1 @c Copyright (C) 1988-2019 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{} 1988-2019 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.
124 Some options take one or more arguments typically separated either
125 by a space or by the equals sign (@samp{=}) from the option name.
126 Unless documented otherwise, an argument can be either numeric or
127 a string. Numeric arguments must typically be small unsigned decimal
128 or hexadecimal integers. Hexadecimal arguments must begin with
129 the @samp{0x} prefix. Arguments to options that specify a size
130 threshold of some sort may be arbitrarily large decimal or hexadecimal
131 integers followed by a byte size suffix designating a multiple of bytes
132 such as @code{kB} and @code{KiB} for kilobyte and kibibyte, respectively,
133 @code{MB} and @code{MiB} for megabyte and mebibyte, @code{GB} and
134 @code{GiB} for gigabyte and gigibyte, and so on. Such arguments are
135 designated by @var{byte-size} in the following text. Refer to the NIST,
136 IEC, and other relevant national and international standards for the full
137 listing and explanation of the binary and decimal byte size prefixes.
141 @xref{Option Index}, for an index to GCC's options.
144 * Option Summary:: Brief list of all options, without explanations.
145 * Overall Options:: Controlling the kind of output:
146 an executable, object files, assembler files,
147 or preprocessed source.
148 * Invoking G++:: Compiling C++ programs.
149 * C Dialect Options:: Controlling the variant of C language compiled.
150 * C++ Dialect Options:: Variations on C++.
151 * Objective-C and Objective-C++ Dialect Options:: Variations on Objective-C
153 * Diagnostic Message Formatting Options:: Controlling how diagnostics should
155 * Warning Options:: How picky should the compiler be?
156 * Debugging Options:: Producing debuggable code.
157 * Optimize Options:: How much optimization?
158 * Instrumentation Options:: Enabling profiling and extra run-time error checking.
159 * Preprocessor Options:: Controlling header files and macro definitions.
160 Also, getting dependency information for Make.
161 * Assembler Options:: Passing options to the assembler.
162 * Link Options:: Specifying libraries and so on.
163 * Directory Options:: Where to find header files and libraries.
164 Where to find the compiler executable files.
165 * Code Gen Options:: Specifying conventions for function calls, data layout
167 * Developer Options:: Printing GCC configuration info, statistics, and
169 * Submodel Options:: Target-specific options, such as compiling for a
170 specific processor variant.
171 * Spec Files:: How to pass switches to sub-processes.
172 * Environment Variables:: Env vars that affect GCC.
173 * Precompiled Headers:: Compiling a header once, and using it many times.
179 @section Option Summary
181 Here is a summary of all the options, grouped by type. Explanations are
182 in the following sections.
185 @item Overall Options
186 @xref{Overall Options,,Options Controlling the Kind of Output}.
187 @gccoptlist{-c -S -E -o @var{file} -x @var{language} @gol
188 -v -### --help@r{[}=@var{class}@r{[},@dots{}@r{]]} --target-help --version @gol
189 -pass-exit-codes -pipe -specs=@var{file} -wrapper @gol
190 @@@var{file} -ffile-prefix-map=@var{old}=@var{new} @gol
191 -fplugin=@var{file} -fplugin-arg-@var{name}=@var{arg} @gol
192 -fdump-ada-spec@r{[}-slim@r{]} -fada-spec-parent=@var{unit} -fdump-go-spec=@var{file}}
194 @item C Language Options
195 @xref{C Dialect Options,,Options Controlling C Dialect}.
196 @gccoptlist{-ansi -std=@var{standard} -fgnu89-inline @gol
197 -fpermitted-flt-eval-methods=@var{standard} @gol
198 -aux-info @var{filename} -fallow-parameterless-variadic-functions @gol
199 -fno-asm -fno-builtin -fno-builtin-@var{function} -fgimple@gol
200 -fhosted -ffreestanding @gol
201 -fopenacc -fopenacc-dim=@var{geom} @gol
202 -fopenmp -fopenmp-simd @gol
203 -fms-extensions -fplan9-extensions -fsso-struct=@var{endianness} @gol
204 -fallow-single-precision -fcond-mismatch -flax-vector-conversions @gol
205 -fsigned-bitfields -fsigned-char @gol
206 -funsigned-bitfields -funsigned-char}
208 @item C++ Language Options
209 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}.
210 @gccoptlist{-fabi-version=@var{n} -fno-access-control @gol
211 -faligned-new=@var{n} -fargs-in-order=@var{n} -fchar8_t -fcheck-new @gol
212 -fconstexpr-depth=@var{n} -fconstexpr-cache-depth=@var{n} @gol
213 -fconstexpr-loop-limit=@var{n} -fconstexpr-ops-limit=@var{n} @gol
214 -fno-elide-constructors @gol
215 -fno-enforce-eh-specs @gol
216 -fno-gnu-keywords @gol
217 -fno-implicit-templates @gol
218 -fno-implicit-inline-templates @gol
219 -fno-implement-inlines -fms-extensions @gol
220 -fnew-inheriting-ctors @gol
221 -fnew-ttp-matching @gol
222 -fno-nonansi-builtins -fnothrow-opt -fno-operator-names @gol
223 -fno-optional-diags -fpermissive @gol
224 -fno-pretty-templates @gol
225 -fno-rtti -fsized-deallocation @gol
226 -ftemplate-backtrace-limit=@var{n} @gol
227 -ftemplate-depth=@var{n} @gol
228 -fno-threadsafe-statics -fuse-cxa-atexit @gol
229 -fno-weak -nostdinc++ @gol
230 -fvisibility-inlines-hidden @gol
231 -fvisibility-ms-compat @gol
232 -fext-numeric-literals @gol
233 -Wabi=@var{n} -Wabi-tag -Wcomma-subscript -Wconversion-null @gol
234 -Wctor-dtor-privacy @gol
235 -Wdelete-non-virtual-dtor -Wdeprecated-copy -Wdeprecated-copy-dtor @gol
236 -Wliteral-suffix @gol
237 -Wmultiple-inheritance -Wno-init-list-lifetime @gol
238 -Wnamespaces -Wnarrowing @gol
239 -Wpessimizing-move -Wredundant-move @gol
240 -Wnoexcept -Wnoexcept-type -Wclass-memaccess @gol
241 -Wnon-virtual-dtor -Wreorder -Wregister @gol
242 -Weffc++ -Wstrict-null-sentinel -Wtemplates @gol
243 -Wno-non-template-friend -Wold-style-cast @gol
244 -Woverloaded-virtual -Wno-pmf-conversions @gol
245 -Wno-class-conversion -Wno-terminate @gol
246 -Wsign-promo -Wvirtual-inheritance -Wvolatile}
248 @item Objective-C and Objective-C++ Language Options
249 @xref{Objective-C and Objective-C++ Dialect Options,,Options Controlling
250 Objective-C and Objective-C++ Dialects}.
251 @gccoptlist{-fconstant-string-class=@var{class-name} @gol
252 -fgnu-runtime -fnext-runtime @gol
253 -fno-nil-receivers @gol
254 -fobjc-abi-version=@var{n} @gol
255 -fobjc-call-cxx-cdtors @gol
256 -fobjc-direct-dispatch @gol
257 -fobjc-exceptions @gol
260 -fobjc-std=objc1 @gol
261 -fno-local-ivars @gol
262 -fivar-visibility=@r{[}public@r{|}protected@r{|}private@r{|}package@r{]} @gol
263 -freplace-objc-classes @gol
266 -Wassign-intercept @gol
267 -Wno-protocol -Wselector @gol
268 -Wstrict-selector-match @gol
269 -Wundeclared-selector}
271 @item Diagnostic Message Formatting Options
272 @xref{Diagnostic Message Formatting Options,,Options to Control Diagnostic Messages Formatting}.
273 @gccoptlist{-fmessage-length=@var{n} @gol
274 -fdiagnostics-show-location=@r{[}once@r{|}every-line@r{]} @gol
275 -fdiagnostics-color=@r{[}auto@r{|}never@r{|}always@r{]} @gol
276 -fdiagnostics-format=@r{[}text@r{|}json@r{]} @gol
277 -fno-diagnostics-show-option -fno-diagnostics-show-caret @gol
278 -fno-diagnostics-show-labels -fno-diagnostics-show-line-numbers @gol
279 -fdiagnostics-minimum-margin-width=@var{width} @gol
280 -fdiagnostics-parseable-fixits -fdiagnostics-generate-patch @gol
281 -fdiagnostics-show-template-tree -fno-elide-type @gol
284 @item Warning Options
285 @xref{Warning Options,,Options to Request or Suppress Warnings}.
286 @gccoptlist{-fsyntax-only -fmax-errors=@var{n} -Wpedantic @gol
287 -pedantic-errors @gol
288 -w -Wextra -Wall -Waddress -Waddress-of-packed-member @gol
289 -Waggregate-return -Waligned-new @gol
290 -Walloc-zero -Walloc-size-larger-than=@var{byte-size} @gol
291 -Walloca -Walloca-larger-than=@var{byte-size} @gol
292 -Wno-aggressive-loop-optimizations -Warray-bounds -Warray-bounds=@var{n} @gol
293 -Wno-attributes -Wattribute-alias=@var{n} @gol
294 -Wbool-compare -Wbool-operation @gol
295 -Wno-builtin-declaration-mismatch @gol
296 -Wno-builtin-macro-redefined -Wc90-c99-compat -Wc99-c11-compat @gol
297 -Wc++-compat -Wc++11-compat -Wc++14-compat -Wc++17-compat @gol
299 -Wcast-align -Wcast-align=strict -Wcast-function-type -Wcast-qual @gol
300 -Wchar-subscripts -Wcatch-value -Wcatch-value=@var{n} @gol
301 -Wclobbered -Wcomment -Wconditionally-supported @gol
302 -Wconversion -Wcoverage-mismatch -Wno-cpp -Wdangling-else -Wdate-time @gol
303 -Wdelete-incomplete @gol
304 -Wno-attribute-warning @gol
305 -Wno-deprecated -Wno-deprecated-declarations -Wno-designated-init @gol
306 -Wdisabled-optimization @gol
307 -Wno-discarded-qualifiers -Wno-discarded-array-qualifiers @gol
308 -Wno-div-by-zero -Wdouble-promotion @gol
309 -Wduplicated-branches -Wduplicated-cond @gol
310 -Wempty-body -Wenum-compare -Wenum-conversion @gol
311 -Wno-endif-labels -Wexpansion-to-defined @gol
312 -Werror -Werror=* -Wextra-semi -Wfatal-errors @gol
313 -Wfloat-equal -Wformat -Wformat=2 @gol
314 -Wno-format-contains-nul -Wno-format-extra-args @gol
315 -Wformat-nonliteral -Wformat-overflow=@var{n} @gol
316 -Wformat-security -Wformat-signedness -Wformat-truncation=@var{n} @gol
317 -Wformat-y2k -Wframe-address @gol
318 -Wframe-larger-than=@var{byte-size} -Wno-free-nonheap-object @gol
319 -Wjump-misses-init @gol
320 -Whsa -Wif-not-aligned @gol
321 -Wignored-qualifiers -Wignored-attributes -Wincompatible-pointer-types @gol
322 -Wimplicit -Wimplicit-fallthrough -Wimplicit-fallthrough=@var{n} @gol
323 -Wimplicit-function-declaration -Wimplicit-int @gol
324 -Winaccessible-base @gol
325 -Winit-self -Winline -Wno-int-conversion -Wint-in-bool-context @gol
326 -Wno-int-to-pointer-cast -Winvalid-memory-model -Wno-invalid-offsetof @gol
327 -Winvalid-pch -Wlarger-than=@var{byte-size} @gol
328 -Wlogical-op -Wlogical-not-parentheses -Wlong-long @gol
329 -Wmain -Wmaybe-uninitialized -Wmemset-elt-size -Wmemset-transposed-args @gol
330 -Wmisleading-indentation -Wmissing-attributes -Wmissing-braces @gol
331 -Wmissing-field-initializers -Wmissing-format-attribute @gol
332 -Wmissing-include-dirs -Wmissing-noreturn -Wmissing-profile @gol
333 -Wno-multichar -Wmultistatement-macros -Wnonnull -Wnonnull-compare @gol
334 -Wnormalized=@r{[}none@r{|}id@r{|}nfc@r{|}nfkc@r{]} @gol
335 -Wnull-dereference -Wodr -Wno-overflow -Wopenmp-simd @gol
336 -Woverride-init-side-effects -Woverlength-strings @gol
337 -Wpacked -Wpacked-bitfield-compat -Wpacked-not-aligned -Wpadded @gol
338 -Wparentheses -Wno-pedantic-ms-format @gol
339 -Wplacement-new -Wplacement-new=@var{n} @gol
340 -Wpointer-arith -Wpointer-compare -Wno-pointer-to-int-cast @gol
341 -Wno-pragmas -Wno-prio-ctor-dtor -Wredundant-decls @gol
342 -Wrestrict -Wno-return-local-addr @gol
343 -Wreturn-type -Wsequence-point -Wshadow -Wno-shadow-ivar @gol
344 -Wshadow=global, -Wshadow=local, -Wshadow=compatible-local @gol
345 -Wshift-overflow -Wshift-overflow=@var{n} @gol
346 -Wshift-count-negative -Wshift-count-overflow -Wshift-negative-value @gol
347 -Wsign-compare -Wsign-conversion -Wfloat-conversion @gol
348 -Wno-scalar-storage-order -Wsizeof-pointer-div @gol
349 -Wsizeof-pointer-memaccess -Wsizeof-array-argument @gol
350 -Wstack-protector -Wstack-usage=@var{byte-size} -Wstrict-aliasing @gol
351 -Wstrict-aliasing=n -Wstrict-overflow -Wstrict-overflow=@var{n} @gol
352 -Wstringop-overflow=@var{n} -Wstringop-truncation -Wsubobject-linkage @gol
353 -Wsuggest-attribute=@r{[}pure@r{|}const@r{|}noreturn@r{|}format@r{|}malloc@r{]} @gol
354 -Wsuggest-final-types @gol -Wsuggest-final-methods -Wsuggest-override @gol
355 -Wswitch -Wswitch-bool -Wswitch-default -Wswitch-enum @gol
356 -Wswitch-unreachable -Wsync-nand @gol
357 -Wsystem-headers -Wtautological-compare -Wtrampolines -Wtrigraphs @gol
358 -Wtype-limits -Wundef @gol
359 -Wuninitialized -Wunknown-pragmas @gol
360 -Wunsuffixed-float-constants -Wunused -Wunused-function @gol
361 -Wunused-label -Wunused-local-typedefs -Wunused-macros @gol
362 -Wunused-parameter -Wno-unused-result @gol
363 -Wunused-value -Wunused-variable @gol
364 -Wunused-const-variable -Wunused-const-variable=@var{n} @gol
365 -Wunused-but-set-parameter -Wunused-but-set-variable @gol
366 -Wuseless-cast -Wvariadic-macros -Wvector-operation-performance @gol
367 -Wvla -Wvla-larger-than=@var{byte-size} -Wvolatile-register-var @gol
369 -Wzero-as-null-pointer-constant}
371 @item C and Objective-C-only Warning Options
372 @gccoptlist{-Wbad-function-cast -Wmissing-declarations @gol
373 -Wmissing-parameter-type -Wmissing-prototypes -Wnested-externs @gol
374 -Wold-style-declaration -Wold-style-definition @gol
375 -Wstrict-prototypes -Wtraditional -Wtraditional-conversion @gol
376 -Wdeclaration-after-statement -Wpointer-sign}
378 @item Debugging Options
379 @xref{Debugging Options,,Options for Debugging Your Program}.
380 @gccoptlist{-g -g@var{level} -gdwarf -gdwarf-@var{version} @gol
381 -ggdb -grecord-gcc-switches -gno-record-gcc-switches @gol
382 -gstabs -gstabs+ -gstrict-dwarf -gno-strict-dwarf @gol
383 -gas-loc-support -gno-as-loc-support @gol
384 -gas-locview-support -gno-as-locview-support @gol
385 -gcolumn-info -gno-column-info @gol
386 -gstatement-frontiers -gno-statement-frontiers @gol
387 -gvariable-location-views -gno-variable-location-views @gol
388 -ginternal-reset-location-views -gno-internal-reset-location-views @gol
389 -ginline-points -gno-inline-points @gol
390 -gvms -gxcoff -gxcoff+ -gz@r{[}=@var{type}@r{]} @gol
391 -gsplit-dwarf -gdescribe-dies -gno-describe-dies @gol
392 -fdebug-prefix-map=@var{old}=@var{new} -fdebug-types-section @gol
393 -fno-eliminate-unused-debug-types @gol
394 -femit-struct-debug-baseonly -femit-struct-debug-reduced @gol
395 -femit-struct-debug-detailed@r{[}=@var{spec-list}@r{]} @gol
396 -fno-eliminate-unused-debug-symbols -femit-class-debug-always @gol
397 -fno-merge-debug-strings -fno-dwarf2-cfi-asm @gol
398 -fvar-tracking -fvar-tracking-assignments}
400 @item Optimization Options
401 @xref{Optimize Options,,Options that Control Optimization}.
402 @gccoptlist{-faggressive-loop-optimizations @gol
403 -falign-functions[=@var{n}[:@var{m}:[@var{n2}[:@var{m2}]]]] @gol
404 -falign-jumps[=@var{n}[:@var{m}:[@var{n2}[:@var{m2}]]]] @gol
405 -falign-labels[=@var{n}[:@var{m}:[@var{n2}[:@var{m2}]]]] @gol
406 -falign-loops[=@var{n}[:@var{m}:[@var{n2}[:@var{m2}]]]] @gol
407 -fassociative-math -fauto-profile -fauto-profile[=@var{path}] @gol
408 -fauto-inc-dec -fbranch-probabilities @gol
410 -fcombine-stack-adjustments -fconserve-stack @gol
411 -fcompare-elim -fcprop-registers -fcrossjumping @gol
412 -fcse-follow-jumps -fcse-skip-blocks -fcx-fortran-rules @gol
413 -fcx-limited-range @gol
414 -fdata-sections -fdce -fdelayed-branch @gol
415 -fdelete-null-pointer-checks -fdevirtualize -fdevirtualize-speculatively @gol
416 -fdevirtualize-at-ltrans -fdse @gol
417 -fearly-inlining -fipa-sra -fexpensive-optimizations -ffat-lto-objects @gol
418 -ffast-math -ffinite-math-only -ffloat-store -fexcess-precision=@var{style} @gol
420 -fforward-propagate -ffp-contract=@var{style} -ffunction-sections @gol
421 -fgcse -fgcse-after-reload -fgcse-las -fgcse-lm -fgraphite-identity @gol
422 -fgcse-sm -fhoist-adjacent-loads -fif-conversion @gol
423 -fif-conversion2 -findirect-inlining @gol
424 -finline-functions -finline-functions-called-once -finline-limit=@var{n} @gol
425 -finline-small-functions -fipa-cp -fipa-cp-clone @gol
426 -fipa-bit-cp -fipa-vrp -fipa-pta -fipa-profile -fipa-pure-const @gol
427 -fipa-reference -fipa-reference-addressable @gol
428 -fipa-stack-alignment -fipa-icf -fira-algorithm=@var{algorithm} @gol
429 -flive-patching=@var{level} @gol
430 -fira-region=@var{region} -fira-hoist-pressure @gol
431 -fira-loop-pressure -fno-ira-share-save-slots @gol
432 -fno-ira-share-spill-slots @gol
433 -fisolate-erroneous-paths-dereference -fisolate-erroneous-paths-attribute @gol
434 -fivopts -fkeep-inline-functions -fkeep-static-functions @gol
435 -fkeep-static-consts -flimit-function-alignment -flive-range-shrinkage @gol
436 -floop-block -floop-interchange -floop-strip-mine @gol
437 -floop-unroll-and-jam -floop-nest-optimize @gol
438 -floop-parallelize-all -flra-remat -flto -flto-compression-level @gol
439 -flto-partition=@var{alg} -fmerge-all-constants @gol
440 -fmerge-constants -fmodulo-sched -fmodulo-sched-allow-regmoves @gol
441 -fmove-loop-invariants -fno-branch-count-reg @gol
442 -fno-defer-pop -fno-fp-int-builtin-inexact -fno-function-cse @gol
443 -fno-guess-branch-probability -fno-inline -fno-math-errno -fno-peephole @gol
444 -fno-peephole2 -fno-printf-return-value -fno-sched-interblock @gol
445 -fno-sched-spec -fno-signed-zeros @gol
446 -fno-toplevel-reorder -fno-trapping-math -fno-zero-initialized-in-bss @gol
447 -fomit-frame-pointer -foptimize-sibling-calls @gol
448 -fpartial-inlining -fpeel-loops -fpredictive-commoning @gol
449 -fprefetch-loop-arrays @gol
450 -fprofile-correction @gol
451 -fprofile-use -fprofile-use=@var{path} -fprofile-values @gol
452 -fprofile-reorder-functions @gol
453 -freciprocal-math -free -frename-registers -freorder-blocks @gol
454 -freorder-blocks-algorithm=@var{algorithm} @gol
455 -freorder-blocks-and-partition -freorder-functions @gol
456 -frerun-cse-after-loop -freschedule-modulo-scheduled-loops @gol
457 -frounding-math -fsave-optimization-record @gol
458 -fsched2-use-superblocks -fsched-pressure @gol
459 -fsched-spec-load -fsched-spec-load-dangerous @gol
460 -fsched-stalled-insns-dep[=@var{n}] -fsched-stalled-insns[=@var{n}] @gol
461 -fsched-group-heuristic -fsched-critical-path-heuristic @gol
462 -fsched-spec-insn-heuristic -fsched-rank-heuristic @gol
463 -fsched-last-insn-heuristic -fsched-dep-count-heuristic @gol
464 -fschedule-fusion @gol
465 -fschedule-insns -fschedule-insns2 -fsection-anchors @gol
466 -fselective-scheduling -fselective-scheduling2 @gol
467 -fsel-sched-pipelining -fsel-sched-pipelining-outer-loops @gol
468 -fsemantic-interposition -fshrink-wrap -fshrink-wrap-separate @gol
469 -fsignaling-nans @gol
470 -fsingle-precision-constant -fsplit-ivs-in-unroller -fsplit-loops@gol
472 -fsplit-wide-types -fsplit-wide-types-early -fssa-backprop -fssa-phiopt @gol
473 -fstdarg-opt -fstore-merging -fstrict-aliasing @gol
474 -fthread-jumps -ftracer -ftree-bit-ccp @gol
475 -ftree-builtin-call-dce -ftree-ccp -ftree-ch @gol
476 -ftree-coalesce-vars -ftree-copy-prop -ftree-dce -ftree-dominator-opts @gol
477 -ftree-dse -ftree-forwprop -ftree-fre -fcode-hoisting @gol
478 -ftree-loop-if-convert -ftree-loop-im @gol
479 -ftree-phiprop -ftree-loop-distribution -ftree-loop-distribute-patterns @gol
480 -ftree-loop-ivcanon -ftree-loop-linear -ftree-loop-optimize @gol
481 -ftree-loop-vectorize @gol
482 -ftree-parallelize-loops=@var{n} -ftree-pre -ftree-partial-pre -ftree-pta @gol
483 -ftree-reassoc -ftree-scev-cprop -ftree-sink -ftree-slsr -ftree-sra @gol
484 -ftree-switch-conversion -ftree-tail-merge @gol
485 -ftree-ter -ftree-vectorize -ftree-vrp -funconstrained-commons @gol
486 -funit-at-a-time -funroll-all-loops -funroll-loops @gol
487 -funsafe-math-optimizations -funswitch-loops @gol
488 -fipa-ra -fvariable-expansion-in-unroller -fvect-cost-model -fvpt @gol
489 -fweb -fwhole-program -fwpa -fuse-linker-plugin @gol
490 --param @var{name}=@var{value}
491 -O -O0 -O1 -O2 -O3 -Os -Ofast -Og}
493 @item Program Instrumentation Options
494 @xref{Instrumentation Options,,Program Instrumentation Options}.
495 @gccoptlist{-p -pg -fprofile-arcs --coverage -ftest-coverage @gol
496 -fprofile-abs-path @gol
497 -fprofile-dir=@var{path} -fprofile-generate -fprofile-generate=@var{path} @gol
498 -fprofile-note=@var{path} -fprofile-update=@var{method} @gol
499 -fprofile-filter-files=@var{regex} -fprofile-exclude-files=@var{regex} @gol
500 -fsanitize=@var{style} -fsanitize-recover -fsanitize-recover=@var{style} @gol
501 -fasan-shadow-offset=@var{number} -fsanitize-sections=@var{s1},@var{s2},... @gol
502 -fsanitize-undefined-trap-on-error -fbounds-check @gol
503 -fcf-protection=@r{[}full@r{|}branch@r{|}return@r{|}none@r{]} @gol
504 -fstack-protector -fstack-protector-all -fstack-protector-strong @gol
505 -fstack-protector-explicit -fstack-check @gol
506 -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol
507 -fno-stack-limit -fsplit-stack @gol
508 -fvtable-verify=@r{[}std@r{|}preinit@r{|}none@r{]} @gol
509 -fvtv-counts -fvtv-debug @gol
510 -finstrument-functions @gol
511 -finstrument-functions-exclude-function-list=@var{sym},@var{sym},@dots{} @gol
512 -finstrument-functions-exclude-file-list=@var{file},@var{file},@dots{}}
514 @item Preprocessor Options
515 @xref{Preprocessor Options,,Options Controlling the Preprocessor}.
516 @gccoptlist{-A@var{question}=@var{answer} @gol
517 -A-@var{question}@r{[}=@var{answer}@r{]} @gol
518 -C -CC -D@var{macro}@r{[}=@var{defn}@r{]} @gol
519 -dD -dI -dM -dN -dU @gol
520 -fdebug-cpp -fdirectives-only -fdollars-in-identifiers @gol
521 -fexec-charset=@var{charset} -fextended-identifiers @gol
522 -finput-charset=@var{charset} -fmacro-prefix-map=@var{old}=@var{new} @gol
523 -fmax-include-depth=@var{depth} @gol
524 -fno-canonical-system-headers -fpch-deps -fpch-preprocess @gol
525 -fpreprocessed -ftabstop=@var{width} -ftrack-macro-expansion @gol
526 -fwide-exec-charset=@var{charset} -fworking-directory @gol
527 -H -imacros @var{file} -include @var{file} @gol
528 -M -MD -MF -MG -MM -MMD -MP -MQ -MT @gol
529 -no-integrated-cpp -P -pthread -remap @gol
530 -traditional -traditional-cpp -trigraphs @gol
531 -U@var{macro} -undef @gol
532 -Wp,@var{option} -Xpreprocessor @var{option}}
534 @item Assembler Options
535 @xref{Assembler Options,,Passing Options to the Assembler}.
536 @gccoptlist{-Wa,@var{option} -Xassembler @var{option}}
539 @xref{Link Options,,Options for Linking}.
540 @gccoptlist{@var{object-file-name} -fuse-ld=@var{linker} -l@var{library} @gol
541 -nostartfiles -nodefaultlibs -nolibc -nostdlib @gol
542 -e @var{entry} --entry=@var{entry} @gol
543 -pie -pthread -r -rdynamic @gol
544 -s -static -static-pie -static-libgcc -static-libstdc++ @gol
545 -static-libasan -static-libtsan -static-liblsan -static-libubsan @gol
546 -shared -shared-libgcc -symbolic @gol
547 -T @var{script} -Wl,@var{option} -Xlinker @var{option} @gol
548 -u @var{symbol} -z @var{keyword}}
550 @item Directory Options
551 @xref{Directory Options,,Options for Directory Search}.
552 @gccoptlist{-B@var{prefix} -I@var{dir} -I- @gol
553 -idirafter @var{dir} @gol
554 -imacros @var{file} -imultilib @var{dir} @gol
555 -iplugindir=@var{dir} -iprefix @var{file} @gol
556 -iquote @var{dir} -isysroot @var{dir} -isystem @var{dir} @gol
557 -iwithprefix @var{dir} -iwithprefixbefore @var{dir} @gol
558 -L@var{dir} -no-canonical-prefixes --no-sysroot-suffix @gol
559 -nostdinc -nostdinc++ --sysroot=@var{dir}}
561 @item Code Generation Options
562 @xref{Code Gen Options,,Options for Code Generation Conventions}.
563 @gccoptlist{-fcall-saved-@var{reg} -fcall-used-@var{reg} @gol
564 -ffixed-@var{reg} -fexceptions @gol
565 -fnon-call-exceptions -fdelete-dead-exceptions -funwind-tables @gol
566 -fasynchronous-unwind-tables @gol
568 -finhibit-size-directive -fno-common -fno-ident @gol
569 -fpcc-struct-return -fpic -fPIC -fpie -fPIE -fno-plt @gol
570 -fno-jump-tables @gol
571 -frecord-gcc-switches @gol
572 -freg-struct-return -fshort-enums -fshort-wchar @gol
573 -fverbose-asm -fpack-struct[=@var{n}] @gol
574 -fleading-underscore -ftls-model=@var{model} @gol
575 -fstack-reuse=@var{reuse_level} @gol
576 -ftrampolines -ftrapv -fwrapv @gol
577 -fvisibility=@r{[}default@r{|}internal@r{|}hidden@r{|}protected@r{]} @gol
578 -fstrict-volatile-bitfields -fsync-libcalls}
580 @item Developer Options
581 @xref{Developer Options,,GCC Developer Options}.
582 @gccoptlist{-d@var{letters} -dumpspecs -dumpmachine -dumpversion @gol
583 -dumpfullversion -fchecking -fchecking=@var{n} -fdbg-cnt-list @gol
584 -fdbg-cnt=@var{counter-value-list} @gol
585 -fdisable-ipa-@var{pass_name} @gol
586 -fdisable-rtl-@var{pass_name} @gol
587 -fdisable-rtl-@var{pass-name}=@var{range-list} @gol
588 -fdisable-tree-@var{pass_name} @gol
589 -fdisable-tree-@var{pass-name}=@var{range-list} @gol
590 -fdump-debug -fdump-earlydebug @gol
591 -fdump-noaddr -fdump-unnumbered -fdump-unnumbered-links @gol
592 -fdump-final-insns@r{[}=@var{file}@r{]} @gol
593 -fdump-ipa-all -fdump-ipa-cgraph -fdump-ipa-inline @gol
595 -fdump-lang-@var{switch} @gol
596 -fdump-lang-@var{switch}-@var{options} @gol
597 -fdump-lang-@var{switch}-@var{options}=@var{filename} @gol
599 -fdump-rtl-@var{pass} -fdump-rtl-@var{pass}=@var{filename} @gol
600 -fdump-statistics @gol
602 -fdump-tree-@var{switch} @gol
603 -fdump-tree-@var{switch}-@var{options} @gol
604 -fdump-tree-@var{switch}-@var{options}=@var{filename} @gol
605 -fcompare-debug@r{[}=@var{opts}@r{]} -fcompare-debug-second @gol
606 -fenable-@var{kind}-@var{pass} @gol
607 -fenable-@var{kind}-@var{pass}=@var{range-list} @gol
608 -fira-verbose=@var{n} @gol
609 -flto-report -flto-report-wpa -fmem-report-wpa @gol
610 -fmem-report -fpre-ipa-mem-report -fpost-ipa-mem-report @gol
611 -fopt-info -fopt-info-@var{options}@r{[}=@var{file}@r{]} @gol
612 -fprofile-report @gol
613 -frandom-seed=@var{string} -fsched-verbose=@var{n} @gol
614 -fsel-sched-verbose -fsel-sched-dump-cfg -fsel-sched-pipelining-verbose @gol
615 -fstats -fstack-usage -ftime-report -ftime-report-details @gol
616 -fvar-tracking-assignments-toggle -gtoggle @gol
617 -print-file-name=@var{library} -print-libgcc-file-name @gol
618 -print-multi-directory -print-multi-lib -print-multi-os-directory @gol
619 -print-prog-name=@var{program} -print-search-dirs -Q @gol
620 -print-sysroot -print-sysroot-headers-suffix @gol
621 -save-temps -save-temps=cwd -save-temps=obj -time@r{[}=@var{file}@r{]}}
623 @item Machine-Dependent Options
624 @xref{Submodel Options,,Machine-Dependent Options}.
625 @c This list is ordered alphanumerically by subsection name.
626 @c Try and put the significant identifier (CPU or system) first,
627 @c so users have a clue at guessing where the ones they want will be.
629 @emph{AArch64 Options}
630 @gccoptlist{-mabi=@var{name} -mbig-endian -mlittle-endian @gol
631 -mgeneral-regs-only @gol
632 -mcmodel=tiny -mcmodel=small -mcmodel=large @gol
633 -mstrict-align -mno-strict-align @gol
634 -momit-leaf-frame-pointer @gol
635 -mtls-dialect=desc -mtls-dialect=traditional @gol
636 -mtls-size=@var{size} @gol
637 -mfix-cortex-a53-835769 -mfix-cortex-a53-843419 @gol
638 -mlow-precision-recip-sqrt -mlow-precision-sqrt -mlow-precision-div @gol
639 -mpc-relative-literal-loads @gol
640 -msign-return-address=@var{scope} @gol
641 -mbranch-protection=@var{none}|@var{standard}|@var{pac-ret}[+@var{leaf}
642 +@var{b-key}]|@var{bti} @gol
643 -march=@var{name} -mcpu=@var{name} -mtune=@var{name} @gol
644 -moverride=@var{string} -mverbose-cost-dump @gol
645 -mstack-protector-guard=@var{guard} -mstack-protector-guard-reg=@var{sysreg} @gol
646 -mstack-protector-guard-offset=@var{offset} -mtrack-speculation @gol
649 @emph{Adapteva Epiphany Options}
650 @gccoptlist{-mhalf-reg-file -mprefer-short-insn-regs @gol
651 -mbranch-cost=@var{num} -mcmove -mnops=@var{num} -msoft-cmpsf @gol
652 -msplit-lohi -mpost-inc -mpost-modify -mstack-offset=@var{num} @gol
653 -mround-nearest -mlong-calls -mshort-calls -msmall16 @gol
654 -mfp-mode=@var{mode} -mvect-double -max-vect-align=@var{num} @gol
655 -msplit-vecmove-early -m1reg-@var{reg}}
657 @emph{AMD GCN Options}
658 @gccoptlist{-march=@var{gpu} -mtune=@var{gpu} -mstack-size=@var{bytes}}
661 @gccoptlist{-mbarrel-shifter -mjli-always @gol
662 -mcpu=@var{cpu} -mA6 -mARC600 -mA7 -mARC700 @gol
663 -mdpfp -mdpfp-compact -mdpfp-fast -mno-dpfp-lrsr @gol
664 -mea -mno-mpy -mmul32x16 -mmul64 -matomic @gol
665 -mnorm -mspfp -mspfp-compact -mspfp-fast -msimd -msoft-float -mswap @gol
666 -mcrc -mdsp-packa -mdvbf -mlock -mmac-d16 -mmac-24 -mrtsc -mswape @gol
667 -mtelephony -mxy -misize -mannotate-align -marclinux -marclinux_prof @gol
668 -mlong-calls -mmedium-calls -msdata -mirq-ctrl-saved @gol
669 -mrgf-banked-regs -mlpc-width=@var{width} -G @var{num} @gol
670 -mvolatile-cache -mtp-regno=@var{regno} @gol
671 -malign-call -mauto-modify-reg -mbbit-peephole -mno-brcc @gol
672 -mcase-vector-pcrel -mcompact-casesi -mno-cond-exec -mearly-cbranchsi @gol
673 -mexpand-adddi -mindexed-loads -mlra -mlra-priority-none @gol
674 -mlra-priority-compact mlra-priority-noncompact -mmillicode @gol
675 -mmixed-code -mq-class -mRcq -mRcw -msize-level=@var{level} @gol
676 -mtune=@var{cpu} -mmultcost=@var{num} -mcode-density-frame @gol
677 -munalign-prob-threshold=@var{probability} -mmpy-option=@var{multo} @gol
678 -mdiv-rem -mcode-density -mll64 -mfpu=@var{fpu} -mrf16 -mbranch-index}
681 @gccoptlist{-mapcs-frame -mno-apcs-frame @gol
682 -mabi=@var{name} @gol
683 -mapcs-stack-check -mno-apcs-stack-check @gol
684 -mapcs-reentrant -mno-apcs-reentrant @gol
685 -mgeneral-regs-only @gol
686 -msched-prolog -mno-sched-prolog @gol
687 -mlittle-endian -mbig-endian @gol
689 -mfloat-abi=@var{name} @gol
690 -mfp16-format=@var{name}
691 -mthumb-interwork -mno-thumb-interwork @gol
692 -mcpu=@var{name} -march=@var{name} -mfpu=@var{name} @gol
693 -mtune=@var{name} -mprint-tune-info @gol
694 -mstructure-size-boundary=@var{n} @gol
695 -mabort-on-noreturn @gol
696 -mlong-calls -mno-long-calls @gol
697 -msingle-pic-base -mno-single-pic-base @gol
698 -mpic-register=@var{reg} @gol
699 -mnop-fun-dllimport @gol
700 -mpoke-function-name @gol
701 -mthumb -marm -mflip-thumb @gol
702 -mtpcs-frame -mtpcs-leaf-frame @gol
703 -mcaller-super-interworking -mcallee-super-interworking @gol
704 -mtp=@var{name} -mtls-dialect=@var{dialect} @gol
705 -mword-relocations @gol
706 -mfix-cortex-m3-ldrd @gol
707 -munaligned-access @gol
708 -mneon-for-64bits @gol
709 -mslow-flash-data @gol
710 -masm-syntax-unified @gol
712 -mverbose-cost-dump @gol
718 @gccoptlist{-mmcu=@var{mcu} -mabsdata -maccumulate-args @gol
719 -mbranch-cost=@var{cost} @gol
720 -mcall-prologues -mgas-isr-prologues -mint8 @gol
721 -mn_flash=@var{size} -mno-interrupts @gol
722 -mmain-is-OS_task -mrelax -mrmw -mstrict-X -mtiny-stack @gol
723 -mfract-convert-truncate @gol
724 -mshort-calls -nodevicelib @gol
725 -Waddr-space-convert -Wmisspelled-isr}
727 @emph{Blackfin Options}
728 @gccoptlist{-mcpu=@var{cpu}@r{[}-@var{sirevision}@r{]} @gol
729 -msim -momit-leaf-frame-pointer -mno-omit-leaf-frame-pointer @gol
730 -mspecld-anomaly -mno-specld-anomaly -mcsync-anomaly -mno-csync-anomaly @gol
731 -mlow-64k -mno-low64k -mstack-check-l1 -mid-shared-library @gol
732 -mno-id-shared-library -mshared-library-id=@var{n} @gol
733 -mleaf-id-shared-library -mno-leaf-id-shared-library @gol
734 -msep-data -mno-sep-data -mlong-calls -mno-long-calls @gol
735 -mfast-fp -minline-plt -mmulticore -mcorea -mcoreb -msdram @gol
739 @gccoptlist{-mbig-endian -mlittle-endian -march=@var{cpu} @gol
740 -msim -msdata=@var{sdata-type}}
743 @gccoptlist{-mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol
744 -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol
745 -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol
746 -mstack-align -mdata-align -mconst-align @gol
747 -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol
748 -melf -maout -melinux -mlinux -sim -sim2 @gol
749 -mmul-bug-workaround -mno-mul-bug-workaround}
752 @gccoptlist{-mmac @gol
753 -mcr16cplus -mcr16c @gol
754 -msim -mint32 -mbit-ops
755 -mdata-model=@var{model}}
758 @gccoptlist{-march=@var{arch} -mcpu=@var{cpu} @gol
759 -mbig-endian -EB -mlittle-endian -EL @gol
760 -mhard-float -msoft-float -mfpu=@var{fpu} -mdouble-float -mfdivdu @gol
761 -melrw -mistack -mmp -mcp -mcache -msecurity -mtrust @gol
762 -mdsp -medsp -mvdsp @gol
763 -mdiv -msmart -mhigh-registers -manchor @gol
764 -mpushpop -mmultiple-stld -mconstpool -mstack-size -mccrt @gol
765 -mbranch-cost=@var{n} -mcse-cc -msched-prolog}
767 @emph{Darwin Options}
768 @gccoptlist{-all_load -allowable_client -arch -arch_errors_fatal @gol
769 -arch_only -bind_at_load -bundle -bundle_loader @gol
770 -client_name -compatibility_version -current_version @gol
772 -dependency-file -dylib_file -dylinker_install_name @gol
773 -dynamic -dynamiclib -exported_symbols_list @gol
774 -filelist -flat_namespace -force_cpusubtype_ALL @gol
775 -force_flat_namespace -headerpad_max_install_names @gol
777 -image_base -init -install_name -keep_private_externs @gol
778 -multi_module -multiply_defined -multiply_defined_unused @gol
779 -noall_load -no_dead_strip_inits_and_terms @gol
780 -nofixprebinding -nomultidefs -noprebind -noseglinkedit @gol
781 -pagezero_size -prebind -prebind_all_twolevel_modules @gol
782 -private_bundle -read_only_relocs -sectalign @gol
783 -sectobjectsymbols -whyload -seg1addr @gol
784 -sectcreate -sectobjectsymbols -sectorder @gol
785 -segaddr -segs_read_only_addr -segs_read_write_addr @gol
786 -seg_addr_table -seg_addr_table_filename -seglinkedit @gol
787 -segprot -segs_read_only_addr -segs_read_write_addr @gol
788 -single_module -static -sub_library -sub_umbrella @gol
789 -twolevel_namespace -umbrella -undefined @gol
790 -unexported_symbols_list -weak_reference_mismatches @gol
791 -whatsloaded -F -gused -gfull -mmacosx-version-min=@var{version} @gol
792 -mkernel -mone-byte-bool}
794 @emph{DEC Alpha Options}
795 @gccoptlist{-mno-fp-regs -msoft-float @gol
796 -mieee -mieee-with-inexact -mieee-conformant @gol
797 -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol
798 -mtrap-precision=@var{mode} -mbuild-constants @gol
799 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} @gol
800 -mbwx -mmax -mfix -mcix @gol
801 -mfloat-vax -mfloat-ieee @gol
802 -mexplicit-relocs -msmall-data -mlarge-data @gol
803 -msmall-text -mlarge-text @gol
804 -mmemory-latency=@var{time}}
807 @gccoptlist{-mbig-endian -mlittle-endian -mkernel=@var{version}
808 -mframe-limit=@var{bytes}}
811 @gccoptlist{-msmall-model -mno-lsim}
814 @gccoptlist{-msim -mlra -mnodiv -mft32b -mcompress -mnopm}
817 @gccoptlist{-mgpr-32 -mgpr-64 -mfpr-32 -mfpr-64 @gol
818 -mhard-float -msoft-float @gol
819 -malloc-cc -mfixed-cc -mdword -mno-dword @gol
820 -mdouble -mno-double @gol
821 -mmedia -mno-media -mmuladd -mno-muladd @gol
822 -mfdpic -minline-plt -mgprel-ro -multilib-library-pic @gol
823 -mlinked-fp -mlong-calls -malign-labels @gol
824 -mlibrary-pic -macc-4 -macc-8 @gol
825 -mpack -mno-pack -mno-eflags -mcond-move -mno-cond-move @gol
826 -moptimize-membar -mno-optimize-membar @gol
827 -mscc -mno-scc -mcond-exec -mno-cond-exec @gol
828 -mvliw-branch -mno-vliw-branch @gol
829 -mmulti-cond-exec -mno-multi-cond-exec -mnested-cond-exec @gol
830 -mno-nested-cond-exec -mtomcat-stats @gol
834 @emph{GNU/Linux Options}
835 @gccoptlist{-mglibc -muclibc -mmusl -mbionic -mandroid @gol
836 -tno-android-cc -tno-android-ld}
838 @emph{H8/300 Options}
839 @gccoptlist{-mrelax -mh -ms -mn -mexr -mno-exr -mint32 -malign-300}
842 @gccoptlist{-march=@var{architecture-type} @gol
843 -mcaller-copies -mdisable-fpregs -mdisable-indexing @gol
844 -mfast-indirect-calls -mgas -mgnu-ld -mhp-ld @gol
845 -mfixed-range=@var{register-range} @gol
846 -mjump-in-delay -mlinker-opt -mlong-calls @gol
847 -mlong-load-store -mno-disable-fpregs @gol
848 -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol
849 -mno-jump-in-delay -mno-long-load-store @gol
850 -mno-portable-runtime -mno-soft-float @gol
851 -mno-space-regs -msoft-float -mpa-risc-1-0 @gol
852 -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol
853 -mschedule=@var{cpu-type} -mspace-regs -msio -mwsio @gol
854 -munix=@var{unix-std} -nolibdld -static -threads}
857 @gccoptlist{-mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol
858 -mvolatile-asm-stop -mregister-names -msdata -mno-sdata @gol
859 -mconstant-gp -mauto-pic -mfused-madd @gol
860 -minline-float-divide-min-latency @gol
861 -minline-float-divide-max-throughput @gol
862 -mno-inline-float-divide @gol
863 -minline-int-divide-min-latency @gol
864 -minline-int-divide-max-throughput @gol
865 -mno-inline-int-divide @gol
866 -minline-sqrt-min-latency -minline-sqrt-max-throughput @gol
867 -mno-inline-sqrt @gol
868 -mdwarf2-asm -mearly-stop-bits @gol
869 -mfixed-range=@var{register-range} -mtls-size=@var{tls-size} @gol
870 -mtune=@var{cpu-type} -milp32 -mlp64 @gol
871 -msched-br-data-spec -msched-ar-data-spec -msched-control-spec @gol
872 -msched-br-in-data-spec -msched-ar-in-data-spec -msched-in-control-spec @gol
873 -msched-spec-ldc -msched-spec-control-ldc @gol
874 -msched-prefer-non-data-spec-insns -msched-prefer-non-control-spec-insns @gol
875 -msched-stop-bits-after-every-cycle -msched-count-spec-in-critical-path @gol
876 -msel-sched-dont-check-control-spec -msched-fp-mem-deps-zero-cost @gol
877 -msched-max-memory-insns-hard-limit -msched-max-memory-insns=@var{max-insns}}
880 @gccoptlist{-mbarrel-shift-enabled -mdivide-enabled -mmultiply-enabled @gol
881 -msign-extend-enabled -muser-enabled}
883 @emph{M32R/D Options}
884 @gccoptlist{-m32r2 -m32rx -m32r @gol
886 -malign-loops -mno-align-loops @gol
887 -missue-rate=@var{number} @gol
888 -mbranch-cost=@var{number} @gol
889 -mmodel=@var{code-size-model-type} @gol
890 -msdata=@var{sdata-type} @gol
891 -mno-flush-func -mflush-func=@var{name} @gol
892 -mno-flush-trap -mflush-trap=@var{number} @gol
896 @gccoptlist{-mcpu=@var{cpu} -msim -memregs=@var{number}}
898 @emph{M680x0 Options}
899 @gccoptlist{-march=@var{arch} -mcpu=@var{cpu} -mtune=@var{tune} @gol
900 -m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol
901 -m68060 -mcpu32 -m5200 -m5206e -m528x -m5307 -m5407 @gol
902 -mcfv4e -mbitfield -mno-bitfield -mc68000 -mc68020 @gol
903 -mnobitfield -mrtd -mno-rtd -mdiv -mno-div -mshort @gol
904 -mno-short -mhard-float -m68881 -msoft-float -mpcrel @gol
905 -malign-int -mstrict-align -msep-data -mno-sep-data @gol
906 -mshared-library-id=n -mid-shared-library -mno-id-shared-library @gol
907 -mxgot -mno-xgot -mlong-jump-table-offsets}
910 @gccoptlist{-mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol
911 -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol
912 -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol
913 -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol
914 -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment}
917 @gccoptlist{-mabsdiff -mall-opts -maverage -mbased=@var{n} -mbitops @gol
918 -mc=@var{n} -mclip -mconfig=@var{name} -mcop -mcop32 -mcop64 -mivc2 @gol
919 -mdc -mdiv -meb -mel -mio-volatile -ml -mleadz -mm -mminmax @gol
920 -mmult -mno-opts -mrepeat -ms -msatur -msdram -msim -msimnovec -mtf @gol
923 @emph{MicroBlaze Options}
924 @gccoptlist{-msoft-float -mhard-float -msmall-divides -mcpu=@var{cpu} @gol
925 -mmemcpy -mxl-soft-mul -mxl-soft-div -mxl-barrel-shift @gol
926 -mxl-pattern-compare -mxl-stack-check -mxl-gp-opt -mno-clearbss @gol
927 -mxl-multiply-high -mxl-float-convert -mxl-float-sqrt @gol
928 -mbig-endian -mlittle-endian -mxl-reorder -mxl-mode-@var{app-model} @gol
929 -mpic-data-is-text-relative}
932 @gccoptlist{-EL -EB -march=@var{arch} -mtune=@var{arch} @gol
933 -mips1 -mips2 -mips3 -mips4 -mips32 -mips32r2 -mips32r3 -mips32r5 @gol
934 -mips32r6 -mips64 -mips64r2 -mips64r3 -mips64r5 -mips64r6 @gol
935 -mips16 -mno-mips16 -mflip-mips16 @gol
936 -minterlink-compressed -mno-interlink-compressed @gol
937 -minterlink-mips16 -mno-interlink-mips16 @gol
938 -mabi=@var{abi} -mabicalls -mno-abicalls @gol
939 -mshared -mno-shared -mplt -mno-plt -mxgot -mno-xgot @gol
940 -mgp32 -mgp64 -mfp32 -mfpxx -mfp64 -mhard-float -msoft-float @gol
941 -mno-float -msingle-float -mdouble-float @gol
942 -modd-spreg -mno-odd-spreg @gol
943 -mabs=@var{mode} -mnan=@var{encoding} @gol
944 -mdsp -mno-dsp -mdspr2 -mno-dspr2 @gol
947 -mvirt -mno-virt @gol
950 -mginv -mno-ginv @gol
951 -mmicromips -mno-micromips @gol
953 -mloongson-mmi -mno-loongson-mmi @gol
954 -mloongson-ext -mno-loongson-ext @gol
955 -mloongson-ext2 -mno-loongson-ext2 @gol
956 -mfpu=@var{fpu-type} @gol
957 -msmartmips -mno-smartmips @gol
958 -mpaired-single -mno-paired-single -mdmx -mno-mdmx @gol
959 -mips3d -mno-mips3d -mmt -mno-mt -mllsc -mno-llsc @gol
960 -mlong64 -mlong32 -msym32 -mno-sym32 @gol
961 -G@var{num} -mlocal-sdata -mno-local-sdata @gol
962 -mextern-sdata -mno-extern-sdata -mgpopt -mno-gopt @gol
963 -membedded-data -mno-embedded-data @gol
964 -muninit-const-in-rodata -mno-uninit-const-in-rodata @gol
965 -mcode-readable=@var{setting} @gol
966 -msplit-addresses -mno-split-addresses @gol
967 -mexplicit-relocs -mno-explicit-relocs @gol
968 -mcheck-zero-division -mno-check-zero-division @gol
969 -mdivide-traps -mdivide-breaks @gol
970 -mload-store-pairs -mno-load-store-pairs @gol
971 -mmemcpy -mno-memcpy -mlong-calls -mno-long-calls @gol
972 -mmad -mno-mad -mimadd -mno-imadd -mfused-madd -mno-fused-madd -nocpp @gol
973 -mfix-24k -mno-fix-24k @gol
974 -mfix-r4000 -mno-fix-r4000 -mfix-r4400 -mno-fix-r4400 @gol
975 -mfix-r5900 -mno-fix-r5900 @gol
976 -mfix-r10000 -mno-fix-r10000 -mfix-rm7000 -mno-fix-rm7000 @gol
977 -mfix-vr4120 -mno-fix-vr4120 @gol
978 -mfix-vr4130 -mno-fix-vr4130 -mfix-sb1 -mno-fix-sb1 @gol
979 -mflush-func=@var{func} -mno-flush-func @gol
980 -mbranch-cost=@var{num} -mbranch-likely -mno-branch-likely @gol
981 -mcompact-branches=@var{policy} @gol
982 -mfp-exceptions -mno-fp-exceptions @gol
983 -mvr4130-align -mno-vr4130-align -msynci -mno-synci @gol
984 -mlxc1-sxc1 -mno-lxc1-sxc1 -mmadd4 -mno-madd4 @gol
985 -mrelax-pic-calls -mno-relax-pic-calls -mmcount-ra-address @gol
986 -mframe-header-opt -mno-frame-header-opt}
989 @gccoptlist{-mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol
990 -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol
991 -melf -mbranch-predict -mno-branch-predict -mbase-addresses @gol
992 -mno-base-addresses -msingle-exit -mno-single-exit}
994 @emph{MN10300 Options}
995 @gccoptlist{-mmult-bug -mno-mult-bug @gol
996 -mno-am33 -mam33 -mam33-2 -mam34 @gol
997 -mtune=@var{cpu-type} @gol
998 -mreturn-pointer-on-d0 @gol
999 -mno-crt0 -mrelax -mliw -msetlb}
1001 @emph{Moxie Options}
1002 @gccoptlist{-meb -mel -mmul.x -mno-crt0}
1004 @emph{MSP430 Options}
1005 @gccoptlist{-msim -masm-hex -mmcu= -mcpu= -mlarge -msmall -mrelax @gol
1007 -mcode-region= -mdata-region= @gol
1008 -msilicon-errata= -msilicon-errata-warn= @gol
1011 @emph{NDS32 Options}
1012 @gccoptlist{-mbig-endian -mlittle-endian @gol
1013 -mreduced-regs -mfull-regs @gol
1014 -mcmov -mno-cmov @gol
1015 -mext-perf -mno-ext-perf @gol
1016 -mext-perf2 -mno-ext-perf2 @gol
1017 -mext-string -mno-ext-string @gol
1018 -mv3push -mno-v3push @gol
1019 -m16bit -mno-16bit @gol
1020 -misr-vector-size=@var{num} @gol
1021 -mcache-block-size=@var{num} @gol
1022 -march=@var{arch} @gol
1023 -mcmodel=@var{code-model} @gol
1024 -mctor-dtor -mrelax}
1026 @emph{Nios II Options}
1027 @gccoptlist{-G @var{num} -mgpopt=@var{option} -mgpopt -mno-gpopt @gol
1028 -mgprel-sec=@var{regexp} -mr0rel-sec=@var{regexp} @gol
1030 -mno-bypass-cache -mbypass-cache @gol
1031 -mno-cache-volatile -mcache-volatile @gol
1032 -mno-fast-sw-div -mfast-sw-div @gol
1033 -mhw-mul -mno-hw-mul -mhw-mulx -mno-hw-mulx -mno-hw-div -mhw-div @gol
1034 -mcustom-@var{insn}=@var{N} -mno-custom-@var{insn} @gol
1035 -mcustom-fpu-cfg=@var{name} @gol
1036 -mhal -msmallc -msys-crt0=@var{name} -msys-lib=@var{name} @gol
1037 -march=@var{arch} -mbmx -mno-bmx -mcdx -mno-cdx}
1039 @emph{Nvidia PTX Options}
1040 @gccoptlist{-m32 -m64 -mmainkernel -moptimize}
1042 @emph{OpenRISC Options}
1043 @gccoptlist{-mboard=@var{name} -mnewlib -mhard-mul -mhard-div @gol
1044 -msoft-mul -msoft-div @gol
1045 -msoft-float -mhard-float -mdouble-float -munordered-float @gol
1046 -mcmov -mror -mrori -msext -msfimm -mshftimm}
1048 @emph{PDP-11 Options}
1049 @gccoptlist{-mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol
1050 -mint32 -mno-int16 -mint16 -mno-int32 @gol
1051 -msplit -munix-asm -mdec-asm -mgnu-asm -mlra}
1053 @emph{picoChip Options}
1054 @gccoptlist{-mae=@var{ae_type} -mvliw-lookahead=@var{N} @gol
1055 -msymbol-as-address -mno-inefficient-warnings}
1057 @emph{PowerPC Options}
1058 See RS/6000 and PowerPC Options.
1061 @gccoptlist{-mmcu=@var{mcu} -minrt -mno-relax -mloop @gol
1062 -mabi=@var{variant} @gol}
1064 @emph{RISC-V Options}
1065 @gccoptlist{-mbranch-cost=@var{N-instruction} @gol
1067 -mabi=@var{ABI-string} @gol
1068 -mfdiv -mno-fdiv @gol
1070 -march=@var{ISA-string} @gol
1071 -mtune=@var{processor-string} @gol
1072 -mpreferred-stack-boundary=@var{num} @gol
1073 -msmall-data-limit=@var{N-bytes} @gol
1074 -msave-restore -mno-save-restore @gol
1075 -mstrict-align -mno-strict-align @gol
1076 -mcmodel=medlow -mcmodel=medany @gol
1077 -mexplicit-relocs -mno-explicit-relocs @gol
1078 -mrelax -mno-relax @gol
1079 -mriscv-attribute -mmo-riscv-attribute @gol
1080 -malign-data=@var{type}}
1083 @gccoptlist{-msim -mmul=none -mmul=g13 -mmul=g14 -mallregs @gol
1084 -mcpu=g10 -mcpu=g13 -mcpu=g14 -mg10 -mg13 -mg14 @gol
1085 -m64bit-doubles -m32bit-doubles -msave-mduc-in-interrupts}
1087 @emph{RS/6000 and PowerPC Options}
1088 @gccoptlist{-mcpu=@var{cpu-type} @gol
1089 -mtune=@var{cpu-type} @gol
1090 -mcmodel=@var{code-model} @gol
1092 -maltivec -mno-altivec @gol
1093 -mpowerpc-gpopt -mno-powerpc-gpopt @gol
1094 -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol
1095 -mmfcrf -mno-mfcrf -mpopcntb -mno-popcntb -mpopcntd -mno-popcntd @gol
1096 -mfprnd -mno-fprnd @gol
1097 -mcmpb -mno-cmpb -mhard-dfp -mno-hard-dfp @gol
1098 -mfull-toc -mminimal-toc -mno-fp-in-toc -mno-sum-in-toc @gol
1099 -m64 -m32 -mxl-compat -mno-xl-compat -mpe @gol
1100 -malign-power -malign-natural @gol
1101 -msoft-float -mhard-float -mmultiple -mno-multiple @gol
1102 -mupdate -mno-update @gol
1103 -mavoid-indexed-addresses -mno-avoid-indexed-addresses @gol
1104 -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol
1105 -mstrict-align -mno-strict-align -mrelocatable @gol
1106 -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol
1107 -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol
1108 -mdynamic-no-pic -mswdiv -msingle-pic-base @gol
1109 -mprioritize-restricted-insns=@var{priority} @gol
1110 -msched-costly-dep=@var{dependence_type} @gol
1111 -minsert-sched-nops=@var{scheme} @gol
1112 -mcall-aixdesc -mcall-eabi -mcall-freebsd @gol
1113 -mcall-linux -mcall-netbsd -mcall-openbsd @gol
1114 -mcall-sysv -mcall-sysv-eabi -mcall-sysv-noeabi @gol
1115 -mtraceback=@var{traceback_type} @gol
1116 -maix-struct-return -msvr4-struct-return @gol
1117 -mabi=@var{abi-type} -msecure-plt -mbss-plt @gol
1118 -mlongcall -mno-longcall -mpltseq -mno-pltseq @gol
1119 -mblock-move-inline-limit=@var{num} @gol
1120 -mblock-compare-inline-limit=@var{num} @gol
1121 -mblock-compare-inline-loop-limit=@var{num} @gol
1122 -mstring-compare-inline-limit=@var{num} @gol
1123 -misel -mno-isel @gol
1124 -mvrsave -mno-vrsave @gol
1125 -mmulhw -mno-mulhw @gol
1126 -mdlmzb -mno-dlmzb @gol
1127 -mprototype -mno-prototype @gol
1128 -msim -mmvme -mads -myellowknife -memb -msdata @gol
1129 -msdata=@var{opt} -mreadonly-in-sdata -mvxworks -G @var{num} @gol
1130 -mrecip -mrecip=@var{opt} -mno-recip -mrecip-precision @gol
1131 -mno-recip-precision @gol
1132 -mveclibabi=@var{type} -mfriz -mno-friz @gol
1133 -mpointers-to-nested-functions -mno-pointers-to-nested-functions @gol
1134 -msave-toc-indirect -mno-save-toc-indirect @gol
1135 -mpower8-fusion -mno-mpower8-fusion -mpower8-vector -mno-power8-vector @gol
1136 -mcrypto -mno-crypto -mhtm -mno-htm @gol
1137 -mquad-memory -mno-quad-memory @gol
1138 -mquad-memory-atomic -mno-quad-memory-atomic @gol
1139 -mcompat-align-parm -mno-compat-align-parm @gol
1140 -mfloat128 -mno-float128 -mfloat128-hardware -mno-float128-hardware @gol
1141 -mgnu-attribute -mno-gnu-attribute @gol
1142 -mstack-protector-guard=@var{guard} -mstack-protector-guard-reg=@var{reg} @gol
1143 -mstack-protector-guard-offset=@var{offset} -mpcrel -mno-pcrel}
1146 @gccoptlist{-m64bit-doubles -m32bit-doubles -fpu -nofpu@gol
1148 -mbig-endian-data -mlittle-endian-data @gol
1151 -mas100-syntax -mno-as100-syntax@gol
1153 -mmax-constant-size=@gol
1156 -mallow-string-insns -mno-allow-string-insns@gol
1158 -mno-warn-multiple-fast-interrupts@gol
1159 -msave-acc-in-interrupts}
1161 @emph{S/390 and zSeries Options}
1162 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
1163 -mhard-float -msoft-float -mhard-dfp -mno-hard-dfp @gol
1164 -mlong-double-64 -mlong-double-128 @gol
1165 -mbackchain -mno-backchain -mpacked-stack -mno-packed-stack @gol
1166 -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol
1167 -m64 -m31 -mdebug -mno-debug -mesa -mzarch @gol
1168 -mhtm -mvx -mzvector @gol
1169 -mtpf-trace -mno-tpf-trace -mfused-madd -mno-fused-madd @gol
1170 -mwarn-framesize -mwarn-dynamicstack -mstack-size -mstack-guard @gol
1171 -mhotpatch=@var{halfwords},@var{halfwords}}
1173 @emph{Score Options}
1174 @gccoptlist{-meb -mel @gol
1178 -mscore5 -mscore5u -mscore7 -mscore7d}
1181 @gccoptlist{-m1 -m2 -m2e @gol
1182 -m2a-nofpu -m2a-single-only -m2a-single -m2a @gol
1184 -m4-nofpu -m4-single-only -m4-single -m4 @gol
1185 -m4a-nofpu -m4a-single-only -m4a-single -m4a -m4al @gol
1186 -mb -ml -mdalign -mrelax @gol
1187 -mbigtable -mfmovd -mrenesas -mno-renesas -mnomacsave @gol
1188 -mieee -mno-ieee -mbitops -misize -minline-ic_invalidate -mpadstruct @gol
1189 -mprefergot -musermode -multcost=@var{number} -mdiv=@var{strategy} @gol
1190 -mdivsi3_libfunc=@var{name} -mfixed-range=@var{register-range} @gol
1191 -maccumulate-outgoing-args @gol
1192 -matomic-model=@var{atomic-model} @gol
1193 -mbranch-cost=@var{num} -mzdcbranch -mno-zdcbranch @gol
1194 -mcbranch-force-delay-slot @gol
1195 -mfused-madd -mno-fused-madd -mfsca -mno-fsca -mfsrra -mno-fsrra @gol
1196 -mpretend-cmove -mtas}
1198 @emph{Solaris 2 Options}
1199 @gccoptlist{-mclear-hwcap -mno-clear-hwcap -mimpure-text -mno-impure-text @gol
1202 @emph{SPARC Options}
1203 @gccoptlist{-mcpu=@var{cpu-type} @gol
1204 -mtune=@var{cpu-type} @gol
1205 -mcmodel=@var{code-model} @gol
1206 -mmemory-model=@var{mem-model} @gol
1207 -m32 -m64 -mapp-regs -mno-app-regs @gol
1208 -mfaster-structs -mno-faster-structs -mflat -mno-flat @gol
1209 -mfpu -mno-fpu -mhard-float -msoft-float @gol
1210 -mhard-quad-float -msoft-quad-float @gol
1211 -mstack-bias -mno-stack-bias @gol
1212 -mstd-struct-return -mno-std-struct-return @gol
1213 -munaligned-doubles -mno-unaligned-doubles @gol
1214 -muser-mode -mno-user-mode @gol
1215 -mv8plus -mno-v8plus -mvis -mno-vis @gol
1216 -mvis2 -mno-vis2 -mvis3 -mno-vis3 @gol
1217 -mvis4 -mno-vis4 -mvis4b -mno-vis4b @gol
1218 -mcbcond -mno-cbcond -mfmaf -mno-fmaf -mfsmuld -mno-fsmuld @gol
1219 -mpopc -mno-popc -msubxc -mno-subxc @gol
1220 -mfix-at697f -mfix-ut699 -mfix-ut700 -mfix-gr712rc @gol
1223 @emph{System V Options}
1224 @gccoptlist{-Qy -Qn -YP,@var{paths} -Ym,@var{dir}}
1226 @emph{TILE-Gx Options}
1227 @gccoptlist{-mcpu=CPU -m32 -m64 -mbig-endian -mlittle-endian @gol
1228 -mcmodel=@var{code-model}}
1230 @emph{TILEPro Options}
1231 @gccoptlist{-mcpu=@var{cpu} -m32}
1234 @gccoptlist{-mlong-calls -mno-long-calls -mep -mno-ep @gol
1235 -mprolog-function -mno-prolog-function -mspace @gol
1236 -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol
1237 -mapp-regs -mno-app-regs @gol
1238 -mdisable-callt -mno-disable-callt @gol
1239 -mv850e2v3 -mv850e2 -mv850e1 -mv850es @gol
1240 -mv850e -mv850 -mv850e3v5 @gol
1251 @gccoptlist{-mg -mgnu -munix}
1253 @emph{Visium Options}
1254 @gccoptlist{-mdebug -msim -mfpu -mno-fpu -mhard-float -msoft-float @gol
1255 -mcpu=@var{cpu-type} -mtune=@var{cpu-type} -msv-mode -muser-mode}
1258 @gccoptlist{-mvms-return-codes -mdebug-main=@var{prefix} -mmalloc64 @gol
1259 -mpointer-size=@var{size}}
1261 @emph{VxWorks Options}
1262 @gccoptlist{-mrtp -non-static -Bstatic -Bdynamic @gol
1263 -Xbind-lazy -Xbind-now}
1266 @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol
1267 -mtune-ctrl=@var{feature-list} -mdump-tune-features -mno-default @gol
1268 -mfpmath=@var{unit} @gol
1269 -masm=@var{dialect} -mno-fancy-math-387 @gol
1270 -mno-fp-ret-in-387 -m80387 -mhard-float -msoft-float @gol
1271 -mno-wide-multiply -mrtd -malign-double @gol
1272 -mpreferred-stack-boundary=@var{num} @gol
1273 -mincoming-stack-boundary=@var{num} @gol
1274 -mcld -mcx16 -msahf -mmovbe -mcrc32 @gol
1275 -mrecip -mrecip=@var{opt} @gol
1276 -mvzeroupper -mprefer-avx128 -mprefer-vector-width=@var{opt} @gol
1277 -mmmx -msse -msse2 -msse3 -mssse3 -msse4.1 -msse4.2 -msse4 -mavx @gol
1278 -mavx2 -mavx512f -mavx512pf -mavx512er -mavx512cd -mavx512vl @gol
1279 -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -msha -maes @gol
1280 -mpclmul -mfsgsbase -mrdrnd -mf16c -mfma -mpconfig -mwbnoinvd @gol
1281 -mptwrite -mprefetchwt1 -mclflushopt -mclwb -mxsavec -mxsaves @gol
1282 -msse4a -m3dnow -m3dnowa -mpopcnt -mabm -mbmi -mtbm -mfma4 -mxop @gol
1283 -madx -mlzcnt -mbmi2 -mfxsr -mxsave -mxsaveopt -mrtm -mhle -mlwp @gol
1284 -mmwaitx -mclzero -mpku -mthreads -mgfni -mvaes -mwaitpkg @gol
1285 -mshstk -mmanual-endbr -mforce-indirect-call -mavx512vbmi2 -mavx512bf16 -menqcmd @gol
1286 -mvpclmulqdq -mavx512bitalg -mmovdiri -mmovdir64b -mavx512vpopcntdq @gol
1287 -mavx5124fmaps -mavx512vnni -mavx5124vnniw -mprfchw -mrdpid @gol
1288 -mrdseed -msgx -mavx512vp2intersect@gol
1289 -mcldemote -mms-bitfields -mno-align-stringops -minline-all-stringops @gol
1290 -minline-stringops-dynamically -mstringop-strategy=@var{alg} @gol
1291 -mmemcpy-strategy=@var{strategy} -mmemset-strategy=@var{strategy} @gol
1292 -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol
1293 -m96bit-long-double -mlong-double-64 -mlong-double-80 -mlong-double-128 @gol
1294 -mregparm=@var{num} -msseregparm @gol
1295 -mveclibabi=@var{type} -mvect8-ret-in-mem @gol
1296 -mpc32 -mpc64 -mpc80 -mstackrealign @gol
1297 -momit-leaf-frame-pointer -mno-red-zone -mno-tls-direct-seg-refs @gol
1298 -mcmodel=@var{code-model} -mabi=@var{name} -maddress-mode=@var{mode} @gol
1299 -m32 -m64 -mx32 -m16 -miamcu -mlarge-data-threshold=@var{num} @gol
1300 -msse2avx -mfentry -mrecord-mcount -mnop-mcount -m8bit-idiv @gol
1301 -minstrument-return=@var{type} -mfentry-name=@var{name} -mfentry-section=@var{name} @gol
1302 -mavx256-split-unaligned-load -mavx256-split-unaligned-store @gol
1303 -malign-data=@var{type} -mstack-protector-guard=@var{guard} @gol
1304 -mstack-protector-guard-reg=@var{reg} @gol
1305 -mstack-protector-guard-offset=@var{offset} @gol
1306 -mstack-protector-guard-symbol=@var{symbol} @gol
1307 -mgeneral-regs-only -mcall-ms2sysv-xlogues @gol
1308 -mindirect-branch=@var{choice} -mfunction-return=@var{choice} @gol
1309 -mindirect-branch-register}
1311 @emph{x86 Windows Options}
1312 @gccoptlist{-mconsole -mcygwin -mno-cygwin -mdll @gol
1313 -mnop-fun-dllimport -mthread @gol
1314 -municode -mwin32 -mwindows -fno-set-stack-executable}
1316 @emph{Xstormy16 Options}
1319 @emph{Xtensa Options}
1320 @gccoptlist{-mconst16 -mno-const16 @gol
1321 -mfused-madd -mno-fused-madd @gol
1323 -mserialize-volatile -mno-serialize-volatile @gol
1324 -mtext-section-literals -mno-text-section-literals @gol
1325 -mauto-litpools -mno-auto-litpools @gol
1326 -mtarget-align -mno-target-align @gol
1327 -mlongcalls -mno-longcalls}
1329 @emph{zSeries Options}
1330 See S/390 and zSeries Options.
1334 @node Overall Options
1335 @section Options Controlling the Kind of Output
1337 Compilation can involve up to four stages: preprocessing, compilation
1338 proper, assembly and linking, always in that order. GCC is capable of
1339 preprocessing and compiling several files either into several
1340 assembler input files, or into one assembler input file; then each
1341 assembler input file produces an object file, and linking combines all
1342 the object files (those newly compiled, and those specified as input)
1343 into an executable file.
1345 @cindex file name suffix
1346 For any given input file, the file name suffix determines what kind of
1347 compilation is done:
1351 C source code that must be preprocessed.
1354 C source code that should not be preprocessed.
1357 C++ source code that should not be preprocessed.
1360 Objective-C source code. Note that you must link with the @file{libobjc}
1361 library to make an Objective-C program work.
1364 Objective-C source code that should not be preprocessed.
1368 Objective-C++ source code. Note that you must link with the @file{libobjc}
1369 library to make an Objective-C++ program work. Note that @samp{.M} refers
1370 to a literal capital M@.
1372 @item @var{file}.mii
1373 Objective-C++ source code that should not be preprocessed.
1376 C, C++, Objective-C or Objective-C++ header file to be turned into a
1377 precompiled header (default), or C, C++ header file to be turned into an
1378 Ada spec (via the @option{-fdump-ada-spec} switch).
1381 @itemx @var{file}.cp
1382 @itemx @var{file}.cxx
1383 @itemx @var{file}.cpp
1384 @itemx @var{file}.CPP
1385 @itemx @var{file}.c++
1387 C++ source code that must be preprocessed. Note that in @samp{.cxx},
1388 the last two letters must both be literally @samp{x}. Likewise,
1389 @samp{.C} refers to a literal capital C@.
1393 Objective-C++ source code that must be preprocessed.
1395 @item @var{file}.mii
1396 Objective-C++ source code that should not be preprocessed.
1400 @itemx @var{file}.hp
1401 @itemx @var{file}.hxx
1402 @itemx @var{file}.hpp
1403 @itemx @var{file}.HPP
1404 @itemx @var{file}.h++
1405 @itemx @var{file}.tcc
1406 C++ header file to be turned into a precompiled header or Ada spec.
1409 @itemx @var{file}.for
1410 @itemx @var{file}.ftn
1411 Fixed form Fortran source code that should not be preprocessed.
1414 @itemx @var{file}.FOR
1415 @itemx @var{file}.fpp
1416 @itemx @var{file}.FPP
1417 @itemx @var{file}.FTN
1418 Fixed form Fortran source code that must be preprocessed (with the traditional
1421 @item @var{file}.f90
1422 @itemx @var{file}.f95
1423 @itemx @var{file}.f03
1424 @itemx @var{file}.f08
1425 Free form Fortran source code that should not be preprocessed.
1427 @item @var{file}.F90
1428 @itemx @var{file}.F95
1429 @itemx @var{file}.F03
1430 @itemx @var{file}.F08
1431 Free form Fortran source code that must be preprocessed (with the
1432 traditional preprocessor).
1437 @item @var{file}.brig
1438 BRIG files (binary representation of HSAIL).
1447 D documentation code (Ddoc).
1449 @item @var{file}.ads
1450 Ada source code file that contains a library unit declaration (a
1451 declaration of a package, subprogram, or generic, or a generic
1452 instantiation), or a library unit renaming declaration (a package,
1453 generic, or subprogram renaming declaration). Such files are also
1456 @item @var{file}.adb
1457 Ada source code file containing a library unit body (a subprogram or
1458 package body). Such files are also called @dfn{bodies}.
1460 @c GCC also knows about some suffixes for languages not yet included:
1468 @itemx @var{file}.sx
1469 Assembler code that must be preprocessed.
1472 An object file to be fed straight into linking.
1473 Any file name with no recognized suffix is treated this way.
1477 You can specify the input language explicitly with the @option{-x} option:
1480 @item -x @var{language}
1481 Specify explicitly the @var{language} for the following input files
1482 (rather than letting the compiler choose a default based on the file
1483 name suffix). This option applies to all following input files until
1484 the next @option{-x} option. Possible values for @var{language} are:
1486 c c-header cpp-output
1487 c++ c++-header c++-cpp-output
1488 objective-c objective-c-header objective-c-cpp-output
1489 objective-c++ objective-c++-header objective-c++-cpp-output
1490 assembler assembler-with-cpp
1493 f77 f77-cpp-input f95 f95-cpp-input
1499 Turn off any specification of a language, so that subsequent files are
1500 handled according to their file name suffixes (as they are if @option{-x}
1501 has not been used at all).
1504 If you only want some of the stages of compilation, you can use
1505 @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and
1506 one of the options @option{-c}, @option{-S}, or @option{-E} to say where
1507 @command{gcc} is to stop. Note that some combinations (for example,
1508 @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all.
1513 Compile or assemble the source files, but do not link. The linking
1514 stage simply is not done. The ultimate output is in the form of an
1515 object file for each source file.
1517 By default, the object file name for a source file is made by replacing
1518 the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}.
1520 Unrecognized input files, not requiring compilation or assembly, are
1525 Stop after the stage of compilation proper; do not assemble. The output
1526 is in the form of an assembler code file for each non-assembler input
1529 By default, the assembler file name for a source file is made by
1530 replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}.
1532 Input files that don't require compilation are ignored.
1536 Stop after the preprocessing stage; do not run the compiler proper. The
1537 output is in the form of preprocessed source code, which is sent to the
1540 Input files that don't require preprocessing are ignored.
1542 @cindex output file option
1545 Place output in file @var{file}. This applies to whatever
1546 sort of output is being produced, whether it be an executable file,
1547 an object file, an assembler file or preprocessed C code.
1549 If @option{-o} is not specified, the default is to put an executable
1550 file in @file{a.out}, the object file for
1551 @file{@var{source}.@var{suffix}} in @file{@var{source}.o}, its
1552 assembler file in @file{@var{source}.s}, a precompiled header file in
1553 @file{@var{source}.@var{suffix}.gch}, and all preprocessed C source on
1558 Print (on standard error output) the commands executed to run the stages
1559 of compilation. Also print the version number of the compiler driver
1560 program and of the preprocessor and the compiler proper.
1564 Like @option{-v} except the commands are not executed and arguments
1565 are quoted unless they contain only alphanumeric characters or @code{./-_}.
1566 This is useful for shell scripts to capture the driver-generated command lines.
1570 Print (on the standard output) a description of the command-line options
1571 understood by @command{gcc}. If the @option{-v} option is also specified
1572 then @option{--help} is also passed on to the various processes
1573 invoked by @command{gcc}, so that they can display the command-line options
1574 they accept. If the @option{-Wextra} option has also been specified
1575 (prior to the @option{--help} option), then command-line options that
1576 have no documentation associated with them are also displayed.
1579 @opindex target-help
1580 Print (on the standard output) a description of target-specific command-line
1581 options for each tool. For some targets extra target-specific
1582 information may also be printed.
1584 @item --help=@{@var{class}@r{|[}^@r{]}@var{qualifier}@}@r{[},@dots{}@r{]}
1585 Print (on the standard output) a description of the command-line
1586 options understood by the compiler that fit into all specified classes
1587 and qualifiers. These are the supported classes:
1590 @item @samp{optimizers}
1591 Display all of the optimization options supported by the
1594 @item @samp{warnings}
1595 Display all of the options controlling warning messages
1596 produced by the compiler.
1599 Display target-specific options. Unlike the
1600 @option{--target-help} option however, target-specific options of the
1601 linker and assembler are not displayed. This is because those
1602 tools do not currently support the extended @option{--help=} syntax.
1605 Display the values recognized by the @option{--param}
1608 @item @var{language}
1609 Display the options supported for @var{language}, where
1610 @var{language} is the name of one of the languages supported in this
1614 Display the options that are common to all languages.
1617 These are the supported qualifiers:
1620 @item @samp{undocumented}
1621 Display only those options that are undocumented.
1624 Display options taking an argument that appears after an equal
1625 sign in the same continuous piece of text, such as:
1626 @samp{--help=target}.
1628 @item @samp{separate}
1629 Display options taking an argument that appears as a separate word
1630 following the original option, such as: @samp{-o output-file}.
1633 Thus for example to display all the undocumented target-specific
1634 switches supported by the compiler, use:
1637 --help=target,undocumented
1640 The sense of a qualifier can be inverted by prefixing it with the
1641 @samp{^} character, so for example to display all binary warning
1642 options (i.e., ones that are either on or off and that do not take an
1643 argument) that have a description, use:
1646 --help=warnings,^joined,^undocumented
1649 The argument to @option{--help=} should not consist solely of inverted
1652 Combining several classes is possible, although this usually
1653 restricts the output so much that there is nothing to display. One
1654 case where it does work, however, is when one of the classes is
1655 @var{target}. For example, to display all the target-specific
1656 optimization options, use:
1659 --help=target,optimizers
1662 The @option{--help=} option can be repeated on the command line. Each
1663 successive use displays its requested class of options, skipping
1664 those that have already been displayed. If @option{--help} is also
1665 specified anywhere on the command line then this takes precedence
1666 over any @option{--help=} option.
1668 If the @option{-Q} option appears on the command line before the
1669 @option{--help=} option, then the descriptive text displayed by
1670 @option{--help=} is changed. Instead of describing the displayed
1671 options, an indication is given as to whether the option is enabled,
1672 disabled or set to a specific value (assuming that the compiler
1673 knows this at the point where the @option{--help=} option is used).
1675 Here is a truncated example from the ARM port of @command{gcc}:
1678 % gcc -Q -mabi=2 --help=target -c
1679 The following options are target specific:
1681 -mabort-on-noreturn [disabled]
1685 The output is sensitive to the effects of previous command-line
1686 options, so for example it is possible to find out which optimizations
1687 are enabled at @option{-O2} by using:
1690 -Q -O2 --help=optimizers
1693 Alternatively you can discover which binary optimizations are enabled
1694 by @option{-O3} by using:
1697 gcc -c -Q -O3 --help=optimizers > /tmp/O3-opts
1698 gcc -c -Q -O2 --help=optimizers > /tmp/O2-opts
1699 diff /tmp/O2-opts /tmp/O3-opts | grep enabled
1704 Display the version number and copyrights of the invoked GCC@.
1706 @item -pass-exit-codes
1707 @opindex pass-exit-codes
1708 Normally the @command{gcc} program exits with the code of 1 if any
1709 phase of the compiler returns a non-success return code. If you specify
1710 @option{-pass-exit-codes}, the @command{gcc} program instead returns with
1711 the numerically highest error produced by any phase returning an error
1712 indication. The C, C++, and Fortran front ends return 4 if an internal
1713 compiler error is encountered.
1717 Use pipes rather than temporary files for communication between the
1718 various stages of compilation. This fails to work on some systems where
1719 the assembler is unable to read from a pipe; but the GNU assembler has
1722 @item -specs=@var{file}
1724 Process @var{file} after the compiler reads in the standard @file{specs}
1725 file, in order to override the defaults which the @command{gcc} driver
1726 program uses when determining what switches to pass to @command{cc1},
1727 @command{cc1plus}, @command{as}, @command{ld}, etc. More than one
1728 @option{-specs=@var{file}} can be specified on the command line, and they
1729 are processed in order, from left to right. @xref{Spec Files}, for
1730 information about the format of the @var{file}.
1734 Invoke all subcommands under a wrapper program. The name of the
1735 wrapper program and its parameters are passed as a comma separated
1739 gcc -c t.c -wrapper gdb,--args
1743 This invokes all subprograms of @command{gcc} under
1744 @samp{gdb --args}, thus the invocation of @command{cc1} is
1745 @samp{gdb --args cc1 @dots{}}.
1747 @item -ffile-prefix-map=@var{old}=@var{new}
1748 @opindex ffile-prefix-map
1749 When compiling files residing in directory @file{@var{old}}, record
1750 any references to them in the result of the compilation as if the
1751 files resided in directory @file{@var{new}} instead. Specifying this
1752 option is equivalent to specifying all the individual
1753 @option{-f*-prefix-map} options. This can be used to make reproducible
1754 builds that are location independent. See also
1755 @option{-fmacro-prefix-map} and @option{-fdebug-prefix-map}.
1757 @item -fplugin=@var{name}.so
1759 Load the plugin code in file @var{name}.so, assumed to be a
1760 shared object to be dlopen'd by the compiler. The base name of
1761 the shared object file is used to identify the plugin for the
1762 purposes of argument parsing (See
1763 @option{-fplugin-arg-@var{name}-@var{key}=@var{value}} below).
1764 Each plugin should define the callback functions specified in the
1767 @item -fplugin-arg-@var{name}-@var{key}=@var{value}
1768 @opindex fplugin-arg
1769 Define an argument called @var{key} with a value of @var{value}
1770 for the plugin called @var{name}.
1772 @item -fdump-ada-spec@r{[}-slim@r{]}
1773 @opindex fdump-ada-spec
1774 For C and C++ source and include files, generate corresponding Ada specs.
1775 @xref{Generating Ada Bindings for C and C++ headers,,, gnat_ugn,
1776 GNAT User's Guide}, which provides detailed documentation on this feature.
1778 @item -fada-spec-parent=@var{unit}
1779 @opindex fada-spec-parent
1780 In conjunction with @option{-fdump-ada-spec@r{[}-slim@r{]}} above, generate
1781 Ada specs as child units of parent @var{unit}.
1783 @item -fdump-go-spec=@var{file}
1784 @opindex fdump-go-spec
1785 For input files in any language, generate corresponding Go
1786 declarations in @var{file}. This generates Go @code{const},
1787 @code{type}, @code{var}, and @code{func} declarations which may be a
1788 useful way to start writing a Go interface to code written in some
1791 @include @value{srcdir}/../libiberty/at-file.texi
1795 @section Compiling C++ Programs
1797 @cindex suffixes for C++ source
1798 @cindex C++ source file suffixes
1799 C++ source files conventionally use one of the suffixes @samp{.C},
1800 @samp{.cc}, @samp{.cpp}, @samp{.CPP}, @samp{.c++}, @samp{.cp}, or
1801 @samp{.cxx}; C++ header files often use @samp{.hh}, @samp{.hpp},
1802 @samp{.H}, or (for shared template code) @samp{.tcc}; and
1803 preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes
1804 files with these names and compiles them as C++ programs even if you
1805 call the compiler the same way as for compiling C programs (usually
1806 with the name @command{gcc}).
1810 However, the use of @command{gcc} does not add the C++ library.
1811 @command{g++} is a program that calls GCC and automatically specifies linking
1812 against the C++ library. It treats @samp{.c},
1813 @samp{.h} and @samp{.i} files as C++ source files instead of C source
1814 files unless @option{-x} is used. This program is also useful when
1815 precompiling a C header file with a @samp{.h} extension for use in C++
1816 compilations. On many systems, @command{g++} is also installed with
1817 the name @command{c++}.
1819 @cindex invoking @command{g++}
1820 When you compile C++ programs, you may specify many of the same
1821 command-line options that you use for compiling programs in any
1822 language; or command-line options meaningful for C and related
1823 languages; or options that are meaningful only for C++ programs.
1824 @xref{C Dialect Options,,Options Controlling C Dialect}, for
1825 explanations of options for languages related to C@.
1826 @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for
1827 explanations of options that are meaningful only for C++ programs.
1829 @node C Dialect Options
1830 @section Options Controlling C Dialect
1831 @cindex dialect options
1832 @cindex language dialect options
1833 @cindex options, dialect
1835 The following options control the dialect of C (or languages derived
1836 from C, such as C++, Objective-C and Objective-C++) that the compiler
1840 @cindex ANSI support
1844 In C mode, this is equivalent to @option{-std=c90}. In C++ mode, it is
1845 equivalent to @option{-std=c++98}.
1847 This turns off certain features of GCC that are incompatible with ISO
1848 C90 (when compiling C code), or of standard C++ (when compiling C++ code),
1849 such as the @code{asm} and @code{typeof} keywords, and
1850 predefined macros such as @code{unix} and @code{vax} that identify the
1851 type of system you are using. It also enables the undesirable and
1852 rarely used ISO trigraph feature. For the C compiler,
1853 it disables recognition of C++ style @samp{//} comments as well as
1854 the @code{inline} keyword.
1856 The alternate keywords @code{__asm__}, @code{__extension__},
1857 @code{__inline__} and @code{__typeof__} continue to work despite
1858 @option{-ansi}. You would not want to use them in an ISO C program, of
1859 course, but it is useful to put them in header files that might be included
1860 in compilations done with @option{-ansi}. Alternate predefined macros
1861 such as @code{__unix__} and @code{__vax__} are also available, with or
1862 without @option{-ansi}.
1864 The @option{-ansi} option does not cause non-ISO programs to be
1865 rejected gratuitously. For that, @option{-Wpedantic} is required in
1866 addition to @option{-ansi}. @xref{Warning Options}.
1868 The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi}
1869 option is used. Some header files may notice this macro and refrain
1870 from declaring certain functions or defining certain macros that the
1871 ISO standard doesn't call for; this is to avoid interfering with any
1872 programs that might use these names for other things.
1874 Functions that are normally built in but do not have semantics
1875 defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in
1876 functions when @option{-ansi} is used. @xref{Other Builtins,,Other
1877 built-in functions provided by GCC}, for details of the functions
1882 Determine the language standard. @xref{Standards,,Language Standards
1883 Supported by GCC}, for details of these standard versions. This option
1884 is currently only supported when compiling C or C++.
1886 The compiler can accept several base standards, such as @samp{c90} or
1887 @samp{c++98}, and GNU dialects of those standards, such as
1888 @samp{gnu90} or @samp{gnu++98}. When a base standard is specified, the
1889 compiler accepts all programs following that standard plus those
1890 using GNU extensions that do not contradict it. For example,
1891 @option{-std=c90} turns off certain features of GCC that are
1892 incompatible with ISO C90, such as the @code{asm} and @code{typeof}
1893 keywords, but not other GNU extensions that do not have a meaning in
1894 ISO C90, such as omitting the middle term of a @code{?:}
1895 expression. On the other hand, when a GNU dialect of a standard is
1896 specified, all features supported by the compiler are enabled, even when
1897 those features change the meaning of the base standard. As a result, some
1898 strict-conforming programs may be rejected. The particular standard
1899 is used by @option{-Wpedantic} to identify which features are GNU
1900 extensions given that version of the standard. For example
1901 @option{-std=gnu90 -Wpedantic} warns about C++ style @samp{//}
1902 comments, while @option{-std=gnu99 -Wpedantic} does not.
1904 A value for this option must be provided; possible values are
1910 Support all ISO C90 programs (certain GNU extensions that conflict
1911 with ISO C90 are disabled). Same as @option{-ansi} for C code.
1913 @item iso9899:199409
1914 ISO C90 as modified in amendment 1.
1920 ISO C99. This standard is substantially completely supported, modulo
1921 bugs and floating-point issues
1922 (mainly but not entirely relating to optional C99 features from
1923 Annexes F and G). See
1924 @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The
1925 names @samp{c9x} and @samp{iso9899:199x} are deprecated.
1930 ISO C11, the 2011 revision of the ISO C standard. This standard is
1931 substantially completely supported, modulo bugs, floating-point issues
1932 (mainly but not entirely relating to optional C11 features from
1933 Annexes F and G) and the optional Annexes K (Bounds-checking
1934 interfaces) and L (Analyzability). The name @samp{c1x} is deprecated.
1940 ISO C17, the 2017 revision of the ISO C standard
1941 (published in 2018). This standard is
1942 same as C11 except for corrections of defects (all of which are also
1943 applied with @option{-std=c11}) and a new value of
1944 @code{__STDC_VERSION__}, and so is supported to the same extent as C11.
1947 The next version of the ISO C standard, still under development. The
1948 support for this version is experimental and incomplete.
1952 GNU dialect of ISO C90 (including some C99 features).
1956 GNU dialect of ISO C99. The name @samp{gnu9x} is deprecated.
1960 GNU dialect of ISO C11.
1961 The name @samp{gnu1x} is deprecated.
1965 GNU dialect of ISO C17. This is the default for C code.
1968 The next version of the ISO C standard, still under development, plus
1969 GNU extensions. The support for this version is experimental and
1974 The 1998 ISO C++ standard plus the 2003 technical corrigendum and some
1975 additional defect reports. Same as @option{-ansi} for C++ code.
1979 GNU dialect of @option{-std=c++98}.
1983 The 2011 ISO C++ standard plus amendments.
1984 The name @samp{c++0x} is deprecated.
1988 GNU dialect of @option{-std=c++11}.
1989 The name @samp{gnu++0x} is deprecated.
1993 The 2014 ISO C++ standard plus amendments.
1994 The name @samp{c++1y} is deprecated.
1998 GNU dialect of @option{-std=c++14}.
1999 This is the default for C++ code.
2000 The name @samp{gnu++1y} is deprecated.
2004 The 2017 ISO C++ standard plus amendments.
2005 The name @samp{c++1z} is deprecated.
2009 GNU dialect of @option{-std=c++17}.
2010 The name @samp{gnu++1z} is deprecated.
2013 The next revision of the ISO C++ standard, tentatively planned for
2014 2020. Support is highly experimental, and will almost certainly
2015 change in incompatible ways in future releases.
2018 GNU dialect of @option{-std=c++2a}. Support is highly experimental,
2019 and will almost certainly change in incompatible ways in future
2023 @item -fgnu89-inline
2024 @opindex fgnu89-inline
2025 The option @option{-fgnu89-inline} tells GCC to use the traditional
2026 GNU semantics for @code{inline} functions when in C99 mode.
2027 @xref{Inline,,An Inline Function is As Fast As a Macro}.
2028 Using this option is roughly equivalent to adding the
2029 @code{gnu_inline} function attribute to all inline functions
2030 (@pxref{Function Attributes}).
2032 The option @option{-fno-gnu89-inline} explicitly tells GCC to use the
2033 C99 semantics for @code{inline} when in C99 or gnu99 mode (i.e., it
2034 specifies the default behavior).
2035 This option is not supported in @option{-std=c90} or
2036 @option{-std=gnu90} mode.
2038 The preprocessor macros @code{__GNUC_GNU_INLINE__} and
2039 @code{__GNUC_STDC_INLINE__} may be used to check which semantics are
2040 in effect for @code{inline} functions. @xref{Common Predefined
2041 Macros,,,cpp,The C Preprocessor}.
2043 @item -fpermitted-flt-eval-methods=@var{style}
2044 @opindex fpermitted-flt-eval-methods
2045 @opindex fpermitted-flt-eval-methods=c11
2046 @opindex fpermitted-flt-eval-methods=ts-18661-3
2047 ISO/IEC TS 18661-3 defines new permissible values for
2048 @code{FLT_EVAL_METHOD} that indicate that operations and constants with
2049 a semantic type that is an interchange or extended format should be
2050 evaluated to the precision and range of that type. These new values are
2051 a superset of those permitted under C99/C11, which does not specify the
2052 meaning of other positive values of @code{FLT_EVAL_METHOD}. As such, code
2053 conforming to C11 may not have been written expecting the possibility of
2056 @option{-fpermitted-flt-eval-methods} specifies whether the compiler
2057 should allow only the values of @code{FLT_EVAL_METHOD} specified in C99/C11,
2058 or the extended set of values specified in ISO/IEC TS 18661-3.
2060 @var{style} is either @code{c11} or @code{ts-18661-3} as appropriate.
2062 The default when in a standards compliant mode (@option{-std=c11} or similar)
2063 is @option{-fpermitted-flt-eval-methods=c11}. The default when in a GNU
2064 dialect (@option{-std=gnu11} or similar) is
2065 @option{-fpermitted-flt-eval-methods=ts-18661-3}.
2067 @item -aux-info @var{filename}
2069 Output to the given filename prototyped declarations for all functions
2070 declared and/or defined in a translation unit, including those in header
2071 files. This option is silently ignored in any language other than C@.
2073 Besides declarations, the file indicates, in comments, the origin of
2074 each declaration (source file and line), whether the declaration was
2075 implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or
2076 @samp{O} for old, respectively, in the first character after the line
2077 number and the colon), and whether it came from a declaration or a
2078 definition (@samp{C} or @samp{F}, respectively, in the following
2079 character). In the case of function definitions, a K&R-style list of
2080 arguments followed by their declarations is also provided, inside
2081 comments, after the declaration.
2083 @item -fallow-parameterless-variadic-functions
2084 @opindex fallow-parameterless-variadic-functions
2085 Accept variadic functions without named parameters.
2087 Although it is possible to define such a function, this is not very
2088 useful as it is not possible to read the arguments. This is only
2089 supported for C as this construct is allowed by C++.
2094 Do not recognize @code{asm}, @code{inline} or @code{typeof} as a
2095 keyword, so that code can use these words as identifiers. You can use
2096 the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__}
2097 instead. @option{-ansi} implies @option{-fno-asm}.
2099 In C++, this switch only affects the @code{typeof} keyword, since
2100 @code{asm} and @code{inline} are standard keywords. You may want to
2101 use the @option{-fno-gnu-keywords} flag instead, which has the same
2102 effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this
2103 switch only affects the @code{asm} and @code{typeof} keywords, since
2104 @code{inline} is a standard keyword in ISO C99.
2107 @itemx -fno-builtin-@var{function}
2108 @opindex fno-builtin
2110 @cindex built-in functions
2111 Don't recognize built-in functions that do not begin with
2112 @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in
2113 functions provided by GCC}, for details of the functions affected,
2114 including those which are not built-in functions when @option{-ansi} or
2115 @option{-std} options for strict ISO C conformance are used because they
2116 do not have an ISO standard meaning.
2118 GCC normally generates special code to handle certain built-in functions
2119 more efficiently; for instance, calls to @code{alloca} may become single
2120 instructions which adjust the stack directly, and calls to @code{memcpy}
2121 may become inline copy loops. The resulting code is often both smaller
2122 and faster, but since the function calls no longer appear as such, you
2123 cannot set a breakpoint on those calls, nor can you change the behavior
2124 of the functions by linking with a different library. In addition,
2125 when a function is recognized as a built-in function, GCC may use
2126 information about that function to warn about problems with calls to
2127 that function, or to generate more efficient code, even if the
2128 resulting code still contains calls to that function. For example,
2129 warnings are given with @option{-Wformat} for bad calls to
2130 @code{printf} when @code{printf} is built in and @code{strlen} is
2131 known not to modify global memory.
2133 With the @option{-fno-builtin-@var{function}} option
2134 only the built-in function @var{function} is
2135 disabled. @var{function} must not begin with @samp{__builtin_}. If a
2136 function is named that is not built-in in this version of GCC, this
2137 option is ignored. There is no corresponding
2138 @option{-fbuiltin-@var{function}} option; if you wish to enable
2139 built-in functions selectively when using @option{-fno-builtin} or
2140 @option{-ffreestanding}, you may define macros such as:
2143 #define abs(n) __builtin_abs ((n))
2144 #define strcpy(d, s) __builtin_strcpy ((d), (s))
2150 Enable parsing of function definitions marked with @code{__GIMPLE}.
2151 This is an experimental feature that allows unit testing of GIMPLE
2156 @cindex hosted environment
2158 Assert that compilation targets a hosted environment. This implies
2159 @option{-fbuiltin}. A hosted environment is one in which the
2160 entire standard library is available, and in which @code{main} has a return
2161 type of @code{int}. Examples are nearly everything except a kernel.
2162 This is equivalent to @option{-fno-freestanding}.
2164 @item -ffreestanding
2165 @opindex ffreestanding
2166 @cindex hosted environment
2168 Assert that compilation targets a freestanding environment. This
2169 implies @option{-fno-builtin}. A freestanding environment
2170 is one in which the standard library may not exist, and program startup may
2171 not necessarily be at @code{main}. The most obvious example is an OS kernel.
2172 This is equivalent to @option{-fno-hosted}.
2174 @xref{Standards,,Language Standards Supported by GCC}, for details of
2175 freestanding and hosted environments.
2179 @cindex OpenACC accelerator programming
2180 Enable handling of OpenACC directives @code{#pragma acc} in C/C++ and
2181 @code{!$acc} in Fortran. When @option{-fopenacc} is specified, the
2182 compiler generates accelerated code according to the OpenACC Application
2183 Programming Interface v2.0 @w{@uref{https://www.openacc.org}}. This option
2184 implies @option{-pthread}, and thus is only supported on targets that
2185 have support for @option{-pthread}.
2187 @item -fopenacc-dim=@var{geom}
2188 @opindex fopenacc-dim
2189 @cindex OpenACC accelerator programming
2190 Specify default compute dimensions for parallel offload regions that do
2191 not explicitly specify. The @var{geom} value is a triple of
2192 ':'-separated sizes, in order 'gang', 'worker' and, 'vector'. A size
2193 can be omitted, to use a target-specific default value.
2197 @cindex OpenMP parallel
2198 Enable handling of OpenMP directives @code{#pragma omp} in C/C++ and
2199 @code{!$omp} in Fortran. When @option{-fopenmp} is specified, the
2200 compiler generates parallel code according to the OpenMP Application
2201 Program Interface v4.5 @w{@uref{https://www.openmp.org}}. This option
2202 implies @option{-pthread}, and thus is only supported on targets that
2203 have support for @option{-pthread}. @option{-fopenmp} implies
2204 @option{-fopenmp-simd}.
2207 @opindex fopenmp-simd
2210 Enable handling of OpenMP's SIMD directives with @code{#pragma omp}
2211 in C/C++ and @code{!$omp} in Fortran. Other OpenMP directives
2216 When the option @option{-fgnu-tm} is specified, the compiler
2217 generates code for the Linux variant of Intel's current Transactional
2218 Memory ABI specification document (Revision 1.1, May 6 2009). This is
2219 an experimental feature whose interface may change in future versions
2220 of GCC, as the official specification changes. Please note that not
2221 all architectures are supported for this feature.
2223 For more information on GCC's support for transactional memory,
2224 @xref{Enabling libitm,,The GNU Transactional Memory Library,libitm,GNU
2225 Transactional Memory Library}.
2227 Note that the transactional memory feature is not supported with
2228 non-call exceptions (@option{-fnon-call-exceptions}).
2230 @item -fms-extensions
2231 @opindex fms-extensions
2232 Accept some non-standard constructs used in Microsoft header files.
2234 In C++ code, this allows member names in structures to be similar
2235 to previous types declarations.
2244 Some cases of unnamed fields in structures and unions are only
2245 accepted with this option. @xref{Unnamed Fields,,Unnamed struct/union
2246 fields within structs/unions}, for details.
2248 Note that this option is off for all targets except for x86
2249 targets using ms-abi.
2251 @item -fplan9-extensions
2252 @opindex fplan9-extensions
2253 Accept some non-standard constructs used in Plan 9 code.
2255 This enables @option{-fms-extensions}, permits passing pointers to
2256 structures with anonymous fields to functions that expect pointers to
2257 elements of the type of the field, and permits referring to anonymous
2258 fields declared using a typedef. @xref{Unnamed Fields,,Unnamed
2259 struct/union fields within structs/unions}, for details. This is only
2260 supported for C, not C++.
2262 @item -fcond-mismatch
2263 @opindex fcond-mismatch
2264 Allow conditional expressions with mismatched types in the second and
2265 third arguments. The value of such an expression is void. This option
2266 is not supported for C++.
2268 @item -flax-vector-conversions
2269 @opindex flax-vector-conversions
2270 Allow implicit conversions between vectors with differing numbers of
2271 elements and/or incompatible element types. This option should not be
2274 @item -funsigned-char
2275 @opindex funsigned-char
2276 Let the type @code{char} be unsigned, like @code{unsigned char}.
2278 Each kind of machine has a default for what @code{char} should
2279 be. It is either like @code{unsigned char} by default or like
2280 @code{signed char} by default.
2282 Ideally, a portable program should always use @code{signed char} or
2283 @code{unsigned char} when it depends on the signedness of an object.
2284 But many programs have been written to use plain @code{char} and
2285 expect it to be signed, or expect it to be unsigned, depending on the
2286 machines they were written for. This option, and its inverse, let you
2287 make such a program work with the opposite default.
2289 The type @code{char} is always a distinct type from each of
2290 @code{signed char} or @code{unsigned char}, even though its behavior
2291 is always just like one of those two.
2294 @opindex fsigned-char
2295 Let the type @code{char} be signed, like @code{signed char}.
2297 Note that this is equivalent to @option{-fno-unsigned-char}, which is
2298 the negative form of @option{-funsigned-char}. Likewise, the option
2299 @option{-fno-signed-char} is equivalent to @option{-funsigned-char}.
2301 @item -fsigned-bitfields
2302 @itemx -funsigned-bitfields
2303 @itemx -fno-signed-bitfields
2304 @itemx -fno-unsigned-bitfields
2305 @opindex fsigned-bitfields
2306 @opindex funsigned-bitfields
2307 @opindex fno-signed-bitfields
2308 @opindex fno-unsigned-bitfields
2309 These options control whether a bit-field is signed or unsigned, when the
2310 declaration does not use either @code{signed} or @code{unsigned}. By
2311 default, such a bit-field is signed, because this is consistent: the
2312 basic integer types such as @code{int} are signed types.
2314 @item -fsso-struct=@var{endianness}
2315 @opindex fsso-struct
2316 Set the default scalar storage order of structures and unions to the
2317 specified endianness. The accepted values are @samp{big-endian},
2318 @samp{little-endian} and @samp{native} for the native endianness of
2319 the target (the default). This option is not supported for C++.
2321 @strong{Warning:} the @option{-fsso-struct} switch causes GCC to generate
2322 code that is not binary compatible with code generated without it if the
2323 specified endianness is not the native endianness of the target.
2326 @node C++ Dialect Options
2327 @section Options Controlling C++ Dialect
2329 @cindex compiler options, C++
2330 @cindex C++ options, command-line
2331 @cindex options, C++
2332 This section describes the command-line options that are only meaningful
2333 for C++ programs. You can also use most of the GNU compiler options
2334 regardless of what language your program is in. For example, you
2335 might compile a file @file{firstClass.C} like this:
2338 g++ -g -fstrict-enums -O -c firstClass.C
2342 In this example, only @option{-fstrict-enums} is an option meant
2343 only for C++ programs; you can use the other options with any
2344 language supported by GCC@.
2346 Some options for compiling C programs, such as @option{-std}, are also
2347 relevant for C++ programs.
2348 @xref{C Dialect Options,,Options Controlling C Dialect}.
2350 Here is a list of options that are @emph{only} for compiling C++ programs:
2354 @item -fabi-version=@var{n}
2355 @opindex fabi-version
2356 Use version @var{n} of the C++ ABI@. The default is version 0.
2358 Version 0 refers to the version conforming most closely to
2359 the C++ ABI specification. Therefore, the ABI obtained using version 0
2360 will change in different versions of G++ as ABI bugs are fixed.
2362 Version 1 is the version of the C++ ABI that first appeared in G++ 3.2.
2364 Version 2 is the version of the C++ ABI that first appeared in G++
2365 3.4, and was the default through G++ 4.9.
2367 Version 3 corrects an error in mangling a constant address as a
2370 Version 4, which first appeared in G++ 4.5, implements a standard
2371 mangling for vector types.
2373 Version 5, which first appeared in G++ 4.6, corrects the mangling of
2374 attribute const/volatile on function pointer types, decltype of a
2375 plain decl, and use of a function parameter in the declaration of
2378 Version 6, which first appeared in G++ 4.7, corrects the promotion
2379 behavior of C++11 scoped enums and the mangling of template argument
2380 packs, const/static_cast, prefix ++ and --, and a class scope function
2381 used as a template argument.
2383 Version 7, which first appeared in G++ 4.8, that treats nullptr_t as a
2384 builtin type and corrects the mangling of lambdas in default argument
2387 Version 8, which first appeared in G++ 4.9, corrects the substitution
2388 behavior of function types with function-cv-qualifiers.
2390 Version 9, which first appeared in G++ 5.2, corrects the alignment of
2393 Version 10, which first appeared in G++ 6.1, adds mangling of
2394 attributes that affect type identity, such as ia32 calling convention
2395 attributes (e.g.@: @samp{stdcall}).
2397 Version 11, which first appeared in G++ 7, corrects the mangling of
2398 sizeof... expressions and operator names. For multiple entities with
2399 the same name within a function, that are declared in different scopes,
2400 the mangling now changes starting with the twelfth occurrence. It also
2401 implies @option{-fnew-inheriting-ctors}.
2403 Version 12, which first appeared in G++ 8, corrects the calling
2404 conventions for empty classes on the x86_64 target and for classes
2405 with only deleted copy/move constructors. It accidentally changes the
2406 calling convention for classes with a deleted copy constructor and a
2407 trivial move constructor.
2409 Version 13, which first appeared in G++ 8.2, fixes the accidental
2410 change in version 12.
2412 See also @option{-Wabi}.
2414 @item -fabi-compat-version=@var{n}
2415 @opindex fabi-compat-version
2416 On targets that support strong aliases, G++
2417 works around mangling changes by creating an alias with the correct
2418 mangled name when defining a symbol with an incorrect mangled name.
2419 This switch specifies which ABI version to use for the alias.
2421 With @option{-fabi-version=0} (the default), this defaults to 11 (GCC 7
2422 compatibility). If another ABI version is explicitly selected, this
2423 defaults to 0. For compatibility with GCC versions 3.2 through 4.9,
2424 use @option{-fabi-compat-version=2}.
2426 If this option is not provided but @option{-Wabi=@var{n}} is, that
2427 version is used for compatibility aliases. If this option is provided
2428 along with @option{-Wabi} (without the version), the version from this
2429 option is used for the warning.
2431 @item -fno-access-control
2432 @opindex fno-access-control
2433 @opindex faccess-control
2434 Turn off all access checking. This switch is mainly useful for working
2435 around bugs in the access control code.
2438 @opindex faligned-new
2439 Enable support for C++17 @code{new} of types that require more
2440 alignment than @code{void* ::operator new(std::size_t)} provides. A
2441 numeric argument such as @code{-faligned-new=32} can be used to
2442 specify how much alignment (in bytes) is provided by that function,
2443 but few users will need to override the default of
2444 @code{alignof(std::max_align_t)}.
2446 This flag is enabled by default for @option{-std=c++17}.
2451 @opindex fno-char8_t
2452 Enable support for @code{char8_t} as adopted for C++2a. This includes
2453 the addition of a new @code{char8_t} fundamental type, changes to the
2454 types of UTF-8 string and character literals, new signatures for
2455 user-defined literals, associated standard library updates, and new
2456 @code{__cpp_char8_t} and @code{__cpp_lib_char8_t} feature test macros.
2458 This option enables functions to be overloaded for ordinary and UTF-8
2462 int f(const char *); // #1
2463 int f(const char8_t *); // #2
2464 int v1 = f("text"); // Calls #1
2465 int v2 = f(u8"text"); // Calls #2
2469 and introduces new signatures for user-defined literals:
2472 int operator""_udl1(char8_t);
2473 int v3 = u8'x'_udl1;
2474 int operator""_udl2(const char8_t*, std::size_t);
2475 int v4 = u8"text"_udl2;
2476 template<typename T, T...> int operator""_udl3();
2477 int v5 = u8"text"_udl3;
2481 The change to the types of UTF-8 string and character literals introduces
2482 incompatibilities with ISO C++11 and later standards. For example, the
2483 following code is well-formed under ISO C++11, but is ill-formed when
2484 @option{-fchar8_t} is specified.
2487 char ca[] = u8"xx"; // error: char-array initialized from wide
2489 const char *cp = u8"xx";// error: invalid conversion from
2490 // `const char8_t*' to `const char*'
2492 auto v = f(u8"xx"); // error: invalid conversion from
2493 // `const char8_t*' to `const char*'
2494 std::string s@{u8"xx"@}; // error: no matching function for call to
2495 // `std::basic_string<char>::basic_string()'
2496 using namespace std::literals;
2497 s = u8"xx"s; // error: conversion from
2498 // `basic_string<char8_t>' to non-scalar
2499 // type `basic_string<char>' requested
2504 Check that the pointer returned by @code{operator new} is non-null
2505 before attempting to modify the storage allocated. This check is
2506 normally unnecessary because the C++ standard specifies that
2507 @code{operator new} only returns @code{0} if it is declared
2508 @code{throw()}, in which case the compiler always checks the
2509 return value even without this option. In all other cases, when
2510 @code{operator new} has a non-empty exception specification, memory
2511 exhaustion is signalled by throwing @code{std::bad_alloc}. See also
2512 @samp{new (nothrow)}.
2516 Enable support for the C++ Extensions for Concepts Technical
2517 Specification, ISO 19217 (2015), which allows code like
2520 template <class T> concept bool Addable = requires (T t) @{ t + t; @};
2521 template <Addable T> T add (T a, T b) @{ return a + b; @}
2524 @item -fconstexpr-depth=@var{n}
2525 @opindex fconstexpr-depth
2526 Set the maximum nested evaluation depth for C++11 constexpr functions
2527 to @var{n}. A limit is needed to detect endless recursion during
2528 constant expression evaluation. The minimum specified by the standard
2531 @item -fconstexpr-cache-depth=@var{n}
2532 @opindex fconstexpr-cache-depth
2533 Set the maximum level of nested evaluation depth for C++11 constexpr
2534 functions that will be cached to @var{n}. This is a heuristic that
2535 trades off compilation speed (when the cache avoids repeated
2536 calculations) against memory consumption (when the cache grows very
2537 large from highly recursive evaluations). The default is 8. Very few
2538 users are likely to want to adjust it, but if your code does heavy
2539 constexpr calculations you might want to experiment to find which
2540 value works best for you.
2542 @item -fconstexpr-loop-limit=@var{n}
2543 @opindex fconstexpr-loop-limit
2544 Set the maximum number of iterations for a loop in C++14 constexpr functions
2545 to @var{n}. A limit is needed to detect infinite loops during
2546 constant expression evaluation. The default is 262144 (1<<18).
2548 @item -fconstexpr-ops-limit=@var{n}
2549 @opindex fconstexpr-ops-limit
2550 Set the maximum number of operations during a single constexpr evaluation.
2551 Even when number of iterations of a single loop is limited with the above limit,
2552 if there are several nested loops and each of them has many iterations but still
2553 smaller than the above limit, or if in a body of some loop or even outside
2554 of a loop too many expressions need to be evaluated, the resulting constexpr
2555 evaluation might take too long.
2556 The default is 33554432 (1<<25).
2558 @item -fno-elide-constructors
2559 @opindex fno-elide-constructors
2560 @opindex felide-constructors
2561 The C++ standard allows an implementation to omit creating a temporary
2562 that is only used to initialize another object of the same type.
2563 Specifying this option disables that optimization, and forces G++ to
2564 call the copy constructor in all cases. This option also causes G++
2565 to call trivial member functions which otherwise would be expanded inline.
2567 In C++17, the compiler is required to omit these temporaries, but this
2568 option still affects trivial member functions.
2570 @item -fno-enforce-eh-specs
2571 @opindex fno-enforce-eh-specs
2572 @opindex fenforce-eh-specs
2573 Don't generate code to check for violation of exception specifications
2574 at run time. This option violates the C++ standard, but may be useful
2575 for reducing code size in production builds, much like defining
2576 @code{NDEBUG}. This does not give user code permission to throw
2577 exceptions in violation of the exception specifications; the compiler
2578 still optimizes based on the specifications, so throwing an
2579 unexpected exception results in undefined behavior at run time.
2581 @item -fextern-tls-init
2582 @itemx -fno-extern-tls-init
2583 @opindex fextern-tls-init
2584 @opindex fno-extern-tls-init
2585 The C++11 and OpenMP standards allow @code{thread_local} and
2586 @code{threadprivate} variables to have dynamic (runtime)
2587 initialization. To support this, any use of such a variable goes
2588 through a wrapper function that performs any necessary initialization.
2589 When the use and definition of the variable are in the same
2590 translation unit, this overhead can be optimized away, but when the
2591 use is in a different translation unit there is significant overhead
2592 even if the variable doesn't actually need dynamic initialization. If
2593 the programmer can be sure that no use of the variable in a
2594 non-defining TU needs to trigger dynamic initialization (either
2595 because the variable is statically initialized, or a use of the
2596 variable in the defining TU will be executed before any uses in
2597 another TU), they can avoid this overhead with the
2598 @option{-fno-extern-tls-init} option.
2600 On targets that support symbol aliases, the default is
2601 @option{-fextern-tls-init}. On targets that do not support symbol
2602 aliases, the default is @option{-fno-extern-tls-init}.
2604 @item -fno-gnu-keywords
2605 @opindex fno-gnu-keywords
2606 @opindex fgnu-keywords
2607 Do not recognize @code{typeof} as a keyword, so that code can use this
2608 word as an identifier. You can use the keyword @code{__typeof__} instead.
2609 This option is implied by the strict ISO C++ dialects: @option{-ansi},
2610 @option{-std=c++98}, @option{-std=c++11}, etc.
2612 @item -fno-implicit-templates
2613 @opindex fno-implicit-templates
2614 @opindex fimplicit-templates
2615 Never emit code for non-inline templates that are instantiated
2616 implicitly (i.e.@: by use); only emit code for explicit instantiations.
2617 If you use this option, you must take care to structure your code to
2618 include all the necessary explicit instantiations to avoid getting
2619 undefined symbols at link time.
2620 @xref{Template Instantiation}, for more information.
2622 @item -fno-implicit-inline-templates
2623 @opindex fno-implicit-inline-templates
2624 @opindex fimplicit-inline-templates
2625 Don't emit code for implicit instantiations of inline templates, either.
2626 The default is to handle inlines differently so that compiles with and
2627 without optimization need the same set of explicit instantiations.
2629 @item -fno-implement-inlines
2630 @opindex fno-implement-inlines
2631 @opindex fimplement-inlines
2632 To save space, do not emit out-of-line copies of inline functions
2633 controlled by @code{#pragma implementation}. This causes linker
2634 errors if these functions are not inlined everywhere they are called.
2636 @item -fms-extensions
2637 @opindex fms-extensions
2638 Disable Wpedantic warnings about constructs used in MFC, such as implicit
2639 int and getting a pointer to member function via non-standard syntax.
2641 @item -fnew-inheriting-ctors
2642 @opindex fnew-inheriting-ctors
2643 Enable the P0136 adjustment to the semantics of C++11 constructor
2644 inheritance. This is part of C++17 but also considered to be a Defect
2645 Report against C++11 and C++14. This flag is enabled by default
2646 unless @option{-fabi-version=10} or lower is specified.
2648 @item -fnew-ttp-matching
2649 @opindex fnew-ttp-matching
2650 Enable the P0522 resolution to Core issue 150, template template
2651 parameters and default arguments: this allows a template with default
2652 template arguments as an argument for a template template parameter
2653 with fewer template parameters. This flag is enabled by default for
2654 @option{-std=c++17}.
2656 @item -fno-nonansi-builtins
2657 @opindex fno-nonansi-builtins
2658 @opindex fnonansi-builtins
2659 Disable built-in declarations of functions that are not mandated by
2660 ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit},
2661 @code{index}, @code{bzero}, @code{conjf}, and other related functions.
2664 @opindex fnothrow-opt
2665 Treat a @code{throw()} exception specification as if it were a
2666 @code{noexcept} specification to reduce or eliminate the text size
2667 overhead relative to a function with no exception specification. If
2668 the function has local variables of types with non-trivial
2669 destructors, the exception specification actually makes the
2670 function smaller because the EH cleanups for those variables can be
2671 optimized away. The semantic effect is that an exception thrown out of
2672 a function with such an exception specification results in a call
2673 to @code{terminate} rather than @code{unexpected}.
2675 @item -fno-operator-names
2676 @opindex fno-operator-names
2677 @opindex foperator-names
2678 Do not treat the operator name keywords @code{and}, @code{bitand},
2679 @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as
2680 synonyms as keywords.
2682 @item -fno-optional-diags
2683 @opindex fno-optional-diags
2684 @opindex foptional-diags
2685 Disable diagnostics that the standard says a compiler does not need to
2686 issue. Currently, the only such diagnostic issued by G++ is the one for
2687 a name having multiple meanings within a class.
2690 @opindex fpermissive
2691 Downgrade some diagnostics about nonconformant code from errors to
2692 warnings. Thus, using @option{-fpermissive} allows some
2693 nonconforming code to compile.
2695 @item -fno-pretty-templates
2696 @opindex fno-pretty-templates
2697 @opindex fpretty-templates
2698 When an error message refers to a specialization of a function
2699 template, the compiler normally prints the signature of the
2700 template followed by the template arguments and any typedefs or
2701 typenames in the signature (e.g.@: @code{void f(T) [with T = int]}
2702 rather than @code{void f(int)}) so that it's clear which template is
2703 involved. When an error message refers to a specialization of a class
2704 template, the compiler omits any template arguments that match
2705 the default template arguments for that template. If either of these
2706 behaviors make it harder to understand the error message rather than
2707 easier, you can use @option{-fno-pretty-templates} to disable them.
2712 Disable generation of information about every class with virtual
2713 functions for use by the C++ run-time type identification features
2714 (@code{dynamic_cast} and @code{typeid}). If you don't use those parts
2715 of the language, you can save some space by using this flag. Note that
2716 exception handling uses the same information, but G++ generates it as
2717 needed. The @code{dynamic_cast} operator can still be used for casts that
2718 do not require run-time type information, i.e.@: casts to @code{void *} or to
2719 unambiguous base classes.
2721 Mixing code compiled with @option{-frtti} with that compiled with
2722 @option{-fno-rtti} may not work. For example, programs may
2723 fail to link if a class compiled with @option{-fno-rtti} is used as a base
2724 for a class compiled with @option{-frtti}.
2726 @item -fsized-deallocation
2727 @opindex fsized-deallocation
2728 Enable the built-in global declarations
2730 void operator delete (void *, std::size_t) noexcept;
2731 void operator delete[] (void *, std::size_t) noexcept;
2733 as introduced in C++14. This is useful for user-defined replacement
2734 deallocation functions that, for example, use the size of the object
2735 to make deallocation faster. Enabled by default under
2736 @option{-std=c++14} and above. The flag @option{-Wsized-deallocation}
2737 warns about places that might want to add a definition.
2739 @item -fstrict-enums
2740 @opindex fstrict-enums
2741 Allow the compiler to optimize using the assumption that a value of
2742 enumerated type can only be one of the values of the enumeration (as
2743 defined in the C++ standard; basically, a value that can be
2744 represented in the minimum number of bits needed to represent all the
2745 enumerators). This assumption may not be valid if the program uses a
2746 cast to convert an arbitrary integer value to the enumerated type.
2748 @item -fstrong-eval-order
2749 @opindex fstrong-eval-order
2750 Evaluate member access, array subscripting, and shift expressions in
2751 left-to-right order, and evaluate assignment in right-to-left order,
2752 as adopted for C++17. Enabled by default with @option{-std=c++17}.
2753 @option{-fstrong-eval-order=some} enables just the ordering of member
2754 access and shift expressions, and is the default without
2755 @option{-std=c++17}.
2757 @item -ftemplate-backtrace-limit=@var{n}
2758 @opindex ftemplate-backtrace-limit
2759 Set the maximum number of template instantiation notes for a single
2760 warning or error to @var{n}. The default value is 10.
2762 @item -ftemplate-depth=@var{n}
2763 @opindex ftemplate-depth
2764 Set the maximum instantiation depth for template classes to @var{n}.
2765 A limit on the template instantiation depth is needed to detect
2766 endless recursions during template class instantiation. ANSI/ISO C++
2767 conforming programs must not rely on a maximum depth greater than 17
2768 (changed to 1024 in C++11). The default value is 900, as the compiler
2769 can run out of stack space before hitting 1024 in some situations.
2771 @item -fno-threadsafe-statics
2772 @opindex fno-threadsafe-statics
2773 @opindex fthreadsafe-statics
2774 Do not emit the extra code to use the routines specified in the C++
2775 ABI for thread-safe initialization of local statics. You can use this
2776 option to reduce code size slightly in code that doesn't need to be
2779 @item -fuse-cxa-atexit
2780 @opindex fuse-cxa-atexit
2781 Register destructors for objects with static storage duration with the
2782 @code{__cxa_atexit} function rather than the @code{atexit} function.
2783 This option is required for fully standards-compliant handling of static
2784 destructors, but only works if your C library supports
2785 @code{__cxa_atexit}.
2787 @item -fno-use-cxa-get-exception-ptr
2788 @opindex fno-use-cxa-get-exception-ptr
2789 @opindex fuse-cxa-get-exception-ptr
2790 Don't use the @code{__cxa_get_exception_ptr} runtime routine. This
2791 causes @code{std::uncaught_exception} to be incorrect, but is necessary
2792 if the runtime routine is not available.
2794 @item -fvisibility-inlines-hidden
2795 @opindex fvisibility-inlines-hidden
2796 This switch declares that the user does not attempt to compare
2797 pointers to inline functions or methods where the addresses of the two functions
2798 are taken in different shared objects.
2800 The effect of this is that GCC may, effectively, mark inline methods with
2801 @code{__attribute__ ((visibility ("hidden")))} so that they do not
2802 appear in the export table of a DSO and do not require a PLT indirection
2803 when used within the DSO@. Enabling this option can have a dramatic effect
2804 on load and link times of a DSO as it massively reduces the size of the
2805 dynamic export table when the library makes heavy use of templates.
2807 The behavior of this switch is not quite the same as marking the
2808 methods as hidden directly, because it does not affect static variables
2809 local to the function or cause the compiler to deduce that
2810 the function is defined in only one shared object.
2812 You may mark a method as having a visibility explicitly to negate the
2813 effect of the switch for that method. For example, if you do want to
2814 compare pointers to a particular inline method, you might mark it as
2815 having default visibility. Marking the enclosing class with explicit
2816 visibility has no effect.
2818 Explicitly instantiated inline methods are unaffected by this option
2819 as their linkage might otherwise cross a shared library boundary.
2820 @xref{Template Instantiation}.
2822 @item -fvisibility-ms-compat
2823 @opindex fvisibility-ms-compat
2824 This flag attempts to use visibility settings to make GCC's C++
2825 linkage model compatible with that of Microsoft Visual Studio.
2827 The flag makes these changes to GCC's linkage model:
2831 It sets the default visibility to @code{hidden}, like
2832 @option{-fvisibility=hidden}.
2835 Types, but not their members, are not hidden by default.
2838 The One Definition Rule is relaxed for types without explicit
2839 visibility specifications that are defined in more than one
2840 shared object: those declarations are permitted if they are
2841 permitted when this option is not used.
2844 In new code it is better to use @option{-fvisibility=hidden} and
2845 export those classes that are intended to be externally visible.
2846 Unfortunately it is possible for code to rely, perhaps accidentally,
2847 on the Visual Studio behavior.
2849 Among the consequences of these changes are that static data members
2850 of the same type with the same name but defined in different shared
2851 objects are different, so changing one does not change the other;
2852 and that pointers to function members defined in different shared
2853 objects may not compare equal. When this flag is given, it is a
2854 violation of the ODR to define types with the same name differently.
2859 Do not use weak symbol support, even if it is provided by the linker.
2860 By default, G++ uses weak symbols if they are available. This
2861 option exists only for testing, and should not be used by end-users;
2862 it results in inferior code and has no benefits. This option may
2863 be removed in a future release of G++.
2867 Do not search for header files in the standard directories specific to
2868 C++, but do still search the other standard directories. (This option
2869 is used when building the C++ library.)
2872 In addition, these optimization, warning, and code generation options
2873 have meanings only for C++ programs:
2876 @item -Wabi @r{(C, Objective-C, C++ and Objective-C++ only)}
2879 Warn when G++ it generates code that is probably not compatible with
2880 the vendor-neutral C++ ABI@. Since G++ now defaults to updating the
2881 ABI with each major release, normally @option{-Wabi} will warn only if
2882 there is a check added later in a release series for an ABI issue
2883 discovered since the initial release. @option{-Wabi} will warn about
2884 more things if an older ABI version is selected (with
2885 @option{-fabi-version=@var{n}}).
2887 @option{-Wabi} can also be used with an explicit version number to
2888 warn about compatibility with a particular @option{-fabi-version}
2889 level, e.g.@: @option{-Wabi=2} to warn about changes relative to
2890 @option{-fabi-version=2}.
2892 If an explicit version number is provided and
2893 @option{-fabi-compat-version} is not specified, the version number
2894 from this option is used for compatibility aliases. If no explicit
2895 version number is provided with this option, but
2896 @option{-fabi-compat-version} is specified, that version number is
2897 used for ABI warnings.
2899 Although an effort has been made to warn about
2900 all such cases, there are probably some cases that are not warned about,
2901 even though G++ is generating incompatible code. There may also be
2902 cases where warnings are emitted even though the code that is generated
2905 You should rewrite your code to avoid these warnings if you are
2906 concerned about the fact that code generated by G++ may not be binary
2907 compatible with code generated by other compilers.
2909 Known incompatibilities in @option{-fabi-version=2} (which was the
2910 default from GCC 3.4 to 4.9) include:
2915 A template with a non-type template parameter of reference type was
2916 mangled incorrectly:
2919 template <int &> struct S @{@};
2923 This was fixed in @option{-fabi-version=3}.
2926 SIMD vector types declared using @code{__attribute ((vector_size))} were
2927 mangled in a non-standard way that does not allow for overloading of
2928 functions taking vectors of different sizes.
2930 The mangling was changed in @option{-fabi-version=4}.
2933 @code{__attribute ((const))} and @code{noreturn} were mangled as type
2934 qualifiers, and @code{decltype} of a plain declaration was folded away.
2936 These mangling issues were fixed in @option{-fabi-version=5}.
2939 Scoped enumerators passed as arguments to a variadic function are
2940 promoted like unscoped enumerators, causing @code{va_arg} to complain.
2941 On most targets this does not actually affect the parameter passing
2942 ABI, as there is no way to pass an argument smaller than @code{int}.
2944 Also, the ABI changed the mangling of template argument packs,
2945 @code{const_cast}, @code{static_cast}, prefix increment/decrement, and
2946 a class scope function used as a template argument.
2948 These issues were corrected in @option{-fabi-version=6}.
2951 Lambdas in default argument scope were mangled incorrectly, and the
2952 ABI changed the mangling of @code{nullptr_t}.
2954 These issues were corrected in @option{-fabi-version=7}.
2957 When mangling a function type with function-cv-qualifiers, the
2958 un-qualified function type was incorrectly treated as a substitution
2961 This was fixed in @option{-fabi-version=8}, the default for GCC 5.1.
2964 @code{decltype(nullptr)} incorrectly had an alignment of 1, leading to
2965 unaligned accesses. Note that this did not affect the ABI of a
2966 function with a @code{nullptr_t} parameter, as parameters have a
2969 This was fixed in @option{-fabi-version=9}, the default for GCC 5.2.
2972 Target-specific attributes that affect the identity of a type, such as
2973 ia32 calling conventions on a function type (stdcall, regparm, etc.),
2974 did not affect the mangled name, leading to name collisions when
2975 function pointers were used as template arguments.
2977 This was fixed in @option{-fabi-version=10}, the default for GCC 6.1.
2981 It also warns about psABI-related changes. The known psABI changes at this
2987 For SysV/x86-64, unions with @code{long double} members are
2988 passed in memory as specified in psABI. For example:
2998 @code{union U} is always passed in memory.
3002 @item -Wabi-tag @r{(C++ and Objective-C++ only)}
3005 Warn when a type with an ABI tag is used in a context that does not
3006 have that ABI tag. See @ref{C++ Attributes} for more information
3009 @item -Wcomma-subscript @r{(C++ and Objective-C++ only)}
3010 @opindex Wcomma-subscript
3011 @opindex Wno-comma-subscript
3012 Warn about uses of a comma expression within a subscripting expression.
3013 This usage was deprecated in C++2a. However, a comma expression wrapped
3014 in @code{( )} is not deprecated. Example:
3018 void f(int *a, int b, int c) @{
3019 a[b,c]; // deprecated
3025 Enabled by default with @option{-std=c++2a}.
3027 @item -Wctor-dtor-privacy @r{(C++ and Objective-C++ only)}
3028 @opindex Wctor-dtor-privacy
3029 @opindex Wno-ctor-dtor-privacy
3030 Warn when a class seems unusable because all the constructors or
3031 destructors in that class are private, and it has neither friends nor
3032 public static member functions. Also warn if there are no non-private
3033 methods, and there's at least one private member function that isn't
3034 a constructor or destructor.
3036 @item -Wdelete-non-virtual-dtor @r{(C++ and Objective-C++ only)}
3037 @opindex Wdelete-non-virtual-dtor
3038 @opindex Wno-delete-non-virtual-dtor
3039 Warn when @code{delete} is used to destroy an instance of a class that
3040 has virtual functions and non-virtual destructor. It is unsafe to delete
3041 an instance of a derived class through a pointer to a base class if the
3042 base class does not have a virtual destructor. This warning is enabled
3045 @item -Wdeprecated-copy @r{(C++ and Objective-C++ only)}
3046 @opindex Wdeprecated-copy
3047 @opindex Wno-deprecated-copy
3048 Warn that the implicit declaration of a copy constructor or copy
3049 assignment operator is deprecated if the class has a user-provided
3050 copy constructor or copy assignment operator, in C++11 and up. This
3051 warning is enabled by @option{-Wextra}. With
3052 @option{-Wdeprecated-copy-dtor}, also deprecate if the class has a
3053 user-provided destructor.
3055 @item -Wno-init-list-lifetime @r{(C++ and Objective-C++ only)}
3056 @opindex Winit-list-lifetime
3057 @opindex Wno-init-list-lifetime
3058 Do not warn about uses of @code{std::initializer_list} that are likely
3059 to result in dangling pointers. Since the underlying array for an
3060 @code{initializer_list} is handled like a normal C++ temporary object,
3061 it is easy to inadvertently keep a pointer to the array past the end
3062 of the array's lifetime. For example:
3066 If a function returns a temporary @code{initializer_list}, or a local
3067 @code{initializer_list} variable, the array's lifetime ends at the end
3068 of the return statement, so the value returned has a dangling pointer.
3071 If a new-expression creates an @code{initializer_list}, the array only
3072 lives until the end of the enclosing full-expression, so the
3073 @code{initializer_list} in the heap has a dangling pointer.
3076 When an @code{initializer_list} variable is assigned from a
3077 brace-enclosed initializer list, the temporary array created for the
3078 right side of the assignment only lives until the end of the
3079 full-expression, so at the next statement the @code{initializer_list}
3080 variable has a dangling pointer.
3083 // li's initial underlying array lives as long as li
3084 std::initializer_list<int> li = @{ 1,2,3 @};
3085 // assignment changes li to point to a temporary array
3087 // now the temporary is gone and li has a dangling pointer
3088 int i = li.begin()[0] // undefined behavior
3092 When a list constructor stores the @code{begin} pointer from the
3093 @code{initializer_list} argument, this doesn't extend the lifetime of
3094 the array, so if a class variable is constructed from a temporary
3095 @code{initializer_list}, the pointer is left dangling by the end of
3096 the variable declaration statement.
3100 @item -Wliteral-suffix @r{(C++ and Objective-C++ only)}
3101 @opindex Wliteral-suffix
3102 @opindex Wno-literal-suffix
3103 Warn when a string or character literal is followed by a ud-suffix which does
3104 not begin with an underscore. As a conforming extension, GCC treats such
3105 suffixes as separate preprocessing tokens in order to maintain backwards
3106 compatibility with code that uses formatting macros from @code{<inttypes.h>}.
3110 #define __STDC_FORMAT_MACROS
3111 #include <inttypes.h>
3116 printf("My int64: %" PRId64"\n", i64);
3120 In this case, @code{PRId64} is treated as a separate preprocessing token.
3122 Additionally, warn when a user-defined literal operator is declared with
3123 a literal suffix identifier that doesn't begin with an underscore. Literal
3124 suffix identifiers that don't begin with an underscore are reserved for
3125 future standardization.
3127 This warning is enabled by default.
3129 @item -Wlto-type-mismatch
3130 @opindex Wlto-type-mismatch
3131 @opindex Wno-lto-type-mismatch
3133 During the link-time optimization warn about type mismatches in
3134 global declarations from different compilation units.
3135 Requires @option{-flto} to be enabled. Enabled by default.
3137 @item -Wno-narrowing @r{(C++ and Objective-C++ only)}
3139 @opindex Wno-narrowing
3140 For C++11 and later standards, narrowing conversions are diagnosed by default,
3141 as required by the standard. A narrowing conversion from a constant produces
3142 an error, and a narrowing conversion from a non-constant produces a warning,
3143 but @option{-Wno-narrowing} suppresses the diagnostic.
3144 Note that this does not affect the meaning of well-formed code;
3145 narrowing conversions are still considered ill-formed in SFINAE contexts.
3147 With @option{-Wnarrowing} in C++98, warn when a narrowing
3148 conversion prohibited by C++11 occurs within
3152 int i = @{ 2.2 @}; // error: narrowing from double to int
3155 This flag is included in @option{-Wall} and @option{-Wc++11-compat}.
3157 @item -Wnoexcept @r{(C++ and Objective-C++ only)}
3159 @opindex Wno-noexcept
3160 Warn when a noexcept-expression evaluates to false because of a call
3161 to a function that does not have a non-throwing exception
3162 specification (i.e. @code{throw()} or @code{noexcept}) but is known by
3163 the compiler to never throw an exception.
3165 @item -Wnoexcept-type @r{(C++ and Objective-C++ only)}
3166 @opindex Wnoexcept-type
3167 @opindex Wno-noexcept-type
3168 Warn if the C++17 feature making @code{noexcept} part of a function
3169 type changes the mangled name of a symbol relative to C++14. Enabled
3170 by @option{-Wabi} and @option{-Wc++17-compat}.
3175 template <class T> void f(T t) @{ t(); @};
3177 void h() @{ f(g); @}
3181 In C++14, @code{f} calls @code{f<void(*)()>}, but in
3182 C++17 it calls @code{f<void(*)()noexcept>}.
3184 @item -Wclass-memaccess @r{(C++ and Objective-C++ only)}
3185 @opindex Wclass-memaccess
3186 @opindex Wno-class-memaccess
3187 Warn when the destination of a call to a raw memory function such as
3188 @code{memset} or @code{memcpy} is an object of class type, and when writing
3189 into such an object might bypass the class non-trivial or deleted constructor
3190 or copy assignment, violate const-correctness or encapsulation, or corrupt
3191 virtual table pointers. Modifying the representation of such objects may
3192 violate invariants maintained by member functions of the class. For example,
3193 the call to @code{memset} below is undefined because it modifies a non-trivial
3194 class object and is, therefore, diagnosed. The safe way to either initialize
3195 or clear the storage of objects of such types is by using the appropriate
3196 constructor or assignment operator, if one is available.
3198 std::string str = "abc";
3199 memset (&str, 0, sizeof str);
3201 The @option{-Wclass-memaccess} option is enabled by @option{-Wall}.
3202 Explicitly casting the pointer to the class object to @code{void *} or
3203 to a type that can be safely accessed by the raw memory function suppresses
3206 @item -Wnon-virtual-dtor @r{(C++ and Objective-C++ only)}
3207 @opindex Wnon-virtual-dtor
3208 @opindex Wno-non-virtual-dtor
3209 Warn when a class has virtual functions and an accessible non-virtual
3210 destructor itself or in an accessible polymorphic base class, in which
3211 case it is possible but unsafe to delete an instance of a derived
3212 class through a pointer to the class itself or base class. This
3213 warning is automatically enabled if @option{-Weffc++} is specified.
