. This works by you enclosing
12927 the declarations you wish to set visibility for with (for example)
12928 @code{#pragma GCC visibility push(hidden)} and
12929 @code{#pragma GCC visibility pop}.
12930 Bear in mind that symbol visibility should be viewed @strong{as
12931 part of the API interface contract} and thus all new code should
12932 always specify visibility when it is not the default; i.e., declarations
12933 only for use within the local DSO should @strong{always} be marked explicitly
12934 as hidden as so to avoid PLT indirection overheads---making this
12935 abundantly clear also aids readability and self-documentation of the code.
12936 Note that due to ISO C++ specification requirements, @code{operator new} and
12937 @code{operator delete} must always be of default visibility.
12939 Be aware that headers from outside your project, in particular system
12940 headers and headers from any other library you use, may not be
12941 expecting to be compiled with visibility other than the default. You
12942 may need to explicitly say @code{#pragma GCC visibility push(default)}
12943 before including any such headers.
12945 @code{extern} declarations are not affected by @option{-fvisibility}, so
12946 a lot of code can be recompiled with @option{-fvisibility=hidden} with
12947 no modifications. However, this means that calls to @code{extern}
12948 functions with no explicit visibility use the PLT, so it is more
12949 effective to use @code{__attribute ((visibility))} and/or
12950 @code{#pragma GCC visibility} to tell the compiler which @code{extern}
12951 declarations should be treated as hidden.
12953 Note that @option{-fvisibility} does affect C++ vague linkage
12954 entities. This means that, for instance, an exception class that is
12955 be thrown between DSOs must be explicitly marked with default
12956 visibility so that the @samp{type_info} nodes are unified between
12959 An overview of these techniques, their benefits and how to use them
12960 is at @uref{http://gcc.gnu.org/@/wiki/@/Visibility}.
12962 @item -fstrict-volatile-bitfields
12963 @opindex fstrict-volatile-bitfields
12964 This option should be used if accesses to volatile bit-fields (or other
12965 structure fields, although the compiler usually honors those types
12966 anyway) should use a single access of the width of the
12967 field's type, aligned to a natural alignment if possible. For
12968 example, targets with memory-mapped peripheral registers might require
12969 all such accesses to be 16 bits wide; with this flag you can
12970 declare all peripheral bit-fields as @code{unsigned short} (assuming short
12971 is 16 bits on these targets) to force GCC to use 16-bit accesses
12972 instead of, perhaps, a more efficient 32-bit access.
12974 If this option is disabled, the compiler uses the most efficient
12975 instruction. In the previous example, that might be a 32-bit load
12976 instruction, even though that accesses bytes that do not contain
12977 any portion of the bit-field, or memory-mapped registers unrelated to
12978 the one being updated.
12980 In some cases, such as when the @code{packed} attribute is applied to a
12981 structure field, it may not be possible to access the field with a single
12982 read or write that is correctly aligned for the target machine. In this
12983 case GCC falls back to generating multiple accesses rather than code that
12984 will fault or truncate the result at run time.
12986 Note: Due to restrictions of the C/C++11 memory model, write accesses are
12987 not allowed to touch non bit-field members. It is therefore recommended
12988 to define all bits of the field's type as bit-field members.
12990 The default value of this option is determined by the application binary
12991 interface for the target processor.
12993 @item -fsync-libcalls
12994 @opindex fsync-libcalls
12995 This option controls whether any out-of-line instance of the @code{__sync}
12996 family of functions may be used to implement the C++11 @code{__atomic}
12997 family of functions.
12999 The default value of this option is enabled, thus the only useful form
13000 of the option is @option{-fno-sync-libcalls}. This option is used in
13001 the implementation of the @file{libatomic} runtime library.
13005 @node Developer Options
13006 @section GCC Developer Options
13007 @cindex developer options
13008 @cindex debugging GCC
13009 @cindex debug dump options
13010 @cindex dump options
13011 @cindex compilation statistics
13013 This section describes command-line options that are primarily of
13014 interest to GCC developers, including options to support compiler
13015 testing and investigation of compiler bugs and compile-time
13016 performance problems. This includes options that produce debug dumps
13017 at various points in the compilation; that print statistics such as
13018 memory use and execution time; and that print information about GCC's
13019 configuration, such as where it searches for libraries. You should
13020 rarely need to use any of these options for ordinary compilation and
13025 @item -d@var{letters}
13026 @itemx -fdump-rtl-@var{pass}
13027 @itemx -fdump-rtl-@var{pass}=@var{filename}
13029 @opindex fdump-rtl-@var{pass}
13030 Says to make debugging dumps during compilation at times specified by
13031 @var{letters}. This is used for debugging the RTL-based passes of the
13032 compiler. The file names for most of the dumps are made by appending
13033 a pass number and a word to the @var{dumpname}, and the files are
13034 created in the directory of the output file. In case of
13035 @option{=@var{filename}} option, the dump is output on the given file
13036 instead of the pass numbered dump files. Note that the pass number is
13037 assigned as passes are registered into the pass manager. Most passes
13038 are registered in the order that they will execute and for these passes
13039 the number corresponds to the pass execution order. However, passes
13040 registered by plugins, passes specific to compilation targets, or
13041 passes that are otherwise registered after all the other passes are
13042 numbered higher than a pass named "final", even if they are executed
13043 earlier. @var{dumpname} is generated from the name of the output
13044 file if explicitly specified and not an executable, otherwise it is
13045 the basename of the source file.
13047 Some @option{-d@var{letters}} switches have different meaning when
13048 @option{-E} is used for preprocessing. @xref{Preprocessor Options},
13049 for information about preprocessor-specific dump options.
13051 Debug dumps can be enabled with a @option{-fdump-rtl} switch or some
13052 @option{-d} option @var{letters}. Here are the possible
13053 letters for use in @var{pass} and @var{letters}, and their meanings:
13057 @item -fdump-rtl-alignments
13058 @opindex fdump-rtl-alignments
13059 Dump after branch alignments have been computed.
13061 @item -fdump-rtl-asmcons
13062 @opindex fdump-rtl-asmcons
13063 Dump after fixing rtl statements that have unsatisfied in/out constraints.
13065 @item -fdump-rtl-auto_inc_dec
13066 @opindex fdump-rtl-auto_inc_dec
13067 Dump after auto-inc-dec discovery. This pass is only run on
13068 architectures that have auto inc or auto dec instructions.
13070 @item -fdump-rtl-barriers
13071 @opindex fdump-rtl-barriers
13072 Dump after cleaning up the barrier instructions.
13074 @item -fdump-rtl-bbpart
13075 @opindex fdump-rtl-bbpart
13076 Dump after partitioning hot and cold basic blocks.
13078 @item -fdump-rtl-bbro
13079 @opindex fdump-rtl-bbro
13080 Dump after block reordering.
13082 @item -fdump-rtl-btl1
13083 @itemx -fdump-rtl-btl2
13084 @opindex fdump-rtl-btl2
13085 @opindex fdump-rtl-btl2
13086 @option{-fdump-rtl-btl1} and @option{-fdump-rtl-btl2} enable dumping
13087 after the two branch
13088 target load optimization passes.
13090 @item -fdump-rtl-bypass
13091 @opindex fdump-rtl-bypass
13092 Dump after jump bypassing and control flow optimizations.
13094 @item -fdump-rtl-combine
13095 @opindex fdump-rtl-combine
13096 Dump after the RTL instruction combination pass.
13098 @item -fdump-rtl-compgotos
13099 @opindex fdump-rtl-compgotos
13100 Dump after duplicating the computed gotos.
13102 @item -fdump-rtl-ce1
13103 @itemx -fdump-rtl-ce2
13104 @itemx -fdump-rtl-ce3
13105 @opindex fdump-rtl-ce1
13106 @opindex fdump-rtl-ce2
13107 @opindex fdump-rtl-ce3
13108 @option{-fdump-rtl-ce1}, @option{-fdump-rtl-ce2}, and
13109 @option{-fdump-rtl-ce3} enable dumping after the three
13110 if conversion passes.
13112 @item -fdump-rtl-cprop_hardreg
13113 @opindex fdump-rtl-cprop_hardreg
13114 Dump after hard register copy propagation.
13116 @item -fdump-rtl-csa
13117 @opindex fdump-rtl-csa
13118 Dump after combining stack adjustments.
13120 @item -fdump-rtl-cse1
13121 @itemx -fdump-rtl-cse2
13122 @opindex fdump-rtl-cse1
13123 @opindex fdump-rtl-cse2
13124 @option{-fdump-rtl-cse1} and @option{-fdump-rtl-cse2} enable dumping after
13125 the two common subexpression elimination passes.
13127 @item -fdump-rtl-dce
13128 @opindex fdump-rtl-dce
13129 Dump after the standalone dead code elimination passes.
13131 @item -fdump-rtl-dbr
13132 @opindex fdump-rtl-dbr
13133 Dump after delayed branch scheduling.
13135 @item -fdump-rtl-dce1
13136 @itemx -fdump-rtl-dce2
13137 @opindex fdump-rtl-dce1
13138 @opindex fdump-rtl-dce2
13139 @option{-fdump-rtl-dce1} and @option{-fdump-rtl-dce2} enable dumping after
13140 the two dead store elimination passes.
13142 @item -fdump-rtl-eh
13143 @opindex fdump-rtl-eh
13144 Dump after finalization of EH handling code.
13146 @item -fdump-rtl-eh_ranges
13147 @opindex fdump-rtl-eh_ranges
13148 Dump after conversion of EH handling range regions.
13150 @item -fdump-rtl-expand
13151 @opindex fdump-rtl-expand
13152 Dump after RTL generation.
13154 @item -fdump-rtl-fwprop1
13155 @itemx -fdump-rtl-fwprop2
13156 @opindex fdump-rtl-fwprop1
13157 @opindex fdump-rtl-fwprop2
13158 @option{-fdump-rtl-fwprop1} and @option{-fdump-rtl-fwprop2} enable
13159 dumping after the two forward propagation passes.
13161 @item -fdump-rtl-gcse1
13162 @itemx -fdump-rtl-gcse2
13163 @opindex fdump-rtl-gcse1
13164 @opindex fdump-rtl-gcse2
13165 @option{-fdump-rtl-gcse1} and @option{-fdump-rtl-gcse2} enable dumping
13166 after global common subexpression elimination.
13168 @item -fdump-rtl-init-regs
13169 @opindex fdump-rtl-init-regs
13170 Dump after the initialization of the registers.
13172 @item -fdump-rtl-initvals
13173 @opindex fdump-rtl-initvals
13174 Dump after the computation of the initial value sets.
13176 @item -fdump-rtl-into_cfglayout
13177 @opindex fdump-rtl-into_cfglayout
13178 Dump after converting to cfglayout mode.
13180 @item -fdump-rtl-ira
13181 @opindex fdump-rtl-ira
13182 Dump after iterated register allocation.
13184 @item -fdump-rtl-jump
13185 @opindex fdump-rtl-jump
13186 Dump after the second jump optimization.
13188 @item -fdump-rtl-loop2
13189 @opindex fdump-rtl-loop2
13190 @option{-fdump-rtl-loop2} enables dumping after the rtl
13191 loop optimization passes.
13193 @item -fdump-rtl-mach
13194 @opindex fdump-rtl-mach
13195 Dump after performing the machine dependent reorganization pass, if that
13198 @item -fdump-rtl-mode_sw
13199 @opindex fdump-rtl-mode_sw
13200 Dump after removing redundant mode switches.
13202 @item -fdump-rtl-rnreg
13203 @opindex fdump-rtl-rnreg
13204 Dump after register renumbering.
13206 @item -fdump-rtl-outof_cfglayout
13207 @opindex fdump-rtl-outof_cfglayout
13208 Dump after converting from cfglayout mode.
13210 @item -fdump-rtl-peephole2
13211 @opindex fdump-rtl-peephole2
13212 Dump after the peephole pass.
13214 @item -fdump-rtl-postreload
13215 @opindex fdump-rtl-postreload
13216 Dump after post-reload optimizations.
13218 @item -fdump-rtl-pro_and_epilogue
13219 @opindex fdump-rtl-pro_and_epilogue
13220 Dump after generating the function prologues and epilogues.
13222 @item -fdump-rtl-sched1
13223 @itemx -fdump-rtl-sched2
13224 @opindex fdump-rtl-sched1
13225 @opindex fdump-rtl-sched2
13226 @option{-fdump-rtl-sched1} and @option{-fdump-rtl-sched2} enable dumping
13227 after the basic block scheduling passes.
13229 @item -fdump-rtl-ree
13230 @opindex fdump-rtl-ree
13231 Dump after sign/zero extension elimination.
13233 @item -fdump-rtl-seqabstr
13234 @opindex fdump-rtl-seqabstr
13235 Dump after common sequence discovery.
13237 @item -fdump-rtl-shorten
13238 @opindex fdump-rtl-shorten
13239 Dump after shortening branches.
13241 @item -fdump-rtl-sibling
13242 @opindex fdump-rtl-sibling
13243 Dump after sibling call optimizations.
13245 @item -fdump-rtl-split1
13246 @itemx -fdump-rtl-split2
13247 @itemx -fdump-rtl-split3
13248 @itemx -fdump-rtl-split4
13249 @itemx -fdump-rtl-split5
13250 @opindex fdump-rtl-split1
13251 @opindex fdump-rtl-split2
13252 @opindex fdump-rtl-split3
13253 @opindex fdump-rtl-split4
13254 @opindex fdump-rtl-split5
13255 These options enable dumping after five rounds of
13256 instruction splitting.
13258 @item -fdump-rtl-sms
13259 @opindex fdump-rtl-sms
13260 Dump after modulo scheduling. This pass is only run on some
13263 @item -fdump-rtl-stack
13264 @opindex fdump-rtl-stack
13265 Dump after conversion from GCC's ``flat register file'' registers to the
13266 x87's stack-like registers. This pass is only run on x86 variants.
13268 @item -fdump-rtl-subreg1
13269 @itemx -fdump-rtl-subreg2
13270 @opindex fdump-rtl-subreg1
13271 @opindex fdump-rtl-subreg2
13272 @option{-fdump-rtl-subreg1} and @option{-fdump-rtl-subreg2} enable dumping after
13273 the two subreg expansion passes.
13275 @item -fdump-rtl-unshare
13276 @opindex fdump-rtl-unshare
13277 Dump after all rtl has been unshared.
13279 @item -fdump-rtl-vartrack
13280 @opindex fdump-rtl-vartrack
13281 Dump after variable tracking.
13283 @item -fdump-rtl-vregs
13284 @opindex fdump-rtl-vregs
13285 Dump after converting virtual registers to hard registers.
13287 @item -fdump-rtl-web
13288 @opindex fdump-rtl-web
13289 Dump after live range splitting.
13291 @item -fdump-rtl-regclass
13292 @itemx -fdump-rtl-subregs_of_mode_init
13293 @itemx -fdump-rtl-subregs_of_mode_finish
13294 @itemx -fdump-rtl-dfinit
13295 @itemx -fdump-rtl-dfinish
13296 @opindex fdump-rtl-regclass
13297 @opindex fdump-rtl-subregs_of_mode_init
13298 @opindex fdump-rtl-subregs_of_mode_finish
13299 @opindex fdump-rtl-dfinit
13300 @opindex fdump-rtl-dfinish
13301 These dumps are defined but always produce empty files.
13304 @itemx -fdump-rtl-all
13306 @opindex fdump-rtl-all
13307 Produce all the dumps listed above.
13311 Annotate the assembler output with miscellaneous debugging information.
13315 Dump all macro definitions, at the end of preprocessing, in addition to
13320 Produce a core dump whenever an error occurs.
13324 Annotate the assembler output with a comment indicating which
13325 pattern and alternative is used. The length of each instruction is
13330 Dump the RTL in the assembler output as a comment before each instruction.
13331 Also turns on @option{-dp} annotation.
13335 Just generate RTL for a function instead of compiling it. Usually used
13336 with @option{-fdump-rtl-expand}.
13339 @item -fdump-noaddr
13340 @opindex fdump-noaddr
13341 When doing debugging dumps, suppress address output. This makes it more
13342 feasible to use diff on debugging dumps for compiler invocations with
13343 different compiler binaries and/or different
13344 text / bss / data / heap / stack / dso start locations.
13347 @opindex freport-bug
13348 Collect and dump debug information into a temporary file if an
13349 internal compiler error (ICE) occurs.
13351 @item -fdump-unnumbered
13352 @opindex fdump-unnumbered
13353 When doing debugging dumps, suppress instruction numbers and address output.
13354 This makes it more feasible to use diff on debugging dumps for compiler
13355 invocations with different options, in particular with and without
13358 @item -fdump-unnumbered-links
13359 @opindex fdump-unnumbered-links
13360 When doing debugging dumps (see @option{-d} option above), suppress
13361 instruction numbers for the links to the previous and next instructions
13364 @item -fdump-ipa-@var{switch}
13366 Control the dumping at various stages of inter-procedural analysis
13367 language tree to a file. The file name is generated by appending a
13368 switch specific suffix to the source file name, and the file is created
13369 in the same directory as the output file. The following dumps are
13374 Enables all inter-procedural analysis dumps.
13377 Dumps information about call-graph optimization, unused function removal,
13378 and inlining decisions.
13381 Dump after function inlining.
13385 @item -fdump-lang-all
13386 @itemx -fdump-lang-@var{switch}
13387 @itemx -fdump-lang-@var{switch}-@var{options}
13388 @itemx -fdump-lang-@var{switch}-@var{options}=@var{filename}
13389 @opindex fdump-lang-all
13390 @opindex fdump-lang
13391 Control the dumping of language-specific information. The @var{options}
13392 and @var{filename} portions behave as described in the
13393 @option{-fdump-tree} option. The following @var{switch} values are
13399 Enable all language-specific dumps.
13402 Dump class hierarchy information. Virtual table information is emitted
13403 unless '@option{slim}' is specified. This option is applicable to C++ only.
13406 Dump the raw internal tree data. This option is applicable to C++ only.
13410 @item -fdump-passes
13411 @opindex fdump-passes
13412 Print on @file{stderr} the list of optimization passes that are turned
13413 on and off by the current command-line options.
13415 @item -fdump-statistics-@var{option}
13416 @opindex fdump-statistics
13417 Enable and control dumping of pass statistics in a separate file. The
13418 file name is generated by appending a suffix ending in
13419 @samp{.statistics} to the source file name, and the file is created in
13420 the same directory as the output file. If the @samp{-@var{option}}
13421 form is used, @samp{-stats} causes counters to be summed over the
13422 whole compilation unit while @samp{-details} dumps every event as
13423 the passes generate them. The default with no option is to sum
13424 counters for each function compiled.
13426 @item -fdump-tree-all
13427 @itemx -fdump-tree-@var{switch}
13428 @itemx -fdump-tree-@var{switch}-@var{options}
13429 @itemx -fdump-tree-@var{switch}-@var{options}=@var{filename}
13430 @opindex fdump-tree-all
13431 @opindex fdump-tree
13432 Control the dumping at various stages of processing the intermediate
13433 language tree to a file. The file name is generated by appending a
13434 switch-specific suffix to the source file name, and the file is
13435 created in the same directory as the output file. In case of
13436 @option{=@var{filename}} option, the dump is output on the given file
13437 instead of the auto named dump files. If the @samp{-@var{options}}
13438 form is used, @var{options} is a list of @samp{-} separated options
13439 which control the details of the dump. Not all options are applicable
13440 to all dumps; those that are not meaningful are ignored. The
13441 following options are available
13445 Print the address of each node. Usually this is not meaningful as it
13446 changes according to the environment and source file. Its primary use
13447 is for tying up a dump file with a debug environment.
13449 If @code{DECL_ASSEMBLER_NAME} has been set for a given decl, use that
13450 in the dump instead of @code{DECL_NAME}. Its primary use is ease of
13451 use working backward from mangled names in the assembly file.
13453 When dumping front-end intermediate representations, inhibit dumping
13454 of members of a scope or body of a function merely because that scope
13455 has been reached. Only dump such items when they are directly reachable
13456 by some other path.
13458 When dumping pretty-printed trees, this option inhibits dumping the
13459 bodies of control structures.
13461 When dumping RTL, print the RTL in slim (condensed) form instead of
13462 the default LISP-like representation.
13464 Print a raw representation of the tree. By default, trees are
13465 pretty-printed into a C-like representation.
13467 Enable more detailed dumps (not honored by every dump option). Also
13468 include information from the optimization passes.
13470 Enable dumping various statistics about the pass (not honored by every dump
13473 Enable showing basic block boundaries (disabled in raw dumps).
13475 For each of the other indicated dump files (@option{-fdump-rtl-@var{pass}}),
13476 dump a representation of the control flow graph suitable for viewing with
13477 GraphViz to @file{@var{file}.@var{passid}.@var{pass}.dot}. Each function in
13478 the file is pretty-printed as a subgraph, so that GraphViz can render them
13479 all in a single plot.
13481 This option currently only works for RTL dumps, and the RTL is always
13482 dumped in slim form.
13484 Enable showing virtual operands for every statement.
13486 Enable showing line numbers for statements.
13488 Enable showing the unique ID (@code{DECL_UID}) for each variable.
13490 Enable showing the tree dump for each statement.
13492 Enable showing the EH region number holding each statement.
13494 Enable showing scalar evolution analysis details.
13496 Enable showing optimization information (only available in certain
13499 Enable showing missed optimization information (only available in certain
13502 Enable other detailed optimization information (only available in
13504 @item =@var{filename}
13505 Instead of an auto named dump file, output into the given file
13506 name. The file names @file{stdout} and @file{stderr} are treated
13507 specially and are considered already open standard streams. For
13511 gcc -O2 -ftree-vectorize -fdump-tree-vect-blocks=foo.dump
13512 -fdump-tree-pre=/dev/stderr file.c
13515 outputs vectorizer dump into @file{foo.dump}, while the PRE dump is
13516 output on to @file{stderr}. If two conflicting dump filenames are
13517 given for the same pass, then the latter option overrides the earlier
13521 Turn on all options, except @option{raw}, @option{slim}, @option{verbose}
13522 and @option{lineno}.
13525 Turn on all optimization options, i.e., @option{optimized},
13526 @option{missed}, and @option{note}.
13529 To determine what tree dumps are available or find the dump for a pass
13530 of interest follow the steps below.
13534 Invoke GCC with @option{-fdump-passes} and in the @file{stderr} output
13535 look for a code that corresponds to the pass you are interested in.
13536 For example, the codes @code{tree-evrp}, @code{tree-vrp1}, and
13537 @code{tree-vrp2} correspond to the three Value Range Propagation passes.
13538 The number at the end distinguishes distinct invocations of the same pass.
13540 To enable the creation of the dump file, append the pass code to
13541 the @option{-fdump-} option prefix and invoke GCC with it. For example,
13542 to enable the dump from the Early Value Range Propagation pass, invoke
13543 GCC with the @option{-fdump-tree-evrp} option. Optionally, you may
13544 specify the name of the dump file. If you don't specify one, GCC
13545 creates as described below.
13547 Find the pass dump in a file whose name is composed of three components
13548 separated by a period: the name of the source file GCC was invoked to
13549 compile, a numeric suffix indicating the pass number followed by the
13550 letter @samp{t} for tree passes (and the letter @samp{r} for RTL passes),
13551 and finally the pass code. For example, the Early VRP pass dump might
13552 be in a file named @file{myfile.c.038t.evrp} in the current working
13553 directory. Note that the numeric codes are not stable and may change
13554 from one version of GCC to another.
13558 @itemx -fopt-info-@var{options}
13559 @itemx -fopt-info-@var{options}=@var{filename}
13561 Controls optimization dumps from various optimization passes. If the
13562 @samp{-@var{options}} form is used, @var{options} is a list of
13563 @samp{-} separated option keywords to select the dump details and
13566 The @var{options} can be divided into two groups: options describing the
13567 verbosity of the dump, and options describing which optimizations
13568 should be included. The options from both the groups can be freely
13569 mixed as they are non-overlapping. However, in case of any conflicts,
13570 the later options override the earlier options on the command
13573 The following options control the dump verbosity:
13577 Print information when an optimization is successfully applied. It is
13578 up to a pass to decide which information is relevant. For example, the
13579 vectorizer passes print the source location of loops which are
13580 successfully vectorized.
13582 Print information about missed optimizations. Individual passes
13583 control which information to include in the output.
13585 Print verbose information about optimizations, such as certain
13586 transformations, more detailed messages about decisions etc.
13588 Print detailed optimization information. This includes
13589 @samp{optimized}, @samp{missed}, and @samp{note}.
13592 One or more of the following option keywords can be used to describe a
13593 group of optimizations:
13597 Enable dumps from all interprocedural optimizations.
13599 Enable dumps from all loop optimizations.
13601 Enable dumps from all inlining optimizations.
13603 Enable dumps from all OMP (Offloading and Multi Processing) optimizations.
13605 Enable dumps from all vectorization optimizations.
13607 Enable dumps from all optimizations. This is a superset of
13608 the optimization groups listed above.
13611 If @var{options} is
13612 omitted, it defaults to @samp{optimized-optall}, which means to dump all
13613 info about successful optimizations from all the passes.
13615 If the @var{filename} is provided, then the dumps from all the
13616 applicable optimizations are concatenated into the @var{filename}.
13617 Otherwise the dump is output onto @file{stderr}. Though multiple
13618 @option{-fopt-info} options are accepted, only one of them can include
13619 a @var{filename}. If other filenames are provided then all but the
13620 first such option are ignored.
13622 Note that the output @var{filename} is overwritten
13623 in case of multiple translation units. If a combined output from
13624 multiple translation units is desired, @file{stderr} should be used
13627 In the following example, the optimization info is output to
13636 gcc -O3 -fopt-info-missed=missed.all
13640 outputs missed optimization report from all the passes into
13641 @file{missed.all}, and this one:
13644 gcc -O2 -ftree-vectorize -fopt-info-vec-missed
13648 prints information about missed optimization opportunities from
13649 vectorization passes on @file{stderr}.
13650 Note that @option{-fopt-info-vec-missed} is equivalent to
13651 @option{-fopt-info-missed-vec}. The order of the optimization group
13652 names and message types listed after @option{-fopt-info} does not matter.
13654 As another example,
13656 gcc -O3 -fopt-info-inline-optimized-missed=inline.txt
13660 outputs information about missed optimizations as well as
13661 optimized locations from all the inlining passes into
13667 gcc -fopt-info-vec-missed=vec.miss -fopt-info-loop-optimized=loop.opt
13671 Here the two output filenames @file{vec.miss} and @file{loop.opt} are
13672 in conflict since only one output file is allowed. In this case, only
13673 the first option takes effect and the subsequent options are
13674 ignored. Thus only @file{vec.miss} is produced which contains
13675 dumps from the vectorizer about missed opportunities.
13677 @item -fsched-verbose=@var{n}
13678 @opindex fsched-verbose
13679 On targets that use instruction scheduling, this option controls the
13680 amount of debugging output the scheduler prints to the dump files.
13682 For @var{n} greater than zero, @option{-fsched-verbose} outputs the
13683 same information as @option{-fdump-rtl-sched1} and @option{-fdump-rtl-sched2}.
13684 For @var{n} greater than one, it also output basic block probabilities,
13685 detailed ready list information and unit/insn info. For @var{n} greater
13686 than two, it includes RTL at abort point, control-flow and regions info.
13687 And for @var{n} over four, @option{-fsched-verbose} also includes
13692 @item -fenable-@var{kind}-@var{pass}
13693 @itemx -fdisable-@var{kind}-@var{pass}=@var{range-list}
13697 This is a set of options that are used to explicitly disable/enable
13698 optimization passes. These options are intended for use for debugging GCC.
13699 Compiler users should use regular options for enabling/disabling
13704 @item -fdisable-ipa-@var{pass}
13705 Disable IPA pass @var{pass}. @var{pass} is the pass name. If the same pass is
13706 statically invoked in the compiler multiple times, the pass name should be
13707 appended with a sequential number starting from 1.
13709 @item -fdisable-rtl-@var{pass}
13710 @itemx -fdisable-rtl-@var{pass}=@var{range-list}
13711 Disable RTL pass @var{pass}. @var{pass} is the pass name. If the same pass is
13712 statically invoked in the compiler multiple times, the pass name should be
13713 appended with a sequential number starting from 1. @var{range-list} is a
13714 comma-separated list of function ranges or assembler names. Each range is a number
13715 pair separated by a colon. The range is inclusive in both ends. If the range
13716 is trivial, the number pair can be simplified as a single number. If the
13717 function's call graph node's @var{uid} falls within one of the specified ranges,
13718 the @var{pass} is disabled for that function. The @var{uid} is shown in the
13719 function header of a dump file, and the pass names can be dumped by using
13720 option @option{-fdump-passes}.
13722 @item -fdisable-tree-@var{pass}
13723 @itemx -fdisable-tree-@var{pass}=@var{range-list}
13724 Disable tree pass @var{pass}. See @option{-fdisable-rtl} for the description of
13727 @item -fenable-ipa-@var{pass}
13728 Enable IPA pass @var{pass}. @var{pass} is the pass name. If the same pass is
13729 statically invoked in the compiler multiple times, the pass name should be
13730 appended with a sequential number starting from 1.
13732 @item -fenable-rtl-@var{pass}
13733 @itemx -fenable-rtl-@var{pass}=@var{range-list}
13734 Enable RTL pass @var{pass}. See @option{-fdisable-rtl} for option argument
13735 description and examples.
13737 @item -fenable-tree-@var{pass}
13738 @itemx -fenable-tree-@var{pass}=@var{range-list}
13739 Enable tree pass @var{pass}. See @option{-fdisable-rtl} for the description
13740 of option arguments.
13744 Here are some examples showing uses of these options.
13748 # disable ccp1 for all functions
13749 -fdisable-tree-ccp1
13750 # disable complete unroll for function whose cgraph node uid is 1
13751 -fenable-tree-cunroll=1
13752 # disable gcse2 for functions at the following ranges [1,1],
13753 # [300,400], and [400,1000]
13754 # disable gcse2 for functions foo and foo2
13755 -fdisable-rtl-gcse2=foo,foo2
13756 # disable early inlining
13757 -fdisable-tree-einline
13758 # disable ipa inlining
13759 -fdisable-ipa-inline
13760 # enable tree full unroll
13761 -fenable-tree-unroll
13766 @itemx -fchecking=@var{n}
13768 @opindex fno-checking
13769 Enable internal consistency checking. The default depends on
13770 the compiler configuration. @option{-fchecking=2} enables further
13771 internal consistency checking that might affect code generation.
13773 @item -frandom-seed=@var{string}
13774 @opindex frandom-seed
13775 This option provides a seed that GCC uses in place of
13776 random numbers in generating certain symbol names
13777 that have to be different in every compiled file. It is also used to
13778 place unique stamps in coverage data files and the object files that
13779 produce them. You can use the @option{-frandom-seed} option to produce
13780 reproducibly identical object files.
13782 The @var{string} can either be a number (decimal, octal or hex) or an
13783 arbitrary string (in which case it's converted to a number by
13786 The @var{string} should be different for every file you compile.
13789 @itemx -save-temps=cwd
13790 @opindex save-temps
13791 Store the usual ``temporary'' intermediate files permanently; place them
13792 in the current directory and name them based on the source file. Thus,
13793 compiling @file{foo.c} with @option{-c -save-temps} produces files
13794 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
13795 preprocessed @file{foo.i} output file even though the compiler now
13796 normally uses an integrated preprocessor.
13798 When used in combination with the @option{-x} command-line option,
13799 @option{-save-temps} is sensible enough to avoid over writing an
13800 input source file with the same extension as an intermediate file.
13801 The corresponding intermediate file may be obtained by renaming the
13802 source file before using @option{-save-temps}.
13804 If you invoke GCC in parallel, compiling several different source
13805 files that share a common base name in different subdirectories or the
13806 same source file compiled for multiple output destinations, it is
13807 likely that the different parallel compilers will interfere with each
13808 other, and overwrite the temporary files. For instance:
13811 gcc -save-temps -o outdir1/foo.o indir1/foo.c&
13812 gcc -save-temps -o outdir2/foo.o indir2/foo.c&
13815 may result in @file{foo.i} and @file{foo.o} being written to
13816 simultaneously by both compilers.
13818 @item -save-temps=obj
13819 @opindex save-temps=obj
13820 Store the usual ``temporary'' intermediate files permanently. If the
13821 @option{-o} option is used, the temporary files are based on the
13822 object file. If the @option{-o} option is not used, the
13823 @option{-save-temps=obj} switch behaves like @option{-save-temps}.
13828 gcc -save-temps=obj -c foo.c
13829 gcc -save-temps=obj -c bar.c -o dir/xbar.o
13830 gcc -save-temps=obj foobar.c -o dir2/yfoobar
13834 creates @file{foo.i}, @file{foo.s}, @file{dir/xbar.i},
13835 @file{dir/xbar.s}, @file{dir2/yfoobar.i}, @file{dir2/yfoobar.s}, and
13836 @file{dir2/yfoobar.o}.
13838 @item -time@r{[}=@var{file}@r{]}
13840 Report the CPU time taken by each subprocess in the compilation
13841 sequence. For C source files, this is the compiler proper and assembler
13842 (plus the linker if linking is done).
13844 Without the specification of an output file, the output looks like this:
13851 The first number on each line is the ``user time'', that is time spent
13852 executing the program itself. The second number is ``system time'',
13853 time spent executing operating system routines on behalf of the program.
13854 Both numbers are in seconds.
13856 With the specification of an output file, the output is appended to the
13857 named file, and it looks like this:
13860 0.12 0.01 cc1 @var{options}
13861 0.00 0.01 as @var{options}
13864 The ``user time'' and the ``system time'' are moved before the program
13865 name, and the options passed to the program are displayed, so that one
13866 can later tell what file was being compiled, and with which options.
13868 @item -fdump-final-insns@r{[}=@var{file}@r{]}
13869 @opindex fdump-final-insns
13870 Dump the final internal representation (RTL) to @var{file}. If the
13871 optional argument is omitted (or if @var{file} is @code{.}), the name
13872 of the dump file is determined by appending @code{.gkd} to the
13873 compilation output file name.
13875 @item -fcompare-debug@r{[}=@var{opts}@r{]}
13876 @opindex fcompare-debug
13877 @opindex fno-compare-debug
13878 If no error occurs during compilation, run the compiler a second time,
13879 adding @var{opts} and @option{-fcompare-debug-second} to the arguments
13880 passed to the second compilation. Dump the final internal
13881 representation in both compilations, and print an error if they differ.
13883 If the equal sign is omitted, the default @option{-gtoggle} is used.
13885 The environment variable @env{GCC_COMPARE_DEBUG}, if defined, non-empty
13886 and nonzero, implicitly enables @option{-fcompare-debug}. If
13887 @env{GCC_COMPARE_DEBUG} is defined to a string starting with a dash,
13888 then it is used for @var{opts}, otherwise the default @option{-gtoggle}
13891 @option{-fcompare-debug=}, with the equal sign but without @var{opts},
13892 is equivalent to @option{-fno-compare-debug}, which disables the dumping
13893 of the final representation and the second compilation, preventing even
13894 @env{GCC_COMPARE_DEBUG} from taking effect.
13896 To verify full coverage during @option{-fcompare-debug} testing, set
13897 @env{GCC_COMPARE_DEBUG} to say @option{-fcompare-debug-not-overridden},
13898 which GCC rejects as an invalid option in any actual compilation
13899 (rather than preprocessing, assembly or linking). To get just a
13900 warning, setting @env{GCC_COMPARE_DEBUG} to @samp{-w%n-fcompare-debug
13901 not overridden} will do.
13903 @item -fcompare-debug-second
13904 @opindex fcompare-debug-second
13905 This option is implicitly passed to the compiler for the second
13906 compilation requested by @option{-fcompare-debug}, along with options to
13907 silence warnings, and omitting other options that would cause
13908 side-effect compiler outputs to files or to the standard output. Dump
13909 files and preserved temporary files are renamed so as to contain the
13910 @code{.gk} additional extension during the second compilation, to avoid
13911 overwriting those generated by the first.
13913 When this option is passed to the compiler driver, it causes the
13914 @emph{first} compilation to be skipped, which makes it useful for little
13915 other than debugging the compiler proper.
13919 Turn off generation of debug info, if leaving out this option
13920 generates it, or turn it on at level 2 otherwise. The position of this
13921 argument in the command line does not matter; it takes effect after all
13922 other options are processed, and it does so only once, no matter how
13923 many times it is given. This is mainly intended to be used with
13924 @option{-fcompare-debug}.
13926 @item -fvar-tracking-assignments-toggle
13927 @opindex fvar-tracking-assignments-toggle
13928 @opindex fno-var-tracking-assignments-toggle
13929 Toggle @option{-fvar-tracking-assignments}, in the same way that
13930 @option{-gtoggle} toggles @option{-g}.
13934 Makes the compiler print out each function name as it is compiled, and
13935 print some statistics about each pass when it finishes.
13937 @item -ftime-report
13938 @opindex ftime-report
13939 Makes the compiler print some statistics about the time consumed by each
13940 pass when it finishes.
13942 @item -ftime-report-details
13943 @opindex ftime-report-details
13944 Record the time consumed by infrastructure parts separately for each pass.
13946 @item -fira-verbose=@var{n}
13947 @opindex fira-verbose
13948 Control the verbosity of the dump file for the integrated register allocator.
13949 The default value is 5. If the value @var{n} is greater or equal to 10,
13950 the dump output is sent to stderr using the same format as @var{n} minus 10.
13953 @opindex flto-report
13954 Prints a report with internal details on the workings of the link-time
13955 optimizer. The contents of this report vary from version to version.
13956 It is meant to be useful to GCC developers when processing object
13957 files in LTO mode (via @option{-flto}).
13959 Disabled by default.
13961 @item -flto-report-wpa
13962 @opindex flto-report-wpa
13963 Like @option{-flto-report}, but only print for the WPA phase of Link
13967 @opindex fmem-report
13968 Makes the compiler print some statistics about permanent memory
13969 allocation when it finishes.
13971 @item -fmem-report-wpa
13972 @opindex fmem-report-wpa
13973 Makes the compiler print some statistics about permanent memory
13974 allocation for the WPA phase only.
13976 @item -fpre-ipa-mem-report
13977 @opindex fpre-ipa-mem-report
13978 @item -fpost-ipa-mem-report
13979 @opindex fpost-ipa-mem-report
13980 Makes the compiler print some statistics about permanent memory
13981 allocation before or after interprocedural optimization.
13983 @item -fprofile-report
13984 @opindex fprofile-report
13985 Makes the compiler print some statistics about consistency of the
13986 (estimated) profile and effect of individual passes.
13988 @item -fstack-usage
13989 @opindex fstack-usage
13990 Makes the compiler output stack usage information for the program, on a
13991 per-function basis. The filename for the dump is made by appending
13992 @file{.su} to the @var{auxname}. @var{auxname} is generated from the name of
13993 the output file, if explicitly specified and it is not an executable,
13994 otherwise it is the basename of the source file. An entry is made up
13999 The name of the function.
14003 One or more qualifiers: @code{static}, @code{dynamic}, @code{bounded}.
14006 The qualifier @code{static} means that the function manipulates the stack
14007 statically: a fixed number of bytes are allocated for the frame on function
14008 entry and released on function exit; no stack adjustments are otherwise made
14009 in the function. The second field is this fixed number of bytes.
14011 The qualifier @code{dynamic} means that the function manipulates the stack
14012 dynamically: in addition to the static allocation described above, stack
14013 adjustments are made in the body of the function, for example to push/pop
14014 arguments around function calls. If the qualifier @code{bounded} is also
14015 present, the amount of these adjustments is bounded at compile time and
14016 the second field is an upper bound of the total amount of stack used by
14017 the function. If it is not present, the amount of these adjustments is
14018 not bounded at compile time and the second field only represents the
14023 Emit statistics about front-end processing at the end of the compilation.
14024 This option is supported only by the C++ front end, and
14025 the information is generally only useful to the G++ development team.
14027 @item -fdbg-cnt-list
14028 @opindex fdbg-cnt-list
14029 Print the name and the counter upper bound for all debug counters.
14032 @item -fdbg-cnt=@var{counter-value-list}
14034 Set the internal debug counter upper bound. @var{counter-value-list}
14035 is a comma-separated list of @var{name}:@var{value} pairs
14036 which sets the upper bound of each debug counter @var{name} to @var{value}.
14037 All debug counters have the initial upper bound of @code{UINT_MAX};
14038 thus @code{dbg_cnt} returns true always unless the upper bound
14039 is set by this option.
14040 For example, with @option{-fdbg-cnt=dce:10,tail_call:0},
14041 @code{dbg_cnt(dce)} returns true only for first 10 invocations.
14043 @item -print-file-name=@var{library}
14044 @opindex print-file-name
14045 Print the full absolute name of the library file @var{library} that
14046 would be used when linking---and don't do anything else. With this
14047 option, GCC does not compile or link anything; it just prints the
14050 @item -print-multi-directory
14051 @opindex print-multi-directory
14052 Print the directory name corresponding to the multilib selected by any
14053 other switches present in the command line. This directory is supposed
14054 to exist in @env{GCC_EXEC_PREFIX}.
14056 @item -print-multi-lib
14057 @opindex print-multi-lib
14058 Print the mapping from multilib directory names to compiler switches
14059 that enable them. The directory name is separated from the switches by
14060 @samp{;}, and each switch starts with an @samp{@@} instead of the
14061 @samp{-}, without spaces between multiple switches. This is supposed to
14062 ease shell processing.
14064 @item -print-multi-os-directory
14065 @opindex print-multi-os-directory
14066 Print the path to OS libraries for the selected
14067 multilib, relative to some @file{lib} subdirectory. If OS libraries are
14068 present in the @file{lib} subdirectory and no multilibs are used, this is
14069 usually just @file{.}, if OS libraries are present in @file{lib@var{suffix}}
14070 sibling directories this prints e.g.@: @file{../lib64}, @file{../lib} or
14071 @file{../lib32}, or if OS libraries are present in @file{lib/@var{subdir}}
14072 subdirectories it prints e.g.@: @file{amd64}, @file{sparcv9} or @file{ev6}.
14074 @item -print-multiarch
14075 @opindex print-multiarch
14076 Print the path to OS libraries for the selected multiarch,
14077 relative to some @file{lib} subdirectory.
14079 @item -print-prog-name=@var{program}
14080 @opindex print-prog-name
14081 Like @option{-print-file-name}, but searches for a program such as @command{cpp}.
14083 @item -print-libgcc-file-name
14084 @opindex print-libgcc-file-name
14085 Same as @option{-print-file-name=libgcc.a}.
14087 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
14088 but you do want to link with @file{libgcc.a}. You can do:
14091 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
14094 @item -print-search-dirs
14095 @opindex print-search-dirs
14096 Print the name of the configured installation directory and a list of
14097 program and library directories @command{gcc} searches---and don't do anything else.
14099 This is useful when @command{gcc} prints the error message
14100 @samp{installation problem, cannot exec cpp0: No such file or directory}.
14101 To resolve this you either need to put @file{cpp0} and the other compiler
14102 components where @command{gcc} expects to find them, or you can set the environment
14103 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
14104 Don't forget the trailing @samp{/}.
14105 @xref{Environment Variables}.
14107 @item -print-sysroot
14108 @opindex print-sysroot
14109 Print the target sysroot directory that is used during
14110 compilation. This is the target sysroot specified either at configure
14111 time or using the @option{--sysroot} option, possibly with an extra
14112 suffix that depends on compilation options. If no target sysroot is
14113 specified, the option prints nothing.
14115 @item -print-sysroot-headers-suffix
14116 @opindex print-sysroot-headers-suffix
14117 Print the suffix added to the target sysroot when searching for
14118 headers, or give an error if the compiler is not configured with such
14119 a suffix---and don't do anything else.
14122 @opindex dumpmachine
14123 Print the compiler's target machine (for example,
14124 @samp{i686-pc-linux-gnu})---and don't do anything else.
14127 @opindex dumpversion
14128 Print the compiler version (for example, @code{3.0}, @code{6.3.0} or @code{7})---and don't do
14129 anything else. This is the compiler version used in filesystem paths,
14130 specs, can be depending on how the compiler has been configured just
14131 a single number (major version), two numbers separated by dot (major and
14132 minor version) or three numbers separated by dots (major, minor and patchlevel
14135 @item -dumpfullversion
14136 @opindex dumpfullversion
14137 Print the full compiler version, always 3 numbers separated by dots,
14138 major, minor and patchlevel version.
14142 Print the compiler's built-in specs---and don't do anything else. (This
14143 is used when GCC itself is being built.) @xref{Spec Files}.
14146 @node Submodel Options
14147 @section Machine-Dependent Options
14148 @cindex submodel options
14149 @cindex specifying hardware config
14150 @cindex hardware models and configurations, specifying
14151 @cindex target-dependent options
14152 @cindex machine-dependent options
14154 Each target machine supported by GCC can have its own options---for
14155 example, to allow you to compile for a particular processor variant or
14156 ABI, or to control optimizations specific to that machine. By
14157 convention, the names of machine-specific options start with
14160 Some configurations of the compiler also support additional target-specific
14161 options, usually for compatibility with other compilers on the same
14164 @c This list is ordered alphanumerically by subsection name.
14165 @c It should be the same order and spelling as these options are listed
14166 @c in Machine Dependent Options
14169 * AArch64 Options::
14170 * Adapteva Epiphany Options::
14174 * Blackfin Options::
14179 * DEC Alpha Options::
14183 * GNU/Linux Options::
14193 * MicroBlaze Options::
14196 * MN10300 Options::
14200 * Nios II Options::
14201 * Nvidia PTX Options::
14203 * picoChip Options::
14204 * PowerPC Options::
14207 * RS/6000 and PowerPC Options::
14209 * S/390 and zSeries Options::
14212 * Solaris 2 Options::
14215 * System V Options::
14216 * TILE-Gx Options::
14217 * TILEPro Options::
14222 * VxWorks Options::
14224 * x86 Windows Options::
14225 * Xstormy16 Options::
14227 * zSeries Options::
14230 @node AArch64 Options
14231 @subsection AArch64 Options
14232 @cindex AArch64 Options
14234 These options are defined for AArch64 implementations:
14238 @item -mabi=@var{name}
14240 Generate code for the specified data model. Permissible values
14241 are @samp{ilp32} for SysV-like data model where int, long int and pointers
14242 are 32 bits, and @samp{lp64} for SysV-like data model where int is 32 bits,
14243 but long int and pointers are 64 bits.
14245 The default depends on the specific target configuration. Note that
14246 the LP64 and ILP32 ABIs are not link-compatible; you must compile your
14247 entire program with the same ABI, and link with a compatible set of libraries.
14250 @opindex mbig-endian
14251 Generate big-endian code. This is the default when GCC is configured for an
14252 @samp{aarch64_be-*-*} target.
14254 @item -mgeneral-regs-only
14255 @opindex mgeneral-regs-only
14256 Generate code which uses only the general-purpose registers. This will prevent
14257 the compiler from using floating-point and Advanced SIMD registers but will not
14258 impose any restrictions on the assembler.
14260 @item -mlittle-endian
14261 @opindex mlittle-endian
14262 Generate little-endian code. This is the default when GCC is configured for an
14263 @samp{aarch64-*-*} but not an @samp{aarch64_be-*-*} target.
14265 @item -mcmodel=tiny
14266 @opindex mcmodel=tiny
14267 Generate code for the tiny code model. The program and its statically defined
14268 symbols must be within 1MB of each other. Programs can be statically or
14269 dynamically linked.
14271 @item -mcmodel=small
14272 @opindex mcmodel=small
14273 Generate code for the small code model. The program and its statically defined
14274 symbols must be within 4GB of each other. Programs can be statically or
14275 dynamically linked. This is the default code model.
14277 @item -mcmodel=large
14278 @opindex mcmodel=large
14279 Generate code for the large code model. This makes no assumptions about
14280 addresses and sizes of sections. Programs can be statically linked only.
14282 @item -mstrict-align
14283 @opindex mstrict-align
14284 Avoid generating memory accesses that may not be aligned on a natural object
14285 boundary as described in the architecture specification.
14287 @item -momit-leaf-frame-pointer
14288 @itemx -mno-omit-leaf-frame-pointer
14289 @opindex momit-leaf-frame-pointer
14290 @opindex mno-omit-leaf-frame-pointer
14291 Omit or keep the frame pointer in leaf functions. The former behavior is the
14294 @item -mtls-dialect=desc
14295 @opindex mtls-dialect=desc
14296 Use TLS descriptors as the thread-local storage mechanism for dynamic accesses
14297 of TLS variables. This is the default.
14299 @item -mtls-dialect=traditional
14300 @opindex mtls-dialect=traditional
14301 Use traditional TLS as the thread-local storage mechanism for dynamic accesses
14304 @item -mtls-size=@var{size}
14306 Specify bit size of immediate TLS offsets. Valid values are 12, 24, 32, 48.
14307 This option requires binutils 2.26 or newer.
14309 @item -mfix-cortex-a53-835769
14310 @itemx -mno-fix-cortex-a53-835769
14311 @opindex mfix-cortex-a53-835769
14312 @opindex mno-fix-cortex-a53-835769
14313 Enable or disable the workaround for the ARM Cortex-A53 erratum number 835769.
14314 This involves inserting a NOP instruction between memory instructions and
14315 64-bit integer multiply-accumulate instructions.
14317 @item -mfix-cortex-a53-843419
14318 @itemx -mno-fix-cortex-a53-843419
14319 @opindex mfix-cortex-a53-843419
14320 @opindex mno-fix-cortex-a53-843419
14321 Enable or disable the workaround for the ARM Cortex-A53 erratum number 843419.
14322 This erratum workaround is made at link time and this will only pass the
14323 corresponding flag to the linker.
14325 @item -mlow-precision-recip-sqrt
14326 @item -mno-low-precision-recip-sqrt
14327 @opindex mlow-precision-recip-sqrt
14328 @opindex mno-low-precision-recip-sqrt
14329 Enable or disable the reciprocal square root approximation.
14330 This option only has an effect if @option{-ffast-math} or
14331 @option{-funsafe-math-optimizations} is used as well. Enabling this reduces
14332 precision of reciprocal square root results to about 16 bits for
14333 single precision and to 32 bits for double precision.
14335 @item -mlow-precision-sqrt
14336 @item -mno-low-precision-sqrt
14337 @opindex -mlow-precision-sqrt
14338 @opindex -mno-low-precision-sqrt
14339 Enable or disable the square root approximation.
14340 This option only has an effect if @option{-ffast-math} or
14341 @option{-funsafe-math-optimizations} is used as well. Enabling this reduces
14342 precision of square root results to about 16 bits for
14343 single precision and to 32 bits for double precision.
14344 If enabled, it implies @option{-mlow-precision-recip-sqrt}.
14346 @item -mlow-precision-div
14347 @item -mno-low-precision-div
14348 @opindex -mlow-precision-div
14349 @opindex -mno-low-precision-div
14350 Enable or disable the division approximation.
14351 This option only has an effect if @option{-ffast-math} or
14352 @option{-funsafe-math-optimizations} is used as well. Enabling this reduces
14353 precision of division results to about 16 bits for
14354 single precision and to 32 bits for double precision.
14356 @item -march=@var{name}
14358 Specify the name of the target architecture and, optionally, one or
14359 more feature modifiers. This option has the form
14360 @option{-march=@var{arch}@r{@{}+@r{[}no@r{]}@var{feature}@r{@}*}}.
14362 The permissible values for @var{arch} are @samp{armv8-a},
14363 @samp{armv8.1-a}, @samp{armv8.2-a}, @samp{armv8.3-a} or @var{native}.
14365 The value @samp{armv8.3-a} implies @samp{armv8.2-a} and enables compiler
14366 support for the ARMv8.3-A architecture extensions.
14368 The value @samp{armv8.2-a} implies @samp{armv8.1-a} and enables compiler
14369 support for the ARMv8.2-A architecture extensions.
14371 The value @samp{armv8.1-a} implies @samp{armv8-a} and enables compiler
14372 support for the ARMv8.1-A architecture extension. In particular, it
14373 enables the @samp{+crc}, @samp{+lse}, and @samp{+rdma} features.
14375 The value @samp{native} is available on native AArch64 GNU/Linux and
14376 causes the compiler to pick the architecture of the host system. This
14377 option has no effect if the compiler is unable to recognize the
14378 architecture of the host system,
14380 The permissible values for @var{feature} are listed in the sub-section
14381 on @ref{aarch64-feature-modifiers,,@option{-march} and @option{-mcpu}
14382 Feature Modifiers}. Where conflicting feature modifiers are
14383 specified, the right-most feature is used.
14385 GCC uses @var{name} to determine what kind of instructions it can emit
14386 when generating assembly code. If @option{-march} is specified
14387 without either of @option{-mtune} or @option{-mcpu} also being
14388 specified, the code is tuned to perform well across a range of target
14389 processors implementing the target architecture.
14391 @item -mtune=@var{name}
14393 Specify the name of the target processor for which GCC should tune the
14394 performance of the code. Permissible values for this option are:
14395 @samp{generic}, @samp{cortex-a35}, @samp{cortex-a53}, @samp{cortex-a55},
14396 @samp{cortex-a57}, @samp{cortex-a72}, @samp{cortex-a73}, @samp{cortex-a75},
14397 @samp{exynos-m1}, @samp{falkor}, @samp{qdf24xx}, @samp{saphira},
14398 @samp{xgene1}, @samp{vulcan}, @samp{thunderx},
14399 @samp{thunderxt88}, @samp{thunderxt88p1}, @samp{thunderxt81},
14400 @samp{thunderxt83}, @samp{thunderx2t99}, @samp{cortex-a57.cortex-a53},
14401 @samp{cortex-a72.cortex-a53}, @samp{cortex-a73.cortex-a35},
14402 @samp{cortex-a73.cortex-a53}, @samp{cortex-a75.cortex-a55},
14405 The values @samp{cortex-a57.cortex-a53}, @samp{cortex-a72.cortex-a53},
14406 @samp{cortex-a73.cortex-a35}, @samp{cortex-a73.cortex-a53},
14407 @samp{cortex-a75.cortex-a55} specify that GCC should tune for a
14410 Additionally on native AArch64 GNU/Linux systems the value
14411 @samp{native} tunes performance to the host system. This option has no effect
14412 if the compiler is unable to recognize the processor of the host system.
14414 Where none of @option{-mtune=}, @option{-mcpu=} or @option{-march=}
14415 are specified, the code is tuned to perform well across a range
14416 of target processors.
14418 This option cannot be suffixed by feature modifiers.
14420 @item -mcpu=@var{name}
14422 Specify the name of the target processor, optionally suffixed by one
14423 or more feature modifiers. This option has the form
14424 @option{-mcpu=@var{cpu}@r{@{}+@r{[}no@r{]}@var{feature}@r{@}*}}, where
14425 the permissible values for @var{cpu} are the same as those available
14426 for @option{-mtune}. The permissible values for @var{feature} are
14427 documented in the sub-section on
14428 @ref{aarch64-feature-modifiers,,@option{-march} and @option{-mcpu}
14429 Feature Modifiers}. Where conflicting feature modifiers are
14430 specified, the right-most feature is used.
14432 GCC uses @var{name} to determine what kind of instructions it can emit when
14433 generating assembly code (as if by @option{-march}) and to determine
14434 the target processor for which to tune for performance (as if
14435 by @option{-mtune}). Where this option is used in conjunction
14436 with @option{-march} or @option{-mtune}, those options take precedence
14437 over the appropriate part of this option.
14439 @item -moverride=@var{string}
14441 Override tuning decisions made by the back-end in response to a
14442 @option{-mtune=} switch. The syntax, semantics, and accepted values
14443 for @var{string} in this option are not guaranteed to be consistent
14446 This option is only intended to be useful when developing GCC.
14448 @item -mpc-relative-literal-loads
14449 @itemx -mno-pc-relative-literal-loads
14450 @opindex mpc-relative-literal-loads
14451 @opindex mno-pc-relative-literal-loads
14452 Enable or disable PC-relative literal loads. With this option literal pools are
14453 accessed using a single instruction and emitted after each function. This
14454 limits the maximum size of functions to 1MB. This is enabled by default for
14455 @option{-mcmodel=tiny}.
14457 @item -msign-return-address=@var{scope}
14458 @opindex msign-return-address
14459 Select the function scope on which return address signing will be applied.
14460 Permissible values are @samp{none}, which disables return address signing,
14461 @samp{non-leaf}, which enables pointer signing for functions which are not leaf
14462 functions, and @samp{all}, which enables pointer signing for all functions. The
14463 default value is @samp{none}.
14467 @subsubsection @option{-march} and @option{-mcpu} Feature Modifiers
14468 @anchor{aarch64-feature-modifiers}
14469 @cindex @option{-march} feature modifiers
14470 @cindex @option{-mcpu} feature modifiers
14471 Feature modifiers used with @option{-march} and @option{-mcpu} can be any of
14472 the following and their inverses @option{no@var{feature}}:
14476 Enable CRC extension. This is on by default for
14477 @option{-march=armv8.1-a}.
14479 Enable Crypto extension. This also enables Advanced SIMD and floating-point
14482 Enable floating-point instructions. This is on by default for all possible
14483 values for options @option{-march} and @option{-mcpu}.
14485 Enable Advanced SIMD instructions. This also enables floating-point
14486 instructions. This is on by default for all possible values for options
14487 @option{-march} and @option{-mcpu}.
14489 Enable Large System Extension instructions. This is on by default for
14490 @option{-march=armv8.1-a}.
14492 Enable Round Double Multiply Accumulate instructions. This is on by default
14493 for @option{-march=armv8.1-a}.
14495 Enable FP16 extension. This also enables floating-point instructions.
14497 Enable the RcPc extension. This does not change code generation from GCC,
14498 but is passed on to the assembler, enabling inline asm statements to use
14499 instructions from the RcPc extension.
14501 Enable the Dot Product extension. This also enables Advanced SIMD instructions.
14505 Feature @option{crypto} implies @option{simd}, which implies @option{fp}.
14506 Conversely, @option{nofp} implies @option{nosimd}, which implies
14509 @node Adapteva Epiphany Options
14510 @subsection Adapteva Epiphany Options
14512 These @samp{-m} options are defined for Adapteva Epiphany:
14515 @item -mhalf-reg-file
14516 @opindex mhalf-reg-file
14517 Don't allocate any register in the range @code{r32}@dots{}@code{r63}.
14518 That allows code to run on hardware variants that lack these registers.
14520 @item -mprefer-short-insn-regs
14521 @opindex mprefer-short-insn-regs
14522 Preferentially allocate registers that allow short instruction generation.
14523 This can result in increased instruction count, so this may either reduce or
14524 increase overall code size.
14526 @item -mbranch-cost=@var{num}
14527 @opindex mbranch-cost
14528 Set the cost of branches to roughly @var{num} ``simple'' instructions.
14529 This cost is only a heuristic and is not guaranteed to produce
14530 consistent results across releases.
14534 Enable the generation of conditional moves.
14536 @item -mnops=@var{num}
14538 Emit @var{num} NOPs before every other generated instruction.
14540 @item -mno-soft-cmpsf
14541 @opindex mno-soft-cmpsf
14542 For single-precision floating-point comparisons, emit an @code{fsub} instruction
14543 and test the flags. This is faster than a software comparison, but can
14544 get incorrect results in the presence of NaNs, or when two different small
14545 numbers are compared such that their difference is calculated as zero.
14546 The default is @option{-msoft-cmpsf}, which uses slower, but IEEE-compliant,
14547 software comparisons.
14549 @item -mstack-offset=@var{num}
14550 @opindex mstack-offset
14551 Set the offset between the top of the stack and the stack pointer.
14552 E.g., a value of 8 means that the eight bytes in the range @code{sp+0@dots{}sp+7}
14553 can be used by leaf functions without stack allocation.
14554 Values other than @samp{8} or @samp{16} are untested and unlikely to work.
14555 Note also that this option changes the ABI; compiling a program with a
14556 different stack offset than the libraries have been compiled with
14557 generally does not work.
14558 This option can be useful if you want to evaluate if a different stack
14559 offset would give you better code, but to actually use a different stack
14560 offset to build working programs, it is recommended to configure the
14561 toolchain with the appropriate @option{--with-stack-offset=@var{num}} option.
14563 @item -mno-round-nearest
14564 @opindex mno-round-nearest
14565 Make the scheduler assume that the rounding mode has been set to
14566 truncating. The default is @option{-mround-nearest}.
14569 @opindex mlong-calls
14570 If not otherwise specified by an attribute, assume all calls might be beyond
14571 the offset range of the @code{b} / @code{bl} instructions, and therefore load the
14572 function address into a register before performing a (otherwise direct) call.
14573 This is the default.
14575 @item -mshort-calls
14576 @opindex short-calls
14577 If not otherwise specified by an attribute, assume all direct calls are
14578 in the range of the @code{b} / @code{bl} instructions, so use these instructions
14579 for direct calls. The default is @option{-mlong-calls}.
14583 Assume addresses can be loaded as 16-bit unsigned values. This does not
14584 apply to function addresses for which @option{-mlong-calls} semantics
14587 @item -mfp-mode=@var{mode}
14589 Set the prevailing mode of the floating-point unit.
14590 This determines the floating-point mode that is provided and expected
14591 at function call and return time. Making this mode match the mode you
14592 predominantly need at function start can make your programs smaller and
14593 faster by avoiding unnecessary mode switches.
14595 @var{mode} can be set to one the following values:
14599 Any mode at function entry is valid, and retained or restored when
14600 the function returns, and when it calls other functions.
14601 This mode is useful for compiling libraries or other compilation units
14602 you might want to incorporate into different programs with different
14603 prevailing FPU modes, and the convenience of being able to use a single
14604 object file outweighs the size and speed overhead for any extra
14605 mode switching that might be needed, compared with what would be needed
14606 with a more specific choice of prevailing FPU mode.
14609 This is the mode used for floating-point calculations with
14610 truncating (i.e.@: round towards zero) rounding mode. That includes
14611 conversion from floating point to integer.
14613 @item round-nearest
14614 This is the mode used for floating-point calculations with
14615 round-to-nearest-or-even rounding mode.
14618 This is the mode used to perform integer calculations in the FPU, e.g.@:
14619 integer multiply, or integer multiply-and-accumulate.
14622 The default is @option{-mfp-mode=caller}
14624 @item -mnosplit-lohi
14625 @itemx -mno-postinc
14626 @itemx -mno-postmodify
14627 @opindex mnosplit-lohi
14628 @opindex mno-postinc
14629 @opindex mno-postmodify
14630 Code generation tweaks that disable, respectively, splitting of 32-bit
14631 loads, generation of post-increment addresses, and generation of
14632 post-modify addresses. The defaults are @option{msplit-lohi},
14633 @option{-mpost-inc}, and @option{-mpost-modify}.
14635 @item -mnovect-double
14636 @opindex mno-vect-double
14637 Change the preferred SIMD mode to SImode. The default is
14638 @option{-mvect-double}, which uses DImode as preferred SIMD mode.
14640 @item -max-vect-align=@var{num}
14641 @opindex max-vect-align
14642 The maximum alignment for SIMD vector mode types.
14643 @var{num} may be 4 or 8. The default is 8.
14644 Note that this is an ABI change, even though many library function
14645 interfaces are unaffected if they don't use SIMD vector modes
14646 in places that affect size and/or alignment of relevant types.