3215 @item -Wregister @r{(C++ and Objective-C++ only)}
3217 @opindex Wno-register
3218 Warn on uses of the @code{register} storage class specifier, except
3219 when it is part of the GNU @ref{Explicit Register Variables} extension.
3220 The use of the @code{register} keyword as storage class specifier has
3221 been deprecated in C++11 and removed in C++17.
3222 Enabled by default with @option{-std=c++17}.
3224 @item -Wreorder @r{(C++ and Objective-C++ only)}
3226 @opindex Wno-reorder
3227 @cindex reordering, warning
3228 @cindex warning for reordering of member initializers
3229 Warn when the order of member initializers given in the code does not
3230 match the order in which they must be executed. For instance:
3236 A(): j (0), i (1) @{ @}
3241 The compiler rearranges the member initializers for @code{i}
3242 and @code{j} to match the declaration order of the members, emitting
3243 a warning to that effect. This warning is enabled by @option{-Wall}.
3245 @item -Wno-pessimizing-move @r{(C++ and Objective-C++ only)}
3246 @opindex Wpessimizing-move
3247 @opindex Wno-pessimizing-move
3248 This warning warns when a call to @code{std::move} prevents copy
3249 elision. A typical scenario when copy elision can occur is when returning in
3250 a function with a class return type, when the expression being returned is the
3251 name of a non-volatile automatic object, and is not a function parameter, and
3252 has the same type as the function return type.
3262 return std::move (t);
3266 But in this example, the @code{std::move} call prevents copy elision.
3268 This warning is enabled by @option{-Wall}.
3270 @item -Wno-redundant-move @r{(C++ and Objective-C++ only)}
3271 @opindex Wredundant-move
3272 @opindex Wno-redundant-move
3273 This warning warns about redundant calls to @code{std::move}; that is, when
3274 a move operation would have been performed even without the @code{std::move}
3275 call. This happens because the compiler is forced to treat the object as if
3276 it were an rvalue in certain situations such as returning a local variable,
3277 where copy elision isn't applicable. Consider:
3286 return std::move (t);
3290 Here, the @code{std::move} call is redundant. Because G++ implements Core
3291 Issue 1579, another example is:
3294 struct T @{ // convertible to U
3304 return std::move (t);
3307 In this example, copy elision isn't applicable because the type of the
3308 expression being returned and the function return type differ, yet G++
3309 treats the return value as if it were designated by an rvalue.
3311 This warning is enabled by @option{-Wextra}.
3313 @item -fext-numeric-literals @r{(C++ and Objective-C++ only)}
3314 @opindex fext-numeric-literals
3315 @opindex fno-ext-numeric-literals
3316 Accept imaginary, fixed-point, or machine-defined
3317 literal number suffixes as GNU extensions.
3318 When this option is turned off these suffixes are treated
3319 as C++11 user-defined literal numeric suffixes.
3320 This is on by default for all pre-C++11 dialects and all GNU dialects:
3321 @option{-std=c++98}, @option{-std=gnu++98}, @option{-std=gnu++11},
3322 @option{-std=gnu++14}.
3323 This option is off by default
3324 for ISO C++11 onwards (@option{-std=c++11}, ...).
3327 The following @option{-W@dots{}} options are not affected by @option{-Wall}.
3330 @item -Weffc++ @r{(C++ and Objective-C++ only)}
3333 Warn about violations of the following style guidelines from Scott Meyers'
3334 @cite{Effective C++} series of books:
3338 Define a copy constructor and an assignment operator for classes
3339 with dynamically-allocated memory.
3342 Prefer initialization to assignment in constructors.
3345 Have @code{operator=} return a reference to @code{*this}.
3348 Don't try to return a reference when you must return an object.
3351 Distinguish between prefix and postfix forms of increment and
3352 decrement operators.
3355 Never overload @code{&&}, @code{||}, or @code{,}.
3359 This option also enables @option{-Wnon-virtual-dtor}, which is also
3360 one of the effective C++ recommendations. However, the check is
3361 extended to warn about the lack of virtual destructor in accessible
3362 non-polymorphic bases classes too.
3364 When selecting this option, be aware that the standard library
3365 headers do not obey all of these guidelines; use @samp{grep -v}
3366 to filter out those warnings.
3368 @item -Wstrict-null-sentinel @r{(C++ and Objective-C++ only)}
3369 @opindex Wstrict-null-sentinel
3370 @opindex Wno-strict-null-sentinel
3371 Warn about the use of an uncasted @code{NULL} as sentinel. When
3372 compiling only with GCC this is a valid sentinel, as @code{NULL} is defined
3373 to @code{__null}. Although it is a null pointer constant rather than a
3374 null pointer, it is guaranteed to be of the same size as a pointer.
3375 But this use is not portable across different compilers.
3377 @item -Wno-non-template-friend @r{(C++ and Objective-C++ only)}
3378 @opindex Wno-non-template-friend
3379 @opindex Wnon-template-friend
3380 Disable warnings when non-template friend functions are declared
3381 within a template. In very old versions of GCC that predate implementation
3382 of the ISO standard, declarations such as
3383 @samp{friend int foo(int)}, where the name of the friend is an unqualified-id,
3384 could be interpreted as a particular specialization of a template
3385 function; the warning exists to diagnose compatibility problems,
3386 and is enabled by default.
3388 @item -Wold-style-cast @r{(C++ and Objective-C++ only)}
3389 @opindex Wold-style-cast
3390 @opindex Wno-old-style-cast
3391 Warn if an old-style (C-style) cast to a non-void type is used within
3392 a C++ program. The new-style casts (@code{dynamic_cast},
3393 @code{static_cast}, @code{reinterpret_cast}, and @code{const_cast}) are
3394 less vulnerable to unintended effects and much easier to search for.
3396 @item -Woverloaded-virtual @r{(C++ and Objective-C++ only)}
3397 @opindex Woverloaded-virtual
3398 @opindex Wno-overloaded-virtual
3399 @cindex overloaded virtual function, warning
3400 @cindex warning for overloaded virtual function
3401 Warn when a function declaration hides virtual functions from a
3402 base class. For example, in:
3409 struct B: public A @{
3414 the @code{A} class version of @code{f} is hidden in @code{B}, and code
3425 @item -Wno-pmf-conversions @r{(C++ and Objective-C++ only)}
3426 @opindex Wno-pmf-conversions
3427 @opindex Wpmf-conversions
3428 Disable the diagnostic for converting a bound pointer to member function
3431 @item -Wsign-promo @r{(C++ and Objective-C++ only)}
3432 @opindex Wsign-promo
3433 @opindex Wno-sign-promo
3434 Warn when overload resolution chooses a promotion from unsigned or
3435 enumerated type to a signed type, over a conversion to an unsigned type of
3436 the same size. Previous versions of G++ tried to preserve
3437 unsignedness, but the standard mandates the current behavior.
3439 @item -Wtemplates @r{(C++ and Objective-C++ only)}
3441 @opindex Wno-templates
3442 Warn when a primary template declaration is encountered. Some coding
3443 rules disallow templates, and this may be used to enforce that rule.
3444 The warning is inactive inside a system header file, such as the STL, so
3445 one can still use the STL. One may also instantiate or specialize
3448 @item -Wmultiple-inheritance @r{(C++ and Objective-C++ only)}
3449 @opindex Wmultiple-inheritance
3450 @opindex Wno-multiple-inheritance
3451 Warn when a class is defined with multiple direct base classes. Some
3452 coding rules disallow multiple inheritance, and this may be used to
3453 enforce that rule. The warning is inactive inside a system header file,
3454 such as the STL, so one can still use the STL. One may also define
3455 classes that indirectly use multiple inheritance.
3457 @item -Wvirtual-inheritance
3458 @opindex Wvirtual-inheritance
3459 @opindex Wno-virtual-inheritance
3460 Warn when a class is defined with a virtual direct base class. Some
3461 coding rules disallow multiple inheritance, and this may be used to
3462 enforce that rule. The warning is inactive inside a system header file,
3463 such as the STL, so one can still use the STL. One may also define
3464 classes that indirectly use virtual inheritance.
3467 @opindex Wnamespaces
3468 @opindex Wno-namespaces
3469 Warn when a namespace definition is opened. Some coding rules disallow
3470 namespaces, and this may be used to enforce that rule. The warning is
3471 inactive inside a system header file, such as the STL, so one can still
3472 use the STL. One may also use using directives and qualified names.
3474 @item -Wno-terminate @r{(C++ and Objective-C++ only)}
3476 @opindex Wno-terminate
3477 Disable the warning about a throw-expression that will immediately
3478 result in a call to @code{terminate}.
3480 @item -Wno-class-conversion @r{(C++ and Objective-C++ only)}
3481 @opindex Wno-class-conversion
3482 @opindex Wclass-conversion
3483 Disable the warning about the case when a conversion function converts an
3484 object to the same type, to a base class of that type, or to void; such
3485 a conversion function will never be called.
3487 @item -Wvolatile @r{(C++ and Objective-C++ only)}
3489 @opindex Wno-volatile
3490 Warn about deprecated uses of the @code{volatile} qualifier. This includes
3491 postfix and prefix @code{++} and @code{--} expressions of
3492 @code{volatile}-qualified types, using simple assignments where the left
3493 operand is a @code{volatile}-qualified non-class type for their value,
3494 compound assignments where the left operand is a @code{volatile}-qualified
3495 non-class type, @code{volatile}-qualified function return type,
3496 @code{volatile}-qualified parameter type, and structured bindings of a
3497 @code{volatile}-qualified type. This usage was deprecated in C++20.
3499 Enabled by default with @option{-std=c++2a}.
3502 @node Objective-C and Objective-C++ Dialect Options
3503 @section Options Controlling Objective-C and Objective-C++ Dialects
3505 @cindex compiler options, Objective-C and Objective-C++
3506 @cindex Objective-C and Objective-C++ options, command-line
3507 @cindex options, Objective-C and Objective-C++
3508 (NOTE: This manual does not describe the Objective-C and Objective-C++
3509 languages themselves. @xref{Standards,,Language Standards
3510 Supported by GCC}, for references.)
3512 This section describes the command-line options that are only meaningful
3513 for Objective-C and Objective-C++ programs. You can also use most of
3514 the language-independent GNU compiler options.
3515 For example, you might compile a file @file{some_class.m} like this:
3518 gcc -g -fgnu-runtime -O -c some_class.m
3522 In this example, @option{-fgnu-runtime} is an option meant only for
3523 Objective-C and Objective-C++ programs; you can use the other options with
3524 any language supported by GCC@.
3526 Note that since Objective-C is an extension of the C language, Objective-C
3527 compilations may also use options specific to the C front-end (e.g.,
3528 @option{-Wtraditional}). Similarly, Objective-C++ compilations may use
3529 C++-specific options (e.g., @option{-Wabi}).
3531 Here is a list of options that are @emph{only} for compiling Objective-C
3532 and Objective-C++ programs:
3535 @item -fconstant-string-class=@var{class-name}
3536 @opindex fconstant-string-class
3537 Use @var{class-name} as the name of the class to instantiate for each
3538 literal string specified with the syntax @code{@@"@dots{}"}. The default
3539 class name is @code{NXConstantString} if the GNU runtime is being used, and
3540 @code{NSConstantString} if the NeXT runtime is being used (see below). The
3541 @option{-fconstant-cfstrings} option, if also present, overrides the
3542 @option{-fconstant-string-class} setting and cause @code{@@"@dots{}"} literals
3543 to be laid out as constant CoreFoundation strings.
3546 @opindex fgnu-runtime
3547 Generate object code compatible with the standard GNU Objective-C
3548 runtime. This is the default for most types of systems.
3550 @item -fnext-runtime
3551 @opindex fnext-runtime
3552 Generate output compatible with the NeXT runtime. This is the default
3553 for NeXT-based systems, including Darwin and Mac OS X@. The macro
3554 @code{__NEXT_RUNTIME__} is predefined if (and only if) this option is
3557 @item -fno-nil-receivers
3558 @opindex fno-nil-receivers
3559 @opindex fnil-receivers
3560 Assume that all Objective-C message dispatches (@code{[receiver
3561 message:arg]}) in this translation unit ensure that the receiver is
3562 not @code{nil}. This allows for more efficient entry points in the
3563 runtime to be used. This option is only available in conjunction with
3564 the NeXT runtime and ABI version 0 or 1.
3566 @item -fobjc-abi-version=@var{n}
3567 @opindex fobjc-abi-version
3568 Use version @var{n} of the Objective-C ABI for the selected runtime.
3569 This option is currently supported only for the NeXT runtime. In that
3570 case, Version 0 is the traditional (32-bit) ABI without support for
3571 properties and other Objective-C 2.0 additions. Version 1 is the
3572 traditional (32-bit) ABI with support for properties and other
3573 Objective-C 2.0 additions. Version 2 is the modern (64-bit) ABI. If
3574 nothing is specified, the default is Version 0 on 32-bit target
3575 machines, and Version 2 on 64-bit target machines.
3577 @item -fobjc-call-cxx-cdtors
3578 @opindex fobjc-call-cxx-cdtors
3579 For each Objective-C class, check if any of its instance variables is a
3580 C++ object with a non-trivial default constructor. If so, synthesize a
3581 special @code{- (id) .cxx_construct} instance method which runs
3582 non-trivial default constructors on any such instance variables, in order,
3583 and then return @code{self}. Similarly, check if any instance variable
3584 is a C++ object with a non-trivial destructor, and if so, synthesize a
3585 special @code{- (void) .cxx_destruct} method which runs
3586 all such default destructors, in reverse order.
3588 The @code{- (id) .cxx_construct} and @code{- (void) .cxx_destruct}
3589 methods thusly generated only operate on instance variables
3590 declared in the current Objective-C class, and not those inherited
3591 from superclasses. It is the responsibility of the Objective-C
3592 runtime to invoke all such methods in an object's inheritance
3593 hierarchy. The @code{- (id) .cxx_construct} methods are invoked
3594 by the runtime immediately after a new object instance is allocated;
3595 the @code{- (void) .cxx_destruct} methods are invoked immediately
3596 before the runtime deallocates an object instance.
3598 As of this writing, only the NeXT runtime on Mac OS X 10.4 and later has
3599 support for invoking the @code{- (id) .cxx_construct} and
3600 @code{- (void) .cxx_destruct} methods.
3602 @item -fobjc-direct-dispatch
3603 @opindex fobjc-direct-dispatch
3604 Allow fast jumps to the message dispatcher. On Darwin this is
3605 accomplished via the comm page.
3607 @item -fobjc-exceptions
3608 @opindex fobjc-exceptions
3609 Enable syntactic support for structured exception handling in
3610 Objective-C, similar to what is offered by C++. This option
3611 is required to use the Objective-C keywords @code{@@try},
3612 @code{@@throw}, @code{@@catch}, @code{@@finally} and
3613 @code{@@synchronized}. This option is available with both the GNU
3614 runtime and the NeXT runtime (but not available in conjunction with
3615 the NeXT runtime on Mac OS X 10.2 and earlier).
3619 Enable garbage collection (GC) in Objective-C and Objective-C++
3620 programs. This option is only available with the NeXT runtime; the
3621 GNU runtime has a different garbage collection implementation that
3622 does not require special compiler flags.
3624 @item -fobjc-nilcheck
3625 @opindex fobjc-nilcheck
3626 For the NeXT runtime with version 2 of the ABI, check for a nil
3627 receiver in method invocations before doing the actual method call.
3628 This is the default and can be disabled using
3629 @option{-fno-objc-nilcheck}. Class methods and super calls are never
3630 checked for nil in this way no matter what this flag is set to.
3631 Currently this flag does nothing when the GNU runtime, or an older
3632 version of the NeXT runtime ABI, is used.
3634 @item -fobjc-std=objc1
3636 Conform to the language syntax of Objective-C 1.0, the language
3637 recognized by GCC 4.0. This only affects the Objective-C additions to
3638 the C/C++ language; it does not affect conformance to C/C++ standards,
3639 which is controlled by the separate C/C++ dialect option flags. When
3640 this option is used with the Objective-C or Objective-C++ compiler,
3641 any Objective-C syntax that is not recognized by GCC 4.0 is rejected.
3642 This is useful if you need to make sure that your Objective-C code can
3643 be compiled with older versions of GCC@.
3645 @item -freplace-objc-classes
3646 @opindex freplace-objc-classes
3647 Emit a special marker instructing @command{ld(1)} not to statically link in
3648 the resulting object file, and allow @command{dyld(1)} to load it in at
3649 run time instead. This is used in conjunction with the Fix-and-Continue
3650 debugging mode, where the object file in question may be recompiled and
3651 dynamically reloaded in the course of program execution, without the need
3652 to restart the program itself. Currently, Fix-and-Continue functionality
3653 is only available in conjunction with the NeXT runtime on Mac OS X 10.3
3658 When compiling for the NeXT runtime, the compiler ordinarily replaces calls
3659 to @code{objc_getClass("@dots{}")} (when the name of the class is known at
3660 compile time) with static class references that get initialized at load time,
3661 which improves run-time performance. Specifying the @option{-fzero-link} flag
3662 suppresses this behavior and causes calls to @code{objc_getClass("@dots{}")}
3663 to be retained. This is useful in Zero-Link debugging mode, since it allows
3664 for individual class implementations to be modified during program execution.
3665 The GNU runtime currently always retains calls to @code{objc_get_class("@dots{}")}
3666 regardless of command-line options.
3668 @item -fno-local-ivars
3669 @opindex fno-local-ivars
3670 @opindex flocal-ivars
3671 By default instance variables in Objective-C can be accessed as if
3672 they were local variables from within the methods of the class they're
3673 declared in. This can lead to shadowing between instance variables
3674 and other variables declared either locally inside a class method or
3675 globally with the same name. Specifying the @option{-fno-local-ivars}
3676 flag disables this behavior thus avoiding variable shadowing issues.
3678 @item -fivar-visibility=@r{[}public@r{|}protected@r{|}private@r{|}package@r{]}
3679 @opindex fivar-visibility
3680 Set the default instance variable visibility to the specified option
3681 so that instance variables declared outside the scope of any access
3682 modifier directives default to the specified visibility.
3686 Dump interface declarations for all classes seen in the source file to a
3687 file named @file{@var{sourcename}.decl}.
3689 @item -Wassign-intercept @r{(Objective-C and Objective-C++ only)}
3690 @opindex Wassign-intercept
3691 @opindex Wno-assign-intercept
3692 Warn whenever an Objective-C assignment is being intercepted by the
3695 @item -Wno-protocol @r{(Objective-C and Objective-C++ only)}
3696 @opindex Wno-protocol
3698 If a class is declared to implement a protocol, a warning is issued for
3699 every method in the protocol that is not implemented by the class. The
3700 default behavior is to issue a warning for every method not explicitly
3701 implemented in the class, even if a method implementation is inherited
3702 from the superclass. If you use the @option{-Wno-protocol} option, then
3703 methods inherited from the superclass are considered to be implemented,
3704 and no warning is issued for them.
3706 @item -Wselector @r{(Objective-C and Objective-C++ only)}
3708 @opindex Wno-selector
3709 Warn if multiple methods of different types for the same selector are
3710 found during compilation. The check is performed on the list of methods
3711 in the final stage of compilation. Additionally, a check is performed
3712 for each selector appearing in a @code{@@selector(@dots{})}
3713 expression, and a corresponding method for that selector has been found
3714 during compilation. Because these checks scan the method table only at
3715 the end of compilation, these warnings are not produced if the final
3716 stage of compilation is not reached, for example because an error is
3717 found during compilation, or because the @option{-fsyntax-only} option is
3720 @item -Wstrict-selector-match @r{(Objective-C and Objective-C++ only)}
3721 @opindex Wstrict-selector-match
3722 @opindex Wno-strict-selector-match
3723 Warn if multiple methods with differing argument and/or return types are
3724 found for a given selector when attempting to send a message using this
3725 selector to a receiver of type @code{id} or @code{Class}. When this flag
3726 is off (which is the default behavior), the compiler omits such warnings
3727 if any differences found are confined to types that share the same size
3730 @item -Wundeclared-selector @r{(Objective-C and Objective-C++ only)}
3731 @opindex Wundeclared-selector
3732 @opindex Wno-undeclared-selector
3733 Warn if a @code{@@selector(@dots{})} expression referring to an
3734 undeclared selector is found. A selector is considered undeclared if no
3735 method with that name has been declared before the
3736 @code{@@selector(@dots{})} expression, either explicitly in an
3737 @code{@@interface} or @code{@@protocol} declaration, or implicitly in
3738 an @code{@@implementation} section. This option always performs its
3739 checks as soon as a @code{@@selector(@dots{})} expression is found,
3740 while @option{-Wselector} only performs its checks in the final stage of
3741 compilation. This also enforces the coding style convention
3742 that methods and selectors must be declared before being used.
3744 @item -print-objc-runtime-info
3745 @opindex print-objc-runtime-info
3746 Generate C header describing the largest structure that is passed by
3751 @node Diagnostic Message Formatting Options
3752 @section Options to Control Diagnostic Messages Formatting
3753 @cindex options to control diagnostics formatting
3754 @cindex diagnostic messages
3755 @cindex message formatting
3757 Traditionally, diagnostic messages have been formatted irrespective of
3758 the output device's aspect (e.g.@: its width, @dots{}). You can use the
3759 options described below
3760 to control the formatting algorithm for diagnostic messages,
3761 e.g.@: how many characters per line, how often source location
3762 information should be reported. Note that some language front ends may not
3763 honor these options.
3766 @item -fmessage-length=@var{n}
3767 @opindex fmessage-length
3768 Try to format error messages so that they fit on lines of about
3769 @var{n} characters. If @var{n} is zero, then no line-wrapping is
3770 done; each error message appears on a single line. This is the
3771 default for all front ends.
3773 Note - this option also affects the display of the @samp{#error} and
3774 @samp{#warning} pre-processor directives, and the @samp{deprecated}
3775 function/type/variable attribute. It does not however affect the
3776 @samp{pragma GCC warning} and @samp{pragma GCC error} pragmas.
3778 @item -fdiagnostics-show-location=once
3779 @opindex fdiagnostics-show-location
3780 Only meaningful in line-wrapping mode. Instructs the diagnostic messages
3781 reporter to emit source location information @emph{once}; that is, in
3782 case the message is too long to fit on a single physical line and has to
3783 be wrapped, the source location won't be emitted (as prefix) again,
3784 over and over, in subsequent continuation lines. This is the default
3787 @item -fdiagnostics-show-location=every-line
3788 Only meaningful in line-wrapping mode. Instructs the diagnostic
3789 messages reporter to emit the same source location information (as
3790 prefix) for physical lines that result from the process of breaking
3791 a message which is too long to fit on a single line.
3793 @item -fdiagnostics-color[=@var{WHEN}]
3794 @itemx -fno-diagnostics-color
3795 @opindex fdiagnostics-color
3796 @cindex highlight, color
3797 @vindex GCC_COLORS @r{environment variable}
3798 Use color in diagnostics. @var{WHEN} is @samp{never}, @samp{always},
3799 or @samp{auto}. The default depends on how the compiler has been configured,
3800 it can be any of the above @var{WHEN} options or also @samp{never}
3801 if @env{GCC_COLORS} environment variable isn't present in the environment,
3802 and @samp{auto} otherwise.
3803 @samp{auto} means to use color only when the standard error is a terminal.
3804 The forms @option{-fdiagnostics-color} and @option{-fno-diagnostics-color} are
3805 aliases for @option{-fdiagnostics-color=always} and
3806 @option{-fdiagnostics-color=never}, respectively.
3808 The colors are defined by the environment variable @env{GCC_COLORS}.
3809 Its value is a colon-separated list of capabilities and Select Graphic
3810 Rendition (SGR) substrings. SGR commands are interpreted by the
3811 terminal or terminal emulator. (See the section in the documentation
3812 of your text terminal for permitted values and their meanings as
3813 character attributes.) These substring values are integers in decimal
3814 representation and can be concatenated with semicolons.
3815 Common values to concatenate include
3817 @samp{4} for underline,
3819 @samp{7} for inverse,
3820 @samp{39} for default foreground color,
3821 @samp{30} to @samp{37} for foreground colors,
3822 @samp{90} to @samp{97} for 16-color mode foreground colors,
3823 @samp{38;5;0} to @samp{38;5;255}
3824 for 88-color and 256-color modes foreground colors,
3825 @samp{49} for default background color,
3826 @samp{40} to @samp{47} for background colors,
3827 @samp{100} to @samp{107} for 16-color mode background colors,
3828 and @samp{48;5;0} to @samp{48;5;255}
3829 for 88-color and 256-color modes background colors.
3831 The default @env{GCC_COLORS} is
3833 error=01;31:warning=01;35:note=01;36:range1=32:range2=34:locus=01:\
3834 quote=01:fixit-insert=32:fixit-delete=31:\
3835 diff-filename=01:diff-hunk=32:diff-delete=31:diff-insert=32:\
3839 where @samp{01;31} is bold red, @samp{01;35} is bold magenta,
3840 @samp{01;36} is bold cyan, @samp{32} is green, @samp{34} is blue,
3841 @samp{01} is bold, and @samp{31} is red.
3842 Setting @env{GCC_COLORS} to the empty string disables colors.
3843 Supported capabilities are as follows.
3847 @vindex error GCC_COLORS @r{capability}
3848 SGR substring for error: markers.
3851 @vindex warning GCC_COLORS @r{capability}
3852 SGR substring for warning: markers.
3855 @vindex note GCC_COLORS @r{capability}
3856 SGR substring for note: markers.
3859 @vindex range1 GCC_COLORS @r{capability}
3860 SGR substring for first additional range.
3863 @vindex range2 GCC_COLORS @r{capability}
3864 SGR substring for second additional range.
3867 @vindex locus GCC_COLORS @r{capability}
3868 SGR substring for location information, @samp{file:line} or
3869 @samp{file:line:column} etc.
3872 @vindex quote GCC_COLORS @r{capability}
3873 SGR substring for information printed within quotes.
3876 @vindex fixit-insert GCC_COLORS @r{capability}
3877 SGR substring for fix-it hints suggesting text to
3878 be inserted or replaced.
3881 @vindex fixit-delete GCC_COLORS @r{capability}
3882 SGR substring for fix-it hints suggesting text to
3885 @item diff-filename=
3886 @vindex diff-filename GCC_COLORS @r{capability}
3887 SGR substring for filename headers within generated patches.
3890 @vindex diff-hunk GCC_COLORS @r{capability}
3891 SGR substring for the starts of hunks within generated patches.
3894 @vindex diff-delete GCC_COLORS @r{capability}
3895 SGR substring for deleted lines within generated patches.
3898 @vindex diff-insert GCC_COLORS @r{capability}
3899 SGR substring for inserted lines within generated patches.
3902 @vindex type-diff GCC_COLORS @r{capability}
3903 SGR substring for highlighting mismatching types within template
3904 arguments in the C++ frontend.
3907 @item -fno-diagnostics-show-option
3908 @opindex fno-diagnostics-show-option
3909 @opindex fdiagnostics-show-option
3910 By default, each diagnostic emitted includes text indicating the
3911 command-line option that directly controls the diagnostic (if such an
3912 option is known to the diagnostic machinery). Specifying the
3913 @option{-fno-diagnostics-show-option} flag suppresses that behavior.
3915 @item -fno-diagnostics-show-caret
3916 @opindex fno-diagnostics-show-caret
3917 @opindex fdiagnostics-show-caret
3918 By default, each diagnostic emitted includes the original source line
3919 and a caret @samp{^} indicating the column. This option suppresses this
3920 information. The source line is truncated to @var{n} characters, if
3921 the @option{-fmessage-length=n} option is given. When the output is done
3922 to the terminal, the width is limited to the width given by the
3923 @env{COLUMNS} environment variable or, if not set, to the terminal width.
3925 @item -fno-diagnostics-show-labels
3926 @opindex fno-diagnostics-show-labels
3927 @opindex fdiagnostics-show-labels
3928 By default, when printing source code (via @option{-fdiagnostics-show-caret}),
3929 diagnostics can label ranges of source code with pertinent information, such
3930 as the types of expressions:
3933 printf ("foo %s bar", long_i + long_j);
3939 This option suppresses the printing of these labels (in the example above,
3940 the vertical bars and the ``char *'' and ``long int'' text).
3942 @item -fno-diagnostics-show-line-numbers
3943 @opindex fno-diagnostics-show-line-numbers
3944 @opindex fdiagnostics-show-line-numbers
3945 By default, when printing source code (via @option{-fdiagnostics-show-caret}),
3946 a left margin is printed, showing line numbers. This option suppresses this
3949 @item -fdiagnostics-minimum-margin-width=@var{width}
3950 @opindex fdiagnostics-minimum-margin-width
3951 This option controls the minimum width of the left margin printed by
3952 @option{-fdiagnostics-show-line-numbers}. It defaults to 6.
3954 @item -fdiagnostics-parseable-fixits
3955 @opindex fdiagnostics-parseable-fixits
3956 Emit fix-it hints in a machine-parseable format, suitable for consumption
3957 by IDEs. For each fix-it, a line will be printed after the relevant
3958 diagnostic, starting with the string ``fix-it:''. For example:
3961 fix-it:"test.c":@{45:3-45:21@}:"gtk_widget_show_all"
3964 The location is expressed as a half-open range, expressed as a count of
3965 bytes, starting at byte 1 for the initial column. In the above example,
3966 bytes 3 through 20 of line 45 of ``test.c'' are to be replaced with the
3970 00000000011111111112222222222
3971 12345678901234567890123456789
3972 gtk_widget_showall (dlg);
3977 The filename and replacement string escape backslash as ``\\", tab as ``\t'',
3978 newline as ``\n'', double quotes as ``\"'', non-printable characters as octal
3979 (e.g. vertical tab as ``\013'').
3981 An empty replacement string indicates that the given range is to be removed.
3982 An empty range (e.g. ``45:3-45:3'') indicates that the string is to
3983 be inserted at the given position.
3985 @item -fdiagnostics-generate-patch
3986 @opindex fdiagnostics-generate-patch
3987 Print fix-it hints to stderr in unified diff format, after any diagnostics
3988 are printed. For example:
3995 void show_cb(GtkDialog *dlg)
3997 - gtk_widget_showall(dlg);
3998 + gtk_widget_show_all(dlg);
4003 The diff may or may not be colorized, following the same rules
4004 as for diagnostics (see @option{-fdiagnostics-color}).
4006 @item -fdiagnostics-show-template-tree
4007 @opindex fdiagnostics-show-template-tree
4009 In the C++ frontend, when printing diagnostics showing mismatching
4010 template types, such as:
4013 could not convert 'std::map<int, std::vector<double> >()'
4014 from 'map<[...],vector<double>>' to 'map<[...],vector<float>>
4017 the @option{-fdiagnostics-show-template-tree} flag enables printing a
4018 tree-like structure showing the common and differing parts of the types,
4028 The parts that differ are highlighted with color (``double'' and
4029 ``float'' in this case).
4031 @item -fno-elide-type
4032 @opindex fno-elide-type
4033 @opindex felide-type
4034 By default when the C++ frontend prints diagnostics showing mismatching
4035 template types, common parts of the types are printed as ``[...]'' to
4036 simplify the error message. For example:
4039 could not convert 'std::map<int, std::vector<double> >()'
4040 from 'map<[...],vector<double>>' to 'map<[...],vector<float>>
4043 Specifying the @option{-fno-elide-type} flag suppresses that behavior.
4044 This flag also affects the output of the
4045 @option{-fdiagnostics-show-template-tree} flag.
4047 @item -fno-show-column
4048 @opindex fno-show-column
4049 @opindex fshow-column
4050 Do not print column numbers in diagnostics. This may be necessary if
4051 diagnostics are being scanned by a program that does not understand the
4052 column numbers, such as @command{dejagnu}.
4054 @item -fdiagnostics-format=@var{FORMAT}
4055 @opindex fdiagnostics-format
4056 Select a different format for printing diagnostics.
4057 @var{FORMAT} is @samp{text} or @samp{json}.
4058 The default is @samp{text}.
4060 The @samp{json} format consists of a top-level JSON array containing JSON
4061 objects representing the diagnostics.
4063 The JSON is emitted as one line, without formatting; the examples below
4064 have been formatted for clarity.
4066 Diagnostics can have child diagnostics. For example, this error and note:
4069 misleading-indentation.c:15:3: warning: this 'if' clause does not
4070 guard... [-Wmisleading-indentation]
4073 misleading-indentation.c:17:5: note: ...this statement, but the latter
4074 is misleadingly indented as if it were guarded by the 'if'
4080 might be printed in JSON form (after formatting) like this:
4090 "file": "misleading-indentation.c",
4095 "file": "misleading-indentation.c",
4100 "message": "this \u2018if\u2019 clause does not guard...",
4101 "option": "-Wmisleading-indentation",
4109 "file": "misleading-indentation.c",
4114 "message": "...this statement, but the latter is @dots{}"
4123 where the @code{note} is a child of the @code{warning}.
4125 A diagnostic has a @code{kind}. If this is @code{warning}, then there is
4126 an @code{option} key describing the command-line option controlling the
4129 A diagnostic can contain zero or more locations. Each location has up
4130 to three positions within it: a @code{caret} position and optional
4131 @code{start} and @code{finish} positions. A location can also have
4132 an optional @code{label} string. For example, this error:
4135 bad-binary-ops.c:64:23: error: invalid operands to binary + (have 'S' @{aka
4136 'struct s'@} and 'T' @{aka 'struct t'@})
4137 64 | return callee_4a () + callee_4b ();
4138 | ~~~~~~~~~~~~ ^ ~~~~~~~~~~~~
4140 | | T @{aka struct t@}
4141 | S @{aka struct s@}
4145 has three locations. Its primary location is at the ``+'' token at column
4146 23. It has two secondary locations, describing the left and right-hand sides
4147 of the expression, which have labels. It might be printed in JSON form as:
4156 "column": 23, "file": "bad-binary-ops.c", "line": 64
4161 "column": 10, "file": "bad-binary-ops.c", "line": 64
4164 "column": 21, "file": "bad-binary-ops.c", "line": 64
4166 "label": "S @{aka struct s@}"
4170 "column": 25, "file": "bad-binary-ops.c", "line": 64
4173 "column": 36, "file": "bad-binary-ops.c", "line": 64
4175 "label": "T @{aka struct t@}"
4178 "message": "invalid operands to binary + @dots{}"
4182 If a diagnostic contains fix-it hints, it has a @code{fixits} array,
4183 consisting of half-open intervals, similar to the output of
4184 @option{-fdiagnostics-parseable-fixits}. For example, this diagnostic
4185 with a replacement fix-it hint:
4188 demo.c:8:15: error: 'struct s' has no member named 'colour'; did you
4190 8 | return ptr->colour;
4196 might be printed in JSON form as:
4231 "message": "\u2018struct s\u2019 has no member named @dots{}"
4236 where the fix-it hint suggests replacing the text from @code{start} up
4237 to but not including @code{next} with @code{string}'s value. Deletions
4238 are expressed via an empty value for @code{string}, insertions by
4239 having @code{start} equal @code{next}.
4243 @node Warning Options
4244 @section Options to Request or Suppress Warnings
4245 @cindex options to control warnings
4246 @cindex warning messages
4247 @cindex messages, warning
4248 @cindex suppressing warnings
4250 Warnings are diagnostic messages that report constructions that
4251 are not inherently erroneous but that are risky or suggest there
4252 may have been an error.
4254 The following language-independent options do not enable specific
4255 warnings but control the kinds of diagnostics produced by GCC@.
4258 @cindex syntax checking
4260 @opindex fsyntax-only
4261 Check the code for syntax errors, but don't do anything beyond that.
4263 @item -fmax-errors=@var{n}
4264 @opindex fmax-errors
4265 Limits the maximum number of error messages to @var{n}, at which point
4266 GCC bails out rather than attempting to continue processing the source
4267 code. If @var{n} is 0 (the default), there is no limit on the number
4268 of error messages produced. If @option{-Wfatal-errors} is also
4269 specified, then @option{-Wfatal-errors} takes precedence over this
4274 Inhibit all warning messages.
4279 Make all warnings into errors.
4284 Make the specified warning into an error. The specifier for a warning
4285 is appended; for example @option{-Werror=switch} turns the warnings
4286 controlled by @option{-Wswitch} into errors. This switch takes a
4287 negative form, to be used to negate @option{-Werror} for specific
4288 warnings; for example @option{-Wno-error=switch} makes
4289 @option{-Wswitch} warnings not be errors, even when @option{-Werror}
4292 The warning message for each controllable warning includes the
4293 option that controls the warning. That option can then be used with
4294 @option{-Werror=} and @option{-Wno-error=} as described above.
4295 (Printing of the option in the warning message can be disabled using the
4296 @option{-fno-diagnostics-show-option} flag.)
4298 Note that specifying @option{-Werror=}@var{foo} automatically implies
4299 @option{-W}@var{foo}. However, @option{-Wno-error=}@var{foo} does not
4302 @item -Wfatal-errors
4303 @opindex Wfatal-errors
4304 @opindex Wno-fatal-errors
4305 This option causes the compiler to abort compilation on the first error
4306 occurred rather than trying to keep going and printing further error
4311 You can request many specific warnings with options beginning with
4312 @samp{-W}, for example @option{-Wimplicit} to request warnings on
4313 implicit declarations. Each of these specific warning options also
4314 has a negative form beginning @samp{-Wno-} to turn off warnings; for
4315 example, @option{-Wno-implicit}. This manual lists only one of the
4316 two forms, whichever is not the default. For further
4317 language-specific options also refer to @ref{C++ Dialect Options} and
4318 @ref{Objective-C and Objective-C++ Dialect Options}.
4320 Some options, such as @option{-Wall} and @option{-Wextra}, turn on other
4321 options, such as @option{-Wunused}, which may turn on further options,
4322 such as @option{-Wunused-value}. The combined effect of positive and
4323 negative forms is that more specific options have priority over less
4324 specific ones, independently of their position in the command-line. For
4325 options of the same specificity, the last one takes effect. Options
4326 enabled or disabled via pragmas (@pxref{Diagnostic Pragmas}) take effect
4327 as if they appeared at the end of the command-line.
4329 When an unrecognized warning option is requested (e.g.,
4330 @option{-Wunknown-warning}), GCC emits a diagnostic stating
4331 that the option is not recognized. However, if the @option{-Wno-} form
4332 is used, the behavior is slightly different: no diagnostic is
4333 produced for @option{-Wno-unknown-warning} unless other diagnostics
4334 are being produced. This allows the use of new @option{-Wno-} options
4335 with old compilers, but if something goes wrong, the compiler
4336 warns that an unrecognized option is present.
4343 @opindex Wno-pedantic
4344 Issue all the warnings demanded by strict ISO C and ISO C++;
4345 reject all programs that use forbidden extensions, and some other
4346 programs that do not follow ISO C and ISO C++. For ISO C, follows the
4347 version of the ISO C standard specified by any @option{-std} option used.
4349 Valid ISO C and ISO C++ programs should compile properly with or without
4350 this option (though a rare few require @option{-ansi} or a
4351 @option{-std} option specifying the required version of ISO C)@. However,
4352 without this option, certain GNU extensions and traditional C and C++
4353 features are supported as well. With this option, they are rejected.
4355 @option{-Wpedantic} does not cause warning messages for use of the
4356 alternate keywords whose names begin and end with @samp{__}. This alternate
4357 format can also be used to disable warnings for non-ISO @samp{__intN} types,
4358 i.e. @samp{__intN__}.
4359 Pedantic warnings are also disabled in the expression that follows
4360 @code{__extension__}. However, only system header files should use
4361 these escape routes; application programs should avoid them.
4362 @xref{Alternate Keywords}.
4364 Some users try to use @option{-Wpedantic} to check programs for strict ISO
4365 C conformance. They soon find that it does not do quite what they want:
4366 it finds some non-ISO practices, but not all---only those for which
4367 ISO C @emph{requires} a diagnostic, and some others for which
4368 diagnostics have been added.
4370 A feature to report any failure to conform to ISO C might be useful in
4371 some instances, but would require considerable additional work and would
4372 be quite different from @option{-Wpedantic}. We don't have plans to
4373 support such a feature in the near future.
4375 Where the standard specified with @option{-std} represents a GNU
4376 extended dialect of C, such as @samp{gnu90} or @samp{gnu99}, there is a
4377 corresponding @dfn{base standard}, the version of ISO C on which the GNU
4378 extended dialect is based. Warnings from @option{-Wpedantic} are given
4379 where they are required by the base standard. (It does not make sense
4380 for such warnings to be given only for features not in the specified GNU
4381 C dialect, since by definition the GNU dialects of C include all
4382 features the compiler supports with the given option, and there would be
4383 nothing to warn about.)
4385 @item -pedantic-errors
4386 @opindex pedantic-errors
4387 Give an error whenever the @dfn{base standard} (see @option{-Wpedantic})
4388 requires a diagnostic, in some cases where there is undefined behavior
4389 at compile-time and in some other cases that do not prevent compilation
4390 of programs that are valid according to the standard. This is not
4391 equivalent to @option{-Werror=pedantic}, since there are errors enabled
4392 by this option and not enabled by the latter and vice versa.
4397 This enables all the warnings about constructions that some users
4398 consider questionable, and that are easy to avoid (or modify to
4399 prevent the warning), even in conjunction with macros. This also
4400 enables some language-specific warnings described in @ref{C++ Dialect
4401 Options} and @ref{Objective-C and Objective-C++ Dialect Options}.
4403 @option{-Wall} turns on the following warning flags:
4405 @gccoptlist{-Waddress @gol
4406 -Warray-bounds=1 @r{(only with} @option{-O2}@r{)} @gol
4408 -Wbool-operation @gol
4409 -Wc++11-compat -Wc++14-compat @gol
4410 -Wcatch-value @r{(C++ and Objective-C++ only)} @gol
4411 -Wchar-subscripts @gol
4413 -Wduplicate-decl-specifier @r{(C and Objective-C only)} @gol
4414 -Wenum-compare @r{(in C/ObjC; this is on by default in C++)} @gol
4415 -Wenum-conversion @r{in C/ObjC;} @gol
4417 -Wint-in-bool-context @gol
4418 -Wimplicit @r{(C and Objective-C only)} @gol
4419 -Wimplicit-int @r{(C and Objective-C only)} @gol
4420 -Wimplicit-function-declaration @r{(C and Objective-C only)} @gol
4421 -Winit-self @r{(only for C++)} @gol
4422 -Wlogical-not-parentheses @gol
4423 -Wmain @r{(only for C/ObjC and unless} @option{-ffreestanding}@r{)} @gol
4424 -Wmaybe-uninitialized @gol
4425 -Wmemset-elt-size @gol
4426 -Wmemset-transposed-args @gol
4427 -Wmisleading-indentation @r{(only for C/C++)} @gol
4428 -Wmissing-attributes @gol
4429 -Wmissing-braces @r{(only for C/ObjC)} @gol
4430 -Wmultistatement-macros @gol
4431 -Wnarrowing @r{(only for C++)} @gol
4433 -Wnonnull-compare @gol
4436 -Wpessimizing-move @r{(only for C++)} @gol
4441 -Wsequence-point @gol
4442 -Wsign-compare @r{(only in C++)} @gol
4443 -Wsizeof-pointer-div @gol
4444 -Wsizeof-pointer-memaccess @gol
4445 -Wstrict-aliasing @gol
4446 -Wstrict-overflow=1 @gol
4448 -Wtautological-compare @gol
4450 -Wuninitialized @gol
4451 -Wunknown-pragmas @gol
4452 -Wunused-function @gol
4455 -Wunused-variable @gol
4456 -Wvolatile-register-var}
4458 Note that some warning flags are not implied by @option{-Wall}. Some of
4459 them warn about constructions that users generally do not consider
4460 questionable, but which occasionally you might wish to check for;
4461 others warn about constructions that are necessary or hard to avoid in
4462 some cases, and there is no simple way to modify the code to suppress
4463 the warning. Some of them are enabled by @option{-Wextra} but many of
4464 them must be enabled individually.
4470 This enables some extra warning flags that are not enabled by
4471 @option{-Wall}. (This option used to be called @option{-W}. The older
4472 name is still supported, but the newer name is more descriptive.)
4474 @gccoptlist{-Wclobbered @gol
4475 -Wcast-function-type @gol
4476 -Wdeprecated-copy @r{(C++ only)} @gol
4478 -Wignored-qualifiers @gol
4479 -Wimplicit-fallthrough=3 @gol
4480 -Wmissing-field-initializers @gol
4481 -Wmissing-parameter-type @r{(C only)} @gol
4482 -Wold-style-declaration @r{(C only)} @gol
4483 -Woverride-init @gol
4484 -Wsign-compare @r{(C only)} @gol
4485 -Wredundant-move @r{(only for C++)} @gol
4487 -Wuninitialized @gol
4488 -Wshift-negative-value @r{(in C++03 and in C99 and newer)} @gol
4489 -Wunused-parameter @r{(only with} @option{-Wunused} @r{or} @option{-Wall}@r{)} @gol
4490 -Wunused-but-set-parameter @r{(only with} @option{-Wunused} @r{or} @option{-Wall}@r{)}}
4493 The option @option{-Wextra} also prints warning messages for the
4499 A pointer is compared against integer zero with @code{<}, @code{<=},
4500 @code{>}, or @code{>=}.
4503 (C++ only) An enumerator and a non-enumerator both appear in a
4504 conditional expression.
4507 (C++ only) Ambiguous virtual bases.
4510 (C++ only) Subscripting an array that has been declared @code{register}.
4513 (C++ only) Taking the address of a variable that has been declared
4517 (C++ only) A base class is not initialized in the copy constructor
4522 @item -Wchar-subscripts
4523 @opindex Wchar-subscripts
4524 @opindex Wno-char-subscripts
4525 Warn if an array subscript has type @code{char}. This is a common cause
4526 of error, as programmers often forget that this type is signed on some
4528 This warning is enabled by @option{-Wall}.
4530 @item -Wno-coverage-mismatch
4531 @opindex Wno-coverage-mismatch
4532 @opindex Wcoverage-mismatch
4533 Warn if feedback profiles do not match when using the
4534 @option{-fprofile-use} option.
4535 If a source file is changed between compiling with @option{-fprofile-generate}
4536 and with @option{-fprofile-use}, the files with the profile feedback can fail
4537 to match the source file and GCC cannot use the profile feedback
4538 information. By default, this warning is enabled and is treated as an
4539 error. @option{-Wno-coverage-mismatch} can be used to disable the
4540 warning or @option{-Wno-error=coverage-mismatch} can be used to
4541 disable the error. Disabling the error for this warning can result in
4542 poorly optimized code and is useful only in the
4543 case of very minor changes such as bug fixes to an existing code-base.
4544 Completely disabling the warning is not recommended.
4547 @r{(C, Objective-C, C++, Objective-C++ and Fortran only)}
4549 Suppress warning messages emitted by @code{#warning} directives.
4551 @item -Wdouble-promotion @r{(C, C++, Objective-C and Objective-C++ only)}
4552 @opindex Wdouble-promotion
4553 @opindex Wno-double-promotion
4554 Give a warning when a value of type @code{float} is implicitly
4555 promoted to @code{double}. CPUs with a 32-bit ``single-precision''
4556 floating-point unit implement @code{float} in hardware, but emulate
4557 @code{double} in software. On such a machine, doing computations
4558 using @code{double} values is much more expensive because of the
4559 overhead required for software emulation.
4561 It is easy to accidentally do computations with @code{double} because
4562 floating-point literals are implicitly of type @code{double}. For
4566 float area(float radius)
4568 return 3.14159 * radius * radius;
4572 the compiler performs the entire computation with @code{double}
4573 because the floating-point literal is a @code{double}.
4575 @item -Wduplicate-decl-specifier @r{(C and Objective-C only)}
4576 @opindex Wduplicate-decl-specifier
4577 @opindex Wno-duplicate-decl-specifier
4578 Warn if a declaration has duplicate @code{const}, @code{volatile},
4579 @code{restrict} or @code{_Atomic} specifier. This warning is enabled by
4583 @itemx -Wformat=@var{n}
4586 @opindex ffreestanding
4587 @opindex fno-builtin
4589 Check calls to @code{printf} and @code{scanf}, etc., to make sure that
4590 the arguments supplied have types appropriate to the format string
4591 specified, and that the conversions specified in the format string make
4592 sense. This includes standard functions, and others specified by format
4593 attributes (@pxref{Function Attributes}), in the @code{printf},
4594 @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension,
4595 not in the C standard) families (or other target-specific families).
4596 Which functions are checked without format attributes having been
4597 specified depends on the standard version selected, and such checks of
4598 functions without the attribute specified are disabled by
4599 @option{-ffreestanding} or @option{-fno-builtin}.
4601 The formats are checked against the format features supported by GNU
4602 libc version 2.2. These include all ISO C90 and C99 features, as well
4603 as features from the Single Unix Specification and some BSD and GNU
4604 extensions. Other library implementations may not support all these
4605 features; GCC does not support warning about features that go beyond a
4606 particular library's limitations. However, if @option{-Wpedantic} is used
4607 with @option{-Wformat}, warnings are given about format features not
4608 in the selected standard version (but not for @code{strfmon} formats,
4609 since those are not in any version of the C standard). @xref{C Dialect
4610 Options,,Options Controlling C Dialect}.
4617 Option @option{-Wformat} is equivalent to @option{-Wformat=1}, and
4618 @option{-Wno-format} is equivalent to @option{-Wformat=0}. Since
4619 @option{-Wformat} also checks for null format arguments for several
4620 functions, @option{-Wformat} also implies @option{-Wnonnull}. Some
4621 aspects of this level of format checking can be disabled by the
4622 options: @option{-Wno-format-contains-nul},
4623 @option{-Wno-format-extra-args}, and @option{-Wno-format-zero-length}.
4624 @option{-Wformat} is enabled by @option{-Wall}.
4626 @item -Wno-format-contains-nul
4627 @opindex Wno-format-contains-nul
4628 @opindex Wformat-contains-nul
4629 If @option{-Wformat} is specified, do not warn about format strings that
4632 @item -Wno-format-extra-args
4633 @opindex Wno-format-extra-args
4634 @opindex Wformat-extra-args
4635 If @option{-Wformat} is specified, do not warn about excess arguments to a
4636 @code{printf} or @code{scanf} format function. The C standard specifies
4637 that such arguments are ignored.
4639 Where the unused arguments lie between used arguments that are
4640 specified with @samp{$} operand number specifications, normally
4641 warnings are still given, since the implementation could not know what
4642 type to pass to @code{va_arg} to skip the unused arguments. However,
4643 in the case of @code{scanf} formats, this option suppresses the
4644 warning if the unused arguments are all pointers, since the Single
4645 Unix Specification says that such unused arguments are allowed.
4647 @item -Wformat-overflow
4648 @itemx -Wformat-overflow=@var{level}
4649 @opindex Wformat-overflow
4650 @opindex Wno-format-overflow
4651 Warn about calls to formatted input/output functions such as @code{sprintf}
4652 and @code{vsprintf} that might overflow the destination buffer. When the
4653 exact number of bytes written by a format directive cannot be determined
4654 at compile-time it is estimated based on heuristics that depend on the
4655 @var{level} argument and on optimization. While enabling optimization
4656 will in most cases improve the accuracy of the warning, it may also
4657 result in false positives.
4660 @item -Wformat-overflow
4661 @itemx -Wformat-overflow=1
4662 @opindex Wformat-overflow
4663 @opindex Wno-format-overflow
4664 Level @var{1} of @option{-Wformat-overflow} enabled by @option{-Wformat}
4665 employs a conservative approach that warns only about calls that most
4666 likely overflow the buffer. At this level, numeric arguments to format
4667 directives with unknown values are assumed to have the value of one, and
4668 strings of unknown length to be empty. Numeric arguments that are known
4669 to be bounded to a subrange of their type, or string arguments whose output
4670 is bounded either by their directive's precision or by a finite set of
4671 string literals, are assumed to take on the value within the range that
4672 results in the most bytes on output. For example, the call to @code{sprintf}
4673 below is diagnosed because even with both @var{a} and @var{b} equal to zero,
4674 the terminating NUL character (@code{'\0'}) appended by the function
4675 to the destination buffer will be written past its end. Increasing
4676 the size of the buffer by a single byte is sufficient to avoid the
4677 warning, though it may not be sufficient to avoid the overflow.
4680 void f (int a, int b)
4683 sprintf (buf, "a = %i, b = %i\n", a, b);
4687 @item -Wformat-overflow=2
4688 Level @var{2} warns also about calls that might overflow the destination
4689 buffer given an argument of sufficient length or magnitude. At level
4690 @var{2}, unknown numeric arguments are assumed to have the minimum
4691 representable value for signed types with a precision greater than 1, and
4692 the maximum representable value otherwise. Unknown string arguments whose
4693 length cannot be assumed to be bounded either by the directive's precision,
4694 or by a finite set of string literals they may evaluate to, or the character
4695 array they may point to, are assumed to be 1 character long.
4697 At level @var{2}, the call in the example above is again diagnosed, but
4698 this time because with @var{a} equal to a 32-bit @code{INT_MIN} the first
4699 @code{%i} directive will write some of its digits beyond the end of
4700 the destination buffer. To make the call safe regardless of the values
4701 of the two variables, the size of the destination buffer must be increased
4702 to at least 34 bytes. GCC includes the minimum size of the buffer in
4703 an informational note following the warning.
4705 An alternative to increasing the size of the destination buffer is to
4706 constrain the range of formatted values. The maximum length of string
4707 arguments can be bounded by specifying the precision in the format
4708 directive. When numeric arguments of format directives can be assumed
4709 to be bounded by less than the precision of their type, choosing
4710 an appropriate length modifier to the format specifier will reduce
4711 the required buffer size. For example, if @var{a} and @var{b} in the
4712 example above can be assumed to be within the precision of
4713 the @code{short int} type then using either the @code{%hi} format
4714 directive or casting the argument to @code{short} reduces the maximum
4715 required size of the buffer to 24 bytes.
4718 void f (int a, int b)
4721 sprintf (buf, "a = %hi, b = %i\n", a, (short)b);
4726 @item -Wno-format-zero-length
4727 @opindex Wno-format-zero-length
4728 @opindex Wformat-zero-length
4729 If @option{-Wformat} is specified, do not warn about zero-length formats.
4730 The C standard specifies that zero-length formats are allowed.
4735 Enable @option{-Wformat} plus additional format checks. Currently
4736 equivalent to @option{-Wformat -Wformat-nonliteral -Wformat-security
4739 @item -Wformat-nonliteral
4740 @opindex Wformat-nonliteral
4741 @opindex Wno-format-nonliteral
4742 If @option{-Wformat} is specified, also warn if the format string is not a
4743 string literal and so cannot be checked, unless the format function
4744 takes its format arguments as a @code{va_list}.
4746 @item -Wformat-security
4747 @opindex Wformat-security
4748 @opindex Wno-format-security
4749 If @option{-Wformat} is specified, also warn about uses of format
4750 functions that represent possible security problems. At present, this
4751 warns about calls to @code{printf} and @code{scanf} functions where the
4752 format string is not a string literal and there are no format arguments,
4753 as in @code{printf (foo);}. This may be a security hole if the format
4754 string came from untrusted input and contains @samp{%n}. (This is
4755 currently a subset of what @option{-Wformat-nonliteral} warns about, but
4756 in future warnings may be added to @option{-Wformat-security} that are not
4757 included in @option{-Wformat-nonliteral}.)
4759 @item -Wformat-signedness
4760 @opindex Wformat-signedness
4761 @opindex Wno-format-signedness
4762 If @option{-Wformat} is specified, also warn if the format string
4763 requires an unsigned argument and the argument is signed and vice versa.
4765 @item -Wformat-truncation
4766 @itemx -Wformat-truncation=@var{level}
4767 @opindex Wformat-truncation
4768 @opindex Wno-format-truncation
4769 Warn about calls to formatted input/output functions such as @code{snprintf}
4770 and @code{vsnprintf} that might result in output truncation. When the exact
4771 number of bytes written by a format directive cannot be determined at
4772 compile-time it is estimated based on heuristics that depend on
4773 the @var{level} argument and on optimization. While enabling optimization
4774 will in most cases improve the accuracy of the warning, it may also result
4775 in false positives. Except as noted otherwise, the option uses the same
4776 logic @option{-Wformat-overflow}.
4779 @item -Wformat-truncation
4780 @itemx -Wformat-truncation=1
4781 @opindex Wformat-truncation
4782 @opindex Wno-format-truncation
4783 Level @var{1} of @option{-Wformat-truncation} enabled by @option{-Wformat}
4784 employs a conservative approach that warns only about calls to bounded
4785 functions whose return value is unused and that will most likely result
4786 in output truncation.
4788 @item -Wformat-truncation=2
4789 Level @var{2} warns also about calls to bounded functions whose return
4790 value is used and that might result in truncation given an argument of
4791 sufficient length or magnitude.
4795 @opindex Wformat-y2k
4796 @opindex Wno-format-y2k
4797 If @option{-Wformat} is specified, also warn about @code{strftime}
4798 formats that may yield only a two-digit year.
4803 @opindex Wno-nonnull
4804 Warn about passing a null pointer for arguments marked as
4805 requiring a non-null value by the @code{nonnull} function attribute.
4807 @option{-Wnonnull} is included in @option{-Wall} and @option{-Wformat}. It
4808 can be disabled with the @option{-Wno-nonnull} option.
4810 @item -Wnonnull-compare
4811 @opindex Wnonnull-compare
4812 @opindex Wno-nonnull-compare
4813 Warn when comparing an argument marked with the @code{nonnull}
4814 function attribute against null inside the function.
4816 @option{-Wnonnull-compare} is included in @option{-Wall}. It
4817 can be disabled with the @option{-Wno-nonnull-compare} option.
4819 @item -Wnull-dereference
4820 @opindex Wnull-dereference
4821 @opindex Wno-null-dereference
4822 Warn if the compiler detects paths that trigger erroneous or
4823 undefined behavior due to dereferencing a null pointer. This option
4824 is only active when @option{-fdelete-null-pointer-checks} is active,
4825 which is enabled by optimizations in most targets. The precision of
4826 the warnings depends on the optimization options used.
4828 @item -Winaccessible-base @r{(C++, Objective-C++ only)}
4829 @opindex Winaccessible-base
4830 @opindex Wno-inaccessible-base
4831 Warn when a base class is inaccessible in a class derived from it due to
4832 ambiguity. The warning is enabled by default. Note the warning for virtual
4833 bases is enabled by the @option{-Wextra} option.
4836 struct A @{ int a; @};
4840 struct C : B, A @{ @};
4844 @item -Winit-self @r{(C, C++, Objective-C and Objective-C++ only)}
4846 @opindex Wno-init-self
4847 Warn about uninitialized variables that are initialized with themselves.
4848 Note this option can only be used with the @option{-Wuninitialized} option.
4850 For example, GCC warns about @code{i} being uninitialized in the
4851 following snippet only when @option{-Winit-self} has been specified:
4862 This warning is enabled by @option{-Wall} in C++.
4864 @item -Wimplicit-int @r{(C and Objective-C only)}
4865 @opindex Wimplicit-int
4866 @opindex Wno-implicit-int
4867 Warn when a declaration does not specify a type.
4868 This warning is enabled by @option{-Wall}.
4870 @item -Wimplicit-function-declaration @r{(C and Objective-C only)}
4871 @opindex Wimplicit-function-declaration
4872 @opindex Wno-implicit-function-declaration
4873 Give a warning whenever a function is used before being declared. In
4874 C99 mode (@option{-std=c99} or @option{-std=gnu99}), this warning is
4875 enabled by default and it is made into an error by
4876 @option{-pedantic-errors}. This warning is also enabled by
4879 @item -Wimplicit @r{(C and Objective-C only)}
4881 @opindex Wno-implicit
4882 Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}.
4883 This warning is enabled by @option{-Wall}.
4885 @item -Wimplicit-fallthrough
4886 @opindex Wimplicit-fallthrough
4887 @opindex Wno-implicit-fallthrough
4888 @option{-Wimplicit-fallthrough} is the same as @option{-Wimplicit-fallthrough=3}
4889 and @option{-Wno-implicit-fallthrough} is the same as
4890 @option{-Wimplicit-fallthrough=0}.
4892 @item -Wimplicit-fallthrough=@var{n}
4893 @opindex Wimplicit-fallthrough=
4894 Warn when a switch case falls through. For example:
4912 This warning does not warn when the last statement of a case cannot
4913 fall through, e.g. when there is a return statement or a call to function
4914 declared with the noreturn attribute. @option{-Wimplicit-fallthrough=}
4915 also takes into account control flow statements, such as ifs, and only
4916 warns when appropriate. E.g.@:
4926 @} else if (i < 1) @{
4936 Since there are occasions where a switch case fall through is desirable,
4937 GCC provides an attribute, @code{__attribute__ ((fallthrough))}, that is
4938 to be used along with a null statement to suppress this warning that
4939 would normally occur:
4947 __attribute__ ((fallthrough));
4954 C++17 provides a standard way to suppress the @option{-Wimplicit-fallthrough}
4955 warning using @code{[[fallthrough]];} instead of the GNU attribute. In C++11
4956 or C++14 users can use @code{[[gnu::fallthrough]];}, which is a GNU extension.
4957 Instead of these attributes, it is also possible to add a fallthrough comment
4958 to silence the warning. The whole body of the C or C++ style comment should
4959 match the given regular expressions listed below. The option argument @var{n}
4960 specifies what kind of comments are accepted:
4964 @item @option{-Wimplicit-fallthrough=0} disables the warning altogether.
4966 @item @option{-Wimplicit-fallthrough=1} matches @code{.*} regular
4967 expression, any comment is used as fallthrough comment.
4969 @item @option{-Wimplicit-fallthrough=2} case insensitively matches
4970 @code{.*falls?[ \t-]*thr(ough|u).*} regular expression.
4972 @item @option{-Wimplicit-fallthrough=3} case sensitively matches one of the
4973 following regular expressions:
4977 @item @code{-fallthrough}
4979 @item @code{@@fallthrough@@}
4981 @item @code{lint -fallthrough[ \t]*}
4983 @item @code{[ \t.!]*(ELSE,? |INTENTIONAL(LY)? )?@*FALL(S | |-)?THR(OUGH|U)[ \t.!]*(-[^\n\r]*)?}
4985 @item @code{[ \t.!]*(Else,? |Intentional(ly)? )?@*Fall((s | |-)[Tt]|t)hr(ough|u)[ \t.!]*(-[^\n\r]*)?}
4987 @item @code{[ \t.!]*([Ee]lse,? |[Ii]ntentional(ly)? )?@*fall(s | |-)?thr(ough|u)[ \t.!]*(-[^\n\r]*)?}
4991 @item @option{-Wimplicit-fallthrough=4} case sensitively matches one of the
4992 following regular expressions:
4996 @item @code{-fallthrough}
4998 @item @code{@@fallthrough@@}
5000 @item @code{lint -fallthrough[ \t]*}
5002 @item @code{[ \t]*FALLTHR(OUGH|U)[ \t]*}
5006 @item @option{-Wimplicit-fallthrough=5} doesn't recognize any comments as
5007 fallthrough comments, only attributes disable the warning.
5011 The comment needs to be followed after optional whitespace and other comments
5012 by @code{case} or @code{default} keywords or by a user label that precedes some
5013 @code{case} or @code{default} label.
5028 The @option{-Wimplicit-fallthrough=3} warning is enabled by @option{-Wextra}.
5030 @item -Wif-not-aligned @r{(C, C++, Objective-C and Objective-C++ only)}
5031 @opindex Wif-not-aligned
5032 @opindex Wno-if-not-aligned
5033 Control if warning triggered by the @code{warn_if_not_aligned} attribute
5034 should be issued. This is enabled by default.
5035 Use @option{-Wno-if-not-aligned} to disable it.
5037 @item -Wignored-qualifiers @r{(C and C++ only)}
5038 @opindex Wignored-qualifiers
5039 @opindex Wno-ignored-qualifiers
5040 Warn if the return type of a function has a type qualifier
5041 such as @code{const}. For ISO C such a type qualifier has no effect,
5042 since the value returned by a function is not an lvalue.
5043 For C++, the warning is only emitted for scalar types or @code{void}.
5044 ISO C prohibits qualified @code{void} return types on function
5045 definitions, so such return types always receive a warning
5046 even without this option.
5048 This warning is also enabled by @option{-Wextra}.
5050 @item -Wignored-attributes @r{(C and C++ only)}
5051 @opindex Wignored-attributes
5052 @opindex Wno-ignored-attributes
5053 Warn when an attribute is ignored. This is different from the
5054 @option{-Wattributes} option in that it warns whenever the compiler decides
5055 to drop an attribute, not that the attribute is either unknown, used in a
5056 wrong place, etc. This warning is enabled by default.
5061 Warn if the type of @code{main} is suspicious. @code{main} should be
5062 a function with external linkage, returning int, taking either zero
5063 arguments, two, or three arguments of appropriate types. This warning
5064 is enabled by default in C++ and is enabled by either @option{-Wall}
5065 or @option{-Wpedantic}.
5067 @item -Wmisleading-indentation @r{(C and C++ only)}
5068 @opindex Wmisleading-indentation
5069 @opindex Wno-misleading-indentation
5070 Warn when the indentation of the code does not reflect the block structure.
5071 Specifically, a warning is issued for @code{if}, @code{else}, @code{while}, and
5072 @code{for} clauses with a guarded statement that does not use braces,
5073 followed by an unguarded statement with the same indentation.
5075 In the following example, the call to ``bar'' is misleadingly indented as
5076 if it were guarded by the ``if'' conditional.
5079 if (some_condition ())
5081 bar (); /* Gotcha: this is not guarded by the "if". */
5084 In the case of mixed tabs and spaces, the warning uses the
5085 @option{-ftabstop=} option to determine if the statements line up
5088 The warning is not issued for code involving multiline preprocessor logic
5089 such as the following example.
5094 #if SOME_CONDITION_THAT_DOES_NOT_HOLD
5100 The warning is not issued after a @code{#line} directive, since this
5101 typically indicates autogenerated code, and no assumptions can be made
5102 about the layout of the file that the directive references.
5104 This warning is enabled by @option{-Wall} in C and C++.
5106 @item -Wmissing-attributes
5107 @opindex Wmissing-attributes
5108 @opindex Wno-missing-attributes
5109 Warn when a declaration of a function is missing one or more attributes
5110 that a related function is declared with and whose absence may adversely
5111 affect the correctness or efficiency of generated code. For example,
5112 the warning is issued for declarations of aliases that use attributes
5113 to specify less restrictive requirements than those of their targets.
5114 This typically represents a potential optimization opportunity.
5115 By contrast, the @option{-Wattribute-alias=2} option controls warnings
5116 issued when the alias is more restrictive than the target, which could
5117 lead to incorrect code generation.
5118 Attributes considered include @code{alloc_align}, @code{alloc_size},
5119 @code{cold}, @code{const}, @code{hot}, @code{leaf}, @code{malloc},
5120 @code{nonnull}, @code{noreturn}, @code{nothrow}, @code{pure},
5121 @code{returns_nonnull}, and @code{returns_twice}.
5123 In C++, the warning is issued when an explicit specialization of a primary
5124 template declared with attribute @code{alloc_align}, @code{alloc_size},
5125 @code{assume_aligned}, @code{format}, @code{format_arg}, @code{malloc},
5126 or @code{nonnull} is declared without it. Attributes @code{deprecated},
5127 @code{error}, and @code{warning} suppress the warning.
5128 (@pxref{Function Attributes}).
5130 You can use the @code{copy} attribute to apply the same
5131 set of attributes to a declaration as that on another declaration without
5132 explicitly enumerating the attributes. This attribute can be applied
5133 to declarations of functions (@pxref{Common Function Attributes}),
5134 variables (@pxref{Common Variable Attributes}), or types
5135 (@pxref{Common Type Attributes}).
5137 @option{-Wmissing-attributes} is enabled by @option{-Wall}.
5139 For example, since the declaration of the primary function template
5140 below makes use of both attribute @code{malloc} and @code{alloc_size}
5141 the declaration of the explicit specialization of the template is
5142 diagnosed because it is missing one of the attributes.
5146 T* __attribute__ ((malloc, alloc_size (1)))
5150 void* __attribute__ ((malloc)) // missing alloc_size
5151 allocate<void> (size_t);
5154 @item -Wmissing-braces
5155 @opindex Wmissing-braces
5156 @opindex Wno-missing-braces
5157 Warn if an aggregate or union initializer is not fully bracketed. In
5158 the following example, the initializer for @code{a} is not fully
5159 bracketed, but that for @code{b} is fully bracketed. This warning is
5160 enabled by @option{-Wall} in C.
5163 int a[2][2] = @{ 0, 1, 2, 3 @};
5164 int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @};
5167 This warning is enabled by @option{-Wall}.
5169 @item -Wmissing-include-dirs @r{(C, C++, Objective-C and Objective-C++ only)}
5170 @opindex Wmissing-include-dirs
5171 @opindex Wno-missing-include-dirs
5172 Warn if a user-supplied include directory does not exist.
5174 @item -Wmissing-profile
5175 @opindex Wmissing-profile
5176 @opindex Wno-missing-profile
5177 Warn if feedback profiles are missing when using the
5178 @option{-fprofile-use} option.
5179 This option diagnoses those cases where a new function or a new file is added
5180 to the user code between compiling with @option{-fprofile-generate} and with
5181 @option{-fprofile-use}, without regenerating the profiles. In these cases, the
5182 profile feedback data files do not contain any profile feedback information for
5183 the newly added function or file respectively. Also, in the case when profile
5184 count data (.gcda) files are removed, GCC cannot use any profile feedback
5185 information. In all these cases, warnings are issued to inform the user that a
5186 profile generation step is due. @option{-Wno-missing-profile} can be used to
5187 disable the warning. Ignoring the warning can result in poorly optimized code.
5188 Completely disabling the warning is not recommended and should be done only
5189 when non-existent profile data is justified.
5191 @item -Wmultistatement-macros
5192 @opindex Wmultistatement-macros
5193 @opindex Wno-multistatement-macros
5194 Warn about unsafe multiple statement macros that appear to be guarded
5195 by a clause such as @code{if}, @code{else}, @code{for}, @code{switch}, or
5196 @code{while}, in which only the first statement is actually guarded after
5197 the macro is expanded.
5202 #define DOIT x++; y++
5207 will increment @code{y} unconditionally, not just when @code{c} holds.
5208 The can usually be fixed by wrapping the macro in a do-while loop:
5210 #define DOIT do @{ x++; y++; @} while (0)
5215 This warning is enabled by @option{-Wall} in C and C++.
5218 @opindex Wparentheses
5219 @opindex Wno-parentheses
5220 Warn if parentheses are omitted in certain contexts, such
5221 as when there is an assignment in a context where a truth value
5222 is expected, or when operators are nested whose precedence people
5223 often get confused about.
5225 Also warn if a comparison like @code{x<=y<=z} appears; this is
5226 equivalent to @code{(x<=y ? 1 : 0) <= z}, which is a different
5227 interpretation from that of ordinary mathematical notation.
5229 Also warn for dangerous uses of the GNU extension to
5230 @code{?:} with omitted middle operand. When the condition
5231 in the @code{?}: operator is a boolean expression, the omitted value is
5232 always 1. Often programmers expect it to be a value computed
5233 inside the conditional expression instead.
5235 For C++ this also warns for some cases of unnecessary parentheses in
5236 declarations, which can indicate an attempt at a function call instead
5240 // Declares a local variable called mymutex.
5241 std::unique_lock<std::mutex> (mymutex);
5242 // User meant std::unique_lock<std::mutex> lock (mymutex);
5246 This warning is enabled by @option{-Wall}.
5248 @item -Wsequence-point
5249 @opindex Wsequence-point
5250 @opindex Wno-sequence-point
5251 Warn about code that may have undefined semantics because of violations
5252 of sequence point rules in the C and C++ standards.
5254 The C and C++ standards define the order in which expressions in a C/C++
5255 program are evaluated in terms of @dfn{sequence points}, which represent
5256 a partial ordering between the execution of parts of the program: those
5257 executed before the sequence point, and those executed after it. These
5258 occur after the evaluation of a full expression (one which is not part
5259 of a larger expression), after the evaluation of the first operand of a
5260 @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a
5261 function is called (but after the evaluation of its arguments and the
5262 expression denoting the called function), and in certain other places.
5263 Other than as expressed by the sequence point rules, the order of
5264 evaluation of subexpressions of an expression is not specified. All
5265 these rules describe only a partial order rather than a total order,
5266 since, for example, if two functions are called within one expression
5267 with no sequence point between them, the order in which the functions
5268 are called is not specified. However, the standards committee have
5269 ruled that function calls do not overlap.
5271 It is not specified when between sequence points modifications to the
5272 values of objects take effect. Programs whose behavior depends on this
5273 have undefined behavior; the C and C++ standards specify that ``Between
5274 the previous and next sequence point an object shall have its stored
5275 value modified at most once by the evaluation of an expression.
5276 Furthermore, the prior value shall be read only to determine the value
5277 to be stored.''. If a program breaks these rules, the results on any
5278 particular implementation are entirely unpredictable.
5280 Examples of code with undefined behavior are @code{a = a++;}, @code{a[n]
5281 = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not
5282 diagnosed by this option, and it may give an occasional false positive
5283 result, but in general it has been found fairly effective at detecting
5284 this sort of problem in programs.
5286 The C++17 standard will define the order of evaluation of operands in
5287 more cases: in particular it requires that the right-hand side of an
5288 assignment be evaluated before the left-hand side, so the above
5289 examples are no longer undefined. But this warning will still warn
5290 about them, to help people avoid writing code that is undefined in C
5291 and earlier revisions of C++.
5293 The standard is worded confusingly, therefore there is some debate
5294 over the precise meaning of the sequence point rules in subtle cases.
5295 Links to discussions of the problem, including proposed formal
5296 definitions, may be found on the GCC readings page, at
5297 @uref{http://gcc.gnu.org/@/readings.html}.
5299 This warning is enabled by @option{-Wall} for C and C++.
5301 @item -Wno-return-local-addr
5302 @opindex Wno-return-local-addr
5303 @opindex Wreturn-local-addr
5304 Do not warn about returning a pointer (or in C++, a reference) to a
5305 variable that goes out of scope after the function returns.
5308 @opindex Wreturn-type
5309 @opindex Wno-return-type
5310 Warn whenever a function is defined with a return type that defaults
5311 to @code{int}. Also warn about any @code{return} statement with no
5312 return value in a function whose return type is not @code{void}
5313 (falling off the end of the function body is considered returning
5316 For C only, warn about a @code{return} statement with an expression in a
5317 function whose return type is @code{void}, unless the expression type is
5318 also @code{void}. As a GNU extension, the latter case is accepted
5319 without a warning unless @option{-Wpedantic} is used. Attempting
5320 to use the return value of a non-@code{void} function other than @code{main}
5321 that flows off the end by reaching the closing curly brace that terminates
5322 the function is undefined.
5324 Unlike in C, in C++, flowing off the end of a non-@code{void} function other
5325 than @code{main} results in undefined behavior even when the value of
5326 the function is not used.
5328 This warning is enabled by default in C++ and by @option{-Wall} otherwise.
5330 @item -Wshift-count-negative
5331 @opindex Wshift-count-negative
5332 @opindex Wno-shift-count-negative
5333 Warn if shift count is negative. This warning is enabled by default.
5335 @item -Wshift-count-overflow
5336 @opindex Wshift-count-overflow
5337 @opindex Wno-shift-count-overflow
5338 Warn if shift count >= width of type. This warning is enabled by default.
5340 @item -Wshift-negative-value
5341 @opindex Wshift-negative-value
5342 @opindex Wno-shift-negative-value
5343 Warn if left shifting a negative value. This warning is enabled by
5344 @option{-Wextra} in C99 and C++11 modes (and newer).
5346 @item -Wshift-overflow
5347 @itemx -Wshift-overflow=@var{n}
5348 @opindex Wshift-overflow
5349 @opindex Wno-shift-overflow
5350 Warn about left shift overflows. This warning is enabled by
5351 default in C99 and C++11 modes (and newer).
5354 @item -Wshift-overflow=1
5355 This is the warning level of @option{-Wshift-overflow} and is enabled
5356 by default in C99 and C++11 modes (and newer). This warning level does
5357 not warn about left-shifting 1 into the sign bit. (However, in C, such
5358 an overflow is still rejected in contexts where an integer constant expression
5359 is required.) No warning is emitted in C++2A mode (and newer), as signed left
5362 @item -Wshift-overflow=2
5363 This warning level also warns about left-shifting 1 into the sign bit,
5364 unless C++14 mode (or newer) is active.
5370 Warn whenever a @code{switch} statement has an index of enumerated type
5371 and lacks a @code{case} for one or more of the named codes of that
5372 enumeration. (The presence of a @code{default} label prevents this
5373 warning.) @code{case} labels outside the enumeration range also
5374 provoke warnings when this option is used (even if there is a
5375 @code{default} label).
5376 This warning is enabled by @option{-Wall}.
5378 @item -Wswitch-default
5379 @opindex Wswitch-default
5380 @opindex Wno-switch-default
5381 Warn whenever a @code{switch} statement does not have a @code{default}
5385 @opindex Wswitch-enum
5386 @opindex Wno-switch-enum
5387 Warn whenever a @code{switch} statement has an index of enumerated type
5388 and lacks a @code{case} for one or more of the named codes of that
5389 enumeration. @code{case} labels outside the enumeration range also
5390 provoke warnings when this option is used. The only difference
5391 between @option{-Wswitch} and this option is that this option gives a
5392 warning about an omitted enumeration code even if there is a
5393 @code{default} label.
5396 @opindex Wswitch-bool
5397 @opindex Wno-switch-bool
5398 Warn whenever a @code{switch} statement has an index of boolean type
5399 and the case values are outside the range of a boolean type.
5400 It is possible to suppress this warning by casting the controlling
5401 expression to a type other than @code{bool}. For example:
5404 switch ((int) (a == 4))
5410 This warning is enabled by default for C and C++ programs.
5412 @item -Wswitch-outside-range
5413 @opindex Wswitch-outside-range
5414 @opindex Wno-switch-outside-range
5415 Warn whenever a @code{switch} case has a value that is outside of its
5416 respective type range. This warning is enabled by default for
5419 @item -Wswitch-unreachable
5420 @opindex Wswitch-unreachable
5421 @opindex Wno-switch-unreachable
5422 Warn whenever a @code{switch} statement contains statements between the
5423 controlling expression and the first case label, which will never be
5424 executed. For example:
5436 @option{-Wswitch-unreachable} does not warn if the statement between the
5437 controlling expression and the first case label is just a declaration:
5450 This warning is enabled by default for C and C++ programs.
5452 @item -Wsync-nand @r{(C and C++ only)}
5454 @opindex Wno-sync-nand
5455 Warn when @code{__sync_fetch_and_nand} and @code{__sync_nand_and_fetch}
5456 built-in functions are used. These functions changed semantics in GCC 4.4.
5458 @item -Wunused-but-set-parameter
5459 @opindex Wunused-but-set-parameter
5460 @opindex Wno-unused-but-set-parameter
5461 Warn whenever a function parameter is assigned to, but otherwise unused
5462 (aside from its declaration).
5464 To suppress this warning use the @code{unused} attribute
5465 (@pxref{Variable Attributes}).
5467 This warning is also enabled by @option{-Wunused} together with
5470 @item -Wunused-but-set-variable
5471 @opindex Wunused-but-set-variable
5472 @opindex Wno-unused-but-set-variable
5473 Warn whenever a local variable is assigned to, but otherwise unused
5474 (aside from its declaration).
5475 This warning is enabled by @option{-Wall}.
5477 To suppress this warning use the @code{unused} attribute
5478 (@pxref{Variable Attributes}).
5480 This warning is also enabled by @option{-Wunused}, which is enabled
5483 @item -Wunused-function
5484 @opindex Wunused-function
5485 @opindex Wno-unused-function
5486 Warn whenever a static function is declared but not defined or a
5487 non-inline static function is unused.
5488 This warning is enabled by @option{-Wall}.
5490 @item -Wunused-label
5491 @opindex Wunused-label
5492 @opindex Wno-unused-label
5493 Warn whenever a label is declared but not used.
5494 This warning is enabled by @option{-Wall}.
5496 To suppress this warning use the @code{unused} attribute
5497 (@pxref{Variable Attributes}).
5499 @item -Wunused-local-typedefs @r{(C, Objective-C, C++ and Objective-C++ only)}
5500 @opindex Wunused-local-typedefs
5501 @opindex Wno-unused-local-typedefs
5502 Warn when a typedef locally defined in a function is not used.
5503 This warning is enabled by @option{-Wall}.
5505 @item -Wunused-parameter
5506 @opindex Wunused-parameter
5507 @opindex Wno-unused-parameter
5508 Warn whenever a function parameter is unused aside from its declaration.
5510 To suppress this warning use the @code{unused} attribute
5511 (@pxref{Variable Attributes}).
5513 @item -Wno-unused-result
5514 @opindex Wunused-result
5515 @opindex Wno-unused-result
5516 Do not warn if a caller of a function marked with attribute
5517 @code{warn_unused_result} (@pxref{Function Attributes}) does not use
5518 its return value. The default is @option{-Wunused-result}.
5520 @item -Wunused-variable
5521 @opindex Wunused-variable
5522 @opindex Wno-unused-variable
5523 Warn whenever a local or static variable is unused aside from its
5524 declaration. This option implies @option{-Wunused-const-variable=1} for C,
5525 but not for C++. This warning is enabled by @option{-Wall}.
5527 To suppress this warning use the @code{unused} attribute
5528 (@pxref{Variable Attributes}).
5530 @item -Wunused-const-variable
5531 @itemx -Wunused-const-variable=@var{n}
5532 @opindex Wunused-const-variable
5533 @opindex Wno-unused-const-variable
5534 Warn whenever a constant static variable is unused aside from its declaration.
5535 @option{-Wunused-const-variable=1} is enabled by @option{-Wunused-variable}
5536 for C, but not for C++. In C this declares variable storage, but in C++ this
5537 is not an error since const variables take the place of @code{#define}s.
5539 To suppress this warning use the @code{unused} attribute
5540 (@pxref{Variable Attributes}).
5543 @item -Wunused-const-variable=1
5544 This is the warning level that is enabled by @option{-Wunused-variable} for
5545 C. It warns only about unused static const variables defined in the main
5546 compilation unit, but not about static const variables declared in any
5549 @item -Wunused-const-variable=2
5550 This warning level also warns for unused constant static variables in
5551 headers (excluding system headers). This is the warning level of
5552 @option{-Wunused-const-variable} and must be explicitly requested since
5553 in C++ this isn't an error and in C it might be harder to clean up all
5557 @item -Wunused-value
5558 @opindex Wunused-value
5559 @opindex Wno-unused-value
5560 Warn whenever a statement computes a result that is explicitly not
5561 used. To suppress this warning cast the unused expression to
5562 @code{void}. This includes an expression-statement or the left-hand
5563 side of a comma expression that contains no side effects. For example,
5564 an expression such as @code{x[i,j]} causes a warning, while
5565 @code{x[(void)i,j]} does not.
5567 This warning is enabled by @option{-Wall}.
5572 All the above @option{-Wunused} options combined.
5574 In order to get a warning about an unused function parameter, you must
5575 either specify @option{-Wextra -Wunused} (note that @option{-Wall} implies
5576 @option{-Wunused}), or separately specify @option{-Wunused-parameter}.
5578 @item -Wuninitialized
5579 @opindex Wuninitialized
5580 @opindex Wno-uninitialized
5581 Warn if an automatic variable is used without first being initialized
5582 or if a variable may be clobbered by a @code{setjmp} call. In C++,
5583 warn if a non-static reference or non-static @code{const} member
5584 appears in a class without constructors.
5586 If you want to warn about code that uses the uninitialized value of the
5587 variable in its own initializer, use the @option{-Winit-self} option.
5589 These warnings occur for individual uninitialized or clobbered
5590 elements of structure, union or array variables as well as for
5591 variables that are uninitialized or clobbered as a whole. They do
5592 not occur for variables or elements declared @code{volatile}. Because
5593 these warnings depend on optimization, the exact variables or elements
5594 for which there are warnings depends on the precise optimization
5595 options and version of GCC used.
5597 Note that there may be no warning about a variable that is used only
5598 to compute a value that itself is never used, because such
5599 computations may be deleted by data flow analysis before the warnings
5602 @item -Winvalid-memory-model
5603 @opindex Winvalid-memory-model
5604 @opindex Wno-invalid-memory-model
5605 Warn for invocations of @ref{__atomic Builtins}, @ref{__sync Builtins},
5606 and the C11 atomic generic functions with a memory consistency argument
5607 that is either invalid for the operation or outside the range of values
5608 of the @code{memory_order} enumeration. For example, since the
5609 @code{__atomic_store} and @code{__atomic_store_n} built-ins are only
5610 defined for the relaxed, release, and sequentially consistent memory
5611 orders the following code is diagnosed:
5616 __atomic_store_n (i, 0, memory_order_consume);
5620 @option{-Winvalid-memory-model} is enabled by default.
5622 @item -Wmaybe-uninitialized
5623 @opindex Wmaybe-uninitialized
5624 @opindex Wno-maybe-uninitialized
5625 For an automatic (i.e.@: local) variable, if there exists a path from the
5626 function entry to a use of the variable that is initialized, but there exist
5627 some other paths for which the variable is not initialized, the compiler
5628 emits a warning if it cannot prove the uninitialized paths are not
5629 executed at run time.
5631 These warnings are only possible in optimizing compilation, because otherwise
5632 GCC does not keep track of the state of variables.
5634 These warnings are made optional because GCC may not be able to determine when
5635 the code is correct in spite of appearing to have an error. Here is one
5636 example of how this can happen:
5656 If the value of @code{y} is always 1, 2 or 3, then @code{x} is
5657 always initialized, but GCC doesn't know this. To suppress the
5658 warning, you need to provide a default case with assert(0) or
5661 @cindex @code{longjmp} warnings
5662 This option also warns when a non-volatile automatic variable might be
5663 changed by a call to @code{longjmp}.
5664 The compiler sees only the calls to @code{setjmp}. It cannot know
5665 where @code{longjmp} will be called; in fact, a signal handler could
5666 call it at any point in the code. As a result, you may get a warning
5667 even when there is in fact no problem because @code{longjmp} cannot
5668 in fact be called at the place that would cause a problem.
5670 Some spurious warnings can be avoided if you declare all the functions
5671 you use that never return as @code{noreturn}. @xref{Function
5674 This warning is enabled by @option{-Wall} or @option{-Wextra}.
5676 @item -Wunknown-pragmas
5677 @opindex Wunknown-pragmas
5678 @opindex Wno-unknown-pragmas
5679 @cindex warning for unknown pragmas
5680 @cindex unknown pragmas, warning
5681 @cindex pragmas, warning of unknown
5682 Warn when a @code{#pragma} directive is encountered that is not understood by
5683 GCC@. If this command-line option is used, warnings are even issued
5684 for unknown pragmas in system header files. This is not the case if
5685 the warnings are only enabled by the @option{-Wall} command-line option.
5688 @opindex Wno-pragmas
5690 Do not warn about misuses of pragmas, such as incorrect parameters,
5691 invalid syntax, or conflicts between pragmas. See also
5692 @option{-Wunknown-pragmas}.
5694 @item -Wno-prio-ctor-dtor
5695 @opindex Wno-prio-ctor-dtor
5696 @opindex Wprio-ctor-dtor
5697 Do not warn if a priority from 0 to 100 is used for constructor or destructor.
5698 The use of constructor and destructor attributes allow you to assign a
5699 priority to the constructor/destructor to control its order of execution
5700 before @code{main} is called or after it returns. The priority values must be
5701 greater than 100 as the compiler reserves priority values between 0--100 for
5704 @item -Wstrict-aliasing
5705 @opindex Wstrict-aliasing
5706 @opindex Wno-strict-aliasing
5707 This option is only active when @option{-fstrict-aliasing} is active.
5708 It warns about code that might break the strict aliasing rules that the
5709 compiler is using for optimization. The warning does not catch all
5710 cases, but does attempt to catch the more common pitfalls. It is
5711 included in @option{-Wall}.
5712 It is equivalent to @option{-Wstrict-aliasing=3}
5714 @item -Wstrict-aliasing=n
5715 @opindex Wstrict-aliasing=n
5716 This option is only active when @option{-fstrict-aliasing} is active.
5717 It warns about code that might break the strict aliasing rules that the
5718 compiler is using for optimization.
5719 Higher levels correspond to higher accuracy (fewer false positives).
5720 Higher levels also correspond to more effort, similar to the way @option{-O}
5722 @option{-Wstrict-aliasing} is equivalent to @option{-Wstrict-aliasing=3}.
5724 Level 1: Most aggressive, quick, least accurate.
5725 Possibly useful when higher levels
5726 do not warn but @option{-fstrict-aliasing} still breaks the code, as it has very few
5727 false negatives. However, it has many false positives.
5728 Warns for all pointer conversions between possibly incompatible types,
5729 even if never dereferenced. Runs in the front end only.
5731 Level 2: Aggressive, quick, not too precise.
5732 May still have many false positives (not as many as level 1 though),
5733 and few false negatives (but possibly more than level 1).
5734 Unlike level 1, it only warns when an address is taken. Warns about
5735 incomplete types. Runs in the front end only.
5737 Level 3 (default for @option{-Wstrict-aliasing}):
5738 Should have very few false positives and few false
5739 negatives. Slightly slower than levels 1 or 2 when optimization is enabled.
5740 Takes care of the common pun+dereference pattern in the front end:
5741 @code{*(int*)&some_float}.
5742 If optimization is enabled, it also runs in the back end, where it deals
5743 with multiple statement cases using flow-sensitive points-to information.
5744 Only warns when the converted pointer is dereferenced.
5745 Does not warn about incomplete types.
5747 @item -Wstrict-overflow
5748 @itemx -Wstrict-overflow=@var{n}
5749 @opindex Wstrict-overflow
5750 @opindex Wno-strict-overflow
5751 This option is only active when signed overflow is undefined.
5752 It warns about cases where the compiler optimizes based on the
5753 assumption that signed overflow does not occur. Note that it does not
5754 warn about all cases where the code might overflow: it only warns
5755 about cases where the compiler implements some optimization. Thus
5756 this warning depends on the optimization level.
5758 An optimization that assumes that signed overflow does not occur is
5759 perfectly safe if the values of the variables involved are such that
5760 overflow never does, in fact, occur. Therefore this warning can
5761 easily give a false positive: a warning about code that is not
5762 actually a problem. To help focus on important issues, several
5763 warning levels are defined. No warnings are issued for the use of
5764 undefined signed overflow when estimating how many iterations a loop
5765 requires, in particular when determining whether a loop will be
5769 @item -Wstrict-overflow=1
5770 Warn about cases that are both questionable and easy to avoid. For
5771 example the compiler simplifies
5772 @code{x + 1 > x} to @code{1}. This level of
5773 @option{-Wstrict-overflow} is enabled by @option{-Wall}; higher levels
5774 are not, and must be explicitly requested.
5776 @item -Wstrict-overflow=2
5777 Also warn about other cases where a comparison is simplified to a
5778 constant. For example: @code{abs (x) >= 0}. This can only be
5779 simplified when signed integer overflow is undefined, because
5780 @code{abs (INT_MIN)} overflows to @code{INT_MIN}, which is less than
5781 zero. @option{-Wstrict-overflow} (with no level) is the same as
5782 @option{-Wstrict-overflow=2}.
5784 @item -Wstrict-overflow=3
5785 Also warn about other cases where a comparison is simplified. For
5786 example: @code{x + 1 > 1} is simplified to @code{x > 0}.
5788 @item -Wstrict-overflow=4
5789 Also warn about other simplifications not covered by the above cases.
5790 For example: @code{(x * 10) / 5} is simplified to @code{x * 2}.
5792 @item -Wstrict-overflow=5
5793 Also warn about cases where the compiler reduces the magnitude of a
5794 constant involved in a comparison. For example: @code{x + 2 > y} is
5795 simplified to @code{x + 1 >= y}. This is reported only at the
5796 highest warning level because this simplification applies to many
5797 comparisons, so this warning level gives a very large number of
5801 @item -Wstringop-overflow
5802 @itemx -Wstringop-overflow=@var{type}
5803 @opindex Wstringop-overflow
5804 @opindex Wno-stringop-overflow
5805 Warn for calls to string manipulation functions such as @code{memcpy} and
5806 @code{strcpy} that are determined to overflow the destination buffer. The
5807 optional argument is one greater than the type of Object Size Checking to
5808 perform to determine the size of the destination. @xref{Object Size Checking}.
5809 The argument is meaningful only for functions that operate on character arrays
5810 but not for raw memory functions like @code{memcpy} which always make use
5811 of Object Size type-0. The option also warns for calls that specify a size
5812 in excess of the largest possible object or at most @code{SIZE_MAX / 2} bytes.
5813 The option produces the best results with optimization enabled but can detect
5814 a small subset of simple buffer overflows even without optimization in
5815 calls to the GCC built-in functions like @code{__builtin_memcpy} that
5816 correspond to the standard functions. In any case, the option warns about
5817 just a subset of buffer overflows detected by the corresponding overflow
5818 checking built-ins. For example, the option will issue a warning for
5819 the @code{strcpy} call below because it copies at least 5 characters
5820 (the string @code{"blue"} including the terminating NUL) into the buffer
5824 enum Color @{ blue, purple, yellow @};
5825 const char* f (enum Color clr)
5827 static char buf [4];
5831 case blue: str = "blue"; break;
5832 case purple: str = "purple"; break;
5833 case yellow: str = "yellow"; break;
5836 return strcpy (buf, str); // warning here
5840 Option @option{-Wstringop-overflow=2} is enabled by default.
5843 @item -Wstringop-overflow
5844 @itemx -Wstringop-overflow=1
5845 @opindex Wstringop-overflow
5846 @opindex Wno-stringop-overflow
5847 The @option{-Wstringop-overflow=1} option uses type-zero Object Size Checking
5848 to determine the sizes of destination objects. This is the default setting
5849 of the option. At this setting the option will not warn for writes past
5850 the end of subobjects of larger objects accessed by pointers unless the
5851 size of the largest surrounding object is known. When the destination may
5852 be one of several objects it is assumed to be the largest one of them. On
5853 Linux systems, when optimization is enabled at this setting the option warns
5854 for the same code as when the @code{_FORTIFY_SOURCE} macro is defined to
5857 @item -Wstringop-overflow=2
5858 The @option{-Wstringop-overflow=2} option uses type-one Object Size Checking
5859 to determine the sizes of destination objects. At this setting the option
5860 will warn about overflows when writing to members of the largest complete
5861 objects whose exact size is known. It will, however, not warn for excessive
5862 writes to the same members of unknown objects referenced by pointers since
5863 they may point to arrays containing unknown numbers of elements.
5865 @item -Wstringop-overflow=3
5866 The @option{-Wstringop-overflow=3} option uses type-two Object Size Checking
5867 to determine the sizes of destination objects. At this setting the option
5868 warns about overflowing the smallest object or data member. This is the
5869 most restrictive setting of the option that may result in warnings for safe
5872 @item -Wstringop-overflow=4
5873 The @option{-Wstringop-overflow=4} option uses type-three Object Size Checking
5874 to determine the sizes of destination objects. At this setting the option
5875 will warn about overflowing any data members, and when the destination is
5876 one of several objects it uses the size of the largest of them to decide
5877 whether to issue a warning. Similarly to @option{-Wstringop-overflow=3} this
5878 setting of the option may result in warnings for benign code.
5881 @item -Wstringop-truncation
5882 @opindex Wstringop-truncation
5883 @opindex Wno-stringop-truncation
5884 Warn for calls to bounded string manipulation functions such as @code{strncat},
5885 @code{strncpy}, and @code{stpncpy} that may either truncate the copied string
5886 or leave the destination unchanged.
5888 In the following example, the call to @code{strncat} specifies a bound that
5889 is less than the length of the source string. As a result, the copy of
5890 the source will be truncated and so the call is diagnosed. To avoid the
5891 warning use @code{bufsize - strlen (buf) - 1)} as the bound.
5894 void append (char *buf, size_t bufsize)
5896 strncat (buf, ".txt", 3);
5900 As another example, the following call to @code{strncpy} results in copying
5901 to @code{d} just the characters preceding the terminating NUL, without
5902 appending the NUL to the end. Assuming the result of @code{strncpy} is
5903 necessarily a NUL-terminated string is a common mistake, and so the call
5904 is diagnosed. To avoid the warning when the result is not expected to be
5905 NUL-terminated, call @code{memcpy} instead.
5908 void copy (char *d, const char *s)
5910 strncpy (d, s, strlen (s));
5914 In the following example, the call to @code{strncpy} specifies the size
5915 of the destination buffer as the bound. If the length of the source
5916 string is equal to or greater than this size the result of the copy will
5917 not be NUL-terminated. Therefore, the call is also diagnosed. To avoid
5918 the warning, specify @code{sizeof buf - 1} as the bound and set the last
5919 element of the buffer to @code{NUL}.
5922 void copy (const char *s)
5925 strncpy (buf, s, sizeof buf);
5930 In situations where a character array is intended to store a sequence
5931 of bytes with no terminating @code{NUL} such an array may be annotated
5932 with attribute @code{nonstring} to avoid this warning. Such arrays,
5933 however, are not suitable arguments to functions that expect
5934 @code{NUL}-terminated strings. To help detect accidental misuses of
5935 such arrays GCC issues warnings unless it can prove that the use is
5936 safe. @xref{Common Variable Attributes}.
5938 @item -Wsuggest-attribute=@r{[}pure@r{|}const@r{|}noreturn@r{|}format@r{|}cold@r{|}malloc@r{]}
5939 @opindex Wsuggest-attribute=
5940 @opindex Wno-suggest-attribute=
5941 Warn for cases where adding an attribute may be beneficial. The
5942 attributes currently supported are listed below.
5945 @item -Wsuggest-attribute=pure
5946 @itemx -Wsuggest-attribute=const
5947 @itemx -Wsuggest-attribute=noreturn
5948 @itemx -Wmissing-noreturn
5949 @itemx -Wsuggest-attribute=malloc
5950 @opindex Wsuggest-attribute=pure
5951 @opindex Wno-suggest-attribute=pure
5952 @opindex Wsuggest-attribute=const
5953 @opindex Wno-suggest-attribute=const
5954 @opindex Wsuggest-attribute=noreturn
5955 @opindex Wno-suggest-attribute=noreturn
5956 @opindex Wmissing-noreturn
5957 @opindex Wno-missing-noreturn
5958 @opindex Wsuggest-attribute=malloc
5959 @opindex Wno-suggest-attribute=malloc
5961 Warn about functions that might be candidates for attributes
5962 @code{pure}, @code{const} or @code{noreturn} or @code{malloc}. The compiler
5963 only warns for functions visible in other compilation units or (in the case of
5964 @code{pure} and @code{const}) if it cannot prove that the function returns
5965 normally. A function returns normally if it doesn't contain an infinite loop or
5966 return abnormally by throwing, calling @code{abort} or trapping. This analysis
5967 requires option @option{-fipa-pure-const}, which is enabled by default at
5968 @option{-O} and higher. Higher optimization levels improve the accuracy
5971 @item -Wsuggest-attribute=format
5972 @itemx -Wmissing-format-attribute
5973 @opindex Wsuggest-attribute=format
5974 @opindex Wmissing-format-attribute
5975 @opindex Wno-suggest-attribute=format
5976 @opindex Wno-missing-format-attribute
5980 Warn about function pointers that might be candidates for @code{format}
5981 attributes. Note these are only possible candidates, not absolute ones.
5982 GCC guesses that function pointers with @code{format} attributes that
5983 are used in assignment, initialization, parameter passing or return
5984 statements should have a corresponding @code{format} attribute in the
5985 resulting type. I.e.@: the left-hand side of the assignment or
5986 initialization, the type of the parameter variable, or the return type
5987 of the containing function respectively should also have a @code{format}
5988 attribute to avoid the warning.
5990 GCC also warns about function definitions that might be
5991 candidates for @code{format} attributes. Again, these are only
5992 possible candidates. GCC guesses that @code{format} attributes
5993 might be appropriate for any function that calls a function like
5994 @code{vprintf} or @code{vscanf}, but this might not always be the
5995 case, and some functions for which @code{format} attributes are
5996 appropriate may not be detected.
5998 @item -Wsuggest-attribute=cold
5999 @opindex Wsuggest-attribute=cold
6000 @opindex Wno-suggest-attribute=cold
6002 Warn about functions that might be candidates for @code{cold} attribute. This
6003 is based on static detection and generally will only warn about functions which
6004 always leads to a call to another @code{cold} function such as wrappers of
6005 C++ @code{throw} or fatal error reporting functions leading to @code{abort}.
6008 @item -Wsuggest-final-types
6009 @opindex Wno-suggest-final-types
6010 @opindex Wsuggest-final-types
6011 Warn about types with virtual methods where code quality would be improved
6012 if the type were declared with the C++11 @code{final} specifier,
6014 declared in an anonymous namespace. This allows GCC to more aggressively
6015 devirtualize the polymorphic calls. This warning is more effective with
6016 link-time optimization,
6017 where the information about the class hierarchy graph is
6020 @item -Wsuggest-final-methods
6021 @opindex Wno-suggest-final-methods
6022 @opindex Wsuggest-final-methods
6023 Warn about virtual methods where code quality would be improved if the method
6024 were declared with the C++11 @code{final} specifier,
6025 or, if possible, its type were
6026 declared in an anonymous namespace or with the @code{final} specifier.
6028 more effective with link-time optimization, where the information about the
6029 class hierarchy graph is more complete. It is recommended to first consider
6030 suggestions of @option{-Wsuggest-final-types} and then rebuild with new
6033 @item -Wsuggest-override
6034 Warn about overriding virtual functions that are not marked with the override
6038 @opindex Wno-alloc-zero
6039 @opindex Walloc-zero
6040 Warn about calls to allocation functions decorated with attribute
6041 @code{alloc_size} that specify zero bytes, including those to the built-in
6042 forms of the functions @code{aligned_alloc}, @code{alloca}, @code{calloc},
6043 @code{malloc}, and @code{realloc}. Because the behavior of these functions
6044 when called with a zero size differs among implementations (and in the case
6045 of @code{realloc} has been deprecated) relying on it may result in subtle
6046 portability bugs and should be avoided.
6048 @item -Walloc-size-larger-than=@var{byte-size}
6049 @opindex Walloc-size-larger-than=
6050 @opindex Wno-alloc-size-larger-than
6051 Warn about calls to functions decorated with attribute @code{alloc_size}
6052 that attempt to allocate objects larger than the specified number of bytes,
6053 or where the result of the size computation in an integer type with infinite
6054 precision would exceed the value of @samp{PTRDIFF_MAX} on the target.
6055 @option{-Walloc-size-larger-than=}@samp{PTRDIFF_MAX} is enabled by default.
6056 Warnings controlled by the option can be disabled either by specifying
6057 @var{byte-size} of @samp{SIZE_MAX} or more or by
6058 @option{-Wno-alloc-size-larger-than}.
6059 @xref{Function Attributes}.
6061 @item -Wno-alloc-size-larger-than
6062 @opindex Wno-alloc-size-larger-than
6063 Disable @option{-Walloc-size-larger-than=} warnings. The option is
6064 equivalent to @option{-Walloc-size-larger-than=}@samp{SIZE_MAX} or
6070 This option warns on all uses of @code{alloca} in the source.
6072 @item -Walloca-larger-than=@var{byte-size}
6073 @opindex Walloca-larger-than=
6074 @opindex Wno-alloca-larger-than
6075 This option warns on calls to @code{alloca} with an integer argument whose
6076 value is either zero, or that is not bounded by a controlling predicate
6077 that limits its value to at most @var{byte-size}. It also warns for calls
6078 to @code{alloca} where the bound value is unknown. Arguments of non-integer
6079 types are considered unbounded even if they appear to be constrained to
6082 For example, a bounded case of @code{alloca} could be:
6085 void func (size_t n)
6096 In the above example, passing @code{-Walloca-larger-than=1000} would not
6097 issue a warning because the call to @code{alloca} is known to be at most
6098 1000 bytes. However, if @code{-Walloca-larger-than=500} were passed,
6099 the compiler would emit a warning.
6101 Unbounded uses, on the other hand, are uses of @code{alloca} with no
6102 controlling predicate constraining its integer argument. For example:
6107 void *p = alloca (n);
6112 If @code{-Walloca-larger-than=500} were passed, the above would trigger
6113 a warning, but this time because of the lack of bounds checking.
6115 Note, that even seemingly correct code involving signed integers could
6119 void func (signed int n)
6129 In the above example, @var{n} could be negative, causing a larger than
6130 expected argument to be implicitly cast into the @code{alloca} call.
6132 This option also warns when @code{alloca} is used in a loop.
6134 @option{-Walloca-larger-than=}@samp{PTRDIFF_MAX} is enabled by default
6135 but is usually only effective when @option{-ftree-vrp} is active (default
6136 for @option{-O2} and above).
6138 See also @option{-Wvla-larger-than=}@samp{byte-size}.
6140 @item -Wno-alloca-larger-than
6141 @opindex Wno-alloca-larger-than
6142 Disable @option{-Walloca-larger-than=} warnings. The option is
6143 equivalent to @option{-Walloca-larger-than=}@samp{SIZE_MAX} or larger.
6145 @item -Warray-bounds
6146 @itemx -Warray-bounds=@var{n}
6147 @opindex Wno-array-bounds
6148 @opindex Warray-bounds
6149 This option is only active when @option{-ftree-vrp} is active
6150 (default for @option{-O2} and above). It warns about subscripts to arrays
6151 that are always out of bounds. This warning is enabled by @option{-Wall}.
6154 @item -Warray-bounds=1
6155 This is the warning level of @option{-Warray-bounds} and is enabled
6156 by @option{-Wall}; higher levels are not, and must be explicitly requested.
6158 @item -Warray-bounds=2
6159 This warning level also warns about out of bounds access for
6160 arrays at the end of a struct and for arrays accessed through
6161 pointers. This warning level may give a larger number of
6162 false positives and is deactivated by default.
6165 @item -Wattribute-alias=@var{n}
6166 @itemx -Wno-attribute-alias
6167 @opindex Wattribute-alias
6168 @opindex Wno-attribute-alias
6169 Warn about declarations using the @code{alias} and similar attributes whose
6170 target is incompatible with the type of the alias.
6171 @xref{Function Attributes,,Declaring Attributes of Functions}.
6174 @item -Wattribute-alias=1
6175 The default warning level of the @option{-Wattribute-alias} option diagnoses
6176 incompatibilities between the type of the alias declaration and that of its
6177 target. Such incompatibilities are typically indicative of bugs.
6179 @item -Wattribute-alias=2
6181 At this level @option{-Wattribute-alias} also diagnoses cases where
6182 the attributes of the alias declaration are more restrictive than the
6183 attributes applied to its target. These mismatches can potentially
6184 result in incorrect code generation. In other cases they may be
6185 benign and could be resolved simply by adding the missing attribute to
6186 the target. For comparison, see the @option{-Wmissing-attributes}
6187 option, which controls diagnostics when the alias declaration is less
6188 restrictive than the target, rather than more restrictive.
6190 Attributes considered include @code{alloc_align}, @code{alloc_size},
6191 @code{cold}, @code{const}, @code{hot}, @code{leaf}, @code{malloc},
6192 @code{nonnull}, @code{noreturn}, @code{nothrow}, @code{pure},
6193 @code{returns_nonnull}, and @code{returns_twice}.
6196 @option{-Wattribute-alias} is equivalent to @option{-Wattribute-alias=1}.
6197 This is the default. You can disable these warnings with either
6198 @option{-Wno-attribute-alias} or @option{-Wattribute-alias=0}.
6200 @item -Wbool-compare
6201 @opindex Wno-bool-compare
6202 @opindex Wbool-compare
6203 Warn about boolean expression compared with an integer value different from
6204 @code{true}/@code{false}. For instance, the following comparison is
6209 if ((n > 1) == 2) @{ @dots{} @}
6211 This warning is enabled by @option{-Wall}.
6213 @item -Wbool-operation
6214 @opindex Wno-bool-operation
6215 @opindex Wbool-operation
6216 Warn about suspicious operations on expressions of a boolean type. For
6217 instance, bitwise negation of a boolean is very likely a bug in the program.
6218 For C, this warning also warns about incrementing or decrementing a boolean,
6219 which rarely makes sense. (In C++, decrementing a boolean is always invalid.
6220 Incrementing a boolean is invalid in C++17, and deprecated otherwise.)
6222 This warning is enabled by @option{-Wall}.
6224 @item -Wduplicated-branches
6225 @opindex Wno-duplicated-branches
6226 @opindex Wduplicated-branches
6227 Warn when an if-else has identical branches. This warning detects cases like
6234 It doesn't warn when both branches contain just a null statement. This warning
6235 also warn for conditional operators:
6237 int i = x ? *p : *p;
6240 @item -Wduplicated-cond
6241 @opindex Wno-duplicated-cond
6242 @opindex Wduplicated-cond
6243 Warn about duplicated conditions in an if-else-if chain. For instance,
6244 warn for the following code:
6246 if (p->q != NULL) @{ @dots{} @}
6247 else if (p->q != NULL) @{ @dots{} @}
6250 @item -Wframe-address
6251 @opindex Wno-frame-address
6252 @opindex Wframe-address
6253 Warn when the @samp{__builtin_frame_address} or @samp{__builtin_return_address}
6254 is called with an argument greater than 0. Such calls may return indeterminate
6255 values or crash the program. The warning is included in @option{-Wall}.
6257 @item -Wno-discarded-qualifiers @r{(C and Objective-C only)}
6258 @opindex Wno-discarded-qualifiers
6259 @opindex Wdiscarded-qualifiers
6260 Do not warn if type qualifiers on pointers are being discarded.
6261 Typically, the compiler warns if a @code{const char *} variable is
6262 passed to a function that takes a @code{char *} parameter. This option
6263 can be used to suppress such a warning.
6265 @item -Wno-discarded-array-qualifiers @r{(C and Objective-C only)}
6266 @opindex Wno-discarded-array-qualifiers
6267 @opindex Wdiscarded-array-qualifiers
6268 Do not warn if type qualifiers on arrays which are pointer targets
6269 are being discarded. Typically, the compiler warns if a
6270 @code{const int (*)[]} variable is passed to a function that
6271 takes a @code{int (*)[]} parameter. This option can be used to
6272 suppress such a warning.
6274 @item -Wno-incompatible-pointer-types @r{(C and Objective-C only)}
6275 @opindex Wno-incompatible-pointer-types
6276 @opindex Wincompatible-pointer-types
6277 Do not warn when there is a conversion between pointers that have incompatible
6278 types. This warning is for cases not covered by @option{-Wno-pointer-sign},
6279 which warns for pointer argument passing or assignment with different
6282 @item -Wno-int-conversion @r{(C and Objective-C only)}
6283 @opindex Wno-int-conversion
6284 @opindex Wint-conversion
6285 Do not warn about incompatible integer to pointer and pointer to integer
6286 conversions. This warning is about implicit conversions; for explicit
6287 conversions the warnings @option{-Wno-int-to-pointer-cast} and
6288 @option{-Wno-pointer-to-int-cast} may be used.
6290 @item -Wno-div-by-zero
6291 @opindex Wno-div-by-zero
6292 @opindex Wdiv-by-zero
6293 Do not warn about compile-time integer division by zero. Floating-point
6294 division by zero is not warned about, as it can be a legitimate way of
6295 obtaining infinities and NaNs.
6297 @item -Wsystem-headers
6298 @opindex Wsystem-headers
6299 @opindex Wno-system-headers
6300 @cindex warnings from system headers
6301 @cindex system headers, warnings from
6302 Print warning messages for constructs found in system header files.
6303 Warnings from system headers are normally suppressed, on the assumption
6304 that they usually do not indicate real problems and would only make the
6305 compiler output harder to read. Using this command-line option tells
6306 GCC to emit warnings from system headers as if they occurred in user
6307 code. However, note that using @option{-Wall} in conjunction with this
6308 option does @emph{not} warn about unknown pragmas in system
6309 headers---for that, @option{-Wunknown-pragmas} must also be used.
6311 @item -Wtautological-compare
6312 @opindex Wtautological-compare
6313 @opindex Wno-tautological-compare
6314 Warn if a self-comparison always evaluates to true or false. This
6315 warning detects various mistakes such as:
6319 if (i > i) @{ @dots{} @}
6322 This warning also warns about bitwise comparisons that always evaluate
6323 to true or false, for instance:
6325 if ((a & 16) == 10) @{ @dots{} @}
6327 will always be false.
6329 This warning is enabled by @option{-Wall}.
6332 @opindex Wtrampolines
6333 @opindex Wno-trampolines
6334 Warn about trampolines generated for pointers to nested functions.
6335 A trampoline is a small piece of data or code that is created at run
6336 time on the stack when the address of a nested function is taken, and is
6337 used to call the nested function indirectly. For some targets, it is
6338 made up of data only and thus requires no special treatment. But, for
6339 most targets, it is made up of code and thus requires the stack to be
6340 made executable in order for the program to work properly.
6343 @opindex Wfloat-equal
6344 @opindex Wno-float-equal
6345 Warn if floating-point values are used in equality comparisons.
6347 The idea behind this is that sometimes it is convenient (for the
6348 programmer) to consider floating-point values as approximations to
6349 infinitely precise real numbers. If you are doing this, then you need
6350 to compute (by analyzing the code, or in some other way) the maximum or
6351 likely maximum error that the computation introduces, and allow for it
6352 when performing comparisons (and when producing output, but that's a
6353 different problem). In particular, instead of testing for equality, you
6354 should check to see whether the two values have ranges that overlap; and
6355 this is done with the relational operators, so equality comparisons are
6358 @item -Wtraditional @r{(C and Objective-C only)}
6359 @opindex Wtraditional
6360 @opindex Wno-traditional
6361 Warn about certain constructs that behave differently in traditional and
6362 ISO C@. Also warn about ISO C constructs that have no traditional C
6363 equivalent, and/or problematic constructs that should be avoided.
6367 Macro parameters that appear within string literals in the macro body.
6368 In traditional C macro replacement takes place within string literals,
6369 but in ISO C it does not.
6372 In traditional C, some preprocessor directives did not exist.
6373 Traditional preprocessors only considered a line to be a directive
6374 if the @samp{#} appeared in column 1 on the line. Therefore
6375 @option{-Wtraditional} warns about directives that traditional C
6376 understands but ignores because the @samp{#} does not appear as the
6377 first character on the line. It also suggests you hide directives like
6378 @code{#pragma} not understood by traditional C by indenting them. Some
6379 traditional implementations do not recognize @code{#elif}, so this option
6380 suggests avoiding it altogether.
6383 A function-like macro that appears without arguments.
6386 The unary plus operator.
6389 The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating-point
6390 constant suffixes. (Traditional C does support the @samp{L} suffix on integer
6391 constants.) Note, these suffixes appear in macros defined in the system
6392 headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}.
6393 Use of these macros in user code might normally lead to spurious
6394 warnings, however GCC's integrated preprocessor has enough context to
6395 avoid warning in these cases.
6398 A function declared external in one block and then used after the end of
6402 A @code{switch} statement has an operand of type @code{long}.
6405 A non-@code{static} function declaration follows a @code{static} one.
6406 This construct is not accepted by some traditional C compilers.
6409 The ISO type of an integer constant has a different width or
6410 signedness from its traditional type. This warning is only issued if
6411 the base of the constant is ten. I.e.@: hexadecimal or octal values, which
6412 typically represent bit patterns, are not warned about.
6415 Usage of ISO string concatenation is detected.
6418 Initialization of automatic aggregates.
6421 Identifier conflicts with labels. Traditional C lacks a separate
6422 namespace for labels.
6425 Initialization of unions. If the initializer is zero, the warning is
6426 omitted. This is done under the assumption that the zero initializer in
6427 user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing
6428 initializer warnings and relies on default initialization to zero in the
6432 Conversions by prototypes between fixed/floating-point values and vice
6433 versa. The absence of these prototypes when compiling with traditional
6434 C causes serious problems. This is a subset of the possible
6435 conversion warnings; for the full set use @option{-Wtraditional-conversion}.
6438 Use of ISO C style function definitions. This warning intentionally is
6439 @emph{not} issued for prototype declarations or variadic functions
6440 because these ISO C features appear in your code when using
6441 libiberty's traditional C compatibility macros, @code{PARAMS} and
6442 @code{VPARAMS}. This warning is also bypassed for nested functions
6443 because that feature is already a GCC extension and thus not relevant to
6444 traditional C compatibility.
6447 @item -Wtraditional-conversion @r{(C and Objective-C only)}
6448 @opindex Wtraditional-conversion
6449 @opindex Wno-traditional-conversion
6450 Warn if a prototype causes a type conversion that is different from what
6451 would happen to the same argument in the absence of a prototype. This
6452 includes conversions of fixed point to floating and vice versa, and
6453 conversions changing the width or signedness of a fixed-point argument
6454 except when the same as the default promotion.
6456 @item -Wdeclaration-after-statement @r{(C and Objective-C only)}
6457 @opindex Wdeclaration-after-statement
6458 @opindex Wno-declaration-after-statement
6459 Warn when a declaration is found after a statement in a block. This
6460 construct, known from C++, was introduced with ISO C99 and is by default
6461 allowed in GCC@. It is not supported by ISO C90. @xref{Mixed Declarations}.
6466 Warn whenever a local variable or type declaration shadows another
6467 variable, parameter, type, class member (in C++), or instance variable
6468 (in Objective-C) or whenever a built-in function is shadowed. Note
6469 that in C++, the compiler warns if a local variable shadows an
6470 explicit typedef, but not if it shadows a struct/class/enum.
6471 Same as @option{-Wshadow=global}.
6473 @item -Wno-shadow-ivar @r{(Objective-C only)}
6474 @opindex Wno-shadow-ivar
6475 @opindex Wshadow-ivar
6476 Do not warn whenever a local variable shadows an instance variable in an
6479 @item -Wshadow=global
6480 @opindex Wshadow=global
6481 The default for @option{-Wshadow}. Warns for any (global) shadowing.
6482 This warning is enabled by @option{-Wshadow=global}.
6484 @item -Wshadow=local
6485 @opindex Wshadow=local
6486 Warn when a local variable shadows another local variable or parameter.
6487 This warning is enabled by @option{-Wshadow=local}.
6489 @item -Wshadow=compatible-local
6490 @opindex Wshadow=compatible-local
6491 Warn when a local variable shadows another local variable or parameter
6492 whose type is compatible with that of the shadowing variable. In C++,
6493 type compatibility here means the type of the shadowing variable can be
6494 converted to that of the shadowed variable. The creation of this flag
6495 (in addition to @option{-Wshadow=local}) is based on the idea that when
6496 a local variable shadows another one of incompatible type, it is most
6497 likely intentional, not a bug or typo, as shown in the following example:
6501 for (SomeIterator i = SomeObj.begin(); i != SomeObj.end(); ++i)
6503 for (int i = 0; i < N; ++i)
6512 Since the two variable @code{i} in the example above have incompatible types,
6513 enabling only @option{-Wshadow=compatible-local} will not emit a warning.
6514 Because their types are incompatible, if a programmer accidentally uses one
6515 in place of the other, type checking will catch that and emit an error or
6516 warning. So not warning (about shadowing) in this case will not lead to
6517 undetected bugs. Use of this flag instead of @option{-Wshadow=local} can
6518 possibly reduce the number of warnings triggered by intentional shadowing.
6519 Note that this does also mean that shadowing @code{const char *i} by
6520 @code{char *i} will not emit a warning.
6522 This warning is enabled by @option{-Wshadow=compatible-local}.
6524 @item -Wlarger-than=@var{byte-size}
6525 @opindex Wlarger-than=
6526 @opindex Wlarger-than-@var{byte-size}
6527 Warn whenever an object is defined whose size exceeds @var{byte-size}.
6528 @option{-Wlarger-than=}@samp{PTRDIFF_MAX} is enabled by default.
6529 Warnings controlled by the option can be disabled either by specifying
6530 @var{byte-size} of @samp{SIZE_MAX} or more or by
6531 @option{-Wno-larger-than}.
6533 @item -Wno-larger-than
6534 @opindex Wno-larger-than
6535 Disable @option{-Wlarger-than=} warnings. The option is equivalent
6536 to @option{-Wlarger-than=}@samp{SIZE_MAX} or larger.
6538 @item -Wframe-larger-than=@var{byte-size}
6539 @opindex Wframe-larger-than=
6540 @opindex Wno-frame-larger-than
6541 Warn if the size of a function frame exceeds @var{byte-size}.
6542 The computation done to determine the stack frame size is approximate
6543 and not conservative.
6544 The actual requirements may be somewhat greater than @var{byte-size}
6545 even if you do not get a warning. In addition, any space allocated
6546 via @code{alloca}, variable-length arrays, or related constructs
6547 is not included by the compiler when determining
6548 whether or not to issue a warning.
6549 @option{-Wframe-larger-than=}@samp{PTRDIFF_MAX} is enabled by default.
6550 Warnings controlled by the option can be disabled either by specifying
6551 @var{byte-size} of @samp{SIZE_MAX} or more or by
6552 @option{-Wno-frame-larger-than}.
6554 @item -Wno-frame-larger-than
6555 @opindex Wno-frame-larger-than
6556 Disable @option{-Wframe-larger-than=} warnings. The option is equivalent
6557 to @option{-Wframe-larger-than=}@samp{SIZE_MAX} or larger.
6559 @item -Wno-free-nonheap-object
6560 @opindex Wno-free-nonheap-object
6561 @opindex Wfree-nonheap-object
6562 Do not warn when attempting to free an object that was not allocated
6565 @item -Wstack-usage=@var{byte-size}
6566 @opindex Wstack-usage
6567 @opindex Wno-stack-usage
6568 Warn if the stack usage of a function might exceed @var{byte-size}.
6569 The computation done to determine the stack usage is conservative.
6570 Any space allocated via @code{alloca}, variable-length arrays, or related
6571 constructs is included by the compiler when determining whether or not to
6574 The message is in keeping with the output of @option{-fstack-usage}.
6578 If the stack usage is fully static but exceeds the specified amount, it's:
6581 warning: stack usage is 1120 bytes
6584 If the stack usage is (partly) dynamic but bounded, it's:
6587 warning: stack usage might be 1648 bytes
6590 If the stack usage is (partly) dynamic and not bounded, it's:
6593 warning: stack usage might be unbounded
6597 @option{-Wstack-usage=}@samp{PTRDIFF_MAX} is enabled by default.
6598 Warnings controlled by the option can be disabled either by specifying
6599 @var{byte-size} of @samp{SIZE_MAX} or more or by
6600 @option{-Wno-stack-usage}.
6602 @item -Wno-stack-usage
6603 @opindex Wno-stack-usage
6604 Disable @option{-Wstack-usage=} warnings. The option is equivalent
6605 to @option{-Wstack-usage=}@samp{SIZE_MAX} or larger.
6607 @item -Wunsafe-loop-optimizations
6608 @opindex Wunsafe-loop-optimizations
6609 @opindex Wno-unsafe-loop-optimizations
6610 Warn if the loop cannot be optimized because the compiler cannot
6611 assume anything on the bounds of the loop indices. With
6612 @option{-funsafe-loop-optimizations} warn if the compiler makes
6615 @item -Wno-pedantic-ms-format @r{(MinGW targets only)}
6616 @opindex Wno-pedantic-ms-format
6617 @opindex Wpedantic-ms-format
6618 When used in combination with @option{-Wformat}
6619 and @option{-pedantic} without GNU extensions, this option
6620 disables the warnings about non-ISO @code{printf} / @code{scanf} format
6621 width specifiers @code{I32}, @code{I64}, and @code{I} used on Windows targets,
6622 which depend on the MS runtime.
6625 @opindex Waligned-new
6626 @opindex Wno-aligned-new
6627 Warn about a new-expression of a type that requires greater alignment
6628 than the @code{alignof(std::max_align_t)} but uses an allocation
6629 function without an explicit alignment parameter. This option is
6630 enabled by @option{-Wall}.
6632 Normally this only warns about global allocation functions, but
6633 @option{-Waligned-new=all} also warns about class member allocation
6636 @item -Wplacement-new
6637 @itemx -Wplacement-new=@var{n}
6638 @opindex Wplacement-new
6639 @opindex Wno-placement-new
6640 Warn about placement new expressions with undefined behavior, such as
6641 constructing an object in a buffer that is smaller than the type of
6642 the object. For example, the placement new expression below is diagnosed
6643 because it attempts to construct an array of 64 integers in a buffer only
6649 This warning is enabled by default.
6652 @item -Wplacement-new=1
6653 This is the default warning level of @option{-Wplacement-new}. At this
6654 level the warning is not issued for some strictly undefined constructs that
6655 GCC allows as extensions for compatibility with legacy code. For example,
6656 the following @code{new} expression is not diagnosed at this level even
6657 though it has undefined behavior according to the C++ standard because
6658 it writes past the end of the one-element array.
6660 struct S @{ int n, a[1]; @};
6661 S *s = (S *)malloc (sizeof *s + 31 * sizeof s->a[0]);
6662 new (s->a)int [32]();
6665 @item -Wplacement-new=2
6666 At this level, in addition to diagnosing all the same constructs as at level
6667 1, a diagnostic is also issued for placement new expressions that construct
6668 an object in the last member of structure whose type is an array of a single
6669 element and whose size is less than the size of the object being constructed.
6670 While the previous example would be diagnosed, the following construct makes
6671 use of the flexible member array extension to avoid the warning at level 2.
6673 struct S @{ int n, a[]; @};
6674 S *s = (S *)malloc (sizeof *s + 32 * sizeof s->a[0]);
6675 new (s->a)int [32]();
6680 @item -Wpointer-arith
6681 @opindex Wpointer-arith
6682 @opindex Wno-pointer-arith
6683 Warn about anything that depends on the ``size of'' a function type or
6684 of @code{void}. GNU C assigns these types a size of 1, for
6685 convenience in calculations with @code{void *} pointers and pointers
6686 to functions. In C++, warn also when an arithmetic operation involves
6687 @code{NULL}. This warning is also enabled by @option{-Wpedantic}.
6689 @item -Wpointer-compare
6690 @opindex Wpointer-compare
6691 @opindex Wno-pointer-compare
6692 Warn if a pointer is compared with a zero character constant. This usually
6693 means that the pointer was meant to be dereferenced. For example:
6696 const char *p = foo ();
6701 Note that the code above is invalid in C++11.
6703 This warning is enabled by default.
6706 @opindex Wtype-limits
6707 @opindex Wno-type-limits
6708 Warn if a comparison is always true or always false due to the limited
6709 range of the data type, but do not warn for constant expressions. For
6710 example, warn if an unsigned variable is compared against zero with
6711 @code{<} or @code{>=}. This warning is also enabled by
6714 @item -Wabsolute-value @r{(C and Objective-C only)}
6715 @opindex Wabsolute-value
6716 @opindex Wno-absolute-value
6717 Warn for calls to standard functions that compute the absolute value
6718 of an argument when a more appropriate standard function is available.
6719 For example, calling @code{abs(3.14)} triggers the warning because the
6720 appropriate function to call to compute the absolute value of a double
6721 argument is @code{fabs}. The option also triggers warnings when the
6722 argument in a call to such a function has an unsigned type. This
6723 warning can be suppressed with an explicit type cast and it is also
6724 enabled by @option{-Wextra}.
6726 @include cppwarnopts.texi
6728 @item -Wbad-function-cast @r{(C and Objective-C only)}
6729 @opindex Wbad-function-cast
6730 @opindex Wno-bad-function-cast
6731 Warn when a function call is cast to a non-matching type.
6732 For example, warn if a call to a function returning an integer type
6733 is cast to a pointer type.
6735 @item -Wc90-c99-compat @r{(C and Objective-C only)}
6736 @opindex Wc90-c99-compat
6737 @opindex Wno-c90-c99-compat
6738 Warn about features not present in ISO C90, but present in ISO C99.
6739 For instance, warn about use of variable length arrays, @code{long long}
6740 type, @code{bool} type, compound literals, designated initializers, and so
6741 on. This option is independent of the standards mode. Warnings are disabled
6742 in the expression that follows @code{__extension__}.
6744 @item -Wc99-c11-compat @r{(C and Objective-C only)}
6745 @opindex Wc99-c11-compat
6746 @opindex Wno-c99-c11-compat
6747 Warn about features not present in ISO C99, but present in ISO C11.
6748 For instance, warn about use of anonymous structures and unions,
6749 @code{_Atomic} type qualifier, @code{_Thread_local} storage-class specifier,
6750 @code{_Alignas} specifier, @code{Alignof} operator, @code{_Generic} keyword,
6751 and so on. This option is independent of the standards mode. Warnings are
6752 disabled in the expression that follows @code{__extension__}.
6754 @item -Wc++-compat @r{(C and Objective-C only)}
6755 @opindex Wc++-compat
6756 @opindex Wno-c++-compat
6757 Warn about ISO C constructs that are outside of the common subset of
6758 ISO C and ISO C++, e.g.@: request for implicit conversion from
6759 @code{void *} to a pointer to non-@code{void} type.
6761 @item -Wc++11-compat @r{(C++ and Objective-C++ only)}
6762 @opindex Wc++11-compat
6763 @opindex Wno-c++11-compat
6764 Warn about C++ constructs whose meaning differs between ISO C++ 1998
6765 and ISO C++ 2011, e.g., identifiers in ISO C++ 1998 that are keywords
6766 in ISO C++ 2011. This warning turns on @option{-Wnarrowing} and is
6767 enabled by @option{-Wall}.
6769 @item -Wc++14-compat @r{(C++ and Objective-C++ only)}
6770 @opindex Wc++14-compat
6771 @opindex Wno-c++14-compat
6772 Warn about C++ constructs whose meaning differs between ISO C++ 2011
6773 and ISO C++ 2014. This warning is enabled by @option{-Wall}.
6775 @item -Wc++17-compat @r{(C++ and Objective-C++ only)}
6776 @opindex Wc++17-compat
6777 @opindex Wno-c++17-compat
6778 Warn about C++ constructs whose meaning differs between ISO C++ 2014
6779 and ISO C++ 2017. This warning is enabled by @option{-Wall}.
6781 @item -Wc++20-compat @r{(C++ and Objective-C++ only)}
6782 @opindex Wc++20-compat
6783 @opindex Wno-c++20-compat
6784 Warn about C++ constructs whose meaning differs between ISO C++ 2017
6785 and ISO C++ 2020. This warning is enabled by @option{-Wall}.
6789 @opindex Wno-cast-qual
6790 Warn whenever a pointer is cast so as to remove a type qualifier from
6791 the target type. For example, warn if a @code{const char *} is cast
6792 to an ordinary @code{char *}.
6794 Also warn when making a cast that introduces a type qualifier in an
6795 unsafe way. For example, casting @code{char **} to @code{const char **}
6796 is unsafe, as in this example:
6799 /* p is char ** value. */
6800 const char **q = (const char **) p;
6801 /* Assignment of readonly string to const char * is OK. */
6803 /* Now char** pointer points to read-only memory. */
6808 @opindex Wcast-align
6809 @opindex Wno-cast-align
6810 Warn whenever a pointer is cast such that the required alignment of the
6811 target is increased. For example, warn if a @code{char *} is cast to
6812 an @code{int *} on machines where integers can only be accessed at
6813 two- or four-byte boundaries.
6815 @item -Wcast-align=strict
6816 @opindex Wcast-align=strict
6817 Warn whenever a pointer is cast such that the required alignment of the
6818 target is increased. For example, warn if a @code{char *} is cast to
6819 an @code{int *} regardless of the target machine.
6821 @item -Wcast-function-type
6822 @opindex Wcast-function-type
6823 @opindex Wno-cast-function-type
6824 Warn when a function pointer is cast to an incompatible function pointer.
6825 In a cast involving function types with a variable argument list only
6826 the types of initial arguments that are provided are considered.
6827 Any parameter of pointer-type matches any other pointer-type. Any benign
6828 differences in integral types are ignored, like @code{int} vs.@: @code{long}
6829 on ILP32 targets. Likewise type qualifiers are ignored. The function
6830 type @code{void (*) (void)} is special and matches everything, which can
6831 be used to suppress this warning.
6832 In a cast involving pointer to member types this warning warns whenever
6833 the type cast is changing the pointer to member type.
6834 This warning is enabled by @option{-Wextra}.
6836 @item -Wwrite-strings
6837 @opindex Wwrite-strings
6838 @opindex Wno-write-strings
6839 When compiling C, give string constants the type @code{const
6840 char[@var{length}]} so that copying the address of one into a
6841 non-@code{const} @code{char *} pointer produces a warning. These
6842 warnings help you find at compile time code that can try to write
6843 into a string constant, but only if you have been very careful about
6844 using @code{const} in declarations and prototypes. Otherwise, it is
6845 just a nuisance. This is why we did not make @option{-Wall} request
6848 When compiling C++, warn about the deprecated conversion from string
6849 literals to @code{char *}. This warning is enabled by default for C++
6853 @itemx -Wcatch-value=@var{n} @r{(C++ and Objective-C++ only)}
6854 @opindex Wcatch-value
6855 @opindex Wno-catch-value
6856 Warn about catch handlers that do not catch via reference.
6857 With @option{-Wcatch-value=1} (or @option{-Wcatch-value} for short)
6858 warn about polymorphic class types that are caught by value.
6859 With @option{-Wcatch-value=2} warn about all class types that are caught
6860 by value. With @option{-Wcatch-value=3} warn about all types that are
6861 not caught by reference. @option{-Wcatch-value} is enabled by @option{-Wall}.
6865 @opindex Wno-clobbered
6866 Warn for variables that might be changed by @code{longjmp} or
6867 @code{vfork}. This warning is also enabled by @option{-Wextra}.
6869 @item -Wconditionally-supported @r{(C++ and Objective-C++ only)}
6870 @opindex Wconditionally-supported
6871 @opindex Wno-conditionally-supported
6872 Warn for conditionally-supported (C++11 [intro.defs]) constructs.
6875 @opindex Wconversion
6876 @opindex Wno-conversion
6877 Warn for implicit conversions that may alter a value. This includes
6878 conversions between real and integer, like @code{abs (x)} when
6879 @code{x} is @code{double}; conversions between signed and unsigned,
6880 like @code{unsigned ui = -1}; and conversions to smaller types, like
6881 @code{sqrtf (M_PI)}. Do not warn for explicit casts like @code{abs
6882 ((int) x)} and @code{ui = (unsigned) -1}, or if the value is not
6883 changed by the conversion like in @code{abs (2.0)}. Warnings about
6884 conversions between signed and unsigned integers can be disabled by
6885 using @option{-Wno-sign-conversion}.
6887 For C++, also warn for confusing overload resolution for user-defined
6888 conversions; and conversions that never use a type conversion
6889 operator: conversions to @code{void}, the same type, a base class or a
6890 reference to them. Warnings about conversions between signed and
6891 unsigned integers are disabled by default in C++ unless
6892 @option{-Wsign-conversion} is explicitly enabled.
6894 @item -Wno-conversion-null @r{(C++ and Objective-C++ only)}
6895 @opindex Wconversion-null
6896 @opindex Wno-conversion-null
6897 Do not warn for conversions between @code{NULL} and non-pointer
6898 types. @option{-Wconversion-null} is enabled by default.
6900 @item -Wzero-as-null-pointer-constant @r{(C++ and Objective-C++ only)}
6901 @opindex Wzero-as-null-pointer-constant
6902 @opindex Wno-zero-as-null-pointer-constant
6903 Warn when a literal @samp{0} is used as null pointer constant. This can
6904 be useful to facilitate the conversion to @code{nullptr} in C++11.
6906 @item -Wsubobject-linkage @r{(C++ and Objective-C++ only)}
6907 @opindex Wsubobject-linkage
6908 @opindex Wno-subobject-linkage
6909 Warn if a class type has a base or a field whose type uses the anonymous
6910 namespace or depends on a type with no linkage. If a type A depends on
6911 a type B with no or internal linkage, defining it in multiple
6912 translation units would be an ODR violation because the meaning of B
6913 is different in each translation unit. If A only appears in a single
6914 translation unit, the best way to silence the warning is to give it
6915 internal linkage by putting it in an anonymous namespace as well. The
6916 compiler doesn't give this warning for types defined in the main .C
6917 file, as those are unlikely to have multiple definitions.
6918 @option{-Wsubobject-linkage} is enabled by default.
6920 @item -Wdangling-else
6921 @opindex Wdangling-else
6922 @opindex Wno-dangling-else
6923 Warn about constructions where there may be confusion to which
6924 @code{if} statement an @code{else} branch belongs. Here is an example of
6939 In C/C++, every @code{else} branch belongs to the innermost possible
6940 @code{if} statement, which in this example is @code{if (b)}. This is
6941 often not what the programmer expected, as illustrated in the above
6942 example by indentation the programmer chose. When there is the
6943 potential for this confusion, GCC issues a warning when this flag
6944 is specified. To eliminate the warning, add explicit braces around
6945 the innermost @code{if} statement so there is no way the @code{else}
6946 can belong to the enclosing @code{if}. The resulting code
6963 This warning is enabled by @option{-Wparentheses}.
6967 @opindex Wno-date-time
6968 Warn when macros @code{__TIME__}, @code{__DATE__} or @code{__TIMESTAMP__}
6969 are encountered as they might prevent bit-wise-identical reproducible
6972 @item -Wdelete-incomplete @r{(C++ and Objective-C++ only)}
6973 @opindex Wdelete-incomplete
6974 @opindex Wno-delete-incomplete
6975 Warn when deleting a pointer to incomplete type, which may cause
6976 undefined behavior at runtime. This warning is enabled by default.
6978 @item -Wuseless-cast @r{(C++ and Objective-C++ only)}
6979 @opindex Wuseless-cast
6980 @opindex Wno-useless-cast
6981 Warn when an expression is casted to its own type.
6984 @opindex Wempty-body
6985 @opindex Wno-empty-body
6986 Warn if an empty body occurs in an @code{if}, @code{else} or @code{do
6987 while} statement. This warning is also enabled by @option{-Wextra}.
6989 @item -Wenum-compare
6990 @opindex Wenum-compare
6991 @opindex Wno-enum-compare
6992 Warn about a comparison between values of different enumerated types.
6993 In C++ enumerated type mismatches in conditional expressions are also
6994 diagnosed and the warning is enabled by default. In C this warning is
6995 enabled by @option{-Wall}.