14648 @item -msplit-vecmove-early
14649 @opindex msplit-vecmove-early
14650 Split vector moves into single word moves before reload. In theory this
14651 can give better register allocation, but so far the reverse seems to be
14652 generally the case.
14654 @item -m1reg-@var{reg}
14656 Specify a register to hold the constant @minus{}1, which makes loading small negative
14657 constants and certain bitmasks faster.
14658 Allowable values for @var{reg} are @samp{r43} and @samp{r63},
14659 which specify use of that register as a fixed register,
14660 and @samp{none}, which means that no register is used for this
14661 purpose. The default is @option{-m1reg-none}.
14666 @subsection ARC Options
14667 @cindex ARC options
14669 The following options control the architecture variant for which code
14672 @c architecture variants
14675 @item -mbarrel-shifter
14676 @opindex mbarrel-shifter
14677 Generate instructions supported by barrel shifter. This is the default
14678 unless @option{-mcpu=ARC601} or @samp{-mcpu=ARCEM} is in effect.
14680 @item -mcpu=@var{cpu}
14682 Set architecture type, register usage, and instruction scheduling
14683 parameters for @var{cpu}. There are also shortcut alias options
14684 available for backward compatibility and convenience. Supported
14685 values for @var{cpu} are
14691 Compile for ARC600. Aliases: @option{-mA6}, @option{-mARC600}.
14695 Compile for ARC601. Alias: @option{-mARC601}.
14700 Compile for ARC700. Aliases: @option{-mA7}, @option{-mARC700}.
14701 This is the default when configured with @option{--with-cpu=arc700}@.
14704 Compile for ARC EM.
14707 Compile for ARC HS.
14710 Compile for ARC EM CPU with no hardware extensions.
14713 Compile for ARC EM4 CPU.
14716 Compile for ARC EM4 DMIPS CPU.
14719 Compile for ARC EM4 DMIPS CPU with the single-precision floating-point
14723 Compile for ARC EM4 DMIPS CPU with single-precision floating-point and
14724 double assist instructions.
14727 Compile for ARC HS CPU with no hardware extensions except the atomic
14731 Compile for ARC HS34 CPU.
14734 Compile for ARC HS38 CPU.
14737 Compile for ARC HS38 CPU with all hardware extensions on.
14740 Compile for ARC 600 CPU with @code{norm} instructions enabled.
14742 @item arc600_mul32x16
14743 Compile for ARC 600 CPU with @code{norm} and 32x16-bit multiply
14744 instructions enabled.
14747 Compile for ARC 600 CPU with @code{norm} and @code{mul64}-family
14748 instructions enabled.
14751 Compile for ARC 601 CPU with @code{norm} instructions enabled.
14753 @item arc601_mul32x16
14754 Compile for ARC 601 CPU with @code{norm} and 32x16-bit multiply
14755 instructions enabled.
14758 Compile for ARC 601 CPU with @code{norm} and @code{mul64}-family
14759 instructions enabled.
14762 Compile for ARC 700 on NPS400 chip.
14768 @itemx -mdpfp-compact
14769 @opindex mdpfp-compact
14770 Generate double-precision FPX instructions, tuned for the compact
14774 @opindex mdpfp-fast
14775 Generate double-precision FPX instructions, tuned for the fast
14778 @item -mno-dpfp-lrsr
14779 @opindex mno-dpfp-lrsr
14780 Disable @code{lr} and @code{sr} instructions from using FPX extension
14785 Generate extended arithmetic instructions. Currently only
14786 @code{divaw}, @code{adds}, @code{subs}, and @code{sat16} are
14787 supported. This is always enabled for @option{-mcpu=ARC700}.
14791 Do not generate @code{mpy}-family instructions for ARC700. This option is
14796 Generate 32x16-bit multiply and multiply-accumulate instructions.
14800 Generate @code{mul64} and @code{mulu64} instructions.
14801 Only valid for @option{-mcpu=ARC600}.
14805 Generate @code{norm} instructions. This is the default if @option{-mcpu=ARC700}
14810 @itemx -mspfp-compact
14811 @opindex mspfp-compact
14812 Generate single-precision FPX instructions, tuned for the compact
14816 @opindex mspfp-fast
14817 Generate single-precision FPX instructions, tuned for the fast
14822 Enable generation of ARC SIMD instructions via target-specific
14823 builtins. Only valid for @option{-mcpu=ARC700}.
14826 @opindex msoft-float
14827 This option ignored; it is provided for compatibility purposes only.
14828 Software floating-point code is emitted by default, and this default
14829 can overridden by FPX options; @option{-mspfp}, @option{-mspfp-compact}, or
14830 @option{-mspfp-fast} for single precision, and @option{-mdpfp},
14831 @option{-mdpfp-compact}, or @option{-mdpfp-fast} for double precision.
14835 Generate @code{swap} instructions.
14839 This enables use of the locked load/store conditional extension to implement
14840 atomic memory built-in functions. Not available for ARC 6xx or ARC
14845 Enable @code{div} and @code{rem} instructions for ARCv2 cores.
14847 @item -mcode-density
14848 @opindex mcode-density
14849 Enable code density instructions for ARC EM.
14850 This option is on by default for ARC HS.
14854 Enable double load/store operations for ARC HS cores.
14856 @item -mtp-regno=@var{regno}
14858 Specify thread pointer register number.
14860 @item -mmpy-option=@var{multo}
14861 @opindex mmpy-option
14862 Compile ARCv2 code with a multiplier design option. You can specify
14863 the option using either a string or numeric value for @var{multo}.
14864 @samp{wlh1} is the default value. The recognized values are:
14869 No multiplier available.
14873 16x16 multiplier, fully pipelined.
14874 The following instructions are enabled: @code{mpyw} and @code{mpyuw}.
14878 32x32 multiplier, fully
14879 pipelined (1 stage). The following instructions are additionally
14880 enabled: @code{mpy}, @code{mpyu}, @code{mpym}, @code{mpymu}, and @code{mpy_s}.
14884 32x32 multiplier, fully pipelined
14885 (2 stages). The following instructions are additionally enabled: @code{mpy},
14886 @code{mpyu}, @code{mpym}, @code{mpymu}, and @code{mpy_s}.
14890 Two 16x16 multipliers, blocking,
14891 sequential. The following instructions are additionally enabled: @code{mpy},
14892 @code{mpyu}, @code{mpym}, @code{mpymu}, and @code{mpy_s}.
14896 One 16x16 multiplier, blocking,
14897 sequential. The following instructions are additionally enabled: @code{mpy},
14898 @code{mpyu}, @code{mpym}, @code{mpymu}, and @code{mpy_s}.
14902 One 32x4 multiplier, blocking,
14903 sequential. The following instructions are additionally enabled: @code{mpy},
14904 @code{mpyu}, @code{mpym}, @code{mpymu}, and @code{mpy_s}.
14908 ARC HS SIMD support.
14912 ARC HS SIMD support.
14916 ARC HS SIMD support.
14920 This option is only available for ARCv2 cores@.
14922 @item -mfpu=@var{fpu}
14924 Enables support for specific floating-point hardware extensions for ARCv2
14925 cores. Supported values for @var{fpu} are:
14930 Enables support for single-precision floating-point hardware
14934 Enables support for double-precision floating-point hardware
14935 extensions. The single-precision floating-point extension is also
14936 enabled. Not available for ARC EM@.
14939 Enables support for double-precision floating-point hardware
14940 extensions using double-precision assist instructions. The single-precision
14941 floating-point extension is also enabled. This option is
14942 only available for ARC EM@.
14945 Enables support for double-precision floating-point hardware
14946 extensions using double-precision assist instructions.
14947 The single-precision floating-point, square-root, and divide
14948 extensions are also enabled. This option is
14949 only available for ARC EM@.
14952 Enables support for double-precision floating-point hardware
14953 extensions using double-precision assist instructions.
14954 The single-precision floating-point and fused multiply and add
14955 hardware extensions are also enabled. This option is
14956 only available for ARC EM@.
14959 Enables support for double-precision floating-point hardware
14960 extensions using double-precision assist instructions.
14961 All single-precision floating-point hardware extensions are also
14962 enabled. This option is only available for ARC EM@.
14965 Enables support for single-precision floating-point, square-root and divide
14966 hardware extensions@.
14969 Enables support for double-precision floating-point, square-root and divide
14970 hardware extensions. This option
14971 includes option @samp{fpus_div}. Not available for ARC EM@.
14974 Enables support for single-precision floating-point and
14975 fused multiply and add hardware extensions@.
14978 Enables support for double-precision floating-point and
14979 fused multiply and add hardware extensions. This option
14980 includes option @samp{fpus_fma}. Not available for ARC EM@.
14983 Enables support for all single-precision floating-point hardware
14987 Enables support for all single- and double-precision floating-point
14988 hardware extensions. Not available for ARC EM@.
14992 @item -mirq-ctrl-saved=@var{register-range}, @var{blink}, @var{lp_count}
14993 @opindex mirq-ctrl-saved
14994 Specifies general-purposes registers that the processor automatically
14995 saves/restores on interrupt entry and exit. @var{register-range} is
14996 specified as two registers separated by a dash. The register range
14997 always starts with @code{r0}, the upper limit is @code{fp} register.
14998 @var{blink} and @var{lp_count} are optional. This option is only
14999 valid for ARC EM and ARC HS cores.
15001 @item -mrgf-banked-regs=@var{number}
15002 @opindex mrgf-banked-regs
15003 Specifies the number of registers replicated in second register bank
15004 on entry to fast interrupt. Fast interrupts are interrupts with the
15005 highest priority level P0. These interrupts save only PC and STATUS32
15006 registers to avoid memory transactions during interrupt entry and exit
15007 sequences. Use this option when you are using fast interrupts in an
15008 ARC V2 family processor. Permitted values are 4, 8, 16, and 32.
15010 @item -mlpc-width=@var{width}
15011 @opindex mlpc-width
15012 Specify the width of the @code{lp_count} register. Valid values for
15013 @var{width} are 8, 16, 20, 24, 28 and 32 bits. The default width is
15014 fixed to 32 bits. If the width is less than 32, the compiler does not
15015 attempt to transform loops in your program to use the zero-delay loop
15016 mechanism unless it is known that the @code{lp_count} register can
15017 hold the required loop-counter value. Depending on the width
15018 specified, the compiler and run-time library might continue to use the
15019 loop mechanism for various needs. This option defines macro
15020 @code{__ARC_LPC_WIDTH__} with the value of @var{width}.
15024 The following options are passed through to the assembler, and also
15025 define preprocessor macro symbols.
15027 @c Flags used by the assembler, but for which we define preprocessor
15028 @c macro symbols as well.
15031 @opindex mdsp-packa
15032 Passed down to the assembler to enable the DSP Pack A extensions.
15033 Also sets the preprocessor symbol @code{__Xdsp_packa}. This option is
15038 Passed down to the assembler to enable the dual Viterbi butterfly
15039 extension. Also sets the preprocessor symbol @code{__Xdvbf}. This
15040 option is deprecated.
15042 @c ARC700 4.10 extension instruction
15045 Passed down to the assembler to enable the locked load/store
15046 conditional extension. Also sets the preprocessor symbol
15051 Passed down to the assembler. Also sets the preprocessor symbol
15052 @code{__Xxmac_d16}. This option is deprecated.
15056 Passed down to the assembler. Also sets the preprocessor symbol
15057 @code{__Xxmac_24}. This option is deprecated.
15059 @c ARC700 4.10 extension instruction
15062 Passed down to the assembler to enable the 64-bit time-stamp counter
15063 extension instruction. Also sets the preprocessor symbol
15064 @code{__Xrtsc}. This option is deprecated.
15066 @c ARC700 4.10 extension instruction
15069 Passed down to the assembler to enable the swap byte ordering
15070 extension instruction. Also sets the preprocessor symbol
15074 @opindex mtelephony
15075 Passed down to the assembler to enable dual- and single-operand
15076 instructions for telephony. Also sets the preprocessor symbol
15077 @code{__Xtelephony}. This option is deprecated.
15081 Passed down to the assembler to enable the XY memory extension. Also
15082 sets the preprocessor symbol @code{__Xxy}.
15086 The following options control how the assembly code is annotated:
15088 @c Assembly annotation options
15092 Annotate assembler instructions with estimated addresses.
15094 @item -mannotate-align
15095 @opindex mannotate-align
15096 Explain what alignment considerations lead to the decision to make an
15097 instruction short or long.
15101 The following options are passed through to the linker:
15103 @c options passed through to the linker
15107 Passed through to the linker, to specify use of the @code{arclinux} emulation.
15108 This option is enabled by default in tool chains built for
15109 @w{@code{arc-linux-uclibc}} and @w{@code{arceb-linux-uclibc}} targets
15110 when profiling is not requested.
15112 @item -marclinux_prof
15113 @opindex marclinux_prof
15114 Passed through to the linker, to specify use of the
15115 @code{arclinux_prof} emulation. This option is enabled by default in
15116 tool chains built for @w{@code{arc-linux-uclibc}} and
15117 @w{@code{arceb-linux-uclibc}} targets when profiling is requested.
15121 The following options control the semantics of generated code:
15123 @c semantically relevant code generation options
15126 @opindex mlong-calls
15127 Generate calls as register indirect calls, thus providing access
15128 to the full 32-bit address range.
15130 @item -mmedium-calls
15131 @opindex mmedium-calls
15132 Don't use less than 25-bit addressing range for calls, which is the
15133 offset available for an unconditional branch-and-link
15134 instruction. Conditional execution of function calls is suppressed, to
15135 allow use of the 25-bit range, rather than the 21-bit range with
15136 conditional branch-and-link. This is the default for tool chains built
15137 for @w{@code{arc-linux-uclibc}} and @w{@code{arceb-linux-uclibc}} targets.
15141 Put definitions of externally-visible data in a small data section if
15142 that data is no bigger than @var{num} bytes. The default value of
15143 @var{num} is 4 for any ARC configuration, or 8 when we have double
15144 load/store operations.
15148 Do not generate sdata references. This is the default for tool chains
15149 built for @w{@code{arc-linux-uclibc}} and @w{@code{arceb-linux-uclibc}}
15152 @item -mvolatile-cache
15153 @opindex mvolatile-cache
15154 Use ordinarily cached memory accesses for volatile references. This is the
15157 @item -mno-volatile-cache
15158 @opindex mno-volatile-cache
15159 Enable cache bypass for volatile references.
15163 The following options fine tune code generation:
15164 @c code generation tuning options
15167 @opindex malign-call
15168 Do alignment optimizations for call instructions.
15170 @item -mauto-modify-reg
15171 @opindex mauto-modify-reg
15172 Enable the use of pre/post modify with register displacement.
15174 @item -mbbit-peephole
15175 @opindex mbbit-peephole
15176 Enable bbit peephole2.
15180 This option disables a target-specific pass in @file{arc_reorg} to
15181 generate compare-and-branch (@code{br@var{cc}}) instructions.
15182 It has no effect on
15183 generation of these instructions driven by the combiner pass.
15185 @item -mcase-vector-pcrel
15186 @opindex mcase-vector-pcrel
15187 Use PC-relative switch case tables to enable case table shortening.
15188 This is the default for @option{-Os}.
15190 @item -mcompact-casesi
15191 @opindex mcompact-casesi
15192 Enable compact @code{casesi} pattern. This is the default for @option{-Os},
15193 and only available for ARCv1 cores.
15195 @item -mno-cond-exec
15196 @opindex mno-cond-exec
15197 Disable the ARCompact-specific pass to generate conditional
15198 execution instructions.
15200 Due to delay slot scheduling and interactions between operand numbers,
15201 literal sizes, instruction lengths, and the support for conditional execution,
15202 the target-independent pass to generate conditional execution is often lacking,
15203 so the ARC port has kept a special pass around that tries to find more
15204 conditional execution generation opportunities after register allocation,
15205 branch shortening, and delay slot scheduling have been done. This pass
15206 generally, but not always, improves performance and code size, at the cost of
15207 extra compilation time, which is why there is an option to switch it off.
15208 If you have a problem with call instructions exceeding their allowable
15209 offset range because they are conditionalized, you should consider using
15210 @option{-mmedium-calls} instead.
15212 @item -mearly-cbranchsi
15213 @opindex mearly-cbranchsi
15214 Enable pre-reload use of the @code{cbranchsi} pattern.
15216 @item -mexpand-adddi
15217 @opindex mexpand-adddi
15218 Expand @code{adddi3} and @code{subdi3} at RTL generation time into
15219 @code{add.f}, @code{adc} etc. This option is deprecated.
15221 @item -mindexed-loads
15222 @opindex mindexed-loads
15223 Enable the use of indexed loads. This can be problematic because some
15224 optimizers then assume that indexed stores exist, which is not
15228 Enable Local Register Allocation. This is still experimental for ARC,
15229 so by default the compiler uses standard reload
15230 (i.e. @option{-mno-lra}).
15232 @item -mlra-priority-none
15233 @opindex mlra-priority-none
15234 Don't indicate any priority for target registers.
15236 @item -mlra-priority-compact
15237 @opindex mlra-priority-compact
15238 Indicate target register priority for r0..r3 / r12..r15.
15240 @item -mlra-priority-noncompact
15241 @opindex mlra-priority-noncompact
15242 Reduce target register priority for r0..r3 / r12..r15.
15244 @item -mno-millicode
15245 @opindex mno-millicode
15246 When optimizing for size (using @option{-Os}), prologues and epilogues
15247 that have to save or restore a large number of registers are often
15248 shortened by using call to a special function in libgcc; this is
15249 referred to as a @emph{millicode} call. As these calls can pose
15250 performance issues, and/or cause linking issues when linking in a
15251 nonstandard way, this option is provided to turn off millicode call
15255 @opindex mmixed-code
15256 Tweak register allocation to help 16-bit instruction generation.
15257 This generally has the effect of decreasing the average instruction size
15258 while increasing the instruction count.
15262 Enable @samp{q} instruction alternatives.
15263 This is the default for @option{-Os}.
15267 Enable @samp{Rcq} constraint handling.
15268 Most short code generation depends on this.
15269 This is the default.
15273 Enable @samp{Rcw} constraint handling.
15274 Most ccfsm condexec mostly depends on this.
15275 This is the default.
15277 @item -msize-level=@var{level}
15278 @opindex msize-level
15279 Fine-tune size optimization with regards to instruction lengths and alignment.
15280 The recognized values for @var{level} are:
15283 No size optimization. This level is deprecated and treated like @samp{1}.
15286 Short instructions are used opportunistically.
15289 In addition, alignment of loops and of code after barriers are dropped.
15292 In addition, optional data alignment is dropped, and the option @option{Os} is enabled.
15296 This defaults to @samp{3} when @option{-Os} is in effect. Otherwise,
15297 the behavior when this is not set is equivalent to level @samp{1}.
15299 @item -mtune=@var{cpu}
15301 Set instruction scheduling parameters for @var{cpu}, overriding any implied
15302 by @option{-mcpu=}.
15304 Supported values for @var{cpu} are
15308 Tune for ARC600 CPU.
15311 Tune for ARC601 CPU.
15314 Tune for ARC700 CPU with standard multiplier block.
15317 Tune for ARC700 CPU with XMAC block.
15320 Tune for ARC725D CPU.
15323 Tune for ARC750D CPU.
15327 @item -mmultcost=@var{num}
15329 Cost to assume for a multiply instruction, with @samp{4} being equal to a
15330 normal instruction.
15332 @item -munalign-prob-threshold=@var{probability}
15333 @opindex munalign-prob-threshold
15334 Set probability threshold for unaligning branches.
15335 When tuning for @samp{ARC700} and optimizing for speed, branches without
15336 filled delay slot are preferably emitted unaligned and long, unless
15337 profiling indicates that the probability for the branch to be taken
15338 is below @var{probability}. @xref{Cross-profiling}.
15339 The default is (REG_BR_PROB_BASE/2), i.e.@: 5000.
15343 The following options are maintained for backward compatibility, but
15344 are now deprecated and will be removed in a future release:
15346 @c Deprecated options
15354 @opindex mbig-endian
15357 Compile code for big-endian targets. Use of these options is now
15358 deprecated. Big-endian code is supported by configuring GCC to build
15359 @w{@code{arceb-elf32}} and @w{@code{arceb-linux-uclibc}} targets,
15360 for which big endian is the default.
15362 @item -mlittle-endian
15363 @opindex mlittle-endian
15366 Compile code for little-endian targets. Use of these options is now
15367 deprecated. Little-endian code is supported by configuring GCC to build
15368 @w{@code{arc-elf32}} and @w{@code{arc-linux-uclibc}} targets,
15369 for which little endian is the default.
15371 @item -mbarrel_shifter
15372 @opindex mbarrel_shifter
15373 Replaced by @option{-mbarrel-shifter}.
15375 @item -mdpfp_compact
15376 @opindex mdpfp_compact
15377 Replaced by @option{-mdpfp-compact}.
15380 @opindex mdpfp_fast
15381 Replaced by @option{-mdpfp-fast}.
15384 @opindex mdsp_packa
15385 Replaced by @option{-mdsp-packa}.
15389 Replaced by @option{-mea}.
15393 Replaced by @option{-mmac-24}.
15397 Replaced by @option{-mmac-d16}.
15399 @item -mspfp_compact
15400 @opindex mspfp_compact
15401 Replaced by @option{-mspfp-compact}.
15404 @opindex mspfp_fast
15405 Replaced by @option{-mspfp-fast}.
15407 @item -mtune=@var{cpu}
15409 Values @samp{arc600}, @samp{arc601}, @samp{arc700} and
15410 @samp{arc700-xmac} for @var{cpu} are replaced by @samp{ARC600},
15411 @samp{ARC601}, @samp{ARC700} and @samp{ARC700-xmac} respectively.
15413 @item -multcost=@var{num}
15415 Replaced by @option{-mmultcost}.
15420 @subsection ARM Options
15421 @cindex ARM options
15423 These @samp{-m} options are defined for the ARM port:
15426 @item -mabi=@var{name}
15428 Generate code for the specified ABI@. Permissible values are: @samp{apcs-gnu},
15429 @samp{atpcs}, @samp{aapcs}, @samp{aapcs-linux} and @samp{iwmmxt}.
15432 @opindex mapcs-frame
15433 Generate a stack frame that is compliant with the ARM Procedure Call
15434 Standard for all functions, even if this is not strictly necessary for
15435 correct execution of the code. Specifying @option{-fomit-frame-pointer}
15436 with this option causes the stack frames not to be generated for
15437 leaf functions. The default is @option{-mno-apcs-frame}.
15438 This option is deprecated.
15442 This is a synonym for @option{-mapcs-frame} and is deprecated.
15445 @c not currently implemented
15446 @item -mapcs-stack-check
15447 @opindex mapcs-stack-check
15448 Generate code to check the amount of stack space available upon entry to
15449 every function (that actually uses some stack space). If there is
15450 insufficient space available then either the function
15451 @code{__rt_stkovf_split_small} or @code{__rt_stkovf_split_big} is
15452 called, depending upon the amount of stack space required. The runtime
15453 system is required to provide these functions. The default is
15454 @option{-mno-apcs-stack-check}, since this produces smaller code.
15456 @c not currently implemented
15457 @item -mapcs-reentrant
15458 @opindex mapcs-reentrant
15459 Generate reentrant, position-independent code. The default is
15460 @option{-mno-apcs-reentrant}.
15463 @item -mthumb-interwork
15464 @opindex mthumb-interwork
15465 Generate code that supports calling between the ARM and Thumb
15466 instruction sets. Without this option, on pre-v5 architectures, the
15467 two instruction sets cannot be reliably used inside one program. The
15468 default is @option{-mno-thumb-interwork}, since slightly larger code
15469 is generated when @option{-mthumb-interwork} is specified. In AAPCS
15470 configurations this option is meaningless.
15472 @item -mno-sched-prolog
15473 @opindex mno-sched-prolog
15474 Prevent the reordering of instructions in the function prologue, or the
15475 merging of those instruction with the instructions in the function's
15476 body. This means that all functions start with a recognizable set
15477 of instructions (or in fact one of a choice from a small set of
15478 different function prologues), and this information can be used to
15479 locate the start of functions inside an executable piece of code. The
15480 default is @option{-msched-prolog}.
15482 @item -mfloat-abi=@var{name}
15483 @opindex mfloat-abi
15484 Specifies which floating-point ABI to use. Permissible values
15485 are: @samp{soft}, @samp{softfp} and @samp{hard}.
15487 Specifying @samp{soft} causes GCC to generate output containing
15488 library calls for floating-point operations.
15489 @samp{softfp} allows the generation of code using hardware floating-point
15490 instructions, but still uses the soft-float calling conventions.
15491 @samp{hard} allows generation of floating-point instructions
15492 and uses FPU-specific calling conventions.
15494 The default depends on the specific target configuration. Note that
15495 the hard-float and soft-float ABIs are not link-compatible; you must
15496 compile your entire program with the same ABI, and link with a
15497 compatible set of libraries.
15499 @item -mlittle-endian
15500 @opindex mlittle-endian
15501 Generate code for a processor running in little-endian mode. This is
15502 the default for all standard configurations.
15505 @opindex mbig-endian
15506 Generate code for a processor running in big-endian mode; the default is
15507 to compile code for a little-endian processor.
15512 When linking a big-endian image select between BE8 and BE32 formats.
15513 The option has no effect for little-endian images and is ignored. The
15514 default is dependent on the selected target architecture. For ARMv6
15515 and later architectures the default is BE8, for older architectures
15516 the default is BE32. BE32 format has been deprecated by ARM.
15518 @item -march=@var{name}@r{[}+extension@dots{}@r{]}
15520 This specifies the name of the target ARM architecture. GCC uses this
15521 name to determine what kind of instructions it can emit when generating
15522 assembly code. This option can be used in conjunction with or instead
15523 of the @option{-mcpu=} option.
15525 Permissible names are:
15527 @samp{armv5t}, @samp{armv5te},
15528 @samp{armv6}, @samp{armv6j}, @samp{armv6k}, @samp{armv6kz}, @samp{armv6t2},
15529 @samp{armv6z}, @samp{armv6zk},
15530 @samp{armv7}, @samp{armv7-a}, @samp{armv7ve},
15531 @samp{armv8-a}, @samp{armv8.1-a}, @samp{armv8.2-a}, @samp{armv8.3-a},
15534 @samp{armv6-m}, @samp{armv6s-m},
15535 @samp{armv7-m}, @samp{armv7e-m},
15536 @samp{armv8-m.base}, @samp{armv8-m.main},
15537 @samp{iwmmxt} and @samp{iwmmxt2}.
15539 Additionally, the following architectures, which lack support for the
15540 Thumb exection state, are recognized but support is deprecated:
15541 @samp{armv2}, @samp{armv2a}, @samp{armv3}, @samp{armv3m},
15542 @samp{armv4}, @samp{armv5} and @samp{armv5e}.
15544 Many of the architectures support extensions. These can be added by
15545 appending @samp{+@var{extension}} to the architecture name. Extension
15546 options are processed in order and capabilities accumulate. An extension
15547 will also enable any necessary base extensions
15548 upon which it depends. For example, the @samp{+crypto} extension
15549 will always enable the @samp{+simd} extension. The exception to the
15550 additive construction is for extensions that are prefixed with
15551 @samp{+no@dots{}}: these extensions disable the specified option and
15552 any other extensions that may depend on the presence of that
15555 For example, @samp{-march=armv7-a+simd+nofp+vfpv4} is equivalent to
15556 writing @samp{-march=armv7-a+vfpv4} since the @samp{+simd} option is
15557 entirely disabled by the @samp{+nofp} option that follows it.
15559 Most extension names are generically named, but have an effect that is
15560 dependent upon the architecture to which it is applied. For example,
15561 the @samp{+simd} option can be applied to both @samp{armv7-a} and
15562 @samp{armv8-a} architectures, but will enable the original ARMv7
15563 Advanced SIMD (Neon) extensions for @samp{armv7-a} and the ARMv8-a
15564 variant for @samp{armv8-a}.
15566 The table below lists the supported extensions for each architecture.
15567 Architectures not mentioned do not support any extensions.
15581 The VFPv2 floating-point instructions. The extension @samp{+vfpv2} can be
15582 used as an alias for this extension.
15585 Disable the floating-point instructions.
15589 The common subset of the ARMv7-A, ARMv7-R and ARMv7-M architectures.
15592 The VFPv3 floating-point instructions, with 16 double-precision
15593 registers. The extension @samp{+vfpv3-d16} can be used as an alias
15594 for this extension. Note that floating-point is not supported by the
15595 base ARMv7-M architecture, but is compatible with both the ARMv7-A and
15596 ARMv7-R architectures.
15599 Disable the floating-point instructions.
15605 The VFPv3 floating-point instructions, with 16 double-precision
15606 registers. The extension @samp{+vfpv3-d16} can be used as an alias
15607 for this extension.
15610 The Advanced SIMD (Neon) v1 and the VFPv3 floating-point instructions.
15611 The extensions @samp{+neon} and @samp{+neon-vfpv3} can be used as aliases
15612 for this extension.
15615 The VFPv3 floating-point instructions, with 32 double-precision
15618 @item +vfpv3-d16-fp16
15619 The VFPv3 floating-point instructions, with 16 double-precision
15620 registers and the half-precision floating-point conversion operations.
15623 The VFPv3 floating-point instructions, with 32 double-precision
15624 registers and the half-precision floating-point conversion operations.
15627 The VFPv4 floating-point instructions, with 16 double-precision
15631 The VFPv4 floating-point instructions, with 32 double-precision
15635 The Advanced SIMD (Neon) v1 and the VFPv3 floating-point instructions, with
15636 the half-precision floating-point conversion operations.
15639 The Advanced SIMD (Neon) v2 and the VFPv4 floating-point instructions.
15642 Disable the Advanced SIMD instructions (does not disable floating point).
15645 Disable the floating-point and Advanced SIMD instructions.
15649 The extended version of the ARMv7-A architecture with support for
15653 The VFPv4 floating-point instructions, with 16 double-precision registers.
15654 The extension @samp{+vfpv4-d16} can be used as an alias for this extension.
15657 The Advanced SIMD (Neon) v2 and the VFPv4 floating-point instructions. The
15658 extension @samp{+neon-vfpv4} can be used as an alias for this extension.
15661 The VFPv3 floating-point instructions, with 16 double-precision
15665 The VFPv3 floating-point instructions, with 32 double-precision
15668 @item +vfpv3-d16-fp16
15669 The VFPv3 floating-point instructions, with 16 double-precision
15670 registers and the half-precision floating-point conversion operations.
15673 The VFPv3 floating-point instructions, with 32 double-precision
15674 registers and the half-precision floating-point conversion operations.
15677 The VFPv4 floating-point instructions, with 16 double-precision
15681 The VFPv4 floating-point instructions, with 32 double-precision
15685 The Advanced SIMD (Neon) v1 and the VFPv3 floating-point instructions.
15686 The extension @samp{+neon-vfpv3} can be used as an alias for this extension.
15689 The Advanced SIMD (Neon) v1 and the VFPv3 floating-point instructions, with
15690 the half-precision floating-point conversion operations.
15693 Disable the Advanced SIMD instructions (does not disable floating point).
15696 Disable the floating-point and Advanced SIMD instructions.
15702 The Cyclic Redundancy Check (CRC) instructions.
15704 The ARMv8 Advanced SIMD and floating-point instructions.
15706 The cryptographic instructions.
15708 Disable the cryptographic isntructions.
15710 Disable the floating-point, Advanced SIMD and cryptographic instructions.
15716 The ARMv8.1 Advanced SIMD and floating-point instructions.
15719 The cryptographic instructions. This also enables the Advanced SIMD and
15720 floating-point instructions.
15723 Disable the cryptographic isntructions.
15726 Disable the floating-point, Advanced SIMD and cryptographic instructions.
15732 The half-precision floating-point data processing instructions.
15733 This also enables the Advanced SIMD and floating-point instructions.
15736 The ARMv8.1 Advanced SIMD and floating-point instructions.
15739 The cryptographic instructions. This also enables the Advanced SIMD and
15740 floating-point instructions.
15743 Enable the Dot Product extension. This also enables Advanced SIMD instructions.
15746 Disable the cryptographic extension.
15749 Disable the floating-point, Advanced SIMD and cryptographic instructions.
15755 The single-precision VFPv3 floating-point instructions. The extension
15756 @samp{+vfpv3xd} can be used as an alias for this extension.
15759 The VFPv3 floating-point instructions with 16 double-precision registers.
15760 The extension +vfpv3-d16 can be used as an alias for this extension.
15763 Disable the floating-point extension.
15766 The ARM-state integer division instructions.
15769 Disable the ARM-state integer division extension.
15775 The single-precision VFPv4 floating-point instructions.
15778 The single-precision FPv5 floating-point instructions.
15781 The single- and double-precision FPv5 floating-point instructions.
15784 Disable the floating-point extensions.
15790 The DSP instructions.
15793 Disable the DSP extension.
15796 The single-precision floating-point instructions.
15799 The single- and double-precision floating-point instructions.
15802 Disable the floating-point extension.
15808 The Cyclic Redundancy Check (CRC) instructions.
15810 The single-precision FPv5 floating-point instructions.
15812 The ARMv8 Advanced SIMD and floating-point instructions.
15814 The cryptographic instructions.
15816 Disable the cryptographic isntructions.
15818 Disable the floating-point, Advanced SIMD and cryptographic instructions.
15823 @option{-march=native} causes the compiler to auto-detect the architecture
15824 of the build computer. At present, this feature is only supported on
15825 GNU/Linux, and not all architectures are recognized. If the auto-detect
15826 is unsuccessful the option has no effect.
15828 @item -mtune=@var{name}
15830 This option specifies the name of the target ARM processor for
15831 which GCC should tune the performance of the code.
15832 For some ARM implementations better performance can be obtained by using
15834 Permissible names are: @samp{arm2}, @samp{arm250},
15835 @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610},
15836 @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm},
15837 @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700},
15838 @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100},
15840 @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm7tdmi-s},
15841 @samp{arm710t}, @samp{arm720t}, @samp{arm740t},
15842 @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100},
15843 @samp{strongarm1110},
15844 @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920},
15845 @samp{arm920t}, @samp{arm922t}, @samp{arm946e-s}, @samp{arm966e-s},
15846 @samp{arm968e-s}, @samp{arm926ej-s}, @samp{arm940t}, @samp{arm9tdmi},
15847 @samp{arm10tdmi}, @samp{arm1020t}, @samp{arm1026ej-s},
15848 @samp{arm10e}, @samp{arm1020e}, @samp{arm1022e},
15849 @samp{arm1136j-s}, @samp{arm1136jf-s}, @samp{mpcore}, @samp{mpcorenovfp},
15850 @samp{arm1156t2-s}, @samp{arm1156t2f-s}, @samp{arm1176jz-s}, @samp{arm1176jzf-s},
15851 @samp{generic-armv7-a}, @samp{cortex-a5}, @samp{cortex-a7}, @samp{cortex-a8},
15852 @samp{cortex-a9}, @samp{cortex-a12}, @samp{cortex-a15}, @samp{cortex-a17},
15853 @samp{cortex-a32}, @samp{cortex-a35}, @samp{cortex-a53}, @samp{cortex-a55},
15854 @samp{cortex-a57}, @samp{cortex-a72}, @samp{cortex-a73}, @samp{cortex-a75},
15855 @samp{cortex-r4}, @samp{cortex-r4f}, @samp{cortex-r5}, @samp{cortex-r7},
15856 @samp{cortex-r8}, @samp{cortex-r52},
15864 @samp{cortex-m0plus},
15865 @samp{cortex-m1.small-multiply},
15866 @samp{cortex-m0.small-multiply},
15867 @samp{cortex-m0plus.small-multiply},
15869 @samp{marvell-pj4},
15870 @samp{xscale}, @samp{iwmmxt}, @samp{iwmmxt2}, @samp{ep9312},
15871 @samp{fa526}, @samp{fa626},
15872 @samp{fa606te}, @samp{fa626te}, @samp{fmp626}, @samp{fa726te},
15875 Additionally, this option can specify that GCC should tune the performance
15876 of the code for a big.LITTLE system. Permissible names are:
15877 @samp{cortex-a15.cortex-a7}, @samp{cortex-a17.cortex-a7},
15878 @samp{cortex-a57.cortex-a53}, @samp{cortex-a72.cortex-a53},
15879 @samp{cortex-a72.cortex-a35}, @samp{cortex-a73.cortex-a53},
15880 @samp{cortex-a75.cortex-a55}.
15882 @option{-mtune=generic-@var{arch}} specifies that GCC should tune the
15883 performance for a blend of processors within architecture @var{arch}.
15884 The aim is to generate code that run well on the current most popular
15885 processors, balancing between optimizations that benefit some CPUs in the
15886 range, and avoiding performance pitfalls of other CPUs. The effects of
15887 this option may change in future GCC versions as CPU models come and go.
15889 @option{-mtune} permits the same extension options as @option{-mcpu}, but
15890 the extension options do not affect the tuning of the generated code.
15892 @option{-mtune=native} causes the compiler to auto-detect the CPU
15893 of the build computer. At present, this feature is only supported on
15894 GNU/Linux, and not all architectures are recognized. If the auto-detect is
15895 unsuccessful the option has no effect.
15897 @item -mcpu=@var{name}@r{[}+extension@dots{}@r{]}
15899 This specifies the name of the target ARM processor. GCC uses this name
15900 to derive the name of the target ARM architecture (as if specified
15901 by @option{-march}) and the ARM processor type for which to tune for
15902 performance (as if specified by @option{-mtune}). Where this option
15903 is used in conjunction with @option{-march} or @option{-mtune},
15904 those options take precedence over the appropriate part of this option.
15906 Many of the supported CPUs implement optional architectural
15907 extensions. Where this is so the architectural extensions are
15908 normally enabled by default. If implementations that lack the
15909 extension exist, then the extension syntax can be used to disable
15910 those extensions that have been omitted. For floating-point and
15911 Advanced SIMD (Neon) instructions, the settings of the options
15912 @option{-mfloat-abi} and @option{-mfpu} must also be considered:
15913 floating-point and Advanced SIMD instructions will only be used if
15914 @option{-mfloat-abi} is not set to @samp{soft}; and any setting of
15915 @option{-mfpu} other than @samp{auto} will override the available
15916 floating-point and SIMD extension instructions.
15918 For example, @samp{cortex-a9} can be found in three major
15919 configurations: integer only, with just a floating-point unit or with
15920 floating-point and Advanced SIMD. The default is to enable all the
15921 instructions, but the extensions @samp{+nosimd} and @samp{+nofp} can
15922 be used to disable just the SIMD or both the SIMD and floating-point
15923 instructions respectively.
15925 Permissible names for this option are the same as those for
15928 The following extension options are common to the listed CPUs:
15932 Disable the DSP instructions on @samp{cortex-m33}.
15935 Disables the floating-point instructions on @samp{arm9e},
15936 @samp{arm946e-s}, @samp{arm966e-s}, @samp{arm968e-s}, @samp{arm10e},
15937 @samp{arm1020e}, @samp{arm1022e}, @samp{arm926ej-s},
15938 @samp{arm1026ej-s}, @samp{cortex-r5}, @samp{cortex-r7}, @samp{cortex-r8},
15939 @samp{cortex-m4}, @samp{cortex-m7} and @samp{cortex-m33}.
15940 Disables the floating-point and SIMD instructions on
15941 @samp{generic-armv7-a}, @samp{cortex-a5}, @samp{cortex-a7},
15942 @samp{cortex-a8}, @samp{cortex-a9}, @samp{cortex-a12},
15943 @samp{cortex-a15}, @samp{cortex-a17}, @samp{cortex-a15.cortex-a7},
15944 @samp{cortex-a17.cortex-a7}, @samp{cortex-a32}, @samp{cortex-a35},
15945 @samp{cortex-a53} and @samp{cortex-a55}.
15948 Disables the double-precision component of the floating-point instructions
15949 on @samp{cortex-r5}, @samp{cortex-r52} and @samp{cortex-m7}.
15952 Disables the SIMD (but not floating-point) instructions on
15953 @samp{generic-armv7-a}, @samp{cortex-a5}, @samp{cortex-a7}
15954 and @samp{cortex-a9}.
15957 Enables the cryptographic instructions on @samp{cortex-a32},
15958 @samp{cortex-a35}, @samp{cortex-a53}, @samp{cortex-a55}, @samp{cortex-a57},
15959 @samp{cortex-a72}, @samp{cortex-a73}, @samp{cortex-a75}, @samp{exynos-m1},
15960 @samp{xgene1}, @samp{cortex-a57.cortex-a53}, @samp{cortex-a72.cortex-a53},
15961 @samp{cortex-a73.cortex-a35}, @samp{cortex-a73.cortex-a53} and
15962 @samp{cortex-a75.cortex-a55}.
15965 Additionally the @samp{generic-armv7-a} pseudo target defaults to
15966 VFPv3 with 16 double-precision registers. It supports the following
15967 extension options: @samp{vfpv3-d16}, @samp{vfpv3},
15968 @samp{vfpv3-d16-fp16}, @samp{vfpv3-fp16}, @samp{vfpv4-d16},
15969 @samp{vfpv4}, @samp{neon}, @samp{neon-vfpv3}, @samp{neon-fp16},
15970 @samp{neon-vfpv4}. The meanings are the same as for the extensions to
15971 @option{-march=armv7-a}.
15973 @option{-mcpu=generic-@var{arch}} is also permissible, and is
15974 equivalent to @option{-march=@var{arch} -mtune=generic-@var{arch}}.
15975 See @option{-mtune} for more information.
15977 @option{-mcpu=native} causes the compiler to auto-detect the CPU
15978 of the build computer. At present, this feature is only supported on
15979 GNU/Linux, and not all architectures are recognized. If the auto-detect
15980 is unsuccessful the option has no effect.
15982 @item -mfpu=@var{name}
15984 This specifies what floating-point hardware (or hardware emulation) is
15985 available on the target. Permissible names are: @samp{auto}, @samp{vfpv2},
15987 @samp{vfpv3-fp16}, @samp{vfpv3-d16}, @samp{vfpv3-d16-fp16}, @samp{vfpv3xd},
15988 @samp{vfpv3xd-fp16}, @samp{neon-vfpv3}, @samp{neon-fp16}, @samp{vfpv4},
15989 @samp{vfpv4-d16}, @samp{fpv4-sp-d16}, @samp{neon-vfpv4},
15990 @samp{fpv5-d16}, @samp{fpv5-sp-d16},
15991 @samp{fp-armv8}, @samp{neon-fp-armv8} and @samp{crypto-neon-fp-armv8}.
15992 Note that @samp{neon} is an alias for @samp{neon-vfpv3} and @samp{vfp}
15993 is an alias for @samp{vfpv2}.
15995 The setting @samp{auto} is the default and is special. It causes the
15996 compiler to select the floating-point and Advanced SIMD instructions
15997 based on the settings of @option{-mcpu} and @option{-march}.
15999 If the selected floating-point hardware includes the NEON extension
16000 (e.g. @option{-mfpu=neon}), note that floating-point
16001 operations are not generated by GCC's auto-vectorization pass unless
16002 @option{-funsafe-math-optimizations} is also specified. This is
16003 because NEON hardware does not fully implement the IEEE 754 standard for
16004 floating-point arithmetic (in particular denormal values are treated as
16005 zero), so the use of NEON instructions may lead to a loss of precision.
16007 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}).
16009 @item -mfp16-format=@var{name}
16010 @opindex mfp16-format
16011 Specify the format of the @code{__fp16} half-precision floating-point type.
16012 Permissible names are @samp{none}, @samp{ieee}, and @samp{alternative};
16013 the default is @samp{none}, in which case the @code{__fp16} type is not
16014 defined. @xref{Half-Precision}, for more information.
16016 @item -mstructure-size-boundary=@var{n}
16017 @opindex mstructure-size-boundary
16018 The sizes of all structures and unions are rounded up to a multiple
16019 of the number of bits set by this option. Permissible values are 8, 32
16020 and 64. The default value varies for different toolchains. For the COFF
16021 targeted toolchain the default value is 8. A value of 64 is only allowed
16022 if the underlying ABI supports it.
16024 Specifying a larger number can produce faster, more efficient code, but
16025 can also increase the size of the program. Different values are potentially
16026 incompatible. Code compiled with one value cannot necessarily expect to
16027 work with code or libraries compiled with another value, if they exchange
16028 information using structures or unions.
16030 This option is deprecated.
16032 @item -mabort-on-noreturn
16033 @opindex mabort-on-noreturn
16034 Generate a call to the function @code{abort} at the end of a
16035 @code{noreturn} function. It is executed if the function tries to
16039 @itemx -mno-long-calls
16040 @opindex mlong-calls
16041 @opindex mno-long-calls
16042 Tells the compiler to perform function calls by first loading the
16043 address of the function into a register and then performing a subroutine
16044 call on this register. This switch is needed if the target function
16045 lies outside of the 64-megabyte addressing range of the offset-based
16046 version of subroutine call instruction.
16048 Even if this switch is enabled, not all function calls are turned
16049 into long calls. The heuristic is that static functions, functions
16050 that have the @code{short_call} attribute, functions that are inside
16051 the scope of a @code{#pragma no_long_calls} directive, and functions whose
16052 definitions have already been compiled within the current compilation
16053 unit are not turned into long calls. The exceptions to this rule are
16054 that weak function definitions, functions with the @code{long_call}
16055 attribute or the @code{section} attribute, and functions that are within
16056 the scope of a @code{#pragma long_calls} directive are always
16057 turned into long calls.
16059 This feature is not enabled by default. Specifying
16060 @option{-mno-long-calls} restores the default behavior, as does
16061 placing the function calls within the scope of a @code{#pragma
16062 long_calls_off} directive. Note these switches have no effect on how
16063 the compiler generates code to handle function calls via function
16066 @item -msingle-pic-base
16067 @opindex msingle-pic-base
16068 Treat the register used for PIC addressing as read-only, rather than
16069 loading it in the prologue for each function. The runtime system is
16070 responsible for initializing this register with an appropriate value
16071 before execution begins.
16073 @item -mpic-register=@var{reg}
16074 @opindex mpic-register
16075 Specify the register to be used for PIC addressing.
16076 For standard PIC base case, the default is any suitable register
16077 determined by compiler. For single PIC base case, the default is
16078 @samp{R9} if target is EABI based or stack-checking is enabled,
16079 otherwise the default is @samp{R10}.
16081 @item -mpic-data-is-text-relative
16082 @opindex mpic-data-is-text-relative
16083 Assume that the displacement between the text and data segments is fixed
16084 at static link time. This permits using PC-relative addressing
16085 operations to access data known to be in the data segment. For
16086 non-VxWorks RTP targets, this option is enabled by default. When
16087 disabled on such targets, it will enable @option{-msingle-pic-base} by
16090 @item -mpoke-function-name
16091 @opindex mpoke-function-name
16092 Write the name of each function into the text section, directly
16093 preceding the function prologue. The generated code is similar to this:
16097 .ascii "arm_poke_function_name", 0
16100 .word 0xff000000 + (t1 - t0)
16101 arm_poke_function_name
16103 stmfd sp!, @{fp, ip, lr, pc@}
16107 When performing a stack backtrace, code can inspect the value of
16108 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
16109 location @code{pc - 12} and the top 8 bits are set, then we know that
16110 there is a function name embedded immediately preceding this location
16111 and has length @code{((pc[-3]) & 0xff000000)}.
16118 Select between generating code that executes in ARM and Thumb
16119 states. The default for most configurations is to generate code
16120 that executes in ARM state, but the default can be changed by
16121 configuring GCC with the @option{--with-mode=}@var{state}
16124 You can also override the ARM and Thumb mode for each function
16125 by using the @code{target("thumb")} and @code{target("arm")} function attributes
16126 (@pxref{ARM Function Attributes}) or pragmas (@pxref{Function Specific Option Pragmas}).
16129 @opindex mtpcs-frame
16130 Generate a stack frame that is compliant with the Thumb Procedure Call
16131 Standard for all non-leaf functions. (A leaf function is one that does
16132 not call any other functions.) The default is @option{-mno-tpcs-frame}.
16134 @item -mtpcs-leaf-frame
16135 @opindex mtpcs-leaf-frame
16136 Generate a stack frame that is compliant with the Thumb Procedure Call
16137 Standard for all leaf functions. (A leaf function is one that does
16138 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
16140 @item -mcallee-super-interworking
16141 @opindex mcallee-super-interworking
16142 Gives all externally visible functions in the file being compiled an ARM
16143 instruction set header which switches to Thumb mode before executing the
16144 rest of the function. This allows these functions to be called from
16145 non-interworking code. This option is not valid in AAPCS configurations
16146 because interworking is enabled by default.
16148 @item -mcaller-super-interworking
16149 @opindex mcaller-super-interworking
16150 Allows calls via function pointers (including virtual functions) to
16151 execute correctly regardless of whether the target code has been
16152 compiled for interworking or not. There is a small overhead in the cost
16153 of executing a function pointer if this option is enabled. This option
16154 is not valid in AAPCS configurations because interworking is enabled
16157 @item -mtp=@var{name}
16159 Specify the access model for the thread local storage pointer. The valid
16160 models are @samp{soft}, which generates calls to @code{__aeabi_read_tp},
16161 @samp{cp15}, which fetches the thread pointer from @code{cp15} directly
16162 (supported in the arm6k architecture), and @samp{auto}, which uses the
16163 best available method for the selected processor. The default setting is
16166 @item -mtls-dialect=@var{dialect}
16167 @opindex mtls-dialect
16168 Specify the dialect to use for accessing thread local storage. Two
16169 @var{dialect}s are supported---@samp{gnu} and @samp{gnu2}. The
16170 @samp{gnu} dialect selects the original GNU scheme for supporting
16171 local and global dynamic TLS models. The @samp{gnu2} dialect
16172 selects the GNU descriptor scheme, which provides better performance
16173 for shared libraries. The GNU descriptor scheme is compatible with
16174 the original scheme, but does require new assembler, linker and
16175 library support. Initial and local exec TLS models are unaffected by
16176 this option and always use the original scheme.
16178 @item -mword-relocations
16179 @opindex mword-relocations
16180 Only generate absolute relocations on word-sized values (i.e. R_ARM_ABS32).
16181 This is enabled by default on targets (uClinux, SymbianOS) where the runtime
16182 loader imposes this restriction, and when @option{-fpic} or @option{-fPIC}
16185 @item -mfix-cortex-m3-ldrd
16186 @opindex mfix-cortex-m3-ldrd
16187 Some Cortex-M3 cores can cause data corruption when @code{ldrd} instructions
16188 with overlapping destination and base registers are used. This option avoids
16189 generating these instructions. This option is enabled by default when
16190 @option{-mcpu=cortex-m3} is specified.
16192 @item -munaligned-access
16193 @itemx -mno-unaligned-access
16194 @opindex munaligned-access
16195 @opindex mno-unaligned-access
16196 Enables (or disables) reading and writing of 16- and 32- bit values
16197 from addresses that are not 16- or 32- bit aligned. By default
16198 unaligned access is disabled for all pre-ARMv6, all ARMv6-M and for
16199 ARMv8-M Baseline architectures, and enabled for all other
16200 architectures. If unaligned access is not enabled then words in packed
16201 data structures are accessed a byte at a time.
16203 The ARM attribute @code{Tag_CPU_unaligned_access} is set in the
16204 generated object file to either true or false, depending upon the
16205 setting of this option. If unaligned access is enabled then the
16206 preprocessor symbol @code{__ARM_FEATURE_UNALIGNED} is also
16209 @item -mneon-for-64bits
16210 @opindex mneon-for-64bits
16211 Enables using Neon to handle scalar 64-bits operations. This is
16212 disabled by default since the cost of moving data from core registers
16215 @item -mslow-flash-data
16216 @opindex mslow-flash-data
16217 Assume loading data from flash is slower than fetching instruction.
16218 Therefore literal load is minimized for better performance.
16219 This option is only supported when compiling for ARMv7 M-profile and
16222 @item -masm-syntax-unified
16223 @opindex masm-syntax-unified
16224 Assume inline assembler is using unified asm syntax. The default is
16225 currently off which implies divided syntax. This option has no impact
16226 on Thumb2. However, this may change in future releases of GCC.
16227 Divided syntax should be considered deprecated.
16229 @item -mrestrict-it
16230 @opindex mrestrict-it
16231 Restricts generation of IT blocks to conform to the rules of ARMv8.
16232 IT blocks can only contain a single 16-bit instruction from a select
16233 set of instructions. This option is on by default for ARMv8 Thumb mode.
16235 @item -mprint-tune-info
16236 @opindex mprint-tune-info
16237 Print CPU tuning information as comment in assembler file. This is
16238 an option used only for regression testing of the compiler and not
16239 intended for ordinary use in compiling code. This option is disabled
16243 @opindex mpure-code
16244 Do not allow constant data to be placed in code sections.
16245 Additionally, when compiling for ELF object format give all text sections the
16246 ELF processor-specific section attribute @code{SHF_ARM_PURECODE}. This option
16247 is only available when generating non-pic code for M-profile targets with the
16252 Generate secure code as per the "ARMv8-M Security Extensions: Requirements on
16253 Development Tools Engineering Specification", which can be found on
16254 @url{http://infocenter.arm.com/help/topic/com.arm.doc.ecm0359818/ECM0359818_armv8m_security_extensions_reqs_on_dev_tools_1_0.pdf}.
16258 @subsection AVR Options
16259 @cindex AVR Options
16261 These options are defined for AVR implementations:
16264 @item -mmcu=@var{mcu}
16266 Specify Atmel AVR instruction set architectures (ISA) or MCU type.
16268 The default for this option is@tie{}@samp{avr2}.
16270 GCC supports the following AVR devices and ISAs:
16272 @include avr-mmcu.texi
16277 Assume that all data in static storage can be accessed by LDS / STS
16278 instructions. This option has only an effect on reduced Tiny devices like
16279 ATtiny40. See also the @code{absdata}
16280 @ref{AVR Variable Attributes,variable attribute}.
16282 @item -maccumulate-args
16283 @opindex maccumulate-args
16284 Accumulate outgoing function arguments and acquire/release the needed
16285 stack space for outgoing function arguments once in function
16286 prologue/epilogue. Without this option, outgoing arguments are pushed
16287 before calling a function and popped afterwards.
16289 Popping the arguments after the function call can be expensive on
16290 AVR so that accumulating the stack space might lead to smaller
16291 executables because arguments need not be removed from the
16292 stack after such a function call.
16294 This option can lead to reduced code size for functions that perform
16295 several calls to functions that get their arguments on the stack like
16296 calls to printf-like functions.
16298 @item -mbranch-cost=@var{cost}
16299 @opindex mbranch-cost
16300 Set the branch costs for conditional branch instructions to
16301 @var{cost}. Reasonable values for @var{cost} are small, non-negative
16302 integers. The default branch cost is 0.
16304 @item -mcall-prologues
16305 @opindex mcall-prologues
16306 Functions prologues/epilogues are expanded as calls to appropriate
16307 subroutines. Code size is smaller.
16309 @item -mgas-isr-prologues
16310 @opindex mgas-isr-prologues
16311 Interrupt service routines (ISRs) may use the @code{__gcc_isr} pseudo
16312 instruction supported by GNU Binutils.
16313 If this option is on, the feature can still be disabled for individual
16314 ISRs by means of the @ref{AVR Function Attributes,,@code{no_gccisr}}
16315 function attribute. This feature is activated per default
16316 if optimization is on (but not with @option{-Og}, @pxref{Optimize Options}),
16317 and if GNU Binutils support @w{@uref{https://sourceware.org/PR21683,PR21683}}.
16321 Assume @code{int} to be 8-bit integer. This affects the sizes of all types: a
16322 @code{char} is 1 byte, an @code{int} is 1 byte, a @code{long} is 2 bytes,
16323 and @code{long long} is 4 bytes. Please note that this option does not
16324 conform to the C standards, but it results in smaller code
16327 @item -mn-flash=@var{num}
16329 Assume that the flash memory has a size of
16330 @var{num} times 64@tie{}KiB.
16332 @item -mno-interrupts
16333 @opindex mno-interrupts
16334 Generated code is not compatible with hardware interrupts.
16335 Code size is smaller.
16339 Try to replace @code{CALL} resp.@: @code{JMP} instruction by the shorter
16340 @code{RCALL} resp.@: @code{RJMP} instruction if applicable.
16341 Setting @option{-mrelax} just adds the @option{--mlink-relax} option to
16342 the assembler's command line and the @option{--relax} option to the
16343 linker's command line.
16345 Jump relaxing is performed by the linker because jump offsets are not
16346 known before code is located. Therefore, the assembler code generated by the
16347 compiler is the same, but the instructions in the executable may
16348 differ from instructions in the assembler code.
16350 Relaxing must be turned on if linker stubs are needed, see the
16351 section on @code{EIND} and linker stubs below.
16355 Assume that the device supports the Read-Modify-Write
16356 instructions @code{XCH}, @code{LAC}, @code{LAS} and @code{LAT}.
16358 @item -mshort-calls
16359 @opindex mshort-calls
16361 Assume that @code{RJMP} and @code{RCALL} can target the whole
16364 This option is used internally for multilib selection. It is
16365 not an optimization option, and you don't need to set it by hand.
16369 Treat the stack pointer register as an 8-bit register,
16370 i.e.@: assume the high byte of the stack pointer is zero.
16371 In general, you don't need to set this option by hand.
16373 This option is used internally by the compiler to select and
16374 build multilibs for architectures @code{avr2} and @code{avr25}.
16375 These architectures mix devices with and without @code{SPH}.
16376 For any setting other than @option{-mmcu=avr2} or @option{-mmcu=avr25}
16377 the compiler driver adds or removes this option from the compiler
16378 proper's command line, because the compiler then knows if the device
16379 or architecture has an 8-bit stack pointer and thus no @code{SPH}
16384 Use address register @code{X} in a way proposed by the hardware. This means
16385 that @code{X} is only used in indirect, post-increment or
16386 pre-decrement addressing.
16388 Without this option, the @code{X} register may be used in the same way
16389 as @code{Y} or @code{Z} which then is emulated by additional
16391 For example, loading a value with @code{X+const} addressing with a
16392 small non-negative @code{const < 64} to a register @var{Rn} is
16396 adiw r26, const ; X += const
16397 ld @var{Rn}, X ; @var{Rn} = *X
16398 sbiw r26, const ; X -= const
16402 @opindex mtiny-stack
16403 Only change the lower 8@tie{}bits of the stack pointer.
16405 @item -mfract-convert-truncate
16406 @opindex mfract-convert-truncate
16407 Allow to use truncation instead of rounding towards zero for fractional fixed-point types.
16410 @opindex nodevicelib
16411 Don't link against AVR-LibC's device specific library @code{lib<mcu>.a}.
16413 @item -Waddr-space-convert
16414 @opindex Waddr-space-convert
16415 Warn about conversions between address spaces in the case where the
16416 resulting address space is not contained in the incoming address space.
16418 @item -Wmisspelled-isr
16419 @opindex Wmisspelled-isr
16420 Warn if the ISR is misspelled, i.e. without __vector prefix.
16421 Enabled by default.
16424 @subsubsection @code{EIND} and Devices with More Than 128 Ki Bytes of Flash
16425 @cindex @code{EIND}
16426 Pointers in the implementation are 16@tie{}bits wide.
16427 The address of a function or label is represented as word address so
16428 that indirect jumps and calls can target any code address in the
16429 range of 64@tie{}Ki words.
16431 In order to facilitate indirect jump on devices with more than 128@tie{}Ki
16432 bytes of program memory space, there is a special function register called
16433 @code{EIND} that serves as most significant part of the target address
16434 when @code{EICALL} or @code{EIJMP} instructions are used.
16436 Indirect jumps and calls on these devices are handled as follows by
16437 the compiler and are subject to some limitations:
16442 The compiler never sets @code{EIND}.
16445 The compiler uses @code{EIND} implicitly in @code{EICALL}/@code{EIJMP}
16446 instructions or might read @code{EIND} directly in order to emulate an
16447 indirect call/jump by means of a @code{RET} instruction.
16450 The compiler assumes that @code{EIND} never changes during the startup
16451 code or during the application. In particular, @code{EIND} is not
16452 saved/restored in function or interrupt service routine
16456 For indirect calls to functions and computed goto, the linker
16457 generates @emph{stubs}. Stubs are jump pads sometimes also called
16458 @emph{trampolines}. Thus, the indirect call/jump jumps to such a stub.
16459 The stub contains a direct jump to the desired address.
16462 Linker relaxation must be turned on so that the linker generates
16463 the stubs correctly in all situations. See the compiler option
16464 @option{-mrelax} and the linker option @option{--relax}.
16465 There are corner cases where the linker is supposed to generate stubs
16466 but aborts without relaxation and without a helpful error message.
16469 The default linker script is arranged for code with @code{EIND = 0}.
16470 If code is supposed to work for a setup with @code{EIND != 0}, a custom
16471 linker script has to be used in order to place the sections whose
16472 name start with @code{.trampolines} into the segment where @code{EIND}
16476 The startup code from libgcc never sets @code{EIND}.
16477 Notice that startup code is a blend of code from libgcc and AVR-LibC.
16478 For the impact of AVR-LibC on @code{EIND}, see the
16479 @w{@uref{http://nongnu.org/avr-libc/user-manual/,AVR-LibC user manual}}.
16482 It is legitimate for user-specific startup code to set up @code{EIND}
16483 early, for example by means of initialization code located in
16484 section @code{.init3}. Such code runs prior to general startup code
16485 that initializes RAM and calls constructors, but after the bit
16486 of startup code from AVR-LibC that sets @code{EIND} to the segment
16487 where the vector table is located.
16489 #include <avr/io.h>
16492 __attribute__((section(".init3"),naked,used,no_instrument_function))
16493 init3_set_eind (void)
16495 __asm volatile ("ldi r24,pm_hh8(__trampolines_start)\n\t"
16496 "out %i0,r24" :: "n" (&EIND) : "r24","memory");
16501 The @code{__trampolines_start} symbol is defined in the linker script.
16504 Stubs are generated automatically by the linker if
16505 the following two conditions are met:
16508 @item The address of a label is taken by means of the @code{gs} modifier
16509 (short for @emph{generate stubs}) like so:
16511 LDI r24, lo8(gs(@var{func}))
16512 LDI r25, hi8(gs(@var{func}))
16514 @item The final location of that label is in a code segment
16515 @emph{outside} the segment where the stubs are located.
16519 The compiler emits such @code{gs} modifiers for code labels in the
16520 following situations:
16522 @item Taking address of a function or code label.
16523 @item Computed goto.
16524 @item If prologue-save function is used, see @option{-mcall-prologues}
16525 command-line option.
16526 @item Switch/case dispatch tables. If you do not want such dispatch
16527 tables you can specify the @option{-fno-jump-tables} command-line option.
16528 @item C and C++ constructors/destructors called during startup/shutdown.
16529 @item If the tools hit a @code{gs()} modifier explained above.
16533 Jumping to non-symbolic addresses like so is @emph{not} supported:
16538 /* Call function at word address 0x2 */
16539 return ((int(*)(void)) 0x2)();
16543 Instead, a stub has to be set up, i.e.@: the function has to be called
16544 through a symbol (@code{func_4} in the example):
16549 extern int func_4 (void);
16551 /* Call function at byte address 0x4 */
16556 and the application be linked with @option{-Wl,--defsym,func_4=0x4}.
16557 Alternatively, @code{func_4} can be defined in the linker script.
16560 @subsubsection Handling of the @code{RAMPD}, @code{RAMPX}, @code{RAMPY} and @code{RAMPZ} Special Function Registers
16561 @cindex @code{RAMPD}
16562 @cindex @code{RAMPX}
16563 @cindex @code{RAMPY}
16564 @cindex @code{RAMPZ}
16565 Some AVR devices support memories larger than the 64@tie{}KiB range
16566 that can be accessed with 16-bit pointers. To access memory locations
16567 outside this 64@tie{}KiB range, the content of a @code{RAMP}
16568 register is used as high part of the address:
16569 The @code{X}, @code{Y}, @code{Z} address register is concatenated
16570 with the @code{RAMPX}, @code{RAMPY}, @code{RAMPZ} special function
16571 register, respectively, to get a wide address. Similarly,
16572 @code{RAMPD} is used together with direct addressing.
16576 The startup code initializes the @code{RAMP} special function
16577 registers with zero.
16580 If a @ref{AVR Named Address Spaces,named address space} other than
16581 generic or @code{__flash} is used, then @code{RAMPZ} is set
16582 as needed before the operation.
16585 If the device supports RAM larger than 64@tie{}KiB and the compiler
16586 needs to change @code{RAMPZ} to accomplish an operation, @code{RAMPZ}
16587 is reset to zero after the operation.
16590 If the device comes with a specific @code{RAMP} register, the ISR
16591 prologue/epilogue saves/restores that SFR and initializes it with
16592 zero in case the ISR code might (implicitly) use it.
16595 RAM larger than 64@tie{}KiB is not supported by GCC for AVR targets.
16596 If you use inline assembler to read from locations outside the
16597 16-bit address range and change one of the @code{RAMP} registers,
16598 you must reset it to zero after the access.
16602 @subsubsection AVR Built-in Macros
16604 GCC defines several built-in macros so that the user code can test
16605 for the presence or absence of features. Almost any of the following
16606 built-in macros are deduced from device capabilities and thus
16607 triggered by the @option{-mmcu=} command-line option.
16609 For even more AVR-specific built-in macros see
16610 @ref{AVR Named Address Spaces} and @ref{AVR Built-in Functions}.
16615 Build-in macro that resolves to a decimal number that identifies the
16616 architecture and depends on the @option{-mmcu=@var{mcu}} option.
16617 Possible values are:
16619 @code{2}, @code{25}, @code{3}, @code{31}, @code{35},
16620 @code{4}, @code{5}, @code{51}, @code{6}
16622 for @var{mcu}=@code{avr2}, @code{avr25}, @code{avr3}, @code{avr31},
16623 @code{avr35}, @code{avr4}, @code{avr5}, @code{avr51}, @code{avr6},
16628 @code{102}, @code{103}, @code{104},
16629 @code{105}, @code{106}, @code{107}
16631 for @var{mcu}=@code{avrtiny},
16632 @code{avrxmega2}, @code{avrxmega3}, @code{avrxmega4},
16633 @code{avrxmega5}, @code{avrxmega6}, @code{avrxmega7}, respectively.
16634 If @var{mcu} specifies a device, this built-in macro is set
16635 accordingly. For example, with @option{-mmcu=atmega8} the macro is
16636 defined to @code{4}.
16638 @item __AVR_@var{Device}__
16639 Setting @option{-mmcu=@var{device}} defines this built-in macro which reflects
16640 the device's name. For example, @option{-mmcu=atmega8} defines the
16641 built-in macro @code{__AVR_ATmega8__}, @option{-mmcu=attiny261a} defines
16642 @code{__AVR_ATtiny261A__}, etc.
16644 The built-in macros' names follow
16645 the scheme @code{__AVR_@var{Device}__} where @var{Device} is
16646 the device name as from the AVR user manual. The difference between
16647 @var{Device} in the built-in macro and @var{device} in
16648 @option{-mmcu=@var{device}} is that the latter is always lowercase.
16650 If @var{device} is not a device but only a core architecture like
16651 @samp{avr51}, this macro is not defined.
16653 @item __AVR_DEVICE_NAME__
16654 Setting @option{-mmcu=@var{device}} defines this built-in macro to
16655 the device's name. For example, with @option{-mmcu=atmega8} the macro
16656 is defined to @code{atmega8}.
16658 If @var{device} is not a device but only a core architecture like
16659 @samp{avr51}, this macro is not defined.
16661 @item __AVR_XMEGA__
16662 The device / architecture belongs to the XMEGA family of devices.
16664 @item __AVR_HAVE_ELPM__
16665 The device has the @code{ELPM} instruction.
16667 @item __AVR_HAVE_ELPMX__
16668 The device has the @code{ELPM R@var{n},Z} and @code{ELPM
16669 R@var{n},Z+} instructions.
16671 @item __AVR_HAVE_MOVW__
16672 The device has the @code{MOVW} instruction to perform 16-bit
16673 register-register moves.
16675 @item __AVR_HAVE_LPMX__
16676 The device has the @code{LPM R@var{n},Z} and
16677 @code{LPM R@var{n},Z+} instructions.
16679 @item __AVR_HAVE_MUL__
16680 The device has a hardware multiplier.
16682 @item __AVR_HAVE_JMP_CALL__
16683 The device has the @code{JMP} and @code{CALL} instructions.
16684 This is the case for devices with more than 8@tie{}KiB of program
16687 @item __AVR_HAVE_EIJMP_EICALL__
16688 @itemx __AVR_3_BYTE_PC__
16689 The device has the @code{EIJMP} and @code{EICALL} instructions.
16690 This is the case for devices with more than 128@tie{}KiB of program memory.
16691 This also means that the program counter
16692 (PC) is 3@tie{}bytes wide.
16694 @item __AVR_2_BYTE_PC__
16695 The program counter (PC) is 2@tie{}bytes wide. This is the case for devices
16696 with up to 128@tie{}KiB of program memory.
16698 @item __AVR_HAVE_8BIT_SP__
16699 @itemx __AVR_HAVE_16BIT_SP__
16700 The stack pointer (SP) register is treated as 8-bit respectively
16701 16-bit register by the compiler.
16702 The definition of these macros is affected by @option{-mtiny-stack}.
16704 @item __AVR_HAVE_SPH__
16706 The device has the SPH (high part of stack pointer) special function
16707 register or has an 8-bit stack pointer, respectively.
16708 The definition of these macros is affected by @option{-mmcu=} and
16709 in the cases of @option{-mmcu=avr2} and @option{-mmcu=avr25} also
16712 @item __AVR_HAVE_RAMPD__
16713 @itemx __AVR_HAVE_RAMPX__
16714 @itemx __AVR_HAVE_RAMPY__
16715 @itemx __AVR_HAVE_RAMPZ__
16716 The device has the @code{RAMPD}, @code{RAMPX}, @code{RAMPY},
16717 @code{RAMPZ} special function register, respectively.
16719 @item __NO_INTERRUPTS__
16720 This macro reflects the @option{-mno-interrupts} command-line option.
16722 @item __AVR_ERRATA_SKIP__
16723 @itemx __AVR_ERRATA_SKIP_JMP_CALL__
16724 Some AVR devices (AT90S8515, ATmega103) must not skip 32-bit
16725 instructions because of a hardware erratum. Skip instructions are
16726 @code{SBRS}, @code{SBRC}, @code{SBIS}, @code{SBIC} and @code{CPSE}.
16727 The second macro is only defined if @code{__AVR_HAVE_JMP_CALL__} is also
16730 @item __AVR_ISA_RMW__
16731 The device has Read-Modify-Write instructions (XCH, LAC, LAS and LAT).
16733 @item __AVR_SFR_OFFSET__=@var{offset}
16734 Instructions that can address I/O special function registers directly
16735 like @code{IN}, @code{OUT}, @code{SBI}, etc.@: may use a different
16736 address as if addressed by an instruction to access RAM like @code{LD}
16737 or @code{STS}. This offset depends on the device architecture and has
16738 to be subtracted from the RAM address in order to get the
16739 respective I/O@tie{}address.
16741 @item __AVR_SHORT_CALLS__
16742 The @option{-mshort-calls} command line option is set.
16744 @item __AVR_PM_BASE_ADDRESS__=@var{addr}
16745 Some devices support reading from flash memory by means of @code{LD*}
16746 instructions. The flash memory is seen in the data address space
16747 at an offset of @code{__AVR_PM_BASE_ADDRESS__}. If this macro
16748 is not defined, this feature is not available. If defined,
16749 the address space is linear and there is no need to put
16750 @code{.rodata} into RAM. This is handled by the default linker
16751 description file, and is currently available for
16752 @code{avrtiny} and @code{avrxmega3}. Even more convenient,
16753 there is no need to use address spaces like @code{__flash} or
16754 features like attribute @code{progmem} and @code{pgm_read_*}.
16756 @item __WITH_AVRLIBC__
16757 The compiler is configured to be used together with AVR-Libc.
16758 See the @option{--with-avrlibc} configure option.
16762 @node Blackfin Options
16763 @subsection Blackfin Options
16764 @cindex Blackfin Options
16767 @item -mcpu=@var{cpu}@r{[}-@var{sirevision}@r{]}
16769 Specifies the name of the target Blackfin processor. Currently, @var{cpu}
16770 can be one of @samp{bf512}, @samp{bf514}, @samp{bf516}, @samp{bf518},
16771 @samp{bf522}, @samp{bf523}, @samp{bf524}, @samp{bf525}, @samp{bf526},
16772 @samp{bf527}, @samp{bf531}, @samp{bf532}, @samp{bf533},
16773 @samp{bf534}, @samp{bf536}, @samp{bf537}, @samp{bf538}, @samp{bf539},
16774 @samp{bf542}, @samp{bf544}, @samp{bf547}, @samp{bf548}, @samp{bf549},
16775 @samp{bf542m}, @samp{bf544m}, @samp{bf547m}, @samp{bf548m}, @samp{bf549m},
16776 @samp{bf561}, @samp{bf592}.
16778 The optional @var{sirevision} specifies the silicon revision of the target
16779 Blackfin processor. Any workarounds available for the targeted silicon revision
16780 are enabled. If @var{sirevision} is @samp{none}, no workarounds are enabled.
16781 If @var{sirevision} is @samp{any}, all workarounds for the targeted processor
16782 are enabled. The @code{__SILICON_REVISION__} macro is defined to two
16783 hexadecimal digits representing the major and minor numbers in the silicon
16784 revision. If @var{sirevision} is @samp{none}, the @code{__SILICON_REVISION__}
16785 is not defined. If @var{sirevision} is @samp{any}, the
16786 @code{__SILICON_REVISION__} is defined to be @code{0xffff}.
16787 If this optional @var{sirevision} is not used, GCC assumes the latest known
16788 silicon revision of the targeted Blackfin processor.
16790 GCC defines a preprocessor macro for the specified @var{cpu}.
16791 For the @samp{bfin-elf} toolchain, this option causes the hardware BSP
16792 provided by libgloss to be linked in if @option{-msim} is not given.
16794 Without this option, @samp{bf532} is used as the processor by default.
16796 Note that support for @samp{bf561} is incomplete. For @samp{bf561},
16797 only the preprocessor macro is defined.
16801 Specifies that the program will be run on the simulator. This causes
16802 the simulator BSP provided by libgloss to be linked in. This option
16803 has effect only for @samp{bfin-elf} toolchain.
16804 Certain other options, such as @option{-mid-shared-library} and
16805 @option{-mfdpic}, imply @option{-msim}.
16807 @item -momit-leaf-frame-pointer
16808 @opindex momit-leaf-frame-pointer
16809 Don't keep the frame pointer in a register for leaf functions. This
16810 avoids the instructions to save, set up and restore frame pointers and
16811 makes an extra register available in leaf functions.
16813 @item -mspecld-anomaly
16814 @opindex mspecld-anomaly
16815 When enabled, the compiler ensures that the generated code does not
16816 contain speculative loads after jump instructions. If this option is used,
16817 @code{__WORKAROUND_SPECULATIVE_LOADS} is defined.
16819 @item -mno-specld-anomaly
16820 @opindex mno-specld-anomaly
16821 Don't generate extra code to prevent speculative loads from occurring.
16823 @item -mcsync-anomaly
16824 @opindex mcsync-anomaly
16825 When enabled, the compiler ensures that the generated code does not
16826 contain CSYNC or SSYNC instructions too soon after conditional branches.
16827 If this option is used, @code{__WORKAROUND_SPECULATIVE_SYNCS} is defined.
16829 @item -mno-csync-anomaly
16830 @opindex mno-csync-anomaly
16831 Don't generate extra code to prevent CSYNC or SSYNC instructions from
16832 occurring too soon after a conditional branch.
16836 When enabled, the compiler is free to take advantage of the knowledge that
16837 the entire program fits into the low 64k of memory.
16840 @opindex mno-low-64k
16841 Assume that the program is arbitrarily large. This is the default.
16843 @item -mstack-check-l1
16844 @opindex mstack-check-l1
16845 Do stack checking using information placed into L1 scratchpad memory by the
16848 @item -mid-shared-library
16849 @opindex mid-shared-library
16850 Generate code that supports shared libraries via the library ID method.
16851 This allows for execute in place and shared libraries in an environment
16852 without virtual memory management. This option implies @option{-fPIC}.
16853 With a @samp{bfin-elf} target, this option implies @option{-msim}.
16855 @item -mno-id-shared-library
16856 @opindex mno-id-shared-library
16857 Generate code that doesn't assume ID-based shared libraries are being used.
16858 This is the default.
16860 @item -mleaf-id-shared-library
16861 @opindex mleaf-id-shared-library
16862 Generate code that supports shared libraries via the library ID method,
16863 but assumes that this library or executable won't link against any other
16864 ID shared libraries. That allows the compiler to use faster code for jumps
16867 @item -mno-leaf-id-shared-library
16868 @opindex mno-leaf-id-shared-library
16869 Do not assume that the code being compiled won't link against any ID shared
16870 libraries. Slower code is generated for jump and call insns.
16872 @item -mshared-library-id=n
16873 @opindex mshared-library-id
16874 Specifies the identification number of the ID-based shared library being
16875 compiled. Specifying a value of 0 generates more compact code; specifying
16876 other values forces the allocation of that number to the current
16877 library but is no more space- or time-efficient than omitting this option.
16881 Generate code that allows the data segment to be located in a different
16882 area of memory from the text segment. This allows for execute in place in
16883 an environment without virtual memory management by eliminating relocations
16884 against the text section.
16886 @item -mno-sep-data
16887 @opindex mno-sep-data
16888 Generate code that assumes that the data segment follows the text segment.
16889 This is the default.
16892 @itemx -mno-long-calls
16893 @opindex mlong-calls
16894 @opindex mno-long-calls
16895 Tells the compiler to perform function calls by first loading the
16896 address of the function into a register and then performing a subroutine
16897 call on this register. This switch is needed if the target function
16898 lies outside of the 24-bit addressing range of the offset-based
16899 version of subroutine call instruction.
16901 This feature is not enabled by default. Specifying
16902 @option{-mno-long-calls} restores the default behavior. Note these
16903 switches have no effect on how the compiler generates code to handle
16904 function calls via function pointers.
16908 Link with the fast floating-point library. This library relaxes some of
16909 the IEEE floating-point standard's rules for checking inputs against
16910 Not-a-Number (NAN), in the interest of performance.
16913 @opindex minline-plt
16914 Enable inlining of PLT entries in function calls to functions that are
16915 not known to bind locally. It has no effect without @option{-mfdpic}.
16918 @opindex mmulticore
16919 Build a standalone application for multicore Blackfin processors.
16920 This option causes proper start files and link scripts supporting
16921 multicore to be used, and defines the macro @code{__BFIN_MULTICORE}.
16922 It can only be used with @option{-mcpu=bf561@r{[}-@var{sirevision}@r{]}}.
16924 This option can be used with @option{-mcorea} or @option{-mcoreb}, which
16925 selects the one-application-per-core programming model. Without
16926 @option{-mcorea} or @option{-mcoreb}, the single-application/dual-core
16927 programming model is used. In this model, the main function of Core B
16928 should be named as @code{coreb_main}.
16930 If this option is not used, the single-core application programming
16935 Build a standalone application for Core A of BF561 when using
16936 the one-application-per-core programming model. Proper start files
16937 and link scripts are used to support Core A, and the macro
16938 @code{__BFIN_COREA} is defined.
16939 This option can only be used in conjunction with @option{-mmulticore}.
16943 Build a standalone application for Core B of BF561 when using
16944 the one-application-per-core programming model. Proper start files
16945 and link scripts are used to support Core B, and the macro
16946 @code{__BFIN_COREB} is defined. When this option is used, @code{coreb_main}
16947 should be used instead of @code{main}.
16948 This option can only be used in conjunction with @option{-mmulticore}.
16952 Build a standalone application for SDRAM. Proper start files and
16953 link scripts are used to put the application into SDRAM, and the macro
16954 @code{__BFIN_SDRAM} is defined.
16955 The loader should initialize SDRAM before loading the application.
16959 Assume that ICPLBs are enabled at run time. This has an effect on certain
16960 anomaly workarounds. For Linux targets, the default is to assume ICPLBs
16961 are enabled; for standalone applications the default is off.
16965 @subsection C6X Options
16966 @cindex C6X Options
16969 @item -march=@var{name}
16971 This specifies the name of the target architecture. GCC uses this
16972 name to determine what kind of instructions it can emit when generating
16973 assembly code. Permissible names are: @samp{c62x},
16974 @samp{c64x}, @samp{c64x+}, @samp{c67x}, @samp{c67x+}, @samp{c674x}.
16977 @opindex mbig-endian
16978 Generate code for a big-endian target.
16980 @item -mlittle-endian
16981 @opindex mlittle-endian
16982 Generate code for a little-endian target. This is the default.
16986 Choose startup files and linker script suitable for the simulator.
16988 @item -msdata=default
16989 @opindex msdata=default
16990 Put small global and static data in the @code{.neardata} section,
16991 which is pointed to by register @code{B14}. Put small uninitialized
16992 global and static data in the @code{.bss} section, which is adjacent
16993 to the @code{.neardata} section. Put small read-only data into the
16994 @code{.rodata} section. The corresponding sections used for large
16995 pieces of data are @code{.fardata}, @code{.far} and @code{.const}.
16998 @opindex msdata=all
16999 Put all data, not just small objects, into the sections reserved for
17000 small data, and use addressing relative to the @code{B14} register to
17004 @opindex msdata=none
17005 Make no use of the sections reserved for small data, and use absolute
17006 addresses to access all data. Put all initialized global and static
17007 data in the @code{.fardata} section, and all uninitialized data in the
17008 @code{.far} section. Put all constant data into the @code{.const}
17013 @subsection CRIS Options
17014 @cindex CRIS Options
17016 These options are defined specifically for the CRIS ports.
17019 @item -march=@var{architecture-type}
17020 @itemx -mcpu=@var{architecture-type}
17023 Generate code for the specified architecture. The choices for
17024 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
17025 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX@.
17026 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
17029 @item -mtune=@var{architecture-type}
17031 Tune to @var{architecture-type} everything applicable about the generated
17032 code, except for the ABI and the set of available instructions. The
17033 choices for @var{architecture-type} are the same as for
17034 @option{-march=@var{architecture-type}}.
17036 @item -mmax-stack-frame=@var{n}
17037 @opindex mmax-stack-frame
17038 Warn when the stack frame of a function exceeds @var{n} bytes.
17044 The options @option{-metrax4} and @option{-metrax100} are synonyms for
17045 @option{-march=v3} and @option{-march=v8} respectively.
17047 @item -mmul-bug-workaround
17048 @itemx -mno-mul-bug-workaround
17049 @opindex mmul-bug-workaround
17050 @opindex mno-mul-bug-workaround
17051 Work around a bug in the @code{muls} and @code{mulu} instructions for CPU
17052 models where it applies. This option is active by default.
17056 Enable CRIS-specific verbose debug-related information in the assembly
17057 code. This option also has the effect of turning off the @samp{#NO_APP}
17058 formatted-code indicator to the assembler at the beginning of the
17063 Do not use condition-code results from previous instruction; always emit
17064 compare and test instructions before use of condition codes.
17066 @item -mno-side-effects
17067 @opindex mno-side-effects
17068 Do not emit instructions with side effects in addressing modes other than
17071 @item -mstack-align
17072 @itemx -mno-stack-align
17073 @itemx -mdata-align
17074 @itemx -mno-data-align
17075 @itemx -mconst-align
17076 @itemx -mno-const-align
17077 @opindex mstack-align
17078 @opindex mno-stack-align
17079 @opindex mdata-align
17080 @opindex mno-data-align
17081 @opindex mconst-align
17082 @opindex mno-const-align
17083 These options (@samp{no-} options) arrange (eliminate arrangements) for the
17084 stack frame, individual data and constants to be aligned for the maximum
17085 single data access size for the chosen CPU model. The default is to
17086 arrange for 32-bit alignment. ABI details such as structure layout are
17087 not affected by these options.
17095 Similar to the stack- data- and const-align options above, these options
17096 arrange for stack frame, writable data and constants to all be 32-bit,
17097 16-bit or 8-bit aligned. The default is 32-bit alignment.
17099 @item -mno-prologue-epilogue
17100 @itemx -mprologue-epilogue
17101 @opindex mno-prologue-epilogue
17102 @opindex mprologue-epilogue
17103 With @option{-mno-prologue-epilogue}, the normal function prologue and
17104 epilogue which set up the stack frame are omitted and no return
17105 instructions or return sequences are generated in the code. Use this
17106 option only together with visual inspection of the compiled code: no
17107 warnings or errors are generated when call-saved registers must be saved,
17108 or storage for local variables needs to be allocated.
17112 @opindex mno-gotplt
17114 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
17115 instruction sequences that load addresses for functions from the PLT part
17116 of the GOT rather than (traditional on other architectures) calls to the
17117 PLT@. The default is @option{-mgotplt}.
17121 Legacy no-op option only recognized with the cris-axis-elf and
17122 cris-axis-linux-gnu targets.
17126 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
17130 This option, recognized for the cris-axis-elf, arranges
17131 to link with input-output functions from a simulator library. Code,
17132 initialized data and zero-initialized data are allocated consecutively.
17136 Like @option{-sim}, but pass linker options to locate initialized data at
17137 0x40000000 and zero-initialized data at 0x80000000.
17141 @subsection CR16 Options
17142 @cindex CR16 Options
17144 These options are defined specifically for the CR16 ports.
17150 Enable the use of multiply-accumulate instructions. Disabled by default.
17154 @opindex mcr16cplus
17156 Generate code for CR16C or CR16C+ architecture. CR16C+ architecture
17161 Links the library libsim.a which is in compatible with simulator. Applicable
17162 to ELF compiler only.
17166 Choose integer type as 32-bit wide.
17170 Generates @code{sbit}/@code{cbit} instructions for bit manipulations.
17172 @item -mdata-model=@var{model}
17173 @opindex mdata-model
17174 Choose a data model. The choices for @var{model} are @samp{near},
17175 @samp{far} or @samp{medium}. @samp{medium} is default.
17176 However, @samp{far} is not valid with @option{-mcr16c}, as the
17177 CR16C architecture does not support the far data model.
17180 @node Darwin Options
17181 @subsection Darwin Options
17182 @cindex Darwin options
17184 These options are defined for all architectures running the Darwin operating
17187 FSF GCC on Darwin does not create ``fat'' object files; it creates
17188 an object file for the single architecture that GCC was built to
17189 target. Apple's GCC on Darwin does create ``fat'' files if multiple
17190 @option{-arch} options are used; it does so by running the compiler or
17191 linker multiple times and joining the results together with
17194 The subtype of the file created (like @samp{ppc7400} or @samp{ppc970} or
17195 @samp{i686}) is determined by the flags that specify the ISA
17196 that GCC is targeting, like @option{-mcpu} or @option{-march}. The
17197 @option{-force_cpusubtype_ALL} option can be used to override this.
17199 The Darwin tools vary in their behavior when presented with an ISA
17200 mismatch. The assembler, @file{as}, only permits instructions to
17201 be used that are valid for the subtype of the file it is generating,
17202 so you cannot put 64-bit instructions in a @samp{ppc750} object file.
17203 The linker for shared libraries, @file{/usr/bin/libtool}, fails
17204 and prints an error if asked to create a shared library with a less
17205 restrictive subtype than its input files (for instance, trying to put
17206 a @samp{ppc970} object file in a @samp{ppc7400} library). The linker
17207 for executables, @command{ld}, quietly gives the executable the most
17208 restrictive subtype of any of its input files.
17213 Add the framework directory @var{dir} to the head of the list of
17214 directories to be searched for header files. These directories are
17215 interleaved with those specified by @option{-I} options and are
17216 scanned in a left-to-right order.
17218 A framework directory is a directory with frameworks in it. A
17219 framework is a directory with a @file{Headers} and/or
17220 @file{PrivateHeaders} directory contained directly in it that ends
17221 in @file{.framework}. The name of a framework is the name of this
17222 directory excluding the @file{.framework}. Headers associated with
17223 the framework are found in one of those two directories, with
17224 @file{Headers} being searched first. A subframework is a framework
17225 directory that is in a framework's @file{Frameworks} directory.
17226 Includes of subframework headers can only appear in a header of a
17227 framework that contains the subframework, or in a sibling subframework
17228 header. Two subframeworks are siblings if they occur in the same
17229 framework. A subframework should not have the same name as a
17230 framework; a warning is issued if this is violated. Currently a
17231 subframework cannot have subframeworks; in the future, the mechanism
17232 may be extended to support this. The standard frameworks can be found
17233 in @file{/System/Library/Frameworks} and
17234 @file{/Library/Frameworks}. An example include looks like
17235 @code{#include <Framework/header.h>}, where @file{Framework} denotes
17236 the name of the framework and @file{header.h} is found in the
17237 @file{PrivateHeaders} or @file{Headers} directory.
17239 @item -iframework@var{dir}
17240 @opindex iframework
17241 Like @option{-F} except the directory is a treated as a system
17242 directory. The main difference between this @option{-iframework} and
17243 @option{-F} is that with @option{-iframework} the compiler does not
17244 warn about constructs contained within header files found via
17245 @var{dir}. This option is valid only for the C family of languages.
17249 Emit debugging information for symbols that are used. For stabs
17250 debugging format, this enables @option{-feliminate-unused-debug-symbols}.
17251 This is by default ON@.
17255 Emit debugging information for all symbols and types.
17257 @item -mmacosx-version-min=@var{version}
17258 The earliest version of MacOS X that this executable will run on
17259 is @var{version}. Typical values of @var{version} include @code{10.1},
17260 @code{10.2}, and @code{10.3.9}.
17262 If the compiler was built to use the system's headers by default,
17263 then the default for this option is the system version on which the
17264 compiler is running, otherwise the default is to make choices that
17265 are compatible with as many systems and code bases as possible.
17269 Enable kernel development mode. The @option{-mkernel} option sets
17270 @option{-static}, @option{-fno-common}, @option{-fno-use-cxa-atexit},
17271 @option{-fno-exceptions}, @option{-fno-non-call-exceptions},
17272 @option{-fapple-kext}, @option{-fno-weak} and @option{-fno-rtti} where
17273 applicable. This mode also sets @option{-mno-altivec},
17274 @option{-msoft-float}, @option{-fno-builtin} and
17275 @option{-mlong-branch} for PowerPC targets.
17277 @item -mone-byte-bool
17278 @opindex mone-byte-bool
17279 Override the defaults for @code{bool} so that @code{sizeof(bool)==1}.
17280 By default @code{sizeof(bool)} is @code{4} when compiling for
17281 Darwin/PowerPC and @code{1} when compiling for Darwin/x86, so this
17282 option has no effect on x86.
17284 @strong{Warning:} The @option{-mone-byte-bool} switch causes GCC
17285 to generate code that is not binary compatible with code generated
17286 without that switch. Using this switch may require recompiling all
17287 other modules in a program, including system libraries. Use this
17288 switch to conform to a non-default data model.
17290 @item -mfix-and-continue
17291 @itemx -ffix-and-continue
17292 @itemx -findirect-data
17293 @opindex mfix-and-continue
17294 @opindex ffix-and-continue
17295 @opindex findirect-data
17296 Generate code suitable for fast turnaround development, such as to
17297 allow GDB to dynamically load @file{.o} files into already-running
17298 programs. @option{-findirect-data} and @option{-ffix-and-continue}
17299 are provided for backwards compatibility.
17303 Loads all members of static archive libraries.
17304 See man ld(1) for more information.
17306 @item -arch_errors_fatal
17307 @opindex arch_errors_fatal
17308 Cause the errors having to do with files that have the wrong architecture
17311 @item -bind_at_load
17312 @opindex bind_at_load
17313 Causes the output file to be marked such that the dynamic linker will
17314 bind all undefined references when the file is loaded or launched.
17318 Produce a Mach-o bundle format file.
17319 See man ld(1) for more information.
17321 @item -bundle_loader @var{executable}
17322 @opindex bundle_loader
17323 This option specifies the @var{executable} that will load the build
17324 output file being linked. See man ld(1) for more information.
17327 @opindex dynamiclib
17328 When passed this option, GCC produces a dynamic library instead of
17329 an executable when linking, using the Darwin @file{libtool} command.
17331 @item -force_cpusubtype_ALL
17332 @opindex force_cpusubtype_ALL
17333 This causes GCC's output file to have the @samp{ALL} subtype, instead of
17334 one controlled by the @option{-mcpu} or @option{-march} option.
17336 @item -allowable_client @var{client_name}
17337 @itemx -client_name
17338 @itemx -compatibility_version
17339 @itemx -current_version
17341 @itemx -dependency-file
17343 @itemx -dylinker_install_name
17345 @itemx -exported_symbols_list
17348 @itemx -flat_namespace
17349 @itemx -force_flat_namespace
17350 @itemx -headerpad_max_install_names
17353 @itemx -install_name
17354 @itemx -keep_private_externs
17355 @itemx -multi_module
17356 @itemx -multiply_defined
17357 @itemx -multiply_defined_unused
17360 @itemx -no_dead_strip_inits_and_terms
17361 @itemx -nofixprebinding
17362 @itemx -nomultidefs
17364 @itemx -noseglinkedit
17365 @itemx -pagezero_size
17367 @itemx -prebind_all_twolevel_modules
17368 @itemx -private_bundle
17370 @itemx -read_only_relocs
17372 @itemx -sectobjectsymbols
17376 @itemx -sectobjectsymbols
17379 @itemx -segs_read_only_addr
17381 @itemx -segs_read_write_addr
17382 @itemx -seg_addr_table
17383 @itemx -seg_addr_table_filename
17384 @itemx -seglinkedit
17386 @itemx -segs_read_only_addr
17387 @itemx -segs_read_write_addr
17388 @itemx -single_module
17390 @itemx -sub_library
17392 @itemx -sub_umbrella
17393 @itemx -twolevel_namespace
17396 @itemx -unexported_symbols_list
17397 @itemx -weak_reference_mismatches
17398 @itemx -whatsloaded
17399 @opindex allowable_client
17400 @opindex client_name
17401 @opindex compatibility_version
17402 @opindex current_version
17403 @opindex dead_strip
17404 @opindex dependency-file
17405 @opindex dylib_file
17406 @opindex dylinker_install_name
17408 @opindex exported_symbols_list
17410 @opindex flat_namespace
17411 @opindex force_flat_namespace
17412 @opindex headerpad_max_install_names
17413 @opindex image_base
17415 @opindex install_name
17416 @opindex keep_private_externs
17417 @opindex multi_module
17418 @opindex multiply_defined
17419 @opindex multiply_defined_unused
17420 @opindex noall_load
17421 @opindex no_dead_strip_inits_and_terms
17422 @opindex nofixprebinding
17423 @opindex nomultidefs
17425 @opindex noseglinkedit
17426 @opindex pagezero_size
17428 @opindex prebind_all_twolevel_modules
17429 @opindex private_bundle
17430 @opindex read_only_relocs
17432 @opindex sectobjectsymbols
17435 @opindex sectcreate
17436 @opindex sectobjectsymbols
17439 @opindex segs_read_only_addr
17440 @opindex segs_read_write_addr
17441 @opindex seg_addr_table
17442 @opindex seg_addr_table_filename
17443 @opindex seglinkedit
17445 @opindex segs_read_only_addr
17446 @opindex segs_read_write_addr
17447 @opindex single_module
17449 @opindex sub_library
17450 @opindex sub_umbrella
17451 @opindex twolevel_namespace
17454 @opindex unexported_symbols_list
17455 @opindex weak_reference_mismatches
17456 @opindex whatsloaded
17457 These options are passed to the Darwin linker. The Darwin linker man page
17458 describes them in detail.
17461 @node DEC Alpha Options
17462 @subsection DEC Alpha Options
17464 These @samp{-m} options are defined for the DEC Alpha implementations:
17467 @item -mno-soft-float
17468 @itemx -msoft-float
17469 @opindex mno-soft-float
17470 @opindex msoft-float
17471 Use (do not use) the hardware floating-point instructions for
17472 floating-point operations. When @option{-msoft-float} is specified,
17473 functions in @file{libgcc.a} are used to perform floating-point
17474 operations. Unless they are replaced by routines that emulate the
17475 floating-point operations, or compiled in such a way as to call such
17476 emulations routines, these routines issue floating-point
17477 operations. If you are compiling for an Alpha without floating-point
17478 operations, you must ensure that the library is built so as not to call
17481 Note that Alpha implementations without floating-point operations are
17482 required to have floating-point registers.
17485 @itemx -mno-fp-regs
17487 @opindex mno-fp-regs
17488 Generate code that uses (does not use) the floating-point register set.
17489 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
17490 register set is not used, floating-point operands are passed in integer
17491 registers as if they were integers and floating-point results are passed
17492 in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence,
17493 so any function with a floating-point argument or return value called by code
17494 compiled with @option{-mno-fp-regs} must also be compiled with that
17497 A typical use of this option is building a kernel that does not use,
17498 and hence need not save and restore, any floating-point registers.
17502 The Alpha architecture implements floating-point hardware optimized for
17503 maximum performance. It is mostly compliant with the IEEE floating-point
17504 standard. However, for full compliance, software assistance is
17505 required. This option generates code fully IEEE-compliant code
17506 @emph{except} that the @var{inexact-flag} is not maintained (see below).
17507 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
17508 defined during compilation. The resulting code is less efficient but is
17509 able to correctly support denormalized numbers and exceptional IEEE
17510 values such as not-a-number and plus/minus infinity. Other Alpha
17511 compilers call this option @option{-ieee_with_no_inexact}.
17513 @item -mieee-with-inexact
17514 @opindex mieee-with-inexact
17515 This is like @option{-mieee} except the generated code also maintains
17516 the IEEE @var{inexact-flag}. Turning on this option causes the
17517 generated code to implement fully-compliant IEEE math. In addition to
17518 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
17519 macro. On some Alpha implementations the resulting code may execute
17520 significantly slower than the code generated by default. Since there is
17521 very little code that depends on the @var{inexact-flag}, you should
17522 normally not specify this option. Other Alpha compilers call this
17523 option @option{-ieee_with_inexact}.
17525 @item -mfp-trap-mode=@var{trap-mode}
17526 @opindex mfp-trap-mode
17527 This option controls what floating-point related traps are enabled.
17528 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
17529 The trap mode can be set to one of four values:
17533 This is the default (normal) setting. The only traps that are enabled
17534 are the ones that cannot be disabled in software (e.g., division by zero
17538 In addition to the traps enabled by @samp{n}, underflow traps are enabled
17542 Like @samp{u}, but the instructions are marked to be safe for software
17543 completion (see Alpha architecture manual for details).
17546 Like @samp{su}, but inexact traps are enabled as well.
17549 @item -mfp-rounding-mode=@var{rounding-mode}
17550 @opindex mfp-rounding-mode
17551 Selects the IEEE rounding mode. Other Alpha compilers call this option
17552 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
17557 Normal IEEE rounding mode. Floating-point numbers are rounded towards
17558 the nearest machine number or towards the even machine number in case
17562 Round towards minus infinity.
17565 Chopped rounding mode. Floating-point numbers are rounded towards zero.
17568 Dynamic rounding mode. A field in the floating-point control register
17569 (@var{fpcr}, see Alpha architecture reference manual) controls the
17570 rounding mode in effect. The C library initializes this register for
17571 rounding towards plus infinity. Thus, unless your program modifies the
17572 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
17575 @item -mtrap-precision=@var{trap-precision}
17576 @opindex mtrap-precision
17577 In the Alpha architecture, floating-point traps are imprecise. This
17578 means without software assistance it is impossible to recover from a
17579 floating trap and program execution normally needs to be terminated.
17580 GCC can generate code that can assist operating system trap handlers
17581 in determining the exact location that caused a floating-point trap.
17582 Depending on the requirements of an application, different levels of
17583 precisions can be selected:
17587 Program precision. This option is the default and means a trap handler
17588 can only identify which program caused a floating-point exception.
17591 Function precision. The trap handler can determine the function that
17592 caused a floating-point exception.
17595 Instruction precision. The trap handler can determine the exact
17596 instruction that caused a floating-point exception.
17599 Other Alpha compilers provide the equivalent options called
17600 @option{-scope_safe} and @option{-resumption_safe}.
17602 @item -mieee-conformant
17603 @opindex mieee-conformant
17604 This option marks the generated code as IEEE conformant. You must not
17605 use this option unless you also specify @option{-mtrap-precision=i} and either
17606 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
17607 is to emit the line @samp{.eflag 48} in the function prologue of the
17608 generated assembly file.
17610 @item -mbuild-constants
17611 @opindex mbuild-constants
17612 Normally GCC examines a 32- or 64-bit integer constant to
17613 see if it can construct it from smaller constants in two or three
17614 instructions. If it cannot, it outputs the constant as a literal and
17615 generates code to load it from the data segment at run time.
17617 Use this option to require GCC to construct @emph{all} integer constants
17618 using code, even if it takes more instructions (the maximum is six).
17620 You typically use this option to build a shared library dynamic
17621 loader. Itself a shared library, it must relocate itself in memory
17622 before it can find the variables and constants in its own data segment.
17640 Indicate whether GCC should generate code to use the optional BWX,
17641 CIX, FIX and MAX instruction sets. The default is to use the instruction
17642 sets supported by the CPU type specified via @option{-mcpu=} option or that
17643 of the CPU on which GCC was built if none is specified.
17646 @itemx -mfloat-ieee
17647 @opindex mfloat-vax
17648 @opindex mfloat-ieee
17649 Generate code that uses (does not use) VAX F and G floating-point
17650 arithmetic instead of IEEE single and double precision.
17652 @item -mexplicit-relocs
17653 @itemx -mno-explicit-relocs
17654 @opindex mexplicit-relocs
17655 @opindex mno-explicit-relocs
17656 Older Alpha assemblers provided no way to generate symbol relocations
17657 except via assembler macros. Use of these macros does not allow
17658 optimal instruction scheduling. GNU binutils as of version 2.12
17659 supports a new syntax that allows the compiler to explicitly mark
17660 which relocations should apply to which instructions. This option
17661 is mostly useful for debugging, as GCC detects the capabilities of
17662 the assembler when it is built and sets the default accordingly.
17665 @itemx -mlarge-data
17666 @opindex msmall-data
17667 @opindex mlarge-data
17668 When @option{-mexplicit-relocs} is in effect, static data is
17669 accessed via @dfn{gp-relative} relocations. When @option{-msmall-data}
17670 is used, objects 8 bytes long or smaller are placed in a @dfn{small data area}
17671 (the @code{.sdata} and @code{.sbss} sections) and are accessed via
17672 16-bit relocations off of the @code{$gp} register. This limits the
17673 size of the small data area to 64KB, but allows the variables to be
17674 directly accessed via a single instruction.
17676 The default is @option{-mlarge-data}. With this option the data area
17677 is limited to just below 2GB@. Programs that require more than 2GB of
17678 data must use @code{malloc} or @code{mmap} to allocate the data in the
17679 heap instead of in the program's data segment.
17681 When generating code for shared libraries, @option{-fpic} implies
17682 @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}.
17685 @itemx -mlarge-text
17686 @opindex msmall-text
17687 @opindex mlarge-text
17688 When @option{-msmall-text} is used, the compiler assumes that the
17689 code of the entire program (or shared library) fits in 4MB, and is
17690 thus reachable with a branch instruction. When @option{-msmall-data}
17691 is used, the compiler can assume that all local symbols share the
17692 same @code{$gp} value, and thus reduce the number of instructions
17693 required for a function call from 4 to 1.
17695 The default is @option{-mlarge-text}.
17697 @item -mcpu=@var{cpu_type}
17699 Set the instruction set and instruction scheduling parameters for
17700 machine type @var{cpu_type}. You can specify either the @samp{EV}
17701 style name or the corresponding chip number. GCC supports scheduling
17702 parameters for the EV4, EV5 and EV6 family of processors and
17703 chooses the default values for the instruction set from the processor
17704 you specify. If you do not specify a processor type, GCC defaults
17705 to the processor on which the compiler was built.
17707 Supported values for @var{cpu_type} are
17713 Schedules as an EV4 and has no instruction set extensions.
17717 Schedules as an EV5 and has no instruction set extensions.
17721 Schedules as an EV5 and supports the BWX extension.
17726 Schedules as an EV5 and supports the BWX and MAX extensions.
17730 Schedules as an EV6 and supports the BWX, FIX, and MAX extensions.
17734 Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions.
17737 Native toolchains also support the value @samp{native},
17738 which selects the best architecture option for the host processor.
17739 @option{-mcpu=native} has no effect if GCC does not recognize
17742 @item -mtune=@var{cpu_type}
17744 Set only the instruction scheduling parameters for machine type
17745 @var{cpu_type}. The instruction set is not changed.
17747 Native toolchains also support the value @samp{native},
17748 which selects the best architecture option for the host processor.
17749 @option{-mtune=native} has no effect if GCC does not recognize
17752 @item -mmemory-latency=@var{time}
17753 @opindex mmemory-latency
17754 Sets the latency the scheduler should assume for typical memory
17755 references as seen by the application. This number is highly
17756 dependent on the memory access patterns used by the application
17757 and the size of the external cache on the machine.
17759 Valid options for @var{time} are
17763 A decimal number representing clock cycles.
17769 The compiler contains estimates of the number of clock cycles for
17770 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
17771 (also called Dcache, Scache, and Bcache), as well as to main memory.
17772 Note that L3 is only valid for EV5.
17778 @subsection FR30 Options
17779 @cindex FR30 Options
17781 These options are defined specifically for the FR30 port.
17785 @item -msmall-model
17786 @opindex msmall-model
17787 Use the small address space model. This can produce smaller code, but
17788 it does assume that all symbolic values and addresses fit into a
17793 Assume that runtime support has been provided and so there is no need
17794 to include the simulator library (@file{libsim.a}) on the linker
17800 @subsection FT32 Options
17801 @cindex FT32 Options
17803 These options are defined specifically for the FT32 port.
17809 Specifies that the program will be run on the simulator. This causes
17810 an alternate runtime startup and library to be linked.
17811 You must not use this option when generating programs that will run on
17812 real hardware; you must provide your own runtime library for whatever
17813 I/O functions are needed.
17817 Enable Local Register Allocation. This is still experimental for FT32,
17818 so by default the compiler uses standard reload.
17822 Do not use div and mod instructions.
17826 Enable use of the extended instructions of the FT32B processor.
17830 Compress all code using the Ft32B code compression scheme.
17834 Do not generate code that reads program memory.
17839 @subsection FRV Options
17840 @cindex FRV Options
17846 Only use the first 32 general-purpose registers.
17851 Use all 64 general-purpose registers.
17856 Use only the first 32 floating-point registers.
17861 Use all 64 floating-point registers.
17864 @opindex mhard-float
17866 Use hardware instructions for floating-point operations.
17869 @opindex msoft-float
17871 Use library routines for floating-point operations.
17876 Dynamically allocate condition code registers.
17881 Do not try to dynamically allocate condition code registers, only
17882 use @code{icc0} and @code{fcc0}.
17887 Change ABI to use double word insns.
17892 Do not use double word instructions.
17897 Use floating-point double instructions.
17900 @opindex mno-double
17902 Do not use floating-point double instructions.
17907 Use media instructions.
17912 Do not use media instructions.
17917 Use multiply and add/subtract instructions.
17920 @opindex mno-muladd
17922 Do not use multiply and add/subtract instructions.
17927 Select the FDPIC ABI, which uses function descriptors to represent
17928 pointers to functions. Without any PIC/PIE-related options, it
17929 implies @option{-fPIE}. With @option{-fpic} or @option{-fpie}, it
17930 assumes GOT entries and small data are within a 12-bit range from the
17931 GOT base address; with @option{-fPIC} or @option{-fPIE}, GOT offsets
17932 are computed with 32 bits.
17933 With a @samp{bfin-elf} target, this option implies @option{-msim}.
17936 @opindex minline-plt
17938 Enable inlining of PLT entries in function calls to functions that are
17939 not known to bind locally. It has no effect without @option{-mfdpic}.
17940 It's enabled by default if optimizing for speed and compiling for
17941 shared libraries (i.e., @option{-fPIC} or @option{-fpic}), or when an
17942 optimization option such as @option{-O3} or above is present in the
17948 Assume a large TLS segment when generating thread-local code.
17953 Do not assume a large TLS segment when generating thread-local code.
17958 Enable the use of @code{GPREL} relocations in the FDPIC ABI for data
17959 that is known to be in read-only sections. It's enabled by default,
17960 except for @option{-fpic} or @option{-fpie}: even though it may help
17961 make the global offset table smaller, it trades 1 instruction for 4.
17962 With @option{-fPIC} or @option{-fPIE}, it trades 3 instructions for 4,
17963 one of which may be shared by multiple symbols, and it avoids the need
17964 for a GOT entry for the referenced symbol, so it's more likely to be a
17965 win. If it is not, @option{-mno-gprel-ro} can be used to disable it.
17967 @item -multilib-library-pic
17968 @opindex multilib-library-pic
17970 Link with the (library, not FD) pic libraries. It's implied by
17971 @option{-mlibrary-pic}, as well as by @option{-fPIC} and
17972 @option{-fpic} without @option{-mfdpic}. You should never have to use
17976 @opindex mlinked-fp
17978 Follow the EABI requirement of always creating a frame pointer whenever
17979 a stack frame is allocated. This option is enabled by default and can
17980 be disabled with @option{-mno-linked-fp}.
17983 @opindex mlong-calls
17985 Use indirect addressing to call functions outside the current
17986 compilation unit. This allows the functions to be placed anywhere
17987 within the 32-bit address space.
17989 @item -malign-labels
17990 @opindex malign-labels
17992 Try to align labels to an 8-byte boundary by inserting NOPs into the
17993 previous packet. This option only has an effect when VLIW packing
17994 is enabled. It doesn't create new packets; it merely adds NOPs to
17997 @item -mlibrary-pic
17998 @opindex mlibrary-pic
18000 Generate position-independent EABI code.
18005 Use only the first four media accumulator registers.
18010 Use all eight media accumulator registers.
18015 Pack VLIW instructions.
18020 Do not pack VLIW instructions.
18023 @opindex mno-eflags
18025 Do not mark ABI switches in e_flags.
18028 @opindex mcond-move
18030 Enable the use of conditional-move instructions (default).
18032 This switch is mainly for debugging the compiler and will likely be removed
18033 in a future version.
18035 @item -mno-cond-move
18036 @opindex mno-cond-move
18038 Disable the use of conditional-move instructions.
18040 This switch is mainly for debugging the compiler and will likely be removed
18041 in a future version.
18046 Enable the use of conditional set instructions (default).
18048 This switch is mainly for debugging the compiler and will likely be removed
18049 in a future version.
18054 Disable the use of conditional set instructions.
18056 This switch is mainly for debugging the compiler and will likely be removed
18057 in a future version.
18060 @opindex mcond-exec
18062 Enable the use of conditional execution (default).
18064 This switch is mainly for debugging the compiler and will likely be removed
18065 in a future version.
18067 @item -mno-cond-exec
18068 @opindex mno-cond-exec
18070 Disable the use of conditional execution.
18072 This switch is mainly for debugging the compiler and will likely be removed
18073 in a future version.
18075 @item -mvliw-branch
18076 @opindex mvliw-branch
18078 Run a pass to pack branches into VLIW instructions (default).
18080 This switch is mainly for debugging the compiler and will likely be removed
18081 in a future version.
18083 @item -mno-vliw-branch
18084 @opindex mno-vliw-branch
18086 Do not run a pass to pack branches into VLIW instructions.
18088 This switch is mainly for debugging the compiler and will likely be removed
18089 in a future version.
18091 @item -mmulti-cond-exec
18092 @opindex mmulti-cond-exec
18094 Enable optimization of @code{&&} and @code{||} in conditional execution
18097 This switch is mainly for debugging the compiler and will likely be removed
18098 in a future version.
18100 @item -mno-multi-cond-exec
18101 @opindex mno-multi-cond-exec
18103 Disable optimization of @code{&&} and @code{||} in conditional execution.
18105 This switch is mainly for debugging the compiler and will likely be removed
18106 in a future version.
18108 @item -mnested-cond-exec
18109 @opindex mnested-cond-exec
18111 Enable nested conditional execution optimizations (default).
18113 This switch is mainly for debugging the compiler and will likely be removed
18114 in a future version.
18116 @item -mno-nested-cond-exec
18117 @opindex mno-nested-cond-exec
18119 Disable nested conditional execution optimizations.
18121 This switch is mainly for debugging the compiler and will likely be removed
18122 in a future version.
18124 @item -moptimize-membar
18125 @opindex moptimize-membar
18127 This switch removes redundant @code{membar} instructions from the
18128 compiler-generated code. It is enabled by default.
18130 @item -mno-optimize-membar
18131 @opindex mno-optimize-membar
18133 This switch disables the automatic removal of redundant @code{membar}
18134 instructions from the generated code.
18136 @item -mtomcat-stats
18137 @opindex mtomcat-stats
18139 Cause gas to print out tomcat statistics.
18141 @item -mcpu=@var{cpu}
18144 Select the processor type for which to generate code. Possible values are
18145 @samp{frv}, @samp{fr550}, @samp{tomcat}, @samp{fr500}, @samp{fr450},
18146 @samp{fr405}, @samp{fr400}, @samp{fr300} and @samp{simple}.
18150 @node GNU/Linux Options
18151 @subsection GNU/Linux Options
18153 These @samp{-m} options are defined for GNU/Linux targets:
18158 Use the GNU C library. This is the default except
18159 on @samp{*-*-linux-*uclibc*}, @samp{*-*-linux-*musl*} and
18160 @samp{*-*-linux-*android*} targets.
18164 Use uClibc C library. This is the default on
18165 @samp{*-*-linux-*uclibc*} targets.
18169 Use the musl C library. This is the default on
18170 @samp{*-*-linux-*musl*} targets.
18174 Use Bionic C library. This is the default on
18175 @samp{*-*-linux-*android*} targets.
18179 Compile code compatible with Android platform. This is the default on
18180 @samp{*-*-linux-*android*} targets.
18182 When compiling, this option enables @option{-mbionic}, @option{-fPIC},
18183 @option{-fno-exceptions} and @option{-fno-rtti} by default. When linking,
18184 this option makes the GCC driver pass Android-specific options to the linker.
18185 Finally, this option causes the preprocessor macro @code{__ANDROID__}
18188 @item -tno-android-cc
18189 @opindex tno-android-cc
18190 Disable compilation effects of @option{-mandroid}, i.e., do not enable
18191 @option{-mbionic}, @option{-fPIC}, @option{-fno-exceptions} and
18192 @option{-fno-rtti} by default.
18194 @item -tno-android-ld
18195 @opindex tno-android-ld
18196 Disable linking effects of @option{-mandroid}, i.e., pass standard Linux
18197 linking options to the linker.
18201 @node H8/300 Options
18202 @subsection H8/300 Options
18204 These @samp{-m} options are defined for the H8/300 implementations:
18209 Shorten some address references at link time, when possible; uses the
18210 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
18211 ld, Using ld}, for a fuller description.
18215 Generate code for the H8/300H@.
18219 Generate code for the H8S@.
18223 Generate code for the H8S and H8/300H in the normal mode. This switch
18224 must be used either with @option{-mh} or @option{-ms}.
18228 Generate code for the H8S/2600. This switch must be used with @option{-ms}.
18232 Extended registers are stored on stack before execution of function
18233 with monitor attribute. Default option is @option{-mexr}.
18234 This option is valid only for H8S targets.
18238 Extended registers are not stored on stack before execution of function
18239 with monitor attribute. Default option is @option{-mno-exr}.
18240 This option is valid only for H8S targets.
18244 Make @code{int} data 32 bits by default.
18247 @opindex malign-300
18248 On the H8/300H and H8S, use the same alignment rules as for the H8/300.
18249 The default for the H8/300H and H8S is to align longs and floats on
18251 @option{-malign-300} causes them to be aligned on 2-byte boundaries.
18252 This option has no effect on the H8/300.
18256 @subsection HPPA Options
18257 @cindex HPPA Options
18259 These @samp{-m} options are defined for the HPPA family of computers:
18262 @item -march=@var{architecture-type}
18264 Generate code for the specified architecture. The choices for
18265 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
18266 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
18267 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
18268 architecture option for your machine. Code compiled for lower numbered
18269 architectures runs on higher numbered architectures, but not the
18272 @item -mpa-risc-1-0
18273 @itemx -mpa-risc-1-1
18274 @itemx -mpa-risc-2-0
18275 @opindex mpa-risc-1-0
18276 @opindex mpa-risc-1-1
18277 @opindex mpa-risc-2-0
18278 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
18280 @item -mcaller-copies
18281 @opindex mcaller-copies
18282 The caller copies function arguments passed by hidden reference. This
18283 option should be used with care as it is not compatible with the default
18284 32-bit runtime. However, only aggregates larger than eight bytes are
18285 passed by hidden reference and the option provides better compatibility
18288 @item -mjump-in-delay
18289 @opindex mjump-in-delay
18290 This option is ignored and provided for compatibility purposes only.
18292 @item -mdisable-fpregs
18293 @opindex mdisable-fpregs
18294 Prevent floating-point registers from being used in any manner. This is
18295 necessary for compiling kernels that perform lazy context switching of
18296 floating-point registers. If you use this option and attempt to perform
18297 floating-point operations, the compiler aborts.
18299 @item -mdisable-indexing
18300 @opindex mdisable-indexing
18301 Prevent the compiler from using indexing address modes. This avoids some
18302 rather obscure problems when compiling MIG generated code under MACH@.
18304 @item -mno-space-regs
18305 @opindex mno-space-regs
18306 Generate code that assumes the target has no space registers. This allows
18307 GCC to generate faster indirect calls and use unscaled index address modes.
18309 Such code is suitable for level 0 PA systems and kernels.
18311 @item -mfast-indirect-calls
18312 @opindex mfast-indirect-calls
18313 Generate code that assumes calls never cross space boundaries. This
18314 allows GCC to emit code that performs faster indirect calls.
18316 This option does not work in the presence of shared libraries or nested
18319 @item -mfixed-range=@var{register-range}
18320 @opindex mfixed-range
18321 Generate code treating the given register range as fixed registers.
18322 A fixed register is one that the register allocator cannot use. This is
18323 useful when compiling kernel code. A register range is specified as
18324 two registers separated by a dash. Multiple register ranges can be
18325 specified separated by a comma.
18327 @item -mlong-load-store
18328 @opindex mlong-load-store
18329 Generate 3-instruction load and store sequences as sometimes required by
18330 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
18333 @item -mportable-runtime
18334 @opindex mportable-runtime
18335 Use the portable calling conventions proposed by HP for ELF systems.
18339 Enable the use of assembler directives only GAS understands.
18341 @item -mschedule=@var{cpu-type}
18343 Schedule code according to the constraints for the machine type
18344 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
18345 @samp{7100}, @samp{7100LC}, @samp{7200}, @samp{7300} and @samp{8000}. Refer
18346 to @file{/usr/lib/sched.models} on an HP-UX system to determine the
18347 proper scheduling option for your machine. The default scheduling is
18351 @opindex mlinker-opt
18352 Enable the optimization pass in the HP-UX linker. Note this makes symbolic
18353 debugging impossible. It also triggers a bug in the HP-UX 8 and HP-UX 9
18354 linkers in which they give bogus error messages when linking some programs.
18357 @opindex msoft-float
18358 Generate output containing library calls for floating point.
18359 @strong{Warning:} the requisite libraries are not available for all HPPA
18360 targets. Normally the facilities of the machine's usual C compiler are
18361 used, but this cannot be done directly in cross-compilation. You must make
18362 your own arrangements to provide suitable library functions for
18365 @option{-msoft-float} changes the calling convention in the output file;
18366 therefore, it is only useful if you compile @emph{all} of a program with
18367 this option. In particular, you need to compile @file{libgcc.a}, the
18368 library that comes with GCC, with @option{-msoft-float} in order for
18373 Generate the predefine, @code{_SIO}, for server IO@. The default is
18374 @option{-mwsio}. This generates the predefines, @code{__hp9000s700},
18375 @code{__hp9000s700__} and @code{_WSIO}, for workstation IO@. These
18376 options are available under HP-UX and HI-UX@.
18380 Use options specific to GNU @command{ld}.
18381 This passes @option{-shared} to @command{ld} when
18382 building a shared library. It is the default when GCC is configured,
18383 explicitly or implicitly, with the GNU linker. This option does not
18384 affect which @command{ld} is called; it only changes what parameters
18385 are passed to that @command{ld}.
18386 The @command{ld} that is called is determined by the
18387 @option{--with-ld} configure option, GCC's program search path, and
18388 finally by the user's @env{PATH}. The linker used by GCC can be printed
18389 using @samp{which `gcc -print-prog-name=ld`}. This option is only available
18390 on the 64-bit HP-UX GCC, i.e.@: configured with @samp{hppa*64*-*-hpux*}.
18394 Use options specific to HP @command{ld}.
18395 This passes @option{-b} to @command{ld} when building
18396 a shared library and passes @option{+Accept TypeMismatch} to @command{ld} on all
18397 links. It is the default when GCC is configured, explicitly or
18398 implicitly, with the HP linker. This option does not affect
18399 which @command{ld} is called; it only changes what parameters are passed to that
18401 The @command{ld} that is called is determined by the @option{--with-ld}
18402 configure option, GCC's program search path, and finally by the user's
18403 @env{PATH}. The linker used by GCC can be printed using @samp{which
18404 `gcc -print-prog-name=ld`}. This option is only available on the 64-bit
18405 HP-UX GCC, i.e.@: configured with @samp{hppa*64*-*-hpux*}.
18408 @opindex mno-long-calls
18409 Generate code that uses long call sequences. This ensures that a call
18410 is always able to reach linker generated stubs. The default is to generate
18411 long calls only when the distance from the call site to the beginning
18412 of the function or translation unit, as the case may be, exceeds a
18413 predefined limit set by the branch type being used. The limits for
18414 normal calls are 7,600,000 and 240,000 bytes, respectively for the
18415 PA 2.0 and PA 1.X architectures. Sibcalls are always limited at
18418 Distances are measured from the beginning of functions when using the
18419 @option{-ffunction-sections} option, or when using the @option{-mgas}
18420 and @option{-mno-portable-runtime} options together under HP-UX with
18423 It is normally not desirable to use this option as it degrades
18424 performance. However, it may be useful in large applications,
18425 particularly when partial linking is used to build the application.
18427 The types of long calls used depends on the capabilities of the
18428 assembler and linker, and the type of code being generated. The
18429 impact on systems that support long absolute calls, and long pic
18430 symbol-difference or pc-relative calls should be relatively small.
18431 However, an indirect call is used on 32-bit ELF systems in pic code
18432 and it is quite long.
18434 @item -munix=@var{unix-std}
18436 Generate compiler predefines and select a startfile for the specified
18437 UNIX standard. The choices for @var{unix-std} are @samp{93}, @samp{95}
18438 and @samp{98}. @samp{93} is supported on all HP-UX versions. @samp{95}
18439 is available on HP-UX 10.10 and later. @samp{98} is available on HP-UX
18440 11.11 and later. The default values are @samp{93} for HP-UX 10.00,
18441 @samp{95} for HP-UX 10.10 though to 11.00, and @samp{98} for HP-UX 11.11
18444 @option{-munix=93} provides the same predefines as GCC 3.3 and 3.4.
18445 @option{-munix=95} provides additional predefines for @code{XOPEN_UNIX}
18446 and @code{_XOPEN_SOURCE_EXTENDED}, and the startfile @file{unix95.o}.
18447 @option{-munix=98} provides additional predefines for @code{_XOPEN_UNIX},
18448 @code{_XOPEN_SOURCE_EXTENDED}, @code{_INCLUDE__STDC_A1_SOURCE} and
18449 @code{_INCLUDE_XOPEN_SOURCE_500}, and the startfile @file{unix98.o}.
18451 It is @emph{important} to note that this option changes the interfaces
18452 for various library routines. It also affects the operational behavior
18453 of the C library. Thus, @emph{extreme} care is needed in using this
18456 Library code that is intended to operate with more than one UNIX
18457 standard must test, set and restore the variable @code{__xpg4_extended_mask}
18458 as appropriate. Most GNU software doesn't provide this capability.
18462 Suppress the generation of link options to search libdld.sl when the
18463 @option{-static} option is specified on HP-UX 10 and later.
18467 The HP-UX implementation of setlocale in libc has a dependency on
18468 libdld.sl. There isn't an archive version of libdld.sl. Thus,
18469 when the @option{-static} option is specified, special link options
18470 are needed to resolve this dependency.
18472 On HP-UX 10 and later, the GCC driver adds the necessary options to
18473 link with libdld.sl when the @option{-static} option is specified.
18474 This causes the resulting binary to be dynamic. On the 64-bit port,
18475 the linkers generate dynamic binaries by default in any case. The
18476 @option{-nolibdld} option can be used to prevent the GCC driver from
18477 adding these link options.
18481 Add support for multithreading with the @dfn{dce thread} library
18482 under HP-UX@. This option sets flags for both the preprocessor and
18486 @node IA-64 Options
18487 @subsection IA-64 Options
18488 @cindex IA-64 Options
18490 These are the @samp{-m} options defined for the Intel IA-64 architecture.
18494 @opindex mbig-endian
18495 Generate code for a big-endian target. This is the default for HP-UX@.
18497 @item -mlittle-endian
18498 @opindex mlittle-endian
18499 Generate code for a little-endian target. This is the default for AIX5
18505 @opindex mno-gnu-as
18506 Generate (or don't) code for the GNU assembler. This is the default.
18507 @c Also, this is the default if the configure option @option{--with-gnu-as}
18513 @opindex mno-gnu-ld
18514 Generate (or don't) code for the GNU linker. This is the default.
18515 @c Also, this is the default if the configure option @option{--with-gnu-ld}
18520 Generate code that does not use a global pointer register. The result
18521 is not position independent code, and violates the IA-64 ABI@.
18523 @item -mvolatile-asm-stop
18524 @itemx -mno-volatile-asm-stop
18525 @opindex mvolatile-asm-stop
18526 @opindex mno-volatile-asm-stop
18527 Generate (or don't) a stop bit immediately before and after volatile asm
18530 @item -mregister-names
18531 @itemx -mno-register-names
18532 @opindex mregister-names
18533 @opindex mno-register-names
18534 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
18535 the stacked registers. This may make assembler output more readable.