6997 @item -Wenum-conversion @r{(C, Objective-C only)}
6998 @opindex Wenum-conversion
6999 @opindex Wno-enum-conversion
7000 Warn when a value of enumerated type is implicitly converted to a
7001 different enumerated type. This warning is enabled by @option{-Wextra}.
7003 @item -Wextra-semi @r{(C++, Objective-C++ only)}
7004 @opindex Wextra-semi
7005 @opindex Wno-extra-semi
7006 Warn about redundant semicolon after in-class function definition.
7008 @item -Wjump-misses-init @r{(C, Objective-C only)}
7009 @opindex Wjump-misses-init
7010 @opindex Wno-jump-misses-init
7011 Warn if a @code{goto} statement or a @code{switch} statement jumps
7012 forward across the initialization of a variable, or jumps backward to a
7013 label after the variable has been initialized. This only warns about
7014 variables that are initialized when they are declared. This warning is
7015 only supported for C and Objective-C; in C++ this sort of branch is an
7018 @option{-Wjump-misses-init} is included in @option{-Wc++-compat}. It
7019 can be disabled with the @option{-Wno-jump-misses-init} option.
7021 @item -Wsign-compare
7022 @opindex Wsign-compare
7023 @opindex Wno-sign-compare
7024 @cindex warning for comparison of signed and unsigned values
7025 @cindex comparison of signed and unsigned values, warning
7026 @cindex signed and unsigned values, comparison warning
7027 Warn when a comparison between signed and unsigned values could produce
7028 an incorrect result when the signed value is converted to unsigned.
7029 In C++, this warning is also enabled by @option{-Wall}. In C, it is
7030 also enabled by @option{-Wextra}.
7032 @item -Wsign-conversion
7033 @opindex Wsign-conversion
7034 @opindex Wno-sign-conversion
7035 Warn for implicit conversions that may change the sign of an integer
7036 value, like assigning a signed integer expression to an unsigned
7037 integer variable. An explicit cast silences the warning. In C, this
7038 option is enabled also by @option{-Wconversion}.
7040 @item -Wfloat-conversion
7041 @opindex Wfloat-conversion
7042 @opindex Wno-float-conversion
7043 Warn for implicit conversions that reduce the precision of a real value.
7044 This includes conversions from real to integer, and from higher precision
7045 real to lower precision real values. This option is also enabled by
7046 @option{-Wconversion}.
7048 @item -Wno-scalar-storage-order
7049 @opindex Wno-scalar-storage-order
7050 @opindex Wscalar-storage-order
7051 Do not warn on suspicious constructs involving reverse scalar storage order.
7053 @item -Wsized-deallocation @r{(C++ and Objective-C++ only)}
7054 @opindex Wsized-deallocation
7055 @opindex Wno-sized-deallocation
7056 Warn about a definition of an unsized deallocation function
7058 void operator delete (void *) noexcept;
7059 void operator delete[] (void *) noexcept;
7061 without a definition of the corresponding sized deallocation function
7063 void operator delete (void *, std::size_t) noexcept;
7064 void operator delete[] (void *, std::size_t) noexcept;
7066 or vice versa. Enabled by @option{-Wextra} along with
7067 @option{-fsized-deallocation}.
7069 @item -Wsizeof-pointer-div
7070 @opindex Wsizeof-pointer-div
7071 @opindex Wno-sizeof-pointer-div
7072 Warn for suspicious divisions of two sizeof expressions that divide
7073 the pointer size by the element size, which is the usual way to compute
7074 the array size but won't work out correctly with pointers. This warning
7075 warns e.g.@: about @code{sizeof (ptr) / sizeof (ptr[0])} if @code{ptr} is
7076 not an array, but a pointer. This warning is enabled by @option{-Wall}.
7078 @item -Wsizeof-pointer-memaccess
7079 @opindex Wsizeof-pointer-memaccess
7080 @opindex Wno-sizeof-pointer-memaccess
7081 Warn for suspicious length parameters to certain string and memory built-in
7082 functions if the argument uses @code{sizeof}. This warning triggers for
7083 example for @code{memset (ptr, 0, sizeof (ptr));} if @code{ptr} is not
7084 an array, but a pointer, and suggests a possible fix, or about
7085 @code{memcpy (&foo, ptr, sizeof (&foo));}. @option{-Wsizeof-pointer-memaccess}
7086 also warns about calls to bounded string copy functions like @code{strncat}
7087 or @code{strncpy} that specify as the bound a @code{sizeof} expression of
7088 the source array. For example, in the following function the call to
7089 @code{strncat} specifies the size of the source string as the bound. That
7090 is almost certainly a mistake and so the call is diagnosed.
7092 void make_file (const char *name)
7094 char path[PATH_MAX];
7095 strncpy (path, name, sizeof path - 1);
7096 strncat (path, ".text", sizeof ".text");
7101 The @option{-Wsizeof-pointer-memaccess} option is enabled by @option{-Wall}.
7103 @item -Wsizeof-array-argument
7104 @opindex Wsizeof-array-argument
7105 @opindex Wno-sizeof-array-argument
7106 Warn when the @code{sizeof} operator is applied to a parameter that is
7107 declared as an array in a function definition. This warning is enabled by
7108 default for C and C++ programs.
7110 @item -Wmemset-elt-size
7111 @opindex Wmemset-elt-size
7112 @opindex Wno-memset-elt-size
7113 Warn for suspicious calls to the @code{memset} built-in function, if the
7114 first argument references an array, and the third argument is a number
7115 equal to the number of elements, but not equal to the size of the array
7116 in memory. This indicates that the user has omitted a multiplication by
7117 the element size. This warning is enabled by @option{-Wall}.
7119 @item -Wmemset-transposed-args
7120 @opindex Wmemset-transposed-args
7121 @opindex Wno-memset-transposed-args
7122 Warn for suspicious calls to the @code{memset} built-in function where
7123 the second argument is not zero and the third argument is zero. For
7124 example, the call @code{memset (buf, sizeof buf, 0)} is diagnosed because
7125 @code{memset (buf, 0, sizeof buf)} was meant instead. The diagnostic
7126 is only emitted if the third argument is a literal zero. Otherwise, if
7127 it is an expression that is folded to zero, or a cast of zero to some
7128 type, it is far less likely that the arguments have been mistakenly
7129 transposed and no warning is emitted. This warning is enabled
7134 @opindex Wno-address
7135 Warn about suspicious uses of memory addresses. These include using
7136 the address of a function in a conditional expression, such as
7137 @code{void func(void); if (func)}, and comparisons against the memory
7138 address of a string literal, such as @code{if (x == "abc")}. Such
7139 uses typically indicate a programmer error: the address of a function
7140 always evaluates to true, so their use in a conditional usually
7141 indicate that the programmer forgot the parentheses in a function
7142 call; and comparisons against string literals result in unspecified
7143 behavior and are not portable in C, so they usually indicate that the
7144 programmer intended to use @code{strcmp}. This warning is enabled by
7147 @item -Waddress-of-packed-member
7148 @opindex Waddress-of-packed-member
7149 @opindex Wno-address-of-packed-member
7150 Warn when the address of packed member of struct or union is taken,
7151 which usually results in an unaligned pointer value. This is
7155 @opindex Wlogical-op
7156 @opindex Wno-logical-op
7157 Warn about suspicious uses of logical operators in expressions.
7158 This includes using logical operators in contexts where a
7159 bit-wise operator is likely to be expected. Also warns when
7160 the operands of a logical operator are the same:
7163 if (a < 0 && a < 0) @{ @dots{} @}
7166 @item -Wlogical-not-parentheses
7167 @opindex Wlogical-not-parentheses
7168 @opindex Wno-logical-not-parentheses
7169 Warn about logical not used on the left hand side operand of a comparison.
7170 This option does not warn if the right operand is considered to be a boolean
7171 expression. Its purpose is to detect suspicious code like the following:
7175 if (!a > 1) @{ @dots{} @}
7178 It is possible to suppress the warning by wrapping the LHS into
7181 if ((!a) > 1) @{ @dots{} @}
7184 This warning is enabled by @option{-Wall}.
7186 @item -Waggregate-return
7187 @opindex Waggregate-return
7188 @opindex Wno-aggregate-return
7189 Warn if any functions that return structures or unions are defined or
7190 called. (In languages where you can return an array, this also elicits
7193 @item -Wno-aggressive-loop-optimizations
7194 @opindex Wno-aggressive-loop-optimizations
7195 @opindex Waggressive-loop-optimizations
7196 Warn if in a loop with constant number of iterations the compiler detects
7197 undefined behavior in some statement during one or more of the iterations.
7199 @item -Wno-attributes
7200 @opindex Wno-attributes
7201 @opindex Wattributes
7202 Do not warn if an unexpected @code{__attribute__} is used, such as
7203 unrecognized attributes, function attributes applied to variables,
7204 etc. This does not stop errors for incorrect use of supported
7207 @item -Wno-builtin-declaration-mismatch
7208 @opindex Wno-builtin-declaration-mismatch
7209 @opindex Wbuiltin-declaration-mismatch
7210 Warn if a built-in function is declared with an incompatible signature
7211 or as a non-function, or when a built-in function declared with a type
7212 that does not include a prototype is called with arguments whose promoted
7213 types do not match those expected by the function. When @option{-Wextra}
7214 is specified, also warn when a built-in function that takes arguments is
7215 declared without a prototype. The @option{-Wno-builtin-declaration-mismatch}
7216 warning is enabled by default. To avoid the warning include the appropriate
7217 header to bring the prototypes of built-in functions into scope.
7219 For example, the call to @code{memset} below is diagnosed by the warning
7220 because the function expects a value of type @code{size_t} as its argument
7221 but the type of @code{32} is @code{int}. With @option{-Wextra},
7222 the declaration of the function is diagnosed as well.
7224 extern void* memset ();
7227 memset (d, '\0', 32);
7231 @item -Wno-builtin-macro-redefined
7232 @opindex Wno-builtin-macro-redefined
7233 @opindex Wbuiltin-macro-redefined
7234 Do not warn if certain built-in macros are redefined. This suppresses
7235 warnings for redefinition of @code{__TIMESTAMP__}, @code{__TIME__},
7236 @code{__DATE__}, @code{__FILE__}, and @code{__BASE_FILE__}.
7238 @item -Wstrict-prototypes @r{(C and Objective-C only)}
7239 @opindex Wstrict-prototypes
7240 @opindex Wno-strict-prototypes
7241 Warn if a function is declared or defined without specifying the
7242 argument types. (An old-style function definition is permitted without
7243 a warning if preceded by a declaration that specifies the argument
7246 @item -Wold-style-declaration @r{(C and Objective-C only)}
7247 @opindex Wold-style-declaration
7248 @opindex Wno-old-style-declaration
7249 Warn for obsolescent usages, according to the C Standard, in a
7250 declaration. For example, warn if storage-class specifiers like
7251 @code{static} are not the first things in a declaration. This warning
7252 is also enabled by @option{-Wextra}.
7254 @item -Wold-style-definition @r{(C and Objective-C only)}
7255 @opindex Wold-style-definition
7256 @opindex Wno-old-style-definition
7257 Warn if an old-style function definition is used. A warning is given
7258 even if there is a previous prototype.
7260 @item -Wmissing-parameter-type @r{(C and Objective-C only)}
7261 @opindex Wmissing-parameter-type
7262 @opindex Wno-missing-parameter-type
7263 A function parameter is declared without a type specifier in K&R-style
7270 This warning is also enabled by @option{-Wextra}.
7272 @item -Wmissing-prototypes @r{(C and Objective-C only)}
7273 @opindex Wmissing-prototypes
7274 @opindex Wno-missing-prototypes
7275 Warn if a global function is defined without a previous prototype
7276 declaration. This warning is issued even if the definition itself
7277 provides a prototype. Use this option to detect global functions
7278 that do not have a matching prototype declaration in a header file.
7279 This option is not valid for C++ because all function declarations
7280 provide prototypes and a non-matching declaration declares an
7281 overload rather than conflict with an earlier declaration.
7282 Use @option{-Wmissing-declarations} to detect missing declarations in C++.
7284 @item -Wmissing-declarations
7285 @opindex Wmissing-declarations
7286 @opindex Wno-missing-declarations
7287 Warn if a global function is defined without a previous declaration.
7288 Do so even if the definition itself provides a prototype.
7289 Use this option to detect global functions that are not declared in
7290 header files. In C, no warnings are issued for functions with previous
7291 non-prototype declarations; use @option{-Wmissing-prototypes} to detect
7292 missing prototypes. In C++, no warnings are issued for function templates,
7293 or for inline functions, or for functions in anonymous namespaces.
7295 @item -Wmissing-field-initializers
7296 @opindex Wmissing-field-initializers
7297 @opindex Wno-missing-field-initializers
7301 Warn if a structure's initializer has some fields missing. For
7302 example, the following code causes such a warning, because
7303 @code{x.h} is implicitly zero:
7306 struct s @{ int f, g, h; @};
7307 struct s x = @{ 3, 4 @};
7310 This option does not warn about designated initializers, so the following
7311 modification does not trigger a warning:
7314 struct s @{ int f, g, h; @};
7315 struct s x = @{ .f = 3, .g = 4 @};
7318 In C this option does not warn about the universal zero initializer
7322 struct s @{ int f, g, h; @};
7323 struct s x = @{ 0 @};
7326 Likewise, in C++ this option does not warn about the empty @{ @}
7327 initializer, for example:
7330 struct s @{ int f, g, h; @};
7334 This warning is included in @option{-Wextra}. To get other @option{-Wextra}
7335 warnings without this one, use @option{-Wextra -Wno-missing-field-initializers}.
7337 @item -Wno-multichar
7338 @opindex Wno-multichar
7340 Do not warn if a multicharacter constant (@samp{'FOOF'}) is used.
7341 Usually they indicate a typo in the user's code, as they have
7342 implementation-defined values, and should not be used in portable code.
7344 @item -Wnormalized=@r{[}none@r{|}id@r{|}nfc@r{|}nfkc@r{]}
7345 @opindex Wnormalized=
7346 @opindex Wnormalized
7347 @opindex Wno-normalized
7350 @cindex character set, input normalization
7351 In ISO C and ISO C++, two identifiers are different if they are
7352 different sequences of characters. However, sometimes when characters
7353 outside the basic ASCII character set are used, you can have two
7354 different character sequences that look the same. To avoid confusion,
7355 the ISO 10646 standard sets out some @dfn{normalization rules} which
7356 when applied ensure that two sequences that look the same are turned into
7357 the same sequence. GCC can warn you if you are using identifiers that
7358 have not been normalized; this option controls that warning.
7360 There are four levels of warning supported by GCC@. The default is
7361 @option{-Wnormalized=nfc}, which warns about any identifier that is
7362 not in the ISO 10646 ``C'' normalized form, @dfn{NFC}. NFC is the
7363 recommended form for most uses. It is equivalent to
7364 @option{-Wnormalized}.
7366 Unfortunately, there are some characters allowed in identifiers by
7367 ISO C and ISO C++ that, when turned into NFC, are not allowed in
7368 identifiers. That is, there's no way to use these symbols in portable
7369 ISO C or C++ and have all your identifiers in NFC@.
7370 @option{-Wnormalized=id} suppresses the warning for these characters.
7371 It is hoped that future versions of the standards involved will correct
7372 this, which is why this option is not the default.
7374 You can switch the warning off for all characters by writing
7375 @option{-Wnormalized=none} or @option{-Wno-normalized}. You should
7376 only do this if you are using some other normalization scheme (like
7377 ``D''), because otherwise you can easily create bugs that are
7378 literally impossible to see.
7380 Some characters in ISO 10646 have distinct meanings but look identical
7381 in some fonts or display methodologies, especially once formatting has
7382 been applied. For instance @code{\u207F}, ``SUPERSCRIPT LATIN SMALL
7383 LETTER N'', displays just like a regular @code{n} that has been
7384 placed in a superscript. ISO 10646 defines the @dfn{NFKC}
7385 normalization scheme to convert all these into a standard form as
7386 well, and GCC warns if your code is not in NFKC if you use
7387 @option{-Wnormalized=nfkc}. This warning is comparable to warning
7388 about every identifier that contains the letter O because it might be
7389 confused with the digit 0, and so is not the default, but may be
7390 useful as a local coding convention if the programming environment
7391 cannot be fixed to display these characters distinctly.
7393 @item -Wno-attribute-warning
7394 @opindex Wno-attribute-warning
7395 @opindex Wattribute-warning
7396 Do not warn about usage of functions (@pxref{Function Attributes})
7397 declared with @code{warning} attribute. By default, this warning is
7398 enabled. @option{-Wno-attribute-warning} can be used to disable the
7399 warning or @option{-Wno-error=attribute-warning} can be used to
7400 disable the error when compiled with @option{-Werror} flag.
7402 @item -Wno-deprecated
7403 @opindex Wno-deprecated
7404 @opindex Wdeprecated
7405 Do not warn about usage of deprecated features. @xref{Deprecated Features}.
7407 @item -Wno-deprecated-declarations
7408 @opindex Wno-deprecated-declarations
7409 @opindex Wdeprecated-declarations
7410 Do not warn about uses of functions (@pxref{Function Attributes}),
7411 variables (@pxref{Variable Attributes}), and types (@pxref{Type
7412 Attributes}) marked as deprecated by using the @code{deprecated}
7416 @opindex Wno-overflow
7418 Do not warn about compile-time overflow in constant expressions.
7423 Warn about One Definition Rule violations during link-time optimization.
7427 @opindex Wopenmp-simd
7428 @opindex Wno-openmp-simd
7429 Warn if the vectorizer cost model overrides the OpenMP
7430 simd directive set by user. The @option{-fsimd-cost-model=unlimited}
7431 option can be used to relax the cost model.
7433 @item -Woverride-init @r{(C and Objective-C only)}
7434 @opindex Woverride-init
7435 @opindex Wno-override-init
7439 Warn if an initialized field without side effects is overridden when
7440 using designated initializers (@pxref{Designated Inits, , Designated
7443 This warning is included in @option{-Wextra}. To get other
7444 @option{-Wextra} warnings without this one, use @option{-Wextra
7445 -Wno-override-init}.
7447 @item -Woverride-init-side-effects @r{(C and Objective-C only)}
7448 @opindex Woverride-init-side-effects
7449 @opindex Wno-override-init-side-effects
7450 Warn if an initialized field with side effects is overridden when
7451 using designated initializers (@pxref{Designated Inits, , Designated
7452 Initializers}). This warning is enabled by default.
7457 Warn if a structure is given the packed attribute, but the packed
7458 attribute has no effect on the layout or size of the structure.
7459 Such structures may be mis-aligned for little benefit. For
7460 instance, in this code, the variable @code{f.x} in @code{struct bar}
7461 is misaligned even though @code{struct bar} does not itself
7462 have the packed attribute:
7469 @} __attribute__((packed));
7477 @item -Wpacked-bitfield-compat
7478 @opindex Wpacked-bitfield-compat
7479 @opindex Wno-packed-bitfield-compat
7480 The 4.1, 4.2 and 4.3 series of GCC ignore the @code{packed} attribute
7481 on bit-fields of type @code{char}. This has been fixed in GCC 4.4 but
7482 the change can lead to differences in the structure layout. GCC
7483 informs you when the offset of such a field has changed in GCC 4.4.
7484 For example there is no longer a 4-bit padding between field @code{a}
7485 and @code{b} in this structure:
7492 @} __attribute__ ((packed));
7495 This warning is enabled by default. Use
7496 @option{-Wno-packed-bitfield-compat} to disable this warning.
7498 @item -Wpacked-not-aligned @r{(C, C++, Objective-C and Objective-C++ only)}
7499 @opindex Wpacked-not-aligned
7500 @opindex Wno-packed-not-aligned
7501 Warn if a structure field with explicitly specified alignment in a
7502 packed struct or union is misaligned. For example, a warning will
7503 be issued on @code{struct S}, like, @code{warning: alignment 1 of
7504 'struct S' is less than 8}, in this code:
7508 struct __attribute__ ((aligned (8))) S8 @{ char a[8]; @};
7509 struct __attribute__ ((packed)) S @{
7515 This warning is enabled by @option{-Wall}.
7520 Warn if padding is included in a structure, either to align an element
7521 of the structure or to align the whole structure. Sometimes when this
7522 happens it is possible to rearrange the fields of the structure to
7523 reduce the padding and so make the structure smaller.
7525 @item -Wredundant-decls
7526 @opindex Wredundant-decls
7527 @opindex Wno-redundant-decls
7528 Warn if anything is declared more than once in the same scope, even in
7529 cases where multiple declaration is valid and changes nothing.
7533 @opindex Wno-restrict
7534 Warn when an object referenced by a @code{restrict}-qualified parameter
7535 (or, in C++, a @code{__restrict}-qualified parameter) is aliased by another
7536 argument, or when copies between such objects overlap. For example,
7537 the call to the @code{strcpy} function below attempts to truncate the string
7538 by replacing its initial characters with the last four. However, because
7539 the call writes the terminating NUL into @code{a[4]}, the copies overlap and
7540 the call is diagnosed.
7545 char a[] = "abcd1234";
7550 The @option{-Wrestrict} option detects some instances of simple overlap
7551 even without optimization but works best at @option{-O2} and above. It
7552 is included in @option{-Wall}.
7554 @item -Wnested-externs @r{(C and Objective-C only)}
7555 @opindex Wnested-externs
7556 @opindex Wno-nested-externs
7557 Warn if an @code{extern} declaration is encountered within a function.
7559 @item -Wno-inherited-variadic-ctor
7560 @opindex Winherited-variadic-ctor
7561 @opindex Wno-inherited-variadic-ctor
7562 Suppress warnings about use of C++11 inheriting constructors when the
7563 base class inherited from has a C variadic constructor; the warning is
7564 on by default because the ellipsis is not inherited.
7569 Warn if a function that is declared as inline cannot be inlined.
7570 Even with this option, the compiler does not warn about failures to
7571 inline functions declared in system headers.
7573 The compiler uses a variety of heuristics to determine whether or not
7574 to inline a function. For example, the compiler takes into account
7575 the size of the function being inlined and the amount of inlining
7576 that has already been done in the current function. Therefore,
7577 seemingly insignificant changes in the source program can cause the
7578 warnings produced by @option{-Winline} to appear or disappear.
7580 @item -Wno-invalid-offsetof @r{(C++ and Objective-C++ only)}
7581 @opindex Wno-invalid-offsetof
7582 @opindex Winvalid-offsetof
7583 Suppress warnings from applying the @code{offsetof} macro to a non-POD
7584 type. According to the 2014 ISO C++ standard, applying @code{offsetof}
7585 to a non-standard-layout type is undefined. In existing C++ implementations,
7586 however, @code{offsetof} typically gives meaningful results.
7587 This flag is for users who are aware that they are
7588 writing nonportable code and who have deliberately chosen to ignore the
7591 The restrictions on @code{offsetof} may be relaxed in a future version
7592 of the C++ standard.
7594 @item -Wint-in-bool-context
7595 @opindex Wint-in-bool-context
7596 @opindex Wno-int-in-bool-context
7597 Warn for suspicious use of integer values where boolean values are expected,
7598 such as conditional expressions (?:) using non-boolean integer constants in
7599 boolean context, like @code{if (a <= b ? 2 : 3)}. Or left shifting of signed
7600 integers in boolean context, like @code{for (a = 0; 1 << a; a++);}. Likewise
7601 for all kinds of multiplications regardless of the data type.
7602 This warning is enabled by @option{-Wall}.
7604 @item -Wno-int-to-pointer-cast
7605 @opindex Wno-int-to-pointer-cast
7606 @opindex Wint-to-pointer-cast
7607 Suppress warnings from casts to pointer type of an integer of a
7608 different size. In C++, casting to a pointer type of smaller size is
7609 an error. @option{Wint-to-pointer-cast} is enabled by default.
7612 @item -Wno-pointer-to-int-cast @r{(C and Objective-C only)}
7613 @opindex Wno-pointer-to-int-cast
7614 @opindex Wpointer-to-int-cast
7615 Suppress warnings from casts from a pointer to an integer type of a
7619 @opindex Winvalid-pch
7620 @opindex Wno-invalid-pch
7621 Warn if a precompiled header (@pxref{Precompiled Headers}) is found in
7622 the search path but cannot be used.
7626 @opindex Wno-long-long
7627 Warn if @code{long long} type is used. This is enabled by either
7628 @option{-Wpedantic} or @option{-Wtraditional} in ISO C90 and C++98
7629 modes. To inhibit the warning messages, use @option{-Wno-long-long}.
7631 @item -Wvariadic-macros
7632 @opindex Wvariadic-macros
7633 @opindex Wno-variadic-macros
7634 Warn if variadic macros are used in ISO C90 mode, or if the GNU
7635 alternate syntax is used in ISO C99 mode. This is enabled by either
7636 @option{-Wpedantic} or @option{-Wtraditional}. To inhibit the warning
7637 messages, use @option{-Wno-variadic-macros}.
7641 @opindex Wno-varargs
7642 Warn upon questionable usage of the macros used to handle variable
7643 arguments like @code{va_start}. This is default. To inhibit the
7644 warning messages, use @option{-Wno-varargs}.
7646 @item -Wvector-operation-performance
7647 @opindex Wvector-operation-performance
7648 @opindex Wno-vector-operation-performance
7649 Warn if vector operation is not implemented via SIMD capabilities of the
7650 architecture. Mainly useful for the performance tuning.
7651 Vector operation can be implemented @code{piecewise}, which means that the
7652 scalar operation is performed on every vector element;
7653 @code{in parallel}, which means that the vector operation is implemented
7654 using scalars of wider type, which normally is more performance efficient;
7655 and @code{as a single scalar}, which means that vector fits into a
7658 @item -Wno-virtual-move-assign
7659 @opindex Wvirtual-move-assign
7660 @opindex Wno-virtual-move-assign
7661 Suppress warnings about inheriting from a virtual base with a
7662 non-trivial C++11 move assignment operator. This is dangerous because
7663 if the virtual base is reachable along more than one path, it is
7664 moved multiple times, which can mean both objects end up in the
7665 moved-from state. If the move assignment operator is written to avoid
7666 moving from a moved-from object, this warning can be disabled.
7671 Warn if a variable-length array is used in the code.
7672 @option{-Wno-vla} prevents the @option{-Wpedantic} warning of
7673 the variable-length array.
7675 @item -Wvla-larger-than=@var{byte-size}
7676 @opindex Wvla-larger-than=
7677 @opindex Wno-vla-larger-than
7678 If this option is used, the compiler will warn for declarations of
7679 variable-length arrays whose size is either unbounded, or bounded
7680 by an argument that allows the array size to exceed @var{byte-size}
7681 bytes. This is similar to how @option{-Walloca-larger-than=}@var{byte-size}
7682 works, but with variable-length arrays.
7684 Note that GCC may optimize small variable-length arrays of a known
7685 value into plain arrays, so this warning may not get triggered for
7688 @option{-Wvla-larger-than=}@samp{PTRDIFF_MAX} is enabled by default but
7689 is typically only effective when @option{-ftree-vrp} is active (default
7690 for @option{-O2} and above).
7692 See also @option{-Walloca-larger-than=@var{byte-size}}.
7694 @item -Wno-vla-larger-than
7695 @opindex Wno-vla-larger-than
7696 Disable @option{-Wvla-larger-than=} warnings. The option is equivalent
7697 to @option{-Wvla-larger-than=}@samp{SIZE_MAX} or larger.
7699 @item -Wvolatile-register-var
7700 @opindex Wvolatile-register-var
7701 @opindex Wno-volatile-register-var
7702 Warn if a register variable is declared volatile. The volatile
7703 modifier does not inhibit all optimizations that may eliminate reads
7704 and/or writes to register variables. This warning is enabled by
7707 @item -Wdisabled-optimization
7708 @opindex Wdisabled-optimization
7709 @opindex Wno-disabled-optimization
7710 Warn if a requested optimization pass is disabled. This warning does
7711 not generally indicate that there is anything wrong with your code; it
7712 merely indicates that GCC's optimizers are unable to handle the code
7713 effectively. Often, the problem is that your code is too big or too
7714 complex; GCC refuses to optimize programs when the optimization
7715 itself is likely to take inordinate amounts of time.
7717 @item -Wpointer-sign @r{(C and Objective-C only)}
7718 @opindex Wpointer-sign
7719 @opindex Wno-pointer-sign
7720 Warn for pointer argument passing or assignment with different signedness.
7721 This option is only supported for C and Objective-C@. It is implied by
7722 @option{-Wall} and by @option{-Wpedantic}, which can be disabled with
7723 @option{-Wno-pointer-sign}.
7725 @item -Wstack-protector
7726 @opindex Wstack-protector
7727 @opindex Wno-stack-protector
7728 This option is only active when @option{-fstack-protector} is active. It
7729 warns about functions that are not protected against stack smashing.
7731 @item -Woverlength-strings
7732 @opindex Woverlength-strings
7733 @opindex Wno-overlength-strings
7734 Warn about string constants that are longer than the ``minimum
7735 maximum'' length specified in the C standard. Modern compilers
7736 generally allow string constants that are much longer than the
7737 standard's minimum limit, but very portable programs should avoid
7738 using longer strings.
7740 The limit applies @emph{after} string constant concatenation, and does
7741 not count the trailing NUL@. In C90, the limit was 509 characters; in
7742 C99, it was raised to 4095. C++98 does not specify a normative
7743 minimum maximum, so we do not diagnose overlength strings in C++@.
7745 This option is implied by @option{-Wpedantic}, and can be disabled with
7746 @option{-Wno-overlength-strings}.
7748 @item -Wunsuffixed-float-constants @r{(C and Objective-C only)}
7749 @opindex Wunsuffixed-float-constants
7750 @opindex Wno-unsuffixed-float-constants
7752 Issue a warning for any floating constant that does not have
7753 a suffix. When used together with @option{-Wsystem-headers} it
7754 warns about such constants in system header files. This can be useful
7755 when preparing code to use with the @code{FLOAT_CONST_DECIMAL64} pragma
7756 from the decimal floating-point extension to C99.
7758 @item -Wno-designated-init @r{(C and Objective-C only)}
7759 Suppress warnings when a positional initializer is used to initialize
7760 a structure that has been marked with the @code{designated_init}
7764 Issue a warning when HSAIL cannot be emitted for the compiled function or
7769 @node Debugging Options
7770 @section Options for Debugging Your Program
7771 @cindex options, debugging
7772 @cindex debugging information options
7774 To tell GCC to emit extra information for use by a debugger, in almost
7775 all cases you need only to add @option{-g} to your other options.
7777 GCC allows you to use @option{-g} with
7778 @option{-O}. The shortcuts taken by optimized code may occasionally
7779 be surprising: some variables you declared may not exist
7780 at all; flow of control may briefly move where you did not expect it;
7781 some statements may not be executed because they compute constant
7782 results or their values are already at hand; some statements may
7783 execute in different places because they have been moved out of loops.
7784 Nevertheless it is possible to debug optimized output. This makes
7785 it reasonable to use the optimizer for programs that might have bugs.
7787 If you are not using some other optimization option, consider
7788 using @option{-Og} (@pxref{Optimize Options}) with @option{-g}.
7789 With no @option{-O} option at all, some compiler passes that collect
7790 information useful for debugging do not run at all, so that
7791 @option{-Og} may result in a better debugging experience.
7796 Produce debugging information in the operating system's native format
7797 (stabs, COFF, XCOFF, or DWARF)@. GDB can work with this debugging
7800 On most systems that use stabs format, @option{-g} enables use of extra
7801 debugging information that only GDB can use; this extra information
7802 makes debugging work better in GDB but probably makes other debuggers
7804 refuse to read the program. If you want to control for certain whether
7805 to generate the extra information, use @option{-gstabs+}, @option{-gstabs},
7806 @option{-gxcoff+}, @option{-gxcoff}, or @option{-gvms} (see below).
7810 Produce debugging information for use by GDB@. This means to use the
7811 most expressive format available (DWARF, stabs, or the native format
7812 if neither of those are supported), including GDB extensions if at all
7816 @itemx -gdwarf-@var{version}
7818 Produce debugging information in DWARF format (if that is supported).
7819 The value of @var{version} may be either 2, 3, 4 or 5; the default version
7820 for most targets is 4. DWARF Version 5 is only experimental.
7822 Note that with DWARF Version 2, some ports require and always
7823 use some non-conflicting DWARF 3 extensions in the unwind tables.
7825 Version 4 may require GDB 7.0 and @option{-fvar-tracking-assignments}
7826 for maximum benefit.
7828 GCC no longer supports DWARF Version 1, which is substantially
7829 different than Version 2 and later. For historical reasons, some
7830 other DWARF-related options such as
7831 @option{-fno-dwarf2-cfi-asm}) retain a reference to DWARF Version 2
7832 in their names, but apply to all currently-supported versions of DWARF.
7836 Produce debugging information in stabs format (if that is supported),
7837 without GDB extensions. This is the format used by DBX on most BSD
7838 systems. On MIPS, Alpha and System V Release 4 systems this option
7839 produces stabs debugging output that is not understood by DBX@.
7840 On System V Release 4 systems this option requires the GNU assembler.
7844 Produce debugging information in stabs format (if that is supported),
7845 using GNU extensions understood only by the GNU debugger (GDB)@. The
7846 use of these extensions is likely to make other debuggers crash or
7847 refuse to read the program.
7851 Produce debugging information in XCOFF format (if that is supported).
7852 This is the format used by the DBX debugger on IBM RS/6000 systems.
7856 Produce debugging information in XCOFF format (if that is supported),
7857 using GNU extensions understood only by the GNU debugger (GDB)@. The
7858 use of these extensions is likely to make other debuggers crash or
7859 refuse to read the program, and may cause assemblers other than the GNU
7860 assembler (GAS) to fail with an error.
7864 Produce debugging information in Alpha/VMS debug format (if that is
7865 supported). This is the format used by DEBUG on Alpha/VMS systems.
7868 @itemx -ggdb@var{level}
7869 @itemx -gstabs@var{level}
7870 @itemx -gxcoff@var{level}
7871 @itemx -gvms@var{level}
7872 Request debugging information and also use @var{level} to specify how
7873 much information. The default level is 2.
7875 Level 0 produces no debug information at all. Thus, @option{-g0} negates
7878 Level 1 produces minimal information, enough for making backtraces in
7879 parts of the program that you don't plan to debug. This includes
7880 descriptions of functions and external variables, and line number
7881 tables, but no information about local variables.
7883 Level 3 includes extra information, such as all the macro definitions
7884 present in the program. Some debuggers support macro expansion when
7885 you use @option{-g3}.
7887 If you use multiple @option{-g} options, with or without level numbers,
7888 the last such option is the one that is effective.
7890 @option{-gdwarf} does not accept a concatenated debug level, to avoid
7891 confusion with @option{-gdwarf-@var{level}}.
7892 Instead use an additional @option{-g@var{level}} option to change the
7893 debug level for DWARF.
7895 @item -fno-eliminate-unused-debug-symbols
7896 @opindex feliminate-unused-debug-symbols
7897 @opindex fno-eliminate-unused-debug-symbols
7898 By default, no debug information is produced for symbols that are not actually
7899 used. Use this option if you want debug information for all symbols.
7901 @item -femit-class-debug-always
7902 @opindex femit-class-debug-always
7903 Instead of emitting debugging information for a C++ class in only one
7904 object file, emit it in all object files using the class. This option
7905 should be used only with debuggers that are unable to handle the way GCC
7906 normally emits debugging information for classes because using this
7907 option increases the size of debugging information by as much as a
7910 @item -fno-merge-debug-strings
7911 @opindex fmerge-debug-strings
7912 @opindex fno-merge-debug-strings
7913 Direct the linker to not merge together strings in the debugging
7914 information that are identical in different object files. Merging is
7915 not supported by all assemblers or linkers. Merging decreases the size
7916 of the debug information in the output file at the cost of increasing
7917 link processing time. Merging is enabled by default.
7919 @item -fdebug-prefix-map=@var{old}=@var{new}
7920 @opindex fdebug-prefix-map
7921 When compiling files residing in directory @file{@var{old}}, record
7922 debugging information describing them as if the files resided in
7923 directory @file{@var{new}} instead. This can be used to replace a
7924 build-time path with an install-time path in the debug info. It can
7925 also be used to change an absolute path to a relative path by using
7926 @file{.} for @var{new}. This can give more reproducible builds, which
7927 are location independent, but may require an extra command to tell GDB
7928 where to find the source files. See also @option{-ffile-prefix-map}.
7930 @item -fvar-tracking
7931 @opindex fvar-tracking
7932 Run variable tracking pass. It computes where variables are stored at each
7933 position in code. Better debugging information is then generated
7934 (if the debugging information format supports this information).
7936 It is enabled by default when compiling with optimization (@option{-Os},
7937 @option{-O}, @option{-O2}, @dots{}), debugging information (@option{-g}) and
7938 the debug info format supports it.
7940 @item -fvar-tracking-assignments
7941 @opindex fvar-tracking-assignments
7942 @opindex fno-var-tracking-assignments
7943 Annotate assignments to user variables early in the compilation and
7944 attempt to carry the annotations over throughout the compilation all the
7945 way to the end, in an attempt to improve debug information while
7946 optimizing. Use of @option{-gdwarf-4} is recommended along with it.
7948 It can be enabled even if var-tracking is disabled, in which case
7949 annotations are created and maintained, but discarded at the end.
7950 By default, this flag is enabled together with @option{-fvar-tracking},
7951 except when selective scheduling is enabled.
7954 @opindex gsplit-dwarf
7955 Separate as much DWARF debugging information as possible into a
7956 separate output file with the extension @file{.dwo}. This option allows
7957 the build system to avoid linking files with debug information. To
7958 be useful, this option requires a debugger capable of reading @file{.dwo}
7961 @item -gdescribe-dies
7962 @opindex gdescribe-dies
7963 Add description attributes to some DWARF DIEs that have no name attribute,
7964 such as artificial variables, external references and call site
7969 Generate DWARF @code{.debug_pubnames} and @code{.debug_pubtypes} sections.
7971 @item -ggnu-pubnames
7972 @opindex ggnu-pubnames
7973 Generate @code{.debug_pubnames} and @code{.debug_pubtypes} sections in a format
7974 suitable for conversion into a GDB@ index. This option is only useful
7975 with a linker that can produce GDB@ index version 7.
7977 @item -fdebug-types-section
7978 @opindex fdebug-types-section
7979 @opindex fno-debug-types-section
7980 When using DWARF Version 4 or higher, type DIEs can be put into
7981 their own @code{.debug_types} section instead of making them part of the
7982 @code{.debug_info} section. It is more efficient to put them in a separate
7983 comdat section since the linker can then remove duplicates.
7984 But not all DWARF consumers support @code{.debug_types} sections yet
7985 and on some objects @code{.debug_types} produces larger instead of smaller
7986 debugging information.
7988 @item -grecord-gcc-switches
7989 @itemx -gno-record-gcc-switches
7990 @opindex grecord-gcc-switches
7991 @opindex gno-record-gcc-switches
7992 This switch causes the command-line options used to invoke the
7993 compiler that may affect code generation to be appended to the
7994 DW_AT_producer attribute in DWARF debugging information. The options
7995 are concatenated with spaces separating them from each other and from
7996 the compiler version.
7997 It is enabled by default.
7998 See also @option{-frecord-gcc-switches} for another
7999 way of storing compiler options into the object file.
8001 @item -gstrict-dwarf
8002 @opindex gstrict-dwarf
8003 Disallow using extensions of later DWARF standard version than selected
8004 with @option{-gdwarf-@var{version}}. On most targets using non-conflicting
8005 DWARF extensions from later standard versions is allowed.
8007 @item -gno-strict-dwarf
8008 @opindex gno-strict-dwarf
8009 Allow using extensions of later DWARF standard version than selected with
8010 @option{-gdwarf-@var{version}}.
8012 @item -gas-loc-support
8013 @opindex gas-loc-support
8014 Inform the compiler that the assembler supports @code{.loc} directives.
8015 It may then use them for the assembler to generate DWARF2+ line number
8018 This is generally desirable, because assembler-generated line-number
8019 tables are a lot more compact than those the compiler can generate
8022 This option will be enabled by default if, at GCC configure time, the
8023 assembler was found to support such directives.
8025 @item -gno-as-loc-support
8026 @opindex gno-as-loc-support
8027 Force GCC to generate DWARF2+ line number tables internally, if DWARF2+
8028 line number tables are to be generated.
8030 @item -gas-locview-support
8031 @opindex gas-locview-support
8032 Inform the compiler that the assembler supports @code{view} assignment
8033 and reset assertion checking in @code{.loc} directives.
8035 This option will be enabled by default if, at GCC configure time, the
8036 assembler was found to support them.
8038 @item -gno-as-locview-support
8039 Force GCC to assign view numbers internally, if
8040 @option{-gvariable-location-views} are explicitly requested.
8043 @itemx -gno-column-info
8044 @opindex gcolumn-info
8045 @opindex gno-column-info
8046 Emit location column information into DWARF debugging information, rather
8047 than just file and line.
8048 This option is enabled by default.
8050 @item -gstatement-frontiers
8051 @itemx -gno-statement-frontiers
8052 @opindex gstatement-frontiers
8053 @opindex gno-statement-frontiers
8054 This option causes GCC to create markers in the internal representation
8055 at the beginning of statements, and to keep them roughly in place
8056 throughout compilation, using them to guide the output of @code{is_stmt}
8057 markers in the line number table. This is enabled by default when
8058 compiling with optimization (@option{-Os}, @option{-O}, @option{-O2},
8059 @dots{}), and outputting DWARF 2 debug information at the normal level.
8061 @item -gvariable-location-views
8062 @itemx -gvariable-location-views=incompat5
8063 @itemx -gno-variable-location-views
8064 @opindex gvariable-location-views
8065 @opindex gvariable-location-views=incompat5
8066 @opindex gno-variable-location-views
8067 Augment variable location lists with progressive view numbers implied
8068 from the line number table. This enables debug information consumers to
8069 inspect state at certain points of the program, even if no instructions
8070 associated with the corresponding source locations are present at that
8071 point. If the assembler lacks support for view numbers in line number
8072 tables, this will cause the compiler to emit the line number table,
8073 which generally makes them somewhat less compact. The augmented line
8074 number tables and location lists are fully backward-compatible, so they
8075 can be consumed by debug information consumers that are not aware of
8076 these augmentations, but they won't derive any benefit from them either.
8078 This is enabled by default when outputting DWARF 2 debug information at
8079 the normal level, as long as there is assembler support,
8080 @option{-fvar-tracking-assignments} is enabled and
8081 @option{-gstrict-dwarf} is not. When assembler support is not
8082 available, this may still be enabled, but it will force GCC to output
8083 internal line number tables, and if
8084 @option{-ginternal-reset-location-views} is not enabled, that will most
8085 certainly lead to silently mismatching location views.
8087 There is a proposed representation for view numbers that is not backward
8088 compatible with the location list format introduced in DWARF 5, that can
8089 be enabled with @option{-gvariable-location-views=incompat5}. This
8090 option may be removed in the future, is only provided as a reference
8091 implementation of the proposed representation. Debug information
8092 consumers are not expected to support this extended format, and they
8093 would be rendered unable to decode location lists using it.
8095 @item -ginternal-reset-location-views
8096 @itemx -gnointernal-reset-location-views
8097 @opindex ginternal-reset-location-views
8098 @opindex gno-internal-reset-location-views
8099 Attempt to determine location views that can be omitted from location
8100 view lists. This requires the compiler to have very accurate insn
8101 length estimates, which isn't always the case, and it may cause
8102 incorrect view lists to be generated silently when using an assembler
8103 that does not support location view lists. The GNU assembler will flag
8104 any such error as a @code{view number mismatch}. This is only enabled
8105 on ports that define a reliable estimation function.
8107 @item -ginline-points
8108 @itemx -gno-inline-points
8109 @opindex ginline-points
8110 @opindex gno-inline-points
8111 Generate extended debug information for inlined functions. Location
8112 view tracking markers are inserted at inlined entry points, so that
8113 address and view numbers can be computed and output in debug
8114 information. This can be enabled independently of location views, in
8115 which case the view numbers won't be output, but it can only be enabled
8116 along with statement frontiers, and it is only enabled by default if
8117 location views are enabled.
8119 @item -gz@r{[}=@var{type}@r{]}
8121 Produce compressed debug sections in DWARF format, if that is supported.
8122 If @var{type} is not given, the default type depends on the capabilities
8123 of the assembler and linker used. @var{type} may be one of
8124 @samp{none} (don't compress debug sections), @samp{zlib} (use zlib
8125 compression in ELF gABI format), or @samp{zlib-gnu} (use zlib
8126 compression in traditional GNU format). If the linker doesn't support
8127 writing compressed debug sections, the option is rejected. Otherwise,
8128 if the assembler does not support them, @option{-gz} is silently ignored
8129 when producing object files.
8131 @item -femit-struct-debug-baseonly
8132 @opindex femit-struct-debug-baseonly
8133 Emit debug information for struct-like types
8134 only when the base name of the compilation source file
8135 matches the base name of file in which the struct is defined.
8137 This option substantially reduces the size of debugging information,
8138 but at significant potential loss in type information to the debugger.
8139 See @option{-femit-struct-debug-reduced} for a less aggressive option.
8140 See @option{-femit-struct-debug-detailed} for more detailed control.
8142 This option works only with DWARF debug output.
8144 @item -femit-struct-debug-reduced
8145 @opindex femit-struct-debug-reduced
8146 Emit debug information for struct-like types
8147 only when the base name of the compilation source file
8148 matches the base name of file in which the type is defined,
8149 unless the struct is a template or defined in a system header.
8151 This option significantly reduces the size of debugging information,
8152 with some potential loss in type information to the debugger.
8153 See @option{-femit-struct-debug-baseonly} for a more aggressive option.
8154 See @option{-femit-struct-debug-detailed} for more detailed control.
8156 This option works only with DWARF debug output.
8158 @item -femit-struct-debug-detailed@r{[}=@var{spec-list}@r{]}
8159 @opindex femit-struct-debug-detailed
8160 Specify the struct-like types
8161 for which the compiler generates debug information.
8162 The intent is to reduce duplicate struct debug information
8163 between different object files within the same program.
8165 This option is a detailed version of
8166 @option{-femit-struct-debug-reduced} and @option{-femit-struct-debug-baseonly},
8167 which serves for most needs.
8169 A specification has the syntax@*
8170 [@samp{dir:}|@samp{ind:}][@samp{ord:}|@samp{gen:}](@samp{any}|@samp{sys}|@samp{base}|@samp{none})
8172 The optional first word limits the specification to
8173 structs that are used directly (@samp{dir:}) or used indirectly (@samp{ind:}).
8174 A struct type is used directly when it is the type of a variable, member.
8175 Indirect uses arise through pointers to structs.
8176 That is, when use of an incomplete struct is valid, the use is indirect.
8178 @samp{struct one direct; struct two * indirect;}.
8180 The optional second word limits the specification to
8181 ordinary structs (@samp{ord:}) or generic structs (@samp{gen:}).
8182 Generic structs are a bit complicated to explain.
8183 For C++, these are non-explicit specializations of template classes,
8184 or non-template classes within the above.
8185 Other programming languages have generics,
8186 but @option{-femit-struct-debug-detailed} does not yet implement them.
8188 The third word specifies the source files for those
8189 structs for which the compiler should emit debug information.
8190 The values @samp{none} and @samp{any} have the normal meaning.
8191 The value @samp{base} means that
8192 the base of name of the file in which the type declaration appears
8193 must match the base of the name of the main compilation file.
8194 In practice, this means that when compiling @file{foo.c}, debug information
8195 is generated for types declared in that file and @file{foo.h},
8196 but not other header files.
8197 The value @samp{sys} means those types satisfying @samp{base}
8198 or declared in system or compiler headers.
8200 You may need to experiment to determine the best settings for your application.
8202 The default is @option{-femit-struct-debug-detailed=all}.
8204 This option works only with DWARF debug output.
8206 @item -fno-dwarf2-cfi-asm
8207 @opindex fdwarf2-cfi-asm
8208 @opindex fno-dwarf2-cfi-asm
8209 Emit DWARF unwind info as compiler generated @code{.eh_frame} section
8210 instead of using GAS @code{.cfi_*} directives.
8212 @item -fno-eliminate-unused-debug-types
8213 @opindex feliminate-unused-debug-types
8214 @opindex fno-eliminate-unused-debug-types
8215 Normally, when producing DWARF output, GCC avoids producing debug symbol
8216 output for types that are nowhere used in the source file being compiled.
8217 Sometimes it is useful to have GCC emit debugging
8218 information for all types declared in a compilation
8219 unit, regardless of whether or not they are actually used
8220 in that compilation unit, for example
8221 if, in the debugger, you want to cast a value to a type that is
8222 not actually used in your program (but is declared). More often,
8223 however, this results in a significant amount of wasted space.
8226 @node Optimize Options
8227 @section Options That Control Optimization
8228 @cindex optimize options
8229 @cindex options, optimization
8231 These options control various sorts of optimizations.
8233 Without any optimization option, the compiler's goal is to reduce the
8234 cost of compilation and to make debugging produce the expected
8235 results. Statements are independent: if you stop the program with a
8236 breakpoint between statements, you can then assign a new value to any
8237 variable or change the program counter to any other statement in the
8238 function and get exactly the results you expect from the source
8241 Turning on optimization flags makes the compiler attempt to improve
8242 the performance and/or code size at the expense of compilation time
8243 and possibly the ability to debug the program.
8245 The compiler performs optimization based on the knowledge it has of the
8246 program. Compiling multiple files at once to a single output file mode allows
8247 the compiler to use information gained from all of the files when compiling
8250 Not all optimizations are controlled directly by a flag. Only
8251 optimizations that have a flag are listed in this section.
8253 Most optimizations are completely disabled at @option{-O0} or if an
8254 @option{-O} level is not set on the command line, even if individual
8255 optimization flags are specified. Similarly, @option{-Og} suppresses
8256 many optimization passes.
8258 Depending on the target and how GCC was configured, a slightly different
8259 set of optimizations may be enabled at each @option{-O} level than
8260 those listed here. You can invoke GCC with @option{-Q --help=optimizers}
8261 to find out the exact set of optimizations that are enabled at each level.
8262 @xref{Overall Options}, for examples.
8269 Optimize. Optimizing compilation takes somewhat more time, and a lot
8270 more memory for a large function.
8272 With @option{-O}, the compiler tries to reduce code size and execution
8273 time, without performing any optimizations that take a great deal of
8276 @c Note that in addition to the default_options_table list in opts.c,
8277 @c several optimization flags default to true but control optimization
8278 @c passes that are explicitly disabled at -O0.
8280 @option{-O} turns on the following optimization flags:
8282 @c Please keep the following list alphabetized.
8283 @gccoptlist{-fauto-inc-dec @gol
8284 -fbranch-count-reg @gol
8285 -fcombine-stack-adjustments @gol
8287 -fcprop-registers @gol
8290 -fdelayed-branch @gol
8292 -fforward-propagate @gol
8293 -fguess-branch-probability @gol
8294 -fif-conversion @gol
8295 -fif-conversion2 @gol
8296 -finline-functions-called-once @gol
8298 -fipa-pure-const @gol
8299 -fipa-reference @gol
8300 -fipa-reference-addressable @gol
8301 -fmerge-constants @gol
8302 -fmove-loop-invariants @gol
8303 -fomit-frame-pointer @gol
8304 -freorder-blocks @gol
8306 -fshrink-wrap-separate @gol
8307 -fsplit-wide-types @gol
8313 -ftree-coalesce-vars @gol
8314 -ftree-copy-prop @gol
8316 -ftree-dominator-opts @gol
8318 -ftree-forwprop @gol
8322 -ftree-scev-cprop @gol
8331 Optimize even more. GCC performs nearly all supported optimizations
8332 that do not involve a space-speed tradeoff.
8333 As compared to @option{-O}, this option increases both compilation time
8334 and the performance of the generated code.
8336 @option{-O2} turns on all optimization flags specified by @option{-O}. It
8337 also turns on the following optimization flags:
8339 @c Please keep the following list alphabetized!
8340 @gccoptlist{-falign-functions -falign-jumps @gol
8341 -falign-labels -falign-loops @gol
8343 -fcode-hoisting @gol
8345 -fcse-follow-jumps -fcse-skip-blocks @gol
8346 -fdelete-null-pointer-checks @gol
8347 -fdevirtualize -fdevirtualize-speculatively @gol
8348 -fexpensive-optimizations @gol
8350 -fgcse -fgcse-lm @gol
8351 -fhoist-adjacent-loads @gol
8352 -finline-functions @gol
8353 -finline-small-functions @gol
8354 -findirect-inlining @gol
8355 -fipa-bit-cp -fipa-cp -fipa-icf @gol
8356 -fipa-ra -fipa-sra -fipa-vrp @gol
8357 -fisolate-erroneous-paths-dereference @gol
8359 -foptimize-sibling-calls @gol
8360 -foptimize-strlen @gol
8361 -fpartial-inlining @gol
8363 -freorder-blocks-algorithm=stc @gol
8364 -freorder-blocks-and-partition -freorder-functions @gol
8365 -frerun-cse-after-loop @gol
8366 -fschedule-insns -fschedule-insns2 @gol
8367 -fsched-interblock -fsched-spec @gol
8368 -fstore-merging @gol
8369 -fstrict-aliasing @gol
8371 -ftree-builtin-call-dce @gol
8373 -ftree-switch-conversion -ftree-tail-merge @gol
8376 Please note the warning under @option{-fgcse} about
8377 invoking @option{-O2} on programs that use computed gotos.
8381 Optimize yet more. @option{-O3} turns on all optimizations specified
8382 by @option{-O2} and also turns on the following optimization flags:
8384 @c Please keep the following list alphabetized!
8385 @gccoptlist{-fgcse-after-reload @gol
8387 -floop-interchange @gol
8388 -floop-unroll-and-jam @gol
8390 -fpredictive-commoning @gol
8392 -ftree-loop-distribute-patterns @gol
8393 -ftree-loop-distribution @gol
8394 -ftree-loop-vectorize @gol
8395 -ftree-partial-pre @gol
8396 -ftree-slp-vectorize @gol
8397 -funswitch-loops @gol
8398 -fvect-cost-model @gol
8399 -fversion-loops-for-strides}
8403 Reduce compilation time and make debugging produce the expected
8404 results. This is the default.
8408 Optimize for size. @option{-Os} enables all @option{-O2} optimizations
8409 except those that often increase code size:
8411 @gccoptlist{-falign-functions -falign-jumps @gol
8412 -falign-labels -falign-loops @gol
8413 -fprefetch-loop-arrays -freorder-blocks-algorithm=stc}
8415 It also enables @option{-finline-functions}, causes the compiler to tune for
8416 code size rather than execution speed, and performs further optimizations
8417 designed to reduce code size.
8421 Disregard strict standards compliance. @option{-Ofast} enables all
8422 @option{-O3} optimizations. It also enables optimizations that are not
8423 valid for all standard-compliant programs.
8424 It turns on @option{-ffast-math} and the Fortran-specific
8425 @option{-fstack-arrays}, unless @option{-fmax-stack-var-size} is
8426 specified, and @option{-fno-protect-parens}.
8430 Optimize debugging experience. @option{-Og} should be the optimization
8431 level of choice for the standard edit-compile-debug cycle, offering
8432 a reasonable level of optimization while maintaining fast compilation
8433 and a good debugging experience. It is a better choice than @option{-O0}
8434 for producing debuggable code because some compiler passes
8435 that collect debug information are disabled at @option{-O0}.
8437 Like @option{-O0}, @option{-Og} completely disables a number of
8438 optimization passes so that individual options controlling them have
8439 no effect. Otherwise @option{-Og} enables all @option{-O1}
8440 optimization flags except for those that may interfere with debugging:
8442 @gccoptlist{-fbranch-count-reg -fdelayed-branch @gol
8443 -fdse -fif-conversion -fif-conversion2 @gol
8444 -finline-functions-called-once @gol
8445 -fmove-loop-invariants -fssa-phiopt @gol
8446 -ftree-bit-ccp -ftree-dse -ftree-pta -ftree-sra}
8450 If you use multiple @option{-O} options, with or without level numbers,
8451 the last such option is the one that is effective.
8453 Options of the form @option{-f@var{flag}} specify machine-independent
8454 flags. Most flags have both positive and negative forms; the negative
8455 form of @option{-ffoo} is @option{-fno-foo}. In the table
8456 below, only one of the forms is listed---the one you typically
8457 use. You can figure out the other form by either removing @samp{no-}
8460 The following options control specific optimizations. They are either
8461 activated by @option{-O} options or are related to ones that are. You
8462 can use the following flags in the rare cases when ``fine-tuning'' of
8463 optimizations to be performed is desired.
8466 @item -fno-defer-pop
8467 @opindex fno-defer-pop
8469 For machines that must pop arguments after a function call, always pop
8470 the arguments as soon as each function returns.
8471 At levels @option{-O1} and higher, @option{-fdefer-pop} is the default;
8472 this allows the compiler to let arguments accumulate on the stack for several
8473 function calls and pop them all at once.
8475 @item -fforward-propagate
8476 @opindex fforward-propagate
8477 Perform a forward propagation pass on RTL@. The pass tries to combine two
8478 instructions and checks if the result can be simplified. If loop unrolling
8479 is active, two passes are performed and the second is scheduled after
8482 This option is enabled by default at optimization levels @option{-O},
8483 @option{-O2}, @option{-O3}, @option{-Os}.
8485 @item -ffp-contract=@var{style}
8486 @opindex ffp-contract
8487 @option{-ffp-contract=off} disables floating-point expression contraction.
8488 @option{-ffp-contract=fast} enables floating-point expression contraction
8489 such as forming of fused multiply-add operations if the target has
8490 native support for them.
8491 @option{-ffp-contract=on} enables floating-point expression contraction
8492 if allowed by the language standard. This is currently not implemented
8493 and treated equal to @option{-ffp-contract=off}.
8495 The default is @option{-ffp-contract=fast}.
8497 @item -fomit-frame-pointer
8498 @opindex fomit-frame-pointer
8499 Omit the frame pointer in functions that don't need one. This avoids the
8500 instructions to save, set up and restore the frame pointer; on many targets
8501 it also makes an extra register available.
8503 On some targets this flag has no effect because the standard calling sequence
8504 always uses a frame pointer, so it cannot be omitted.
8506 Note that @option{-fno-omit-frame-pointer} doesn't guarantee the frame pointer
8507 is used in all functions. Several targets always omit the frame pointer in
8510 Enabled by default at @option{-O} and higher.
8512 @item -foptimize-sibling-calls
8513 @opindex foptimize-sibling-calls
8514 Optimize sibling and tail recursive calls.
8516 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
8518 @item -foptimize-strlen
8519 @opindex foptimize-strlen
8520 Optimize various standard C string functions (e.g.@: @code{strlen},
8521 @code{strchr} or @code{strcpy}) and
8522 their @code{_FORTIFY_SOURCE} counterparts into faster alternatives.
8524 Enabled at levels @option{-O2}, @option{-O3}.
8529 Do not expand any functions inline apart from those marked with
8530 the @code{always_inline} attribute. This is the default when not
8533 Single functions can be exempted from inlining by marking them
8534 with the @code{noinline} attribute.
8536 @item -finline-small-functions
8537 @opindex finline-small-functions
8538 Integrate functions into their callers when their body is smaller than expected
8539 function call code (so overall size of program gets smaller). The compiler
8540 heuristically decides which functions are simple enough to be worth integrating
8541 in this way. This inlining applies to all functions, even those not declared
8544 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
8546 @item -findirect-inlining
8547 @opindex findirect-inlining
8548 Inline also indirect calls that are discovered to be known at compile
8549 time thanks to previous inlining. This option has any effect only
8550 when inlining itself is turned on by the @option{-finline-functions}
8551 or @option{-finline-small-functions} options.
8553 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
8555 @item -finline-functions
8556 @opindex finline-functions
8557 Consider all functions for inlining, even if they are not declared inline.
8558 The compiler heuristically decides which functions are worth integrating
8561 If all calls to a given function are integrated, and the function is
8562 declared @code{static}, then the function is normally not output as
8563 assembler code in its own right.
8565 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}. Also enabled
8566 by @option{-fprofile-use} and @option{-fauto-profile}.
8568 @item -finline-functions-called-once
8569 @opindex finline-functions-called-once
8570 Consider all @code{static} functions called once for inlining into their
8571 caller even if they are not marked @code{inline}. If a call to a given
8572 function is integrated, then the function is not output as assembler code
8575 Enabled at levels @option{-O1}, @option{-O2}, @option{-O3} and @option{-Os},
8576 but not @option{-Og}.
8578 @item -fearly-inlining
8579 @opindex fearly-inlining
8580 Inline functions marked by @code{always_inline} and functions whose body seems
8581 smaller than the function call overhead early before doing
8582 @option{-fprofile-generate} instrumentation and real inlining pass. Doing so
8583 makes profiling significantly cheaper and usually inlining faster on programs
8584 having large chains of nested wrapper functions.
8590 Perform interprocedural scalar replacement of aggregates, removal of
8591 unused parameters and replacement of parameters passed by reference
8592 by parameters passed by value.
8594 Enabled at levels @option{-O2}, @option{-O3} and @option{-Os}.
8596 @item -finline-limit=@var{n}
8597 @opindex finline-limit
8598 By default, GCC limits the size of functions that can be inlined. This flag
8599 allows coarse control of this limit. @var{n} is the size of functions that
8600 can be inlined in number of pseudo instructions.
8602 Inlining is actually controlled by a number of parameters, which may be
8603 specified individually by using @option{--param @var{name}=@var{value}}.
8604 The @option{-finline-limit=@var{n}} option sets some of these parameters
8608 @item max-inline-insns-single
8609 is set to @var{n}/2.
8610 @item max-inline-insns-auto
8611 is set to @var{n}/2.
8614 See below for a documentation of the individual
8615 parameters controlling inlining and for the defaults of these parameters.
8617 @emph{Note:} there may be no value to @option{-finline-limit} that results
8618 in default behavior.
8620 @emph{Note:} pseudo instruction represents, in this particular context, an
8621 abstract measurement of function's size. In no way does it represent a count
8622 of assembly instructions and as such its exact meaning might change from one
8623 release to an another.
8625 @item -fno-keep-inline-dllexport
8626 @opindex fno-keep-inline-dllexport
8627 @opindex fkeep-inline-dllexport
8628 This is a more fine-grained version of @option{-fkeep-inline-functions},
8629 which applies only to functions that are declared using the @code{dllexport}
8630 attribute or declspec. @xref{Function Attributes,,Declaring Attributes of
8633 @item -fkeep-inline-functions
8634 @opindex fkeep-inline-functions
8635 In C, emit @code{static} functions that are declared @code{inline}
8636 into the object file, even if the function has been inlined into all
8637 of its callers. This switch does not affect functions using the
8638 @code{extern inline} extension in GNU C90@. In C++, emit any and all
8639 inline functions into the object file.
8641 @item -fkeep-static-functions
8642 @opindex fkeep-static-functions
8643 Emit @code{static} functions into the object file, even if the function
8646 @item -fkeep-static-consts
8647 @opindex fkeep-static-consts
8648 Emit variables declared @code{static const} when optimization isn't turned
8649 on, even if the variables aren't referenced.
8651 GCC enables this option by default. If you want to force the compiler to
8652 check if a variable is referenced, regardless of whether or not
8653 optimization is turned on, use the @option{-fno-keep-static-consts} option.
8655 @item -fmerge-constants
8656 @opindex fmerge-constants
8657 Attempt to merge identical constants (string constants and floating-point
8658 constants) across compilation units.
8660 This option is the default for optimized compilation if the assembler and
8661 linker support it. Use @option{-fno-merge-constants} to inhibit this
8664 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
8666 @item -fmerge-all-constants
8667 @opindex fmerge-all-constants
8668 Attempt to merge identical constants and identical variables.
8670 This option implies @option{-fmerge-constants}. In addition to
8671 @option{-fmerge-constants} this considers e.g.@: even constant initialized
8672 arrays or initialized constant variables with integral or floating-point
8673 types. Languages like C or C++ require each variable, including multiple
8674 instances of the same variable in recursive calls, to have distinct locations,
8675 so using this option results in non-conforming
8678 @item -fmodulo-sched
8679 @opindex fmodulo-sched
8680 Perform swing modulo scheduling immediately before the first scheduling
8681 pass. This pass looks at innermost loops and reorders their
8682 instructions by overlapping different iterations.
8684 @item -fmodulo-sched-allow-regmoves
8685 @opindex fmodulo-sched-allow-regmoves
8686 Perform more aggressive SMS-based modulo scheduling with register moves
8687 allowed. By setting this flag certain anti-dependences edges are
8688 deleted, which triggers the generation of reg-moves based on the
8689 life-range analysis. This option is effective only with
8690 @option{-fmodulo-sched} enabled.
8692 @item -fno-branch-count-reg
8693 @opindex fno-branch-count-reg
8694 @opindex fbranch-count-reg
8695 Disable the optimization pass that scans for opportunities to use
8696 ``decrement and branch'' instructions on a count register instead of
8697 instruction sequences that decrement a register, compare it against zero, and
8698 then branch based upon the result. This option is only meaningful on
8699 architectures that support such instructions, which include x86, PowerPC,
8700 IA-64 and S/390. Note that the @option{-fno-branch-count-reg} option
8701 doesn't remove the decrement and branch instructions from the generated
8702 instruction stream introduced by other optimization passes.
8704 The default is @option{-fbranch-count-reg} at @option{-O1} and higher,
8705 except for @option{-Og}.
8707 @item -fno-function-cse
8708 @opindex fno-function-cse
8709 @opindex ffunction-cse
8710 Do not put function addresses in registers; make each instruction that
8711 calls a constant function contain the function's address explicitly.
8713 This option results in less efficient code, but some strange hacks
8714 that alter the assembler output may be confused by the optimizations
8715 performed when this option is not used.
8717 The default is @option{-ffunction-cse}
8719 @item -fno-zero-initialized-in-bss
8720 @opindex fno-zero-initialized-in-bss
8721 @opindex fzero-initialized-in-bss
8722 If the target supports a BSS section, GCC by default puts variables that
8723 are initialized to zero into BSS@. This can save space in the resulting
8726 This option turns off this behavior because some programs explicitly
8727 rely on variables going to the data section---e.g., so that the
8728 resulting executable can find the beginning of that section and/or make
8729 assumptions based on that.
8731 The default is @option{-fzero-initialized-in-bss}.
8733 @item -fthread-jumps
8734 @opindex fthread-jumps
8735 Perform optimizations that check to see if a jump branches to a
8736 location where another comparison subsumed by the first is found. If
8737 so, the first branch is redirected to either the destination of the
8738 second branch or a point immediately following it, depending on whether
8739 the condition is known to be true or false.
8741 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
8743 @item -fsplit-wide-types
8744 @opindex fsplit-wide-types
8745 When using a type that occupies multiple registers, such as @code{long
8746 long} on a 32-bit system, split the registers apart and allocate them
8747 independently. This normally generates better code for those types,
8748 but may make debugging more difficult.
8750 Enabled at levels @option{-O}, @option{-O2}, @option{-O3},
8753 @item -fsplit-wide-types-early
8754 @opindex fsplit-wide-types-early
8755 Fully split wide types early, instead of very late.
8756 This option has no effect unless @option{-fsplit-wide-types} is turned on.
8758 This is the default on some targets.
8760 @item -fcse-follow-jumps
8761 @opindex fcse-follow-jumps
8762 In common subexpression elimination (CSE), scan through jump instructions
8763 when the target of the jump is not reached by any other path. For
8764 example, when CSE encounters an @code{if} statement with an
8765 @code{else} clause, CSE follows the jump when the condition
8768 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
8770 @item -fcse-skip-blocks
8771 @opindex fcse-skip-blocks
8772 This is similar to @option{-fcse-follow-jumps}, but causes CSE to
8773 follow jumps that conditionally skip over blocks. When CSE
8774 encounters a simple @code{if} statement with no else clause,
8775 @option{-fcse-skip-blocks} causes CSE to follow the jump around the
8776 body of the @code{if}.
8778 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
8780 @item -frerun-cse-after-loop
8781 @opindex frerun-cse-after-loop
8782 Re-run common subexpression elimination after loop optimizations are
8785 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
8789 Perform a global common subexpression elimination pass.
8790 This pass also performs global constant and copy propagation.
8792 @emph{Note:} When compiling a program using computed gotos, a GCC
8793 extension, you may get better run-time performance if you disable
8794 the global common subexpression elimination pass by adding
8795 @option{-fno-gcse} to the command line.
8797 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
8801 When @option{-fgcse-lm} is enabled, global common subexpression elimination
8802 attempts to move loads that are only killed by stores into themselves. This
8803 allows a loop containing a load/store sequence to be changed to a load outside
8804 the loop, and a copy/store within the loop.
8806 Enabled by default when @option{-fgcse} is enabled.
8810 When @option{-fgcse-sm} is enabled, a store motion pass is run after
8811 global common subexpression elimination. This pass attempts to move
8812 stores out of loops. When used in conjunction with @option{-fgcse-lm},
8813 loops containing a load/store sequence can be changed to a load before
8814 the loop and a store after the loop.
8816 Not enabled at any optimization level.
8820 When @option{-fgcse-las} is enabled, the global common subexpression
8821 elimination pass eliminates redundant loads that come after stores to the
8822 same memory location (both partial and full redundancies).
8824 Not enabled at any optimization level.
8826 @item -fgcse-after-reload
8827 @opindex fgcse-after-reload
8828 When @option{-fgcse-after-reload} is enabled, a redundant load elimination
8829 pass is performed after reload. The purpose of this pass is to clean up
8832 Enabled by @option{-fprofile-use} and @option{-fauto-profile}.
8834 @item -faggressive-loop-optimizations
8835 @opindex faggressive-loop-optimizations
8836 This option tells the loop optimizer to use language constraints to
8837 derive bounds for the number of iterations of a loop. This assumes that
8838 loop code does not invoke undefined behavior by for example causing signed
8839 integer overflows or out-of-bound array accesses. The bounds for the
8840 number of iterations of a loop are used to guide loop unrolling and peeling
8841 and loop exit test optimizations.
8842 This option is enabled by default.
8844 @item -funconstrained-commons
8845 @opindex funconstrained-commons
8846 This option tells the compiler that variables declared in common blocks
8847 (e.g.@: Fortran) may later be overridden with longer trailing arrays. This
8848 prevents certain optimizations that depend on knowing the array bounds.
8850 @item -fcrossjumping
8851 @opindex fcrossjumping
8852 Perform cross-jumping transformation.
8853 This transformation unifies equivalent code and saves code size. The
8854 resulting code may or may not perform better than without cross-jumping.
8856 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
8858 @item -fauto-inc-dec
8859 @opindex fauto-inc-dec
8860 Combine increments or decrements of addresses with memory accesses.
8861 This pass is always skipped on architectures that do not have
8862 instructions to support this. Enabled by default at @option{-O} and
8863 higher on architectures that support this.
8867 Perform dead code elimination (DCE) on RTL@.
8868 Enabled by default at @option{-O} and higher.
8872 Perform dead store elimination (DSE) on RTL@.
8873 Enabled by default at @option{-O} and higher.
8875 @item -fif-conversion
8876 @opindex fif-conversion
8877 Attempt to transform conditional jumps into branch-less equivalents. This
8878 includes use of conditional moves, min, max, set flags and abs instructions, and
8879 some tricks doable by standard arithmetics. The use of conditional execution
8880 on chips where it is available is controlled by @option{-fif-conversion2}.
8882 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}, but
8883 not with @option{-Og}.
8885 @item -fif-conversion2
8886 @opindex fif-conversion2
8887 Use conditional execution (where available) to transform conditional jumps into
8888 branch-less equivalents.
8890 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}, but
8891 not with @option{-Og}.
8893 @item -fdeclone-ctor-dtor
8894 @opindex fdeclone-ctor-dtor
8895 The C++ ABI requires multiple entry points for constructors and
8896 destructors: one for a base subobject, one for a complete object, and
8897 one for a virtual destructor that calls operator delete afterwards.
8898 For a hierarchy with virtual bases, the base and complete variants are
8899 clones, which means two copies of the function. With this option, the
8900 base and complete variants are changed to be thunks that call a common
8903 Enabled by @option{-Os}.
8905 @item -fdelete-null-pointer-checks
8906 @opindex fdelete-null-pointer-checks
8907 Assume that programs cannot safely dereference null pointers, and that
8908 no code or data element resides at address zero.
8909 This option enables simple constant
8910 folding optimizations at all optimization levels. In addition, other
8911 optimization passes in GCC use this flag to control global dataflow
8912 analyses that eliminate useless checks for null pointers; these assume
8913 that a memory access to address zero always results in a trap, so
8914 that if a pointer is checked after it has already been dereferenced,
8917 Note however that in some environments this assumption is not true.
8918 Use @option{-fno-delete-null-pointer-checks} to disable this optimization
8919 for programs that depend on that behavior.
8921 This option is enabled by default on most targets. On Nios II ELF, it
8922 defaults to off. On AVR, CR16, and MSP430, this option is completely disabled.
8924 Passes that use the dataflow information
8925 are enabled independently at different optimization levels.
8927 @item -fdevirtualize
8928 @opindex fdevirtualize
8929 Attempt to convert calls to virtual functions to direct calls. This
8930 is done both within a procedure and interprocedurally as part of
8931 indirect inlining (@option{-findirect-inlining}) and interprocedural constant
8932 propagation (@option{-fipa-cp}).
8933 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
8935 @item -fdevirtualize-speculatively
8936 @opindex fdevirtualize-speculatively
8937 Attempt to convert calls to virtual functions to speculative direct calls.
8938 Based on the analysis of the type inheritance graph, determine for a given call
8939 the set of likely targets. If the set is small, preferably of size 1, change
8940 the call into a conditional deciding between direct and indirect calls. The
8941 speculative calls enable more optimizations, such as inlining. When they seem
8942 useless after further optimization, they are converted back into original form.
8944 @item -fdevirtualize-at-ltrans
8945 @opindex fdevirtualize-at-ltrans
8946 Stream extra information needed for aggressive devirtualization when running
8947 the link-time optimizer in local transformation mode.
8948 This option enables more devirtualization but
8949 significantly increases the size of streamed data. For this reason it is
8950 disabled by default.
8952 @item -fexpensive-optimizations
8953 @opindex fexpensive-optimizations
8954 Perform a number of minor optimizations that are relatively expensive.
8956 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
8960 Attempt to remove redundant extension instructions. This is especially
8961 helpful for the x86-64 architecture, which implicitly zero-extends in 64-bit
8962 registers after writing to their lower 32-bit half.
8964 Enabled for Alpha, AArch64 and x86 at levels @option{-O2},
8965 @option{-O3}, @option{-Os}.
8967 @item -fno-lifetime-dse
8968 @opindex fno-lifetime-dse
8969 @opindex flifetime-dse
8970 In C++ the value of an object is only affected by changes within its
8971 lifetime: when the constructor begins, the object has an indeterminate
8972 value, and any changes during the lifetime of the object are dead when
8973 the object is destroyed. Normally dead store elimination will take
8974 advantage of this; if your code relies on the value of the object
8975 storage persisting beyond the lifetime of the object, you can use this
8976 flag to disable this optimization. To preserve stores before the
8977 constructor starts (e.g.@: because your operator new clears the object
8978 storage) but still treat the object as dead after the destructor you,
8979 can use @option{-flifetime-dse=1}. The default behavior can be
8980 explicitly selected with @option{-flifetime-dse=2}.
8981 @option{-flifetime-dse=0} is equivalent to @option{-fno-lifetime-dse}.
8983 @item -flive-range-shrinkage
8984 @opindex flive-range-shrinkage
8985 Attempt to decrease register pressure through register live range
8986 shrinkage. This is helpful for fast processors with small or moderate
8989 @item -fira-algorithm=@var{algorithm}
8990 @opindex fira-algorithm
8991 Use the specified coloring algorithm for the integrated register
8992 allocator. The @var{algorithm} argument can be @samp{priority}, which
8993 specifies Chow's priority coloring, or @samp{CB}, which specifies
8994 Chaitin-Briggs coloring. Chaitin-Briggs coloring is not implemented
8995 for all architectures, but for those targets that do support it, it is
8996 the default because it generates better code.
8998 @item -fira-region=@var{region}
8999 @opindex fira-region
9000 Use specified regions for the integrated register allocator. The
9001 @var{region} argument should be one of the following:
9006 Use all loops as register allocation regions.
9007 This can give the best results for machines with a small and/or
9008 irregular register set.
9011 Use all loops except for loops with small register pressure
9012 as the regions. This value usually gives
9013 the best results in most cases and for most architectures,
9014 and is enabled by default when compiling with optimization for speed
9015 (@option{-O}, @option{-O2}, @dots{}).
9018 Use all functions as a single region.
9019 This typically results in the smallest code size, and is enabled by default for
9020 @option{-Os} or @option{-O0}.
9024 @item -fira-hoist-pressure
9025 @opindex fira-hoist-pressure
9026 Use IRA to evaluate register pressure in the code hoisting pass for
9027 decisions to hoist expressions. This option usually results in smaller
9028 code, but it can slow the compiler down.
9030 This option is enabled at level @option{-Os} for all targets.
9032 @item -fira-loop-pressure
9033 @opindex fira-loop-pressure
9034 Use IRA to evaluate register pressure in loops for decisions to move
9035 loop invariants. This option usually results in generation
9036 of faster and smaller code on machines with large register files (>= 32
9037 registers), but it can slow the compiler down.
9039 This option is enabled at level @option{-O3} for some targets.
9041 @item -fno-ira-share-save-slots
9042 @opindex fno-ira-share-save-slots
9043 @opindex fira-share-save-slots
9044 Disable sharing of stack slots used for saving call-used hard
9045 registers living through a call. Each hard register gets a
9046 separate stack slot, and as a result function stack frames are
9049 @item -fno-ira-share-spill-slots
9050 @opindex fno-ira-share-spill-slots
9051 @opindex fira-share-spill-slots
9052 Disable sharing of stack slots allocated for pseudo-registers. Each
9053 pseudo-register that does not get a hard register gets a separate
9054 stack slot, and as a result function stack frames are larger.
9058 Enable CFG-sensitive rematerialization in LRA. Instead of loading
9059 values of spilled pseudos, LRA tries to rematerialize (recalculate)
9060 values if it is profitable.
9062 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
9064 @item -fdelayed-branch
9065 @opindex fdelayed-branch
9066 If supported for the target machine, attempt to reorder instructions
9067 to exploit instruction slots available after delayed branch
9070 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os},
9071 but not at @option{-Og}.
9073 @item -fschedule-insns
9074 @opindex fschedule-insns
9075 If supported for the target machine, attempt to reorder instructions to
9076 eliminate execution stalls due to required data being unavailable. This
9077 helps machines that have slow floating point or memory load instructions
9078 by allowing other instructions to be issued until the result of the load
9079 or floating-point instruction is required.
9081 Enabled at levels @option{-O2}, @option{-O3}.
9083 @item -fschedule-insns2
9084 @opindex fschedule-insns2
9085 Similar to @option{-fschedule-insns}, but requests an additional pass of
9086 instruction scheduling after register allocation has been done. This is
9087 especially useful on machines with a relatively small number of
9088 registers and where memory load instructions take more than one cycle.
9090 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
9092 @item -fno-sched-interblock
9093 @opindex fno-sched-interblock
9094 @opindex fsched-interblock
9095 Disable instruction scheduling across basic blocks, which
9096 is normally enabled when scheduling before register allocation, i.e.@:
9097 with @option{-fschedule-insns} or at @option{-O2} or higher.
9099 @item -fno-sched-spec
9100 @opindex fno-sched-spec
9101 @opindex fsched-spec
9102 Disable speculative motion of non-load instructions, which
9103 is normally enabled when scheduling before register allocation, i.e.@:
9104 with @option{-fschedule-insns} or at @option{-O2} or higher.
9106 @item -fsched-pressure
9107 @opindex fsched-pressure
9108 Enable register pressure sensitive insn scheduling before register
9109 allocation. This only makes sense when scheduling before register
9110 allocation is enabled, i.e.@: with @option{-fschedule-insns} or at
9111 @option{-O2} or higher. Usage of this option can improve the
9112 generated code and decrease its size by preventing register pressure
9113 increase above the number of available hard registers and subsequent
9114 spills in register allocation.
9116 @item -fsched-spec-load
9117 @opindex fsched-spec-load
9118 Allow speculative motion of some load instructions. This only makes
9119 sense when scheduling before register allocation, i.e.@: with
9120 @option{-fschedule-insns} or at @option{-O2} or higher.
9122 @item -fsched-spec-load-dangerous
9123 @opindex fsched-spec-load-dangerous
9124 Allow speculative motion of more load instructions. This only makes
9125 sense when scheduling before register allocation, i.e.@: with
9126 @option{-fschedule-insns} or at @option{-O2} or higher.
9128 @item -fsched-stalled-insns
9129 @itemx -fsched-stalled-insns=@var{n}
9130 @opindex fsched-stalled-insns
9131 Define how many insns (if any) can be moved prematurely from the queue
9132 of stalled insns into the ready list during the second scheduling pass.
9133 @option{-fno-sched-stalled-insns} means that no insns are moved
9134 prematurely, @option{-fsched-stalled-insns=0} means there is no limit
9135 on how many queued insns can be moved prematurely.
9136 @option{-fsched-stalled-insns} without a value is equivalent to
9137 @option{-fsched-stalled-insns=1}.
9139 @item -fsched-stalled-insns-dep
9140 @itemx -fsched-stalled-insns-dep=@var{n}
9141 @opindex fsched-stalled-insns-dep
9142 Define how many insn groups (cycles) are examined for a dependency
9143 on a stalled insn that is a candidate for premature removal from the queue
9144 of stalled insns. This has an effect only during the second scheduling pass,
9145 and only if @option{-fsched-stalled-insns} is used.
9146 @option{-fno-sched-stalled-insns-dep} is equivalent to
9147 @option{-fsched-stalled-insns-dep=0}.
9148 @option{-fsched-stalled-insns-dep} without a value is equivalent to
9149 @option{-fsched-stalled-insns-dep=1}.
9151 @item -fsched2-use-superblocks
9152 @opindex fsched2-use-superblocks
9153 When scheduling after register allocation, use superblock scheduling.
9154 This allows motion across basic block boundaries,
9155 resulting in faster schedules. This option is experimental, as not all machine
9156 descriptions used by GCC model the CPU closely enough to avoid unreliable
9157 results from the algorithm.
9159 This only makes sense when scheduling after register allocation, i.e.@: with
9160 @option{-fschedule-insns2} or at @option{-O2} or higher.
9162 @item -fsched-group-heuristic
9163 @opindex fsched-group-heuristic
9164 Enable the group heuristic in the scheduler. This heuristic favors
9165 the instruction that belongs to a schedule group. This is enabled
9166 by default when scheduling is enabled, i.e.@: with @option{-fschedule-insns}
9167 or @option{-fschedule-insns2} or at @option{-O2} or higher.
9169 @item -fsched-critical-path-heuristic
9170 @opindex fsched-critical-path-heuristic
9171 Enable the critical-path heuristic in the scheduler. This heuristic favors
9172 instructions on the critical path. This is enabled by default when
9173 scheduling is enabled, i.e.@: with @option{-fschedule-insns}
9174 or @option{-fschedule-insns2} or at @option{-O2} or higher.
9176 @item -fsched-spec-insn-heuristic
9177 @opindex fsched-spec-insn-heuristic
9178 Enable the speculative instruction heuristic in the scheduler. This
9179 heuristic favors speculative instructions with greater dependency weakness.
9180 This is enabled by default when scheduling is enabled, i.e.@:
9181 with @option{-fschedule-insns} or @option{-fschedule-insns2}
9182 or at @option{-O2} or higher.
9184 @item -fsched-rank-heuristic
9185 @opindex fsched-rank-heuristic
9186 Enable the rank heuristic in the scheduler. This heuristic favors
9187 the instruction belonging to a basic block with greater size or frequency.
9188 This is enabled by default when scheduling is enabled, i.e.@:
9189 with @option{-fschedule-insns} or @option{-fschedule-insns2} or
9190 at @option{-O2} or higher.
9192 @item -fsched-last-insn-heuristic
9193 @opindex fsched-last-insn-heuristic
9194 Enable the last-instruction heuristic in the scheduler. This heuristic
9195 favors the instruction that is less dependent on the last instruction
9196 scheduled. This is enabled by default when scheduling is enabled,
9197 i.e.@: with @option{-fschedule-insns} or @option{-fschedule-insns2} or
9198 at @option{-O2} or higher.
9200 @item -fsched-dep-count-heuristic
9201 @opindex fsched-dep-count-heuristic
9202 Enable the dependent-count heuristic in the scheduler. This heuristic
9203 favors the instruction that has more instructions depending on it.
9204 This is enabled by default when scheduling is enabled, i.e.@:
9205 with @option{-fschedule-insns} or @option{-fschedule-insns2} or
9206 at @option{-O2} or higher.
9208 @item -freschedule-modulo-scheduled-loops
9209 @opindex freschedule-modulo-scheduled-loops
9210 Modulo scheduling is performed before traditional scheduling. If a loop
9211 is modulo scheduled, later scheduling passes may change its schedule.
9212 Use this option to control that behavior.
9214 @item -fselective-scheduling
9215 @opindex fselective-scheduling
9216 Schedule instructions using selective scheduling algorithm. Selective
9217 scheduling runs instead of the first scheduler pass.
9219 @item -fselective-scheduling2
9220 @opindex fselective-scheduling2
9221 Schedule instructions using selective scheduling algorithm. Selective
9222 scheduling runs instead of the second scheduler pass.
9224 @item -fsel-sched-pipelining
9225 @opindex fsel-sched-pipelining
9226 Enable software pipelining of innermost loops during selective scheduling.
9227 This option has no effect unless one of @option{-fselective-scheduling} or
9228 @option{-fselective-scheduling2} is turned on.
9230 @item -fsel-sched-pipelining-outer-loops
9231 @opindex fsel-sched-pipelining-outer-loops
9232 When pipelining loops during selective scheduling, also pipeline outer loops.
9233 This option has no effect unless @option{-fsel-sched-pipelining} is turned on.
9235 @item -fsemantic-interposition
9236 @opindex fsemantic-interposition
9237 Some object formats, like ELF, allow interposing of symbols by the
9239 This means that for symbols exported from the DSO, the compiler cannot perform
9240 interprocedural propagation, inlining and other optimizations in anticipation
9241 that the function or variable in question may change. While this feature is
9242 useful, for example, to rewrite memory allocation functions by a debugging
9243 implementation, it is expensive in the terms of code quality.
9244 With @option{-fno-semantic-interposition} the compiler assumes that
9245 if interposition happens for functions the overwriting function will have
9246 precisely the same semantics (and side effects).
9247 Similarly if interposition happens
9248 for variables, the constructor of the variable will be the same. The flag
9249 has no effect for functions explicitly declared inline
9250 (where it is never allowed for interposition to change semantics)
9251 and for symbols explicitly declared weak.
9254 @opindex fshrink-wrap
9255 Emit function prologues only before parts of the function that need it,
9256 rather than at the top of the function. This flag is enabled by default at
9257 @option{-O} and higher.
9259 @item -fshrink-wrap-separate
9260 @opindex fshrink-wrap-separate
9261 Shrink-wrap separate parts of the prologue and epilogue separately, so that
9262 those parts are only executed when needed.
9263 This option is on by default, but has no effect unless @option{-fshrink-wrap}
9264 is also turned on and the target supports this.
9266 @item -fcaller-saves
9267 @opindex fcaller-saves
9268 Enable allocation of values to registers that are clobbered by
9269 function calls, by emitting extra instructions to save and restore the
9270 registers around such calls. Such allocation is done only when it
9271 seems to result in better code.
9273 This option is always enabled by default on certain machines, usually
9274 those which have no call-preserved registers to use instead.
9276 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
9278 @item -fcombine-stack-adjustments
9279 @opindex fcombine-stack-adjustments
9280 Tracks stack adjustments (pushes and pops) and stack memory references
9281 and then tries to find ways to combine them.
9283 Enabled by default at @option{-O1} and higher.
9287 Use caller save registers for allocation if those registers are not used by
9288 any called function. In that case it is not necessary to save and restore
9289 them around calls. This is only possible if called functions are part of
9290 same compilation unit as current function and they are compiled before it.
9292 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}, however the option
9293 is disabled if generated code will be instrumented for profiling
9294 (@option{-p}, or @option{-pg}) or if callee's register usage cannot be known
9295 exactly (this happens on targets that do not expose prologues
9296 and epilogues in RTL).
9298 @item -fconserve-stack
9299 @opindex fconserve-stack
9300 Attempt to minimize stack usage. The compiler attempts to use less
9301 stack space, even if that makes the program slower. This option
9302 implies setting the @option{large-stack-frame} parameter to 100
9303 and the @option{large-stack-frame-growth} parameter to 400.
9305 @item -ftree-reassoc
9306 @opindex ftree-reassoc
9307 Perform reassociation on trees. This flag is enabled by default
9308 at @option{-O} and higher.
9310 @item -fcode-hoisting
9311 @opindex fcode-hoisting
9312 Perform code hoisting. Code hoisting tries to move the
9313 evaluation of expressions executed on all paths to the function exit
9314 as early as possible. This is especially useful as a code size
9315 optimization, but it often helps for code speed as well.
9316 This flag is enabled by default at @option{-O2} and higher.
9320 Perform partial redundancy elimination (PRE) on trees. This flag is
9321 enabled by default at @option{-O2} and @option{-O3}.
9323 @item -ftree-partial-pre
9324 @opindex ftree-partial-pre
9325 Make partial redundancy elimination (PRE) more aggressive. This flag is
9326 enabled by default at @option{-O3}.
9328 @item -ftree-forwprop
9329 @opindex ftree-forwprop
9330 Perform forward propagation on trees. This flag is enabled by default
9331 at @option{-O} and higher.
9335 Perform full redundancy elimination (FRE) on trees. The difference
9336 between FRE and PRE is that FRE only considers expressions
9337 that are computed on all paths leading to the redundant computation.
9338 This analysis is faster than PRE, though it exposes fewer redundancies.
9339 This flag is enabled by default at @option{-O} and higher.
9341 @item -ftree-phiprop
9342 @opindex ftree-phiprop
9343 Perform hoisting of loads from conditional pointers on trees. This
9344 pass is enabled by default at @option{-O} and higher.
9346 @item -fhoist-adjacent-loads
9347 @opindex fhoist-adjacent-loads
9348 Speculatively hoist loads from both branches of an if-then-else if the
9349 loads are from adjacent locations in the same structure and the target
9350 architecture has a conditional move instruction. This flag is enabled
9351 by default at @option{-O2} and higher.
9353 @item -ftree-copy-prop
9354 @opindex ftree-copy-prop
9355 Perform copy propagation on trees. This pass eliminates unnecessary
9356 copy operations. This flag is enabled by default at @option{-O} and
9359 @item -fipa-pure-const
9360 @opindex fipa-pure-const
9361 Discover which functions are pure or constant.
9362 Enabled by default at @option{-O} and higher.
9364 @item -fipa-reference
9365 @opindex fipa-reference
9366 Discover which static variables do not escape the
9368 Enabled by default at @option{-O} and higher.
9370 @item -fipa-reference-addressable
9371 @opindex fipa-reference-addressable
9372 Discover read-only, write-only and non-addressable static variables.
9373 Enabled by default at @option{-O} and higher.
9375 @item -fipa-stack-alignment
9376 @opindex fipa-stack-alignment
9377 Reduce stack alignment on call sites if possible.
9382 Perform interprocedural pointer analysis and interprocedural modification
9383 and reference analysis. This option can cause excessive memory and
9384 compile-time usage on large compilation units. It is not enabled by
9385 default at any optimization level.
9388 @opindex fipa-profile
9389 Perform interprocedural profile propagation. The functions called only from
9390 cold functions are marked as cold. Also functions executed once (such as
9391 @code{cold}, @code{noreturn}, static constructors or destructors) are identified. Cold
9392 functions and loop less parts of functions executed once are then optimized for
9394 Enabled by default at @option{-O} and higher.
9398 Perform interprocedural constant propagation.
9399 This optimization analyzes the program to determine when values passed
9400 to functions are constants and then optimizes accordingly.
9401 This optimization can substantially increase performance
9402 if the application has constants passed to functions.
9403 This flag is enabled by default at @option{-O2}, @option{-Os} and @option{-O3}.
9404 It is also enabled by @option{-fprofile-use} and @option{-fauto-profile}.
9406 @item -fipa-cp-clone
9407 @opindex fipa-cp-clone
9408 Perform function cloning to make interprocedural constant propagation stronger.
9409 When enabled, interprocedural constant propagation performs function cloning
9410 when externally visible function can be called with constant arguments.
9411 Because this optimization can create multiple copies of functions,
9412 it may significantly increase code size
9413 (see @option{--param ipcp-unit-growth=@var{value}}).
9414 This flag is enabled by default at @option{-O3}.
9415 It is also enabled by @option{-fprofile-use} and @option{-fauto-profile}.
9418 @opindex fipa-bit-cp
9419 When enabled, perform interprocedural bitwise constant
9420 propagation. This flag is enabled by default at @option{-O2} and
9421 by @option{-fprofile-use} and @option{-fauto-profile}.
9422 It requires that @option{-fipa-cp} is enabled.
9426 When enabled, perform interprocedural propagation of value
9427 ranges. This flag is enabled by default at @option{-O2}. It requires
9428 that @option{-fipa-cp} is enabled.
9432 Perform Identical Code Folding for functions and read-only variables.
9433 The optimization reduces code size and may disturb unwind stacks by replacing
9434 a function by equivalent one with a different name. The optimization works
9435 more effectively with link-time optimization enabled.
9437 Although the behavior is similar to the Gold Linker's ICF optimization, GCC ICF
9438 works on different levels and thus the optimizations are not same - there are
9439 equivalences that are found only by GCC and equivalences found only by Gold.
9441 This flag is enabled by default at @option{-O2} and @option{-Os}.
9443 @item -flive-patching=@var{level}
9444 @opindex flive-patching
9445 Control GCC's optimizations to produce output suitable for live-patching.
9447 If the compiler's optimization uses a function's body or information extracted
9448 from its body to optimize/change another function, the latter is called an
9449 impacted function of the former. If a function is patched, its impacted
9450 functions should be patched too.
9452 The impacted functions are determined by the compiler's interprocedural
9453 optimizations. For example, a caller is impacted when inlining a function
9455 cloning a function and changing its caller to call this new clone,
9456 or extracting a function's pureness/constness information to optimize
9457 its direct or indirect callers, etc.
9459 Usually, the more IPA optimizations enabled, the larger the number of
9460 impacted functions for each function. In order to control the number of
9461 impacted functions and more easily compute the list of impacted function,
9462 IPA optimizations can be partially enabled at two different levels.
9464 The @var{level} argument should be one of the following:
9470 Only enable inlining and cloning optimizations, which includes inlining,
9471 cloning, interprocedural scalar replacement of aggregates and partial inlining.
9472 As a result, when patching a function, all its callers and its clones'
9473 callers are impacted, therefore need to be patched as well.
9475 @option{-flive-patching=inline-clone} disables the following optimization flags:
9476 @gccoptlist{-fwhole-program -fipa-pta -fipa-reference -fipa-ra @gol
9477 -fipa-icf -fipa-icf-functions -fipa-icf-variables @gol
9478 -fipa-bit-cp -fipa-vrp -fipa-pure-const -fipa-reference-addressable @gol
9479 -fipa-stack-alignment}
9481 @item inline-only-static
9483 Only enable inlining of static functions.
9484 As a result, when patching a static function, all its callers are impacted
9485 and so need to be patched as well.
9487 In addition to all the flags that @option{-flive-patching=inline-clone}
9489 @option{-flive-patching=inline-only-static} disables the following additional
9491 @gccoptlist{-fipa-cp-clone -fipa-sra -fpartial-inlining -fipa-cp}
9495 When @option{-flive-patching} is specified without any value, the default value
9496 is @var{inline-clone}.
9498 This flag is disabled by default.
9500 Note that @option{-flive-patching} is not supported with link-time optimization
9503 @item -fisolate-erroneous-paths-dereference
9504 @opindex fisolate-erroneous-paths-dereference
9505 Detect paths that trigger erroneous or undefined behavior due to
9506 dereferencing a null pointer. Isolate those paths from the main control
9507 flow and turn the statement with erroneous or undefined behavior into a trap.
9508 This flag is enabled by default at @option{-O2} and higher and depends on
9509 @option{-fdelete-null-pointer-checks} also being enabled.
9511 @item -fisolate-erroneous-paths-attribute
9512 @opindex fisolate-erroneous-paths-attribute
9513 Detect paths that trigger erroneous or undefined behavior due to a null value
9514 being used in a way forbidden by a @code{returns_nonnull} or @code{nonnull}
9515 attribute. Isolate those paths from the main control flow and turn the
9516 statement with erroneous or undefined behavior into a trap. This is not
9517 currently enabled, but may be enabled by @option{-O2} in the future.
9521 Perform forward store motion on trees. This flag is
9522 enabled by default at @option{-O} and higher.
9524 @item -ftree-bit-ccp
9525 @opindex ftree-bit-ccp
9526 Perform sparse conditional bit constant propagation on trees and propagate
9527 pointer alignment information.
9528 This pass only operates on local scalar variables and is enabled by default
9529 at @option{-O1} and higher, except for @option{-Og}.
9530 It requires that @option{-ftree-ccp} is enabled.
9534 Perform sparse conditional constant propagation (CCP) on trees. This
9535 pass only operates on local scalar variables and is enabled by default
9536 at @option{-O} and higher.
9538 @item -fssa-backprop
9539 @opindex fssa-backprop
9540 Propagate information about uses of a value up the definition chain
9541 in order to simplify the definitions. For example, this pass strips
9542 sign operations if the sign of a value never matters. The flag is
9543 enabled by default at @option{-O} and higher.
9546 @opindex fssa-phiopt
9547 Perform pattern matching on SSA PHI nodes to optimize conditional
9548 code. This pass is enabled by default at @option{-O1} and higher,
9549 except for @option{-Og}.
9551 @item -ftree-switch-conversion
9552 @opindex ftree-switch-conversion
9553 Perform conversion of simple initializations in a switch to
9554 initializations from a scalar array. This flag is enabled by default
9555 at @option{-O2} and higher.
9557 @item -ftree-tail-merge
9558 @opindex ftree-tail-merge
9559 Look for identical code sequences. When found, replace one with a jump to the
9560 other. This optimization is known as tail merging or cross jumping. This flag
9561 is enabled by default at @option{-O2} and higher. The compilation time
9563 be limited using @option{max-tail-merge-comparisons} parameter and
9564 @option{max-tail-merge-iterations} parameter.
9568 Perform dead code elimination (DCE) on trees. This flag is enabled by
9569 default at @option{-O} and higher.
9571 @item -ftree-builtin-call-dce
9572 @opindex ftree-builtin-call-dce
9573 Perform conditional dead code elimination (DCE) for calls to built-in functions
9574 that may set @code{errno} but are otherwise free of side effects. This flag is
9575 enabled by default at @option{-O2} and higher if @option{-Os} is not also
9578 @item -ffinite-loops
9579 @opindex ffinite-loops
9580 @opindex fno-finite-loops
9581 Assume that a loop with an exit will eventually take the exit and not loop
9582 indefinitely. This allows the compiler to remove loops that otherwise have
9583 no side-effects, not considering eventual endless looping as such.
9585 This option is enabled by default at @option{-O2}.
9587 @item -ftree-dominator-opts
9588 @opindex ftree-dominator-opts
9589 Perform a variety of simple scalar cleanups (constant/copy
9590 propagation, redundancy elimination, range propagation and expression
9591 simplification) based on a dominator tree traversal. This also
9592 performs jump threading (to reduce jumps to jumps). This flag is
9593 enabled by default at @option{-O} and higher.
9597 Perform dead store elimination (DSE) on trees. A dead store is a store into
9598 a memory location that is later overwritten by another store without
9599 any intervening loads. In this case the earlier store can be deleted. This
9600 flag is enabled by default at @option{-O} and higher.
9604 Perform loop header copying on trees. This is beneficial since it increases
9605 effectiveness of code motion optimizations. It also saves one jump. This flag
9606 is enabled by default at @option{-O} and higher. It is not enabled
9607 for @option{-Os}, since it usually increases code size.
9609 @item -ftree-loop-optimize
9610 @opindex ftree-loop-optimize
9611 Perform loop optimizations on trees. This flag is enabled by default
9612 at @option{-O} and higher.
9614 @item -ftree-loop-linear
9615 @itemx -floop-strip-mine
9617 @opindex ftree-loop-linear
9618 @opindex floop-strip-mine
9619 @opindex floop-block
9620 Perform loop nest optimizations. Same as
9621 @option{-floop-nest-optimize}. To use this code transformation, GCC has
9622 to be configured with @option{--with-isl} to enable the Graphite loop
9623 transformation infrastructure.
9625 @item -fgraphite-identity
9626 @opindex fgraphite-identity
9627 Enable the identity transformation for graphite. For every SCoP we generate
9628 the polyhedral representation and transform it back to gimple. Using
9629 @option{-fgraphite-identity} we can check the costs or benefits of the
9630 GIMPLE -> GRAPHITE -> GIMPLE transformation. Some minimal optimizations
9631 are also performed by the code generator isl, like index splitting and
9632 dead code elimination in loops.
9634 @item -floop-nest-optimize
9635 @opindex floop-nest-optimize
9636 Enable the isl based loop nest optimizer. This is a generic loop nest
9637 optimizer based on the Pluto optimization algorithms. It calculates a loop
9638 structure optimized for data-locality and parallelism. This option
9641 @item -floop-parallelize-all
9642 @opindex floop-parallelize-all
9643 Use the Graphite data dependence analysis to identify loops that can
9644 be parallelized. Parallelize all the loops that can be analyzed to
9645 not contain loop carried dependences without checking that it is
9646 profitable to parallelize the loops.
9648 @item -ftree-coalesce-vars
9649 @opindex ftree-coalesce-vars
9650 While transforming the program out of the SSA representation, attempt to
9651 reduce copying by coalescing versions of different user-defined
9652 variables, instead of just compiler temporaries. This may severely
9653 limit the ability to debug an optimized program compiled with
9654 @option{-fno-var-tracking-assignments}. In the negated form, this flag
9655 prevents SSA coalescing of user variables. This option is enabled by
9656 default if optimization is enabled, and it does very little otherwise.
9658 @item -ftree-loop-if-convert
9659 @opindex ftree-loop-if-convert
9660 Attempt to transform conditional jumps in the innermost loops to
9661 branch-less equivalents. The intent is to remove control-flow from
9662 the innermost loops in order to improve the ability of the
9663 vectorization pass to handle these loops. This is enabled by default
9664 if vectorization is enabled.
9666 @item -ftree-loop-distribution
9667 @opindex ftree-loop-distribution
9668 Perform loop distribution. This flag can improve cache performance on
9669 big loop bodies and allow further loop optimizations, like
9670 parallelization or vectorization, to take place. For example, the loop
9686 This flag is enabled by default at @option{-O3}.
9687 It is also enabled by @option{-fprofile-use} and @option{-fauto-profile}.
9689 @item -ftree-loop-distribute-patterns
9690 @opindex ftree-loop-distribute-patterns
9691 Perform loop distribution of patterns that can be code generated with
9692 calls to a library. This flag is enabled by default at @option{-O3}, and
9693 by @option{-fprofile-use} and @option{-fauto-profile}.
9695 This pass distributes the initialization loops and generates a call to
9696 memset zero. For example, the loop
9712 and the initialization loop is transformed into a call to memset zero.
9713 This flag is enabled by default at @option{-O3}.
9714 It is also enabled by @option{-fprofile-use} and @option{-fauto-profile}.
9716 @item -floop-interchange
9717 @opindex floop-interchange
9718 Perform loop interchange outside of graphite. This flag can improve cache
9719 performance on loop nest and allow further loop optimizations, like
9720 vectorization, to take place. For example, the loop
9722 for (int i = 0; i < N; i++)
9723 for (int j = 0; j < N; j++)
9724 for (int k = 0; k < N; k++)
9725 c[i][j] = c[i][j] + a[i][k]*b[k][j];
9729 for (int i = 0; i < N; i++)
9730 for (int k = 0; k < N; k++)
9731 for (int j = 0; j < N; j++)
9732 c[i][j] = c[i][j] + a[i][k]*b[k][j];
9734 This flag is enabled by default at @option{-O3}.
9735 It is also enabled by @option{-fprofile-use} and @option{-fauto-profile}.
9737 @item -floop-unroll-and-jam
9738 @opindex floop-unroll-and-jam
9739 Apply unroll and jam transformations on feasible loops. In a loop
9740 nest this unrolls the outer loop by some factor and fuses the resulting
9741 multiple inner loops. This flag is enabled by default at @option{-O3}.
9742 It is also enabled by @option{-fprofile-use} and @option{-fauto-profile}.
9744 @item -ftree-loop-im
9745 @opindex ftree-loop-im
9746 Perform loop invariant motion on trees. This pass moves only invariants that
9747 are hard to handle at RTL level (function calls, operations that expand to
9748 nontrivial sequences of insns). With @option{-funswitch-loops} it also moves
9749 operands of conditions that are invariant out of the loop, so that we can use
9750 just trivial invariantness analysis in loop unswitching. The pass also includes
9753 @item -ftree-loop-ivcanon
9754 @opindex ftree-loop-ivcanon
9755 Create a canonical counter for number of iterations in loops for which
9756 determining number of iterations requires complicated analysis. Later
9757 optimizations then may determine the number easily. Useful especially
9758 in connection with unrolling.
9760 @item -ftree-scev-cprop
9761 @opindex ftree-scev-cprop
9762 Perform final value replacement. If a variable is modified in a loop
9763 in such a way that its value when exiting the loop can be determined using
9764 only its initial value and the number of loop iterations, replace uses of
9765 the final value by such a computation, provided it is sufficiently cheap.
9766 This reduces data dependencies and may allow further simplifications.
9767 Enabled by default at @option{-O} and higher.
9771 Perform induction variable optimizations (strength reduction, induction
9772 variable merging and induction variable elimination) on trees.
9774 @item -ftree-parallelize-loops=n
9775 @opindex ftree-parallelize-loops
9776 Parallelize loops, i.e., split their iteration space to run in n threads.
9777 This is only possible for loops whose iterations are independent
9778 and can be arbitrarily reordered. The optimization is only
9779 profitable on multiprocessor machines, for loops that are CPU-intensive,
9780 rather than constrained e.g.@: by memory bandwidth. This option
9781 implies @option{-pthread}, and thus is only supported on targets
9782 that have support for @option{-pthread}.
9786 Perform function-local points-to analysis on trees. This flag is
9787 enabled by default at @option{-O1} and higher, except for @option{-Og}.
9791 Perform scalar replacement of aggregates. This pass replaces structure
9792 references with scalars to prevent committing structures to memory too
9793 early. This flag is enabled by default at @option{-O1} and higher,
9794 except for @option{-Og}.
9796 @item -fstore-merging
9797 @opindex fstore-merging
9798 Perform merging of narrow stores to consecutive memory addresses. This pass
9799 merges contiguous stores of immediate values narrower than a word into fewer
9800 wider stores to reduce the number of instructions. This is enabled by default
9801 at @option{-O2} and higher as well as @option{-Os}.
9805 Perform temporary expression replacement during the SSA->normal phase. Single
9806 use/single def temporaries are replaced at their use location with their
9807 defining expression. This results in non-GIMPLE code, but gives the expanders
9808 much more complex trees to work on resulting in better RTL generation. This is
9809 enabled by default at @option{-O} and higher.
9813 Perform straight-line strength reduction on trees. This recognizes related
9814 expressions involving multiplications and replaces them by less expensive
9815 calculations when possible. This is enabled by default at @option{-O} and
9818 @item -ftree-vectorize
9819 @opindex ftree-vectorize
9820 Perform vectorization on trees. This flag enables @option{-ftree-loop-vectorize}
9821 and @option{-ftree-slp-vectorize} if not explicitly specified.
9823 @item -ftree-loop-vectorize
9824 @opindex ftree-loop-vectorize
9825 Perform loop vectorization on trees. This flag is enabled by default at
9826 @option{-O3} and by @option{-ftree-vectorize}, @option{-fprofile-use},
9827 and @option{-fauto-profile}.
9829 @item -ftree-slp-vectorize
9830 @opindex ftree-slp-vectorize
9831 Perform basic block vectorization on trees. This flag is enabled by default at
9832 @option{-O3} and by @option{-ftree-vectorize}, @option{-fprofile-use},
9833 and @option{-fauto-profile}.
9835 @item -fvect-cost-model=@var{model}
9836 @opindex fvect-cost-model
9837 Alter the cost model used for vectorization. The @var{model} argument
9838 should be one of @samp{unlimited}, @samp{dynamic} or @samp{cheap}.
9839 With the @samp{unlimited} model the vectorized code-path is assumed
9840 to be profitable while with the @samp{dynamic} model a runtime check
9841 guards the vectorized code-path to enable it only for iteration
9842 counts that will likely execute faster than when executing the original
9843 scalar loop. The @samp{cheap} model disables vectorization of
9844 loops where doing so would be cost prohibitive for example due to
9845 required runtime checks for data dependence or alignment but otherwise
9846 is equal to the @samp{dynamic} model.
9847 The default cost model depends on other optimization flags and is
9848 either @samp{dynamic} or @samp{cheap}.
9850 @item -fsimd-cost-model=@var{model}
9851 @opindex fsimd-cost-model
9852 Alter the cost model used for vectorization of loops marked with the OpenMP
9853 simd directive. The @var{model} argument should be one of
9854 @samp{unlimited}, @samp{dynamic}, @samp{cheap}. All values of @var{model}
9855 have the same meaning as described in @option{-fvect-cost-model} and by
9856 default a cost model defined with @option{-fvect-cost-model} is used.
9860 Perform Value Range Propagation on trees. This is similar to the
9861 constant propagation pass, but instead of values, ranges of values are
9862 propagated. This allows the optimizers to remove unnecessary range
9863 checks like array bound checks and null pointer checks. This is
9864 enabled by default at @option{-O2} and higher. Null pointer check
9865 elimination is only done if @option{-fdelete-null-pointer-checks} is
9869 @opindex fsplit-paths
9870 Split paths leading to loop backedges. This can improve dead code
9871 elimination and common subexpression elimination. This is enabled by
9872 default at @option{-O3} and above.
9874 @item -fsplit-ivs-in-unroller
9875 @opindex fsplit-ivs-in-unroller
9876 Enables expression of values of induction variables in later iterations
9877 of the unrolled loop using the value in the first iteration. This breaks
9878 long dependency chains, thus improving efficiency of the scheduling passes.
9880 A combination of @option{-fweb} and CSE is often sufficient to obtain the
9881 same effect. However, that is not reliable in cases where the loop body
9882 is more complicated than a single basic block. It also does not work at all
9883 on some architectures due to restrictions in the CSE pass.
9885 This optimization is enabled by default.
9887 @item -fvariable-expansion-in-unroller
9888 @opindex fvariable-expansion-in-unroller
9889 With this option, the compiler creates multiple copies of some
9890 local variables when unrolling a loop, which can result in superior code.
9892 This optimization is enabled by default for PowerPC targets, but disabled
9893 by default otherwise.
9895 @item -fpartial-inlining
9896 @opindex fpartial-inlining
9897 Inline parts of functions. This option has any effect only
9898 when inlining itself is turned on by the @option{-finline-functions}
9899 or @option{-finline-small-functions} options.
9901 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
9903 @item -fpredictive-commoning
9904 @opindex fpredictive-commoning
9905 Perform predictive commoning optimization, i.e., reusing computations
9906 (especially memory loads and stores) performed in previous
9907 iterations of loops.
9909 This option is enabled at level @option{-O3}.
9910 It is also enabled by @option{-fprofile-use} and @option{-fauto-profile}.
9912 @item -fprefetch-loop-arrays
9913 @opindex fprefetch-loop-arrays
9914 If supported by the target machine, generate instructions to prefetch
9915 memory to improve the performance of loops that access large arrays.
9917 This option may generate better or worse code; results are highly
9918 dependent on the structure of loops within the source code.
9920 Disabled at level @option{-Os}.
9922 @item -fno-printf-return-value
9923 @opindex fno-printf-return-value
9924 @opindex fprintf-return-value
9925 Do not substitute constants for known return value of formatted output
9926 functions such as @code{sprintf}, @code{snprintf}, @code{vsprintf}, and
9927 @code{vsnprintf} (but not @code{printf} of @code{fprintf}). This
9928 transformation allows GCC to optimize or even eliminate branches based
9929 on the known return value of these functions called with arguments that
9930 are either constant, or whose values are known to be in a range that
9931 makes determining the exact return value possible. For example, when
9932 @option{-fprintf-return-value} is in effect, both the branch and the
9933 body of the @code{if} statement (but not the call to @code{snprint})
9934 can be optimized away when @code{i} is a 32-bit or smaller integer
9935 because the return value is guaranteed to be at most 8.
9939 if (snprintf (buf, "%08x", i) >= sizeof buf)
9943 The @option{-fprintf-return-value} option relies on other optimizations
9944 and yields best results with @option{-O2} and above. It works in tandem
9945 with the @option{-Wformat-overflow} and @option{-Wformat-truncation}
9946 options. The @option{-fprintf-return-value} option is enabled by default.
9949 @itemx -fno-peephole2
9950 @opindex fno-peephole
9952 @opindex fno-peephole2
9954 Disable any machine-specific peephole optimizations. The difference
9955 between @option{-fno-peephole} and @option{-fno-peephole2} is in how they
9956 are implemented in the compiler; some targets use one, some use the
9957 other, a few use both.
9959 @option{-fpeephole} is enabled by default.
9960 @option{-fpeephole2} enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
9962 @item -fno-guess-branch-probability
9963 @opindex fno-guess-branch-probability
9964 @opindex fguess-branch-probability
9965 Do not guess branch probabilities using heuristics.
9967 GCC uses heuristics to guess branch probabilities if they are
9968 not provided by profiling feedback (@option{-fprofile-arcs}). These
9969 heuristics are based on the control flow graph. If some branch probabilities
9970 are specified by @code{__builtin_expect}, then the heuristics are
9971 used to guess branch probabilities for the rest of the control flow graph,
9972 taking the @code{__builtin_expect} info into account. The interactions
9973 between the heuristics and @code{__builtin_expect} can be complex, and in
9974 some cases, it may be useful to disable the heuristics so that the effects
9975 of @code{__builtin_expect} are easier to understand.
9977 It is also possible to specify expected probability of the expression
9978 with @code{__builtin_expect_with_probability} built-in function.
9980 The default is @option{-fguess-branch-probability} at levels
9981 @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
9983 @item -freorder-blocks
9984 @opindex freorder-blocks
9985 Reorder basic blocks in the compiled function in order to reduce number of
9986 taken branches and improve code locality.
9988 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
9990 @item -freorder-blocks-algorithm=@var{algorithm}
9991 @opindex freorder-blocks-algorithm
9992 Use the specified algorithm for basic block reordering. The
9993 @var{algorithm} argument can be @samp{simple}, which does not increase
9994 code size (except sometimes due to secondary effects like alignment),
9995 or @samp{stc}, the ``software trace cache'' algorithm, which tries to
9996 put all often executed code together, minimizing the number of branches
9997 executed by making extra copies of code.
9999 The default is @samp{simple} at levels @option{-O}, @option{-Os}, and
10000 @samp{stc} at levels @option{-O2}, @option{-O3}.
10002 @item -freorder-blocks-and-partition
10003 @opindex freorder-blocks-and-partition
10004 In addition to reordering basic blocks in the compiled function, in order
10005 to reduce number of taken branches, partitions hot and cold basic blocks
10006 into separate sections of the assembly and @file{.o} files, to improve
10007 paging and cache locality performance.
10009 This optimization is automatically turned off in the presence of
10010 exception handling or unwind tables (on targets using setjump/longjump or target specific scheme), for linkonce sections, for functions with a user-defined
10011 section attribute and on any architecture that does not support named
10012 sections. When @option{-fsplit-stack} is used this option is not
10013 enabled by default (to avoid linker errors), but may be enabled
10014 explicitly (if using a working linker).
10016 Enabled for x86 at levels @option{-O2}, @option{-O3}, @option{-Os}.
10018 @item -freorder-functions
10019 @opindex freorder-functions
10020 Reorder functions in the object file in order to
10021 improve code locality. This is implemented by using special
10022 subsections @code{.text.hot} for most frequently executed functions and
10023 @code{.text.unlikely} for unlikely executed functions. Reordering is done by
10024 the linker so object file format must support named sections and linker must
10025 place them in a reasonable way.
10027 This option isn't effective unless you either provide profile feedback
10028 (see @option{-fprofile-arcs} for details) or manually annotate functions with
10029 @code{hot} or @code{cold} attributes (@pxref{Common Function Attributes}).
10031 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
10033 @item -fstrict-aliasing
10034 @opindex fstrict-aliasing
10035 Allow the compiler to assume the strictest aliasing rules applicable to
10036 the language being compiled. For C (and C++), this activates
10037 optimizations based on the type of expressions. In particular, an
10038 object of one type is assumed never to reside at the same address as an
10039 object of a different type, unless the types are almost the same. For
10040 example, an @code{unsigned int} can alias an @code{int}, but not a
10041 @code{void*} or a @code{double}. A character type may alias any other
10044 @anchor{Type-punning}Pay special attention to code like this:
10057 The practice of reading from a different union member than the one most
10058 recently written to (called ``type-punning'') is common. Even with
10059 @option{-fstrict-aliasing}, type-punning is allowed, provided the memory
10060 is accessed through the union type. So, the code above works as
10061 expected. @xref{Structures unions enumerations and bit-fields
10062 implementation}. However, this code might not:
10073 Similarly, access by taking the address, casting the resulting pointer
10074 and dereferencing the result has undefined behavior, even if the cast
10075 uses a union type, e.g.:
10079 return ((union a_union *) &d)->i;
10083 The @option{-fstrict-aliasing} option is enabled at levels
10084 @option{-O2}, @option{-O3}, @option{-Os}.
10086 @item -falign-functions
10087 @itemx -falign-functions=@var{n}
10088 @itemx -falign-functions=@var{n}:@var{m}
10089 @itemx -falign-functions=@var{n}:@var{m}:@var{n2}
10090 @itemx -falign-functions=@var{n}:@var{m}:@var{n2}:@var{m2}
10091 @opindex falign-functions
10092 Align the start of functions to the next power-of-two greater than
10093 @var{n}, skipping up to @var{m}-1 bytes. This ensures that at least
10094 the first @var{m} bytes of the function can be fetched by the CPU
10095 without crossing an @var{n}-byte alignment boundary.
10097 If @var{m} is not specified, it defaults to @var{n}.
10099 Examples: @option{-falign-functions=32} aligns functions to the next
10100 32-byte boundary, @option{-falign-functions=24} aligns to the next
10101 32-byte boundary only if this can be done by skipping 23 bytes or less,
10102 @option{-falign-functions=32:7} aligns to the next
10103 32-byte boundary only if this can be done by skipping 6 bytes or less.
10105 The second pair of @var{n2}:@var{m2} values allows you to specify
10106 a secondary alignment: @option{-falign-functions=64:7:32:3} aligns to
10107 the next 64-byte boundary if this can be done by skipping 6 bytes or less,
10108 otherwise aligns to the next 32-byte boundary if this can be done
10109 by skipping 2 bytes or less.
10110 If @var{m2} is not specified, it defaults to @var{n2}.
10112 Some assemblers only support this flag when @var{n} is a power of two;
10113 in that case, it is rounded up.
10115 @option{-fno-align-functions} and @option{-falign-functions=1} are
10116 equivalent and mean that functions are not aligned.
10118 If @var{n} is not specified or is zero, use a machine-dependent default.
10119 The maximum allowed @var{n} option value is 65536.
10121 Enabled at levels @option{-O2}, @option{-O3}.
10123 @item -flimit-function-alignment
10124 If this option is enabled, the compiler tries to avoid unnecessarily
10125 overaligning functions. It attempts to instruct the assembler to align
10126 by the amount specified by @option{-falign-functions}, but not to
10127 skip more bytes than the size of the function.
10129 @item -falign-labels
10130 @itemx -falign-labels=@var{n}
10131 @itemx -falign-labels=@var{n}:@var{m}
10132 @itemx -falign-labels=@var{n}:@var{m}:@var{n2}
10133 @itemx -falign-labels=@var{n}:@var{m}:@var{n2}:@var{m2}
10134 @opindex falign-labels
10135 Align all branch targets to a power-of-two boundary.
10137 Parameters of this option are analogous to the @option{-falign-functions} option.
10138 @option{-fno-align-labels} and @option{-falign-labels=1} are
10139 equivalent and mean that labels are not aligned.
10141 If @option{-falign-loops} or @option{-falign-jumps} are applicable and
10142 are greater than this value, then their values are used instead.
10144 If @var{n} is not specified or is zero, use a machine-dependent default
10145 which is very likely to be @samp{1}, meaning no alignment.
10146 The maximum allowed @var{n} option value is 65536.
10148 Enabled at levels @option{-O2}, @option{-O3}.
10150 @item -falign-loops
10151 @itemx -falign-loops=@var{n}
10152 @itemx -falign-loops=@var{n}:@var{m}
10153 @itemx -falign-loops=@var{n}:@var{m}:@var{n2}
10154 @itemx -falign-loops=@var{n}:@var{m}:@var{n2}:@var{m2}
10155 @opindex falign-loops
10156 Align loops to a power-of-two boundary. If the loops are executed
10157 many times, this makes up for any execution of the dummy padding
10160 Parameters of this option are analogous to the @option{-falign-functions} option.
10161 @option{-fno-align-loops} and @option{-falign-loops=1} are
10162 equivalent and mean that loops are not aligned.
10163 The maximum allowed @var{n} option value is 65536.
10165 If @var{n} is not specified or is zero, use a machine-dependent default.
10167 Enabled at levels @option{-O2}, @option{-O3}.
10169 @item -falign-jumps
10170 @itemx -falign-jumps=@var{n}
10171 @itemx -falign-jumps=@var{n}:@var{m}
10172 @itemx -falign-jumps=@var{n}:@var{m}:@var{n2}
10173 @itemx -falign-jumps=@var{n}:@var{m}:@var{n2}:@var{m2}
10174 @opindex falign-jumps
10175 Align branch targets to a power-of-two boundary, for branch targets
10176 where the targets can only be reached by jumping. In this case,
10177 no dummy operations need be executed.
10179 Parameters of this option are analogous to the @option{-falign-functions} option.
10180 @option{-fno-align-jumps} and @option{-falign-jumps=1} are
10181 equivalent and mean that loops are not aligned.
10183 If @var{n} is not specified or is zero, use a machine-dependent default.
10184 The maximum allowed @var{n} option value is 65536.
10186 Enabled at levels @option{-O2}, @option{-O3}.
10188 @item -funit-at-a-time
10189 @opindex funit-at-a-time
10190 This option is left for compatibility reasons. @option{-funit-at-a-time}
10191 has no effect, while @option{-fno-unit-at-a-time} implies
10192 @option{-fno-toplevel-reorder} and @option{-fno-section-anchors}.
10194 Enabled by default.
10196 @item -fno-toplevel-reorder
10197 @opindex fno-toplevel-reorder
10198 @opindex ftoplevel-reorder
10199 Do not reorder top-level functions, variables, and @code{asm}
10200 statements. Output them in the same order that they appear in the
10201 input file. When this option is used, unreferenced static variables
10202 are not removed. This option is intended to support existing code
10203 that relies on a particular ordering. For new code, it is better to
10204 use attributes when possible.
10206 @option{-ftoplevel-reorder} is the default at @option{-O1} and higher, and
10207 also at @option{-O0} if @option{-fsection-anchors} is explicitly requested.
10208 Additionally @option{-fno-toplevel-reorder} implies
10209 @option{-fno-section-anchors}.
10213 Constructs webs as commonly used for register allocation purposes and assign
10214 each web individual pseudo register. This allows the register allocation pass
10215 to operate on pseudos directly, but also strengthens several other optimization
10216 passes, such as CSE, loop optimizer and trivial dead code remover. It can,
10217 however, make debugging impossible, since variables no longer stay in a
10220 Enabled by default with @option{-funroll-loops}.
10222 @item -fwhole-program
10223 @opindex fwhole-program
10224 Assume that the current compilation unit represents the whole program being
10225 compiled. All public functions and variables with the exception of @code{main}
10226 and those merged by attribute @code{externally_visible} become static functions
10227 and in effect are optimized more aggressively by interprocedural optimizers.
10229 This option should not be used in combination with @option{-flto}.
10230 Instead relying on a linker plugin should provide safer and more precise
10233 @item -flto[=@var{n}]
10235 This option runs the standard link-time optimizer. When invoked
10236 with source code, it generates GIMPLE (one of GCC's internal
10237 representations) and writes it to special ELF sections in the object
10238 file. When the object files are linked together, all the function
10239 bodies are read from these ELF sections and instantiated as if they
10240 had been part of the same translation unit.
10242 To use the link-time optimizer, @option{-flto} and optimization
10243 options should be specified at compile time and during the final link.
10244 It is recommended that you compile all the files participating in the
10245 same link with the same options and also specify those options at
10250 gcc -c -O2 -flto foo.c
10251 gcc -c -O2 -flto bar.c
10252 gcc -o myprog -flto -O2 foo.o bar.o
10255 The first two invocations to GCC save a bytecode representation
10256 of GIMPLE into special ELF sections inside @file{foo.o} and
10257 @file{bar.o}. The final invocation reads the GIMPLE bytecode from
10258 @file{foo.o} and @file{bar.o}, merges the two files into a single
10259 internal image, and compiles the result as usual. Since both
10260 @file{foo.o} and @file{bar.o} are merged into a single image, this
10261 causes all the interprocedural analyses and optimizations in GCC to
10262 work across the two files as if they were a single one. This means,
10263 for example, that the inliner is able to inline functions in
10264 @file{bar.o} into functions in @file{foo.o} and vice-versa.
10266 Another (simpler) way to enable link-time optimization is:
10269 gcc -o myprog -flto -O2 foo.c bar.c
10272 The above generates bytecode for @file{foo.c} and @file{bar.c},
10273 merges them together into a single GIMPLE representation and optimizes
10274 them as usual to produce @file{myprog}.
10276 The important thing to keep in mind is that to enable link-time
10277 optimizations you need to use the GCC driver to perform the link step.
10278 GCC automatically performs link-time optimization if any of the
10279 objects involved were compiled with the @option{-flto} command-line option.
10280 You can always override
10281 the automatic decision to do link-time optimization
10282 by passing @option{-fno-lto} to the link command.
10284 To make whole program optimization effective, it is necessary to make
10285 certain whole program assumptions. The compiler needs to know
10286 what functions and variables can be accessed by libraries and runtime
10287 outside of the link-time optimized unit. When supported by the linker,
10288 the linker plugin (see @option{-fuse-linker-plugin}) passes information
10289 to the compiler about used and externally visible symbols. When
10290 the linker plugin is not available, @option{-fwhole-program} should be
10291 used to allow the compiler to make these assumptions, which leads
10292 to more aggressive optimization decisions.
10294 When a file is compiled with @option{-flto} without
10295 @option{-fuse-linker-plugin}, the generated object file is larger than
10296 a regular object file because it contains GIMPLE bytecodes and the usual
10297 final code (see @option{-ffat-lto-objects}. This means that
10298 object files with LTO information can be linked as normal object
10299 files; if @option{-fno-lto} is passed to the linker, no
10300 interprocedural optimizations are applied. Note that when
10301 @option{-fno-fat-lto-objects} is enabled the compile stage is faster
10302 but you cannot perform a regular, non-LTO link on them.
10304 When producing the final binary, GCC only
10305 applies link-time optimizations to those files that contain bytecode.
10306 Therefore, you can mix and match object files and libraries with
10307 GIMPLE bytecodes and final object code. GCC automatically selects
10308 which files to optimize in LTO mode and which files to link without
10309 further processing.
10311 Generally, options specified at link time override those
10312 specified at compile time, although in some cases GCC attempts to infer
10313 link-time options from the settings used to compile the input files.
10315 If you do not specify an optimization level option @option{-O} at
10316 link time, then GCC uses the highest optimization level
10317 used when compiling the object files. Note that it is generally
10318 ineffective to specify an optimization level option only at link time and
10319 not at compile time, for two reasons. First, compiling without
10320 optimization suppresses compiler passes that gather information
10321 needed for effective optimization at link time. Second, some early
10322 optimization passes can be performed only at compile time and
10325 There are some code generation flags preserved by GCC when
10326 generating bytecodes, as they need to be used during the final link.
10327 Currently, the following options and their settings are taken from
10328 the first object file that explicitly specifies them:
10329 @option{-fPIC}, @option{-fpic}, @option{-fpie}, @option{-fcommon},
10330 @option{-fexceptions}, @option{-fnon-call-exceptions}, @option{-fgnu-tm}
10331 and all the @option{-m} target flags.
10333 Certain ABI-changing flags are required to match in all compilation units,
10334 and trying to override this at link time with a conflicting value
10335 is ignored. This includes options such as @option{-freg-struct-return}
10336 and @option{-fpcc-struct-return}.
10338 Other options such as @option{-ffp-contract}, @option{-fno-strict-overflow},
10339 @option{-fwrapv}, @option{-fno-trapv} or @option{-fno-strict-aliasing}
10340 are passed through to the link stage and merged conservatively for
10341 conflicting translation units. Specifically
10342 @option{-fno-strict-overflow}, @option{-fwrapv} and @option{-fno-trapv} take
10343 precedence; and for example @option{-ffp-contract=off} takes precedence
10344 over @option{-ffp-contract=fast}. You can override them at link time.
10346 To enable debug info generation you need to supply @option{-g} at
10347 compile-time. If any of the input files at link time were built
10348 with debug info generation enabled the link will enable debug info
10349 generation as well. Any elaborate debug info settings
10350 like the dwarf level @option{-gdwarf-5} need to be explicitly repeated
10351 at the linker command line and mixing different settings in different
10352 translation units is discouraged.
10354 If LTO encounters objects with C linkage declared with incompatible
10355 types in separate translation units to be linked together (undefined
10356 behavior according to ISO C99 6.2.7), a non-fatal diagnostic may be
10357 issued. The behavior is still undefined at run time. Similar
10358 diagnostics may be raised for other languages.
10360 Another feature of LTO is that it is possible to apply interprocedural
10361 optimizations on files written in different languages:
10365 g++ -c -flto bar.cc
10366 gfortran -c -flto baz.f90
10367 g++ -o myprog -flto -O3 foo.o bar.o baz.o -lgfortran
10370 Notice that the final link is done with @command{g++} to get the C++
10371 runtime libraries and @option{-lgfortran} is added to get the Fortran
10372 runtime libraries. In general, when mixing languages in LTO mode, you
10373 should use the same link command options as when mixing languages in a
10374 regular (non-LTO) compilation.
10376 If object files containing GIMPLE bytecode are stored in a library archive, say
10377 @file{libfoo.a}, it is possible to extract and use them in an LTO link if you
10378 are using a linker with plugin support. To create static libraries suitable
10379 for LTO, use @command{gcc-ar} and @command{gcc-ranlib} instead of @command{ar}
10380 and @command{ranlib};
10381 to show the symbols of object files with GIMPLE bytecode, use
10382 @command{gcc-nm}. Those commands require that @command{ar}, @command{ranlib}
10383 and @command{nm} have been compiled with plugin support. At link time, use the
10384 flag @option{-fuse-linker-plugin} to ensure that the library participates in
10385 the LTO optimization process:
10388 gcc -o myprog -O2 -flto -fuse-linker-plugin a.o b.o -lfoo
10391 With the linker plugin enabled, the linker extracts the needed
10392 GIMPLE files from @file{libfoo.a} and passes them on to the running GCC
10393 to make them part of the aggregated GIMPLE image to be optimized.
10395 If you are not using a linker with plugin support and/or do not
10396 enable the linker plugin, then the objects inside @file{libfoo.a}
10397 are extracted and linked as usual, but they do not participate
10398 in the LTO optimization process. In order to make a static library suitable
10399 for both LTO optimization and usual linkage, compile its object files with
10400 @option{-flto} @option{-ffat-lto-objects}.
10402 Link-time optimizations do not require the presence of the whole program to
10403 operate. If the program does not require any symbols to be exported, it is
10404 possible to combine @option{-flto} and @option{-fwhole-program} to allow
10405 the interprocedural optimizers to use more aggressive assumptions which may
10406 lead to improved optimization opportunities.
10407 Use of @option{-fwhole-program} is not needed when linker plugin is
10408 active (see @option{-fuse-linker-plugin}).
10410 The current implementation of LTO makes no
10411 attempt to generate bytecode that is portable between different
10412 types of hosts. The bytecode files are versioned and there is a
10413 strict version check, so bytecode files generated in one version of
10414 GCC do not work with an older or newer version of GCC.
10416 Link-time optimization does not work well with generation of debugging
10417 information on systems other than those using a combination of ELF and
10420 If you specify the optional @var{n}, the optimization and code
10421 generation done at link time is executed in parallel using @var{n}
10422 parallel jobs by utilizing an installed @command{make} program. The
10423 environment variable @env{MAKE} may be used to override the program
10426 You can also specify @option{-flto=jobserver} to use GNU make's
10427 job server mode to determine the number of parallel jobs. This
10428 is useful when the Makefile calling GCC is already executing in parallel.
10429 You must prepend a @samp{+} to the command recipe in the parent Makefile
10430 for this to work. This option likely only works if @env{MAKE} is
10431 GNU make. Even without the option value, GCC tries to automatically
10432 detect a running GNU make's job server.
10434 Use @option{-flto=auto} to use GNU make's job server, if available,
10435 or otherwise fall back to autodetection of the number of CPU threads
10436 present in your system.
10438 @item -flto-partition=@var{alg}
10439 @opindex flto-partition
10440 Specify the partitioning algorithm used by the link-time optimizer.
10441 The value is either @samp{1to1} to specify a partitioning mirroring
10442 the original source files or @samp{balanced} to specify partitioning
10443 into equally sized chunks (whenever possible) or @samp{max} to create
10444 new partition for every symbol where possible. Specifying @samp{none}
10445 as an algorithm disables partitioning and streaming completely.
10446 The default value is @samp{balanced}. While @samp{1to1} can be used
10447 as an workaround for various code ordering issues, the @samp{max}
10448 partitioning is intended for internal testing only.
10449 The value @samp{one} specifies that exactly one partition should be
10450 used while the value @samp{none} bypasses partitioning and executes
10451 the link-time optimization step directly from the WPA phase.
10453 @item -flto-compression-level=@var{n}
10454 @opindex flto-compression-level
10455 This option specifies the level of compression used for intermediate
10456 language written to LTO object files, and is only meaningful in
10457 conjunction with LTO mode (@option{-flto}). Valid
10458 values are 0 (no compression) to 9 (maximum compression). Values
10459 outside this range are clamped to either 0 or 9. If the option is not
10460 given, a default balanced compression setting is used.
10462 @item -fuse-linker-plugin
10463 @opindex fuse-linker-plugin
10464 Enables the use of a linker plugin during link-time optimization. This
10465 option relies on plugin support in the linker, which is available in gold
10466 or in GNU ld 2.21 or newer.
10468 This option enables the extraction of object files with GIMPLE bytecode out
10469 of library archives. This improves the quality of optimization by exposing
10470 more code to the link-time optimizer. This information specifies what
10471 symbols can be accessed externally (by non-LTO object or during dynamic
10472 linking). Resulting code quality improvements on binaries (and shared
10473 libraries that use hidden visibility) are similar to @option{-fwhole-program}.
10474 See @option{-flto} for a description of the effect of this flag and how to
10477 This option is enabled by default when LTO support in GCC is enabled
10478 and GCC was configured for use with
10479 a linker supporting plugins (GNU ld 2.21 or newer or gold).
10481 @item -ffat-lto-objects
10482 @opindex ffat-lto-objects
10483 Fat LTO objects are object files that contain both the intermediate language
10484 and the object code. This makes them usable for both LTO linking and normal
10485 linking. This option is effective only when compiling with @option{-flto}
10486 and is ignored at link time.
10488 @option{-fno-fat-lto-objects} improves compilation time over plain LTO, but
10489 requires the complete toolchain to be aware of LTO. It requires a linker with
10490 linker plugin support for basic functionality. Additionally,
10491 @command{nm}, @command{ar} and @command{ranlib}
10492 need to support linker plugins to allow a full-featured build environment
10493 (capable of building static libraries etc). GCC provides the @command{gcc-ar},
10494 @command{gcc-nm}, @command{gcc-ranlib} wrappers to pass the right options
10495 to these tools. With non fat LTO makefiles need to be modified to use them.
10497 Note that modern binutils provide plugin auto-load mechanism.
10498 Installing the linker plugin into @file{$libdir/bfd-plugins} has the same
10499 effect as usage of the command wrappers (@command{gcc-ar}, @command{gcc-nm} and
10500 @command{gcc-ranlib}).
10502 The default is @option{-fno-fat-lto-objects} on targets with linker plugin
10505 @item -fcompare-elim
10506 @opindex fcompare-elim
10507 After register allocation and post-register allocation instruction splitting,
10508 identify arithmetic instructions that compute processor flags similar to a
10509 comparison operation based on that arithmetic. If possible, eliminate the
10510 explicit comparison operation.
10512 This pass only applies to certain targets that cannot explicitly represent
10513 the comparison operation before register allocation is complete.
10515 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
10517 @item -fcprop-registers
10518 @opindex fcprop-registers
10519 After register allocation and post-register allocation instruction splitting,
10520 perform a copy-propagation pass to try to reduce scheduling dependencies
10521 and occasionally eliminate the copy.
10523 Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
10525 @item -fprofile-correction
10526 @opindex fprofile-correction
10527 Profiles collected using an instrumented binary for multi-threaded programs may
10528 be inconsistent due to missed counter updates. When this option is specified,
10529 GCC uses heuristics to correct or smooth out such inconsistencies. By
10530 default, GCC emits an error message when an inconsistent profile is detected.