18541 Disable (or enable) optimizations that use the small data section. This may
18542 be useful for working around optimizer bugs.
18544 @item -mconstant-gp
18545 @opindex mconstant-gp
18546 Generate code that uses a single constant global pointer value. This is
18547 useful when compiling kernel code.
18551 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
18552 This is useful when compiling firmware code.
18554 @item -minline-float-divide-min-latency
18555 @opindex minline-float-divide-min-latency
18556 Generate code for inline divides of floating-point values
18557 using the minimum latency algorithm.
18559 @item -minline-float-divide-max-throughput
18560 @opindex minline-float-divide-max-throughput
18561 Generate code for inline divides of floating-point values
18562 using the maximum throughput algorithm.
18564 @item -mno-inline-float-divide
18565 @opindex mno-inline-float-divide
18566 Do not generate inline code for divides of floating-point values.
18568 @item -minline-int-divide-min-latency
18569 @opindex minline-int-divide-min-latency
18570 Generate code for inline divides of integer values
18571 using the minimum latency algorithm.
18573 @item -minline-int-divide-max-throughput
18574 @opindex minline-int-divide-max-throughput
18575 Generate code for inline divides of integer values
18576 using the maximum throughput algorithm.
18578 @item -mno-inline-int-divide
18579 @opindex mno-inline-int-divide
18580 Do not generate inline code for divides of integer values.
18582 @item -minline-sqrt-min-latency
18583 @opindex minline-sqrt-min-latency
18584 Generate code for inline square roots
18585 using the minimum latency algorithm.
18587 @item -minline-sqrt-max-throughput
18588 @opindex minline-sqrt-max-throughput
18589 Generate code for inline square roots
18590 using the maximum throughput algorithm.
18592 @item -mno-inline-sqrt
18593 @opindex mno-inline-sqrt
18594 Do not generate inline code for @code{sqrt}.
18597 @itemx -mno-fused-madd
18598 @opindex mfused-madd
18599 @opindex mno-fused-madd
18600 Do (don't) generate code that uses the fused multiply/add or multiply/subtract
18601 instructions. The default is to use these instructions.
18603 @item -mno-dwarf2-asm
18604 @itemx -mdwarf2-asm
18605 @opindex mno-dwarf2-asm
18606 @opindex mdwarf2-asm
18607 Don't (or do) generate assembler code for the DWARF line number debugging
18608 info. This may be useful when not using the GNU assembler.
18610 @item -mearly-stop-bits
18611 @itemx -mno-early-stop-bits
18612 @opindex mearly-stop-bits
18613 @opindex mno-early-stop-bits
18614 Allow stop bits to be placed earlier than immediately preceding the
18615 instruction that triggered the stop bit. This can improve instruction
18616 scheduling, but does not always do so.
18618 @item -mfixed-range=@var{register-range}
18619 @opindex mfixed-range
18620 Generate code treating the given register range as fixed registers.
18621 A fixed register is one that the register allocator cannot use. This is
18622 useful when compiling kernel code. A register range is specified as
18623 two registers separated by a dash. Multiple register ranges can be
18624 specified separated by a comma.
18626 @item -mtls-size=@var{tls-size}
18628 Specify bit size of immediate TLS offsets. Valid values are 14, 22, and
18631 @item -mtune=@var{cpu-type}
18633 Tune the instruction scheduling for a particular CPU, Valid values are
18634 @samp{itanium}, @samp{itanium1}, @samp{merced}, @samp{itanium2},
18635 and @samp{mckinley}.
18641 Generate code for a 32-bit or 64-bit environment.
18642 The 32-bit environment sets int, long and pointer to 32 bits.
18643 The 64-bit environment sets int to 32 bits and long and pointer
18644 to 64 bits. These are HP-UX specific flags.
18646 @item -mno-sched-br-data-spec
18647 @itemx -msched-br-data-spec
18648 @opindex mno-sched-br-data-spec
18649 @opindex msched-br-data-spec
18650 (Dis/En)able data speculative scheduling before reload.
18651 This results in generation of @code{ld.a} instructions and
18652 the corresponding check instructions (@code{ld.c} / @code{chk.a}).
18653 The default setting is disabled.
18655 @item -msched-ar-data-spec
18656 @itemx -mno-sched-ar-data-spec
18657 @opindex msched-ar-data-spec
18658 @opindex mno-sched-ar-data-spec
18659 (En/Dis)able data speculative scheduling after reload.
18660 This results in generation of @code{ld.a} instructions and
18661 the corresponding check instructions (@code{ld.c} / @code{chk.a}).
18662 The default setting is enabled.
18664 @item -mno-sched-control-spec
18665 @itemx -msched-control-spec
18666 @opindex mno-sched-control-spec
18667 @opindex msched-control-spec
18668 (Dis/En)able control speculative scheduling. This feature is
18669 available only during region scheduling (i.e.@: before reload).
18670 This results in generation of the @code{ld.s} instructions and
18671 the corresponding check instructions @code{chk.s}.
18672 The default setting is disabled.
18674 @item -msched-br-in-data-spec
18675 @itemx -mno-sched-br-in-data-spec
18676 @opindex msched-br-in-data-spec
18677 @opindex mno-sched-br-in-data-spec
18678 (En/Dis)able speculative scheduling of the instructions that
18679 are dependent on the data speculative loads before reload.
18680 This is effective only with @option{-msched-br-data-spec} enabled.
18681 The default setting is enabled.
18683 @item -msched-ar-in-data-spec
18684 @itemx -mno-sched-ar-in-data-spec
18685 @opindex msched-ar-in-data-spec
18686 @opindex mno-sched-ar-in-data-spec
18687 (En/Dis)able speculative scheduling of the instructions that
18688 are dependent on the data speculative loads after reload.
18689 This is effective only with @option{-msched-ar-data-spec} enabled.
18690 The default setting is enabled.
18692 @item -msched-in-control-spec
18693 @itemx -mno-sched-in-control-spec
18694 @opindex msched-in-control-spec
18695 @opindex mno-sched-in-control-spec
18696 (En/Dis)able speculative scheduling of the instructions that
18697 are dependent on the control speculative loads.
18698 This is effective only with @option{-msched-control-spec} enabled.
18699 The default setting is enabled.
18701 @item -mno-sched-prefer-non-data-spec-insns
18702 @itemx -msched-prefer-non-data-spec-insns
18703 @opindex mno-sched-prefer-non-data-spec-insns
18704 @opindex msched-prefer-non-data-spec-insns
18705 If enabled, data-speculative instructions are chosen for schedule
18706 only if there are no other choices at the moment. This makes
18707 the use of the data speculation much more conservative.
18708 The default setting is disabled.
18710 @item -mno-sched-prefer-non-control-spec-insns
18711 @itemx -msched-prefer-non-control-spec-insns
18712 @opindex mno-sched-prefer-non-control-spec-insns
18713 @opindex msched-prefer-non-control-spec-insns
18714 If enabled, control-speculative instructions are chosen for schedule
18715 only if there are no other choices at the moment. This makes
18716 the use of the control speculation much more conservative.
18717 The default setting is disabled.
18719 @item -mno-sched-count-spec-in-critical-path
18720 @itemx -msched-count-spec-in-critical-path
18721 @opindex mno-sched-count-spec-in-critical-path
18722 @opindex msched-count-spec-in-critical-path
18723 If enabled, speculative dependencies are considered during
18724 computation of the instructions priorities. This makes the use of the
18725 speculation a bit more conservative.
18726 The default setting is disabled.
18728 @item -msched-spec-ldc
18729 @opindex msched-spec-ldc
18730 Use a simple data speculation check. This option is on by default.
18732 @item -msched-control-spec-ldc
18733 @opindex msched-spec-ldc
18734 Use a simple check for control speculation. This option is on by default.
18736 @item -msched-stop-bits-after-every-cycle
18737 @opindex msched-stop-bits-after-every-cycle
18738 Place a stop bit after every cycle when scheduling. This option is on
18741 @item -msched-fp-mem-deps-zero-cost
18742 @opindex msched-fp-mem-deps-zero-cost
18743 Assume that floating-point stores and loads are not likely to cause a conflict
18744 when placed into the same instruction group. This option is disabled by
18747 @item -msel-sched-dont-check-control-spec
18748 @opindex msel-sched-dont-check-control-spec
18749 Generate checks for control speculation in selective scheduling.
18750 This flag is disabled by default.
18752 @item -msched-max-memory-insns=@var{max-insns}
18753 @opindex msched-max-memory-insns
18754 Limit on the number of memory insns per instruction group, giving lower
18755 priority to subsequent memory insns attempting to schedule in the same
18756 instruction group. Frequently useful to prevent cache bank conflicts.
18757 The default value is 1.
18759 @item -msched-max-memory-insns-hard-limit
18760 @opindex msched-max-memory-insns-hard-limit
18761 Makes the limit specified by @option{msched-max-memory-insns} a hard limit,
18762 disallowing more than that number in an instruction group.
18763 Otherwise, the limit is ``soft'', meaning that non-memory operations
18764 are preferred when the limit is reached, but memory operations may still
18770 @subsection LM32 Options
18771 @cindex LM32 options
18773 These @option{-m} options are defined for the LatticeMico32 architecture:
18776 @item -mbarrel-shift-enabled
18777 @opindex mbarrel-shift-enabled
18778 Enable barrel-shift instructions.
18780 @item -mdivide-enabled
18781 @opindex mdivide-enabled
18782 Enable divide and modulus instructions.
18784 @item -mmultiply-enabled
18785 @opindex multiply-enabled
18786 Enable multiply instructions.
18788 @item -msign-extend-enabled
18789 @opindex msign-extend-enabled
18790 Enable sign extend instructions.
18792 @item -muser-enabled
18793 @opindex muser-enabled
18794 Enable user-defined instructions.
18799 @subsection M32C Options
18800 @cindex M32C options
18803 @item -mcpu=@var{name}
18805 Select the CPU for which code is generated. @var{name} may be one of
18806 @samp{r8c} for the R8C/Tiny series, @samp{m16c} for the M16C (up to
18807 /60) series, @samp{m32cm} for the M16C/80 series, or @samp{m32c} for
18808 the M32C/80 series.
18812 Specifies that the program will be run on the simulator. This causes
18813 an alternate runtime library to be linked in which supports, for
18814 example, file I/O@. You must not use this option when generating
18815 programs that will run on real hardware; you must provide your own
18816 runtime library for whatever I/O functions are needed.
18818 @item -memregs=@var{number}
18820 Specifies the number of memory-based pseudo-registers GCC uses
18821 during code generation. These pseudo-registers are used like real
18822 registers, so there is a tradeoff between GCC's ability to fit the
18823 code into available registers, and the performance penalty of using
18824 memory instead of registers. Note that all modules in a program must
18825 be compiled with the same value for this option. Because of that, you
18826 must not use this option with GCC's default runtime libraries.
18830 @node M32R/D Options
18831 @subsection M32R/D Options
18832 @cindex M32R/D options
18834 These @option{-m} options are defined for Renesas M32R/D architectures:
18839 Generate code for the M32R/2@.
18843 Generate code for the M32R/X@.
18847 Generate code for the M32R@. This is the default.
18849 @item -mmodel=small
18850 @opindex mmodel=small
18851 Assume all objects live in the lower 16MB of memory (so that their addresses
18852 can be loaded with the @code{ld24} instruction), and assume all subroutines
18853 are reachable with the @code{bl} instruction.
18854 This is the default.
18856 The addressability of a particular object can be set with the
18857 @code{model} attribute.
18859 @item -mmodel=medium
18860 @opindex mmodel=medium
18861 Assume objects may be anywhere in the 32-bit address space (the compiler
18862 generates @code{seth/add3} instructions to load their addresses), and
18863 assume all subroutines are reachable with the @code{bl} instruction.
18865 @item -mmodel=large
18866 @opindex mmodel=large
18867 Assume objects may be anywhere in the 32-bit address space (the compiler
18868 generates @code{seth/add3} instructions to load their addresses), and
18869 assume subroutines may not be reachable with the @code{bl} instruction
18870 (the compiler generates the much slower @code{seth/add3/jl}
18871 instruction sequence).
18874 @opindex msdata=none
18875 Disable use of the small data area. Variables are put into
18876 one of @code{.data}, @code{.bss}, or @code{.rodata} (unless the
18877 @code{section} attribute has been specified).
18878 This is the default.
18880 The small data area consists of sections @code{.sdata} and @code{.sbss}.
18881 Objects may be explicitly put in the small data area with the
18882 @code{section} attribute using one of these sections.
18884 @item -msdata=sdata
18885 @opindex msdata=sdata
18886 Put small global and static data in the small data area, but do not
18887 generate special code to reference them.
18890 @opindex msdata=use
18891 Put small global and static data in the small data area, and generate
18892 special instructions to reference them.
18896 @cindex smaller data references
18897 Put global and static objects less than or equal to @var{num} bytes
18898 into the small data or BSS sections instead of the normal data or BSS
18899 sections. The default value of @var{num} is 8.
18900 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
18901 for this option to have any effect.
18903 All modules should be compiled with the same @option{-G @var{num}} value.
18904 Compiling with different values of @var{num} may or may not work; if it
18905 doesn't the linker gives an error message---incorrect code is not
18910 Makes the M32R-specific code in the compiler display some statistics
18911 that might help in debugging programs.
18913 @item -malign-loops
18914 @opindex malign-loops
18915 Align all loops to a 32-byte boundary.
18917 @item -mno-align-loops
18918 @opindex mno-align-loops
18919 Do not enforce a 32-byte alignment for loops. This is the default.
18921 @item -missue-rate=@var{number}
18922 @opindex missue-rate=@var{number}
18923 Issue @var{number} instructions per cycle. @var{number} can only be 1
18926 @item -mbranch-cost=@var{number}
18927 @opindex mbranch-cost=@var{number}
18928 @var{number} can only be 1 or 2. If it is 1 then branches are
18929 preferred over conditional code, if it is 2, then the opposite applies.
18931 @item -mflush-trap=@var{number}
18932 @opindex mflush-trap=@var{number}
18933 Specifies the trap number to use to flush the cache. The default is
18934 12. Valid numbers are between 0 and 15 inclusive.
18936 @item -mno-flush-trap
18937 @opindex mno-flush-trap
18938 Specifies that the cache cannot be flushed by using a trap.
18940 @item -mflush-func=@var{name}
18941 @opindex mflush-func=@var{name}
18942 Specifies the name of the operating system function to call to flush
18943 the cache. The default is @samp{_flush_cache}, but a function call
18944 is only used if a trap is not available.
18946 @item -mno-flush-func
18947 @opindex mno-flush-func
18948 Indicates that there is no OS function for flushing the cache.
18952 @node M680x0 Options
18953 @subsection M680x0 Options
18954 @cindex M680x0 options
18956 These are the @samp{-m} options defined for M680x0 and ColdFire processors.
18957 The default settings depend on which architecture was selected when
18958 the compiler was configured; the defaults for the most common choices
18962 @item -march=@var{arch}
18964 Generate code for a specific M680x0 or ColdFire instruction set
18965 architecture. Permissible values of @var{arch} for M680x0
18966 architectures are: @samp{68000}, @samp{68010}, @samp{68020},
18967 @samp{68030}, @samp{68040}, @samp{68060} and @samp{cpu32}. ColdFire
18968 architectures are selected according to Freescale's ISA classification
18969 and the permissible values are: @samp{isaa}, @samp{isaaplus},
18970 @samp{isab} and @samp{isac}.
18972 GCC defines a macro @code{__mcf@var{arch}__} whenever it is generating
18973 code for a ColdFire target. The @var{arch} in this macro is one of the
18974 @option{-march} arguments given above.
18976 When used together, @option{-march} and @option{-mtune} select code
18977 that runs on a family of similar processors but that is optimized
18978 for a particular microarchitecture.
18980 @item -mcpu=@var{cpu}
18982 Generate code for a specific M680x0 or ColdFire processor.
18983 The M680x0 @var{cpu}s are: @samp{68000}, @samp{68010}, @samp{68020},
18984 @samp{68030}, @samp{68040}, @samp{68060}, @samp{68302}, @samp{68332}
18985 and @samp{cpu32}. The ColdFire @var{cpu}s are given by the table
18986 below, which also classifies the CPUs into families:
18988 @multitable @columnfractions 0.20 0.80
18989 @item @strong{Family} @tab @strong{@samp{-mcpu} arguments}
18990 @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}
18991 @item @samp{5206} @tab @samp{5202} @samp{5204} @samp{5206}
18992 @item @samp{5206e} @tab @samp{5206e}
18993 @item @samp{5208} @tab @samp{5207} @samp{5208}
18994 @item @samp{5211a} @tab @samp{5210a} @samp{5211a}
18995 @item @samp{5213} @tab @samp{5211} @samp{5212} @samp{5213}
18996 @item @samp{5216} @tab @samp{5214} @samp{5216}
18997 @item @samp{52235} @tab @samp{52230} @samp{52231} @samp{52232} @samp{52233} @samp{52234} @samp{52235}
18998 @item @samp{5225} @tab @samp{5224} @samp{5225}
18999 @item @samp{52259} @tab @samp{52252} @samp{52254} @samp{52255} @samp{52256} @samp{52258} @samp{52259}
19000 @item @samp{5235} @tab @samp{5232} @samp{5233} @samp{5234} @samp{5235} @samp{523x}
19001 @item @samp{5249} @tab @samp{5249}
19002 @item @samp{5250} @tab @samp{5250}
19003 @item @samp{5271} @tab @samp{5270} @samp{5271}
19004 @item @samp{5272} @tab @samp{5272}
19005 @item @samp{5275} @tab @samp{5274} @samp{5275}
19006 @item @samp{5282} @tab @samp{5280} @samp{5281} @samp{5282} @samp{528x}
19007 @item @samp{53017} @tab @samp{53011} @samp{53012} @samp{53013} @samp{53014} @samp{53015} @samp{53016} @samp{53017}
19008 @item @samp{5307} @tab @samp{5307}
19009 @item @samp{5329} @tab @samp{5327} @samp{5328} @samp{5329} @samp{532x}
19010 @item @samp{5373} @tab @samp{5372} @samp{5373} @samp{537x}
19011 @item @samp{5407} @tab @samp{5407}
19012 @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}
19015 @option{-mcpu=@var{cpu}} overrides @option{-march=@var{arch}} if
19016 @var{arch} is compatible with @var{cpu}. Other combinations of
19017 @option{-mcpu} and @option{-march} are rejected.
19019 GCC defines the macro @code{__mcf_cpu_@var{cpu}} when ColdFire target
19020 @var{cpu} is selected. It also defines @code{__mcf_family_@var{family}},
19021 where the value of @var{family} is given by the table above.
19023 @item -mtune=@var{tune}
19025 Tune the code for a particular microarchitecture within the
19026 constraints set by @option{-march} and @option{-mcpu}.
19027 The M680x0 microarchitectures are: @samp{68000}, @samp{68010},
19028 @samp{68020}, @samp{68030}, @samp{68040}, @samp{68060}
19029 and @samp{cpu32}. The ColdFire microarchitectures
19030 are: @samp{cfv1}, @samp{cfv2}, @samp{cfv3}, @samp{cfv4} and @samp{cfv4e}.
19032 You can also use @option{-mtune=68020-40} for code that needs
19033 to run relatively well on 68020, 68030 and 68040 targets.
19034 @option{-mtune=68020-60} is similar but includes 68060 targets
19035 as well. These two options select the same tuning decisions as
19036 @option{-m68020-40} and @option{-m68020-60} respectively.
19038 GCC defines the macros @code{__mc@var{arch}} and @code{__mc@var{arch}__}
19039 when tuning for 680x0 architecture @var{arch}. It also defines
19040 @code{mc@var{arch}} unless either @option{-ansi} or a non-GNU @option{-std}
19041 option is used. If GCC is tuning for a range of architectures,
19042 as selected by @option{-mtune=68020-40} or @option{-mtune=68020-60},
19043 it defines the macros for every architecture in the range.
19045 GCC also defines the macro @code{__m@var{uarch}__} when tuning for
19046 ColdFire microarchitecture @var{uarch}, where @var{uarch} is one
19047 of the arguments given above.
19053 Generate output for a 68000. This is the default
19054 when the compiler is configured for 68000-based systems.
19055 It is equivalent to @option{-march=68000}.
19057 Use this option for microcontrollers with a 68000 or EC000 core,
19058 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
19062 Generate output for a 68010. This is the default
19063 when the compiler is configured for 68010-based systems.
19064 It is equivalent to @option{-march=68010}.
19070 Generate output for a 68020. This is the default
19071 when the compiler is configured for 68020-based systems.
19072 It is equivalent to @option{-march=68020}.
19076 Generate output for a 68030. This is the default when the compiler is
19077 configured for 68030-based systems. It is equivalent to
19078 @option{-march=68030}.
19082 Generate output for a 68040. This is the default when the compiler is
19083 configured for 68040-based systems. It is equivalent to
19084 @option{-march=68040}.
19086 This option inhibits the use of 68881/68882 instructions that have to be
19087 emulated by software on the 68040. Use this option if your 68040 does not
19088 have code to emulate those instructions.
19092 Generate output for a 68060. This is the default when the compiler is
19093 configured for 68060-based systems. It is equivalent to
19094 @option{-march=68060}.
19096 This option inhibits the use of 68020 and 68881/68882 instructions that
19097 have to be emulated by software on the 68060. Use this option if your 68060
19098 does not have code to emulate those instructions.
19102 Generate output for a CPU32. This is the default
19103 when the compiler is configured for CPU32-based systems.
19104 It is equivalent to @option{-march=cpu32}.
19106 Use this option for microcontrollers with a
19107 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
19108 68336, 68340, 68341, 68349 and 68360.
19112 Generate output for a 520X ColdFire CPU@. This is the default
19113 when the compiler is configured for 520X-based systems.
19114 It is equivalent to @option{-mcpu=5206}, and is now deprecated
19115 in favor of that option.
19117 Use this option for microcontroller with a 5200 core, including
19118 the MCF5202, MCF5203, MCF5204 and MCF5206.
19122 Generate output for a 5206e ColdFire CPU@. The option is now
19123 deprecated in favor of the equivalent @option{-mcpu=5206e}.
19127 Generate output for a member of the ColdFire 528X family.
19128 The option is now deprecated in favor of the equivalent
19129 @option{-mcpu=528x}.
19133 Generate output for a ColdFire 5307 CPU@. The option is now deprecated
19134 in favor of the equivalent @option{-mcpu=5307}.
19138 Generate output for a ColdFire 5407 CPU@. The option is now deprecated
19139 in favor of the equivalent @option{-mcpu=5407}.
19143 Generate output for a ColdFire V4e family CPU (e.g.@: 547x/548x).
19144 This includes use of hardware floating-point instructions.
19145 The option is equivalent to @option{-mcpu=547x}, and is now
19146 deprecated in favor of that option.
19150 Generate output for a 68040, without using any of the new instructions.
19151 This results in code that can run relatively efficiently on either a
19152 68020/68881 or a 68030 or a 68040. The generated code does use the
19153 68881 instructions that are emulated on the 68040.
19155 The option is equivalent to @option{-march=68020} @option{-mtune=68020-40}.
19159 Generate output for a 68060, without using any of the new instructions.
19160 This results in code that can run relatively efficiently on either a
19161 68020/68881 or a 68030 or a 68040. The generated code does use the
19162 68881 instructions that are emulated on the 68060.
19164 The option is equivalent to @option{-march=68020} @option{-mtune=68020-60}.
19168 @opindex mhard-float
19170 Generate floating-point instructions. This is the default for 68020
19171 and above, and for ColdFire devices that have an FPU@. It defines the
19172 macro @code{__HAVE_68881__} on M680x0 targets and @code{__mcffpu__}
19173 on ColdFire targets.
19176 @opindex msoft-float
19177 Do not generate floating-point instructions; use library calls instead.
19178 This is the default for 68000, 68010, and 68832 targets. It is also
19179 the default for ColdFire devices that have no FPU.
19185 Generate (do not generate) ColdFire hardware divide and remainder
19186 instructions. If @option{-march} is used without @option{-mcpu},
19187 the default is ``on'' for ColdFire architectures and ``off'' for M680x0
19188 architectures. Otherwise, the default is taken from the target CPU
19189 (either the default CPU, or the one specified by @option{-mcpu}). For
19190 example, the default is ``off'' for @option{-mcpu=5206} and ``on'' for
19191 @option{-mcpu=5206e}.
19193 GCC defines the macro @code{__mcfhwdiv__} when this option is enabled.
19197 Consider type @code{int} to be 16 bits wide, like @code{short int}.
19198 Additionally, parameters passed on the stack are also aligned to a
19199 16-bit boundary even on targets whose API mandates promotion to 32-bit.
19203 Do not consider type @code{int} to be 16 bits wide. This is the default.
19206 @itemx -mno-bitfield
19207 @opindex mnobitfield
19208 @opindex mno-bitfield
19209 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
19210 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
19214 Do use the bit-field instructions. The @option{-m68020} option implies
19215 @option{-mbitfield}. This is the default if you use a configuration
19216 designed for a 68020.
19220 Use a different function-calling convention, in which functions
19221 that take a fixed number of arguments return with the @code{rtd}
19222 instruction, which pops their arguments while returning. This
19223 saves one instruction in the caller since there is no need to pop
19224 the arguments there.
19226 This calling convention is incompatible with the one normally
19227 used on Unix, so you cannot use it if you need to call libraries
19228 compiled with the Unix compiler.
19230 Also, you must provide function prototypes for all functions that
19231 take variable numbers of arguments (including @code{printf});
19232 otherwise incorrect code is generated for calls to those
19235 In addition, seriously incorrect code results if you call a
19236 function with too many arguments. (Normally, extra arguments are
19237 harmlessly ignored.)
19239 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
19240 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
19244 Do not use the calling conventions selected by @option{-mrtd}.
19245 This is the default.
19248 @itemx -mno-align-int
19249 @opindex malign-int
19250 @opindex mno-align-int
19251 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
19252 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
19253 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
19254 Aligning variables on 32-bit boundaries produces code that runs somewhat
19255 faster on processors with 32-bit busses at the expense of more memory.
19257 @strong{Warning:} if you use the @option{-malign-int} switch, GCC
19258 aligns structures containing the above types differently than
19259 most published application binary interface specifications for the m68k.
19263 Use the pc-relative addressing mode of the 68000 directly, instead of
19264 using a global offset table. At present, this option implies @option{-fpic},
19265 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
19266 not presently supported with @option{-mpcrel}, though this could be supported for
19267 68020 and higher processors.
19269 @item -mno-strict-align
19270 @itemx -mstrict-align
19271 @opindex mno-strict-align
19272 @opindex mstrict-align
19273 Do not (do) assume that unaligned memory references are handled by
19277 Generate code that allows the data segment to be located in a different
19278 area of memory from the text segment. This allows for execute-in-place in
19279 an environment without virtual memory management. This option implies
19282 @item -mno-sep-data
19283 Generate code that assumes that the data segment follows the text segment.
19284 This is the default.
19286 @item -mid-shared-library
19287 Generate code that supports shared libraries via the library ID method.
19288 This allows for execute-in-place and shared libraries in an environment
19289 without virtual memory management. This option implies @option{-fPIC}.
19291 @item -mno-id-shared-library
19292 Generate code that doesn't assume ID-based shared libraries are being used.
19293 This is the default.
19295 @item -mshared-library-id=n
19296 Specifies the identification number of the ID-based shared library being
19297 compiled. Specifying a value of 0 generates more compact code; specifying
19298 other values forces the allocation of that number to the current
19299 library, but is no more space- or time-efficient than omitting this option.
19305 When generating position-independent code for ColdFire, generate code
19306 that works if the GOT has more than 8192 entries. This code is
19307 larger and slower than code generated without this option. On M680x0
19308 processors, this option is not needed; @option{-fPIC} suffices.
19310 GCC normally uses a single instruction to load values from the GOT@.
19311 While this is relatively efficient, it only works if the GOT
19312 is smaller than about 64k. Anything larger causes the linker
19313 to report an error such as:
19315 @cindex relocation truncated to fit (ColdFire)
19317 relocation truncated to fit: R_68K_GOT16O foobar
19320 If this happens, you should recompile your code with @option{-mxgot}.
19321 It should then work with very large GOTs. However, code generated with
19322 @option{-mxgot} is less efficient, since it takes 4 instructions to fetch
19323 the value of a global symbol.
19325 Note that some linkers, including newer versions of the GNU linker,
19326 can create multiple GOTs and sort GOT entries. If you have such a linker,
19327 you should only need to use @option{-mxgot} when compiling a single
19328 object file that accesses more than 8192 GOT entries. Very few do.
19330 These options have no effect unless GCC is generating
19331 position-independent code.
19333 @item -mlong-jump-table-offsets
19334 @opindex mlong-jump-table-offsets
19335 Use 32-bit offsets in @code{switch} tables. The default is to use
19340 @node MCore Options
19341 @subsection MCore Options
19342 @cindex MCore options
19344 These are the @samp{-m} options defined for the Motorola M*Core
19350 @itemx -mno-hardlit
19352 @opindex mno-hardlit
19353 Inline constants into the code stream if it can be done in two
19354 instructions or less.
19360 Use the divide instruction. (Enabled by default).
19362 @item -mrelax-immediate
19363 @itemx -mno-relax-immediate
19364 @opindex mrelax-immediate
19365 @opindex mno-relax-immediate
19366 Allow arbitrary-sized immediates in bit operations.
19368 @item -mwide-bitfields
19369 @itemx -mno-wide-bitfields
19370 @opindex mwide-bitfields
19371 @opindex mno-wide-bitfields
19372 Always treat bit-fields as @code{int}-sized.
19374 @item -m4byte-functions
19375 @itemx -mno-4byte-functions
19376 @opindex m4byte-functions
19377 @opindex mno-4byte-functions
19378 Force all functions to be aligned to a 4-byte boundary.
19380 @item -mcallgraph-data
19381 @itemx -mno-callgraph-data
19382 @opindex mcallgraph-data
19383 @opindex mno-callgraph-data
19384 Emit callgraph information.
19387 @itemx -mno-slow-bytes
19388 @opindex mslow-bytes
19389 @opindex mno-slow-bytes
19390 Prefer word access when reading byte quantities.
19392 @item -mlittle-endian
19393 @itemx -mbig-endian
19394 @opindex mlittle-endian
19395 @opindex mbig-endian
19396 Generate code for a little-endian target.
19402 Generate code for the 210 processor.
19406 Assume that runtime support has been provided and so omit the
19407 simulator library (@file{libsim.a)} from the linker command line.
19409 @item -mstack-increment=@var{size}
19410 @opindex mstack-increment
19411 Set the maximum amount for a single stack increment operation. Large
19412 values can increase the speed of programs that contain functions
19413 that need a large amount of stack space, but they can also trigger a
19414 segmentation fault if the stack is extended too much. The default
19420 @subsection MeP Options
19421 @cindex MeP options
19427 Enables the @code{abs} instruction, which is the absolute difference
19428 between two registers.
19432 Enables all the optional instructions---average, multiply, divide, bit
19433 operations, leading zero, absolute difference, min/max, clip, and
19439 Enables the @code{ave} instruction, which computes the average of two
19442 @item -mbased=@var{n}
19444 Variables of size @var{n} bytes or smaller are placed in the
19445 @code{.based} section by default. Based variables use the @code{$tp}
19446 register as a base register, and there is a 128-byte limit to the
19447 @code{.based} section.
19451 Enables the bit operation instructions---bit test (@code{btstm}), set
19452 (@code{bsetm}), clear (@code{bclrm}), invert (@code{bnotm}), and
19453 test-and-set (@code{tas}).
19455 @item -mc=@var{name}
19457 Selects which section constant data is placed in. @var{name} may
19458 be @samp{tiny}, @samp{near}, or @samp{far}.
19462 Enables the @code{clip} instruction. Note that @option{-mclip} is not
19463 useful unless you also provide @option{-mminmax}.
19465 @item -mconfig=@var{name}
19467 Selects one of the built-in core configurations. Each MeP chip has
19468 one or more modules in it; each module has a core CPU and a variety of
19469 coprocessors, optional instructions, and peripherals. The
19470 @code{MeP-Integrator} tool, not part of GCC, provides these
19471 configurations through this option; using this option is the same as
19472 using all the corresponding command-line options. The default
19473 configuration is @samp{default}.
19477 Enables the coprocessor instructions. By default, this is a 32-bit
19478 coprocessor. Note that the coprocessor is normally enabled via the
19479 @option{-mconfig=} option.
19483 Enables the 32-bit coprocessor's instructions.
19487 Enables the 64-bit coprocessor's instructions.
19491 Enables IVC2 scheduling. IVC2 is a 64-bit VLIW coprocessor.
19495 Causes constant variables to be placed in the @code{.near} section.
19499 Enables the @code{div} and @code{divu} instructions.
19503 Generate big-endian code.
19507 Generate little-endian code.
19509 @item -mio-volatile
19510 @opindex mio-volatile
19511 Tells the compiler that any variable marked with the @code{io}
19512 attribute is to be considered volatile.
19516 Causes variables to be assigned to the @code{.far} section by default.
19520 Enables the @code{leadz} (leading zero) instruction.
19524 Causes variables to be assigned to the @code{.near} section by default.
19528 Enables the @code{min} and @code{max} instructions.
19532 Enables the multiplication and multiply-accumulate instructions.
19536 Disables all the optional instructions enabled by @option{-mall-opts}.
19540 Enables the @code{repeat} and @code{erepeat} instructions, used for
19541 low-overhead looping.
19545 Causes all variables to default to the @code{.tiny} section. Note
19546 that there is a 65536-byte limit to this section. Accesses to these
19547 variables use the @code{%gp} base register.
19551 Enables the saturation instructions. Note that the compiler does not
19552 currently generate these itself, but this option is included for
19553 compatibility with other tools, like @code{as}.
19557 Link the SDRAM-based runtime instead of the default ROM-based runtime.
19561 Link the simulator run-time libraries.
19565 Link the simulator runtime libraries, excluding built-in support
19566 for reset and exception vectors and tables.
19570 Causes all functions to default to the @code{.far} section. Without
19571 this option, functions default to the @code{.near} section.
19573 @item -mtiny=@var{n}
19575 Variables that are @var{n} bytes or smaller are allocated to the
19576 @code{.tiny} section. These variables use the @code{$gp} base
19577 register. The default for this option is 4, but note that there's a
19578 65536-byte limit to the @code{.tiny} section.
19582 @node MicroBlaze Options
19583 @subsection MicroBlaze Options
19584 @cindex MicroBlaze Options
19589 @opindex msoft-float
19590 Use software emulation for floating point (default).
19593 @opindex mhard-float
19594 Use hardware floating-point instructions.
19598 Do not optimize block moves, use @code{memcpy}.
19600 @item -mno-clearbss
19601 @opindex mno-clearbss
19602 This option is deprecated. Use @option{-fno-zero-initialized-in-bss} instead.
19604 @item -mcpu=@var{cpu-type}
19606 Use features of, and schedule code for, the given CPU.
19607 Supported values are in the format @samp{v@var{X}.@var{YY}.@var{Z}},
19608 where @var{X} is a major version, @var{YY} is the minor version, and
19609 @var{Z} is compatibility code. Example values are @samp{v3.00.a},
19610 @samp{v4.00.b}, @samp{v5.00.a}, @samp{v5.00.b}, @samp{v5.00.b}, @samp{v6.00.a}.
19612 @item -mxl-soft-mul
19613 @opindex mxl-soft-mul
19614 Use software multiply emulation (default).
19616 @item -mxl-soft-div
19617 @opindex mxl-soft-div
19618 Use software emulation for divides (default).
19620 @item -mxl-barrel-shift
19621 @opindex mxl-barrel-shift
19622 Use the hardware barrel shifter.
19624 @item -mxl-pattern-compare
19625 @opindex mxl-pattern-compare
19626 Use pattern compare instructions.
19628 @item -msmall-divides
19629 @opindex msmall-divides
19630 Use table lookup optimization for small signed integer divisions.
19632 @item -mxl-stack-check
19633 @opindex mxl-stack-check
19634 This option is deprecated. Use @option{-fstack-check} instead.
19637 @opindex mxl-gp-opt
19638 Use GP-relative @code{.sdata}/@code{.sbss} sections.
19640 @item -mxl-multiply-high
19641 @opindex mxl-multiply-high
19642 Use multiply high instructions for high part of 32x32 multiply.
19644 @item -mxl-float-convert
19645 @opindex mxl-float-convert
19646 Use hardware floating-point conversion instructions.
19648 @item -mxl-float-sqrt
19649 @opindex mxl-float-sqrt
19650 Use hardware floating-point square root instruction.
19653 @opindex mbig-endian
19654 Generate code for a big-endian target.
19656 @item -mlittle-endian
19657 @opindex mlittle-endian
19658 Generate code for a little-endian target.
19661 @opindex mxl-reorder
19662 Use reorder instructions (swap and byte reversed load/store).
19664 @item -mxl-mode-@var{app-model}
19665 Select application model @var{app-model}. Valid models are
19668 normal executable (default), uses startup code @file{crt0.o}.
19671 for use with Xilinx Microprocessor Debugger (XMD) based
19672 software intrusive debug agent called xmdstub. This uses startup file
19673 @file{crt1.o} and sets the start address of the program to 0x800.
19676 for applications that are loaded using a bootloader.
19677 This model uses startup file @file{crt2.o} which does not contain a processor
19678 reset vector handler. This is suitable for transferring control on a
19679 processor reset to the bootloader rather than the application.
19682 for applications that do not require any of the
19683 MicroBlaze vectors. This option may be useful for applications running
19684 within a monitoring application. This model uses @file{crt3.o} as a startup file.
19687 Option @option{-xl-mode-@var{app-model}} is a deprecated alias for
19688 @option{-mxl-mode-@var{app-model}}.
19693 @subsection MIPS Options
19694 @cindex MIPS options
19700 Generate big-endian code.
19704 Generate little-endian code. This is the default for @samp{mips*el-*-*}
19707 @item -march=@var{arch}
19709 Generate code that runs on @var{arch}, which can be the name of a
19710 generic MIPS ISA, or the name of a particular processor.
19712 @samp{mips1}, @samp{mips2}, @samp{mips3}, @samp{mips4},
19713 @samp{mips32}, @samp{mips32r2}, @samp{mips32r3}, @samp{mips32r5},
19714 @samp{mips32r6}, @samp{mips64}, @samp{mips64r2}, @samp{mips64r3},
19715 @samp{mips64r5} and @samp{mips64r6}.
19716 The processor names are:
19717 @samp{4kc}, @samp{4km}, @samp{4kp}, @samp{4ksc},
19718 @samp{4kec}, @samp{4kem}, @samp{4kep}, @samp{4ksd},
19719 @samp{5kc}, @samp{5kf},
19721 @samp{24kc}, @samp{24kf2_1}, @samp{24kf1_1},
19722 @samp{24kec}, @samp{24kef2_1}, @samp{24kef1_1},
19723 @samp{34kc}, @samp{34kf2_1}, @samp{34kf1_1}, @samp{34kn},
19724 @samp{74kc}, @samp{74kf2_1}, @samp{74kf1_1}, @samp{74kf3_2},
19725 @samp{1004kc}, @samp{1004kf2_1}, @samp{1004kf1_1},
19728 @samp{loongson2e}, @samp{loongson2f}, @samp{loongson3a},
19730 @samp{m14k}, @samp{m14kc}, @samp{m14ke}, @samp{m14kec},
19731 @samp{m5100}, @samp{m5101},
19732 @samp{octeon}, @samp{octeon+}, @samp{octeon2}, @samp{octeon3},
19735 @samp{r2000}, @samp{r3000}, @samp{r3900}, @samp{r4000}, @samp{r4400},
19736 @samp{r4600}, @samp{r4650}, @samp{r4700}, @samp{r6000}, @samp{r8000},
19737 @samp{rm7000}, @samp{rm9000},
19738 @samp{r10000}, @samp{r12000}, @samp{r14000}, @samp{r16000},
19741 @samp{vr4100}, @samp{vr4111}, @samp{vr4120}, @samp{vr4130}, @samp{vr4300},
19742 @samp{vr5000}, @samp{vr5400}, @samp{vr5500},
19743 @samp{xlr} and @samp{xlp}.
19744 The special value @samp{from-abi} selects the
19745 most compatible architecture for the selected ABI (that is,
19746 @samp{mips1} for 32-bit ABIs and @samp{mips3} for 64-bit ABIs)@.
19748 The native Linux/GNU toolchain also supports the value @samp{native},
19749 which selects the best architecture option for the host processor.
19750 @option{-march=native} has no effect if GCC does not recognize
19753 In processor names, a final @samp{000} can be abbreviated as @samp{k}
19754 (for example, @option{-march=r2k}). Prefixes are optional, and
19755 @samp{vr} may be written @samp{r}.
19757 Names of the form @samp{@var{n}f2_1} refer to processors with
19758 FPUs clocked at half the rate of the core, names of the form
19759 @samp{@var{n}f1_1} refer to processors with FPUs clocked at the same
19760 rate as the core, and names of the form @samp{@var{n}f3_2} refer to
19761 processors with FPUs clocked a ratio of 3:2 with respect to the core.
19762 For compatibility reasons, @samp{@var{n}f} is accepted as a synonym
19763 for @samp{@var{n}f2_1} while @samp{@var{n}x} and @samp{@var{b}fx} are
19764 accepted as synonyms for @samp{@var{n}f1_1}.
19766 GCC defines two macros based on the value of this option. The first
19767 is @code{_MIPS_ARCH}, which gives the name of target architecture, as
19768 a string. The second has the form @code{_MIPS_ARCH_@var{foo}},
19769 where @var{foo} is the capitalized value of @code{_MIPS_ARCH}@.
19770 For example, @option{-march=r2000} sets @code{_MIPS_ARCH}
19771 to @code{"r2000"} and defines the macro @code{_MIPS_ARCH_R2000}.
19773 Note that the @code{_MIPS_ARCH} macro uses the processor names given
19774 above. In other words, it has the full prefix and does not
19775 abbreviate @samp{000} as @samp{k}. In the case of @samp{from-abi},
19776 the macro names the resolved architecture (either @code{"mips1"} or
19777 @code{"mips3"}). It names the default architecture when no
19778 @option{-march} option is given.
19780 @item -mtune=@var{arch}
19782 Optimize for @var{arch}. Among other things, this option controls
19783 the way instructions are scheduled, and the perceived cost of arithmetic
19784 operations. The list of @var{arch} values is the same as for
19787 When this option is not used, GCC optimizes for the processor
19788 specified by @option{-march}. By using @option{-march} and
19789 @option{-mtune} together, it is possible to generate code that
19790 runs on a family of processors, but optimize the code for one
19791 particular member of that family.
19793 @option{-mtune} defines the macros @code{_MIPS_TUNE} and
19794 @code{_MIPS_TUNE_@var{foo}}, which work in the same way as the
19795 @option{-march} ones described above.
19799 Equivalent to @option{-march=mips1}.
19803 Equivalent to @option{-march=mips2}.
19807 Equivalent to @option{-march=mips3}.
19811 Equivalent to @option{-march=mips4}.
19815 Equivalent to @option{-march=mips32}.
19819 Equivalent to @option{-march=mips32r3}.
19823 Equivalent to @option{-march=mips32r5}.
19827 Equivalent to @option{-march=mips32r6}.
19831 Equivalent to @option{-march=mips64}.
19835 Equivalent to @option{-march=mips64r2}.
19839 Equivalent to @option{-march=mips64r3}.
19843 Equivalent to @option{-march=mips64r5}.
19847 Equivalent to @option{-march=mips64r6}.
19852 @opindex mno-mips16
19853 Generate (do not generate) MIPS16 code. If GCC is targeting a
19854 MIPS32 or MIPS64 architecture, it makes use of the MIPS16e ASE@.
19856 MIPS16 code generation can also be controlled on a per-function basis
19857 by means of @code{mips16} and @code{nomips16} attributes.
19858 @xref{Function Attributes}, for more information.
19860 @item -mflip-mips16
19861 @opindex mflip-mips16
19862 Generate MIPS16 code on alternating functions. This option is provided
19863 for regression testing of mixed MIPS16/non-MIPS16 code generation, and is
19864 not intended for ordinary use in compiling user code.
19866 @item -minterlink-compressed
19867 @item -mno-interlink-compressed
19868 @opindex minterlink-compressed
19869 @opindex mno-interlink-compressed
19870 Require (do not require) that code using the standard (uncompressed) MIPS ISA
19871 be link-compatible with MIPS16 and microMIPS code, and vice versa.
19873 For example, code using the standard ISA encoding cannot jump directly
19874 to MIPS16 or microMIPS code; it must either use a call or an indirect jump.
19875 @option{-minterlink-compressed} therefore disables direct jumps unless GCC
19876 knows that the target of the jump is not compressed.
19878 @item -minterlink-mips16
19879 @itemx -mno-interlink-mips16
19880 @opindex minterlink-mips16
19881 @opindex mno-interlink-mips16
19882 Aliases of @option{-minterlink-compressed} and
19883 @option{-mno-interlink-compressed}. These options predate the microMIPS ASE
19884 and are retained for backwards compatibility.
19896 Generate code for the given ABI@.
19898 Note that the EABI has a 32-bit and a 64-bit variant. GCC normally
19899 generates 64-bit code when you select a 64-bit architecture, but you
19900 can use @option{-mgp32} to get 32-bit code instead.
19902 For information about the O64 ABI, see
19903 @uref{http://gcc.gnu.org/@/projects/@/mipso64-abi.html}.
19905 GCC supports a variant of the o32 ABI in which floating-point registers
19906 are 64 rather than 32 bits wide. You can select this combination with
19907 @option{-mabi=32} @option{-mfp64}. This ABI relies on the @code{mthc1}
19908 and @code{mfhc1} instructions and is therefore only supported for
19909 MIPS32R2, MIPS32R3 and MIPS32R5 processors.
19911 The register assignments for arguments and return values remain the
19912 same, but each scalar value is passed in a single 64-bit register
19913 rather than a pair of 32-bit registers. For example, scalar
19914 floating-point values are returned in @samp{$f0} only, not a
19915 @samp{$f0}/@samp{$f1} pair. The set of call-saved registers also
19916 remains the same in that the even-numbered double-precision registers
19919 Two additional variants of the o32 ABI are supported to enable
19920 a transition from 32-bit to 64-bit registers. These are FPXX
19921 (@option{-mfpxx}) and FP64A (@option{-mfp64} @option{-mno-odd-spreg}).
19922 The FPXX extension mandates that all code must execute correctly
19923 when run using 32-bit or 64-bit registers. The code can be interlinked
19924 with either FP32 or FP64, but not both.
19925 The FP64A extension is similar to the FP64 extension but forbids the
19926 use of odd-numbered single-precision registers. This can be used
19927 in conjunction with the @code{FRE} mode of FPUs in MIPS32R5
19928 processors and allows both FP32 and FP64A code to interlink and
19929 run in the same process without changing FPU modes.
19932 @itemx -mno-abicalls
19934 @opindex mno-abicalls
19935 Generate (do not generate) code that is suitable for SVR4-style
19936 dynamic objects. @option{-mabicalls} is the default for SVR4-based
19941 Generate (do not generate) code that is fully position-independent,
19942 and that can therefore be linked into shared libraries. This option
19943 only affects @option{-mabicalls}.
19945 All @option{-mabicalls} code has traditionally been position-independent,
19946 regardless of options like @option{-fPIC} and @option{-fpic}. However,
19947 as an extension, the GNU toolchain allows executables to use absolute
19948 accesses for locally-binding symbols. It can also use shorter GP
19949 initialization sequences and generate direct calls to locally-defined
19950 functions. This mode is selected by @option{-mno-shared}.
19952 @option{-mno-shared} depends on binutils 2.16 or higher and generates
19953 objects that can only be linked by the GNU linker. However, the option
19954 does not affect the ABI of the final executable; it only affects the ABI
19955 of relocatable objects. Using @option{-mno-shared} generally makes
19956 executables both smaller and quicker.
19958 @option{-mshared} is the default.
19964 Assume (do not assume) that the static and dynamic linkers
19965 support PLTs and copy relocations. This option only affects
19966 @option{-mno-shared -mabicalls}. For the n64 ABI, this option
19967 has no effect without @option{-msym32}.
19969 You can make @option{-mplt} the default by configuring
19970 GCC with @option{--with-mips-plt}. The default is
19971 @option{-mno-plt} otherwise.
19977 Lift (do not lift) the usual restrictions on the size of the global
19980 GCC normally uses a single instruction to load values from the GOT@.
19981 While this is relatively efficient, it only works if the GOT
19982 is smaller than about 64k. Anything larger causes the linker
19983 to report an error such as:
19985 @cindex relocation truncated to fit (MIPS)
19987 relocation truncated to fit: R_MIPS_GOT16 foobar
19990 If this happens, you should recompile your code with @option{-mxgot}.
19991 This works with very large GOTs, although the code is also
19992 less efficient, since it takes three instructions to fetch the
19993 value of a global symbol.
19995 Note that some linkers can create multiple GOTs. If you have such a
19996 linker, you should only need to use @option{-mxgot} when a single object
19997 file accesses more than 64k's worth of GOT entries. Very few do.
19999 These options have no effect unless GCC is generating position
20004 Assume that general-purpose registers are 32 bits wide.
20008 Assume that general-purpose registers are 64 bits wide.
20012 Assume that floating-point registers are 32 bits wide.
20016 Assume that floating-point registers are 64 bits wide.
20020 Do not assume the width of floating-point registers.
20023 @opindex mhard-float
20024 Use floating-point coprocessor instructions.
20027 @opindex msoft-float
20028 Do not use floating-point coprocessor instructions. Implement
20029 floating-point calculations using library calls instead.
20033 Equivalent to @option{-msoft-float}, but additionally asserts that the
20034 program being compiled does not perform any floating-point operations.
20035 This option is presently supported only by some bare-metal MIPS
20036 configurations, where it may select a special set of libraries
20037 that lack all floating-point support (including, for example, the
20038 floating-point @code{printf} formats).
20039 If code compiled with @option{-mno-float} accidentally contains
20040 floating-point operations, it is likely to suffer a link-time
20041 or run-time failure.
20043 @item -msingle-float
20044 @opindex msingle-float
20045 Assume that the floating-point coprocessor only supports single-precision
20048 @item -mdouble-float
20049 @opindex mdouble-float
20050 Assume that the floating-point coprocessor supports double-precision
20051 operations. This is the default.
20054 @itemx -mno-odd-spreg
20055 @opindex modd-spreg
20056 @opindex mno-odd-spreg
20057 Enable the use of odd-numbered single-precision floating-point registers
20058 for the o32 ABI. This is the default for processors that are known to
20059 support these registers. When using the o32 FPXX ABI, @option{-mno-odd-spreg}
20063 @itemx -mabs=legacy
20065 @opindex mabs=legacy
20066 These options control the treatment of the special not-a-number (NaN)
20067 IEEE 754 floating-point data with the @code{abs.@i{fmt}} and
20068 @code{neg.@i{fmt}} machine instructions.
20070 By default or when @option{-mabs=legacy} is used the legacy
20071 treatment is selected. In this case these instructions are considered
20072 arithmetic and avoided where correct operation is required and the
20073 input operand might be a NaN. A longer sequence of instructions that
20074 manipulate the sign bit of floating-point datum manually is used
20075 instead unless the @option{-ffinite-math-only} option has also been
20078 The @option{-mabs=2008} option selects the IEEE 754-2008 treatment. In
20079 this case these instructions are considered non-arithmetic and therefore
20080 operating correctly in all cases, including in particular where the
20081 input operand is a NaN. These instructions are therefore always used
20082 for the respective operations.
20085 @itemx -mnan=legacy
20087 @opindex mnan=legacy
20088 These options control the encoding of the special not-a-number (NaN)
20089 IEEE 754 floating-point data.
20091 The @option{-mnan=legacy} option selects the legacy encoding. In this
20092 case quiet NaNs (qNaNs) are denoted by the first bit of their trailing
20093 significand field being 0, whereas signaling NaNs (sNaNs) are denoted
20094 by the first bit of their trailing significand field being 1.
20096 The @option{-mnan=2008} option selects the IEEE 754-2008 encoding. In
20097 this case qNaNs are denoted by the first bit of their trailing
20098 significand field being 1, whereas sNaNs are denoted by the first bit of
20099 their trailing significand field being 0.
20101 The default is @option{-mnan=legacy} unless GCC has been configured with
20102 @option{--with-nan=2008}.
20108 Use (do not use) @samp{ll}, @samp{sc}, and @samp{sync} instructions to
20109 implement atomic memory built-in functions. When neither option is
20110 specified, GCC uses the instructions if the target architecture
20113 @option{-mllsc} is useful if the runtime environment can emulate the
20114 instructions and @option{-mno-llsc} can be useful when compiling for
20115 nonstandard ISAs. You can make either option the default by
20116 configuring GCC with @option{--with-llsc} and @option{--without-llsc}
20117 respectively. @option{--with-llsc} is the default for some
20118 configurations; see the installation documentation for details.
20124 Use (do not use) revision 1 of the MIPS DSP ASE@.
20125 @xref{MIPS DSP Built-in Functions}. This option defines the
20126 preprocessor macro @code{__mips_dsp}. It also defines
20127 @code{__mips_dsp_rev} to 1.
20133 Use (do not use) revision 2 of the MIPS DSP ASE@.
20134 @xref{MIPS DSP Built-in Functions}. This option defines the
20135 preprocessor macros @code{__mips_dsp} and @code{__mips_dspr2}.
20136 It also defines @code{__mips_dsp_rev} to 2.
20139 @itemx -mno-smartmips
20140 @opindex msmartmips
20141 @opindex mno-smartmips
20142 Use (do not use) the MIPS SmartMIPS ASE.
20144 @item -mpaired-single
20145 @itemx -mno-paired-single
20146 @opindex mpaired-single
20147 @opindex mno-paired-single
20148 Use (do not use) paired-single floating-point instructions.
20149 @xref{MIPS Paired-Single Support}. This option requires
20150 hardware floating-point support to be enabled.
20156 Use (do not use) MIPS Digital Media Extension instructions.
20157 This option can only be used when generating 64-bit code and requires
20158 hardware floating-point support to be enabled.
20163 @opindex mno-mips3d
20164 Use (do not use) the MIPS-3D ASE@. @xref{MIPS-3D Built-in Functions}.
20165 The option @option{-mips3d} implies @option{-mpaired-single}.
20168 @itemx -mno-micromips
20169 @opindex mmicromips
20170 @opindex mno-mmicromips
20171 Generate (do not generate) microMIPS code.
20173 MicroMIPS code generation can also be controlled on a per-function basis
20174 by means of @code{micromips} and @code{nomicromips} attributes.
20175 @xref{Function Attributes}, for more information.
20181 Use (do not use) MT Multithreading instructions.
20187 Use (do not use) the MIPS MCU ASE instructions.
20193 Use (do not use) the MIPS Enhanced Virtual Addressing instructions.
20199 Use (do not use) the MIPS Virtualization (VZ) instructions.
20205 Use (do not use) the MIPS eXtended Physical Address (XPA) instructions.
20209 Force @code{long} types to be 64 bits wide. See @option{-mlong32} for
20210 an explanation of the default and the way that the pointer size is
20215 Force @code{long}, @code{int}, and pointer types to be 32 bits wide.
20217 The default size of @code{int}s, @code{long}s and pointers depends on
20218 the ABI@. All the supported ABIs use 32-bit @code{int}s. The n64 ABI
20219 uses 64-bit @code{long}s, as does the 64-bit EABI; the others use
20220 32-bit @code{long}s. Pointers are the same size as @code{long}s,
20221 or the same size as integer registers, whichever is smaller.
20227 Assume (do not assume) that all symbols have 32-bit values, regardless
20228 of the selected ABI@. This option is useful in combination with
20229 @option{-mabi=64} and @option{-mno-abicalls} because it allows GCC
20230 to generate shorter and faster references to symbolic addresses.
20234 Put definitions of externally-visible data in a small data section
20235 if that data is no bigger than @var{num} bytes. GCC can then generate
20236 more efficient accesses to the data; see @option{-mgpopt} for details.
20238 The default @option{-G} option depends on the configuration.
20240 @item -mlocal-sdata
20241 @itemx -mno-local-sdata
20242 @opindex mlocal-sdata
20243 @opindex mno-local-sdata
20244 Extend (do not extend) the @option{-G} behavior to local data too,
20245 such as to static variables in C@. @option{-mlocal-sdata} is the
20246 default for all configurations.
20248 If the linker complains that an application is using too much small data,
20249 you might want to try rebuilding the less performance-critical parts with
20250 @option{-mno-local-sdata}. You might also want to build large
20251 libraries with @option{-mno-local-sdata}, so that the libraries leave
20252 more room for the main program.
20254 @item -mextern-sdata
20255 @itemx -mno-extern-sdata
20256 @opindex mextern-sdata
20257 @opindex mno-extern-sdata
20258 Assume (do not assume) that externally-defined data is in
20259 a small data section if the size of that data is within the @option{-G} limit.
20260 @option{-mextern-sdata} is the default for all configurations.
20262 If you compile a module @var{Mod} with @option{-mextern-sdata} @option{-G
20263 @var{num}} @option{-mgpopt}, and @var{Mod} references a variable @var{Var}
20264 that is no bigger than @var{num} bytes, you must make sure that @var{Var}
20265 is placed in a small data section. If @var{Var} is defined by another
20266 module, you must either compile that module with a high-enough
20267 @option{-G} setting or attach a @code{section} attribute to @var{Var}'s
20268 definition. If @var{Var} is common, you must link the application
20269 with a high-enough @option{-G} setting.
20271 The easiest way of satisfying these restrictions is to compile
20272 and link every module with the same @option{-G} option. However,
20273 you may wish to build a library that supports several different
20274 small data limits. You can do this by compiling the library with
20275 the highest supported @option{-G} setting and additionally using
20276 @option{-mno-extern-sdata} to stop the library from making assumptions
20277 about externally-defined data.
20283 Use (do not use) GP-relative accesses for symbols that are known to be
20284 in a small data section; see @option{-G}, @option{-mlocal-sdata} and
20285 @option{-mextern-sdata}. @option{-mgpopt} is the default for all
20288 @option{-mno-gpopt} is useful for cases where the @code{$gp} register
20289 might not hold the value of @code{_gp}. For example, if the code is
20290 part of a library that might be used in a boot monitor, programs that
20291 call boot monitor routines pass an unknown value in @code{$gp}.
20292 (In such situations, the boot monitor itself is usually compiled
20293 with @option{-G0}.)
20295 @option{-mno-gpopt} implies @option{-mno-local-sdata} and
20296 @option{-mno-extern-sdata}.
20298 @item -membedded-data
20299 @itemx -mno-embedded-data
20300 @opindex membedded-data
20301 @opindex mno-embedded-data
20302 Allocate variables to the read-only data section first if possible, then
20303 next in the small data section if possible, otherwise in data. This gives
20304 slightly slower code than the default, but reduces the amount of RAM required
20305 when executing, and thus may be preferred for some embedded systems.
20307 @item -muninit-const-in-rodata
20308 @itemx -mno-uninit-const-in-rodata
20309 @opindex muninit-const-in-rodata
20310 @opindex mno-uninit-const-in-rodata
20311 Put uninitialized @code{const} variables in the read-only data section.
20312 This option is only meaningful in conjunction with @option{-membedded-data}.
20314 @item -mcode-readable=@var{setting}
20315 @opindex mcode-readable
20316 Specify whether GCC may generate code that reads from executable sections.
20317 There are three possible settings:
20320 @item -mcode-readable=yes
20321 Instructions may freely access executable sections. This is the
20324 @item -mcode-readable=pcrel
20325 MIPS16 PC-relative load instructions can access executable sections,
20326 but other instructions must not do so. This option is useful on 4KSc
20327 and 4KSd processors when the code TLBs have the Read Inhibit bit set.
20328 It is also useful on processors that can be configured to have a dual
20329 instruction/data SRAM interface and that, like the M4K, automatically
20330 redirect PC-relative loads to the instruction RAM.
20332 @item -mcode-readable=no
20333 Instructions must not access executable sections. This option can be
20334 useful on targets that are configured to have a dual instruction/data
20335 SRAM interface but that (unlike the M4K) do not automatically redirect
20336 PC-relative loads to the instruction RAM.
20339 @item -msplit-addresses
20340 @itemx -mno-split-addresses
20341 @opindex msplit-addresses
20342 @opindex mno-split-addresses
20343 Enable (disable) use of the @code{%hi()} and @code{%lo()} assembler
20344 relocation operators. This option has been superseded by
20345 @option{-mexplicit-relocs} but is retained for backwards compatibility.
20347 @item -mexplicit-relocs
20348 @itemx -mno-explicit-relocs
20349 @opindex mexplicit-relocs
20350 @opindex mno-explicit-relocs
20351 Use (do not use) assembler relocation operators when dealing with symbolic
20352 addresses. The alternative, selected by @option{-mno-explicit-relocs},
20353 is to use assembler macros instead.
20355 @option{-mexplicit-relocs} is the default if GCC was configured
20356 to use an assembler that supports relocation operators.
20358 @item -mcheck-zero-division
20359 @itemx -mno-check-zero-division
20360 @opindex mcheck-zero-division
20361 @opindex mno-check-zero-division
20362 Trap (do not trap) on integer division by zero.
20364 The default is @option{-mcheck-zero-division}.
20366 @item -mdivide-traps
20367 @itemx -mdivide-breaks
20368 @opindex mdivide-traps
20369 @opindex mdivide-breaks
20370 MIPS systems check for division by zero by generating either a
20371 conditional trap or a break instruction. Using traps results in
20372 smaller code, but is only supported on MIPS II and later. Also, some
20373 versions of the Linux kernel have a bug that prevents trap from
20374 generating the proper signal (@code{SIGFPE}). Use @option{-mdivide-traps} to
20375 allow conditional traps on architectures that support them and
20376 @option{-mdivide-breaks} to force the use of breaks.
20378 The default is usually @option{-mdivide-traps}, but this can be
20379 overridden at configure time using @option{--with-divide=breaks}.
20380 Divide-by-zero checks can be completely disabled using
20381 @option{-mno-check-zero-division}.
20383 @item -mload-store-pairs
20384 @itemx -mno-load-store-pairs
20385 @opindex mload-store-pairs
20386 @opindex mno-load-store-pairs
20387 Enable (disable) an optimization that pairs consecutive load or store
20388 instructions to enable load/store bonding. This option is enabled by
20389 default but only takes effect when the selected architecture is known
20390 to support bonding.
20395 @opindex mno-memcpy
20396 Force (do not force) the use of @code{memcpy} for non-trivial block
20397 moves. The default is @option{-mno-memcpy}, which allows GCC to inline
20398 most constant-sized copies.
20401 @itemx -mno-long-calls
20402 @opindex mlong-calls
20403 @opindex mno-long-calls
20404 Disable (do not disable) use of the @code{jal} instruction. Calling
20405 functions using @code{jal} is more efficient but requires the caller
20406 and callee to be in the same 256 megabyte segment.
20408 This option has no effect on abicalls code. The default is
20409 @option{-mno-long-calls}.
20415 Enable (disable) use of the @code{mad}, @code{madu} and @code{mul}
20416 instructions, as provided by the R4650 ISA@.
20422 Enable (disable) use of the @code{madd} and @code{msub} integer
20423 instructions. The default is @option{-mimadd} on architectures
20424 that support @code{madd} and @code{msub} except for the 74k
20425 architecture where it was found to generate slower code.
20428 @itemx -mno-fused-madd
20429 @opindex mfused-madd
20430 @opindex mno-fused-madd
20431 Enable (disable) use of the floating-point multiply-accumulate
20432 instructions, when they are available. The default is
20433 @option{-mfused-madd}.
20435 On the R8000 CPU when multiply-accumulate instructions are used,
20436 the intermediate product is calculated to infinite precision
20437 and is not subject to the FCSR Flush to Zero bit. This may be
20438 undesirable in some circumstances. On other processors the result
20439 is numerically identical to the equivalent computation using
20440 separate multiply, add, subtract and negate instructions.
20444 Tell the MIPS assembler to not run its preprocessor over user
20445 assembler files (with a @samp{.s} suffix) when assembling them.
20450 @opindex mno-fix-24k
20451 Work around the 24K E48 (lost data on stores during refill) errata.
20452 The workarounds are implemented by the assembler rather than by GCC@.
20455 @itemx -mno-fix-r4000
20456 @opindex mfix-r4000
20457 @opindex mno-fix-r4000
20458 Work around certain R4000 CPU errata:
20461 A double-word or a variable shift may give an incorrect result if executed
20462 immediately after starting an integer division.
20464 A double-word or a variable shift may give an incorrect result if executed
20465 while an integer multiplication is in progress.
20467 An integer division may give an incorrect result if started in a delay slot
20468 of a taken branch or a jump.
20472 @itemx -mno-fix-r4400
20473 @opindex mfix-r4400
20474 @opindex mno-fix-r4400
20475 Work around certain R4400 CPU errata:
20478 A double-word or a variable shift may give an incorrect result if executed
20479 immediately after starting an integer division.
20483 @itemx -mno-fix-r10000
20484 @opindex mfix-r10000
20485 @opindex mno-fix-r10000
20486 Work around certain R10000 errata:
20489 @code{ll}/@code{sc} sequences may not behave atomically on revisions
20490 prior to 3.0. They may deadlock on revisions 2.6 and earlier.
20493 This option can only be used if the target architecture supports
20494 branch-likely instructions. @option{-mfix-r10000} is the default when
20495 @option{-march=r10000} is used; @option{-mno-fix-r10000} is the default
20499 @itemx -mno-fix-rm7000
20500 @opindex mfix-rm7000
20501 Work around the RM7000 @code{dmult}/@code{dmultu} errata. The
20502 workarounds are implemented by the assembler rather than by GCC@.
20505 @itemx -mno-fix-vr4120
20506 @opindex mfix-vr4120
20507 Work around certain VR4120 errata:
20510 @code{dmultu} does not always produce the correct result.
20512 @code{div} and @code{ddiv} do not always produce the correct result if one
20513 of the operands is negative.
20515 The workarounds for the division errata rely on special functions in
20516 @file{libgcc.a}. At present, these functions are only provided by
20517 the @code{mips64vr*-elf} configurations.
20519 Other VR4120 errata require a NOP to be inserted between certain pairs of
20520 instructions. These errata are handled by the assembler, not by GCC itself.
20523 @opindex mfix-vr4130
20524 Work around the VR4130 @code{mflo}/@code{mfhi} errata. The
20525 workarounds are implemented by the assembler rather than by GCC,
20526 although GCC avoids using @code{mflo} and @code{mfhi} if the
20527 VR4130 @code{macc}, @code{macchi}, @code{dmacc} and @code{dmacchi}
20528 instructions are available instead.
20531 @itemx -mno-fix-sb1
20533 Work around certain SB-1 CPU core errata.
20534 (This flag currently works around the SB-1 revision 2
20535 ``F1'' and ``F2'' floating-point errata.)
20537 @item -mr10k-cache-barrier=@var{setting}
20538 @opindex mr10k-cache-barrier
20539 Specify whether GCC should insert cache barriers to avoid the
20540 side-effects of speculation on R10K processors.
20542 In common with many processors, the R10K tries to predict the outcome
20543 of a conditional branch and speculatively executes instructions from
20544 the ``taken'' branch. It later aborts these instructions if the
20545 predicted outcome is wrong. However, on the R10K, even aborted
20546 instructions can have side effects.
20548 This problem only affects kernel stores and, depending on the system,
20549 kernel loads. As an example, a speculatively-executed store may load
20550 the target memory into cache and mark the cache line as dirty, even if
20551 the store itself is later aborted. If a DMA operation writes to the
20552 same area of memory before the ``dirty'' line is flushed, the cached
20553 data overwrites the DMA-ed data. See the R10K processor manual
20554 for a full description, including other potential problems.
20556 One workaround is to insert cache barrier instructions before every memory
20557 access that might be speculatively executed and that might have side
20558 effects even if aborted. @option{-mr10k-cache-barrier=@var{setting}}
20559 controls GCC's implementation of this workaround. It assumes that
20560 aborted accesses to any byte in the following regions does not have
20565 the memory occupied by the current function's stack frame;
20568 the memory occupied by an incoming stack argument;
20571 the memory occupied by an object with a link-time-constant address.
20574 It is the kernel's responsibility to ensure that speculative
20575 accesses to these regions are indeed safe.
20577 If the input program contains a function declaration such as:
20583 then the implementation of @code{foo} must allow @code{j foo} and
20584 @code{jal foo} to be executed speculatively. GCC honors this
20585 restriction for functions it compiles itself. It expects non-GCC
20586 functions (such as hand-written assembly code) to do the same.
20588 The option has three forms:
20591 @item -mr10k-cache-barrier=load-store
20592 Insert a cache barrier before a load or store that might be
20593 speculatively executed and that might have side effects even
20596 @item -mr10k-cache-barrier=store
20597 Insert a cache barrier before a store that might be speculatively
20598 executed and that might have side effects even if aborted.
20600 @item -mr10k-cache-barrier=none
20601 Disable the insertion of cache barriers. This is the default setting.
20604 @item -mflush-func=@var{func}
20605 @itemx -mno-flush-func
20606 @opindex mflush-func
20607 Specifies the function to call to flush the I and D caches, or to not
20608 call any such function. If called, the function must take the same
20609 arguments as the common @code{_flush_func}, that is, the address of the
20610 memory range for which the cache is being flushed, the size of the
20611 memory range, and the number 3 (to flush both caches). The default
20612 depends on the target GCC was configured for, but commonly is either
20613 @code{_flush_func} or @code{__cpu_flush}.
20615 @item mbranch-cost=@var{num}
20616 @opindex mbranch-cost
20617 Set the cost of branches to roughly @var{num} ``simple'' instructions.
20618 This cost is only a heuristic and is not guaranteed to produce
20619 consistent results across releases. A zero cost redundantly selects
20620 the default, which is based on the @option{-mtune} setting.
20622 @item -mbranch-likely
20623 @itemx -mno-branch-likely
20624 @opindex mbranch-likely
20625 @opindex mno-branch-likely
20626 Enable or disable use of Branch Likely instructions, regardless of the
20627 default for the selected architecture. By default, Branch Likely
20628 instructions may be generated if they are supported by the selected
20629 architecture. An exception is for the MIPS32 and MIPS64 architectures
20630 and processors that implement those architectures; for those, Branch
20631 Likely instructions are not be generated by default because the MIPS32
20632 and MIPS64 architectures specifically deprecate their use.
20634 @item -mcompact-branches=never
20635 @itemx -mcompact-branches=optimal
20636 @itemx -mcompact-branches=always
20637 @opindex mcompact-branches=never
20638 @opindex mcompact-branches=optimal
20639 @opindex mcompact-branches=always
20640 These options control which form of branches will be generated. The
20641 default is @option{-mcompact-branches=optimal}.
20643 The @option{-mcompact-branches=never} option ensures that compact branch
20644 instructions will never be generated.
20646 The @option{-mcompact-branches=always} option ensures that a compact
20647 branch instruction will be generated if available. If a compact branch
20648 instruction is not available, a delay slot form of the branch will be
20651 This option is supported from MIPS Release 6 onwards.
20653 The @option{-mcompact-branches=optimal} option will cause a delay slot
20654 branch to be used if one is available in the current ISA and the delay
20655 slot is successfully filled. If the delay slot is not filled, a compact
20656 branch will be chosen if one is available.
20658 @item -mfp-exceptions
20659 @itemx -mno-fp-exceptions
20660 @opindex mfp-exceptions
20661 Specifies whether FP exceptions are enabled. This affects how
20662 FP instructions are scheduled for some processors.
20663 The default is that FP exceptions are
20666 For instance, on the SB-1, if FP exceptions are disabled, and we are emitting
20667 64-bit code, then we can use both FP pipes. Otherwise, we can only use one
20670 @item -mvr4130-align
20671 @itemx -mno-vr4130-align
20672 @opindex mvr4130-align
20673 The VR4130 pipeline is two-way superscalar, but can only issue two
20674 instructions together if the first one is 8-byte aligned. When this
20675 option is enabled, GCC aligns pairs of instructions that it
20676 thinks should execute in parallel.
20678 This option only has an effect when optimizing for the VR4130.
20679 It normally makes code faster, but at the expense of making it bigger.
20680 It is enabled by default at optimization level @option{-O3}.
20685 Enable (disable) generation of @code{synci} instructions on
20686 architectures that support it. The @code{synci} instructions (if
20687 enabled) are generated when @code{__builtin___clear_cache} is
20690 This option defaults to @option{-mno-synci}, but the default can be
20691 overridden by configuring GCC with @option{--with-synci}.
20693 When compiling code for single processor systems, it is generally safe
20694 to use @code{synci}. However, on many multi-core (SMP) systems, it
20695 does not invalidate the instruction caches on all cores and may lead
20696 to undefined behavior.
20698 @item -mrelax-pic-calls
20699 @itemx -mno-relax-pic-calls
20700 @opindex mrelax-pic-calls
20701 Try to turn PIC calls that are normally dispatched via register
20702 @code{$25} into direct calls. This is only possible if the linker can
20703 resolve the destination at link time and if the destination is within
20704 range for a direct call.
20706 @option{-mrelax-pic-calls} is the default if GCC was configured to use
20707 an assembler and a linker that support the @code{.reloc} assembly
20708 directive and @option{-mexplicit-relocs} is in effect. With
20709 @option{-mno-explicit-relocs}, this optimization can be performed by the
20710 assembler and the linker alone without help from the compiler.
20712 @item -mmcount-ra-address
20713 @itemx -mno-mcount-ra-address
20714 @opindex mmcount-ra-address
20715 @opindex mno-mcount-ra-address
20716 Emit (do not emit) code that allows @code{_mcount} to modify the
20717 calling function's return address. When enabled, this option extends
20718 the usual @code{_mcount} interface with a new @var{ra-address}
20719 parameter, which has type @code{intptr_t *} and is passed in register
20720 @code{$12}. @code{_mcount} can then modify the return address by
20721 doing both of the following:
20724 Returning the new address in register @code{$31}.
20726 Storing the new address in @code{*@var{ra-address}},
20727 if @var{ra-address} is nonnull.
20730 The default is @option{-mno-mcount-ra-address}.
20732 @item -mframe-header-opt
20733 @itemx -mno-frame-header-opt
20734 @opindex mframe-header-opt
20735 Enable (disable) frame header optimization in the o32 ABI. When using the
20736 o32 ABI, calling functions will allocate 16 bytes on the stack for the called
20737 function to write out register arguments. When enabled, this optimization
20738 will suppress the allocation of the frame header if it can be determined that
20741 This optimization is off by default at all optimization levels.
20744 @itemx -mno-lxc1-sxc1
20745 @opindex mlxc1-sxc1
20746 When applicable, enable (disable) the generation of @code{lwxc1},
20747 @code{swxc1}, @code{ldxc1}, @code{sdxc1} instructions. Enabled by default.
20752 When applicable, enable (disable) the generation of 4-operand @code{madd.s},
20753 @code{madd.d} and related instructions. Enabled by default.
20758 @subsection MMIX Options
20759 @cindex MMIX Options
20761 These options are defined for the MMIX:
20765 @itemx -mno-libfuncs
20767 @opindex mno-libfuncs
20768 Specify that intrinsic library functions are being compiled, passing all
20769 values in registers, no matter the size.
20772 @itemx -mno-epsilon
20774 @opindex mno-epsilon
20775 Generate floating-point comparison instructions that compare with respect
20776 to the @code{rE} epsilon register.
20778 @item -mabi=mmixware
20780 @opindex mabi=mmixware
20782 Generate code that passes function parameters and return values that (in
20783 the called function) are seen as registers @code{$0} and up, as opposed to
20784 the GNU ABI which uses global registers @code{$231} and up.
20786 @item -mzero-extend
20787 @itemx -mno-zero-extend
20788 @opindex mzero-extend
20789 @opindex mno-zero-extend
20790 When reading data from memory in sizes shorter than 64 bits, use (do not
20791 use) zero-extending load instructions by default, rather than
20792 sign-extending ones.
20795 @itemx -mno-knuthdiv
20797 @opindex mno-knuthdiv
20798 Make the result of a division yielding a remainder have the same sign as
20799 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
20800 remainder follows the sign of the dividend. Both methods are
20801 arithmetically valid, the latter being almost exclusively used.
20803 @item -mtoplevel-symbols
20804 @itemx -mno-toplevel-symbols
20805 @opindex mtoplevel-symbols
20806 @opindex mno-toplevel-symbols
20807 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
20808 code can be used with the @code{PREFIX} assembly directive.
20812 Generate an executable in the ELF format, rather than the default
20813 @samp{mmo} format used by the @command{mmix} simulator.
20815 @item -mbranch-predict
20816 @itemx -mno-branch-predict
20817 @opindex mbranch-predict
20818 @opindex mno-branch-predict
20819 Use (do not use) the probable-branch instructions, when static branch
20820 prediction indicates a probable branch.
20822 @item -mbase-addresses
20823 @itemx -mno-base-addresses
20824 @opindex mbase-addresses
20825 @opindex mno-base-addresses
20826 Generate (do not generate) code that uses @emph{base addresses}. Using a
20827 base address automatically generates a request (handled by the assembler
20828 and the linker) for a constant to be set up in a global register. The
20829 register is used for one or more base address requests within the range 0
20830 to 255 from the value held in the register. The generally leads to short
20831 and fast code, but the number of different data items that can be
20832 addressed is limited. This means that a program that uses lots of static
20833 data may require @option{-mno-base-addresses}.
20835 @item -msingle-exit
20836 @itemx -mno-single-exit
20837 @opindex msingle-exit
20838 @opindex mno-single-exit
20839 Force (do not force) generated code to have a single exit point in each
20843 @node MN10300 Options
20844 @subsection MN10300 Options
20845 @cindex MN10300 options
20847 These @option{-m} options are defined for Matsushita MN10300 architectures:
20852 Generate code to avoid bugs in the multiply instructions for the MN10300
20853 processors. This is the default.
20855 @item -mno-mult-bug
20856 @opindex mno-mult-bug
20857 Do not generate code to avoid bugs in the multiply instructions for the
20858 MN10300 processors.
20862 Generate code using features specific to the AM33 processor.
20866 Do not generate code using features specific to the AM33 processor. This
20871 Generate code using features specific to the AM33/2.0 processor.
20875 Generate code using features specific to the AM34 processor.
20877 @item -mtune=@var{cpu-type}
20879 Use the timing characteristics of the indicated CPU type when
20880 scheduling instructions. This does not change the targeted processor
20881 type. The CPU type must be one of @samp{mn10300}, @samp{am33},
20882 @samp{am33-2} or @samp{am34}.
20884 @item -mreturn-pointer-on-d0
20885 @opindex mreturn-pointer-on-d0
20886 When generating a function that returns a pointer, return the pointer
20887 in both @code{a0} and @code{d0}. Otherwise, the pointer is returned
20888 only in @code{a0}, and attempts to call such functions without a prototype
20889 result in errors. Note that this option is on by default; use
20890 @option{-mno-return-pointer-on-d0} to disable it.
20894 Do not link in the C run-time initialization object file.
20898 Indicate to the linker that it should perform a relaxation optimization pass
20899 to shorten branches, calls and absolute memory addresses. This option only
20900 has an effect when used on the command line for the final link step.
20902 This option makes symbolic debugging impossible.
20906 Allow the compiler to generate @emph{Long Instruction Word}
20907 instructions if the target is the @samp{AM33} or later. This is the
20908 default. This option defines the preprocessor macro @code{__LIW__}.
20912 Do not allow the compiler to generate @emph{Long Instruction Word}
20913 instructions. This option defines the preprocessor macro
20918 Allow the compiler to generate the @emph{SETLB} and @emph{Lcc}
20919 instructions if the target is the @samp{AM33} or later. This is the
20920 default. This option defines the preprocessor macro @code{__SETLB__}.
20924 Do not allow the compiler to generate @emph{SETLB} or @emph{Lcc}
20925 instructions. This option defines the preprocessor macro
20926 @code{__NO_SETLB__}.
20930 @node Moxie Options
20931 @subsection Moxie Options
20932 @cindex Moxie Options
20938 Generate big-endian code. This is the default for @samp{moxie-*-*}
20943 Generate little-endian code.
20947 Generate mul.x and umul.x instructions. This is the default for
20948 @samp{moxiebox-*-*} configurations.
20952 Do not link in the C run-time initialization object file.
20956 @node MSP430 Options
20957 @subsection MSP430 Options
20958 @cindex MSP430 Options
20960 These options are defined for the MSP430:
20966 Force assembly output to always use hex constants. Normally such
20967 constants are signed decimals, but this option is available for
20968 testsuite and/or aesthetic purposes.
20972 Select the MCU to target. This is used to create a C preprocessor
20973 symbol based upon the MCU name, converted to upper case and pre- and
20974 post-fixed with @samp{__}. This in turn is used by the
20975 @file{msp430.h} header file to select an MCU-specific supplementary
20978 The option also sets the ISA to use. If the MCU name is one that is
20979 known to only support the 430 ISA then that is selected, otherwise the
20980 430X ISA is selected. A generic MCU name of @samp{msp430} can also be
20981 used to select the 430 ISA. Similarly the generic @samp{msp430x} MCU
20982 name selects the 430X ISA.
20984 In addition an MCU-specific linker script is added to the linker
20985 command line. The script's name is the name of the MCU with
20986 @file{.ld} appended. Thus specifying @option{-mmcu=xxx} on the @command{gcc}
20987 command line defines the C preprocessor symbol @code{__XXX__} and
20988 cause the linker to search for a script called @file{xxx.ld}.
20990 This option is also passed on to the assembler.
20993 @itemx -mno-warn-mcu
20995 @opindex mno-warn-mcu
20996 This option enables or disables warnings about conflicts between the
20997 MCU name specified by the @option{-mmcu} option and the ISA set by the
20998 @option{-mcpu} option and/or the hardware multiply support set by the
20999 @option{-mhwmult} option. It also toggles warnings about unrecognized
21000 MCU names. This option is on by default.
21004 Specifies the ISA to use. Accepted values are @samp{msp430},
21005 @samp{msp430x} and @samp{msp430xv2}. This option is deprecated. The
21006 @option{-mmcu=} option should be used to select the ISA.
21010 Link to the simulator runtime libraries and linker script. Overrides
21011 any scripts that would be selected by the @option{-mmcu=} option.
21015 Use large-model addressing (20-bit pointers, 32-bit @code{size_t}).
21019 Use small-model addressing (16-bit pointers, 16-bit @code{size_t}).
21023 This option is passed to the assembler and linker, and allows the
21024 linker to perform certain optimizations that cannot be done until
21029 Describes the type of hardware multiply supported by the target.
21030 Accepted values are @samp{none} for no hardware multiply, @samp{16bit}
21031 for the original 16-bit-only multiply supported by early MCUs.
21032 @samp{32bit} for the 16/32-bit multiply supported by later MCUs and
21033 @samp{f5series} for the 16/32-bit multiply supported by F5-series MCUs.
21034 A value of @samp{auto} can also be given. This tells GCC to deduce
21035 the hardware multiply support based upon the MCU name provided by the
21036 @option{-mmcu} option. If no @option{-mmcu} option is specified or if
21037 the MCU name is not recognized then no hardware multiply support is
21038 assumed. @code{auto} is the default setting.
21040 Hardware multiplies are normally performed by calling a library
21041 routine. This saves space in the generated code. When compiling at
21042 @option{-O3} or higher however the hardware multiplier is invoked
21043 inline. This makes for bigger, but faster code.
21045 The hardware multiply routines disable interrupts whilst running and
21046 restore the previous interrupt state when they finish. This makes
21047 them safe to use inside interrupt handlers as well as in normal code.
21051 Enable the use of a minimum runtime environment - no static
21052 initializers or constructors. This is intended for memory-constrained
21053 devices. The compiler includes special symbols in some objects
21054 that tell the linker and runtime which code fragments are required.
21056 @item -mcode-region=
21057 @itemx -mdata-region=
21058 @opindex mcode-region
21059 @opindex mdata-region
21060 These options tell the compiler where to place functions and data that
21061 do not have one of the @code{lower}, @code{upper}, @code{either} or
21062 @code{section} attributes. Possible values are @code{lower},
21063 @code{upper}, @code{either} or @code{any}. The first three behave
21064 like the corresponding attribute. The fourth possible value -
21065 @code{any} - is the default. It leaves placement entirely up to the
21066 linker script and how it assigns the standard sections
21067 (@code{.text}, @code{.data}, etc) to the memory regions.
21069 @item -msilicon-errata=
21070 @opindex msilicon-errata
21071 This option passes on a request to assembler to enable the fixes for
21072 the named silicon errata.
21074 @item -msilicon-errata-warn=
21075 @opindex msilicon-errata-warn
21076 This option passes on a request to the assembler to enable warning
21077 messages when a silicon errata might need to be applied.
21081 @node NDS32 Options
21082 @subsection NDS32 Options
21083 @cindex NDS32 Options
21085 These options are defined for NDS32 implementations:
21090 @opindex mbig-endian
21091 Generate code in big-endian mode.
21093 @item -mlittle-endian
21094 @opindex mlittle-endian
21095 Generate code in little-endian mode.
21097 @item -mreduced-regs
21098 @opindex mreduced-regs
21099 Use reduced-set registers for register allocation.
21102 @opindex mfull-regs
21103 Use full-set registers for register allocation.
21107 Generate conditional move instructions.
21111 Do not generate conditional move instructions.
21115 Generate performance extension instructions.
21117 @item -mno-perf-ext
21118 @opindex mno-perf-ext
21119 Do not generate performance extension instructions.
21123 Generate v3 push25/pop25 instructions.
21126 @opindex mno-v3push
21127 Do not generate v3 push25/pop25 instructions.
21131 Generate 16-bit instructions.
21134 @opindex mno-16-bit
21135 Do not generate 16-bit instructions.
21137 @item -misr-vector-size=@var{num}
21138 @opindex misr-vector-size
21139 Specify the size of each interrupt vector, which must be 4 or 16.
21141 @item -mcache-block-size=@var{num}
21142 @opindex mcache-block-size
21143 Specify the size of each cache block,
21144 which must be a power of 2 between 4 and 512.
21146 @item -march=@var{arch}
21148 Specify the name of the target architecture.
21150 @item -mcmodel=@var{code-model}
21152 Set the code model to one of
21155 All the data and read-only data segments must be within 512KB addressing space.
21156 The text segment must be within 16MB addressing space.
21157 @item @samp{medium}
21158 The data segment must be within 512KB while the read-only data segment can be
21159 within 4GB addressing space. The text segment should be still within 16MB
21162 All the text and data segments can be within 4GB addressing space.
21166 @opindex mctor-dtor
21167 Enable constructor/destructor feature.
21171 Guide linker to relax instructions.
21175 @node Nios II Options
21176 @subsection Nios II Options
21177 @cindex Nios II options
21178 @cindex Altera Nios II options
21180 These are the options defined for the Altera Nios II processor.
21186 @cindex smaller data references
21187 Put global and static objects less than or equal to @var{num} bytes
21188 into the small data or BSS sections instead of the normal data or BSS
21189 sections. The default value of @var{num} is 8.
21191 @item -mgpopt=@var{option}
21196 Generate (do not generate) GP-relative accesses. The following
21197 @var{option} names are recognized:
21202 Do not generate GP-relative accesses.
21205 Generate GP-relative accesses for small data objects that are not
21206 external, weak, or uninitialized common symbols.
21207 Also use GP-relative addressing for objects that
21208 have been explicitly placed in a small data section via a @code{section}
21212 As for @samp{local}, but also generate GP-relative accesses for
21213 small data objects that are external, weak, or common. If you use this option,
21214 you must ensure that all parts of your program (including libraries) are
21215 compiled with the same @option{-G} setting.
21218 Generate GP-relative accesses for all data objects in the program. If you
21219 use this option, the entire data and BSS segments
21220 of your program must fit in 64K of memory and you must use an appropriate
21221 linker script to allocate them within the addressable range of the
21225 Generate GP-relative addresses for function pointers as well as data
21226 pointers. If you use this option, the entire text, data, and BSS segments
21227 of your program must fit in 64K of memory and you must use an appropriate
21228 linker script to allocate them within the addressable range of the
21233 @option{-mgpopt} is equivalent to @option{-mgpopt=local}, and
21234 @option{-mno-gpopt} is equivalent to @option{-mgpopt=none}.
21236 The default is @option{-mgpopt} except when @option{-fpic} or
21237 @option{-fPIC} is specified to generate position-independent code.
21238 Note that the Nios II ABI does not permit GP-relative accesses from
21241 You may need to specify @option{-mno-gpopt} explicitly when building
21242 programs that include large amounts of small data, including large
21243 GOT data sections. In this case, the 16-bit offset for GP-relative
21244 addressing may not be large enough to allow access to the entire
21245 small data section.
21247 @item -mgprel-sec=@var{regexp}
21248 @opindex mgprel-sec
21249 This option specifies additional section names that can be accessed via
21250 GP-relative addressing. It is most useful in conjunction with
21251 @code{section} attributes on variable declarations
21252 (@pxref{Common Variable Attributes}) and a custom linker script.
21253 The @var{regexp} is a POSIX Extended Regular Expression.
21255 This option does not affect the behavior of the @option{-G} option, and
21256 and the specified sections are in addition to the standard @code{.sdata}
21257 and @code{.sbss} small-data sections that are recognized by @option{-mgpopt}.
21259 @item -mr0rel-sec=@var{regexp}
21260 @opindex mr0rel-sec
21261 This option specifies names of sections that can be accessed via a
21262 16-bit offset from @code{r0}; that is, in the low 32K or high 32K
21263 of the 32-bit address space. It is most useful in conjunction with
21264 @code{section} attributes on variable declarations
21265 (@pxref{Common Variable Attributes}) and a custom linker script.
21266 The @var{regexp} is a POSIX Extended Regular Expression.
21268 In contrast to the use of GP-relative addressing for small data,
21269 zero-based addressing is never generated by default and there are no
21270 conventional section names used in standard linker scripts for sections
21271 in the low or high areas of memory.
21277 Generate little-endian (default) or big-endian (experimental) code,
21280 @item -march=@var{arch}
21282 This specifies the name of the target Nios II architecture. GCC uses this
21283 name to determine what kind of instructions it can emit when generating
21284 assembly code. Permissible names are: @samp{r1}, @samp{r2}.
21286 The preprocessor macro @code{__nios2_arch__} is available to programs,
21287 with value 1 or 2, indicating the targeted ISA level.
21289 @item -mbypass-cache
21290 @itemx -mno-bypass-cache
21291 @opindex mno-bypass-cache
21292 @opindex mbypass-cache
21293 Force all load and store instructions to always bypass cache by
21294 using I/O variants of the instructions. The default is not to
21297 @item -mno-cache-volatile
21298 @itemx -mcache-volatile
21299 @opindex mcache-volatile
21300 @opindex mno-cache-volatile
21301 Volatile memory access bypass the cache using the I/O variants of
21302 the load and store instructions. The default is not to bypass the cache.
21304 @item -mno-fast-sw-div
21305 @itemx -mfast-sw-div
21306 @opindex mno-fast-sw-div
21307 @opindex mfast-sw-div
21308 Do not use table-based fast divide for small numbers. The default
21309 is to use the fast divide at @option{-O3} and above.
21313 @itemx -mno-hw-mulx
21317 @opindex mno-hw-mul
21319 @opindex mno-hw-mulx
21321 @opindex mno-hw-div
21323 Enable or disable emitting @code{mul}, @code{mulx} and @code{div} family of
21324 instructions by the compiler. The default is to emit @code{mul}
21325 and not emit @code{div} and @code{mulx}.
21331 Enable or disable generation of Nios II R2 BMX (bit manipulation) and
21332 CDX (code density) instructions. Enabling these instructions also
21333 requires @option{-march=r2}. Since these instructions are optional
21334 extensions to the R2 architecture, the default is not to emit them.
21336 @item -mcustom-@var{insn}=@var{N}
21337 @itemx -mno-custom-@var{insn}
21338 @opindex mcustom-@var{insn}
21339 @opindex mno-custom-@var{insn}
21340 Each @option{-mcustom-@var{insn}=@var{N}} option enables use of a
21341 custom instruction with encoding @var{N} when generating code that uses
21342 @var{insn}. For example, @option{-mcustom-fadds=253} generates custom
21343 instruction 253 for single-precision floating-point add operations instead
21344 of the default behavior of using a library call.
21346 The following values of @var{insn} are supported. Except as otherwise
21347 noted, floating-point operations are expected to be implemented with
21348 normal IEEE 754 semantics and correspond directly to the C operators or the
21349 equivalent GCC built-in functions (@pxref{Other Builtins}).
21351 Single-precision floating point:
21354 @item @samp{fadds}, @samp{fsubs}, @samp{fdivs}, @samp{fmuls}
21355 Binary arithmetic operations.
21361 Unary absolute value.
21363 @item @samp{fcmpeqs}, @samp{fcmpges}, @samp{fcmpgts}, @samp{fcmples}, @samp{fcmplts}, @samp{fcmpnes}
21364 Comparison operations.
21366 @item @samp{fmins}, @samp{fmaxs}
21367 Floating-point minimum and maximum. These instructions are only
21368 generated if @option{-ffinite-math-only} is specified.
21370 @item @samp{fsqrts}
21371 Unary square root operation.
21373 @item @samp{fcoss}, @samp{fsins}, @samp{ftans}, @samp{fatans}, @samp{fexps}, @samp{flogs}
21374 Floating-point trigonometric and exponential functions. These instructions
21375 are only generated if @option{-funsafe-math-optimizations} is also specified.
21379 Double-precision floating point:
21382 @item @samp{faddd}, @samp{fsubd}, @samp{fdivd}, @samp{fmuld}
21383 Binary arithmetic operations.
21389 Unary absolute value.
21391 @item @samp{fcmpeqd}, @samp{fcmpged}, @samp{fcmpgtd}, @samp{fcmpled}, @samp{fcmpltd}, @samp{fcmpned}
21392 Comparison operations.
21394 @item @samp{fmind}, @samp{fmaxd}
21395 Double-precision minimum and maximum. These instructions are only
21396 generated if @option{-ffinite-math-only} is specified.
21398 @item @samp{fsqrtd}
21399 Unary square root operation.
21401 @item @samp{fcosd}, @samp{fsind}, @samp{ftand}, @samp{fatand}, @samp{fexpd}, @samp{flogd}
21402 Double-precision trigonometric and exponential functions. These instructions
21403 are only generated if @option{-funsafe-math-optimizations} is also specified.
21409 @item @samp{fextsd}
21410 Conversion from single precision to double precision.
21412 @item @samp{ftruncds}
21413 Conversion from double precision to single precision.
21415 @item @samp{fixsi}, @samp{fixsu}, @samp{fixdi}, @samp{fixdu}
21416 Conversion from floating point to signed or unsigned integer types, with
21417 truncation towards zero.
21420 Conversion from single-precision floating point to signed integer,
21421 rounding to the nearest integer and ties away from zero.
21422 This corresponds to the @code{__builtin_lroundf} function when
21423 @option{-fno-math-errno} is used.
21425 @item @samp{floatis}, @samp{floatus}, @samp{floatid}, @samp{floatud}
21426 Conversion from signed or unsigned integer types to floating-point types.
21430 In addition, all of the following transfer instructions for internal
21431 registers X and Y must be provided to use any of the double-precision
21432 floating-point instructions. Custom instructions taking two
21433 double-precision source operands expect the first operand in the
21434 64-bit register X. The other operand (or only operand of a unary
21435 operation) is given to the custom arithmetic instruction with the
21436 least significant half in source register @var{src1} and the most
21437 significant half in @var{src2}. A custom instruction that returns a
21438 double-precision result returns the most significant 32 bits in the
21439 destination register and the other half in 32-bit register Y.
21440 GCC automatically generates the necessary code sequences to write
21441 register X and/or read register Y when double-precision floating-point
21442 instructions are used.
21447 Write @var{src1} into the least significant half of X and @var{src2} into
21448 the most significant half of X.
21451 Write @var{src1} into Y.
21453 @item @samp{frdxhi}, @samp{frdxlo}
21454 Read the most or least (respectively) significant half of X and store it in
21458 Read the value of Y and store it into @var{dest}.
21461 Note that you can gain more local control over generation of Nios II custom
21462 instructions by using the @code{target("custom-@var{insn}=@var{N}")}
21463 and @code{target("no-custom-@var{insn}")} function attributes
21464 (@pxref{Function Attributes})
21465 or pragmas (@pxref{Function Specific Option Pragmas}).
21467 @item -mcustom-fpu-cfg=@var{name}
21468 @opindex mcustom-fpu-cfg
21470 This option enables a predefined, named set of custom instruction encodings
21471 (see @option{-mcustom-@var{insn}} above).
21472 Currently, the following sets are defined:
21474 @option{-mcustom-fpu-cfg=60-1} is equivalent to:
21475 @gccoptlist{-mcustom-fmuls=252 @gol
21476 -mcustom-fadds=253 @gol
21477 -mcustom-fsubs=254 @gol
21478 -fsingle-precision-constant}
21480 @option{-mcustom-fpu-cfg=60-2} is equivalent to:
21481 @gccoptlist{-mcustom-fmuls=252 @gol
21482 -mcustom-fadds=253 @gol
21483 -mcustom-fsubs=254 @gol
21484 -mcustom-fdivs=255 @gol
21485 -fsingle-precision-constant}
21487 @option{-mcustom-fpu-cfg=72-3} is equivalent to:
21488 @gccoptlist{-mcustom-floatus=243 @gol
21489 -mcustom-fixsi=244 @gol
21490 -mcustom-floatis=245 @gol
21491 -mcustom-fcmpgts=246 @gol
21492 -mcustom-fcmples=249 @gol
21493 -mcustom-fcmpeqs=250 @gol
21494 -mcustom-fcmpnes=251 @gol
21495 -mcustom-fmuls=252 @gol
21496 -mcustom-fadds=253 @gol
21497 -mcustom-fsubs=254 @gol
21498 -mcustom-fdivs=255 @gol
21499 -fsingle-precision-constant}
21501 Custom instruction assignments given by individual
21502 @option{-mcustom-@var{insn}=} options override those given by
21503 @option{-mcustom-fpu-cfg=}, regardless of the
21504 order of the options on the command line.
21506 Note that you can gain more local control over selection of a FPU
21507 configuration by using the @code{target("custom-fpu-cfg=@var{name}")}
21508 function attribute (@pxref{Function Attributes})
21509 or pragma (@pxref{Function Specific Option Pragmas}).
21513 These additional @samp{-m} options are available for the Altera Nios II
21514 ELF (bare-metal) target:
21520 Link with HAL BSP. This suppresses linking with the GCC-provided C runtime
21521 startup and termination code, and is typically used in conjunction with
21522 @option{-msys-crt0=} to specify the location of the alternate startup code
21523 provided by the HAL BSP.
21527 Link with a limited version of the C library, @option{-lsmallc}, rather than
21530 @item -msys-crt0=@var{startfile}
21532 @var{startfile} is the file name of the startfile (crt0) to use
21533 when linking. This option is only useful in conjunction with @option{-mhal}.
21535 @item -msys-lib=@var{systemlib}
21537 @var{systemlib} is the library name of the library that provides
21538 low-level system calls required by the C library,
21539 e.g. @code{read} and @code{write}.
21540 This option is typically used to link with a library provided by a HAL BSP.
21544 @node Nvidia PTX Options
21545 @subsection Nvidia PTX Options
21546 @cindex Nvidia PTX options
21547 @cindex nvptx options
21549 These options are defined for Nvidia PTX:
21557 Generate code for 32-bit or 64-bit ABI.
21560 @opindex mmainkernel
21561 Link in code for a __main kernel. This is for stand-alone instead of
21562 offloading execution.
21566 Apply partitioned execution optimizations. This is the default when any
21567 level of optimization is selected.
21570 @opindex msoft-stack
21571 Generate code that does not use @code{.local} memory
21572 directly for stack storage. Instead, a per-warp stack pointer is
21573 maintained explicitly. This enables variable-length stack allocation (with
21574 variable-length arrays or @code{alloca}), and when global memory is used for
21575 underlying storage, makes it possible to access automatic variables from other
21576 threads, or with atomic instructions. This code generation variant is used
21577 for OpenMP offloading, but the option is exposed on its own for the purpose
21578 of testing the compiler; to generate code suitable for linking into programs
21579 using OpenMP offloading, use option @option{-mgomp}.
21581 @item -muniform-simt
21582 @opindex muniform-simt
21583 Switch to code generation variant that allows to execute all threads in each
21584 warp, while maintaining memory state and side effects as if only one thread
21585 in each warp was active outside of OpenMP SIMD regions. All atomic operations
21586 and calls to runtime (malloc, free, vprintf) are conditionally executed (iff
21587 current lane index equals the master lane index), and the register being
21588 assigned is copied via a shuffle instruction from the master lane. Outside of
21589 SIMD regions lane 0 is the master; inside, each thread sees itself as the
21590 master. Shared memory array @code{int __nvptx_uni[]} stores all-zeros or
21591 all-ones bitmasks for each warp, indicating current mode (0 outside of SIMD
21592 regions). Each thread can bitwise-and the bitmask at position @code{tid.y}
21593 with current lane index to compute the master lane index.
21597 Generate code for use in OpenMP offloading: enables @option{-msoft-stack} and
21598 @option{-muniform-simt} options, and selects corresponding multilib variant.
21602 @node PDP-11 Options
21603 @subsection PDP-11 Options
21604 @cindex PDP-11 Options
21606 These options are defined for the PDP-11:
21611 Use hardware FPP floating point. This is the default. (FIS floating
21612 point on the PDP-11/40 is not supported.)
21615 @opindex msoft-float
21616 Do not use hardware floating point.
21620 Return floating-point results in ac0 (fr0 in Unix assembler syntax).
21624 Return floating-point results in memory. This is the default.
21628 Generate code for a PDP-11/40.
21632 Generate code for a PDP-11/45. This is the default.
21636 Generate code for a PDP-11/10.
21638 @item -mbcopy-builtin
21639 @opindex mbcopy-builtin
21640 Use inline @code{movmemhi} patterns for copying memory. This is the
21645 Do not use inline @code{movmemhi} patterns for copying memory.
21651 Use 16-bit @code{int}. This is the default.
21657 Use 32-bit @code{int}.
21660 @itemx -mno-float32
21662 @opindex mno-float32
21663 Use 64-bit @code{float}. This is the default.
21666 @itemx -mno-float64
21668 @opindex mno-float64
21669 Use 32-bit @code{float}.
21673 Use @code{abshi2} pattern. This is the default.
21677 Do not use @code{abshi2} pattern.
21679 @item -mbranch-expensive
21680 @opindex mbranch-expensive
21681 Pretend that branches are expensive. This is for experimenting with
21682 code generation only.
21684 @item -mbranch-cheap
21685 @opindex mbranch-cheap
21686 Do not pretend that branches are expensive. This is the default.
21690 Use Unix assembler syntax. This is the default when configured for
21691 @samp{pdp11-*-bsd}.
21695 Use DEC assembler syntax. This is the default when configured for any
21696 PDP-11 target other than @samp{pdp11-*-bsd}.
21699 @node picoChip Options
21700 @subsection picoChip Options
21701 @cindex picoChip options
21703 These @samp{-m} options are defined for picoChip implementations:
21707 @item -mae=@var{ae_type}
21709 Set the instruction set, register set, and instruction scheduling
21710 parameters for array element type @var{ae_type}. Supported values
21711 for @var{ae_type} are @samp{ANY}, @samp{MUL}, and @samp{MAC}.
21713 @option{-mae=ANY} selects a completely generic AE type. Code
21714 generated with this option runs on any of the other AE types. The
21715 code is not as efficient as it would be if compiled for a specific
21716 AE type, and some types of operation (e.g., multiplication) do not
21717 work properly on all types of AE.
21719 @option{-mae=MUL} selects a MUL AE type. This is the most useful AE type
21720 for compiled code, and is the default.
21722 @option{-mae=MAC} selects a DSP-style MAC AE. Code compiled with this
21723 option may suffer from poor performance of byte (char) manipulation,
21724 since the DSP AE does not provide hardware support for byte load/stores.
21726 @item -msymbol-as-address
21727 Enable the compiler to directly use a symbol name as an address in a
21728 load/store instruction, without first loading it into a
21729 register. Typically, the use of this option generates larger
21730 programs, which run faster than when the option isn't used. However, the
21731 results vary from program to program, so it is left as a user option,
21732 rather than being permanently enabled.
21734 @item -mno-inefficient-warnings
21735 Disables warnings about the generation of inefficient code. These
21736 warnings can be generated, for example, when compiling code that
21737 performs byte-level memory operations on the MAC AE type. The MAC AE has
21738 no hardware support for byte-level memory operations, so all byte
21739 load/stores must be synthesized from word load/store operations. This is
21740 inefficient and a warning is generated to indicate
21741 that you should rewrite the code to avoid byte operations, or to target
21742 an AE type that has the necessary hardware support. This option disables
21747 @node PowerPC Options
21748 @subsection PowerPC Options
21749 @cindex PowerPC options
21751 These are listed under @xref{RS/6000 and PowerPC Options}.
21753 @node RISC-V Options
21754 @subsection RISC-V Options
21755 @cindex RISC-V Options
21757 These command-line options are defined for RISC-V targets:
21760 @item -mbranch-cost=@var{n}
21761 @opindex mbranch-cost
21762 Set the cost of branches to roughly @var{n} instructions.
21767 Don't optimize block moves.
21772 When generating PIC code, allow the use of PLTs. Ignored for non-PIC.
21774 @item -mabi=@var{ABI-string}
21776 @item -mabi=@var{ABI-string}
21778 Specify integer and floating-point calling convention. @var{ABI-string}
21779 contains two parts: the size of integer types and the registers used for
21780 floating-point types. For example @samp{-march=rv64ifd -mabi=lp64d} means that
21781 @samp{long} and pointers are 64-bit (implicitly defining @samp{int} to be
21782 32-bit), and that floating-point values up to 64 bits wide are passed in F
21783 registers. Contrast this with @samp{-march=rv64ifd -mabi=lp64f}, which still
21784 allows the compiler to generate code that uses the F and D extensions but only
21785 allows floating-point values up to 32 bits long to be passed in registers; or
21786 @samp{-march=rv64ifd -mabi=lp64}, in which no floating-point arguments will be
21787 passed in registers.
21789 The default for this argument is system dependent, users who want a specific
21790 calling convention should specify one explicitly. The valid calling
21791 conventions are: @samp{ilp32}, @samp{ilp32f}, @samp{ilp32d}, @samp{lp64},
21792 @samp{lp64f}, and @samp{lp64d}. Some calling conventions are impossible to
21793 implement on some ISAs: for example, @samp{-march=rv32if -mabi=ilp32d} is
21794 invalid because the ABI requires 64-bit values be passed in F registers, but F
21795 registers are only 32 bits wide.
21800 Use hardware floating-point divide and square root instructions. This requires
21801 the F or D extensions for floating-point registers.
21806 Use hardware instructions for integer division. This requires the M extension.
21808 @item -march=@var{ISA-string}
21810 Generate code for given RISC-V ISA (e.g.@ @samp{rv64im}). ISA strings must be
21811 lower-case. Examples include @samp{rv64i}, @samp{rv32g}, and @samp{rv32imaf}.
21813 @item -mtune=@var{processor-string}
21815 Optimize the output for the given processor, specified by microarchitecture
21818 @item -msmall-data-limit=@var{n}
21819 @opindex msmall-data-limit
21820 Put global and static data smaller than @var{n} bytes into a special section
21823 @item -msave-restore
21824 @itemx -mno-save-restore
21825 @opindex msave-restore
21826 Use smaller but slower prologue and epilogue code.
21828 @item -mstrict-align
21829 @itemx -mno-strict-align
21830 @opindex mstrict-align
21831 Do not generate unaligned memory accesses.
21833 @item -mcmodel=medlow
21834 @opindex mcmodel=medlow
21835 Generate code for the medium-low code model. The program and its statically
21836 defined symbols must lie within a single 2 GiB address range and must lie
21837 between absolute addresses @minus{}2 GiB and +2 GiB. Programs can be
21838 statically or dynamically linked. This is the default code model.
21840 @item -mcmodel=medany
21841 @opindex mcmodel=medany
21842 Generate code for the medium-any code model. The program and its statically
21843 defined symbols must be within any single 2 GiB address range. Programs can be
21844 statically or dynamically linked.
21849 @subsection RL78 Options
21850 @cindex RL78 Options
21856 Links in additional target libraries to support operation within a
21865 Specifies the type of hardware multiplication and division support to
21866 be used. The simplest is @code{none}, which uses software for both
21867 multiplication and division. This is the default. The @code{g13}
21868 value is for the hardware multiply/divide peripheral found on the
21869 RL78/G13 (S2 core) targets. The @code{g14} value selects the use of
21870 the multiplication and division instructions supported by the RL78/G14
21871 (S3 core) parts. The value @code{rl78} is an alias for @code{g14} and
21872 the value @code{mg10} is an alias for @code{none}.
21874 In addition a C preprocessor macro is defined, based upon the setting
21875 of this option. Possible values are: @code{__RL78_MUL_NONE__},
21876 @code{__RL78_MUL_G13__} or @code{__RL78_MUL_G14__}.
21883 Specifies the RL78 core to target. The default is the G14 core, also
21884 known as an S3 core or just RL78. The G13 or S2 core does not have
21885 multiply or divide instructions, instead it uses a hardware peripheral
21886 for these operations. The G10 or S1 core does not have register
21887 banks, so it uses a different calling convention.
21889 If this option is set it also selects the type of hardware multiply
21890 support to use, unless this is overridden by an explicit
21891 @option{-mmul=none} option on the command line. Thus specifying
21892 @option{-mcpu=g13} enables the use of the G13 hardware multiply
21893 peripheral and specifying @option{-mcpu=g10} disables the use of
21894 hardware multiplications altogether.
21896 Note, although the RL78/G14 core is the default target, specifying
21897 @option{-mcpu=g14} or @option{-mcpu=rl78} on the command line does
21898 change the behavior of the toolchain since it also enables G14
21899 hardware multiply support. If these options are not specified on the
21900 command line then software multiplication routines will be used even
21901 though the code targets the RL78 core. This is for backwards
21902 compatibility with older toolchains which did not have hardware
21903 multiply and divide support.
21905 In addition a C preprocessor macro is defined, based upon the setting
21906 of this option. Possible values are: @code{__RL78_G10__},
21907 @code{__RL78_G13__} or @code{__RL78_G14__}.
21917 These are aliases for the corresponding @option{-mcpu=} option. They
21918 are provided for backwards compatibility.
21922 Allow the compiler to use all of the available registers. By default
21923 registers @code{r24..r31} are reserved for use in interrupt handlers.
21924 With this option enabled these registers can be used in ordinary
21927 @item -m64bit-doubles
21928 @itemx -m32bit-doubles
21929 @opindex m64bit-doubles
21930 @opindex m32bit-doubles
21931 Make the @code{double} data type be 64 bits (@option{-m64bit-doubles})
21932 or 32 bits (@option{-m32bit-doubles}) in size. The default is
21933 @option{-m32bit-doubles}.
21935 @item -msave-mduc-in-interrupts
21936 @item -mno-save-mduc-in-interrupts
21937 @opindex msave-mduc-in-interrupts
21938 @opindex mno-save-mduc-in-interrupts
21939 Specifies that interrupt handler functions should preserve the
21940 MDUC registers. This is only necessary if normal code might use
21941 the MDUC registers, for example because it performs multiplication
21942 and division operations. The default is to ignore the MDUC registers
21943 as this makes the interrupt handlers faster. The target option -mg13
21944 needs to be passed for this to work as this feature is only available
21945 on the G13 target (S2 core). The MDUC registers will only be saved
21946 if the interrupt handler performs a multiplication or division
21947 operation or it calls another function.
21951 @node RS/6000 and PowerPC Options
21952 @subsection IBM RS/6000 and PowerPC Options
21953 @cindex RS/6000 and PowerPC Options
21954 @cindex IBM RS/6000 and PowerPC Options
21956 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
21958 @item -mpowerpc-gpopt
21959 @itemx -mno-powerpc-gpopt
21960 @itemx -mpowerpc-gfxopt
21961 @itemx -mno-powerpc-gfxopt
21964 @itemx -mno-powerpc64
21968 @itemx -mno-popcntb
21970 @itemx -mno-popcntd
21979 @itemx -mno-hard-dfp
21980 @opindex mpowerpc-gpopt
21981 @opindex mno-powerpc-gpopt
21982 @opindex mpowerpc-gfxopt
21983 @opindex mno-powerpc-gfxopt
21984 @opindex mpowerpc64
21985 @opindex mno-powerpc64
21989 @opindex mno-popcntb
21991 @opindex mno-popcntd
21997 @opindex mno-mfpgpr
21999 @opindex mno-hard-dfp
22000 You use these options to specify which instructions are available on the
22001 processor you are using. The default value of these options is
22002 determined when configuring GCC@. Specifying the
22003 @option{-mcpu=@var{cpu_type}} overrides the specification of these
22004 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
22005 rather than the options listed above.
22007 Specifying @option{-mpowerpc-gpopt} allows
22008 GCC to use the optional PowerPC architecture instructions in the
22009 General Purpose group, including floating-point square root. Specifying
22010 @option{-mpowerpc-gfxopt} allows GCC to
22011 use the optional PowerPC architecture instructions in the Graphics
22012 group, including floating-point select.
22014 The @option{-mmfcrf} option allows GCC to generate the move from
22015 condition register field instruction implemented on the POWER4
22016 processor and other processors that support the PowerPC V2.01
22018 The @option{-mpopcntb} option allows GCC to generate the popcount and
22019 double-precision FP reciprocal estimate instruction implemented on the
22020 POWER5 processor and other processors that support the PowerPC V2.02
22022 The @option{-mpopcntd} option allows GCC to generate the popcount
22023 instruction implemented on the POWER7 processor and other processors
22024 that support the PowerPC V2.06 architecture.
22025 The @option{-mfprnd} option allows GCC to generate the FP round to
22026 integer instructions implemented on the POWER5+ processor and other
22027 processors that support the PowerPC V2.03 architecture.
22028 The @option{-mcmpb} option allows GCC to generate the compare bytes
22029 instruction implemented on the POWER6 processor and other processors
22030 that support the PowerPC V2.05 architecture.
22031 The @option{-mmfpgpr} option allows GCC to generate the FP move to/from
22032 general-purpose register instructions implemented on the POWER6X
22033 processor and other processors that support the extended PowerPC V2.05
22035 The @option{-mhard-dfp} option allows GCC to generate the decimal
22036 floating-point instructions implemented on some POWER processors.
22038 The @option{-mpowerpc64} option allows GCC to generate the additional
22039 64-bit instructions that are found in the full PowerPC64 architecture
22040 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
22041 @option{-mno-powerpc64}.
22043 @item -mcpu=@var{cpu_type}
22045 Set architecture type, register usage, and
22046 instruction scheduling parameters for machine type @var{cpu_type}.
22047 Supported values for @var{cpu_type} are @samp{401}, @samp{403},
22048 @samp{405}, @samp{405fp}, @samp{440}, @samp{440fp}, @samp{464}, @samp{464fp},
22049 @samp{476}, @samp{476fp}, @samp{505}, @samp{601}, @samp{602}, @samp{603},
22050 @samp{603e}, @samp{604}, @samp{604e}, @samp{620}, @samp{630}, @samp{740},
22051 @samp{7400}, @samp{7450}, @samp{750}, @samp{801}, @samp{821}, @samp{823},
22052 @samp{860}, @samp{970}, @samp{8540}, @samp{a2}, @samp{e300c2},
22053 @samp{e300c3}, @samp{e500mc}, @samp{e500mc64}, @samp{e5500},
22054 @samp{e6500}, @samp{ec603e}, @samp{G3}, @samp{G4}, @samp{G5},
22055 @samp{titan}, @samp{power3}, @samp{power4}, @samp{power5}, @samp{power5+},
22056 @samp{power6}, @samp{power6x}, @samp{power7}, @samp{power8},
22057 @samp{power9}, @samp{powerpc}, @samp{powerpc64}, @samp{powerpc64le},
22060 @option{-mcpu=powerpc}, @option{-mcpu=powerpc64}, and
22061 @option{-mcpu=powerpc64le} specify pure 32-bit PowerPC (either
22062 endian), 64-bit big endian PowerPC and 64-bit little endian PowerPC
22063 architecture machine types, with an appropriate, generic processor
22064 model assumed for scheduling purposes.
22066 The other options specify a specific processor. Code generated under
22067 those options runs best on that processor, and may not run at all on
22070 The @option{-mcpu} options automatically enable or disable the
22073 @gccoptlist{-maltivec -mfprnd -mhard-float -mmfcrf -mmultiple @gol
22074 -mpopcntb -mpopcntd -mpowerpc64 @gol
22075 -mpowerpc-gpopt -mpowerpc-gfxopt -msingle-float -mdouble-float @gol
22076 -msimple-fpu -mstring -mmulhw -mdlmzb -mmfpgpr -mvsx @gol
22077 -mcrypto -mdirect-move -mhtm -mpower8-fusion -mpower8-vector @gol
22078 -mquad-memory -mquad-memory-atomic -mfloat128 -mfloat128-hardware}
22080 The particular options set for any particular CPU varies between
22081 compiler versions, depending on what setting seems to produce optimal
22082 code for that CPU; it doesn't necessarily reflect the actual hardware's
22083 capabilities. If you wish to set an individual option to a particular
22084 value, you may specify it after the @option{-mcpu} option, like
22085 @option{-mcpu=970 -mno-altivec}.
22087 On AIX, the @option{-maltivec} and @option{-mpowerpc64} options are
22088 not enabled or disabled by the @option{-mcpu} option at present because
22089 AIX does not have full support for these options. You may still
22090 enable or disable them individually if you're sure it'll work in your
22093 @item -mtune=@var{cpu_type}
22095 Set the instruction scheduling parameters for machine type
22096 @var{cpu_type}, but do not set the architecture type or register usage,
22097 as @option{-mcpu=@var{cpu_type}} does. The same
22098 values for @var{cpu_type} are used for @option{-mtune} as for
22099 @option{-mcpu}. If both are specified, the code generated uses the
22100 architecture and registers set by @option{-mcpu}, but the
22101 scheduling parameters set by @option{-mtune}.
22103 @item -mcmodel=small
22104 @opindex mcmodel=small
22105 Generate PowerPC64 code for the small model: The TOC is limited to
22108 @item -mcmodel=medium
22109 @opindex mcmodel=medium
22110 Generate PowerPC64 code for the medium model: The TOC and other static
22111 data may be up to a total of 4G in size. This is the default for 64-bit
22114 @item -mcmodel=large
22115 @opindex mcmodel=large
22116 Generate PowerPC64 code for the large model: The TOC may be up to 4G
22117 in size. Other data and code is only limited by the 64-bit address
22121 @itemx -mno-altivec
22123 @opindex mno-altivec
22124 Generate code that uses (does not use) AltiVec instructions, and also
22125 enable the use of built-in functions that allow more direct access to
22126 the AltiVec instruction set. You may also need to set
22127 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
22130 When @option{-maltivec} is used, rather than @option{-maltivec=le} or
22131 @option{-maltivec=be}, the element order for AltiVec intrinsics such
22132 as @code{vec_splat}, @code{vec_extract}, and @code{vec_insert}
22133 match array element order corresponding to the endianness of the
22134 target. That is, element zero identifies the leftmost element in a
22135 vector register when targeting a big-endian platform, and identifies
22136 the rightmost element in a vector register when targeting a
22137 little-endian platform.
22140 @opindex maltivec=be
22141 Generate AltiVec instructions using big-endian element order,
22142 regardless of whether the target is big- or little-endian. This is
22143 the default when targeting a big-endian platform.
22145 The element order is used to interpret element numbers in AltiVec
22146 intrinsics such as @code{vec_splat}, @code{vec_extract}, and
22147 @code{vec_insert}. By default, these match array element order
22148 corresponding to the endianness for the target.
22151 @opindex maltivec=le
22152 Generate AltiVec instructions using little-endian element order,
22153 regardless of whether the target is big- or little-endian. This is
22154 the default when targeting a little-endian platform. This option is
22155 currently ignored when targeting a big-endian platform.
22157 The element order is used to interpret element numbers in AltiVec
22158 intrinsics such as @code{vec_splat}, @code{vec_extract}, and
22159 @code{vec_insert}. By default, these match array element order
22160 corresponding to the endianness for the target.
22165 @opindex mno-vrsave
22166 Generate VRSAVE instructions when generating AltiVec code.
22169 @opindex msecure-plt
22170 Generate code that allows @command{ld} and @command{ld.so}
22171 to build executables and shared
22172 libraries with non-executable @code{.plt} and @code{.got} sections.
22174 32-bit SYSV ABI option.
22178 Generate code that uses a BSS @code{.plt} section that @command{ld.so}
22180 requires @code{.plt} and @code{.got}
22181 sections that are both writable and executable.
22182 This is a PowerPC 32-bit SYSV ABI option.
22188 This switch enables or disables the generation of ISEL instructions.
22190 @item -misel=@var{yes/no}
22191 This switch has been deprecated. Use @option{-misel} and
22192 @option{-mno-isel} instead.
22198 This switch enables or disables the generation of SPE simd
22204 @opindex mno-paired
22205 This switch enables or disables the generation of PAIRED simd
22208 @item -mspe=@var{yes/no}
22209 This option has been deprecated. Use @option{-mspe} and
22210 @option{-mno-spe} instead.
22216 Generate code that uses (does not use) vector/scalar (VSX)
22217 instructions, and also enable the use of built-in functions that allow
22218 more direct access to the VSX instruction set.
22223 @opindex mno-crypto
22224 Enable the use (disable) of the built-in functions that allow direct
22225 access to the cryptographic instructions that were added in version
22226 2.07 of the PowerPC ISA.
22228 @item -mdirect-move
22229 @itemx -mno-direct-move
22230 @opindex mdirect-move
22231 @opindex mno-direct-move
22232 Generate code that uses (does not use) the instructions to move data
22233 between the general purpose registers and the vector/scalar (VSX)
22234 registers that were added in version 2.07 of the PowerPC ISA.
22240 Enable (disable) the use of the built-in functions that allow direct
22241 access to the Hardware Transactional Memory (HTM) instructions that
22242 were added in version 2.07 of the PowerPC ISA.
22244 @item -mpower8-fusion
22245 @itemx -mno-power8-fusion
22246 @opindex mpower8-fusion
22247 @opindex mno-power8-fusion
22248 Generate code that keeps (does not keeps) some integer operations
22249 adjacent so that the instructions can be fused together on power8 and
22252 @item -mpower8-vector
22253 @itemx -mno-power8-vector
22254 @opindex mpower8-vector
22255 @opindex mno-power8-vector
22256 Generate code that uses (does not use) the vector and scalar
22257 instructions that were added in version 2.07 of the PowerPC ISA. Also
22258 enable the use of built-in functions that allow more direct access to
22259 the vector instructions.
22261 @item -mquad-memory
22262 @itemx -mno-quad-memory
22263 @opindex mquad-memory
22264 @opindex mno-quad-memory
22265 Generate code that uses (does not use) the non-atomic quad word memory
22266 instructions. The @option{-mquad-memory} option requires use of
22269 @item -mquad-memory-atomic
22270 @itemx -mno-quad-memory-atomic
22271 @opindex mquad-memory-atomic
22272 @opindex mno-quad-memory-atomic
22273 Generate code that uses (does not use) the atomic quad word memory
22274 instructions. The @option{-mquad-memory-atomic} option requires use of
22278 @itemx -mno-float128
22280 @opindex mno-float128
22281 Enable/disable the @var{__float128} keyword for IEEE 128-bit floating point
22282 and use either software emulation for IEEE 128-bit floating point or
22283 hardware instructions.
22285 The VSX instruction set (@option{-mvsx}, @option{-mcpu=power7},
22286 @option{-mcpu=power8}), or @option{-mcpu=power9} must be enabled to
22287 use the IEEE 128-bit floating point support. The IEEE 128-bit
22288 floating point support only works on PowerPC Linux systems.
22290 The default for @option{-mfloat128} is enabled on PowerPC Linux
22291 systems using the VSX instruction set, and disabled on other systems.
22293 If you use the ISA 3.0 instruction set (@option{-mpower9-vector} or
22294 @option{-mcpu=power9}) on a 64-bit system, the IEEE 128-bit floating
22295 point support will also enable the generation of ISA 3.0 IEEE 128-bit
22296 floating point instructions. Otherwise, if you do not specify to
22297 generate ISA 3.0 instructions or you are targeting a 32-bit big endian
22298 system, IEEE 128-bit floating point will be done with software
22301 @item -mfloat128-hardware
22302 @itemx -mno-float128-hardware
22303 @opindex mfloat128-hardware
22304 @opindex mno-float128-hardware
22305 Enable/disable using ISA 3.0 hardware instructions to support the
22306 @var{__float128} data type.
22308 The default for @option{-mfloat128-hardware} is enabled on PowerPC
22309 Linux systems using the ISA 3.0 instruction set, and disabled on other
22312 @item -mfloat-gprs=@var{yes/single/double/no}
22313 @itemx -mfloat-gprs
22314 @opindex mfloat-gprs
22315 This switch enables or disables the generation of floating-point
22316 operations on the general-purpose registers for architectures that
22319 The argument @samp{yes} or @samp{single} enables the use of
22320 single-precision floating-point operations.
22322 The argument @samp{double} enables the use of single and
22323 double-precision floating-point operations.
22325 The argument @samp{no} disables floating-point operations on the
22326 general-purpose registers.
22328 This option is currently only available on the MPC854x.
22334 Generate code for 32-bit or 64-bit environments of Darwin and SVR4
22335 targets (including GNU/Linux). The 32-bit environment sets int, long
22336 and pointer to 32 bits and generates code that runs on any PowerPC
22337 variant. The 64-bit environment sets int to 32 bits and long and
22338 pointer to 64 bits, and generates code for PowerPC64, as for
22339 @option{-mpowerpc64}.