10532 This option is enabled by @option{-fauto-profile}.
10534 @item -fprofile-use
10535 @itemx -fprofile-use=@var{path}
10536 @opindex fprofile-use
10537 Enable profile feedback-directed optimizations,
10538 and the following optimizations, many of which
10539 are generally profitable only with profile feedback available:
10541 @gccoptlist{-fbranch-probabilities -fprofile-values @gol
10542 -funroll-loops -fpeel-loops -ftracer -fvpt @gol
10543 -finline-functions -fipa-cp -fipa-cp-clone -fipa-bit-cp @gol
10544 -fpredictive-commoning -fsplit-loops -funswitch-loops @gol
10545 -fgcse-after-reload -ftree-loop-vectorize -ftree-slp-vectorize @gol
10546 -fvect-cost-model=dynamic -ftree-loop-distribute-patterns @gol
10547 -fprofile-reorder-functions}
10549 Before you can use this option, you must first generate profiling information.
10550 @xref{Instrumentation Options}, for information about the
10551 @option{-fprofile-generate} option.
10553 By default, GCC emits an error message if the feedback profiles do not
10554 match the source code. This error can be turned into a warning by using
10555 @option{-Wno-error=coverage-mismatch}. Note this may result in poorly
10556 optimized code. Additionally, by default, GCC also emits a warning message if
10557 the feedback profiles do not exist (see @option{-Wmissing-profile}).
10559 If @var{path} is specified, GCC looks at the @var{path} to find
10560 the profile feedback data files. See @option{-fprofile-dir}.
10562 @item -fauto-profile
10563 @itemx -fauto-profile=@var{path}
10564 @opindex fauto-profile
10565 Enable sampling-based feedback-directed optimizations,
10566 and the following optimizations,
10567 many of which are generally profitable only with profile feedback available:
10569 @gccoptlist{-fbranch-probabilities -fprofile-values @gol
10570 -funroll-loops -fpeel-loops -ftracer -fvpt @gol
10571 -finline-functions -fipa-cp -fipa-cp-clone -fipa-bit-cp @gol
10572 -fpredictive-commoning -fsplit-loops -funswitch-loops @gol
10573 -fgcse-after-reload -ftree-loop-vectorize -ftree-slp-vectorize @gol
10574 -fvect-cost-model=dynamic -ftree-loop-distribute-patterns @gol
10575 -fprofile-correction}
10577 @var{path} is the name of a file containing AutoFDO profile information.
10578 If omitted, it defaults to @file{fbdata.afdo} in the current directory.
10580 Producing an AutoFDO profile data file requires running your program
10581 with the @command{perf} utility on a supported GNU/Linux target system.
10582 For more information, see @uref{https://perf.wiki.kernel.org/}.
10586 perf record -e br_inst_retired:near_taken -b -o perf.data \
10590 Then use the @command{create_gcov} tool to convert the raw profile data
10591 to a format that can be used by GCC.@ You must also supply the
10592 unstripped binary for your program to this tool.
10593 See @uref{https://github.com/google/autofdo}.
10597 create_gcov --binary=your_program.unstripped --profile=perf.data \
10598 --gcov=profile.afdo
10602 The following options control compiler behavior regarding floating-point
10603 arithmetic. These options trade off between speed and
10604 correctness. All must be specifically enabled.
10607 @item -ffloat-store
10608 @opindex ffloat-store
10609 Do not store floating-point variables in registers, and inhibit other
10610 options that might change whether a floating-point value is taken from a
10611 register or memory.
10613 @cindex floating-point precision
10614 This option prevents undesirable excess precision on machines such as
10615 the 68000 where the floating registers (of the 68881) keep more
10616 precision than a @code{double} is supposed to have. Similarly for the
10617 x86 architecture. For most programs, the excess precision does only
10618 good, but a few programs rely on the precise definition of IEEE floating
10619 point. Use @option{-ffloat-store} for such programs, after modifying
10620 them to store all pertinent intermediate computations into variables.
10622 @item -fexcess-precision=@var{style}
10623 @opindex fexcess-precision
10624 This option allows further control over excess precision on machines
10625 where floating-point operations occur in a format with more precision or
10626 range than the IEEE standard and interchange floating-point types. By
10627 default, @option{-fexcess-precision=fast} is in effect; this means that
10628 operations may be carried out in a wider precision than the types specified
10629 in the source if that would result in faster code, and it is unpredictable
10630 when rounding to the types specified in the source code takes place.
10631 When compiling C, if @option{-fexcess-precision=standard} is specified then
10632 excess precision follows the rules specified in ISO C99; in particular,
10633 both casts and assignments cause values to be rounded to their
10634 semantic types (whereas @option{-ffloat-store} only affects
10635 assignments). This option is enabled by default for C if a strict
10636 conformance option such as @option{-std=c99} is used.
10637 @option{-ffast-math} enables @option{-fexcess-precision=fast} by default
10638 regardless of whether a strict conformance option is used.
10641 @option{-fexcess-precision=standard} is not implemented for languages
10642 other than C. On the x86, it has no effect if @option{-mfpmath=sse}
10643 or @option{-mfpmath=sse+387} is specified; in the former case, IEEE
10644 semantics apply without excess precision, and in the latter, rounding
10648 @opindex ffast-math
10649 Sets the options @option{-fno-math-errno}, @option{-funsafe-math-optimizations},
10650 @option{-ffinite-math-only}, @option{-fno-rounding-math},
10651 @option{-fno-signaling-nans}, @option{-fcx-limited-range} and
10652 @option{-fexcess-precision=fast}.
10654 This option causes the preprocessor macro @code{__FAST_MATH__} to be defined.
10656 This option is not turned on by any @option{-O} option besides
10657 @option{-Ofast} since it can result in incorrect output for programs
10658 that depend on an exact implementation of IEEE or ISO rules/specifications
10659 for math functions. It may, however, yield faster code for programs
10660 that do not require the guarantees of these specifications.
10662 @item -fno-math-errno
10663 @opindex fno-math-errno
10664 @opindex fmath-errno
10665 Do not set @code{errno} after calling math functions that are executed
10666 with a single instruction, e.g., @code{sqrt}. A program that relies on
10667 IEEE exceptions for math error handling may want to use this flag
10668 for speed while maintaining IEEE arithmetic compatibility.
10670 This option is not turned on by any @option{-O} option since
10671 it can result in incorrect output for programs that depend on
10672 an exact implementation of IEEE or ISO rules/specifications for
10673 math functions. It may, however, yield faster code for programs
10674 that do not require the guarantees of these specifications.
10676 The default is @option{-fmath-errno}.
10678 On Darwin systems, the math library never sets @code{errno}. There is
10679 therefore no reason for the compiler to consider the possibility that
10680 it might, and @option{-fno-math-errno} is the default.
10682 @item -funsafe-math-optimizations
10683 @opindex funsafe-math-optimizations
10685 Allow optimizations for floating-point arithmetic that (a) assume
10686 that arguments and results are valid and (b) may violate IEEE or
10687 ANSI standards. When used at link time, it may include libraries
10688 or startup files that change the default FPU control word or other
10689 similar optimizations.
10691 This option is not turned on by any @option{-O} option since
10692 it can result in incorrect output for programs that depend on
10693 an exact implementation of IEEE or ISO rules/specifications for
10694 math functions. It may, however, yield faster code for programs
10695 that do not require the guarantees of these specifications.
10696 Enables @option{-fno-signed-zeros}, @option{-fno-trapping-math},
10697 @option{-fassociative-math} and @option{-freciprocal-math}.
10699 The default is @option{-fno-unsafe-math-optimizations}.
10701 @item -fassociative-math
10702 @opindex fassociative-math
10704 Allow re-association of operands in series of floating-point operations.
10705 This violates the ISO C and C++ language standard by possibly changing
10706 computation result. NOTE: re-ordering may change the sign of zero as
10707 well as ignore NaNs and inhibit or create underflow or overflow (and
10708 thus cannot be used on code that relies on rounding behavior like
10709 @code{(x + 2**52) - 2**52}. May also reorder floating-point comparisons
10710 and thus may not be used when ordered comparisons are required.
10711 This option requires that both @option{-fno-signed-zeros} and
10712 @option{-fno-trapping-math} be in effect. Moreover, it doesn't make
10713 much sense with @option{-frounding-math}. For Fortran the option
10714 is automatically enabled when both @option{-fno-signed-zeros} and
10715 @option{-fno-trapping-math} are in effect.
10717 The default is @option{-fno-associative-math}.
10719 @item -freciprocal-math
10720 @opindex freciprocal-math
10722 Allow the reciprocal of a value to be used instead of dividing by
10723 the value if this enables optimizations. For example @code{x / y}
10724 can be replaced with @code{x * (1/y)}, which is useful if @code{(1/y)}
10725 is subject to common subexpression elimination. Note that this loses
10726 precision and increases the number of flops operating on the value.
10728 The default is @option{-fno-reciprocal-math}.
10730 @item -ffinite-math-only
10731 @opindex ffinite-math-only
10732 Allow optimizations for floating-point arithmetic that assume
10733 that arguments and results are not NaNs or +-Infs.
10735 This option is not turned on by any @option{-O} option since
10736 it can result in incorrect output for programs that depend on
10737 an exact implementation of IEEE or ISO rules/specifications for
10738 math functions. It may, however, yield faster code for programs
10739 that do not require the guarantees of these specifications.
10741 The default is @option{-fno-finite-math-only}.
10743 @item -fno-signed-zeros
10744 @opindex fno-signed-zeros
10745 @opindex fsigned-zeros
10746 Allow optimizations for floating-point arithmetic that ignore the
10747 signedness of zero. IEEE arithmetic specifies the behavior of
10748 distinct +0.0 and @minus{}0.0 values, which then prohibits simplification
10749 of expressions such as x+0.0 or 0.0*x (even with @option{-ffinite-math-only}).
10750 This option implies that the sign of a zero result isn't significant.
10752 The default is @option{-fsigned-zeros}.
10754 @item -fno-trapping-math
10755 @opindex fno-trapping-math
10756 @opindex ftrapping-math
10757 Compile code assuming that floating-point operations cannot generate
10758 user-visible traps. These traps include division by zero, overflow,
10759 underflow, inexact result and invalid operation. This option requires
10760 that @option{-fno-signaling-nans} be in effect. Setting this option may
10761 allow faster code if one relies on ``non-stop'' IEEE arithmetic, for example.
10763 This option should never be turned on by any @option{-O} option since
10764 it can result in incorrect output for programs that depend on
10765 an exact implementation of IEEE or ISO rules/specifications for
10768 The default is @option{-ftrapping-math}.
10770 @item -frounding-math
10771 @opindex frounding-math
10772 Disable transformations and optimizations that assume default floating-point
10773 rounding behavior. This is round-to-zero for all floating point
10774 to integer conversions, and round-to-nearest for all other arithmetic
10775 truncations. This option should be specified for programs that change
10776 the FP rounding mode dynamically, or that may be executed with a
10777 non-default rounding mode. This option disables constant folding of
10778 floating-point expressions at compile time (which may be affected by
10779 rounding mode) and arithmetic transformations that are unsafe in the
10780 presence of sign-dependent rounding modes.
10782 The default is @option{-fno-rounding-math}.
10784 This option is experimental and does not currently guarantee to
10785 disable all GCC optimizations that are affected by rounding mode.
10786 Future versions of GCC may provide finer control of this setting
10787 using C99's @code{FENV_ACCESS} pragma. This command-line option
10788 will be used to specify the default state for @code{FENV_ACCESS}.
10790 @item -fsignaling-nans
10791 @opindex fsignaling-nans
10792 Compile code assuming that IEEE signaling NaNs may generate user-visible
10793 traps during floating-point operations. Setting this option disables
10794 optimizations that may change the number of exceptions visible with
10795 signaling NaNs. This option implies @option{-ftrapping-math}.
10797 This option causes the preprocessor macro @code{__SUPPORT_SNAN__} to
10800 The default is @option{-fno-signaling-nans}.
10802 This option is experimental and does not currently guarantee to
10803 disable all GCC optimizations that affect signaling NaN behavior.
10805 @item -fno-fp-int-builtin-inexact
10806 @opindex fno-fp-int-builtin-inexact
10807 @opindex ffp-int-builtin-inexact
10808 Do not allow the built-in functions @code{ceil}, @code{floor},
10809 @code{round} and @code{trunc}, and their @code{float} and @code{long
10810 double} variants, to generate code that raises the ``inexact''
10811 floating-point exception for noninteger arguments. ISO C99 and C11
10812 allow these functions to raise the ``inexact'' exception, but ISO/IEC
10813 TS 18661-1:2014, the C bindings to IEEE 754-2008, does not allow these
10814 functions to do so.
10816 The default is @option{-ffp-int-builtin-inexact}, allowing the
10817 exception to be raised. This option does nothing unless
10818 @option{-ftrapping-math} is in effect.
10820 Even if @option{-fno-fp-int-builtin-inexact} is used, if the functions
10821 generate a call to a library function then the ``inexact'' exception
10822 may be raised if the library implementation does not follow TS 18661.
10824 @item -fsingle-precision-constant
10825 @opindex fsingle-precision-constant
10826 Treat floating-point constants as single precision instead of
10827 implicitly converting them to double-precision constants.
10829 @item -fcx-limited-range
10830 @opindex fcx-limited-range
10831 When enabled, this option states that a range reduction step is not
10832 needed when performing complex division. Also, there is no checking
10833 whether the result of a complex multiplication or division is @code{NaN
10834 + I*NaN}, with an attempt to rescue the situation in that case. The
10835 default is @option{-fno-cx-limited-range}, but is enabled by
10836 @option{-ffast-math}.
10838 This option controls the default setting of the ISO C99
10839 @code{CX_LIMITED_RANGE} pragma. Nevertheless, the option applies to
10842 @item -fcx-fortran-rules
10843 @opindex fcx-fortran-rules
10844 Complex multiplication and division follow Fortran rules. Range
10845 reduction is done as part of complex division, but there is no checking
10846 whether the result of a complex multiplication or division is @code{NaN
10847 + I*NaN}, with an attempt to rescue the situation in that case.
10849 The default is @option{-fno-cx-fortran-rules}.
10853 The following options control optimizations that may improve
10854 performance, but are not enabled by any @option{-O} options. This
10855 section includes experimental options that may produce broken code.
10858 @item -fbranch-probabilities
10859 @opindex fbranch-probabilities
10860 After running a program compiled with @option{-fprofile-arcs}
10861 (@pxref{Instrumentation Options}),
10862 you can compile it a second time using
10863 @option{-fbranch-probabilities}, to improve optimizations based on
10864 the number of times each branch was taken. When a program
10865 compiled with @option{-fprofile-arcs} exits, it saves arc execution
10866 counts to a file called @file{@var{sourcename}.gcda} for each source
10867 file. The information in this data file is very dependent on the
10868 structure of the generated code, so you must use the same source code
10869 and the same optimization options for both compilations.
10871 With @option{-fbranch-probabilities}, GCC puts a
10872 @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}.
10873 These can be used to improve optimization. Currently, they are only
10874 used in one place: in @file{reorg.c}, instead of guessing which path a
10875 branch is most likely to take, the @samp{REG_BR_PROB} values are used to
10876 exactly determine which path is taken more often.
10878 Enabled by @option{-fprofile-use} and @option{-fauto-profile}.
10880 @item -fprofile-values
10881 @opindex fprofile-values
10882 If combined with @option{-fprofile-arcs}, it adds code so that some
10883 data about values of expressions in the program is gathered.
10885 With @option{-fbranch-probabilities}, it reads back the data gathered
10886 from profiling values of expressions for usage in optimizations.
10888 Enabled by @option{-fprofile-generate}, @option{-fprofile-use}, and
10889 @option{-fauto-profile}.
10891 @item -fprofile-reorder-functions
10892 @opindex fprofile-reorder-functions
10893 Function reordering based on profile instrumentation collects
10894 first time of execution of a function and orders these functions
10895 in ascending order.
10897 Enabled with @option{-fprofile-use}.
10901 If combined with @option{-fprofile-arcs}, this option instructs the compiler
10902 to add code to gather information about values of expressions.
10904 With @option{-fbranch-probabilities}, it reads back the data gathered
10905 and actually performs the optimizations based on them.
10906 Currently the optimizations include specialization of division operations
10907 using the knowledge about the value of the denominator.
10909 Enabled with @option{-fprofile-use} and @option{-fauto-profile}.
10911 @item -frename-registers
10912 @opindex frename-registers
10913 Attempt to avoid false dependencies in scheduled code by making use
10914 of registers left over after register allocation. This optimization
10915 most benefits processors with lots of registers. Depending on the
10916 debug information format adopted by the target, however, it can
10917 make debugging impossible, since variables no longer stay in
10918 a ``home register''.
10920 Enabled by default with @option{-funroll-loops}.
10922 @item -fschedule-fusion
10923 @opindex fschedule-fusion
10924 Performs a target dependent pass over the instruction stream to schedule
10925 instructions of same type together because target machine can execute them
10926 more efficiently if they are adjacent to each other in the instruction flow.
10928 Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
10932 Perform tail duplication to enlarge superblock size. This transformation
10933 simplifies the control flow of the function allowing other optimizations to do
10936 Enabled by @option{-fprofile-use} and @option{-fauto-profile}.
10938 @item -funroll-loops
10939 @opindex funroll-loops
10940 Unroll loops whose number of iterations can be determined at compile time or
10941 upon entry to the loop. @option{-funroll-loops} implies
10942 @option{-frerun-cse-after-loop}, @option{-fweb} and @option{-frename-registers}.
10943 It also turns on complete loop peeling (i.e.@: complete removal of loops with
10944 a small constant number of iterations). This option makes code larger, and may
10945 or may not make it run faster.
10947 Enabled by @option{-fprofile-use} and @option{-fauto-profile}.
10949 @item -funroll-all-loops
10950 @opindex funroll-all-loops
10951 Unroll all loops, even if their number of iterations is uncertain when
10952 the loop is entered. This usually makes programs run more slowly.
10953 @option{-funroll-all-loops} implies the same options as
10954 @option{-funroll-loops}.
10957 @opindex fpeel-loops
10958 Peels loops for which there is enough information that they do not
10959 roll much (from profile feedback or static analysis). It also turns on
10960 complete loop peeling (i.e.@: complete removal of loops with small constant
10961 number of iterations).
10963 Enabled by @option{-O3}, @option{-fprofile-use}, and @option{-fauto-profile}.
10965 @item -fmove-loop-invariants
10966 @opindex fmove-loop-invariants
10967 Enables the loop invariant motion pass in the RTL loop optimizer. Enabled
10968 at level @option{-O1} and higher, except for @option{-Og}.
10970 @item -fsplit-loops
10971 @opindex fsplit-loops
10972 Split a loop into two if it contains a condition that's always true
10973 for one side of the iteration space and false for the other.
10975 Enabled by @option{-fprofile-use} and @option{-fauto-profile}.
10977 @item -funswitch-loops
10978 @opindex funswitch-loops
10979 Move branches with loop invariant conditions out of the loop, with duplicates
10980 of the loop on both branches (modified according to result of the condition).
10982 Enabled by @option{-fprofile-use} and @option{-fauto-profile}.
10984 @item -fversion-loops-for-strides
10985 @opindex fversion-loops-for-strides
10986 If a loop iterates over an array with a variable stride, create another
10987 version of the loop that assumes the stride is always one. For example:
10990 for (int i = 0; i < n; ++i)
10991 x[i * stride] = @dots{};
10998 for (int i = 0; i < n; ++i)
11001 for (int i = 0; i < n; ++i)
11002 x[i * stride] = @dots{};
11005 This is particularly useful for assumed-shape arrays in Fortran where
11006 (for example) it allows better vectorization assuming contiguous accesses.
11007 This flag is enabled by default at @option{-O3}.
11008 It is also enabled by @option{-fprofile-use} and @option{-fauto-profile}.
11010 @item -ffunction-sections
11011 @itemx -fdata-sections
11012 @opindex ffunction-sections
11013 @opindex fdata-sections
11014 Place each function or data item into its own section in the output
11015 file if the target supports arbitrary sections. The name of the
11016 function or the name of the data item determines the section's name
11017 in the output file.
11019 Use these options on systems where the linker can perform optimizations to
11020 improve locality of reference in the instruction space. Most systems using the
11021 ELF object format have linkers with such optimizations. On AIX, the linker
11022 rearranges sections (CSECTs) based on the call graph. The performance impact
11025 Together with a linker garbage collection (linker @option{--gc-sections}
11026 option) these options may lead to smaller statically-linked executables (after
11029 On ELF/DWARF systems these options do not degenerate the quality of the debug
11030 information. There could be issues with other object files/debug info formats.
11032 Only use these options when there are significant benefits from doing so. When
11033 you specify these options, the assembler and linker create larger object and
11034 executable files and are also slower. These options affect code generation.
11035 They prevent optimizations by the compiler and assembler using relative
11036 locations inside a translation unit since the locations are unknown until
11037 link time. An example of such an optimization is relaxing calls to short call
11041 @opindex fstdarg-opt
11042 Optimize the prologue of variadic argument functions with respect to usage of
11045 @item -fsection-anchors
11046 @opindex fsection-anchors
11047 Try to reduce the number of symbolic address calculations by using
11048 shared ``anchor'' symbols to address nearby objects. This transformation
11049 can help to reduce the number of GOT entries and GOT accesses on some
11052 For example, the implementation of the following function @code{foo}:
11055 static int a, b, c;
11056 int foo (void) @{ return a + b + c; @}
11060 usually calculates the addresses of all three variables, but if you
11061 compile it with @option{-fsection-anchors}, it accesses the variables
11062 from a common anchor point instead. The effect is similar to the
11063 following pseudocode (which isn't valid C):
11068 register int *xr = &x;
11069 return xr[&a - &x] + xr[&b - &x] + xr[&c - &x];
11073 Not all targets support this option.
11075 @item --param @var{name}=@var{value}
11077 In some places, GCC uses various constants to control the amount of
11078 optimization that is done. For example, GCC does not inline functions
11079 that contain more than a certain number of instructions. You can
11080 control some of these constants on the command line using the
11081 @option{--param} option.
11083 The names of specific parameters, and the meaning of the values, are
11084 tied to the internals of the compiler, and are subject to change
11085 without notice in future releases.
11087 In order to get minimal, maximal and default value of a parameter,
11088 one can use @option{--help=param -Q} options.
11090 In each case, the @var{value} is an integer. The allowable choices for
11094 @item predictable-branch-outcome
11095 When branch is predicted to be taken with probability lower than this threshold
11096 (in percent), then it is considered well predictable.
11098 @item max-rtl-if-conversion-insns
11099 RTL if-conversion tries to remove conditional branches around a block and
11100 replace them with conditionally executed instructions. This parameter
11101 gives the maximum number of instructions in a block which should be
11102 considered for if-conversion. The compiler will
11103 also use other heuristics to decide whether if-conversion is likely to be
11106 @item max-rtl-if-conversion-predictable-cost
11107 @itemx max-rtl-if-conversion-unpredictable-cost
11108 RTL if-conversion will try to remove conditional branches around a block
11109 and replace them with conditionally executed instructions. These parameters
11110 give the maximum permissible cost for the sequence that would be generated
11111 by if-conversion depending on whether the branch is statically determined
11112 to be predictable or not. The units for this parameter are the same as
11113 those for the GCC internal seq_cost metric. The compiler will try to
11114 provide a reasonable default for this parameter using the BRANCH_COST
11117 @item max-crossjump-edges
11118 The maximum number of incoming edges to consider for cross-jumping.
11119 The algorithm used by @option{-fcrossjumping} is @math{O(N^2)} in
11120 the number of edges incoming to each block. Increasing values mean
11121 more aggressive optimization, making the compilation time increase with
11122 probably small improvement in executable size.
11124 @item min-crossjump-insns
11125 The minimum number of instructions that must be matched at the end
11126 of two blocks before cross-jumping is performed on them. This
11127 value is ignored in the case where all instructions in the block being
11128 cross-jumped from are matched.
11130 @item max-grow-copy-bb-insns
11131 The maximum code size expansion factor when copying basic blocks
11132 instead of jumping. The expansion is relative to a jump instruction.
11134 @item max-goto-duplication-insns
11135 The maximum number of instructions to duplicate to a block that jumps
11136 to a computed goto. To avoid @math{O(N^2)} behavior in a number of
11137 passes, GCC factors computed gotos early in the compilation process,
11138 and unfactors them as late as possible. Only computed jumps at the
11139 end of a basic blocks with no more than max-goto-duplication-insns are
11142 @item max-delay-slot-insn-search
11143 The maximum number of instructions to consider when looking for an
11144 instruction to fill a delay slot. If more than this arbitrary number of
11145 instructions are searched, the time savings from filling the delay slot
11146 are minimal, so stop searching. Increasing values mean more
11147 aggressive optimization, making the compilation time increase with probably
11148 small improvement in execution time.
11150 @item max-delay-slot-live-search
11151 When trying to fill delay slots, the maximum number of instructions to
11152 consider when searching for a block with valid live register
11153 information. Increasing this arbitrarily chosen value means more
11154 aggressive optimization, increasing the compilation time. This parameter
11155 should be removed when the delay slot code is rewritten to maintain the
11156 control-flow graph.
11158 @item max-gcse-memory
11159 The approximate maximum amount of memory that can be allocated in
11160 order to perform the global common subexpression elimination
11161 optimization. If more memory than specified is required, the
11162 optimization is not done.
11164 @item max-gcse-insertion-ratio
11165 If the ratio of expression insertions to deletions is larger than this value
11166 for any expression, then RTL PRE inserts or removes the expression and thus
11167 leaves partially redundant computations in the instruction stream.
11169 @item max-pending-list-length
11170 The maximum number of pending dependencies scheduling allows
11171 before flushing the current state and starting over. Large functions
11172 with few branches or calls can create excessively large lists which
11173 needlessly consume memory and resources.
11175 @item max-modulo-backtrack-attempts
11176 The maximum number of backtrack attempts the scheduler should make
11177 when modulo scheduling a loop. Larger values can exponentially increase
11180 @item max-inline-insns-single
11181 @item max-inline-insns-single-O2
11182 Several parameters control the tree inliner used in GCC@. This number sets the
11183 maximum number of instructions (counted in GCC's internal representation) in a
11184 single function that the tree inliner considers for inlining. This only
11185 affects functions declared inline and methods implemented in a class
11188 For functions compiled with optimization levels
11189 @option{-O3} and @option{-Ofast} parameter @option{max-inline-insns-single} is
11190 applied. In other cases @option{max-inline-insns-single-O2} is applied.
11193 @item max-inline-insns-auto
11194 @item max-inline-insns-auto-O2
11195 When you use @option{-finline-functions} (included in @option{-O3}),
11196 a lot of functions that would otherwise not be considered for inlining
11197 by the compiler are investigated. To those functions, a different
11198 (more restrictive) limit compared to functions declared inline can
11201 For functions compiled with optimization levels
11202 @option{-O3} and @option{-Ofast} parameter @option{max-inline-insns-auto} is
11203 applied. In other cases @option{max-inline-insns-auto-O2} is applied.
11205 @item max-inline-insns-small
11206 This is bound applied to calls which are considered relevant with
11207 @option{-finline-small-functions}.
11209 @item max-inline-insns-size
11210 This is bound applied to calls which are optimized for size. Small growth
11211 may be desirable to anticipate optimization oppurtunities exposed by inlining.
11213 @item uninlined-function-insns
11214 Number of instructions accounted by inliner for function overhead such as
11215 function prologue and epilogue.
11217 @item uninlined-function-time
11218 Extra time accounted by inliner for function overhead such as time needed to
11219 execute function prologue and epilogue
11221 @item inline-heuristics-hint-percent
11222 @item inline-heuristics-hint-percent-O2
11223 The scale (in percents) applied to @option{inline-insns-single},
11224 @option{inline-insns-single-O2}, @option{inline-insns-auto},
11225 @option{inline-insns-auto-O2} when inline heuristics hints that inlining is
11226 very profitable (will enable later optimizations).
11228 For functions compiled with optimization levels
11229 @option{-O3} and @option{-Ofast} parameter
11230 @option{inline-heuristics-hint-percent} is applied. In other cases
11231 @option{inline-heuristics-hint-percent-O2} is applied.
11233 @item uninlined-thunk-insns
11234 @item uninlined-thunk-time
11235 Same as @option{--param uninlined-function-insns} and
11236 @option{--param uninlined-function-time} but applied to function thunks
11238 @item inline-min-speedup
11239 @item inline-min-speedup-O2
11240 When estimated performance improvement of caller + callee runtime exceeds this
11241 threshold (in percent), the function can be inlined regardless of the limit on
11242 @option{--param max-inline-insns-single} and @option{--param
11243 max-inline-insns-auto}.
11245 For functions compiled with optimization levels
11246 @option{-O3} and @option{-Ofast} parameter @option{inline-min-speedup} is
11247 applied. In other cases @option{inline-min-speedup-O2} is applied.
11249 @item large-function-insns
11250 The limit specifying really large functions. For functions larger than this
11251 limit after inlining, inlining is constrained by
11252 @option{--param large-function-growth}. This parameter is useful primarily
11253 to avoid extreme compilation time caused by non-linear algorithms used by the
11256 @item large-function-growth
11257 Specifies maximal growth of large function caused by inlining in percents.
11258 For example, parameter value 100 limits large function growth to 2.0 times
11261 @item large-unit-insns
11262 The limit specifying large translation unit. Growth caused by inlining of
11263 units larger than this limit is limited by @option{--param inline-unit-growth}.
11264 For small units this might be too tight.
11265 For example, consider a unit consisting of function A
11266 that is inline and B that just calls A three times. If B is small relative to
11267 A, the growth of unit is 300\% and yet such inlining is very sane. For very
11268 large units consisting of small inlineable functions, however, the overall unit
11269 growth limit is needed to avoid exponential explosion of code size. Thus for
11270 smaller units, the size is increased to @option{--param large-unit-insns}
11271 before applying @option{--param inline-unit-growth}.
11273 @item inline-unit-growth
11274 Specifies maximal overall growth of the compilation unit caused by inlining.
11275 For example, parameter value 20 limits unit growth to 1.2 times the original
11276 size. Cold functions (either marked cold via an attribute or by profile
11277 feedback) are not accounted into the unit size.
11279 @item ipcp-unit-growth
11280 Specifies maximal overall growth of the compilation unit caused by
11281 interprocedural constant propagation. For example, parameter value 10 limits
11282 unit growth to 1.1 times the original size.
11284 @item large-stack-frame
11285 The limit specifying large stack frames. While inlining the algorithm is trying
11286 to not grow past this limit too much.
11288 @item large-stack-frame-growth
11289 Specifies maximal growth of large stack frames caused by inlining in percents.
11290 For example, parameter value 1000 limits large stack frame growth to 11 times
11293 @item max-inline-insns-recursive
11294 @itemx max-inline-insns-recursive-auto
11295 Specifies the maximum number of instructions an out-of-line copy of a
11296 self-recursive inline
11297 function can grow into by performing recursive inlining.
11299 @option{--param max-inline-insns-recursive} applies to functions
11301 For functions not declared inline, recursive inlining
11302 happens only when @option{-finline-functions} (included in @option{-O3}) is
11303 enabled; @option{--param max-inline-insns-recursive-auto} applies instead.
11305 @item max-inline-recursive-depth
11306 @itemx max-inline-recursive-depth-auto
11307 Specifies the maximum recursion depth used for recursive inlining.
11309 @option{--param max-inline-recursive-depth} applies to functions
11310 declared inline. For functions not declared inline, recursive inlining
11311 happens only when @option{-finline-functions} (included in @option{-O3}) is
11312 enabled; @option{--param max-inline-recursive-depth-auto} applies instead.
11314 @item min-inline-recursive-probability
11315 Recursive inlining is profitable only for function having deep recursion
11316 in average and can hurt for function having little recursion depth by
11317 increasing the prologue size or complexity of function body to other
11320 When profile feedback is available (see @option{-fprofile-generate}) the actual
11321 recursion depth can be guessed from the probability that function recurses
11322 via a given call expression. This parameter limits inlining only to call
11323 expressions whose probability exceeds the given threshold (in percents).
11325 @item early-inlining-insns
11326 @item early-inlining-insns-O2
11327 Specify growth that the early inliner can make. In effect it increases
11328 the amount of inlining for code having a large abstraction penalty.
11330 For functions compiled with optimization levels
11331 @option{-O3} and @option{-Ofast} parameter @option{early-inlining-insns} is
11332 applied. In other cases @option{early-inlining-insns-O2} is applied.
11334 @item max-early-inliner-iterations
11335 Limit of iterations of the early inliner. This basically bounds
11336 the number of nested indirect calls the early inliner can resolve.
11337 Deeper chains are still handled by late inlining.
11339 @item comdat-sharing-probability
11340 Probability (in percent) that C++ inline function with comdat visibility
11341 are shared across multiple compilation units.
11343 @item profile-func-internal-id
11344 A parameter to control whether to use function internal id in profile
11345 database lookup. If the value is 0, the compiler uses an id that
11346 is based on function assembler name and filename, which makes old profile
11347 data more tolerant to source changes such as function reordering etc.
11349 @item min-vect-loop-bound
11350 The minimum number of iterations under which loops are not vectorized
11351 when @option{-ftree-vectorize} is used. The number of iterations after
11352 vectorization needs to be greater than the value specified by this option
11353 to allow vectorization.
11355 @item gcse-cost-distance-ratio
11356 Scaling factor in calculation of maximum distance an expression
11357 can be moved by GCSE optimizations. This is currently supported only in the
11358 code hoisting pass. The bigger the ratio, the more aggressive code hoisting
11359 is with simple expressions, i.e., the expressions that have cost
11360 less than @option{gcse-unrestricted-cost}. Specifying 0 disables
11361 hoisting of simple expressions.
11363 @item gcse-unrestricted-cost
11364 Cost, roughly measured as the cost of a single typical machine
11365 instruction, at which GCSE optimizations do not constrain
11366 the distance an expression can travel. This is currently
11367 supported only in the code hoisting pass. The lesser the cost,
11368 the more aggressive code hoisting is. Specifying 0
11369 allows all expressions to travel unrestricted distances.
11371 @item max-hoist-depth
11372 The depth of search in the dominator tree for expressions to hoist.
11373 This is used to avoid quadratic behavior in hoisting algorithm.
11374 The value of 0 does not limit on the search, but may slow down compilation
11377 @item max-tail-merge-comparisons
11378 The maximum amount of similar bbs to compare a bb with. This is used to
11379 avoid quadratic behavior in tree tail merging.
11381 @item max-tail-merge-iterations
11382 The maximum amount of iterations of the pass over the function. This is used to
11383 limit compilation time in tree tail merging.
11385 @item store-merging-allow-unaligned
11386 Allow the store merging pass to introduce unaligned stores if it is legal to
11389 @item max-stores-to-merge
11390 The maximum number of stores to attempt to merge into wider stores in the store
11393 @item max-unrolled-insns
11394 The maximum number of instructions that a loop may have to be unrolled.
11395 If a loop is unrolled, this parameter also determines how many times
11396 the loop code is unrolled.
11398 @item max-average-unrolled-insns
11399 The maximum number of instructions biased by probabilities of their execution
11400 that a loop may have to be unrolled. If a loop is unrolled,
11401 this parameter also determines how many times the loop code is unrolled.
11403 @item max-unroll-times
11404 The maximum number of unrollings of a single loop.
11406 @item max-peeled-insns
11407 The maximum number of instructions that a loop may have to be peeled.
11408 If a loop is peeled, this parameter also determines how many times
11409 the loop code is peeled.
11411 @item max-peel-times
11412 The maximum number of peelings of a single loop.
11414 @item max-peel-branches
11415 The maximum number of branches on the hot path through the peeled sequence.
11417 @item max-completely-peeled-insns
11418 The maximum number of insns of a completely peeled loop.
11420 @item max-completely-peel-times
11421 The maximum number of iterations of a loop to be suitable for complete peeling.
11423 @item max-completely-peel-loop-nest-depth
11424 The maximum depth of a loop nest suitable for complete peeling.
11426 @item max-unswitch-insns
11427 The maximum number of insns of an unswitched loop.
11429 @item max-unswitch-level
11430 The maximum number of branches unswitched in a single loop.
11432 @item lim-expensive
11433 The minimum cost of an expensive expression in the loop invariant motion.
11435 @item iv-consider-all-candidates-bound
11436 Bound on number of candidates for induction variables, below which
11437 all candidates are considered for each use in induction variable
11438 optimizations. If there are more candidates than this,
11439 only the most relevant ones are considered to avoid quadratic time complexity.
11441 @item iv-max-considered-uses
11442 The induction variable optimizations give up on loops that contain more
11443 induction variable uses.
11445 @item iv-always-prune-cand-set-bound
11446 If the number of candidates in the set is smaller than this value,
11447 always try to remove unnecessary ivs from the set
11448 when adding a new one.
11450 @item avg-loop-niter
11451 Average number of iterations of a loop.
11453 @item dse-max-object-size
11454 Maximum size (in bytes) of objects tracked bytewise by dead store elimination.
11455 Larger values may result in larger compilation times.
11457 @item dse-max-alias-queries-per-store
11458 Maximum number of queries into the alias oracle per store.
11459 Larger values result in larger compilation times and may result in more
11460 removed dead stores.
11462 @item scev-max-expr-size
11463 Bound on size of expressions used in the scalar evolutions analyzer.
11464 Large expressions slow the analyzer.
11466 @item scev-max-expr-complexity
11467 Bound on the complexity of the expressions in the scalar evolutions analyzer.
11468 Complex expressions slow the analyzer.
11470 @item max-tree-if-conversion-phi-args
11471 Maximum number of arguments in a PHI supported by TREE if conversion
11472 unless the loop is marked with simd pragma.
11474 @item vect-max-version-for-alignment-checks
11475 The maximum number of run-time checks that can be performed when
11476 doing loop versioning for alignment in the vectorizer.
11478 @item vect-max-version-for-alias-checks
11479 The maximum number of run-time checks that can be performed when
11480 doing loop versioning for alias in the vectorizer.
11482 @item vect-max-peeling-for-alignment
11483 The maximum number of loop peels to enhance access alignment
11484 for vectorizer. Value -1 means no limit.
11486 @item max-iterations-to-track
11487 The maximum number of iterations of a loop the brute-force algorithm
11488 for analysis of the number of iterations of the loop tries to evaluate.
11490 @item hot-bb-count-fraction
11491 The denominator n of fraction 1/n of the maximal execution count of a
11492 basic block in the entire program that a basic block needs to at least
11493 have in order to be considered hot. The default is 10000, which means
11494 that a basic block is considered hot if its execution count is greater
11495 than 1/10000 of the maximal execution count. 0 means that it is never
11496 considered hot. Used in non-LTO mode.
11498 @item hot-bb-count-ws-permille
11499 The number of most executed permilles, ranging from 0 to 1000, of the
11500 profiled execution of the entire program to which the execution count
11501 of a basic block must be part of in order to be considered hot. The
11502 default is 990, which means that a basic block is considered hot if
11503 its execution count contributes to the upper 990 permilles, or 99.0%,
11504 of the profiled execution of the entire program. 0 means that it is
11505 never considered hot. Used in LTO mode.
11507 @item hot-bb-frequency-fraction
11508 The denominator n of fraction 1/n of the execution frequency of the
11509 entry block of a function that a basic block of this function needs
11510 to at least have in order to be considered hot. The default is 1000,
11511 which means that a basic block is considered hot in a function if it
11512 is executed more frequently than 1/1000 of the frequency of the entry
11513 block of the function. 0 means that it is never considered hot.
11515 @item unlikely-bb-count-fraction
11516 The denominator n of fraction 1/n of the number of profiled runs of
11517 the entire program below which the execution count of a basic block
11518 must be in order for the basic block to be considered unlikely executed.
11519 The default is 20, which means that a basic block is considered unlikely
11520 executed if it is executed in fewer than 1/20, or 5%, of the runs of
11521 the program. 0 means that it is always considered unlikely executed.
11523 @item max-predicted-iterations
11524 The maximum number of loop iterations we predict statically. This is useful
11525 in cases where a function contains a single loop with known bound and
11526 another loop with unknown bound.
11527 The known number of iterations is predicted correctly, while
11528 the unknown number of iterations average to roughly 10. This means that the
11529 loop without bounds appears artificially cold relative to the other one.
11531 @item builtin-expect-probability
11532 Control the probability of the expression having the specified value. This
11533 parameter takes a percentage (i.e.@: 0 ... 100) as input.
11535 @item builtin-string-cmp-inline-length
11536 The maximum length of a constant string for a builtin string cmp call
11537 eligible for inlining.
11539 @item align-threshold
11541 Select fraction of the maximal frequency of executions of a basic block in
11542 a function to align the basic block.
11544 @item align-loop-iterations
11546 A loop expected to iterate at least the selected number of iterations is
11549 @item tracer-dynamic-coverage
11550 @itemx tracer-dynamic-coverage-feedback
11552 This value is used to limit superblock formation once the given percentage of
11553 executed instructions is covered. This limits unnecessary code size
11556 The @option{tracer-dynamic-coverage-feedback} parameter
11557 is used only when profile
11558 feedback is available. The real profiles (as opposed to statically estimated
11559 ones) are much less balanced allowing the threshold to be larger value.
11561 @item tracer-max-code-growth
11562 Stop tail duplication once code growth has reached given percentage. This is
11563 a rather artificial limit, as most of the duplicates are eliminated later in
11564 cross jumping, so it may be set to much higher values than is the desired code
11567 @item tracer-min-branch-ratio
11569 Stop reverse growth when the reverse probability of best edge is less than this
11570 threshold (in percent).
11572 @item tracer-min-branch-probability
11573 @itemx tracer-min-branch-probability-feedback
11575 Stop forward growth if the best edge has probability lower than this
11578 Similarly to @option{tracer-dynamic-coverage} two parameters are
11579 provided. @option{tracer-min-branch-probability-feedback} is used for
11580 compilation with profile feedback and @option{tracer-min-branch-probability}
11581 compilation without. The value for compilation with profile feedback
11582 needs to be more conservative (higher) in order to make tracer
11585 @item stack-clash-protection-guard-size
11586 Specify the size of the operating system provided stack guard as
11587 2 raised to @var{num} bytes. Higher values may reduce the
11588 number of explicit probes, but a value larger than the operating system
11589 provided guard will leave code vulnerable to stack clash style attacks.
11591 @item stack-clash-protection-probe-interval
11592 Stack clash protection involves probing stack space as it is allocated. This
11593 param controls the maximum distance between probes into the stack as 2 raised
11594 to @var{num} bytes. Higher values may reduce the number of explicit probes, but a value
11595 larger than the operating system provided guard will leave code vulnerable to
11596 stack clash style attacks.
11598 @item max-cse-path-length
11600 The maximum number of basic blocks on path that CSE considers.
11602 @item max-cse-insns
11603 The maximum number of instructions CSE processes before flushing.
11605 @item ggc-min-expand
11607 GCC uses a garbage collector to manage its own memory allocation. This
11608 parameter specifies the minimum percentage by which the garbage
11609 collector's heap should be allowed to expand between collections.
11610 Tuning this may improve compilation speed; it has no effect on code
11613 The default is 30% + 70% * (RAM/1GB) with an upper bound of 100% when
11614 RAM >= 1GB@. If @code{getrlimit} is available, the notion of ``RAM'' is
11615 the smallest of actual RAM and @code{RLIMIT_DATA} or @code{RLIMIT_AS}. If
11616 GCC is not able to calculate RAM on a particular platform, the lower
11617 bound of 30% is used. Setting this parameter and
11618 @option{ggc-min-heapsize} to zero causes a full collection to occur at
11619 every opportunity. This is extremely slow, but can be useful for
11622 @item ggc-min-heapsize
11624 Minimum size of the garbage collector's heap before it begins bothering
11625 to collect garbage. The first collection occurs after the heap expands
11626 by @option{ggc-min-expand}% beyond @option{ggc-min-heapsize}. Again,
11627 tuning this may improve compilation speed, and has no effect on code
11630 The default is the smaller of RAM/8, RLIMIT_RSS, or a limit that
11631 tries to ensure that RLIMIT_DATA or RLIMIT_AS are not exceeded, but
11632 with a lower bound of 4096 (four megabytes) and an upper bound of
11633 131072 (128 megabytes). If GCC is not able to calculate RAM on a
11634 particular platform, the lower bound is used. Setting this parameter
11635 very large effectively disables garbage collection. Setting this
11636 parameter and @option{ggc-min-expand} to zero causes a full collection
11637 to occur at every opportunity.
11639 @item max-reload-search-insns
11640 The maximum number of instruction reload should look backward for equivalent
11641 register. Increasing values mean more aggressive optimization, making the
11642 compilation time increase with probably slightly better performance.
11644 @item max-cselib-memory-locations
11645 The maximum number of memory locations cselib should take into account.
11646 Increasing values mean more aggressive optimization, making the compilation time
11647 increase with probably slightly better performance.
11649 @item max-sched-ready-insns
11650 The maximum number of instructions ready to be issued the scheduler should
11651 consider at any given time during the first scheduling pass. Increasing
11652 values mean more thorough searches, making the compilation time increase
11653 with probably little benefit.
11655 @item max-sched-region-blocks
11656 The maximum number of blocks in a region to be considered for
11657 interblock scheduling.
11659 @item max-pipeline-region-blocks
11660 The maximum number of blocks in a region to be considered for
11661 pipelining in the selective scheduler.
11663 @item max-sched-region-insns
11664 The maximum number of insns in a region to be considered for
11665 interblock scheduling.
11667 @item max-pipeline-region-insns
11668 The maximum number of insns in a region to be considered for
11669 pipelining in the selective scheduler.
11671 @item min-spec-prob
11672 The minimum probability (in percents) of reaching a source block
11673 for interblock speculative scheduling.
11675 @item max-sched-extend-regions-iters
11676 The maximum number of iterations through CFG to extend regions.
11677 A value of 0 disables region extensions.
11679 @item max-sched-insn-conflict-delay
11680 The maximum conflict delay for an insn to be considered for speculative motion.
11682 @item sched-spec-prob-cutoff
11683 The minimal probability of speculation success (in percents), so that
11684 speculative insns are scheduled.
11686 @item sched-state-edge-prob-cutoff
11687 The minimum probability an edge must have for the scheduler to save its
11690 @item sched-mem-true-dep-cost
11691 Minimal distance (in CPU cycles) between store and load targeting same
11694 @item selsched-max-lookahead
11695 The maximum size of the lookahead window of selective scheduling. It is a
11696 depth of search for available instructions.
11698 @item selsched-max-sched-times
11699 The maximum number of times that an instruction is scheduled during
11700 selective scheduling. This is the limit on the number of iterations
11701 through which the instruction may be pipelined.
11703 @item selsched-insns-to-rename
11704 The maximum number of best instructions in the ready list that are considered
11705 for renaming in the selective scheduler.
11708 The minimum value of stage count that swing modulo scheduler
11711 @item max-last-value-rtl
11712 The maximum size measured as number of RTLs that can be recorded in an expression
11713 in combiner for a pseudo register as last known value of that register.
11715 @item max-combine-insns
11716 The maximum number of instructions the RTL combiner tries to combine.
11718 @item integer-share-limit
11719 Small integer constants can use a shared data structure, reducing the
11720 compiler's memory usage and increasing its speed. This sets the maximum
11721 value of a shared integer constant.
11723 @item ssp-buffer-size
11724 The minimum size of buffers (i.e.@: arrays) that receive stack smashing
11725 protection when @option{-fstack-protection} is used.
11727 @item min-size-for-stack-sharing
11728 The minimum size of variables taking part in stack slot sharing when not
11731 @item max-jump-thread-duplication-stmts
11732 Maximum number of statements allowed in a block that needs to be
11733 duplicated when threading jumps.
11735 @item max-fields-for-field-sensitive
11736 Maximum number of fields in a structure treated in
11737 a field sensitive manner during pointer analysis.
11739 @item prefetch-latency
11740 Estimate on average number of instructions that are executed before
11741 prefetch finishes. The distance prefetched ahead is proportional
11742 to this constant. Increasing this number may also lead to less
11743 streams being prefetched (see @option{simultaneous-prefetches}).
11745 @item simultaneous-prefetches
11746 Maximum number of prefetches that can run at the same time.
11748 @item l1-cache-line-size
11749 The size of cache line in L1 data cache, in bytes.
11751 @item l1-cache-size
11752 The size of L1 data cache, in kilobytes.
11754 @item l2-cache-size
11755 The size of L2 data cache, in kilobytes.
11757 @item prefetch-dynamic-strides
11758 Whether the loop array prefetch pass should issue software prefetch hints
11759 for strides that are non-constant. In some cases this may be
11760 beneficial, though the fact the stride is non-constant may make it
11761 hard to predict when there is clear benefit to issuing these hints.
11763 Set to 1 if the prefetch hints should be issued for non-constant
11764 strides. Set to 0 if prefetch hints should be issued only for strides that
11765 are known to be constant and below @option{prefetch-minimum-stride}.
11767 @item prefetch-minimum-stride
11768 Minimum constant stride, in bytes, to start using prefetch hints for. If
11769 the stride is less than this threshold, prefetch hints will not be issued.
11771 This setting is useful for processors that have hardware prefetchers, in
11772 which case there may be conflicts between the hardware prefetchers and
11773 the software prefetchers. If the hardware prefetchers have a maximum
11774 stride they can handle, it should be used here to improve the use of
11775 software prefetchers.
11777 A value of -1 means we don't have a threshold and therefore
11778 prefetch hints can be issued for any constant stride.
11780 This setting is only useful for strides that are known and constant.
11782 @item loop-interchange-max-num-stmts
11783 The maximum number of stmts in a loop to be interchanged.
11785 @item loop-interchange-stride-ratio
11786 The minimum ratio between stride of two loops for interchange to be profitable.
11788 @item min-insn-to-prefetch-ratio
11789 The minimum ratio between the number of instructions and the
11790 number of prefetches to enable prefetching in a loop.
11792 @item prefetch-min-insn-to-mem-ratio
11793 The minimum ratio between the number of instructions and the
11794 number of memory references to enable prefetching in a loop.
11796 @item use-canonical-types
11797 Whether the compiler should use the ``canonical'' type system.
11798 Should always be 1, which uses a more efficient internal
11799 mechanism for comparing types in C++ and Objective-C++. However, if
11800 bugs in the canonical type system are causing compilation failures,
11801 set this value to 0 to disable canonical types.
11803 @item switch-conversion-max-branch-ratio
11804 Switch initialization conversion refuses to create arrays that are
11805 bigger than @option{switch-conversion-max-branch-ratio} times the number of
11806 branches in the switch.
11808 @item max-partial-antic-length
11809 Maximum length of the partial antic set computed during the tree
11810 partial redundancy elimination optimization (@option{-ftree-pre}) when
11811 optimizing at @option{-O3} and above. For some sorts of source code
11812 the enhanced partial redundancy elimination optimization can run away,
11813 consuming all of the memory available on the host machine. This
11814 parameter sets a limit on the length of the sets that are computed,
11815 which prevents the runaway behavior. Setting a value of 0 for
11816 this parameter allows an unlimited set length.
11818 @item rpo-vn-max-loop-depth
11819 Maximum loop depth that is value-numbered optimistically.
11820 When the limit hits the innermost
11821 @var{rpo-vn-max-loop-depth} loops and the outermost loop in the
11822 loop nest are value-numbered optimistically and the remaining ones not.
11824 @item sccvn-max-alias-queries-per-access
11825 Maximum number of alias-oracle queries we perform when looking for
11826 redundancies for loads and stores. If this limit is hit the search
11827 is aborted and the load or store is not considered redundant. The
11828 number of queries is algorithmically limited to the number of
11829 stores on all paths from the load to the function entry.
11831 @item ira-max-loops-num
11832 IRA uses regional register allocation by default. If a function
11833 contains more loops than the number given by this parameter, only at most
11834 the given number of the most frequently-executed loops form regions
11835 for regional register allocation.
11837 @item ira-max-conflict-table-size
11838 Although IRA uses a sophisticated algorithm to compress the conflict
11839 table, the table can still require excessive amounts of memory for
11840 huge functions. If the conflict table for a function could be more
11841 than the size in MB given by this parameter, the register allocator
11842 instead uses a faster, simpler, and lower-quality
11843 algorithm that does not require building a pseudo-register conflict table.
11845 @item ira-loop-reserved-regs
11846 IRA can be used to evaluate more accurate register pressure in loops
11847 for decisions to move loop invariants (see @option{-O3}). The number
11848 of available registers reserved for some other purposes is given
11849 by this parameter. Default of the parameter
11850 is the best found from numerous experiments.
11852 @item lra-inheritance-ebb-probability-cutoff
11853 LRA tries to reuse values reloaded in registers in subsequent insns.
11854 This optimization is called inheritance. EBB is used as a region to
11855 do this optimization. The parameter defines a minimal fall-through
11856 edge probability in percentage used to add BB to inheritance EBB in
11857 LRA. The default value was chosen
11858 from numerous runs of SPEC2000 on x86-64.
11860 @item loop-invariant-max-bbs-in-loop
11861 Loop invariant motion can be very expensive, both in compilation time and
11862 in amount of needed compile-time memory, with very large loops. Loops
11863 with more basic blocks than this parameter won't have loop invariant
11864 motion optimization performed on them.
11866 @item loop-max-datarefs-for-datadeps
11867 Building data dependencies is expensive for very large loops. This
11868 parameter limits the number of data references in loops that are
11869 considered for data dependence analysis. These large loops are no
11870 handled by the optimizations using loop data dependencies.
11872 @item max-vartrack-size
11873 Sets a maximum number of hash table slots to use during variable
11874 tracking dataflow analysis of any function. If this limit is exceeded
11875 with variable tracking at assignments enabled, analysis for that
11876 function is retried without it, after removing all debug insns from
11877 the function. If the limit is exceeded even without debug insns, var
11878 tracking analysis is completely disabled for the function. Setting
11879 the parameter to zero makes it unlimited.
11881 @item max-vartrack-expr-depth
11882 Sets a maximum number of recursion levels when attempting to map
11883 variable names or debug temporaries to value expressions. This trades
11884 compilation time for more complete debug information. If this is set too
11885 low, value expressions that are available and could be represented in
11886 debug information may end up not being used; setting this higher may
11887 enable the compiler to find more complex debug expressions, but compile
11888 time and memory use may grow.
11890 @item max-debug-marker-count
11891 Sets a threshold on the number of debug markers (e.g.@: begin stmt
11892 markers) to avoid complexity explosion at inlining or expanding to RTL.
11893 If a function has more such gimple stmts than the set limit, such stmts
11894 will be dropped from the inlined copy of a function, and from its RTL
11897 @item min-nondebug-insn-uid
11898 Use uids starting at this parameter for nondebug insns. The range below
11899 the parameter is reserved exclusively for debug insns created by
11900 @option{-fvar-tracking-assignments}, but debug insns may get
11901 (non-overlapping) uids above it if the reserved range is exhausted.
11903 @item ipa-sra-ptr-growth-factor
11904 IPA-SRA replaces a pointer to an aggregate with one or more new
11905 parameters only when their cumulative size is less or equal to
11906 @option{ipa-sra-ptr-growth-factor} times the size of the original
11909 @item ipa-sra-max-replacements
11910 Maximum pieces of an aggregate that IPA-SRA tracks. As a
11911 consequence, it is also the maximum number of replacements of a formal
11914 @item sra-max-scalarization-size-Ospeed
11915 @itemx sra-max-scalarization-size-Osize
11916 The two Scalar Reduction of Aggregates passes (SRA and IPA-SRA) aim to
11917 replace scalar parts of aggregates with uses of independent scalar
11918 variables. These parameters control the maximum size, in storage units,
11919 of aggregate which is considered for replacement when compiling for
11921 (@option{sra-max-scalarization-size-Ospeed}) or size
11922 (@option{sra-max-scalarization-size-Osize}) respectively.
11924 @item tm-max-aggregate-size
11925 When making copies of thread-local variables in a transaction, this
11926 parameter specifies the size in bytes after which variables are
11927 saved with the logging functions as opposed to save/restore code
11928 sequence pairs. This option only applies when using
11931 @item graphite-max-nb-scop-params
11932 To avoid exponential effects in the Graphite loop transforms, the
11933 number of parameters in a Static Control Part (SCoP) is bounded.
11934 A value of zero can be used to lift
11935 the bound. A variable whose value is unknown at compilation time and
11936 defined outside a SCoP is a parameter of the SCoP.
11938 @item loop-block-tile-size
11939 Loop blocking or strip mining transforms, enabled with
11940 @option{-floop-block} or @option{-floop-strip-mine}, strip mine each
11941 loop in the loop nest by a given number of iterations. The strip
11942 length can be changed using the @option{loop-block-tile-size}
11945 @item ipa-cp-value-list-size
11946 IPA-CP attempts to track all possible values and types passed to a function's
11947 parameter in order to propagate them and perform devirtualization.
11948 @option{ipa-cp-value-list-size} is the maximum number of values and types it
11949 stores per one formal parameter of a function.
11951 @item ipa-cp-eval-threshold
11952 IPA-CP calculates its own score of cloning profitability heuristics
11953 and performs those cloning opportunities with scores that exceed
11954 @option{ipa-cp-eval-threshold}.
11956 @item ipa-cp-recursion-penalty
11957 Percentage penalty the recursive functions will receive when they
11958 are evaluated for cloning.
11960 @item ipa-cp-single-call-penalty
11961 Percentage penalty functions containing a single call to another
11962 function will receive when they are evaluated for cloning.
11964 @item ipa-max-agg-items
11965 IPA-CP is also capable to propagate a number of scalar values passed
11966 in an aggregate. @option{ipa-max-agg-items} controls the maximum
11967 number of such values per one parameter.
11969 @item ipa-cp-loop-hint-bonus
11970 When IPA-CP determines that a cloning candidate would make the number
11971 of iterations of a loop known, it adds a bonus of
11972 @option{ipa-cp-loop-hint-bonus} to the profitability score of
11975 @item ipa-max-aa-steps
11976 During its analysis of function bodies, IPA-CP employs alias analysis
11977 in order to track values pointed to by function parameters. In order
11978 not spend too much time analyzing huge functions, it gives up and
11979 consider all memory clobbered after examining
11980 @option{ipa-max-aa-steps} statements modifying memory.
11982 @item ipa-max-switch-predicate-bounds
11983 Maximal number of boundary endpoints of case ranges of switch statement.
11984 For switch exceeding this limit, IPA-CP will not construct cloning cost
11985 predicate, which is used to estimate cloning benefit, for default case
11986 of the switch statement.
11988 @item lto-partitions
11989 Specify desired number of partitions produced during WHOPR compilation.
11990 The number of partitions should exceed the number of CPUs used for compilation.
11992 @item lto-min-partition
11993 Size of minimal partition for WHOPR (in estimated instructions).
11994 This prevents expenses of splitting very small programs into too many
11997 @item lto-max-partition
11998 Size of max partition for WHOPR (in estimated instructions).
11999 to provide an upper bound for individual size of partition.
12000 Meant to be used only with balanced partitioning.
12002 @item lto-max-streaming-parallelism
12003 Maximal number of parallel processes used for LTO streaming.
12005 @item cxx-max-namespaces-for-diagnostic-help
12006 The maximum number of namespaces to consult for suggestions when C++
12007 name lookup fails for an identifier.
12009 @item sink-frequency-threshold
12010 The maximum relative execution frequency (in percents) of the target block
12011 relative to a statement's original block to allow statement sinking of a
12012 statement. Larger numbers result in more aggressive statement sinking.
12013 A small positive adjustment is applied for
12014 statements with memory operands as those are even more profitable so sink.
12016 @item max-stores-to-sink
12017 The maximum number of conditional store pairs that can be sunk. Set to 0
12018 if either vectorization (@option{-ftree-vectorize}) or if-conversion
12019 (@option{-ftree-loop-if-convert}) is disabled.
12021 @item allow-store-data-races
12022 Allow optimizers to introduce new data races on stores.
12023 Set to 1 to allow, otherwise to 0.
12025 @item case-values-threshold
12026 The smallest number of different values for which it is best to use a
12027 jump-table instead of a tree of conditional branches. If the value is
12028 0, use the default for the machine.
12030 @item jump-table-max-growth-ratio-for-size
12031 The maximum code size growth ratio when expanding
12032 into a jump table (in percent). The parameter is used when
12033 optimizing for size.
12035 @item jump-table-max-growth-ratio-for-speed
12036 The maximum code size growth ratio when expanding
12037 into a jump table (in percent). The parameter is used when
12038 optimizing for speed.
12040 @item tree-reassoc-width
12041 Set the maximum number of instructions executed in parallel in
12042 reassociated tree. This parameter overrides target dependent
12043 heuristics used by default if has non zero value.
12045 @item sched-pressure-algorithm
12046 Choose between the two available implementations of
12047 @option{-fsched-pressure}. Algorithm 1 is the original implementation
12048 and is the more likely to prevent instructions from being reordered.
12049 Algorithm 2 was designed to be a compromise between the relatively
12050 conservative approach taken by algorithm 1 and the rather aggressive
12051 approach taken by the default scheduler. It relies more heavily on
12052 having a regular register file and accurate register pressure classes.
12053 See @file{haifa-sched.c} in the GCC sources for more details.
12055 The default choice depends on the target.
12057 @item max-slsr-cand-scan
12058 Set the maximum number of existing candidates that are considered when
12059 seeking a basis for a new straight-line strength reduction candidate.
12062 Enable buffer overflow detection for global objects. This kind
12063 of protection is enabled by default if you are using
12064 @option{-fsanitize=address} option.
12065 To disable global objects protection use @option{--param asan-globals=0}.
12068 Enable buffer overflow detection for stack objects. This kind of
12069 protection is enabled by default when using @option{-fsanitize=address}.
12070 To disable stack protection use @option{--param asan-stack=0} option.
12072 @item asan-instrument-reads
12073 Enable buffer overflow detection for memory reads. This kind of
12074 protection is enabled by default when using @option{-fsanitize=address}.
12075 To disable memory reads protection use
12076 @option{--param asan-instrument-reads=0}.
12078 @item asan-instrument-writes
12079 Enable buffer overflow detection for memory writes. This kind of
12080 protection is enabled by default when using @option{-fsanitize=address}.
12081 To disable memory writes protection use
12082 @option{--param asan-instrument-writes=0} option.
12084 @item asan-memintrin
12085 Enable detection for built-in functions. This kind of protection
12086 is enabled by default when using @option{-fsanitize=address}.
12087 To disable built-in functions protection use
12088 @option{--param asan-memintrin=0}.
12090 @item asan-use-after-return
12091 Enable detection of use-after-return. This kind of protection
12092 is enabled by default when using the @option{-fsanitize=address} option.
12093 To disable it use @option{--param asan-use-after-return=0}.
12095 Note: By default the check is disabled at run time. To enable it,
12096 add @code{detect_stack_use_after_return=1} to the environment variable
12097 @env{ASAN_OPTIONS}.
12099 @item asan-instrumentation-with-call-threshold
12100 If number of memory accesses in function being instrumented
12101 is greater or equal to this number, use callbacks instead of inline checks.
12102 E.g. to disable inline code use
12103 @option{--param asan-instrumentation-with-call-threshold=0}.
12105 @item use-after-scope-direct-emission-threshold
12106 If the size of a local variable in bytes is smaller or equal to this
12107 number, directly poison (or unpoison) shadow memory instead of using
12108 run-time callbacks.
12110 @item max-fsm-thread-path-insns
12111 Maximum number of instructions to copy when duplicating blocks on a
12112 finite state automaton jump thread path.
12114 @item max-fsm-thread-length
12115 Maximum number of basic blocks on a finite state automaton jump thread
12118 @item max-fsm-thread-paths
12119 Maximum number of new jump thread paths to create for a finite state
12122 @item parloops-chunk-size
12123 Chunk size of omp schedule for loops parallelized by parloops.
12125 @item parloops-schedule
12126 Schedule type of omp schedule for loops parallelized by parloops (static,
12127 dynamic, guided, auto, runtime).
12129 @item parloops-min-per-thread
12130 The minimum number of iterations per thread of an innermost parallelized
12131 loop for which the parallelized variant is preferred over the single threaded
12132 one. Note that for a parallelized loop nest the
12133 minimum number of iterations of the outermost loop per thread is two.
12135 @item max-ssa-name-query-depth
12136 Maximum depth of recursion when querying properties of SSA names in things
12137 like fold routines. One level of recursion corresponds to following a
12140 @item hsa-gen-debug-stores
12141 Enable emission of special debug stores within HSA kernels which are
12142 then read and reported by libgomp plugin. Generation of these stores
12143 is disabled by default, use @option{--param hsa-gen-debug-stores=1} to
12146 @item max-speculative-devirt-maydefs
12147 The maximum number of may-defs we analyze when looking for a must-def
12148 specifying the dynamic type of an object that invokes a virtual call
12149 we may be able to devirtualize speculatively.
12151 @item max-vrp-switch-assertions
12152 The maximum number of assertions to add along the default edge of a switch
12153 statement during VRP.
12155 @item unroll-jam-min-percent
12156 The minimum percentage of memory references that must be optimized
12157 away for the unroll-and-jam transformation to be considered profitable.
12159 @item unroll-jam-max-unroll
12160 The maximum number of times the outer loop should be unrolled by
12161 the unroll-and-jam transformation.
12163 @item max-rtl-if-conversion-unpredictable-cost
12164 Maximum permissible cost for the sequence that would be generated
12165 by the RTL if-conversion pass for a branch that is considered unpredictable.
12167 @item max-variable-expansions-in-unroller
12168 If @option{-fvariable-expansion-in-unroller} is used, the maximum number
12169 of times that an individual variable will be expanded during loop unrolling.
12171 @item tracer-min-branch-probability-feedback
12172 Stop forward growth if the probability of best edge is less than
12173 this threshold (in percent). Used when profile feedback is available.
12175 @item partial-inlining-entry-probability
12176 Maximum probability of the entry BB of split region
12177 (in percent relative to entry BB of the function)
12178 to make partial inlining happen.
12180 @item max-tracked-strlens
12181 Maximum number of strings for which strlen optimization pass will
12182 track string lengths.
12184 @item gcse-after-reload-partial-fraction
12185 The threshold ratio for performing partial redundancy
12186 elimination after reload.
12188 @item gcse-after-reload-critical-fraction
12189 The threshold ratio of critical edges execution count that
12190 permit performing redundancy elimination after reload.
12192 @item max-loop-header-insns
12193 The maximum number of insns in loop header duplicated
12194 by the copy loop headers pass.
12196 @item vect-epilogues-nomask
12197 Enable loop epilogue vectorization using smaller vector size.
12199 @item slp-max-insns-in-bb
12200 Maximum number of instructions in basic block to be
12201 considered for SLP vectorization.
12203 @item avoid-fma-max-bits
12204 Maximum number of bits for which we avoid creating FMAs.
12206 @item sms-loop-average-count-threshold
12207 A threshold on the average loop count considered by the swing modulo scheduler.
12209 @item sms-dfa-history
12210 The number of cycles the swing modulo scheduler considers when checking
12211 conflicts using DFA.
12213 @item max-inline-insns-recursive-auto
12214 The maximum number of instructions non-inline function
12215 can grow to via recursive inlining.
12217 @item graphite-allow-codegen-errors
12218 Whether codegen errors should be ICEs when @option{-fchecking}.
12220 @item sms-max-ii-factor
12221 A factor for tuning the upper bound that swing modulo scheduler
12222 uses for scheduling a loop.
12224 @item lra-max-considered-reload-pseudos
12225 The max number of reload pseudos which are considered during
12226 spilling a non-reload pseudo.
12228 @item max-pow-sqrt-depth
12229 Maximum depth of sqrt chains to use when synthesizing exponentiation
12230 by a real constant.
12232 @item max-dse-active-local-stores
12233 Maximum number of active local stores in RTL dead store elimination.
12235 @item asan-instrument-allocas
12236 Enable asan allocas/VLAs protection.
12238 @item max-iterations-computation-cost
12239 Bound on the cost of an expression to compute the number of iterations.
12241 @item max-isl-operations
12242 Maximum number of isl operations, 0 means unlimited.
12244 @item graphite-max-arrays-per-scop
12245 Maximum number of arrays per scop.
12247 @item max-vartrack-reverse-op-size
12248 Max. size of loc list for which reverse ops should be added.
12250 @item tracer-dynamic-coverage-feedback
12251 The percentage of function, weighted by execution frequency,
12252 that must be covered by trace formation.
12253 Used when profile feedback is available.
12255 @item max-inline-recursive-depth-auto
12256 The maximum depth of recursive inlining for non-inline functions.
12258 @item fsm-scale-path-stmts
12259 Scale factor to apply to the number of statements in a threading path
12260 when comparing to the number of (scaled) blocks.
12262 @item fsm-maximum-phi-arguments
12263 Maximum number of arguments a PHI may have before the FSM threader
12264 will not try to thread through its block.
12266 @item uninit-control-dep-attempts
12267 Maximum number of nested calls to search for control dependencies
12268 during uninitialized variable analysis.
12270 @item max-once-peeled-insns
12271 The maximum number of insns of a peeled loop that rolls only once.
12273 @item sra-max-scalarization-size-Osize
12274 Maximum size, in storage units, of an aggregate
12275 which should be considered for scalarization when compiling for size.
12277 @item fsm-scale-path-blocks
12278 Scale factor to apply to the number of blocks in a threading path
12279 when comparing to the number of (scaled) statements.
12281 @item sched-autopref-queue-depth
12282 Hardware autoprefetcher scheduler model control flag.
12283 Number of lookahead cycles the model looks into; at '
12284 ' only enable instruction sorting heuristic.
12286 @item loop-versioning-max-inner-insns
12287 The maximum number of instructions that an inner loop can have
12288 before the loop versioning pass considers it too big to copy.
12290 @item loop-versioning-max-outer-insns
12291 The maximum number of instructions that an outer loop can have
12292 before the loop versioning pass considers it too big to copy,
12293 discounting any instructions in inner loops that directly benefit
12296 @item ssa-name-def-chain-limit
12297 The maximum number of SSA_NAME assignments to follow in determining
12298 a property of a variable such as its value. This limits the number
12299 of iterations or recursive calls GCC performs when optimizing certain
12300 statements or when determining their validity prior to issuing
12306 @node Instrumentation Options
12307 @section Program Instrumentation Options
12308 @cindex instrumentation options
12309 @cindex program instrumentation options
12310 @cindex run-time error checking options
12311 @cindex profiling options
12312 @cindex options, program instrumentation
12313 @cindex options, run-time error checking
12314 @cindex options, profiling
12316 GCC supports a number of command-line options that control adding
12317 run-time instrumentation to the code it normally generates.
12318 For example, one purpose of instrumentation is collect profiling
12319 statistics for use in finding program hot spots, code coverage
12320 analysis, or profile-guided optimizations.
12321 Another class of program instrumentation is adding run-time checking
12322 to detect programming errors like invalid pointer
12323 dereferences or out-of-bounds array accesses, as well as deliberately
12324 hostile attacks such as stack smashing or C++ vtable hijacking.
12325 There is also a general hook which can be used to implement other
12326 forms of tracing or function-level instrumentation for debug or
12327 program analysis purposes.
12330 @cindex @command{prof}
12331 @cindex @command{gprof}
12336 Generate extra code to write profile information suitable for the
12337 analysis program @command{prof} (for @option{-p}) or @command{gprof}
12338 (for @option{-pg}). You must use this option when compiling
12339 the source files you want data about, and you must also use it when
12342 You can use the function attribute @code{no_instrument_function} to
12343 suppress profiling of individual functions when compiling with these options.
12344 @xref{Common Function Attributes}.
12346 @item -fprofile-arcs
12347 @opindex fprofile-arcs
12348 Add code so that program flow @dfn{arcs} are instrumented. During
12349 execution the program records how many times each branch and call is
12350 executed and how many times it is taken or returns. On targets that support
12351 constructors with priority support, profiling properly handles constructors,
12352 destructors and C++ constructors (and destructors) of classes which are used
12353 as a type of a global variable.
12356 program exits it saves this data to a file called
12357 @file{@var{auxname}.gcda} for each source file. The data may be used for
12358 profile-directed optimizations (@option{-fbranch-probabilities}), or for
12359 test coverage analysis (@option{-ftest-coverage}). Each object file's
12360 @var{auxname} is generated from the name of the output file, if
12361 explicitly specified and it is not the final executable, otherwise it is
12362 the basename of the source file. In both cases any suffix is removed
12363 (e.g.@: @file{foo.gcda} for input file @file{dir/foo.c}, or
12364 @file{dir/foo.gcda} for output file specified as @option{-o dir/foo.o}).
12365 @xref{Cross-profiling}.
12367 @cindex @command{gcov}
12371 This option is used to compile and link code instrumented for coverage
12372 analysis. The option is a synonym for @option{-fprofile-arcs}
12373 @option{-ftest-coverage} (when compiling) and @option{-lgcov} (when
12374 linking). See the documentation for those options for more details.
12379 Compile the source files with @option{-fprofile-arcs} plus optimization
12380 and code generation options. For test coverage analysis, use the
12381 additional @option{-ftest-coverage} option. You do not need to profile
12382 every source file in a program.
12385 Compile the source files additionally with @option{-fprofile-abs-path}
12386 to create absolute path names in the @file{.gcno} files. This allows
12387 @command{gcov} to find the correct sources in projects where compilations
12388 occur with different working directories.
12391 Link your object files with @option{-lgcov} or @option{-fprofile-arcs}
12392 (the latter implies the former).
12395 Run the program on a representative workload to generate the arc profile
12396 information. This may be repeated any number of times. You can run
12397 concurrent instances of your program, and provided that the file system
12398 supports locking, the data files will be correctly updated. Unless
12399 a strict ISO C dialect option is in effect, @code{fork} calls are
12400 detected and correctly handled without double counting.
12403 For profile-directed optimizations, compile the source files again with
12404 the same optimization and code generation options plus
12405 @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that
12406 Control Optimization}).
12409 For test coverage analysis, use @command{gcov} to produce human readable
12410 information from the @file{.gcno} and @file{.gcda} files. Refer to the
12411 @command{gcov} documentation for further information.
12415 With @option{-fprofile-arcs}, for each function of your program GCC
12416 creates a program flow graph, then finds a spanning tree for the graph.
12417 Only arcs that are not on the spanning tree have to be instrumented: the
12418 compiler adds code to count the number of times that these arcs are
12419 executed. When an arc is the only exit or only entrance to a block, the
12420 instrumentation code can be added to the block; otherwise, a new basic
12421 block must be created to hold the instrumentation code.
12424 @item -ftest-coverage
12425 @opindex ftest-coverage
12426 Produce a notes file that the @command{gcov} code-coverage utility
12427 (@pxref{Gcov,, @command{gcov}---a Test Coverage Program}) can use to
12428 show program coverage. Each source file's note file is called
12429 @file{@var{auxname}.gcno}. Refer to the @option{-fprofile-arcs} option
12430 above for a description of @var{auxname} and instructions on how to
12431 generate test coverage data. Coverage data matches the source files
12432 more closely if you do not optimize.
12434 @item -fprofile-abs-path
12435 @opindex fprofile-abs-path
12436 Automatically convert relative source file names to absolute path names
12437 in the @file{.gcno} files. This allows @command{gcov} to find the correct
12438 sources in projects where compilations occur with different working
12441 @item -fprofile-dir=@var{path}
12442 @opindex fprofile-dir
12444 Set the directory to search for the profile data files in to @var{path}.
12445 This option affects only the profile data generated by
12446 @option{-fprofile-generate}, @option{-ftest-coverage}, @option{-fprofile-arcs}
12447 and used by @option{-fprofile-use} and @option{-fbranch-probabilities}
12448 and its related options. Both absolute and relative paths can be used.
12449 By default, GCC uses the current directory as @var{path}, thus the
12450 profile data file appears in the same directory as the object file.
12451 In order to prevent the file name clashing, if the object file name is
12452 not an absolute path, we mangle the absolute path of the
12453 @file{@var{sourcename}.gcda} file and use it as the file name of a
12454 @file{.gcda} file. See similar option @option{-fprofile-note}.
12456 When an executable is run in a massive parallel environment, it is recommended
12457 to save profile to different folders. That can be done with variables
12458 in @var{path} that are exported during run-time:
12466 value of environment variable @var{VAR}
12470 @item -fprofile-generate
12471 @itemx -fprofile-generate=@var{path}
12472 @opindex fprofile-generate
12474 Enable options usually used for instrumenting application to produce
12475 profile useful for later recompilation with profile feedback based
12476 optimization. You must use @option{-fprofile-generate} both when
12477 compiling and when linking your program.
12479 The following options are enabled:
12480 @option{-fprofile-arcs}, @option{-fprofile-values},
12481 @option{-finline-functions}, and @option{-fipa-bit-cp}.
12483 If @var{path} is specified, GCC looks at the @var{path} to find
12484 the profile feedback data files. See @option{-fprofile-dir}.
12486 To optimize the program based on the collected profile information, use
12487 @option{-fprofile-use}. @xref{Optimize Options}, for more information.
12489 @item -fprofile-note=@var{path}
12490 @opindex fprofile-note
12492 If @var{path} is specified, GCC saves @file{.gcno} file into @var{path}
12493 location. If you combine the option with multiple source files,
12494 the @file{.gcno} file will be overwritten.
12496 @item -fprofile-update=@var{method}
12497 @opindex fprofile-update
12499 Alter the update method for an application instrumented for profile
12500 feedback based optimization. The @var{method} argument should be one of
12501 @samp{single}, @samp{atomic} or @samp{prefer-atomic}.
12502 The first one is useful for single-threaded applications,
12503 while the second one prevents profile corruption by emitting thread-safe code.
12505 @strong{Warning:} When an application does not properly join all threads
12506 (or creates an detached thread), a profile file can be still corrupted.
12508 Using @samp{prefer-atomic} would be transformed either to @samp{atomic},
12509 when supported by a target, or to @samp{single} otherwise. The GCC driver
12510 automatically selects @samp{prefer-atomic} when @option{-pthread}
12511 is present in the command line.
12513 @item -fprofile-filter-files=@var{regex}
12514 @opindex fprofile-filter-files
12516 Instrument only functions from files where names match
12517 any regular expression (separated by a semi-colon).
12519 For example, @option{-fprofile-filter-files=main.c;module.*.c} will instrument
12520 only @file{main.c} and all C files starting with 'module'.
12522 @item -fprofile-exclude-files=@var{regex}
12523 @opindex fprofile-exclude-files
12525 Instrument only functions from files where names do not match
12526 all the regular expressions (separated by a semi-colon).
12528 For example, @option{-fprofile-exclude-files=/usr/*} will prevent instrumentation
12529 of all files that are located in @file{/usr/} folder.
12531 @item -fsanitize=address
12532 @opindex fsanitize=address
12533 Enable AddressSanitizer, a fast memory error detector.
12534 Memory access instructions are instrumented to detect
12535 out-of-bounds and use-after-free bugs.
12536 The option enables @option{-fsanitize-address-use-after-scope}.
12537 See @uref{https://github.com/google/sanitizers/wiki/AddressSanitizer} for
12538 more details. The run-time behavior can be influenced using the
12539 @env{ASAN_OPTIONS} environment variable. When set to @code{help=1},
12540 the available options are shown at startup of the instrumented program. See
12541 @url{https://github.com/google/sanitizers/wiki/AddressSanitizerFlags#run-time-flags}
12542 for a list of supported options.
12543 The option cannot be combined with @option{-fsanitize=thread}.
12545 @item -fsanitize=kernel-address
12546 @opindex fsanitize=kernel-address
12547 Enable AddressSanitizer for Linux kernel.
12548 See @uref{https://github.com/google/kasan/wiki} for more details.
12550 @item -fsanitize=pointer-compare
12551 @opindex fsanitize=pointer-compare
12552 Instrument comparison operation (<, <=, >, >=) with pointer operands.
12553 The option must be combined with either @option{-fsanitize=kernel-address} or
12554 @option{-fsanitize=address}
12555 The option cannot be combined with @option{-fsanitize=thread}.
12556 Note: By default the check is disabled at run time. To enable it,
12557 add @code{detect_invalid_pointer_pairs=2} to the environment variable
12558 @env{ASAN_OPTIONS}. Using @code{detect_invalid_pointer_pairs=1} detects
12559 invalid operation only when both pointers are non-null.
12561 @item -fsanitize=pointer-subtract
12562 @opindex fsanitize=pointer-subtract
12563 Instrument subtraction with pointer operands.
12564 The option must be combined with either @option{-fsanitize=kernel-address} or
12565 @option{-fsanitize=address}
12566 The option cannot be combined with @option{-fsanitize=thread}.
12567 Note: By default the check is disabled at run time. To enable it,
12568 add @code{detect_invalid_pointer_pairs=2} to the environment variable
12569 @env{ASAN_OPTIONS}. Using @code{detect_invalid_pointer_pairs=1} detects
12570 invalid operation only when both pointers are non-null.
12572 @item -fsanitize=thread
12573 @opindex fsanitize=thread
12574 Enable ThreadSanitizer, a fast data race detector.
12575 Memory access instructions are instrumented to detect
12576 data race bugs. See @uref{https://github.com/google/sanitizers/wiki#threadsanitizer} for more
12577 details. The run-time behavior can be influenced using the @env{TSAN_OPTIONS}
12578 environment variable; see
12579 @url{https://github.com/google/sanitizers/wiki/ThreadSanitizerFlags} for a list of
12581 The option cannot be combined with @option{-fsanitize=address},
12582 @option{-fsanitize=leak}.
12584 Note that sanitized atomic builtins cannot throw exceptions when
12585 operating on invalid memory addresses with non-call exceptions
12586 (@option{-fnon-call-exceptions}).
12588 @item -fsanitize=leak
12589 @opindex fsanitize=leak
12590 Enable LeakSanitizer, a memory leak detector.
12591 This option only matters for linking of executables and
12592 the executable is linked against a library that overrides @code{malloc}
12593 and other allocator functions. See
12594 @uref{https://github.com/google/sanitizers/wiki/AddressSanitizerLeakSanitizer} for more
12595 details. The run-time behavior can be influenced using the
12596 @env{LSAN_OPTIONS} environment variable.
12597 The option cannot be combined with @option{-fsanitize=thread}.
12599 @item -fsanitize=undefined
12600 @opindex fsanitize=undefined
12601 Enable UndefinedBehaviorSanitizer, a fast undefined behavior detector.
12602 Various computations are instrumented to detect undefined behavior
12603 at runtime. Current suboptions are:
12607 @item -fsanitize=shift
12608 @opindex fsanitize=shift
12609 This option enables checking that the result of a shift operation is
12610 not undefined. Note that what exactly is considered undefined differs
12611 slightly between C and C++, as well as between ISO C90 and C99, etc.
12612 This option has two suboptions, @option{-fsanitize=shift-base} and
12613 @option{-fsanitize=shift-exponent}.
12615 @item -fsanitize=shift-exponent
12616 @opindex fsanitize=shift-exponent
12617 This option enables checking that the second argument of a shift operation
12618 is not negative and is smaller than the precision of the promoted first
12621 @item -fsanitize=shift-base
12622 @opindex fsanitize=shift-base
12623 If the second argument of a shift operation is within range, check that the
12624 result of a shift operation is not undefined. Note that what exactly is
12625 considered undefined differs slightly between C and C++, as well as between
12626 ISO C90 and C99, etc.
12628 @item -fsanitize=integer-divide-by-zero
12629 @opindex fsanitize=integer-divide-by-zero
12630 Detect integer division by zero as well as @code{INT_MIN / -1} division.
12632 @item -fsanitize=unreachable
12633 @opindex fsanitize=unreachable
12634 With this option, the compiler turns the @code{__builtin_unreachable}
12635 call into a diagnostics message call instead. When reaching the
12636 @code{__builtin_unreachable} call, the behavior is undefined.
12638 @item -fsanitize=vla-bound
12639 @opindex fsanitize=vla-bound
12640 This option instructs the compiler to check that the size of a variable
12641 length array is positive.
12643 @item -fsanitize=null
12644 @opindex fsanitize=null
12645 This option enables pointer checking. Particularly, the application
12646 built with this option turned on will issue an error message when it
12647 tries to dereference a NULL pointer, or if a reference (possibly an
12648 rvalue reference) is bound to a NULL pointer, or if a method is invoked
12649 on an object pointed by a NULL pointer.
12651 @item -fsanitize=return
12652 @opindex fsanitize=return
12653 This option enables return statement checking. Programs
12654 built with this option turned on will issue an error message
12655 when the end of a non-void function is reached without actually
12656 returning a value. This option works in C++ only.
12658 @item -fsanitize=signed-integer-overflow
12659 @opindex fsanitize=signed-integer-overflow
12660 This option enables signed integer overflow checking. We check that
12661 the result of @code{+}, @code{*}, and both unary and binary @code{-}
12662 does not overflow in the signed arithmetics. Note, integer promotion
12663 rules must be taken into account. That is, the following is not an
12666 signed char a = SCHAR_MAX;
12670 @item -fsanitize=bounds
12671 @opindex fsanitize=bounds
12672 This option enables instrumentation of array bounds. Various out of bounds
12673 accesses are detected. Flexible array members, flexible array member-like
12674 arrays, and initializers of variables with static storage are not instrumented.
12676 @item -fsanitize=bounds-strict
12677 @opindex fsanitize=bounds-strict
12678 This option enables strict instrumentation of array bounds. Most out of bounds
12679 accesses are detected, including flexible array members and flexible array
12680 member-like arrays. Initializers of variables with static storage are not
12683 @item -fsanitize=alignment
12684 @opindex fsanitize=alignment
12686 This option enables checking of alignment of pointers when they are
12687 dereferenced, or when a reference is bound to insufficiently aligned target,
12688 or when a method or constructor is invoked on insufficiently aligned object.
12690 @item -fsanitize=object-size
12691 @opindex fsanitize=object-size
12692 This option enables instrumentation of memory references using the
12693 @code{__builtin_object_size} function. Various out of bounds pointer
12694 accesses are detected.
12696 @item -fsanitize=float-divide-by-zero
12697 @opindex fsanitize=float-divide-by-zero
12698 Detect floating-point division by zero. Unlike other similar options,
12699 @option{-fsanitize=float-divide-by-zero} is not enabled by
12700 @option{-fsanitize=undefined}, since floating-point division by zero can
12701 be a legitimate way of obtaining infinities and NaNs.
12703 @item -fsanitize=float-cast-overflow
12704 @opindex fsanitize=float-cast-overflow
12705 This option enables floating-point type to integer conversion checking.
12706 We check that the result of the conversion does not overflow.
12707 Unlike other similar options, @option{-fsanitize=float-cast-overflow} is
12708 not enabled by @option{-fsanitize=undefined}.
12709 This option does not work well with @code{FE_INVALID} exceptions enabled.
12711 @item -fsanitize=nonnull-attribute
12712 @opindex fsanitize=nonnull-attribute
12714 This option enables instrumentation of calls, checking whether null values
12715 are not passed to arguments marked as requiring a non-null value by the
12716 @code{nonnull} function attribute.
12718 @item -fsanitize=returns-nonnull-attribute
12719 @opindex fsanitize=returns-nonnull-attribute
12721 This option enables instrumentation of return statements in functions
12722 marked with @code{returns_nonnull} function attribute, to detect returning
12723 of null values from such functions.
12725 @item -fsanitize=bool
12726 @opindex fsanitize=bool
12728 This option enables instrumentation of loads from bool. If a value other
12729 than 0/1 is loaded, a run-time error is issued.
12731 @item -fsanitize=enum
12732 @opindex fsanitize=enum
12734 This option enables instrumentation of loads from an enum type. If
12735 a value outside the range of values for the enum type is loaded,
12736 a run-time error is issued.
12738 @item -fsanitize=vptr
12739 @opindex fsanitize=vptr
12741 This option enables instrumentation of C++ member function calls, member
12742 accesses and some conversions between pointers to base and derived classes,
12743 to verify the referenced object has the correct dynamic type.
12745 @item -fsanitize=pointer-overflow
12746 @opindex fsanitize=pointer-overflow
12748 This option enables instrumentation of pointer arithmetics. If the pointer
12749 arithmetics overflows, a run-time error is issued.
12751 @item -fsanitize=builtin
12752 @opindex fsanitize=builtin
12754 This option enables instrumentation of arguments to selected builtin
12755 functions. If an invalid value is passed to such arguments, a run-time
12756 error is issued. E.g.@ passing 0 as the argument to @code{__builtin_ctz}
12757 or @code{__builtin_clz} invokes undefined behavior and is diagnosed
12762 While @option{-ftrapv} causes traps for signed overflows to be emitted,
12763 @option{-fsanitize=undefined} gives a diagnostic message.
12764 This currently works only for the C family of languages.
12766 @item -fno-sanitize=all
12767 @opindex fno-sanitize=all
12769 This option disables all previously enabled sanitizers.
12770 @option{-fsanitize=all} is not allowed, as some sanitizers cannot be used
12773 @item -fasan-shadow-offset=@var{number}
12774 @opindex fasan-shadow-offset
12775 This option forces GCC to use custom shadow offset in AddressSanitizer checks.
12776 It is useful for experimenting with different shadow memory layouts in
12777 Kernel AddressSanitizer.
12779 @item -fsanitize-sections=@var{s1},@var{s2},...
12780 @opindex fsanitize-sections
12781 Sanitize global variables in selected user-defined sections. @var{si} may
12784 @item -fsanitize-recover@r{[}=@var{opts}@r{]}
12785 @opindex fsanitize-recover
12786 @opindex fno-sanitize-recover
12787 @option{-fsanitize-recover=} controls error recovery mode for sanitizers
12788 mentioned in comma-separated list of @var{opts}. Enabling this option
12789 for a sanitizer component causes it to attempt to continue
12790 running the program as if no error happened. This means multiple
12791 runtime errors can be reported in a single program run, and the exit
12792 code of the program may indicate success even when errors
12793 have been reported. The @option{-fno-sanitize-recover=} option
12794 can be used to alter
12795 this behavior: only the first detected error is reported
12796 and program then exits with a non-zero exit code.
12798 Currently this feature only works for @option{-fsanitize=undefined} (and its suboptions
12799 except for @option{-fsanitize=unreachable} and @option{-fsanitize=return}),
12800 @option{-fsanitize=float-cast-overflow}, @option{-fsanitize=float-divide-by-zero},
12801 @option{-fsanitize=bounds-strict},
12802 @option{-fsanitize=kernel-address} and @option{-fsanitize=address}.
12803 For these sanitizers error recovery is turned on by default,
12804 except @option{-fsanitize=address}, for which this feature is experimental.
12805 @option{-fsanitize-recover=all} and @option{-fno-sanitize-recover=all} is also
12806 accepted, the former enables recovery for all sanitizers that support it,
12807 the latter disables recovery for all sanitizers that support it.
12809 Even if a recovery mode is turned on the compiler side, it needs to be also
12810 enabled on the runtime library side, otherwise the failures are still fatal.
12811 The runtime library defaults to @code{halt_on_error=0} for
12812 ThreadSanitizer and UndefinedBehaviorSanitizer, while default value for
12813 AddressSanitizer is @code{halt_on_error=1}. This can be overridden through
12814 setting the @code{halt_on_error} flag in the corresponding environment variable.
12816 Syntax without an explicit @var{opts} parameter is deprecated. It is
12817 equivalent to specifying an @var{opts} list of:
12820 undefined,float-cast-overflow,float-divide-by-zero,bounds-strict
12823 @item -fsanitize-address-use-after-scope
12824 @opindex fsanitize-address-use-after-scope
12825 Enable sanitization of local variables to detect use-after-scope bugs.
12826 The option sets @option{-fstack-reuse} to @samp{none}.
12828 @item -fsanitize-undefined-trap-on-error
12829 @opindex fsanitize-undefined-trap-on-error
12830 The @option{-fsanitize-undefined-trap-on-error} option instructs the compiler to
12831 report undefined behavior using @code{__builtin_trap} rather than
12832 a @code{libubsan} library routine. The advantage of this is that the
12833 @code{libubsan} library is not needed and is not linked in, so this
12834 is usable even in freestanding environments.
12836 @item -fsanitize-coverage=trace-pc
12837 @opindex fsanitize-coverage=trace-pc
12838 Enable coverage-guided fuzzing code instrumentation.
12839 Inserts a call to @code{__sanitizer_cov_trace_pc} into every basic block.
12841 @item -fsanitize-coverage=trace-cmp
12842 @opindex fsanitize-coverage=trace-cmp
12843 Enable dataflow guided fuzzing code instrumentation.
12844 Inserts a call to @code{__sanitizer_cov_trace_cmp1},
12845 @code{__sanitizer_cov_trace_cmp2}, @code{__sanitizer_cov_trace_cmp4} or
12846 @code{__sanitizer_cov_trace_cmp8} for integral comparison with both operands
12847 variable or @code{__sanitizer_cov_trace_const_cmp1},
12848 @code{__sanitizer_cov_trace_const_cmp2},
12849 @code{__sanitizer_cov_trace_const_cmp4} or
12850 @code{__sanitizer_cov_trace_const_cmp8} for integral comparison with one
12851 operand constant, @code{__sanitizer_cov_trace_cmpf} or
12852 @code{__sanitizer_cov_trace_cmpd} for float or double comparisons and
12853 @code{__sanitizer_cov_trace_switch} for switch statements.
12855 @item -fcf-protection=@r{[}full@r{|}branch@r{|}return@r{|}none@r{]}
12856 @opindex fcf-protection
12857 Enable code instrumentation of control-flow transfers to increase
12858 program security by checking that target addresses of control-flow
12859 transfer instructions (such as indirect function call, function return,
12860 indirect jump) are valid. This prevents diverting the flow of control
12861 to an unexpected target. This is intended to protect against such
12862 threats as Return-oriented Programming (ROP), and similarly
12863 call/jmp-oriented programming (COP/JOP).
12865 The value @code{branch} tells the compiler to implement checking of
12866 validity of control-flow transfer at the point of indirect branch
12867 instructions, i.e.@: call/jmp instructions. The value @code{return}
12868 implements checking of validity at the point of returning from a
12869 function. The value @code{full} is an alias for specifying both
12870 @code{branch} and @code{return}. The value @code{none} turns off
12873 The macro @code{__CET__} is defined when @option{-fcf-protection} is
12874 used. The first bit of @code{__CET__} is set to 1 for the value
12875 @code{branch} and the second bit of @code{__CET__} is set to 1 for
12878 You can also use the @code{nocf_check} attribute to identify
12879 which functions and calls should be skipped from instrumentation
12880 (@pxref{Function Attributes}).
12882 Currently the x86 GNU/Linux target provides an implementation based
12883 on Intel Control-flow Enforcement Technology (CET).
12885 @item -fstack-protector
12886 @opindex fstack-protector
12887 Emit extra code to check for buffer overflows, such as stack smashing
12888 attacks. This is done by adding a guard variable to functions with
12889 vulnerable objects. This includes functions that call @code{alloca}, and
12890 functions with buffers larger than 8 bytes. The guards are initialized
12891 when a function is entered and then checked when the function exits.
12892 If a guard check fails, an error message is printed and the program exits.
12894 @item -fstack-protector-all
12895 @opindex fstack-protector-all
12896 Like @option{-fstack-protector} except that all functions are protected.
12898 @item -fstack-protector-strong
12899 @opindex fstack-protector-strong
12900 Like @option{-fstack-protector} but includes additional functions to
12901 be protected --- those that have local array definitions, or have
12902 references to local frame addresses.
12904 @item -fstack-protector-explicit
12905 @opindex fstack-protector-explicit
12906 Like @option{-fstack-protector} but only protects those functions which
12907 have the @code{stack_protect} attribute.
12909 @item -fstack-check
12910 @opindex fstack-check
12911 Generate code to verify that you do not go beyond the boundary of the
12912 stack. You should specify this flag if you are running in an
12913 environment with multiple threads, but you only rarely need to specify it in
12914 a single-threaded environment since stack overflow is automatically
12915 detected on nearly all systems if there is only one stack.
12917 Note that this switch does not actually cause checking to be done; the
12918 operating system or the language runtime must do that. The switch causes
12919 generation of code to ensure that they see the stack being extended.
12921 You can additionally specify a string parameter: @samp{no} means no
12922 checking, @samp{generic} means force the use of old-style checking,
12923 @samp{specific} means use the best checking method and is equivalent
12924 to bare @option{-fstack-check}.
12926 Old-style checking is a generic mechanism that requires no specific
12927 target support in the compiler but comes with the following drawbacks:
12931 Modified allocation strategy for large objects: they are always
12932 allocated dynamically if their size exceeds a fixed threshold. Note this
12933 may change the semantics of some code.
12936 Fixed limit on the size of the static frame of functions: when it is
12937 topped by a particular function, stack checking is not reliable and
12938 a warning is issued by the compiler.
12941 Inefficiency: because of both the modified allocation strategy and the
12942 generic implementation, code performance is hampered.
12945 Note that old-style stack checking is also the fallback method for
12946 @samp{specific} if no target support has been added in the compiler.
12948 @samp{-fstack-check=} is designed for Ada's needs to detect infinite recursion
12949 and stack overflows. @samp{specific} is an excellent choice when compiling
12950 Ada code. It is not generally sufficient to protect against stack-clash
12951 attacks. To protect against those you want @samp{-fstack-clash-protection}.
12953 @item -fstack-clash-protection
12954 @opindex fstack-clash-protection
12955 Generate code to prevent stack clash style attacks. When this option is
12956 enabled, the compiler will only allocate one page of stack space at a time
12957 and each page is accessed immediately after allocation. Thus, it prevents
12958 allocations from jumping over any stack guard page provided by the
12961 Most targets do not fully support stack clash protection. However, on
12962 those targets @option{-fstack-clash-protection} will protect dynamic stack
12963 allocations. @option{-fstack-clash-protection} may also provide limited
12964 protection for static stack allocations if the target supports
12965 @option{-fstack-check=specific}.
12967 @item -fstack-limit-register=@var{reg}
12968 @itemx -fstack-limit-symbol=@var{sym}
12969 @itemx -fno-stack-limit
12970 @opindex fstack-limit-register
12971 @opindex fstack-limit-symbol
12972 @opindex fno-stack-limit
12973 Generate code to ensure that the stack does not grow beyond a certain value,
12974 either the value of a register or the address of a symbol. If a larger
12975 stack is required, a signal is raised at run time. For most targets,
12976 the signal is raised before the stack overruns the boundary, so
12977 it is possible to catch the signal without taking special precautions.
12979 For instance, if the stack starts at absolute address @samp{0x80000000}
12980 and grows downwards, you can use the flags
12981 @option{-fstack-limit-symbol=__stack_limit} and
12982 @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit
12983 of 128KB@. Note that this may only work with the GNU linker.
12985 You can locally override stack limit checking by using the
12986 @code{no_stack_limit} function attribute (@pxref{Function Attributes}).
12988 @item -fsplit-stack
12989 @opindex fsplit-stack
12990 Generate code to automatically split the stack before it overflows.
12991 The resulting program has a discontiguous stack which can only
12992 overflow if the program is unable to allocate any more memory. This
12993 is most useful when running threaded programs, as it is no longer
12994 necessary to calculate a good stack size to use for each thread. This
12995 is currently only implemented for the x86 targets running
12998 When code compiled with @option{-fsplit-stack} calls code compiled
12999 without @option{-fsplit-stack}, there may not be much stack space
13000 available for the latter code to run. If compiling all code,
13001 including library code, with @option{-fsplit-stack} is not an option,
13002 then the linker can fix up these calls so that the code compiled
13003 without @option{-fsplit-stack} always has a large stack. Support for
13004 this is implemented in the gold linker in GNU binutils release 2.21
13007 @item -fvtable-verify=@r{[}std@r{|}preinit@r{|}none@r{]}
13008 @opindex fvtable-verify
13009 This option is only available when compiling C++ code.
13010 It turns on (or off, if using @option{-fvtable-verify=none}) the security
13011 feature that verifies at run time, for every virtual call, that
13012 the vtable pointer through which the call is made is valid for the type of
13013 the object, and has not been corrupted or overwritten. If an invalid vtable
13014 pointer is detected at run time, an error is reported and execution of the
13015 program is immediately halted.
13017 This option causes run-time data structures to be built at program startup,
13018 which are used for verifying the vtable pointers.
13019 The options @samp{std} and @samp{preinit}
13020 control the timing of when these data structures are built. In both cases the
13021 data structures are built before execution reaches @code{main}. Using
13022 @option{-fvtable-verify=std} causes the data structures to be built after
13023 shared libraries have been loaded and initialized.
13024 @option{-fvtable-verify=preinit} causes them to be built before shared
13025 libraries have been loaded and initialized.
13027 If this option appears multiple times in the command line with different
13028 values specified, @samp{none} takes highest priority over both @samp{std} and
13029 @samp{preinit}; @samp{preinit} takes priority over @samp{std}.
13032 @opindex fvtv-debug
13033 When used in conjunction with @option{-fvtable-verify=std} or
13034 @option{-fvtable-verify=preinit}, causes debug versions of the
13035 runtime functions for the vtable verification feature to be called.
13036 This flag also causes the compiler to log information about which
13037 vtable pointers it finds for each class.
13038 This information is written to a file named @file{vtv_set_ptr_data.log}
13039 in the directory named by the environment variable @env{VTV_LOGS_DIR}
13040 if that is defined or the current working directory otherwise.
13042 Note: This feature @emph{appends} data to the log file. If you want a fresh log
13043 file, be sure to delete any existing one.
13046 @opindex fvtv-counts
13047 This is a debugging flag. When used in conjunction with
13048 @option{-fvtable-verify=std} or @option{-fvtable-verify=preinit}, this
13049 causes the compiler to keep track of the total number of virtual calls
13050 it encounters and the number of verifications it inserts. It also
13051 counts the number of calls to certain run-time library functions
13052 that it inserts and logs this information for each compilation unit.
13053 The compiler writes this information to a file named
13054 @file{vtv_count_data.log} in the directory named by the environment
13055 variable @env{VTV_LOGS_DIR} if that is defined or the current working
13056 directory otherwise. It also counts the size of the vtable pointer sets
13057 for each class, and writes this information to @file{vtv_class_set_sizes.log}
13058 in the same directory.
13060 Note: This feature @emph{appends} data to the log files. To get fresh log
13061 files, be sure to delete any existing ones.
13063 @item -finstrument-functions
13064 @opindex finstrument-functions
13065 Generate instrumentation calls for entry and exit to functions. Just
13066 after function entry and just before function exit, the following
13067 profiling functions are called with the address of the current
13068 function and its call site. (On some platforms,
13069 @code{__builtin_return_address} does not work beyond the current
13070 function, so the call site information may not be available to the
13071 profiling functions otherwise.)
13074 void __cyg_profile_func_enter (void *this_fn,
13076 void __cyg_profile_func_exit (void *this_fn,
13080 The first argument is the address of the start of the current function,
13081 which may be looked up exactly in the symbol table.
13083 This instrumentation is also done for functions expanded inline in other
13084 functions. The profiling calls indicate where, conceptually, the
13085 inline function is entered and exited. This means that addressable
13086 versions of such functions must be available. If all your uses of a
13087 function are expanded inline, this may mean an additional expansion of
13088 code size. If you use @code{extern inline} in your C code, an
13089 addressable version of such functions must be provided. (This is
13090 normally the case anyway, but if you get lucky and the optimizer always
13091 expands the functions inline, you might have gotten away without
13092 providing static copies.)
13094 A function may be given the attribute @code{no_instrument_function}, in
13095 which case this instrumentation is not done. This can be used, for
13096 example, for the profiling functions listed above, high-priority
13097 interrupt routines, and any functions from which the profiling functions
13098 cannot safely be called (perhaps signal handlers, if the profiling
13099 routines generate output or allocate memory).
13100 @xref{Common Function Attributes}.
13102 @item -finstrument-functions-exclude-file-list=@var{file},@var{file},@dots{}
13103 @opindex finstrument-functions-exclude-file-list
13105 Set the list of functions that are excluded from instrumentation (see
13106 the description of @option{-finstrument-functions}). If the file that
13107 contains a function definition matches with one of @var{file}, then
13108 that function is not instrumented. The match is done on substrings:
13109 if the @var{file} parameter is a substring of the file name, it is
13110 considered to be a match.
13115 -finstrument-functions-exclude-file-list=/bits/stl,include/sys
13119 excludes any inline function defined in files whose pathnames
13120 contain @file{/bits/stl} or @file{include/sys}.
13122 If, for some reason, you want to include letter @samp{,} in one of
13123 @var{sym}, write @samp{\,}. For example,
13124 @option{-finstrument-functions-exclude-file-list='\,\,tmp'}
13125 (note the single quote surrounding the option).
13127 @item -finstrument-functions-exclude-function-list=@var{sym},@var{sym},@dots{}
13128 @opindex finstrument-functions-exclude-function-list
13130 This is similar to @option{-finstrument-functions-exclude-file-list},
13131 but this option sets the list of function names to be excluded from
13132 instrumentation. The function name to be matched is its user-visible
13133 name, such as @code{vector<int> blah(const vector<int> &)}, not the
13134 internal mangled name (e.g., @code{_Z4blahRSt6vectorIiSaIiEE}). The
13135 match is done on substrings: if the @var{sym} parameter is a substring
13136 of the function name, it is considered to be a match. For C99 and C++
13137 extended identifiers, the function name must be given in UTF-8, not
13138 using universal character names.
13140 @item -fpatchable-function-entry=@var{N}[,@var{M}]
13141 @opindex fpatchable-function-entry
13142 Generate @var{N} NOPs right at the beginning
13143 of each function, with the function entry point before the @var{M}th NOP.
13144 If @var{M} is omitted, it defaults to @code{0} so the
13145 function entry points to the address just at the first NOP.
13146 The NOP instructions reserve extra space which can be used to patch in
13147 any desired instrumentation at run time, provided that the code segment
13148 is writable. The amount of space is controllable indirectly via
13149 the number of NOPs; the NOP instruction used corresponds to the instruction
13150 emitted by the internal GCC back-end interface @code{gen_nop}. This behavior
13151 is target-specific and may also depend on the architecture variant and/or
13152 other compilation options.
13154 For run-time identification, the starting addresses of these areas,
13155 which correspond to their respective function entries minus @var{M},
13156 are additionally collected in the @code{__patchable_function_entries}
13157 section of the resulting binary.
13159 Note that the value of @code{__attribute__ ((patchable_function_entry
13160 (N,M)))} takes precedence over command-line option
13161 @option{-fpatchable-function-entry=N,M}. This can be used to increase
13162 the area size or to remove it completely on a single function.
13163 If @code{N=0}, no pad location is recorded.
13165 The NOP instructions are inserted at---and maybe before, depending on
13166 @var{M}---the function entry address, even before the prologue.
13171 @node Preprocessor Options
13172 @section Options Controlling the Preprocessor
13173 @cindex preprocessor options
13174 @cindex options, preprocessor
13176 These options control the C preprocessor, which is run on each C source
13177 file before actual compilation.
13179 If you use the @option{-E} option, nothing is done except preprocessing.
13180 Some of these options make sense only together with @option{-E} because
13181 they cause the preprocessor output to be unsuitable for actual
13184 In addition to the options listed here, there are a number of options
13185 to control search paths for include files documented in
13186 @ref{Directory Options}.
13187 Options to control preprocessor diagnostics are listed in
13188 @ref{Warning Options}.
13191 @include cppopts.texi
13193 @item -Wp,@var{option}
13195 You can use @option{-Wp,@var{option}} to bypass the compiler driver
13196 and pass @var{option} directly through to the preprocessor. If
13197 @var{option} contains commas, it is split into multiple options at the
13198 commas. However, many options are modified, translated or interpreted
13199 by the compiler driver before being passed to the preprocessor, and
13200 @option{-Wp} forcibly bypasses this phase. The preprocessor's direct
13201 interface is undocumented and subject to change, so whenever possible
13202 you should avoid using @option{-Wp} and let the driver handle the
13205 @item -Xpreprocessor @var{option}
13206 @opindex Xpreprocessor
13207 Pass @var{option} as an option to the preprocessor. You can use this to
13208 supply system-specific preprocessor options that GCC does not
13211 If you want to pass an option that takes an argument, you must use
13212 @option{-Xpreprocessor} twice, once for the option and once for the argument.
13214 @item -no-integrated-cpp
13215 @opindex no-integrated-cpp
13216 Perform preprocessing as a separate pass before compilation.
13217 By default, GCC performs preprocessing as an integrated part of
13218 input tokenization and parsing.
13219 If this option is provided, the appropriate language front end
13220 (@command{cc1}, @command{cc1plus}, or @command{cc1obj} for C, C++,
13221 and Objective-C, respectively) is instead invoked twice,
13222 once for preprocessing only and once for actual compilation
13223 of the preprocessed input.
13224 This option may be useful in conjunction with the @option{-B} or
13225 @option{-wrapper} options to specify an alternate preprocessor or
13226 perform additional processing of the program source between
13227 normal preprocessing and compilation.
13231 @node Assembler Options
13232 @section Passing Options to the Assembler
13234 @c prevent bad page break with this line
13235 You can pass options to the assembler.
13238 @item -Wa,@var{option}
13240 Pass @var{option} as an option to the assembler. If @var{option}
13241 contains commas, it is split into multiple options at the commas.
13243 @item -Xassembler @var{option}
13244 @opindex Xassembler
13245 Pass @var{option} as an option to the assembler. You can use this to
13246 supply system-specific assembler options that GCC does not
13249 If you want to pass an option that takes an argument, you must use
13250 @option{-Xassembler} twice, once for the option and once for the argument.
13255 @section Options for Linking
13256 @cindex link options
13257 @cindex options, linking
13259 These options come into play when the compiler links object files into
13260 an executable output file. They are meaningless if the compiler is
13261 not doing a link step.
13265 @item @var{object-file-name}
13266 A file name that does not end in a special recognized suffix is
13267 considered to name an object file or library. (Object files are
13268 distinguished from libraries by the linker according to the file
13269 contents.) If linking is done, these object files are used as input
13278 If any of these options is used, then the linker is not run, and
13279 object file names should not be used as arguments. @xref{Overall
13282 @item -flinker-output=@var{type}
13283 @opindex flinker-output
13284 This option controls code generation of the link-time optimizer. By
13285 default the linker output is automatically determined by the linker
13286 plugin. For debugging the compiler and if incremental linking with a
13287 non-LTO object file is desired, it may be useful to control the type
13290 If @var{type} is @samp{exec}, code generation produces a static
13291 binary. In this case @option{-fpic} and @option{-fpie} are both
13294 If @var{type} is @samp{dyn}, code generation produces a shared
13295 library. In this case @option{-fpic} or @option{-fPIC} is preserved,
13296 but not enabled automatically. This allows to build shared libraries
13297 without position-independent code on architectures where this is
13298 possible, i.e.@: on x86.
13300 If @var{type} is @samp{pie}, code generation produces an @option{-fpie}
13301 executable. This results in similar optimizations as @samp{exec}
13302 except that @option{-fpie} is not disabled if specified at compilation
13305 If @var{type} is @samp{rel}, the compiler assumes that incremental linking is
13306 done. The sections containing intermediate code for link-time optimization are
13307 merged, pre-optimized, and output to the resulting object file. In addition, if
13308 @option{-ffat-lto-objects} is specified, binary code is produced for future
13309 non-LTO linking. The object file produced by incremental linking is smaller
13310 than a static library produced from the same object files. At link time the
13311 result of incremental linking also loads faster than a static
13312 library assuming that the majority of objects in the library are used.
13314 Finally @samp{nolto-rel} configures the compiler for incremental linking where
13315 code generation is forced, a final binary is produced, and the intermediate
13316 code for later link-time optimization is stripped. When multiple object files
13317 are linked together the resulting code is better optimized than with
13318 link-time optimizations disabled (for example, cross-module inlining
13319 happens), but most of benefits of whole program optimizations are lost.
13321 During the incremental link (by @option{-r}) the linker plugin defaults to
13322 @option{rel}. With current interfaces to GNU Binutils it is however not
13323 possible to incrementally link LTO objects and non-LTO objects into a single
13324 mixed object file. If any of object files in incremental link cannot
13325 be used for link-time optimization, the linker plugin issues a warning and
13326 uses @samp{nolto-rel}. To maintain whole program optimization, it is
13327 recommended to link such objects into static library instead. Alternatively it
13328 is possible to use H.J. Lu's binutils with support for mixed objects.
13331 @opindex fuse-ld=bfd
13332 Use the @command{bfd} linker instead of the default linker.
13334 @item -fuse-ld=gold
13335 @opindex fuse-ld=gold
13336 Use the @command{gold} linker instead of the default linker.
13339 @opindex fuse-ld=lld
13340 Use the LLVM @command{lld} linker instead of the default linker.
13343 @item -l@var{library}
13344 @itemx -l @var{library}
13346 Search the library named @var{library} when linking. (The second
13347 alternative with the library as a separate argument is only for
13348 POSIX compliance and is not recommended.)
13350 The @option{-l} option is passed directly to the linker by GCC. Refer
13351 to your linker documentation for exact details. The general
13352 description below applies to the GNU linker.
13354 The linker searches a standard list of directories for the library.
13355 The directories searched include several standard system directories
13356 plus any that you specify with @option{-L}.
13358 Static libraries are archives of object files, and have file names
13359 like @file{lib@var{library}.a}. Some targets also support shared
13360 libraries, which typically have names like @file{lib@var{library}.so}.
13361 If both static and shared libraries are found, the linker gives
13362 preference to linking with the shared library unless the
13363 @option{-static} option is used.
13365 It makes a difference where in the command you write this option; the
13366 linker searches and processes libraries and object files in the order they
13367 are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z}
13368 after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers
13369 to functions in @samp{z}, those functions may not be loaded.
13373 You need this special case of the @option{-l} option in order to
13374 link an Objective-C or Objective-C++ program.
13376 @item -nostartfiles
13377 @opindex nostartfiles
13378 Do not use the standard system startup files when linking.
13379 The standard system libraries are used normally, unless @option{-nostdlib},
13380 @option{-nolibc}, or @option{-nodefaultlibs} is used.
13382 @item -nodefaultlibs
13383 @opindex nodefaultlibs
13384 Do not use the standard system libraries when linking.
13385 Only the libraries you specify are passed to the linker, and options
13386 specifying linkage of the system libraries, such as @option{-static-libgcc}
13387 or @option{-shared-libgcc}, are ignored.
13388 The standard startup files are used normally, unless @option{-nostartfiles}
13391 The compiler may generate calls to @code{memcmp},
13392 @code{memset}, @code{memcpy} and @code{memmove}.
13393 These entries are usually resolved by entries in
13394 libc. These entry points should be supplied through some other
13395 mechanism when this option is specified.
13399 Do not use the C library or system libraries tightly coupled with it when
13400 linking. Still link with the startup files, @file{libgcc} or toolchain
13401 provided language support libraries such as @file{libgnat}, @file{libgfortran}
13402 or @file{libstdc++} unless options preventing their inclusion are used as
13403 well. This typically removes @option{-lc} from the link command line, as well
13404 as system libraries that normally go with it and become meaningless when
13405 absence of a C library is assumed, for example @option{-lpthread} or
13406 @option{-lm} in some configurations. This is intended for bare-board
13407 targets when there is indeed no C library available.
13411 Do not use the standard system startup files or libraries when linking.
13412 No startup files and only the libraries you specify are passed to
13413 the linker, and options specifying linkage of the system libraries, such as
13414 @option{-static-libgcc} or @option{-shared-libgcc}, are ignored.
13416 The compiler may generate calls to @code{memcmp}, @code{memset},
13417 @code{memcpy} and @code{memmove}.
13418 These entries are usually resolved by entries in
13419 libc. These entry points should be supplied through some other
13420 mechanism when this option is specified.
13422 @cindex @option{-lgcc}, use with @option{-nostdlib}
13423 @cindex @option{-nostdlib} and unresolved references
13424 @cindex unresolved references and @option{-nostdlib}
13425 @cindex @option{-lgcc}, use with @option{-nodefaultlibs}
13426 @cindex @option{-nodefaultlibs} and unresolved references
13427 @cindex unresolved references and @option{-nodefaultlibs}
13428 One of the standard libraries bypassed by @option{-nostdlib} and
13429 @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines
13430 which GCC uses to overcome shortcomings of particular machines, or special
13431 needs for some languages.
13432 (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler
13433 Collection (GCC) Internals},
13434 for more discussion of @file{libgcc.a}.)
13435 In most cases, you need @file{libgcc.a} even when you want to avoid
13436 other standard libraries. In other words, when you specify @option{-nostdlib}
13437 or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well.
13438 This ensures that you have no unresolved references to internal GCC
13439 library subroutines.
13440 (An example of such an internal subroutine is @code{__main}, used to ensure C++
13441 constructors are called; @pxref{Collect2,,@code{collect2}, gccint,
13442 GNU Compiler Collection (GCC) Internals}.)
13444 @item -e @var{entry}
13445 @itemx --entry=@var{entry}
13449 Specify that the program entry point is @var{entry}. The argument is
13450 interpreted by the linker; the GNU linker accepts either a symbol name
13455 Produce a dynamically linked position independent executable on targets
13456 that support it. For predictable results, you must also specify the same
13457 set of options used for compilation (@option{-fpie}, @option{-fPIE},
13458 or model suboptions) when you specify this linker option.
13462 Don't produce a dynamically linked position independent executable.
13465 @opindex static-pie
13466 Produce a static position independent executable on targets that support
13467 it. A static position independent executable is similar to a static
13468 executable, but can be loaded at any address without a dynamic linker.
13469 For predictable results, you must also specify the same set of options
13470 used for compilation (@option{-fpie}, @option{-fPIE}, or model
13471 suboptions) when you specify this linker option.
13475 Link with the POSIX threads library. This option is supported on
13476 GNU/Linux targets, most other Unix derivatives, and also on
13477 x86 Cygwin and MinGW targets. On some targets this option also sets
13478 flags for the preprocessor, so it should be used consistently for both
13479 compilation and linking.
13483 Produce a relocatable object as output. This is also known as partial
13488 Pass the flag @option{-export-dynamic} to the ELF linker, on targets
13489 that support it. This instructs the linker to add all symbols, not
13490 only used ones, to the dynamic symbol table. This option is needed
13491 for some uses of @code{dlopen} or to allow obtaining backtraces
13492 from within a program.
13496 Remove all symbol table and relocation information from the executable.
13500 On systems that support dynamic linking, this overrides @option{-pie}
13501 and prevents linking with the shared libraries. On other systems, this
13502 option has no effect.
13506 Produce a shared object which can then be linked with other objects to
13507 form an executable. Not all systems support this option. For predictable
13508 results, you must also specify the same set of options used for compilation
13509 (@option{-fpic}, @option{-fPIC}, or model suboptions) when
13510 you specify this linker option.@footnote{On some systems, @samp{gcc -shared}
13511 needs to build supplementary stub code for constructors to work. On
13512 multi-libbed systems, @samp{gcc -shared} must select the correct support
13513 libraries to link against. Failing to supply the correct flags may lead
13514 to subtle defects. Supplying them in cases where they are not necessary
13517 @item -shared-libgcc
13518 @itemx -static-libgcc
13519 @opindex shared-libgcc
13520 @opindex static-libgcc
13521 On systems that provide @file{libgcc} as a shared library, these options
13522 force the use of either the shared or static version, respectively.
13523 If no shared version of @file{libgcc} was built when the compiler was
13524 configured, these options have no effect.
13526 There are several situations in which an application should use the
13527 shared @file{libgcc} instead of the static version. The most common
13528 of these is when the application wishes to throw and catch exceptions
13529 across different shared libraries. In that case, each of the libraries
13530 as well as the application itself should use the shared @file{libgcc}.
13532 Therefore, the G++ driver automatically adds @option{-shared-libgcc}
13533 whenever you build a shared library or a main executable, because C++
13534 programs typically use exceptions, so this is the right thing to do.
13536 If, instead, you use the GCC driver to create shared libraries, you may
13537 find that they are not always linked with the shared @file{libgcc}.
13538 If GCC finds, at its configuration time, that you have a non-GNU linker
13539 or a GNU linker that does not support option @option{--eh-frame-hdr},
13540 it links the shared version of @file{libgcc} into shared libraries
13541 by default. Otherwise, it takes advantage of the linker and optimizes
13542 away the linking with the shared version of @file{libgcc}, linking with
13543 the static version of libgcc by default. This allows exceptions to
13544 propagate through such shared libraries, without incurring relocation
13545 costs at library load time.
13547 However, if a library or main executable is supposed to throw or catch
13548 exceptions, you must link it using the G++ driver, or using the option
13549 @option{-shared-libgcc}, such that it is linked with the shared
13552 @item -static-libasan
13553 @opindex static-libasan
13554 When the @option{-fsanitize=address} option is used to link a program,
13555 the GCC driver automatically links against @option{libasan}. If
13556 @file{libasan} is available as a shared library, and the @option{-static}
13557 option is not used, then this links against the shared version of
13558 @file{libasan}. The @option{-static-libasan} option directs the GCC
13559 driver to link @file{libasan} statically, without necessarily linking
13560 other libraries statically.
13562 @item -static-libtsan
13563 @opindex static-libtsan
13564 When the @option{-fsanitize=thread} option is used to link a program,
13565 the GCC driver automatically links against @option{libtsan}. If
13566 @file{libtsan} is available as a shared library, and the @option{-static}
13567 option is not used, then this links against the shared version of
13568 @file{libtsan}. The @option{-static-libtsan} option directs the GCC
13569 driver to link @file{libtsan} statically, without necessarily linking
13570 other libraries statically.
13572 @item -static-liblsan
13573 @opindex static-liblsan
13574 When the @option{-fsanitize=leak} option is used to link a program,
13575 the GCC driver automatically links against @option{liblsan}. If
13576 @file{liblsan} is available as a shared library, and the @option{-static}
13577 option is not used, then this links against the shared version of
13578 @file{liblsan}. The @option{-static-liblsan} option directs the GCC
13579 driver to link @file{liblsan} statically, without necessarily linking
13580 other libraries statically.
13582 @item -static-libubsan
13583 @opindex static-libubsan
13584 When the @option{-fsanitize=undefined} option is used to link a program,
13585 the GCC driver automatically links against @option{libubsan}. If
13586 @file{libubsan} is available as a shared library, and the @option{-static}
13587 option is not used, then this links against the shared version of
13588 @file{libubsan}. The @option{-static-libubsan} option directs the GCC
13589 driver to link @file{libubsan} statically, without necessarily linking
13590 other libraries statically.
13592 @item -static-libstdc++
13593 @opindex static-libstdc++
13594 When the @command{g++} program is used to link a C++ program, it
13595 normally automatically links against @option{libstdc++}. If
13596 @file{libstdc++} is available as a shared library, and the
13597 @option{-static} option is not used, then this links against the
13598 shared version of @file{libstdc++}. That is normally fine. However, it
13599 is sometimes useful to freeze the version of @file{libstdc++} used by
13600 the program without going all the way to a fully static link. The
13601 @option{-static-libstdc++} option directs the @command{g++} driver to
13602 link @file{libstdc++} statically, without necessarily linking other
13603 libraries statically.
13607 Bind references to global symbols when building a shared object. Warn
13608 about any unresolved references (unless overridden by the link editor
13609 option @option{-Xlinker -z -Xlinker defs}). Only a few systems support
13612 @item -T @var{script}
13614 @cindex linker script
13615 Use @var{script} as the linker script. This option is supported by most
13616 systems using the GNU linker. On some targets, such as bare-board
13617 targets without an operating system, the @option{-T} option may be required
13618 when linking to avoid references to undefined symbols.
13620 @item -Xlinker @var{option}
13622 Pass @var{option} as an option to the linker. You can use this to
13623 supply system-specific linker options that GCC does not recognize.
13625 If you want to pass an option that takes a separate argument, you must use
13626 @option{-Xlinker} twice, once for the option and once for the argument.
13627 For example, to pass @option{-assert definitions}, you must write
13628 @option{-Xlinker -assert -Xlinker definitions}. It does not work to write
13629 @option{-Xlinker "-assert definitions"}, because this passes the entire
13630 string as a single argument, which is not what the linker expects.
13632 When using the GNU linker, it is usually more convenient to pass
13633 arguments to linker options using the @option{@var{option}=@var{value}}
13634 syntax than as separate arguments. For example, you can specify
13635 @option{-Xlinker -Map=output.map} rather than
13636 @option{-Xlinker -Map -Xlinker output.map}. Other linkers may not support
13637 this syntax for command-line options.
13639 @item -Wl,@var{option}
13641 Pass @var{option} as an option to the linker. If @var{option} contains
13642 commas, it is split into multiple options at the commas. You can use this
13643 syntax to pass an argument to the option.
13644 For example, @option{-Wl,-Map,output.map} passes @option{-Map output.map} to the
13645 linker. When using the GNU linker, you can also get the same effect with
13646 @option{-Wl,-Map=output.map}.
13648 @item -u @var{symbol}
13650 Pretend the symbol @var{symbol} is undefined, to force linking of
13651 library modules to define it. You can use @option{-u} multiple times with
13652 different symbols to force loading of additional library modules.
13654 @item -z @var{keyword}
13656 @option{-z} is passed directly on to the linker along with the keyword
13657 @var{keyword}. See the section in the documentation of your linker for
13658 permitted values and their meanings.
13661 @node Directory Options
13662 @section Options for Directory Search
13663 @cindex directory options
13664 @cindex options, directory search
13665 @cindex search path
13667 These options specify directories to search for header files, for
13668 libraries and for parts of the compiler:
13671 @include cppdiropts.texi
13673 @item -iplugindir=@var{dir}
13674 @opindex iplugindir=
13675 Set the directory to search for plugins that are passed
13676 by @option{-fplugin=@var{name}} instead of
13677 @option{-fplugin=@var{path}/@var{name}.so}. This option is not meant
13678 to be used by the user, but only passed by the driver.
13682 Add directory @var{dir} to the list of directories to be searched
13685 @item -B@var{prefix}
13687 This option specifies where to find the executables, libraries,
13688 include files, and data files of the compiler itself.
13690 The compiler driver program runs one or more of the subprograms
13691 @command{cpp}, @command{cc1}, @command{as} and @command{ld}. It tries
13692 @var{prefix} as a prefix for each program it tries to run, both with and
13693 without @samp{@var{machine}/@var{version}/} for the corresponding target
13694 machine and compiler version.
13696 For each subprogram to be run, the compiler driver first tries the
13697 @option{-B} prefix, if any. If that name is not found, or if @option{-B}
13698 is not specified, the driver tries two standard prefixes,
13699 @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc/}. If neither of
13700 those results in a file name that is found, the unmodified program
13701 name is searched for using the directories specified in your
13702 @env{PATH} environment variable.
13704 The compiler checks to see if the path provided by @option{-B}
13705 refers to a directory, and if necessary it adds a directory
13706 separator character at the end of the path.
13708 @option{-B} prefixes that effectively specify directory names also apply
13709 to libraries in the linker, because the compiler translates these
13710 options into @option{-L} options for the linker. They also apply to
13711 include files in the preprocessor, because the compiler translates these
13712 options into @option{-isystem} options for the preprocessor. In this case,
13713 the compiler appends @samp{include} to the prefix.
13715 The runtime support file @file{libgcc.a} can also be searched for using
13716 the @option{-B} prefix, if needed. If it is not found there, the two
13717 standard prefixes above are tried, and that is all. The file is left
13718 out of the link if it is not found by those means.
13720 Another way to specify a prefix much like the @option{-B} prefix is to use
13721 the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment
13724 As a special kludge, if the path provided by @option{-B} is
13725 @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to
13726 9, then it is replaced by @file{[dir/]include}. This is to help
13727 with boot-strapping the compiler.
13729 @item -no-canonical-prefixes
13730 @opindex no-canonical-prefixes
13731 Do not expand any symbolic links, resolve references to @samp{/../}
13732 or @samp{/./}, or make the path absolute when generating a relative
13735 @item --sysroot=@var{dir}
13737 Use @var{dir} as the logical root directory for headers and libraries.
13738 For example, if the compiler normally searches for headers in
13739 @file{/usr/include} and libraries in @file{/usr/lib}, it instead
13740 searches @file{@var{dir}/usr/include} and @file{@var{dir}/usr/lib}.
13742 If you use both this option and the @option{-isysroot} option, then
13743 the @option{--sysroot} option applies to libraries, but the
13744 @option{-isysroot} option applies to header files.
13746 The GNU linker (beginning with version 2.16) has the necessary support
13747 for this option. If your linker does not support this option, the
13748 header file aspect of @option{--sysroot} still works, but the
13749 library aspect does not.
13751 @item --no-sysroot-suffix
13752 @opindex no-sysroot-suffix
13753 For some targets, a suffix is added to the root directory specified
13754 with @option{--sysroot}, depending on the other options used, so that
13755 headers may for example be found in
13756 @file{@var{dir}/@var{suffix}/usr/include} instead of
13757 @file{@var{dir}/usr/include}. This option disables the addition of
13762 @node Code Gen Options
13763 @section Options for Code Generation Conventions
13764 @cindex code generation conventions
13765 @cindex options, code generation
13766 @cindex run-time options
13768 These machine-independent options control the interface conventions
13769 used in code generation.
13771 Most of them have both positive and negative forms; the negative form
13772 of @option{-ffoo} is @option{-fno-foo}. In the table below, only
13773 one of the forms is listed---the one that is not the default. You
13774 can figure out the other form by either removing @samp{no-} or adding
13778 @item -fstack-reuse=@var{reuse-level}
13779 @opindex fstack_reuse
13780 This option controls stack space reuse for user declared local/auto variables
13781 and compiler generated temporaries. @var{reuse_level} can be @samp{all},
13782 @samp{named_vars}, or @samp{none}. @samp{all} enables stack reuse for all
13783 local variables and temporaries, @samp{named_vars} enables the reuse only for
13784 user defined local variables with names, and @samp{none} disables stack reuse
13785 completely. The default value is @samp{all}. The option is needed when the
13786 program extends the lifetime of a scoped local variable or a compiler generated
13787 temporary beyond the end point defined by the language. When a lifetime of
13788 a variable ends, and if the variable lives in memory, the optimizing compiler
13789 has the freedom to reuse its stack space with other temporaries or scoped
13790 local variables whose live range does not overlap with it. Legacy code extending
13791 local lifetime is likely to break with the stack reuse optimization.
13810 if (*p == 10) // out of scope use of local1
13821 A(int k) : i(k), j(k) @{ @}
13828 void foo(const A& ar)
13835 foo(A(10)); // temp object's lifetime ends when foo returns
13841 ap->i+= 10; // ap references out of scope temp whose space
13842 // is reused with a. What is the value of ap->i?
13847 The lifetime of a compiler generated temporary is well defined by the C++
13848 standard. When a lifetime of a temporary ends, and if the temporary lives
13849 in memory, the optimizing compiler has the freedom to reuse its stack
13850 space with other temporaries or scoped local variables whose live range
13851 does not overlap with it. However some of the legacy code relies on
13852 the behavior of older compilers in which temporaries' stack space is
13853 not reused, the aggressive stack reuse can lead to runtime errors. This
13854 option is used to control the temporary stack reuse optimization.
13858 This option generates traps for signed overflow on addition, subtraction,
13859 multiplication operations.
13860 The options @option{-ftrapv} and @option{-fwrapv} override each other, so using
13861 @option{-ftrapv} @option{-fwrapv} on the command-line results in
13862 @option{-fwrapv} being effective. Note that only active options override, so
13863 using @option{-ftrapv} @option{-fwrapv} @option{-fno-wrapv} on the command-line
13864 results in @option{-ftrapv} being effective.
13868 This option instructs the compiler to assume that signed arithmetic
13869 overflow of addition, subtraction and multiplication wraps around
13870 using twos-complement representation. This flag enables some optimizations
13871 and disables others.
13872 The options @option{-ftrapv} and @option{-fwrapv} override each other, so using
13873 @option{-ftrapv} @option{-fwrapv} on the command-line results in
13874 @option{-fwrapv} being effective. Note that only active options override, so
13875 using @option{-ftrapv} @option{-fwrapv} @option{-fno-wrapv} on the command-line
13876 results in @option{-ftrapv} being effective.
13878 @item -fwrapv-pointer
13879 @opindex fwrapv-pointer
13880 This option instructs the compiler to assume that pointer arithmetic
13881 overflow on addition and subtraction wraps around using twos-complement
13882 representation. This flag disables some optimizations which assume
13883 pointer overflow is invalid.
13885 @item -fstrict-overflow
13886 @opindex fstrict-overflow
13887 This option implies @option{-fno-wrapv} @option{-fno-wrapv-pointer} and when
13888 negated implies @option{-fwrapv} @option{-fwrapv-pointer}.
13891 @opindex fexceptions
13892 Enable exception handling. Generates extra code needed to propagate
13893 exceptions. For some targets, this implies GCC generates frame
13894 unwind information for all functions, which can produce significant data
13895 size overhead, although it does not affect execution. If you do not
13896 specify this option, GCC enables it by default for languages like
13897 C++ that normally require exception handling, and disables it for
13898 languages like C that do not normally require it. However, you may need
13899 to enable this option when compiling C code that needs to interoperate
13900 properly with exception handlers written in C++. You may also wish to
13901 disable this option if you are compiling older C++ programs that don't
13902 use exception handling.
13904 @item -fnon-call-exceptions
13905 @opindex fnon-call-exceptions
13906 Generate code that allows trapping instructions to throw exceptions.
13907 Note that this requires platform-specific runtime support that does
13908 not exist everywhere. Moreover, it only allows @emph{trapping}
13909 instructions to throw exceptions, i.e.@: memory references or floating-point
13910 instructions. It does not allow exceptions to be thrown from
13911 arbitrary signal handlers such as @code{SIGALRM}.
13913 @item -fdelete-dead-exceptions
13914 @opindex fdelete-dead-exceptions
13915 Consider that instructions that may throw exceptions but don't otherwise
13916 contribute to the execution of the program can be optimized away.
13917 This option is enabled by default for the Ada front end, as permitted by
13918 the Ada language specification.
13919 Optimization passes that cause dead exceptions to be removed are enabled independently at different optimization levels.
13921 @item -funwind-tables
13922 @opindex funwind-tables
13923 Similar to @option{-fexceptions}, except that it just generates any needed
13924 static data, but does not affect the generated code in any other way.
13925 You normally do not need to enable this option; instead, a language processor
13926 that needs this handling enables it on your behalf.
13928 @item -fasynchronous-unwind-tables
13929 @opindex fasynchronous-unwind-tables
13930 Generate unwind table in DWARF format, if supported by target machine. The
13931 table is exact at each instruction boundary, so it can be used for stack
13932 unwinding from asynchronous events (such as debugger or garbage collector).
13934 @item -fno-gnu-unique
13935 @opindex fno-gnu-unique
13936 @opindex fgnu-unique
13937 On systems with recent GNU assembler and C library, the C++ compiler
13938 uses the @code{STB_GNU_UNIQUE} binding to make sure that definitions
13939 of template static data members and static local variables in inline
13940 functions are unique even in the presence of @code{RTLD_LOCAL}; this
13941 is necessary to avoid problems with a library used by two different
13942 @code{RTLD_LOCAL} plugins depending on a definition in one of them and
13943 therefore disagreeing with the other one about the binding of the
13944 symbol. But this causes @code{dlclose} to be ignored for affected
13945 DSOs; if your program relies on reinitialization of a DSO via
13946 @code{dlclose} and @code{dlopen}, you can use
13947 @option{-fno-gnu-unique}.
13949 @item -fpcc-struct-return
13950 @opindex fpcc-struct-return
13951 Return ``short'' @code{struct} and @code{union} values in memory like
13952 longer ones, rather than in registers. This convention is less
13953 efficient, but it has the advantage of allowing intercallability between
13954 GCC-compiled files and files compiled with other compilers, particularly
13955 the Portable C Compiler (pcc).
13957 The precise convention for returning structures in memory depends
13958 on the target configuration macros.
13960 Short structures and unions are those whose size and alignment match
13961 that of some integer type.
13963 @strong{Warning:} code compiled with the @option{-fpcc-struct-return}
13964 switch is not binary compatible with code compiled with the
13965 @option{-freg-struct-return} switch.
13966 Use it to conform to a non-default application binary interface.
13968 @item -freg-struct-return
13969 @opindex freg-struct-return
13970 Return @code{struct} and @code{union} values in registers when possible.
13971 This is more efficient for small structures than
13972 @option{-fpcc-struct-return}.
13974 If you specify neither @option{-fpcc-struct-return} nor
13975 @option{-freg-struct-return}, GCC defaults to whichever convention is
13976 standard for the target. If there is no standard convention, GCC
13977 defaults to @option{-fpcc-struct-return}, except on targets where GCC is
13978 the principal compiler. In those cases, we can choose the standard, and
13979 we chose the more efficient register return alternative.
13981 @strong{Warning:} code compiled with the @option{-freg-struct-return}
13982 switch is not binary compatible with code compiled with the
13983 @option{-fpcc-struct-return} switch.
13984 Use it to conform to a non-default application binary interface.
13986 @item -fshort-enums
13987 @opindex fshort-enums
13988 Allocate to an @code{enum} type only as many bytes as it needs for the
13989 declared range of possible values. Specifically, the @code{enum} type
13990 is equivalent to the smallest integer type that has enough room.
13992 @strong{Warning:} the @option{-fshort-enums} switch causes GCC to generate
13993 code that is not binary compatible with code generated without that switch.
13994 Use it to conform to a non-default application binary interface.
13996 @item -fshort-wchar
13997 @opindex fshort-wchar
13998 Override the underlying type for @code{wchar_t} to be @code{short
13999 unsigned int} instead of the default for the target. This option is
14000 useful for building programs to run under WINE@.
14002 @strong{Warning:} the @option{-fshort-wchar} switch causes GCC to generate
14003 code that is not binary compatible with code generated without that switch.
14004 Use it to conform to a non-default application binary interface.
14007 @opindex fno-common
14009 @cindex tentative definitions
14010 In C code, this option controls the placement of global variables
14011 defined without an initializer, known as @dfn{tentative definitions}
14012 in the C standard. Tentative definitions are distinct from declarations
14013 of a variable with the @code{extern} keyword, which do not allocate storage.
14015 Unix C compilers have traditionally allocated storage for
14016 uninitialized global variables in a common block. This allows the
14017 linker to resolve all tentative definitions of the same variable
14018 in different compilation units to the same object, or to a non-tentative
14020 This is the behavior specified by @option{-fcommon}, and is the default for
14021 GCC on most targets.
14022 On the other hand, this behavior is not required by ISO
14023 C, and on some targets may carry a speed or code size penalty on
14024 variable references.
14026 The @option{-fno-common} option specifies that the compiler should instead
14027 place uninitialized global variables in the BSS section of the object file.
14028 This inhibits the merging of tentative definitions by the linker so
14029 you get a multiple-definition error if the same
14030 variable is defined in more than one compilation unit.
14031 Compiling with @option{-fno-common} is useful on targets for which
14032 it provides better performance, or if you wish to verify that the
14033 program will work on other systems that always treat uninitialized
14034 variable definitions this way.
14039 Ignore the @code{#ident} directive.
14041 @item -finhibit-size-directive
14042 @opindex finhibit-size-directive
14043 Don't output a @code{.size} assembler directive, or anything else that
14044 would cause trouble if the function is split in the middle, and the
14045 two halves are placed at locations far apart in memory. This option is
14046 used when compiling @file{crtstuff.c}; you should not need to use it
14049 @item -fverbose-asm
14050 @opindex fverbose-asm
14051 Put extra commentary information in the generated assembly code to
14052 make it more readable. This option is generally only of use to those
14053 who actually need to read the generated assembly code (perhaps while
14054 debugging the compiler itself).
14056 @option{-fno-verbose-asm}, the default, causes the
14057 extra information to be omitted and is useful when comparing two assembler
14060 The added comments include:
14065 information on the compiler version and command-line options,
14068 the source code lines associated with the assembly instructions,
14069 in the form FILENAME:LINENUMBER:CONTENT OF LINE,
14072 hints on which high-level expressions correspond to
14073 the various assembly instruction operands.
14077 For example, given this C source file:
14085 for (i = 0; i < n; i++)
14092 compiling to (x86_64) assembly via @option{-S} and emitting the result
14093 direct to stdout via @option{-o} @option{-}
14096 gcc -S test.c -fverbose-asm -Os -o -
14099 gives output similar to this:
14103 # GNU C11 (GCC) version 7.0.0 20160809 (experimental) (x86_64-pc-linux-gnu)
14110 .type test, @@function
14114 # test.c:4: int total = 0;
14115 xorl %eax, %eax # <retval>
14116 # test.c:6: for (i = 0; i < n; i++)
14117 xorl %edx, %edx # i
14119 # test.c:6: for (i = 0; i < n; i++)
14120 cmpl %edi, %edx # n, i
14122 # test.c:7: total += i * i;
14123 movl %edx, %ecx # i, tmp92
14124 imull %edx, %ecx # i, tmp92
14125 # test.c:6: for (i = 0; i < n; i++)
14127 # test.c:7: total += i * i;
14128 addl %ecx, %eax # tmp92, <retval>
14136 .ident "GCC: (GNU) 7.0.0 20160809 (experimental)"
14137 .section .note.GNU-stack,"",@@progbits
14140 The comments are intended for humans rather than machines and hence the
14141 precise format of the comments is subject to change.
14143 @item -frecord-gcc-switches
14144 @opindex frecord-gcc-switches
14145 This switch causes the command line used to invoke the
14146 compiler to be recorded into the object file that is being created.
14147 This switch is only implemented on some targets and the exact format
14148 of the recording is target and binary file format dependent, but it
14149 usually takes the form of a section containing ASCII text. This
14150 switch is related to the @option{-fverbose-asm} switch, but that
14151 switch only records information in the assembler output file as
14152 comments, so it never reaches the object file.
14153 See also @option{-grecord-gcc-switches} for another
14154 way of storing compiler options into the object file.
14158 @cindex global offset table
14160 Generate position-independent code (PIC) suitable for use in a shared
14161 library, if supported for the target machine. Such code accesses all
14162 constant addresses through a global offset table (GOT)@. The dynamic
14163 loader resolves the GOT entries when the program starts (the dynamic
14164 loader is not part of GCC; it is part of the operating system). If
14165 the GOT size for the linked executable exceeds a machine-specific
14166 maximum size, you get an error message from the linker indicating that
14167 @option{-fpic} does not work; in that case, recompile with @option{-fPIC}
14168 instead. (These maximums are 8k on the SPARC, 28k on AArch64 and 32k
14169 on the m68k and RS/6000. The x86 has no such limit.)
14171 Position-independent code requires special support, and therefore works
14172 only on certain machines. For the x86, GCC supports PIC for System V
14173 but not for the Sun 386i. Code generated for the IBM RS/6000 is always
14174 position-independent.
14176 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
14181 If supported for the target machine, emit position-independent code,
14182 suitable for dynamic linking and avoiding any limit on the size of the
14183 global offset table. This option makes a difference on AArch64, m68k,
14184 PowerPC and SPARC@.
14186 Position-independent code requires special support, and therefore works
14187 only on certain machines.
14189 When this flag is set, the macros @code{__pic__} and @code{__PIC__}
14196 These options are similar to @option{-fpic} and @option{-fPIC}, but the
14197 generated position-independent code can be only linked into executables.
14198 Usually these options are used to compile code that will be linked using
14199 the @option{-pie} GCC option.
14201 @option{-fpie} and @option{-fPIE} both define the macros
14202 @code{__pie__} and @code{__PIE__}. The macros have the value 1
14203 for @option{-fpie} and 2 for @option{-fPIE}.
14208 Do not use the PLT for external function calls in position-independent code.
14209 Instead, load the callee address at call sites from the GOT and branch to it.
14210 This leads to more efficient code by eliminating PLT stubs and exposing
14211 GOT loads to optimizations. On architectures such as 32-bit x86 where
14212 PLT stubs expect the GOT pointer in a specific register, this gives more
14213 register allocation freedom to the compiler.
14214 Lazy binding requires use of the PLT;
14215 with @option{-fno-plt} all external symbols are resolved at load time.
14217 Alternatively, the function attribute @code{noplt} can be used to avoid calls
14218 through the PLT for specific external functions.
14220 In position-dependent code, a few targets also convert calls to
14221 functions that are marked to not use the PLT to use the GOT instead.
14223 @item -fno-jump-tables
14224 @opindex fno-jump-tables
14225 @opindex fjump-tables
14226 Do not use jump tables for switch statements even where it would be
14227 more efficient than other code generation strategies. This option is
14228 of use in conjunction with @option{-fpic} or @option{-fPIC} for
14229 building code that forms part of a dynamic linker and cannot
14230 reference the address of a jump table. On some targets, jump tables
14231 do not require a GOT and this option is not needed.
14233 @item -ffixed-@var{reg}
14235 Treat the register named @var{reg} as a fixed register; generated code
14236 should never refer to it (except perhaps as a stack pointer, frame
14237 pointer or in some other fixed role).
14239 @var{reg} must be the name of a register. The register names accepted
14240 are machine-specific and are defined in the @code{REGISTER_NAMES}
14241 macro in the machine description macro file.
14243 This flag does not have a negative form, because it specifies a
14246 @item -fcall-used-@var{reg}
14247 @opindex fcall-used
14248 Treat the register named @var{reg} as an allocable register that is
14249 clobbered by function calls. It may be allocated for temporaries or
14250 variables that do not live across a call. Functions compiled this way
14251 do not save and restore the register @var{reg}.
14253 It is an error to use this flag with the frame pointer or stack pointer.
14254 Use of this flag for other registers that have fixed pervasive roles in
14255 the machine's execution model produces disastrous results.
14257 This flag does not have a negative form, because it specifies a
14260 @item -fcall-saved-@var{reg}
14261 @opindex fcall-saved
14262 Treat the register named @var{reg} as an allocable register saved by
14263 functions. It may be allocated even for temporaries or variables that
14264 live across a call. Functions compiled this way save and restore
14265 the register @var{reg} if they use it.
14267 It is an error to use this flag with the frame pointer or stack pointer.
14268 Use of this flag for other registers that have fixed pervasive roles in
14269 the machine's execution model produces disastrous results.
14271 A different sort of disaster results from the use of this flag for
14272 a register in which function values may be returned.
14274 This flag does not have a negative form, because it specifies a
14277 @item -fpack-struct[=@var{n}]
14278 @opindex fpack-struct
14279 Without a value specified, pack all structure members together without
14280 holes. When a value is specified (which must be a small power of two), pack
14281 structure members according to this value, representing the maximum
14282 alignment (that is, objects with default alignment requirements larger than
14283 this are output potentially unaligned at the next fitting location.
14285 @strong{Warning:} the @option{-fpack-struct} switch causes GCC to generate
14286 code that is not binary compatible with code generated without that switch.
14287 Additionally, it makes the code suboptimal.
14288 Use it to conform to a non-default application binary interface.
14290 @item -fleading-underscore
14291 @opindex fleading-underscore
14292 This option and its counterpart, @option{-fno-leading-underscore}, forcibly
14293 change the way C symbols are represented in the object file. One use
14294 is to help link with legacy assembly code.
14296 @strong{Warning:} the @option{-fleading-underscore} switch causes GCC to
14297 generate code that is not binary compatible with code generated without that
14298 switch. Use it to conform to a non-default application binary interface.
14299 Not all targets provide complete support for this switch.
14301 @item -ftls-model=@var{model}
14302 @opindex ftls-model
14303 Alter the thread-local storage model to be used (@pxref{Thread-Local}).
14304 The @var{model} argument should be one of @samp{global-dynamic},
14305 @samp{local-dynamic}, @samp{initial-exec} or @samp{local-exec}.
14306 Note that the choice is subject to optimization: the compiler may use
14307 a more efficient model for symbols not visible outside of the translation
14308 unit, or if @option{-fpic} is not given on the command line.
14310 The default without @option{-fpic} is @samp{initial-exec}; with
14311 @option{-fpic} the default is @samp{global-dynamic}.
14313 @item -ftrampolines
14314 @opindex ftrampolines
14315 For targets that normally need trampolines for nested functions, always
14316 generate them instead of using descriptors. Otherwise, for targets that
14317 do not need them, like for example HP-PA or IA-64, do nothing.
14319 A trampoline is a small piece of code that is created at run time on the
14320 stack when the address of a nested function is taken, and is used to call
14321 the nested function indirectly. Therefore, it requires the stack to be
14322 made executable in order for the program to work properly.
14324 @option{-fno-trampolines} is enabled by default on a language by language
14325 basis to let the compiler avoid generating them, if it computes that this
14326 is safe, and replace them with descriptors. Descriptors are made up of data
14327 only, but the generated code must be prepared to deal with them. As of this
14328 writing, @option{-fno-trampolines} is enabled by default only for Ada.
14330 Moreover, code compiled with @option{-ftrampolines} and code compiled with
14331 @option{-fno-trampolines} are not binary compatible if nested functions are
14332 present. This option must therefore be used on a program-wide basis and be
14333 manipulated with extreme care.
14335 @item -fvisibility=@r{[}default@r{|}internal@r{|}hidden@r{|}protected@r{]}
14336 @opindex fvisibility
14337 Set the default ELF image symbol visibility to the specified option---all
14338 symbols are marked with this unless overridden within the code.
14339 Using this feature can very substantially improve linking and
14340 load times of shared object libraries, produce more optimized
14341 code, provide near-perfect API export and prevent symbol clashes.
14342 It is @strong{strongly} recommended that you use this in any shared objects
14345 Despite the nomenclature, @samp{default} always means public; i.e.,
14346 available to be linked against from outside the shared object.
14347 @samp{protected} and @samp{internal} are pretty useless in real-world
14348 usage so the only other commonly used option is @samp{hidden}.
14349 The default if @option{-fvisibility} isn't specified is
14350 @samp{default}, i.e., make every symbol public.
14352 A good explanation of the benefits offered by ensuring ELF
14353 symbols have the correct visibility is given by ``How To Write
14354 Shared Libraries'' by Ulrich Drepper (which can be found at
14355 @w{@uref{https://www.akkadia.org/drepper/}})---however a superior
14356 solution made possible by this option to marking things hidden when
14357 the default is public is to make the default hidden and mark things
14358 public. This is the norm with DLLs on Windows and with @option{-fvisibility=hidden}
14359 and @code{__attribute__ ((visibility("default")))} instead of
14360 @code{__declspec(dllexport)} you get almost identical semantics with
14361 identical syntax. This is a great boon to those working with
14362 cross-platform projects.
14364 For those adding visibility support to existing code, you may find
14365 @code{#pragma GCC visibility} of use. This works by you enclosing
14366 the declarations you wish to set visibility for with (for example)
14367 @code{#pragma GCC visibility push(hidden)} and
14368 @code{#pragma GCC visibility pop}.
14369 Bear in mind that symbol visibility should be viewed @strong{as
14370 part of the API interface contract} and thus all new code should
14371 always specify visibility when it is not the default; i.e., declarations
14372 only for use within the local DSO should @strong{always} be marked explicitly
14373 as hidden as so to avoid PLT indirection overheads---making this
14374 abundantly clear also aids readability and self-documentation of the code.
14375 Note that due to ISO C++ specification requirements, @code{operator new} and
14376 @code{operator delete} must always be of default visibility.
14378 Be aware that headers from outside your project, in particular system
14379 headers and headers from any other library you use, may not be
14380 expecting to be compiled with visibility other than the default. You
14381 may need to explicitly say @code{#pragma GCC visibility push(default)}
14382 before including any such headers.
14384 @code{extern} declarations are not affected by @option{-fvisibility}, so
14385 a lot of code can be recompiled with @option{-fvisibility=hidden} with
14386 no modifications. However, this means that calls to @code{extern}
14387 functions with no explicit visibility use the PLT, so it is more
14388 effective to use @code{__attribute ((visibility))} and/or
14389 @code{#pragma GCC visibility} to tell the compiler which @code{extern}
14390 declarations should be treated as hidden.
14392 Note that @option{-fvisibility} does affect C++ vague linkage
14393 entities. This means that, for instance, an exception class that is
14394 be thrown between DSOs must be explicitly marked with default
14395 visibility so that the @samp{type_info} nodes are unified between
14398 An overview of these techniques, their benefits and how to use them
14399 is at @uref{http://gcc.gnu.org/@/wiki/@/Visibility}.
14401 @item -fstrict-volatile-bitfields
14402 @opindex fstrict-volatile-bitfields
14403 This option should be used if accesses to volatile bit-fields (or other
14404 structure fields, although the compiler usually honors those types
14405 anyway) should use a single access of the width of the
14406 field's type, aligned to a natural alignment if possible. For
14407 example, targets with memory-mapped peripheral registers might require
14408 all such accesses to be 16 bits wide; with this flag you can
14409 declare all peripheral bit-fields as @code{unsigned short} (assuming short
14410 is 16 bits on these targets) to force GCC to use 16-bit accesses
14411 instead of, perhaps, a more efficient 32-bit access.
14413 If this option is disabled, the compiler uses the most efficient
14414 instruction. In the previous example, that might be a 32-bit load
14415 instruction, even though that accesses bytes that do not contain
14416 any portion of the bit-field, or memory-mapped registers unrelated to
14417 the one being updated.
14419 In some cases, such as when the @code{packed} attribute is applied to a
14420 structure field, it may not be possible to access the field with a single
14421 read or write that is correctly aligned for the target machine. In this
14422 case GCC falls back to generating multiple accesses rather than code that
14423 will fault or truncate the result at run time.
14425 Note: Due to restrictions of the C/C++11 memory model, write accesses are
14426 not allowed to touch non bit-field members. It is therefore recommended
14427 to define all bits of the field's type as bit-field members.
14429 The default value of this option is determined by the application binary
14430 interface for the target processor.
14432 @item -fsync-libcalls
14433 @opindex fsync-libcalls
14434 This option controls whether any out-of-line instance of the @code{__sync}
14435 family of functions may be used to implement the C++11 @code{__atomic}
14436 family of functions.
14438 The default value of this option is enabled, thus the only useful form
14439 of the option is @option{-fno-sync-libcalls}. This option is used in
14440 the implementation of the @file{libatomic} runtime library.
14444 @node Developer Options
14445 @section GCC Developer Options
14446 @cindex developer options
14447 @cindex debugging GCC
14448 @cindex debug dump options
14449 @cindex dump options
14450 @cindex compilation statistics
14452 This section describes command-line options that are primarily of
14453 interest to GCC developers, including options to support compiler
14454 testing and investigation of compiler bugs and compile-time
14455 performance problems. This includes options that produce debug dumps
14456 at various points in the compilation; that print statistics such as
14457 memory use and execution time; and that print information about GCC's
14458 configuration, such as where it searches for libraries. You should
14459 rarely need to use any of these options for ordinary compilation and
14462 Many developer options that cause GCC to dump output to a file take an
14463 optional @samp{=@var{filename}} suffix. You can specify @samp{stdout}
14464 or @samp{-} to dump to standard output, and @samp{stderr} for standard
14467 If @samp{=@var{filename}} is omitted, a default dump file name is
14468 constructed by concatenating the base dump file name, a pass number,
14469 phase letter, and pass name. The base dump file name is the name of
14470 output file produced by the compiler if explicitly specified and not
14471 an executable; otherwise it is the source file name.
14472 The pass number is determined by the order passes are registered with
14473 the compiler's pass manager.
14474 This is generally the same as the order of execution, but passes
14475 registered by plugins, target-specific passes, or passes that are
14476 otherwise registered late are numbered higher than the pass named
14477 @samp{final}, even if they are executed earlier. The phase letter is
14478 one of @samp{i} (inter-procedural analysis), @samp{l}
14479 (language-specific), @samp{r} (RTL), or @samp{t} (tree).
14480 The files are created in the directory of the output file.
14484 @item -d@var{letters}
14485 @itemx -fdump-rtl-@var{pass}
14486 @itemx -fdump-rtl-@var{pass}=@var{filename}
14488 @opindex fdump-rtl-@var{pass}
14489 Says to make debugging dumps during compilation at times specified by
14490 @var{letters}. This is used for debugging the RTL-based passes of the
14493 Some @option{-d@var{letters}} switches have different meaning when
14494 @option{-E} is used for preprocessing. @xref{Preprocessor Options},
14495 for information about preprocessor-specific dump options.
14497 Debug dumps can be enabled with a @option{-fdump-rtl} switch or some
14498 @option{-d} option @var{letters}. Here are the possible
14499 letters for use in @var{pass} and @var{letters}, and their meanings:
14503 @item -fdump-rtl-alignments
14504 @opindex fdump-rtl-alignments
14505 Dump after branch alignments have been computed.
14507 @item -fdump-rtl-asmcons
14508 @opindex fdump-rtl-asmcons
14509 Dump after fixing rtl statements that have unsatisfied in/out constraints.
14511 @item -fdump-rtl-auto_inc_dec
14512 @opindex fdump-rtl-auto_inc_dec
14513 Dump after auto-inc-dec discovery. This pass is only run on
14514 architectures that have auto inc or auto dec instructions.
14516 @item -fdump-rtl-barriers
14517 @opindex fdump-rtl-barriers
14518 Dump after cleaning up the barrier instructions.
14520 @item -fdump-rtl-bbpart
14521 @opindex fdump-rtl-bbpart
14522 Dump after partitioning hot and cold basic blocks.
14524 @item -fdump-rtl-bbro
14525 @opindex fdump-rtl-bbro
14526 Dump after block reordering.
14528 @item -fdump-rtl-btl1
14529 @itemx -fdump-rtl-btl2
14530 @opindex fdump-rtl-btl2
14531 @opindex fdump-rtl-btl2
14532 @option{-fdump-rtl-btl1} and @option{-fdump-rtl-btl2} enable dumping
14533 after the two branch
14534 target load optimization passes.
14536 @item -fdump-rtl-bypass
14537 @opindex fdump-rtl-bypass
14538 Dump after jump bypassing and control flow optimizations.
14540 @item -fdump-rtl-combine
14541 @opindex fdump-rtl-combine
14542 Dump after the RTL instruction combination pass.
14544 @item -fdump-rtl-compgotos
14545 @opindex fdump-rtl-compgotos
14546 Dump after duplicating the computed gotos.
14548 @item -fdump-rtl-ce1
14549 @itemx -fdump-rtl-ce2
14550 @itemx -fdump-rtl-ce3
14551 @opindex fdump-rtl-ce1
14552 @opindex fdump-rtl-ce2
14553 @opindex fdump-rtl-ce3
14554 @option{-fdump-rtl-ce1}, @option{-fdump-rtl-ce2}, and
14555 @option{-fdump-rtl-ce3} enable dumping after the three
14556 if conversion passes.
14558 @item -fdump-rtl-cprop_hardreg
14559 @opindex fdump-rtl-cprop_hardreg
14560 Dump after hard register copy propagation.
14562 @item -fdump-rtl-csa
14563 @opindex fdump-rtl-csa
14564 Dump after combining stack adjustments.
14566 @item -fdump-rtl-cse1
14567 @itemx -fdump-rtl-cse2
14568 @opindex fdump-rtl-cse1
14569 @opindex fdump-rtl-cse2
14570 @option{-fdump-rtl-cse1} and @option{-fdump-rtl-cse2} enable dumping after
14571 the two common subexpression elimination passes.
14573 @item -fdump-rtl-dce
14574 @opindex fdump-rtl-dce
14575 Dump after the standalone dead code elimination passes.
14577 @item -fdump-rtl-dbr
14578 @opindex fdump-rtl-dbr
14579 Dump after delayed branch scheduling.
14581 @item -fdump-rtl-dce1
14582 @itemx -fdump-rtl-dce2
14583 @opindex fdump-rtl-dce1
14584 @opindex fdump-rtl-dce2
14585 @option{-fdump-rtl-dce1} and @option{-fdump-rtl-dce2} enable dumping after
14586 the two dead store elimination passes.
14588 @item -fdump-rtl-eh
14589 @opindex fdump-rtl-eh
14590 Dump after finalization of EH handling code.
14592 @item -fdump-rtl-eh_ranges
14593 @opindex fdump-rtl-eh_ranges
14594 Dump after conversion of EH handling range regions.
14596 @item -fdump-rtl-expand
14597 @opindex fdump-rtl-expand
14598 Dump after RTL generation.
14600 @item -fdump-rtl-fwprop1
14601 @itemx -fdump-rtl-fwprop2
14602 @opindex fdump-rtl-fwprop1
14603 @opindex fdump-rtl-fwprop2
14604 @option{-fdump-rtl-fwprop1} and @option{-fdump-rtl-fwprop2} enable
14605 dumping after the two forward propagation passes.
14607 @item -fdump-rtl-gcse1
14608 @itemx -fdump-rtl-gcse2
14609 @opindex fdump-rtl-gcse1
14610 @opindex fdump-rtl-gcse2
14611 @option{-fdump-rtl-gcse1} and @option{-fdump-rtl-gcse2} enable dumping
14612 after global common subexpression elimination.
14614 @item -fdump-rtl-init-regs
14615 @opindex fdump-rtl-init-regs
14616 Dump after the initialization of the registers.
14618 @item -fdump-rtl-initvals
14619 @opindex fdump-rtl-initvals
14620 Dump after the computation of the initial value sets.
14622 @item -fdump-rtl-into_cfglayout
14623 @opindex fdump-rtl-into_cfglayout
14624 Dump after converting to cfglayout mode.
14626 @item -fdump-rtl-ira
14627 @opindex fdump-rtl-ira
14628 Dump after iterated register allocation.
14630 @item -fdump-rtl-jump
14631 @opindex fdump-rtl-jump
14632 Dump after the second jump optimization.
14634 @item -fdump-rtl-loop2
14635 @opindex fdump-rtl-loop2
14636 @option{-fdump-rtl-loop2} enables dumping after the rtl
14637 loop optimization passes.
14639 @item -fdump-rtl-mach
14640 @opindex fdump-rtl-mach
14641 Dump after performing the machine dependent reorganization pass, if that
14644 @item -fdump-rtl-mode_sw
14645 @opindex fdump-rtl-mode_sw
14646 Dump after removing redundant mode switches.
14648 @item -fdump-rtl-rnreg
14649 @opindex fdump-rtl-rnreg
14650 Dump after register renumbering.
14652 @item -fdump-rtl-outof_cfglayout
14653 @opindex fdump-rtl-outof_cfglayout
14654 Dump after converting from cfglayout mode.
14656 @item -fdump-rtl-peephole2
14657 @opindex fdump-rtl-peephole2
14658 Dump after the peephole pass.
14660 @item -fdump-rtl-postreload
14661 @opindex fdump-rtl-postreload
14662 Dump after post-reload optimizations.
14664 @item -fdump-rtl-pro_and_epilogue
14665 @opindex fdump-rtl-pro_and_epilogue
14666 Dump after generating the function prologues and epilogues.
14668 @item -fdump-rtl-sched1
14669 @itemx -fdump-rtl-sched2
14670 @opindex fdump-rtl-sched1
14671 @opindex fdump-rtl-sched2
14672 @option{-fdump-rtl-sched1} and @option{-fdump-rtl-sched2} enable dumping
14673 after the basic block scheduling passes.
14675 @item -fdump-rtl-ree
14676 @opindex fdump-rtl-ree
14677 Dump after sign/zero extension elimination.
14679 @item -fdump-rtl-seqabstr
14680 @opindex fdump-rtl-seqabstr
14681 Dump after common sequence discovery.
14683 @item -fdump-rtl-shorten
14684 @opindex fdump-rtl-shorten
14685 Dump after shortening branches.
14687 @item -fdump-rtl-sibling
14688 @opindex fdump-rtl-sibling
14689 Dump after sibling call optimizations.
14691 @item -fdump-rtl-split1
14692 @itemx -fdump-rtl-split2
14693 @itemx -fdump-rtl-split3
14694 @itemx -fdump-rtl-split4
14695 @itemx -fdump-rtl-split5
14696 @opindex fdump-rtl-split1
14697 @opindex fdump-rtl-split2
14698 @opindex fdump-rtl-split3
14699 @opindex fdump-rtl-split4
14700 @opindex fdump-rtl-split5
14701 These options enable dumping after five rounds of
14702 instruction splitting.
14704 @item -fdump-rtl-sms
14705 @opindex fdump-rtl-sms
14706 Dump after modulo scheduling. This pass is only run on some
14709 @item -fdump-rtl-stack
14710 @opindex fdump-rtl-stack
14711 Dump after conversion from GCC's ``flat register file'' registers to the
14712 x87's stack-like registers. This pass is only run on x86 variants.
14714 @item -fdump-rtl-subreg1
14715 @itemx -fdump-rtl-subreg2
14716 @opindex fdump-rtl-subreg1
14717 @opindex fdump-rtl-subreg2
14718 @option{-fdump-rtl-subreg1} and @option{-fdump-rtl-subreg2} enable dumping after
14719 the two subreg expansion passes.
14721 @item -fdump-rtl-unshare
14722 @opindex fdump-rtl-unshare
14723 Dump after all rtl has been unshared.
14725 @item -fdump-rtl-vartrack
14726 @opindex fdump-rtl-vartrack
14727 Dump after variable tracking.
14729 @item -fdump-rtl-vregs
14730 @opindex fdump-rtl-vregs
14731 Dump after converting virtual registers to hard registers.
14733 @item -fdump-rtl-web
14734 @opindex fdump-rtl-web
14735 Dump after live range splitting.
14737 @item -fdump-rtl-regclass
14738 @itemx -fdump-rtl-subregs_of_mode_init
14739 @itemx -fdump-rtl-subregs_of_mode_finish
14740 @itemx -fdump-rtl-dfinit
14741 @itemx -fdump-rtl-dfinish
14742 @opindex fdump-rtl-regclass
14743 @opindex fdump-rtl-subregs_of_mode_init
14744 @opindex fdump-rtl-subregs_of_mode_finish
14745 @opindex fdump-rtl-dfinit
14746 @opindex fdump-rtl-dfinish
14747 These dumps are defined but always produce empty files.
14750 @itemx -fdump-rtl-all
14752 @opindex fdump-rtl-all
14753 Produce all the dumps listed above.
14757 Annotate the assembler output with miscellaneous debugging information.
14761 Dump all macro definitions, at the end of preprocessing, in addition to
14766 Produce a core dump whenever an error occurs.
14770 Annotate the assembler output with a comment indicating which
14771 pattern and alternative is used. The length and cost of each instruction are
14776 Dump the RTL in the assembler output as a comment before each instruction.
14777 Also turns on @option{-dp} annotation.
14781 Just generate RTL for a function instead of compiling it. Usually used
14782 with @option{-fdump-rtl-expand}.
14786 @opindex fdump-debug
14787 Dump debugging information generated during the debug
14790 @item -fdump-earlydebug
14791 @opindex fdump-earlydebug
14792 Dump debugging information generated during the early debug
14795 @item -fdump-noaddr
14796 @opindex fdump-noaddr
14797 When doing debugging dumps, suppress address output. This makes it more
14798 feasible to use diff on debugging dumps for compiler invocations with
14799 different compiler binaries and/or different
14800 text / bss / data / heap / stack / dso start locations.
14803 @opindex freport-bug
14804 Collect and dump debug information into a temporary file if an
14805 internal compiler error (ICE) occurs.
14807 @item -fdump-unnumbered
14808 @opindex fdump-unnumbered
14809 When doing debugging dumps, suppress instruction numbers and address output.
14810 This makes it more feasible to use diff on debugging dumps for compiler
14811 invocations with different options, in particular with and without
14814 @item -fdump-unnumbered-links
14815 @opindex fdump-unnumbered-links
14816 When doing debugging dumps (see @option{-d} option above), suppress
14817 instruction numbers for the links to the previous and next instructions
14820 @item -fdump-ipa-@var{switch}
14821 @itemx -fdump-ipa-@var{switch}-@var{options}
14823 Control the dumping at various stages of inter-procedural analysis
14824 language tree to a file. The file name is generated by appending a
14825 switch specific suffix to the source file name, and the file is created
14826 in the same directory as the output file. The following dumps are
14831 Enables all inter-procedural analysis dumps.
14834 Dumps information about call-graph optimization, unused function removal,
14835 and inlining decisions.
14838 Dump after function inlining.
14842 Additionally, the options @option{-optimized}, @option{-missed},
14843 @option{-note}, and @option{-all} can be provided, with the same meaning
14844 as for @option{-fopt-info}, defaulting to @option{-optimized}.
14846 For example, @option{-fdump-ipa-inline-optimized-missed} will emit
14847 information on callsites that were inlined, along with callsites
14848 that were not inlined.
14850 By default, the dump will contain messages about successful
14851 optimizations (equivalent to @option{-optimized}) together with
14852 low-level details about the analysis.
14854 @item -fdump-lang-all
14855 @itemx -fdump-lang-@var{switch}
14856 @itemx -fdump-lang-@var{switch}-@var{options}
14857 @itemx -fdump-lang-@var{switch}-@var{options}=@var{filename}
14858 @opindex fdump-lang-all
14859 @opindex fdump-lang
14860 Control the dumping of language-specific information. The @var{options}
14861 and @var{filename} portions behave as described in the
14862 @option{-fdump-tree} option. The following @var{switch} values are
14868 Enable all language-specific dumps.
14871 Dump class hierarchy information. Virtual table information is emitted
14872 unless '@option{slim}' is specified. This option is applicable to C++ only.
14875 Dump the raw internal tree data. This option is applicable to C++ only.
14879 @item -fdump-passes
14880 @opindex fdump-passes
14881 Print on @file{stderr} the list of optimization passes that are turned
14882 on and off by the current command-line options.
14884 @item -fdump-statistics-@var{option}
14885 @opindex fdump-statistics
14886 Enable and control dumping of pass statistics in a separate file. The
14887 file name is generated by appending a suffix ending in
14888 @samp{.statistics} to the source file name, and the file is created in
14889 the same directory as the output file. If the @samp{-@var{option}}
14890 form is used, @samp{-stats} causes counters to be summed over the
14891 whole compilation unit while @samp{-details} dumps every event as
14892 the passes generate them. The default with no option is to sum
14893 counters for each function compiled.
14895 @item -fdump-tree-all
14896 @itemx -fdump-tree-@var{switch}
14897 @itemx -fdump-tree-@var{switch}-@var{options}
14898 @itemx -fdump-tree-@var{switch}-@var{options}=@var{filename}
14899 @opindex fdump-tree-all
14900 @opindex fdump-tree
14901 Control the dumping at various stages of processing the intermediate
14902 language tree to a file. If the @samp{-@var{options}}
14903 form is used, @var{options} is a list of @samp{-} separated options
14904 which control the details of the dump. Not all options are applicable
14905 to all dumps; those that are not meaningful are ignored. The
14906 following options are available
14910 Print the address of each node. Usually this is not meaningful as it
14911 changes according to the environment and source file. Its primary use
14912 is for tying up a dump file with a debug environment.
14914 If @code{DECL_ASSEMBLER_NAME} has been set for a given decl, use that
14915 in the dump instead of @code{DECL_NAME}. Its primary use is ease of
14916 use working backward from mangled names in the assembly file.
14918 When dumping front-end intermediate representations, inhibit dumping
14919 of members of a scope or body of a function merely because that scope
14920 has been reached. Only dump such items when they are directly reachable
14921 by some other path.
14923 When dumping pretty-printed trees, this option inhibits dumping the
14924 bodies of control structures.
14926 When dumping RTL, print the RTL in slim (condensed) form instead of
14927 the default LISP-like representation.
14929 Print a raw representation of the tree. By default, trees are
14930 pretty-printed into a C-like representation.
14932 Enable more detailed dumps (not honored by every dump option). Also
14933 include information from the optimization passes.
14935 Enable dumping various statistics about the pass (not honored by every dump
14938 Enable showing basic block boundaries (disabled in raw dumps).
14940 For each of the other indicated dump files (@option{-fdump-rtl-@var{pass}}),
14941 dump a representation of the control flow graph suitable for viewing with
14942 GraphViz to @file{@var{file}.@var{passid}.@var{pass}.dot}. Each function in
14943 the file is pretty-printed as a subgraph, so that GraphViz can render them
14944 all in a single plot.
14946 This option currently only works for RTL dumps, and the RTL is always
14947 dumped in slim form.
14949 Enable showing virtual operands for every statement.
14951 Enable showing line numbers for statements.
14953 Enable showing the unique ID (@code{DECL_UID}) for each variable.
14955 Enable showing the tree dump for each statement.
14957 Enable showing the EH region number holding each statement.
14959 Enable showing scalar evolution analysis details.
14961 Enable showing optimization information (only available in certain
14964 Enable showing missed optimization information (only available in certain
14967 Enable other detailed optimization information (only available in
14970 Turn on all options, except @option{raw}, @option{slim}, @option{verbose}
14971 and @option{lineno}.
14973 Turn on all optimization options, i.e., @option{optimized},
14974 @option{missed}, and @option{note}.
14977 To determine what tree dumps are available or find the dump for a pass
14978 of interest follow the steps below.
14982 Invoke GCC with @option{-fdump-passes} and in the @file{stderr} output
14983 look for a code that corresponds to the pass you are interested in.
14984 For example, the codes @code{tree-evrp}, @code{tree-vrp1}, and
14985 @code{tree-vrp2} correspond to the three Value Range Propagation passes.
14986 The number at the end distinguishes distinct invocations of the same pass.
14988 To enable the creation of the dump file, append the pass code to
14989 the @option{-fdump-} option prefix and invoke GCC with it. For example,
14990 to enable the dump from the Early Value Range Propagation pass, invoke
14991 GCC with the @option{-fdump-tree-evrp} option. Optionally, you may
14992 specify the name of the dump file. If you don't specify one, GCC
14993 creates as described below.
14995 Find the pass dump in a file whose name is composed of three components
14996 separated by a period: the name of the source file GCC was invoked to
14997 compile, a numeric suffix indicating the pass number followed by the
14998 letter @samp{t} for tree passes (and the letter @samp{r} for RTL passes),
14999 and finally the pass code. For example, the Early VRP pass dump might
15000 be in a file named @file{myfile.c.038t.evrp} in the current working
15001 directory. Note that the numeric codes are not stable and may change
15002 from one version of GCC to another.
15006 @itemx -fopt-info-@var{options}
15007 @itemx -fopt-info-@var{options}=@var{filename}
15009 Controls optimization dumps from various optimization passes. If the
15010 @samp{-@var{options}} form is used, @var{options} is a list of
15011 @samp{-} separated option keywords to select the dump details and
15014 The @var{options} can be divided into three groups:
15017 options describing what kinds of messages should be emitted,
15019 options describing the verbosity of the dump, and
15021 options describing which optimizations should be included.
15023 The options from each group can be freely mixed as they are
15024 non-overlapping. However, in case of any conflicts,
15025 the later options override the earlier options on the command
15028 The following options control which kinds of messages should be emitted:
15032 Print information when an optimization is successfully applied. It is
15033 up to a pass to decide which information is relevant. For example, the
15034 vectorizer passes print the source location of loops which are
15035 successfully vectorized.
15037 Print information about missed optimizations. Individual passes
15038 control which information to include in the output.
15040 Print verbose information about optimizations, such as certain
15041 transformations, more detailed messages about decisions etc.
15043 Print detailed optimization information. This includes
15044 @samp{optimized}, @samp{missed}, and @samp{note}.
15047 The following option controls the dump verbosity:
15051 By default, only ``high-level'' messages are emitted. This option enables
15052 additional, more detailed, messages, which are likely to only be of interest
15056 One or more of the following option keywords can be used to describe a
15057 group of optimizations:
15061 Enable dumps from all interprocedural optimizations.
15063 Enable dumps from all loop optimizations.
15065 Enable dumps from all inlining optimizations.
15067 Enable dumps from all OMP (Offloading and Multi Processing) optimizations.
15069 Enable dumps from all vectorization optimizations.
15071 Enable dumps from all optimizations. This is a superset of
15072 the optimization groups listed above.
15075 If @var{options} is
15076 omitted, it defaults to @samp{optimized-optall}, which means to dump messages
15077 about successful optimizations from all the passes, omitting messages
15078 that are treated as ``internals''.
15080 If the @var{filename} is provided, then the dumps from all the
15081 applicable optimizations are concatenated into the @var{filename}.
15082 Otherwise the dump is output onto @file{stderr}. Though multiple
15083 @option{-fopt-info} options are accepted, only one of them can include
15084 a @var{filename}. If other filenames are provided then all but the
15085 first such option are ignored.
15087 Note that the output @var{filename} is overwritten
15088 in case of multiple translation units. If a combined output from
15089 multiple translation units is desired, @file{stderr} should be used
15092 In the following example, the optimization info is output to
15101 gcc -O3 -fopt-info-missed=missed.all
15105 outputs missed optimization report from all the passes into
15106 @file{missed.all}, and this one:
15109 gcc -O2 -ftree-vectorize -fopt-info-vec-missed
15113 prints information about missed optimization opportunities from
15114 vectorization passes on @file{stderr}.
15115 Note that @option{-fopt-info-vec-missed} is equivalent to
15116 @option{-fopt-info-missed-vec}. The order of the optimization group
15117 names and message types listed after @option{-fopt-info} does not matter.
15119 As another example,
15121 gcc -O3 -fopt-info-inline-optimized-missed=inline.txt
15125 outputs information about missed optimizations as well as
15126 optimized locations from all the inlining passes into
15132 gcc -fopt-info-vec-missed=vec.miss -fopt-info-loop-optimized=loop.opt
15136 Here the two output filenames @file{vec.miss} and @file{loop.opt} are
15137 in conflict since only one output file is allowed. In this case, only
15138 the first option takes effect and the subsequent options are
15139 ignored. Thus only @file{vec.miss} is produced which contains
15140 dumps from the vectorizer about missed opportunities.
15142 @item -fsave-optimization-record
15143 @opindex fsave-optimization-record
15144 Write a SRCFILE.opt-record.json.gz file detailing what optimizations
15145 were performed, for those optimizations that support @option{-fopt-info}.
15147 This option is experimental and the format of the data within the
15148 compressed JSON file is subject to change.
15150 It is roughly equivalent to a machine-readable version of
15151 @option{-fopt-info-all}, as a collection of messages with source file,
15152 line number and column number, with the following additional data for
15158 the execution count of the code being optimized, along with metadata about
15159 whether this was from actual profile data, or just an estimate, allowing
15160 consumers to prioritize messages by code hotness,
15163 the function name of the code being optimized, where applicable,
15166 the ``inlining chain'' for the code being optimized, so that when
15167 a function is inlined into several different places (which might
15168 themselves be inlined), the reader can distinguish between the copies,
15171 objects identifying those parts of the message that refer to expressions,
15172 statements or symbol-table nodes, which of these categories they are, and,
15173 when available, their source code location,
15176 the GCC pass that emitted the message, and
15179 the location in GCC's own code from which the message was emitted
15183 Additionally, some messages are logically nested within other
15184 messages, reflecting implementation details of the optimization
15187 @item -fsched-verbose=@var{n}
15188 @opindex fsched-verbose
15189 On targets that use instruction scheduling, this option controls the
15190 amount of debugging output the scheduler prints to the dump files.
15192 For @var{n} greater than zero, @option{-fsched-verbose} outputs the
15193 same information as @option{-fdump-rtl-sched1} and @option{-fdump-rtl-sched2}.
15194 For @var{n} greater than one, it also output basic block probabilities,
15195 detailed ready list information and unit/insn info. For @var{n} greater
15196 than two, it includes RTL at abort point, control-flow and regions info.
15197 And for @var{n} over four, @option{-fsched-verbose} also includes
15202 @item -fenable-@var{kind}-@var{pass}
15203 @itemx -fdisable-@var{kind}-@var{pass}=@var{range-list}
15207 This is a set of options that are used to explicitly disable/enable
15208 optimization passes. These options are intended for use for debugging GCC.
15209 Compiler users should use regular options for enabling/disabling
15214 @item -fdisable-ipa-@var{pass}
15215 Disable IPA pass @var{pass}. @var{pass} is the pass name. If the same pass is
15216 statically invoked in the compiler multiple times, the pass name should be
15217 appended with a sequential number starting from 1.
15219 @item -fdisable-rtl-@var{pass}
15220 @itemx -fdisable-rtl-@var{pass}=@var{range-list}
15221 Disable RTL pass @var{pass}. @var{pass} is the pass name. If the same pass is
15222 statically invoked in the compiler multiple times, the pass name should be
15223 appended with a sequential number starting from 1. @var{range-list} is a
15224 comma-separated list of function ranges or assembler names. Each range is a number
15225 pair separated by a colon. The range is inclusive in both ends. If the range
15226 is trivial, the number pair can be simplified as a single number. If the
15227 function's call graph node's @var{uid} falls within one of the specified ranges,
15228 the @var{pass} is disabled for that function. The @var{uid} is shown in the
15229 function header of a dump file, and the pass names can be dumped by using
15230 option @option{-fdump-passes}.
15232 @item -fdisable-tree-@var{pass}
15233 @itemx -fdisable-tree-@var{pass}=@var{range-list}
15234 Disable tree pass @var{pass}. See @option{-fdisable-rtl} for the description of
15237 @item -fenable-ipa-@var{pass}
15238 Enable IPA pass @var{pass}. @var{pass} is the pass name. If the same pass is
15239 statically invoked in the compiler multiple times, the pass name should be
15240 appended with a sequential number starting from 1.
15242 @item -fenable-rtl-@var{pass}
15243 @itemx -fenable-rtl-@var{pass}=@var{range-list}
15244 Enable RTL pass @var{pass}. See @option{-fdisable-rtl} for option argument
15245 description and examples.
15247 @item -fenable-tree-@var{pass}
15248 @itemx -fenable-tree-@var{pass}=@var{range-list}
15249 Enable tree pass @var{pass}. See @option{-fdisable-rtl} for the description
15250 of option arguments.
15254 Here are some examples showing uses of these options.
15258 # disable ccp1 for all functions
15259 -fdisable-tree-ccp1
15260 # disable complete unroll for function whose cgraph node uid is 1
15261 -fenable-tree-cunroll=1
15262 # disable gcse2 for functions at the following ranges [1,1],
15263 # [300,400], and [400,1000]
15264 # disable gcse2 for functions foo and foo2
15265 -fdisable-rtl-gcse2=foo,foo2
15266 # disable early inlining
15267 -fdisable-tree-einline
15268 # disable ipa inlining
15269 -fdisable-ipa-inline
15270 # enable tree full unroll
15271 -fenable-tree-unroll
15276 @itemx -fchecking=@var{n}
15278 @opindex fno-checking
15279 Enable internal consistency checking. The default depends on
15280 the compiler configuration. @option{-fchecking=2} enables further
15281 internal consistency checking that might affect code generation.
15283 @item -frandom-seed=@var{string}
15284 @opindex frandom-seed
15285 This option provides a seed that GCC uses in place of
15286 random numbers in generating certain symbol names
15287 that have to be different in every compiled file. It is also used to
15288 place unique stamps in coverage data files and the object files that
15289 produce them. You can use the @option{-frandom-seed} option to produce
15290 reproducibly identical object files.
15292 The @var{string} can either be a number (decimal, octal or hex) or an
15293 arbitrary string (in which case it's converted to a number by
15296 The @var{string} should be different for every file you compile.
15299 @itemx -save-temps=cwd
15300 @opindex save-temps
15301 Store the usual ``temporary'' intermediate files permanently; place them
15302 in the current directory and name them based on the source file. Thus,
15303 compiling @file{foo.c} with @option{-c -save-temps} produces files
15304 @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a
15305 preprocessed @file{foo.i} output file even though the compiler now
15306 normally uses an integrated preprocessor.
15308 When used in combination with the @option{-x} command-line option,
15309 @option{-save-temps} is sensible enough to avoid over writing an
15310 input source file with the same extension as an intermediate file.
15311 The corresponding intermediate file may be obtained by renaming the
15312 source file before using @option{-save-temps}.
15314 If you invoke GCC in parallel, compiling several different source
15315 files that share a common base name in different subdirectories or the
15316 same source file compiled for multiple output destinations, it is
15317 likely that the different parallel compilers will interfere with each
15318 other, and overwrite the temporary files. For instance:
15321 gcc -save-temps -o outdir1/foo.o indir1/foo.c&
15322 gcc -save-temps -o outdir2/foo.o indir2/foo.c&
15325 may result in @file{foo.i} and @file{foo.o} being written to
15326 simultaneously by both compilers.
15328 @item -save-temps=obj
15329 @opindex save-temps=obj
15330 Store the usual ``temporary'' intermediate files permanently. If the
15331 @option{-o} option is used, the temporary files are based on the
15332 object file. If the @option{-o} option is not used, the
15333 @option{-save-temps=obj} switch behaves like @option{-save-temps}.
15338 gcc -save-temps=obj -c foo.c
15339 gcc -save-temps=obj -c bar.c -o dir/xbar.o
15340 gcc -save-temps=obj foobar.c -o dir2/yfoobar
15344 creates @file{foo.i}, @file{foo.s}, @file{dir/xbar.i},
15345 @file{dir/xbar.s}, @file{dir2/yfoobar.i}, @file{dir2/yfoobar.s}, and
15346 @file{dir2/yfoobar.o}.
15348 @item -time@r{[}=@var{file}@r{]}
15350 Report the CPU time taken by each subprocess in the compilation
15351 sequence. For C source files, this is the compiler proper and assembler
15352 (plus the linker if linking is done).
15354 Without the specification of an output file, the output looks like this:
15361 The first number on each line is the ``user time'', that is time spent
15362 executing the program itself. The second number is ``system time'',
15363 time spent executing operating system routines on behalf of the program.
15364 Both numbers are in seconds.
15366 With the specification of an output file, the output is appended to the
15367 named file, and it looks like this:
15370 0.12 0.01 cc1 @var{options}
15371 0.00 0.01 as @var{options}
15374 The ``user time'' and the ``system time'' are moved before the program
15375 name, and the options passed to the program are displayed, so that one
15376 can later tell what file was being compiled, and with which options.
15378 @item -fdump-final-insns@r{[}=@var{file}@r{]}
15379 @opindex fdump-final-insns
15380 Dump the final internal representation (RTL) to @var{file}. If the
15381 optional argument is omitted (or if @var{file} is @code{.}), the name
15382 of the dump file is determined by appending @code{.gkd} to the
15383 compilation output file name.
15385 @item -fcompare-debug@r{[}=@var{opts}@r{]}
15386 @opindex fcompare-debug
15387 @opindex fno-compare-debug
15388 If no error occurs during compilation, run the compiler a second time,
15389 adding @var{opts} and @option{-fcompare-debug-second} to the arguments
15390 passed to the second compilation. Dump the final internal
15391 representation in both compilations, and print an error if they differ.
15393 If the equal sign is omitted, the default @option{-gtoggle} is used.
15395 The environment variable @env{GCC_COMPARE_DEBUG}, if defined, non-empty
15396 and nonzero, implicitly enables @option{-fcompare-debug}. If
15397 @env{GCC_COMPARE_DEBUG} is defined to a string starting with a dash,
15398 then it is used for @var{opts}, otherwise the default @option{-gtoggle}
15401 @option{-fcompare-debug=}, with the equal sign but without @var{opts},
15402 is equivalent to @option{-fno-compare-debug}, which disables the dumping
15403 of the final representation and the second compilation, preventing even
15404 @env{GCC_COMPARE_DEBUG} from taking effect.
15406 To verify full coverage during @option{-fcompare-debug} testing, set
15407 @env{GCC_COMPARE_DEBUG} to say @option{-fcompare-debug-not-overridden},
15408 which GCC rejects as an invalid option in any actual compilation
15409 (rather than preprocessing, assembly or linking). To get just a
15410 warning, setting @env{GCC_COMPARE_DEBUG} to @samp{-w%n-fcompare-debug
15411 not overridden} will do.
15413 @item -fcompare-debug-second
15414 @opindex fcompare-debug-second
15415 This option is implicitly passed to the compiler for the second
15416 compilation requested by @option{-fcompare-debug}, along with options to
15417 silence warnings, and omitting other options that would cause the compiler
15418 to produce output to files or to standard output as a side effect. Dump
15419 files and preserved temporary files are renamed so as to contain the
15420 @code{.gk} additional extension during the second compilation, to avoid
15421 overwriting those generated by the first.
15423 When this option is passed to the compiler driver, it causes the
15424 @emph{first} compilation to be skipped, which makes it useful for little
15425 other than debugging the compiler proper.
15429 Turn off generation of debug info, if leaving out this option
15430 generates it, or turn it on at level 2 otherwise. The position of this
15431 argument in the command line does not matter; it takes effect after all
15432 other options are processed, and it does so only once, no matter how
15433 many times it is given. This is mainly intended to be used with
15434 @option{-fcompare-debug}.
15436 @item -fvar-tracking-assignments-toggle
15437 @opindex fvar-tracking-assignments-toggle
15438 @opindex fno-var-tracking-assignments-toggle
15439 Toggle @option{-fvar-tracking-assignments}, in the same way that
15440 @option{-gtoggle} toggles @option{-g}.
15444 Makes the compiler print out each function name as it is compiled, and
15445 print some statistics about each pass when it finishes.
15447 @item -ftime-report
15448 @opindex ftime-report
15449 Makes the compiler print some statistics about the time consumed by each
15450 pass when it finishes.
15452 @item -ftime-report-details
15453 @opindex ftime-report-details
15454 Record the time consumed by infrastructure parts separately for each pass.
15456 @item -fira-verbose=@var{n}
15457 @opindex fira-verbose
15458 Control the verbosity of the dump file for the integrated register allocator.
15459 The default value is 5. If the value @var{n} is greater or equal to 10,
15460 the dump output is sent to stderr using the same format as @var{n} minus 10.
15463 @opindex flto-report
15464 Prints a report with internal details on the workings of the link-time
15465 optimizer. The contents of this report vary from version to version.
15466 It is meant to be useful to GCC developers when processing object
15467 files in LTO mode (via @option{-flto}).
15469 Disabled by default.
15471 @item -flto-report-wpa
15472 @opindex flto-report-wpa
15473 Like @option{-flto-report}, but only print for the WPA phase of link-time
15477 @opindex fmem-report
15478 Makes the compiler print some statistics about permanent memory
15479 allocation when it finishes.
15481 @item -fmem-report-wpa
15482 @opindex fmem-report-wpa
15483 Makes the compiler print some statistics about permanent memory
15484 allocation for the WPA phase only.
15486 @item -fpre-ipa-mem-report
15487 @opindex fpre-ipa-mem-report
15488 @item -fpost-ipa-mem-report
15489 @opindex fpost-ipa-mem-report
15490 Makes the compiler print some statistics about permanent memory
15491 allocation before or after interprocedural optimization.
15493 @item -fprofile-report
15494 @opindex fprofile-report
15495 Makes the compiler print some statistics about consistency of the
15496 (estimated) profile and effect of individual passes.
15498 @item -fstack-usage
15499 @opindex fstack-usage
15500 Makes the compiler output stack usage information for the program, on a
15501 per-function basis. The filename for the dump is made by appending
15502 @file{.su} to the @var{auxname}. @var{auxname} is generated from the name of
15503 the output file, if explicitly specified and it is not an executable,
15504 otherwise it is the basename of the source file. An entry is made up
15509 The name of the function.
15513 One or more qualifiers: @code{static}, @code{dynamic}, @code{bounded}.
15516 The qualifier @code{static} means that the function manipulates the stack
15517 statically: a fixed number of bytes are allocated for the frame on function
15518 entry and released on function exit; no stack adjustments are otherwise made
15519 in the function. The second field is this fixed number of bytes.
15521 The qualifier @code{dynamic} means that the function manipulates the stack
15522 dynamically: in addition to the static allocation described above, stack
15523 adjustments are made in the body of the function, for example to push/pop
15524 arguments around function calls. If the qualifier @code{bounded} is also
15525 present, the amount of these adjustments is bounded at compile time and
15526 the second field is an upper bound of the total amount of stack used by
15527 the function. If it is not present, the amount of these adjustments is
15528 not bounded at compile time and the second field only represents the
15533 Emit statistics about front-end processing at the end of the compilation.
15534 This option is supported only by the C++ front end, and
15535 the information is generally only useful to the G++ development team.
15537 @item -fdbg-cnt-list
15538 @opindex fdbg-cnt-list
15539 Print the name and the counter upper bound for all debug counters.
15542 @item -fdbg-cnt=@var{counter-value-list}
15544 Set the internal debug counter lower and upper bound. @var{counter-value-list}
15545 is a comma-separated list of @var{name}:@var{lower_bound}:@var{upper_bound}
15546 tuples which sets the lower and the upper bound of each debug
15547 counter @var{name}. The @var{lower_bound} is optional and is zero
15548 initialized if not set.
15549 All debug counters have the initial upper bound of @code{UINT_MAX};
15550 thus @code{dbg_cnt} returns true always unless the upper bound
15551 is set by this option.
15552 For example, with @option{-fdbg-cnt=dce:2:4,tail_call:10},
15553 @code{dbg_cnt(dce)} returns true only for third and fourth invocation.
15554 For @code{dbg_cnt(tail_call)} true is returned for first 10 invocations.
15556 @item -print-file-name=@var{library}
15557 @opindex print-file-name
15558 Print the full absolute name of the library file @var{library} that
15559 would be used when linking---and don't do anything else. With this
15560 option, GCC does not compile or link anything; it just prints the
15563 @item -print-multi-directory
15564 @opindex print-multi-directory
15565 Print the directory name corresponding to the multilib selected by any
15566 other switches present in the command line. This directory is supposed
15567 to exist in @env{GCC_EXEC_PREFIX}.
15569 @item -print-multi-lib
15570 @opindex print-multi-lib
15571 Print the mapping from multilib directory names to compiler switches
15572 that enable them. The directory name is separated from the switches by
15573 @samp{;}, and each switch starts with an @samp{@@} instead of the
15574 @samp{-}, without spaces between multiple switches. This is supposed to
15575 ease shell processing.
15577 @item -print-multi-os-directory
15578 @opindex print-multi-os-directory
15579 Print the path to OS libraries for the selected
15580 multilib, relative to some @file{lib} subdirectory. If OS libraries are
15581 present in the @file{lib} subdirectory and no multilibs are used, this is
15582 usually just @file{.}, if OS libraries are present in @file{lib@var{suffix}}
15583 sibling directories this prints e.g.@: @file{../lib64}, @file{../lib} or
15584 @file{../lib32}, or if OS libraries are present in @file{lib/@var{subdir}}
15585 subdirectories it prints e.g.@: @file{amd64}, @file{sparcv9} or @file{ev6}.
15587 @item -print-multiarch
15588 @opindex print-multiarch
15589 Print the path to OS libraries for the selected multiarch,
15590 relative to some @file{lib} subdirectory.
15592 @item -print-prog-name=@var{program}
15593 @opindex print-prog-name
15594 Like @option{-print-file-name}, but searches for a program such as @command{cpp}.
15596 @item -print-libgcc-file-name
15597 @opindex print-libgcc-file-name
15598 Same as @option{-print-file-name=libgcc.a}.
15600 This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs}
15601 but you do want to link with @file{libgcc.a}. You can do:
15604 gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name`
15607 @item -print-search-dirs
15608 @opindex print-search-dirs
15609 Print the name of the configured installation directory and a list of
15610 program and library directories @command{gcc} searches---and don't do anything else.
15612 This is useful when @command{gcc} prints the error message
15613 @samp{installation problem, cannot exec cpp0: No such file or directory}.
15614 To resolve this you either need to put @file{cpp0} and the other compiler
15615 components where @command{gcc} expects to find them, or you can set the environment
15616 variable @env{GCC_EXEC_PREFIX} to the directory where you installed them.
15617 Don't forget the trailing @samp{/}.
15618 @xref{Environment Variables}.
15620 @item -print-sysroot
15621 @opindex print-sysroot
15622 Print the target sysroot directory that is used during
15623 compilation. This is the target sysroot specified either at configure
15624 time or using the @option{--sysroot} option, possibly with an extra
15625 suffix that depends on compilation options. If no target sysroot is
15626 specified, the option prints nothing.
15628 @item -print-sysroot-headers-suffix
15629 @opindex print-sysroot-headers-suffix
15630 Print the suffix added to the target sysroot when searching for
15631 headers, or give an error if the compiler is not configured with such
15632 a suffix---and don't do anything else.
15635 @opindex dumpmachine
15636 Print the compiler's target machine (for example,
15637 @samp{i686-pc-linux-gnu})---and don't do anything else.
15640 @opindex dumpversion
15641 Print the compiler version (for example, @code{3.0}, @code{6.3.0} or @code{7})---and don't do
15642 anything else. This is the compiler version used in filesystem paths and
15643 specs. Depending on how the compiler has been configured it can be just
15644 a single number (major version), two numbers separated by a dot (major and
15645 minor version) or three numbers separated by dots (major, minor and patchlevel
15648 @item -dumpfullversion
15649 @opindex dumpfullversion
15650 Print the full compiler version---and don't do anything else. The output is
15651 always three numbers separated by dots, major, minor and patchlevel version.
15655 Print the compiler's built-in specs---and don't do anything else. (This
15656 is used when GCC itself is being built.) @xref{Spec Files}.
15659 @node Submodel Options
15660 @section Machine-Dependent Options
15661 @cindex submodel options
15662 @cindex specifying hardware config
15663 @cindex hardware models and configurations, specifying
15664 @cindex target-dependent options
15665 @cindex machine-dependent options
15667 Each target machine supported by GCC can have its own options---for
15668 example, to allow you to compile for a particular processor variant or
15669 ABI, or to control optimizations specific to that machine. By
15670 convention, the names of machine-specific options start with
15673 Some configurations of the compiler also support additional target-specific
15674 options, usually for compatibility with other compilers on the same
15677 @c This list is ordered alphanumerically by subsection name.
15678 @c It should be the same order and spelling as these options are listed
15679 @c in Machine Dependent Options
15682 * AArch64 Options::
15683 * Adapteva Epiphany Options::
15684 * AMD GCN Options::
15688 * Blackfin Options::
15694 * DEC Alpha Options::
15699 * GNU/Linux Options::
15709 * MicroBlaze Options::
15712 * MN10300 Options::
15716 * Nios II Options::
15717 * Nvidia PTX Options::
15718 * OpenRISC Options::
15720 * picoChip Options::
15721 * PowerPC Options::
15725 * RS/6000 and PowerPC Options::
15727 * S/390 and zSeries Options::
15730 * Solaris 2 Options::
15732 * System V Options::
15733 * TILE-Gx Options::
15734 * TILEPro Options::
15739 * VxWorks Options::
15741 * x86 Windows Options::
15742 * Xstormy16 Options::
15744 * zSeries Options::
15747 @node AArch64 Options
15748 @subsection AArch64 Options
15749 @cindex AArch64 Options
15751 These options are defined for AArch64 implementations:
15755 @item -mabi=@var{name}
15757 Generate code for the specified data model. Permissible values
15758 are @samp{ilp32} for SysV-like data model where int, long int and pointers
15759 are 32 bits, and @samp{lp64} for SysV-like data model where int is 32 bits,
15760 but long int and pointers are 64 bits.
15762 The default depends on the specific target configuration. Note that
15763 the LP64 and ILP32 ABIs are not link-compatible; you must compile your
15764 entire program with the same ABI, and link with a compatible set of libraries.
15767 @opindex mbig-endian
15768 Generate big-endian code. This is the default when GCC is configured for an
15769 @samp{aarch64_be-*-*} target.
15771 @item -mgeneral-regs-only
15772 @opindex mgeneral-regs-only
15773 Generate code which uses only the general-purpose registers. This will prevent
15774 the compiler from using floating-point and Advanced SIMD registers but will not
15775 impose any restrictions on the assembler.
15777 @item -mlittle-endian
15778 @opindex mlittle-endian
15779 Generate little-endian code. This is the default when GCC is configured for an
15780 @samp{aarch64-*-*} but not an @samp{aarch64_be-*-*} target.
15782 @item -mcmodel=tiny
15783 @opindex mcmodel=tiny
15784 Generate code for the tiny code model. The program and its statically defined
15785 symbols must be within 1MB of each other. Programs can be statically or
15786 dynamically linked.
15788 @item -mcmodel=small
15789 @opindex mcmodel=small
15790 Generate code for the small code model. The program and its statically defined
15791 symbols must be within 4GB of each other. Programs can be statically or
15792 dynamically linked. This is the default code model.
15794 @item -mcmodel=large
15795 @opindex mcmodel=large
15796 Generate code for the large code model. This makes no assumptions about
15797 addresses and sizes of sections. Programs can be statically linked only.
15799 @item -mstrict-align
15800 @itemx -mno-strict-align
15801 @opindex mstrict-align
15802 @opindex mno-strict-align
15803 Avoid or allow generating memory accesses that may not be aligned on a natural
15804 object boundary as described in the architecture specification.
15806 @item -momit-leaf-frame-pointer
15807 @itemx -mno-omit-leaf-frame-pointer
15808 @opindex momit-leaf-frame-pointer
15809 @opindex mno-omit-leaf-frame-pointer
15810 Omit or keep the frame pointer in leaf functions. The former behavior is the
15813 @item -mstack-protector-guard=@var{guard}
15814 @itemx -mstack-protector-guard-reg=@var{reg}
15815 @itemx -mstack-protector-guard-offset=@var{offset}
15816 @opindex mstack-protector-guard
15817 @opindex mstack-protector-guard-reg
15818 @opindex mstack-protector-guard-offset
15819 Generate stack protection code using canary at @var{guard}. Supported
15820 locations are @samp{global} for a global canary or @samp{sysreg} for a
15821 canary in an appropriate system register.
15823 With the latter choice the options
15824 @option{-mstack-protector-guard-reg=@var{reg}} and
15825 @option{-mstack-protector-guard-offset=@var{offset}} furthermore specify
15826 which system register to use as base register for reading the canary,
15827 and from what offset from that base register. There is no default
15828 register or offset as this is entirely for use within the Linux
15831 @item -mstack-protector-guard=@var{guard}
15832 @itemx -mstack-protector-guard-reg=@var{reg}
15833 @itemx -mstack-protector-guard-offset=@var{offset}
15834 @opindex mstack-protector-guard
15835 @opindex mstack-protector-guard-reg
15836 @opindex mstack-protector-guard-offset
15837 Generate stack protection code using canary at @var{guard}. Supported
15838 locations are @samp{global} for a global canary or @samp{sysreg} for a
15839 canary in an appropriate system register.
15841 With the latter choice the options
15842 @option{-mstack-protector-guard-reg=@var{reg}} and
15843 @option{-mstack-protector-guard-offset=@var{offset}} furthermore specify
15844 which system register to use as base register for reading the canary,
15845 and from what offset from that base register. There is no default
15846 register or offset as this is entirely for use within the Linux
15849 @item -mtls-dialect=desc
15850 @opindex mtls-dialect=desc
15851 Use TLS descriptors as the thread-local storage mechanism for dynamic accesses
15852 of TLS variables. This is the default.
15854 @item -mtls-dialect=traditional
15855 @opindex mtls-dialect=traditional
15856 Use traditional TLS as the thread-local storage mechanism for dynamic accesses
15859 @item -mtls-size=@var{size}
15861 Specify bit size of immediate TLS offsets. Valid values are 12, 24, 32, 48.
15862 This option requires binutils 2.26 or newer.
15864 @item -mfix-cortex-a53-835769
15865 @itemx -mno-fix-cortex-a53-835769
15866 @opindex mfix-cortex-a53-835769
15867 @opindex mno-fix-cortex-a53-835769
15868 Enable or disable the workaround for the ARM Cortex-A53 erratum number 835769.
15869 This involves inserting a NOP instruction between memory instructions and
15870 64-bit integer multiply-accumulate instructions.
15872 @item -mfix-cortex-a53-843419
15873 @itemx -mno-fix-cortex-a53-843419
15874 @opindex mfix-cortex-a53-843419
15875 @opindex mno-fix-cortex-a53-843419
15876 Enable or disable the workaround for the ARM Cortex-A53 erratum number 843419.
15877 This erratum workaround is made at link time and this will only pass the
15878 corresponding flag to the linker.
15880 @item -mlow-precision-recip-sqrt
15881 @itemx -mno-low-precision-recip-sqrt
15882 @opindex mlow-precision-recip-sqrt
15883 @opindex mno-low-precision-recip-sqrt
15884 Enable or disable the reciprocal square root approximation.
15885 This option only has an effect if @option{-ffast-math} or
15886 @option{-funsafe-math-optimizations} is used as well. Enabling this reduces
15887 precision of reciprocal square root results to about 16 bits for
15888 single precision and to 32 bits for double precision.
15890 @item -mlow-precision-sqrt
15891 @itemx -mno-low-precision-sqrt
15892 @opindex mlow-precision-sqrt
15893 @opindex mno-low-precision-sqrt
15894 Enable or disable the square root approximation.
15895 This option only has an effect if @option{-ffast-math} or
15896 @option{-funsafe-math-optimizations} is used as well. Enabling this reduces
15897 precision of square root results to about 16 bits for
15898 single precision and to 32 bits for double precision.
15899 If enabled, it implies @option{-mlow-precision-recip-sqrt}.
15901 @item -mlow-precision-div
15902 @itemx -mno-low-precision-div
15903 @opindex mlow-precision-div
15904 @opindex mno-low-precision-div
15905 Enable or disable the division approximation.
15906 This option only has an effect if @option{-ffast-math} or
15907 @option{-funsafe-math-optimizations} is used as well. Enabling this reduces
15908 precision of division results to about 16 bits for
15909 single precision and to 32 bits for double precision.
15911 @item -mtrack-speculation
15912 @itemx -mno-track-speculation
15913 Enable or disable generation of additional code to track speculative
15914 execution through conditional branches. The tracking state can then
15915 be used by the compiler when expanding calls to
15916 @code{__builtin_speculation_safe_copy} to permit a more efficient code
15917 sequence to be generated.
15919 @item -moutline-atomics
15920 @itemx -mno-outline-atomics
15921 Enable or disable calls to out-of-line helpers to implement atomic operations.
15922 These helpers will, at runtime, determine if the LSE instructions from
15923 ARMv8.1-A can be used; if not, they will use the load/store-exclusive
15924 instructions that are present in the base ARMv8.0 ISA.
15926 This option is only applicable when compiling for the base ARMv8.0
15927 instruction set. If using a later revision, e.g. @option{-march=armv8.1-a}
15928 or @option{-march=armv8-a+lse}, the ARMv8.1-Atomics instructions will be
15929 used directly. The same applies when using @option{-mcpu=} when the
15930 selected cpu supports the @samp{lse} feature.
15932 @item -march=@var{name}
15934 Specify the name of the target architecture and, optionally, one or
15935 more feature modifiers. This option has the form
15936 @option{-march=@var{arch}@r{@{}+@r{[}no@r{]}@var{feature}@r{@}*}}.
15938 The permissible values for @var{arch} are @samp{armv8-a},
15939 @samp{armv8.1-a}, @samp{armv8.2-a}, @samp{armv8.3-a}, @samp{armv8.4-a},
15940 @samp{armv8.5-a} or @var{native}.
15942 The value @samp{armv8.5-a} implies @samp{armv8.4-a} and enables compiler
15943 support for the ARMv8.5-A architecture extensions.
15945 The value @samp{armv8.4-a} implies @samp{armv8.3-a} and enables compiler
15946 support for the ARMv8.4-A architecture extensions.
15948 The value @samp{armv8.3-a} implies @samp{armv8.2-a} and enables compiler
15949 support for the ARMv8.3-A architecture extensions.
15951 The value @samp{armv8.2-a} implies @samp{armv8.1-a} and enables compiler
15952 support for the ARMv8.2-A architecture extensions.
15954 The value @samp{armv8.1-a} implies @samp{armv8-a} and enables compiler
15955 support for the ARMv8.1-A architecture extension. In particular, it
15956 enables the @samp{+crc}, @samp{+lse}, and @samp{+rdma} features.
15958 The value @samp{native} is available on native AArch64 GNU/Linux and
15959 causes the compiler to pick the architecture of the host system. This
15960 option has no effect if the compiler is unable to recognize the
15961 architecture of the host system,
15963 The permissible values for @var{feature} are listed in the sub-section
15964 on @ref{aarch64-feature-modifiers,,@option{-march} and @option{-mcpu}
15965 Feature Modifiers}. Where conflicting feature modifiers are
15966 specified, the right-most feature is used.
15968 GCC uses @var{name} to determine what kind of instructions it can emit
15969 when generating assembly code. If @option{-march} is specified
15970 without either of @option{-mtune} or @option{-mcpu} also being
15971 specified, the code is tuned to perform well across a range of target
15972 processors implementing the target architecture.
15974 @item -mtune=@var{name}
15976 Specify the name of the target processor for which GCC should tune the
15977 performance of the code. Permissible values for this option are:
15978 @samp{generic}, @samp{cortex-a35}, @samp{cortex-a53}, @samp{cortex-a55},
15979 @samp{cortex-a57}, @samp{cortex-a72}, @samp{cortex-a73}, @samp{cortex-a75},
15980 @samp{cortex-a76}, @samp{cortex-a76ae}, @samp{cortex-a77},
15981 @samp{cortex-a65}, @samp{cortex-a65ae}, @samp{cortex-a34},
15982 @samp{ares}, @samp{exynos-m1}, @samp{emag}, @samp{falkor},
15983 @samp{neoverse-e1},@samp{neoverse-n1},@samp{qdf24xx}, @samp{saphira},
15984 @samp{phecda}, @samp{xgene1}, @samp{vulcan}, @samp{octeontx},
15985 @samp{octeontx81}, @samp{octeontx83}, @samp{thunderx}, @samp{thunderxt88},
15986 @samp{thunderxt88p1}, @samp{thunderxt81}, @samp{tsv110},
15987 @samp{thunderxt83}, @samp{thunderx2t99},
15988 @samp{cortex-a57.cortex-a53}, @samp{cortex-a72.cortex-a53},
15989 @samp{cortex-a73.cortex-a35}, @samp{cortex-a73.cortex-a53},
15990 @samp{cortex-a75.cortex-a55}, @samp{cortex-a76.cortex-a55}
15993 The values @samp{cortex-a57.cortex-a53}, @samp{cortex-a72.cortex-a53},
15994 @samp{cortex-a73.cortex-a35}, @samp{cortex-a73.cortex-a53},
15995 @samp{cortex-a75.cortex-a55}, @samp{cortex-a76.cortex-a55} specify that GCC
15996 should tune for a big.LITTLE system.
15998 Additionally on native AArch64 GNU/Linux systems the value
15999 @samp{native} tunes performance to the host system. This option has no effect
16000 if the compiler is unable to recognize the processor of the host system.
16002 Where none of @option{-mtune=}, @option{-mcpu=} or @option{-march=}
16003 are specified, the code is tuned to perform well across a range
16004 of target processors.
16006 This option cannot be suffixed by feature modifiers.
16008 @item -mcpu=@var{name}
16010 Specify the name of the target processor, optionally suffixed by one
16011 or more feature modifiers. This option has the form
16012 @option{-mcpu=@var{cpu}@r{@{}+@r{[}no@r{]}@var{feature}@r{@}*}}, where
16013 the permissible values for @var{cpu} are the same as those available
16014 for @option{-mtune}. The permissible values for @var{feature} are
16015 documented in the sub-section on
16016 @ref{aarch64-feature-modifiers,,@option{-march} and @option{-mcpu}
16017 Feature Modifiers}. Where conflicting feature modifiers are
16018 specified, the right-most feature is used.
16020 GCC uses @var{name} to determine what kind of instructions it can emit when
16021 generating assembly code (as if by @option{-march}) and to determine
16022 the target processor for which to tune for performance (as if
16023 by @option{-mtune}). Where this option is used in conjunction
16024 with @option{-march} or @option{-mtune}, those options take precedence
16025 over the appropriate part of this option.
16027 @item -moverride=@var{string}
16029 Override tuning decisions made by the back-end in response to a
16030 @option{-mtune=} switch. The syntax, semantics, and accepted values
16031 for @var{string} in this option are not guaranteed to be consistent
16034 This option is only intended to be useful when developing GCC.
16036 @item -mverbose-cost-dump
16037 @opindex mverbose-cost-dump
16038 Enable verbose cost model dumping in the debug dump files. This option is
16039 provided for use in debugging the compiler.
16041 @item -mpc-relative-literal-loads
16042 @itemx -mno-pc-relative-literal-loads
16043 @opindex mpc-relative-literal-loads
16044 @opindex mno-pc-relative-literal-loads
16045 Enable or disable PC-relative literal loads. With this option literal pools are
16046 accessed using a single instruction and emitted after each function. This
16047 limits the maximum size of functions to 1MB. This is enabled by default for
16048 @option{-mcmodel=tiny}.
16050 @item -msign-return-address=@var{scope}
16051 @opindex msign-return-address
16052 Select the function scope on which return address signing will be applied.
16053 Permissible values are @samp{none}, which disables return address signing,
16054 @samp{non-leaf}, which enables pointer signing for functions which are not leaf
16055 functions, and @samp{all}, which enables pointer signing for all functions. The
16056 default value is @samp{none}. This option has been deprecated by
16057 -mbranch-protection.
16059 @item -mbranch-protection=@var{none}|@var{standard}|@var{pac-ret}[+@var{leaf}+@var{b-key}]|@var{bti}
16060 @opindex mbranch-protection
16061 Select the branch protection features to use.
16062 @samp{none} is the default and turns off all types of branch protection.
16063 @samp{standard} turns on all types of branch protection features. If a feature
16064 has additional tuning options, then @samp{standard} sets it to its standard
16066 @samp{pac-ret[+@var{leaf}]} turns on return address signing to its standard
16067 level: signing functions that save the return address to memory (non-leaf
16068 functions will practically always do this) using the a-key. The optional
16069 argument @samp{leaf} can be used to extend the signing to include leaf
16070 functions. The optional argument @samp{b-key} can be used to sign the functions
16071 with the B-key instead of the A-key.
16072 @samp{bti} turns on branch target identification mechanism.
16074 @item -msve-vector-bits=@var{bits}
16075 @opindex msve-vector-bits
16076 Specify the number of bits in an SVE vector register. This option only has
16077 an effect when SVE is enabled.
16079 GCC supports two forms of SVE code generation: ``vector-length
16080 agnostic'' output that works with any size of vector register and
16081 ``vector-length specific'' output that allows GCC to make assumptions
16082 about the vector length when it is useful for optimization reasons.
16083 The possible values of @samp{bits} are: @samp{scalable}, @samp{128},
16084 @samp{256}, @samp{512}, @samp{1024} and @samp{2048}.
16085 Specifying @samp{scalable} selects vector-length agnostic
16086 output. At present @samp{-msve-vector-bits=128} also generates vector-length
16087 agnostic output. All other values generate vector-length specific code.
16088 The behavior of these values may change in future releases and no value except
16089 @samp{scalable} should be relied on for producing code that is portable across
16090 different hardware SVE vector lengths.
16092 The default is @samp{-msve-vector-bits=scalable}, which produces
16093 vector-length agnostic code.
16096 @subsubsection @option{-march} and @option{-mcpu} Feature Modifiers
16097 @anchor{aarch64-feature-modifiers}
16098 @cindex @option{-march} feature modifiers
16099 @cindex @option{-mcpu} feature modifiers
16100 Feature modifiers used with @option{-march} and @option{-mcpu} can be any of
16101 the following and their inverses @option{no@var{feature}}:
16105 Enable CRC extension. This is on by default for
16106 @option{-march=armv8.1-a}.
16108 Enable Crypto extension. This also enables Advanced SIMD and floating-point
16111 Enable floating-point instructions. This is on by default for all possible
16112 values for options @option{-march} and @option{-mcpu}.
16114 Enable Advanced SIMD instructions. This also enables floating-point
16115 instructions. This is on by default for all possible values for options
16116 @option{-march} and @option{-mcpu}.
16118 Enable Scalable Vector Extension instructions. This also enables Advanced
16119 SIMD and floating-point instructions.
16121 Enable Large System Extension instructions. This is on by default for
16122 @option{-march=armv8.1-a}.
16124 Enable Round Double Multiply Accumulate instructions. This is on by default
16125 for @option{-march=armv8.1-a}.
16127 Enable FP16 extension. This also enables floating-point instructions.
16129 Enable FP16 fmla extension. This also enables FP16 extensions and
16130 floating-point instructions. This option is enabled by default for @option{-march=armv8.4-a}. Use of this option with architectures prior to Armv8.2-A is not supported.
16133 Enable the RcPc extension. This does not change code generation from GCC,
16134 but is passed on to the assembler, enabling inline asm statements to use
16135 instructions from the RcPc extension.
16137 Enable the Dot Product extension. This also enables Advanced SIMD instructions.
16139 Enable the Armv8-a aes and pmull crypto extension. This also enables Advanced
16142 Enable the Armv8-a sha2 crypto extension. This also enables Advanced SIMD instructions.
16144 Enable the sha512 and sha3 crypto extension. This also enables Advanced SIMD
16145 instructions. Use of this option with architectures prior to Armv8.2-A is not supported.
16147 Enable the sm3 and sm4 crypto extension. This also enables Advanced SIMD instructions.
16148 Use of this option with architectures prior to Armv8.2-A is not supported.
16150 Enable the Statistical Profiling extension. This option is only to enable the
16151 extension at the assembler level and does not affect code generation.
16153 Enable the Armv8.5-a Random Number instructions. This option is only to
16154 enable the extension at the assembler level and does not affect code
16157 Enable the Armv8.5-a Memory Tagging Extensions. This option is only to
16158 enable the extension at the assembler level and does not affect code
16161 Enable the Armv8-a Speculation Barrier instruction. This option is only to
16162 enable the extension at the assembler level and does not affect code
16163 generation. This option is enabled by default for @option{-march=armv8.5-a}.
16165 Enable the Armv8-a Speculative Store Bypass Safe instruction. This option
16166 is only to enable the extension at the assembler level and does not affect code
16167 generation. This option is enabled by default for @option{-march=armv8.5-a}.
16169 Enable the Armv8-a Execution and Data Prediction Restriction instructions.
16170 This option is only to enable the extension at the assembler level and does
16171 not affect code generation. This option is enabled by default for
16173 Enable the Armv8-a Scalable Vector Extension 2. This also enables SVE
16176 Enable SVE2 bitperm instructions. This also enables SVE2 instructions.
16178 Enable SVE2 sm4 instructions. This also enables SVE2 instructions.
16180 Enable SVE2 aes instructions. This also enables SVE2 instructions.
16182 Enable SVE2 sha3 instructions. This also enables SVE2 instructions.
16183 @option{-march=armv8.5-a}.
16185 Enable the Transactional Memory Extension.
16189 Feature @option{crypto} implies @option{aes}, @option{sha2}, and @option{simd},
16190 which implies @option{fp}.
16191 Conversely, @option{nofp} implies @option{nosimd}, which implies
16192 @option{nocrypto}, @option{noaes} and @option{nosha2}.
16194 @node Adapteva Epiphany Options
16195 @subsection Adapteva Epiphany Options
16197 These @samp{-m} options are defined for Adapteva Epiphany:
16200 @item -mhalf-reg-file
16201 @opindex mhalf-reg-file
16202 Don't allocate any register in the range @code{r32}@dots{}@code{r63}.
16203 That allows code to run on hardware variants that lack these registers.
16205 @item -mprefer-short-insn-regs
16206 @opindex mprefer-short-insn-regs
16207 Preferentially allocate registers that allow short instruction generation.
16208 This can result in increased instruction count, so this may either reduce or
16209 increase overall code size.
16211 @item -mbranch-cost=@var{num}
16212 @opindex mbranch-cost
16213 Set the cost of branches to roughly @var{num} ``simple'' instructions.
16214 This cost is only a heuristic and is not guaranteed to produce
16215 consistent results across releases.
16219 Enable the generation of conditional moves.
16221 @item -mnops=@var{num}
16223 Emit @var{num} NOPs before every other generated instruction.
16225 @item -mno-soft-cmpsf
16226 @opindex mno-soft-cmpsf
16227 @opindex msoft-cmpsf
16228 For single-precision floating-point comparisons, emit an @code{fsub} instruction
16229 and test the flags. This is faster than a software comparison, but can
16230 get incorrect results in the presence of NaNs, or when two different small
16231 numbers are compared such that their difference is calculated as zero.
16232 The default is @option{-msoft-cmpsf}, which uses slower, but IEEE-compliant,
16233 software comparisons.
16235 @item -mstack-offset=@var{num}
16236 @opindex mstack-offset
16237 Set the offset between the top of the stack and the stack pointer.
16238 E.g., a value of 8 means that the eight bytes in the range @code{sp+0@dots{}sp+7}
16239 can be used by leaf functions without stack allocation.
16240 Values other than @samp{8} or @samp{16} are untested and unlikely to work.
16241 Note also that this option changes the ABI; compiling a program with a
16242 different stack offset than the libraries have been compiled with
16243 generally does not work.
16244 This option can be useful if you want to evaluate if a different stack
16245 offset would give you better code, but to actually use a different stack
16246 offset to build working programs, it is recommended to configure the
16247 toolchain with the appropriate @option{--with-stack-offset=@var{num}} option.
16249 @item -mno-round-nearest
16250 @opindex mno-round-nearest
16251 @opindex mround-nearest
16252 Make the scheduler assume that the rounding mode has been set to
16253 truncating. The default is @option{-mround-nearest}.
16256 @opindex mlong-calls
16257 If not otherwise specified by an attribute, assume all calls might be beyond
16258 the offset range of the @code{b} / @code{bl} instructions, and therefore load the
16259 function address into a register before performing a (otherwise direct) call.
16260 This is the default.
16262 @item -mshort-calls
16263 @opindex short-calls
16264 If not otherwise specified by an attribute, assume all direct calls are
16265 in the range of the @code{b} / @code{bl} instructions, so use these instructions
16266 for direct calls. The default is @option{-mlong-calls}.
16270 Assume addresses can be loaded as 16-bit unsigned values. This does not
16271 apply to function addresses for which @option{-mlong-calls} semantics
16274 @item -mfp-mode=@var{mode}
16276 Set the prevailing mode of the floating-point unit.
16277 This determines the floating-point mode that is provided and expected
16278 at function call and return time. Making this mode match the mode you
16279 predominantly need at function start can make your programs smaller and
16280 faster by avoiding unnecessary mode switches.
16282 @var{mode} can be set to one the following values:
16286 Any mode at function entry is valid, and retained or restored when
16287 the function returns, and when it calls other functions.
16288 This mode is useful for compiling libraries or other compilation units
16289 you might want to incorporate into different programs with different
16290 prevailing FPU modes, and the convenience of being able to use a single
16291 object file outweighs the size and speed overhead for any extra
16292 mode switching that might be needed, compared with what would be needed
16293 with a more specific choice of prevailing FPU mode.
16296 This is the mode used for floating-point calculations with
16297 truncating (i.e.@: round towards zero) rounding mode. That includes
16298 conversion from floating point to integer.
16300 @item round-nearest
16301 This is the mode used for floating-point calculations with
16302 round-to-nearest-or-even rounding mode.
16305 This is the mode used to perform integer calculations in the FPU, e.g.@:
16306 integer multiply, or integer multiply-and-accumulate.
16309 The default is @option{-mfp-mode=caller}
16311 @item -mno-split-lohi
16312 @itemx -mno-postinc
16313 @itemx -mno-postmodify
16314 @opindex mno-split-lohi
16315 @opindex msplit-lohi
16316 @opindex mno-postinc
16318 @opindex mno-postmodify
16319 @opindex mpostmodify
16320 Code generation tweaks that disable, respectively, splitting of 32-bit
16321 loads, generation of post-increment addresses, and generation of
16322 post-modify addresses. The defaults are @option{msplit-lohi},
16323 @option{-mpost-inc}, and @option{-mpost-modify}.
16325 @item -mnovect-double
16326 @opindex mno-vect-double
16327 @opindex mvect-double
16328 Change the preferred SIMD mode to SImode. The default is
16329 @option{-mvect-double}, which uses DImode as preferred SIMD mode.
16331 @item -max-vect-align=@var{num}
16332 @opindex max-vect-align
16333 The maximum alignment for SIMD vector mode types.
16334 @var{num} may be 4 or 8. The default is 8.
16335 Note that this is an ABI change, even though many library function
16336 interfaces are unaffected if they don't use SIMD vector modes
16337 in places that affect size and/or alignment of relevant types.
16339 @item -msplit-vecmove-early
16340 @opindex msplit-vecmove-early
16341 Split vector moves into single word moves before reload. In theory this
16342 can give better register allocation, but so far the reverse seems to be
16343 generally the case.
16345 @item -m1reg-@var{reg}
16347 Specify a register to hold the constant @minus{}1, which makes loading small negative
16348 constants and certain bitmasks faster.
16349 Allowable values for @var{reg} are @samp{r43} and @samp{r63},
16350 which specify use of that register as a fixed register,
16351 and @samp{none}, which means that no register is used for this
16352 purpose. The default is @option{-m1reg-none}.
16356 @node AMD GCN Options
16357 @subsection AMD GCN Options
16358 @cindex AMD GCN Options
16360 These options are defined specifically for the AMD GCN port.
16364 @item -march=@var{gpu}
16366 @itemx -mtune=@var{gpu}
16368 Set architecture type or tuning for @var{gpu}. Supported values for @var{gpu}
16374 Compile for GCN3 Fiji devices (gfx803).
16377 Compile for GCN5 Vega 10 devices (gfx900).
16380 Compile for GCN5 Vega 20 devices (gfx906).
16384 @item -mstack-size=@var{bytes}
16385 @opindex mstack-size
16386 Specify how many @var{bytes} of stack space will be requested for each GPU
16387 thread (wave-front). Beware that there may be many threads and limited memory
16388 available. The size of the stack allocation may also have an impact on
16389 run-time performance. The default is 32KB when using OpenACC or OpenMP, and
16395 @subsection ARC Options
16396 @cindex ARC options
16398 The following options control the architecture variant for which code
16401 @c architecture variants
16404 @item -mbarrel-shifter
16405 @opindex mbarrel-shifter
16406 Generate instructions supported by barrel shifter. This is the default
16407 unless @option{-mcpu=ARC601} or @samp{-mcpu=ARCEM} is in effect.
16410 @opindex mjli-alawys
16411 Force to call a function using jli_s instruction. This option is
16412 valid only for ARCv2 architecture.
16414 @item -mcpu=@var{cpu}
16416 Set architecture type, register usage, and instruction scheduling
16417 parameters for @var{cpu}. There are also shortcut alias options
16418 available for backward compatibility and convenience. Supported
16419 values for @var{cpu} are
16425 Compile for ARC600. Aliases: @option{-mA6}, @option{-mARC600}.
16429 Compile for ARC601. Alias: @option{-mARC601}.
16434 Compile for ARC700. Aliases: @option{-mA7}, @option{-mARC700}.
16435 This is the default when configured with @option{--with-cpu=arc700}@.
16438 Compile for ARC EM.
16441 Compile for ARC HS.
16444 Compile for ARC EM CPU with no hardware extensions.
16447 Compile for ARC EM4 CPU.
16450 Compile for ARC EM4 DMIPS CPU.
16453 Compile for ARC EM4 DMIPS CPU with the single-precision floating-point
16457 Compile for ARC EM4 DMIPS CPU with single-precision floating-point and
16458 double assist instructions.
16461 Compile for ARC HS CPU with no hardware extensions except the atomic
16465 Compile for ARC HS34 CPU.
16468 Compile for ARC HS38 CPU.
16471 Compile for ARC HS38 CPU with all hardware extensions on.
16474 Compile for ARC 600 CPU with @code{norm} instructions enabled.
16476 @item arc600_mul32x16
16477 Compile for ARC 600 CPU with @code{norm} and 32x16-bit multiply
16478 instructions enabled.
16481 Compile for ARC 600 CPU with @code{norm} and @code{mul64}-family
16482 instructions enabled.
16485 Compile for ARC 601 CPU with @code{norm} instructions enabled.
16487 @item arc601_mul32x16
16488 Compile for ARC 601 CPU with @code{norm} and 32x16-bit multiply
16489 instructions enabled.
16492 Compile for ARC 601 CPU with @code{norm} and @code{mul64}-family
16493 instructions enabled.
16496 Compile for ARC 700 on NPS400 chip.
16499 Compile for ARC EM minimalist configuration featuring reduced register
16506 @itemx -mdpfp-compact
16507 @opindex mdpfp-compact
16508 Generate double-precision FPX instructions, tuned for the compact
16512 @opindex mdpfp-fast
16513 Generate double-precision FPX instructions, tuned for the fast
16516 @item -mno-dpfp-lrsr
16517 @opindex mno-dpfp-lrsr
16518 Disable @code{lr} and @code{sr} instructions from using FPX extension
16523 Generate extended arithmetic instructions. Currently only
16524 @code{divaw}, @code{adds}, @code{subs}, and @code{sat16} are
16525 supported. This is always enabled for @option{-mcpu=ARC700}.
16530 Do not generate @code{mpy}-family instructions for ARC700. This option is
16535 Generate 32x16-bit multiply and multiply-accumulate instructions.
16539 Generate @code{mul64} and @code{mulu64} instructions.
16540 Only valid for @option{-mcpu=ARC600}.
16544 Generate @code{norm} instructions. This is the default if @option{-mcpu=ARC700}
16549 @itemx -mspfp-compact
16550 @opindex mspfp-compact
16551 Generate single-precision FPX instructions, tuned for the compact
16555 @opindex mspfp-fast
16556 Generate single-precision FPX instructions, tuned for the fast
16561 Enable generation of ARC SIMD instructions via target-specific
16562 builtins. Only valid for @option{-mcpu=ARC700}.
16565 @opindex msoft-float
16566 This option ignored; it is provided for compatibility purposes only.
16567 Software floating-point code is emitted by default, and this default
16568 can overridden by FPX options; @option{-mspfp}, @option{-mspfp-compact}, or
16569 @option{-mspfp-fast} for single precision, and @option{-mdpfp},
16570 @option{-mdpfp-compact}, or @option{-mdpfp-fast} for double precision.
16574 Generate @code{swap} instructions.
16578 This enables use of the locked load/store conditional extension to implement
16579 atomic memory built-in functions. Not available for ARC 6xx or ARC
16584 Enable @code{div} and @code{rem} instructions for ARCv2 cores.
16586 @item -mcode-density
16587 @opindex mcode-density
16588 Enable code density instructions for ARC EM.
16589 This option is on by default for ARC HS.
16593 Enable double load/store operations for ARC HS cores.
16595 @item -mtp-regno=@var{regno}
16597 Specify thread pointer register number.
16599 @item -mmpy-option=@var{multo}
16600 @opindex mmpy-option
16601 Compile ARCv2 code with a multiplier design option. You can specify
16602 the option using either a string or numeric value for @var{multo}.
16603 @samp{wlh1} is the default value. The recognized values are:
16608 No multiplier available.
16612 16x16 multiplier, fully pipelined.
16613 The following instructions are enabled: @code{mpyw} and @code{mpyuw}.
16617 32x32 multiplier, fully
16618 pipelined (1 stage). The following instructions are additionally
16619 enabled: @code{mpy}, @code{mpyu}, @code{mpym}, @code{mpymu}, and @code{mpy_s}.
16623 32x32 multiplier, fully pipelined
16624 (2 stages). The following instructions are additionally enabled: @code{mpy},
16625 @code{mpyu}, @code{mpym}, @code{mpymu}, and @code{mpy_s}.
16629 Two 16x16 multipliers, blocking,
16630 sequential. The following instructions are additionally enabled: @code{mpy},
16631 @code{mpyu}, @code{mpym}, @code{mpymu}, and @code{mpy_s}.
16635 One 16x16 multiplier, blocking,
16636 sequential. The following instructions are additionally enabled: @code{mpy},
16637 @code{mpyu}, @code{mpym}, @code{mpymu}, and @code{mpy_s}.
16641 One 32x4 multiplier, blocking,
16642 sequential. The following instructions are additionally enabled: @code{mpy},
16643 @code{mpyu}, @code{mpym}, @code{mpymu}, and @code{mpy_s}.
16647 ARC HS SIMD support.
16651 ARC HS SIMD support.
16655 ARC HS SIMD support.
16659 This option is only available for ARCv2 cores@.
16661 @item -mfpu=@var{fpu}
16663 Enables support for specific floating-point hardware extensions for ARCv2
16664 cores. Supported values for @var{fpu} are:
16669 Enables support for single-precision floating-point hardware
16673 Enables support for double-precision floating-point hardware
16674 extensions. The single-precision floating-point extension is also
16675 enabled. Not available for ARC EM@.
16678 Enables support for double-precision floating-point hardware
16679 extensions using double-precision assist instructions. The single-precision
16680 floating-point extension is also enabled. This option is
16681 only available for ARC EM@.
16684 Enables support for double-precision floating-point hardware
16685 extensions using double-precision assist instructions.
16686 The single-precision floating-point, square-root, and divide
16687 extensions are also enabled. This option is
16688 only available for ARC EM@.
16691 Enables support for double-precision floating-point hardware
16692 extensions using double-precision assist instructions.
16693 The single-precision floating-point and fused multiply and add
16694 hardware extensions are also enabled. This option is
16695 only available for ARC EM@.
16698 Enables support for double-precision floating-point hardware
16699 extensions using double-precision assist instructions.
16700 All single-precision floating-point hardware extensions are also
16701 enabled. This option is only available for ARC EM@.
16704 Enables support for single-precision floating-point, square-root and divide
16705 hardware extensions@.
16708 Enables support for double-precision floating-point, square-root and divide
16709 hardware extensions. This option
16710 includes option @samp{fpus_div}. Not available for ARC EM@.
16713 Enables support for single-precision floating-point and
16714 fused multiply and add hardware extensions@.
16717 Enables support for double-precision floating-point and
16718 fused multiply and add hardware extensions. This option
16719 includes option @samp{fpus_fma}. Not available for ARC EM@.
16722 Enables support for all single-precision floating-point hardware
16726 Enables support for all single- and double-precision floating-point
16727 hardware extensions. Not available for ARC EM@.
16731 @item -mirq-ctrl-saved=@var{register-range}, @var{blink}, @var{lp_count}
16732 @opindex mirq-ctrl-saved
16733 Specifies general-purposes registers that the processor automatically
16734 saves/restores on interrupt entry and exit. @var{register-range} is
16735 specified as two registers separated by a dash. The register range
16736 always starts with @code{r0}, the upper limit is @code{fp} register.
16737 @var{blink} and @var{lp_count} are optional. This option is only
16738 valid for ARC EM and ARC HS cores.
16740 @item -mrgf-banked-regs=@var{number}
16741 @opindex mrgf-banked-regs
16742 Specifies the number of registers replicated in second register bank
16743 on entry to fast interrupt. Fast interrupts are interrupts with the
16744 highest priority level P0. These interrupts save only PC and STATUS32
16745 registers to avoid memory transactions during interrupt entry and exit
16746 sequences. Use this option when you are using fast interrupts in an
16747 ARC V2 family processor. Permitted values are 4, 8, 16, and 32.
16749 @item -mlpc-width=@var{width}
16750 @opindex mlpc-width
16751 Specify the width of the @code{lp_count} register. Valid values for
16752 @var{width} are 8, 16, 20, 24, 28 and 32 bits. The default width is
16753 fixed to 32 bits. If the width is less than 32, the compiler does not
16754 attempt to transform loops in your program to use the zero-delay loop
16755 mechanism unless it is known that the @code{lp_count} register can
16756 hold the required loop-counter value. Depending on the width
16757 specified, the compiler and run-time library might continue to use the
16758 loop mechanism for various needs. This option defines macro
16759 @code{__ARC_LPC_WIDTH__} with the value of @var{width}.
16763 This option instructs the compiler to generate code for a 16-entry
16764 register file. This option defines the @code{__ARC_RF16__}
16765 preprocessor macro.
16767 @item -mbranch-index
16768 @opindex mbranch-index
16769 Enable use of @code{bi} or @code{bih} instructions to implement jump
16774 The following options are passed through to the assembler, and also
16775 define preprocessor macro symbols.
16777 @c Flags used by the assembler, but for which we define preprocessor
16778 @c macro symbols as well.
16781 @opindex mdsp-packa
16782 Passed down to the assembler to enable the DSP Pack A extensions.
16783 Also sets the preprocessor symbol @code{__Xdsp_packa}. This option is
16788 Passed down to the assembler to enable the dual Viterbi butterfly
16789 extension. Also sets the preprocessor symbol @code{__Xdvbf}. This
16790 option is deprecated.
16792 @c ARC700 4.10 extension instruction
16795 Passed down to the assembler to enable the locked load/store
16796 conditional extension. Also sets the preprocessor symbol
16801 Passed down to the assembler. Also sets the preprocessor symbol
16802 @code{__Xxmac_d16}. This option is deprecated.
16806 Passed down to the assembler. Also sets the preprocessor symbol
16807 @code{__Xxmac_24}. This option is deprecated.
16809 @c ARC700 4.10 extension instruction
16812 Passed down to the assembler to enable the 64-bit time-stamp counter
16813 extension instruction. Also sets the preprocessor symbol
16814 @code{__Xrtsc}. This option is deprecated.
16816 @c ARC700 4.10 extension instruction
16819 Passed down to the assembler to enable the swap byte ordering
16820 extension instruction. Also sets the preprocessor symbol
16824 @opindex mtelephony
16825 Passed down to the assembler to enable dual- and single-operand
16826 instructions for telephony. Also sets the preprocessor symbol
16827 @code{__Xtelephony}. This option is deprecated.
16831 Passed down to the assembler to enable the XY memory extension. Also
16832 sets the preprocessor symbol @code{__Xxy}.
16836 The following options control how the assembly code is annotated:
16838 @c Assembly annotation options
16842 Annotate assembler instructions with estimated addresses.
16844 @item -mannotate-align
16845 @opindex mannotate-align
16846 Explain what alignment considerations lead to the decision to make an
16847 instruction short or long.
16851 The following options are passed through to the linker:
16853 @c options passed through to the linker
16857 Passed through to the linker, to specify use of the @code{arclinux} emulation.
16858 This option is enabled by default in tool chains built for
16859 @w{@code{arc-linux-uclibc}} and @w{@code{arceb-linux-uclibc}} targets
16860 when profiling is not requested.
16862 @item -marclinux_prof
16863 @opindex marclinux_prof
16864 Passed through to the linker, to specify use of the
16865 @code{arclinux_prof} emulation. This option is enabled by default in
16866 tool chains built for @w{@code{arc-linux-uclibc}} and
16867 @w{@code{arceb-linux-uclibc}} targets when profiling is requested.
16871 The following options control the semantics of generated code:
16873 @c semantically relevant code generation options
16876 @opindex mlong-calls
16877 Generate calls as register indirect calls, thus providing access
16878 to the full 32-bit address range.
16880 @item -mmedium-calls
16881 @opindex mmedium-calls
16882 Don't use less than 25-bit addressing range for calls, which is the
16883 offset available for an unconditional branch-and-link
16884 instruction. Conditional execution of function calls is suppressed, to
16885 allow use of the 25-bit range, rather than the 21-bit range with
16886 conditional branch-and-link. This is the default for tool chains built
16887 for @w{@code{arc-linux-uclibc}} and @w{@code{arceb-linux-uclibc}} targets.
16891 Put definitions of externally-visible data in a small data section if
16892 that data is no bigger than @var{num} bytes. The default value of
16893 @var{num} is 4 for any ARC configuration, or 8 when we have double
16894 load/store operations.
16899 Do not generate sdata references. This is the default for tool chains
16900 built for @w{@code{arc-linux-uclibc}} and @w{@code{arceb-linux-uclibc}}
16903 @item -mvolatile-cache
16904 @opindex mvolatile-cache
16905 Use ordinarily cached memory accesses for volatile references. This is the
16908 @item -mno-volatile-cache
16909 @opindex mno-volatile-cache
16910 @opindex mvolatile-cache
16911 Enable cache bypass for volatile references.
16915 The following options fine tune code generation:
16916 @c code generation tuning options
16919 @opindex malign-call
16920 Do alignment optimizations for call instructions.
16922 @item -mauto-modify-reg
16923 @opindex mauto-modify-reg
16924 Enable the use of pre/post modify with register displacement.
16926 @item -mbbit-peephole
16927 @opindex mbbit-peephole
16928 Enable bbit peephole2.
16932 This option disables a target-specific pass in @file{arc_reorg} to
16933 generate compare-and-branch (@code{br@var{cc}}) instructions.
16934 It has no effect on
16935 generation of these instructions driven by the combiner pass.
16937 @item -mcase-vector-pcrel
16938 @opindex mcase-vector-pcrel
16939 Use PC-relative switch case tables to enable case table shortening.
16940 This is the default for @option{-Os}.
16942 @item -mcompact-casesi
16943 @opindex mcompact-casesi
16944 Enable compact @code{casesi} pattern. This is the default for @option{-Os},
16945 and only available for ARCv1 cores. This option is deprecated.
16947 @item -mno-cond-exec
16948 @opindex mno-cond-exec
16949 Disable the ARCompact-specific pass to generate conditional
16950 execution instructions.
16952 Due to delay slot scheduling and interactions between operand numbers,
16953 literal sizes, instruction lengths, and the support for conditional execution,
16954 the target-independent pass to generate conditional execution is often lacking,
16955 so the ARC port has kept a special pass around that tries to find more
16956 conditional execution generation opportunities after register allocation,
16957 branch shortening, and delay slot scheduling have been done. This pass
16958 generally, but not always, improves performance and code size, at the cost of
16959 extra compilation time, which is why there is an option to switch it off.
16960 If you have a problem with call instructions exceeding their allowable
16961 offset range because they are conditionalized, you should consider using
16962 @option{-mmedium-calls} instead.
16964 @item -mearly-cbranchsi
16965 @opindex mearly-cbranchsi
16966 Enable pre-reload use of the @code{cbranchsi} pattern.
16968 @item -mexpand-adddi
16969 @opindex mexpand-adddi
16970 Expand @code{adddi3} and @code{subdi3} at RTL generation time into
16971 @code{add.f}, @code{adc} etc. This option is deprecated.
16973 @item -mindexed-loads
16974 @opindex mindexed-loads
16975 Enable the use of indexed loads. This can be problematic because some
16976 optimizers then assume that indexed stores exist, which is not
16981 Enable Local Register Allocation. This is still experimental for ARC,
16982 so by default the compiler uses standard reload
16983 (i.e.@: @option{-mno-lra}).
16985 @item -mlra-priority-none
16986 @opindex mlra-priority-none
16987 Don't indicate any priority for target registers.
16989 @item -mlra-priority-compact
16990 @opindex mlra-priority-compact
16991 Indicate target register priority for r0..r3 / r12..r15.
16993 @item -mlra-priority-noncompact
16994 @opindex mlra-priority-noncompact
16995 Reduce target register priority for r0..r3 / r12..r15.
16998 @opindex mmillicode
16999 When optimizing for size (using @option{-Os}), prologues and epilogues
17000 that have to save or restore a large number of registers are often
17001 shortened by using call to a special function in libgcc; this is
17002 referred to as a @emph{millicode} call. As these calls can pose
17003 performance issues, and/or cause linking issues when linking in a
17004 nonstandard way, this option is provided to turn on or off millicode
17007 @item -mcode-density-frame
17008 @opindex mcode-density-frame
17009 This option enable the compiler to emit @code{enter} and @code{leave}
17010 instructions. These instructions are only valid for CPUs with
17011 code-density feature.
17014 @opindex mmixed-code
17015 Tweak register allocation to help 16-bit instruction generation.
17016 This generally has the effect of decreasing the average instruction size
17017 while increasing the instruction count.
17021 Enable @samp{q} instruction alternatives.
17022 This is the default for @option{-Os}.
17026 Enable @samp{Rcq} constraint handling.
17027 Most short code generation depends on this.
17028 This is the default.
17032 Enable @samp{Rcw} constraint handling.
17033 Most ccfsm condexec mostly depends on this.
17034 This is the default.
17036 @item -msize-level=@var{level}
17037 @opindex msize-level
17038 Fine-tune size optimization with regards to instruction lengths and alignment.
17039 The recognized values for @var{level} are:
17042 No size optimization. This level is deprecated and treated like @samp{1}.
17045 Short instructions are used opportunistically.
17048 In addition, alignment of loops and of code after barriers are dropped.
17051 In addition, optional data alignment is dropped, and the option @option{Os} is enabled.
17055 This defaults to @samp{3} when @option{-Os} is in effect. Otherwise,
17056 the behavior when this is not set is equivalent to level @samp{1}.
17058 @item -mtune=@var{cpu}
17060 Set instruction scheduling parameters for @var{cpu}, overriding any implied
17061 by @option{-mcpu=}.
17063 Supported values for @var{cpu} are
17067 Tune for ARC600 CPU.
17070 Tune for ARC601 CPU.
17073 Tune for ARC700 CPU with standard multiplier block.
17076 Tune for ARC700 CPU with XMAC block.
17079 Tune for ARC725D CPU.
17082 Tune for ARC750D CPU.
17086 @item -mmultcost=@var{num}
17088 Cost to assume for a multiply instruction, with @samp{4} being equal to a
17089 normal instruction.
17091 @item -munalign-prob-threshold=@var{probability}
17092 @opindex munalign-prob-threshold
17093 Set probability threshold for unaligning branches.
17094 When tuning for @samp{ARC700} and optimizing for speed, branches without
17095 filled delay slot are preferably emitted unaligned and long, unless
17096 profiling indicates that the probability for the branch to be taken
17097 is below @var{probability}. @xref{Cross-profiling}.
17098 The default is (REG_BR_PROB_BASE/2), i.e.@: 5000.
17102 The following options are maintained for backward compatibility, but
17103 are now deprecated and will be removed in a future release:
17105 @c Deprecated options
17113 @opindex mbig-endian
17116 Compile code for big-endian targets. Use of these options is now
17117 deprecated. Big-endian code is supported by configuring GCC to build
17118 @w{@code{arceb-elf32}} and @w{@code{arceb-linux-uclibc}} targets,
17119 for which big endian is the default.
17121 @item -mlittle-endian
17122 @opindex mlittle-endian
17125 Compile code for little-endian targets. Use of these options is now
17126 deprecated. Little-endian code is supported by configuring GCC to build
17127 @w{@code{arc-elf32}} and @w{@code{arc-linux-uclibc}} targets,
17128 for which little endian is the default.
17130 @item -mbarrel_shifter
17131 @opindex mbarrel_shifter
17132 Replaced by @option{-mbarrel-shifter}.
17134 @item -mdpfp_compact
17135 @opindex mdpfp_compact
17136 Replaced by @option{-mdpfp-compact}.
17139 @opindex mdpfp_fast
17140 Replaced by @option{-mdpfp-fast}.
17143 @opindex mdsp_packa
17144 Replaced by @option{-mdsp-packa}.
17148 Replaced by @option{-mea}.
17152 Replaced by @option{-mmac-24}.
17156 Replaced by @option{-mmac-d16}.
17158 @item -mspfp_compact
17159 @opindex mspfp_compact
17160 Replaced by @option{-mspfp-compact}.
17163 @opindex mspfp_fast
17164 Replaced by @option{-mspfp-fast}.
17166 @item -mtune=@var{cpu}
17168 Values @samp{arc600}, @samp{arc601}, @samp{arc700} and
17169 @samp{arc700-xmac} for @var{cpu} are replaced by @samp{ARC600},
17170 @samp{ARC601}, @samp{ARC700} and @samp{ARC700-xmac} respectively.
17172 @item -multcost=@var{num}
17174 Replaced by @option{-mmultcost}.
17179 @subsection ARM Options
17180 @cindex ARM options
17182 These @samp{-m} options are defined for the ARM port:
17185 @item -mabi=@var{name}
17187 Generate code for the specified ABI@. Permissible values are: @samp{apcs-gnu},
17188 @samp{atpcs}, @samp{aapcs}, @samp{aapcs-linux} and @samp{iwmmxt}.
17191 @opindex mapcs-frame
17192 Generate a stack frame that is compliant with the ARM Procedure Call
17193 Standard for all functions, even if this is not strictly necessary for
17194 correct execution of the code. Specifying @option{-fomit-frame-pointer}
17195 with this option causes the stack frames not to be generated for
17196 leaf functions. The default is @option{-mno-apcs-frame}.
17197 This option is deprecated.
17201 This is a synonym for @option{-mapcs-frame} and is deprecated.
17204 @c not currently implemented
17205 @item -mapcs-stack-check
17206 @opindex mapcs-stack-check
17207 Generate code to check the amount of stack space available upon entry to
17208 every function (that actually uses some stack space). If there is
17209 insufficient space available then either the function
17210 @code{__rt_stkovf_split_small} or @code{__rt_stkovf_split_big} is
17211 called, depending upon the amount of stack space required. The runtime
17212 system is required to provide these functions. The default is
17213 @option{-mno-apcs-stack-check}, since this produces smaller code.
17215 @c not currently implemented
17216 @item -mapcs-reentrant
17217 @opindex mapcs-reentrant
17218 Generate reentrant, position-independent code. The default is
17219 @option{-mno-apcs-reentrant}.
17222 @item -mthumb-interwork
17223 @opindex mthumb-interwork
17224 Generate code that supports calling between the ARM and Thumb
17225 instruction sets. Without this option, on pre-v5 architectures, the
17226 two instruction sets cannot be reliably used inside one program. The
17227 default is @option{-mno-thumb-interwork}, since slightly larger code
17228 is generated when @option{-mthumb-interwork} is specified. In AAPCS
17229 configurations this option is meaningless.
17231 @item -mno-sched-prolog
17232 @opindex mno-sched-prolog
17233 @opindex msched-prolog
17234 Prevent the reordering of instructions in the function prologue, or the
17235 merging of those instruction with the instructions in the function's
17236 body. This means that all functions start with a recognizable set
17237 of instructions (or in fact one of a choice from a small set of
17238 different function prologues), and this information can be used to
17239 locate the start of functions inside an executable piece of code. The
17240 default is @option{-msched-prolog}.
17242 @item -mfloat-abi=@var{name}
17243 @opindex mfloat-abi
17244 Specifies which floating-point ABI to use. Permissible values
17245 are: @samp{soft}, @samp{softfp} and @samp{hard}.
17247 Specifying @samp{soft} causes GCC to generate output containing
17248 library calls for floating-point operations.
17249 @samp{softfp} allows the generation of code using hardware floating-point
17250 instructions, but still uses the soft-float calling conventions.
17251 @samp{hard} allows generation of floating-point instructions
17252 and uses FPU-specific calling conventions.
17254 The default depends on the specific target configuration. Note that
17255 the hard-float and soft-float ABIs are not link-compatible; you must
17256 compile your entire program with the same ABI, and link with a
17257 compatible set of libraries.
17259 @item -mgeneral-regs-only
17260 @opindex mgeneral-regs-only
17261 Generate code which uses only the general-purpose registers. This will prevent
17262 the compiler from using floating-point and Advanced SIMD registers but will not
17263 impose any restrictions on the assembler.
17265 @item -mlittle-endian
17266 @opindex mlittle-endian
17267 Generate code for a processor running in little-endian mode. This is
17268 the default for all standard configurations.
17271 @opindex mbig-endian
17272 Generate code for a processor running in big-endian mode; the default is
17273 to compile code for a little-endian processor.
17278 When linking a big-endian image select between BE8 and BE32 formats.
17279 The option has no effect for little-endian images and is ignored. The
17280 default is dependent on the selected target architecture. For ARMv6
17281 and later architectures the default is BE8, for older architectures
17282 the default is BE32. BE32 format has been deprecated by ARM.
17284 @item -march=@var{name}@r{[}+extension@dots{}@r{]}
17286 This specifies the name of the target ARM architecture. GCC uses this
17287 name to determine what kind of instructions it can emit when generating
17288 assembly code. This option can be used in conjunction with or instead
17289 of the @option{-mcpu=} option.
17291 Permissible names are:
17293 @samp{armv5t}, @samp{armv5te},
17294 @samp{armv6}, @samp{armv6j}, @samp{armv6k}, @samp{armv6kz}, @samp{armv6t2},
17295 @samp{armv6z}, @samp{armv6zk},
17296 @samp{armv7}, @samp{armv7-a}, @samp{armv7ve},
17297 @samp{armv8-a}, @samp{armv8.1-a}, @samp{armv8.2-a}, @samp{armv8.3-a},
17302 @samp{armv6-m}, @samp{armv6s-m},
17303 @samp{armv7-m}, @samp{armv7e-m},
17304 @samp{armv8-m.base}, @samp{armv8-m.main},
17305 @samp{iwmmxt} and @samp{iwmmxt2}.
17307 Additionally, the following architectures, which lack support for the
17308 Thumb execution state, are recognized but support is deprecated: @samp{armv4}.
17310 Many of the architectures support extensions. These can be added by
17311 appending @samp{+@var{extension}} to the architecture name. Extension
17312 options are processed in order and capabilities accumulate. An extension
17313 will also enable any necessary base extensions
17314 upon which it depends. For example, the @samp{+crypto} extension
17315 will always enable the @samp{+simd} extension. The exception to the
17316 additive construction is for extensions that are prefixed with
17317 @samp{+no@dots{}}: these extensions disable the specified option and
17318 any other extensions that may depend on the presence of that
17321 For example, @samp{-march=armv7-a+simd+nofp+vfpv4} is equivalent to
17322 writing @samp{-march=armv7-a+vfpv4} since the @samp{+simd} option is
17323 entirely disabled by the @samp{+nofp} option that follows it.
17325 Most extension names are generically named, but have an effect that is
17326 dependent upon the architecture to which it is applied. For example,
17327 the @samp{+simd} option can be applied to both @samp{armv7-a} and
17328 @samp{armv8-a} architectures, but will enable the original ARMv7-A
17329 Advanced SIMD (Neon) extensions for @samp{armv7-a} and the ARMv8-A
17330 variant for @samp{armv8-a}.
17332 The table below lists the supported extensions for each architecture.
17333 Architectures not mentioned do not support any extensions.
17346 The VFPv2 floating-point instructions. The extension @samp{+vfpv2} can be
17347 used as an alias for this extension.
17350 Disable the floating-point instructions.
17354 The common subset of the ARMv7-A, ARMv7-R and ARMv7-M architectures.
17357 The VFPv3 floating-point instructions, with 16 double-precision
17358 registers. The extension @samp{+vfpv3-d16} can be used as an alias
17359 for this extension. Note that floating-point is not supported by the
17360 base ARMv7-M architecture, but is compatible with both the ARMv7-A and
17361 ARMv7-R architectures.
17364 Disable the floating-point instructions.
17370 The multiprocessing extension.
17373 The security extension.
17376 The VFPv3 floating-point instructions, with 16 double-precision
17377 registers. The extension @samp{+vfpv3-d16} can be used as an alias
17378 for this extension.
17381 The Advanced SIMD (Neon) v1 and the VFPv3 floating-point instructions.
17382 The extensions @samp{+neon} and @samp{+neon-vfpv3} can be used as aliases
17383 for this extension.
17386 The VFPv3 floating-point instructions, with 32 double-precision
17389 @item +vfpv3-d16-fp16
17390 The VFPv3 floating-point instructions, with 16 double-precision
17391 registers and the half-precision floating-point conversion operations.
17394 The VFPv3 floating-point instructions, with 32 double-precision
17395 registers and the half-precision floating-point conversion operations.
17398 The VFPv4 floating-point instructions, with 16 double-precision
17402 The VFPv4 floating-point instructions, with 32 double-precision
17406 The Advanced SIMD (Neon) v1 and the VFPv3 floating-point instructions, with
17407 the half-precision floating-point conversion operations.
17410 The Advanced SIMD (Neon) v2 and the VFPv4 floating-point instructions.
17413 Disable the Advanced SIMD instructions (does not disable floating point).
17416 Disable the floating-point and Advanced SIMD instructions.
17420 The extended version of the ARMv7-A architecture with support for
17424 The VFPv4 floating-point instructions, with 16 double-precision registers.
17425 The extension @samp{+vfpv4-d16} can be used as an alias for this extension.
17428 The Advanced SIMD (Neon) v2 and the VFPv4 floating-point instructions. The
17429 extension @samp{+neon-vfpv4} can be used as an alias for this extension.
17432 The VFPv3 floating-point instructions, with 16 double-precision
17436 The VFPv3 floating-point instructions, with 32 double-precision
17439 @item +vfpv3-d16-fp16
17440 The VFPv3 floating-point instructions, with 16 double-precision
17441 registers and the half-precision floating-point conversion operations.
17444 The VFPv3 floating-point instructions, with 32 double-precision
17445 registers and the half-precision floating-point conversion operations.
17448 The VFPv4 floating-point instructions, with 16 double-precision
17452 The VFPv4 floating-point instructions, with 32 double-precision
17456 The Advanced SIMD (Neon) v1 and the VFPv3 floating-point instructions.
17457 The extension @samp{+neon-vfpv3} can be used as an alias for this extension.
17460 The Advanced SIMD (Neon) v1 and the VFPv3 floating-point instructions, with
17461 the half-precision floating-point conversion operations.
17464 Disable the Advanced SIMD instructions (does not disable floating point).
17467 Disable the floating-point and Advanced SIMD instructions.
17473 The Cyclic Redundancy Check (CRC) instructions.
17475 The ARMv8-A Advanced SIMD and floating-point instructions.
17477 The cryptographic instructions.
17479 Disable the cryptographic instructions.
17481 Disable the floating-point, Advanced SIMD and cryptographic instructions.
17483 Speculation Barrier Instruction.
17485 Execution and Data Prediction Restriction Instructions.
17491 The ARMv8.1-A Advanced SIMD and floating-point instructions.
17494 The cryptographic instructions. This also enables the Advanced SIMD and
17495 floating-point instructions.
17498 Disable the cryptographic instructions.
17501 Disable the floating-point, Advanced SIMD and cryptographic instructions.
17504 Speculation Barrier Instruction.
17507 Execution and Data Prediction Restriction Instructions.
17514 The half-precision floating-point data processing instructions.
17515 This also enables the Advanced SIMD and floating-point instructions.
17518 The half-precision floating-point fmla extension. This also enables
17519 the half-precision floating-point extension and Advanced SIMD and
17520 floating-point instructions.
17523 The ARMv8.1-A Advanced SIMD and floating-point instructions.
17526 The cryptographic instructions. This also enables the Advanced SIMD and
17527 floating-point instructions.
17530 Enable the Dot Product extension. This also enables Advanced SIMD instructions.
17533 Disable the cryptographic extension.
17536 Disable the floating-point, Advanced SIMD and cryptographic instructions.
17539 Speculation Barrier Instruction.
17542 Execution and Data Prediction Restriction Instructions.
17548 The half-precision floating-point data processing instructions.
17549 This also enables the Advanced SIMD and floating-point instructions as well
17550 as the Dot Product extension and the half-precision floating-point fmla
17554 The ARMv8.3-A Advanced SIMD and floating-point instructions as well as the
17555 Dot Product extension.
17558 The cryptographic instructions. This also enables the Advanced SIMD and
17559 floating-point instructions as well as the Dot Product extension.
17562 Disable the cryptographic extension.
17565 Disable the floating-point, Advanced SIMD and cryptographic instructions.
17568 Speculation Barrier Instruction.
17571 Execution and Data Prediction Restriction Instructions.
17577 The half-precision floating-point data processing instructions.
17578 This also enables the Advanced SIMD and floating-point instructions as well
17579 as the Dot Product extension and the half-precision floating-point fmla
17583 The ARMv8.3-A Advanced SIMD and floating-point instructions as well as the
17584 Dot Product extension.
17587 The cryptographic instructions. This also enables the Advanced SIMD and
17588 floating-point instructions as well as the Dot Product extension.
17591 Disable the cryptographic extension.
17594 Disable the floating-point, Advanced SIMD and cryptographic instructions.
17600 The single-precision VFPv3 floating-point instructions. The extension
17601 @samp{+vfpv3xd} can be used as an alias for this extension.
17604 The VFPv3 floating-point instructions with 16 double-precision registers.
17605 The extension +vfpv3-d16 can be used as an alias for this extension.
17607 @item +vfpv3xd-d16-fp16
17608 The single-precision VFPv3 floating-point instructions with 16 double-precision
17609 registers and the half-precision floating-point conversion operations.
17611 @item +vfpv3-d16-fp16
17612 The VFPv3 floating-point instructions with 16 double-precision
17613 registers and the half-precision floating-point conversion operations.
17616 Disable the floating-point extension.
17619 The ARM-state integer division instructions.
17622 Disable the ARM-state integer division extension.
17628 The single-precision VFPv4 floating-point instructions.
17631 The single-precision FPv5 floating-point instructions.
17634 The single- and double-precision FPv5 floating-point instructions.
17637 Disable the floating-point extensions.
17643 The DSP instructions.
17646 Disable the DSP extension.
17649 The single-precision floating-point instructions.
17652 The single- and double-precision floating-point instructions.
17655 Disable the floating-point extension.
17661 The Cyclic Redundancy Check (CRC) instructions.
17663 The single-precision FPv5 floating-point instructions.
17665 The ARMv8-A Advanced SIMD and floating-point instructions.
17667 The cryptographic instructions.
17669 Disable the cryptographic instructions.
17671 Disable the floating-point, Advanced SIMD and cryptographic instructions.
17676 @option{-march=native} causes the compiler to auto-detect the architecture
17677 of the build computer. At present, this feature is only supported on
17678 GNU/Linux, and not all architectures are recognized. If the auto-detect
17679 is unsuccessful the option has no effect.
17681 @item -mtune=@var{name}
17683 This option specifies the name of the target ARM processor for
17684 which GCC should tune the performance of the code.
17685 For some ARM implementations better performance can be obtained by using
17687 Permissible names are: @samp{arm7tdmi}, @samp{arm7tdmi-s}, @samp{arm710t},
17688 @samp{arm720t}, @samp{arm740t}, @samp{strongarm}, @samp{strongarm110},
17689 @samp{strongarm1100}, 0@samp{strongarm1110}, @samp{arm8}, @samp{arm810},
17690 @samp{arm9}, @samp{arm9e}, @samp{arm920}, @samp{arm920t}, @samp{arm922t},
17691 @samp{arm946e-s}, @samp{arm966e-s}, @samp{arm968e-s}, @samp{arm926ej-s},
17692 @samp{arm940t}, @samp{arm9tdmi}, @samp{arm10tdmi}, @samp{arm1020t},
17693 @samp{arm1026ej-s}, @samp{arm10e}, @samp{arm1020e}, @samp{arm1022e},
17694 @samp{arm1136j-s}, @samp{arm1136jf-s}, @samp{mpcore}, @samp{mpcorenovfp},
17695 @samp{arm1156t2-s}, @samp{arm1156t2f-s}, @samp{arm1176jz-s}, @samp{arm1176jzf-s},
17696 @samp{generic-armv7-a}, @samp{cortex-a5}, @samp{cortex-a7}, @samp{cortex-a8},
17697 @samp{cortex-a9}, @samp{cortex-a12}, @samp{cortex-a15}, @samp{cortex-a17},
17698 @samp{cortex-a32}, @samp{cortex-a35}, @samp{cortex-a53}, @samp{cortex-a55},
17699 @samp{cortex-a57}, @samp{cortex-a72}, @samp{cortex-a73}, @samp{cortex-a75},
17700 @samp{cortex-a76}, @samp{cortex-a76ae}, @samp{cortex-a77},
17701 @samp{ares}, @samp{cortex-r4}, @samp{cortex-r4f},
17702 @samp{cortex-r5}, @samp{cortex-r7}, @samp{cortex-r8}, @samp{cortex-r52},
17703 @samp{cortex-m0}, @samp{cortex-m0plus}, @samp{cortex-m1}, @samp{cortex-m3},
17704 @samp{cortex-m4}, @samp{cortex-m7}, @samp{cortex-m23}, @samp{cortex-m33},
17705 @samp{cortex-m35p},
17706 @samp{cortex-m1.small-multiply}, @samp{cortex-m0.small-multiply},
17707 @samp{cortex-m0plus.small-multiply}, @samp{exynos-m1}, @samp{marvell-pj4},
17708 @samp{neoverse-n1}, @samp{xscale}, @samp{iwmmxt}, @samp{iwmmxt2},
17709 @samp{ep9312}, @samp{fa526}, @samp{fa626}, @samp{fa606te}, @samp{fa626te},
17710 @samp{fmp626}, @samp{fa726te}, @samp{xgene1}.
17712 Additionally, this option can specify that GCC should tune the performance
17713 of the code for a big.LITTLE system. Permissible names are:
17714 @samp{cortex-a15.cortex-a7}, @samp{cortex-a17.cortex-a7},
17715 @samp{cortex-a57.cortex-a53}, @samp{cortex-a72.cortex-a53},
17716 @samp{cortex-a72.cortex-a35}, @samp{cortex-a73.cortex-a53},
17717 @samp{cortex-a75.cortex-a55}, @samp{cortex-a76.cortex-a55}.
17719 @option{-mtune=generic-@var{arch}} specifies that GCC should tune the
17720 performance for a blend of processors within architecture @var{arch}.
17721 The aim is to generate code that run well on the current most popular
17722 processors, balancing between optimizations that benefit some CPUs in the
17723 range, and avoiding performance pitfalls of other CPUs. The effects of
17724 this option may change in future GCC versions as CPU models come and go.
17726 @option{-mtune} permits the same extension options as @option{-mcpu}, but
17727 the extension options do not affect the tuning of the generated code.
17729 @option{-mtune=native} causes the compiler to auto-detect the CPU
17730 of the build computer. At present, this feature is only supported on
17731 GNU/Linux, and not all architectures are recognized. If the auto-detect is
17732 unsuccessful the option has no effect.
17734 @item -mcpu=@var{name}@r{[}+extension@dots{}@r{]}
17736 This specifies the name of the target ARM processor. GCC uses this name
17737 to derive the name of the target ARM architecture (as if specified
17738 by @option{-march}) and the ARM processor type for which to tune for
17739 performance (as if specified by @option{-mtune}). Where this option
17740 is used in conjunction with @option{-march} or @option{-mtune},
17741 those options take precedence over the appropriate part of this option.
17743 Many of the supported CPUs implement optional architectural
17744 extensions. Where this is so the architectural extensions are
17745 normally enabled by default. If implementations that lack the
17746 extension exist, then the extension syntax can be used to disable
17747 those extensions that have been omitted. For floating-point and
17748 Advanced SIMD (Neon) instructions, the settings of the options
17749 @option{-mfloat-abi} and @option{-mfpu} must also be considered:
17750 floating-point and Advanced SIMD instructions will only be used if
17751 @option{-mfloat-abi} is not set to @samp{soft}; and any setting of
17752 @option{-mfpu} other than @samp{auto} will override the available
17753 floating-point and SIMD extension instructions.
17755 For example, @samp{cortex-a9} can be found in three major
17756 configurations: integer only, with just a floating-point unit or with
17757 floating-point and Advanced SIMD. The default is to enable all the
17758 instructions, but the extensions @samp{+nosimd} and @samp{+nofp} can
17759 be used to disable just the SIMD or both the SIMD and floating-point
17760 instructions respectively.
17762 Permissible names for this option are the same as those for
17765 The following extension options are common to the listed CPUs:
17769 Disable the DSP instructions on @samp{cortex-m33}, @samp{cortex-m35p}.
17772 Disables the floating-point instructions on @samp{arm9e},
17773 @samp{arm946e-s}, @samp{arm966e-s}, @samp{arm968e-s}, @samp{arm10e},
17774 @samp{arm1020e}, @samp{arm1022e}, @samp{arm926ej-s},
17775 @samp{arm1026ej-s}, @samp{cortex-r5}, @samp{cortex-r7}, @samp{cortex-r8},
17776 @samp{cortex-m4}, @samp{cortex-m7}, @samp{cortex-m33} and @samp{cortex-m35p}.
17777 Disables the floating-point and SIMD instructions on
17778 @samp{generic-armv7-a}, @samp{cortex-a5}, @samp{cortex-a7},
17779 @samp{cortex-a8}, @samp{cortex-a9}, @samp{cortex-a12},
17780 @samp{cortex-a15}, @samp{cortex-a17}, @samp{cortex-a15.cortex-a7},
17781 @samp{cortex-a17.cortex-a7}, @samp{cortex-a32}, @samp{cortex-a35},
17782 @samp{cortex-a53} and @samp{cortex-a55}.
17785 Disables the double-precision component of the floating-point instructions
17786 on @samp{cortex-r5}, @samp{cortex-r7}, @samp{cortex-r8}, @samp{cortex-r52} and
17790 Disables the SIMD (but not floating-point) instructions on
17791 @samp{generic-armv7-a}, @samp{cortex-a5}, @samp{cortex-a7}
17792 and @samp{cortex-a9}.
17795 Enables the cryptographic instructions on @samp{cortex-a32},
17796 @samp{cortex-a35}, @samp{cortex-a53}, @samp{cortex-a55}, @samp{cortex-a57},
17797 @samp{cortex-a72}, @samp{cortex-a73}, @samp{cortex-a75}, @samp{exynos-m1},
17798 @samp{xgene1}, @samp{cortex-a57.cortex-a53}, @samp{cortex-a72.cortex-a53},
17799 @samp{cortex-a73.cortex-a35}, @samp{cortex-a73.cortex-a53} and
17800 @samp{cortex-a75.cortex-a55}.
17803 Additionally the @samp{generic-armv7-a} pseudo target defaults to
17804 VFPv3 with 16 double-precision registers. It supports the following
17805 extension options: @samp{mp}, @samp{sec}, @samp{vfpv3-d16},
17806 @samp{vfpv3}, @samp{vfpv3-d16-fp16}, @samp{vfpv3-fp16},
17807 @samp{vfpv4-d16}, @samp{vfpv4}, @samp{neon}, @samp{neon-vfpv3},
17808 @samp{neon-fp16}, @samp{neon-vfpv4}. The meanings are the same as for
17809 the extensions to @option{-march=armv7-a}.
17811 @option{-mcpu=generic-@var{arch}} is also permissible, and is
17812 equivalent to @option{-march=@var{arch} -mtune=generic-@var{arch}}.
17813 See @option{-mtune} for more information.
17815 @option{-mcpu=native} causes the compiler to auto-detect the CPU
17816 of the build computer. At present, this feature is only supported on
17817 GNU/Linux, and not all architectures are recognized. If the auto-detect
17818 is unsuccessful the option has no effect.
17820 @item -mfpu=@var{name}
17822 This specifies what floating-point hardware (or hardware emulation) is
17823 available on the target. Permissible names are: @samp{auto}, @samp{vfpv2},
17825 @samp{vfpv3-fp16}, @samp{vfpv3-d16}, @samp{vfpv3-d16-fp16}, @samp{vfpv3xd},
17826 @samp{vfpv3xd-fp16}, @samp{neon-vfpv3}, @samp{neon-fp16}, @samp{vfpv4},
17827 @samp{vfpv4-d16}, @samp{fpv4-sp-d16}, @samp{neon-vfpv4},
17828 @samp{fpv5-d16}, @samp{fpv5-sp-d16},
17829 @samp{fp-armv8}, @samp{neon-fp-armv8} and @samp{crypto-neon-fp-armv8}.
17830 Note that @samp{neon} is an alias for @samp{neon-vfpv3} and @samp{vfp}
17831 is an alias for @samp{vfpv2}.
17833 The setting @samp{auto} is the default and is special. It causes the
17834 compiler to select the floating-point and Advanced SIMD instructions
17835 based on the settings of @option{-mcpu} and @option{-march}.
17837 If the selected floating-point hardware includes the NEON extension
17838 (e.g.@: @option{-mfpu=neon}), note that floating-point
17839 operations are not generated by GCC's auto-vectorization pass unless
17840 @option{-funsafe-math-optimizations} is also specified. This is
17841 because NEON hardware does not fully implement the IEEE 754 standard for
17842 floating-point arithmetic (in particular denormal values are treated as
17843 zero), so the use of NEON instructions may lead to a loss of precision.
17845 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}).
17847 @item -mfp16-format=@var{name}
17848 @opindex mfp16-format
17849 Specify the format of the @code{__fp16} half-precision floating-point type.
17850 Permissible names are @samp{none}, @samp{ieee}, and @samp{alternative};
17851 the default is @samp{none}, in which case the @code{__fp16} type is not
17852 defined. @xref{Half-Precision}, for more information.
17854 @item -mstructure-size-boundary=@var{n}
17855 @opindex mstructure-size-boundary
17856 The sizes of all structures and unions are rounded up to a multiple
17857 of the number of bits set by this option. Permissible values are 8, 32
17858 and 64. The default value varies for different toolchains. For the COFF
17859 targeted toolchain the default value is 8. A value of 64 is only allowed
17860 if the underlying ABI supports it.
17862 Specifying a larger number can produce faster, more efficient code, but
17863 can also increase the size of the program. Different values are potentially
17864 incompatible. Code compiled with one value cannot necessarily expect to
17865 work with code or libraries compiled with another value, if they exchange
17866 information using structures or unions.
17868 This option is deprecated.
17870 @item -mabort-on-noreturn
17871 @opindex mabort-on-noreturn
17872 Generate a call to the function @code{abort} at the end of a
17873 @code{noreturn} function. It is executed if the function tries to
17877 @itemx -mno-long-calls
17878 @opindex mlong-calls
17879 @opindex mno-long-calls
17880 Tells the compiler to perform function calls by first loading the
17881 address of the function into a register and then performing a subroutine
17882 call on this register. This switch is needed if the target function
17883 lies outside of the 64-megabyte addressing range of the offset-based
17884 version of subroutine call instruction.
17886 Even if this switch is enabled, not all function calls are turned
17887 into long calls. The heuristic is that static functions, functions
17888 that have the @code{short_call} attribute, functions that are inside
17889 the scope of a @code{#pragma no_long_calls} directive, and functions whose
17890 definitions have already been compiled within the current compilation
17891 unit are not turned into long calls. The exceptions to this rule are
17892 that weak function definitions, functions with the @code{long_call}
17893 attribute or the @code{section} attribute, and functions that are within
17894 the scope of a @code{#pragma long_calls} directive are always
17895 turned into long calls.
17897 This feature is not enabled by default. Specifying
17898 @option{-mno-long-calls} restores the default behavior, as does
17899 placing the function calls within the scope of a @code{#pragma
17900 long_calls_off} directive. Note these switches have no effect on how
17901 the compiler generates code to handle function calls via function
17904 @item -msingle-pic-base
17905 @opindex msingle-pic-base
17906 Treat the register used for PIC addressing as read-only, rather than
17907 loading it in the prologue for each function. The runtime system is
17908 responsible for initializing this register with an appropriate value
17909 before execution begins.
17911 @item -mpic-register=@var{reg}
17912 @opindex mpic-register
17913 Specify the register to be used for PIC addressing.
17914 For standard PIC base case, the default is any suitable register
17915 determined by compiler. For single PIC base case, the default is
17916 @samp{R9} if target is EABI based or stack-checking is enabled,
17917 otherwise the default is @samp{R10}.
17919 @item -mpic-data-is-text-relative
17920 @opindex mpic-data-is-text-relative
17921 Assume that the displacement between the text and data segments is fixed
17922 at static link time. This permits using PC-relative addressing
17923 operations to access data known to be in the data segment. For
17924 non-VxWorks RTP targets, this option is enabled by default. When
17925 disabled on such targets, it will enable @option{-msingle-pic-base} by
17928 @item -mpoke-function-name
17929 @opindex mpoke-function-name
17930 Write the name of each function into the text section, directly
17931 preceding the function prologue. The generated code is similar to this:
17935 .ascii "arm_poke_function_name", 0
17938 .word 0xff000000 + (t1 - t0)
17939 arm_poke_function_name
17941 stmfd sp!, @{fp, ip, lr, pc@}
17945 When performing a stack backtrace, code can inspect the value of
17946 @code{pc} stored at @code{fp + 0}. If the trace function then looks at
17947 location @code{pc - 12} and the top 8 bits are set, then we know that
17948 there is a function name embedded immediately preceding this location
17949 and has length @code{((pc[-3]) & 0xff000000)}.
17956 Select between generating code that executes in ARM and Thumb
17957 states. The default for most configurations is to generate code
17958 that executes in ARM state, but the default can be changed by
17959 configuring GCC with the @option{--with-mode=}@var{state}
17962 You can also override the ARM and Thumb mode for each function
17963 by using the @code{target("thumb")} and @code{target("arm")} function attributes
17964 (@pxref{ARM Function Attributes}) or pragmas (@pxref{Function Specific Option Pragmas}).
17967 @opindex mflip-thumb
17968 Switch ARM/Thumb modes on alternating functions.
17969 This option is provided for regression testing of mixed Thumb/ARM code
17970 generation, and is not intended for ordinary use in compiling code.
17973 @opindex mtpcs-frame
17974 Generate a stack frame that is compliant with the Thumb Procedure Call
17975 Standard for all non-leaf functions. (A leaf function is one that does
17976 not call any other functions.) The default is @option{-mno-tpcs-frame}.
17978 @item -mtpcs-leaf-frame
17979 @opindex mtpcs-leaf-frame
17980 Generate a stack frame that is compliant with the Thumb Procedure Call
17981 Standard for all leaf functions. (A leaf function is one that does
17982 not call any other functions.) The default is @option{-mno-apcs-leaf-frame}.
17984 @item -mcallee-super-interworking
17985 @opindex mcallee-super-interworking
17986 Gives all externally visible functions in the file being compiled an ARM
17987 instruction set header which switches to Thumb mode before executing the
17988 rest of the function. This allows these functions to be called from
17989 non-interworking code. This option is not valid in AAPCS configurations
17990 because interworking is enabled by default.