22342 @itemx -mno-fp-in-toc
22343 @itemx -mno-sum-in-toc
22344 @itemx -mminimal-toc
22346 @opindex mno-fp-in-toc
22347 @opindex mno-sum-in-toc
22348 @opindex mminimal-toc
22349 Modify generation of the TOC (Table Of Contents), which is created for
22350 every executable file. The @option{-mfull-toc} option is selected by
22351 default. In that case, GCC allocates at least one TOC entry for
22352 each unique non-automatic variable reference in your program. GCC
22353 also places floating-point constants in the TOC@. However, only
22354 16,384 entries are available in the TOC@.
22356 If you receive a linker error message that saying you have overflowed
22357 the available TOC space, you can reduce the amount of TOC space used
22358 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
22359 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
22360 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
22361 generate code to calculate the sum of an address and a constant at
22362 run time instead of putting that sum into the TOC@. You may specify one
22363 or both of these options. Each causes GCC to produce very slightly
22364 slower and larger code at the expense of conserving TOC space.
22366 If you still run out of space in the TOC even when you specify both of
22367 these options, specify @option{-mminimal-toc} instead. This option causes
22368 GCC to make only one TOC entry for every file. When you specify this
22369 option, GCC produces code that is slower and larger but which
22370 uses extremely little TOC space. You may wish to use this option
22371 only on files that contain less frequently-executed code.
22377 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
22378 @code{long} type, and the infrastructure needed to support them.
22379 Specifying @option{-maix64} implies @option{-mpowerpc64},
22380 while @option{-maix32} disables the 64-bit ABI and
22381 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
22384 @itemx -mno-xl-compat
22385 @opindex mxl-compat
22386 @opindex mno-xl-compat
22387 Produce code that conforms more closely to IBM XL compiler semantics
22388 when using AIX-compatible ABI@. Pass floating-point arguments to
22389 prototyped functions beyond the register save area (RSA) on the stack
22390 in addition to argument FPRs. Do not assume that most significant
22391 double in 128-bit long double value is properly rounded when comparing
22392 values and converting to double. Use XL symbol names for long double
22395 The AIX calling convention was extended but not initially documented to
22396 handle an obscure K&R C case of calling a function that takes the
22397 address of its arguments with fewer arguments than declared. IBM XL
22398 compilers access floating-point arguments that do not fit in the
22399 RSA from the stack when a subroutine is compiled without
22400 optimization. Because always storing floating-point arguments on the
22401 stack is inefficient and rarely needed, this option is not enabled by
22402 default and only is necessary when calling subroutines compiled by IBM
22403 XL compilers without optimization.
22407 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
22408 application written to use message passing with special startup code to
22409 enable the application to run. The system must have PE installed in the
22410 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
22411 must be overridden with the @option{-specs=} option to specify the
22412 appropriate directory location. The Parallel Environment does not
22413 support threads, so the @option{-mpe} option and the @option{-pthread}
22414 option are incompatible.
22416 @item -malign-natural
22417 @itemx -malign-power
22418 @opindex malign-natural
22419 @opindex malign-power
22420 On AIX, 32-bit Darwin, and 64-bit PowerPC GNU/Linux, the option
22421 @option{-malign-natural} overrides the ABI-defined alignment of larger
22422 types, such as floating-point doubles, on their natural size-based boundary.
22423 The option @option{-malign-power} instructs GCC to follow the ABI-specified
22424 alignment rules. GCC defaults to the standard alignment defined in the ABI@.
22426 On 64-bit Darwin, natural alignment is the default, and @option{-malign-power}
22430 @itemx -mhard-float
22431 @opindex msoft-float
22432 @opindex mhard-float
22433 Generate code that does not use (uses) the floating-point register set.
22434 Software floating-point emulation is provided if you use the
22435 @option{-msoft-float} option, and pass the option to GCC when linking.
22437 @item -msingle-float
22438 @itemx -mdouble-float
22439 @opindex msingle-float
22440 @opindex mdouble-float
22441 Generate code for single- or double-precision floating-point operations.
22442 @option{-mdouble-float} implies @option{-msingle-float}.
22445 @opindex msimple-fpu
22446 Do not generate @code{sqrt} and @code{div} instructions for hardware
22447 floating-point unit.
22449 @item -mfpu=@var{name}
22451 Specify type of floating-point unit. Valid values for @var{name} are
22452 @samp{sp_lite} (equivalent to @option{-msingle-float -msimple-fpu}),
22453 @samp{dp_lite} (equivalent to @option{-mdouble-float -msimple-fpu}),
22454 @samp{sp_full} (equivalent to @option{-msingle-float}),
22455 and @samp{dp_full} (equivalent to @option{-mdouble-float}).
22458 @opindex mxilinx-fpu
22459 Perform optimizations for the floating-point unit on Xilinx PPC 405/440.
22462 @itemx -mno-multiple
22464 @opindex mno-multiple
22465 Generate code that uses (does not use) the load multiple word
22466 instructions and the store multiple word instructions. These
22467 instructions are generated by default on POWER systems, and not
22468 generated on PowerPC systems. Do not use @option{-mmultiple} on little-endian
22469 PowerPC systems, since those instructions do not work when the
22470 processor is in little-endian mode. The exceptions are PPC740 and
22471 PPC750 which permit these instructions in little-endian mode.
22476 @opindex mno-string
22477 Generate code that uses (does not use) the load string instructions
22478 and the store string word instructions to save multiple registers and
22479 do small block moves. These instructions are generated by default on
22480 POWER systems, and not generated on PowerPC systems. Do not use
22481 @option{-mstring} on little-endian PowerPC systems, since those
22482 instructions do not work when the processor is in little-endian mode.
22483 The exceptions are PPC740 and PPC750 which permit these instructions
22484 in little-endian mode.
22489 @opindex mno-update
22490 Generate code that uses (does not use) the load or store instructions
22491 that update the base register to the address of the calculated memory
22492 location. These instructions are generated by default. If you use
22493 @option{-mno-update}, there is a small window between the time that the
22494 stack pointer is updated and the address of the previous frame is
22495 stored, which means code that walks the stack frame across interrupts or
22496 signals may get corrupted data.
22498 @item -mavoid-indexed-addresses
22499 @itemx -mno-avoid-indexed-addresses
22500 @opindex mavoid-indexed-addresses
22501 @opindex mno-avoid-indexed-addresses
22502 Generate code that tries to avoid (not avoid) the use of indexed load
22503 or store instructions. These instructions can incur a performance
22504 penalty on Power6 processors in certain situations, such as when
22505 stepping through large arrays that cross a 16M boundary. This option
22506 is enabled by default when targeting Power6 and disabled otherwise.
22509 @itemx -mno-fused-madd
22510 @opindex mfused-madd
22511 @opindex mno-fused-madd
22512 Generate code that uses (does not use) the floating-point multiply and
22513 accumulate instructions. These instructions are generated by default
22514 if hardware floating point is used. The machine-dependent
22515 @option{-mfused-madd} option is now mapped to the machine-independent
22516 @option{-ffp-contract=fast} option, and @option{-mno-fused-madd} is
22517 mapped to @option{-ffp-contract=off}.
22523 Generate code that uses (does not use) the half-word multiply and
22524 multiply-accumulate instructions on the IBM 405, 440, 464 and 476 processors.
22525 These instructions are generated by default when targeting those
22532 Generate code that uses (does not use) the string-search @samp{dlmzb}
22533 instruction on the IBM 405, 440, 464 and 476 processors. This instruction is
22534 generated by default when targeting those processors.
22536 @item -mno-bit-align
22538 @opindex mno-bit-align
22539 @opindex mbit-align
22540 On System V.4 and embedded PowerPC systems do not (do) force structures
22541 and unions that contain bit-fields to be aligned to the base type of the
22544 For example, by default a structure containing nothing but 8
22545 @code{unsigned} bit-fields of length 1 is aligned to a 4-byte
22546 boundary and has a size of 4 bytes. By using @option{-mno-bit-align},
22547 the structure is aligned to a 1-byte boundary and is 1 byte in
22550 @item -mno-strict-align
22551 @itemx -mstrict-align
22552 @opindex mno-strict-align
22553 @opindex mstrict-align
22554 On System V.4 and embedded PowerPC systems do not (do) assume that
22555 unaligned memory references are handled by the system.
22557 @item -mrelocatable
22558 @itemx -mno-relocatable
22559 @opindex mrelocatable
22560 @opindex mno-relocatable
22561 Generate code that allows (does not allow) a static executable to be
22562 relocated to a different address at run time. A simple embedded
22563 PowerPC system loader should relocate the entire contents of
22564 @code{.got2} and 4-byte locations listed in the @code{.fixup} section,
22565 a table of 32-bit addresses generated by this option. For this to
22566 work, all objects linked together must be compiled with
22567 @option{-mrelocatable} or @option{-mrelocatable-lib}.
22568 @option{-mrelocatable} code aligns the stack to an 8-byte boundary.
22570 @item -mrelocatable-lib
22571 @itemx -mno-relocatable-lib
22572 @opindex mrelocatable-lib
22573 @opindex mno-relocatable-lib
22574 Like @option{-mrelocatable}, @option{-mrelocatable-lib} generates a
22575 @code{.fixup} section to allow static executables to be relocated at
22576 run time, but @option{-mrelocatable-lib} does not use the smaller stack
22577 alignment of @option{-mrelocatable}. Objects compiled with
22578 @option{-mrelocatable-lib} may be linked with objects compiled with
22579 any combination of the @option{-mrelocatable} options.
22585 On System V.4 and embedded PowerPC systems do not (do) assume that
22586 register 2 contains a pointer to a global area pointing to the addresses
22587 used in the program.
22590 @itemx -mlittle-endian
22592 @opindex mlittle-endian
22593 On System V.4 and embedded PowerPC systems compile code for the
22594 processor in little-endian mode. The @option{-mlittle-endian} option is
22595 the same as @option{-mlittle}.
22598 @itemx -mbig-endian
22600 @opindex mbig-endian
22601 On System V.4 and embedded PowerPC systems compile code for the
22602 processor in big-endian mode. The @option{-mbig-endian} option is
22603 the same as @option{-mbig}.
22605 @item -mdynamic-no-pic
22606 @opindex mdynamic-no-pic
22607 On Darwin and Mac OS X systems, compile code so that it is not
22608 relocatable, but that its external references are relocatable. The
22609 resulting code is suitable for applications, but not shared
22612 @item -msingle-pic-base
22613 @opindex msingle-pic-base
22614 Treat the register used for PIC addressing as read-only, rather than
22615 loading it in the prologue for each function. The runtime system is
22616 responsible for initializing this register with an appropriate value
22617 before execution begins.
22619 @item -mprioritize-restricted-insns=@var{priority}
22620 @opindex mprioritize-restricted-insns
22621 This option controls the priority that is assigned to
22622 dispatch-slot restricted instructions during the second scheduling
22623 pass. The argument @var{priority} takes the value @samp{0}, @samp{1},
22624 or @samp{2} to assign no, highest, or second-highest (respectively)
22625 priority to dispatch-slot restricted
22628 @item -msched-costly-dep=@var{dependence_type}
22629 @opindex msched-costly-dep
22630 This option controls which dependences are considered costly
22631 by the target during instruction scheduling. The argument
22632 @var{dependence_type} takes one of the following values:
22636 No dependence is costly.
22639 All dependences are costly.
22641 @item @samp{true_store_to_load}
22642 A true dependence from store to load is costly.
22644 @item @samp{store_to_load}
22645 Any dependence from store to load is costly.
22648 Any dependence for which the latency is greater than or equal to
22649 @var{number} is costly.
22652 @item -minsert-sched-nops=@var{scheme}
22653 @opindex minsert-sched-nops
22654 This option controls which NOP insertion scheme is used during
22655 the second scheduling pass. The argument @var{scheme} takes one of the
22663 Pad with NOPs any dispatch group that has vacant issue slots,
22664 according to the scheduler's grouping.
22666 @item @samp{regroup_exact}
22667 Insert NOPs to force costly dependent insns into
22668 separate groups. Insert exactly as many NOPs as needed to force an insn
22669 to a new group, according to the estimated processor grouping.
22672 Insert NOPs to force costly dependent insns into
22673 separate groups. Insert @var{number} NOPs to force an insn to a new group.
22677 @opindex mcall-sysv
22678 On System V.4 and embedded PowerPC systems compile code using calling
22679 conventions that adhere to the March 1995 draft of the System V
22680 Application Binary Interface, PowerPC processor supplement. This is the
22681 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
22683 @item -mcall-sysv-eabi
22685 @opindex mcall-sysv-eabi
22686 @opindex mcall-eabi
22687 Specify both @option{-mcall-sysv} and @option{-meabi} options.
22689 @item -mcall-sysv-noeabi
22690 @opindex mcall-sysv-noeabi
22691 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
22693 @item -mcall-aixdesc
22695 On System V.4 and embedded PowerPC systems compile code for the AIX
22699 @opindex mcall-linux
22700 On System V.4 and embedded PowerPC systems compile code for the
22701 Linux-based GNU system.
22703 @item -mcall-freebsd
22704 @opindex mcall-freebsd
22705 On System V.4 and embedded PowerPC systems compile code for the
22706 FreeBSD operating system.
22708 @item -mcall-netbsd
22709 @opindex mcall-netbsd
22710 On System V.4 and embedded PowerPC systems compile code for the
22711 NetBSD operating system.
22713 @item -mcall-openbsd
22714 @opindex mcall-netbsd
22715 On System V.4 and embedded PowerPC systems compile code for the
22716 OpenBSD operating system.
22718 @item -maix-struct-return
22719 @opindex maix-struct-return
22720 Return all structures in memory (as specified by the AIX ABI)@.
22722 @item -msvr4-struct-return
22723 @opindex msvr4-struct-return
22724 Return structures smaller than 8 bytes in registers (as specified by the
22727 @item -mabi=@var{abi-type}
22729 Extend the current ABI with a particular extension, or remove such extension.
22730 Valid values are @samp{altivec}, @samp{no-altivec}, @samp{spe},
22731 @samp{no-spe}, @samp{ibmlongdouble}, @samp{ieeelongdouble},
22732 @samp{elfv1}, @samp{elfv2}@.
22736 Extend the current ABI with SPE ABI extensions. This does not change
22737 the default ABI, instead it adds the SPE ABI extensions to the current
22741 @opindex mabi=no-spe
22742 Disable Book-E SPE ABI extensions for the current ABI@.
22744 @item -mabi=ibmlongdouble
22745 @opindex mabi=ibmlongdouble
22746 Change the current ABI to use IBM extended-precision long double.
22747 This is not likely to work if your system defaults to using IEEE
22748 extended-precision long double. If you change the long double type
22749 from IEEE extended-precision, the compiler will issue a warning unless
22750 you use the @option{-Wno-psabi} option.
22752 @item -mabi=ieeelongdouble
22753 @opindex mabi=ieeelongdouble
22754 Change the current ABI to use IEEE extended-precision long double.
22755 This is not likely to work if your system defaults to using IBM
22756 extended-precision long double. If you change the long double type
22757 from IBM extended-precision, the compiler will issue a warning unless
22758 you use the @option{-Wno-psabi} option.
22761 @opindex mabi=elfv1
22762 Change the current ABI to use the ELFv1 ABI.
22763 This is the default ABI for big-endian PowerPC 64-bit Linux.
22764 Overriding the default ABI requires special system support and is
22765 likely to fail in spectacular ways.
22768 @opindex mabi=elfv2
22769 Change the current ABI to use the ELFv2 ABI.
22770 This is the default ABI for little-endian PowerPC 64-bit Linux.
22771 Overriding the default ABI requires special system support and is
22772 likely to fail in spectacular ways.
22774 @item -mgnu-attribute
22775 @itemx -mno-gnu-attribute
22776 @opindex mgnu-attribute
22777 @opindex mno-gnu-attribute
22778 Emit .gnu_attribute assembly directives to set tag/value pairs in a
22779 .gnu.attributes section that specify ABI variations in function
22780 parameters or return values.
22783 @itemx -mno-prototype
22784 @opindex mprototype
22785 @opindex mno-prototype
22786 On System V.4 and embedded PowerPC systems assume that all calls to
22787 variable argument functions are properly prototyped. Otherwise, the
22788 compiler must insert an instruction before every non-prototyped call to
22789 set or clear bit 6 of the condition code register (@code{CR}) to
22790 indicate whether floating-point values are passed in the floating-point
22791 registers in case the function takes variable arguments. With
22792 @option{-mprototype}, only calls to prototyped variable argument functions
22793 set or clear the bit.
22797 On embedded PowerPC systems, assume that the startup module is called
22798 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
22799 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}
22804 On embedded PowerPC systems, assume that the startup module is called
22805 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
22810 On embedded PowerPC systems, assume that the startup module is called
22811 @file{crt0.o} and the standard C libraries are @file{libads.a} and
22814 @item -myellowknife
22815 @opindex myellowknife
22816 On embedded PowerPC systems, assume that the startup module is called
22817 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
22822 On System V.4 and embedded PowerPC systems, specify that you are
22823 compiling for a VxWorks system.
22827 On embedded PowerPC systems, set the @code{PPC_EMB} bit in the ELF flags
22828 header to indicate that @samp{eabi} extended relocations are used.
22834 On System V.4 and embedded PowerPC systems do (do not) adhere to the
22835 Embedded Applications Binary Interface (EABI), which is a set of
22836 modifications to the System V.4 specifications. Selecting @option{-meabi}
22837 means that the stack is aligned to an 8-byte boundary, a function
22838 @code{__eabi} is called from @code{main} to set up the EABI
22839 environment, and the @option{-msdata} option can use both @code{r2} and
22840 @code{r13} to point to two separate small data areas. Selecting
22841 @option{-mno-eabi} means that the stack is aligned to a 16-byte boundary,
22842 no EABI initialization function is called from @code{main}, and the
22843 @option{-msdata} option only uses @code{r13} to point to a single
22844 small data area. The @option{-meabi} option is on by default if you
22845 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
22848 @opindex msdata=eabi
22849 On System V.4 and embedded PowerPC systems, put small initialized
22850 @code{const} global and static data in the @code{.sdata2} section, which
22851 is pointed to by register @code{r2}. Put small initialized
22852 non-@code{const} global and static data in the @code{.sdata} section,
22853 which is pointed to by register @code{r13}. Put small uninitialized
22854 global and static data in the @code{.sbss} section, which is adjacent to
22855 the @code{.sdata} section. The @option{-msdata=eabi} option is
22856 incompatible with the @option{-mrelocatable} option. The
22857 @option{-msdata=eabi} option also sets the @option{-memb} option.
22860 @opindex msdata=sysv
22861 On System V.4 and embedded PowerPC systems, put small global and static
22862 data in the @code{.sdata} section, which is pointed to by register
22863 @code{r13}. Put small uninitialized global and static data in the
22864 @code{.sbss} section, which is adjacent to the @code{.sdata} section.
22865 The @option{-msdata=sysv} option is incompatible with the
22866 @option{-mrelocatable} option.
22868 @item -msdata=default
22870 @opindex msdata=default
22872 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
22873 compile code the same as @option{-msdata=eabi}, otherwise compile code the
22874 same as @option{-msdata=sysv}.
22877 @opindex msdata=data
22878 On System V.4 and embedded PowerPC systems, put small global
22879 data in the @code{.sdata} section. Put small uninitialized global
22880 data in the @code{.sbss} section. Do not use register @code{r13}
22881 to address small data however. This is the default behavior unless
22882 other @option{-msdata} options are used.
22886 @opindex msdata=none
22888 On embedded PowerPC systems, put all initialized global and static data
22889 in the @code{.data} section, and all uninitialized data in the
22890 @code{.bss} section.
22892 @item -mblock-move-inline-limit=@var{num}
22893 @opindex mblock-move-inline-limit
22894 Inline all block moves (such as calls to @code{memcpy} or structure
22895 copies) less than or equal to @var{num} bytes. The minimum value for
22896 @var{num} is 32 bytes on 32-bit targets and 64 bytes on 64-bit
22897 targets. The default value is target-specific.
22901 @cindex smaller data references (PowerPC)
22902 @cindex .sdata/.sdata2 references (PowerPC)
22903 On embedded PowerPC systems, put global and static items less than or
22904 equal to @var{num} bytes into the small data or BSS sections instead of
22905 the normal data or BSS section. By default, @var{num} is 8. The
22906 @option{-G @var{num}} switch is also passed to the linker.
22907 All modules should be compiled with the same @option{-G @var{num}} value.
22910 @itemx -mno-regnames
22912 @opindex mno-regnames
22913 On System V.4 and embedded PowerPC systems do (do not) emit register
22914 names in the assembly language output using symbolic forms.
22917 @itemx -mno-longcall
22919 @opindex mno-longcall
22920 By default assume that all calls are far away so that a longer and more
22921 expensive calling sequence is required. This is required for calls
22922 farther than 32 megabytes (33,554,432 bytes) from the current location.
22923 A short call is generated if the compiler knows
22924 the call cannot be that far away. This setting can be overridden by
22925 the @code{shortcall} function attribute, or by @code{#pragma
22928 Some linkers are capable of detecting out-of-range calls and generating
22929 glue code on the fly. On these systems, long calls are unnecessary and
22930 generate slower code. As of this writing, the AIX linker can do this,
22931 as can the GNU linker for PowerPC/64. It is planned to add this feature
22932 to the GNU linker for 32-bit PowerPC systems as well.
22934 On Darwin/PPC systems, @code{#pragma longcall} generates @code{jbsr
22935 callee, L42}, plus a @dfn{branch island} (glue code). The two target
22936 addresses represent the callee and the branch island. The
22937 Darwin/PPC linker prefers the first address and generates a @code{bl
22938 callee} if the PPC @code{bl} instruction reaches the callee directly;
22939 otherwise, the linker generates @code{bl L42} to call the branch
22940 island. The branch island is appended to the body of the
22941 calling function; it computes the full 32-bit address of the callee
22944 On Mach-O (Darwin) systems, this option directs the compiler emit to
22945 the glue for every direct call, and the Darwin linker decides whether
22946 to use or discard it.
22948 In the future, GCC may ignore all longcall specifications
22949 when the linker is known to generate glue.
22951 @item -mtls-markers
22952 @itemx -mno-tls-markers
22953 @opindex mtls-markers
22954 @opindex mno-tls-markers
22955 Mark (do not mark) calls to @code{__tls_get_addr} with a relocation
22956 specifying the function argument. The relocation allows the linker to
22957 reliably associate function call with argument setup instructions for
22958 TLS optimization, which in turn allows GCC to better schedule the
22964 This option enables use of the reciprocal estimate and
22965 reciprocal square root estimate instructions with additional
22966 Newton-Raphson steps to increase precision instead of doing a divide or
22967 square root and divide for floating-point arguments. You should use
22968 the @option{-ffast-math} option when using @option{-mrecip} (or at
22969 least @option{-funsafe-math-optimizations},
22970 @option{-ffinite-math-only}, @option{-freciprocal-math} and
22971 @option{-fno-trapping-math}). Note that while the throughput of the
22972 sequence is generally higher than the throughput of the non-reciprocal
22973 instruction, the precision of the sequence can be decreased by up to 2
22974 ulp (i.e.@: the inverse of 1.0 equals 0.99999994) for reciprocal square
22977 @item -mrecip=@var{opt}
22978 @opindex mrecip=opt
22979 This option controls which reciprocal estimate instructions
22980 may be used. @var{opt} is a comma-separated list of options, which may
22981 be preceded by a @code{!} to invert the option:
22986 Enable all estimate instructions.
22989 Enable the default instructions, equivalent to @option{-mrecip}.
22992 Disable all estimate instructions, equivalent to @option{-mno-recip}.
22995 Enable the reciprocal approximation instructions for both
22996 single and double precision.
22999 Enable the single-precision reciprocal approximation instructions.
23002 Enable the double-precision reciprocal approximation instructions.
23005 Enable the reciprocal square root approximation instructions for both
23006 single and double precision.
23009 Enable the single-precision reciprocal square root approximation instructions.
23012 Enable the double-precision reciprocal square root approximation instructions.
23016 So, for example, @option{-mrecip=all,!rsqrtd} enables
23017 all of the reciprocal estimate instructions, except for the
23018 @code{FRSQRTE}, @code{XSRSQRTEDP}, and @code{XVRSQRTEDP} instructions
23019 which handle the double-precision reciprocal square root calculations.
23021 @item -mrecip-precision
23022 @itemx -mno-recip-precision
23023 @opindex mrecip-precision
23024 Assume (do not assume) that the reciprocal estimate instructions
23025 provide higher-precision estimates than is mandated by the PowerPC
23026 ABI. Selecting @option{-mcpu=power6}, @option{-mcpu=power7} or
23027 @option{-mcpu=power8} automatically selects @option{-mrecip-precision}.
23028 The double-precision square root estimate instructions are not generated by
23029 default on low-precision machines, since they do not provide an
23030 estimate that converges after three steps.
23032 @item -mveclibabi=@var{type}
23033 @opindex mveclibabi
23034 Specifies the ABI type to use for vectorizing intrinsics using an
23035 external library. The only type supported at present is @samp{mass},
23036 which specifies to use IBM's Mathematical Acceleration Subsystem
23037 (MASS) libraries for vectorizing intrinsics using external libraries.
23038 GCC currently emits calls to @code{acosd2}, @code{acosf4},
23039 @code{acoshd2}, @code{acoshf4}, @code{asind2}, @code{asinf4},
23040 @code{asinhd2}, @code{asinhf4}, @code{atan2d2}, @code{atan2f4},
23041 @code{atand2}, @code{atanf4}, @code{atanhd2}, @code{atanhf4},
23042 @code{cbrtd2}, @code{cbrtf4}, @code{cosd2}, @code{cosf4},
23043 @code{coshd2}, @code{coshf4}, @code{erfcd2}, @code{erfcf4},
23044 @code{erfd2}, @code{erff4}, @code{exp2d2}, @code{exp2f4},
23045 @code{expd2}, @code{expf4}, @code{expm1d2}, @code{expm1f4},
23046 @code{hypotd2}, @code{hypotf4}, @code{lgammad2}, @code{lgammaf4},
23047 @code{log10d2}, @code{log10f4}, @code{log1pd2}, @code{log1pf4},
23048 @code{log2d2}, @code{log2f4}, @code{logd2}, @code{logf4},
23049 @code{powd2}, @code{powf4}, @code{sind2}, @code{sinf4}, @code{sinhd2},
23050 @code{sinhf4}, @code{sqrtd2}, @code{sqrtf4}, @code{tand2},
23051 @code{tanf4}, @code{tanhd2}, and @code{tanhf4} when generating code
23052 for power7. Both @option{-ftree-vectorize} and
23053 @option{-funsafe-math-optimizations} must also be enabled. The MASS
23054 libraries must be specified at link time.
23059 Generate (do not generate) the @code{friz} instruction when the
23060 @option{-funsafe-math-optimizations} option is used to optimize
23061 rounding of floating-point values to 64-bit integer and back to floating
23062 point. The @code{friz} instruction does not return the same value if
23063 the floating-point number is too large to fit in an integer.
23065 @item -mpointers-to-nested-functions
23066 @itemx -mno-pointers-to-nested-functions
23067 @opindex mpointers-to-nested-functions
23068 Generate (do not generate) code to load up the static chain register
23069 (@code{r11}) when calling through a pointer on AIX and 64-bit Linux
23070 systems where a function pointer points to a 3-word descriptor giving
23071 the function address, TOC value to be loaded in register @code{r2}, and
23072 static chain value to be loaded in register @code{r11}. The
23073 @option{-mpointers-to-nested-functions} is on by default. You cannot
23074 call through pointers to nested functions or pointers
23075 to functions compiled in other languages that use the static chain if
23076 you use @option{-mno-pointers-to-nested-functions}.
23078 @item -msave-toc-indirect
23079 @itemx -mno-save-toc-indirect
23080 @opindex msave-toc-indirect
23081 Generate (do not generate) code to save the TOC value in the reserved
23082 stack location in the function prologue if the function calls through
23083 a pointer on AIX and 64-bit Linux systems. If the TOC value is not
23084 saved in the prologue, it is saved just before the call through the
23085 pointer. The @option{-mno-save-toc-indirect} option is the default.
23087 @item -mcompat-align-parm
23088 @itemx -mno-compat-align-parm
23089 @opindex mcompat-align-parm
23090 Generate (do not generate) code to pass structure parameters with a
23091 maximum alignment of 64 bits, for compatibility with older versions
23094 Older versions of GCC (prior to 4.9.0) incorrectly did not align a
23095 structure parameter on a 128-bit boundary when that structure contained
23096 a member requiring 128-bit alignment. This is corrected in more
23097 recent versions of GCC. This option may be used to generate code
23098 that is compatible with functions compiled with older versions of
23101 The @option{-mno-compat-align-parm} option is the default.
23103 @item -mstack-protector-guard=@var{guard}
23104 @itemx -mstack-protector-guard-reg=@var{reg}
23105 @itemx -mstack-protector-guard-offset=@var{offset}
23106 @itemx -mstack-protector-guard-symbol=@var{symbol}
23107 @opindex mstack-protector-guard
23108 @opindex mstack-protector-guard-reg
23109 @opindex mstack-protector-guard-offset
23110 @opindex mstack-protector-guard-symbol
23111 Generate stack protection code using canary at @var{guard}. Supported
23112 locations are @samp{global} for global canary or @samp{tls} for per-thread
23113 canary in the TLS block (the default with GNU libc version 2.4 or later).
23115 With the latter choice the options
23116 @option{-mstack-protector-guard-reg=@var{reg}} and
23117 @option{-mstack-protector-guard-offset=@var{offset}} furthermore specify
23118 which register to use as base register for reading the canary, and from what
23119 offset from that base register. The default for those is as specified in the
23120 relevant ABI. @option{-mstack-protector-guard-symbol=@var{symbol}} overrides
23121 the offset with a symbol reference to a canary in the TLS block.
23125 @subsection RX Options
23128 These command-line options are defined for RX targets:
23131 @item -m64bit-doubles
23132 @itemx -m32bit-doubles
23133 @opindex m64bit-doubles
23134 @opindex m32bit-doubles
23135 Make the @code{double} data type be 64 bits (@option{-m64bit-doubles})
23136 or 32 bits (@option{-m32bit-doubles}) in size. The default is
23137 @option{-m32bit-doubles}. @emph{Note} RX floating-point hardware only
23138 works on 32-bit values, which is why the default is
23139 @option{-m32bit-doubles}.
23145 Enables (@option{-fpu}) or disables (@option{-nofpu}) the use of RX
23146 floating-point hardware. The default is enabled for the RX600
23147 series and disabled for the RX200 series.
23149 Floating-point instructions are only generated for 32-bit floating-point
23150 values, however, so the FPU hardware is not used for doubles if the
23151 @option{-m64bit-doubles} option is used.
23153 @emph{Note} If the @option{-fpu} option is enabled then
23154 @option{-funsafe-math-optimizations} is also enabled automatically.
23155 This is because the RX FPU instructions are themselves unsafe.
23157 @item -mcpu=@var{name}
23159 Selects the type of RX CPU to be targeted. Currently three types are
23160 supported, the generic @samp{RX600} and @samp{RX200} series hardware and
23161 the specific @samp{RX610} CPU. The default is @samp{RX600}.
23163 The only difference between @samp{RX600} and @samp{RX610} is that the
23164 @samp{RX610} does not support the @code{MVTIPL} instruction.
23166 The @samp{RX200} series does not have a hardware floating-point unit
23167 and so @option{-nofpu} is enabled by default when this type is
23170 @item -mbig-endian-data
23171 @itemx -mlittle-endian-data
23172 @opindex mbig-endian-data
23173 @opindex mlittle-endian-data
23174 Store data (but not code) in the big-endian format. The default is
23175 @option{-mlittle-endian-data}, i.e.@: to store data in the little-endian
23178 @item -msmall-data-limit=@var{N}
23179 @opindex msmall-data-limit
23180 Specifies the maximum size in bytes of global and static variables
23181 which can be placed into the small data area. Using the small data
23182 area can lead to smaller and faster code, but the size of area is
23183 limited and it is up to the programmer to ensure that the area does
23184 not overflow. Also when the small data area is used one of the RX's
23185 registers (usually @code{r13}) is reserved for use pointing to this
23186 area, so it is no longer available for use by the compiler. This
23187 could result in slower and/or larger code if variables are pushed onto
23188 the stack instead of being held in this register.
23190 Note, common variables (variables that have not been initialized) and
23191 constants are not placed into the small data area as they are assigned
23192 to other sections in the output executable.
23194 The default value is zero, which disables this feature. Note, this
23195 feature is not enabled by default with higher optimization levels
23196 (@option{-O2} etc) because of the potentially detrimental effects of
23197 reserving a register. It is up to the programmer to experiment and
23198 discover whether this feature is of benefit to their program. See the
23199 description of the @option{-mpid} option for a description of how the
23200 actual register to hold the small data area pointer is chosen.
23206 Use the simulator runtime. The default is to use the libgloss
23207 board-specific runtime.
23209 @item -mas100-syntax
23210 @itemx -mno-as100-syntax
23211 @opindex mas100-syntax
23212 @opindex mno-as100-syntax
23213 When generating assembler output use a syntax that is compatible with
23214 Renesas's AS100 assembler. This syntax can also be handled by the GAS
23215 assembler, but it has some restrictions so it is not generated by default.
23217 @item -mmax-constant-size=@var{N}
23218 @opindex mmax-constant-size
23219 Specifies the maximum size, in bytes, of a constant that can be used as
23220 an operand in a RX instruction. Although the RX instruction set does
23221 allow constants of up to 4 bytes in length to be used in instructions,
23222 a longer value equates to a longer instruction. Thus in some
23223 circumstances it can be beneficial to restrict the size of constants
23224 that are used in instructions. Constants that are too big are instead
23225 placed into a constant pool and referenced via register indirection.
23227 The value @var{N} can be between 0 and 4. A value of 0 (the default)
23228 or 4 means that constants of any size are allowed.
23232 Enable linker relaxation. Linker relaxation is a process whereby the
23233 linker attempts to reduce the size of a program by finding shorter
23234 versions of various instructions. Disabled by default.
23236 @item -mint-register=@var{N}
23237 @opindex mint-register
23238 Specify the number of registers to reserve for fast interrupt handler
23239 functions. The value @var{N} can be between 0 and 4. A value of 1
23240 means that register @code{r13} is reserved for the exclusive use
23241 of fast interrupt handlers. A value of 2 reserves @code{r13} and
23242 @code{r12}. A value of 3 reserves @code{r13}, @code{r12} and
23243 @code{r11}, and a value of 4 reserves @code{r13} through @code{r10}.
23244 A value of 0, the default, does not reserve any registers.
23246 @item -msave-acc-in-interrupts
23247 @opindex msave-acc-in-interrupts
23248 Specifies that interrupt handler functions should preserve the
23249 accumulator register. This is only necessary if normal code might use
23250 the accumulator register, for example because it performs 64-bit
23251 multiplications. The default is to ignore the accumulator as this
23252 makes the interrupt handlers faster.
23258 Enables the generation of position independent data. When enabled any
23259 access to constant data is done via an offset from a base address
23260 held in a register. This allows the location of constant data to be
23261 determined at run time without requiring the executable to be
23262 relocated, which is a benefit to embedded applications with tight
23263 memory constraints. Data that can be modified is not affected by this
23266 Note, using this feature reserves a register, usually @code{r13}, for
23267 the constant data base address. This can result in slower and/or
23268 larger code, especially in complicated functions.
23270 The actual register chosen to hold the constant data base address
23271 depends upon whether the @option{-msmall-data-limit} and/or the
23272 @option{-mint-register} command-line options are enabled. Starting
23273 with register @code{r13} and proceeding downwards, registers are
23274 allocated first to satisfy the requirements of @option{-mint-register},
23275 then @option{-mpid} and finally @option{-msmall-data-limit}. Thus it
23276 is possible for the small data area register to be @code{r8} if both
23277 @option{-mint-register=4} and @option{-mpid} are specified on the
23280 By default this feature is not enabled. The default can be restored
23281 via the @option{-mno-pid} command-line option.
23283 @item -mno-warn-multiple-fast-interrupts
23284 @itemx -mwarn-multiple-fast-interrupts
23285 @opindex mno-warn-multiple-fast-interrupts
23286 @opindex mwarn-multiple-fast-interrupts
23287 Prevents GCC from issuing a warning message if it finds more than one
23288 fast interrupt handler when it is compiling a file. The default is to
23289 issue a warning for each extra fast interrupt handler found, as the RX
23290 only supports one such interrupt.
23292 @item -mallow-string-insns
23293 @itemx -mno-allow-string-insns
23294 @opindex mallow-string-insns
23295 @opindex mno-allow-string-insns
23296 Enables or disables the use of the string manipulation instructions
23297 @code{SMOVF}, @code{SCMPU}, @code{SMOVB}, @code{SMOVU}, @code{SUNTIL}
23298 @code{SWHILE} and also the @code{RMPA} instruction. These
23299 instructions may prefetch data, which is not safe to do if accessing
23300 an I/O register. (See section 12.2.7 of the RX62N Group User's Manual
23301 for more information).
23303 The default is to allow these instructions, but it is not possible for
23304 GCC to reliably detect all circumstances where a string instruction
23305 might be used to access an I/O register, so their use cannot be
23306 disabled automatically. Instead it is reliant upon the programmer to
23307 use the @option{-mno-allow-string-insns} option if their program
23308 accesses I/O space.
23310 When the instructions are enabled GCC defines the C preprocessor
23311 symbol @code{__RX_ALLOW_STRING_INSNS__}, otherwise it defines the
23312 symbol @code{__RX_DISALLOW_STRING_INSNS__}.
23318 Use only (or not only) @code{JSR} instructions to access functions.
23319 This option can be used when code size exceeds the range of @code{BSR}
23320 instructions. Note that @option{-mno-jsr} does not mean to not use
23321 @code{JSR} but instead means that any type of branch may be used.
23324 @emph{Note:} The generic GCC command-line option @option{-ffixed-@var{reg}}
23325 has special significance to the RX port when used with the
23326 @code{interrupt} function attribute. This attribute indicates a
23327 function intended to process fast interrupts. GCC ensures
23328 that it only uses the registers @code{r10}, @code{r11}, @code{r12}
23329 and/or @code{r13} and only provided that the normal use of the
23330 corresponding registers have been restricted via the
23331 @option{-ffixed-@var{reg}} or @option{-mint-register} command-line
23334 @node S/390 and zSeries Options
23335 @subsection S/390 and zSeries Options
23336 @cindex S/390 and zSeries Options
23338 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
23342 @itemx -msoft-float
23343 @opindex mhard-float
23344 @opindex msoft-float
23345 Use (do not use) the hardware floating-point instructions and registers
23346 for floating-point operations. When @option{-msoft-float} is specified,
23347 functions in @file{libgcc.a} are used to perform floating-point
23348 operations. When @option{-mhard-float} is specified, the compiler
23349 generates IEEE floating-point instructions. This is the default.
23352 @itemx -mno-hard-dfp
23354 @opindex mno-hard-dfp
23355 Use (do not use) the hardware decimal-floating-point instructions for
23356 decimal-floating-point operations. When @option{-mno-hard-dfp} is
23357 specified, functions in @file{libgcc.a} are used to perform
23358 decimal-floating-point operations. When @option{-mhard-dfp} is
23359 specified, the compiler generates decimal-floating-point hardware
23360 instructions. This is the default for @option{-march=z9-ec} or higher.
23362 @item -mlong-double-64
23363 @itemx -mlong-double-128
23364 @opindex mlong-double-64
23365 @opindex mlong-double-128
23366 These switches control the size of @code{long double} type. A size
23367 of 64 bits makes the @code{long double} type equivalent to the @code{double}
23368 type. This is the default.
23371 @itemx -mno-backchain
23372 @opindex mbackchain
23373 @opindex mno-backchain
23374 Store (do not store) the address of the caller's frame as backchain pointer
23375 into the callee's stack frame.
23376 A backchain may be needed to allow debugging using tools that do not understand
23377 DWARF call frame information.
23378 When @option{-mno-packed-stack} is in effect, the backchain pointer is stored
23379 at the bottom of the stack frame; when @option{-mpacked-stack} is in effect,
23380 the backchain is placed into the topmost word of the 96/160 byte register
23383 In general, code compiled with @option{-mbackchain} is call-compatible with
23384 code compiled with @option{-mmo-backchain}; however, use of the backchain
23385 for debugging purposes usually requires that the whole binary is built with
23386 @option{-mbackchain}. Note that the combination of @option{-mbackchain},
23387 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
23388 to build a linux kernel use @option{-msoft-float}.
23390 The default is to not maintain the backchain.
23392 @item -mpacked-stack
23393 @itemx -mno-packed-stack
23394 @opindex mpacked-stack
23395 @opindex mno-packed-stack
23396 Use (do not use) the packed stack layout. When @option{-mno-packed-stack} is
23397 specified, the compiler uses the all fields of the 96/160 byte register save
23398 area only for their default purpose; unused fields still take up stack space.
23399 When @option{-mpacked-stack} is specified, register save slots are densely
23400 packed at the top of the register save area; unused space is reused for other
23401 purposes, allowing for more efficient use of the available stack space.
23402 However, when @option{-mbackchain} is also in effect, the topmost word of
23403 the save area is always used to store the backchain, and the return address
23404 register is always saved two words below the backchain.
23406 As long as the stack frame backchain is not used, code generated with
23407 @option{-mpacked-stack} is call-compatible with code generated with
23408 @option{-mno-packed-stack}. Note that some non-FSF releases of GCC 2.95 for
23409 S/390 or zSeries generated code that uses the stack frame backchain at run
23410 time, not just for debugging purposes. Such code is not call-compatible
23411 with code compiled with @option{-mpacked-stack}. Also, note that the
23412 combination of @option{-mbackchain},
23413 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
23414 to build a linux kernel use @option{-msoft-float}.
23416 The default is to not use the packed stack layout.
23419 @itemx -mno-small-exec
23420 @opindex msmall-exec
23421 @opindex mno-small-exec
23422 Generate (or do not generate) code using the @code{bras} instruction
23423 to do subroutine calls.
23424 This only works reliably if the total executable size does not
23425 exceed 64k. The default is to use the @code{basr} instruction instead,
23426 which does not have this limitation.
23432 When @option{-m31} is specified, generate code compliant to the
23433 GNU/Linux for S/390 ABI@. When @option{-m64} is specified, generate
23434 code compliant to the GNU/Linux for zSeries ABI@. This allows GCC in
23435 particular to generate 64-bit instructions. For the @samp{s390}
23436 targets, the default is @option{-m31}, while the @samp{s390x}
23437 targets default to @option{-m64}.
23443 When @option{-mzarch} is specified, generate code using the
23444 instructions available on z/Architecture.
23445 When @option{-mesa} is specified, generate code using the
23446 instructions available on ESA/390. Note that @option{-mesa} is
23447 not possible with @option{-m64}.
23448 When generating code compliant to the GNU/Linux for S/390 ABI,
23449 the default is @option{-mesa}. When generating code compliant
23450 to the GNU/Linux for zSeries ABI, the default is @option{-mzarch}.
23456 The @option{-mhtm} option enables a set of builtins making use of
23457 instructions available with the transactional execution facility
23458 introduced with the IBM zEnterprise EC12 machine generation
23459 @ref{S/390 System z Built-in Functions}.
23460 @option{-mhtm} is enabled by default when using @option{-march=zEC12}.
23466 When @option{-mvx} is specified, generate code using the instructions
23467 available with the vector extension facility introduced with the IBM
23468 z13 machine generation.
23469 This option changes the ABI for some vector type values with regard to
23470 alignment and calling conventions. In case vector type values are
23471 being used in an ABI-relevant context a GAS @samp{.gnu_attribute}
23472 command will be added to mark the resulting binary with the ABI used.
23473 @option{-mvx} is enabled by default when using @option{-march=z13}.
23476 @itemx -mno-zvector
23478 @opindex mno-zvector
23479 The @option{-mzvector} option enables vector language extensions and
23480 builtins using instructions available with the vector extension
23481 facility introduced with the IBM z13 machine generation.
23482 This option adds support for @samp{vector} to be used as a keyword to
23483 define vector type variables and arguments. @samp{vector} is only
23484 available when GNU extensions are enabled. It will not be expanded
23485 when requesting strict standard compliance e.g. with @option{-std=c99}.
23486 In addition to the GCC low-level builtins @option{-mzvector} enables
23487 a set of builtins added for compatibility with AltiVec-style
23488 implementations like Power and Cell. In order to make use of these
23489 builtins the header file @file{vecintrin.h} needs to be included.
23490 @option{-mzvector} is disabled by default.
23496 Generate (or do not generate) code using the @code{mvcle} instruction
23497 to perform block moves. When @option{-mno-mvcle} is specified,
23498 use a @code{mvc} loop instead. This is the default unless optimizing for
23505 Print (or do not print) additional debug information when compiling.
23506 The default is to not print debug information.
23508 @item -march=@var{cpu-type}
23510 Generate code that runs on @var{cpu-type}, which is the name of a
23511 system representing a certain processor type. Possible values for
23512 @var{cpu-type} are @samp{z900}/@samp{arch5}, @samp{z990}/@samp{arch6},
23513 @samp{z9-109}, @samp{z9-ec}/@samp{arch7}, @samp{z10}/@samp{arch8},
23514 @samp{z196}/@samp{arch9}, @samp{zEC12}, @samp{z13}/@samp{arch11}, and
23517 The default is @option{-march=z900}. @samp{g5}/@samp{arch3} and
23518 @samp{g6} are deprecated and will be removed with future releases.
23520 Specifying @samp{native} as cpu type can be used to select the best
23521 architecture option for the host processor.
23522 @option{-march=native} has no effect if GCC does not recognize the
23525 @item -mtune=@var{cpu-type}
23527 Tune to @var{cpu-type} everything applicable about the generated code,
23528 except for the ABI and the set of available instructions.
23529 The list of @var{cpu-type} values is the same as for @option{-march}.
23530 The default is the value used for @option{-march}.
23533 @itemx -mno-tpf-trace
23534 @opindex mtpf-trace
23535 @opindex mno-tpf-trace
23536 Generate code that adds (does not add) in TPF OS specific branches to trace
23537 routines in the operating system. This option is off by default, even
23538 when compiling for the TPF OS@.
23541 @itemx -mno-fused-madd
23542 @opindex mfused-madd
23543 @opindex mno-fused-madd
23544 Generate code that uses (does not use) the floating-point multiply and
23545 accumulate instructions. These instructions are generated by default if
23546 hardware floating point is used.
23548 @item -mwarn-framesize=@var{framesize}
23549 @opindex mwarn-framesize
23550 Emit a warning if the current function exceeds the given frame size. Because
23551 this is a compile-time check it doesn't need to be a real problem when the program
23552 runs. It is intended to identify functions that most probably cause
23553 a stack overflow. It is useful to be used in an environment with limited stack
23554 size e.g.@: the linux kernel.
23556 @item -mwarn-dynamicstack
23557 @opindex mwarn-dynamicstack
23558 Emit a warning if the function calls @code{alloca} or uses dynamically-sized
23559 arrays. This is generally a bad idea with a limited stack size.
23561 @item -mstack-guard=@var{stack-guard}
23562 @itemx -mstack-size=@var{stack-size}
23563 @opindex mstack-guard
23564 @opindex mstack-size
23565 If these options are provided the S/390 back end emits additional instructions in
23566 the function prologue that trigger a trap if the stack size is @var{stack-guard}
23567 bytes above the @var{stack-size} (remember that the stack on S/390 grows downward).
23568 If the @var{stack-guard} option is omitted the smallest power of 2 larger than
23569 the frame size of the compiled function is chosen.
23570 These options are intended to be used to help debugging stack overflow problems.
23571 The additionally emitted code causes only little overhead and hence can also be
23572 used in production-like systems without greater performance degradation. The given
23573 values have to be exact powers of 2 and @var{stack-size} has to be greater than
23574 @var{stack-guard} without exceeding 64k.
23575 In order to be efficient the extra code makes the assumption that the stack starts
23576 at an address aligned to the value given by @var{stack-size}.
23577 The @var{stack-guard} option can only be used in conjunction with @var{stack-size}.
23579 @item -mhotpatch=@var{pre-halfwords},@var{post-halfwords}
23581 If the hotpatch option is enabled, a ``hot-patching'' function
23582 prologue is generated for all functions in the compilation unit.
23583 The funtion label is prepended with the given number of two-byte
23584 NOP instructions (@var{pre-halfwords}, maximum 1000000). After
23585 the label, 2 * @var{post-halfwords} bytes are appended, using the
23586 largest NOP like instructions the architecture allows (maximum
23589 If both arguments are zero, hotpatching is disabled.
23591 This option can be overridden for individual functions with the
23592 @code{hotpatch} attribute.
23595 @node Score Options
23596 @subsection Score Options
23597 @cindex Score Options
23599 These options are defined for Score implementations:
23604 Compile code for big-endian mode. This is the default.
23608 Compile code for little-endian mode.
23612 Disable generation of @code{bcnz} instructions.
23616 Enable generation of unaligned load and store instructions.
23620 Enable the use of multiply-accumulate instructions. Disabled by default.
23624 Specify the SCORE5 as the target architecture.
23628 Specify the SCORE5U of the target architecture.
23632 Specify the SCORE7 as the target architecture. This is the default.
23636 Specify the SCORE7D as the target architecture.
23640 @subsection SH Options
23642 These @samp{-m} options are defined for the SH implementations:
23647 Generate code for the SH1.
23651 Generate code for the SH2.
23654 Generate code for the SH2e.
23658 Generate code for the SH2a without FPU, or for a SH2a-FPU in such a way
23659 that the floating-point unit is not used.
23661 @item -m2a-single-only
23662 @opindex m2a-single-only
23663 Generate code for the SH2a-FPU, in such a way that no double-precision
23664 floating-point operations are used.
23667 @opindex m2a-single
23668 Generate code for the SH2a-FPU assuming the floating-point unit is in
23669 single-precision mode by default.
23673 Generate code for the SH2a-FPU assuming the floating-point unit is in
23674 double-precision mode by default.
23678 Generate code for the SH3.
23682 Generate code for the SH3e.
23686 Generate code for the SH4 without a floating-point unit.
23688 @item -m4-single-only
23689 @opindex m4-single-only
23690 Generate code for the SH4 with a floating-point unit that only
23691 supports single-precision arithmetic.
23695 Generate code for the SH4 assuming the floating-point unit is in
23696 single-precision mode by default.
23700 Generate code for the SH4.
23704 Generate code for SH4-100.
23706 @item -m4-100-nofpu
23707 @opindex m4-100-nofpu
23708 Generate code for SH4-100 in such a way that the
23709 floating-point unit is not used.
23711 @item -m4-100-single
23712 @opindex m4-100-single
23713 Generate code for SH4-100 assuming the floating-point unit is in
23714 single-precision mode by default.
23716 @item -m4-100-single-only
23717 @opindex m4-100-single-only
23718 Generate code for SH4-100 in such a way that no double-precision
23719 floating-point operations are used.
23723 Generate code for SH4-200.
23725 @item -m4-200-nofpu
23726 @opindex m4-200-nofpu
23727 Generate code for SH4-200 without in such a way that the
23728 floating-point unit is not used.
23730 @item -m4-200-single
23731 @opindex m4-200-single
23732 Generate code for SH4-200 assuming the floating-point unit is in
23733 single-precision mode by default.
23735 @item -m4-200-single-only
23736 @opindex m4-200-single-only
23737 Generate code for SH4-200 in such a way that no double-precision
23738 floating-point operations are used.
23742 Generate code for SH4-300.
23744 @item -m4-300-nofpu
23745 @opindex m4-300-nofpu
23746 Generate code for SH4-300 without in such a way that the
23747 floating-point unit is not used.
23749 @item -m4-300-single
23750 @opindex m4-300-single
23751 Generate code for SH4-300 in such a way that no double-precision
23752 floating-point operations are used.
23754 @item -m4-300-single-only
23755 @opindex m4-300-single-only
23756 Generate code for SH4-300 in such a way that no double-precision
23757 floating-point operations are used.
23761 Generate code for SH4-340 (no MMU, no FPU).
23765 Generate code for SH4-500 (no FPU). Passes @option{-isa=sh4-nofpu} to the
23770 Generate code for the SH4al-dsp, or for a SH4a in such a way that the
23771 floating-point unit is not used.
23773 @item -m4a-single-only
23774 @opindex m4a-single-only
23775 Generate code for the SH4a, in such a way that no double-precision
23776 floating-point operations are used.
23779 @opindex m4a-single
23780 Generate code for the SH4a assuming the floating-point unit is in
23781 single-precision mode by default.
23785 Generate code for the SH4a.
23789 Same as @option{-m4a-nofpu}, except that it implicitly passes
23790 @option{-dsp} to the assembler. GCC doesn't generate any DSP
23791 instructions at the moment.
23795 Compile code for the processor in big-endian mode.
23799 Compile code for the processor in little-endian mode.
23803 Align doubles at 64-bit boundaries. Note that this changes the calling
23804 conventions, and thus some functions from the standard C library do
23805 not work unless you recompile it first with @option{-mdalign}.
23809 Shorten some address references at link time, when possible; uses the
23810 linker option @option{-relax}.
23814 Use 32-bit offsets in @code{switch} tables. The default is to use
23819 Enable the use of bit manipulation instructions on SH2A.
23823 Enable the use of the instruction @code{fmovd}. Check @option{-mdalign} for
23824 alignment constraints.
23828 Comply with the calling conventions defined by Renesas.
23831 @opindex mno-renesas
23832 Comply with the calling conventions defined for GCC before the Renesas
23833 conventions were available. This option is the default for all
23834 targets of the SH toolchain.
23837 @opindex mnomacsave
23838 Mark the @code{MAC} register as call-clobbered, even if
23839 @option{-mrenesas} is given.
23845 Control the IEEE compliance of floating-point comparisons, which affects the
23846 handling of cases where the result of a comparison is unordered. By default
23847 @option{-mieee} is implicitly enabled. If @option{-ffinite-math-only} is
23848 enabled @option{-mno-ieee} is implicitly set, which results in faster
23849 floating-point greater-equal and less-equal comparisons. The implicit settings
23850 can be overridden by specifying either @option{-mieee} or @option{-mno-ieee}.
23852 @item -minline-ic_invalidate
23853 @opindex minline-ic_invalidate
23854 Inline code to invalidate instruction cache entries after setting up
23855 nested function trampolines.
23856 This option has no effect if @option{-musermode} is in effect and the selected
23857 code generation option (e.g. @option{-m4}) does not allow the use of the @code{icbi}
23859 If the selected code generation option does not allow the use of the @code{icbi}
23860 instruction, and @option{-musermode} is not in effect, the inlined code
23861 manipulates the instruction cache address array directly with an associative
23862 write. This not only requires privileged mode at run time, but it also
23863 fails if the cache line had been mapped via the TLB and has become unmapped.
23867 Dump instruction size and location in the assembly code.
23870 @opindex mpadstruct
23871 This option is deprecated. It pads structures to multiple of 4 bytes,
23872 which is incompatible with the SH ABI@.
23874 @item -matomic-model=@var{model}
23875 @opindex matomic-model=@var{model}
23876 Sets the model of atomic operations and additional parameters as a comma
23877 separated list. For details on the atomic built-in functions see
23878 @ref{__atomic Builtins}. The following models and parameters are supported:
23883 Disable compiler generated atomic sequences and emit library calls for atomic
23884 operations. This is the default if the target is not @code{sh*-*-linux*}.
23887 Generate GNU/Linux compatible gUSA software atomic sequences for the atomic
23888 built-in functions. The generated atomic sequences require additional support
23889 from the interrupt/exception handling code of the system and are only suitable
23890 for SH3* and SH4* single-core systems. This option is enabled by default when
23891 the target is @code{sh*-*-linux*} and SH3* or SH4*. When the target is SH4A,
23892 this option also partially utilizes the hardware atomic instructions
23893 @code{movli.l} and @code{movco.l} to create more efficient code, unless
23894 @samp{strict} is specified.
23897 Generate software atomic sequences that use a variable in the thread control
23898 block. This is a variation of the gUSA sequences which can also be used on
23899 SH1* and SH2* targets. The generated atomic sequences require additional
23900 support from the interrupt/exception handling code of the system and are only
23901 suitable for single-core systems. When using this model, the @samp{gbr-offset=}
23902 parameter has to be specified as well.
23905 Generate software atomic sequences that temporarily disable interrupts by
23906 setting @code{SR.IMASK = 1111}. This model works only when the program runs
23907 in privileged mode and is only suitable for single-core systems. Additional
23908 support from the interrupt/exception handling code of the system is not
23909 required. This model is enabled by default when the target is
23910 @code{sh*-*-linux*} and SH1* or SH2*.
23913 Generate hardware atomic sequences using the @code{movli.l} and @code{movco.l}
23914 instructions only. This is only available on SH4A and is suitable for
23915 multi-core systems. Since the hardware instructions support only 32 bit atomic
23916 variables access to 8 or 16 bit variables is emulated with 32 bit accesses.
23917 Code compiled with this option is also compatible with other software
23918 atomic model interrupt/exception handling systems if executed on an SH4A
23919 system. Additional support from the interrupt/exception handling code of the
23920 system is not required for this model.
23923 This parameter specifies the offset in bytes of the variable in the thread
23924 control block structure that should be used by the generated atomic sequences
23925 when the @samp{soft-tcb} model has been selected. For other models this
23926 parameter is ignored. The specified value must be an integer multiple of four
23927 and in the range 0-1020.
23930 This parameter prevents mixed usage of multiple atomic models, even if they
23931 are compatible, and makes the compiler generate atomic sequences of the
23932 specified model only.
23938 Generate the @code{tas.b} opcode for @code{__atomic_test_and_set}.
23939 Notice that depending on the particular hardware and software configuration
23940 this can degrade overall performance due to the operand cache line flushes
23941 that are implied by the @code{tas.b} instruction. On multi-core SH4A
23942 processors the @code{tas.b} instruction must be used with caution since it
23943 can result in data corruption for certain cache configurations.
23946 @opindex mprefergot
23947 When generating position-independent code, emit function calls using
23948 the Global Offset Table instead of the Procedure Linkage Table.
23951 @itemx -mno-usermode
23953 @opindex mno-usermode
23954 Don't allow (allow) the compiler generating privileged mode code. Specifying
23955 @option{-musermode} also implies @option{-mno-inline-ic_invalidate} if the
23956 inlined code would not work in user mode. @option{-musermode} is the default
23957 when the target is @code{sh*-*-linux*}. If the target is SH1* or SH2*
23958 @option{-musermode} has no effect, since there is no user mode.
23960 @item -multcost=@var{number}
23961 @opindex multcost=@var{number}
23962 Set the cost to assume for a multiply insn.
23964 @item -mdiv=@var{strategy}
23965 @opindex mdiv=@var{strategy}
23966 Set the division strategy to be used for integer division operations.
23967 @var{strategy} can be one of:
23972 Calls a library function that uses the single-step division instruction
23973 @code{div1} to perform the operation. Division by zero calculates an
23974 unspecified result and does not trap. This is the default except for SH4,
23975 SH2A and SHcompact.
23978 Calls a library function that performs the operation in double precision
23979 floating point. Division by zero causes a floating-point exception. This is
23980 the default for SHcompact with FPU. Specifying this for targets that do not
23981 have a double precision FPU defaults to @code{call-div1}.
23984 Calls a library function that uses a lookup table for small divisors and
23985 the @code{div1} instruction with case distinction for larger divisors. Division
23986 by zero calculates an unspecified result and does not trap. This is the default
23987 for SH4. Specifying this for targets that do not have dynamic shift
23988 instructions defaults to @code{call-div1}.
23992 When a division strategy has not been specified the default strategy is
23993 selected based on the current target. For SH2A the default strategy is to
23994 use the @code{divs} and @code{divu} instructions instead of library function
23997 @item -maccumulate-outgoing-args
23998 @opindex maccumulate-outgoing-args
23999 Reserve space once for outgoing arguments in the function prologue rather
24000 than around each call. Generally beneficial for performance and size. Also
24001 needed for unwinding to avoid changing the stack frame around conditional code.
24003 @item -mdivsi3_libfunc=@var{name}
24004 @opindex mdivsi3_libfunc=@var{name}
24005 Set the name of the library function used for 32-bit signed division to
24007 This only affects the name used in the @samp{call} division strategies, and
24008 the compiler still expects the same sets of input/output/clobbered registers as
24009 if this option were not present.
24011 @item -mfixed-range=@var{register-range}
24012 @opindex mfixed-range
24013 Generate code treating the given register range as fixed registers.
24014 A fixed register is one that the register allocator can not use. This is
24015 useful when compiling kernel code. A register range is specified as
24016 two registers separated by a dash. Multiple register ranges can be
24017 specified separated by a comma.
24019 @item -mbranch-cost=@var{num}
24020 @opindex mbranch-cost=@var{num}
24021 Assume @var{num} to be the cost for a branch instruction. Higher numbers
24022 make the compiler try to generate more branch-free code if possible.
24023 If not specified the value is selected depending on the processor type that
24024 is being compiled for.
24027 @itemx -mno-zdcbranch
24028 @opindex mzdcbranch
24029 @opindex mno-zdcbranch
24030 Assume (do not assume) that zero displacement conditional branch instructions
24031 @code{bt} and @code{bf} are fast. If @option{-mzdcbranch} is specified, the
24032 compiler prefers zero displacement branch code sequences. This is
24033 enabled by default when generating code for SH4 and SH4A. It can be explicitly
24034 disabled by specifying @option{-mno-zdcbranch}.
24036 @item -mcbranch-force-delay-slot
24037 @opindex mcbranch-force-delay-slot
24038 Force the usage of delay slots for conditional branches, which stuffs the delay
24039 slot with a @code{nop} if a suitable instruction cannot be found. By default
24040 this option is disabled. It can be enabled to work around hardware bugs as
24041 found in the original SH7055.
24044 @itemx -mno-fused-madd
24045 @opindex mfused-madd
24046 @opindex mno-fused-madd
24047 Generate code that uses (does not use) the floating-point multiply and
24048 accumulate instructions. These instructions are generated by default
24049 if hardware floating point is used. The machine-dependent
24050 @option{-mfused-madd} option is now mapped to the machine-independent
24051 @option{-ffp-contract=fast} option, and @option{-mno-fused-madd} is
24052 mapped to @option{-ffp-contract=off}.
24058 Allow or disallow the compiler to emit the @code{fsca} instruction for sine
24059 and cosine approximations. The option @option{-mfsca} must be used in
24060 combination with @option{-funsafe-math-optimizations}. It is enabled by default
24061 when generating code for SH4A. Using @option{-mno-fsca} disables sine and cosine
24062 approximations even if @option{-funsafe-math-optimizations} is in effect.
24068 Allow or disallow the compiler to emit the @code{fsrra} instruction for
24069 reciprocal square root approximations. The option @option{-mfsrra} must be used
24070 in combination with @option{-funsafe-math-optimizations} and
24071 @option{-ffinite-math-only}. It is enabled by default when generating code for
24072 SH4A. Using @option{-mno-fsrra} disables reciprocal square root approximations
24073 even if @option{-funsafe-math-optimizations} and @option{-ffinite-math-only} are
24076 @item -mpretend-cmove
24077 @opindex mpretend-cmove
24078 Prefer zero-displacement conditional branches for conditional move instruction
24079 patterns. This can result in faster code on the SH4 processor.