17992 @item -mcaller-super-interworking
17993 @opindex mcaller-super-interworking
17994 Allows calls via function pointers (including virtual functions) to
17995 execute correctly regardless of whether the target code has been
17996 compiled for interworking or not. There is a small overhead in the cost
17997 of executing a function pointer if this option is enabled. This option
17998 is not valid in AAPCS configurations because interworking is enabled
18001 @item -mtp=@var{name}
18003 Specify the access model for the thread local storage pointer. The valid
18004 models are @samp{soft}, which generates calls to @code{__aeabi_read_tp},
18005 @samp{cp15}, which fetches the thread pointer from @code{cp15} directly
18006 (supported in the arm6k architecture), and @samp{auto}, which uses the
18007 best available method for the selected processor. The default setting is
18010 @item -mtls-dialect=@var{dialect}
18011 @opindex mtls-dialect
18012 Specify the dialect to use for accessing thread local storage. Two
18013 @var{dialect}s are supported---@samp{gnu} and @samp{gnu2}. The
18014 @samp{gnu} dialect selects the original GNU scheme for supporting
18015 local and global dynamic TLS models. The @samp{gnu2} dialect
18016 selects the GNU descriptor scheme, which provides better performance
18017 for shared libraries. The GNU descriptor scheme is compatible with
18018 the original scheme, but does require new assembler, linker and
18019 library support. Initial and local exec TLS models are unaffected by
18020 this option and always use the original scheme.
18022 @item -mword-relocations
18023 @opindex mword-relocations
18024 Only generate absolute relocations on word-sized values (i.e.@: R_ARM_ABS32).
18025 This is enabled by default on targets (uClinux, SymbianOS) where the runtime
18026 loader imposes this restriction, and when @option{-fpic} or @option{-fPIC}
18027 is specified. This option conflicts with @option{-mslow-flash-data}.
18029 @item -mfix-cortex-m3-ldrd
18030 @opindex mfix-cortex-m3-ldrd
18031 Some Cortex-M3 cores can cause data corruption when @code{ldrd} instructions
18032 with overlapping destination and base registers are used. This option avoids
18033 generating these instructions. This option is enabled by default when
18034 @option{-mcpu=cortex-m3} is specified.
18036 @item -munaligned-access
18037 @itemx -mno-unaligned-access
18038 @opindex munaligned-access
18039 @opindex mno-unaligned-access
18040 Enables (or disables) reading and writing of 16- and 32- bit values
18041 from addresses that are not 16- or 32- bit aligned. By default
18042 unaligned access is disabled for all pre-ARMv6, all ARMv6-M and for
18043 ARMv8-M Baseline architectures, and enabled for all other
18044 architectures. If unaligned access is not enabled then words in packed
18045 data structures are accessed a byte at a time.
18047 The ARM attribute @code{Tag_CPU_unaligned_access} is set in the
18048 generated object file to either true or false, depending upon the
18049 setting of this option. If unaligned access is enabled then the
18050 preprocessor symbol @code{__ARM_FEATURE_UNALIGNED} is also
18053 @item -mneon-for-64bits
18054 @opindex mneon-for-64bits
18055 This option is deprecated and has no effect.
18057 @item -mslow-flash-data
18058 @opindex mslow-flash-data
18059 Assume loading data from flash is slower than fetching instruction.
18060 Therefore literal load is minimized for better performance.
18061 This option is only supported when compiling for ARMv7 M-profile and
18062 off by default. It conflicts with @option{-mword-relocations}.
18064 @item -masm-syntax-unified
18065 @opindex masm-syntax-unified
18066 Assume inline assembler is using unified asm syntax. The default is
18067 currently off which implies divided syntax. This option has no impact
18068 on Thumb2. However, this may change in future releases of GCC.
18069 Divided syntax should be considered deprecated.
18071 @item -mrestrict-it
18072 @opindex mrestrict-it
18073 Restricts generation of IT blocks to conform to the rules of ARMv8-A.
18074 IT blocks can only contain a single 16-bit instruction from a select
18075 set of instructions. This option is on by default for ARMv8-A Thumb mode.
18077 @item -mprint-tune-info
18078 @opindex mprint-tune-info
18079 Print CPU tuning information as comment in assembler file. This is
18080 an option used only for regression testing of the compiler and not
18081 intended for ordinary use in compiling code. This option is disabled
18084 @item -mverbose-cost-dump
18085 @opindex mverbose-cost-dump
18086 Enable verbose cost model dumping in the debug dump files. This option is
18087 provided for use in debugging the compiler.
18090 @opindex mpure-code
18091 Do not allow constant data to be placed in code sections.
18092 Additionally, when compiling for ELF object format give all text sections the
18093 ELF processor-specific section attribute @code{SHF_ARM_PURECODE}. This option
18094 is only available when generating non-pic code for M-profile targets with the
18099 Generate secure code as per the "ARMv8-M Security Extensions: Requirements on
18100 Development Tools Engineering Specification", which can be found on
18101 @url{http://infocenter.arm.com/help/topic/com.arm.doc.ecm0359818/ECM0359818_armv8m_security_extensions_reqs_on_dev_tools_1_0.pdf}.
18107 Select the FDPIC ABI, which uses 64-bit function descriptors to
18108 represent pointers to functions. When the compiler is configured for
18109 @code{arm-*-uclinuxfdpiceabi} targets, this option is on by default
18110 and implies @option{-fPIE} if none of the PIC/PIE-related options is
18111 provided. On other targets, it only enables the FDPIC-specific code
18112 generation features, and the user should explicitly provide the
18113 PIC/PIE-related options as needed.
18115 Note that static linking is not supported because it would still
18116 involve the dynamic linker when the program self-relocates. If such
18117 behavior is acceptable, use -static and -Wl,-dynamic-linker options.
18119 The opposite @option{-mno-fdpic} option is useful (and required) to
18120 build the Linux kernel using the same (@code{arm-*-uclinuxfdpiceabi})
18121 toolchain as the one used to build the userland programs.
18126 @subsection AVR Options
18127 @cindex AVR Options
18129 These options are defined for AVR implementations:
18132 @item -mmcu=@var{mcu}
18134 Specify Atmel AVR instruction set architectures (ISA) or MCU type.
18136 The default for this option is@tie{}@samp{avr2}.
18138 GCC supports the following AVR devices and ISAs:
18140 @include avr-mmcu.texi
18145 Assume that all data in static storage can be accessed by LDS / STS
18146 instructions. This option has only an effect on reduced Tiny devices like
18147 ATtiny40. See also the @code{absdata}
18148 @ref{AVR Variable Attributes,variable attribute}.
18150 @item -maccumulate-args
18151 @opindex maccumulate-args
18152 Accumulate outgoing function arguments and acquire/release the needed
18153 stack space for outgoing function arguments once in function
18154 prologue/epilogue. Without this option, outgoing arguments are pushed
18155 before calling a function and popped afterwards.
18157 Popping the arguments after the function call can be expensive on
18158 AVR so that accumulating the stack space might lead to smaller
18159 executables because arguments need not be removed from the
18160 stack after such a function call.
18162 This option can lead to reduced code size for functions that perform
18163 several calls to functions that get their arguments on the stack like
18164 calls to printf-like functions.
18166 @item -mbranch-cost=@var{cost}
18167 @opindex mbranch-cost
18168 Set the branch costs for conditional branch instructions to
18169 @var{cost}. Reasonable values for @var{cost} are small, non-negative
18170 integers. The default branch cost is 0.
18172 @item -mcall-prologues
18173 @opindex mcall-prologues
18174 Functions prologues/epilogues are expanded as calls to appropriate
18175 subroutines. Code size is smaller.
18177 @item -mgas-isr-prologues
18178 @opindex mgas-isr-prologues
18179 Interrupt service routines (ISRs) may use the @code{__gcc_isr} pseudo
18180 instruction supported by GNU Binutils.
18181 If this option is on, the feature can still be disabled for individual
18182 ISRs by means of the @ref{AVR Function Attributes,,@code{no_gccisr}}
18183 function attribute. This feature is activated per default
18184 if optimization is on (but not with @option{-Og}, @pxref{Optimize Options}),
18185 and if GNU Binutils support @w{@uref{https://sourceware.org/PR21683,PR21683}}.
18189 Assume @code{int} to be 8-bit integer. This affects the sizes of all types: a
18190 @code{char} is 1 byte, an @code{int} is 1 byte, a @code{long} is 2 bytes,
18191 and @code{long long} is 4 bytes. Please note that this option does not
18192 conform to the C standards, but it results in smaller code
18195 @item -mmain-is-OS_task
18196 @opindex mmain-is-OS_task
18197 Do not save registers in @code{main}. The effect is the same like
18198 attaching attribute @ref{AVR Function Attributes,,@code{OS_task}}
18199 to @code{main}. It is activated per default if optimization is on.
18201 @item -mn-flash=@var{num}
18203 Assume that the flash memory has a size of
18204 @var{num} times 64@tie{}KiB.
18206 @item -mno-interrupts
18207 @opindex mno-interrupts
18208 Generated code is not compatible with hardware interrupts.
18209 Code size is smaller.
18213 Try to replace @code{CALL} resp.@: @code{JMP} instruction by the shorter
18214 @code{RCALL} resp.@: @code{RJMP} instruction if applicable.
18215 Setting @option{-mrelax} just adds the @option{--mlink-relax} option to
18216 the assembler's command line and the @option{--relax} option to the
18217 linker's command line.
18219 Jump relaxing is performed by the linker because jump offsets are not
18220 known before code is located. Therefore, the assembler code generated by the
18221 compiler is the same, but the instructions in the executable may
18222 differ from instructions in the assembler code.
18224 Relaxing must be turned on if linker stubs are needed, see the
18225 section on @code{EIND} and linker stubs below.
18229 Assume that the device supports the Read-Modify-Write
18230 instructions @code{XCH}, @code{LAC}, @code{LAS} and @code{LAT}.
18232 @item -mshort-calls
18233 @opindex mshort-calls
18235 Assume that @code{RJMP} and @code{RCALL} can target the whole
18238 This option is used internally for multilib selection. It is
18239 not an optimization option, and you don't need to set it by hand.
18243 Treat the stack pointer register as an 8-bit register,
18244 i.e.@: assume the high byte of the stack pointer is zero.
18245 In general, you don't need to set this option by hand.
18247 This option is used internally by the compiler to select and
18248 build multilibs for architectures @code{avr2} and @code{avr25}.
18249 These architectures mix devices with and without @code{SPH}.
18250 For any setting other than @option{-mmcu=avr2} or @option{-mmcu=avr25}
18251 the compiler driver adds or removes this option from the compiler
18252 proper's command line, because the compiler then knows if the device
18253 or architecture has an 8-bit stack pointer and thus no @code{SPH}
18258 Use address register @code{X} in a way proposed by the hardware. This means
18259 that @code{X} is only used in indirect, post-increment or
18260 pre-decrement addressing.
18262 Without this option, the @code{X} register may be used in the same way
18263 as @code{Y} or @code{Z} which then is emulated by additional
18265 For example, loading a value with @code{X+const} addressing with a
18266 small non-negative @code{const < 64} to a register @var{Rn} is
18270 adiw r26, const ; X += const
18271 ld @var{Rn}, X ; @var{Rn} = *X
18272 sbiw r26, const ; X -= const
18276 @opindex mtiny-stack
18277 Only change the lower 8@tie{}bits of the stack pointer.
18279 @item -mfract-convert-truncate
18280 @opindex mfract-convert-truncate
18281 Allow to use truncation instead of rounding towards zero for fractional fixed-point types.
18284 @opindex nodevicelib
18285 Don't link against AVR-LibC's device specific library @code{lib<mcu>.a}.
18287 @item -Waddr-space-convert
18288 @opindex Waddr-space-convert
18289 @opindex Wno-addr-space-convert
18290 Warn about conversions between address spaces in the case where the
18291 resulting address space is not contained in the incoming address space.
18293 @item -Wmisspelled-isr
18294 @opindex Wmisspelled-isr
18295 @opindex Wno-misspelled-isr
18296 Warn if the ISR is misspelled, i.e.@: without __vector prefix.
18297 Enabled by default.
18300 @subsubsection @code{EIND} and Devices with More Than 128 Ki Bytes of Flash
18301 @cindex @code{EIND}
18302 Pointers in the implementation are 16@tie{}bits wide.
18303 The address of a function or label is represented as word address so
18304 that indirect jumps and calls can target any code address in the
18305 range of 64@tie{}Ki words.
18307 In order to facilitate indirect jump on devices with more than 128@tie{}Ki
18308 bytes of program memory space, there is a special function register called
18309 @code{EIND} that serves as most significant part of the target address
18310 when @code{EICALL} or @code{EIJMP} instructions are used.
18312 Indirect jumps and calls on these devices are handled as follows by
18313 the compiler and are subject to some limitations:
18318 The compiler never sets @code{EIND}.
18321 The compiler uses @code{EIND} implicitly in @code{EICALL}/@code{EIJMP}
18322 instructions or might read @code{EIND} directly in order to emulate an
18323 indirect call/jump by means of a @code{RET} instruction.
18326 The compiler assumes that @code{EIND} never changes during the startup
18327 code or during the application. In particular, @code{EIND} is not
18328 saved/restored in function or interrupt service routine
18332 For indirect calls to functions and computed goto, the linker
18333 generates @emph{stubs}. Stubs are jump pads sometimes also called
18334 @emph{trampolines}. Thus, the indirect call/jump jumps to such a stub.
18335 The stub contains a direct jump to the desired address.
18338 Linker relaxation must be turned on so that the linker generates
18339 the stubs correctly in all situations. See the compiler option
18340 @option{-mrelax} and the linker option @option{--relax}.
18341 There are corner cases where the linker is supposed to generate stubs
18342 but aborts without relaxation and without a helpful error message.
18345 The default linker script is arranged for code with @code{EIND = 0}.
18346 If code is supposed to work for a setup with @code{EIND != 0}, a custom
18347 linker script has to be used in order to place the sections whose
18348 name start with @code{.trampolines} into the segment where @code{EIND}
18352 The startup code from libgcc never sets @code{EIND}.
18353 Notice that startup code is a blend of code from libgcc and AVR-LibC.
18354 For the impact of AVR-LibC on @code{EIND}, see the
18355 @w{@uref{http://nongnu.org/avr-libc/user-manual/,AVR-LibC user manual}}.
18358 It is legitimate for user-specific startup code to set up @code{EIND}
18359 early, for example by means of initialization code located in
18360 section @code{.init3}. Such code runs prior to general startup code
18361 that initializes RAM and calls constructors, but after the bit
18362 of startup code from AVR-LibC that sets @code{EIND} to the segment
18363 where the vector table is located.
18365 #include <avr/io.h>
18368 __attribute__((section(".init3"),naked,used,no_instrument_function))
18369 init3_set_eind (void)
18371 __asm volatile ("ldi r24,pm_hh8(__trampolines_start)\n\t"
18372 "out %i0,r24" :: "n" (&EIND) : "r24","memory");
18377 The @code{__trampolines_start} symbol is defined in the linker script.
18380 Stubs are generated automatically by the linker if
18381 the following two conditions are met:
18384 @item The address of a label is taken by means of the @code{gs} modifier
18385 (short for @emph{generate stubs}) like so:
18387 LDI r24, lo8(gs(@var{func}))
18388 LDI r25, hi8(gs(@var{func}))
18390 @item The final location of that label is in a code segment
18391 @emph{outside} the segment where the stubs are located.
18395 The compiler emits such @code{gs} modifiers for code labels in the
18396 following situations:
18398 @item Taking address of a function or code label.
18399 @item Computed goto.
18400 @item If prologue-save function is used, see @option{-mcall-prologues}
18401 command-line option.
18402 @item Switch/case dispatch tables. If you do not want such dispatch
18403 tables you can specify the @option{-fno-jump-tables} command-line option.
18404 @item C and C++ constructors/destructors called during startup/shutdown.
18405 @item If the tools hit a @code{gs()} modifier explained above.
18409 Jumping to non-symbolic addresses like so is @emph{not} supported:
18414 /* Call function at word address 0x2 */
18415 return ((int(*)(void)) 0x2)();
18419 Instead, a stub has to be set up, i.e.@: the function has to be called
18420 through a symbol (@code{func_4} in the example):
18425 extern int func_4 (void);
18427 /* Call function at byte address 0x4 */
18432 and the application be linked with @option{-Wl,--defsym,func_4=0x4}.
18433 Alternatively, @code{func_4} can be defined in the linker script.
18436 @subsubsection Handling of the @code{RAMPD}, @code{RAMPX}, @code{RAMPY} and @code{RAMPZ} Special Function Registers
18437 @cindex @code{RAMPD}
18438 @cindex @code{RAMPX}
18439 @cindex @code{RAMPY}
18440 @cindex @code{RAMPZ}
18441 Some AVR devices support memories larger than the 64@tie{}KiB range
18442 that can be accessed with 16-bit pointers. To access memory locations
18443 outside this 64@tie{}KiB range, the content of a @code{RAMP}
18444 register is used as high part of the address:
18445 The @code{X}, @code{Y}, @code{Z} address register is concatenated
18446 with the @code{RAMPX}, @code{RAMPY}, @code{RAMPZ} special function
18447 register, respectively, to get a wide address. Similarly,
18448 @code{RAMPD} is used together with direct addressing.
18452 The startup code initializes the @code{RAMP} special function
18453 registers with zero.
18456 If a @ref{AVR Named Address Spaces,named address space} other than
18457 generic or @code{__flash} is used, then @code{RAMPZ} is set
18458 as needed before the operation.
18461 If the device supports RAM larger than 64@tie{}KiB and the compiler
18462 needs to change @code{RAMPZ} to accomplish an operation, @code{RAMPZ}
18463 is reset to zero after the operation.
18466 If the device comes with a specific @code{RAMP} register, the ISR
18467 prologue/epilogue saves/restores that SFR and initializes it with
18468 zero in case the ISR code might (implicitly) use it.
18471 RAM larger than 64@tie{}KiB is not supported by GCC for AVR targets.
18472 If you use inline assembler to read from locations outside the
18473 16-bit address range and change one of the @code{RAMP} registers,
18474 you must reset it to zero after the access.
18478 @subsubsection AVR Built-in Macros
18480 GCC defines several built-in macros so that the user code can test
18481 for the presence or absence of features. Almost any of the following
18482 built-in macros are deduced from device capabilities and thus
18483 triggered by the @option{-mmcu=} command-line option.
18485 For even more AVR-specific built-in macros see
18486 @ref{AVR Named Address Spaces} and @ref{AVR Built-in Functions}.
18491 Build-in macro that resolves to a decimal number that identifies the
18492 architecture and depends on the @option{-mmcu=@var{mcu}} option.
18493 Possible values are:
18495 @code{2}, @code{25}, @code{3}, @code{31}, @code{35},
18496 @code{4}, @code{5}, @code{51}, @code{6}
18498 for @var{mcu}=@code{avr2}, @code{avr25}, @code{avr3}, @code{avr31},
18499 @code{avr35}, @code{avr4}, @code{avr5}, @code{avr51}, @code{avr6},
18504 @code{102}, @code{103}, @code{104},
18505 @code{105}, @code{106}, @code{107}
18507 for @var{mcu}=@code{avrtiny},
18508 @code{avrxmega2}, @code{avrxmega3}, @code{avrxmega4},
18509 @code{avrxmega5}, @code{avrxmega6}, @code{avrxmega7}, respectively.
18510 If @var{mcu} specifies a device, this built-in macro is set
18511 accordingly. For example, with @option{-mmcu=atmega8} the macro is
18512 defined to @code{4}.
18514 @item __AVR_@var{Device}__
18515 Setting @option{-mmcu=@var{device}} defines this built-in macro which reflects
18516 the device's name. For example, @option{-mmcu=atmega8} defines the
18517 built-in macro @code{__AVR_ATmega8__}, @option{-mmcu=attiny261a} defines
18518 @code{__AVR_ATtiny261A__}, etc.
18520 The built-in macros' names follow
18521 the scheme @code{__AVR_@var{Device}__} where @var{Device} is
18522 the device name as from the AVR user manual. The difference between
18523 @var{Device} in the built-in macro and @var{device} in
18524 @option{-mmcu=@var{device}} is that the latter is always lowercase.
18526 If @var{device} is not a device but only a core architecture like
18527 @samp{avr51}, this macro is not defined.
18529 @item __AVR_DEVICE_NAME__
18530 Setting @option{-mmcu=@var{device}} defines this built-in macro to
18531 the device's name. For example, with @option{-mmcu=atmega8} the macro
18532 is defined to @code{atmega8}.
18534 If @var{device} is not a device but only a core architecture like
18535 @samp{avr51}, this macro is not defined.
18537 @item __AVR_XMEGA__
18538 The device / architecture belongs to the XMEGA family of devices.
18540 @item __AVR_HAVE_ELPM__
18541 The device has the @code{ELPM} instruction.
18543 @item __AVR_HAVE_ELPMX__
18544 The device has the @code{ELPM R@var{n},Z} and @code{ELPM
18545 R@var{n},Z+} instructions.
18547 @item __AVR_HAVE_MOVW__
18548 The device has the @code{MOVW} instruction to perform 16-bit
18549 register-register moves.
18551 @item __AVR_HAVE_LPMX__
18552 The device has the @code{LPM R@var{n},Z} and
18553 @code{LPM R@var{n},Z+} instructions.
18555 @item __AVR_HAVE_MUL__
18556 The device has a hardware multiplier.
18558 @item __AVR_HAVE_JMP_CALL__
18559 The device has the @code{JMP} and @code{CALL} instructions.
18560 This is the case for devices with more than 8@tie{}KiB of program
18563 @item __AVR_HAVE_EIJMP_EICALL__
18564 @itemx __AVR_3_BYTE_PC__
18565 The device has the @code{EIJMP} and @code{EICALL} instructions.
18566 This is the case for devices with more than 128@tie{}KiB of program memory.
18567 This also means that the program counter
18568 (PC) is 3@tie{}bytes wide.
18570 @item __AVR_2_BYTE_PC__
18571 The program counter (PC) is 2@tie{}bytes wide. This is the case for devices
18572 with up to 128@tie{}KiB of program memory.
18574 @item __AVR_HAVE_8BIT_SP__
18575 @itemx __AVR_HAVE_16BIT_SP__
18576 The stack pointer (SP) register is treated as 8-bit respectively
18577 16-bit register by the compiler.
18578 The definition of these macros is affected by @option{-mtiny-stack}.
18580 @item __AVR_HAVE_SPH__
18582 The device has the SPH (high part of stack pointer) special function
18583 register or has an 8-bit stack pointer, respectively.
18584 The definition of these macros is affected by @option{-mmcu=} and
18585 in the cases of @option{-mmcu=avr2} and @option{-mmcu=avr25} also
18588 @item __AVR_HAVE_RAMPD__
18589 @itemx __AVR_HAVE_RAMPX__
18590 @itemx __AVR_HAVE_RAMPY__
18591 @itemx __AVR_HAVE_RAMPZ__
18592 The device has the @code{RAMPD}, @code{RAMPX}, @code{RAMPY},
18593 @code{RAMPZ} special function register, respectively.
18595 @item __NO_INTERRUPTS__
18596 This macro reflects the @option{-mno-interrupts} command-line option.
18598 @item __AVR_ERRATA_SKIP__
18599 @itemx __AVR_ERRATA_SKIP_JMP_CALL__
18600 Some AVR devices (AT90S8515, ATmega103) must not skip 32-bit
18601 instructions because of a hardware erratum. Skip instructions are
18602 @code{SBRS}, @code{SBRC}, @code{SBIS}, @code{SBIC} and @code{CPSE}.
18603 The second macro is only defined if @code{__AVR_HAVE_JMP_CALL__} is also
18606 @item __AVR_ISA_RMW__
18607 The device has Read-Modify-Write instructions (XCH, LAC, LAS and LAT).
18609 @item __AVR_SFR_OFFSET__=@var{offset}
18610 Instructions that can address I/O special function registers directly
18611 like @code{IN}, @code{OUT}, @code{SBI}, etc.@: may use a different
18612 address as if addressed by an instruction to access RAM like @code{LD}
18613 or @code{STS}. This offset depends on the device architecture and has
18614 to be subtracted from the RAM address in order to get the
18615 respective I/O@tie{}address.
18617 @item __AVR_SHORT_CALLS__
18618 The @option{-mshort-calls} command line option is set.
18620 @item __AVR_PM_BASE_ADDRESS__=@var{addr}
18621 Some devices support reading from flash memory by means of @code{LD*}
18622 instructions. The flash memory is seen in the data address space
18623 at an offset of @code{__AVR_PM_BASE_ADDRESS__}. If this macro
18624 is not defined, this feature is not available. If defined,
18625 the address space is linear and there is no need to put
18626 @code{.rodata} into RAM. This is handled by the default linker
18627 description file, and is currently available for
18628 @code{avrtiny} and @code{avrxmega3}. Even more convenient,
18629 there is no need to use address spaces like @code{__flash} or
18630 features like attribute @code{progmem} and @code{pgm_read_*}.
18632 @item __WITH_AVRLIBC__
18633 The compiler is configured to be used together with AVR-Libc.
18634 See the @option{--with-avrlibc} configure option.
18638 @node Blackfin Options
18639 @subsection Blackfin Options
18640 @cindex Blackfin Options
18643 @item -mcpu=@var{cpu}@r{[}-@var{sirevision}@r{]}
18645 Specifies the name of the target Blackfin processor. Currently, @var{cpu}
18646 can be one of @samp{bf512}, @samp{bf514}, @samp{bf516}, @samp{bf518},
18647 @samp{bf522}, @samp{bf523}, @samp{bf524}, @samp{bf525}, @samp{bf526},
18648 @samp{bf527}, @samp{bf531}, @samp{bf532}, @samp{bf533},
18649 @samp{bf534}, @samp{bf536}, @samp{bf537}, @samp{bf538}, @samp{bf539},
18650 @samp{bf542}, @samp{bf544}, @samp{bf547}, @samp{bf548}, @samp{bf549},
18651 @samp{bf542m}, @samp{bf544m}, @samp{bf547m}, @samp{bf548m}, @samp{bf549m},
18652 @samp{bf561}, @samp{bf592}.
18654 The optional @var{sirevision} specifies the silicon revision of the target
18655 Blackfin processor. Any workarounds available for the targeted silicon revision
18656 are enabled. If @var{sirevision} is @samp{none}, no workarounds are enabled.
18657 If @var{sirevision} is @samp{any}, all workarounds for the targeted processor
18658 are enabled. The @code{__SILICON_REVISION__} macro is defined to two
18659 hexadecimal digits representing the major and minor numbers in the silicon
18660 revision. If @var{sirevision} is @samp{none}, the @code{__SILICON_REVISION__}
18661 is not defined. If @var{sirevision} is @samp{any}, the
18662 @code{__SILICON_REVISION__} is defined to be @code{0xffff}.
18663 If this optional @var{sirevision} is not used, GCC assumes the latest known
18664 silicon revision of the targeted Blackfin processor.
18666 GCC defines a preprocessor macro for the specified @var{cpu}.
18667 For the @samp{bfin-elf} toolchain, this option causes the hardware BSP
18668 provided by libgloss to be linked in if @option{-msim} is not given.
18670 Without this option, @samp{bf532} is used as the processor by default.
18672 Note that support for @samp{bf561} is incomplete. For @samp{bf561},
18673 only the preprocessor macro is defined.
18677 Specifies that the program will be run on the simulator. This causes
18678 the simulator BSP provided by libgloss to be linked in. This option
18679 has effect only for @samp{bfin-elf} toolchain.
18680 Certain other options, such as @option{-mid-shared-library} and
18681 @option{-mfdpic}, imply @option{-msim}.
18683 @item -momit-leaf-frame-pointer
18684 @opindex momit-leaf-frame-pointer
18685 Don't keep the frame pointer in a register for leaf functions. This
18686 avoids the instructions to save, set up and restore frame pointers and
18687 makes an extra register available in leaf functions.
18689 @item -mspecld-anomaly
18690 @opindex mspecld-anomaly
18691 When enabled, the compiler ensures that the generated code does not
18692 contain speculative loads after jump instructions. If this option is used,
18693 @code{__WORKAROUND_SPECULATIVE_LOADS} is defined.
18695 @item -mno-specld-anomaly
18696 @opindex mno-specld-anomaly
18697 @opindex mspecld-anomaly
18698 Don't generate extra code to prevent speculative loads from occurring.
18700 @item -mcsync-anomaly
18701 @opindex mcsync-anomaly
18702 When enabled, the compiler ensures that the generated code does not
18703 contain CSYNC or SSYNC instructions too soon after conditional branches.
18704 If this option is used, @code{__WORKAROUND_SPECULATIVE_SYNCS} is defined.
18706 @item -mno-csync-anomaly
18707 @opindex mno-csync-anomaly
18708 @opindex mcsync-anomaly
18709 Don't generate extra code to prevent CSYNC or SSYNC instructions from
18710 occurring too soon after a conditional branch.
18714 When enabled, the compiler is free to take advantage of the knowledge that
18715 the entire program fits into the low 64k of memory.
18718 @opindex mno-low64k
18719 Assume that the program is arbitrarily large. This is the default.
18721 @item -mstack-check-l1
18722 @opindex mstack-check-l1
18723 Do stack checking using information placed into L1 scratchpad memory by the
18726 @item -mid-shared-library
18727 @opindex mid-shared-library
18728 Generate code that supports shared libraries via the library ID method.
18729 This allows for execute in place and shared libraries in an environment
18730 without virtual memory management. This option implies @option{-fPIC}.
18731 With a @samp{bfin-elf} target, this option implies @option{-msim}.
18733 @item -mno-id-shared-library
18734 @opindex mno-id-shared-library
18735 @opindex mid-shared-library
18736 Generate code that doesn't assume ID-based shared libraries are being used.
18737 This is the default.
18739 @item -mleaf-id-shared-library
18740 @opindex mleaf-id-shared-library
18741 Generate code that supports shared libraries via the library ID method,
18742 but assumes that this library or executable won't link against any other
18743 ID shared libraries. That allows the compiler to use faster code for jumps
18746 @item -mno-leaf-id-shared-library
18747 @opindex mno-leaf-id-shared-library
18748 @opindex mleaf-id-shared-library
18749 Do not assume that the code being compiled won't link against any ID shared
18750 libraries. Slower code is generated for jump and call insns.
18752 @item -mshared-library-id=n
18753 @opindex mshared-library-id
18754 Specifies the identification number of the ID-based shared library being
18755 compiled. Specifying a value of 0 generates more compact code; specifying
18756 other values forces the allocation of that number to the current
18757 library but is no more space- or time-efficient than omitting this option.
18761 Generate code that allows the data segment to be located in a different
18762 area of memory from the text segment. This allows for execute in place in
18763 an environment without virtual memory management by eliminating relocations
18764 against the text section.
18766 @item -mno-sep-data
18767 @opindex mno-sep-data
18769 Generate code that assumes that the data segment follows the text segment.
18770 This is the default.
18773 @itemx -mno-long-calls
18774 @opindex mlong-calls
18775 @opindex mno-long-calls
18776 Tells the compiler to perform function calls by first loading the
18777 address of the function into a register and then performing a subroutine
18778 call on this register. This switch is needed if the target function
18779 lies outside of the 24-bit addressing range of the offset-based
18780 version of subroutine call instruction.
18782 This feature is not enabled by default. Specifying
18783 @option{-mno-long-calls} restores the default behavior. Note these
18784 switches have no effect on how the compiler generates code to handle
18785 function calls via function pointers.
18789 Link with the fast floating-point library. This library relaxes some of
18790 the IEEE floating-point standard's rules for checking inputs against
18791 Not-a-Number (NAN), in the interest of performance.
18794 @opindex minline-plt
18795 Enable inlining of PLT entries in function calls to functions that are
18796 not known to bind locally. It has no effect without @option{-mfdpic}.
18799 @opindex mmulticore
18800 Build a standalone application for multicore Blackfin processors.
18801 This option causes proper start files and link scripts supporting
18802 multicore to be used, and defines the macro @code{__BFIN_MULTICORE}.
18803 It can only be used with @option{-mcpu=bf561@r{[}-@var{sirevision}@r{]}}.
18805 This option can be used with @option{-mcorea} or @option{-mcoreb}, which
18806 selects the one-application-per-core programming model. Without
18807 @option{-mcorea} or @option{-mcoreb}, the single-application/dual-core
18808 programming model is used. In this model, the main function of Core B
18809 should be named as @code{coreb_main}.
18811 If this option is not used, the single-core application programming
18816 Build a standalone application for Core A of BF561 when using
18817 the one-application-per-core programming model. Proper start files
18818 and link scripts are used to support Core A, and the macro
18819 @code{__BFIN_COREA} is defined.
18820 This option can only be used in conjunction with @option{-mmulticore}.
18824 Build a standalone application for Core B of BF561 when using
18825 the one-application-per-core programming model. Proper start files
18826 and link scripts are used to support Core B, and the macro
18827 @code{__BFIN_COREB} is defined. When this option is used, @code{coreb_main}
18828 should be used instead of @code{main}.
18829 This option can only be used in conjunction with @option{-mmulticore}.
18833 Build a standalone application for SDRAM. Proper start files and
18834 link scripts are used to put the application into SDRAM, and the macro
18835 @code{__BFIN_SDRAM} is defined.
18836 The loader should initialize SDRAM before loading the application.
18840 Assume that ICPLBs are enabled at run time. This has an effect on certain
18841 anomaly workarounds. For Linux targets, the default is to assume ICPLBs
18842 are enabled; for standalone applications the default is off.
18846 @subsection C6X Options
18847 @cindex C6X Options
18850 @item -march=@var{name}
18852 This specifies the name of the target architecture. GCC uses this
18853 name to determine what kind of instructions it can emit when generating
18854 assembly code. Permissible names are: @samp{c62x},
18855 @samp{c64x}, @samp{c64x+}, @samp{c67x}, @samp{c67x+}, @samp{c674x}.
18858 @opindex mbig-endian
18859 Generate code for a big-endian target.
18861 @item -mlittle-endian
18862 @opindex mlittle-endian
18863 Generate code for a little-endian target. This is the default.
18867 Choose startup files and linker script suitable for the simulator.
18869 @item -msdata=default
18870 @opindex msdata=default
18871 Put small global and static data in the @code{.neardata} section,
18872 which is pointed to by register @code{B14}. Put small uninitialized
18873 global and static data in the @code{.bss} section, which is adjacent
18874 to the @code{.neardata} section. Put small read-only data into the
18875 @code{.rodata} section. The corresponding sections used for large
18876 pieces of data are @code{.fardata}, @code{.far} and @code{.const}.
18879 @opindex msdata=all
18880 Put all data, not just small objects, into the sections reserved for
18881 small data, and use addressing relative to the @code{B14} register to
18885 @opindex msdata=none
18886 Make no use of the sections reserved for small data, and use absolute
18887 addresses to access all data. Put all initialized global and static
18888 data in the @code{.fardata} section, and all uninitialized data in the
18889 @code{.far} section. Put all constant data into the @code{.const}
18894 @subsection CRIS Options
18895 @cindex CRIS Options
18897 These options are defined specifically for the CRIS ports.
18900 @item -march=@var{architecture-type}
18901 @itemx -mcpu=@var{architecture-type}
18904 Generate code for the specified architecture. The choices for
18905 @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for
18906 respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX@.
18907 Default is @samp{v0} except for cris-axis-linux-gnu, where the default is
18910 @item -mtune=@var{architecture-type}
18912 Tune to @var{architecture-type} everything applicable about the generated
18913 code, except for the ABI and the set of available instructions. The
18914 choices for @var{architecture-type} are the same as for
18915 @option{-march=@var{architecture-type}}.
18917 @item -mmax-stack-frame=@var{n}
18918 @opindex mmax-stack-frame
18919 Warn when the stack frame of a function exceeds @var{n} bytes.
18925 The options @option{-metrax4} and @option{-metrax100} are synonyms for
18926 @option{-march=v3} and @option{-march=v8} respectively.
18928 @item -mmul-bug-workaround
18929 @itemx -mno-mul-bug-workaround
18930 @opindex mmul-bug-workaround
18931 @opindex mno-mul-bug-workaround
18932 Work around a bug in the @code{muls} and @code{mulu} instructions for CPU
18933 models where it applies. This option is active by default.
18937 Enable CRIS-specific verbose debug-related information in the assembly
18938 code. This option also has the effect of turning off the @samp{#NO_APP}
18939 formatted-code indicator to the assembler at the beginning of the
18944 Do not use condition-code results from previous instruction; always emit
18945 compare and test instructions before use of condition codes.
18947 @item -mno-side-effects
18948 @opindex mno-side-effects
18949 @opindex mside-effects
18950 Do not emit instructions with side effects in addressing modes other than
18953 @item -mstack-align
18954 @itemx -mno-stack-align
18955 @itemx -mdata-align
18956 @itemx -mno-data-align
18957 @itemx -mconst-align
18958 @itemx -mno-const-align
18959 @opindex mstack-align
18960 @opindex mno-stack-align
18961 @opindex mdata-align
18962 @opindex mno-data-align
18963 @opindex mconst-align
18964 @opindex mno-const-align
18965 These options (@samp{no-} options) arrange (eliminate arrangements) for the
18966 stack frame, individual data and constants to be aligned for the maximum
18967 single data access size for the chosen CPU model. The default is to
18968 arrange for 32-bit alignment. ABI details such as structure layout are
18969 not affected by these options.
18977 Similar to the stack- data- and const-align options above, these options
18978 arrange for stack frame, writable data and constants to all be 32-bit,
18979 16-bit or 8-bit aligned. The default is 32-bit alignment.
18981 @item -mno-prologue-epilogue
18982 @itemx -mprologue-epilogue
18983 @opindex mno-prologue-epilogue
18984 @opindex mprologue-epilogue
18985 With @option{-mno-prologue-epilogue}, the normal function prologue and
18986 epilogue which set up the stack frame are omitted and no return
18987 instructions or return sequences are generated in the code. Use this
18988 option only together with visual inspection of the compiled code: no
18989 warnings or errors are generated when call-saved registers must be saved,
18990 or storage for local variables needs to be allocated.
18994 @opindex mno-gotplt
18996 With @option{-fpic} and @option{-fPIC}, don't generate (do generate)
18997 instruction sequences that load addresses for functions from the PLT part
18998 of the GOT rather than (traditional on other architectures) calls to the
18999 PLT@. The default is @option{-mgotplt}.
19003 Legacy no-op option only recognized with the cris-axis-elf and
19004 cris-axis-linux-gnu targets.
19008 Legacy no-op option only recognized with the cris-axis-linux-gnu target.
19012 This option, recognized for the cris-axis-elf, arranges
19013 to link with input-output functions from a simulator library. Code,
19014 initialized data and zero-initialized data are allocated consecutively.
19018 Like @option{-sim}, but pass linker options to locate initialized data at
19019 0x40000000 and zero-initialized data at 0x80000000.
19023 @subsection CR16 Options
19024 @cindex CR16 Options
19026 These options are defined specifically for the CR16 ports.
19032 Enable the use of multiply-accumulate instructions. Disabled by default.
19036 @opindex mcr16cplus
19038 Generate code for CR16C or CR16C+ architecture. CR16C+ architecture
19043 Links the library libsim.a which is in compatible with simulator. Applicable
19044 to ELF compiler only.
19048 Choose integer type as 32-bit wide.
19052 Generates @code{sbit}/@code{cbit} instructions for bit manipulations.
19054 @item -mdata-model=@var{model}
19055 @opindex mdata-model
19056 Choose a data model. The choices for @var{model} are @samp{near},
19057 @samp{far} or @samp{medium}. @samp{medium} is default.
19058 However, @samp{far} is not valid with @option{-mcr16c}, as the
19059 CR16C architecture does not support the far data model.
19062 @node C-SKY Options
19063 @subsection C-SKY Options
19064 @cindex C-SKY Options
19066 GCC supports these options when compiling for C-SKY V2 processors.
19070 @item -march=@var{arch}
19072 Specify the C-SKY target architecture. Valid values for @var{arch} are:
19073 @samp{ck801}, @samp{ck802}, @samp{ck803}, @samp{ck807}, and @samp{ck810}.
19074 The default is @samp{ck810}.
19076 @item -mcpu=@var{cpu}
19078 Specify the C-SKY target processor. Valid values for @var{cpu} are:
19079 @samp{ck801}, @samp{ck801t},
19080 @samp{ck802}, @samp{ck802t}, @samp{ck802j},
19081 @samp{ck803}, @samp{ck803h}, @samp{ck803t}, @samp{ck803ht},
19082 @samp{ck803f}, @samp{ck803fh}, @samp{ck803e}, @samp{ck803eh},
19083 @samp{ck803et}, @samp{ck803eht}, @samp{ck803ef}, @samp{ck803efh},
19084 @samp{ck803ft}, @samp{ck803eft}, @samp{ck803efht}, @samp{ck803r1},
19085 @samp{ck803hr1}, @samp{ck803tr1}, @samp{ck803htr1}, @samp{ck803fr1},
19086 @samp{ck803fhr1}, @samp{ck803er1}, @samp{ck803ehr1}, @samp{ck803etr1},
19087 @samp{ck803ehtr1}, @samp{ck803efr1}, @samp{ck803efhr1}, @samp{ck803ftr1},
19088 @samp{ck803eftr1}, @samp{ck803efhtr1},
19089 @samp{ck803s}, @samp{ck803st}, @samp{ck803se}, @samp{ck803sf},
19090 @samp{ck803sef}, @samp{ck803seft},
19091 @samp{ck807e}, @samp{ck807ef}, @samp{ck807}, @samp{ck807f},
19092 @samp{ck810e}, @samp{ck810et}, @samp{ck810ef}, @samp{ck810eft},
19093 @samp{ck810}, @samp{ck810v}, @samp{ck810f}, @samp{ck810t}, @samp{ck810fv},
19094 @samp{ck810tv}, @samp{ck810ft}, and @samp{ck810ftv}.
19097 @opindex mbig-endian
19100 @itemx -mlittle-endian
19101 @opindex mlittle-endian
19105 Select big- or little-endian code. The default is little-endian.
19108 @opindex mhard-float
19109 @itemx -msoft-float
19110 @opindex msoft-float
19112 Select hardware or software floating-point implementations.
19113 The default is soft float.
19115 @item -mdouble-float
19116 @itemx -mno-double-float
19117 @opindex mdouble-float
19118 When @option{-mhard-float} is in effect, enable generation of
19119 double-precision float instructions. This is the default except
19120 when compiling for CK803.
19125 When @option{-mhard-float} is in effect, enable generation of
19126 @code{frecipd}, @code{fsqrtd}, and @code{fdivd} instructions.
19127 This is the default except when compiling for CK803.
19129 @item -mfpu=@var{fpu}
19131 Select the floating-point processor. This option can only be used with
19132 @option{-mhard-float}.
19133 Values for @var{fpu} are
19134 @samp{fpv2_sf} (equivalent to @samp{-mno-double-float -mno-fdivdu}),
19135 @samp{fpv2} (@samp{-mdouble-float -mno-divdu}), and
19136 @samp{fpv2_divd} (@samp{-mdouble-float -mdivdu}).
19141 Enable the extended @code{lrw} instruction. This option defaults to on
19142 for CK801 and off otherwise.
19147 Enable interrupt stack instructions; the default is off.
19149 The @option{-mistack} option is required to handle the
19150 @code{interrupt} and @code{isr} function attributes
19151 (@pxref{C-SKY Function Attributes}).
19155 Enable multiprocessor instructions; the default is off.
19159 Enable coprocessor instructions; the default is off.
19163 Enable coprocessor instructions; the default is off.
19167 Enable C-SKY security instructions; the default is off.
19171 Enable C-SKY trust instructions; the default is off.
19179 Enable C-SKY DSP, Enhanced DSP, or Vector DSP instructions, respectively.
19180 All of these options default to off.
19185 Generate divide instructions. Default is off.
19190 Generate code for Smart Mode, using only registers numbered 0-7 to allow
19191 use of 16-bit instructions. This option is ignored for CK801 where this
19192 is the required behavior, and it defaults to on for CK802.
19193 For other targets, the default is off.
19195 @item -mhigh-registers
19196 @itemx -mno-high-registers
19197 @opindex mhigh-registers
19198 Generate code using the high registers numbered 16-31. This option
19199 is not supported on CK801, CK802, or CK803, and is enabled by default
19200 for other processors.
19205 Generate code using global anchor symbol addresses.
19208 @itemx -mno-pushpop
19210 Generate code using @code{push} and @code{pop} instructions. This option
19213 @item -mmultiple-stld
19215 @itemx -mno-multiple-stld
19217 @opindex mmultiple-stld
19218 Generate code using @code{stm} and @code{ldm} instructions. This option
19219 isn't supported on CK801 but is enabled by default on other processors.
19222 @itemx -mno-constpool
19223 @opindex mconstpool
19224 Create constant pools in the compiler instead of deferring it to the
19225 assembler. This option is the default and required for correct code
19226 generation on CK801 and CK802, and is optional on other processors.
19229 @item -mno-stack-size
19230 @opindex mstack-size
19231 Emit @code{.stack_size} directives for each function in the assembly
19232 output. This option defaults to off.
19237 Generate code for the C-SKY compiler runtime instead of libgcc. This
19238 option defaults to off.
19240 @item -mbranch-cost=@var{n}
19241 @opindex mbranch-cost=
19242 Set the branch costs to roughly @code{n} instructions. The default is 1.
19244 @item -msched-prolog
19245 @itemx -mno-sched-prolog
19246 @opindex msched-prolog
19247 Permit scheduling of function prologue and epilogue sequences. Using
19248 this option can result in code that is not compliant with the C-SKY V2 ABI
19249 prologue requirements and that cannot be debugged or backtraced.
19250 It is disabled by default.
19254 @node Darwin Options
19255 @subsection Darwin Options
19256 @cindex Darwin options
19258 These options are defined for all architectures running the Darwin operating
19261 FSF GCC on Darwin does not create ``fat'' object files; it creates
19262 an object file for the single architecture that GCC was built to
19263 target. Apple's GCC on Darwin does create ``fat'' files if multiple
19264 @option{-arch} options are used; it does so by running the compiler or
19265 linker multiple times and joining the results together with
19268 The subtype of the file created (like @samp{ppc7400} or @samp{ppc970} or
19269 @samp{i686}) is determined by the flags that specify the ISA
19270 that GCC is targeting, like @option{-mcpu} or @option{-march}. The
19271 @option{-force_cpusubtype_ALL} option can be used to override this.
19273 The Darwin tools vary in their behavior when presented with an ISA
19274 mismatch. The assembler, @file{as}, only permits instructions to
19275 be used that are valid for the subtype of the file it is generating,
19276 so you cannot put 64-bit instructions in a @samp{ppc750} object file.
19277 The linker for shared libraries, @file{/usr/bin/libtool}, fails
19278 and prints an error if asked to create a shared library with a less
19279 restrictive subtype than its input files (for instance, trying to put
19280 a @samp{ppc970} object file in a @samp{ppc7400} library). The linker
19281 for executables, @command{ld}, quietly gives the executable the most
19282 restrictive subtype of any of its input files.
19287 Add the framework directory @var{dir} to the head of the list of
19288 directories to be searched for header files. These directories are
19289 interleaved with those specified by @option{-I} options and are
19290 scanned in a left-to-right order.
19292 A framework directory is a directory with frameworks in it. A
19293 framework is a directory with a @file{Headers} and/or
19294 @file{PrivateHeaders} directory contained directly in it that ends
19295 in @file{.framework}. The name of a framework is the name of this
19296 directory excluding the @file{.framework}. Headers associated with
19297 the framework are found in one of those two directories, with
19298 @file{Headers} being searched first. A subframework is a framework
19299 directory that is in a framework's @file{Frameworks} directory.
19300 Includes of subframework headers can only appear in a header of a
19301 framework that contains the subframework, or in a sibling subframework
19302 header. Two subframeworks are siblings if they occur in the same
19303 framework. A subframework should not have the same name as a
19304 framework; a warning is issued if this is violated. Currently a
19305 subframework cannot have subframeworks; in the future, the mechanism
19306 may be extended to support this. The standard frameworks can be found
19307 in @file{/System/Library/Frameworks} and
19308 @file{/Library/Frameworks}. An example include looks like
19309 @code{#include <Framework/header.h>}, where @file{Framework} denotes
19310 the name of the framework and @file{header.h} is found in the
19311 @file{PrivateHeaders} or @file{Headers} directory.
19313 @item -iframework@var{dir}
19314 @opindex iframework
19315 Like @option{-F} except the directory is a treated as a system
19316 directory. The main difference between this @option{-iframework} and
19317 @option{-F} is that with @option{-iframework} the compiler does not
19318 warn about constructs contained within header files found via
19319 @var{dir}. This option is valid only for the C family of languages.
19323 Emit debugging information for symbols that are used. For stabs
19324 debugging format, this enables @option{-feliminate-unused-debug-symbols}.
19325 This is by default ON@.
19329 Emit debugging information for all symbols and types.
19331 @item -mmacosx-version-min=@var{version}
19332 The earliest version of MacOS X that this executable will run on
19333 is @var{version}. Typical values of @var{version} include @code{10.1},
19334 @code{10.2}, and @code{10.3.9}.
19336 If the compiler was built to use the system's headers by default,
19337 then the default for this option is the system version on which the
19338 compiler is running, otherwise the default is to make choices that
19339 are compatible with as many systems and code bases as possible.
19343 Enable kernel development mode. The @option{-mkernel} option sets
19344 @option{-static}, @option{-fno-common}, @option{-fno-use-cxa-atexit},
19345 @option{-fno-exceptions}, @option{-fno-non-call-exceptions},
19346 @option{-fapple-kext}, @option{-fno-weak} and @option{-fno-rtti} where
19347 applicable. This mode also sets @option{-mno-altivec},
19348 @option{-msoft-float}, @option{-fno-builtin} and
19349 @option{-mlong-branch} for PowerPC targets.
19351 @item -mone-byte-bool
19352 @opindex mone-byte-bool
19353 Override the defaults for @code{bool} so that @code{sizeof(bool)==1}.
19354 By default @code{sizeof(bool)} is @code{4} when compiling for
19355 Darwin/PowerPC and @code{1} when compiling for Darwin/x86, so this
19356 option has no effect on x86.
19358 @strong{Warning:} The @option{-mone-byte-bool} switch causes GCC
19359 to generate code that is not binary compatible with code generated
19360 without that switch. Using this switch may require recompiling all
19361 other modules in a program, including system libraries. Use this
19362 switch to conform to a non-default data model.
19364 @item -mfix-and-continue
19365 @itemx -ffix-and-continue
19366 @itemx -findirect-data
19367 @opindex mfix-and-continue
19368 @opindex ffix-and-continue
19369 @opindex findirect-data
19370 Generate code suitable for fast turnaround development, such as to
19371 allow GDB to dynamically load @file{.o} files into already-running
19372 programs. @option{-findirect-data} and @option{-ffix-and-continue}
19373 are provided for backwards compatibility.
19377 Loads all members of static archive libraries.
19378 See man ld(1) for more information.
19380 @item -arch_errors_fatal
19381 @opindex arch_errors_fatal
19382 Cause the errors having to do with files that have the wrong architecture
19385 @item -bind_at_load
19386 @opindex bind_at_load
19387 Causes the output file to be marked such that the dynamic linker will
19388 bind all undefined references when the file is loaded or launched.
19392 Produce a Mach-o bundle format file.
19393 See man ld(1) for more information.
19395 @item -bundle_loader @var{executable}
19396 @opindex bundle_loader
19397 This option specifies the @var{executable} that will load the build
19398 output file being linked. See man ld(1) for more information.
19401 @opindex dynamiclib
19402 When passed this option, GCC produces a dynamic library instead of
19403 an executable when linking, using the Darwin @file{libtool} command.
19405 @item -force_cpusubtype_ALL
19406 @opindex force_cpusubtype_ALL
19407 This causes GCC's output file to have the @samp{ALL} subtype, instead of
19408 one controlled by the @option{-mcpu} or @option{-march} option.
19410 @item -allowable_client @var{client_name}
19411 @itemx -client_name
19412 @itemx -compatibility_version
19413 @itemx -current_version
19415 @itemx -dependency-file
19417 @itemx -dylinker_install_name
19419 @itemx -exported_symbols_list
19422 @itemx -flat_namespace
19423 @itemx -force_flat_namespace
19424 @itemx -headerpad_max_install_names
19427 @itemx -install_name
19428 @itemx -keep_private_externs
19429 @itemx -multi_module
19430 @itemx -multiply_defined
19431 @itemx -multiply_defined_unused
19434 @itemx -no_dead_strip_inits_and_terms
19435 @itemx -nofixprebinding
19436 @itemx -nomultidefs
19438 @itemx -noseglinkedit
19439 @itemx -pagezero_size
19441 @itemx -prebind_all_twolevel_modules
19442 @itemx -private_bundle
19444 @itemx -read_only_relocs
19446 @itemx -sectobjectsymbols
19450 @itemx -sectobjectsymbols
19453 @itemx -segs_read_only_addr
19455 @itemx -segs_read_write_addr
19456 @itemx -seg_addr_table
19457 @itemx -seg_addr_table_filename
19458 @itemx -seglinkedit
19460 @itemx -segs_read_only_addr
19461 @itemx -segs_read_write_addr
19462 @itemx -single_module
19464 @itemx -sub_library
19466 @itemx -sub_umbrella
19467 @itemx -twolevel_namespace
19470 @itemx -unexported_symbols_list
19471 @itemx -weak_reference_mismatches
19472 @itemx -whatsloaded
19473 @opindex allowable_client
19474 @opindex client_name
19475 @opindex compatibility_version
19476 @opindex current_version
19477 @opindex dead_strip
19478 @opindex dependency-file
19479 @opindex dylib_file
19480 @opindex dylinker_install_name
19482 @opindex exported_symbols_list
19484 @opindex flat_namespace
19485 @opindex force_flat_namespace
19486 @opindex headerpad_max_install_names
19487 @opindex image_base
19489 @opindex install_name
19490 @opindex keep_private_externs
19491 @opindex multi_module
19492 @opindex multiply_defined
19493 @opindex multiply_defined_unused
19494 @opindex noall_load
19495 @opindex no_dead_strip_inits_and_terms
19496 @opindex nofixprebinding
19497 @opindex nomultidefs
19499 @opindex noseglinkedit
19500 @opindex pagezero_size
19502 @opindex prebind_all_twolevel_modules
19503 @opindex private_bundle
19504 @opindex read_only_relocs
19506 @opindex sectobjectsymbols
19509 @opindex sectcreate
19510 @opindex sectobjectsymbols
19513 @opindex segs_read_only_addr
19514 @opindex segs_read_write_addr
19515 @opindex seg_addr_table
19516 @opindex seg_addr_table_filename
19517 @opindex seglinkedit
19519 @opindex segs_read_only_addr
19520 @opindex segs_read_write_addr
19521 @opindex single_module
19523 @opindex sub_library
19524 @opindex sub_umbrella
19525 @opindex twolevel_namespace
19528 @opindex unexported_symbols_list
19529 @opindex weak_reference_mismatches
19530 @opindex whatsloaded
19531 These options are passed to the Darwin linker. The Darwin linker man page
19532 describes them in detail.
19535 @node DEC Alpha Options
19536 @subsection DEC Alpha Options
19538 These @samp{-m} options are defined for the DEC Alpha implementations:
19541 @item -mno-soft-float
19542 @itemx -msoft-float
19543 @opindex mno-soft-float
19544 @opindex msoft-float
19545 Use (do not use) the hardware floating-point instructions for
19546 floating-point operations. When @option{-msoft-float} is specified,
19547 functions in @file{libgcc.a} are used to perform floating-point
19548 operations. Unless they are replaced by routines that emulate the
19549 floating-point operations, or compiled in such a way as to call such
19550 emulations routines, these routines issue floating-point
19551 operations. If you are compiling for an Alpha without floating-point
19552 operations, you must ensure that the library is built so as not to call
19555 Note that Alpha implementations without floating-point operations are
19556 required to have floating-point registers.
19559 @itemx -mno-fp-regs
19561 @opindex mno-fp-regs
19562 Generate code that uses (does not use) the floating-point register set.
19563 @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point
19564 register set is not used, floating-point operands are passed in integer
19565 registers as if they were integers and floating-point results are passed
19566 in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence,
19567 so any function with a floating-point argument or return value called by code
19568 compiled with @option{-mno-fp-regs} must also be compiled with that
19571 A typical use of this option is building a kernel that does not use,
19572 and hence need not save and restore, any floating-point registers.
19576 The Alpha architecture implements floating-point hardware optimized for
19577 maximum performance. It is mostly compliant with the IEEE floating-point
19578 standard. However, for full compliance, software assistance is
19579 required. This option generates code fully IEEE-compliant code
19580 @emph{except} that the @var{inexact-flag} is not maintained (see below).
19581 If this option is turned on, the preprocessor macro @code{_IEEE_FP} is
19582 defined during compilation. The resulting code is less efficient but is
19583 able to correctly support denormalized numbers and exceptional IEEE
19584 values such as not-a-number and plus/minus infinity. Other Alpha
19585 compilers call this option @option{-ieee_with_no_inexact}.
19587 @item -mieee-with-inexact
19588 @opindex mieee-with-inexact
19589 This is like @option{-mieee} except the generated code also maintains
19590 the IEEE @var{inexact-flag}. Turning on this option causes the
19591 generated code to implement fully-compliant IEEE math. In addition to
19592 @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor
19593 macro. On some Alpha implementations the resulting code may execute
19594 significantly slower than the code generated by default. Since there is
19595 very little code that depends on the @var{inexact-flag}, you should
19596 normally not specify this option. Other Alpha compilers call this
19597 option @option{-ieee_with_inexact}.
19599 @item -mfp-trap-mode=@var{trap-mode}
19600 @opindex mfp-trap-mode
19601 This option controls what floating-point related traps are enabled.
19602 Other Alpha compilers call this option @option{-fptm @var{trap-mode}}.
19603 The trap mode can be set to one of four values:
19607 This is the default (normal) setting. The only traps that are enabled
19608 are the ones that cannot be disabled in software (e.g., division by zero
19612 In addition to the traps enabled by @samp{n}, underflow traps are enabled
19616 Like @samp{u}, but the instructions are marked to be safe for software
19617 completion (see Alpha architecture manual for details).
19620 Like @samp{su}, but inexact traps are enabled as well.
19623 @item -mfp-rounding-mode=@var{rounding-mode}
19624 @opindex mfp-rounding-mode
19625 Selects the IEEE rounding mode. Other Alpha compilers call this option
19626 @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one
19631 Normal IEEE rounding mode. Floating-point numbers are rounded towards
19632 the nearest machine number or towards the even machine number in case
19636 Round towards minus infinity.
19639 Chopped rounding mode. Floating-point numbers are rounded towards zero.
19642 Dynamic rounding mode. A field in the floating-point control register
19643 (@var{fpcr}, see Alpha architecture reference manual) controls the
19644 rounding mode in effect. The C library initializes this register for
19645 rounding towards plus infinity. Thus, unless your program modifies the
19646 @var{fpcr}, @samp{d} corresponds to round towards plus infinity.
19649 @item -mtrap-precision=@var{trap-precision}
19650 @opindex mtrap-precision
19651 In the Alpha architecture, floating-point traps are imprecise. This
19652 means without software assistance it is impossible to recover from a
19653 floating trap and program execution normally needs to be terminated.
19654 GCC can generate code that can assist operating system trap handlers
19655 in determining the exact location that caused a floating-point trap.
19656 Depending on the requirements of an application, different levels of
19657 precisions can be selected:
19661 Program precision. This option is the default and means a trap handler
19662 can only identify which program caused a floating-point exception.
19665 Function precision. The trap handler can determine the function that
19666 caused a floating-point exception.
19669 Instruction precision. The trap handler can determine the exact
19670 instruction that caused a floating-point exception.
19673 Other Alpha compilers provide the equivalent options called
19674 @option{-scope_safe} and @option{-resumption_safe}.
19676 @item -mieee-conformant
19677 @opindex mieee-conformant
19678 This option marks the generated code as IEEE conformant. You must not
19679 use this option unless you also specify @option{-mtrap-precision=i} and either
19680 @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect
19681 is to emit the line @samp{.eflag 48} in the function prologue of the
19682 generated assembly file.
19684 @item -mbuild-constants
19685 @opindex mbuild-constants
19686 Normally GCC examines a 32- or 64-bit integer constant to
19687 see if it can construct it from smaller constants in two or three
19688 instructions. If it cannot, it outputs the constant as a literal and
19689 generates code to load it from the data segment at run time.
19691 Use this option to require GCC to construct @emph{all} integer constants
19692 using code, even if it takes more instructions (the maximum is six).
19694 You typically use this option to build a shared library dynamic
19695 loader. Itself a shared library, it must relocate itself in memory
19696 before it can find the variables and constants in its own data segment.
19714 Indicate whether GCC should generate code to use the optional BWX,
19715 CIX, FIX and MAX instruction sets. The default is to use the instruction
19716 sets supported by the CPU type specified via @option{-mcpu=} option or that
19717 of the CPU on which GCC was built if none is specified.
19720 @itemx -mfloat-ieee
19721 @opindex mfloat-vax
19722 @opindex mfloat-ieee
19723 Generate code that uses (does not use) VAX F and G floating-point
19724 arithmetic instead of IEEE single and double precision.
19726 @item -mexplicit-relocs
19727 @itemx -mno-explicit-relocs
19728 @opindex mexplicit-relocs
19729 @opindex mno-explicit-relocs
19730 Older Alpha assemblers provided no way to generate symbol relocations
19731 except via assembler macros. Use of these macros does not allow
19732 optimal instruction scheduling. GNU binutils as of version 2.12
19733 supports a new syntax that allows the compiler to explicitly mark
19734 which relocations should apply to which instructions. This option
19735 is mostly useful for debugging, as GCC detects the capabilities of
19736 the assembler when it is built and sets the default accordingly.
19739 @itemx -mlarge-data
19740 @opindex msmall-data
19741 @opindex mlarge-data
19742 When @option{-mexplicit-relocs} is in effect, static data is
19743 accessed via @dfn{gp-relative} relocations. When @option{-msmall-data}
19744 is used, objects 8 bytes long or smaller are placed in a @dfn{small data area}
19745 (the @code{.sdata} and @code{.sbss} sections) and are accessed via
19746 16-bit relocations off of the @code{$gp} register. This limits the
19747 size of the small data area to 64KB, but allows the variables to be
19748 directly accessed via a single instruction.
19750 The default is @option{-mlarge-data}. With this option the data area
19751 is limited to just below 2GB@. Programs that require more than 2GB of
19752 data must use @code{malloc} or @code{mmap} to allocate the data in the
19753 heap instead of in the program's data segment.
19755 When generating code for shared libraries, @option{-fpic} implies
19756 @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}.
19759 @itemx -mlarge-text
19760 @opindex msmall-text
19761 @opindex mlarge-text
19762 When @option{-msmall-text} is used, the compiler assumes that the
19763 code of the entire program (or shared library) fits in 4MB, and is
19764 thus reachable with a branch instruction. When @option{-msmall-data}
19765 is used, the compiler can assume that all local symbols share the
19766 same @code{$gp} value, and thus reduce the number of instructions
19767 required for a function call from 4 to 1.
19769 The default is @option{-mlarge-text}.
19771 @item -mcpu=@var{cpu_type}
19773 Set the instruction set and instruction scheduling parameters for
19774 machine type @var{cpu_type}. You can specify either the @samp{EV}
19775 style name or the corresponding chip number. GCC supports scheduling
19776 parameters for the EV4, EV5 and EV6 family of processors and
19777 chooses the default values for the instruction set from the processor
19778 you specify. If you do not specify a processor type, GCC defaults
19779 to the processor on which the compiler was built.
19781 Supported values for @var{cpu_type} are
19787 Schedules as an EV4 and has no instruction set extensions.
19791 Schedules as an EV5 and has no instruction set extensions.
19795 Schedules as an EV5 and supports the BWX extension.
19800 Schedules as an EV5 and supports the BWX and MAX extensions.
19804 Schedules as an EV6 and supports the BWX, FIX, and MAX extensions.
19808 Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions.
19811 Native toolchains also support the value @samp{native},
19812 which selects the best architecture option for the host processor.
19813 @option{-mcpu=native} has no effect if GCC does not recognize
19816 @item -mtune=@var{cpu_type}
19818 Set only the instruction scheduling parameters for machine type
19819 @var{cpu_type}. The instruction set is not changed.
19821 Native toolchains also support the value @samp{native},
19822 which selects the best architecture option for the host processor.
19823 @option{-mtune=native} has no effect if GCC does not recognize
19826 @item -mmemory-latency=@var{time}
19827 @opindex mmemory-latency
19828 Sets the latency the scheduler should assume for typical memory
19829 references as seen by the application. This number is highly
19830 dependent on the memory access patterns used by the application
19831 and the size of the external cache on the machine.
19833 Valid options for @var{time} are
19837 A decimal number representing clock cycles.
19843 The compiler contains estimates of the number of clock cycles for
19844 ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches
19845 (also called Dcache, Scache, and Bcache), as well as to main memory.
19846 Note that L3 is only valid for EV5.
19852 @subsection eBPF Options
19853 @cindex eBPF Options
19856 @item -mframe-limit=@var{bytes}
19857 This specifies the hard limit for frame sizes, in bytes. Currently,
19858 the value that can be specified should be less than or equal to
19859 @samp{32767}. Defaults to whatever limit is imposed by the version of
19860 the Linux kernel targeted.
19862 @item -mkernel=@var{version}
19864 This specifies the minimum version of the kernel that will run the
19865 compiled program. GCC uses this version to determine which
19866 instructions to use, what kernel helpers to allow, etc. Currently,
19867 @var{version} can be one of @samp{4.0}, @samp{4.1}, @samp{4.2},
19868 @samp{4.3}, @samp{4.4}, @samp{4.5}, @samp{4.6}, @samp{4.7},
19869 @samp{4.8}, @samp{4.9}, @samp{4.10}, @samp{4.11}, @samp{4.12},
19870 @samp{4.13}, @samp{4.14}, @samp{4.15}, @samp{4.16}, @samp{4.17},
19871 @samp{4.18}, @samp{4.19}, @samp{4.20}, @samp{5.0}, @samp{5.1},
19872 @samp{5.2}, @samp{latest} and @samp{native}.
19875 @opindex mbig-endian
19876 Generate code for a big-endian target.
19878 @item -mlittle-endian
19879 @opindex mlittle-endian
19880 Generate code for a little-endian target. This is the default.
19884 @subsection FR30 Options
19885 @cindex FR30 Options
19887 These options are defined specifically for the FR30 port.
19891 @item -msmall-model
19892 @opindex msmall-model
19893 Use the small address space model. This can produce smaller code, but
19894 it does assume that all symbolic values and addresses fit into a
19899 Assume that runtime support has been provided and so there is no need
19900 to include the simulator library (@file{libsim.a}) on the linker
19906 @subsection FT32 Options
19907 @cindex FT32 Options
19909 These options are defined specifically for the FT32 port.
19915 Specifies that the program will be run on the simulator. This causes
19916 an alternate runtime startup and library to be linked.
19917 You must not use this option when generating programs that will run on
19918 real hardware; you must provide your own runtime library for whatever
19919 I/O functions are needed.
19923 Enable Local Register Allocation. This is still experimental for FT32,
19924 so by default the compiler uses standard reload.
19928 Do not use div and mod instructions.
19932 Enable use of the extended instructions of the FT32B processor.
19936 Compress all code using the Ft32B code compression scheme.
19940 Do not generate code that reads program memory.
19945 @subsection FRV Options
19946 @cindex FRV Options
19952 Only use the first 32 general-purpose registers.
19957 Use all 64 general-purpose registers.
19962 Use only the first 32 floating-point registers.
19967 Use all 64 floating-point registers.
19970 @opindex mhard-float
19972 Use hardware instructions for floating-point operations.
19975 @opindex msoft-float
19977 Use library routines for floating-point operations.
19982 Dynamically allocate condition code registers.
19987 Do not try to dynamically allocate condition code registers, only
19988 use @code{icc0} and @code{fcc0}.
19993 Change ABI to use double word insns.
19999 Do not use double word instructions.
20004 Use floating-point double instructions.
20007 @opindex mno-double
20009 Do not use floating-point double instructions.
20014 Use media instructions.
20019 Do not use media instructions.
20024 Use multiply and add/subtract instructions.
20027 @opindex mno-muladd
20029 Do not use multiply and add/subtract instructions.
20034 Select the FDPIC ABI, which uses function descriptors to represent
20035 pointers to functions. Without any PIC/PIE-related options, it
20036 implies @option{-fPIE}. With @option{-fpic} or @option{-fpie}, it
20037 assumes GOT entries and small data are within a 12-bit range from the
20038 GOT base address; with @option{-fPIC} or @option{-fPIE}, GOT offsets
20039 are computed with 32 bits.
20040 With a @samp{bfin-elf} target, this option implies @option{-msim}.
20043 @opindex minline-plt
20045 Enable inlining of PLT entries in function calls to functions that are
20046 not known to bind locally. It has no effect without @option{-mfdpic}.
20047 It's enabled by default if optimizing for speed and compiling for
20048 shared libraries (i.e., @option{-fPIC} or @option{-fpic}), or when an
20049 optimization option such as @option{-O3} or above is present in the
20055 Assume a large TLS segment when generating thread-local code.
20060 Do not assume a large TLS segment when generating thread-local code.
20065 Enable the use of @code{GPREL} relocations in the FDPIC ABI for data
20066 that is known to be in read-only sections. It's enabled by default,
20067 except for @option{-fpic} or @option{-fpie}: even though it may help
20068 make the global offset table smaller, it trades 1 instruction for 4.
20069 With @option{-fPIC} or @option{-fPIE}, it trades 3 instructions for 4,
20070 one of which may be shared by multiple symbols, and it avoids the need
20071 for a GOT entry for the referenced symbol, so it's more likely to be a
20072 win. If it is not, @option{-mno-gprel-ro} can be used to disable it.
20074 @item -multilib-library-pic
20075 @opindex multilib-library-pic
20077 Link with the (library, not FD) pic libraries. It's implied by
20078 @option{-mlibrary-pic}, as well as by @option{-fPIC} and
20079 @option{-fpic} without @option{-mfdpic}. You should never have to use
20083 @opindex mlinked-fp
20085 Follow the EABI requirement of always creating a frame pointer whenever
20086 a stack frame is allocated. This option is enabled by default and can
20087 be disabled with @option{-mno-linked-fp}.
20090 @opindex mlong-calls
20092 Use indirect addressing to call functions outside the current
20093 compilation unit. This allows the functions to be placed anywhere
20094 within the 32-bit address space.
20096 @item -malign-labels
20097 @opindex malign-labels
20099 Try to align labels to an 8-byte boundary by inserting NOPs into the
20100 previous packet. This option only has an effect when VLIW packing
20101 is enabled. It doesn't create new packets; it merely adds NOPs to
20104 @item -mlibrary-pic
20105 @opindex mlibrary-pic
20107 Generate position-independent EABI code.
20112 Use only the first four media accumulator registers.
20117 Use all eight media accumulator registers.
20122 Pack VLIW instructions.
20127 Do not pack VLIW instructions.
20130 @opindex mno-eflags
20132 Do not mark ABI switches in e_flags.
20135 @opindex mcond-move
20137 Enable the use of conditional-move instructions (default).
20139 This switch is mainly for debugging the compiler and will likely be removed
20140 in a future version.
20142 @item -mno-cond-move
20143 @opindex mno-cond-move
20145 Disable the use of conditional-move instructions.
20147 This switch is mainly for debugging the compiler and will likely be removed
20148 in a future version.
20153 Enable the use of conditional set instructions (default).
20155 This switch is mainly for debugging the compiler and will likely be removed
20156 in a future version.
20161 Disable the use of conditional set instructions.
20163 This switch is mainly for debugging the compiler and will likely be removed
20164 in a future version.
20167 @opindex mcond-exec
20169 Enable the use of conditional execution (default).
20171 This switch is mainly for debugging the compiler and will likely be removed
20172 in a future version.
20174 @item -mno-cond-exec
20175 @opindex mno-cond-exec
20177 Disable the use of conditional execution.
20179 This switch is mainly for debugging the compiler and will likely be removed
20180 in a future version.
20182 @item -mvliw-branch
20183 @opindex mvliw-branch
20185 Run a pass to pack branches into VLIW instructions (default).
20187 This switch is mainly for debugging the compiler and will likely be removed
20188 in a future version.
20190 @item -mno-vliw-branch
20191 @opindex mno-vliw-branch
20193 Do not run a pass to pack branches into VLIW instructions.
20195 This switch is mainly for debugging the compiler and will likely be removed
20196 in a future version.
20198 @item -mmulti-cond-exec
20199 @opindex mmulti-cond-exec
20201 Enable optimization of @code{&&} and @code{||} in conditional execution
20204 This switch is mainly for debugging the compiler and will likely be removed
20205 in a future version.
20207 @item -mno-multi-cond-exec
20208 @opindex mno-multi-cond-exec
20210 Disable optimization of @code{&&} and @code{||} in conditional execution.
20212 This switch is mainly for debugging the compiler and will likely be removed
20213 in a future version.
20215 @item -mnested-cond-exec
20216 @opindex mnested-cond-exec
20218 Enable nested conditional execution optimizations (default).
20220 This switch is mainly for debugging the compiler and will likely be removed
20221 in a future version.
20223 @item -mno-nested-cond-exec
20224 @opindex mno-nested-cond-exec
20226 Disable nested conditional execution optimizations.
20228 This switch is mainly for debugging the compiler and will likely be removed
20229 in a future version.
20231 @item -moptimize-membar
20232 @opindex moptimize-membar
20234 This switch removes redundant @code{membar} instructions from the
20235 compiler-generated code. It is enabled by default.
20237 @item -mno-optimize-membar
20238 @opindex mno-optimize-membar
20239 @opindex moptimize-membar
20241 This switch disables the automatic removal of redundant @code{membar}
20242 instructions from the generated code.
20244 @item -mtomcat-stats
20245 @opindex mtomcat-stats
20247 Cause gas to print out tomcat statistics.
20249 @item -mcpu=@var{cpu}
20252 Select the processor type for which to generate code. Possible values are
20253 @samp{frv}, @samp{fr550}, @samp{tomcat}, @samp{fr500}, @samp{fr450},
20254 @samp{fr405}, @samp{fr400}, @samp{fr300} and @samp{simple}.
20258 @node GNU/Linux Options
20259 @subsection GNU/Linux Options
20261 These @samp{-m} options are defined for GNU/Linux targets:
20266 Use the GNU C library. This is the default except
20267 on @samp{*-*-linux-*uclibc*}, @samp{*-*-linux-*musl*} and
20268 @samp{*-*-linux-*android*} targets.
20272 Use uClibc C library. This is the default on
20273 @samp{*-*-linux-*uclibc*} targets.
20277 Use the musl C library. This is the default on
20278 @samp{*-*-linux-*musl*} targets.
20282 Use Bionic C library. This is the default on
20283 @samp{*-*-linux-*android*} targets.
20287 Compile code compatible with Android platform. This is the default on
20288 @samp{*-*-linux-*android*} targets.
20290 When compiling, this option enables @option{-mbionic}, @option{-fPIC},
20291 @option{-fno-exceptions} and @option{-fno-rtti} by default. When linking,
20292 this option makes the GCC driver pass Android-specific options to the linker.
20293 Finally, this option causes the preprocessor macro @code{__ANDROID__}
20296 @item -tno-android-cc
20297 @opindex tno-android-cc
20298 Disable compilation effects of @option{-mandroid}, i.e., do not enable
20299 @option{-mbionic}, @option{-fPIC}, @option{-fno-exceptions} and
20300 @option{-fno-rtti} by default.
20302 @item -tno-android-ld
20303 @opindex tno-android-ld
20304 Disable linking effects of @option{-mandroid}, i.e., pass standard Linux
20305 linking options to the linker.
20309 @node H8/300 Options
20310 @subsection H8/300 Options
20312 These @samp{-m} options are defined for the H8/300 implementations:
20317 Shorten some address references at link time, when possible; uses the
20318 linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300,
20319 ld, Using ld}, for a fuller description.
20323 Generate code for the H8/300H@.
20327 Generate code for the H8S@.
20331 Generate code for the H8S and H8/300H in the normal mode. This switch
20332 must be used either with @option{-mh} or @option{-ms}.
20336 Generate code for the H8S/2600. This switch must be used with @option{-ms}.
20340 Extended registers are stored on stack before execution of function
20341 with monitor attribute. Default option is @option{-mexr}.
20342 This option is valid only for H8S targets.
20347 Extended registers are not stored on stack before execution of function
20348 with monitor attribute. Default option is @option{-mno-exr}.
20349 This option is valid only for H8S targets.
20353 Make @code{int} data 32 bits by default.
20356 @opindex malign-300
20357 On the H8/300H and H8S, use the same alignment rules as for the H8/300.
20358 The default for the H8/300H and H8S is to align longs and floats on
20360 @option{-malign-300} causes them to be aligned on 2-byte boundaries.
20361 This option has no effect on the H8/300.
20365 @subsection HPPA Options
20366 @cindex HPPA Options
20368 These @samp{-m} options are defined for the HPPA family of computers:
20371 @item -march=@var{architecture-type}
20373 Generate code for the specified architecture. The choices for
20374 @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA
20375 1.1, and @samp{2.0} for PA 2.0 processors. Refer to
20376 @file{/usr/lib/sched.models} on an HP-UX system to determine the proper
20377 architecture option for your machine. Code compiled for lower numbered
20378 architectures runs on higher numbered architectures, but not the
20381 @item -mpa-risc-1-0
20382 @itemx -mpa-risc-1-1
20383 @itemx -mpa-risc-2-0
20384 @opindex mpa-risc-1-0
20385 @opindex mpa-risc-1-1
20386 @opindex mpa-risc-2-0
20387 Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively.
20389 @item -mcaller-copies
20390 @opindex mcaller-copies
20391 The caller copies function arguments passed by hidden reference. This
20392 option should be used with care as it is not compatible with the default
20393 32-bit runtime. However, only aggregates larger than eight bytes are
20394 passed by hidden reference and the option provides better compatibility
20397 @item -mjump-in-delay
20398 @opindex mjump-in-delay
20399 This option is ignored and provided for compatibility purposes only.
20401 @item -mdisable-fpregs
20402 @opindex mdisable-fpregs
20403 Prevent floating-point registers from being used in any manner. This is
20404 necessary for compiling kernels that perform lazy context switching of
20405 floating-point registers. If you use this option and attempt to perform
20406 floating-point operations, the compiler aborts.
20408 @item -mdisable-indexing
20409 @opindex mdisable-indexing
20410 Prevent the compiler from using indexing address modes. This avoids some
20411 rather obscure problems when compiling MIG generated code under MACH@.
20413 @item -mno-space-regs
20414 @opindex mno-space-regs
20415 @opindex mspace-regs
20416 Generate code that assumes the target has no space registers. This allows
20417 GCC to generate faster indirect calls and use unscaled index address modes.
20419 Such code is suitable for level 0 PA systems and kernels.
20421 @item -mfast-indirect-calls
20422 @opindex mfast-indirect-calls
20423 Generate code that assumes calls never cross space boundaries. This
20424 allows GCC to emit code that performs faster indirect calls.
20426 This option does not work in the presence of shared libraries or nested
20429 @item -mfixed-range=@var{register-range}
20430 @opindex mfixed-range
20431 Generate code treating the given register range as fixed registers.
20432 A fixed register is one that the register allocator cannot use. This is
20433 useful when compiling kernel code. A register range is specified as
20434 two registers separated by a dash. Multiple register ranges can be
20435 specified separated by a comma.
20437 @item -mlong-load-store
20438 @opindex mlong-load-store
20439 Generate 3-instruction load and store sequences as sometimes required by
20440 the HP-UX 10 linker. This is equivalent to the @samp{+k} option to
20443 @item -mportable-runtime
20444 @opindex mportable-runtime
20445 Use the portable calling conventions proposed by HP for ELF systems.
20449 Enable the use of assembler directives only GAS understands.
20451 @item -mschedule=@var{cpu-type}
20453 Schedule code according to the constraints for the machine type
20454 @var{cpu-type}. The choices for @var{cpu-type} are @samp{700}
20455 @samp{7100}, @samp{7100LC}, @samp{7200}, @samp{7300} and @samp{8000}. Refer
20456 to @file{/usr/lib/sched.models} on an HP-UX system to determine the
20457 proper scheduling option for your machine. The default scheduling is
20461 @opindex mlinker-opt
20462 Enable the optimization pass in the HP-UX linker. Note this makes symbolic
20463 debugging impossible. It also triggers a bug in the HP-UX 8 and HP-UX 9
20464 linkers in which they give bogus error messages when linking some programs.
20467 @opindex msoft-float
20468 Generate output containing library calls for floating point.
20469 @strong{Warning:} the requisite libraries are not available for all HPPA
20470 targets. Normally the facilities of the machine's usual C compiler are
20471 used, but this cannot be done directly in cross-compilation. You must make
20472 your own arrangements to provide suitable library functions for
20475 @option{-msoft-float} changes the calling convention in the output file;
20476 therefore, it is only useful if you compile @emph{all} of a program with
20477 this option. In particular, you need to compile @file{libgcc.a}, the
20478 library that comes with GCC, with @option{-msoft-float} in order for
20483 Generate the predefine, @code{_SIO}, for server IO@. The default is
20484 @option{-mwsio}. This generates the predefines, @code{__hp9000s700},
20485 @code{__hp9000s700__} and @code{_WSIO}, for workstation IO@. These
20486 options are available under HP-UX and HI-UX@.
20490 Use options specific to GNU @command{ld}.
20491 This passes @option{-shared} to @command{ld} when
20492 building a shared library. It is the default when GCC is configured,
20493 explicitly or implicitly, with the GNU linker. This option does not
20494 affect which @command{ld} is called; it only changes what parameters
20495 are passed to that @command{ld}.
20496 The @command{ld} that is called is determined by the
20497 @option{--with-ld} configure option, GCC's program search path, and
20498 finally by the user's @env{PATH}. The linker used by GCC can be printed
20499 using @samp{which `gcc -print-prog-name=ld`}. This option is only available
20500 on the 64-bit HP-UX GCC, i.e.@: configured with @samp{hppa*64*-*-hpux*}.
20504 Use options specific to HP @command{ld}.
20505 This passes @option{-b} to @command{ld} when building
20506 a shared library and passes @option{+Accept TypeMismatch} to @command{ld} on all
20507 links. It is the default when GCC is configured, explicitly or
20508 implicitly, with the HP linker. This option does not affect
20509 which @command{ld} is called; it only changes what parameters are passed to that
20511 The @command{ld} that is called is determined by the @option{--with-ld}
20512 configure option, GCC's program search path, and finally by the user's
20513 @env{PATH}. The linker used by GCC can be printed using @samp{which
20514 `gcc -print-prog-name=ld`}. This option is only available on the 64-bit
20515 HP-UX GCC, i.e.@: configured with @samp{hppa*64*-*-hpux*}.
20518 @opindex mno-long-calls
20519 @opindex mlong-calls
20520 Generate code that uses long call sequences. This ensures that a call
20521 is always able to reach linker generated stubs. The default is to generate
20522 long calls only when the distance from the call site to the beginning
20523 of the function or translation unit, as the case may be, exceeds a
20524 predefined limit set by the branch type being used. The limits for
20525 normal calls are 7,600,000 and 240,000 bytes, respectively for the
20526 PA 2.0 and PA 1.X architectures. Sibcalls are always limited at
20529 Distances are measured from the beginning of functions when using the
20530 @option{-ffunction-sections} option, or when using the @option{-mgas}
20531 and @option{-mno-portable-runtime} options together under HP-UX with
20534 It is normally not desirable to use this option as it degrades
20535 performance. However, it may be useful in large applications,
20536 particularly when partial linking is used to build the application.
20538 The types of long calls used depends on the capabilities of the
20539 assembler and linker, and the type of code being generated. The
20540 impact on systems that support long absolute calls, and long pic
20541 symbol-difference or pc-relative calls should be relatively small.
20542 However, an indirect call is used on 32-bit ELF systems in pic code
20543 and it is quite long.
20545 @item -munix=@var{unix-std}
20547 Generate compiler predefines and select a startfile for the specified
20548 UNIX standard. The choices for @var{unix-std} are @samp{93}, @samp{95}
20549 and @samp{98}. @samp{93} is supported on all HP-UX versions. @samp{95}
20550 is available on HP-UX 10.10 and later. @samp{98} is available on HP-UX
20551 11.11 and later. The default values are @samp{93} for HP-UX 10.00,
20552 @samp{95} for HP-UX 10.10 though to 11.00, and @samp{98} for HP-UX 11.11
20555 @option{-munix=93} provides the same predefines as GCC 3.3 and 3.4.
20556 @option{-munix=95} provides additional predefines for @code{XOPEN_UNIX}
20557 and @code{_XOPEN_SOURCE_EXTENDED}, and the startfile @file{unix95.o}.
20558 @option{-munix=98} provides additional predefines for @code{_XOPEN_UNIX},
20559 @code{_XOPEN_SOURCE_EXTENDED}, @code{_INCLUDE__STDC_A1_SOURCE} and
20560 @code{_INCLUDE_XOPEN_SOURCE_500}, and the startfile @file{unix98.o}.
20562 It is @emph{important} to note that this option changes the interfaces
20563 for various library routines. It also affects the operational behavior
20564 of the C library. Thus, @emph{extreme} care is needed in using this
20567 Library code that is intended to operate with more than one UNIX
20568 standard must test, set and restore the variable @code{__xpg4_extended_mask}
20569 as appropriate. Most GNU software doesn't provide this capability.
20573 Suppress the generation of link options to search libdld.sl when the
20574 @option{-static} option is specified on HP-UX 10 and later.
20578 The HP-UX implementation of setlocale in libc has a dependency on
20579 libdld.sl. There isn't an archive version of libdld.sl. Thus,
20580 when the @option{-static} option is specified, special link options
20581 are needed to resolve this dependency.
20583 On HP-UX 10 and later, the GCC driver adds the necessary options to
20584 link with libdld.sl when the @option{-static} option is specified.
20585 This causes the resulting binary to be dynamic. On the 64-bit port,
20586 the linkers generate dynamic binaries by default in any case. The
20587 @option{-nolibdld} option can be used to prevent the GCC driver from
20588 adding these link options.
20592 Add support for multithreading with the @dfn{dce thread} library
20593 under HP-UX@. This option sets flags for both the preprocessor and
20597 @node IA-64 Options
20598 @subsection IA-64 Options
20599 @cindex IA-64 Options
20601 These are the @samp{-m} options defined for the Intel IA-64 architecture.
20605 @opindex mbig-endian
20606 Generate code for a big-endian target. This is the default for HP-UX@.
20608 @item -mlittle-endian
20609 @opindex mlittle-endian
20610 Generate code for a little-endian target. This is the default for AIX5
20616 @opindex mno-gnu-as
20617 Generate (or don't) code for the GNU assembler. This is the default.
20618 @c Also, this is the default if the configure option @option{--with-gnu-as}
20624 @opindex mno-gnu-ld
20625 Generate (or don't) code for the GNU linker. This is the default.
20626 @c Also, this is the default if the configure option @option{--with-gnu-ld}
20631 Generate code that does not use a global pointer register. The result
20632 is not position independent code, and violates the IA-64 ABI@.
20634 @item -mvolatile-asm-stop
20635 @itemx -mno-volatile-asm-stop
20636 @opindex mvolatile-asm-stop
20637 @opindex mno-volatile-asm-stop
20638 Generate (or don't) a stop bit immediately before and after volatile asm
20641 @item -mregister-names
20642 @itemx -mno-register-names
20643 @opindex mregister-names
20644 @opindex mno-register-names
20645 Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for
20646 the stacked registers. This may make assembler output more readable.
20652 Disable (or enable) optimizations that use the small data section. This may
20653 be useful for working around optimizer bugs.
20655 @item -mconstant-gp
20656 @opindex mconstant-gp
20657 Generate code that uses a single constant global pointer value. This is
20658 useful when compiling kernel code.
20662 Generate code that is self-relocatable. This implies @option{-mconstant-gp}.
20663 This is useful when compiling firmware code.
20665 @item -minline-float-divide-min-latency
20666 @opindex minline-float-divide-min-latency
20667 Generate code for inline divides of floating-point values
20668 using the minimum latency algorithm.
20670 @item -minline-float-divide-max-throughput
20671 @opindex minline-float-divide-max-throughput
20672 Generate code for inline divides of floating-point values
20673 using the maximum throughput algorithm.
20675 @item -mno-inline-float-divide
20676 @opindex mno-inline-float-divide
20677 Do not generate inline code for divides of floating-point values.
20679 @item -minline-int-divide-min-latency
20680 @opindex minline-int-divide-min-latency
20681 Generate code for inline divides of integer values
20682 using the minimum latency algorithm.
20684 @item -minline-int-divide-max-throughput
20685 @opindex minline-int-divide-max-throughput
20686 Generate code for inline divides of integer values
20687 using the maximum throughput algorithm.
20689 @item -mno-inline-int-divide
20690 @opindex mno-inline-int-divide
20691 @opindex minline-int-divide
20692 Do not generate inline code for divides of integer values.
20694 @item -minline-sqrt-min-latency
20695 @opindex minline-sqrt-min-latency
20696 Generate code for inline square roots
20697 using the minimum latency algorithm.
20699 @item -minline-sqrt-max-throughput
20700 @opindex minline-sqrt-max-throughput
20701 Generate code for inline square roots
20702 using the maximum throughput algorithm.
20704 @item -mno-inline-sqrt
20705 @opindex mno-inline-sqrt
20706 Do not generate inline code for @code{sqrt}.
20709 @itemx -mno-fused-madd
20710 @opindex mfused-madd
20711 @opindex mno-fused-madd
20712 Do (don't) generate code that uses the fused multiply/add or multiply/subtract
20713 instructions. The default is to use these instructions.
20715 @item -mno-dwarf2-asm
20716 @itemx -mdwarf2-asm
20717 @opindex mno-dwarf2-asm
20718 @opindex mdwarf2-asm
20719 Don't (or do) generate assembler code for the DWARF line number debugging
20720 info. This may be useful when not using the GNU assembler.
20722 @item -mearly-stop-bits
20723 @itemx -mno-early-stop-bits
20724 @opindex mearly-stop-bits
20725 @opindex mno-early-stop-bits
20726 Allow stop bits to be placed earlier than immediately preceding the
20727 instruction that triggered the stop bit. This can improve instruction
20728 scheduling, but does not always do so.
20730 @item -mfixed-range=@var{register-range}
20731 @opindex mfixed-range
20732 Generate code treating the given register range as fixed registers.
20733 A fixed register is one that the register allocator cannot use. This is
20734 useful when compiling kernel code. A register range is specified as
20735 two registers separated by a dash. Multiple register ranges can be
20736 specified separated by a comma.
20738 @item -mtls-size=@var{tls-size}
20740 Specify bit size of immediate TLS offsets. Valid values are 14, 22, and
20743 @item -mtune=@var{cpu-type}
20745 Tune the instruction scheduling for a particular CPU, Valid values are
20746 @samp{itanium}, @samp{itanium1}, @samp{merced}, @samp{itanium2},
20747 and @samp{mckinley}.
20753 Generate code for a 32-bit or 64-bit environment.
20754 The 32-bit environment sets int, long and pointer to 32 bits.
20755 The 64-bit environment sets int to 32 bits and long and pointer
20756 to 64 bits. These are HP-UX specific flags.
20758 @item -mno-sched-br-data-spec
20759 @itemx -msched-br-data-spec
20760 @opindex mno-sched-br-data-spec
20761 @opindex msched-br-data-spec
20762 (Dis/En)able data speculative scheduling before reload.
20763 This results in generation of @code{ld.a} instructions and
20764 the corresponding check instructions (@code{ld.c} / @code{chk.a}).
20765 The default setting is disabled.
20767 @item -msched-ar-data-spec
20768 @itemx -mno-sched-ar-data-spec
20769 @opindex msched-ar-data-spec
20770 @opindex mno-sched-ar-data-spec
20771 (En/Dis)able data speculative scheduling after reload.
20772 This results in generation of @code{ld.a} instructions and
20773 the corresponding check instructions (@code{ld.c} / @code{chk.a}).
20774 The default setting is enabled.
20776 @item -mno-sched-control-spec
20777 @itemx -msched-control-spec
20778 @opindex mno-sched-control-spec
20779 @opindex msched-control-spec
20780 (Dis/En)able control speculative scheduling. This feature is
20781 available only during region scheduling (i.e.@: before reload).
20782 This results in generation of the @code{ld.s} instructions and
20783 the corresponding check instructions @code{chk.s}.
20784 The default setting is disabled.
20786 @item -msched-br-in-data-spec
20787 @itemx -mno-sched-br-in-data-spec
20788 @opindex msched-br-in-data-spec
20789 @opindex mno-sched-br-in-data-spec
20790 (En/Dis)able speculative scheduling of the instructions that
20791 are dependent on the data speculative loads before reload.
20792 This is effective only with @option{-msched-br-data-spec} enabled.
20793 The default setting is enabled.
20795 @item -msched-ar-in-data-spec
20796 @itemx -mno-sched-ar-in-data-spec
20797 @opindex msched-ar-in-data-spec
20798 @opindex mno-sched-ar-in-data-spec
20799 (En/Dis)able speculative scheduling of the instructions that
20800 are dependent on the data speculative loads after reload.
20801 This is effective only with @option{-msched-ar-data-spec} enabled.
20802 The default setting is enabled.
20804 @item -msched-in-control-spec
20805 @itemx -mno-sched-in-control-spec
20806 @opindex msched-in-control-spec
20807 @opindex mno-sched-in-control-spec
20808 (En/Dis)able speculative scheduling of the instructions that
20809 are dependent on the control speculative loads.
20810 This is effective only with @option{-msched-control-spec} enabled.
20811 The default setting is enabled.
20813 @item -mno-sched-prefer-non-data-spec-insns
20814 @itemx -msched-prefer-non-data-spec-insns
20815 @opindex mno-sched-prefer-non-data-spec-insns
20816 @opindex msched-prefer-non-data-spec-insns
20817 If enabled, data-speculative instructions are chosen for schedule
20818 only if there are no other choices at the moment. This makes
20819 the use of the data speculation much more conservative.
20820 The default setting is disabled.
20822 @item -mno-sched-prefer-non-control-spec-insns
20823 @itemx -msched-prefer-non-control-spec-insns
20824 @opindex mno-sched-prefer-non-control-spec-insns
20825 @opindex msched-prefer-non-control-spec-insns
20826 If enabled, control-speculative instructions are chosen for schedule
20827 only if there are no other choices at the moment. This makes
20828 the use of the control speculation much more conservative.
20829 The default setting is disabled.
20831 @item -mno-sched-count-spec-in-critical-path
20832 @itemx -msched-count-spec-in-critical-path
20833 @opindex mno-sched-count-spec-in-critical-path
20834 @opindex msched-count-spec-in-critical-path
20835 If enabled, speculative dependencies are considered during
20836 computation of the instructions priorities. This makes the use of the
20837 speculation a bit more conservative.
20838 The default setting is disabled.
20840 @item -msched-spec-ldc
20841 @opindex msched-spec-ldc
20842 Use a simple data speculation check. This option is on by default.
20844 @item -msched-control-spec-ldc
20845 @opindex msched-spec-ldc
20846 Use a simple check for control speculation. This option is on by default.
20848 @item -msched-stop-bits-after-every-cycle
20849 @opindex msched-stop-bits-after-every-cycle
20850 Place a stop bit after every cycle when scheduling. This option is on
20853 @item -msched-fp-mem-deps-zero-cost
20854 @opindex msched-fp-mem-deps-zero-cost
20855 Assume that floating-point stores and loads are not likely to cause a conflict
20856 when placed into the same instruction group. This option is disabled by
20859 @item -msel-sched-dont-check-control-spec
20860 @opindex msel-sched-dont-check-control-spec
20861 Generate checks for control speculation in selective scheduling.
20862 This flag is disabled by default.
20864 @item -msched-max-memory-insns=@var{max-insns}
20865 @opindex msched-max-memory-insns
20866 Limit on the number of memory insns per instruction group, giving lower
20867 priority to subsequent memory insns attempting to schedule in the same
20868 instruction group. Frequently useful to prevent cache bank conflicts.
20869 The default value is 1.
20871 @item -msched-max-memory-insns-hard-limit
20872 @opindex msched-max-memory-insns-hard-limit
20873 Makes the limit specified by @option{msched-max-memory-insns} a hard limit,
20874 disallowing more than that number in an instruction group.
20875 Otherwise, the limit is ``soft'', meaning that non-memory operations
20876 are preferred when the limit is reached, but memory operations may still
20882 @subsection LM32 Options
20883 @cindex LM32 options
20885 These @option{-m} options are defined for the LatticeMico32 architecture:
20888 @item -mbarrel-shift-enabled
20889 @opindex mbarrel-shift-enabled
20890 Enable barrel-shift instructions.
20892 @item -mdivide-enabled
20893 @opindex mdivide-enabled
20894 Enable divide and modulus instructions.
20896 @item -mmultiply-enabled
20897 @opindex multiply-enabled
20898 Enable multiply instructions.
20900 @item -msign-extend-enabled
20901 @opindex msign-extend-enabled
20902 Enable sign extend instructions.
20904 @item -muser-enabled
20905 @opindex muser-enabled
20906 Enable user-defined instructions.
20911 @subsection M32C Options
20912 @cindex M32C options
20915 @item -mcpu=@var{name}
20917 Select the CPU for which code is generated. @var{name} may be one of
20918 @samp{r8c} for the R8C/Tiny series, @samp{m16c} for the M16C (up to
20919 /60) series, @samp{m32cm} for the M16C/80 series, or @samp{m32c} for
20920 the M32C/80 series.
20924 Specifies that the program will be run on the simulator. This causes
20925 an alternate runtime library to be linked in which supports, for
20926 example, file I/O@. You must not use this option when generating
20927 programs that will run on real hardware; you must provide your own
20928 runtime library for whatever I/O functions are needed.
20930 @item -memregs=@var{number}
20932 Specifies the number of memory-based pseudo-registers GCC uses
20933 during code generation. These pseudo-registers are used like real
20934 registers, so there is a tradeoff between GCC's ability to fit the
20935 code into available registers, and the performance penalty of using
20936 memory instead of registers. Note that all modules in a program must
20937 be compiled with the same value for this option. Because of that, you
20938 must not use this option with GCC's default runtime libraries.
20942 @node M32R/D Options
20943 @subsection M32R/D Options
20944 @cindex M32R/D options
20946 These @option{-m} options are defined for Renesas M32R/D architectures:
20951 Generate code for the M32R/2@.
20955 Generate code for the M32R/X@.
20959 Generate code for the M32R@. This is the default.
20961 @item -mmodel=small
20962 @opindex mmodel=small
20963 Assume all objects live in the lower 16MB of memory (so that their addresses
20964 can be loaded with the @code{ld24} instruction), and assume all subroutines
20965 are reachable with the @code{bl} instruction.
20966 This is the default.
20968 The addressability of a particular object can be set with the
20969 @code{model} attribute.
20971 @item -mmodel=medium
20972 @opindex mmodel=medium
20973 Assume objects may be anywhere in the 32-bit address space (the compiler
20974 generates @code{seth/add3} instructions to load their addresses), and
20975 assume all subroutines are reachable with the @code{bl} instruction.
20977 @item -mmodel=large
20978 @opindex mmodel=large
20979 Assume objects may be anywhere in the 32-bit address space (the compiler
20980 generates @code{seth/add3} instructions to load their addresses), and
20981 assume subroutines may not be reachable with the @code{bl} instruction
20982 (the compiler generates the much slower @code{seth/add3/jl}
20983 instruction sequence).
20986 @opindex msdata=none
20987 Disable use of the small data area. Variables are put into
20988 one of @code{.data}, @code{.bss}, or @code{.rodata} (unless the
20989 @code{section} attribute has been specified).
20990 This is the default.
20992 The small data area consists of sections @code{.sdata} and @code{.sbss}.
20993 Objects may be explicitly put in the small data area with the
20994 @code{section} attribute using one of these sections.
20996 @item -msdata=sdata
20997 @opindex msdata=sdata
20998 Put small global and static data in the small data area, but do not
20999 generate special code to reference them.
21002 @opindex msdata=use
21003 Put small global and static data in the small data area, and generate
21004 special instructions to reference them.
21008 @cindex smaller data references
21009 Put global and static objects less than or equal to @var{num} bytes
21010 into the small data or BSS sections instead of the normal data or BSS
21011 sections. The default value of @var{num} is 8.
21012 The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use}
21013 for this option to have any effect.
21015 All modules should be compiled with the same @option{-G @var{num}} value.
21016 Compiling with different values of @var{num} may or may not work; if it
21017 doesn't the linker gives an error message---incorrect code is not
21022 Makes the M32R-specific code in the compiler display some statistics
21023 that might help in debugging programs.
21025 @item -malign-loops
21026 @opindex malign-loops
21027 Align all loops to a 32-byte boundary.
21029 @item -mno-align-loops
21030 @opindex mno-align-loops
21031 Do not enforce a 32-byte alignment for loops. This is the default.
21033 @item -missue-rate=@var{number}
21034 @opindex missue-rate=@var{number}
21035 Issue @var{number} instructions per cycle. @var{number} can only be 1
21038 @item -mbranch-cost=@var{number}
21039 @opindex mbranch-cost=@var{number}
21040 @var{number} can only be 1 or 2. If it is 1 then branches are
21041 preferred over conditional code, if it is 2, then the opposite applies.
21043 @item -mflush-trap=@var{number}
21044 @opindex mflush-trap=@var{number}
21045 Specifies the trap number to use to flush the cache. The default is
21046 12. Valid numbers are between 0 and 15 inclusive.
21048 @item -mno-flush-trap
21049 @opindex mno-flush-trap
21050 Specifies that the cache cannot be flushed by using a trap.
21052 @item -mflush-func=@var{name}
21053 @opindex mflush-func=@var{name}
21054 Specifies the name of the operating system function to call to flush
21055 the cache. The default is @samp{_flush_cache}, but a function call
21056 is only used if a trap is not available.
21058 @item -mno-flush-func
21059 @opindex mno-flush-func
21060 Indicates that there is no OS function for flushing the cache.
21064 @node M680x0 Options
21065 @subsection M680x0 Options
21066 @cindex M680x0 options
21068 These are the @samp{-m} options defined for M680x0 and ColdFire processors.
21069 The default settings depend on which architecture was selected when
21070 the compiler was configured; the defaults for the most common choices
21074 @item -march=@var{arch}
21076 Generate code for a specific M680x0 or ColdFire instruction set
21077 architecture. Permissible values of @var{arch} for M680x0
21078 architectures are: @samp{68000}, @samp{68010}, @samp{68020},
21079 @samp{68030}, @samp{68040}, @samp{68060} and @samp{cpu32}. ColdFire
21080 architectures are selected according to Freescale's ISA classification
21081 and the permissible values are: @samp{isaa}, @samp{isaaplus},
21082 @samp{isab} and @samp{isac}.
21084 GCC defines a macro @code{__mcf@var{arch}__} whenever it is generating
21085 code for a ColdFire target. The @var{arch} in this macro is one of the
21086 @option{-march} arguments given above.
21088 When used together, @option{-march} and @option{-mtune} select code
21089 that runs on a family of similar processors but that is optimized
21090 for a particular microarchitecture.
21092 @item -mcpu=@var{cpu}
21094 Generate code for a specific M680x0 or ColdFire processor.
21095 The M680x0 @var{cpu}s are: @samp{68000}, @samp{68010}, @samp{68020},
21096 @samp{68030}, @samp{68040}, @samp{68060}, @samp{68302}, @samp{68332}
21097 and @samp{cpu32}. The ColdFire @var{cpu}s are given by the table
21098 below, which also classifies the CPUs into families:
21100 @multitable @columnfractions 0.20 0.80
21101 @item @strong{Family} @tab @strong{@samp{-mcpu} arguments}
21102 @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}
21103 @item @samp{5206} @tab @samp{5202} @samp{5204} @samp{5206}
21104 @item @samp{5206e} @tab @samp{5206e}
21105 @item @samp{5208} @tab @samp{5207} @samp{5208}
21106 @item @samp{5211a} @tab @samp{5210a} @samp{5211a}
21107 @item @samp{5213} @tab @samp{5211} @samp{5212} @samp{5213}
21108 @item @samp{5216} @tab @samp{5214} @samp{5216}
21109 @item @samp{52235} @tab @samp{52230} @samp{52231} @samp{52232} @samp{52233} @samp{52234} @samp{52235}
21110 @item @samp{5225} @tab @samp{5224} @samp{5225}
21111 @item @samp{52259} @tab @samp{52252} @samp{52254} @samp{52255} @samp{52256} @samp{52258} @samp{52259}
21112 @item @samp{5235} @tab @samp{5232} @samp{5233} @samp{5234} @samp{5235} @samp{523x}
21113 @item @samp{5249} @tab @samp{5249}
21114 @item @samp{5250} @tab @samp{5250}
21115 @item @samp{5271} @tab @samp{5270} @samp{5271}
21116 @item @samp{5272} @tab @samp{5272}
21117 @item @samp{5275} @tab @samp{5274} @samp{5275}
21118 @item @samp{5282} @tab @samp{5280} @samp{5281} @samp{5282} @samp{528x}
21119 @item @samp{53017} @tab @samp{53011} @samp{53012} @samp{53013} @samp{53014} @samp{53015} @samp{53016} @samp{53017}
21120 @item @samp{5307} @tab @samp{5307}
21121 @item @samp{5329} @tab @samp{5327} @samp{5328} @samp{5329} @samp{532x}
21122 @item @samp{5373} @tab @samp{5372} @samp{5373} @samp{537x}
21123 @item @samp{5407} @tab @samp{5407}
21124 @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}
21127 @option{-mcpu=@var{cpu}} overrides @option{-march=@var{arch}} if
21128 @var{arch} is compatible with @var{cpu}. Other combinations of
21129 @option{-mcpu} and @option{-march} are rejected.
21131 GCC defines the macro @code{__mcf_cpu_@var{cpu}} when ColdFire target
21132 @var{cpu} is selected. It also defines @code{__mcf_family_@var{family}},
21133 where the value of @var{family} is given by the table above.
21135 @item -mtune=@var{tune}
21137 Tune the code for a particular microarchitecture within the
21138 constraints set by @option{-march} and @option{-mcpu}.
21139 The M680x0 microarchitectures are: @samp{68000}, @samp{68010},
21140 @samp{68020}, @samp{68030}, @samp{68040}, @samp{68060}
21141 and @samp{cpu32}. The ColdFire microarchitectures
21142 are: @samp{cfv1}, @samp{cfv2}, @samp{cfv3}, @samp{cfv4} and @samp{cfv4e}.
21144 You can also use @option{-mtune=68020-40} for code that needs
21145 to run relatively well on 68020, 68030 and 68040 targets.
21146 @option{-mtune=68020-60} is similar but includes 68060 targets
21147 as well. These two options select the same tuning decisions as
21148 @option{-m68020-40} and @option{-m68020-60} respectively.
21150 GCC defines the macros @code{__mc@var{arch}} and @code{__mc@var{arch}__}
21151 when tuning for 680x0 architecture @var{arch}. It also defines
21152 @code{mc@var{arch}} unless either @option{-ansi} or a non-GNU @option{-std}
21153 option is used. If GCC is tuning for a range of architectures,
21154 as selected by @option{-mtune=68020-40} or @option{-mtune=68020-60},
21155 it defines the macros for every architecture in the range.
21157 GCC also defines the macro @code{__m@var{uarch}__} when tuning for
21158 ColdFire microarchitecture @var{uarch}, where @var{uarch} is one
21159 of the arguments given above.
21165 Generate output for a 68000. This is the default
21166 when the compiler is configured for 68000-based systems.
21167 It is equivalent to @option{-march=68000}.
21169 Use this option for microcontrollers with a 68000 or EC000 core,
21170 including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356.
21174 Generate output for a 68010. This is the default
21175 when the compiler is configured for 68010-based systems.
21176 It is equivalent to @option{-march=68010}.
21182 Generate output for a 68020. This is the default
21183 when the compiler is configured for 68020-based systems.
21184 It is equivalent to @option{-march=68020}.
21188 Generate output for a 68030. This is the default when the compiler is
21189 configured for 68030-based systems. It is equivalent to
21190 @option{-march=68030}.
21194 Generate output for a 68040. This is the default when the compiler is
21195 configured for 68040-based systems. It is equivalent to
21196 @option{-march=68040}.
21198 This option inhibits the use of 68881/68882 instructions that have to be
21199 emulated by software on the 68040. Use this option if your 68040 does not
21200 have code to emulate those instructions.
21204 Generate output for a 68060. This is the default when the compiler is
21205 configured for 68060-based systems. It is equivalent to
21206 @option{-march=68060}.
21208 This option inhibits the use of 68020 and 68881/68882 instructions that
21209 have to be emulated by software on the 68060. Use this option if your 68060
21210 does not have code to emulate those instructions.
21214 Generate output for a CPU32. This is the default
21215 when the compiler is configured for CPU32-based systems.
21216 It is equivalent to @option{-march=cpu32}.
21218 Use this option for microcontrollers with a
21219 CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334,
21220 68336, 68340, 68341, 68349 and 68360.
21224 Generate output for a 520X ColdFire CPU@. This is the default
21225 when the compiler is configured for 520X-based systems.
21226 It is equivalent to @option{-mcpu=5206}, and is now deprecated
21227 in favor of that option.
21229 Use this option for microcontroller with a 5200 core, including
21230 the MCF5202, MCF5203, MCF5204 and MCF5206.
21234 Generate output for a 5206e ColdFire CPU@. The option is now
21235 deprecated in favor of the equivalent @option{-mcpu=5206e}.
21239 Generate output for a member of the ColdFire 528X family.
21240 The option is now deprecated in favor of the equivalent
21241 @option{-mcpu=528x}.
21245 Generate output for a ColdFire 5307 CPU@. The option is now deprecated
21246 in favor of the equivalent @option{-mcpu=5307}.
21250 Generate output for a ColdFire 5407 CPU@. The option is now deprecated
21251 in favor of the equivalent @option{-mcpu=5407}.
21255 Generate output for a ColdFire V4e family CPU (e.g.@: 547x/548x).
21256 This includes use of hardware floating-point instructions.
21257 The option is equivalent to @option{-mcpu=547x}, and is now
21258 deprecated in favor of that option.
21262 Generate output for a 68040, without using any of the new instructions.
21263 This results in code that can run relatively efficiently on either a
21264 68020/68881 or a 68030 or a 68040. The generated code does use the
21265 68881 instructions that are emulated on the 68040.
21267 The option is equivalent to @option{-march=68020} @option{-mtune=68020-40}.
21271 Generate output for a 68060, without using any of the new instructions.
21272 This results in code that can run relatively efficiently on either a
21273 68020/68881 or a 68030 or a 68040. The generated code does use the
21274 68881 instructions that are emulated on the 68060.
21276 The option is equivalent to @option{-march=68020} @option{-mtune=68020-60}.
21280 @opindex mhard-float
21282 Generate floating-point instructions. This is the default for 68020
21283 and above, and for ColdFire devices that have an FPU@. It defines the
21284 macro @code{__HAVE_68881__} on M680x0 targets and @code{__mcffpu__}
21285 on ColdFire targets.
21288 @opindex msoft-float
21289 Do not generate floating-point instructions; use library calls instead.
21290 This is the default for 68000, 68010, and 68832 targets. It is also
21291 the default for ColdFire devices that have no FPU.
21297 Generate (do not generate) ColdFire hardware divide and remainder
21298 instructions. If @option{-march} is used without @option{-mcpu},
21299 the default is ``on'' for ColdFire architectures and ``off'' for M680x0
21300 architectures. Otherwise, the default is taken from the target CPU
21301 (either the default CPU, or the one specified by @option{-mcpu}). For
21302 example, the default is ``off'' for @option{-mcpu=5206} and ``on'' for
21303 @option{-mcpu=5206e}.
21305 GCC defines the macro @code{__mcfhwdiv__} when this option is enabled.
21309 Consider type @code{int} to be 16 bits wide, like @code{short int}.
21310 Additionally, parameters passed on the stack are also aligned to a
21311 16-bit boundary even on targets whose API mandates promotion to 32-bit.
21315 Do not consider type @code{int} to be 16 bits wide. This is the default.
21318 @itemx -mno-bitfield
21319 @opindex mnobitfield
21320 @opindex mno-bitfield
21321 Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32}
21322 and @option{-m5200} options imply @w{@option{-mnobitfield}}.
21326 Do use the bit-field instructions. The @option{-m68020} option implies
21327 @option{-mbitfield}. This is the default if you use a configuration
21328 designed for a 68020.
21332 Use a different function-calling convention, in which functions
21333 that take a fixed number of arguments return with the @code{rtd}
21334 instruction, which pops their arguments while returning. This
21335 saves one instruction in the caller since there is no need to pop
21336 the arguments there.
21338 This calling convention is incompatible with the one normally
21339 used on Unix, so you cannot use it if you need to call libraries
21340 compiled with the Unix compiler.
21342 Also, you must provide function prototypes for all functions that
21343 take variable numbers of arguments (including @code{printf});
21344 otherwise incorrect code is generated for calls to those
21347 In addition, seriously incorrect code results if you call a
21348 function with too many arguments. (Normally, extra arguments are
21349 harmlessly ignored.)
21351 The @code{rtd} instruction is supported by the 68010, 68020, 68030,
21352 68040, 68060 and CPU32 processors, but not by the 68000 or 5200.
21354 The default is @option{-mno-rtd}.
21357 @itemx -mno-align-int
21358 @opindex malign-int
21359 @opindex mno-align-int
21360 Control whether GCC aligns @code{int}, @code{long}, @code{long long},
21361 @code{float}, @code{double}, and @code{long double} variables on a 32-bit
21362 boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}).
21363 Aligning variables on 32-bit boundaries produces code that runs somewhat
21364 faster on processors with 32-bit busses at the expense of more memory.
21366 @strong{Warning:} if you use the @option{-malign-int} switch, GCC
21367 aligns structures containing the above types differently than
21368 most published application binary interface specifications for the m68k.
21372 Use the pc-relative addressing mode of the 68000 directly, instead of
21373 using a global offset table. At present, this option implies @option{-fpic},
21374 allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is
21375 not presently supported with @option{-mpcrel}, though this could be supported for
21376 68020 and higher processors.
21378 @item -mno-strict-align
21379 @itemx -mstrict-align
21380 @opindex mno-strict-align
21381 @opindex mstrict-align
21382 Do not (do) assume that unaligned memory references are handled by
21386 Generate code that allows the data segment to be located in a different
21387 area of memory from the text segment. This allows for execute-in-place in
21388 an environment without virtual memory management. This option implies
21391 @item -mno-sep-data
21392 Generate code that assumes that the data segment follows the text segment.
21393 This is the default.
21395 @item -mid-shared-library
21396 Generate code that supports shared libraries via the library ID method.
21397 This allows for execute-in-place and shared libraries in an environment
21398 without virtual memory management. This option implies @option{-fPIC}.
21400 @item -mno-id-shared-library
21401 Generate code that doesn't assume ID-based shared libraries are being used.
21402 This is the default.
21404 @item -mshared-library-id=n
21405 Specifies the identification number of the ID-based shared library being
21406 compiled. Specifying a value of 0 generates more compact code; specifying
21407 other values forces the allocation of that number to the current
21408 library, but is no more space- or time-efficient than omitting this option.
21414 When generating position-independent code for ColdFire, generate code
21415 that works if the GOT has more than 8192 entries. This code is
21416 larger and slower than code generated without this option. On M680x0
21417 processors, this option is not needed; @option{-fPIC} suffices.
21419 GCC normally uses a single instruction to load values from the GOT@.
21420 While this is relatively efficient, it only works if the GOT
21421 is smaller than about 64k. Anything larger causes the linker
21422 to report an error such as:
21424 @cindex relocation truncated to fit (ColdFire)
21426 relocation truncated to fit: R_68K_GOT16O foobar
21429 If this happens, you should recompile your code with @option{-mxgot}.
21430 It should then work with very large GOTs. However, code generated with
21431 @option{-mxgot} is less efficient, since it takes 4 instructions to fetch
21432 the value of a global symbol.
21434 Note that some linkers, including newer versions of the GNU linker,
21435 can create multiple GOTs and sort GOT entries. If you have such a linker,
21436 you should only need to use @option{-mxgot} when compiling a single
21437 object file that accesses more than 8192 GOT entries. Very few do.
21439 These options have no effect unless GCC is generating
21440 position-independent code.
21442 @item -mlong-jump-table-offsets
21443 @opindex mlong-jump-table-offsets
21444 Use 32-bit offsets in @code{switch} tables. The default is to use
21449 @node MCore Options
21450 @subsection MCore Options
21451 @cindex MCore options
21453 These are the @samp{-m} options defined for the Motorola M*Core
21459 @itemx -mno-hardlit
21461 @opindex mno-hardlit
21462 Inline constants into the code stream if it can be done in two
21463 instructions or less.
21469 Use the divide instruction. (Enabled by default).
21471 @item -mrelax-immediate
21472 @itemx -mno-relax-immediate
21473 @opindex mrelax-immediate
21474 @opindex mno-relax-immediate
21475 Allow arbitrary-sized immediates in bit operations.
21477 @item -mwide-bitfields
21478 @itemx -mno-wide-bitfields
21479 @opindex mwide-bitfields
21480 @opindex mno-wide-bitfields
21481 Always treat bit-fields as @code{int}-sized.
21483 @item -m4byte-functions
21484 @itemx -mno-4byte-functions
21485 @opindex m4byte-functions
21486 @opindex mno-4byte-functions
21487 Force all functions to be aligned to a 4-byte boundary.
21489 @item -mcallgraph-data
21490 @itemx -mno-callgraph-data
21491 @opindex mcallgraph-data
21492 @opindex mno-callgraph-data
21493 Emit callgraph information.
21496 @itemx -mno-slow-bytes
21497 @opindex mslow-bytes
21498 @opindex mno-slow-bytes
21499 Prefer word access when reading byte quantities.
21501 @item -mlittle-endian
21502 @itemx -mbig-endian
21503 @opindex mlittle-endian
21504 @opindex mbig-endian
21505 Generate code for a little-endian target.
21511 Generate code for the 210 processor.
21515 Assume that runtime support has been provided and so omit the
21516 simulator library (@file{libsim.a)} from the linker command line.
21518 @item -mstack-increment=@var{size}
21519 @opindex mstack-increment
21520 Set the maximum amount for a single stack increment operation. Large
21521 values can increase the speed of programs that contain functions
21522 that need a large amount of stack space, but they can also trigger a
21523 segmentation fault if the stack is extended too much. The default
21529 @subsection MeP Options
21530 @cindex MeP options
21536 Enables the @code{abs} instruction, which is the absolute difference
21537 between two registers.
21541 Enables all the optional instructions---average, multiply, divide, bit
21542 operations, leading zero, absolute difference, min/max, clip, and
21548 Enables the @code{ave} instruction, which computes the average of two
21551 @item -mbased=@var{n}
21553 Variables of size @var{n} bytes or smaller are placed in the
21554 @code{.based} section by default. Based variables use the @code{$tp}
21555 register as a base register, and there is a 128-byte limit to the
21556 @code{.based} section.
21560 Enables the bit operation instructions---bit test (@code{btstm}), set
21561 (@code{bsetm}), clear (@code{bclrm}), invert (@code{bnotm}), and
21562 test-and-set (@code{tas}).
21564 @item -mc=@var{name}
21566 Selects which section constant data is placed in. @var{name} may
21567 be @samp{tiny}, @samp{near}, or @samp{far}.
21571 Enables the @code{clip} instruction. Note that @option{-mclip} is not
21572 useful unless you also provide @option{-mminmax}.
21574 @item -mconfig=@var{name}
21576 Selects one of the built-in core configurations. Each MeP chip has
21577 one or more modules in it; each module has a core CPU and a variety of
21578 coprocessors, optional instructions, and peripherals. The
21579 @code{MeP-Integrator} tool, not part of GCC, provides these
21580 configurations through this option; using this option is the same as
21581 using all the corresponding command-line options. The default
21582 configuration is @samp{default}.
21586 Enables the coprocessor instructions. By default, this is a 32-bit
21587 coprocessor. Note that the coprocessor is normally enabled via the
21588 @option{-mconfig=} option.
21592 Enables the 32-bit coprocessor's instructions.
21596 Enables the 64-bit coprocessor's instructions.
21600 Enables IVC2 scheduling. IVC2 is a 64-bit VLIW coprocessor.
21604 Causes constant variables to be placed in the @code{.near} section.
21608 Enables the @code{div} and @code{divu} instructions.
21612 Generate big-endian code.
21616 Generate little-endian code.
21618 @item -mio-volatile
21619 @opindex mio-volatile
21620 Tells the compiler that any variable marked with the @code{io}
21621 attribute is to be considered volatile.
21625 Causes variables to be assigned to the @code{.far} section by default.
21629 Enables the @code{leadz} (leading zero) instruction.
21633 Causes variables to be assigned to the @code{.near} section by default.
21637 Enables the @code{min} and @code{max} instructions.
21641 Enables the multiplication and multiply-accumulate instructions.
21645 Disables all the optional instructions enabled by @option{-mall-opts}.
21649 Enables the @code{repeat} and @code{erepeat} instructions, used for
21650 low-overhead looping.
21654 Causes all variables to default to the @code{.tiny} section. Note
21655 that there is a 65536-byte limit to this section. Accesses to these
21656 variables use the @code{%gp} base register.
21660 Enables the saturation instructions. Note that the compiler does not
21661 currently generate these itself, but this option is included for
21662 compatibility with other tools, like @code{as}.
21666 Link the SDRAM-based runtime instead of the default ROM-based runtime.
21670 Link the simulator run-time libraries.
21674 Link the simulator runtime libraries, excluding built-in support
21675 for reset and exception vectors and tables.
21679 Causes all functions to default to the @code{.far} section. Without
21680 this option, functions default to the @code{.near} section.
21682 @item -mtiny=@var{n}
21684 Variables that are @var{n} bytes or smaller are allocated to the
21685 @code{.tiny} section. These variables use the @code{$gp} base
21686 register. The default for this option is 4, but note that there's a
21687 65536-byte limit to the @code{.tiny} section.
21691 @node MicroBlaze Options
21692 @subsection MicroBlaze Options
21693 @cindex MicroBlaze Options
21698 @opindex msoft-float
21699 Use software emulation for floating point (default).
21702 @opindex mhard-float
21703 Use hardware floating-point instructions.
21707 Do not optimize block moves, use @code{memcpy}.
21709 @item -mno-clearbss
21710 @opindex mno-clearbss
21711 This option is deprecated. Use @option{-fno-zero-initialized-in-bss} instead.
21713 @item -mcpu=@var{cpu-type}
21715 Use features of, and schedule code for, the given CPU.
21716 Supported values are in the format @samp{v@var{X}.@var{YY}.@var{Z}},
21717 where @var{X} is a major version, @var{YY} is the minor version, and
21718 @var{Z} is compatibility code. Example values are @samp{v3.00.a},
21719 @samp{v4.00.b}, @samp{v5.00.a}, @samp{v5.00.b}, @samp{v6.00.a}.
21721 @item -mxl-soft-mul
21722 @opindex mxl-soft-mul
21723 Use software multiply emulation (default).
21725 @item -mxl-soft-div
21726 @opindex mxl-soft-div
21727 Use software emulation for divides (default).
21729 @item -mxl-barrel-shift
21730 @opindex mxl-barrel-shift
21731 Use the hardware barrel shifter.
21733 @item -mxl-pattern-compare
21734 @opindex mxl-pattern-compare
21735 Use pattern compare instructions.
21737 @item -msmall-divides
21738 @opindex msmall-divides
21739 Use table lookup optimization for small signed integer divisions.
21741 @item -mxl-stack-check
21742 @opindex mxl-stack-check
21743 This option is deprecated. Use @option{-fstack-check} instead.
21746 @opindex mxl-gp-opt
21747 Use GP-relative @code{.sdata}/@code{.sbss} sections.
21749 @item -mxl-multiply-high
21750 @opindex mxl-multiply-high
21751 Use multiply high instructions for high part of 32x32 multiply.
21753 @item -mxl-float-convert
21754 @opindex mxl-float-convert
21755 Use hardware floating-point conversion instructions.
21757 @item -mxl-float-sqrt
21758 @opindex mxl-float-sqrt
21759 Use hardware floating-point square root instruction.
21762 @opindex mbig-endian
21763 Generate code for a big-endian target.
21765 @item -mlittle-endian
21766 @opindex mlittle-endian
21767 Generate code for a little-endian target.
21770 @opindex mxl-reorder
21771 Use reorder instructions (swap and byte reversed load/store).
21773 @item -mxl-mode-@var{app-model}
21774 Select application model @var{app-model}. Valid models are
21777 normal executable (default), uses startup code @file{crt0.o}.
21779 @item -mpic-data-is-text-relative
21780 @opindex mpic-data-is-text-relative
21781 Assume that the displacement between the text and data segments is fixed
21782 at static link time. This allows data to be referenced by offset from start of
21783 text address instead of GOT since PC-relative addressing is not supported.
21786 for use with Xilinx Microprocessor Debugger (XMD) based
21787 software intrusive debug agent called xmdstub. This uses startup file
21788 @file{crt1.o} and sets the start address of the program to 0x800.
21791 for applications that are loaded using a bootloader.
21792 This model uses startup file @file{crt2.o} which does not contain a processor
21793 reset vector handler. This is suitable for transferring control on a
21794 processor reset to the bootloader rather than the application.
21797 for applications that do not require any of the
21798 MicroBlaze vectors. This option may be useful for applications running
21799 within a monitoring application. This model uses @file{crt3.o} as a startup file.
21802 Option @option{-xl-mode-@var{app-model}} is a deprecated alias for
21803 @option{-mxl-mode-@var{app-model}}.
21808 @subsection MIPS Options
21809 @cindex MIPS options
21815 Generate big-endian code.
21819 Generate little-endian code. This is the default for @samp{mips*el-*-*}
21822 @item -march=@var{arch}
21824 Generate code that runs on @var{arch}, which can be the name of a
21825 generic MIPS ISA, or the name of a particular processor.
21827 @samp{mips1}, @samp{mips2}, @samp{mips3}, @samp{mips4},
21828 @samp{mips32}, @samp{mips32r2}, @samp{mips32r3}, @samp{mips32r5},
21829 @samp{mips32r6}, @samp{mips64}, @samp{mips64r2}, @samp{mips64r3},
21830 @samp{mips64r5} and @samp{mips64r6}.
21831 The processor names are:
21832 @samp{4kc}, @samp{4km}, @samp{4kp}, @samp{4ksc},
21833 @samp{4kec}, @samp{4kem}, @samp{4kep}, @samp{4ksd},
21834 @samp{5kc}, @samp{5kf},
21836 @samp{24kc}, @samp{24kf2_1}, @samp{24kf1_1},
21837 @samp{24kec}, @samp{24kef2_1}, @samp{24kef1_1},
21838 @samp{34kc}, @samp{34kf2_1}, @samp{34kf1_1}, @samp{34kn},
21839 @samp{74kc}, @samp{74kf2_1}, @samp{74kf1_1}, @samp{74kf3_2},
21840 @samp{1004kc}, @samp{1004kf2_1}, @samp{1004kf1_1},
21841 @samp{i6400}, @samp{i6500},
21843 @samp{loongson2e}, @samp{loongson2f}, @samp{loongson3a}, @samp{gs464},
21844 @samp{gs464e}, @samp{gs264e},
21846 @samp{m14k}, @samp{m14kc}, @samp{m14ke}, @samp{m14kec},
21847 @samp{m5100}, @samp{m5101},
21848 @samp{octeon}, @samp{octeon+}, @samp{octeon2}, @samp{octeon3},
21850 @samp{p5600}, @samp{p6600},
21851 @samp{r2000}, @samp{r3000}, @samp{r3900}, @samp{r4000}, @samp{r4400},
21852 @samp{r4600}, @samp{r4650}, @samp{r4700}, @samp{r5900},
21853 @samp{r6000}, @samp{r8000},
21854 @samp{rm7000}, @samp{rm9000},
21855 @samp{r10000}, @samp{r12000}, @samp{r14000}, @samp{r16000},
21858 @samp{vr4100}, @samp{vr4111}, @samp{vr4120}, @samp{vr4130}, @samp{vr4300},
21859 @samp{vr5000}, @samp{vr5400}, @samp{vr5500},
21860 @samp{xlr} and @samp{xlp}.
21861 The special value @samp{from-abi} selects the
21862 most compatible architecture for the selected ABI (that is,
21863 @samp{mips1} for 32-bit ABIs and @samp{mips3} for 64-bit ABIs)@.
21865 The native Linux/GNU toolchain also supports the value @samp{native},
21866 which selects the best architecture option for the host processor.
21867 @option{-march=native} has no effect if GCC does not recognize
21870 In processor names, a final @samp{000} can be abbreviated as @samp{k}
21871 (for example, @option{-march=r2k}). Prefixes are optional, and
21872 @samp{vr} may be written @samp{r}.
21874 Names of the form @samp{@var{n}f2_1} refer to processors with
21875 FPUs clocked at half the rate of the core, names of the form
21876 @samp{@var{n}f1_1} refer to processors with FPUs clocked at the same
21877 rate as the core, and names of the form @samp{@var{n}f3_2} refer to
21878 processors with FPUs clocked a ratio of 3:2 with respect to the core.
21879 For compatibility reasons, @samp{@var{n}f} is accepted as a synonym
21880 for @samp{@var{n}f2_1} while @samp{@var{n}x} and @samp{@var{b}fx} are
21881 accepted as synonyms for @samp{@var{n}f1_1}.
21883 GCC defines two macros based on the value of this option. The first
21884 is @code{_MIPS_ARCH}, which gives the name of target architecture, as
21885 a string. The second has the form @code{_MIPS_ARCH_@var{foo}},
21886 where @var{foo} is the capitalized value of @code{_MIPS_ARCH}@.
21887 For example, @option{-march=r2000} sets @code{_MIPS_ARCH}
21888 to @code{"r2000"} and defines the macro @code{_MIPS_ARCH_R2000}.
21890 Note that the @code{_MIPS_ARCH} macro uses the processor names given
21891 above. In other words, it has the full prefix and does not
21892 abbreviate @samp{000} as @samp{k}. In the case of @samp{from-abi},
21893 the macro names the resolved architecture (either @code{"mips1"} or
21894 @code{"mips3"}). It names the default architecture when no
21895 @option{-march} option is given.
21897 @item -mtune=@var{arch}
21899 Optimize for @var{arch}. Among other things, this option controls
21900 the way instructions are scheduled, and the perceived cost of arithmetic
21901 operations. The list of @var{arch} values is the same as for
21904 When this option is not used, GCC optimizes for the processor
21905 specified by @option{-march}. By using @option{-march} and
21906 @option{-mtune} together, it is possible to generate code that
21907 runs on a family of processors, but optimize the code for one
21908 particular member of that family.
21910 @option{-mtune} defines the macros @code{_MIPS_TUNE} and
21911 @code{_MIPS_TUNE_@var{foo}}, which work in the same way as the
21912 @option{-march} ones described above.
21916 Equivalent to @option{-march=mips1}.
21920 Equivalent to @option{-march=mips2}.
21924 Equivalent to @option{-march=mips3}.
21928 Equivalent to @option{-march=mips4}.
21932 Equivalent to @option{-march=mips32}.
21936 Equivalent to @option{-march=mips32r3}.
21940 Equivalent to @option{-march=mips32r5}.
21944 Equivalent to @option{-march=mips32r6}.
21948 Equivalent to @option{-march=mips64}.
21952 Equivalent to @option{-march=mips64r2}.
21956 Equivalent to @option{-march=mips64r3}.
21960 Equivalent to @option{-march=mips64r5}.
21964 Equivalent to @option{-march=mips64r6}.
21969 @opindex mno-mips16
21970 Generate (do not generate) MIPS16 code. If GCC is targeting a
21971 MIPS32 or MIPS64 architecture, it makes use of the MIPS16e ASE@.
21973 MIPS16 code generation can also be controlled on a per-function basis
21974 by means of @code{mips16} and @code{nomips16} attributes.
21975 @xref{Function Attributes}, for more information.
21977 @item -mflip-mips16
21978 @opindex mflip-mips16
21979 Generate MIPS16 code on alternating functions. This option is provided
21980 for regression testing of mixed MIPS16/non-MIPS16 code generation, and is
21981 not intended for ordinary use in compiling user code.
21983 @item -minterlink-compressed
21984 @itemx -mno-interlink-compressed
21985 @opindex minterlink-compressed
21986 @opindex mno-interlink-compressed
21987 Require (do not require) that code using the standard (uncompressed) MIPS ISA
21988 be link-compatible with MIPS16 and microMIPS code, and vice versa.
21990 For example, code using the standard ISA encoding cannot jump directly
21991 to MIPS16 or microMIPS code; it must either use a call or an indirect jump.
21992 @option{-minterlink-compressed} therefore disables direct jumps unless GCC
21993 knows that the target of the jump is not compressed.
21995 @item -minterlink-mips16
21996 @itemx -mno-interlink-mips16
21997 @opindex minterlink-mips16
21998 @opindex mno-interlink-mips16
21999 Aliases of @option{-minterlink-compressed} and
22000 @option{-mno-interlink-compressed}. These options predate the microMIPS ASE
22001 and are retained for backwards compatibility.
22013 Generate code for the given ABI@.
22015 Note that the EABI has a 32-bit and a 64-bit variant. GCC normally
22016 generates 64-bit code when you select a 64-bit architecture, but you
22017 can use @option{-mgp32} to get 32-bit code instead.
22019 For information about the O64 ABI, see
22020 @uref{http://gcc.gnu.org/@/projects/@/mipso64-abi.html}.
22022 GCC supports a variant of the o32 ABI in which floating-point registers
22023 are 64 rather than 32 bits wide. You can select this combination with
22024 @option{-mabi=32} @option{-mfp64}. This ABI relies on the @code{mthc1}
22025 and @code{mfhc1} instructions and is therefore only supported for
22026 MIPS32R2, MIPS32R3 and MIPS32R5 processors.
22028 The register assignments for arguments and return values remain the
22029 same, but each scalar value is passed in a single 64-bit register
22030 rather than a pair of 32-bit registers. For example, scalar
22031 floating-point values are returned in @samp{$f0} only, not a
22032 @samp{$f0}/@samp{$f1} pair. The set of call-saved registers also
22033 remains the same in that the even-numbered double-precision registers
22036 Two additional variants of the o32 ABI are supported to enable
22037 a transition from 32-bit to 64-bit registers. These are FPXX
22038 (@option{-mfpxx}) and FP64A (@option{-mfp64} @option{-mno-odd-spreg}).
22039 The FPXX extension mandates that all code must execute correctly
22040 when run using 32-bit or 64-bit registers. The code can be interlinked
22041 with either FP32 or FP64, but not both.
22042 The FP64A extension is similar to the FP64 extension but forbids the
22043 use of odd-numbered single-precision registers. This can be used
22044 in conjunction with the @code{FRE} mode of FPUs in MIPS32R5
22045 processors and allows both FP32 and FP64A code to interlink and
22046 run in the same process without changing FPU modes.
22049 @itemx -mno-abicalls
22051 @opindex mno-abicalls
22052 Generate (do not generate) code that is suitable for SVR4-style
22053 dynamic objects. @option{-mabicalls} is the default for SVR4-based
22058 Generate (do not generate) code that is fully position-independent,
22059 and that can therefore be linked into shared libraries. This option
22060 only affects @option{-mabicalls}.
22062 All @option{-mabicalls} code has traditionally been position-independent,
22063 regardless of options like @option{-fPIC} and @option{-fpic}. However,
22064 as an extension, the GNU toolchain allows executables to use absolute
22065 accesses for locally-binding symbols. It can also use shorter GP
22066 initialization sequences and generate direct calls to locally-defined
22067 functions. This mode is selected by @option{-mno-shared}.
22069 @option{-mno-shared} depends on binutils 2.16 or higher and generates
22070 objects that can only be linked by the GNU linker. However, the option
22071 does not affect the ABI of the final executable; it only affects the ABI
22072 of relocatable objects. Using @option{-mno-shared} generally makes
22073 executables both smaller and quicker.
22075 @option{-mshared} is the default.
22081 Assume (do not assume) that the static and dynamic linkers
22082 support PLTs and copy relocations. This option only affects
22083 @option{-mno-shared -mabicalls}. For the n64 ABI, this option
22084 has no effect without @option{-msym32}.
22086 You can make @option{-mplt} the default by configuring
22087 GCC with @option{--with-mips-plt}. The default is
22088 @option{-mno-plt} otherwise.
22094 Lift (do not lift) the usual restrictions on the size of the global
22097 GCC normally uses a single instruction to load values from the GOT@.
22098 While this is relatively efficient, it only works if the GOT
22099 is smaller than about 64k. Anything larger causes the linker
22100 to report an error such as:
22102 @cindex relocation truncated to fit (MIPS)
22104 relocation truncated to fit: R_MIPS_GOT16 foobar
22107 If this happens, you should recompile your code with @option{-mxgot}.
22108 This works with very large GOTs, although the code is also
22109 less efficient, since it takes three instructions to fetch the
22110 value of a global symbol.
22112 Note that some linkers can create multiple GOTs. If you have such a
22113 linker, you should only need to use @option{-mxgot} when a single object
22114 file accesses more than 64k's worth of GOT entries. Very few do.
22116 These options have no effect unless GCC is generating position
22121 Assume that general-purpose registers are 32 bits wide.
22125 Assume that general-purpose registers are 64 bits wide.
22129 Assume that floating-point registers are 32 bits wide.
22133 Assume that floating-point registers are 64 bits wide.
22137 Do not assume the width of floating-point registers.
22140 @opindex mhard-float
22141 Use floating-point coprocessor instructions.
22144 @opindex msoft-float
22145 Do not use floating-point coprocessor instructions. Implement
22146 floating-point calculations using library calls instead.
22150 Equivalent to @option{-msoft-float}, but additionally asserts that the
22151 program being compiled does not perform any floating-point operations.
22152 This option is presently supported only by some bare-metal MIPS
22153 configurations, where it may select a special set of libraries
22154 that lack all floating-point support (including, for example, the
22155 floating-point @code{printf} formats).