24083 Generate code using the FDPIC ABI.
24087 @node Solaris 2 Options
24088 @subsection Solaris 2 Options
24089 @cindex Solaris 2 options
24091 These @samp{-m} options are supported on Solaris 2:
24094 @item -mclear-hwcap
24095 @opindex mclear-hwcap
24096 @option{-mclear-hwcap} tells the compiler to remove the hardware
24097 capabilities generated by the Solaris assembler. This is only necessary
24098 when object files use ISA extensions not supported by the current
24099 machine, but check at runtime whether or not to use them.
24101 @item -mimpure-text
24102 @opindex mimpure-text
24103 @option{-mimpure-text}, used in addition to @option{-shared}, tells
24104 the compiler to not pass @option{-z text} to the linker when linking a
24105 shared object. Using this option, you can link position-dependent
24106 code into a shared object.
24108 @option{-mimpure-text} suppresses the ``relocations remain against
24109 allocatable but non-writable sections'' linker error message.
24110 However, the necessary relocations trigger copy-on-write, and the
24111 shared object is not actually shared across processes. Instead of
24112 using @option{-mimpure-text}, you should compile all source code with
24113 @option{-fpic} or @option{-fPIC}.
24117 These switches are supported in addition to the above on Solaris 2:
24122 This is a synonym for @option{-pthread}.
24125 @node SPARC Options
24126 @subsection SPARC Options
24127 @cindex SPARC options
24129 These @samp{-m} options are supported on the SPARC:
24132 @item -mno-app-regs
24134 @opindex mno-app-regs
24136 Specify @option{-mapp-regs} to generate output using the global registers
24137 2 through 4, which the SPARC SVR4 ABI reserves for applications. Like the
24138 global register 1, each global register 2 through 4 is then treated as an
24139 allocable register that is clobbered by function calls. This is the default.
24141 To be fully SVR4 ABI-compliant at the cost of some performance loss,
24142 specify @option{-mno-app-regs}. You should compile libraries and system
24143 software with this option.
24149 With @option{-mflat}, the compiler does not generate save/restore instructions
24150 and uses a ``flat'' or single register window model. This model is compatible
24151 with the regular register window model. The local registers and the input
24152 registers (0--5) are still treated as ``call-saved'' registers and are
24153 saved on the stack as needed.
24155 With @option{-mno-flat} (the default), the compiler generates save/restore
24156 instructions (except for leaf functions). This is the normal operating mode.
24159 @itemx -mhard-float
24161 @opindex mhard-float
24162 Generate output containing floating-point instructions. This is the
24166 @itemx -msoft-float
24168 @opindex msoft-float
24169 Generate output containing library calls for floating point.
24170 @strong{Warning:} the requisite libraries are not available for all SPARC
24171 targets. Normally the facilities of the machine's usual C compiler are
24172 used, but this cannot be done directly in cross-compilation. You must make
24173 your own arrangements to provide suitable library functions for
24174 cross-compilation. The embedded targets @samp{sparc-*-aout} and
24175 @samp{sparclite-*-*} do provide software floating-point support.
24177 @option{-msoft-float} changes the calling convention in the output file;
24178 therefore, it is only useful if you compile @emph{all} of a program with
24179 this option. In particular, you need to compile @file{libgcc.a}, the
24180 library that comes with GCC, with @option{-msoft-float} in order for
24183 @item -mhard-quad-float
24184 @opindex mhard-quad-float
24185 Generate output containing quad-word (long double) floating-point
24188 @item -msoft-quad-float
24189 @opindex msoft-quad-float
24190 Generate output containing library calls for quad-word (long double)
24191 floating-point instructions. The functions called are those specified
24192 in the SPARC ABI@. This is the default.
24194 As of this writing, there are no SPARC implementations that have hardware
24195 support for the quad-word floating-point instructions. They all invoke
24196 a trap handler for one of these instructions, and then the trap handler
24197 emulates the effect of the instruction. Because of the trap handler overhead,
24198 this is much slower than calling the ABI library routines. Thus the
24199 @option{-msoft-quad-float} option is the default.
24201 @item -mno-unaligned-doubles
24202 @itemx -munaligned-doubles
24203 @opindex mno-unaligned-doubles
24204 @opindex munaligned-doubles
24205 Assume that doubles have 8-byte alignment. This is the default.
24207 With @option{-munaligned-doubles}, GCC assumes that doubles have 8-byte
24208 alignment only if they are contained in another type, or if they have an
24209 absolute address. Otherwise, it assumes they have 4-byte alignment.
24210 Specifying this option avoids some rare compatibility problems with code
24211 generated by other compilers. It is not the default because it results
24212 in a performance loss, especially for floating-point code.
24215 @itemx -mno-user-mode
24216 @opindex muser-mode
24217 @opindex mno-user-mode
24218 Do not generate code that can only run in supervisor mode. This is relevant
24219 only for the @code{casa} instruction emitted for the LEON3 processor. This
24222 @item -mfaster-structs
24223 @itemx -mno-faster-structs
24224 @opindex mfaster-structs
24225 @opindex mno-faster-structs
24226 With @option{-mfaster-structs}, the compiler assumes that structures
24227 should have 8-byte alignment. This enables the use of pairs of
24228 @code{ldd} and @code{std} instructions for copies in structure
24229 assignment, in place of twice as many @code{ld} and @code{st} pairs.
24230 However, the use of this changed alignment directly violates the SPARC
24231 ABI@. Thus, it's intended only for use on targets where the developer
24232 acknowledges that their resulting code is not directly in line with
24233 the rules of the ABI@.
24235 @item -mstd-struct-return
24236 @itemx -mno-std-struct-return
24237 @opindex mstd-struct-return
24238 @opindex mno-std-struct-return
24239 With @option{-mstd-struct-return}, the compiler generates checking code
24240 in functions returning structures or unions to detect size mismatches
24241 between the two sides of function calls, as per the 32-bit ABI@.
24243 The default is @option{-mno-std-struct-return}. This option has no effect
24250 Enable Local Register Allocation. This is the default for SPARC since GCC 7
24251 so @option{-mno-lra} needs to be passed to get old Reload.
24253 @item -mcpu=@var{cpu_type}
24255 Set the instruction set, register set, and instruction scheduling parameters
24256 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
24257 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{hypersparc},
24258 @samp{leon}, @samp{leon3}, @samp{leon3v7}, @samp{sparclite}, @samp{f930},
24259 @samp{f934}, @samp{sparclite86x}, @samp{sparclet}, @samp{tsc701}, @samp{v9},
24260 @samp{ultrasparc}, @samp{ultrasparc3}, @samp{niagara}, @samp{niagara2},
24261 @samp{niagara3}, @samp{niagara4}, @samp{niagara7} and @samp{m8}.
24263 Native Solaris and GNU/Linux toolchains also support the value @samp{native},
24264 which selects the best architecture option for the host processor.
24265 @option{-mcpu=native} has no effect if GCC does not recognize
24268 Default instruction scheduling parameters are used for values that select
24269 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
24270 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
24272 Here is a list of each supported architecture and their supported
24280 supersparc, hypersparc, leon, leon3
24283 f930, f934, sparclite86x
24289 ultrasparc, ultrasparc3, niagara, niagara2, niagara3, niagara4,
24293 By default (unless configured otherwise), GCC generates code for the V7
24294 variant of the SPARC architecture. With @option{-mcpu=cypress}, the compiler
24295 additionally optimizes it for the Cypress CY7C602 chip, as used in the
24296 SPARCStation/SPARCServer 3xx series. This is also appropriate for the older
24297 SPARCStation 1, 2, IPX etc.
24299 With @option{-mcpu=v8}, GCC generates code for the V8 variant of the SPARC
24300 architecture. The only difference from V7 code is that the compiler emits
24301 the integer multiply and integer divide instructions which exist in SPARC-V8
24302 but not in SPARC-V7. With @option{-mcpu=supersparc}, the compiler additionally
24303 optimizes it for the SuperSPARC chip, as used in the SPARCStation 10, 1000 and
24306 With @option{-mcpu=sparclite}, GCC generates code for the SPARClite variant of
24307 the SPARC architecture. This adds the integer multiply, integer divide step
24308 and scan (@code{ffs}) instructions which exist in SPARClite but not in SPARC-V7.
24309 With @option{-mcpu=f930}, the compiler additionally optimizes it for the
24310 Fujitsu MB86930 chip, which is the original SPARClite, with no FPU@. With
24311 @option{-mcpu=f934}, the compiler additionally optimizes it for the Fujitsu
24312 MB86934 chip, which is the more recent SPARClite with FPU@.
24314 With @option{-mcpu=sparclet}, GCC generates code for the SPARClet variant of
24315 the SPARC architecture. This adds the integer multiply, multiply/accumulate,
24316 integer divide step and scan (@code{ffs}) instructions which exist in SPARClet
24317 but not in SPARC-V7. With @option{-mcpu=tsc701}, the compiler additionally
24318 optimizes it for the TEMIC SPARClet chip.
24320 With @option{-mcpu=v9}, GCC generates code for the V9 variant of the SPARC
24321 architecture. This adds 64-bit integer and floating-point move instructions,
24322 3 additional floating-point condition code registers and conditional move
24323 instructions. With @option{-mcpu=ultrasparc}, the compiler additionally
24324 optimizes it for the Sun UltraSPARC I/II/IIi chips. With
24325 @option{-mcpu=ultrasparc3}, the compiler additionally optimizes it for the
24326 Sun UltraSPARC III/III+/IIIi/IIIi+/IV/IV+ chips. With
24327 @option{-mcpu=niagara}, the compiler additionally optimizes it for
24328 Sun UltraSPARC T1 chips. With @option{-mcpu=niagara2}, the compiler
24329 additionally optimizes it for Sun UltraSPARC T2 chips. With
24330 @option{-mcpu=niagara3}, the compiler additionally optimizes it for Sun
24331 UltraSPARC T3 chips. With @option{-mcpu=niagara4}, the compiler
24332 additionally optimizes it for Sun UltraSPARC T4 chips. With
24333 @option{-mcpu=niagara7}, the compiler additionally optimizes it for
24334 Oracle SPARC M7 chips. With @option{-mcpu=m8}, the compiler
24335 additionally optimizes it for Oracle M8 chips.
24337 @item -mtune=@var{cpu_type}
24339 Set the instruction scheduling parameters for machine type
24340 @var{cpu_type}, but do not set the instruction set or register set that the
24341 option @option{-mcpu=@var{cpu_type}} does.
24343 The same values for @option{-mcpu=@var{cpu_type}} can be used for
24344 @option{-mtune=@var{cpu_type}}, but the only useful values are those
24345 that select a particular CPU implementation. Those are
24346 @samp{cypress}, @samp{supersparc}, @samp{hypersparc}, @samp{leon},
24347 @samp{leon3}, @samp{leon3v7}, @samp{f930}, @samp{f934},
24348 @samp{sparclite86x}, @samp{tsc701}, @samp{ultrasparc},
24349 @samp{ultrasparc3}, @samp{niagara}, @samp{niagara2}, @samp{niagara3},
24350 @samp{niagara4}, @samp{niagara7} and @samp{m8}. With native Solaris
24351 and GNU/Linux toolchains, @samp{native} can also be used.
24356 @opindex mno-v8plus
24357 With @option{-mv8plus}, GCC generates code for the SPARC-V8+ ABI@. The
24358 difference from the V8 ABI is that the global and out registers are
24359 considered 64 bits wide. This is enabled by default on Solaris in 32-bit
24360 mode for all SPARC-V9 processors.
24366 With @option{-mvis}, GCC generates code that takes advantage of the UltraSPARC
24367 Visual Instruction Set extensions. The default is @option{-mno-vis}.
24373 With @option{-mvis2}, GCC generates code that takes advantage of
24374 version 2.0 of the UltraSPARC Visual Instruction Set extensions. The
24375 default is @option{-mvis2} when targeting a cpu that supports such
24376 instructions, such as UltraSPARC-III and later. Setting @option{-mvis2}
24377 also sets @option{-mvis}.
24383 With @option{-mvis3}, GCC generates code that takes advantage of
24384 version 3.0 of the UltraSPARC Visual Instruction Set extensions. The
24385 default is @option{-mvis3} when targeting a cpu that supports such
24386 instructions, such as niagara-3 and later. Setting @option{-mvis3}
24387 also sets @option{-mvis2} and @option{-mvis}.
24393 With @option{-mvis4}, GCC generates code that takes advantage of
24394 version 4.0 of the UltraSPARC Visual Instruction Set extensions. The
24395 default is @option{-mvis4} when targeting a cpu that supports such
24396 instructions, such as niagara-7 and later. Setting @option{-mvis4}
24397 also sets @option{-mvis3}, @option{-mvis2} and @option{-mvis}.
24403 With @option{-mvis4b}, GCC generates code that takes advantage of
24404 version 4.0 of the UltraSPARC Visual Instruction Set extensions, plus
24405 the additional VIS instructions introduced in the Oracle SPARC
24406 Architecture 2017. The default is @option{-mvis4b} when targeting a
24407 cpu that supports such instructions, such as m8 and later. Setting
24408 @option{-mvis4b} also sets @option{-mvis4}, @option{-mvis3},
24409 @option{-mvis2} and @option{-mvis}.
24414 @opindex mno-cbcond
24415 With @option{-mcbcond}, GCC generates code that takes advantage of the UltraSPARC
24416 Compare-and-Branch-on-Condition instructions. The default is @option{-mcbcond}
24417 when targeting a CPU that supports such instructions, such as Niagara-4 and
24424 With @option{-mfmaf}, GCC generates code that takes advantage of the UltraSPARC
24425 Fused Multiply-Add Floating-point instructions. The default is @option{-mfmaf}
24426 when targeting a CPU that supports such instructions, such as Niagara-3 and
24432 @opindex mno-fsmuld
24433 With @option{-mfsmuld}, GCC generates code that takes advantage of the
24434 Floating-point Multiply Single to Double (FsMULd) instruction. The default is
24435 @option{-mfsmuld} when targeting a CPU supporting the architecture versions V8
24436 or V9 with FPU except @option{-mcpu=leon}.
24442 With @option{-mpopc}, GCC generates code that takes advantage of the UltraSPARC
24443 Population Count instruction. The default is @option{-mpopc}
24444 when targeting a CPU that supports such an instruction, such as Niagara-2 and
24451 With @option{-msubxc}, GCC generates code that takes advantage of the UltraSPARC
24452 Subtract-Extended-with-Carry instruction. The default is @option{-msubxc}
24453 when targeting a CPU that supports such an instruction, such as Niagara-7 and
24457 @opindex mfix-at697f
24458 Enable the documented workaround for the single erratum of the Atmel AT697F
24459 processor (which corresponds to erratum #13 of the AT697E processor).
24462 @opindex mfix-ut699
24463 Enable the documented workarounds for the floating-point errata and the data
24464 cache nullify errata of the UT699 processor.
24467 @opindex mfix-ut700
24468 Enable the documented workaround for the back-to-back store errata of
24469 the UT699E/UT700 processor.
24471 @item -mfix-gr712rc
24472 @opindex mfix-gr712rc
24473 Enable the documented workaround for the back-to-back store errata of
24474 the GR712RC processor.
24477 These @samp{-m} options are supported in addition to the above
24478 on SPARC-V9 processors in 64-bit environments:
24485 Generate code for a 32-bit or 64-bit environment.
24486 The 32-bit environment sets int, long and pointer to 32 bits.
24487 The 64-bit environment sets int to 32 bits and long and pointer
24490 @item -mcmodel=@var{which}
24492 Set the code model to one of
24496 The Medium/Low code model: 64-bit addresses, programs
24497 must be linked in the low 32 bits of memory. Programs can be statically
24498 or dynamically linked.
24501 The Medium/Middle code model: 64-bit addresses, programs
24502 must be linked in the low 44 bits of memory, the text and data segments must
24503 be less than 2GB in size and the data segment must be located within 2GB of
24507 The Medium/Anywhere code model: 64-bit addresses, programs
24508 may be linked anywhere in memory, the text and data segments must be less
24509 than 2GB in size and the data segment must be located within 2GB of the
24513 The Medium/Anywhere code model for embedded systems:
24514 64-bit addresses, the text and data segments must be less than 2GB in
24515 size, both starting anywhere in memory (determined at link time). The
24516 global register %g4 points to the base of the data segment. Programs
24517 are statically linked and PIC is not supported.
24520 @item -mmemory-model=@var{mem-model}
24521 @opindex mmemory-model
24522 Set the memory model in force on the processor to one of
24526 The default memory model for the processor and operating system.
24529 Relaxed Memory Order
24532 Partial Store Order
24538 Sequential Consistency
24541 These memory models are formally defined in Appendix D of the SPARC-V9
24542 architecture manual, as set in the processor's @code{PSTATE.MM} field.
24545 @itemx -mno-stack-bias
24546 @opindex mstack-bias
24547 @opindex mno-stack-bias
24548 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
24549 frame pointer if present, are offset by @minus{}2047 which must be added back
24550 when making stack frame references. This is the default in 64-bit mode.
24551 Otherwise, assume no such offset is present.
24555 @subsection SPU Options
24556 @cindex SPU options
24558 These @samp{-m} options are supported on the SPU:
24562 @itemx -merror-reloc
24563 @opindex mwarn-reloc
24564 @opindex merror-reloc
24566 The loader for SPU does not handle dynamic relocations. By default, GCC
24567 gives an error when it generates code that requires a dynamic
24568 relocation. @option{-mno-error-reloc} disables the error,
24569 @option{-mwarn-reloc} generates a warning instead.
24572 @itemx -munsafe-dma
24574 @opindex munsafe-dma
24576 Instructions that initiate or test completion of DMA must not be
24577 reordered with respect to loads and stores of the memory that is being
24579 With @option{-munsafe-dma} you must use the @code{volatile} keyword to protect
24580 memory accesses, but that can lead to inefficient code in places where the
24581 memory is known to not change. Rather than mark the memory as volatile,
24582 you can use @option{-msafe-dma} to tell the compiler to treat
24583 the DMA instructions as potentially affecting all memory.
24585 @item -mbranch-hints
24586 @opindex mbranch-hints
24588 By default, GCC generates a branch hint instruction to avoid
24589 pipeline stalls for always-taken or probably-taken branches. A hint
24590 is not generated closer than 8 instructions away from its branch.
24591 There is little reason to disable them, except for debugging purposes,
24592 or to make an object a little bit smaller.
24596 @opindex msmall-mem
24597 @opindex mlarge-mem
24599 By default, GCC generates code assuming that addresses are never larger
24600 than 18 bits. With @option{-mlarge-mem} code is generated that assumes
24601 a full 32-bit address.
24606 By default, GCC links against startup code that assumes the SPU-style
24607 main function interface (which has an unconventional parameter list).
24608 With @option{-mstdmain}, GCC links your program against startup
24609 code that assumes a C99-style interface to @code{main}, including a
24610 local copy of @code{argv} strings.
24612 @item -mfixed-range=@var{register-range}
24613 @opindex mfixed-range
24614 Generate code treating the given register range as fixed registers.
24615 A fixed register is one that the register allocator cannot use. This is
24616 useful when compiling kernel code. A register range is specified as
24617 two registers separated by a dash. Multiple register ranges can be
24618 specified separated by a comma.
24624 Compile code assuming that pointers to the PPU address space accessed
24625 via the @code{__ea} named address space qualifier are either 32 or 64
24626 bits wide. The default is 32 bits. As this is an ABI-changing option,
24627 all object code in an executable must be compiled with the same setting.
24629 @item -maddress-space-conversion
24630 @itemx -mno-address-space-conversion
24631 @opindex maddress-space-conversion
24632 @opindex mno-address-space-conversion
24633 Allow/disallow treating the @code{__ea} address space as superset
24634 of the generic address space. This enables explicit type casts
24635 between @code{__ea} and generic pointer as well as implicit
24636 conversions of generic pointers to @code{__ea} pointers. The
24637 default is to allow address space pointer conversions.
24639 @item -mcache-size=@var{cache-size}
24640 @opindex mcache-size
24641 This option controls the version of libgcc that the compiler links to an
24642 executable and selects a software-managed cache for accessing variables
24643 in the @code{__ea} address space with a particular cache size. Possible
24644 options for @var{cache-size} are @samp{8}, @samp{16}, @samp{32}, @samp{64}
24645 and @samp{128}. The default cache size is 64KB.
24647 @item -matomic-updates
24648 @itemx -mno-atomic-updates
24649 @opindex matomic-updates
24650 @opindex mno-atomic-updates
24651 This option controls the version of libgcc that the compiler links to an
24652 executable and selects whether atomic updates to the software-managed
24653 cache of PPU-side variables are used. If you use atomic updates, changes
24654 to a PPU variable from SPU code using the @code{__ea} named address space
24655 qualifier do not interfere with changes to other PPU variables residing
24656 in the same cache line from PPU code. If you do not use atomic updates,
24657 such interference may occur; however, writing back cache lines is
24658 more efficient. The default behavior is to use atomic updates.
24661 @itemx -mdual-nops=@var{n}
24662 @opindex mdual-nops
24663 By default, GCC inserts NOPs to increase dual issue when it expects
24664 it to increase performance. @var{n} can be a value from 0 to 10. A
24665 smaller @var{n} inserts fewer NOPs. 10 is the default, 0 is the
24666 same as @option{-mno-dual-nops}. Disabled with @option{-Os}.
24668 @item -mhint-max-nops=@var{n}
24669 @opindex mhint-max-nops
24670 Maximum number of NOPs to insert for a branch hint. A branch hint must
24671 be at least 8 instructions away from the branch it is affecting. GCC
24672 inserts up to @var{n} NOPs to enforce this, otherwise it does not
24673 generate the branch hint.
24675 @item -mhint-max-distance=@var{n}
24676 @opindex mhint-max-distance
24677 The encoding of the branch hint instruction limits the hint to be within
24678 256 instructions of the branch it is affecting. By default, GCC makes
24679 sure it is within 125.
24682 @opindex msafe-hints
24683 Work around a hardware bug that causes the SPU to stall indefinitely.
24684 By default, GCC inserts the @code{hbrp} instruction to make sure
24685 this stall won't happen.
24689 @node System V Options
24690 @subsection Options for System V
24692 These additional options are available on System V Release 4 for
24693 compatibility with other compilers on those systems:
24698 Create a shared object.
24699 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
24703 Identify the versions of each tool used by the compiler, in a
24704 @code{.ident} assembler directive in the output.
24708 Refrain from adding @code{.ident} directives to the output file (this is
24711 @item -YP,@var{dirs}
24713 Search the directories @var{dirs}, and no others, for libraries
24714 specified with @option{-l}.
24716 @item -Ym,@var{dir}
24718 Look in the directory @var{dir} to find the M4 preprocessor.
24719 The assembler uses this option.
24720 @c This is supposed to go with a -Yd for predefined M4 macro files, but
24721 @c the generic assembler that comes with Solaris takes just -Ym.
24724 @node TILE-Gx Options
24725 @subsection TILE-Gx Options
24726 @cindex TILE-Gx options
24728 These @samp{-m} options are supported on the TILE-Gx:
24731 @item -mcmodel=small
24732 @opindex mcmodel=small
24733 Generate code for the small model. The distance for direct calls is
24734 limited to 500M in either direction. PC-relative addresses are 32
24735 bits. Absolute addresses support the full address range.
24737 @item -mcmodel=large
24738 @opindex mcmodel=large
24739 Generate code for the large model. There is no limitation on call
24740 distance, pc-relative addresses, or absolute addresses.
24742 @item -mcpu=@var{name}
24744 Selects the type of CPU to be targeted. Currently the only supported
24745 type is @samp{tilegx}.
24751 Generate code for a 32-bit or 64-bit environment. The 32-bit
24752 environment sets int, long, and pointer to 32 bits. The 64-bit
24753 environment sets int to 32 bits and long and pointer to 64 bits.
24756 @itemx -mlittle-endian
24757 @opindex mbig-endian
24758 @opindex mlittle-endian
24759 Generate code in big/little endian mode, respectively.
24762 @node TILEPro Options
24763 @subsection TILEPro Options
24764 @cindex TILEPro options
24766 These @samp{-m} options are supported on the TILEPro:
24769 @item -mcpu=@var{name}
24771 Selects the type of CPU to be targeted. Currently the only supported
24772 type is @samp{tilepro}.
24776 Generate code for a 32-bit environment, which sets int, long, and
24777 pointer to 32 bits. This is the only supported behavior so the flag
24778 is essentially ignored.
24782 @subsection V850 Options
24783 @cindex V850 Options
24785 These @samp{-m} options are defined for V850 implementations:
24789 @itemx -mno-long-calls
24790 @opindex mlong-calls
24791 @opindex mno-long-calls
24792 Treat all calls as being far away (near). If calls are assumed to be
24793 far away, the compiler always loads the function's address into a
24794 register, and calls indirect through the pointer.
24800 Do not optimize (do optimize) basic blocks that use the same index
24801 pointer 4 or more times to copy pointer into the @code{ep} register, and
24802 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
24803 option is on by default if you optimize.
24805 @item -mno-prolog-function
24806 @itemx -mprolog-function
24807 @opindex mno-prolog-function
24808 @opindex mprolog-function
24809 Do not use (do use) external functions to save and restore registers
24810 at the prologue and epilogue of a function. The external functions
24811 are slower, but use less code space if more than one function saves
24812 the same number of registers. The @option{-mprolog-function} option
24813 is on by default if you optimize.
24817 Try to make the code as small as possible. At present, this just turns
24818 on the @option{-mep} and @option{-mprolog-function} options.
24820 @item -mtda=@var{n}
24822 Put static or global variables whose size is @var{n} bytes or less into
24823 the tiny data area that register @code{ep} points to. The tiny data
24824 area can hold up to 256 bytes in total (128 bytes for byte references).
24826 @item -msda=@var{n}
24828 Put static or global variables whose size is @var{n} bytes or less into
24829 the small data area that register @code{gp} points to. The small data
24830 area can hold up to 64 kilobytes.
24832 @item -mzda=@var{n}
24834 Put static or global variables whose size is @var{n} bytes or less into
24835 the first 32 kilobytes of memory.
24839 Specify that the target processor is the V850.
24843 Specify that the target processor is the V850E3V5. The preprocessor
24844 constant @code{__v850e3v5__} is defined if this option is used.
24848 Specify that the target processor is the V850E3V5. This is an alias for
24849 the @option{-mv850e3v5} option.
24853 Specify that the target processor is the V850E2V3. The preprocessor
24854 constant @code{__v850e2v3__} is defined if this option is used.
24858 Specify that the target processor is the V850E2. The preprocessor
24859 constant @code{__v850e2__} is defined if this option is used.
24863 Specify that the target processor is the V850E1. The preprocessor
24864 constants @code{__v850e1__} and @code{__v850e__} are defined if
24865 this option is used.
24869 Specify that the target processor is the V850ES. This is an alias for
24870 the @option{-mv850e1} option.
24874 Specify that the target processor is the V850E@. The preprocessor
24875 constant @code{__v850e__} is defined if this option is used.
24877 If neither @option{-mv850} nor @option{-mv850e} nor @option{-mv850e1}
24878 nor @option{-mv850e2} nor @option{-mv850e2v3} nor @option{-mv850e3v5}
24879 are defined then a default target processor is chosen and the
24880 relevant @samp{__v850*__} preprocessor constant is defined.
24882 The preprocessor constants @code{__v850} and @code{__v851__} are always
24883 defined, regardless of which processor variant is the target.
24885 @item -mdisable-callt
24886 @itemx -mno-disable-callt
24887 @opindex mdisable-callt
24888 @opindex mno-disable-callt
24889 This option suppresses generation of the @code{CALLT} instruction for the
24890 v850e, v850e1, v850e2, v850e2v3 and v850e3v5 flavors of the v850
24893 This option is enabled by default when the RH850 ABI is
24894 in use (see @option{-mrh850-abi}), and disabled by default when the
24895 GCC ABI is in use. If @code{CALLT} instructions are being generated
24896 then the C preprocessor symbol @code{__V850_CALLT__} is defined.
24902 Pass on (or do not pass on) the @option{-mrelax} command-line option
24906 @itemx -mno-long-jumps
24907 @opindex mlong-jumps
24908 @opindex mno-long-jumps
24909 Disable (or re-enable) the generation of PC-relative jump instructions.
24912 @itemx -mhard-float
24913 @opindex msoft-float
24914 @opindex mhard-float
24915 Disable (or re-enable) the generation of hardware floating point
24916 instructions. This option is only significant when the target
24917 architecture is @samp{V850E2V3} or higher. If hardware floating point
24918 instructions are being generated then the C preprocessor symbol
24919 @code{__FPU_OK__} is defined, otherwise the symbol
24920 @code{__NO_FPU__} is defined.
24924 Enables the use of the e3v5 LOOP instruction. The use of this
24925 instruction is not enabled by default when the e3v5 architecture is
24926 selected because its use is still experimental.
24930 @opindex mrh850-abi
24932 Enables support for the RH850 version of the V850 ABI. This is the
24933 default. With this version of the ABI the following rules apply:
24937 Integer sized structures and unions are returned via a memory pointer
24938 rather than a register.
24941 Large structures and unions (more than 8 bytes in size) are passed by
24945 Functions are aligned to 16-bit boundaries.
24948 The @option{-m8byte-align} command-line option is supported.
24951 The @option{-mdisable-callt} command-line option is enabled by
24952 default. The @option{-mno-disable-callt} command-line option is not
24956 When this version of the ABI is enabled the C preprocessor symbol
24957 @code{__V850_RH850_ABI__} is defined.
24961 Enables support for the old GCC version of the V850 ABI. With this
24962 version of the ABI the following rules apply:
24966 Integer sized structures and unions are returned in register @code{r10}.
24969 Large structures and unions (more than 8 bytes in size) are passed by
24973 Functions are aligned to 32-bit boundaries, unless optimizing for
24977 The @option{-m8byte-align} command-line option is not supported.
24980 The @option{-mdisable-callt} command-line option is supported but not
24981 enabled by default.
24984 When this version of the ABI is enabled the C preprocessor symbol
24985 @code{__V850_GCC_ABI__} is defined.
24987 @item -m8byte-align
24988 @itemx -mno-8byte-align
24989 @opindex m8byte-align
24990 @opindex mno-8byte-align
24991 Enables support for @code{double} and @code{long long} types to be
24992 aligned on 8-byte boundaries. The default is to restrict the
24993 alignment of all objects to at most 4-bytes. When
24994 @option{-m8byte-align} is in effect the C preprocessor symbol
24995 @code{__V850_8BYTE_ALIGN__} is defined.
24998 @opindex mbig-switch
24999 Generate code suitable for big switch tables. Use this option only if
25000 the assembler/linker complain about out of range branches within a switch
25005 This option causes r2 and r5 to be used in the code generated by
25006 the compiler. This setting is the default.
25008 @item -mno-app-regs
25009 @opindex mno-app-regs
25010 This option causes r2 and r5 to be treated as fixed registers.
25015 @subsection VAX Options
25016 @cindex VAX options
25018 These @samp{-m} options are defined for the VAX:
25023 Do not output certain jump instructions (@code{aobleq} and so on)
25024 that the Unix assembler for the VAX cannot handle across long
25029 Do output those jump instructions, on the assumption that the
25030 GNU assembler is being used.
25034 Output code for G-format floating-point numbers instead of D-format.
25037 @node Visium Options
25038 @subsection Visium Options
25039 @cindex Visium options
25045 A program which performs file I/O and is destined to run on an MCM target
25046 should be linked with this option. It causes the libraries libc.a and
25047 libdebug.a to be linked. The program should be run on the target under
25048 the control of the GDB remote debugging stub.
25052 A program which performs file I/O and is destined to run on the simulator
25053 should be linked with option. This causes libraries libc.a and libsim.a to
25057 @itemx -mhard-float
25059 @opindex mhard-float
25060 Generate code containing floating-point instructions. This is the
25064 @itemx -msoft-float
25066 @opindex msoft-float
25067 Generate code containing library calls for floating-point.
25069 @option{-msoft-float} changes the calling convention in the output file;
25070 therefore, it is only useful if you compile @emph{all} of a program with
25071 this option. In particular, you need to compile @file{libgcc.a}, the
25072 library that comes with GCC, with @option{-msoft-float} in order for
25075 @item -mcpu=@var{cpu_type}
25077 Set the instruction set, register set, and instruction scheduling parameters
25078 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
25079 @samp{mcm}, @samp{gr5} and @samp{gr6}.
25081 @samp{mcm} is a synonym of @samp{gr5} present for backward compatibility.
25083 By default (unless configured otherwise), GCC generates code for the GR5
25084 variant of the Visium architecture.
25086 With @option{-mcpu=gr6}, GCC generates code for the GR6 variant of the Visium
25087 architecture. The only difference from GR5 code is that the compiler will
25088 generate block move instructions.
25090 @item -mtune=@var{cpu_type}
25092 Set the instruction scheduling parameters for machine type @var{cpu_type},
25093 but do not set the instruction set or register set that the option
25094 @option{-mcpu=@var{cpu_type}} would.
25098 Generate code for the supervisor mode, where there are no restrictions on
25099 the access to general registers. This is the default.
25102 @opindex muser-mode
25103 Generate code for the user mode, where the access to some general registers
25104 is forbidden: on the GR5, registers r24 to r31 cannot be accessed in this
25105 mode; on the GR6, only registers r29 to r31 are affected.
25109 @subsection VMS Options
25111 These @samp{-m} options are defined for the VMS implementations:
25114 @item -mvms-return-codes
25115 @opindex mvms-return-codes
25116 Return VMS condition codes from @code{main}. The default is to return POSIX-style
25117 condition (e.g.@ error) codes.
25119 @item -mdebug-main=@var{prefix}
25120 @opindex mdebug-main=@var{prefix}
25121 Flag the first routine whose name starts with @var{prefix} as the main
25122 routine for the debugger.
25126 Default to 64-bit memory allocation routines.
25128 @item -mpointer-size=@var{size}
25129 @opindex mpointer-size=@var{size}
25130 Set the default size of pointers. Possible options for @var{size} are
25131 @samp{32} or @samp{short} for 32 bit pointers, @samp{64} or @samp{long}
25132 for 64 bit pointers, and @samp{no} for supporting only 32 bit pointers.
25133 The later option disables @code{pragma pointer_size}.
25136 @node VxWorks Options
25137 @subsection VxWorks Options
25138 @cindex VxWorks Options
25140 The options in this section are defined for all VxWorks targets.
25141 Options specific to the target hardware are listed with the other
25142 options for that target.
25147 GCC can generate code for both VxWorks kernels and real time processes
25148 (RTPs). This option switches from the former to the latter. It also
25149 defines the preprocessor macro @code{__RTP__}.
25152 @opindex non-static
25153 Link an RTP executable against shared libraries rather than static
25154 libraries. The options @option{-static} and @option{-shared} can
25155 also be used for RTPs (@pxref{Link Options}); @option{-static}
25162 These options are passed down to the linker. They are defined for
25163 compatibility with Diab.
25166 @opindex Xbind-lazy
25167 Enable lazy binding of function calls. This option is equivalent to
25168 @option{-Wl,-z,now} and is defined for compatibility with Diab.
25172 Disable lazy binding of function calls. This option is the default and
25173 is defined for compatibility with Diab.
25177 @subsection x86 Options
25178 @cindex x86 Options
25180 These @samp{-m} options are defined for the x86 family of computers.
25184 @item -march=@var{cpu-type}
25186 Generate instructions for the machine type @var{cpu-type}. In contrast to
25187 @option{-mtune=@var{cpu-type}}, which merely tunes the generated code
25188 for the specified @var{cpu-type}, @option{-march=@var{cpu-type}} allows GCC
25189 to generate code that may not run at all on processors other than the one
25190 indicated. Specifying @option{-march=@var{cpu-type}} implies
25191 @option{-mtune=@var{cpu-type}}.
25193 The choices for @var{cpu-type} are:
25197 This selects the CPU to generate code for at compilation time by determining
25198 the processor type of the compiling machine. Using @option{-march=native}
25199 enables all instruction subsets supported by the local machine (hence
25200 the result might not run on different machines). Using @option{-mtune=native}
25201 produces code optimized for the local machine under the constraints
25202 of the selected instruction set.
25205 Original Intel i386 CPU@.
25208 Intel i486 CPU@. (No scheduling is implemented for this chip.)
25212 Intel Pentium CPU with no MMX support.
25215 Intel Lakemont MCU, based on Intel Pentium CPU.
25218 Intel Pentium MMX CPU, based on Pentium core with MMX instruction set support.
25221 Intel Pentium Pro CPU@.
25224 When used with @option{-march}, the Pentium Pro
25225 instruction set is used, so the code runs on all i686 family chips.
25226 When used with @option{-mtune}, it has the same meaning as @samp{generic}.
25229 Intel Pentium II CPU, based on Pentium Pro core with MMX instruction set
25234 Intel Pentium III CPU, based on Pentium Pro core with MMX and SSE instruction
25238 Intel Pentium M; low-power version of Intel Pentium III CPU
25239 with MMX, SSE and SSE2 instruction set support. Used by Centrino notebooks.
25243 Intel Pentium 4 CPU with MMX, SSE and SSE2 instruction set support.
25246 Improved version of Intel Pentium 4 CPU with MMX, SSE, SSE2 and SSE3 instruction
25250 Improved version of Intel Pentium 4 CPU with 64-bit extensions, MMX, SSE,
25251 SSE2 and SSE3 instruction set support.
25254 Intel Core 2 CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3 and SSSE3
25255 instruction set support.
25258 Intel Nehalem CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3, SSSE3,
25259 SSE4.1, SSE4.2 and POPCNT instruction set support.
25262 Intel Westmere CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3, SSSE3,
25263 SSE4.1, SSE4.2, POPCNT, AES and PCLMUL instruction set support.
25266 Intel Sandy Bridge CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3, SSSE3,
25267 SSE4.1, SSE4.2, POPCNT, AVX, AES and PCLMUL instruction set support.
25270 Intel Ivy Bridge CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3, SSSE3,
25271 SSE4.1, SSE4.2, POPCNT, AVX, AES, PCLMUL, FSGSBASE, RDRND and F16C
25272 instruction set support.
25275 Intel Haswell CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3, SSSE3,
25276 SSE4.1, SSE4.2, POPCNT, AVX, AVX2, AES, PCLMUL, FSGSBASE, RDRND, FMA,
25277 BMI, BMI2 and F16C instruction set support.
25280 Intel Broadwell CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3, SSSE3,
25281 SSE4.1, SSE4.2, POPCNT, AVX, AVX2, AES, PCLMUL, FSGSBASE, RDRND, FMA,
25282 BMI, BMI2, F16C, RDSEED, ADCX and PREFETCHW instruction set support.
25285 Intel Skylake CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3, SSSE3,
25286 SSE4.1, SSE4.2, POPCNT, AVX, AVX2, AES, PCLMUL, FSGSBASE, RDRND, FMA,
25287 BMI, BMI2, F16C, RDSEED, ADCX, PREFETCHW, CLFLUSHOPT, XSAVEC and
25288 XSAVES instruction set support.
25291 Intel Bonnell CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3 and SSSE3
25292 instruction set support.
25295 Intel Silvermont CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3, SSSE3,
25296 SSE4.1, SSE4.2, POPCNT, AES, PCLMUL and RDRND instruction set support.
25299 Intel Knight's Landing CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3,
25300 SSSE3, SSE4.1, SSE4.2, POPCNT, AVX, AVX2, AES, PCLMUL, FSGSBASE, RDRND, FMA,
25301 BMI, BMI2, F16C, RDSEED, ADCX, PREFETCHW, AVX512F, AVX512PF, AVX512ER and
25302 AVX512CD instruction set support.
25305 Intel Knights Mill CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3,
25306 SSSE3, SSE4.1, SSE4.2, POPCNT, AVX, AVX2, AES, PCLMUL, FSGSBASE, RDRND, FMA,
25307 BMI, BMI2, F16C, RDSEED, ADCX, PREFETCHW, AVX512F, AVX512PF, AVX512ER, AVX512CD,
25308 AVX5124VNNIW, AVX5124FMAPS and AVX512VPOPCNTDQ instruction set support.
25310 @item skylake-avx512
25311 Intel Skylake Server CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3,
25312 SSSE3, SSE4.1, SSE4.2, POPCNT, PKU, AVX, AVX2, AES, PCLMUL, FSGSBASE, RDRND, FMA,
25313 BMI, BMI2, F16C, RDSEED, ADCX, PREFETCHW, CLFLUSHOPT, XSAVEC, XSAVES, AVX512F,
25314 AVX512VL, AVX512BW, AVX512DQ and AVX512CD instruction set support.
25317 AMD K6 CPU with MMX instruction set support.
25321 Improved versions of AMD K6 CPU with MMX and 3DNow!@: instruction set support.
25324 @itemx athlon-tbird
25325 AMD Athlon CPU with MMX, 3dNOW!, enhanced 3DNow!@: and SSE prefetch instructions
25331 Improved AMD Athlon CPU with MMX, 3DNow!, enhanced 3DNow!@: and full SSE
25332 instruction set support.
25338 Processors based on the AMD K8 core with x86-64 instruction set support,
25339 including the AMD Opteron, Athlon 64, and Athlon 64 FX processors.
25340 (This supersets MMX, SSE, SSE2, 3DNow!, enhanced 3DNow!@: and 64-bit
25341 instruction set extensions.)
25344 @itemx opteron-sse3
25345 @itemx athlon64-sse3
25346 Improved versions of AMD K8 cores with SSE3 instruction set support.
25350 CPUs based on AMD Family 10h cores with x86-64 instruction set support. (This
25351 supersets MMX, SSE, SSE2, SSE3, SSE4A, 3DNow!, enhanced 3DNow!, ABM and 64-bit
25352 instruction set extensions.)
25355 CPUs based on AMD Family 15h cores with x86-64 instruction set support. (This
25356 supersets FMA4, AVX, XOP, LWP, AES, PCL_MUL, CX16, MMX, SSE, SSE2, SSE3, SSE4A,
25357 SSSE3, SSE4.1, SSE4.2, ABM and 64-bit instruction set extensions.)
25359 AMD Family 15h core based CPUs with x86-64 instruction set support. (This
25360 supersets BMI, TBM, F16C, FMA, FMA4, AVX, XOP, LWP, AES, PCL_MUL, CX16, MMX,
25361 SSE, SSE2, SSE3, SSE4A, SSSE3, SSE4.1, SSE4.2, ABM and 64-bit instruction set
25364 AMD Family 15h core based CPUs with x86-64 instruction set support. (This
25365 supersets BMI, TBM, F16C, FMA, FMA4, FSGSBASE, AVX, XOP, LWP, AES,
25366 PCL_MUL, CX16, MMX, SSE, SSE2, SSE3, SSE4A, SSSE3, SSE4.1, SSE4.2, ABM and
25367 64-bit instruction set extensions.
25369 AMD Family 15h core based CPUs with x86-64 instruction set support. (This
25370 supersets BMI, BMI2, TBM, F16C, FMA, FMA4, FSGSBASE, AVX, AVX2, XOP, LWP,
25371 AES, PCL_MUL, CX16, MOVBE, MMX, SSE, SSE2, SSE3, SSE4A, SSSE3, SSE4.1,
25372 SSE4.2, ABM and 64-bit instruction set extensions.
25375 AMD Family 17h core based CPUs with x86-64 instruction set support. (This
25376 supersets BMI, BMI2, F16C, FMA, FSGSBASE, AVX, AVX2, ADCX, RDSEED, MWAITX,
25377 SHA, CLZERO, AES, PCL_MUL, CX16, MOVBE, MMX, SSE, SSE2, SSE3, SSE4A, SSSE3,
25378 SSE4.1, SSE4.2, ABM, XSAVEC, XSAVES, CLFLUSHOPT, POPCNT, and 64-bit
25379 instruction set extensions.
25382 CPUs based on AMD Family 14h cores with x86-64 instruction set support. (This
25383 supersets MMX, SSE, SSE2, SSE3, SSSE3, SSE4A, CX16, ABM and 64-bit
25384 instruction set extensions.)
25387 CPUs based on AMD Family 16h cores with x86-64 instruction set support. This
25388 includes MOVBE, F16C, BMI, AVX, PCL_MUL, AES, SSE4.2, SSE4.1, CX16, ABM,
25389 SSE4A, SSSE3, SSE3, SSE2, SSE, MMX and 64-bit instruction set extensions.
25392 IDT WinChip C6 CPU, dealt in same way as i486 with additional MMX instruction
25396 IDT WinChip 2 CPU, dealt in same way as i486 with additional MMX and 3DNow!@:
25397 instruction set support.
25400 VIA C3 CPU with MMX and 3DNow!@: instruction set support.
25401 (No scheduling is implemented for this chip.)
25404 VIA C3-2 (Nehemiah/C5XL) CPU with MMX and SSE instruction set support.
25405 (No scheduling is implemented for this chip.)
25408 VIA C7 (Esther) CPU with MMX, SSE, SSE2 and SSE3 instruction set support.
25409 (No scheduling is implemented for this chip.)
25412 VIA Eden Samuel 2 CPU with MMX and 3DNow!@: instruction set support.
25413 (No scheduling is implemented for this chip.)
25416 VIA Eden Nehemiah CPU with MMX and SSE instruction set support.
25417 (No scheduling is implemented for this chip.)
25420 VIA Eden Esther CPU with MMX, SSE, SSE2 and SSE3 instruction set support.
25421 (No scheduling is implemented for this chip.)
25424 VIA Eden X2 CPU with x86-64, MMX, SSE, SSE2 and SSE3 instruction set support.
25425 (No scheduling is implemented for this chip.)
25428 VIA Eden X4 CPU with x86-64, MMX, SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2,
25429 AVX and AVX2 instruction set support.
25430 (No scheduling is implemented for this chip.)
25433 Generic VIA Nano CPU with x86-64, MMX, SSE, SSE2, SSE3 and SSSE3
25434 instruction set support.
25435 (No scheduling is implemented for this chip.)
25438 VIA Nano 1xxx CPU with x86-64, MMX, SSE, SSE2, SSE3 and SSSE3
25439 instruction set support.
25440 (No scheduling is implemented for this chip.)
25443 VIA Nano 2xxx CPU with x86-64, MMX, SSE, SSE2, SSE3 and SSSE3
25444 instruction set support.
25445 (No scheduling is implemented for this chip.)
25448 VIA Nano 3xxx CPU with x86-64, MMX, SSE, SSE2, SSE3, SSSE3 and SSE4.1
25449 instruction set support.
25450 (No scheduling is implemented for this chip.)
25453 VIA Nano Dual Core CPU with x86-64, MMX, SSE, SSE2, SSE3, SSSE3 and SSE4.1
25454 instruction set support.
25455 (No scheduling is implemented for this chip.)
25458 VIA Nano Quad Core CPU with x86-64, MMX, SSE, SSE2, SSE3, SSSE3 and SSE4.1
25459 instruction set support.
25460 (No scheduling is implemented for this chip.)
25463 AMD Geode embedded processor with MMX and 3DNow!@: instruction set support.
25466 @item -mtune=@var{cpu-type}
25468 Tune to @var{cpu-type} everything applicable about the generated code, except
25469 for the ABI and the set of available instructions.
25470 While picking a specific @var{cpu-type} schedules things appropriately
25471 for that particular chip, the compiler does not generate any code that
25472 cannot run on the default machine type unless you use a
25473 @option{-march=@var{cpu-type}} option.
25474 For example, if GCC is configured for i686-pc-linux-gnu
25475 then @option{-mtune=pentium4} generates code that is tuned for Pentium 4
25476 but still runs on i686 machines.
25478 The choices for @var{cpu-type} are the same as for @option{-march}.
25479 In addition, @option{-mtune} supports 2 extra choices for @var{cpu-type}:
25483 Produce code optimized for the most common IA32/@/AMD64/@/EM64T processors.
25484 If you know the CPU on which your code will run, then you should use
25485 the corresponding @option{-mtune} or @option{-march} option instead of
25486 @option{-mtune=generic}. But, if you do not know exactly what CPU users
25487 of your application will have, then you should use this option.
25489 As new processors are deployed in the marketplace, the behavior of this
25490 option will change. Therefore, if you upgrade to a newer version of
25491 GCC, code generation controlled by this option will change to reflect
25493 that are most common at the time that version of GCC is released.
25495 There is no @option{-march=generic} option because @option{-march}
25496 indicates the instruction set the compiler can use, and there is no
25497 generic instruction set applicable to all processors. In contrast,
25498 @option{-mtune} indicates the processor (or, in this case, collection of
25499 processors) for which the code is optimized.
25502 Produce code optimized for the most current Intel processors, which are
25503 Haswell and Silvermont for this version of GCC. If you know the CPU
25504 on which your code will run, then you should use the corresponding
25505 @option{-mtune} or @option{-march} option instead of @option{-mtune=intel}.
25506 But, if you want your application performs better on both Haswell and
25507 Silvermont, then you should use this option.
25509 As new Intel processors are deployed in the marketplace, the behavior of
25510 this option will change. Therefore, if you upgrade to a newer version of
25511 GCC, code generation controlled by this option will change to reflect
25512 the most current Intel processors at the time that version of GCC is
25515 There is no @option{-march=intel} option because @option{-march} indicates
25516 the instruction set the compiler can use, and there is no common
25517 instruction set applicable to all processors. In contrast,
25518 @option{-mtune} indicates the processor (or, in this case, collection of
25519 processors) for which the code is optimized.
25522 @item -mcpu=@var{cpu-type}
25524 A deprecated synonym for @option{-mtune}.
25526 @item -mfpmath=@var{unit}
25528 Generate floating-point arithmetic for selected unit @var{unit}. The choices
25529 for @var{unit} are:
25533 Use the standard 387 floating-point coprocessor present on the majority of chips and
25534 emulated otherwise. Code compiled with this option runs almost everywhere.
25535 The temporary results are computed in 80-bit precision instead of the precision
25536 specified by the type, resulting in slightly different results compared to most
25537 of other chips. See @option{-ffloat-store} for more detailed description.
25539 This is the default choice for non-Darwin x86-32 targets.
25542 Use scalar floating-point instructions present in the SSE instruction set.
25543 This instruction set is supported by Pentium III and newer chips,
25544 and in the AMD line
25545 by Athlon-4, Athlon XP and Athlon MP chips. The earlier version of the SSE
25546 instruction set supports only single-precision arithmetic, thus the double and
25547 extended-precision arithmetic are still done using 387. A later version, present
25548 only in Pentium 4 and AMD x86-64 chips, supports double-precision
25551 For the x86-32 compiler, you must use @option{-march=@var{cpu-type}}, @option{-msse}
25552 or @option{-msse2} switches to enable SSE extensions and make this option
25553 effective. For the x86-64 compiler, these extensions are enabled by default.
25555 The resulting code should be considerably faster in the majority of cases and avoid
25556 the numerical instability problems of 387 code, but may break some existing
25557 code that expects temporaries to be 80 bits.
25559 This is the default choice for the x86-64 compiler, Darwin x86-32 targets,
25560 and the default choice for x86-32 targets with the SSE2 instruction set
25561 when @option{-ffast-math} is enabled.
25566 Attempt to utilize both instruction sets at once. This effectively doubles the
25567 amount of available registers, and on chips with separate execution units for
25568 387 and SSE the execution resources too. Use this option with care, as it is
25569 still experimental, because the GCC register allocator does not model separate
25570 functional units well, resulting in unstable performance.
25573 @item -masm=@var{dialect}
25574 @opindex masm=@var{dialect}
25575 Output assembly instructions using selected @var{dialect}. Also affects
25576 which dialect is used for basic @code{asm} (@pxref{Basic Asm}) and
25577 extended @code{asm} (@pxref{Extended Asm}). Supported choices (in dialect
25578 order) are @samp{att} or @samp{intel}. The default is @samp{att}. Darwin does
25579 not support @samp{intel}.
25582 @itemx -mno-ieee-fp
25584 @opindex mno-ieee-fp
25585 Control whether or not the compiler uses IEEE floating-point
25586 comparisons. These correctly handle the case where the result of a
25587 comparison is unordered.
25592 @opindex mhard-float
25593 Generate output containing 80387 instructions for floating point.
25598 @opindex msoft-float
25599 Generate output containing library calls for floating point.
25601 @strong{Warning:} the requisite libraries are not part of GCC@.
25602 Normally the facilities of the machine's usual C compiler are used, but
25603 this cannot be done directly in cross-compilation. You must make your
25604 own arrangements to provide suitable library functions for
25607 On machines where a function returns floating-point results in the 80387
25608 register stack, some floating-point opcodes may be emitted even if
25609 @option{-msoft-float} is used.
25611 @item -mno-fp-ret-in-387
25612 @opindex mno-fp-ret-in-387
25613 Do not use the FPU registers for return values of functions.
25615 The usual calling convention has functions return values of types
25616 @code{float} and @code{double} in an FPU register, even if there
25617 is no FPU@. The idea is that the operating system should emulate
25620 The option @option{-mno-fp-ret-in-387} causes such values to be returned
25621 in ordinary CPU registers instead.
25623 @item -mno-fancy-math-387
25624 @opindex mno-fancy-math-387
25625 Some 387 emulators do not support the @code{sin}, @code{cos} and
25626 @code{sqrt} instructions for the 387. Specify this option to avoid
25627 generating those instructions. This option is the default on
25628 OpenBSD and NetBSD@. This option is overridden when @option{-march}
25629 indicates that the target CPU always has an FPU and so the
25630 instruction does not need emulation. These
25631 instructions are not generated unless you also use the
25632 @option{-funsafe-math-optimizations} switch.
25634 @item -malign-double
25635 @itemx -mno-align-double
25636 @opindex malign-double
25637 @opindex mno-align-double
25638 Control whether GCC aligns @code{double}, @code{long double}, and
25639 @code{long long} variables on a two-word boundary or a one-word
25640 boundary. Aligning @code{double} variables on a two-word boundary
25641 produces code that runs somewhat faster on a Pentium at the
25642 expense of more memory.
25644 On x86-64, @option{-malign-double} is enabled by default.
25646 @strong{Warning:} if you use the @option{-malign-double} switch,
25647 structures containing the above types are aligned differently than
25648 the published application binary interface specifications for the x86-32
25649 and are not binary compatible with structures in code compiled
25650 without that switch.
25652 @item -m96bit-long-double
25653 @itemx -m128bit-long-double
25654 @opindex m96bit-long-double
25655 @opindex m128bit-long-double
25656 These switches control the size of @code{long double} type. The x86-32
25657 application binary interface specifies the size to be 96 bits,
25658 so @option{-m96bit-long-double} is the default in 32-bit mode.
25660 Modern architectures (Pentium and newer) prefer @code{long double}
25661 to be aligned to an 8- or 16-byte boundary. In arrays or structures
25662 conforming to the ABI, this is not possible. So specifying
25663 @option{-m128bit-long-double} aligns @code{long double}
25664 to a 16-byte boundary by padding the @code{long double} with an additional
25667 In the x86-64 compiler, @option{-m128bit-long-double} is the default choice as
25668 its ABI specifies that @code{long double} is aligned on 16-byte boundary.
25670 Notice that neither of these options enable any extra precision over the x87
25671 standard of 80 bits for a @code{long double}.
25673 @strong{Warning:} if you override the default value for your target ABI, this
25674 changes the size of
25675 structures and arrays containing @code{long double} variables,
25676 as well as modifying the function calling convention for functions taking
25677 @code{long double}. Hence they are not binary-compatible
25678 with code compiled without that switch.
25680 @item -mlong-double-64
25681 @itemx -mlong-double-80
25682 @itemx -mlong-double-128
25683 @opindex mlong-double-64
25684 @opindex mlong-double-80
25685 @opindex mlong-double-128
25686 These switches control the size of @code{long double} type. A size
25687 of 64 bits makes the @code{long double} type equivalent to the @code{double}
25688 type. This is the default for 32-bit Bionic C library. A size
25689 of 128 bits makes the @code{long double} type equivalent to the
25690 @code{__float128} type. This is the default for 64-bit Bionic C library.
25692 @strong{Warning:} if you override the default value for your target ABI, this
25693 changes the size of
25694 structures and arrays containing @code{long double} variables,
25695 as well as modifying the function calling convention for functions taking
25696 @code{long double}. Hence they are not binary-compatible
25697 with code compiled without that switch.
25699 @item -malign-data=@var{type}
25700 @opindex malign-data
25701 Control how GCC aligns variables. Supported values for @var{type} are
25702 @samp{compat} uses increased alignment value compatible uses GCC 4.8
25703 and earlier, @samp{abi} uses alignment value as specified by the
25704 psABI, and @samp{cacheline} uses increased alignment value to match
25705 the cache line size. @samp{compat} is the default.
25707 @item -mlarge-data-threshold=@var{threshold}
25708 @opindex mlarge-data-threshold
25709 When @option{-mcmodel=medium} is specified, data objects larger than
25710 @var{threshold} are placed in the large data section. This value must be the
25711 same across all objects linked into the binary, and defaults to 65535.
25715 Use a different function-calling convention, in which functions that
25716 take a fixed number of arguments return with the @code{ret @var{num}}
25717 instruction, which pops their arguments while returning. This saves one
25718 instruction in the caller since there is no need to pop the arguments
25721 You can specify that an individual function is called with this calling
25722 sequence with the function attribute @code{stdcall}. You can also
25723 override the @option{-mrtd} option by using the function attribute
25724 @code{cdecl}. @xref{Function Attributes}.
25726 @strong{Warning:} this calling convention is incompatible with the one
25727 normally used on Unix, so you cannot use it if you need to call
25728 libraries compiled with the Unix compiler.
25730 Also, you must provide function prototypes for all functions that
25731 take variable numbers of arguments (including @code{printf});
25732 otherwise incorrect code is generated for calls to those
25735 In addition, seriously incorrect code results if you call a
25736 function with too many arguments. (Normally, extra arguments are
25737 harmlessly ignored.)
25739 @item -mregparm=@var{num}
25741 Control how many registers are used to pass integer arguments. By
25742 default, no registers are used to pass arguments, and at most 3
25743 registers can be used. You can control this behavior for a specific
25744 function by using the function attribute @code{regparm}.
25745 @xref{Function Attributes}.
25747 @strong{Warning:} if you use this switch, and
25748 @var{num} is nonzero, then you must build all modules with the same
25749 value, including any libraries. This includes the system libraries and
25753 @opindex msseregparm
25754 Use SSE register passing conventions for float and double arguments
25755 and return values. You can control this behavior for a specific
25756 function by using the function attribute @code{sseregparm}.
25757 @xref{Function Attributes}.
25759 @strong{Warning:} if you use this switch then you must build all
25760 modules with the same value, including any libraries. This includes
25761 the system libraries and startup modules.
25763 @item -mvect8-ret-in-mem
25764 @opindex mvect8-ret-in-mem
25765 Return 8-byte vectors in memory instead of MMX registers. This is the
25766 default on Solaris@tie{}8 and 9 and VxWorks to match the ABI of the Sun
25767 Studio compilers until version 12. Later compiler versions (starting
25768 with Studio 12 Update@tie{}1) follow the ABI used by other x86 targets, which
25769 is the default on Solaris@tie{}10 and later. @emph{Only} use this option if
25770 you need to remain compatible with existing code produced by those
25771 previous compiler versions or older versions of GCC@.
25780 Set 80387 floating-point precision to 32, 64 or 80 bits. When @option{-mpc32}
25781 is specified, the significands of results of floating-point operations are
25782 rounded to 24 bits (single precision); @option{-mpc64} rounds the
25783 significands of results of floating-point operations to 53 bits (double
25784 precision) and @option{-mpc80} rounds the significands of results of
25785 floating-point operations to 64 bits (extended double precision), which is
25786 the default. When this option is used, floating-point operations in higher
25787 precisions are not available to the programmer without setting the FPU
25788 control word explicitly.
25790 Setting the rounding of floating-point operations to less than the default
25791 80 bits can speed some programs by 2% or more. Note that some mathematical
25792 libraries assume that extended-precision (80-bit) floating-point operations
25793 are enabled by default; routines in such libraries could suffer significant
25794 loss of accuracy, typically through so-called ``catastrophic cancellation'',
25795 when this option is used to set the precision to less than extended precision.
25797 @item -mstackrealign
25798 @opindex mstackrealign
25799 Realign the stack at entry. On the x86, the @option{-mstackrealign}
25800 option generates an alternate prologue and epilogue that realigns the
25801 run-time stack if necessary. This supports mixing legacy codes that keep
25802 4-byte stack alignment with modern codes that keep 16-byte stack alignment for
25803 SSE compatibility. See also the attribute @code{force_align_arg_pointer},
25804 applicable to individual functions.
25806 @item -mpreferred-stack-boundary=@var{num}
25807 @opindex mpreferred-stack-boundary
25808 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
25809 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
25810 the default is 4 (16 bytes or 128 bits).
25812 @strong{Warning:} When generating code for the x86-64 architecture with
25813 SSE extensions disabled, @option{-mpreferred-stack-boundary=3} can be
25814 used to keep the stack boundary aligned to 8 byte boundary. Since
25815 x86-64 ABI require 16 byte stack alignment, this is ABI incompatible and
25816 intended to be used in controlled environment where stack space is
25817 important limitation. This option leads to wrong code when functions
25818 compiled with 16 byte stack alignment (such as functions from a standard
25819 library) are called with misaligned stack. In this case, SSE
25820 instructions may lead to misaligned memory access traps. In addition,
25821 variable arguments are handled incorrectly for 16 byte aligned
25822 objects (including x87 long double and __int128), leading to wrong
25823 results. You must build all modules with
25824 @option{-mpreferred-stack-boundary=3}, including any libraries. This
25825 includes the system libraries and startup modules.
25827 @item -mincoming-stack-boundary=@var{num}
25828 @opindex mincoming-stack-boundary
25829 Assume the incoming stack is aligned to a 2 raised to @var{num} byte
25830 boundary. If @option{-mincoming-stack-boundary} is not specified,
25831 the one specified by @option{-mpreferred-stack-boundary} is used.
25833 On Pentium and Pentium Pro, @code{double} and @code{long double} values
25834 should be aligned to an 8-byte boundary (see @option{-malign-double}) or
25835 suffer significant run time performance penalties. On Pentium III, the
25836 Streaming SIMD Extension (SSE) data type @code{__m128} may not work
25837 properly if it is not 16-byte aligned.
25839 To ensure proper alignment of this values on the stack, the stack boundary
25840 must be as aligned as that required by any value stored on the stack.
25841 Further, every function must be generated such that it keeps the stack
25842 aligned. Thus calling a function compiled with a higher preferred
25843 stack boundary from a function compiled with a lower preferred stack
25844 boundary most likely misaligns the stack. It is recommended that
25845 libraries that use callbacks always use the default setting.
25847 This extra alignment does consume extra stack space, and generally
25848 increases code size. Code that is sensitive to stack space usage, such
25849 as embedded systems and operating system kernels, may want to reduce the
25850 preferred alignment to @option{-mpreferred-stack-boundary=2}.