22156 If code compiled with @option{-mno-float} accidentally contains
22157 floating-point operations, it is likely to suffer a link-time
22158 or run-time failure.
22160 @item -msingle-float
22161 @opindex msingle-float
22162 Assume that the floating-point coprocessor only supports single-precision
22165 @item -mdouble-float
22166 @opindex mdouble-float
22167 Assume that the floating-point coprocessor supports double-precision
22168 operations. This is the default.
22171 @itemx -mno-odd-spreg
22172 @opindex modd-spreg
22173 @opindex mno-odd-spreg
22174 Enable the use of odd-numbered single-precision floating-point registers
22175 for the o32 ABI. This is the default for processors that are known to
22176 support these registers. When using the o32 FPXX ABI, @option{-mno-odd-spreg}
22180 @itemx -mabs=legacy
22182 @opindex mabs=legacy
22183 These options control the treatment of the special not-a-number (NaN)
22184 IEEE 754 floating-point data with the @code{abs.@i{fmt}} and
22185 @code{neg.@i{fmt}} machine instructions.
22187 By default or when @option{-mabs=legacy} is used the legacy
22188 treatment is selected. In this case these instructions are considered
22189 arithmetic and avoided where correct operation is required and the
22190 input operand might be a NaN. A longer sequence of instructions that
22191 manipulate the sign bit of floating-point datum manually is used
22192 instead unless the @option{-ffinite-math-only} option has also been
22195 The @option{-mabs=2008} option selects the IEEE 754-2008 treatment. In
22196 this case these instructions are considered non-arithmetic and therefore
22197 operating correctly in all cases, including in particular where the
22198 input operand is a NaN. These instructions are therefore always used
22199 for the respective operations.
22202 @itemx -mnan=legacy
22204 @opindex mnan=legacy
22205 These options control the encoding of the special not-a-number (NaN)
22206 IEEE 754 floating-point data.
22208 The @option{-mnan=legacy} option selects the legacy encoding. In this
22209 case quiet NaNs (qNaNs) are denoted by the first bit of their trailing
22210 significand field being 0, whereas signaling NaNs (sNaNs) are denoted
22211 by the first bit of their trailing significand field being 1.
22213 The @option{-mnan=2008} option selects the IEEE 754-2008 encoding. In
22214 this case qNaNs are denoted by the first bit of their trailing
22215 significand field being 1, whereas sNaNs are denoted by the first bit of
22216 their trailing significand field being 0.
22218 The default is @option{-mnan=legacy} unless GCC has been configured with
22219 @option{--with-nan=2008}.
22225 Use (do not use) @samp{ll}, @samp{sc}, and @samp{sync} instructions to
22226 implement atomic memory built-in functions. When neither option is
22227 specified, GCC uses the instructions if the target architecture
22230 @option{-mllsc} is useful if the runtime environment can emulate the
22231 instructions and @option{-mno-llsc} can be useful when compiling for
22232 nonstandard ISAs. You can make either option the default by
22233 configuring GCC with @option{--with-llsc} and @option{--without-llsc}
22234 respectively. @option{--with-llsc} is the default for some
22235 configurations; see the installation documentation for details.
22241 Use (do not use) revision 1 of the MIPS DSP ASE@.
22242 @xref{MIPS DSP Built-in Functions}. This option defines the
22243 preprocessor macro @code{__mips_dsp}. It also defines
22244 @code{__mips_dsp_rev} to 1.
22250 Use (do not use) revision 2 of the MIPS DSP ASE@.
22251 @xref{MIPS DSP Built-in Functions}. This option defines the
22252 preprocessor macros @code{__mips_dsp} and @code{__mips_dspr2}.
22253 It also defines @code{__mips_dsp_rev} to 2.
22256 @itemx -mno-smartmips
22257 @opindex msmartmips
22258 @opindex mno-smartmips
22259 Use (do not use) the MIPS SmartMIPS ASE.
22261 @item -mpaired-single
22262 @itemx -mno-paired-single
22263 @opindex mpaired-single
22264 @opindex mno-paired-single
22265 Use (do not use) paired-single floating-point instructions.
22266 @xref{MIPS Paired-Single Support}. This option requires
22267 hardware floating-point support to be enabled.
22273 Use (do not use) MIPS Digital Media Extension instructions.
22274 This option can only be used when generating 64-bit code and requires
22275 hardware floating-point support to be enabled.
22280 @opindex mno-mips3d
22281 Use (do not use) the MIPS-3D ASE@. @xref{MIPS-3D Built-in Functions}.
22282 The option @option{-mips3d} implies @option{-mpaired-single}.
22285 @itemx -mno-micromips
22286 @opindex mmicromips
22287 @opindex mno-mmicromips
22288 Generate (do not generate) microMIPS code.
22290 MicroMIPS code generation can also be controlled on a per-function basis
22291 by means of @code{micromips} and @code{nomicromips} attributes.
22292 @xref{Function Attributes}, for more information.
22298 Use (do not use) MT Multithreading instructions.
22304 Use (do not use) the MIPS MCU ASE instructions.
22310 Use (do not use) the MIPS Enhanced Virtual Addressing instructions.
22316 Use (do not use) the MIPS Virtualization (VZ) instructions.
22322 Use (do not use) the MIPS eXtended Physical Address (XPA) instructions.
22328 Use (do not use) the MIPS Cyclic Redundancy Check (CRC) instructions.
22334 Use (do not use) the MIPS Global INValidate (GINV) instructions.
22336 @item -mloongson-mmi
22337 @itemx -mno-loongson-mmi
22338 @opindex mloongson-mmi
22339 @opindex mno-loongson-mmi
22340 Use (do not use) the MIPS Loongson MultiMedia extensions Instructions (MMI).
22342 @item -mloongson-ext
22343 @itemx -mno-loongson-ext
22344 @opindex mloongson-ext
22345 @opindex mno-loongson-ext
22346 Use (do not use) the MIPS Loongson EXTensions (EXT) instructions.
22348 @item -mloongson-ext2
22349 @itemx -mno-loongson-ext2
22350 @opindex mloongson-ext2
22351 @opindex mno-loongson-ext2
22352 Use (do not use) the MIPS Loongson EXTensions r2 (EXT2) instructions.
22356 Force @code{long} types to be 64 bits wide. See @option{-mlong32} for
22357 an explanation of the default and the way that the pointer size is
22362 Force @code{long}, @code{int}, and pointer types to be 32 bits wide.
22364 The default size of @code{int}s, @code{long}s and pointers depends on
22365 the ABI@. All the supported ABIs use 32-bit @code{int}s. The n64 ABI
22366 uses 64-bit @code{long}s, as does the 64-bit EABI; the others use
22367 32-bit @code{long}s. Pointers are the same size as @code{long}s,
22368 or the same size as integer registers, whichever is smaller.
22374 Assume (do not assume) that all symbols have 32-bit values, regardless
22375 of the selected ABI@. This option is useful in combination with
22376 @option{-mabi=64} and @option{-mno-abicalls} because it allows GCC
22377 to generate shorter and faster references to symbolic addresses.
22381 Put definitions of externally-visible data in a small data section
22382 if that data is no bigger than @var{num} bytes. GCC can then generate
22383 more efficient accesses to the data; see @option{-mgpopt} for details.
22385 The default @option{-G} option depends on the configuration.
22387 @item -mlocal-sdata
22388 @itemx -mno-local-sdata
22389 @opindex mlocal-sdata
22390 @opindex mno-local-sdata
22391 Extend (do not extend) the @option{-G} behavior to local data too,
22392 such as to static variables in C@. @option{-mlocal-sdata} is the
22393 default for all configurations.
22395 If the linker complains that an application is using too much small data,
22396 you might want to try rebuilding the less performance-critical parts with
22397 @option{-mno-local-sdata}. You might also want to build large
22398 libraries with @option{-mno-local-sdata}, so that the libraries leave
22399 more room for the main program.
22401 @item -mextern-sdata
22402 @itemx -mno-extern-sdata
22403 @opindex mextern-sdata
22404 @opindex mno-extern-sdata
22405 Assume (do not assume) that externally-defined data is in
22406 a small data section if the size of that data is within the @option{-G} limit.
22407 @option{-mextern-sdata} is the default for all configurations.
22409 If you compile a module @var{Mod} with @option{-mextern-sdata} @option{-G
22410 @var{num}} @option{-mgpopt}, and @var{Mod} references a variable @var{Var}
22411 that is no bigger than @var{num} bytes, you must make sure that @var{Var}
22412 is placed in a small data section. If @var{Var} is defined by another
22413 module, you must either compile that module with a high-enough
22414 @option{-G} setting or attach a @code{section} attribute to @var{Var}'s
22415 definition. If @var{Var} is common, you must link the application
22416 with a high-enough @option{-G} setting.
22418 The easiest way of satisfying these restrictions is to compile
22419 and link every module with the same @option{-G} option. However,
22420 you may wish to build a library that supports several different
22421 small data limits. You can do this by compiling the library with
22422 the highest supported @option{-G} setting and additionally using
22423 @option{-mno-extern-sdata} to stop the library from making assumptions
22424 about externally-defined data.
22430 Use (do not use) GP-relative accesses for symbols that are known to be
22431 in a small data section; see @option{-G}, @option{-mlocal-sdata} and
22432 @option{-mextern-sdata}. @option{-mgpopt} is the default for all
22435 @option{-mno-gpopt} is useful for cases where the @code{$gp} register
22436 might not hold the value of @code{_gp}. For example, if the code is
22437 part of a library that might be used in a boot monitor, programs that
22438 call boot monitor routines pass an unknown value in @code{$gp}.
22439 (In such situations, the boot monitor itself is usually compiled
22440 with @option{-G0}.)
22442 @option{-mno-gpopt} implies @option{-mno-local-sdata} and
22443 @option{-mno-extern-sdata}.
22445 @item -membedded-data
22446 @itemx -mno-embedded-data
22447 @opindex membedded-data
22448 @opindex mno-embedded-data
22449 Allocate variables to the read-only data section first if possible, then
22450 next in the small data section if possible, otherwise in data. This gives
22451 slightly slower code than the default, but reduces the amount of RAM required
22452 when executing, and thus may be preferred for some embedded systems.
22454 @item -muninit-const-in-rodata
22455 @itemx -mno-uninit-const-in-rodata
22456 @opindex muninit-const-in-rodata
22457 @opindex mno-uninit-const-in-rodata
22458 Put uninitialized @code{const} variables in the read-only data section.
22459 This option is only meaningful in conjunction with @option{-membedded-data}.
22461 @item -mcode-readable=@var{setting}
22462 @opindex mcode-readable
22463 Specify whether GCC may generate code that reads from executable sections.
22464 There are three possible settings:
22467 @item -mcode-readable=yes
22468 Instructions may freely access executable sections. This is the
22471 @item -mcode-readable=pcrel
22472 MIPS16 PC-relative load instructions can access executable sections,
22473 but other instructions must not do so. This option is useful on 4KSc
22474 and 4KSd processors when the code TLBs have the Read Inhibit bit set.
22475 It is also useful on processors that can be configured to have a dual
22476 instruction/data SRAM interface and that, like the M4K, automatically
22477 redirect PC-relative loads to the instruction RAM.
22479 @item -mcode-readable=no
22480 Instructions must not access executable sections. This option can be
22481 useful on targets that are configured to have a dual instruction/data
22482 SRAM interface but that (unlike the M4K) do not automatically redirect
22483 PC-relative loads to the instruction RAM.
22486 @item -msplit-addresses
22487 @itemx -mno-split-addresses
22488 @opindex msplit-addresses
22489 @opindex mno-split-addresses
22490 Enable (disable) use of the @code{%hi()} and @code{%lo()} assembler
22491 relocation operators. This option has been superseded by
22492 @option{-mexplicit-relocs} but is retained for backwards compatibility.
22494 @item -mexplicit-relocs
22495 @itemx -mno-explicit-relocs
22496 @opindex mexplicit-relocs
22497 @opindex mno-explicit-relocs
22498 Use (do not use) assembler relocation operators when dealing with symbolic
22499 addresses. The alternative, selected by @option{-mno-explicit-relocs},
22500 is to use assembler macros instead.
22502 @option{-mexplicit-relocs} is the default if GCC was configured
22503 to use an assembler that supports relocation operators.
22505 @item -mcheck-zero-division
22506 @itemx -mno-check-zero-division
22507 @opindex mcheck-zero-division
22508 @opindex mno-check-zero-division
22509 Trap (do not trap) on integer division by zero.
22511 The default is @option{-mcheck-zero-division}.
22513 @item -mdivide-traps
22514 @itemx -mdivide-breaks
22515 @opindex mdivide-traps
22516 @opindex mdivide-breaks
22517 MIPS systems check for division by zero by generating either a
22518 conditional trap or a break instruction. Using traps results in
22519 smaller code, but is only supported on MIPS II and later. Also, some
22520 versions of the Linux kernel have a bug that prevents trap from
22521 generating the proper signal (@code{SIGFPE}). Use @option{-mdivide-traps} to
22522 allow conditional traps on architectures that support them and
22523 @option{-mdivide-breaks} to force the use of breaks.
22525 The default is usually @option{-mdivide-traps}, but this can be
22526 overridden at configure time using @option{--with-divide=breaks}.
22527 Divide-by-zero checks can be completely disabled using
22528 @option{-mno-check-zero-division}.
22530 @item -mload-store-pairs
22531 @itemx -mno-load-store-pairs
22532 @opindex mload-store-pairs
22533 @opindex mno-load-store-pairs
22534 Enable (disable) an optimization that pairs consecutive load or store
22535 instructions to enable load/store bonding. This option is enabled by
22536 default but only takes effect when the selected architecture is known
22537 to support bonding.
22542 @opindex mno-memcpy
22543 Force (do not force) the use of @code{memcpy} for non-trivial block
22544 moves. The default is @option{-mno-memcpy}, which allows GCC to inline
22545 most constant-sized copies.
22548 @itemx -mno-long-calls
22549 @opindex mlong-calls
22550 @opindex mno-long-calls
22551 Disable (do not disable) use of the @code{jal} instruction. Calling
22552 functions using @code{jal} is more efficient but requires the caller
22553 and callee to be in the same 256 megabyte segment.
22555 This option has no effect on abicalls code. The default is
22556 @option{-mno-long-calls}.
22562 Enable (disable) use of the @code{mad}, @code{madu} and @code{mul}
22563 instructions, as provided by the R4650 ISA@.
22569 Enable (disable) use of the @code{madd} and @code{msub} integer
22570 instructions. The default is @option{-mimadd} on architectures
22571 that support @code{madd} and @code{msub} except for the 74k
22572 architecture where it was found to generate slower code.
22575 @itemx -mno-fused-madd
22576 @opindex mfused-madd
22577 @opindex mno-fused-madd
22578 Enable (disable) use of the floating-point multiply-accumulate
22579 instructions, when they are available. The default is
22580 @option{-mfused-madd}.
22582 On the R8000 CPU when multiply-accumulate instructions are used,
22583 the intermediate product is calculated to infinite precision
22584 and is not subject to the FCSR Flush to Zero bit. This may be
22585 undesirable in some circumstances. On other processors the result
22586 is numerically identical to the equivalent computation using
22587 separate multiply, add, subtract and negate instructions.
22591 Tell the MIPS assembler to not run its preprocessor over user
22592 assembler files (with a @samp{.s} suffix) when assembling them.
22595 @itemx -mno-fix-24k
22597 @opindex mno-fix-24k
22598 Work around the 24K E48 (lost data on stores during refill) errata.
22599 The workarounds are implemented by the assembler rather than by GCC@.
22602 @itemx -mno-fix-r4000
22603 @opindex mfix-r4000
22604 @opindex mno-fix-r4000
22605 Work around certain R4000 CPU errata:
22608 A double-word or a variable shift may give an incorrect result if executed
22609 immediately after starting an integer division.
22611 A double-word or a variable shift may give an incorrect result if executed
22612 while an integer multiplication is in progress.
22614 An integer division may give an incorrect result if started in a delay slot
22615 of a taken branch or a jump.
22619 @itemx -mno-fix-r4400
22620 @opindex mfix-r4400
22621 @opindex mno-fix-r4400
22622 Work around certain R4400 CPU errata:
22625 A double-word or a variable shift may give an incorrect result if executed
22626 immediately after starting an integer division.
22630 @itemx -mno-fix-r10000
22631 @opindex mfix-r10000
22632 @opindex mno-fix-r10000
22633 Work around certain R10000 errata:
22636 @code{ll}/@code{sc} sequences may not behave atomically on revisions
22637 prior to 3.0. They may deadlock on revisions 2.6 and earlier.
22640 This option can only be used if the target architecture supports
22641 branch-likely instructions. @option{-mfix-r10000} is the default when
22642 @option{-march=r10000} is used; @option{-mno-fix-r10000} is the default
22646 @itemx -mno-fix-r5900
22647 @opindex mfix-r5900
22648 Do not attempt to schedule the preceding instruction into the delay slot
22649 of a branch instruction placed at the end of a short loop of six
22650 instructions or fewer and always schedule a @code{nop} instruction there
22651 instead. The short loop bug under certain conditions causes loops to
22652 execute only once or twice, due to a hardware bug in the R5900 chip. The
22653 workaround is implemented by the assembler rather than by GCC@.
22656 @itemx -mno-fix-rm7000
22657 @opindex mfix-rm7000
22658 Work around the RM7000 @code{dmult}/@code{dmultu} errata. The
22659 workarounds are implemented by the assembler rather than by GCC@.
22662 @itemx -mno-fix-vr4120
22663 @opindex mfix-vr4120
22664 Work around certain VR4120 errata:
22667 @code{dmultu} does not always produce the correct result.
22669 @code{div} and @code{ddiv} do not always produce the correct result if one
22670 of the operands is negative.
22672 The workarounds for the division errata rely on special functions in
22673 @file{libgcc.a}. At present, these functions are only provided by
22674 the @code{mips64vr*-elf} configurations.
22676 Other VR4120 errata require a NOP to be inserted between certain pairs of
22677 instructions. These errata are handled by the assembler, not by GCC itself.
22680 @opindex mfix-vr4130
22681 Work around the VR4130 @code{mflo}/@code{mfhi} errata. The
22682 workarounds are implemented by the assembler rather than by GCC,
22683 although GCC avoids using @code{mflo} and @code{mfhi} if the
22684 VR4130 @code{macc}, @code{macchi}, @code{dmacc} and @code{dmacchi}
22685 instructions are available instead.
22688 @itemx -mno-fix-sb1
22690 Work around certain SB-1 CPU core errata.
22691 (This flag currently works around the SB-1 revision 2
22692 ``F1'' and ``F2'' floating-point errata.)
22694 @item -mr10k-cache-barrier=@var{setting}
22695 @opindex mr10k-cache-barrier
22696 Specify whether GCC should insert cache barriers to avoid the
22697 side effects of speculation on R10K processors.
22699 In common with many processors, the R10K tries to predict the outcome
22700 of a conditional branch and speculatively executes instructions from
22701 the ``taken'' branch. It later aborts these instructions if the
22702 predicted outcome is wrong. However, on the R10K, even aborted
22703 instructions can have side effects.
22705 This problem only affects kernel stores and, depending on the system,
22706 kernel loads. As an example, a speculatively-executed store may load
22707 the target memory into cache and mark the cache line as dirty, even if
22708 the store itself is later aborted. If a DMA operation writes to the
22709 same area of memory before the ``dirty'' line is flushed, the cached
22710 data overwrites the DMA-ed data. See the R10K processor manual
22711 for a full description, including other potential problems.
22713 One workaround is to insert cache barrier instructions before every memory
22714 access that might be speculatively executed and that might have side
22715 effects even if aborted. @option{-mr10k-cache-barrier=@var{setting}}
22716 controls GCC's implementation of this workaround. It assumes that
22717 aborted accesses to any byte in the following regions does not have
22722 the memory occupied by the current function's stack frame;
22725 the memory occupied by an incoming stack argument;
22728 the memory occupied by an object with a link-time-constant address.
22731 It is the kernel's responsibility to ensure that speculative
22732 accesses to these regions are indeed safe.
22734 If the input program contains a function declaration such as:
22740 then the implementation of @code{foo} must allow @code{j foo} and
22741 @code{jal foo} to be executed speculatively. GCC honors this
22742 restriction for functions it compiles itself. It expects non-GCC
22743 functions (such as hand-written assembly code) to do the same.
22745 The option has three forms:
22748 @item -mr10k-cache-barrier=load-store
22749 Insert a cache barrier before a load or store that might be
22750 speculatively executed and that might have side effects even
22753 @item -mr10k-cache-barrier=store
22754 Insert a cache barrier before a store that might be speculatively
22755 executed and that might have side effects even if aborted.
22757 @item -mr10k-cache-barrier=none
22758 Disable the insertion of cache barriers. This is the default setting.
22761 @item -mflush-func=@var{func}
22762 @itemx -mno-flush-func
22763 @opindex mflush-func
22764 Specifies the function to call to flush the I and D caches, or to not
22765 call any such function. If called, the function must take the same
22766 arguments as the common @code{_flush_func}, that is, the address of the
22767 memory range for which the cache is being flushed, the size of the
22768 memory range, and the number 3 (to flush both caches). The default
22769 depends on the target GCC was configured for, but commonly is either
22770 @code{_flush_func} or @code{__cpu_flush}.
22772 @item mbranch-cost=@var{num}
22773 @opindex mbranch-cost
22774 Set the cost of branches to roughly @var{num} ``simple'' instructions.
22775 This cost is only a heuristic and is not guaranteed to produce
22776 consistent results across releases. A zero cost redundantly selects
22777 the default, which is based on the @option{-mtune} setting.
22779 @item -mbranch-likely
22780 @itemx -mno-branch-likely
22781 @opindex mbranch-likely
22782 @opindex mno-branch-likely
22783 Enable or disable use of Branch Likely instructions, regardless of the
22784 default for the selected architecture. By default, Branch Likely
22785 instructions may be generated if they are supported by the selected
22786 architecture. An exception is for the MIPS32 and MIPS64 architectures
22787 and processors that implement those architectures; for those, Branch
22788 Likely instructions are not be generated by default because the MIPS32
22789 and MIPS64 architectures specifically deprecate their use.
22791 @item -mcompact-branches=never
22792 @itemx -mcompact-branches=optimal
22793 @itemx -mcompact-branches=always
22794 @opindex mcompact-branches=never
22795 @opindex mcompact-branches=optimal
22796 @opindex mcompact-branches=always
22797 These options control which form of branches will be generated. The
22798 default is @option{-mcompact-branches=optimal}.
22800 The @option{-mcompact-branches=never} option ensures that compact branch
22801 instructions will never be generated.
22803 The @option{-mcompact-branches=always} option ensures that a compact
22804 branch instruction will be generated if available. If a compact branch
22805 instruction is not available, a delay slot form of the branch will be
22808 This option is supported from MIPS Release 6 onwards.
22810 The @option{-mcompact-branches=optimal} option will cause a delay slot
22811 branch to be used if one is available in the current ISA and the delay
22812 slot is successfully filled. If the delay slot is not filled, a compact
22813 branch will be chosen if one is available.
22815 @item -mfp-exceptions
22816 @itemx -mno-fp-exceptions
22817 @opindex mfp-exceptions
22818 Specifies whether FP exceptions are enabled. This affects how
22819 FP instructions are scheduled for some processors.
22820 The default is that FP exceptions are
22823 For instance, on the SB-1, if FP exceptions are disabled, and we are emitting
22824 64-bit code, then we can use both FP pipes. Otherwise, we can only use one
22827 @item -mvr4130-align
22828 @itemx -mno-vr4130-align
22829 @opindex mvr4130-align
22830 The VR4130 pipeline is two-way superscalar, but can only issue two
22831 instructions together if the first one is 8-byte aligned. When this
22832 option is enabled, GCC aligns pairs of instructions that it
22833 thinks should execute in parallel.
22835 This option only has an effect when optimizing for the VR4130.
22836 It normally makes code faster, but at the expense of making it bigger.
22837 It is enabled by default at optimization level @option{-O3}.
22842 Enable (disable) generation of @code{synci} instructions on
22843 architectures that support it. The @code{synci} instructions (if
22844 enabled) are generated when @code{__builtin___clear_cache} is
22847 This option defaults to @option{-mno-synci}, but the default can be
22848 overridden by configuring GCC with @option{--with-synci}.
22850 When compiling code for single processor systems, it is generally safe
22851 to use @code{synci}. However, on many multi-core (SMP) systems, it
22852 does not invalidate the instruction caches on all cores and may lead
22853 to undefined behavior.
22855 @item -mrelax-pic-calls
22856 @itemx -mno-relax-pic-calls
22857 @opindex mrelax-pic-calls
22858 Try to turn PIC calls that are normally dispatched via register
22859 @code{$25} into direct calls. This is only possible if the linker can
22860 resolve the destination at link time and if the destination is within
22861 range for a direct call.
22863 @option{-mrelax-pic-calls} is the default if GCC was configured to use
22864 an assembler and a linker that support the @code{.reloc} assembly
22865 directive and @option{-mexplicit-relocs} is in effect. With
22866 @option{-mno-explicit-relocs}, this optimization can be performed by the
22867 assembler and the linker alone without help from the compiler.
22869 @item -mmcount-ra-address
22870 @itemx -mno-mcount-ra-address
22871 @opindex mmcount-ra-address
22872 @opindex mno-mcount-ra-address
22873 Emit (do not emit) code that allows @code{_mcount} to modify the
22874 calling function's return address. When enabled, this option extends
22875 the usual @code{_mcount} interface with a new @var{ra-address}
22876 parameter, which has type @code{intptr_t *} and is passed in register
22877 @code{$12}. @code{_mcount} can then modify the return address by
22878 doing both of the following:
22881 Returning the new address in register @code{$31}.
22883 Storing the new address in @code{*@var{ra-address}},
22884 if @var{ra-address} is nonnull.
22887 The default is @option{-mno-mcount-ra-address}.
22889 @item -mframe-header-opt
22890 @itemx -mno-frame-header-opt
22891 @opindex mframe-header-opt
22892 Enable (disable) frame header optimization in the o32 ABI. When using the
22893 o32 ABI, calling functions will allocate 16 bytes on the stack for the called
22894 function to write out register arguments. When enabled, this optimization
22895 will suppress the allocation of the frame header if it can be determined that
22898 This optimization is off by default at all optimization levels.
22901 @itemx -mno-lxc1-sxc1
22902 @opindex mlxc1-sxc1
22903 When applicable, enable (disable) the generation of @code{lwxc1},
22904 @code{swxc1}, @code{ldxc1}, @code{sdxc1} instructions. Enabled by default.
22909 When applicable, enable (disable) the generation of 4-operand @code{madd.s},
22910 @code{madd.d} and related instructions. Enabled by default.
22915 @subsection MMIX Options
22916 @cindex MMIX Options
22918 These options are defined for the MMIX:
22922 @itemx -mno-libfuncs
22924 @opindex mno-libfuncs
22925 Specify that intrinsic library functions are being compiled, passing all
22926 values in registers, no matter the size.
22929 @itemx -mno-epsilon
22931 @opindex mno-epsilon
22932 Generate floating-point comparison instructions that compare with respect
22933 to the @code{rE} epsilon register.
22935 @item -mabi=mmixware
22937 @opindex mabi=mmixware
22939 Generate code that passes function parameters and return values that (in
22940 the called function) are seen as registers @code{$0} and up, as opposed to
22941 the GNU ABI which uses global registers @code{$231} and up.
22943 @item -mzero-extend
22944 @itemx -mno-zero-extend
22945 @opindex mzero-extend
22946 @opindex mno-zero-extend
22947 When reading data from memory in sizes shorter than 64 bits, use (do not
22948 use) zero-extending load instructions by default, rather than
22949 sign-extending ones.
22952 @itemx -mno-knuthdiv
22954 @opindex mno-knuthdiv
22955 Make the result of a division yielding a remainder have the same sign as
22956 the divisor. With the default, @option{-mno-knuthdiv}, the sign of the
22957 remainder follows the sign of the dividend. Both methods are
22958 arithmetically valid, the latter being almost exclusively used.
22960 @item -mtoplevel-symbols
22961 @itemx -mno-toplevel-symbols
22962 @opindex mtoplevel-symbols
22963 @opindex mno-toplevel-symbols
22964 Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly
22965 code can be used with the @code{PREFIX} assembly directive.
22969 Generate an executable in the ELF format, rather than the default
22970 @samp{mmo} format used by the @command{mmix} simulator.
22972 @item -mbranch-predict
22973 @itemx -mno-branch-predict
22974 @opindex mbranch-predict
22975 @opindex mno-branch-predict
22976 Use (do not use) the probable-branch instructions, when static branch
22977 prediction indicates a probable branch.
22979 @item -mbase-addresses
22980 @itemx -mno-base-addresses
22981 @opindex mbase-addresses
22982 @opindex mno-base-addresses
22983 Generate (do not generate) code that uses @emph{base addresses}. Using a
22984 base address automatically generates a request (handled by the assembler
22985 and the linker) for a constant to be set up in a global register. The
22986 register is used for one or more base address requests within the range 0
22987 to 255 from the value held in the register. The generally leads to short
22988 and fast code, but the number of different data items that can be
22989 addressed is limited. This means that a program that uses lots of static
22990 data may require @option{-mno-base-addresses}.
22992 @item -msingle-exit
22993 @itemx -mno-single-exit
22994 @opindex msingle-exit
22995 @opindex mno-single-exit
22996 Force (do not force) generated code to have a single exit point in each
23000 @node MN10300 Options
23001 @subsection MN10300 Options
23002 @cindex MN10300 options
23004 These @option{-m} options are defined for Matsushita MN10300 architectures:
23009 Generate code to avoid bugs in the multiply instructions for the MN10300
23010 processors. This is the default.
23012 @item -mno-mult-bug
23013 @opindex mno-mult-bug
23014 Do not generate code to avoid bugs in the multiply instructions for the
23015 MN10300 processors.
23019 Generate code using features specific to the AM33 processor.
23023 Do not generate code using features specific to the AM33 processor. This
23028 Generate code using features specific to the AM33/2.0 processor.
23032 Generate code using features specific to the AM34 processor.
23034 @item -mtune=@var{cpu-type}
23036 Use the timing characteristics of the indicated CPU type when
23037 scheduling instructions. This does not change the targeted processor
23038 type. The CPU type must be one of @samp{mn10300}, @samp{am33},
23039 @samp{am33-2} or @samp{am34}.
23041 @item -mreturn-pointer-on-d0
23042 @opindex mreturn-pointer-on-d0
23043 When generating a function that returns a pointer, return the pointer
23044 in both @code{a0} and @code{d0}. Otherwise, the pointer is returned
23045 only in @code{a0}, and attempts to call such functions without a prototype
23046 result in errors. Note that this option is on by default; use
23047 @option{-mno-return-pointer-on-d0} to disable it.
23051 Do not link in the C run-time initialization object file.
23055 Indicate to the linker that it should perform a relaxation optimization pass
23056 to shorten branches, calls and absolute memory addresses. This option only
23057 has an effect when used on the command line for the final link step.
23059 This option makes symbolic debugging impossible.
23063 Allow the compiler to generate @emph{Long Instruction Word}
23064 instructions if the target is the @samp{AM33} or later. This is the
23065 default. This option defines the preprocessor macro @code{__LIW__}.
23069 Do not allow the compiler to generate @emph{Long Instruction Word}
23070 instructions. This option defines the preprocessor macro
23075 Allow the compiler to generate the @emph{SETLB} and @emph{Lcc}
23076 instructions if the target is the @samp{AM33} or later. This is the
23077 default. This option defines the preprocessor macro @code{__SETLB__}.
23081 Do not allow the compiler to generate @emph{SETLB} or @emph{Lcc}
23082 instructions. This option defines the preprocessor macro
23083 @code{__NO_SETLB__}.
23087 @node Moxie Options
23088 @subsection Moxie Options
23089 @cindex Moxie Options
23095 Generate big-endian code. This is the default for @samp{moxie-*-*}
23100 Generate little-endian code.
23104 Generate mul.x and umul.x instructions. This is the default for
23105 @samp{moxiebox-*-*} configurations.
23109 Do not link in the C run-time initialization object file.
23113 @node MSP430 Options
23114 @subsection MSP430 Options
23115 @cindex MSP430 Options
23117 These options are defined for the MSP430:
23123 Force assembly output to always use hex constants. Normally such
23124 constants are signed decimals, but this option is available for
23125 testsuite and/or aesthetic purposes.
23129 Select the MCU to target. This is used to create a C preprocessor
23130 symbol based upon the MCU name, converted to upper case and pre- and
23131 post-fixed with @samp{__}. This in turn is used by the
23132 @file{msp430.h} header file to select an MCU-specific supplementary
23135 The option also sets the ISA to use. If the MCU name is one that is
23136 known to only support the 430 ISA then that is selected, otherwise the
23137 430X ISA is selected. A generic MCU name of @samp{msp430} can also be
23138 used to select the 430 ISA. Similarly the generic @samp{msp430x} MCU
23139 name selects the 430X ISA.
23141 In addition an MCU-specific linker script is added to the linker
23142 command line. The script's name is the name of the MCU with
23143 @file{.ld} appended. Thus specifying @option{-mmcu=xxx} on the @command{gcc}
23144 command line defines the C preprocessor symbol @code{__XXX__} and
23145 cause the linker to search for a script called @file{xxx.ld}.
23147 The ISA and hardware multiply supported for the different MCUs is hard-coded
23148 into GCC. However, an external @samp{devices.csv} file can be used to
23149 extend device support beyond those that have been hard-coded.
23151 GCC searches for the @samp{devices.csv} file on the paths specified
23152 with the @code{-I} and @code{-L} options.
23155 @itemx -mno-warn-mcu
23157 @opindex mno-warn-mcu
23158 This option enables or disables warnings about conflicts between the
23159 MCU name specified by the @option{-mmcu} option and the ISA set by the
23160 @option{-mcpu} option and/or the hardware multiply support set by the
23161 @option{-mhwmult} option. It also toggles warnings about unrecognized
23162 MCU names. This option is on by default.
23166 Specifies the ISA to use. Accepted values are @samp{msp430},
23167 @samp{msp430x} and @samp{msp430xv2}. This option is deprecated. The
23168 @option{-mmcu=} option should be used to select the ISA.
23172 Link to the simulator runtime libraries and linker script. Overrides
23173 any scripts that would be selected by the @option{-mmcu=} option.
23177 Use large-model addressing (20-bit pointers, 32-bit @code{size_t}).
23181 Use small-model addressing (16-bit pointers, 16-bit @code{size_t}).
23185 This option is passed to the assembler and linker, and allows the
23186 linker to perform certain optimizations that cannot be done until
23191 Describes the type of hardware multiply supported by the target.
23192 Accepted values are @samp{none} for no hardware multiply, @samp{16bit}
23193 for the original 16-bit-only multiply supported by early MCUs.
23194 @samp{32bit} for the 16/32-bit multiply supported by later MCUs and
23195 @samp{f5series} for the 16/32-bit multiply supported by F5-series MCUs.
23196 A value of @samp{auto} can also be given. This tells GCC to deduce
23197 the hardware multiply support based upon the MCU name provided by the
23198 @option{-mmcu} option. If no @option{-mmcu} option is specified or if
23199 the MCU name is not recognized then no hardware multiply support is
23200 assumed. @code{auto} is the default setting.
23202 Hardware multiplies are normally performed by calling a library
23203 routine. This saves space in the generated code. When compiling at
23204 @option{-O3} or higher however the hardware multiplier is invoked
23205 inline. This makes for bigger, but faster code.
23207 The hardware multiply routines disable interrupts whilst running and
23208 restore the previous interrupt state when they finish. This makes
23209 them safe to use inside interrupt handlers as well as in normal code.
23213 Enable the use of a minimum runtime environment - no static
23214 initializers or constructors. This is intended for memory-constrained
23215 devices. The compiler includes special symbols in some objects
23216 that tell the linker and runtime which code fragments are required.
23218 @item -mcode-region=
23219 @itemx -mdata-region=
23220 @opindex mcode-region
23221 @opindex mdata-region
23222 These options tell the compiler where to place functions and data that
23223 do not have one of the @code{lower}, @code{upper}, @code{either} or
23224 @code{section} attributes. Possible values are @code{lower},
23225 @code{upper}, @code{either} or @code{any}. The first three behave
23226 like the corresponding attribute. The fourth possible value -
23227 @code{any} - is the default. It leaves placement entirely up to the
23228 linker script and how it assigns the standard sections
23229 (@code{.text}, @code{.data}, etc) to the memory regions.
23231 @item -msilicon-errata=
23232 @opindex msilicon-errata
23233 This option passes on a request to assembler to enable the fixes for
23234 the named silicon errata.
23236 @item -msilicon-errata-warn=
23237 @opindex msilicon-errata-warn
23238 This option passes on a request to the assembler to enable warning
23239 messages when a silicon errata might need to be applied.
23241 @item -mwarn-devices-csv
23242 @itemx -mno-warn-devices-csv
23243 @opindex mwarn-devices-csv
23244 @opindex mno-warn-devices-csv
23245 Warn if @samp{devices.csv} is not found or there are problem parsing it
23250 @node NDS32 Options
23251 @subsection NDS32 Options
23252 @cindex NDS32 Options
23254 These options are defined for NDS32 implementations:
23259 @opindex mbig-endian
23260 Generate code in big-endian mode.
23262 @item -mlittle-endian
23263 @opindex mlittle-endian
23264 Generate code in little-endian mode.
23266 @item -mreduced-regs
23267 @opindex mreduced-regs
23268 Use reduced-set registers for register allocation.
23271 @opindex mfull-regs
23272 Use full-set registers for register allocation.
23276 Generate conditional move instructions.
23280 Do not generate conditional move instructions.
23284 Generate performance extension instructions.
23286 @item -mno-ext-perf
23287 @opindex mno-ext-perf
23288 Do not generate performance extension instructions.
23291 @opindex mext-perf2
23292 Generate performance extension 2 instructions.
23294 @item -mno-ext-perf2
23295 @opindex mno-ext-perf2
23296 Do not generate performance extension 2 instructions.
23299 @opindex mext-string
23300 Generate string extension instructions.
23302 @item -mno-ext-string
23303 @opindex mno-ext-string
23304 Do not generate string extension instructions.
23308 Generate v3 push25/pop25 instructions.
23311 @opindex mno-v3push
23312 Do not generate v3 push25/pop25 instructions.
23316 Generate 16-bit instructions.
23319 @opindex mno-16-bit
23320 Do not generate 16-bit instructions.
23322 @item -misr-vector-size=@var{num}
23323 @opindex misr-vector-size
23324 Specify the size of each interrupt vector, which must be 4 or 16.
23326 @item -mcache-block-size=@var{num}
23327 @opindex mcache-block-size
23328 Specify the size of each cache block,
23329 which must be a power of 2 between 4 and 512.
23331 @item -march=@var{arch}
23333 Specify the name of the target architecture.
23335 @item -mcmodel=@var{code-model}
23337 Set the code model to one of
23340 All the data and read-only data segments must be within 512KB addressing space.
23341 The text segment must be within 16MB addressing space.
23342 @item @samp{medium}
23343 The data segment must be within 512KB while the read-only data segment can be
23344 within 4GB addressing space. The text segment should be still within 16MB
23347 All the text and data segments can be within 4GB addressing space.
23351 @opindex mctor-dtor
23352 Enable constructor/destructor feature.
23356 Guide linker to relax instructions.
23360 @node Nios II Options
23361 @subsection Nios II Options
23362 @cindex Nios II options
23363 @cindex Altera Nios II options
23365 These are the options defined for the Altera Nios II processor.
23371 @cindex smaller data references
23372 Put global and static objects less than or equal to @var{num} bytes
23373 into the small data or BSS sections instead of the normal data or BSS
23374 sections. The default value of @var{num} is 8.
23376 @item -mgpopt=@var{option}
23381 Generate (do not generate) GP-relative accesses. The following
23382 @var{option} names are recognized:
23387 Do not generate GP-relative accesses.
23390 Generate GP-relative accesses for small data objects that are not
23391 external, weak, or uninitialized common symbols.
23392 Also use GP-relative addressing for objects that
23393 have been explicitly placed in a small data section via a @code{section}
23397 As for @samp{local}, but also generate GP-relative accesses for
23398 small data objects that are external, weak, or common. If you use this option,
23399 you must ensure that all parts of your program (including libraries) are
23400 compiled with the same @option{-G} setting.
23403 Generate GP-relative accesses for all data objects in the program. If you
23404 use this option, the entire data and BSS segments
23405 of your program must fit in 64K of memory and you must use an appropriate
23406 linker script to allocate them within the addressable range of the
23410 Generate GP-relative addresses for function pointers as well as data
23411 pointers. If you use this option, the entire text, data, and BSS segments
23412 of your program must fit in 64K of memory and you must use an appropriate
23413 linker script to allocate them within the addressable range of the
23418 @option{-mgpopt} is equivalent to @option{-mgpopt=local}, and
23419 @option{-mno-gpopt} is equivalent to @option{-mgpopt=none}.
23421 The default is @option{-mgpopt} except when @option{-fpic} or
23422 @option{-fPIC} is specified to generate position-independent code.
23423 Note that the Nios II ABI does not permit GP-relative accesses from
23426 You may need to specify @option{-mno-gpopt} explicitly when building
23427 programs that include large amounts of small data, including large
23428 GOT data sections. In this case, the 16-bit offset for GP-relative
23429 addressing may not be large enough to allow access to the entire
23430 small data section.
23432 @item -mgprel-sec=@var{regexp}
23433 @opindex mgprel-sec
23434 This option specifies additional section names that can be accessed via
23435 GP-relative addressing. It is most useful in conjunction with
23436 @code{section} attributes on variable declarations
23437 (@pxref{Common Variable Attributes}) and a custom linker script.
23438 The @var{regexp} is a POSIX Extended Regular Expression.
23440 This option does not affect the behavior of the @option{-G} option, and
23441 the specified sections are in addition to the standard @code{.sdata}
23442 and @code{.sbss} small-data sections that are recognized by @option{-mgpopt}.
23444 @item -mr0rel-sec=@var{regexp}
23445 @opindex mr0rel-sec
23446 This option specifies names of sections that can be accessed via a
23447 16-bit offset from @code{r0}; that is, in the low 32K or high 32K
23448 of the 32-bit address space. It is most useful in conjunction with
23449 @code{section} attributes on variable declarations
23450 (@pxref{Common Variable Attributes}) and a custom linker script.
23451 The @var{regexp} is a POSIX Extended Regular Expression.
23453 In contrast to the use of GP-relative addressing for small data,
23454 zero-based addressing is never generated by default and there are no
23455 conventional section names used in standard linker scripts for sections
23456 in the low or high areas of memory.
23462 Generate little-endian (default) or big-endian (experimental) code,
23465 @item -march=@var{arch}
23467 This specifies the name of the target Nios II architecture. GCC uses this
23468 name to determine what kind of instructions it can emit when generating
23469 assembly code. Permissible names are: @samp{r1}, @samp{r2}.
23471 The preprocessor macro @code{__nios2_arch__} is available to programs,
23472 with value 1 or 2, indicating the targeted ISA level.
23474 @item -mbypass-cache
23475 @itemx -mno-bypass-cache
23476 @opindex mno-bypass-cache
23477 @opindex mbypass-cache
23478 Force all load and store instructions to always bypass cache by
23479 using I/O variants of the instructions. The default is not to
23482 @item -mno-cache-volatile
23483 @itemx -mcache-volatile
23484 @opindex mcache-volatile
23485 @opindex mno-cache-volatile
23486 Volatile memory access bypass the cache using the I/O variants of
23487 the load and store instructions. The default is not to bypass the cache.
23489 @item -mno-fast-sw-div
23490 @itemx -mfast-sw-div
23491 @opindex mno-fast-sw-div
23492 @opindex mfast-sw-div
23493 Do not use table-based fast divide for small numbers. The default
23494 is to use the fast divide at @option{-O3} and above.
23498 @itemx -mno-hw-mulx
23502 @opindex mno-hw-mul
23504 @opindex mno-hw-mulx
23506 @opindex mno-hw-div
23508 Enable or disable emitting @code{mul}, @code{mulx} and @code{div} family of
23509 instructions by the compiler. The default is to emit @code{mul}
23510 and not emit @code{div} and @code{mulx}.
23516 Enable or disable generation of Nios II R2 BMX (bit manipulation) and
23517 CDX (code density) instructions. Enabling these instructions also
23518 requires @option{-march=r2}. Since these instructions are optional
23519 extensions to the R2 architecture, the default is not to emit them.
23521 @item -mcustom-@var{insn}=@var{N}
23522 @itemx -mno-custom-@var{insn}
23523 @opindex mcustom-@var{insn}
23524 @opindex mno-custom-@var{insn}
23525 Each @option{-mcustom-@var{insn}=@var{N}} option enables use of a
23526 custom instruction with encoding @var{N} when generating code that uses
23527 @var{insn}. For example, @option{-mcustom-fadds=253} generates custom
23528 instruction 253 for single-precision floating-point add operations instead
23529 of the default behavior of using a library call.
23531 The following values of @var{insn} are supported. Except as otherwise
23532 noted, floating-point operations are expected to be implemented with
23533 normal IEEE 754 semantics and correspond directly to the C operators or the
23534 equivalent GCC built-in functions (@pxref{Other Builtins}).
23536 Single-precision floating point:
23539 @item @samp{fadds}, @samp{fsubs}, @samp{fdivs}, @samp{fmuls}
23540 Binary arithmetic operations.
23546 Unary absolute value.
23548 @item @samp{fcmpeqs}, @samp{fcmpges}, @samp{fcmpgts}, @samp{fcmples}, @samp{fcmplts}, @samp{fcmpnes}
23549 Comparison operations.
23551 @item @samp{fmins}, @samp{fmaxs}
23552 Floating-point minimum and maximum. These instructions are only
23553 generated if @option{-ffinite-math-only} is specified.
23555 @item @samp{fsqrts}
23556 Unary square root operation.
23558 @item @samp{fcoss}, @samp{fsins}, @samp{ftans}, @samp{fatans}, @samp{fexps}, @samp{flogs}
23559 Floating-point trigonometric and exponential functions. These instructions
23560 are only generated if @option{-funsafe-math-optimizations} is also specified.
23564 Double-precision floating point:
23567 @item @samp{faddd}, @samp{fsubd}, @samp{fdivd}, @samp{fmuld}
23568 Binary arithmetic operations.
23574 Unary absolute value.
23576 @item @samp{fcmpeqd}, @samp{fcmpged}, @samp{fcmpgtd}, @samp{fcmpled}, @samp{fcmpltd}, @samp{fcmpned}
23577 Comparison operations.
23579 @item @samp{fmind}, @samp{fmaxd}
23580 Double-precision minimum and maximum. These instructions are only
23581 generated if @option{-ffinite-math-only} is specified.
23583 @item @samp{fsqrtd}
23584 Unary square root operation.
23586 @item @samp{fcosd}, @samp{fsind}, @samp{ftand}, @samp{fatand}, @samp{fexpd}, @samp{flogd}
23587 Double-precision trigonometric and exponential functions. These instructions
23588 are only generated if @option{-funsafe-math-optimizations} is also specified.
23594 @item @samp{fextsd}
23595 Conversion from single precision to double precision.
23597 @item @samp{ftruncds}
23598 Conversion from double precision to single precision.
23600 @item @samp{fixsi}, @samp{fixsu}, @samp{fixdi}, @samp{fixdu}
23601 Conversion from floating point to signed or unsigned integer types, with
23602 truncation towards zero.
23605 Conversion from single-precision floating point to signed integer,
23606 rounding to the nearest integer and ties away from zero.
23607 This corresponds to the @code{__builtin_lroundf} function when
23608 @option{-fno-math-errno} is used.
23610 @item @samp{floatis}, @samp{floatus}, @samp{floatid}, @samp{floatud}
23611 Conversion from signed or unsigned integer types to floating-point types.
23615 In addition, all of the following transfer instructions for internal
23616 registers X and Y must be provided to use any of the double-precision
23617 floating-point instructions. Custom instructions taking two
23618 double-precision source operands expect the first operand in the
23619 64-bit register X. The other operand (or only operand of a unary
23620 operation) is given to the custom arithmetic instruction with the
23621 least significant half in source register @var{src1} and the most
23622 significant half in @var{src2}. A custom instruction that returns a
23623 double-precision result returns the most significant 32 bits in the
23624 destination register and the other half in 32-bit register Y.
23625 GCC automatically generates the necessary code sequences to write
23626 register X and/or read register Y when double-precision floating-point
23627 instructions are used.
23632 Write @var{src1} into the least significant half of X and @var{src2} into
23633 the most significant half of X.
23636 Write @var{src1} into Y.
23638 @item @samp{frdxhi}, @samp{frdxlo}
23639 Read the most or least (respectively) significant half of X and store it in
23643 Read the value of Y and store it into @var{dest}.
23646 Note that you can gain more local control over generation of Nios II custom
23647 instructions by using the @code{target("custom-@var{insn}=@var{N}")}
23648 and @code{target("no-custom-@var{insn}")} function attributes
23649 (@pxref{Function Attributes})
23650 or pragmas (@pxref{Function Specific Option Pragmas}).
23652 @item -mcustom-fpu-cfg=@var{name}
23653 @opindex mcustom-fpu-cfg
23655 This option enables a predefined, named set of custom instruction encodings
23656 (see @option{-mcustom-@var{insn}} above).
23657 Currently, the following sets are defined:
23659 @option{-mcustom-fpu-cfg=60-1} is equivalent to:
23660 @gccoptlist{-mcustom-fmuls=252 @gol
23661 -mcustom-fadds=253 @gol
23662 -mcustom-fsubs=254 @gol
23663 -fsingle-precision-constant}
23665 @option{-mcustom-fpu-cfg=60-2} is equivalent to:
23666 @gccoptlist{-mcustom-fmuls=252 @gol
23667 -mcustom-fadds=253 @gol
23668 -mcustom-fsubs=254 @gol
23669 -mcustom-fdivs=255 @gol
23670 -fsingle-precision-constant}
23672 @option{-mcustom-fpu-cfg=72-3} is equivalent to:
23673 @gccoptlist{-mcustom-floatus=243 @gol
23674 -mcustom-fixsi=244 @gol
23675 -mcustom-floatis=245 @gol
23676 -mcustom-fcmpgts=246 @gol
23677 -mcustom-fcmples=249 @gol
23678 -mcustom-fcmpeqs=250 @gol
23679 -mcustom-fcmpnes=251 @gol
23680 -mcustom-fmuls=252 @gol
23681 -mcustom-fadds=253 @gol
23682 -mcustom-fsubs=254 @gol
23683 -mcustom-fdivs=255 @gol
23684 -fsingle-precision-constant}
23686 Custom instruction assignments given by individual
23687 @option{-mcustom-@var{insn}=} options override those given by
23688 @option{-mcustom-fpu-cfg=}, regardless of the
23689 order of the options on the command line.
23691 Note that you can gain more local control over selection of a FPU
23692 configuration by using the @code{target("custom-fpu-cfg=@var{name}")}
23693 function attribute (@pxref{Function Attributes})
23694 or pragma (@pxref{Function Specific Option Pragmas}).
23698 These additional @samp{-m} options are available for the Altera Nios II
23699 ELF (bare-metal) target:
23705 Link with HAL BSP. This suppresses linking with the GCC-provided C runtime
23706 startup and termination code, and is typically used in conjunction with
23707 @option{-msys-crt0=} to specify the location of the alternate startup code
23708 provided by the HAL BSP.
23712 Link with a limited version of the C library, @option{-lsmallc}, rather than
23715 @item -msys-crt0=@var{startfile}
23717 @var{startfile} is the file name of the startfile (crt0) to use
23718 when linking. This option is only useful in conjunction with @option{-mhal}.
23720 @item -msys-lib=@var{systemlib}
23722 @var{systemlib} is the library name of the library that provides
23723 low-level system calls required by the C library,
23724 e.g.@: @code{read} and @code{write}.
23725 This option is typically used to link with a library provided by a HAL BSP.
23729 @node Nvidia PTX Options
23730 @subsection Nvidia PTX Options
23731 @cindex Nvidia PTX options
23732 @cindex nvptx options
23734 These options are defined for Nvidia PTX:
23742 Generate code for 32-bit or 64-bit ABI.
23744 @item -misa=@var{ISA-string}
23746 Generate code for given the specified PTX ISA (e.g.@: @samp{sm_35}). ISA
23747 strings must be lower-case. Valid ISA strings include @samp{sm_30} and
23748 @samp{sm_35}. The default ISA is sm_30.
23751 @opindex mmainkernel
23752 Link in code for a __main kernel. This is for stand-alone instead of
23753 offloading execution.
23757 Apply partitioned execution optimizations. This is the default when any
23758 level of optimization is selected.
23761 @opindex msoft-stack
23762 Generate code that does not use @code{.local} memory
23763 directly for stack storage. Instead, a per-warp stack pointer is
23764 maintained explicitly. This enables variable-length stack allocation (with
23765 variable-length arrays or @code{alloca}), and when global memory is used for
23766 underlying storage, makes it possible to access automatic variables from other
23767 threads, or with atomic instructions. This code generation variant is used
23768 for OpenMP offloading, but the option is exposed on its own for the purpose
23769 of testing the compiler; to generate code suitable for linking into programs
23770 using OpenMP offloading, use option @option{-mgomp}.
23772 @item -muniform-simt
23773 @opindex muniform-simt
23774 Switch to code generation variant that allows to execute all threads in each
23775 warp, while maintaining memory state and side effects as if only one thread
23776 in each warp was active outside of OpenMP SIMD regions. All atomic operations
23777 and calls to runtime (malloc, free, vprintf) are conditionally executed (iff
23778 current lane index equals the master lane index), and the register being
23779 assigned is copied via a shuffle instruction from the master lane. Outside of
23780 SIMD regions lane 0 is the master; inside, each thread sees itself as the
23781 master. Shared memory array @code{int __nvptx_uni[]} stores all-zeros or
23782 all-ones bitmasks for each warp, indicating current mode (0 outside of SIMD
23783 regions). Each thread can bitwise-and the bitmask at position @code{tid.y}
23784 with current lane index to compute the master lane index.
23788 Generate code for use in OpenMP offloading: enables @option{-msoft-stack} and
23789 @option{-muniform-simt} options, and selects corresponding multilib variant.
23793 @node OpenRISC Options
23794 @subsection OpenRISC Options
23795 @cindex OpenRISC Options
23797 These options are defined for OpenRISC:
23801 @item -mboard=@var{name}
23803 Configure a board specific runtime. This will be passed to the linker for
23804 newlib board library linking. The default is @code{or1ksim}.
23808 This option is ignored; it is for compatibility purposes only. This used to
23809 select linker and preprocessor options for use with newlib.
23815 Select software or hardware divide (@code{l.div}, @code{l.divu}) instructions.
23816 This default is hardware divide.
23822 Select software or hardware multiply (@code{l.mul}, @code{l.muli}) instructions.
23823 This default is hardware multiply.
23826 @itemx -mhard-float
23827 @opindex msoft-float
23828 @opindex mhard-float
23829 Select software or hardware for floating point operations.
23830 The default is software.
23832 @item -mdouble-float
23833 @opindex mdouble-float
23834 When @option{-mhard-float} is selected, enables generation of double-precision
23835 floating point instructions. By default functions from @file{libgcc} are used
23836 to perform double-precision floating point operations.
23838 @item -munordered-float
23839 @opindex munordered-float
23840 When @option{-mhard-float} is selected, enables generation of unordered
23841 floating point compare and set flag (@code{lf.sfun*}) instructions. By default
23842 functions from @file{libgcc} are used to perform unordered floating point
23843 compare and set flag operations.
23847 Enable generation of conditional move (@code{l.cmov}) instructions. By
23848 default the equivalent will be generated using using set and branch.
23852 Enable generation of rotate right (@code{l.ror}) instructions. By default
23853 functions from @file{libgcc} are used to perform rotate right operations.
23857 Enable generation of rotate right with immediate (@code{l.rori}) instructions.
23858 By default functions from @file{libgcc} are used to perform rotate right with
23859 immediate operations.
23863 Enable generation of sign extension (@code{l.ext*}) instructions. By default
23864 memory loads are used to perform sign extension.
23868 Enable generation of compare and set flag with immediate (@code{l.sf*i})
23869 instructions. By default extra instructions will be generated to store the
23870 immediate to a register first.
23874 Enable generation of shift with immediate (@code{l.srai}, @code{l.srli},
23875 @code{l.slli}) instructions. By default extra instructions will be generated
23876 to store the immediate to a register first.
23881 @node PDP-11 Options
23882 @subsection PDP-11 Options
23883 @cindex PDP-11 Options
23885 These options are defined for the PDP-11:
23890 Use hardware FPP floating point. This is the default. (FIS floating
23891 point on the PDP-11/40 is not supported.) Implies -m45.
23894 @opindex msoft-float
23895 Do not use hardware floating point.
23899 Return floating-point results in ac0 (fr0 in Unix assembler syntax).
23903 Return floating-point results in memory. This is the default.
23907 Generate code for a PDP-11/40. Implies -msoft-float -mno-split.
23911 Generate code for a PDP-11/45. This is the default.
23915 Generate code for a PDP-11/10. Implies -msoft-float -mno-split.
23921 Use 16-bit @code{int}. This is the default.
23927 Use 32-bit @code{int}.
23931 Target has split instruction and data space. Implies -m45.
23935 Use Unix assembler syntax.
23939 Use DEC assembler syntax.
23943 Use GNU assembler syntax. This is the default.
23947 Use the new LRA register allocator. By default, the old ``reload''
23951 @node picoChip Options
23952 @subsection picoChip Options
23953 @cindex picoChip options
23955 These @samp{-m} options are defined for picoChip implementations:
23959 @item -mae=@var{ae_type}
23961 Set the instruction set, register set, and instruction scheduling
23962 parameters for array element type @var{ae_type}. Supported values
23963 for @var{ae_type} are @samp{ANY}, @samp{MUL}, and @samp{MAC}.
23965 @option{-mae=ANY} selects a completely generic AE type. Code
23966 generated with this option runs on any of the other AE types. The
23967 code is not as efficient as it would be if compiled for a specific
23968 AE type, and some types of operation (e.g., multiplication) do not
23969 work properly on all types of AE.
23971 @option{-mae=MUL} selects a MUL AE type. This is the most useful AE type
23972 for compiled code, and is the default.
23974 @option{-mae=MAC} selects a DSP-style MAC AE. Code compiled with this
23975 option may suffer from poor performance of byte (char) manipulation,
23976 since the DSP AE does not provide hardware support for byte load/stores.
23978 @item -msymbol-as-address
23979 Enable the compiler to directly use a symbol name as an address in a
23980 load/store instruction, without first loading it into a
23981 register. Typically, the use of this option generates larger
23982 programs, which run faster than when the option isn't used. However, the
23983 results vary from program to program, so it is left as a user option,
23984 rather than being permanently enabled.
23986 @item -mno-inefficient-warnings
23987 Disables warnings about the generation of inefficient code. These
23988 warnings can be generated, for example, when compiling code that
23989 performs byte-level memory operations on the MAC AE type. The MAC AE has
23990 no hardware support for byte-level memory operations, so all byte
23991 load/stores must be synthesized from word load/store operations. This is
23992 inefficient and a warning is generated to indicate
23993 that you should rewrite the code to avoid byte operations, or to target
23994 an AE type that has the necessary hardware support. This option disables
23999 @node PowerPC Options
24000 @subsection PowerPC Options
24001 @cindex PowerPC options
24003 These are listed under @xref{RS/6000 and PowerPC Options}.
24006 @subsection PRU Options
24007 @cindex PRU Options
24009 These command-line options are defined for PRU target:
24014 Link with a minimum runtime environment, with no support for static
24015 initializers and constructors. Using this option can significantly reduce
24016 the size of the final ELF binary. Beware that the compiler could still
24017 generate code with static initializers and constructors. It is up to the
24018 programmer to ensure that the source program will not use those features.
24020 @item -mmcu=@var{mcu}
24022 Specify the PRU MCU variant to use. Check Newlib for the exact list of
24027 Make GCC pass the @option{--no-relax} command-line option to the linker
24028 instead of the @option{--relax} option.
24032 Allow (or do not allow) GCC to use the LOOP instruction.
24034 @item -mabi=@var{variant}
24036 Specify the ABI variant to output code for. @option{-mabi=ti} selects the
24037 unmodified TI ABI while @option{-mabi=gnu} selects a GNU variant that copes
24038 more naturally with certain GCC assumptions. These are the differences:
24041 @item Function Pointer Size
24042 TI ABI specifies that function (code) pointers are 16-bit, whereas GNU
24043 supports only 32-bit data and code pointers.
24045 @item Optional Return Value Pointer
24046 Function return values larger than 64 bits are passed by using a hidden
24047 pointer as the first argument of the function. TI ABI, though, mandates that
24048 the pointer can be NULL in case the caller is not using the returned value.
24049 GNU always passes and expects a valid return value pointer.
24053 The current @option{-mabi=ti} implementation simply raises a compile error
24054 when any of the above code constructs is detected. As a consequence
24055 the standard C library cannot be built and it is omitted when linking with
24058 Relaxation is a GNU feature and for safety reasons is disabled when using
24059 @option{-mabi=ti}. The TI toolchain does not emit relocations for QBBx
24060 instructions, so the GNU linker cannot adjust them when shortening adjacent
24061 LDI32 pseudo instructions.
24065 @node RISC-V Options
24066 @subsection RISC-V Options
24067 @cindex RISC-V Options
24069 These command-line options are defined for RISC-V targets:
24072 @item -mbranch-cost=@var{n}
24073 @opindex mbranch-cost
24074 Set the cost of branches to roughly @var{n} instructions.
24079 When generating PIC code, do or don't allow the use of PLTs. Ignored for
24080 non-PIC. The default is @option{-mplt}.
24082 @item -mabi=@var{ABI-string}
24084 Specify integer and floating-point calling convention. @var{ABI-string}
24085 contains two parts: the size of integer types and the registers used for
24086 floating-point types. For example @samp{-march=rv64ifd -mabi=lp64d} means that
24087 @samp{long} and pointers are 64-bit (implicitly defining @samp{int} to be
24088 32-bit), and that floating-point values up to 64 bits wide are passed in F
24089 registers. Contrast this with @samp{-march=rv64ifd -mabi=lp64f}, which still
24090 allows the compiler to generate code that uses the F and D extensions but only
24091 allows floating-point values up to 32 bits long to be passed in registers; or
24092 @samp{-march=rv64ifd -mabi=lp64}, in which no floating-point arguments will be
24093 passed in registers.
24095 The default for this argument is system dependent, users who want a specific
24096 calling convention should specify one explicitly. The valid calling
24097 conventions are: @samp{ilp32}, @samp{ilp32f}, @samp{ilp32d}, @samp{lp64},
24098 @samp{lp64f}, and @samp{lp64d}. Some calling conventions are impossible to
24099 implement on some ISAs: for example, @samp{-march=rv32if -mabi=ilp32d} is
24100 invalid because the ABI requires 64-bit values be passed in F registers, but F
24101 registers are only 32 bits wide. There is also the @samp{ilp32e} ABI that can
24102 only be used with the @samp{rv32e} architecture. This ABI is not well
24103 specified at present, and is subject to change.
24108 Do or don't use hardware floating-point divide and square root instructions.
24109 This requires the F or D extensions for floating-point registers. The default
24110 is to use them if the specified architecture has these instructions.
24115 Do or don't use hardware instructions for integer division. This requires the
24116 M extension. The default is to use them if the specified architecture has
24117 these instructions.
24119 @item -march=@var{ISA-string}
24121 Generate code for given RISC-V ISA (e.g.@: @samp{rv64im}). ISA strings must be
24122 lower-case. Examples include @samp{rv64i}, @samp{rv32g}, @samp{rv32e}, and
24125 @item -mtune=@var{processor-string}
24127 Optimize the output for the given processor, specified by microarchitecture
24128 name. Permissible values for this option are: @samp{rocket},
24129 @samp{sifive-3-series}, @samp{sifive-5-series}, @samp{sifive-7-series},
24132 When @option{-mtune=} is not specified, the default is @samp{rocket}.
24134 The @samp{size} choice is not intended for use by end-users. This is used
24135 when @option{-Os} is specified. It overrides the instruction cost info
24136 provided by @option{-mtune=}, but does not override the pipeline info. This
24137 helps reduce code size while still giving good performance.
24139 @item -mpreferred-stack-boundary=@var{num}
24140 @opindex mpreferred-stack-boundary
24141 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
24142 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
24143 the default is 4 (16 bytes or 128-bits).
24145 @strong{Warning:} If you use this switch, then you must build all modules with
24146 the same value, including any libraries. This includes the system libraries
24147 and startup modules.
24149 @item -msmall-data-limit=@var{n}
24150 @opindex msmall-data-limit
24151 Put global and static data smaller than @var{n} bytes into a special section
24154 @item -msave-restore
24155 @itemx -mno-save-restore
24156 @opindex msave-restore
24157 Do or don't use smaller but slower prologue and epilogue code that uses
24158 library function calls. The default is to use fast inline prologues and
24161 @item -mstrict-align
24162 @itemx -mno-strict-align
24163 @opindex mstrict-align
24164 Do not or do generate unaligned memory accesses. The default is set depending
24165 on whether the processor we are optimizing for supports fast unaligned access
24168 @item -mcmodel=medlow
24169 @opindex mcmodel=medlow
24170 Generate code for the medium-low code model. The program and its statically
24171 defined symbols must lie within a single 2 GiB address range and must lie
24172 between absolute addresses @minus{}2 GiB and +2 GiB. Programs can be
24173 statically or dynamically linked. This is the default code model.
24175 @item -mcmodel=medany
24176 @opindex mcmodel=medany
24177 Generate code for the medium-any code model. The program and its statically
24178 defined symbols must be within any single 2 GiB address range. Programs can be
24179 statically or dynamically linked.
24181 @item -mexplicit-relocs
24182 @itemx -mno-exlicit-relocs
24183 Use or do not use assembler relocation operators when dealing with symbolic
24184 addresses. The alternative is to use assembler macros instead, which may
24185 limit optimization.
24189 Take advantage of linker relaxations to reduce the number of instructions
24190 required to materialize symbol addresses. The default is to take advantage of
24191 linker relaxations.
24193 @item -memit-attribute
24194 @itemx -mno-emit-attribute
24195 Emit (do not emit) RISC-V attribute to record extra information into ELF
24196 objects. This feature requires at least binutils 2.32.
24198 @item -malign-data=@var{type}
24199 @opindex malign-data
24200 Control how GCC aligns variables and constants of array, structure, or union
24201 types. Supported values for @var{type} are @samp{xlen} which uses x register
24202 width as the alignment value, and @samp{natural} which uses natural alignment.
24203 @samp{xlen} is the default.
24207 @subsection RL78 Options
24208 @cindex RL78 Options
24214 Links in additional target libraries to support operation within a
24223 Specifies the type of hardware multiplication and division support to
24224 be used. The simplest is @code{none}, which uses software for both
24225 multiplication and division. This is the default. The @code{g13}
24226 value is for the hardware multiply/divide peripheral found on the
24227 RL78/G13 (S2 core) targets. The @code{g14} value selects the use of
24228 the multiplication and division instructions supported by the RL78/G14
24229 (S3 core) parts. The value @code{rl78} is an alias for @code{g14} and
24230 the value @code{mg10} is an alias for @code{none}.
24232 In addition a C preprocessor macro is defined, based upon the setting
24233 of this option. Possible values are: @code{__RL78_MUL_NONE__},
24234 @code{__RL78_MUL_G13__} or @code{__RL78_MUL_G14__}.
24241 Specifies the RL78 core to target. The default is the G14 core, also
24242 known as an S3 core or just RL78. The G13 or S2 core does not have
24243 multiply or divide instructions, instead it uses a hardware peripheral
24244 for these operations. The G10 or S1 core does not have register
24245 banks, so it uses a different calling convention.
24247 If this option is set it also selects the type of hardware multiply
24248 support to use, unless this is overridden by an explicit
24249 @option{-mmul=none} option on the command line. Thus specifying
24250 @option{-mcpu=g13} enables the use of the G13 hardware multiply
24251 peripheral and specifying @option{-mcpu=g10} disables the use of
24252 hardware multiplications altogether.
24254 Note, although the RL78/G14 core is the default target, specifying
24255 @option{-mcpu=g14} or @option{-mcpu=rl78} on the command line does
24256 change the behavior of the toolchain since it also enables G14
24257 hardware multiply support. If these options are not specified on the
24258 command line then software multiplication routines will be used even
24259 though the code targets the RL78 core. This is for backwards
24260 compatibility with older toolchains which did not have hardware
24261 multiply and divide support.
24263 In addition a C preprocessor macro is defined, based upon the setting
24264 of this option. Possible values are: @code{__RL78_G10__},
24265 @code{__RL78_G13__} or @code{__RL78_G14__}.
24275 These are aliases for the corresponding @option{-mcpu=} option. They
24276 are provided for backwards compatibility.
24280 Allow the compiler to use all of the available registers. By default
24281 registers @code{r24..r31} are reserved for use in interrupt handlers.
24282 With this option enabled these registers can be used in ordinary
24285 @item -m64bit-doubles
24286 @itemx -m32bit-doubles
24287 @opindex m64bit-doubles
24288 @opindex m32bit-doubles
24289 Make the @code{double} data type be 64 bits (@option{-m64bit-doubles})
24290 or 32 bits (@option{-m32bit-doubles}) in size. The default is
24291 @option{-m32bit-doubles}.
24293 @item -msave-mduc-in-interrupts
24294 @itemx -mno-save-mduc-in-interrupts
24295 @opindex msave-mduc-in-interrupts
24296 @opindex mno-save-mduc-in-interrupts
24297 Specifies that interrupt handler functions should preserve the
24298 MDUC registers. This is only necessary if normal code might use
24299 the MDUC registers, for example because it performs multiplication
24300 and division operations. The default is to ignore the MDUC registers
24301 as this makes the interrupt handlers faster. The target option -mg13
24302 needs to be passed for this to work as this feature is only available
24303 on the G13 target (S2 core). The MDUC registers will only be saved
24304 if the interrupt handler performs a multiplication or division
24305 operation or it calls another function.
24309 @node RS/6000 and PowerPC Options
24310 @subsection IBM RS/6000 and PowerPC Options
24311 @cindex RS/6000 and PowerPC Options
24312 @cindex IBM RS/6000 and PowerPC Options
24314 These @samp{-m} options are defined for the IBM RS/6000 and PowerPC:
24316 @item -mpowerpc-gpopt
24317 @itemx -mno-powerpc-gpopt
24318 @itemx -mpowerpc-gfxopt
24319 @itemx -mno-powerpc-gfxopt
24322 @itemx -mno-powerpc64
24326 @itemx -mno-popcntb
24328 @itemx -mno-popcntd
24335 @itemx -mno-hard-dfp
24336 @opindex mpowerpc-gpopt
24337 @opindex mno-powerpc-gpopt
24338 @opindex mpowerpc-gfxopt
24339 @opindex mno-powerpc-gfxopt
24340 @opindex mpowerpc64
24341 @opindex mno-powerpc64
24345 @opindex mno-popcntb
24347 @opindex mno-popcntd
24353 @opindex mno-hard-dfp
24354 You use these options to specify which instructions are available on the
24355 processor you are using. The default value of these options is
24356 determined when configuring GCC@. Specifying the
24357 @option{-mcpu=@var{cpu_type}} overrides the specification of these
24358 options. We recommend you use the @option{-mcpu=@var{cpu_type}} option
24359 rather than the options listed above.
24361 Specifying @option{-mpowerpc-gpopt} allows
24362 GCC to use the optional PowerPC architecture instructions in the
24363 General Purpose group, including floating-point square root. Specifying
24364 @option{-mpowerpc-gfxopt} allows GCC to
24365 use the optional PowerPC architecture instructions in the Graphics
24366 group, including floating-point select.
24368 The @option{-mmfcrf} option allows GCC to generate the move from
24369 condition register field instruction implemented on the POWER4
24370 processor and other processors that support the PowerPC V2.01
24372 The @option{-mpopcntb} option allows GCC to generate the popcount and
24373 double-precision FP reciprocal estimate instruction implemented on the
24374 POWER5 processor and other processors that support the PowerPC V2.02
24376 The @option{-mpopcntd} option allows GCC to generate the popcount
24377 instruction implemented on the POWER7 processor and other processors
24378 that support the PowerPC V2.06 architecture.
24379 The @option{-mfprnd} option allows GCC to generate the FP round to
24380 integer instructions implemented on the POWER5+ processor and other
24381 processors that support the PowerPC V2.03 architecture.
24382 The @option{-mcmpb} option allows GCC to generate the compare bytes
24383 instruction implemented on the POWER6 processor and other processors
24384 that support the PowerPC V2.05 architecture.
24385 The @option{-mhard-dfp} option allows GCC to generate the decimal
24386 floating-point instructions implemented on some POWER processors.
24388 The @option{-mpowerpc64} option allows GCC to generate the additional
24389 64-bit instructions that are found in the full PowerPC64 architecture
24390 and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to
24391 @option{-mno-powerpc64}.
24393 @item -mcpu=@var{cpu_type}
24395 Set architecture type, register usage, and
24396 instruction scheduling parameters for machine type @var{cpu_type}.
24397 Supported values for @var{cpu_type} are @samp{401}, @samp{403},
24398 @samp{405}, @samp{405fp}, @samp{440}, @samp{440fp}, @samp{464}, @samp{464fp},
24399 @samp{476}, @samp{476fp}, @samp{505}, @samp{601}, @samp{602}, @samp{603},
24400 @samp{603e}, @samp{604}, @samp{604e}, @samp{620}, @samp{630}, @samp{740},
24401 @samp{7400}, @samp{7450}, @samp{750}, @samp{801}, @samp{821}, @samp{823},
24402 @samp{860}, @samp{970}, @samp{8540}, @samp{a2}, @samp{e300c2},
24403 @samp{e300c3}, @samp{e500mc}, @samp{e500mc64}, @samp{e5500},
24404 @samp{e6500}, @samp{ec603e}, @samp{G3}, @samp{G4}, @samp{G5},
24405 @samp{titan}, @samp{power3}, @samp{power4}, @samp{power5}, @samp{power5+},
24406 @samp{power6}, @samp{power6x}, @samp{power7}, @samp{power8},
24407 @samp{power9}, @samp{future}, @samp{powerpc}, @samp{powerpc64},
24408 @samp{powerpc64le}, @samp{rs64}, and @samp{native}.
24410 @option{-mcpu=powerpc}, @option{-mcpu=powerpc64}, and
24411 @option{-mcpu=powerpc64le} specify pure 32-bit PowerPC (either
24412 endian), 64-bit big endian PowerPC and 64-bit little endian PowerPC
24413 architecture machine types, with an appropriate, generic processor
24414 model assumed for scheduling purposes.
24416 Specifying @samp{native} as cpu type detects and selects the
24417 architecture option that corresponds to the host processor of the
24418 system performing the compilation.
24419 @option{-mcpu=native} has no effect if GCC does not recognize the
24422 The other options specify a specific processor. Code generated under
24423 those options runs best on that processor, and may not run at all on
24426 The @option{-mcpu} options automatically enable or disable the
24429 @gccoptlist{-maltivec -mfprnd -mhard-float -mmfcrf -mmultiple @gol
24430 -mpopcntb -mpopcntd -mpowerpc64 @gol
24431 -mpowerpc-gpopt -mpowerpc-gfxopt @gol
24432 -mmulhw -mdlmzb -mmfpgpr -mvsx @gol
24433 -mcrypto -mhtm -mpower8-fusion -mpower8-vector @gol
24434 -mquad-memory -mquad-memory-atomic -mfloat128 -mfloat128-hardware}
24436 The particular options set for any particular CPU varies between
24437 compiler versions, depending on what setting seems to produce optimal
24438 code for that CPU; it doesn't necessarily reflect the actual hardware's
24439 capabilities. If you wish to set an individual option to a particular
24440 value, you may specify it after the @option{-mcpu} option, like
24441 @option{-mcpu=970 -mno-altivec}.
24443 On AIX, the @option{-maltivec} and @option{-mpowerpc64} options are
24444 not enabled or disabled by the @option{-mcpu} option at present because
24445 AIX does not have full support for these options. You may still
24446 enable or disable them individually if you're sure it'll work in your
24449 @item -mtune=@var{cpu_type}
24451 Set the instruction scheduling parameters for machine type
24452 @var{cpu_type}, but do not set the architecture type or register usage,
24453 as @option{-mcpu=@var{cpu_type}} does. The same
24454 values for @var{cpu_type} are used for @option{-mtune} as for
24455 @option{-mcpu}. If both are specified, the code generated uses the
24456 architecture and registers set by @option{-mcpu}, but the
24457 scheduling parameters set by @option{-mtune}.
24459 @item -mcmodel=small
24460 @opindex mcmodel=small
24461 Generate PowerPC64 code for the small model: The TOC is limited to
24464 @item -mcmodel=medium
24465 @opindex mcmodel=medium
24466 Generate PowerPC64 code for the medium model: The TOC and other static
24467 data may be up to a total of 4G in size. This is the default for 64-bit
24470 @item -mcmodel=large
24471 @opindex mcmodel=large
24472 Generate PowerPC64 code for the large model: The TOC may be up to 4G
24473 in size. Other data and code is only limited by the 64-bit address
24477 @itemx -mno-altivec
24479 @opindex mno-altivec
24480 Generate code that uses (does not use) AltiVec instructions, and also
24481 enable the use of built-in functions that allow more direct access to
24482 the AltiVec instruction set. You may also need to set
24483 @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI
24486 When @option{-maltivec} is used, the element order for AltiVec intrinsics
24487 such as @code{vec_splat}, @code{vec_extract}, and @code{vec_insert}
24488 match array element order corresponding to the endianness of the
24489 target. That is, element zero identifies the leftmost element in a
24490 vector register when targeting a big-endian platform, and identifies
24491 the rightmost element in a vector register when targeting a
24492 little-endian platform.
24497 @opindex mno-vrsave
24498 Generate VRSAVE instructions when generating AltiVec code.
24501 @opindex msecure-plt
24502 Generate code that allows @command{ld} and @command{ld.so}
24503 to build executables and shared
24504 libraries with non-executable @code{.plt} and @code{.got} sections.
24506 32-bit SYSV ABI option.
24510 Generate code that uses a BSS @code{.plt} section that @command{ld.so}
24512 requires @code{.plt} and @code{.got}
24513 sections that are both writable and executable.
24514 This is a PowerPC 32-bit SYSV ABI option.
24520 This switch enables or disables the generation of ISEL instructions.
24526 Generate code that uses (does not use) vector/scalar (VSX)
24527 instructions, and also enable the use of built-in functions that allow
24528 more direct access to the VSX instruction set.
24533 @opindex mno-crypto
24534 Enable the use (disable) of the built-in functions that allow direct
24535 access to the cryptographic instructions that were added in version
24536 2.07 of the PowerPC ISA.
24542 Enable (disable) the use of the built-in functions that allow direct
24543 access to the Hardware Transactional Memory (HTM) instructions that
24544 were added in version 2.07 of the PowerPC ISA.
24546 @item -mpower8-fusion
24547 @itemx -mno-power8-fusion
24548 @opindex mpower8-fusion
24549 @opindex mno-power8-fusion
24550 Generate code that keeps (does not keeps) some integer operations
24551 adjacent so that the instructions can be fused together on power8 and
24554 @item -mpower8-vector
24555 @itemx -mno-power8-vector
24556 @opindex mpower8-vector
24557 @opindex mno-power8-vector
24558 Generate code that uses (does not use) the vector and scalar
24559 instructions that were added in version 2.07 of the PowerPC ISA. Also
24560 enable the use of built-in functions that allow more direct access to
24561 the vector instructions.
24563 @item -mquad-memory
24564 @itemx -mno-quad-memory
24565 @opindex mquad-memory
24566 @opindex mno-quad-memory
24567 Generate code that uses (does not use) the non-atomic quad word memory
24568 instructions. The @option{-mquad-memory} option requires use of
24571 @item -mquad-memory-atomic
24572 @itemx -mno-quad-memory-atomic
24573 @opindex mquad-memory-atomic
24574 @opindex mno-quad-memory-atomic
24575 Generate code that uses (does not use) the atomic quad word memory
24576 instructions. The @option{-mquad-memory-atomic} option requires use of
24580 @itemx -mno-float128
24582 @opindex mno-float128
24583 Enable/disable the @var{__float128} keyword for IEEE 128-bit floating point
24584 and use either software emulation for IEEE 128-bit floating point or
24585 hardware instructions.
24587 The VSX instruction set (@option{-mvsx}, @option{-mcpu=power7},
24588 @option{-mcpu=power8}), or @option{-mcpu=power9} must be enabled to
24589 use the IEEE 128-bit floating point support. The IEEE 128-bit
24590 floating point support only works on PowerPC Linux systems.
24592 The default for @option{-mfloat128} is enabled on PowerPC Linux
24593 systems using the VSX instruction set, and disabled on other systems.
24595 If you use the ISA 3.0 instruction set (@option{-mpower9-vector} or
24596 @option{-mcpu=power9}) on a 64-bit system, the IEEE 128-bit floating
24597 point support will also enable the generation of ISA 3.0 IEEE 128-bit
24598 floating point instructions. Otherwise, if you do not specify to
24599 generate ISA 3.0 instructions or you are targeting a 32-bit big endian
24600 system, IEEE 128-bit floating point will be done with software
24603 @item -mfloat128-hardware
24604 @itemx -mno-float128-hardware
24605 @opindex mfloat128-hardware
24606 @opindex mno-float128-hardware
24607 Enable/disable using ISA 3.0 hardware instructions to support the
24608 @var{__float128} data type.
24610 The default for @option{-mfloat128-hardware} is enabled on PowerPC
24611 Linux systems using the ISA 3.0 instruction set, and disabled on other
24618 Generate code for 32-bit or 64-bit environments of Darwin and SVR4
24619 targets (including GNU/Linux). The 32-bit environment sets int, long
24620 and pointer to 32 bits and generates code that runs on any PowerPC
24621 variant. The 64-bit environment sets int to 32 bits and long and
24622 pointer to 64 bits, and generates code for PowerPC64, as for
24623 @option{-mpowerpc64}.
24626 @itemx -mno-fp-in-toc
24627 @itemx -mno-sum-in-toc
24628 @itemx -mminimal-toc
24630 @opindex mno-fp-in-toc
24631 @opindex mno-sum-in-toc
24632 @opindex mminimal-toc
24633 Modify generation of the TOC (Table Of Contents), which is created for
24634 every executable file. The @option{-mfull-toc} option is selected by
24635 default. In that case, GCC allocates at least one TOC entry for
24636 each unique non-automatic variable reference in your program. GCC
24637 also places floating-point constants in the TOC@. However, only
24638 16,384 entries are available in the TOC@.
24640 If you receive a linker error message that saying you have overflowed
24641 the available TOC space, you can reduce the amount of TOC space used
24642 with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options.
24643 @option{-mno-fp-in-toc} prevents GCC from putting floating-point
24644 constants in the TOC and @option{-mno-sum-in-toc} forces GCC to
24645 generate code to calculate the sum of an address and a constant at
24646 run time instead of putting that sum into the TOC@. You may specify one
24647 or both of these options. Each causes GCC to produce very slightly
24648 slower and larger code at the expense of conserving TOC space.
24650 If you still run out of space in the TOC even when you specify both of
24651 these options, specify @option{-mminimal-toc} instead. This option causes
24652 GCC to make only one TOC entry for every file. When you specify this
24653 option, GCC produces code that is slower and larger but which
24654 uses extremely little TOC space. You may wish to use this option
24655 only on files that contain less frequently-executed code.
24661 Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit
24662 @code{long} type, and the infrastructure needed to support them.
24663 Specifying @option{-maix64} implies @option{-mpowerpc64},
24664 while @option{-maix32} disables the 64-bit ABI and
24665 implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}.
24668 @itemx -mno-xl-compat
24669 @opindex mxl-compat
24670 @opindex mno-xl-compat
24671 Produce code that conforms more closely to IBM XL compiler semantics
24672 when using AIX-compatible ABI@. Pass floating-point arguments to
24673 prototyped functions beyond the register save area (RSA) on the stack
24674 in addition to argument FPRs. Do not assume that most significant
24675 double in 128-bit long double value is properly rounded when comparing
24676 values and converting to double. Use XL symbol names for long double
24679 The AIX calling convention was extended but not initially documented to
24680 handle an obscure K&R C case of calling a function that takes the
24681 address of its arguments with fewer arguments than declared. IBM XL
24682 compilers access floating-point arguments that do not fit in the
24683 RSA from the stack when a subroutine is compiled without
24684 optimization. Because always storing floating-point arguments on the
24685 stack is inefficient and rarely needed, this option is not enabled by
24686 default and only is necessary when calling subroutines compiled by IBM
24687 XL compilers without optimization.
24691 Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an
24692 application written to use message passing with special startup code to
24693 enable the application to run. The system must have PE installed in the
24694 standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file
24695 must be overridden with the @option{-specs=} option to specify the
24696 appropriate directory location. The Parallel Environment does not
24697 support threads, so the @option{-mpe} option and the @option{-pthread}
24698 option are incompatible.
24700 @item -malign-natural
24701 @itemx -malign-power
24702 @opindex malign-natural
24703 @opindex malign-power
24704 On AIX, 32-bit Darwin, and 64-bit PowerPC GNU/Linux, the option
24705 @option{-malign-natural} overrides the ABI-defined alignment of larger
24706 types, such as floating-point doubles, on their natural size-based boundary.
24707 The option @option{-malign-power} instructs GCC to follow the ABI-specified
24708 alignment rules. GCC defaults to the standard alignment defined in the ABI@.
24710 On 64-bit Darwin, natural alignment is the default, and @option{-malign-power}
24714 @itemx -mhard-float
24715 @opindex msoft-float
24716 @opindex mhard-float
24717 Generate code that does not use (uses) the floating-point register set.
24718 Software floating-point emulation is provided if you use the
24719 @option{-msoft-float} option, and pass the option to GCC when linking.
24722 @itemx -mno-multiple
24724 @opindex mno-multiple
24725 Generate code that uses (does not use) the load multiple word
24726 instructions and the store multiple word instructions. These
24727 instructions are generated by default on POWER systems, and not
24728 generated on PowerPC systems. Do not use @option{-mmultiple} on little-endian
24729 PowerPC systems, since those instructions do not work when the
24730 processor is in little-endian mode. The exceptions are PPC740 and
24731 PPC750 which permit these instructions in little-endian mode.
24736 @opindex mno-update
24737 Generate code that uses (does not use) the load or store instructions
24738 that update the base register to the address of the calculated memory
24739 location. These instructions are generated by default. If you use
24740 @option{-mno-update}, there is a small window between the time that the
24741 stack pointer is updated and the address of the previous frame is
24742 stored, which means code that walks the stack frame across interrupts or
24743 signals may get corrupted data.
24745 @item -mavoid-indexed-addresses
24746 @itemx -mno-avoid-indexed-addresses
24747 @opindex mavoid-indexed-addresses
24748 @opindex mno-avoid-indexed-addresses
24749 Generate code that tries to avoid (not avoid) the use of indexed load
24750 or store instructions. These instructions can incur a performance
24751 penalty on Power6 processors in certain situations, such as when
24752 stepping through large arrays that cross a 16M boundary. This option
24753 is enabled by default when targeting Power6 and disabled otherwise.
24756 @itemx -mno-fused-madd
24757 @opindex mfused-madd
24758 @opindex mno-fused-madd
24759 Generate code that uses (does not use) the floating-point multiply and
24760 accumulate instructions. These instructions are generated by default
24761 if hardware floating point is used. The machine-dependent
24762 @option{-mfused-madd} option is now mapped to the machine-independent
24763 @option{-ffp-contract=fast} option, and @option{-mno-fused-madd} is
24764 mapped to @option{-ffp-contract=off}.
24770 Generate code that uses (does not use) the half-word multiply and
24771 multiply-accumulate instructions on the IBM 405, 440, 464 and 476 processors.
24772 These instructions are generated by default when targeting those
24779 Generate code that uses (does not use) the string-search @samp{dlmzb}
24780 instruction on the IBM 405, 440, 464 and 476 processors. This instruction is
24781 generated by default when targeting those processors.
24783 @item -mno-bit-align
24785 @opindex mno-bit-align
24786 @opindex mbit-align
24787 On System V.4 and embedded PowerPC systems do not (do) force structures
24788 and unions that contain bit-fields to be aligned to the base type of the
24791 For example, by default a structure containing nothing but 8
24792 @code{unsigned} bit-fields of length 1 is aligned to a 4-byte
24793 boundary and has a size of 4 bytes. By using @option{-mno-bit-align},
24794 the structure is aligned to a 1-byte boundary and is 1 byte in
24797 @item -mno-strict-align
24798 @itemx -mstrict-align
24799 @opindex mno-strict-align
24800 @opindex mstrict-align
24801 On System V.4 and embedded PowerPC systems do not (do) assume that
24802 unaligned memory references are handled by the system.
24804 @item -mrelocatable
24805 @itemx -mno-relocatable
24806 @opindex mrelocatable
24807 @opindex mno-relocatable
24808 Generate code that allows (does not allow) a static executable to be
24809 relocated to a different address at run time. A simple embedded
24810 PowerPC system loader should relocate the entire contents of
24811 @code{.got2} and 4-byte locations listed in the @code{.fixup} section,
24812 a table of 32-bit addresses generated by this option. For this to
24813 work, all objects linked together must be compiled with
24814 @option{-mrelocatable} or @option{-mrelocatable-lib}.
24815 @option{-mrelocatable} code aligns the stack to an 8-byte boundary.
24817 @item -mrelocatable-lib
24818 @itemx -mno-relocatable-lib
24819 @opindex mrelocatable-lib
24820 @opindex mno-relocatable-lib
24821 Like @option{-mrelocatable}, @option{-mrelocatable-lib} generates a
24822 @code{.fixup} section to allow static executables to be relocated at
24823 run time, but @option{-mrelocatable-lib} does not use the smaller stack
24824 alignment of @option{-mrelocatable}. Objects compiled with
24825 @option{-mrelocatable-lib} may be linked with objects compiled with
24826 any combination of the @option{-mrelocatable} options.
24832 On System V.4 and embedded PowerPC systems do not (do) assume that
24833 register 2 contains a pointer to a global area pointing to the addresses
24834 used in the program.
24837 @itemx -mlittle-endian
24839 @opindex mlittle-endian
24840 On System V.4 and embedded PowerPC systems compile code for the
24841 processor in little-endian mode. The @option{-mlittle-endian} option is
24842 the same as @option{-mlittle}.
24845 @itemx -mbig-endian
24847 @opindex mbig-endian
24848 On System V.4 and embedded PowerPC systems compile code for the
24849 processor in big-endian mode. The @option{-mbig-endian} option is
24850 the same as @option{-mbig}.
24852 @item -mdynamic-no-pic
24853 @opindex mdynamic-no-pic
24854 On Darwin and Mac OS X systems, compile code so that it is not
24855 relocatable, but that its external references are relocatable. The
24856 resulting code is suitable for applications, but not shared
24859 @item -msingle-pic-base
24860 @opindex msingle-pic-base
24861 Treat the register used for PIC addressing as read-only, rather than
24862 loading it in the prologue for each function. The runtime system is
24863 responsible for initializing this register with an appropriate value
24864 before execution begins.
24866 @item -mprioritize-restricted-insns=@var{priority}
24867 @opindex mprioritize-restricted-insns
24868 This option controls the priority that is assigned to
24869 dispatch-slot restricted instructions during the second scheduling
24870 pass. The argument @var{priority} takes the value @samp{0}, @samp{1},
24871 or @samp{2} to assign no, highest, or second-highest (respectively)
24872 priority to dispatch-slot restricted
24875 @item -msched-costly-dep=@var{dependence_type}
24876 @opindex msched-costly-dep
24877 This option controls which dependences are considered costly
24878 by the target during instruction scheduling. The argument
24879 @var{dependence_type} takes one of the following values:
24883 No dependence is costly.
24886 All dependences are costly.
24888 @item @samp{true_store_to_load}
24889 A true dependence from store to load is costly.
24891 @item @samp{store_to_load}
24892 Any dependence from store to load is costly.
24895 Any dependence for which the latency is greater than or equal to
24896 @var{number} is costly.
24899 @item -minsert-sched-nops=@var{scheme}
24900 @opindex minsert-sched-nops
24901 This option controls which NOP insertion scheme is used during
24902 the second scheduling pass. The argument @var{scheme} takes one of the
24910 Pad with NOPs any dispatch group that has vacant issue slots,
24911 according to the scheduler's grouping.
24913 @item @samp{regroup_exact}
24914 Insert NOPs to force costly dependent insns into
24915 separate groups. Insert exactly as many NOPs as needed to force an insn
24916 to a new group, according to the estimated processor grouping.
24919 Insert NOPs to force costly dependent insns into
24920 separate groups. Insert @var{number} NOPs to force an insn to a new group.
24924 @opindex mcall-sysv
24925 On System V.4 and embedded PowerPC systems compile code using calling
24926 conventions that adhere to the March 1995 draft of the System V
24927 Application Binary Interface, PowerPC processor supplement. This is the
24928 default unless you configured GCC using @samp{powerpc-*-eabiaix}.
24930 @item -mcall-sysv-eabi
24932 @opindex mcall-sysv-eabi
24933 @opindex mcall-eabi
24934 Specify both @option{-mcall-sysv} and @option{-meabi} options.
24936 @item -mcall-sysv-noeabi
24937 @opindex mcall-sysv-noeabi
24938 Specify both @option{-mcall-sysv} and @option{-mno-eabi} options.
24940 @item -mcall-aixdesc
24942 On System V.4 and embedded PowerPC systems compile code for the AIX
24946 @opindex mcall-linux
24947 On System V.4 and embedded PowerPC systems compile code for the
24948 Linux-based GNU system.
24950 @item -mcall-freebsd
24951 @opindex mcall-freebsd
24952 On System V.4 and embedded PowerPC systems compile code for the
24953 FreeBSD operating system.
24955 @item -mcall-netbsd
24956 @opindex mcall-netbsd
24957 On System V.4 and embedded PowerPC systems compile code for the
24958 NetBSD operating system.
24960 @item -mcall-openbsd
24961 @opindex mcall-netbsd
24962 On System V.4 and embedded PowerPC systems compile code for the
24963 OpenBSD operating system.
24965 @item -mtraceback=@var{traceback_type}
24966 @opindex mtraceback
24967 Select the type of traceback table. Valid values for @var{traceback_type}
24968 are @samp{full}, @samp{part}, and @samp{no}.
24970 @item -maix-struct-return
24971 @opindex maix-struct-return
24972 Return all structures in memory (as specified by the AIX ABI)@.
24974 @item -msvr4-struct-return
24975 @opindex msvr4-struct-return
24976 Return structures smaller than 8 bytes in registers (as specified by the
24979 @item -mabi=@var{abi-type}
24981 Extend the current ABI with a particular extension, or remove such extension.
24982 Valid values are @samp{altivec}, @samp{no-altivec},
24983 @samp{ibmlongdouble}, @samp{ieeelongdouble},
24984 @samp{elfv1}, @samp{elfv2}@.
24986 @item -mabi=ibmlongdouble
24987 @opindex mabi=ibmlongdouble
24988 Change the current ABI to use IBM extended-precision long double.
24989 This is not likely to work if your system defaults to using IEEE
24990 extended-precision long double. If you change the long double type
24991 from IEEE extended-precision, the compiler will issue a warning unless
24992 you use the @option{-Wno-psabi} option. Requires @option{-mlong-double-128}
24995 @item -mabi=ieeelongdouble
24996 @opindex mabi=ieeelongdouble
24997 Change the current ABI to use IEEE extended-precision long double.
24998 This is not likely to work if your system defaults to using IBM
24999 extended-precision long double. If you change the long double type
25000 from IBM extended-precision, the compiler will issue a warning unless
25001 you use the @option{-Wno-psabi} option. Requires @option{-mlong-double-128}
25005 @opindex mabi=elfv1
25006 Change the current ABI to use the ELFv1 ABI.
25007 This is the default ABI for big-endian PowerPC 64-bit Linux.
25008 Overriding the default ABI requires special system support and is
25009 likely to fail in spectacular ways.
25012 @opindex mabi=elfv2
25013 Change the current ABI to use the ELFv2 ABI.
25014 This is the default ABI for little-endian PowerPC 64-bit Linux.
25015 Overriding the default ABI requires special system support and is
25016 likely to fail in spectacular ways.
25018 @item -mgnu-attribute
25019 @itemx -mno-gnu-attribute
25020 @opindex mgnu-attribute
25021 @opindex mno-gnu-attribute
25022 Emit .gnu_attribute assembly directives to set tag/value pairs in a
25023 .gnu.attributes section that specify ABI variations in function
25024 parameters or return values.
25027 @itemx -mno-prototype
25028 @opindex mprototype
25029 @opindex mno-prototype
25030 On System V.4 and embedded PowerPC systems assume that all calls to
25031 variable argument functions are properly prototyped. Otherwise, the
25032 compiler must insert an instruction before every non-prototyped call to
25033 set or clear bit 6 of the condition code register (@code{CR}) to
25034 indicate whether floating-point values are passed in the floating-point
25035 registers in case the function takes variable arguments. With
25036 @option{-mprototype}, only calls to prototyped variable argument functions
25037 set or clear the bit.
25041 On embedded PowerPC systems, assume that the startup module is called
25042 @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and
25043 @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}
25048 On embedded PowerPC systems, assume that the startup module is called
25049 @file{crt0.o} and the standard C libraries are @file{libmvme.a} and
25054 On embedded PowerPC systems, assume that the startup module is called
25055 @file{crt0.o} and the standard C libraries are @file{libads.a} and
25058 @item -myellowknife
25059 @opindex myellowknife
25060 On embedded PowerPC systems, assume that the startup module is called
25061 @file{crt0.o} and the standard C libraries are @file{libyk.a} and
25066 On System V.4 and embedded PowerPC systems, specify that you are
25067 compiling for a VxWorks system.
25071 On embedded PowerPC systems, set the @code{PPC_EMB} bit in the ELF flags
25072 header to indicate that @samp{eabi} extended relocations are used.
25078 On System V.4 and embedded PowerPC systems do (do not) adhere to the
25079 Embedded Applications Binary Interface (EABI), which is a set of
25080 modifications to the System V.4 specifications. Selecting @option{-meabi}
25081 means that the stack is aligned to an 8-byte boundary, a function
25082 @code{__eabi} is called from @code{main} to set up the EABI
25083 environment, and the @option{-msdata} option can use both @code{r2} and
25084 @code{r13} to point to two separate small data areas. Selecting
25085 @option{-mno-eabi} means that the stack is aligned to a 16-byte boundary,
25086 no EABI initialization function is called from @code{main}, and the
25087 @option{-msdata} option only uses @code{r13} to point to a single
25088 small data area. The @option{-meabi} option is on by default if you
25089 configured GCC using one of the @samp{powerpc*-*-eabi*} options.
25092 @opindex msdata=eabi
25093 On System V.4 and embedded PowerPC systems, put small initialized
25094 @code{const} global and static data in the @code{.sdata2} section, which
25095 is pointed to by register @code{r2}. Put small initialized
25096 non-@code{const} global and static data in the @code{.sdata} section,
25097 which is pointed to by register @code{r13}. Put small uninitialized
25098 global and static data in the @code{.sbss} section, which is adjacent to
25099 the @code{.sdata} section. The @option{-msdata=eabi} option is
25100 incompatible with the @option{-mrelocatable} option. The
25101 @option{-msdata=eabi} option also sets the @option{-memb} option.
25104 @opindex msdata=sysv
25105 On System V.4 and embedded PowerPC systems, put small global and static
25106 data in the @code{.sdata} section, which is pointed to by register
25107 @code{r13}. Put small uninitialized global and static data in the
25108 @code{.sbss} section, which is adjacent to the @code{.sdata} section.
25109 The @option{-msdata=sysv} option is incompatible with the
25110 @option{-mrelocatable} option.
25112 @item -msdata=default
25114 @opindex msdata=default
25116 On System V.4 and embedded PowerPC systems, if @option{-meabi} is used,
25117 compile code the same as @option{-msdata=eabi}, otherwise compile code the
25118 same as @option{-msdata=sysv}.
25121 @opindex msdata=data
25122 On System V.4 and embedded PowerPC systems, put small global
25123 data in the @code{.sdata} section. Put small uninitialized global
25124 data in the @code{.sbss} section. Do not use register @code{r13}
25125 to address small data however. This is the default behavior unless
25126 other @option{-msdata} options are used.
25130 @opindex msdata=none
25132 On embedded PowerPC systems, put all initialized global and static data
25133 in the @code{.data} section, and all uninitialized data in the
25134 @code{.bss} section.
25136 @item -mreadonly-in-sdata
25137 @opindex mreadonly-in-sdata
25138 @opindex mno-readonly-in-sdata
25139 Put read-only objects in the @code{.sdata} section as well. This is the
25142 @item -mblock-move-inline-limit=@var{num}
25143 @opindex mblock-move-inline-limit
25144 Inline all block moves (such as calls to @code{memcpy} or structure
25145 copies) less than or equal to @var{num} bytes. The minimum value for
25146 @var{num} is 32 bytes on 32-bit targets and 64 bytes on 64-bit
25147 targets. The default value is target-specific.
25149 @item -mblock-compare-inline-limit=@var{num}
25150 @opindex mblock-compare-inline-limit
25151 Generate non-looping inline code for all block compares (such as calls
25152 to @code{memcmp} or structure compares) less than or equal to @var{num}
25153 bytes. If @var{num} is 0, all inline expansion (non-loop and loop) of
25154 block compare is disabled. The default value is target-specific.
25156 @item -mblock-compare-inline-loop-limit=@var{num}
25157 @opindex mblock-compare-inline-loop-limit
25158 Generate an inline expansion using loop code for all block compares that
25159 are less than or equal to @var{num} bytes, but greater than the limit
25160 for non-loop inline block compare expansion. If the block length is not
25161 constant, at most @var{num} bytes will be compared before @code{memcmp}
25162 is called to compare the remainder of the block. The default value is
25165 @item -mstring-compare-inline-limit=@var{num}
25166 @opindex mstring-compare-inline-limit
25167 Compare at most @var{num} string bytes with inline code.
25168 If the difference or end of string is not found at the
25169 end of the inline compare a call to @code{strcmp} or @code{strncmp} will
25170 take care of the rest of the comparison. The default is 64 bytes.
25174 @cindex smaller data references (PowerPC)
25175 @cindex .sdata/.sdata2 references (PowerPC)
25176 On embedded PowerPC systems, put global and static items less than or
25177 equal to @var{num} bytes into the small data or BSS sections instead of
25178 the normal data or BSS section. By default, @var{num} is 8. The
25179 @option{-G @var{num}} switch is also passed to the linker.
25180 All modules should be compiled with the same @option{-G @var{num}} value.
25183 @itemx -mno-regnames
25185 @opindex mno-regnames
25186 On System V.4 and embedded PowerPC systems do (do not) emit register
25187 names in the assembly language output using symbolic forms.
25190 @itemx -mno-longcall
25192 @opindex mno-longcall
25193 By default assume that all calls are far away so that a longer and more
25194 expensive calling sequence is required. This is required for calls
25195 farther than 32 megabytes (33,554,432 bytes) from the current location.
25196 A short call is generated if the compiler knows
25197 the call cannot be that far away. This setting can be overridden by
25198 the @code{shortcall} function attribute, or by @code{#pragma
25201 Some linkers are capable of detecting out-of-range calls and generating
25202 glue code on the fly. On these systems, long calls are unnecessary and
25203 generate slower code. As of this writing, the AIX linker can do this,
25204 as can the GNU linker for PowerPC/64. It is planned to add this feature
25205 to the GNU linker for 32-bit PowerPC systems as well.
25207 On PowerPC64 ELFv2 and 32-bit PowerPC systems with newer GNU linkers,
25208 GCC can generate long calls using an inline PLT call sequence (see
25209 @option{-mpltseq}). PowerPC with @option{-mbss-plt} and PowerPC64
25210 ELFv1 (big-endian) do not support inline PLT calls.
25212 On Darwin/PPC systems, @code{#pragma longcall} generates @code{jbsr
25213 callee, L42}, plus a @dfn{branch island} (glue code). The two target
25214 addresses represent the callee and the branch island. The
25215 Darwin/PPC linker prefers the first address and generates a @code{bl
25216 callee} if the PPC @code{bl} instruction reaches the callee directly;
25217 otherwise, the linker generates @code{bl L42} to call the branch
25218 island. The branch island is appended to the body of the
25219 calling function; it computes the full 32-bit address of the callee
25222 On Mach-O (Darwin) systems, this option directs the compiler emit to
25223 the glue for every direct call, and the Darwin linker decides whether
25224 to use or discard it.
25226 In the future, GCC may ignore all longcall specifications
25227 when the linker is known to generate glue.
25232 @opindex mno-pltseq
25233 Implement (do not implement) -fno-plt and long calls using an inline
25234 PLT call sequence that supports lazy linking and long calls to
25235 functions in dlopen'd shared libraries. Inline PLT calls are only
25236 supported on PowerPC64 ELFv2 and 32-bit PowerPC systems with newer GNU
25237 linkers, and are enabled by default if the support is detected when
25238 configuring GCC, and, in the case of 32-bit PowerPC, if GCC is
25239 configured with @option{--enable-secureplt}. @option{-mpltseq} code
25240 and @option{-mbss-plt} 32-bit PowerPC relocatable objects may not be
25243 @item -mtls-markers
25244 @itemx -mno-tls-markers
25245 @opindex mtls-markers
25246 @opindex mno-tls-markers
25247 Mark (do not mark) calls to @code{__tls_get_addr} with a relocation
25248 specifying the function argument. The relocation allows the linker to
25249 reliably associate function call with argument setup instructions for
25250 TLS optimization, which in turn allows GCC to better schedule the
25256 This option enables use of the reciprocal estimate and
25257 reciprocal square root estimate instructions with additional
25258 Newton-Raphson steps to increase precision instead of doing a divide or
25259 square root and divide for floating-point arguments. You should use
25260 the @option{-ffast-math} option when using @option{-mrecip} (or at
25261 least @option{-funsafe-math-optimizations},
25262 @option{-ffinite-math-only}, @option{-freciprocal-math} and
25263 @option{-fno-trapping-math}). Note that while the throughput of the
25264 sequence is generally higher than the throughput of the non-reciprocal
25265 instruction, the precision of the sequence can be decreased by up to 2
25266 ulp (i.e.@: the inverse of 1.0 equals 0.99999994) for reciprocal square
25269 @item -mrecip=@var{opt}
25270 @opindex mrecip=opt
25271 This option controls which reciprocal estimate instructions
25272 may be used. @var{opt} is a comma-separated list of options, which may
25273 be preceded by a @code{!} to invert the option:
25278 Enable all estimate instructions.
25281 Enable the default instructions, equivalent to @option{-mrecip}.
25284 Disable all estimate instructions, equivalent to @option{-mno-recip}.
25287 Enable the reciprocal approximation instructions for both
25288 single and double precision.
25291 Enable the single-precision reciprocal approximation instructions.
25294 Enable the double-precision reciprocal approximation instructions.
25297 Enable the reciprocal square root approximation instructions for both
25298 single and double precision.
25301 Enable the single-precision reciprocal square root approximation instructions.
25304 Enable the double-precision reciprocal square root approximation instructions.
25308 So, for example, @option{-mrecip=all,!rsqrtd} enables
25309 all of the reciprocal estimate instructions, except for the
25310 @code{FRSQRTE}, @code{XSRSQRTEDP}, and @code{XVRSQRTEDP} instructions
25311 which handle the double-precision reciprocal square root calculations.
25313 @item -mrecip-precision
25314 @itemx -mno-recip-precision
25315 @opindex mrecip-precision
25316 Assume (do not assume) that the reciprocal estimate instructions
25317 provide higher-precision estimates than is mandated by the PowerPC
25318 ABI. Selecting @option{-mcpu=power6}, @option{-mcpu=power7} or
25319 @option{-mcpu=power8} automatically selects @option{-mrecip-precision}.
25320 The double-precision square root estimate instructions are not generated by
25321 default on low-precision machines, since they do not provide an
25322 estimate that converges after three steps.
25324 @item -mveclibabi=@var{type}
25325 @opindex mveclibabi
25326 Specifies the ABI type to use for vectorizing intrinsics using an
25327 external library. The only type supported at present is @samp{mass},
25328 which specifies to use IBM's Mathematical Acceleration Subsystem
25329 (MASS) libraries for vectorizing intrinsics using external libraries.
25330 GCC currently emits calls to @code{acosd2}, @code{acosf4},
25331 @code{acoshd2}, @code{acoshf4}, @code{asind2}, @code{asinf4},
25332 @code{asinhd2}, @code{asinhf4}, @code{atan2d2}, @code{atan2f4},
25333 @code{atand2}, @code{atanf4}, @code{atanhd2}, @code{atanhf4},
25334 @code{cbrtd2}, @code{cbrtf4}, @code{cosd2}, @code{cosf4},
25335 @code{coshd2}, @code{coshf4}, @code{erfcd2}, @code{erfcf4},
25336 @code{erfd2}, @code{erff4}, @code{exp2d2}, @code{exp2f4},
25337 @code{expd2}, @code{expf4}, @code{expm1d2}, @code{expm1f4},
25338 @code{hypotd2}, @code{hypotf4}, @code{lgammad2}, @code{lgammaf4},
25339 @code{log10d2}, @code{log10f4}, @code{log1pd2}, @code{log1pf4},
25340 @code{log2d2}, @code{log2f4}, @code{logd2}, @code{logf4},
25341 @code{powd2}, @code{powf4}, @code{sind2}, @code{sinf4}, @code{sinhd2},
25342 @code{sinhf4}, @code{sqrtd2}, @code{sqrtf4}, @code{tand2},
25343 @code{tanf4}, @code{tanhd2}, and @code{tanhf4} when generating code
25344 for power7. Both @option{-ftree-vectorize} and
25345 @option{-funsafe-math-optimizations} must also be enabled. The MASS
25346 libraries must be specified at link time.
25351 Generate (do not generate) the @code{friz} instruction when the
25352 @option{-funsafe-math-optimizations} option is used to optimize
25353 rounding of floating-point values to 64-bit integer and back to floating
25354 point. The @code{friz} instruction does not return the same value if
25355 the floating-point number is too large to fit in an integer.
25357 @item -mpointers-to-nested-functions
25358 @itemx -mno-pointers-to-nested-functions
25359 @opindex mpointers-to-nested-functions
25360 Generate (do not generate) code to load up the static chain register
25361 (@code{r11}) when calling through a pointer on AIX and 64-bit Linux
25362 systems where a function pointer points to a 3-word descriptor giving
25363 the function address, TOC value to be loaded in register @code{r2}, and
25364 static chain value to be loaded in register @code{r11}. The
25365 @option{-mpointers-to-nested-functions} is on by default. You cannot
25366 call through pointers to nested functions or pointers
25367 to functions compiled in other languages that use the static chain if
25368 you use @option{-mno-pointers-to-nested-functions}.
25370 @item -msave-toc-indirect
25371 @itemx -mno-save-toc-indirect
25372 @opindex msave-toc-indirect
25373 Generate (do not generate) code to save the TOC value in the reserved
25374 stack location in the function prologue if the function calls through
25375 a pointer on AIX and 64-bit Linux systems. If the TOC value is not
25376 saved in the prologue, it is saved just before the call through the
25377 pointer. The @option{-mno-save-toc-indirect} option is the default.
25379 @item -mcompat-align-parm
25380 @itemx -mno-compat-align-parm
25381 @opindex mcompat-align-parm
25382 Generate (do not generate) code to pass structure parameters with a
25383 maximum alignment of 64 bits, for compatibility with older versions
25386 Older versions of GCC (prior to 4.9.0) incorrectly did not align a
25387 structure parameter on a 128-bit boundary when that structure contained
25388 a member requiring 128-bit alignment. This is corrected in more
25389 recent versions of GCC. This option may be used to generate code
25390 that is compatible with functions compiled with older versions of
25393 The @option{-mno-compat-align-parm} option is the default.
25395 @item -mstack-protector-guard=@var{guard}
25396 @itemx -mstack-protector-guard-reg=@var{reg}
25397 @itemx -mstack-protector-guard-offset=@var{offset}
25398 @itemx -mstack-protector-guard-symbol=@var{symbol}
25399 @opindex mstack-protector-guard
25400 @opindex mstack-protector-guard-reg
25401 @opindex mstack-protector-guard-offset
25402 @opindex mstack-protector-guard-symbol
25403 Generate stack protection code using canary at @var{guard}. Supported
25404 locations are @samp{global} for global canary or @samp{tls} for per-thread
25405 canary in the TLS block (the default with GNU libc version 2.4 or later).
25407 With the latter choice the options
25408 @option{-mstack-protector-guard-reg=@var{reg}} and
25409 @option{-mstack-protector-guard-offset=@var{offset}} furthermore specify
25410 which register to use as base register for reading the canary, and from what
25411 offset from that base register. The default for those is as specified in the
25412 relevant ABI. @option{-mstack-protector-guard-symbol=@var{symbol}} overrides
25413 the offset with a symbol reference to a canary in the TLS block.
25419 Generate (do not generate) pc-relative addressing when the option
25420 @option{-mcpu=future} is used.
25424 @subsection RX Options
25427 These command-line options are defined for RX targets:
25430 @item -m64bit-doubles
25431 @itemx -m32bit-doubles
25432 @opindex m64bit-doubles
25433 @opindex m32bit-doubles
25434 Make the @code{double} data type be 64 bits (@option{-m64bit-doubles})
25435 or 32 bits (@option{-m32bit-doubles}) in size. The default is
25436 @option{-m32bit-doubles}. @emph{Note} RX floating-point hardware only
25437 works on 32-bit values, which is why the default is
25438 @option{-m32bit-doubles}.
25444 Enables (@option{-fpu}) or disables (@option{-nofpu}) the use of RX
25445 floating-point hardware. The default is enabled for the RX600
25446 series and disabled for the RX200 series.
25448 Floating-point instructions are only generated for 32-bit floating-point
25449 values, however, so the FPU hardware is not used for doubles if the
25450 @option{-m64bit-doubles} option is used.
25452 @emph{Note} If the @option{-fpu} option is enabled then
25453 @option{-funsafe-math-optimizations} is also enabled automatically.
25454 This is because the RX FPU instructions are themselves unsafe.
25456 @item -mcpu=@var{name}
25458 Selects the type of RX CPU to be targeted. Currently three types are
25459 supported, the generic @samp{RX600} and @samp{RX200} series hardware and
25460 the specific @samp{RX610} CPU. The default is @samp{RX600}.
25462 The only difference between @samp{RX600} and @samp{RX610} is that the
25463 @samp{RX610} does not support the @code{MVTIPL} instruction.
25465 The @samp{RX200} series does not have a hardware floating-point unit
25466 and so @option{-nofpu} is enabled by default when this type is
25469 @item -mbig-endian-data
25470 @itemx -mlittle-endian-data
25471 @opindex mbig-endian-data
25472 @opindex mlittle-endian-data
25473 Store data (but not code) in the big-endian format. The default is
25474 @option{-mlittle-endian-data}, i.e.@: to store data in the little-endian
25477 @item -msmall-data-limit=@var{N}
25478 @opindex msmall-data-limit
25479 Specifies the maximum size in bytes of global and static variables
25480 which can be placed into the small data area. Using the small data
25481 area can lead to smaller and faster code, but the size of area is
25482 limited and it is up to the programmer to ensure that the area does
25483 not overflow. Also when the small data area is used one of the RX's
25484 registers (usually @code{r13}) is reserved for use pointing to this
25485 area, so it is no longer available for use by the compiler. This
25486 could result in slower and/or larger code if variables are pushed onto
25487 the stack instead of being held in this register.
25489 Note, common variables (variables that have not been initialized) and
25490 constants are not placed into the small data area as they are assigned
25491 to other sections in the output executable.
25493 The default value is zero, which disables this feature. Note, this
25494 feature is not enabled by default with higher optimization levels
25495 (@option{-O2} etc) because of the potentially detrimental effects of
25496 reserving a register. It is up to the programmer to experiment and
25497 discover whether this feature is of benefit to their program. See the
25498 description of the @option{-mpid} option for a description of how the
25499 actual register to hold the small data area pointer is chosen.
25505 Use the simulator runtime. The default is to use the libgloss
25506 board-specific runtime.
25508 @item -mas100-syntax
25509 @itemx -mno-as100-syntax
25510 @opindex mas100-syntax
25511 @opindex mno-as100-syntax
25512 When generating assembler output use a syntax that is compatible with
25513 Renesas's AS100 assembler. This syntax can also be handled by the GAS
25514 assembler, but it has some restrictions so it is not generated by default.
25516 @item -mmax-constant-size=@var{N}
25517 @opindex mmax-constant-size
25518 Specifies the maximum size, in bytes, of a constant that can be used as
25519 an operand in a RX instruction. Although the RX instruction set does
25520 allow constants of up to 4 bytes in length to be used in instructions,
25521 a longer value equates to a longer instruction. Thus in some
25522 circumstances it can be beneficial to restrict the size of constants
25523 that are used in instructions. Constants that are too big are instead
25524 placed into a constant pool and referenced via register indirection.
25526 The value @var{N} can be between 0 and 4. A value of 0 (the default)
25527 or 4 means that constants of any size are allowed.
25531 Enable linker relaxation. Linker relaxation is a process whereby the
25532 linker attempts to reduce the size of a program by finding shorter
25533 versions of various instructions. Disabled by default.
25535 @item -mint-register=@var{N}
25536 @opindex mint-register
25537 Specify the number of registers to reserve for fast interrupt handler
25538 functions. The value @var{N} can be between 0 and 4. A value of 1
25539 means that register @code{r13} is reserved for the exclusive use
25540 of fast interrupt handlers. A value of 2 reserves @code{r13} and
25541 @code{r12}. A value of 3 reserves @code{r13}, @code{r12} and
25542 @code{r11}, and a value of 4 reserves @code{r13} through @code{r10}.
25543 A value of 0, the default, does not reserve any registers.
25545 @item -msave-acc-in-interrupts
25546 @opindex msave-acc-in-interrupts
25547 Specifies that interrupt handler functions should preserve the
25548 accumulator register. This is only necessary if normal code might use
25549 the accumulator register, for example because it performs 64-bit
25550 multiplications. The default is to ignore the accumulator as this
25551 makes the interrupt handlers faster.
25557 Enables the generation of position independent data. When enabled any
25558 access to constant data is done via an offset from a base address
25559 held in a register. This allows the location of constant data to be
25560 determined at run time without requiring the executable to be
25561 relocated, which is a benefit to embedded applications with tight
25562 memory constraints. Data that can be modified is not affected by this
25565 Note, using this feature reserves a register, usually @code{r13}, for
25566 the constant data base address. This can result in slower and/or
25567 larger code, especially in complicated functions.
25569 The actual register chosen to hold the constant data base address
25570 depends upon whether the @option{-msmall-data-limit} and/or the
25571 @option{-mint-register} command-line options are enabled. Starting
25572 with register @code{r13} and proceeding downwards, registers are
25573 allocated first to satisfy the requirements of @option{-mint-register},
25574 then @option{-mpid} and finally @option{-msmall-data-limit}. Thus it
25575 is possible for the small data area register to be @code{r8} if both
25576 @option{-mint-register=4} and @option{-mpid} are specified on the
25579 By default this feature is not enabled. The default can be restored
25580 via the @option{-mno-pid} command-line option.
25582 @item -mno-warn-multiple-fast-interrupts
25583 @itemx -mwarn-multiple-fast-interrupts
25584 @opindex mno-warn-multiple-fast-interrupts
25585 @opindex mwarn-multiple-fast-interrupts
25586 Prevents GCC from issuing a warning message if it finds more than one
25587 fast interrupt handler when it is compiling a file. The default is to
25588 issue a warning for each extra fast interrupt handler found, as the RX
25589 only supports one such interrupt.
25591 @item -mallow-string-insns
25592 @itemx -mno-allow-string-insns
25593 @opindex mallow-string-insns
25594 @opindex mno-allow-string-insns
25595 Enables or disables the use of the string manipulation instructions
25596 @code{SMOVF}, @code{SCMPU}, @code{SMOVB}, @code{SMOVU}, @code{SUNTIL}
25597 @code{SWHILE} and also the @code{RMPA} instruction. These
25598 instructions may prefetch data, which is not safe to do if accessing
25599 an I/O register. (See section 12.2.7 of the RX62N Group User's Manual
25600 for more information).
25602 The default is to allow these instructions, but it is not possible for
25603 GCC to reliably detect all circumstances where a string instruction
25604 might be used to access an I/O register, so their use cannot be
25605 disabled automatically. Instead it is reliant upon the programmer to
25606 use the @option{-mno-allow-string-insns} option if their program
25607 accesses I/O space.
25609 When the instructions are enabled GCC defines the C preprocessor
25610 symbol @code{__RX_ALLOW_STRING_INSNS__}, otherwise it defines the
25611 symbol @code{__RX_DISALLOW_STRING_INSNS__}.
25617 Use only (or not only) @code{JSR} instructions to access functions.
25618 This option can be used when code size exceeds the range of @code{BSR}
25619 instructions. Note that @option{-mno-jsr} does not mean to not use
25620 @code{JSR} but instead means that any type of branch may be used.
25623 @emph{Note:} The generic GCC command-line option @option{-ffixed-@var{reg}}
25624 has special significance to the RX port when used with the
25625 @code{interrupt} function attribute. This attribute indicates a
25626 function intended to process fast interrupts. GCC ensures
25627 that it only uses the registers @code{r10}, @code{r11}, @code{r12}
25628 and/or @code{r13} and only provided that the normal use of the
25629 corresponding registers have been restricted via the
25630 @option{-ffixed-@var{reg}} or @option{-mint-register} command-line
25633 @node S/390 and zSeries Options
25634 @subsection S/390 and zSeries Options
25635 @cindex S/390 and zSeries Options
25637 These are the @samp{-m} options defined for the S/390 and zSeries architecture.
25641 @itemx -msoft-float
25642 @opindex mhard-float
25643 @opindex msoft-float
25644 Use (do not use) the hardware floating-point instructions and registers
25645 for floating-point operations. When @option{-msoft-float} is specified,
25646 functions in @file{libgcc.a} are used to perform floating-point
25647 operations. When @option{-mhard-float} is specified, the compiler
25648 generates IEEE floating-point instructions. This is the default.
25651 @itemx -mno-hard-dfp
25653 @opindex mno-hard-dfp
25654 Use (do not use) the hardware decimal-floating-point instructions for
25655 decimal-floating-point operations. When @option{-mno-hard-dfp} is
25656 specified, functions in @file{libgcc.a} are used to perform
25657 decimal-floating-point operations. When @option{-mhard-dfp} is
25658 specified, the compiler generates decimal-floating-point hardware
25659 instructions. This is the default for @option{-march=z9-ec} or higher.
25661 @item -mlong-double-64
25662 @itemx -mlong-double-128
25663 @opindex mlong-double-64
25664 @opindex mlong-double-128
25665 These switches control the size of @code{long double} type. A size
25666 of 64 bits makes the @code{long double} type equivalent to the @code{double}
25667 type. This is the default.
25670 @itemx -mno-backchain
25671 @opindex mbackchain
25672 @opindex mno-backchain
25673 Store (do not store) the address of the caller's frame as backchain pointer
25674 into the callee's stack frame.
25675 A backchain may be needed to allow debugging using tools that do not understand
25676 DWARF call frame information.
25677 When @option{-mno-packed-stack} is in effect, the backchain pointer is stored
25678 at the bottom of the stack frame; when @option{-mpacked-stack} is in effect,
25679 the backchain is placed into the topmost word of the 96/160 byte register
25682 In general, code compiled with @option{-mbackchain} is call-compatible with
25683 code compiled with @option{-mmo-backchain}; however, use of the backchain
25684 for debugging purposes usually requires that the whole binary is built with
25685 @option{-mbackchain}. Note that the combination of @option{-mbackchain},
25686 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
25687 to build a linux kernel use @option{-msoft-float}.
25689 The default is to not maintain the backchain.
25691 @item -mpacked-stack
25692 @itemx -mno-packed-stack
25693 @opindex mpacked-stack
25694 @opindex mno-packed-stack
25695 Use (do not use) the packed stack layout. When @option{-mno-packed-stack} is
25696 specified, the compiler uses the all fields of the 96/160 byte register save
25697 area only for their default purpose; unused fields still take up stack space.
25698 When @option{-mpacked-stack} is specified, register save slots are densely
25699 packed at the top of the register save area; unused space is reused for other
25700 purposes, allowing for more efficient use of the available stack space.
25701 However, when @option{-mbackchain} is also in effect, the topmost word of
25702 the save area is always used to store the backchain, and the return address
25703 register is always saved two words below the backchain.
25705 As long as the stack frame backchain is not used, code generated with
25706 @option{-mpacked-stack} is call-compatible with code generated with
25707 @option{-mno-packed-stack}. Note that some non-FSF releases of GCC 2.95 for
25708 S/390 or zSeries generated code that uses the stack frame backchain at run
25709 time, not just for debugging purposes. Such code is not call-compatible
25710 with code compiled with @option{-mpacked-stack}. Also, note that the
25711 combination of @option{-mbackchain},
25712 @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order
25713 to build a linux kernel use @option{-msoft-float}.
25715 The default is to not use the packed stack layout.
25718 @itemx -mno-small-exec
25719 @opindex msmall-exec
25720 @opindex mno-small-exec
25721 Generate (or do not generate) code using the @code{bras} instruction
25722 to do subroutine calls.
25723 This only works reliably if the total executable size does not
25724 exceed 64k. The default is to use the @code{basr} instruction instead,
25725 which does not have this limitation.
25731 When @option{-m31} is specified, generate code compliant to the
25732 GNU/Linux for S/390 ABI@. When @option{-m64} is specified, generate
25733 code compliant to the GNU/Linux for zSeries ABI@. This allows GCC in
25734 particular to generate 64-bit instructions. For the @samp{s390}
25735 targets, the default is @option{-m31}, while the @samp{s390x}
25736 targets default to @option{-m64}.
25742 When @option{-mzarch} is specified, generate code using the
25743 instructions available on z/Architecture.
25744 When @option{-mesa} is specified, generate code using the
25745 instructions available on ESA/390. Note that @option{-mesa} is
25746 not possible with @option{-m64}.
25747 When generating code compliant to the GNU/Linux for S/390 ABI,
25748 the default is @option{-mesa}. When generating code compliant
25749 to the GNU/Linux for zSeries ABI, the default is @option{-mzarch}.
25755 The @option{-mhtm} option enables a set of builtins making use of
25756 instructions available with the transactional execution facility
25757 introduced with the IBM zEnterprise EC12 machine generation
25758 @ref{S/390 System z Built-in Functions}.
25759 @option{-mhtm} is enabled by default when using @option{-march=zEC12}.
25765 When @option{-mvx} is specified, generate code using the instructions
25766 available with the vector extension facility introduced with the IBM
25767 z13 machine generation.
25768 This option changes the ABI for some vector type values with regard to
25769 alignment and calling conventions. In case vector type values are
25770 being used in an ABI-relevant context a GAS @samp{.gnu_attribute}
25771 command will be added to mark the resulting binary with the ABI used.
25772 @option{-mvx} is enabled by default when using @option{-march=z13}.
25775 @itemx -mno-zvector
25777 @opindex mno-zvector
25778 The @option{-mzvector} option enables vector language extensions and
25779 builtins using instructions available with the vector extension
25780 facility introduced with the IBM z13 machine generation.
25781 This option adds support for @samp{vector} to be used as a keyword to
25782 define vector type variables and arguments. @samp{vector} is only
25783 available when GNU extensions are enabled. It will not be expanded
25784 when requesting strict standard compliance e.g.@: with @option{-std=c99}.
25785 In addition to the GCC low-level builtins @option{-mzvector} enables
25786 a set of builtins added for compatibility with AltiVec-style
25787 implementations like Power and Cell. In order to make use of these
25788 builtins the header file @file{vecintrin.h} needs to be included.
25789 @option{-mzvector} is disabled by default.
25795 Generate (or do not generate) code using the @code{mvcle} instruction
25796 to perform block moves. When @option{-mno-mvcle} is specified,
25797 use a @code{mvc} loop instead. This is the default unless optimizing for
25804 Print (or do not print) additional debug information when compiling.
25805 The default is to not print debug information.
25807 @item -march=@var{cpu-type}
25809 Generate code that runs on @var{cpu-type}, which is the name of a
25810 system representing a certain processor type. Possible values for
25811 @var{cpu-type} are @samp{z900}/@samp{arch5}, @samp{z990}/@samp{arch6},
25812 @samp{z9-109}, @samp{z9-ec}/@samp{arch7}, @samp{z10}/@samp{arch8},
25813 @samp{z196}/@samp{arch9}, @samp{zEC12}, @samp{z13}/@samp{arch11},
25814 @samp{z14}/@samp{arch12}, and @samp{native}.
25816 The default is @option{-march=z900}.
25818 Specifying @samp{native} as cpu type can be used to select the best
25819 architecture option for the host processor.
25820 @option{-march=native} has no effect if GCC does not recognize the
25823 @item -mtune=@var{cpu-type}
25825 Tune to @var{cpu-type} everything applicable about the generated code,
25826 except for the ABI and the set of available instructions.
25827 The list of @var{cpu-type} values is the same as for @option{-march}.
25828 The default is the value used for @option{-march}.
25831 @itemx -mno-tpf-trace
25832 @opindex mtpf-trace
25833 @opindex mno-tpf-trace
25834 Generate code that adds (does not add) in TPF OS specific branches to trace
25835 routines in the operating system. This option is off by default, even
25836 when compiling for the TPF OS@.
25839 @itemx -mno-fused-madd
25840 @opindex mfused-madd
25841 @opindex mno-fused-madd
25842 Generate code that uses (does not use) the floating-point multiply and
25843 accumulate instructions. These instructions are generated by default if
25844 hardware floating point is used.
25846 @item -mwarn-framesize=@var{framesize}
25847 @opindex mwarn-framesize
25848 Emit a warning if the current function exceeds the given frame size. Because
25849 this is a compile-time check it doesn't need to be a real problem when the program
25850 runs. It is intended to identify functions that most probably cause
25851 a stack overflow. It is useful to be used in an environment with limited stack
25852 size e.g.@: the linux kernel.
25854 @item -mwarn-dynamicstack
25855 @opindex mwarn-dynamicstack
25856 Emit a warning if the function calls @code{alloca} or uses dynamically-sized
25857 arrays. This is generally a bad idea with a limited stack size.
25859 @item -mstack-guard=@var{stack-guard}
25860 @itemx -mstack-size=@var{stack-size}
25861 @opindex mstack-guard
25862 @opindex mstack-size
25863 If these options are provided the S/390 back end emits additional instructions in
25864 the function prologue that trigger a trap if the stack size is @var{stack-guard}
25865 bytes above the @var{stack-size} (remember that the stack on S/390 grows downward).
25866 If the @var{stack-guard} option is omitted the smallest power of 2 larger than
25867 the frame size of the compiled function is chosen.
25868 These options are intended to be used to help debugging stack overflow problems.
25869 The additionally emitted code causes only little overhead and hence can also be
25870 used in production-like systems without greater performance degradation. The given
25871 values have to be exact powers of 2 and @var{stack-size} has to be greater than
25872 @var{stack-guard} without exceeding 64k.
25873 In order to be efficient the extra code makes the assumption that the stack starts
25874 at an address aligned to the value given by @var{stack-size}.
25875 The @var{stack-guard} option can only be used in conjunction with @var{stack-size}.
25877 @item -mhotpatch=@var{pre-halfwords},@var{post-halfwords}
25879 If the hotpatch option is enabled, a ``hot-patching'' function
25880 prologue is generated for all functions in the compilation unit.
25881 The funtion label is prepended with the given number of two-byte
25882 NOP instructions (@var{pre-halfwords}, maximum 1000000). After
25883 the label, 2 * @var{post-halfwords} bytes are appended, using the
25884 largest NOP like instructions the architecture allows (maximum
25887 If both arguments are zero, hotpatching is disabled.
25889 This option can be overridden for individual functions with the
25890 @code{hotpatch} attribute.
25893 @node Score Options
25894 @subsection Score Options
25895 @cindex Score Options
25897 These options are defined for Score implementations:
25902 Compile code for big-endian mode. This is the default.
25906 Compile code for little-endian mode.
25910 Disable generation of @code{bcnz} instructions.
25914 Enable generation of unaligned load and store instructions.
25918 Enable the use of multiply-accumulate instructions. Disabled by default.
25922 Specify the SCORE5 as the target architecture.
25926 Specify the SCORE5U of the target architecture.
25930 Specify the SCORE7 as the target architecture. This is the default.
25934 Specify the SCORE7D as the target architecture.
25938 @subsection SH Options
25940 These @samp{-m} options are defined for the SH implementations:
25945 Generate code for the SH1.
25949 Generate code for the SH2.
25952 Generate code for the SH2e.
25956 Generate code for the SH2a without FPU, or for a SH2a-FPU in such a way
25957 that the floating-point unit is not used.
25959 @item -m2a-single-only
25960 @opindex m2a-single-only
25961 Generate code for the SH2a-FPU, in such a way that no double-precision
25962 floating-point operations are used.
25965 @opindex m2a-single
25966 Generate code for the SH2a-FPU assuming the floating-point unit is in
25967 single-precision mode by default.
25971 Generate code for the SH2a-FPU assuming the floating-point unit is in
25972 double-precision mode by default.
25976 Generate code for the SH3.
25980 Generate code for the SH3e.
25984 Generate code for the SH4 without a floating-point unit.
25986 @item -m4-single-only
25987 @opindex m4-single-only
25988 Generate code for the SH4 with a floating-point unit that only
25989 supports single-precision arithmetic.
25993 Generate code for the SH4 assuming the floating-point unit is in
25994 single-precision mode by default.
25998 Generate code for the SH4.
26002 Generate code for SH4-100.
26004 @item -m4-100-nofpu
26005 @opindex m4-100-nofpu
26006 Generate code for SH4-100 in such a way that the
26007 floating-point unit is not used.
26009 @item -m4-100-single
26010 @opindex m4-100-single
26011 Generate code for SH4-100 assuming the floating-point unit is in
26012 single-precision mode by default.
26014 @item -m4-100-single-only
26015 @opindex m4-100-single-only
26016 Generate code for SH4-100 in such a way that no double-precision
26017 floating-point operations are used.
26021 Generate code for SH4-200.
26023 @item -m4-200-nofpu
26024 @opindex m4-200-nofpu
26025 Generate code for SH4-200 without in such a way that the
26026 floating-point unit is not used.
26028 @item -m4-200-single
26029 @opindex m4-200-single
26030 Generate code for SH4-200 assuming the floating-point unit is in
26031 single-precision mode by default.
26033 @item -m4-200-single-only
26034 @opindex m4-200-single-only
26035 Generate code for SH4-200 in such a way that no double-precision
26036 floating-point operations are used.
26040 Generate code for SH4-300.
26042 @item -m4-300-nofpu
26043 @opindex m4-300-nofpu
26044 Generate code for SH4-300 without in such a way that the
26045 floating-point unit is not used.
26047 @item -m4-300-single
26048 @opindex m4-300-single
26049 Generate code for SH4-300 in such a way that no double-precision
26050 floating-point operations are used.
26052 @item -m4-300-single-only
26053 @opindex m4-300-single-only
26054 Generate code for SH4-300 in such a way that no double-precision
26055 floating-point operations are used.
26059 Generate code for SH4-340 (no MMU, no FPU).
26063 Generate code for SH4-500 (no FPU). Passes @option{-isa=sh4-nofpu} to the
26068 Generate code for the SH4al-dsp, or for a SH4a in such a way that the
26069 floating-point unit is not used.
26071 @item -m4a-single-only
26072 @opindex m4a-single-only
26073 Generate code for the SH4a, in such a way that no double-precision
26074 floating-point operations are used.
26077 @opindex m4a-single
26078 Generate code for the SH4a assuming the floating-point unit is in
26079 single-precision mode by default.
26083 Generate code for the SH4a.
26087 Same as @option{-m4a-nofpu}, except that it implicitly passes
26088 @option{-dsp} to the assembler. GCC doesn't generate any DSP
26089 instructions at the moment.
26093 Compile code for the processor in big-endian mode.
26097 Compile code for the processor in little-endian mode.
26101 Align doubles at 64-bit boundaries. Note that this changes the calling
26102 conventions, and thus some functions from the standard C library do
26103 not work unless you recompile it first with @option{-mdalign}.
26107 Shorten some address references at link time, when possible; uses the
26108 linker option @option{-relax}.
26112 Use 32-bit offsets in @code{switch} tables. The default is to use
26117 Enable the use of bit manipulation instructions on SH2A.
26121 Enable the use of the instruction @code{fmovd}. Check @option{-mdalign} for
26122 alignment constraints.
26126 Comply with the calling conventions defined by Renesas.
26129 @opindex mno-renesas
26130 Comply with the calling conventions defined for GCC before the Renesas
26131 conventions were available. This option is the default for all
26132 targets of the SH toolchain.
26135 @opindex mnomacsave
26136 Mark the @code{MAC} register as call-clobbered, even if
26137 @option{-mrenesas} is given.
26143 Control the IEEE compliance of floating-point comparisons, which affects the
26144 handling of cases where the result of a comparison is unordered. By default
26145 @option{-mieee} is implicitly enabled. If @option{-ffinite-math-only} is
26146 enabled @option{-mno-ieee} is implicitly set, which results in faster
26147 floating-point greater-equal and less-equal comparisons. The implicit settings
26148 can be overridden by specifying either @option{-mieee} or @option{-mno-ieee}.
26150 @item -minline-ic_invalidate
26151 @opindex minline-ic_invalidate
26152 Inline code to invalidate instruction cache entries after setting up
26153 nested function trampolines.
26154 This option has no effect if @option{-musermode} is in effect and the selected
26155 code generation option (e.g.@: @option{-m4}) does not allow the use of the @code{icbi}
26157 If the selected code generation option does not allow the use of the @code{icbi}
26158 instruction, and @option{-musermode} is not in effect, the inlined code
26159 manipulates the instruction cache address array directly with an associative
26160 write. This not only requires privileged mode at run time, but it also
26161 fails if the cache line had been mapped via the TLB and has become unmapped.