25907 @itemx -mavx512ifma
25908 @opindex mavx512ifma
25910 @itemx -mavx512vbmi
25911 @opindex mavx512vbmi
25923 @opindex mclfushopt
25940 @itemx -mprefetchwt1
25941 @opindex mprefetchwt1
26005 @itemx -mavx512vbmi2
26006 @opindex mavx512vbmi2
26010 These switches enable the use of instructions in the MMX, SSE,
26011 SSE2, SSE3, SSSE3, SSE4.1, AVX, AVX2, AVX512F, AVX512PF, AVX512ER, AVX512CD,
26012 SHA, AES, PCLMUL, FSGSBASE, RDRND, F16C, FMA, SSE4A, FMA4, XOP, LWP, ABM,
26013 AVX512VL, AVX512BW, AVX512DQ, AVX512IFMA, AVX512VBMI, BMI, BMI2,
26014 FXSR, XSAVE, XSAVEOPT, LZCNT, RTM, MPX, MWAITX, PKU, IBT, SHSTK, AVX512VBMI2,
26015 GFNI, 3DNow!@: or enhanced 3DNow!@: extended instruction sets. Each has a
26016 corresponding @option{-mno-} option to disable use of these instructions.
26018 These extensions are also available as built-in functions: see
26019 @ref{x86 Built-in Functions}, for details of the functions enabled and
26020 disabled by these switches.
26022 To generate SSE/SSE2 instructions automatically from floating-point
26023 code (as opposed to 387 instructions), see @option{-mfpmath=sse}.
26025 GCC depresses SSEx instructions when @option{-mavx} is used. Instead, it
26026 generates new AVX instructions or AVX equivalence for all SSEx instructions
26029 These options enable GCC to use these extended instructions in
26030 generated code, even without @option{-mfpmath=sse}. Applications that
26031 perform run-time CPU detection must compile separate files for each
26032 supported architecture, using the appropriate flags. In particular,
26033 the file containing the CPU detection code should be compiled without
26036 The @option{-mcet} option turns on the @option{-mibt} and @option{-mshstk}
26037 options. The @option{-mibt} option enables indirect branch tracking support
26038 and the @option{-mshstk} option enables shadow stack support from
26039 Intel Control-flow Enforcement Technology (CET). The compiler also provides
26040 a number of built-in functions for fine-grained control in a CET-based
26041 application. See @xref{x86 Built-in Functions}, for more information.
26043 @item -mdump-tune-features
26044 @opindex mdump-tune-features
26045 This option instructs GCC to dump the names of the x86 performance
26046 tuning features and default settings. The names can be used in
26047 @option{-mtune-ctrl=@var{feature-list}}.
26049 @item -mtune-ctrl=@var{feature-list}
26050 @opindex mtune-ctrl=@var{feature-list}
26051 This option is used to do fine grain control of x86 code generation features.
26052 @var{feature-list} is a comma separated list of @var{feature} names. See also
26053 @option{-mdump-tune-features}. When specified, the @var{feature} is turned
26054 on if it is not preceded with @samp{^}, otherwise, it is turned off.
26055 @option{-mtune-ctrl=@var{feature-list}} is intended to be used by GCC
26056 developers. Using it may lead to code paths not covered by testing and can
26057 potentially result in compiler ICEs or runtime errors.
26060 @opindex mno-default
26061 This option instructs GCC to turn off all tunable features. See also
26062 @option{-mtune-ctrl=@var{feature-list}} and @option{-mdump-tune-features}.
26066 This option instructs GCC to emit a @code{cld} instruction in the prologue
26067 of functions that use string instructions. String instructions depend on
26068 the DF flag to select between autoincrement or autodecrement mode. While the
26069 ABI specifies the DF flag to be cleared on function entry, some operating
26070 systems violate this specification by not clearing the DF flag in their
26071 exception dispatchers. The exception handler can be invoked with the DF flag
26072 set, which leads to wrong direction mode when string instructions are used.
26073 This option can be enabled by default on 32-bit x86 targets by configuring
26074 GCC with the @option{--enable-cld} configure option. Generation of @code{cld}
26075 instructions can be suppressed with the @option{-mno-cld} compiler option
26079 @opindex mvzeroupper
26080 This option instructs GCC to emit a @code{vzeroupper} instruction
26081 before a transfer of control flow out of the function to minimize
26082 the AVX to SSE transition penalty as well as remove unnecessary @code{zeroupper}
26085 @item -mprefer-avx128
26086 @opindex mprefer-avx128
26087 This option instructs GCC to use 128-bit AVX instructions instead of
26088 256-bit AVX instructions in the auto-vectorizer.
26090 @item -mprefer-avx256
26091 @opindex mprefer-avx256
26092 This option instructs GCC to use 256-bit AVX instructions instead of
26093 512-bit AVX instructions in the auto-vectorizer.
26097 This option enables GCC to generate @code{CMPXCHG16B} instructions in 64-bit
26098 code to implement compare-and-exchange operations on 16-byte aligned 128-bit
26099 objects. This is useful for atomic updates of data structures exceeding one
26100 machine word in size. The compiler uses this instruction to implement
26101 @ref{__sync Builtins}. However, for @ref{__atomic Builtins} operating on
26102 128-bit integers, a library call is always used.
26106 This option enables generation of @code{SAHF} instructions in 64-bit code.
26107 Early Intel Pentium 4 CPUs with Intel 64 support,
26108 prior to the introduction of Pentium 4 G1 step in December 2005,
26109 lacked the @code{LAHF} and @code{SAHF} instructions
26110 which are supported by AMD64.
26111 These are load and store instructions, respectively, for certain status flags.
26112 In 64-bit mode, the @code{SAHF} instruction is used to optimize @code{fmod},
26113 @code{drem}, and @code{remainder} built-in functions;
26114 see @ref{Other Builtins} for details.
26118 This option enables use of the @code{movbe} instruction to implement
26119 @code{__builtin_bswap32} and @code{__builtin_bswap64}.
26123 This option tells the compiler to use indirect branch tracking support
26124 (for indirect calls and jumps) from x86 Control-flow Enforcement
26125 Technology (CET). The option has effect only if the
26126 @option{-fcf-protection=full} or @option{-fcf-protection=branch} option
26127 is specified. The option @option{-mibt} is on by default when the
26128 @code{-mcet} option is specified.
26132 This option tells the compiler to use shadow stack support (return
26133 address tracking) from x86 Control-flow Enforcement Technology (CET).
26134 The option has effect only if the @option{-fcf-protection=full} or
26135 @option{-fcf-protection=return} option is specified. The option
26136 @option{-mshstk} is on by default when the @option{-mcet} option is
26141 This option enables built-in functions @code{__builtin_ia32_crc32qi},
26142 @code{__builtin_ia32_crc32hi}, @code{__builtin_ia32_crc32si} and
26143 @code{__builtin_ia32_crc32di} to generate the @code{crc32} machine instruction.
26147 This option enables use of @code{RCPSS} and @code{RSQRTSS} instructions
26148 (and their vectorized variants @code{RCPPS} and @code{RSQRTPS})
26149 with an additional Newton-Raphson step
26150 to increase precision instead of @code{DIVSS} and @code{SQRTSS}
26151 (and their vectorized
26152 variants) for single-precision floating-point arguments. These instructions
26153 are generated only when @option{-funsafe-math-optimizations} is enabled
26154 together with @option{-ffinite-math-only} and @option{-fno-trapping-math}.
26155 Note that while the throughput of the sequence is higher than the throughput
26156 of the non-reciprocal instruction, the precision of the sequence can be
26157 decreased by up to 2 ulp (i.e. the inverse of 1.0 equals 0.99999994).
26159 Note that GCC implements @code{1.0f/sqrtf(@var{x})} in terms of @code{RSQRTSS}
26160 (or @code{RSQRTPS}) already with @option{-ffast-math} (or the above option
26161 combination), and doesn't need @option{-mrecip}.
26163 Also note that GCC emits the above sequence with additional Newton-Raphson step
26164 for vectorized single-float division and vectorized @code{sqrtf(@var{x})}
26165 already with @option{-ffast-math} (or the above option combination), and
26166 doesn't need @option{-mrecip}.
26168 @item -mrecip=@var{opt}
26169 @opindex mrecip=opt
26170 This option controls which reciprocal estimate instructions
26171 may be used. @var{opt} is a comma-separated list of options, which may
26172 be preceded by a @samp{!} to invert the option:
26176 Enable all estimate instructions.
26179 Enable the default instructions, equivalent to @option{-mrecip}.
26182 Disable all estimate instructions, equivalent to @option{-mno-recip}.
26185 Enable the approximation for scalar division.
26188 Enable the approximation for vectorized division.
26191 Enable the approximation for scalar square root.
26194 Enable the approximation for vectorized square root.
26197 So, for example, @option{-mrecip=all,!sqrt} enables
26198 all of the reciprocal approximations, except for square root.
26200 @item -mveclibabi=@var{type}
26201 @opindex mveclibabi
26202 Specifies the ABI type to use for vectorizing intrinsics using an
26203 external library. Supported values for @var{type} are @samp{svml}
26204 for the Intel short
26205 vector math library and @samp{acml} for the AMD math core library.
26206 To use this option, both @option{-ftree-vectorize} and
26207 @option{-funsafe-math-optimizations} have to be enabled, and an SVML or ACML
26208 ABI-compatible library must be specified at link time.
26210 GCC currently emits calls to @code{vmldExp2},
26211 @code{vmldLn2}, @code{vmldLog102}, @code{vmldLog102}, @code{vmldPow2},
26212 @code{vmldTanh2}, @code{vmldTan2}, @code{vmldAtan2}, @code{vmldAtanh2},
26213 @code{vmldCbrt2}, @code{vmldSinh2}, @code{vmldSin2}, @code{vmldAsinh2},
26214 @code{vmldAsin2}, @code{vmldCosh2}, @code{vmldCos2}, @code{vmldAcosh2},
26215 @code{vmldAcos2}, @code{vmlsExp4}, @code{vmlsLn4}, @code{vmlsLog104},
26216 @code{vmlsLog104}, @code{vmlsPow4}, @code{vmlsTanh4}, @code{vmlsTan4},
26217 @code{vmlsAtan4}, @code{vmlsAtanh4}, @code{vmlsCbrt4}, @code{vmlsSinh4},
26218 @code{vmlsSin4}, @code{vmlsAsinh4}, @code{vmlsAsin4}, @code{vmlsCosh4},
26219 @code{vmlsCos4}, @code{vmlsAcosh4} and @code{vmlsAcos4} for corresponding
26220 function type when @option{-mveclibabi=svml} is used, and @code{__vrd2_sin},
26221 @code{__vrd2_cos}, @code{__vrd2_exp}, @code{__vrd2_log}, @code{__vrd2_log2},
26222 @code{__vrd2_log10}, @code{__vrs4_sinf}, @code{__vrs4_cosf},
26223 @code{__vrs4_expf}, @code{__vrs4_logf}, @code{__vrs4_log2f},
26224 @code{__vrs4_log10f} and @code{__vrs4_powf} for the corresponding function type
26225 when @option{-mveclibabi=acml} is used.
26227 @item -mabi=@var{name}
26229 Generate code for the specified calling convention. Permissible values
26230 are @samp{sysv} for the ABI used on GNU/Linux and other systems, and
26231 @samp{ms} for the Microsoft ABI. The default is to use the Microsoft
26232 ABI when targeting Microsoft Windows and the SysV ABI on all other systems.
26233 You can control this behavior for specific functions by
26234 using the function attributes @code{ms_abi} and @code{sysv_abi}.
26235 @xref{Function Attributes}.
26237 @item -mforce-indirect-call
26238 @opindex mforce-indirect-call
26239 Force all calls to functions to be indirect. This is useful
26240 when using Intel Processor Trace where it generates more precise timing
26241 information for function calls.
26243 @item -mcall-ms2sysv-xlogues
26244 @opindex mcall-ms2sysv-xlogues
26245 @opindex mno-call-ms2sysv-xlogues
26246 Due to differences in 64-bit ABIs, any Microsoft ABI function that calls a
26247 System V ABI function must consider RSI, RDI and XMM6-15 as clobbered. By
26248 default, the code for saving and restoring these registers is emitted inline,
26249 resulting in fairly lengthy prologues and epilogues. Using
26250 @option{-mcall-ms2sysv-xlogues} emits prologues and epilogues that
26251 use stubs in the static portion of libgcc to perform these saves and restores,
26252 thus reducing function size at the cost of a few extra instructions.
26254 @item -mtls-dialect=@var{type}
26255 @opindex mtls-dialect
26256 Generate code to access thread-local storage using the @samp{gnu} or
26257 @samp{gnu2} conventions. @samp{gnu} is the conservative default;
26258 @samp{gnu2} is more efficient, but it may add compile- and run-time
26259 requirements that cannot be satisfied on all systems.
26262 @itemx -mno-push-args
26263 @opindex mpush-args
26264 @opindex mno-push-args
26265 Use PUSH operations to store outgoing parameters. This method is shorter
26266 and usually equally fast as method using SUB/MOV operations and is enabled
26267 by default. In some cases disabling it may improve performance because of
26268 improved scheduling and reduced dependencies.
26270 @item -maccumulate-outgoing-args
26271 @opindex maccumulate-outgoing-args
26272 If enabled, the maximum amount of space required for outgoing arguments is
26273 computed in the function prologue. This is faster on most modern CPUs
26274 because of reduced dependencies, improved scheduling and reduced stack usage
26275 when the preferred stack boundary is not equal to 2. The drawback is a notable
26276 increase in code size. This switch implies @option{-mno-push-args}.
26280 Support thread-safe exception handling on MinGW. Programs that rely
26281 on thread-safe exception handling must compile and link all code with the
26282 @option{-mthreads} option. When compiling, @option{-mthreads} defines
26283 @option{-D_MT}; when linking, it links in a special thread helper library
26284 @option{-lmingwthrd} which cleans up per-thread exception-handling data.
26286 @item -mms-bitfields
26287 @itemx -mno-ms-bitfields
26288 @opindex mms-bitfields
26289 @opindex mno-ms-bitfields
26291 Enable/disable bit-field layout compatible with the native Microsoft
26294 If @code{packed} is used on a structure, or if bit-fields are used,
26295 it may be that the Microsoft ABI lays out the structure differently
26296 than the way GCC normally does. Particularly when moving packed
26297 data between functions compiled with GCC and the native Microsoft compiler
26298 (either via function call or as data in a file), it may be necessary to access
26301 This option is enabled by default for Microsoft Windows
26302 targets. This behavior can also be controlled locally by use of variable
26303 or type attributes. For more information, see @ref{x86 Variable Attributes}
26304 and @ref{x86 Type Attributes}.
26306 The Microsoft structure layout algorithm is fairly simple with the exception
26307 of the bit-field packing.
26308 The padding and alignment of members of structures and whether a bit-field
26309 can straddle a storage-unit boundary are determine by these rules:
26312 @item Structure members are stored sequentially in the order in which they are
26313 declared: the first member has the lowest memory address and the last member
26316 @item Every data object has an alignment requirement. The alignment requirement
26317 for all data except structures, unions, and arrays is either the size of the
26318 object or the current packing size (specified with either the
26319 @code{aligned} attribute or the @code{pack} pragma),
26320 whichever is less. For structures, unions, and arrays,
26321 the alignment requirement is the largest alignment requirement of its members.
26322 Every object is allocated an offset so that:
26325 offset % alignment_requirement == 0
26328 @item Adjacent bit-fields are packed into the same 1-, 2-, or 4-byte allocation
26329 unit if the integral types are the same size and if the next bit-field fits
26330 into the current allocation unit without crossing the boundary imposed by the
26331 common alignment requirements of the bit-fields.
26334 MSVC interprets zero-length bit-fields in the following ways:
26337 @item If a zero-length bit-field is inserted between two bit-fields that
26338 are normally coalesced, the bit-fields are not coalesced.
26345 unsigned long bf_1 : 12;
26347 unsigned long bf_2 : 12;
26352 The size of @code{t1} is 8 bytes with the zero-length bit-field. If the
26353 zero-length bit-field were removed, @code{t1}'s size would be 4 bytes.
26355 @item If a zero-length bit-field is inserted after a bit-field, @code{foo}, and the
26356 alignment of the zero-length bit-field is greater than the member that follows it,
26357 @code{bar}, @code{bar} is aligned as the type of the zero-length bit-field.
26378 For @code{t2}, @code{bar} is placed at offset 2, rather than offset 1.
26379 Accordingly, the size of @code{t2} is 4. For @code{t3}, the zero-length
26380 bit-field does not affect the alignment of @code{bar} or, as a result, the size
26383 Taking this into account, it is important to note the following:
26386 @item If a zero-length bit-field follows a normal bit-field, the type of the
26387 zero-length bit-field may affect the alignment of the structure as whole. For
26388 example, @code{t2} has a size of 4 bytes, since the zero-length bit-field follows a
26389 normal bit-field, and is of type short.
26391 @item Even if a zero-length bit-field is not followed by a normal bit-field, it may
26392 still affect the alignment of the structure:
26403 Here, @code{t4} takes up 4 bytes.
26406 @item Zero-length bit-fields following non-bit-field members are ignored:
26418 Here, @code{t5} takes up 2 bytes.
26422 @item -mno-align-stringops
26423 @opindex mno-align-stringops
26424 Do not align the destination of inlined string operations. This switch reduces
26425 code size and improves performance in case the destination is already aligned,
26426 but GCC doesn't know about it.
26428 @item -minline-all-stringops
26429 @opindex minline-all-stringops
26430 By default GCC inlines string operations only when the destination is
26431 known to be aligned to least a 4-byte boundary.
26432 This enables more inlining and increases code
26433 size, but may improve performance of code that depends on fast
26434 @code{memcpy}, @code{strlen},
26435 and @code{memset} for short lengths.
26437 @item -minline-stringops-dynamically
26438 @opindex minline-stringops-dynamically
26439 For string operations of unknown size, use run-time checks with
26440 inline code for small blocks and a library call for large blocks.
26442 @item -mstringop-strategy=@var{alg}
26443 @opindex mstringop-strategy=@var{alg}
26444 Override the internal decision heuristic for the particular algorithm to use
26445 for inlining string operations. The allowed values for @var{alg} are:
26451 Expand using i386 @code{rep} prefix of the specified size.
26455 @itemx unrolled_loop
26456 Expand into an inline loop.
26459 Always use a library call.
26462 @item -mmemcpy-strategy=@var{strategy}
26463 @opindex mmemcpy-strategy=@var{strategy}
26464 Override the internal decision heuristic to decide if @code{__builtin_memcpy}
26465 should be inlined and what inline algorithm to use when the expected size
26466 of the copy operation is known. @var{strategy}
26467 is a comma-separated list of @var{alg}:@var{max_size}:@var{dest_align} triplets.
26468 @var{alg} is specified in @option{-mstringop-strategy}, @var{max_size} specifies
26469 the max byte size with which inline algorithm @var{alg} is allowed. For the last
26470 triplet, the @var{max_size} must be @code{-1}. The @var{max_size} of the triplets
26471 in the list must be specified in increasing order. The minimal byte size for
26472 @var{alg} is @code{0} for the first triplet and @code{@var{max_size} + 1} of the
26475 @item -mmemset-strategy=@var{strategy}
26476 @opindex mmemset-strategy=@var{strategy}
26477 The option is similar to @option{-mmemcpy-strategy=} except that it is to control
26478 @code{__builtin_memset} expansion.
26480 @item -momit-leaf-frame-pointer
26481 @opindex momit-leaf-frame-pointer
26482 Don't keep the frame pointer in a register for leaf functions. This
26483 avoids the instructions to save, set up, and restore frame pointers and
26484 makes an extra register available in leaf functions. The option
26485 @option{-fomit-leaf-frame-pointer} removes the frame pointer for leaf functions,
26486 which might make debugging harder.
26488 @item -mtls-direct-seg-refs
26489 @itemx -mno-tls-direct-seg-refs
26490 @opindex mtls-direct-seg-refs
26491 Controls whether TLS variables may be accessed with offsets from the
26492 TLS segment register (@code{%gs} for 32-bit, @code{%fs} for 64-bit),
26493 or whether the thread base pointer must be added. Whether or not this
26494 is valid depends on the operating system, and whether it maps the
26495 segment to cover the entire TLS area.
26497 For systems that use the GNU C Library, the default is on.
26500 @itemx -mno-sse2avx
26502 Specify that the assembler should encode SSE instructions with VEX
26503 prefix. The option @option{-mavx} turns this on by default.
26508 If profiling is active (@option{-pg}), put the profiling
26509 counter call before the prologue.
26510 Note: On x86 architectures the attribute @code{ms_hook_prologue}
26511 isn't possible at the moment for @option{-mfentry} and @option{-pg}.
26513 @item -mrecord-mcount
26514 @itemx -mno-record-mcount
26515 @opindex mrecord-mcount
26516 If profiling is active (@option{-pg}), generate a __mcount_loc section
26517 that contains pointers to each profiling call. This is useful for
26518 automatically patching and out calls.
26521 @itemx -mno-nop-mcount
26522 @opindex mnop-mcount
26523 If profiling is active (@option{-pg}), generate the calls to
26524 the profiling functions as NOPs. This is useful when they
26525 should be patched in later dynamically. This is likely only
26526 useful together with @option{-mrecord-mcount}.
26528 @item -mskip-rax-setup
26529 @itemx -mno-skip-rax-setup
26530 @opindex mskip-rax-setup
26531 When generating code for the x86-64 architecture with SSE extensions
26532 disabled, @option{-mskip-rax-setup} can be used to skip setting up RAX
26533 register when there are no variable arguments passed in vector registers.
26535 @strong{Warning:} Since RAX register is used to avoid unnecessarily
26536 saving vector registers on stack when passing variable arguments, the
26537 impacts of this option are callees may waste some stack space,
26538 misbehave or jump to a random location. GCC 4.4 or newer don't have
26539 those issues, regardless the RAX register value.
26542 @itemx -mno-8bit-idiv
26543 @opindex m8bit-idiv
26544 On some processors, like Intel Atom, 8-bit unsigned integer divide is
26545 much faster than 32-bit/64-bit integer divide. This option generates a
26546 run-time check. If both dividend and divisor are within range of 0
26547 to 255, 8-bit unsigned integer divide is used instead of
26548 32-bit/64-bit integer divide.
26550 @item -mavx256-split-unaligned-load
26551 @itemx -mavx256-split-unaligned-store
26552 @opindex mavx256-split-unaligned-load
26553 @opindex mavx256-split-unaligned-store
26554 Split 32-byte AVX unaligned load and store.
26556 @item -mstack-protector-guard=@var{guard}
26557 @itemx -mstack-protector-guard-reg=@var{reg}
26558 @itemx -mstack-protector-guard-offset=@var{offset}
26559 @opindex mstack-protector-guard
26560 @opindex mstack-protector-guard-reg
26561 @opindex mstack-protector-guard-offset
26562 Generate stack protection code using canary at @var{guard}. Supported
26563 locations are @samp{global} for global canary or @samp{tls} for per-thread
26564 canary in the TLS block (the default). This option has effect only when
26565 @option{-fstack-protector} or @option{-fstack-protector-all} is specified.
26567 With the latter choice the options
26568 @option{-mstack-protector-guard-reg=@var{reg}} and
26569 @option{-mstack-protector-guard-offset=@var{offset}} furthermore specify
26570 which segment register (@code{%fs} or @code{%gs}) to use as base register
26571 for reading the canary, and from what offset from that base register.
26572 The default for those is as specified in the relevant ABI.
26574 @item -mmitigate-rop
26575 @opindex mmitigate-rop
26576 Try to avoid generating code sequences that contain unintended return
26577 opcodes, to mitigate against certain forms of attack. At the moment,
26578 this option is limited in what it can do and should not be relied
26579 on to provide serious protection.
26581 @item -mgeneral-regs-only
26582 @opindex mgeneral-regs-only
26583 Generate code that uses only the general-purpose registers. This
26584 prevents the compiler from using floating-point, vector, mask and bound
26589 These @samp{-m} switches are supported in addition to the above
26590 on x86-64 processors in 64-bit environments.
26603 Generate code for a 16-bit, 32-bit or 64-bit environment.
26604 The @option{-m32} option sets @code{int}, @code{long}, and pointer types
26606 generates code that runs on any i386 system.
26608 The @option{-m64} option sets @code{int} to 32 bits and @code{long} and pointer
26609 types to 64 bits, and generates code for the x86-64 architecture.
26610 For Darwin only the @option{-m64} option also turns off the @option{-fno-pic}
26611 and @option{-mdynamic-no-pic} options.
26613 The @option{-mx32} option sets @code{int}, @code{long}, and pointer types
26615 generates code for the x86-64 architecture.
26617 The @option{-m16} option is the same as @option{-m32}, except for that
26618 it outputs the @code{.code16gcc} assembly directive at the beginning of
26619 the assembly output so that the binary can run in 16-bit mode.
26621 The @option{-miamcu} option generates code which conforms to Intel MCU
26622 psABI. It requires the @option{-m32} option to be turned on.
26624 @item -mno-red-zone
26625 @opindex mno-red-zone
26626 Do not use a so-called ``red zone'' for x86-64 code. The red zone is mandated
26627 by the x86-64 ABI; it is a 128-byte area beyond the location of the
26628 stack pointer that is not modified by signal or interrupt handlers
26629 and therefore can be used for temporary data without adjusting the stack
26630 pointer. The flag @option{-mno-red-zone} disables this red zone.
26632 @item -mcmodel=small
26633 @opindex mcmodel=small
26634 Generate code for the small code model: the program and its symbols must
26635 be linked in the lower 2 GB of the address space. Pointers are 64 bits.
26636 Programs can be statically or dynamically linked. This is the default
26639 @item -mcmodel=kernel
26640 @opindex mcmodel=kernel
26641 Generate code for the kernel code model. The kernel runs in the
26642 negative 2 GB of the address space.
26643 This model has to be used for Linux kernel code.
26645 @item -mcmodel=medium
26646 @opindex mcmodel=medium
26647 Generate code for the medium model: the program is linked in the lower 2
26648 GB of the address space. Small symbols are also placed there. Symbols
26649 with sizes larger than @option{-mlarge-data-threshold} are put into
26650 large data or BSS sections and can be located above 2GB. Programs can
26651 be statically or dynamically linked.
26653 @item -mcmodel=large
26654 @opindex mcmodel=large
26655 Generate code for the large model. This model makes no assumptions
26656 about addresses and sizes of sections.
26658 @item -maddress-mode=long
26659 @opindex maddress-mode=long
26660 Generate code for long address mode. This is only supported for 64-bit
26661 and x32 environments. It is the default address mode for 64-bit
26664 @item -maddress-mode=short
26665 @opindex maddress-mode=short
26666 Generate code for short address mode. This is only supported for 32-bit
26667 and x32 environments. It is the default address mode for 32-bit and
26671 @node x86 Windows Options
26672 @subsection x86 Windows Options
26673 @cindex x86 Windows Options
26674 @cindex Windows Options for x86
26676 These additional options are available for Microsoft Windows targets:
26682 specifies that a console application is to be generated, by
26683 instructing the linker to set the PE header subsystem type
26684 required for console applications.
26685 This option is available for Cygwin and MinGW targets and is
26686 enabled by default on those targets.
26690 This option is available for Cygwin and MinGW targets. It
26691 specifies that a DLL---a dynamic link library---is to be
26692 generated, enabling the selection of the required runtime
26693 startup object and entry point.
26695 @item -mnop-fun-dllimport
26696 @opindex mnop-fun-dllimport
26697 This option is available for Cygwin and MinGW targets. It
26698 specifies that the @code{dllimport} attribute should be ignored.
26702 This option is available for MinGW targets. It specifies
26703 that MinGW-specific thread support is to be used.
26707 This option is available for MinGW-w64 targets. It causes
26708 the @code{UNICODE} preprocessor macro to be predefined, and
26709 chooses Unicode-capable runtime startup code.
26713 This option is available for Cygwin and MinGW targets. It
26714 specifies that the typical Microsoft Windows predefined macros are to
26715 be set in the pre-processor, but does not influence the choice
26716 of runtime library/startup code.
26720 This option is available for Cygwin and MinGW targets. It
26721 specifies that a GUI application is to be generated by
26722 instructing the linker to set the PE header subsystem type
26725 @item -fno-set-stack-executable
26726 @opindex fno-set-stack-executable
26727 This option is available for MinGW targets. It specifies that
26728 the executable flag for the stack used by nested functions isn't
26729 set. This is necessary for binaries running in kernel mode of
26730 Microsoft Windows, as there the User32 API, which is used to set executable
26731 privileges, isn't available.
26733 @item -fwritable-relocated-rdata
26734 @opindex fno-writable-relocated-rdata
26735 This option is available for MinGW and Cygwin targets. It specifies
26736 that relocated-data in read-only section is put into the @code{.data}
26737 section. This is a necessary for older runtimes not supporting
26738 modification of @code{.rdata} sections for pseudo-relocation.
26740 @item -mpe-aligned-commons
26741 @opindex mpe-aligned-commons
26742 This option is available for Cygwin and MinGW targets. It
26743 specifies that the GNU extension to the PE file format that
26744 permits the correct alignment of COMMON variables should be
26745 used when generating code. It is enabled by default if
26746 GCC detects that the target assembler found during configuration
26747 supports the feature.
26750 See also under @ref{x86 Options} for standard options.
26752 @node Xstormy16 Options
26753 @subsection Xstormy16 Options
26754 @cindex Xstormy16 Options
26756 These options are defined for Xstormy16:
26761 Choose startup files and linker script suitable for the simulator.
26764 @node Xtensa Options
26765 @subsection Xtensa Options
26766 @cindex Xtensa Options
26768 These options are supported for Xtensa targets:
26772 @itemx -mno-const16
26774 @opindex mno-const16
26775 Enable or disable use of @code{CONST16} instructions for loading
26776 constant values. The @code{CONST16} instruction is currently not a
26777 standard option from Tensilica. When enabled, @code{CONST16}
26778 instructions are always used in place of the standard @code{L32R}
26779 instructions. The use of @code{CONST16} is enabled by default only if
26780 the @code{L32R} instruction is not available.
26783 @itemx -mno-fused-madd
26784 @opindex mfused-madd
26785 @opindex mno-fused-madd
26786 Enable or disable use of fused multiply/add and multiply/subtract
26787 instructions in the floating-point option. This has no effect if the
26788 floating-point option is not also enabled. Disabling fused multiply/add
26789 and multiply/subtract instructions forces the compiler to use separate
26790 instructions for the multiply and add/subtract operations. This may be
26791 desirable in some cases where strict IEEE 754-compliant results are
26792 required: the fused multiply add/subtract instructions do not round the
26793 intermediate result, thereby producing results with @emph{more} bits of
26794 precision than specified by the IEEE standard. Disabling fused multiply
26795 add/subtract instructions also ensures that the program output is not
26796 sensitive to the compiler's ability to combine multiply and add/subtract
26799 @item -mserialize-volatile
26800 @itemx -mno-serialize-volatile
26801 @opindex mserialize-volatile
26802 @opindex mno-serialize-volatile
26803 When this option is enabled, GCC inserts @code{MEMW} instructions before
26804 @code{volatile} memory references to guarantee sequential consistency.
26805 The default is @option{-mserialize-volatile}. Use
26806 @option{-mno-serialize-volatile} to omit the @code{MEMW} instructions.
26808 @item -mforce-no-pic
26809 @opindex mforce-no-pic
26810 For targets, like GNU/Linux, where all user-mode Xtensa code must be
26811 position-independent code (PIC), this option disables PIC for compiling
26814 @item -mtext-section-literals
26815 @itemx -mno-text-section-literals
26816 @opindex mtext-section-literals
26817 @opindex mno-text-section-literals
26818 These options control the treatment of literal pools. The default is
26819 @option{-mno-text-section-literals}, which places literals in a separate
26820 section in the output file. This allows the literal pool to be placed
26821 in a data RAM/ROM, and it also allows the linker to combine literal
26822 pools from separate object files to remove redundant literals and
26823 improve code size. With @option{-mtext-section-literals}, the literals
26824 are interspersed in the text section in order to keep them as close as
26825 possible to their references. This may be necessary for large assembly
26826 files. Literals for each function are placed right before that function.
26828 @item -mauto-litpools
26829 @itemx -mno-auto-litpools
26830 @opindex mauto-litpools
26831 @opindex mno-auto-litpools
26832 These options control the treatment of literal pools. The default is
26833 @option{-mno-auto-litpools}, which places literals in a separate
26834 section in the output file unless @option{-mtext-section-literals} is
26835 used. With @option{-mauto-litpools} the literals are interspersed in
26836 the text section by the assembler. Compiler does not produce explicit
26837 @code{.literal} directives and loads literals into registers with
26838 @code{MOVI} instructions instead of @code{L32R} to let the assembler
26839 do relaxation and place literals as necessary. This option allows
26840 assembler to create several literal pools per function and assemble
26841 very big functions, which may not be possible with
26842 @option{-mtext-section-literals}.
26844 @item -mtarget-align
26845 @itemx -mno-target-align
26846 @opindex mtarget-align
26847 @opindex mno-target-align
26848 When this option is enabled, GCC instructs the assembler to
26849 automatically align instructions to reduce branch penalties at the
26850 expense of some code density. The assembler attempts to widen density
26851 instructions to align branch targets and the instructions following call
26852 instructions. If there are not enough preceding safe density
26853 instructions to align a target, no widening is performed. The
26854 default is @option{-mtarget-align}. These options do not affect the
26855 treatment of auto-aligned instructions like @code{LOOP}, which the
26856 assembler always aligns, either by widening density instructions or
26857 by inserting NOP instructions.
26860 @itemx -mno-longcalls
26861 @opindex mlongcalls
26862 @opindex mno-longcalls
26863 When this option is enabled, GCC instructs the assembler to translate
26864 direct calls to indirect calls unless it can determine that the target
26865 of a direct call is in the range allowed by the call instruction. This
26866 translation typically occurs for calls to functions in other source
26867 files. Specifically, the assembler translates a direct @code{CALL}
26868 instruction into an @code{L32R} followed by a @code{CALLX} instruction.
26869 The default is @option{-mno-longcalls}. This option should be used in
26870 programs where the call target can potentially be out of range. This
26871 option is implemented in the assembler, not the compiler, so the
26872 assembly code generated by GCC still shows direct call
26873 instructions---look at the disassembled object code to see the actual
26874 instructions. Note that the assembler uses an indirect call for
26875 every cross-file call, not just those that really are out of range.
26878 @node zSeries Options
26879 @subsection zSeries Options
26880 @cindex zSeries options
26882 These are listed under @xref{S/390 and zSeries Options}.
26888 @section Specifying Subprocesses and the Switches to Pass to Them
26891 @command{gcc} is a driver program. It performs its job by invoking a
26892 sequence of other programs to do the work of compiling, assembling and
26893 linking. GCC interprets its command-line parameters and uses these to
26894 deduce which programs it should invoke, and which command-line options
26895 it ought to place on their command lines. This behavior is controlled
26896 by @dfn{spec strings}. In most cases there is one spec string for each
26897 program that GCC can invoke, but a few programs have multiple spec
26898 strings to control their behavior. The spec strings built into GCC can
26899 be overridden by using the @option{-specs=} command-line switch to specify
26902 @dfn{Spec files} are plain-text files that are used to construct spec
26903 strings. They consist of a sequence of directives separated by blank
26904 lines. The type of directive is determined by the first non-whitespace
26905 character on the line, which can be one of the following:
26908 @item %@var{command}
26909 Issues a @var{command} to the spec file processor. The commands that can
26913 @item %include <@var{file}>
26914 @cindex @code{%include}
26915 Search for @var{file} and insert its text at the current point in the
26918 @item %include_noerr <@var{file}>
26919 @cindex @code{%include_noerr}
26920 Just like @samp{%include}, but do not generate an error message if the include
26921 file cannot be found.
26923 @item %rename @var{old_name} @var{new_name}
26924 @cindex @code{%rename}
26925 Rename the spec string @var{old_name} to @var{new_name}.
26929 @item *[@var{spec_name}]:
26930 This tells the compiler to create, override or delete the named spec
26931 string. All lines after this directive up to the next directive or
26932 blank line are considered to be the text for the spec string. If this
26933 results in an empty string then the spec is deleted. (Or, if the
26934 spec did not exist, then nothing happens.) Otherwise, if the spec
26935 does not currently exist a new spec is created. If the spec does
26936 exist then its contents are overridden by the text of this
26937 directive, unless the first character of that text is the @samp{+}
26938 character, in which case the text is appended to the spec.
26940 @item [@var{suffix}]:
26941 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
26942 and up to the next directive or blank line are considered to make up the
26943 spec string for the indicated suffix. When the compiler encounters an
26944 input file with the named suffix, it processes the spec string in
26945 order to work out how to compile that file. For example:
26949 z-compile -input %i
26952 This says that any input file whose name ends in @samp{.ZZ} should be
26953 passed to the program @samp{z-compile}, which should be invoked with the
26954 command-line switch @option{-input} and with the result of performing the
26955 @samp{%i} substitution. (See below.)
26957 As an alternative to providing a spec string, the text following a
26958 suffix directive can be one of the following:
26961 @item @@@var{language}
26962 This says that the suffix is an alias for a known @var{language}. This is
26963 similar to using the @option{-x} command-line switch to GCC to specify a
26964 language explicitly. For example:
26971 Says that .ZZ files are, in fact, C++ source files.
26974 This causes an error messages saying:
26977 @var{name} compiler not installed on this system.
26981 GCC already has an extensive list of suffixes built into it.
26982 This directive adds an entry to the end of the list of suffixes, but
26983 since the list is searched from the end backwards, it is effectively
26984 possible to override earlier entries using this technique.
26988 GCC has the following spec strings built into it. Spec files can
26989 override these strings or create their own. Note that individual
26990 targets can also add their own spec strings to this list.
26993 asm Options to pass to the assembler
26994 asm_final Options to pass to the assembler post-processor
26995 cpp Options to pass to the C preprocessor
26996 cc1 Options to pass to the C compiler
26997 cc1plus Options to pass to the C++ compiler
26998 endfile Object files to include at the end of the link
26999 link Options to pass to the linker
27000 lib Libraries to include on the command line to the linker
27001 libgcc Decides which GCC support library to pass to the linker
27002 linker Sets the name of the linker
27003 predefines Defines to be passed to the C preprocessor
27004 signed_char Defines to pass to CPP to say whether @code{char} is signed
27006 startfile Object files to include at the start of the link
27009 Here is a small example of a spec file:
27012 %rename lib old_lib
27015 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
27018 This example renames the spec called @samp{lib} to @samp{old_lib} and
27019 then overrides the previous definition of @samp{lib} with a new one.
27020 The new definition adds in some extra command-line options before
27021 including the text of the old definition.
27023 @dfn{Spec strings} are a list of command-line options to be passed to their
27024 corresponding program. In addition, the spec strings can contain
27025 @samp{%}-prefixed sequences to substitute variable text or to
27026 conditionally insert text into the command line. Using these constructs
27027 it is possible to generate quite complex command lines.
27029 Here is a table of all defined @samp{%}-sequences for spec
27030 strings. Note that spaces are not generated automatically around the
27031 results of expanding these sequences. Therefore you can concatenate them
27032 together or combine them with constant text in a single argument.
27036 Substitute one @samp{%} into the program name or argument.
27039 Substitute the name of the input file being processed.
27042 Substitute the basename of the input file being processed.
27043 This is the substring up to (and not including) the last period
27044 and not including the directory.
27047 This is the same as @samp{%b}, but include the file suffix (text after
27051 Marks the argument containing or following the @samp{%d} as a
27052 temporary file name, so that that file is deleted if GCC exits
27053 successfully. Unlike @samp{%g}, this contributes no text to the
27056 @item %g@var{suffix}
27057 Substitute a file name that has suffix @var{suffix} and is chosen
27058 once per compilation, and mark the argument in the same way as
27059 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
27060 name is now chosen in a way that is hard to predict even when previously
27061 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
27062 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
27063 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
27064 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
27065 was simply substituted with a file name chosen once per compilation,
27066 without regard to any appended suffix (which was therefore treated
27067 just like ordinary text), making such attacks more likely to succeed.
27069 @item %u@var{suffix}
27070 Like @samp{%g}, but generates a new temporary file name
27071 each time it appears instead of once per compilation.
27073 @item %U@var{suffix}
27074 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
27075 new one if there is no such last file name. In the absence of any
27076 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
27077 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
27078 involves the generation of two distinct file names, one
27079 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
27080 simply substituted with a file name chosen for the previous @samp{%u},
27081 without regard to any appended suffix.
27083 @item %j@var{suffix}
27084 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
27085 writable, and if @option{-save-temps} is not used;
27086 otherwise, substitute the name
27087 of a temporary file, just like @samp{%u}. This temporary file is not
27088 meant for communication between processes, but rather as a junk
27089 disposal mechanism.
27091 @item %|@var{suffix}
27092 @itemx %m@var{suffix}
27093 Like @samp{%g}, except if @option{-pipe} is in effect. In that case
27094 @samp{%|} substitutes a single dash and @samp{%m} substitutes nothing at
27095 all. These are the two most common ways to instruct a program that it
27096 should read from standard input or write to standard output. If you
27097 need something more elaborate you can use an @samp{%@{pipe:@code{X}@}}
27098 construct: see for example @file{f/lang-specs.h}.
27100 @item %.@var{SUFFIX}
27101 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
27102 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
27103 terminated by the next space or %.
27106 Marks the argument containing or following the @samp{%w} as the
27107 designated output file of this compilation. This puts the argument
27108 into the sequence of arguments that @samp{%o} substitutes.
27111 Substitutes the names of all the output files, with spaces
27112 automatically placed around them. You should write spaces
27113 around the @samp{%o} as well or the results are undefined.
27114 @samp{%o} is for use in the specs for running the linker.
27115 Input files whose names have no recognized suffix are not compiled
27116 at all, but they are included among the output files, so they are
27120 Substitutes the suffix for object files. Note that this is
27121 handled specially when it immediately follows @samp{%g, %u, or %U},
27122 because of the need for those to form complete file names. The
27123 handling is such that @samp{%O} is treated exactly as if it had already
27124 been substituted, except that @samp{%g, %u, and %U} do not currently
27125 support additional @var{suffix} characters following @samp{%O} as they do
27126 following, for example, @samp{.o}.
27129 Substitutes the standard macro predefinitions for the
27130 current target machine. Use this when running @command{cpp}.
27133 Like @samp{%p}, but puts @samp{__} before and after the name of each
27134 predefined macro, except for macros that start with @samp{__} or with
27135 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
27139 Substitute any of @option{-iprefix} (made from @env{GCC_EXEC_PREFIX}),
27140 @option{-isysroot} (made from @env{TARGET_SYSTEM_ROOT}),
27141 @option{-isystem} (made from @env{COMPILER_PATH} and @option{-B} options)
27142 and @option{-imultilib} as necessary.
27145 Current argument is the name of a library or startup file of some sort.
27146 Search for that file in a standard list of directories and substitute
27147 the full name found. The current working directory is included in the
27148 list of directories scanned.
27151 Current argument is the name of a linker script. Search for that file
27152 in the current list of directories to scan for libraries. If the file
27153 is located insert a @option{--script} option into the command line
27154 followed by the full path name found. If the file is not found then
27155 generate an error message. Note: the current working directory is not
27159 Print @var{str} as an error message. @var{str} is terminated by a newline.
27160 Use this when inconsistent options are detected.
27162 @item %(@var{name})
27163 Substitute the contents of spec string @var{name} at this point.
27165 @item %x@{@var{option}@}
27166 Accumulate an option for @samp{%X}.
27169 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
27173 Output the accumulated assembler options specified by @option{-Wa}.
27176 Output the accumulated preprocessor options specified by @option{-Wp}.
27179 Process the @code{asm} spec. This is used to compute the
27180 switches to be passed to the assembler.
27183 Process the @code{asm_final} spec. This is a spec string for
27184 passing switches to an assembler post-processor, if such a program is
27188 Process the @code{link} spec. This is the spec for computing the
27189 command line passed to the linker. Typically it makes use of the
27190 @samp{%L %G %S %D and %E} sequences.
27193 Dump out a @option{-L} option for each directory that GCC believes might
27194 contain startup files. If the target supports multilibs then the
27195 current multilib directory is prepended to each of these paths.
27198 Process the @code{lib} spec. This is a spec string for deciding which
27199 libraries are included on the command line to the linker.
27202 Process the @code{libgcc} spec. This is a spec string for deciding
27203 which GCC support library is included on the command line to the linker.
27206 Process the @code{startfile} spec. This is a spec for deciding which
27207 object files are the first ones passed to the linker. Typically
27208 this might be a file named @file{crt0.o}.
27211 Process the @code{endfile} spec. This is a spec string that specifies
27212 the last object files that are passed to the linker.
27215 Process the @code{cpp} spec. This is used to construct the arguments
27216 to be passed to the C preprocessor.
27219 Process the @code{cc1} spec. This is used to construct the options to be
27220 passed to the actual C compiler (@command{cc1}).
27223 Process the @code{cc1plus} spec. This is used to construct the options to be
27224 passed to the actual C++ compiler (@command{cc1plus}).
27227 Substitute the variable part of a matched option. See below.
27228 Note that each comma in the substituted string is replaced by
27232 Remove all occurrences of @code{-S} from the command line. Note---this
27233 command is position dependent. @samp{%} commands in the spec string
27234 before this one see @code{-S}, @samp{%} commands in the spec string
27235 after this one do not.
27237 @item %:@var{function}(@var{args})
27238 Call the named function @var{function}, passing it @var{args}.
27239 @var{args} is first processed as a nested spec string, then split
27240 into an argument vector in the usual fashion. The function returns
27241 a string which is processed as if it had appeared literally as part
27242 of the current spec.
27244 The following built-in spec functions are provided:
27247 @item @code{getenv}
27248 The @code{getenv} spec function takes two arguments: an environment
27249 variable name and a string. If the environment variable is not
27250 defined, a fatal error is issued. Otherwise, the return value is the
27251 value of the environment variable concatenated with the string. For
27252 example, if @env{TOPDIR} is defined as @file{/path/to/top}, then:
27255 %:getenv(TOPDIR /include)
27258 expands to @file{/path/to/top/include}.
27260 @item @code{if-exists}
27261 The @code{if-exists} spec function takes one argument, an absolute
27262 pathname to a file. If the file exists, @code{if-exists} returns the
27263 pathname. Here is a small example of its usage:
27267 crt0%O%s %:if-exists(crti%O%s) crtbegin%O%s
27270 @item @code{if-exists-else}
27271 The @code{if-exists-else} spec function is similar to the @code{if-exists}
27272 spec function, except that it takes two arguments. The first argument is
27273 an absolute pathname to a file. If the file exists, @code{if-exists-else}
27274 returns the pathname. If it does not exist, it returns the second argument.
27275 This way, @code{if-exists-else} can be used to select one file or another,
27276 based on the existence of the first. Here is a small example of its usage:
27280 crt0%O%s %:if-exists(crti%O%s) \
27281 %:if-exists-else(crtbeginT%O%s crtbegin%O%s)
27284 @item @code{replace-outfile}
27285 The @code{replace-outfile} spec function takes two arguments. It looks for the
27286 first argument in the outfiles array and replaces it with the second argument. Here
27287 is a small example of its usage:
27290 %@{fgnu-runtime:%:replace-outfile(-lobjc -lobjc-gnu)@}
27293 @item @code{remove-outfile}
27294 The @code{remove-outfile} spec function takes one argument. It looks for the
27295 first argument in the outfiles array and removes it. Here is a small example
27299 %:remove-outfile(-lm)
27302 @item @code{pass-through-libs}
27303 The @code{pass-through-libs} spec function takes any number of arguments. It
27304 finds any @option{-l} options and any non-options ending in @file{.a} (which it
27305 assumes are the names of linker input library archive files) and returns a
27306 result containing all the found arguments each prepended by
27307 @option{-plugin-opt=-pass-through=} and joined by spaces. This list is
27308 intended to be passed to the LTO linker plugin.
27311 %:pass-through-libs(%G %L %G)
27314 @item @code{print-asm-header}
27315 The @code{print-asm-header} function takes no arguments and simply
27316 prints a banner like:
27322 Use "-Wa,OPTION" to pass "OPTION" to the assembler.
27325 It is used to separate compiler options from assembler options
27326 in the @option{--target-help} output.
27330 Substitutes the @code{-S} switch, if that switch is given to GCC@.
27331 If that switch is not specified, this substitutes nothing. Note that
27332 the leading dash is omitted when specifying this option, and it is
27333 automatically inserted if the substitution is performed. Thus the spec
27334 string @samp{%@{foo@}} matches the command-line option @option{-foo}
27335 and outputs the command-line option @option{-foo}.
27338 Like %@{@code{S}@} but mark last argument supplied within as a file to be
27339 deleted on failure.
27342 Substitutes all the switches specified to GCC whose names start
27343 with @code{-S}, but which also take an argument. This is used for
27344 switches like @option{-o}, @option{-D}, @option{-I}, etc.
27345 GCC considers @option{-o foo} as being
27346 one switch whose name starts with @samp{o}. %@{o*@} substitutes this
27347 text, including the space. Thus two arguments are generated.
27350 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
27351 (the order of @code{S} and @code{T} in the spec is not significant).
27352 There can be any number of ampersand-separated variables; for each the
27353 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
27356 Substitutes @code{X}, if the @option{-S} switch is given to GCC@.
27359 Substitutes @code{X}, if the @option{-S} switch is @emph{not} given to GCC@.
27362 Substitutes @code{X} if one or more switches whose names start with
27363 @code{-S} are specified to GCC@. Normally @code{X} is substituted only
27364 once, no matter how many such switches appeared. However, if @code{%*}
27365 appears somewhere in @code{X}, then @code{X} is substituted once
27366 for each matching switch, with the @code{%*} replaced by the part of
27367 that switch matching the @code{*}.
27369 If @code{%*} appears as the last part of a spec sequence then a space
27370 is added after the end of the last substitution. If there is more
27371 text in the sequence, however, then a space is not generated. This
27372 allows the @code{%*} substitution to be used as part of a larger
27373 string. For example, a spec string like this:
27376 %@{mcu=*:--script=%*/memory.ld@}
27380 when matching an option like @option{-mcu=newchip} produces:
27383 --script=newchip/memory.ld
27387 Substitutes @code{X}, if processing a file with suffix @code{S}.
27390 Substitutes @code{X}, if @emph{not} processing a file with suffix @code{S}.
27393 Substitutes @code{X}, if processing a file for language @code{S}.
27396 Substitutes @code{X}, if not processing a file for language @code{S}.
27399 Substitutes @code{X} if either @code{-S} or @code{-P} is given to
27400 GCC@. This may be combined with @samp{!}, @samp{.}, @samp{,}, and
27401 @code{*} sequences as well, although they have a stronger binding than
27402 the @samp{|}. If @code{%*} appears in @code{X}, all of the
27403 alternatives must be starred, and only the first matching alternative
27406 For example, a spec string like this:
27409 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
27413 outputs the following command-line options from the following input
27414 command-line options:
27419 -d fred.c -foo -baz -boggle
27420 -d jim.d -bar -baz -boggle
27423 @item %@{S:X; T:Y; :D@}
27425 If @code{S} is given to GCC, substitutes @code{X}; else if @code{T} is
27426 given to GCC, substitutes @code{Y}; else substitutes @code{D}. There can
27427 be as many clauses as you need. This may be combined with @code{.},
27428 @code{,}, @code{!}, @code{|}, and @code{*} as needed.
27433 The switch matching text @code{S} in a @samp{%@{S@}}, @samp{%@{S:X@}}
27434 or similar construct can use a backslash to ignore the special meaning
27435 of the character following it, thus allowing literal matching of a
27436 character that is otherwise specially treated. For example,
27437 @samp{%@{std=iso9899\:1999:X@}} substitutes @code{X} if the
27438 @option{-std=iso9899:1999} option is given.
27440 The conditional text @code{X} in a @samp{%@{S:X@}} or similar
27441 construct may contain other nested @samp{%} constructs or spaces, or
27442 even newlines. They are processed as usual, as described above.
27443 Trailing white space in @code{X} is ignored. White space may also
27444 appear anywhere on the left side of the colon in these constructs,
27445 except between @code{.} or @code{*} and the corresponding word.
27447 The @option{-O}, @option{-f}, @option{-m}, and @option{-W} switches are
27448 handled specifically in these constructs. If another value of
27449 @option{-O} or the negated form of a @option{-f}, @option{-m}, or
27450 @option{-W} switch is found later in the command line, the earlier
27451 switch value is ignored, except with @{@code{S}*@} where @code{S} is
27452 just one letter, which passes all matching options.
27454 The character @samp{|} at the beginning of the predicate text is used to
27455 indicate that a command should be piped to the following command, but
27456 only if @option{-pipe} is specified.
27458 It is built into GCC which switches take arguments and which do not.
27459 (You might think it would be useful to generalize this to allow each
27460 compiler's spec to say which switches take arguments. But this cannot
27461 be done in a consistent fashion. GCC cannot even decide which input
27462 files have been specified without knowing which switches take arguments,
27463 and it must know which input files to compile in order to tell which
27466 GCC also knows implicitly that arguments starting in @option{-l} are to be
27467 treated as compiler output files, and passed to the linker in their
27468 proper position among the other output files.
27470 @node Environment Variables
27471 @section Environment Variables Affecting GCC
27472 @cindex environment variables
27474 @c man begin ENVIRONMENT
27475 This section describes several environment variables that affect how GCC
27476 operates. Some of them work by specifying directories or prefixes to use
27477 when searching for various kinds of files. Some are used to specify other
27478 aspects of the compilation environment.
27480 Note that you can also specify places to search using options such as
27481 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
27482 take precedence over places specified using environment variables, which
27483 in turn take precedence over those specified by the configuration of GCC@.
27484 @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint,
27485 GNU Compiler Collection (GCC) Internals}.
27490 @c @itemx LC_COLLATE
27492 @c @itemx LC_MONETARY
27493 @c @itemx LC_NUMERIC
27498 @c @findex LC_COLLATE
27499 @findex LC_MESSAGES
27500 @c @findex LC_MONETARY
27501 @c @findex LC_NUMERIC
27505 These environment variables control the way that GCC uses
27506 localization information which allows GCC to work with different
27507 national conventions. GCC inspects the locale categories
27508 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
27509 so. These locale categories can be set to any value supported by your
27510 installation. A typical value is @samp{en_GB.UTF-8} for English in the United
27511 Kingdom encoded in UTF-8.
27513 The @env{LC_CTYPE} environment variable specifies character
27514 classification. GCC uses it to determine the character boundaries in
27515 a string; this is needed for some multibyte encodings that contain quote
27516 and escape characters that are otherwise interpreted as a string
27519 The @env{LC_MESSAGES} environment variable specifies the language to
27520 use in diagnostic messages.
27522 If the @env{LC_ALL} environment variable is set, it overrides the value
27523 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
27524 and @env{LC_MESSAGES} default to the value of the @env{LANG}
27525 environment variable. If none of these variables are set, GCC
27526 defaults to traditional C English behavior.
27530 If @env{TMPDIR} is set, it specifies the directory to use for temporary
27531 files. GCC uses temporary files to hold the output of one stage of
27532 compilation which is to be used as input to the next stage: for example,
27533 the output of the preprocessor, which is the input to the compiler
27536 @item GCC_COMPARE_DEBUG
27537 @findex GCC_COMPARE_DEBUG
27538 Setting @env{GCC_COMPARE_DEBUG} is nearly equivalent to passing
27539 @option{-fcompare-debug} to the compiler driver. See the documentation
27540 of this option for more details.
27542 @item GCC_EXEC_PREFIX
27543 @findex GCC_EXEC_PREFIX
27544 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
27545 names of the subprograms executed by the compiler. No slash is added
27546 when this prefix is combined with the name of a subprogram, but you can
27547 specify a prefix that ends with a slash if you wish.
27549 If @env{GCC_EXEC_PREFIX} is not set, GCC attempts to figure out
27550 an appropriate prefix to use based on the pathname it is invoked with.
27552 If GCC cannot find the subprogram using the specified prefix, it
27553 tries looking in the usual places for the subprogram.
27555 The default value of @env{GCC_EXEC_PREFIX} is
27556 @file{@var{prefix}/lib/gcc/} where @var{prefix} is the prefix to
27557 the installed compiler. In many cases @var{prefix} is the value
27558 of @code{prefix} when you ran the @file{configure} script.
27560 Other prefixes specified with @option{-B} take precedence over this prefix.
27562 This prefix is also used for finding files such as @file{crt0.o} that are
27565 In addition, the prefix is used in an unusual way in finding the
27566 directories to search for header files. For each of the standard
27567 directories whose name normally begins with @samp{/usr/local/lib/gcc}
27568 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
27569 replacing that beginning with the specified prefix to produce an
27570 alternate directory name. Thus, with @option{-Bfoo/}, GCC searches
27571 @file{foo/bar} just before it searches the standard directory
27572 @file{/usr/local/lib/bar}.
27573 If a standard directory begins with the configured
27574 @var{prefix} then the value of @var{prefix} is replaced by
27575 @env{GCC_EXEC_PREFIX} when looking for header files.
27577 @item COMPILER_PATH
27578 @findex COMPILER_PATH
27579 The value of @env{COMPILER_PATH} is a colon-separated list of
27580 directories, much like @env{PATH}. GCC tries the directories thus
27581 specified when searching for subprograms, if it cannot find the
27582 subprograms using @env{GCC_EXEC_PREFIX}.
27585 @findex LIBRARY_PATH
27586 The value of @env{LIBRARY_PATH} is a colon-separated list of
27587 directories, much like @env{PATH}. When configured as a native compiler,
27588 GCC tries the directories thus specified when searching for special
27589 linker files, if it cannot find them using @env{GCC_EXEC_PREFIX}. Linking
27590 using GCC also uses these directories when searching for ordinary
27591 libraries for the @option{-l} option (but directories specified with
27592 @option{-L} come first).
27596 @cindex locale definition
27597 This variable is used to pass locale information to the compiler. One way in
27598 which this information is used is to determine the character set to be used
27599 when character literals, string literals and comments are parsed in C and C++.
27600 When the compiler is configured to allow multibyte characters,
27601 the following values for @env{LANG} are recognized:
27605 Recognize JIS characters.
27607 Recognize SJIS characters.
27609 Recognize EUCJP characters.
27612 If @env{LANG} is not defined, or if it has some other value, then the
27613 compiler uses @code{mblen} and @code{mbtowc} as defined by the default locale to
27614 recognize and translate multibyte characters.
27618 Some additional environment variables affect the behavior of the
27621 @include cppenv.texi
27625 @node Precompiled Headers
27626 @section Using Precompiled Headers
27627 @cindex precompiled headers
27628 @cindex speed of compilation
27630 Often large projects have many header files that are included in every
27631 source file. The time the compiler takes to process these header files
27632 over and over again can account for nearly all of the time required to
27633 build the project. To make builds faster, GCC allows you to
27634 @dfn{precompile} a header file.
27636 To create a precompiled header file, simply compile it as you would any
27637 other file, if necessary using the @option{-x} option to make the driver
27638 treat it as a C or C++ header file. You may want to use a
27639 tool like @command{make} to keep the precompiled header up-to-date when
27640 the headers it contains change.
27642 A precompiled header file is searched for when @code{#include} is
27643 seen in the compilation. As it searches for the included file
27644 (@pxref{Search Path,,Search Path,cpp,The C Preprocessor}) the
27645 compiler looks for a precompiled header in each directory just before it
27646 looks for the include file in that directory. The name searched for is
27647 the name specified in the @code{#include} with @samp{.gch} appended. If
27648 the precompiled header file cannot be used, it is ignored.
27650 For instance, if you have @code{#include "all.h"}, and you have
27651 @file{all.h.gch} in the same directory as @file{all.h}, then the
27652 precompiled header file is used if possible, and the original
27653 header is used otherwise.
27655 Alternatively, you might decide to put the precompiled header file in a
27656 directory and use @option{-I} to ensure that directory is searched
27657 before (or instead of) the directory containing the original header.
27658 Then, if you want to check that the precompiled header file is always
27659 used, you can put a file of the same name as the original header in this
27660 directory containing an @code{#error} command.
27662 This also works with @option{-include}. So yet another way to use
27663 precompiled headers, good for projects not designed with precompiled
27664 header files in mind, is to simply take most of the header files used by
27665 a project, include them from another header file, precompile that header
27666 file, and @option{-include} the precompiled header. If the header files
27667 have guards against multiple inclusion, they are skipped because
27668 they've already been included (in the precompiled header).
27670 If you need to precompile the same header file for different
27671 languages, targets, or compiler options, you can instead make a
27672 @emph{directory} named like @file{all.h.gch}, and put each precompiled
27673 header in the directory, perhaps using @option{-o}. It doesn't matter
27674 what you call the files in the directory; every precompiled header in
27675 the directory is considered. The first precompiled header
27676 encountered in the directory that is valid for this compilation is
27677 used; they're searched in no particular order.
27679 There are many other possibilities, limited only by your imagination,
27680 good sense, and the constraints of your build system.
27682 A precompiled header file can be used only when these conditions apply:
27686 Only one precompiled header can be used in a particular compilation.
27689 A precompiled header cannot be used once the first C token is seen. You
27690 can have preprocessor directives before a precompiled header; you cannot
27691 include a precompiled header from inside another header.
27694 The precompiled header file must be produced for the same language as
27695 the current compilation. You cannot use a C precompiled header for a C++
27699 The precompiled header file must have been produced by the same compiler
27700 binary as the current compilation is using.
27703 Any macros defined before the precompiled header is included must
27704 either be defined in the same way as when the precompiled header was
27705 generated, or must not affect the precompiled header, which usually
27706 means that they don't appear in the precompiled header at all.
27708 The @option{-D} option is one way to define a macro before a
27709 precompiled header is included; using a @code{#define} can also do it.
27710 There are also some options that define macros implicitly, like
27711 @option{-O} and @option{-Wdeprecated}; the same rule applies to macros
27714 @item If debugging information is output when using the precompiled
27715 header, using @option{-g} or similar, the same kind of debugging information
27716 must have been output when building the precompiled header. However,
27717 a precompiled header built using @option{-g} can be used in a compilation
27718 when no debugging information is being output.
27720 @item The same @option{-m} options must generally be used when building
27721 and using the precompiled header. @xref{Submodel Options},
27722 for any cases where this rule is relaxed.
27724 @item Each of the following options must be the same when building and using
27725 the precompiled header:
27727 @gccoptlist{-fexceptions}
27730 Some other command-line options starting with @option{-f},
27731 @option{-p}, or @option{-O} must be defined in the same way as when
27732 the precompiled header was generated. At present, it's not clear
27733 which options are safe to change and which are not; the safest choice
27734 is to use exactly the same options when generating and using the
27735 precompiled header. The following are known to be safe:
27737 @gccoptlist{-fmessage-length= -fpreprocessed -fsched-interblock @gol
27738 -fsched-spec -fsched-spec-load -fsched-spec-load-dangerous @gol
27739 -fsched-verbose=@var{number} -fschedule-insns -fvisibility= @gol
27744 For all of these except the last, the compiler automatically
27745 ignores the precompiled header if the conditions aren't met. If you
27746 find an option combination that doesn't work and doesn't cause the
27747 precompiled header to be ignored, please consider filing a bug report,
27750 If you do use differing options when generating and using the
27751 precompiled header, the actual behavior is a mixture of the
27752 behavior for the options. For instance, if you use @option{-g} to
27753 generate the precompiled header but not when using it, you may or may
27754 not get debugging information for routines in the precompiled header.