26165 Dump instruction size and location in the assembly code.
26168 @opindex mpadstruct
26169 This option is deprecated. It pads structures to multiple of 4 bytes,
26170 which is incompatible with the SH ABI@.
26172 @item -matomic-model=@var{model}
26173 @opindex matomic-model=@var{model}
26174 Sets the model of atomic operations and additional parameters as a comma
26175 separated list. For details on the atomic built-in functions see
26176 @ref{__atomic Builtins}. The following models and parameters are supported:
26181 Disable compiler generated atomic sequences and emit library calls for atomic
26182 operations. This is the default if the target is not @code{sh*-*-linux*}.
26185 Generate GNU/Linux compatible gUSA software atomic sequences for the atomic
26186 built-in functions. The generated atomic sequences require additional support
26187 from the interrupt/exception handling code of the system and are only suitable
26188 for SH3* and SH4* single-core systems. This option is enabled by default when
26189 the target is @code{sh*-*-linux*} and SH3* or SH4*. When the target is SH4A,
26190 this option also partially utilizes the hardware atomic instructions
26191 @code{movli.l} and @code{movco.l} to create more efficient code, unless
26192 @samp{strict} is specified.
26195 Generate software atomic sequences that use a variable in the thread control
26196 block. This is a variation of the gUSA sequences which can also be used on
26197 SH1* and SH2* targets. The generated atomic sequences require additional
26198 support from the interrupt/exception handling code of the system and are only
26199 suitable for single-core systems. When using this model, the @samp{gbr-offset=}
26200 parameter has to be specified as well.
26203 Generate software atomic sequences that temporarily disable interrupts by
26204 setting @code{SR.IMASK = 1111}. This model works only when the program runs
26205 in privileged mode and is only suitable for single-core systems. Additional
26206 support from the interrupt/exception handling code of the system is not
26207 required. This model is enabled by default when the target is
26208 @code{sh*-*-linux*} and SH1* or SH2*.
26211 Generate hardware atomic sequences using the @code{movli.l} and @code{movco.l}
26212 instructions only. This is only available on SH4A and is suitable for
26213 multi-core systems. Since the hardware instructions support only 32 bit atomic
26214 variables access to 8 or 16 bit variables is emulated with 32 bit accesses.
26215 Code compiled with this option is also compatible with other software
26216 atomic model interrupt/exception handling systems if executed on an SH4A
26217 system. Additional support from the interrupt/exception handling code of the
26218 system is not required for this model.
26221 This parameter specifies the offset in bytes of the variable in the thread
26222 control block structure that should be used by the generated atomic sequences
26223 when the @samp{soft-tcb} model has been selected. For other models this
26224 parameter is ignored. The specified value must be an integer multiple of four
26225 and in the range 0-1020.
26228 This parameter prevents mixed usage of multiple atomic models, even if they
26229 are compatible, and makes the compiler generate atomic sequences of the
26230 specified model only.
26236 Generate the @code{tas.b} opcode for @code{__atomic_test_and_set}.
26237 Notice that depending on the particular hardware and software configuration
26238 this can degrade overall performance due to the operand cache line flushes
26239 that are implied by the @code{tas.b} instruction. On multi-core SH4A
26240 processors the @code{tas.b} instruction must be used with caution since it
26241 can result in data corruption for certain cache configurations.
26244 @opindex mprefergot
26245 When generating position-independent code, emit function calls using
26246 the Global Offset Table instead of the Procedure Linkage Table.
26249 @itemx -mno-usermode
26251 @opindex mno-usermode
26252 Don't allow (allow) the compiler generating privileged mode code. Specifying
26253 @option{-musermode} also implies @option{-mno-inline-ic_invalidate} if the
26254 inlined code would not work in user mode. @option{-musermode} is the default
26255 when the target is @code{sh*-*-linux*}. If the target is SH1* or SH2*
26256 @option{-musermode} has no effect, since there is no user mode.
26258 @item -multcost=@var{number}
26259 @opindex multcost=@var{number}
26260 Set the cost to assume for a multiply insn.
26262 @item -mdiv=@var{strategy}
26263 @opindex mdiv=@var{strategy}
26264 Set the division strategy to be used for integer division operations.
26265 @var{strategy} can be one of:
26270 Calls a library function that uses the single-step division instruction
26271 @code{div1} to perform the operation. Division by zero calculates an
26272 unspecified result and does not trap. This is the default except for SH4,
26273 SH2A and SHcompact.
26276 Calls a library function that performs the operation in double precision
26277 floating point. Division by zero causes a floating-point exception. This is
26278 the default for SHcompact with FPU. Specifying this for targets that do not
26279 have a double precision FPU defaults to @code{call-div1}.
26282 Calls a library function that uses a lookup table for small divisors and
26283 the @code{div1} instruction with case distinction for larger divisors. Division
26284 by zero calculates an unspecified result and does not trap. This is the default
26285 for SH4. Specifying this for targets that do not have dynamic shift
26286 instructions defaults to @code{call-div1}.
26290 When a division strategy has not been specified the default strategy is
26291 selected based on the current target. For SH2A the default strategy is to
26292 use the @code{divs} and @code{divu} instructions instead of library function
26295 @item -maccumulate-outgoing-args
26296 @opindex maccumulate-outgoing-args
26297 Reserve space once for outgoing arguments in the function prologue rather
26298 than around each call. Generally beneficial for performance and size. Also
26299 needed for unwinding to avoid changing the stack frame around conditional code.
26301 @item -mdivsi3_libfunc=@var{name}
26302 @opindex mdivsi3_libfunc=@var{name}
26303 Set the name of the library function used for 32-bit signed division to
26305 This only affects the name used in the @samp{call} division strategies, and
26306 the compiler still expects the same sets of input/output/clobbered registers as
26307 if this option were not present.
26309 @item -mfixed-range=@var{register-range}
26310 @opindex mfixed-range
26311 Generate code treating the given register range as fixed registers.
26312 A fixed register is one that the register allocator cannot use. This is
26313 useful when compiling kernel code. A register range is specified as
26314 two registers separated by a dash. Multiple register ranges can be
26315 specified separated by a comma.
26317 @item -mbranch-cost=@var{num}
26318 @opindex mbranch-cost=@var{num}
26319 Assume @var{num} to be the cost for a branch instruction. Higher numbers
26320 make the compiler try to generate more branch-free code if possible.
26321 If not specified the value is selected depending on the processor type that
26322 is being compiled for.
26325 @itemx -mno-zdcbranch
26326 @opindex mzdcbranch
26327 @opindex mno-zdcbranch
26328 Assume (do not assume) that zero displacement conditional branch instructions
26329 @code{bt} and @code{bf} are fast. If @option{-mzdcbranch} is specified, the
26330 compiler prefers zero displacement branch code sequences. This is
26331 enabled by default when generating code for SH4 and SH4A. It can be explicitly
26332 disabled by specifying @option{-mno-zdcbranch}.
26334 @item -mcbranch-force-delay-slot
26335 @opindex mcbranch-force-delay-slot
26336 Force the usage of delay slots for conditional branches, which stuffs the delay
26337 slot with a @code{nop} if a suitable instruction cannot be found. By default
26338 this option is disabled. It can be enabled to work around hardware bugs as
26339 found in the original SH7055.
26342 @itemx -mno-fused-madd
26343 @opindex mfused-madd
26344 @opindex mno-fused-madd
26345 Generate code that uses (does not use) the floating-point multiply and
26346 accumulate instructions. These instructions are generated by default
26347 if hardware floating point is used. The machine-dependent
26348 @option{-mfused-madd} option is now mapped to the machine-independent
26349 @option{-ffp-contract=fast} option, and @option{-mno-fused-madd} is
26350 mapped to @option{-ffp-contract=off}.
26356 Allow or disallow the compiler to emit the @code{fsca} instruction for sine
26357 and cosine approximations. The option @option{-mfsca} must be used in
26358 combination with @option{-funsafe-math-optimizations}. It is enabled by default
26359 when generating code for SH4A. Using @option{-mno-fsca} disables sine and cosine
26360 approximations even if @option{-funsafe-math-optimizations} is in effect.
26366 Allow or disallow the compiler to emit the @code{fsrra} instruction for
26367 reciprocal square root approximations. The option @option{-mfsrra} must be used
26368 in combination with @option{-funsafe-math-optimizations} and
26369 @option{-ffinite-math-only}. It is enabled by default when generating code for
26370 SH4A. Using @option{-mno-fsrra} disables reciprocal square root approximations
26371 even if @option{-funsafe-math-optimizations} and @option{-ffinite-math-only} are
26374 @item -mpretend-cmove
26375 @opindex mpretend-cmove
26376 Prefer zero-displacement conditional branches for conditional move instruction
26377 patterns. This can result in faster code on the SH4 processor.
26381 Generate code using the FDPIC ABI.
26385 @node Solaris 2 Options
26386 @subsection Solaris 2 Options
26387 @cindex Solaris 2 options
26389 These @samp{-m} options are supported on Solaris 2:
26392 @item -mclear-hwcap
26393 @opindex mclear-hwcap
26394 @option{-mclear-hwcap} tells the compiler to remove the hardware
26395 capabilities generated by the Solaris assembler. This is only necessary
26396 when object files use ISA extensions not supported by the current
26397 machine, but check at runtime whether or not to use them.
26399 @item -mimpure-text
26400 @opindex mimpure-text
26401 @option{-mimpure-text}, used in addition to @option{-shared}, tells
26402 the compiler to not pass @option{-z text} to the linker when linking a
26403 shared object. Using this option, you can link position-dependent
26404 code into a shared object.
26406 @option{-mimpure-text} suppresses the ``relocations remain against
26407 allocatable but non-writable sections'' linker error message.
26408 However, the necessary relocations trigger copy-on-write, and the
26409 shared object is not actually shared across processes. Instead of
26410 using @option{-mimpure-text}, you should compile all source code with
26411 @option{-fpic} or @option{-fPIC}.
26415 These switches are supported in addition to the above on Solaris 2:
26420 This is a synonym for @option{-pthread}.
26423 @node SPARC Options
26424 @subsection SPARC Options
26425 @cindex SPARC options
26427 These @samp{-m} options are supported on the SPARC:
26430 @item -mno-app-regs
26432 @opindex mno-app-regs
26434 Specify @option{-mapp-regs} to generate output using the global registers
26435 2 through 4, which the SPARC SVR4 ABI reserves for applications. Like the
26436 global register 1, each global register 2 through 4 is then treated as an
26437 allocable register that is clobbered by function calls. This is the default.
26439 To be fully SVR4 ABI-compliant at the cost of some performance loss,
26440 specify @option{-mno-app-regs}. You should compile libraries and system
26441 software with this option.
26447 With @option{-mflat}, the compiler does not generate save/restore instructions
26448 and uses a ``flat'' or single register window model. This model is compatible
26449 with the regular register window model. The local registers and the input
26450 registers (0--5) are still treated as ``call-saved'' registers and are
26451 saved on the stack as needed.
26453 With @option{-mno-flat} (the default), the compiler generates save/restore
26454 instructions (except for leaf functions). This is the normal operating mode.
26457 @itemx -mhard-float
26459 @opindex mhard-float
26460 Generate output containing floating-point instructions. This is the
26464 @itemx -msoft-float
26466 @opindex msoft-float
26467 Generate output containing library calls for floating point.
26468 @strong{Warning:} the requisite libraries are not available for all SPARC
26469 targets. Normally the facilities of the machine's usual C compiler are
26470 used, but this cannot be done directly in cross-compilation. You must make
26471 your own arrangements to provide suitable library functions for
26472 cross-compilation. The embedded targets @samp{sparc-*-aout} and
26473 @samp{sparclite-*-*} do provide software floating-point support.
26475 @option{-msoft-float} changes the calling convention in the output file;
26476 therefore, it is only useful if you compile @emph{all} of a program with
26477 this option. In particular, you need to compile @file{libgcc.a}, the
26478 library that comes with GCC, with @option{-msoft-float} in order for
26481 @item -mhard-quad-float
26482 @opindex mhard-quad-float
26483 Generate output containing quad-word (long double) floating-point
26486 @item -msoft-quad-float
26487 @opindex msoft-quad-float
26488 Generate output containing library calls for quad-word (long double)
26489 floating-point instructions. The functions called are those specified
26490 in the SPARC ABI@. This is the default.
26492 As of this writing, there are no SPARC implementations that have hardware
26493 support for the quad-word floating-point instructions. They all invoke
26494 a trap handler for one of these instructions, and then the trap handler
26495 emulates the effect of the instruction. Because of the trap handler overhead,
26496 this is much slower than calling the ABI library routines. Thus the
26497 @option{-msoft-quad-float} option is the default.
26499 @item -mno-unaligned-doubles
26500 @itemx -munaligned-doubles
26501 @opindex mno-unaligned-doubles
26502 @opindex munaligned-doubles
26503 Assume that doubles have 8-byte alignment. This is the default.
26505 With @option{-munaligned-doubles}, GCC assumes that doubles have 8-byte
26506 alignment only if they are contained in another type, or if they have an
26507 absolute address. Otherwise, it assumes they have 4-byte alignment.
26508 Specifying this option avoids some rare compatibility problems with code
26509 generated by other compilers. It is not the default because it results
26510 in a performance loss, especially for floating-point code.
26513 @itemx -mno-user-mode
26514 @opindex muser-mode
26515 @opindex mno-user-mode
26516 Do not generate code that can only run in supervisor mode. This is relevant
26517 only for the @code{casa} instruction emitted for the LEON3 processor. This
26520 @item -mfaster-structs
26521 @itemx -mno-faster-structs
26522 @opindex mfaster-structs
26523 @opindex mno-faster-structs
26524 With @option{-mfaster-structs}, the compiler assumes that structures
26525 should have 8-byte alignment. This enables the use of pairs of
26526 @code{ldd} and @code{std} instructions for copies in structure
26527 assignment, in place of twice as many @code{ld} and @code{st} pairs.
26528 However, the use of this changed alignment directly violates the SPARC
26529 ABI@. Thus, it's intended only for use on targets where the developer
26530 acknowledges that their resulting code is not directly in line with
26531 the rules of the ABI@.
26533 @item -mstd-struct-return
26534 @itemx -mno-std-struct-return
26535 @opindex mstd-struct-return
26536 @opindex mno-std-struct-return
26537 With @option{-mstd-struct-return}, the compiler generates checking code
26538 in functions returning structures or unions to detect size mismatches
26539 between the two sides of function calls, as per the 32-bit ABI@.
26541 The default is @option{-mno-std-struct-return}. This option has no effect
26548 Enable Local Register Allocation. This is the default for SPARC since GCC 7
26549 so @option{-mno-lra} needs to be passed to get old Reload.
26551 @item -mcpu=@var{cpu_type}
26553 Set the instruction set, register set, and instruction scheduling parameters
26554 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
26555 @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{hypersparc},
26556 @samp{leon}, @samp{leon3}, @samp{leon3v7}, @samp{sparclite}, @samp{f930},
26557 @samp{f934}, @samp{sparclite86x}, @samp{sparclet}, @samp{tsc701}, @samp{v9},
26558 @samp{ultrasparc}, @samp{ultrasparc3}, @samp{niagara}, @samp{niagara2},
26559 @samp{niagara3}, @samp{niagara4}, @samp{niagara7} and @samp{m8}.
26561 Native Solaris and GNU/Linux toolchains also support the value @samp{native},
26562 which selects the best architecture option for the host processor.
26563 @option{-mcpu=native} has no effect if GCC does not recognize
26566 Default instruction scheduling parameters are used for values that select
26567 an architecture and not an implementation. These are @samp{v7}, @samp{v8},
26568 @samp{sparclite}, @samp{sparclet}, @samp{v9}.
26570 Here is a list of each supported architecture and their supported
26578 supersparc, hypersparc, leon, leon3
26581 f930, f934, sparclite86x
26587 ultrasparc, ultrasparc3, niagara, niagara2, niagara3, niagara4,
26591 By default (unless configured otherwise), GCC generates code for the V7
26592 variant of the SPARC architecture. With @option{-mcpu=cypress}, the compiler
26593 additionally optimizes it for the Cypress CY7C602 chip, as used in the
26594 SPARCStation/SPARCServer 3xx series. This is also appropriate for the older
26595 SPARCStation 1, 2, IPX etc.
26597 With @option{-mcpu=v8}, GCC generates code for the V8 variant of the SPARC
26598 architecture. The only difference from V7 code is that the compiler emits
26599 the integer multiply and integer divide instructions which exist in SPARC-V8
26600 but not in SPARC-V7. With @option{-mcpu=supersparc}, the compiler additionally
26601 optimizes it for the SuperSPARC chip, as used in the SPARCStation 10, 1000 and
26604 With @option{-mcpu=sparclite}, GCC generates code for the SPARClite variant of
26605 the SPARC architecture. This adds the integer multiply, integer divide step
26606 and scan (@code{ffs}) instructions which exist in SPARClite but not in SPARC-V7.
26607 With @option{-mcpu=f930}, the compiler additionally optimizes it for the
26608 Fujitsu MB86930 chip, which is the original SPARClite, with no FPU@. With
26609 @option{-mcpu=f934}, the compiler additionally optimizes it for the Fujitsu
26610 MB86934 chip, which is the more recent SPARClite with FPU@.
26612 With @option{-mcpu=sparclet}, GCC generates code for the SPARClet variant of
26613 the SPARC architecture. This adds the integer multiply, multiply/accumulate,
26614 integer divide step and scan (@code{ffs}) instructions which exist in SPARClet
26615 but not in SPARC-V7. With @option{-mcpu=tsc701}, the compiler additionally
26616 optimizes it for the TEMIC SPARClet chip.
26618 With @option{-mcpu=v9}, GCC generates code for the V9 variant of the SPARC
26619 architecture. This adds 64-bit integer and floating-point move instructions,
26620 3 additional floating-point condition code registers and conditional move
26621 instructions. With @option{-mcpu=ultrasparc}, the compiler additionally
26622 optimizes it for the Sun UltraSPARC I/II/IIi chips. With
26623 @option{-mcpu=ultrasparc3}, the compiler additionally optimizes it for the
26624 Sun UltraSPARC III/III+/IIIi/IIIi+/IV/IV+ chips. With
26625 @option{-mcpu=niagara}, the compiler additionally optimizes it for
26626 Sun UltraSPARC T1 chips. With @option{-mcpu=niagara2}, the compiler
26627 additionally optimizes it for Sun UltraSPARC T2 chips. With
26628 @option{-mcpu=niagara3}, the compiler additionally optimizes it for Sun
26629 UltraSPARC T3 chips. With @option{-mcpu=niagara4}, the compiler
26630 additionally optimizes it for Sun UltraSPARC T4 chips. With
26631 @option{-mcpu=niagara7}, the compiler additionally optimizes it for
26632 Oracle SPARC M7 chips. With @option{-mcpu=m8}, the compiler
26633 additionally optimizes it for Oracle M8 chips.
26635 @item -mtune=@var{cpu_type}
26637 Set the instruction scheduling parameters for machine type
26638 @var{cpu_type}, but do not set the instruction set or register set that the
26639 option @option{-mcpu=@var{cpu_type}} does.
26641 The same values for @option{-mcpu=@var{cpu_type}} can be used for
26642 @option{-mtune=@var{cpu_type}}, but the only useful values are those
26643 that select a particular CPU implementation. Those are
26644 @samp{cypress}, @samp{supersparc}, @samp{hypersparc}, @samp{leon},
26645 @samp{leon3}, @samp{leon3v7}, @samp{f930}, @samp{f934},
26646 @samp{sparclite86x}, @samp{tsc701}, @samp{ultrasparc},
26647 @samp{ultrasparc3}, @samp{niagara}, @samp{niagara2}, @samp{niagara3},
26648 @samp{niagara4}, @samp{niagara7} and @samp{m8}. With native Solaris
26649 and GNU/Linux toolchains, @samp{native} can also be used.
26654 @opindex mno-v8plus
26655 With @option{-mv8plus}, GCC generates code for the SPARC-V8+ ABI@. The
26656 difference from the V8 ABI is that the global and out registers are
26657 considered 64 bits wide. This is enabled by default on Solaris in 32-bit
26658 mode for all SPARC-V9 processors.
26664 With @option{-mvis}, GCC generates code that takes advantage of the UltraSPARC
26665 Visual Instruction Set extensions. The default is @option{-mno-vis}.
26671 With @option{-mvis2}, GCC generates code that takes advantage of
26672 version 2.0 of the UltraSPARC Visual Instruction Set extensions. The
26673 default is @option{-mvis2} when targeting a cpu that supports such
26674 instructions, such as UltraSPARC-III and later. Setting @option{-mvis2}
26675 also sets @option{-mvis}.
26681 With @option{-mvis3}, GCC generates code that takes advantage of
26682 version 3.0 of the UltraSPARC Visual Instruction Set extensions. The
26683 default is @option{-mvis3} when targeting a cpu that supports such
26684 instructions, such as niagara-3 and later. Setting @option{-mvis3}
26685 also sets @option{-mvis2} and @option{-mvis}.
26691 With @option{-mvis4}, GCC generates code that takes advantage of
26692 version 4.0 of the UltraSPARC Visual Instruction Set extensions. The
26693 default is @option{-mvis4} when targeting a cpu that supports such
26694 instructions, such as niagara-7 and later. Setting @option{-mvis4}
26695 also sets @option{-mvis3}, @option{-mvis2} and @option{-mvis}.
26701 With @option{-mvis4b}, GCC generates code that takes advantage of
26702 version 4.0 of the UltraSPARC Visual Instruction Set extensions, plus
26703 the additional VIS instructions introduced in the Oracle SPARC
26704 Architecture 2017. The default is @option{-mvis4b} when targeting a
26705 cpu that supports such instructions, such as m8 and later. Setting
26706 @option{-mvis4b} also sets @option{-mvis4}, @option{-mvis3},
26707 @option{-mvis2} and @option{-mvis}.
26712 @opindex mno-cbcond
26713 With @option{-mcbcond}, GCC generates code that takes advantage of the UltraSPARC
26714 Compare-and-Branch-on-Condition instructions. The default is @option{-mcbcond}
26715 when targeting a CPU that supports such instructions, such as Niagara-4 and
26722 With @option{-mfmaf}, GCC generates code that takes advantage of the UltraSPARC
26723 Fused Multiply-Add Floating-point instructions. The default is @option{-mfmaf}
26724 when targeting a CPU that supports such instructions, such as Niagara-3 and
26730 @opindex mno-fsmuld
26731 With @option{-mfsmuld}, GCC generates code that takes advantage of the
26732 Floating-point Multiply Single to Double (FsMULd) instruction. The default is
26733 @option{-mfsmuld} when targeting a CPU supporting the architecture versions V8
26734 or V9 with FPU except @option{-mcpu=leon}.
26740 With @option{-mpopc}, GCC generates code that takes advantage of the UltraSPARC
26741 Population Count instruction. The default is @option{-mpopc}
26742 when targeting a CPU that supports such an instruction, such as Niagara-2 and
26749 With @option{-msubxc}, GCC generates code that takes advantage of the UltraSPARC
26750 Subtract-Extended-with-Carry instruction. The default is @option{-msubxc}
26751 when targeting a CPU that supports such an instruction, such as Niagara-7 and
26755 @opindex mfix-at697f
26756 Enable the documented workaround for the single erratum of the Atmel AT697F
26757 processor (which corresponds to erratum #13 of the AT697E processor).
26760 @opindex mfix-ut699
26761 Enable the documented workarounds for the floating-point errata and the data
26762 cache nullify errata of the UT699 processor.
26765 @opindex mfix-ut700
26766 Enable the documented workaround for the back-to-back store errata of
26767 the UT699E/UT700 processor.
26769 @item -mfix-gr712rc
26770 @opindex mfix-gr712rc
26771 Enable the documented workaround for the back-to-back store errata of
26772 the GR712RC processor.
26775 These @samp{-m} options are supported in addition to the above
26776 on SPARC-V9 processors in 64-bit environments:
26783 Generate code for a 32-bit or 64-bit environment.
26784 The 32-bit environment sets int, long and pointer to 32 bits.
26785 The 64-bit environment sets int to 32 bits and long and pointer
26788 @item -mcmodel=@var{which}
26790 Set the code model to one of
26794 The Medium/Low code model: 64-bit addresses, programs
26795 must be linked in the low 32 bits of memory. Programs can be statically
26796 or dynamically linked.
26799 The Medium/Middle code model: 64-bit addresses, programs
26800 must be linked in the low 44 bits of memory, the text and data segments must
26801 be less than 2GB in size and the data segment must be located within 2GB of
26805 The Medium/Anywhere code model: 64-bit addresses, programs
26806 may be linked anywhere in memory, the text and data segments must be less
26807 than 2GB in size and the data segment must be located within 2GB of the
26811 The Medium/Anywhere code model for embedded systems:
26812 64-bit addresses, the text and data segments must be less than 2GB in
26813 size, both starting anywhere in memory (determined at link time). The
26814 global register %g4 points to the base of the data segment. Programs
26815 are statically linked and PIC is not supported.
26818 @item -mmemory-model=@var{mem-model}
26819 @opindex mmemory-model
26820 Set the memory model in force on the processor to one of
26824 The default memory model for the processor and operating system.
26827 Relaxed Memory Order
26830 Partial Store Order
26836 Sequential Consistency
26839 These memory models are formally defined in Appendix D of the SPARC-V9
26840 architecture manual, as set in the processor's @code{PSTATE.MM} field.
26843 @itemx -mno-stack-bias
26844 @opindex mstack-bias
26845 @opindex mno-stack-bias
26846 With @option{-mstack-bias}, GCC assumes that the stack pointer, and
26847 frame pointer if present, are offset by @minus{}2047 which must be added back
26848 when making stack frame references. This is the default in 64-bit mode.
26849 Otherwise, assume no such offset is present.
26852 @node System V Options
26853 @subsection Options for System V
26855 These additional options are available on System V Release 4 for
26856 compatibility with other compilers on those systems:
26861 Create a shared object.
26862 It is recommended that @option{-symbolic} or @option{-shared} be used instead.
26866 Identify the versions of each tool used by the compiler, in a
26867 @code{.ident} assembler directive in the output.
26871 Refrain from adding @code{.ident} directives to the output file (this is
26874 @item -YP,@var{dirs}
26876 Search the directories @var{dirs}, and no others, for libraries
26877 specified with @option{-l}.
26879 @item -Ym,@var{dir}
26881 Look in the directory @var{dir} to find the M4 preprocessor.
26882 The assembler uses this option.
26883 @c This is supposed to go with a -Yd for predefined M4 macro files, but
26884 @c the generic assembler that comes with Solaris takes just -Ym.
26887 @node TILE-Gx Options
26888 @subsection TILE-Gx Options
26889 @cindex TILE-Gx options
26891 These @samp{-m} options are supported on the TILE-Gx:
26894 @item -mcmodel=small
26895 @opindex mcmodel=small
26896 Generate code for the small model. The distance for direct calls is
26897 limited to 500M in either direction. PC-relative addresses are 32
26898 bits. Absolute addresses support the full address range.
26900 @item -mcmodel=large
26901 @opindex mcmodel=large
26902 Generate code for the large model. There is no limitation on call
26903 distance, pc-relative addresses, or absolute addresses.
26905 @item -mcpu=@var{name}
26907 Selects the type of CPU to be targeted. Currently the only supported
26908 type is @samp{tilegx}.
26914 Generate code for a 32-bit or 64-bit environment. The 32-bit
26915 environment sets int, long, and pointer to 32 bits. The 64-bit
26916 environment sets int to 32 bits and long and pointer to 64 bits.
26919 @itemx -mlittle-endian
26920 @opindex mbig-endian
26921 @opindex mlittle-endian
26922 Generate code in big/little endian mode, respectively.
26925 @node TILEPro Options
26926 @subsection TILEPro Options
26927 @cindex TILEPro options
26929 These @samp{-m} options are supported on the TILEPro:
26932 @item -mcpu=@var{name}
26934 Selects the type of CPU to be targeted. Currently the only supported
26935 type is @samp{tilepro}.
26939 Generate code for a 32-bit environment, which sets int, long, and
26940 pointer to 32 bits. This is the only supported behavior so the flag
26941 is essentially ignored.
26945 @subsection V850 Options
26946 @cindex V850 Options
26948 These @samp{-m} options are defined for V850 implementations:
26952 @itemx -mno-long-calls
26953 @opindex mlong-calls
26954 @opindex mno-long-calls
26955 Treat all calls as being far away (near). If calls are assumed to be
26956 far away, the compiler always loads the function's address into a
26957 register, and calls indirect through the pointer.
26963 Do not optimize (do optimize) basic blocks that use the same index
26964 pointer 4 or more times to copy pointer into the @code{ep} register, and
26965 use the shorter @code{sld} and @code{sst} instructions. The @option{-mep}
26966 option is on by default if you optimize.
26968 @item -mno-prolog-function
26969 @itemx -mprolog-function
26970 @opindex mno-prolog-function
26971 @opindex mprolog-function
26972 Do not use (do use) external functions to save and restore registers
26973 at the prologue and epilogue of a function. The external functions
26974 are slower, but use less code space if more than one function saves
26975 the same number of registers. The @option{-mprolog-function} option
26976 is on by default if you optimize.
26980 Try to make the code as small as possible. At present, this just turns
26981 on the @option{-mep} and @option{-mprolog-function} options.
26983 @item -mtda=@var{n}
26985 Put static or global variables whose size is @var{n} bytes or less into
26986 the tiny data area that register @code{ep} points to. The tiny data
26987 area can hold up to 256 bytes in total (128 bytes for byte references).
26989 @item -msda=@var{n}
26991 Put static or global variables whose size is @var{n} bytes or less into
26992 the small data area that register @code{gp} points to. The small data
26993 area can hold up to 64 kilobytes.
26995 @item -mzda=@var{n}
26997 Put static or global variables whose size is @var{n} bytes or less into
26998 the first 32 kilobytes of memory.
27002 Specify that the target processor is the V850.
27006 Specify that the target processor is the V850E3V5. The preprocessor
27007 constant @code{__v850e3v5__} is defined if this option is used.
27011 Specify that the target processor is the V850E3V5. This is an alias for
27012 the @option{-mv850e3v5} option.
27016 Specify that the target processor is the V850E2V3. The preprocessor
27017 constant @code{__v850e2v3__} is defined if this option is used.
27021 Specify that the target processor is the V850E2. The preprocessor
27022 constant @code{__v850e2__} is defined if this option is used.
27026 Specify that the target processor is the V850E1. The preprocessor
27027 constants @code{__v850e1__} and @code{__v850e__} are defined if
27028 this option is used.
27032 Specify that the target processor is the V850ES. This is an alias for
27033 the @option{-mv850e1} option.
27037 Specify that the target processor is the V850E@. The preprocessor
27038 constant @code{__v850e__} is defined if this option is used.
27040 If neither @option{-mv850} nor @option{-mv850e} nor @option{-mv850e1}
27041 nor @option{-mv850e2} nor @option{-mv850e2v3} nor @option{-mv850e3v5}
27042 are defined then a default target processor is chosen and the
27043 relevant @samp{__v850*__} preprocessor constant is defined.
27045 The preprocessor constants @code{__v850} and @code{__v851__} are always
27046 defined, regardless of which processor variant is the target.
27048 @item -mdisable-callt
27049 @itemx -mno-disable-callt
27050 @opindex mdisable-callt
27051 @opindex mno-disable-callt
27052 This option suppresses generation of the @code{CALLT} instruction for the
27053 v850e, v850e1, v850e2, v850e2v3 and v850e3v5 flavors of the v850
27056 This option is enabled by default when the RH850 ABI is
27057 in use (see @option{-mrh850-abi}), and disabled by default when the
27058 GCC ABI is in use. If @code{CALLT} instructions are being generated
27059 then the C preprocessor symbol @code{__V850_CALLT__} is defined.
27065 Pass on (or do not pass on) the @option{-mrelax} command-line option
27069 @itemx -mno-long-jumps
27070 @opindex mlong-jumps
27071 @opindex mno-long-jumps
27072 Disable (or re-enable) the generation of PC-relative jump instructions.
27075 @itemx -mhard-float
27076 @opindex msoft-float
27077 @opindex mhard-float
27078 Disable (or re-enable) the generation of hardware floating point
27079 instructions. This option is only significant when the target
27080 architecture is @samp{V850E2V3} or higher. If hardware floating point
27081 instructions are being generated then the C preprocessor symbol
27082 @code{__FPU_OK__} is defined, otherwise the symbol
27083 @code{__NO_FPU__} is defined.
27087 Enables the use of the e3v5 LOOP instruction. The use of this
27088 instruction is not enabled by default when the e3v5 architecture is
27089 selected because its use is still experimental.
27093 @opindex mrh850-abi
27095 Enables support for the RH850 version of the V850 ABI. This is the
27096 default. With this version of the ABI the following rules apply:
27100 Integer sized structures and unions are returned via a memory pointer
27101 rather than a register.
27104 Large structures and unions (more than 8 bytes in size) are passed by
27108 Functions are aligned to 16-bit boundaries.
27111 The @option{-m8byte-align} command-line option is supported.
27114 The @option{-mdisable-callt} command-line option is enabled by
27115 default. The @option{-mno-disable-callt} command-line option is not
27119 When this version of the ABI is enabled the C preprocessor symbol
27120 @code{__V850_RH850_ABI__} is defined.
27124 Enables support for the old GCC version of the V850 ABI. With this
27125 version of the ABI the following rules apply:
27129 Integer sized structures and unions are returned in register @code{r10}.
27132 Large structures and unions (more than 8 bytes in size) are passed by
27136 Functions are aligned to 32-bit boundaries, unless optimizing for
27140 The @option{-m8byte-align} command-line option is not supported.
27143 The @option{-mdisable-callt} command-line option is supported but not
27144 enabled by default.
27147 When this version of the ABI is enabled the C preprocessor symbol
27148 @code{__V850_GCC_ABI__} is defined.
27150 @item -m8byte-align
27151 @itemx -mno-8byte-align
27152 @opindex m8byte-align
27153 @opindex mno-8byte-align
27154 Enables support for @code{double} and @code{long long} types to be
27155 aligned on 8-byte boundaries. The default is to restrict the
27156 alignment of all objects to at most 4-bytes. When
27157 @option{-m8byte-align} is in effect the C preprocessor symbol
27158 @code{__V850_8BYTE_ALIGN__} is defined.
27161 @opindex mbig-switch
27162 Generate code suitable for big switch tables. Use this option only if
27163 the assembler/linker complain about out of range branches within a switch
27168 This option causes r2 and r5 to be used in the code generated by
27169 the compiler. This setting is the default.
27171 @item -mno-app-regs
27172 @opindex mno-app-regs
27173 This option causes r2 and r5 to be treated as fixed registers.
27178 @subsection VAX Options
27179 @cindex VAX options
27181 These @samp{-m} options are defined for the VAX:
27186 Do not output certain jump instructions (@code{aobleq} and so on)
27187 that the Unix assembler for the VAX cannot handle across long
27192 Do output those jump instructions, on the assumption that the
27193 GNU assembler is being used.
27197 Output code for G-format floating-point numbers instead of D-format.
27200 @node Visium Options
27201 @subsection Visium Options
27202 @cindex Visium options
27208 A program which performs file I/O and is destined to run on an MCM target
27209 should be linked with this option. It causes the libraries libc.a and
27210 libdebug.a to be linked. The program should be run on the target under
27211 the control of the GDB remote debugging stub.
27215 A program which performs file I/O and is destined to run on the simulator
27216 should be linked with option. This causes libraries libc.a and libsim.a to
27220 @itemx -mhard-float
27222 @opindex mhard-float
27223 Generate code containing floating-point instructions. This is the
27227 @itemx -msoft-float
27229 @opindex msoft-float
27230 Generate code containing library calls for floating-point.
27232 @option{-msoft-float} changes the calling convention in the output file;
27233 therefore, it is only useful if you compile @emph{all} of a program with
27234 this option. In particular, you need to compile @file{libgcc.a}, the
27235 library that comes with GCC, with @option{-msoft-float} in order for
27238 @item -mcpu=@var{cpu_type}
27240 Set the instruction set, register set, and instruction scheduling parameters
27241 for machine type @var{cpu_type}. Supported values for @var{cpu_type} are
27242 @samp{mcm}, @samp{gr5} and @samp{gr6}.
27244 @samp{mcm} is a synonym of @samp{gr5} present for backward compatibility.
27246 By default (unless configured otherwise), GCC generates code for the GR5
27247 variant of the Visium architecture.
27249 With @option{-mcpu=gr6}, GCC generates code for the GR6 variant of the Visium
27250 architecture. The only difference from GR5 code is that the compiler will
27251 generate block move instructions.
27253 @item -mtune=@var{cpu_type}
27255 Set the instruction scheduling parameters for machine type @var{cpu_type},
27256 but do not set the instruction set or register set that the option
27257 @option{-mcpu=@var{cpu_type}} would.
27261 Generate code for the supervisor mode, where there are no restrictions on
27262 the access to general registers. This is the default.
27265 @opindex muser-mode
27266 Generate code for the user mode, where the access to some general registers
27267 is forbidden: on the GR5, registers r24 to r31 cannot be accessed in this
27268 mode; on the GR6, only registers r29 to r31 are affected.
27272 @subsection VMS Options
27274 These @samp{-m} options are defined for the VMS implementations:
27277 @item -mvms-return-codes
27278 @opindex mvms-return-codes
27279 Return VMS condition codes from @code{main}. The default is to return POSIX-style
27280 condition (e.g.@: error) codes.
27282 @item -mdebug-main=@var{prefix}
27283 @opindex mdebug-main=@var{prefix}
27284 Flag the first routine whose name starts with @var{prefix} as the main
27285 routine for the debugger.
27289 Default to 64-bit memory allocation routines.
27291 @item -mpointer-size=@var{size}
27292 @opindex mpointer-size=@var{size}
27293 Set the default size of pointers. Possible options for @var{size} are
27294 @samp{32} or @samp{short} for 32 bit pointers, @samp{64} or @samp{long}
27295 for 64 bit pointers, and @samp{no} for supporting only 32 bit pointers.
27296 The later option disables @code{pragma pointer_size}.
27299 @node VxWorks Options
27300 @subsection VxWorks Options
27301 @cindex VxWorks Options
27303 The options in this section are defined for all VxWorks targets.
27304 Options specific to the target hardware are listed with the other
27305 options for that target.
27310 GCC can generate code for both VxWorks kernels and real time processes
27311 (RTPs). This option switches from the former to the latter. It also
27312 defines the preprocessor macro @code{__RTP__}.
27315 @opindex non-static
27316 Link an RTP executable against shared libraries rather than static
27317 libraries. The options @option{-static} and @option{-shared} can
27318 also be used for RTPs (@pxref{Link Options}); @option{-static}
27325 These options are passed down to the linker. They are defined for
27326 compatibility with Diab.
27329 @opindex Xbind-lazy
27330 Enable lazy binding of function calls. This option is equivalent to
27331 @option{-Wl,-z,now} and is defined for compatibility with Diab.
27335 Disable lazy binding of function calls. This option is the default and
27336 is defined for compatibility with Diab.
27340 @subsection x86 Options
27341 @cindex x86 Options
27343 These @samp{-m} options are defined for the x86 family of computers.
27347 @item -march=@var{cpu-type}
27349 Generate instructions for the machine type @var{cpu-type}. In contrast to
27350 @option{-mtune=@var{cpu-type}}, which merely tunes the generated code
27351 for the specified @var{cpu-type}, @option{-march=@var{cpu-type}} allows GCC
27352 to generate code that may not run at all on processors other than the one
27353 indicated. Specifying @option{-march=@var{cpu-type}} implies
27354 @option{-mtune=@var{cpu-type}}.
27356 The choices for @var{cpu-type} are:
27360 This selects the CPU to generate code for at compilation time by determining
27361 the processor type of the compiling machine. Using @option{-march=native}
27362 enables all instruction subsets supported by the local machine (hence
27363 the result might not run on different machines). Using @option{-mtune=native}
27364 produces code optimized for the local machine under the constraints
27365 of the selected instruction set.
27368 A generic CPU with 64-bit extensions.
27371 Original Intel i386 CPU@.
27374 Intel i486 CPU@. (No scheduling is implemented for this chip.)
27378 Intel Pentium CPU with no MMX support.
27381 Intel Lakemont MCU, based on Intel Pentium CPU.
27384 Intel Pentium MMX CPU, based on Pentium core with MMX instruction set support.
27387 Intel Pentium Pro CPU@.
27390 When used with @option{-march}, the Pentium Pro
27391 instruction set is used, so the code runs on all i686 family chips.
27392 When used with @option{-mtune}, it has the same meaning as @samp{generic}.
27395 Intel Pentium II CPU, based on Pentium Pro core with MMX instruction set
27400 Intel Pentium III CPU, based on Pentium Pro core with MMX and SSE instruction
27404 Intel Pentium M; low-power version of Intel Pentium III CPU
27405 with MMX, SSE and SSE2 instruction set support. Used by Centrino notebooks.
27409 Intel Pentium 4 CPU with MMX, SSE and SSE2 instruction set support.
27412 Improved version of Intel Pentium 4 CPU with MMX, SSE, SSE2 and SSE3 instruction
27416 Improved version of Intel Pentium 4 CPU with 64-bit extensions, MMX, SSE,
27417 SSE2 and SSE3 instruction set support.
27420 Intel Core 2 CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3 and SSSE3
27421 instruction set support.
27424 Intel Nehalem CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3, SSSE3,
27425 SSE4.1, SSE4.2 and POPCNT instruction set support.
27428 Intel Westmere CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3, SSSE3,
27429 SSE4.1, SSE4.2, POPCNT, AES and PCLMUL instruction set support.
27432 Intel Sandy Bridge CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3, SSSE3,
27433 SSE4.1, SSE4.2, POPCNT, AVX, AES and PCLMUL instruction set support.
27436 Intel Ivy Bridge CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3, SSSE3,
27437 SSE4.1, SSE4.2, POPCNT, AVX, AES, PCLMUL, FSGSBASE, RDRND and F16C
27438 instruction set support.
27441 Intel Haswell CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3, SSSE3,
27442 SSE4.1, SSE4.2, POPCNT, AVX, AVX2, AES, PCLMUL, FSGSBASE, RDRND, FMA,
27443 BMI, BMI2 and F16C instruction set support.
27446 Intel Broadwell CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3, SSSE3,
27447 SSE4.1, SSE4.2, POPCNT, AVX, AVX2, AES, PCLMUL, FSGSBASE, RDRND, FMA,
27448 BMI, BMI2, F16C, RDSEED, ADCX and PREFETCHW instruction set support.
27451 Intel Skylake CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3, SSSE3,
27452 SSE4.1, SSE4.2, POPCNT, AVX, AVX2, AES, PCLMUL, FSGSBASE, RDRND, FMA,
27453 BMI, BMI2, F16C, RDSEED, ADCX, PREFETCHW, CLFLUSHOPT, XSAVEC and
27454 XSAVES instruction set support.
27457 Intel Bonnell CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3 and SSSE3
27458 instruction set support.
27461 Intel Silvermont CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3, SSSE3,
27462 SSE4.1, SSE4.2, POPCNT, AES, PCLMUL and RDRND instruction set support.
27465 Intel Goldmont CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3, SSSE3,
27466 SSE4.1, SSE4.2, POPCNT, AES, PCLMUL, RDRND, XSAVE, XSAVEOPT and FSGSBASE
27467 instruction set support.
27469 @item goldmont-plus
27470 Intel Goldmont Plus CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3,
27471 SSSE3, SSE4.1, SSE4.2, POPCNT, AES, PCLMUL, RDRND, XSAVE, XSAVEOPT, FSGSBASE,
27472 PTWRITE, RDPID, SGX and UMIP instruction set support.
27475 Intel Tremont CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3, SSSE3,
27476 SSE4.1, SSE4.2, POPCNT, AES, PCLMUL, RDRND, XSAVE, XSAVEOPT, FSGSBASE, PTWRITE,
27477 RDPID, SGX, UMIP, GFNI-SSE, CLWB and ENCLV instruction set support.
27480 Intel Knight's Landing CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3,
27481 SSSE3, SSE4.1, SSE4.2, POPCNT, AVX, AVX2, AES, PCLMUL, FSGSBASE, RDRND, FMA,
27482 BMI, BMI2, F16C, RDSEED, ADCX, PREFETCHW, AVX512F, AVX512PF, AVX512ER and
27483 AVX512CD instruction set support.
27486 Intel Knights Mill CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3,
27487 SSSE3, SSE4.1, SSE4.2, POPCNT, AVX, AVX2, AES, PCLMUL, FSGSBASE, RDRND, FMA,
27488 BMI, BMI2, F16C, RDSEED, ADCX, PREFETCHW, AVX512F, AVX512PF, AVX512ER, AVX512CD,
27489 AVX5124VNNIW, AVX5124FMAPS and AVX512VPOPCNTDQ instruction set support.
27491 @item skylake-avx512
27492 Intel Skylake Server CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3,
27493 SSSE3, SSE4.1, SSE4.2, POPCNT, PKU, AVX, AVX2, AES, PCLMUL, FSGSBASE, RDRND, FMA,
27494 BMI, BMI2, F16C, RDSEED, ADCX, PREFETCHW, CLFLUSHOPT, XSAVEC, XSAVES, AVX512F,
27495 CLWB, AVX512VL, AVX512BW, AVX512DQ and AVX512CD instruction set support.
27498 Intel Cannonlake Server CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2,
27499 SSE3, SSSE3, SSE4.1, SSE4.2, POPCNT, PKU, AVX, AVX2, AES, PCLMUL, FSGSBASE,
27500 RDRND, FMA, BMI, BMI2, F16C, RDSEED, ADCX, PREFETCHW, CLFLUSHOPT, XSAVEC,
27501 XSAVES, AVX512F, AVX512VL, AVX512BW, AVX512DQ, AVX512CD, AVX512VBMI,
27502 AVX512IFMA, SHA and UMIP instruction set support.
27504 @item icelake-client
27505 Intel Icelake Client CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2,
27506 SSE3, SSSE3, SSE4.1, SSE4.2, POPCNT, PKU, AVX, AVX2, AES, PCLMUL, FSGSBASE,
27507 RDRND, FMA, BMI, BMI2, F16C, RDSEED, ADCX, PREFETCHW, CLFLUSHOPT, XSAVEC,
27508 XSAVES, AVX512F, AVX512VL, AVX512BW, AVX512DQ, AVX512CD, AVX512VBMI,
27509 AVX512IFMA, SHA, CLWB, UMIP, RDPID, GFNI, AVX512VBMI2, AVX512VPOPCNTDQ,
27510 AVX512BITALG, AVX512VNNI, VPCLMULQDQ, VAES instruction set support.
27512 @item icelake-server
27513 Intel Icelake Server CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2,
27514 SSE3, SSSE3, SSE4.1, SSE4.2, POPCNT, PKU, AVX, AVX2, AES, PCLMUL, FSGSBASE,
27515 RDRND, FMA, BMI, BMI2, F16C, RDSEED, ADCX, PREFETCHW, CLFLUSHOPT, XSAVEC,
27516 XSAVES, AVX512F, AVX512VL, AVX512BW, AVX512DQ, AVX512CD, AVX512VBMI,
27517 AVX512IFMA, SHA, CLWB, UMIP, RDPID, GFNI, AVX512VBMI2, AVX512VPOPCNTDQ,
27518 AVX512BITALG, AVX512VNNI, VPCLMULQDQ, VAES, PCONFIG and WBNOINVD instruction
27522 Intel Cascadelake CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3, SSSE3,
27523 SSE4.1, SSE4.2, POPCNT, PKU, AVX, AVX2, AES, PCLMUL, FSGSBASE, RDRND, FMA, BMI,
27524 BMI2, F16C, RDSEED, ADCX, PREFETCHW, CLFLUSHOPT, XSAVEC, XSAVES, AVX512F, CLWB,
27525 AVX512VL, AVX512BW, AVX512DQ, AVX512CD and AVX512VNNI instruction set support.
27528 Intel cooperlake CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3, SSSE3,
27529 SSE4.1, SSE4.2, POPCNT, PKU, AVX, AVX2, AES, PCLMUL, FSGSBASE, RDRND, FMA, BMI,
27530 BMI2, F16C, RDSEED, ADCX, PREFETCHW, CLFLUSHOPT, XSAVEC, XSAVES, AVX512F, CLWB,
27531 AVX512VL, AVX512BW, AVX512DQ, AVX512CD, AVX512VNNI and AVX512BF16 instruction
27535 Intel Tigerlake CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3, SSSE3,
27536 SSE4.1, SSE4.2, POPCNT, PKU, AVX, AVX2, AES, PCLMUL, FSGSBASE, RDRND, FMA, BMI,
27537 BMI2, F16C, RDSEED, ADCX, PREFETCHW, CLFLUSHOPT, XSAVEC, XSAVES, AVX512F,
27538 AVX512VL, AVX512BW, AVX512DQ, AVX512CD, AVX512VBMI, AVX512IFMA, SHA, CLWB, UMIP,
27539 RDPID, GFNI, AVX512VBMI2, AVX512VPOPCNTDQ, AVX512BITALG, AVX512VNNI, VPCLMULQDQ,
27540 VAES, PCONFIG, WBNOINVD, MOVDIRI, MOVDIR64B and AVX512VP2INTERSECT instruction
27544 AMD K6 CPU with MMX instruction set support.
27548 Improved versions of AMD K6 CPU with MMX and 3DNow!@: instruction set support.
27551 @itemx athlon-tbird
27552 AMD Athlon CPU with MMX, 3dNOW!, enhanced 3DNow!@: and SSE prefetch instructions
27558 Improved AMD Athlon CPU with MMX, 3DNow!, enhanced 3DNow!@: and full SSE
27559 instruction set support.
27565 Processors based on the AMD K8 core with x86-64 instruction set support,
27566 including the AMD Opteron, Athlon 64, and Athlon 64 FX processors.
27567 (This supersets MMX, SSE, SSE2, 3DNow!, enhanced 3DNow!@: and 64-bit
27568 instruction set extensions.)
27571 @itemx opteron-sse3
27572 @itemx athlon64-sse3
27573 Improved versions of AMD K8 cores with SSE3 instruction set support.
27577 CPUs based on AMD Family 10h cores with x86-64 instruction set support. (This
27578 supersets MMX, SSE, SSE2, SSE3, SSE4A, 3DNow!, enhanced 3DNow!, ABM and 64-bit
27579 instruction set extensions.)
27582 CPUs based on AMD Family 15h cores with x86-64 instruction set support. (This
27583 supersets FMA4, AVX, XOP, LWP, AES, PCLMUL, CX16, MMX, SSE, SSE2, SSE3, SSE4A,
27584 SSSE3, SSE4.1, SSE4.2, ABM and 64-bit instruction set extensions.)
27586 AMD Family 15h core based CPUs with x86-64 instruction set support. (This
27587 supersets BMI, TBM, F16C, FMA, FMA4, AVX, XOP, LWP, AES, PCLMUL, CX16, MMX,
27588 SSE, SSE2, SSE3, SSE4A, SSSE3, SSE4.1, SSE4.2, ABM and 64-bit instruction set
27591 AMD Family 15h core based CPUs with x86-64 instruction set support. (This
27592 supersets BMI, TBM, F16C, FMA, FMA4, FSGSBASE, AVX, XOP, LWP, AES,
27593 PCLMUL, CX16, MMX, SSE, SSE2, SSE3, SSE4A, SSSE3, SSE4.1, SSE4.2, ABM and
27594 64-bit instruction set extensions.
27596 AMD Family 15h core based CPUs with x86-64 instruction set support. (This
27597 supersets BMI, BMI2, TBM, F16C, FMA, FMA4, FSGSBASE, AVX, AVX2, XOP, LWP,
27598 AES, PCLMUL, CX16, MOVBE, MMX, SSE, SSE2, SSE3, SSE4A, SSSE3, SSE4.1,
27599 SSE4.2, ABM and 64-bit instruction set extensions.
27602 AMD Family 17h core based CPUs with x86-64 instruction set support. (This
27603 supersets BMI, BMI2, F16C, FMA, FSGSBASE, AVX, AVX2, ADCX, RDSEED, MWAITX,
27604 SHA, CLZERO, AES, PCLMUL, CX16, MOVBE, MMX, SSE, SSE2, SSE3, SSE4A, SSSE3,
27605 SSE4.1, SSE4.2, ABM, XSAVEC, XSAVES, CLFLUSHOPT, POPCNT, and 64-bit
27606 instruction set extensions.
27608 AMD Family 17h core based CPUs with x86-64 instruction set support. (This
27609 supersets BMI, BMI2, ,CLWB, F16C, FMA, FSGSBASE, AVX, AVX2, ADCX, RDSEED,
27610 MWAITX, SHA, CLZERO, AES, PCLMUL, CX16, MOVBE, MMX, SSE, SSE2, SSE3, SSE4A,
27611 SSSE3, SSE4.1, SSE4.2, ABM, XSAVEC, XSAVES, CLFLUSHOPT, POPCNT, and 64-bit
27612 instruction set extensions.)
27616 CPUs based on AMD Family 14h cores with x86-64 instruction set support. (This
27617 supersets MMX, SSE, SSE2, SSE3, SSSE3, SSE4A, CX16, ABM and 64-bit
27618 instruction set extensions.)
27621 CPUs based on AMD Family 16h cores with x86-64 instruction set support. This
27622 includes MOVBE, F16C, BMI, AVX, PCLMUL, AES, SSE4.2, SSE4.1, CX16, ABM,
27623 SSE4A, SSSE3, SSE3, SSE2, SSE, MMX and 64-bit instruction set extensions.
27626 IDT WinChip C6 CPU, dealt in same way as i486 with additional MMX instruction
27630 IDT WinChip 2 CPU, dealt in same way as i486 with additional MMX and 3DNow!@:
27631 instruction set support.
27634 VIA C3 CPU with MMX and 3DNow!@: instruction set support.
27635 (No scheduling is implemented for this chip.)
27638 VIA C3-2 (Nehemiah/C5XL) CPU with MMX and SSE instruction set support.
27639 (No scheduling is implemented for this chip.)
27642 VIA C7 (Esther) CPU with MMX, SSE, SSE2 and SSE3 instruction set support.
27643 (No scheduling is implemented for this chip.)
27646 VIA Eden Samuel 2 CPU with MMX and 3DNow!@: instruction set support.
27647 (No scheduling is implemented for this chip.)
27650 VIA Eden Nehemiah CPU with MMX and SSE instruction set support.
27651 (No scheduling is implemented for this chip.)
27654 VIA Eden Esther CPU with MMX, SSE, SSE2 and SSE3 instruction set support.
27655 (No scheduling is implemented for this chip.)
27658 VIA Eden X2 CPU with x86-64, MMX, SSE, SSE2 and SSE3 instruction set support.
27659 (No scheduling is implemented for this chip.)
27662 VIA Eden X4 CPU with x86-64, MMX, SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2,
27663 AVX and AVX2 instruction set support.
27664 (No scheduling is implemented for this chip.)
27667 Generic VIA Nano CPU with x86-64, MMX, SSE, SSE2, SSE3 and SSSE3
27668 instruction set support.
27669 (No scheduling is implemented for this chip.)
27672 VIA Nano 1xxx CPU with x86-64, MMX, SSE, SSE2, SSE3 and SSSE3
27673 instruction set support.
27674 (No scheduling is implemented for this chip.)
27677 VIA Nano 2xxx CPU with x86-64, MMX, SSE, SSE2, SSE3 and SSSE3
27678 instruction set support.
27679 (No scheduling is implemented for this chip.)
27682 VIA Nano 3xxx CPU with x86-64, MMX, SSE, SSE2, SSE3, SSSE3 and SSE4.1
27683 instruction set support.
27684 (No scheduling is implemented for this chip.)
27687 VIA Nano Dual Core CPU with x86-64, MMX, SSE, SSE2, SSE3, SSSE3 and SSE4.1
27688 instruction set support.
27689 (No scheduling is implemented for this chip.)
27692 VIA Nano Quad Core CPU with x86-64, MMX, SSE, SSE2, SSE3, SSSE3 and SSE4.1
27693 instruction set support.
27694 (No scheduling is implemented for this chip.)
27697 AMD Geode embedded processor with MMX and 3DNow!@: instruction set support.
27700 @item -mtune=@var{cpu-type}
27702 Tune to @var{cpu-type} everything applicable about the generated code, except
27703 for the ABI and the set of available instructions.
27704 While picking a specific @var{cpu-type} schedules things appropriately
27705 for that particular chip, the compiler does not generate any code that
27706 cannot run on the default machine type unless you use a
27707 @option{-march=@var{cpu-type}} option.
27708 For example, if GCC is configured for i686-pc-linux-gnu
27709 then @option{-mtune=pentium4} generates code that is tuned for Pentium 4
27710 but still runs on i686 machines.
27712 The choices for @var{cpu-type} are the same as for @option{-march}.
27713 In addition, @option{-mtune} supports 2 extra choices for @var{cpu-type}:
27717 Produce code optimized for the most common IA32/@/AMD64/@/EM64T processors.
27718 If you know the CPU on which your code will run, then you should use
27719 the corresponding @option{-mtune} or @option{-march} option instead of
27720 @option{-mtune=generic}. But, if you do not know exactly what CPU users
27721 of your application will have, then you should use this option.
27723 As new processors are deployed in the marketplace, the behavior of this
27724 option will change. Therefore, if you upgrade to a newer version of
27725 GCC, code generation controlled by this option will change to reflect
27727 that are most common at the time that version of GCC is released.
27729 There is no @option{-march=generic} option because @option{-march}
27730 indicates the instruction set the compiler can use, and there is no
27731 generic instruction set applicable to all processors. In contrast,
27732 @option{-mtune} indicates the processor (or, in this case, collection of
27733 processors) for which the code is optimized.
27736 Produce code optimized for the most current Intel processors, which are
27737 Haswell and Silvermont for this version of GCC. If you know the CPU
27738 on which your code will run, then you should use the corresponding
27739 @option{-mtune} or @option{-march} option instead of @option{-mtune=intel}.
27740 But, if you want your application performs better on both Haswell and
27741 Silvermont, then you should use this option.
27743 As new Intel processors are deployed in the marketplace, the behavior of
27744 this option will change. Therefore, if you upgrade to a newer version of
27745 GCC, code generation controlled by this option will change to reflect
27746 the most current Intel processors at the time that version of GCC is
27749 There is no @option{-march=intel} option because @option{-march} indicates
27750 the instruction set the compiler can use, and there is no common
27751 instruction set applicable to all processors. In contrast,
27752 @option{-mtune} indicates the processor (or, in this case, collection of
27753 processors) for which the code is optimized.
27756 @item -mcpu=@var{cpu-type}
27758 A deprecated synonym for @option{-mtune}.
27760 @item -mfpmath=@var{unit}
27762 Generate floating-point arithmetic for selected unit @var{unit}. The choices
27763 for @var{unit} are:
27767 Use the standard 387 floating-point coprocessor present on the majority of chips and
27768 emulated otherwise. Code compiled with this option runs almost everywhere.
27769 The temporary results are computed in 80-bit precision instead of the precision
27770 specified by the type, resulting in slightly different results compared to most
27771 of other chips. See @option{-ffloat-store} for more detailed description.
27773 This is the default choice for non-Darwin x86-32 targets.
27776 Use scalar floating-point instructions present in the SSE instruction set.
27777 This instruction set is supported by Pentium III and newer chips,
27778 and in the AMD line
27779 by Athlon-4, Athlon XP and Athlon MP chips. The earlier version of the SSE
27780 instruction set supports only single-precision arithmetic, thus the double and
27781 extended-precision arithmetic are still done using 387. A later version, present
27782 only in Pentium 4 and AMD x86-64 chips, supports double-precision
27785 For the x86-32 compiler, you must use @option{-march=@var{cpu-type}}, @option{-msse}
27786 or @option{-msse2} switches to enable SSE extensions and make this option
27787 effective. For the x86-64 compiler, these extensions are enabled by default.
27789 The resulting code should be considerably faster in the majority of cases and avoid
27790 the numerical instability problems of 387 code, but may break some existing
27791 code that expects temporaries to be 80 bits.
27793 This is the default choice for the x86-64 compiler, Darwin x86-32 targets,
27794 and the default choice for x86-32 targets with the SSE2 instruction set
27795 when @option{-ffast-math} is enabled.
27800 Attempt to utilize both instruction sets at once. This effectively doubles the
27801 amount of available registers, and on chips with separate execution units for
27802 387 and SSE the execution resources too. Use this option with care, as it is
27803 still experimental, because the GCC register allocator does not model separate
27804 functional units well, resulting in unstable performance.
27807 @item -masm=@var{dialect}
27808 @opindex masm=@var{dialect}
27809 Output assembly instructions using selected @var{dialect}. Also affects
27810 which dialect is used for basic @code{asm} (@pxref{Basic Asm}) and
27811 extended @code{asm} (@pxref{Extended Asm}). Supported choices (in dialect
27812 order) are @samp{att} or @samp{intel}. The default is @samp{att}. Darwin does
27813 not support @samp{intel}.
27816 @itemx -mno-ieee-fp
27818 @opindex mno-ieee-fp
27819 Control whether or not the compiler uses IEEE floating-point
27820 comparisons. These correctly handle the case where the result of a
27821 comparison is unordered.
27824 @itemx -mhard-float
27826 @opindex mhard-float
27827 Generate output containing 80387 instructions for floating point.
27830 @itemx -msoft-float
27832 @opindex msoft-float
27833 Generate output containing library calls for floating point.
27835 @strong{Warning:} the requisite libraries are not part of GCC@.
27836 Normally the facilities of the machine's usual C compiler are used, but
27837 this cannot be done directly in cross-compilation. You must make your
27838 own arrangements to provide suitable library functions for
27841 On machines where a function returns floating-point results in the 80387
27842 register stack, some floating-point opcodes may be emitted even if
27843 @option{-msoft-float} is used.
27845 @item -mno-fp-ret-in-387
27846 @opindex mno-fp-ret-in-387
27847 @opindex mfp-ret-in-387
27848 Do not use the FPU registers for return values of functions.
27850 The usual calling convention has functions return values of types
27851 @code{float} and @code{double} in an FPU register, even if there
27852 is no FPU@. The idea is that the operating system should emulate
27855 The option @option{-mno-fp-ret-in-387} causes such values to be returned
27856 in ordinary CPU registers instead.
27858 @item -mno-fancy-math-387
27859 @opindex mno-fancy-math-387
27860 @opindex mfancy-math-387
27861 Some 387 emulators do not support the @code{sin}, @code{cos} and
27862 @code{sqrt} instructions for the 387. Specify this option to avoid
27863 generating those instructions.
27864 This option is overridden when @option{-march}
27865 indicates that the target CPU always has an FPU and so the
27866 instruction does not need emulation. These
27867 instructions are not generated unless you also use the
27868 @option{-funsafe-math-optimizations} switch.
27870 @item -malign-double
27871 @itemx -mno-align-double
27872 @opindex malign-double
27873 @opindex mno-align-double
27874 Control whether GCC aligns @code{double}, @code{long double}, and
27875 @code{long long} variables on a two-word boundary or a one-word
27876 boundary. Aligning @code{double} variables on a two-word boundary
27877 produces code that runs somewhat faster on a Pentium at the
27878 expense of more memory.
27880 On x86-64, @option{-malign-double} is enabled by default.
27882 @strong{Warning:} if you use the @option{-malign-double} switch,
27883 structures containing the above types are aligned differently than
27884 the published application binary interface specifications for the x86-32
27885 and are not binary compatible with structures in code compiled
27886 without that switch.
27888 @item -m96bit-long-double
27889 @itemx -m128bit-long-double
27890 @opindex m96bit-long-double
27891 @opindex m128bit-long-double
27892 These switches control the size of @code{long double} type. The x86-32
27893 application binary interface specifies the size to be 96 bits,
27894 so @option{-m96bit-long-double} is the default in 32-bit mode.
27896 Modern architectures (Pentium and newer) prefer @code{long double}
27897 to be aligned to an 8- or 16-byte boundary. In arrays or structures
27898 conforming to the ABI, this is not possible. So specifying
27899 @option{-m128bit-long-double} aligns @code{long double}
27900 to a 16-byte boundary by padding the @code{long double} with an additional
27903 In the x86-64 compiler, @option{-m128bit-long-double} is the default choice as
27904 its ABI specifies that @code{long double} is aligned on 16-byte boundary.
27906 Notice that neither of these options enable any extra precision over the x87
27907 standard of 80 bits for a @code{long double}.
27909 @strong{Warning:} if you override the default value for your target ABI, this
27910 changes the size of
27911 structures and arrays containing @code{long double} variables,
27912 as well as modifying the function calling convention for functions taking
27913 @code{long double}. Hence they are not binary-compatible
27914 with code compiled without that switch.
27916 @item -mlong-double-64
27917 @itemx -mlong-double-80
27918 @itemx -mlong-double-128
27919 @opindex mlong-double-64
27920 @opindex mlong-double-80
27921 @opindex mlong-double-128
27922 These switches control the size of @code{long double} type. A size
27923 of 64 bits makes the @code{long double} type equivalent to the @code{double}
27924 type. This is the default for 32-bit Bionic C library. A size
27925 of 128 bits makes the @code{long double} type equivalent to the
27926 @code{__float128} type. This is the default for 64-bit Bionic C library.
27928 @strong{Warning:} if you override the default value for your target ABI, this
27929 changes the size of
27930 structures and arrays containing @code{long double} variables,
27931 as well as modifying the function calling convention for functions taking
27932 @code{long double}. Hence they are not binary-compatible
27933 with code compiled without that switch.
27935 @item -malign-data=@var{type}
27936 @opindex malign-data
27937 Control how GCC aligns variables. Supported values for @var{type} are
27938 @samp{compat} uses increased alignment value compatible uses GCC 4.8
27939 and earlier, @samp{abi} uses alignment value as specified by the
27940 psABI, and @samp{cacheline} uses increased alignment value to match
27941 the cache line size. @samp{compat} is the default.
27943 @item -mlarge-data-threshold=@var{threshold}
27944 @opindex mlarge-data-threshold
27945 When @option{-mcmodel=medium} is specified, data objects larger than
27946 @var{threshold} are placed in the large data section. This value must be the
27947 same across all objects linked into the binary, and defaults to 65535.
27951 Use a different function-calling convention, in which functions that
27952 take a fixed number of arguments return with the @code{ret @var{num}}
27953 instruction, which pops their arguments while returning. This saves one
27954 instruction in the caller since there is no need to pop the arguments
27957 You can specify that an individual function is called with this calling
27958 sequence with the function attribute @code{stdcall}. You can also
27959 override the @option{-mrtd} option by using the function attribute
27960 @code{cdecl}. @xref{Function Attributes}.
27962 @strong{Warning:} this calling convention is incompatible with the one
27963 normally used on Unix, so you cannot use it if you need to call
27964 libraries compiled with the Unix compiler.
27966 Also, you must provide function prototypes for all functions that
27967 take variable numbers of arguments (including @code{printf});
27968 otherwise incorrect code is generated for calls to those
27971 In addition, seriously incorrect code results if you call a
27972 function with too many arguments. (Normally, extra arguments are
27973 harmlessly ignored.)
27975 @item -mregparm=@var{num}
27977 Control how many registers are used to pass integer arguments. By
27978 default, no registers are used to pass arguments, and at most 3
27979 registers can be used. You can control this behavior for a specific
27980 function by using the function attribute @code{regparm}.
27981 @xref{Function Attributes}.
27983 @strong{Warning:} if you use this switch, and
27984 @var{num} is nonzero, then you must build all modules with the same
27985 value, including any libraries. This includes the system libraries and
27989 @opindex msseregparm
27990 Use SSE register passing conventions for float and double arguments
27991 and return values. You can control this behavior for a specific
27992 function by using the function attribute @code{sseregparm}.
27993 @xref{Function Attributes}.
27995 @strong{Warning:} if you use this switch then you must build all
27996 modules with the same value, including any libraries. This includes
27997 the system libraries and startup modules.
27999 @item -mvect8-ret-in-mem
28000 @opindex mvect8-ret-in-mem
28001 Return 8-byte vectors in memory instead of MMX registers. This is the
28002 default on VxWorks to match the ABI of the Sun Studio compilers until
28003 version 12. @emph{Only} use this option if you need to remain
28004 compatible with existing code produced by those previous compiler
28005 versions or older versions of GCC@.
28014 Set 80387 floating-point precision to 32, 64 or 80 bits. When @option{-mpc32}
28015 is specified, the significands of results of floating-point operations are
28016 rounded to 24 bits (single precision); @option{-mpc64} rounds the
28017 significands of results of floating-point operations to 53 bits (double
28018 precision) and @option{-mpc80} rounds the significands of results of
28019 floating-point operations to 64 bits (extended double precision), which is
28020 the default. When this option is used, floating-point operations in higher
28021 precisions are not available to the programmer without setting the FPU
28022 control word explicitly.
28024 Setting the rounding of floating-point operations to less than the default
28025 80 bits can speed some programs by 2% or more. Note that some mathematical
28026 libraries assume that extended-precision (80-bit) floating-point operations
28027 are enabled by default; routines in such libraries could suffer significant
28028 loss of accuracy, typically through so-called ``catastrophic cancellation'',
28029 when this option is used to set the precision to less than extended precision.
28031 @item -mstackrealign
28032 @opindex mstackrealign
28033 Realign the stack at entry. On the x86, the @option{-mstackrealign}
28034 option generates an alternate prologue and epilogue that realigns the
28035 run-time stack if necessary. This supports mixing legacy codes that keep
28036 4-byte stack alignment with modern codes that keep 16-byte stack alignment for
28037 SSE compatibility. See also the attribute @code{force_align_arg_pointer},
28038 applicable to individual functions.
28040 @item -mpreferred-stack-boundary=@var{num}
28041 @opindex mpreferred-stack-boundary
28042 Attempt to keep the stack boundary aligned to a 2 raised to @var{num}
28043 byte boundary. If @option{-mpreferred-stack-boundary} is not specified,
28044 the default is 4 (16 bytes or 128 bits).
28046 @strong{Warning:} When generating code for the x86-64 architecture with
28047 SSE extensions disabled, @option{-mpreferred-stack-boundary=3} can be
28048 used to keep the stack boundary aligned to 8 byte boundary. Since
28049 x86-64 ABI require 16 byte stack alignment, this is ABI incompatible and
28050 intended to be used in controlled environment where stack space is
28051 important limitation. This option leads to wrong code when functions
28052 compiled with 16 byte stack alignment (such as functions from a standard
28053 library) are called with misaligned stack. In this case, SSE
28054 instructions may lead to misaligned memory access traps. In addition,
28055 variable arguments are handled incorrectly for 16 byte aligned
28056 objects (including x87 long double and __int128), leading to wrong
28057 results. You must build all modules with
28058 @option{-mpreferred-stack-boundary=3}, including any libraries. This
28059 includes the system libraries and startup modules.
28061 @item -mincoming-stack-boundary=@var{num}
28062 @opindex mincoming-stack-boundary
28063 Assume the incoming stack is aligned to a 2 raised to @var{num} byte
28064 boundary. If @option{-mincoming-stack-boundary} is not specified,
28065 the one specified by @option{-mpreferred-stack-boundary} is used.
28067 On Pentium and Pentium Pro, @code{double} and @code{long double} values
28068 should be aligned to an 8-byte boundary (see @option{-malign-double}) or
28069 suffer significant run time performance penalties. On Pentium III, the
28070 Streaming SIMD Extension (SSE) data type @code{__m128} may not work
28071 properly if it is not 16-byte aligned.
28073 To ensure proper alignment of this values on the stack, the stack boundary
28074 must be as aligned as that required by any value stored on the stack.
28075 Further, every function must be generated such that it keeps the stack
28076 aligned. Thus calling a function compiled with a higher preferred
28077 stack boundary from a function compiled with a lower preferred stack
28078 boundary most likely misaligns the stack. It is recommended that
28079 libraries that use callbacks always use the default setting.
28081 This extra alignment does consume extra stack space, and generally
28082 increases code size. Code that is sensitive to stack space usage, such
28083 as embedded systems and operating system kernels, may want to reduce the
28084 preferred alignment to @option{-mpreferred-stack-boundary=2}.
28141 @itemx -mavx512ifma
28142 @opindex mavx512ifma
28144 @itemx -mavx512vbmi
28145 @opindex mavx512vbmi
28156 @itemx -mclflushopt
28157 @opindex mclflushopt
28192 @itemx -mprefetchwt1
28193 @opindex mprefetchwt1
28264 @itemx -mavx512vbmi2
28265 @opindex mavx512vbmi2
28267 @itemx -mavx512bf16
28268 @opindex mavx512bf16
28279 @itemx -mvpclmulqdq
28280 @opindex mvpclmulqdq
28282 @itemx -mavx512bitalg
28283 @opindex mavx512bitalg
28289 @opindex mmovdir64b
28294 @itemx -mavx512vpopcntdq
28295 @opindex mavx512vpopcntdq
28297 @itemx -mavx512vp2intersect
28298 @opindex mavx512vp2intersect
28300 @itemx -mavx5124fmaps
28301 @opindex mavx5124fmaps
28303 @itemx -mavx512vnni
28304 @opindex mavx512vnni
28306 @itemx -mavx5124vnniw
28307 @opindex mavx5124vnniw
28311 These switches enable the use of instructions in the MMX, SSE,
28312 SSE2, SSE3, SSSE3, SSE4, SSE4A, SSE4.1, SSE4.2, AVX, AVX2, AVX512F, AVX512PF,
28313 AVX512ER, AVX512CD, AVX512VL, AVX512BW, AVX512DQ, AVX512IFMA, AVX512VBMI, SHA,
28314 AES, PCLMUL, CLFLUSHOPT, CLWB, FSGSBASE, PTWRITE, RDRND, F16C, FMA, PCONFIG,
28315 WBNOINVD, FMA4, PREFETCHW, RDPID, PREFETCHWT1, RDSEED, SGX, XOP, LWP,
28316 3DNow!@:, enhanced 3DNow!@:, POPCNT, ABM, ADX, BMI, BMI2, LZCNT, FXSR, XSAVE,
28317 XSAVEOPT, XSAVEC, XSAVES, RTM, HLE, TBM, MWAITX, CLZERO, PKU, AVX512VBMI2,
28318 GFNI, VAES, WAITPKG, VPCLMULQDQ, AVX512BITALG, MOVDIRI, MOVDIR64B, AVX512BF16,
28319 ENQCMD, AVX512VPOPCNTDQ, AVX5124FMAPS, AVX512VNNI, AVX5124VNNIW, or CLDEMOTE
28320 extended instruction sets. Each has a corresponding @option{-mno-} option to
28321 disable use of these instructions.
28323 These extensions are also available as built-in functions: see
28324 @ref{x86 Built-in Functions}, for details of the functions enabled and
28325 disabled by these switches.
28327 To generate SSE/SSE2 instructions automatically from floating-point
28328 code (as opposed to 387 instructions), see @option{-mfpmath=sse}.
28330 GCC depresses SSEx instructions when @option{-mavx} is used. Instead, it
28331 generates new AVX instructions or AVX equivalence for all SSEx instructions
28334 These options enable GCC to use these extended instructions in
28335 generated code, even without @option{-mfpmath=sse}. Applications that
28336 perform run-time CPU detection must compile separate files for each
28337 supported architecture, using the appropriate flags. In particular,
28338 the file containing the CPU detection code should be compiled without
28341 @item -mdump-tune-features
28342 @opindex mdump-tune-features
28343 This option instructs GCC to dump the names of the x86 performance
28344 tuning features and default settings. The names can be used in
28345 @option{-mtune-ctrl=@var{feature-list}}.
28347 @item -mtune-ctrl=@var{feature-list}
28348 @opindex mtune-ctrl=@var{feature-list}
28349 This option is used to do fine grain control of x86 code generation features.
28350 @var{feature-list} is a comma separated list of @var{feature} names. See also
28351 @option{-mdump-tune-features}. When specified, the @var{feature} is turned
28352 on if it is not preceded with @samp{^}, otherwise, it is turned off.
28353 @option{-mtune-ctrl=@var{feature-list}} is intended to be used by GCC
28354 developers. Using it may lead to code paths not covered by testing and can
28355 potentially result in compiler ICEs or runtime errors.
28358 @opindex mno-default
28359 This option instructs GCC to turn off all tunable features. See also
28360 @option{-mtune-ctrl=@var{feature-list}} and @option{-mdump-tune-features}.
28364 This option instructs GCC to emit a @code{cld} instruction in the prologue
28365 of functions that use string instructions. String instructions depend on
28366 the DF flag to select between autoincrement or autodecrement mode. While the
28367 ABI specifies the DF flag to be cleared on function entry, some operating
28368 systems violate this specification by not clearing the DF flag in their
28369 exception dispatchers. The exception handler can be invoked with the DF flag
28370 set, which leads to wrong direction mode when string instructions are used.
28371 This option can be enabled by default on 32-bit x86 targets by configuring
28372 GCC with the @option{--enable-cld} configure option. Generation of @code{cld}
28373 instructions can be suppressed with the @option{-mno-cld} compiler option
28377 @opindex mvzeroupper
28378 This option instructs GCC to emit a @code{vzeroupper} instruction
28379 before a transfer of control flow out of the function to minimize
28380 the AVX to SSE transition penalty as well as remove unnecessary @code{zeroupper}
28383 @item -mprefer-avx128
28384 @opindex mprefer-avx128
28385 This option instructs GCC to use 128-bit AVX instructions instead of
28386 256-bit AVX instructions in the auto-vectorizer.
28388 @item -mprefer-vector-width=@var{opt}
28389 @opindex mprefer-vector-width
28390 This option instructs GCC to use @var{opt}-bit vector width in instructions
28391 instead of default on the selected platform.
28395 No extra limitations applied to GCC other than defined by the selected platform.
28398 Prefer 128-bit vector width for instructions.
28401 Prefer 256-bit vector width for instructions.
28404 Prefer 512-bit vector width for instructions.
28409 This option enables GCC to generate @code{CMPXCHG16B} instructions in 64-bit
28410 code to implement compare-and-exchange operations on 16-byte aligned 128-bit
28411 objects. This is useful for atomic updates of data structures exceeding one
28412 machine word in size. The compiler uses this instruction to implement
28413 @ref{__sync Builtins}. However, for @ref{__atomic Builtins} operating on
28414 128-bit integers, a library call is always used.
28418 This option enables generation of @code{SAHF} instructions in 64-bit code.
28419 Early Intel Pentium 4 CPUs with Intel 64 support,
28420 prior to the introduction of Pentium 4 G1 step in December 2005,
28421 lacked the @code{LAHF} and @code{SAHF} instructions
28422 which are supported by AMD64.
28423 These are load and store instructions, respectively, for certain status flags.
28424 In 64-bit mode, the @code{SAHF} instruction is used to optimize @code{fmod},
28425 @code{drem}, and @code{remainder} built-in functions;
28426 see @ref{Other Builtins} for details.
28430 This option enables use of the @code{movbe} instruction to implement
28431 @code{__builtin_bswap32} and @code{__builtin_bswap64}.
28435 The @option{-mshstk} option enables shadow stack built-in functions
28436 from x86 Control-flow Enforcement Technology (CET).
28440 This option enables built-in functions @code{__builtin_ia32_crc32qi},
28441 @code{__builtin_ia32_crc32hi}, @code{__builtin_ia32_crc32si} and
28442 @code{__builtin_ia32_crc32di} to generate the @code{crc32} machine instruction.
28446 This option enables use of @code{RCPSS} and @code{RSQRTSS} instructions
28447 (and their vectorized variants @code{RCPPS} and @code{RSQRTPS})
28448 with an additional Newton-Raphson step
28449 to increase precision instead of @code{DIVSS} and @code{SQRTSS}
28450 (and their vectorized
28451 variants) for single-precision floating-point arguments. These instructions
28452 are generated only when @option{-funsafe-math-optimizations} is enabled
28453 together with @option{-ffinite-math-only} and @option{-fno-trapping-math}.
28454 Note that while the throughput of the sequence is higher than the throughput
28455 of the non-reciprocal instruction, the precision of the sequence can be
28456 decreased by up to 2 ulp (i.e.@: the inverse of 1.0 equals 0.99999994).
28458 Note that GCC implements @code{1.0f/sqrtf(@var{x})} in terms of @code{RSQRTSS}
28459 (or @code{RSQRTPS}) already with @option{-ffast-math} (or the above option
28460 combination), and doesn't need @option{-mrecip}.
28462 Also note that GCC emits the above sequence with additional Newton-Raphson step
28463 for vectorized single-float division and vectorized @code{sqrtf(@var{x})}
28464 already with @option{-ffast-math} (or the above option combination), and
28465 doesn't need @option{-mrecip}.
28467 @item -mrecip=@var{opt}
28468 @opindex mrecip=opt
28469 This option controls which reciprocal estimate instructions
28470 may be used. @var{opt} is a comma-separated list of options, which may
28471 be preceded by a @samp{!} to invert the option:
28475 Enable all estimate instructions.
28478 Enable the default instructions, equivalent to @option{-mrecip}.
28481 Disable all estimate instructions, equivalent to @option{-mno-recip}.
28484 Enable the approximation for scalar division.
28487 Enable the approximation for vectorized division.
28490 Enable the approximation for scalar square root.
28493 Enable the approximation for vectorized square root.
28496 So, for example, @option{-mrecip=all,!sqrt} enables
28497 all of the reciprocal approximations, except for square root.
28499 @item -mveclibabi=@var{type}
28500 @opindex mveclibabi
28501 Specifies the ABI type to use for vectorizing intrinsics using an
28502 external library. Supported values for @var{type} are @samp{svml}
28503 for the Intel short
28504 vector math library and @samp{acml} for the AMD math core library.
28505 To use this option, both @option{-ftree-vectorize} and
28506 @option{-funsafe-math-optimizations} have to be enabled, and an SVML or ACML
28507 ABI-compatible library must be specified at link time.
28509 GCC currently emits calls to @code{vmldExp2},
28510 @code{vmldLn2}, @code{vmldLog102}, @code{vmldPow2},
28511 @code{vmldTanh2}, @code{vmldTan2}, @code{vmldAtan2}, @code{vmldAtanh2},
28512 @code{vmldCbrt2}, @code{vmldSinh2}, @code{vmldSin2}, @code{vmldAsinh2},
28513 @code{vmldAsin2}, @code{vmldCosh2}, @code{vmldCos2}, @code{vmldAcosh2},
28514 @code{vmldAcos2}, @code{vmlsExp4}, @code{vmlsLn4},
28515 @code{vmlsLog104}, @code{vmlsPow4}, @code{vmlsTanh4}, @code{vmlsTan4},
28516 @code{vmlsAtan4}, @code{vmlsAtanh4}, @code{vmlsCbrt4}, @code{vmlsSinh4},
28517 @code{vmlsSin4}, @code{vmlsAsinh4}, @code{vmlsAsin4}, @code{vmlsCosh4},
28518 @code{vmlsCos4}, @code{vmlsAcosh4} and @code{vmlsAcos4} for corresponding
28519 function type when @option{-mveclibabi=svml} is used, and @code{__vrd2_sin},
28520 @code{__vrd2_cos}, @code{__vrd2_exp}, @code{__vrd2_log}, @code{__vrd2_log2},
28521 @code{__vrd2_log10}, @code{__vrs4_sinf}, @code{__vrs4_cosf},
28522 @code{__vrs4_expf}, @code{__vrs4_logf}, @code{__vrs4_log2f},
28523 @code{__vrs4_log10f} and @code{__vrs4_powf} for the corresponding function type
28524 when @option{-mveclibabi=acml} is used.
28526 @item -mabi=@var{name}
28528 Generate code for the specified calling convention. Permissible values
28529 are @samp{sysv} for the ABI used on GNU/Linux and other systems, and
28530 @samp{ms} for the Microsoft ABI. The default is to use the Microsoft
28531 ABI when targeting Microsoft Windows and the SysV ABI on all other systems.
28532 You can control this behavior for specific functions by
28533 using the function attributes @code{ms_abi} and @code{sysv_abi}.
28534 @xref{Function Attributes}.
28536 @item -mforce-indirect-call
28537 @opindex mforce-indirect-call
28538 Force all calls to functions to be indirect. This is useful
28539 when using Intel Processor Trace where it generates more precise timing
28540 information for function calls.
28542 @item -mmanual-endbr
28543 @opindex mmanual-endbr
28544 Insert ENDBR instruction at function entry only via the @code{cf_check}
28545 function attribute. This is useful when used with the option
28546 @option{-fcf-protection=branch} to control ENDBR insertion at the
28549 @item -mcall-ms2sysv-xlogues
28550 @opindex mcall-ms2sysv-xlogues
28551 @opindex mno-call-ms2sysv-xlogues
28552 Due to differences in 64-bit ABIs, any Microsoft ABI function that calls a
28553 System V ABI function must consider RSI, RDI and XMM6-15 as clobbered. By
28554 default, the code for saving and restoring these registers is emitted inline,
28555 resulting in fairly lengthy prologues and epilogues. Using
28556 @option{-mcall-ms2sysv-xlogues} emits prologues and epilogues that
28557 use stubs in the static portion of libgcc to perform these saves and restores,
28558 thus reducing function size at the cost of a few extra instructions.
28560 @item -mtls-dialect=@var{type}
28561 @opindex mtls-dialect
28562 Generate code to access thread-local storage using the @samp{gnu} or
28563 @samp{gnu2} conventions. @samp{gnu} is the conservative default;
28564 @samp{gnu2} is more efficient, but it may add compile- and run-time
28565 requirements that cannot be satisfied on all systems.
28568 @itemx -mno-push-args
28569 @opindex mpush-args
28570 @opindex mno-push-args
28571 Use PUSH operations to store outgoing parameters. This method is shorter
28572 and usually equally fast as method using SUB/MOV operations and is enabled
28573 by default. In some cases disabling it may improve performance because of
28574 improved scheduling and reduced dependencies.
28576 @item -maccumulate-outgoing-args
28577 @opindex maccumulate-outgoing-args
28578 If enabled, the maximum amount of space required for outgoing arguments is
28579 computed in the function prologue. This is faster on most modern CPUs
28580 because of reduced dependencies, improved scheduling and reduced stack usage
28581 when the preferred stack boundary is not equal to 2. The drawback is a notable
28582 increase in code size. This switch implies @option{-mno-push-args}.
28586 Support thread-safe exception handling on MinGW. Programs that rely
28587 on thread-safe exception handling must compile and link all code with the
28588 @option{-mthreads} option. When compiling, @option{-mthreads} defines
28589 @option{-D_MT}; when linking, it links in a special thread helper library
28590 @option{-lmingwthrd} which cleans up per-thread exception-handling data.
28592 @item -mms-bitfields
28593 @itemx -mno-ms-bitfields
28594 @opindex mms-bitfields
28595 @opindex mno-ms-bitfields
28597 Enable/disable bit-field layout compatible with the native Microsoft
28600 If @code{packed} is used on a structure, or if bit-fields are used,
28601 it may be that the Microsoft ABI lays out the structure differently
28602 than the way GCC normally does. Particularly when moving packed
28603 data between functions compiled with GCC and the native Microsoft compiler
28604 (either via function call or as data in a file), it may be necessary to access
28607 This option is enabled by default for Microsoft Windows
28608 targets. This behavior can also be controlled locally by use of variable
28609 or type attributes. For more information, see @ref{x86 Variable Attributes}
28610 and @ref{x86 Type Attributes}.
28612 The Microsoft structure layout algorithm is fairly simple with the exception
28613 of the bit-field packing.
28614 The padding and alignment of members of structures and whether a bit-field
28615 can straddle a storage-unit boundary are determine by these rules:
28618 @item Structure members are stored sequentially in the order in which they are
28619 declared: the first member has the lowest memory address and the last member
28622 @item Every data object has an alignment requirement. The alignment requirement
28623 for all data except structures, unions, and arrays is either the size of the
28624 object or the current packing size (specified with either the
28625 @code{aligned} attribute or the @code{pack} pragma),
28626 whichever is less. For structures, unions, and arrays,
28627 the alignment requirement is the largest alignment requirement of its members.
28628 Every object is allocated an offset so that:
28631 offset % alignment_requirement == 0
28634 @item Adjacent bit-fields are packed into the same 1-, 2-, or 4-byte allocation
28635 unit if the integral types are the same size and if the next bit-field fits
28636 into the current allocation unit without crossing the boundary imposed by the
28637 common alignment requirements of the bit-fields.
28640 MSVC interprets zero-length bit-fields in the following ways:
28643 @item If a zero-length bit-field is inserted between two bit-fields that
28644 are normally coalesced, the bit-fields are not coalesced.
28651 unsigned long bf_1 : 12;
28653 unsigned long bf_2 : 12;
28658 The size of @code{t1} is 8 bytes with the zero-length bit-field. If the
28659 zero-length bit-field were removed, @code{t1}'s size would be 4 bytes.
28661 @item If a zero-length bit-field is inserted after a bit-field, @code{foo}, and the
28662 alignment of the zero-length bit-field is greater than the member that follows it,
28663 @code{bar}, @code{bar} is aligned as the type of the zero-length bit-field.
28684 For @code{t2}, @code{bar} is placed at offset 2, rather than offset 1.
28685 Accordingly, the size of @code{t2} is 4. For @code{t3}, the zero-length
28686 bit-field does not affect the alignment of @code{bar} or, as a result, the size
28689 Taking this into account, it is important to note the following:
28692 @item If a zero-length bit-field follows a normal bit-field, the type of the
28693 zero-length bit-field may affect the alignment of the structure as whole. For
28694 example, @code{t2} has a size of 4 bytes, since the zero-length bit-field follows a
28695 normal bit-field, and is of type short.
28697 @item Even if a zero-length bit-field is not followed by a normal bit-field, it may
28698 still affect the alignment of the structure:
28709 Here, @code{t4} takes up 4 bytes.
28712 @item Zero-length bit-fields following non-bit-field members are ignored:
28724 Here, @code{t5} takes up 2 bytes.
28728 @item -mno-align-stringops
28729 @opindex mno-align-stringops
28730 @opindex malign-stringops
28731 Do not align the destination of inlined string operations. This switch reduces
28732 code size and improves performance in case the destination is already aligned,
28733 but GCC doesn't know about it.
28735 @item -minline-all-stringops
28736 @opindex minline-all-stringops
28737 By default GCC inlines string operations only when the destination is
28738 known to be aligned to least a 4-byte boundary.
28739 This enables more inlining and increases code
28740 size, but may improve performance of code that depends on fast
28741 @code{memcpy} and @code{memset} for short lengths.
28742 The option enables inline expansion of @code{strlen} for all
28743 pointer alignments.
28745 @item -minline-stringops-dynamically
28746 @opindex minline-stringops-dynamically
28747 For string operations of unknown size, use run-time checks with
28748 inline code for small blocks and a library call for large blocks.
28750 @item -mstringop-strategy=@var{alg}
28751 @opindex mstringop-strategy=@var{alg}
28752 Override the internal decision heuristic for the particular algorithm to use
28753 for inlining string operations. The allowed values for @var{alg} are:
28759 Expand using i386 @code{rep} prefix of the specified size.
28763 @itemx unrolled_loop
28764 Expand into an inline loop.
28767 Always use a library call.
28770 @item -mmemcpy-strategy=@var{strategy}
28771 @opindex mmemcpy-strategy=@var{strategy}
28772 Override the internal decision heuristic to decide if @code{__builtin_memcpy}
28773 should be inlined and what inline algorithm to use when the expected size
28774 of the copy operation is known. @var{strategy}
28775 is a comma-separated list of @var{alg}:@var{max_size}:@var{dest_align} triplets.
28776 @var{alg} is specified in @option{-mstringop-strategy}, @var{max_size} specifies
28777 the max byte size with which inline algorithm @var{alg} is allowed. For the last
28778 triplet, the @var{max_size} must be @code{-1}. The @var{max_size} of the triplets
28779 in the list must be specified in increasing order. The minimal byte size for
28780 @var{alg} is @code{0} for the first triplet and @code{@var{max_size} + 1} of the
28783 @item -mmemset-strategy=@var{strategy}
28784 @opindex mmemset-strategy=@var{strategy}
28785 The option is similar to @option{-mmemcpy-strategy=} except that it is to control
28786 @code{__builtin_memset} expansion.
28788 @item -momit-leaf-frame-pointer
28789 @opindex momit-leaf-frame-pointer
28790 Don't keep the frame pointer in a register for leaf functions. This
28791 avoids the instructions to save, set up, and restore frame pointers and
28792 makes an extra register available in leaf functions. The option
28793 @option{-fomit-leaf-frame-pointer} removes the frame pointer for leaf functions,
28794 which might make debugging harder.
28796 @item -mtls-direct-seg-refs
28797 @itemx -mno-tls-direct-seg-refs
28798 @opindex mtls-direct-seg-refs
28799 Controls whether TLS variables may be accessed with offsets from the
28800 TLS segment register (@code{%gs} for 32-bit, @code{%fs} for 64-bit),
28801 or whether the thread base pointer must be added. Whether or not this
28802 is valid depends on the operating system, and whether it maps the
28803 segment to cover the entire TLS area.
28805 For systems that use the GNU C Library, the default is on.
28808 @itemx -mno-sse2avx
28810 Specify that the assembler should encode SSE instructions with VEX
28811 prefix. The option @option{-mavx} turns this on by default.
28816 If profiling is active (@option{-pg}), put the profiling
28817 counter call before the prologue.
28818 Note: On x86 architectures the attribute @code{ms_hook_prologue}
28819 isn't possible at the moment for @option{-mfentry} and @option{-pg}.
28821 @item -mrecord-mcount
28822 @itemx -mno-record-mcount
28823 @opindex mrecord-mcount
28824 If profiling is active (@option{-pg}), generate a __mcount_loc section
28825 that contains pointers to each profiling call. This is useful for
28826 automatically patching and out calls.
28829 @itemx -mno-nop-mcount
28830 @opindex mnop-mcount
28831 If profiling is active (@option{-pg}), generate the calls to
28832 the profiling functions as NOPs. This is useful when they
28833 should be patched in later dynamically. This is likely only
28834 useful together with @option{-mrecord-mcount}.
28836 @item -minstrument-return=@var{type}
28837 @opindex minstrument-return
28838 Instrument function exit in -pg -mfentry instrumented functions with
28839 call to specified function. This only instruments true returns ending
28840 with ret, but not sibling calls ending with jump. Valid types
28841 are @var{none} to not instrument, @var{call} to generate a call to __return__,
28842 or @var{nop5} to generate a 5 byte nop.
28844 @item -mrecord-return
28845 @itemx -mno-record-return
28846 @opindex mrecord-return
28847 Generate a __return_loc section pointing to all return instrumentation code.
28849 @item -mfentry-name=@var{name}
28850 @opindex mfentry-name
28851 Set name of __fentry__ symbol called at function entry for -pg -mfentry functions.
28853 @item -mfentry-section=@var{name}
28854 @opindex mfentry-section
28855 Set name of section to record -mrecord-mcount calls (default __mcount_loc).
28857 @item -mskip-rax-setup
28858 @itemx -mno-skip-rax-setup
28859 @opindex mskip-rax-setup
28860 When generating code for the x86-64 architecture with SSE extensions
28861 disabled, @option{-mskip-rax-setup} can be used to skip setting up RAX
28862 register when there are no variable arguments passed in vector registers.
28864 @strong{Warning:} Since RAX register is used to avoid unnecessarily
28865 saving vector registers on stack when passing variable arguments, the
28866 impacts of this option are callees may waste some stack space,
28867 misbehave or jump to a random location. GCC 4.4 or newer don't have
28868 those issues, regardless the RAX register value.
28871 @itemx -mno-8bit-idiv
28872 @opindex m8bit-idiv
28873 On some processors, like Intel Atom, 8-bit unsigned integer divide is
28874 much faster than 32-bit/64-bit integer divide. This option generates a
28875 run-time check. If both dividend and divisor are within range of 0
28876 to 255, 8-bit unsigned integer divide is used instead of
28877 32-bit/64-bit integer divide.
28879 @item -mavx256-split-unaligned-load
28880 @itemx -mavx256-split-unaligned-store
28881 @opindex mavx256-split-unaligned-load
28882 @opindex mavx256-split-unaligned-store
28883 Split 32-byte AVX unaligned load and store.
28885 @item -mstack-protector-guard=@var{guard}
28886 @itemx -mstack-protector-guard-reg=@var{reg}
28887 @itemx -mstack-protector-guard-offset=@var{offset}
28888 @opindex mstack-protector-guard
28889 @opindex mstack-protector-guard-reg
28890 @opindex mstack-protector-guard-offset
28891 Generate stack protection code using canary at @var{guard}. Supported
28892 locations are @samp{global} for global canary or @samp{tls} for per-thread
28893 canary in the TLS block (the default). This option has effect only when
28894 @option{-fstack-protector} or @option{-fstack-protector-all} is specified.
28896 With the latter choice the options
28897 @option{-mstack-protector-guard-reg=@var{reg}} and
28898 @option{-mstack-protector-guard-offset=@var{offset}} furthermore specify
28899 which segment register (@code{%fs} or @code{%gs}) to use as base register
28900 for reading the canary, and from what offset from that base register.
28901 The default for those is as specified in the relevant ABI.
28903 @item -mgeneral-regs-only
28904 @opindex mgeneral-regs-only
28905 Generate code that uses only the general-purpose registers. This
28906 prevents the compiler from using floating-point, vector, mask and bound
28909 @item -mindirect-branch=@var{choice}
28910 @opindex mindirect-branch
28911 Convert indirect call and jump with @var{choice}. The default is
28912 @samp{keep}, which keeps indirect call and jump unmodified.
28913 @samp{thunk} converts indirect call and jump to call and return thunk.
28914 @samp{thunk-inline} converts indirect call and jump to inlined call
28915 and return thunk. @samp{thunk-extern} converts indirect call and jump
28916 to external call and return thunk provided in a separate object file.
28917 You can control this behavior for a specific function by using the
28918 function attribute @code{indirect_branch}. @xref{Function Attributes}.
28920 Note that @option{-mcmodel=large} is incompatible with
28921 @option{-mindirect-branch=thunk} and
28922 @option{-mindirect-branch=thunk-extern} since the thunk function may
28923 not be reachable in the large code model.
28925 Note that @option{-mindirect-branch=thunk-extern} is incompatible with
28926 @option{-fcf-protection=branch} since the external thunk cannot be modified
28927 to disable control-flow check.
28929 @item -mfunction-return=@var{choice}
28930 @opindex mfunction-return
28931 Convert function return with @var{choice}. The default is @samp{keep},
28932 which keeps function return unmodified. @samp{thunk} converts function
28933 return to call and return thunk. @samp{thunk-inline} converts function
28934 return to inlined call and return thunk. @samp{thunk-extern} converts
28935 function return to external call and return thunk provided in a separate
28936 object file. You can control this behavior for a specific function by
28937 using the function attribute @code{function_return}.
28938 @xref{Function Attributes}.
28940 Note that @option{-mcmodel=large} is incompatible with
28941 @option{-mfunction-return=thunk} and
28942 @option{-mfunction-return=thunk-extern} since the thunk function may
28943 not be reachable in the large code model.
28946 @item -mindirect-branch-register
28947 @opindex mindirect-branch-register
28948 Force indirect call and jump via register.
28952 These @samp{-m} switches are supported in addition to the above
28953 on x86-64 processors in 64-bit environments.
28966 Generate code for a 16-bit, 32-bit or 64-bit environment.
28967 The @option{-m32} option sets @code{int}, @code{long}, and pointer types
28969 generates code that runs on any i386 system.
28971 The @option{-m64} option sets @code{int} to 32 bits and @code{long} and pointer
28972 types to 64 bits, and generates code for the x86-64 architecture.
28973 For Darwin only the @option{-m64} option also turns off the @option{-fno-pic}
28974 and @option{-mdynamic-no-pic} options.
28976 The @option{-mx32} option sets @code{int}, @code{long}, and pointer types
28978 generates code for the x86-64 architecture.
28980 The @option{-m16} option is the same as @option{-m32}, except for that
28981 it outputs the @code{.code16gcc} assembly directive at the beginning of
28982 the assembly output so that the binary can run in 16-bit mode.
28984 The @option{-miamcu} option generates code which conforms to Intel MCU
28985 psABI. It requires the @option{-m32} option to be turned on.
28987 @item -mno-red-zone
28988 @opindex mno-red-zone
28990 Do not use a so-called ``red zone'' for x86-64 code. The red zone is mandated
28991 by the x86-64 ABI; it is a 128-byte area beyond the location of the
28992 stack pointer that is not modified by signal or interrupt handlers
28993 and therefore can be used for temporary data without adjusting the stack
28994 pointer. The flag @option{-mno-red-zone} disables this red zone.
28996 @item -mcmodel=small
28997 @opindex mcmodel=small
28998 Generate code for the small code model: the program and its symbols must
28999 be linked in the lower 2 GB of the address space. Pointers are 64 bits.
29000 Programs can be statically or dynamically linked. This is the default
29003 @item -mcmodel=kernel
29004 @opindex mcmodel=kernel
29005 Generate code for the kernel code model. The kernel runs in the
29006 negative 2 GB of the address space.
29007 This model has to be used for Linux kernel code.
29009 @item -mcmodel=medium
29010 @opindex mcmodel=medium
29011 Generate code for the medium model: the program is linked in the lower 2
29012 GB of the address space. Small symbols are also placed there. Symbols
29013 with sizes larger than @option{-mlarge-data-threshold} are put into
29014 large data or BSS sections and can be located above 2GB. Programs can
29015 be statically or dynamically linked.
29017 @item -mcmodel=large
29018 @opindex mcmodel=large
29019 Generate code for the large model. This model makes no assumptions
29020 about addresses and sizes of sections.
29022 @item -maddress-mode=long
29023 @opindex maddress-mode=long
29024 Generate code for long address mode. This is only supported for 64-bit
29025 and x32 environments. It is the default address mode for 64-bit
29028 @item -maddress-mode=short
29029 @opindex maddress-mode=short
29030 Generate code for short address mode. This is only supported for 32-bit
29031 and x32 environments. It is the default address mode for 32-bit and
29035 @node x86 Windows Options
29036 @subsection x86 Windows Options
29037 @cindex x86 Windows Options
29038 @cindex Windows Options for x86
29040 These additional options are available for Microsoft Windows targets:
29046 specifies that a console application is to be generated, by
29047 instructing the linker to set the PE header subsystem type
29048 required for console applications.
29049 This option is available for Cygwin and MinGW targets and is
29050 enabled by default on those targets.
29054 This option is available for Cygwin and MinGW targets. It
29055 specifies that a DLL---a dynamic link library---is to be
29056 generated, enabling the selection of the required runtime
29057 startup object and entry point.
29059 @item -mnop-fun-dllimport
29060 @opindex mnop-fun-dllimport
29061 This option is available for Cygwin and MinGW targets. It
29062 specifies that the @code{dllimport} attribute should be ignored.
29066 This option is available for MinGW targets. It specifies
29067 that MinGW-specific thread support is to be used.
29071 This option is available for MinGW-w64 targets. It causes
29072 the @code{UNICODE} preprocessor macro to be predefined, and
29073 chooses Unicode-capable runtime startup code.
29077 This option is available for Cygwin and MinGW targets. It
29078 specifies that the typical Microsoft Windows predefined macros are to
29079 be set in the pre-processor, but does not influence the choice
29080 of runtime library/startup code.
29084 This option is available for Cygwin and MinGW targets. It
29085 specifies that a GUI application is to be generated by
29086 instructing the linker to set the PE header subsystem type
29089 @item -fno-set-stack-executable
29090 @opindex fno-set-stack-executable
29091 @opindex fset-stack-executable
29092 This option is available for MinGW targets. It specifies that
29093 the executable flag for the stack used by nested functions isn't
29094 set. This is necessary for binaries running in kernel mode of
29095 Microsoft Windows, as there the User32 API, which is used to set executable
29096 privileges, isn't available.
29098 @item -fwritable-relocated-rdata
29099 @opindex fno-writable-relocated-rdata
29100 @opindex fwritable-relocated-rdata
29101 This option is available for MinGW and Cygwin targets. It specifies
29102 that relocated-data in read-only section is put into the @code{.data}
29103 section. This is a necessary for older runtimes not supporting
29104 modification of @code{.rdata} sections for pseudo-relocation.
29106 @item -mpe-aligned-commons
29107 @opindex mpe-aligned-commons
29108 This option is available for Cygwin and MinGW targets. It
29109 specifies that the GNU extension to the PE file format that
29110 permits the correct alignment of COMMON variables should be
29111 used when generating code. It is enabled by default if
29112 GCC detects that the target assembler found during configuration
29113 supports the feature.
29116 See also under @ref{x86 Options} for standard options.
29118 @node Xstormy16 Options
29119 @subsection Xstormy16 Options
29120 @cindex Xstormy16 Options
29122 These options are defined for Xstormy16:
29127 Choose startup files and linker script suitable for the simulator.
29130 @node Xtensa Options
29131 @subsection Xtensa Options
29132 @cindex Xtensa Options
29134 These options are supported for Xtensa targets:
29138 @itemx -mno-const16
29140 @opindex mno-const16
29141 Enable or disable use of @code{CONST16} instructions for loading
29142 constant values. The @code{CONST16} instruction is currently not a
29143 standard option from Tensilica. When enabled, @code{CONST16}
29144 instructions are always used in place of the standard @code{L32R}
29145 instructions. The use of @code{CONST16} is enabled by default only if
29146 the @code{L32R} instruction is not available.
29149 @itemx -mno-fused-madd
29150 @opindex mfused-madd
29151 @opindex mno-fused-madd
29152 Enable or disable use of fused multiply/add and multiply/subtract
29153 instructions in the floating-point option. This has no effect if the
29154 floating-point option is not also enabled. Disabling fused multiply/add
29155 and multiply/subtract instructions forces the compiler to use separate
29156 instructions for the multiply and add/subtract operations. This may be
29157 desirable in some cases where strict IEEE 754-compliant results are
29158 required: the fused multiply add/subtract instructions do not round the
29159 intermediate result, thereby producing results with @emph{more} bits of
29160 precision than specified by the IEEE standard. Disabling fused multiply
29161 add/subtract instructions also ensures that the program output is not
29162 sensitive to the compiler's ability to combine multiply and add/subtract
29165 @item -mserialize-volatile
29166 @itemx -mno-serialize-volatile
29167 @opindex mserialize-volatile
29168 @opindex mno-serialize-volatile
29169 When this option is enabled, GCC inserts @code{MEMW} instructions before
29170 @code{volatile} memory references to guarantee sequential consistency.
29171 The default is @option{-mserialize-volatile}. Use
29172 @option{-mno-serialize-volatile} to omit the @code{MEMW} instructions.
29174 @item -mforce-no-pic
29175 @opindex mforce-no-pic
29176 For targets, like GNU/Linux, where all user-mode Xtensa code must be
29177 position-independent code (PIC), this option disables PIC for compiling
29180 @item -mtext-section-literals
29181 @itemx -mno-text-section-literals
29182 @opindex mtext-section-literals
29183 @opindex mno-text-section-literals
29184 These options control the treatment of literal pools. The default is
29185 @option{-mno-text-section-literals}, which places literals in a separate
29186 section in the output file. This allows the literal pool to be placed
29187 in a data RAM/ROM, and it also allows the linker to combine literal
29188 pools from separate object files to remove redundant literals and
29189 improve code size. With @option{-mtext-section-literals}, the literals
29190 are interspersed in the text section in order to keep them as close as
29191 possible to their references. This may be necessary for large assembly
29192 files. Literals for each function are placed right before that function.
29194 @item -mauto-litpools
29195 @itemx -mno-auto-litpools
29196 @opindex mauto-litpools
29197 @opindex mno-auto-litpools
29198 These options control the treatment of literal pools. The default is
29199 @option{-mno-auto-litpools}, which places literals in a separate
29200 section in the output file unless @option{-mtext-section-literals} is
29201 used. With @option{-mauto-litpools} the literals are interspersed in
29202 the text section by the assembler. Compiler does not produce explicit
29203 @code{.literal} directives and loads literals into registers with
29204 @code{MOVI} instructions instead of @code{L32R} to let the assembler
29205 do relaxation and place literals as necessary. This option allows
29206 assembler to create several literal pools per function and assemble
29207 very big functions, which may not be possible with
29208 @option{-mtext-section-literals}.
29210 @item -mtarget-align
29211 @itemx -mno-target-align
29212 @opindex mtarget-align
29213 @opindex mno-target-align
29214 When this option is enabled, GCC instructs the assembler to
29215 automatically align instructions to reduce branch penalties at the
29216 expense of some code density. The assembler attempts to widen density
29217 instructions to align branch targets and the instructions following call
29218 instructions. If there are not enough preceding safe density
29219 instructions to align a target, no widening is performed. The
29220 default is @option{-mtarget-align}. These options do not affect the
29221 treatment of auto-aligned instructions like @code{LOOP}, which the
29222 assembler always aligns, either by widening density instructions or
29223 by inserting NOP instructions.
29226 @itemx -mno-longcalls
29227 @opindex mlongcalls
29228 @opindex mno-longcalls
29229 When this option is enabled, GCC instructs the assembler to translate
29230 direct calls to indirect calls unless it can determine that the target
29231 of a direct call is in the range allowed by the call instruction. This
29232 translation typically occurs for calls to functions in other source
29233 files. Specifically, the assembler translates a direct @code{CALL}
29234 instruction into an @code{L32R} followed by a @code{CALLX} instruction.
29235 The default is @option{-mno-longcalls}. This option should be used in
29236 programs where the call target can potentially be out of range. This
29237 option is implemented in the assembler, not the compiler, so the
29238 assembly code generated by GCC still shows direct call
29239 instructions---look at the disassembled object code to see the actual
29240 instructions. Note that the assembler uses an indirect call for
29241 every cross-file call, not just those that really are out of range.
29244 @node zSeries Options
29245 @subsection zSeries Options
29246 @cindex zSeries options
29248 These are listed under @xref{S/390 and zSeries Options}.
29254 @section Specifying Subprocesses and the Switches to Pass to Them
29257 @command{gcc} is a driver program. It performs its job by invoking a
29258 sequence of other programs to do the work of compiling, assembling and
29259 linking. GCC interprets its command-line parameters and uses these to
29260 deduce which programs it should invoke, and which command-line options
29261 it ought to place on their command lines. This behavior is controlled
29262 by @dfn{spec strings}. In most cases there is one spec string for each
29263 program that GCC can invoke, but a few programs have multiple spec
29264 strings to control their behavior. The spec strings built into GCC can
29265 be overridden by using the @option{-specs=} command-line switch to specify
29268 @dfn{Spec files} are plain-text files that are used to construct spec
29269 strings. They consist of a sequence of directives separated by blank
29270 lines. The type of directive is determined by the first non-whitespace
29271 character on the line, which can be one of the following:
29274 @item %@var{command}
29275 Issues a @var{command} to the spec file processor. The commands that can
29279 @item %include <@var{file}>
29280 @cindex @code{%include}
29281 Search for @var{file} and insert its text at the current point in the
29284 @item %include_noerr <@var{file}>
29285 @cindex @code{%include_noerr}
29286 Just like @samp{%include}, but do not generate an error message if the include
29287 file cannot be found.
29289 @item %rename @var{old_name} @var{new_name}
29290 @cindex @code{%rename}
29291 Rename the spec string @var{old_name} to @var{new_name}.
29295 @item *[@var{spec_name}]:
29296 This tells the compiler to create, override or delete the named spec
29297 string. All lines after this directive up to the next directive or
29298 blank line are considered to be the text for the spec string. If this
29299 results in an empty string then the spec is deleted. (Or, if the
29300 spec did not exist, then nothing happens.) Otherwise, if the spec
29301 does not currently exist a new spec is created. If the spec does
29302 exist then its contents are overridden by the text of this
29303 directive, unless the first character of that text is the @samp{+}
29304 character, in which case the text is appended to the spec.
29306 @item [@var{suffix}]:
29307 Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive
29308 and up to the next directive or blank line are considered to make up the
29309 spec string for the indicated suffix. When the compiler encounters an
29310 input file with the named suffix, it processes the spec string in
29311 order to work out how to compile that file. For example:
29315 z-compile -input %i
29318 This says that any input file whose name ends in @samp{.ZZ} should be
29319 passed to the program @samp{z-compile}, which should be invoked with the
29320 command-line switch @option{-input} and with the result of performing the
29321 @samp{%i} substitution. (See below.)
29323 As an alternative to providing a spec string, the text following a
29324 suffix directive can be one of the following:
29327 @item @@@var{language}
29328 This says that the suffix is an alias for a known @var{language}. This is
29329 similar to using the @option{-x} command-line switch to GCC to specify a
29330 language explicitly. For example:
29337 Says that .ZZ files are, in fact, C++ source files.
29340 This causes an error messages saying:
29343 @var{name} compiler not installed on this system.
29347 GCC already has an extensive list of suffixes built into it.
29348 This directive adds an entry to the end of the list of suffixes, but
29349 since the list is searched from the end backwards, it is effectively
29350 possible to override earlier entries using this technique.
29354 GCC has the following spec strings built into it. Spec files can
29355 override these strings or create their own. Note that individual
29356 targets can also add their own spec strings to this list.
29359 asm Options to pass to the assembler
29360 asm_final Options to pass to the assembler post-processor
29361 cpp Options to pass to the C preprocessor
29362 cc1 Options to pass to the C compiler
29363 cc1plus Options to pass to the C++ compiler
29364 endfile Object files to include at the end of the link
29365 link Options to pass to the linker
29366 lib Libraries to include on the command line to the linker
29367 libgcc Decides which GCC support library to pass to the linker
29368 linker Sets the name of the linker
29369 predefines Defines to be passed to the C preprocessor
29370 signed_char Defines to pass to CPP to say whether @code{char} is signed
29372 startfile Object files to include at the start of the link
29375 Here is a small example of a spec file:
29378 %rename lib old_lib
29381 --start-group -lgcc -lc -leval1 --end-group %(old_lib)
29384 This example renames the spec called @samp{lib} to @samp{old_lib} and
29385 then overrides the previous definition of @samp{lib} with a new one.
29386 The new definition adds in some extra command-line options before
29387 including the text of the old definition.
29389 @dfn{Spec strings} are a list of command-line options to be passed to their
29390 corresponding program. In addition, the spec strings can contain
29391 @samp{%}-prefixed sequences to substitute variable text or to
29392 conditionally insert text into the command line. Using these constructs
29393 it is possible to generate quite complex command lines.
29395 Here is a table of all defined @samp{%}-sequences for spec
29396 strings. Note that spaces are not generated automatically around the
29397 results of expanding these sequences. Therefore you can concatenate them
29398 together or combine them with constant text in a single argument.
29402 Substitute one @samp{%} into the program name or argument.
29405 Substitute the name of the input file being processed.
29408 Substitute the basename of the input file being processed.
29409 This is the substring up to (and not including) the last period
29410 and not including the directory.
29413 This is the same as @samp{%b}, but include the file suffix (text after
29417 Marks the argument containing or following the @samp{%d} as a
29418 temporary file name, so that that file is deleted if GCC exits
29419 successfully. Unlike @samp{%g}, this contributes no text to the
29422 @item %g@var{suffix}
29423 Substitute a file name that has suffix @var{suffix} and is chosen
29424 once per compilation, and mark the argument in the same way as
29425 @samp{%d}. To reduce exposure to denial-of-service attacks, the file
29426 name is now chosen in a way that is hard to predict even when previously
29427 chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s}
29428 might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches
29429 the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is
29430 treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g}
29431 was simply substituted with a file name chosen once per compilation,
29432 without regard to any appended suffix (which was therefore treated
29433 just like ordinary text), making such attacks more likely to succeed.
29435 @item %u@var{suffix}
29436 Like @samp{%g}, but generates a new temporary file name
29437 each time it appears instead of once per compilation.
29439 @item %U@var{suffix}
29440 Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a
29441 new one if there is no such last file name. In the absence of any
29442 @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share
29443 the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s}
29444 involves the generation of two distinct file names, one
29445 for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was
29446 simply substituted with a file name chosen for the previous @samp{%u},
29447 without regard to any appended suffix.
29449 @item %j@var{suffix}
29450 Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is
29451 writable, and if @option{-save-temps} is not used;
29452 otherwise, substitute the name
29453 of a temporary file, just like @samp{%u}. This temporary file is not
29454 meant for communication between processes, but rather as a junk
29455 disposal mechanism.
29457 @item %|@var{suffix}
29458 @itemx %m@var{suffix}
29459 Like @samp{%g}, except if @option{-pipe} is in effect. In that case
29460 @samp{%|} substitutes a single dash and @samp{%m} substitutes nothing at
29461 all. These are the two most common ways to instruct a program that it
29462 should read from standard input or write to standard output. If you
29463 need something more elaborate you can use an @samp{%@{pipe:@code{X}@}}
29464 construct: see for example @file{gcc/fortran/lang-specs.h}.
29466 @item %.@var{SUFFIX}
29467 Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args
29468 when it is subsequently output with @samp{%*}. @var{SUFFIX} is
29469 terminated by the next space or %.
29472 Marks the argument containing or following the @samp{%w} as the
29473 designated output file of this compilation. This puts the argument
29474 into the sequence of arguments that @samp{%o} substitutes.
29477 Substitutes the names of all the output files, with spaces
29478 automatically placed around them. You should write spaces
29479 around the @samp{%o} as well or the results are undefined.
29480 @samp{%o} is for use in the specs for running the linker.
29481 Input files whose names have no recognized suffix are not compiled
29482 at all, but they are included among the output files, so they are
29486 Substitutes the suffix for object files. Note that this is
29487 handled specially when it immediately follows @samp{%g, %u, or %U},
29488 because of the need for those to form complete file names. The
29489 handling is such that @samp{%O} is treated exactly as if it had already
29490 been substituted, except that @samp{%g, %u, and %U} do not currently
29491 support additional @var{suffix} characters following @samp{%O} as they do
29492 following, for example, @samp{.o}.
29495 Substitutes the standard macro predefinitions for the
29496 current target machine. Use this when running @command{cpp}.
29499 Like @samp{%p}, but puts @samp{__} before and after the name of each
29500 predefined macro, except for macros that start with @samp{__} or with
29501 @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO
29505 Substitute any of @option{-iprefix} (made from @env{GCC_EXEC_PREFIX}),
29506 @option{-isysroot} (made from @env{TARGET_SYSTEM_ROOT}),
29507 @option{-isystem} (made from @env{COMPILER_PATH} and @option{-B} options)
29508 and @option{-imultilib} as necessary.
29511 Current argument is the name of a library or startup file of some sort.
29512 Search for that file in a standard list of directories and substitute
29513 the full name found. The current working directory is included in the
29514 list of directories scanned.
29517 Current argument is the name of a linker script. Search for that file
29518 in the current list of directories to scan for libraries. If the file
29519 is located insert a @option{--script} option into the command line
29520 followed by the full path name found. If the file is not found then
29521 generate an error message. Note: the current working directory is not
29525 Print @var{str} as an error message. @var{str} is terminated by a newline.
29526 Use this when inconsistent options are detected.
29528 @item %(@var{name})
29529 Substitute the contents of spec string @var{name} at this point.
29531 @item %x@{@var{option}@}
29532 Accumulate an option for @samp{%X}.
29535 Output the accumulated linker options specified by @option{-Wl} or a @samp{%x}
29539 Output the accumulated assembler options specified by @option{-Wa}.
29542 Output the accumulated preprocessor options specified by @option{-Wp}.
29545 Process the @code{asm} spec. This is used to compute the
29546 switches to be passed to the assembler.
29549 Process the @code{asm_final} spec. This is a spec string for
29550 passing switches to an assembler post-processor, if such a program is
29554 Process the @code{link} spec. This is the spec for computing the
29555 command line passed to the linker. Typically it makes use of the
29556 @samp{%L %G %S %D and %E} sequences.
29559 Dump out a @option{-L} option for each directory that GCC believes might
29560 contain startup files. If the target supports multilibs then the
29561 current multilib directory is prepended to each of these paths.
29564 Process the @code{lib} spec. This is a spec string for deciding which
29565 libraries are included on the command line to the linker.
29568 Process the @code{libgcc} spec. This is a spec string for deciding
29569 which GCC support library is included on the command line to the linker.
29572 Process the @code{startfile} spec. This is a spec for deciding which
29573 object files are the first ones passed to the linker. Typically
29574 this might be a file named @file{crt0.o}.
29577 Process the @code{endfile} spec. This is a spec string that specifies
29578 the last object files that are passed to the linker.
29581 Process the @code{cpp} spec. This is used to construct the arguments
29582 to be passed to the C preprocessor.
29585 Process the @code{cc1} spec. This is used to construct the options to be
29586 passed to the actual C compiler (@command{cc1}).
29589 Process the @code{cc1plus} spec. This is used to construct the options to be
29590 passed to the actual C++ compiler (@command{cc1plus}).
29593 Substitute the variable part of a matched option. See below.
29594 Note that each comma in the substituted string is replaced by
29598 Remove all occurrences of @code{-S} from the command line. Note---this
29599 command is position dependent. @samp{%} commands in the spec string
29600 before this one see @code{-S}, @samp{%} commands in the spec string
29601 after this one do not.
29603 @item %:@var{function}(@var{args})
29604 Call the named function @var{function}, passing it @var{args}.
29605 @var{args} is first processed as a nested spec string, then split
29606 into an argument vector in the usual fashion. The function returns
29607 a string which is processed as if it had appeared literally as part
29608 of the current spec.
29610 The following built-in spec functions are provided:
29613 @item @code{getenv}
29614 The @code{getenv} spec function takes two arguments: an environment
29615 variable name and a string. If the environment variable is not
29616 defined, a fatal error is issued. Otherwise, the return value is the
29617 value of the environment variable concatenated with the string. For
29618 example, if @env{TOPDIR} is defined as @file{/path/to/top}, then:
29621 %:getenv(TOPDIR /include)
29624 expands to @file{/path/to/top/include}.
29626 @item @code{if-exists}
29627 The @code{if-exists} spec function takes one argument, an absolute
29628 pathname to a file. If the file exists, @code{if-exists} returns the
29629 pathname. Here is a small example of its usage:
29633 crt0%O%s %:if-exists(crti%O%s) crtbegin%O%s
29636 @item @code{if-exists-else}
29637 The @code{if-exists-else} spec function is similar to the @code{if-exists}
29638 spec function, except that it takes two arguments. The first argument is
29639 an absolute pathname to a file. If the file exists, @code{if-exists-else}
29640 returns the pathname. If it does not exist, it returns the second argument.
29641 This way, @code{if-exists-else} can be used to select one file or another,
29642 based on the existence of the first. Here is a small example of its usage:
29646 crt0%O%s %:if-exists(crti%O%s) \
29647 %:if-exists-else(crtbeginT%O%s crtbegin%O%s)
29650 @item @code{replace-outfile}
29651 The @code{replace-outfile} spec function takes two arguments. It looks for the
29652 first argument in the outfiles array and replaces it with the second argument. Here
29653 is a small example of its usage:
29656 %@{fgnu-runtime:%:replace-outfile(-lobjc -lobjc-gnu)@}
29659 @item @code{remove-outfile}
29660 The @code{remove-outfile} spec function takes one argument. It looks for the
29661 first argument in the outfiles array and removes it. Here is a small example
29665 %:remove-outfile(-lm)
29668 @item @code{pass-through-libs}
29669 The @code{pass-through-libs} spec function takes any number of arguments. It
29670 finds any @option{-l} options and any non-options ending in @file{.a} (which it
29671 assumes are the names of linker input library archive files) and returns a
29672 result containing all the found arguments each prepended by
29673 @option{-plugin-opt=-pass-through=} and joined by spaces. This list is
29674 intended to be passed to the LTO linker plugin.
29677 %:pass-through-libs(%G %L %G)
29680 @item @code{print-asm-header}
29681 The @code{print-asm-header} function takes no arguments and simply
29682 prints a banner like:
29688 Use "-Wa,OPTION" to pass "OPTION" to the assembler.
29691 It is used to separate compiler options from assembler options
29692 in the @option{--target-help} output.
29696 Substitutes the @code{-S} switch, if that switch is given to GCC@.
29697 If that switch is not specified, this substitutes nothing. Note that
29698 the leading dash is omitted when specifying this option, and it is
29699 automatically inserted if the substitution is performed. Thus the spec
29700 string @samp{%@{foo@}} matches the command-line option @option{-foo}
29701 and outputs the command-line option @option{-foo}.
29704 Like %@{@code{S}@} but mark last argument supplied within as a file to be
29705 deleted on failure.
29708 Substitutes all the switches specified to GCC whose names start
29709 with @code{-S}, but which also take an argument. This is used for
29710 switches like @option{-o}, @option{-D}, @option{-I}, etc.
29711 GCC considers @option{-o foo} as being
29712 one switch whose name starts with @samp{o}. %@{o*@} substitutes this
29713 text, including the space. Thus two arguments are generated.
29716 Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options
29717 (the order of @code{S} and @code{T} in the spec is not significant).
29718 There can be any number of ampersand-separated variables; for each the
29719 wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}.
29722 Substitutes @code{X}, if the @option{-S} switch is given to GCC@.
29725 Substitutes @code{X}, if the @option{-S} switch is @emph{not} given to GCC@.
29728 Substitutes @code{X} if one or more switches whose names start with
29729 @code{-S} are specified to GCC@. Normally @code{X} is substituted only
29730 once, no matter how many such switches appeared. However, if @code{%*}
29731 appears somewhere in @code{X}, then @code{X} is substituted once
29732 for each matching switch, with the @code{%*} replaced by the part of
29733 that switch matching the @code{*}.
29735 If @code{%*} appears as the last part of a spec sequence then a space
29736 is added after the end of the last substitution. If there is more
29737 text in the sequence, however, then a space is not generated. This
29738 allows the @code{%*} substitution to be used as part of a larger
29739 string. For example, a spec string like this:
29742 %@{mcu=*:--script=%*/memory.ld@}
29746 when matching an option like @option{-mcu=newchip} produces:
29749 --script=newchip/memory.ld
29753 Substitutes @code{X}, if processing a file with suffix @code{S}.
29756 Substitutes @code{X}, if @emph{not} processing a file with suffix @code{S}.
29759 Substitutes @code{X}, if processing a file for language @code{S}.
29762 Substitutes @code{X}, if not processing a file for language @code{S}.
29765 Substitutes @code{X} if either @code{-S} or @code{-P} is given to
29766 GCC@. This may be combined with @samp{!}, @samp{.}, @samp{,}, and
29767 @code{*} sequences as well, although they have a stronger binding than
29768 the @samp{|}. If @code{%*} appears in @code{X}, all of the
29769 alternatives must be starred, and only the first matching alternative
29772 For example, a spec string like this:
29775 %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@}
29779 outputs the following command-line options from the following input
29780 command-line options:
29785 -d fred.c -foo -baz -boggle
29786 -d jim.d -bar -baz -boggle
29789 @item %@{S:X; T:Y; :D@}
29791 If @code{S} is given to GCC, substitutes @code{X}; else if @code{T} is
29792 given to GCC, substitutes @code{Y}; else substitutes @code{D}. There can
29793 be as many clauses as you need. This may be combined with @code{.},
29794 @code{,}, @code{!}, @code{|}, and @code{*} as needed.
29799 The switch matching text @code{S} in a @samp{%@{S@}}, @samp{%@{S:X@}}
29800 or similar construct can use a backslash to ignore the special meaning
29801 of the character following it, thus allowing literal matching of a
29802 character that is otherwise specially treated. For example,
29803 @samp{%@{std=iso9899\:1999:X@}} substitutes @code{X} if the
29804 @option{-std=iso9899:1999} option is given.
29806 The conditional text @code{X} in a @samp{%@{S:X@}} or similar
29807 construct may contain other nested @samp{%} constructs or spaces, or
29808 even newlines. They are processed as usual, as described above.
29809 Trailing white space in @code{X} is ignored. White space may also
29810 appear anywhere on the left side of the colon in these constructs,
29811 except between @code{.} or @code{*} and the corresponding word.
29813 The @option{-O}, @option{-f}, @option{-m}, and @option{-W} switches are
29814 handled specifically in these constructs. If another value of
29815 @option{-O} or the negated form of a @option{-f}, @option{-m}, or
29816 @option{-W} switch is found later in the command line, the earlier
29817 switch value is ignored, except with @{@code{S}*@} where @code{S} is
29818 just one letter, which passes all matching options.
29820 The character @samp{|} at the beginning of the predicate text is used to
29821 indicate that a command should be piped to the following command, but
29822 only if @option{-pipe} is specified.
29824 It is built into GCC which switches take arguments and which do not.
29825 (You might think it would be useful to generalize this to allow each
29826 compiler's spec to say which switches take arguments. But this cannot
29827 be done in a consistent fashion. GCC cannot even decide which input
29828 files have been specified without knowing which switches take arguments,
29829 and it must know which input files to compile in order to tell which
29832 GCC also knows implicitly that arguments starting in @option{-l} are to be
29833 treated as compiler output files, and passed to the linker in their
29834 proper position among the other output files.
29836 @node Environment Variables
29837 @section Environment Variables Affecting GCC
29838 @cindex environment variables
29840 @c man begin ENVIRONMENT
29841 This section describes several environment variables that affect how GCC
29842 operates. Some of them work by specifying directories or prefixes to use
29843 when searching for various kinds of files. Some are used to specify other
29844 aspects of the compilation environment.
29846 Note that you can also specify places to search using options such as
29847 @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These
29848 take precedence over places specified using environment variables, which
29849 in turn take precedence over those specified by the configuration of GCC@.
29850 @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint,
29851 GNU Compiler Collection (GCC) Internals}.
29856 @c @itemx LC_COLLATE
29858 @c @itemx LC_MONETARY
29859 @c @itemx LC_NUMERIC
29864 @c @findex LC_COLLATE
29865 @findex LC_MESSAGES
29866 @c @findex LC_MONETARY
29867 @c @findex LC_NUMERIC
29871 These environment variables control the way that GCC uses
29872 localization information which allows GCC to work with different
29873 national conventions. GCC inspects the locale categories
29874 @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do
29875 so. These locale categories can be set to any value supported by your
29876 installation. A typical value is @samp{en_GB.UTF-8} for English in the United
29877 Kingdom encoded in UTF-8.
29879 The @env{LC_CTYPE} environment variable specifies character
29880 classification. GCC uses it to determine the character boundaries in
29881 a string; this is needed for some multibyte encodings that contain quote
29882 and escape characters that are otherwise interpreted as a string
29885 The @env{LC_MESSAGES} environment variable specifies the language to
29886 use in diagnostic messages.
29888 If the @env{LC_ALL} environment variable is set, it overrides the value
29889 of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE}
29890 and @env{LC_MESSAGES} default to the value of the @env{LANG}
29891 environment variable. If none of these variables are set, GCC
29892 defaults to traditional C English behavior.
29896 If @env{TMPDIR} is set, it specifies the directory to use for temporary
29897 files. GCC uses temporary files to hold the output of one stage of
29898 compilation which is to be used as input to the next stage: for example,
29899 the output of the preprocessor, which is the input to the compiler
29902 @item GCC_COMPARE_DEBUG
29903 @findex GCC_COMPARE_DEBUG
29904 Setting @env{GCC_COMPARE_DEBUG} is nearly equivalent to passing
29905 @option{-fcompare-debug} to the compiler driver. See the documentation
29906 of this option for more details.
29908 @item GCC_EXEC_PREFIX
29909 @findex GCC_EXEC_PREFIX
29910 If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the
29911 names of the subprograms executed by the compiler. No slash is added
29912 when this prefix is combined with the name of a subprogram, but you can
29913 specify a prefix that ends with a slash if you wish.
29915 If @env{GCC_EXEC_PREFIX} is not set, GCC attempts to figure out
29916 an appropriate prefix to use based on the pathname it is invoked with.
29918 If GCC cannot find the subprogram using the specified prefix, it
29919 tries looking in the usual places for the subprogram.
29921 The default value of @env{GCC_EXEC_PREFIX} is
29922 @file{@var{prefix}/lib/gcc/} where @var{prefix} is the prefix to
29923 the installed compiler. In many cases @var{prefix} is the value
29924 of @code{prefix} when you ran the @file{configure} script.
29926 Other prefixes specified with @option{-B} take precedence over this prefix.
29928 This prefix is also used for finding files such as @file{crt0.o} that are
29931 In addition, the prefix is used in an unusual way in finding the
29932 directories to search for header files. For each of the standard
29933 directories whose name normally begins with @samp{/usr/local/lib/gcc}
29934 (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries
29935 replacing that beginning with the specified prefix to produce an
29936 alternate directory name. Thus, with @option{-Bfoo/}, GCC searches
29937 @file{foo/bar} just before it searches the standard directory
29938 @file{/usr/local/lib/bar}.
29939 If a standard directory begins with the configured
29940 @var{prefix} then the value of @var{prefix} is replaced by
29941 @env{GCC_EXEC_PREFIX} when looking for header files.
29943 @item COMPILER_PATH
29944 @findex COMPILER_PATH
29945 The value of @env{COMPILER_PATH} is a colon-separated list of
29946 directories, much like @env{PATH}. GCC tries the directories thus
29947 specified when searching for subprograms, if it cannot find the
29948 subprograms using @env{GCC_EXEC_PREFIX}.
29951 @findex LIBRARY_PATH
29952 The value of @env{LIBRARY_PATH} is a colon-separated list of
29953 directories, much like @env{PATH}. When configured as a native compiler,
29954 GCC tries the directories thus specified when searching for special
29955 linker files, if it cannot find them using @env{GCC_EXEC_PREFIX}. Linking
29956 using GCC also uses these directories when searching for ordinary
29957 libraries for the @option{-l} option (but directories specified with
29958 @option{-L} come first).
29962 @cindex locale definition
29963 This variable is used to pass locale information to the compiler. One way in
29964 which this information is used is to determine the character set to be used
29965 when character literals, string literals and comments are parsed in C and C++.
29966 When the compiler is configured to allow multibyte characters,
29967 the following values for @env{LANG} are recognized:
29971 Recognize JIS characters.
29973 Recognize SJIS characters.
29975 Recognize EUCJP characters.
29978 If @env{LANG} is not defined, or if it has some other value, then the
29979 compiler uses @code{mblen} and @code{mbtowc} as defined by the default locale to
29980 recognize and translate multibyte characters.
29984 Some additional environment variables affect the behavior of the
29987 @include cppenv.texi
29991 @node Precompiled Headers
29992 @section Using Precompiled Headers
29993 @cindex precompiled headers
29994 @cindex speed of compilation
29996 Often large projects have many header files that are included in every
29997 source file. The time the compiler takes to process these header files
29998 over and over again can account for nearly all of the time required to
29999 build the project. To make builds faster, GCC allows you to
30000 @dfn{precompile} a header file.
30002 To create a precompiled header file, simply compile it as you would any
30003 other file, if necessary using the @option{-x} option to make the driver
30004 treat it as a C or C++ header file. You may want to use a
30005 tool like @command{make} to keep the precompiled header up-to-date when
30006 the headers it contains change.
30008 A precompiled header file is searched for when @code{#include} is
30009 seen in the compilation. As it searches for the included file
30010 (@pxref{Search Path,,Search Path,cpp,The C Preprocessor}) the
30011 compiler looks for a precompiled header in each directory just before it
30012 looks for the include file in that directory. The name searched for is
30013 the name specified in the @code{#include} with @samp{.gch} appended. If
30014 the precompiled header file cannot be used, it is ignored.
30016 For instance, if you have @code{#include "all.h"}, and you have
30017 @file{all.h.gch} in the same directory as @file{all.h}, then the
30018 precompiled header file is used if possible, and the original
30019 header is used otherwise.
30021 Alternatively, you might decide to put the precompiled header file in a
30022 directory and use @option{-I} to ensure that directory is searched
30023 before (or instead of) the directory containing the original header.
30024 Then, if you want to check that the precompiled header file is always
30025 used, you can put a file of the same name as the original header in this
30026 directory containing an @code{#error} command.
30028 This also works with @option{-include}. So yet another way to use
30029 precompiled headers, good for projects not designed with precompiled
30030 header files in mind, is to simply take most of the header files used by
30031 a project, include them from another header file, precompile that header
30032 file, and @option{-include} the precompiled header. If the header files
30033 have guards against multiple inclusion, they are skipped because
30034 they've already been included (in the precompiled header).
30036 If you need to precompile the same header file for different
30037 languages, targets, or compiler options, you can instead make a
30038 @emph{directory} named like @file{all.h.gch}, and put each precompiled
30039 header in the directory, perhaps using @option{-o}. It doesn't matter
30040 what you call the files in the directory; every precompiled header in
30041 the directory is considered. The first precompiled header
30042 encountered in the directory that is valid for this compilation is
30043 used; they're searched in no particular order.
30045 There are many other possibilities, limited only by your imagination,
30046 good sense, and the constraints of your build system.
30048 A precompiled header file can be used only when these conditions apply:
30052 Only one precompiled header can be used in a particular compilation.
30055 A precompiled header cannot be used once the first C token is seen. You
30056 can have preprocessor directives before a precompiled header; you cannot
30057 include a precompiled header from inside another header.
30060 The precompiled header file must be produced for the same language as
30061 the current compilation. You cannot use a C precompiled header for a C++
30065 The precompiled header file must have been produced by the same compiler
30066 binary as the current compilation is using.
30069 Any macros defined before the precompiled header is included must
30070 either be defined in the same way as when the precompiled header was
30071 generated, or must not affect the precompiled header, which usually
30072 means that they don't appear in the precompiled header at all.
30074 The @option{-D} option is one way to define a macro before a
30075 precompiled header is included; using a @code{#define} can also do it.
30076 There are also some options that define macros implicitly, like
30077 @option{-O} and @option{-Wdeprecated}; the same rule applies to macros
30080 @item If debugging information is output when using the precompiled
30081 header, using @option{-g} or similar, the same kind of debugging information
30082 must have been output when building the precompiled header. However,
30083 a precompiled header built using @option{-g} can be used in a compilation
30084 when no debugging information is being output.
30086 @item The same @option{-m} options must generally be used when building
30087 and using the precompiled header. @xref{Submodel Options},
30088 for any cases where this rule is relaxed.
30090 @item Each of the following options must be the same when building and using
30091 the precompiled header:
30093 @gccoptlist{-fexceptions}
30096 Some other command-line options starting with @option{-f},
30097 @option{-p}, or @option{-O} must be defined in the same way as when
30098 the precompiled header was generated. At present, it's not clear
30099 which options are safe to change and which are not; the safest choice
30100 is to use exactly the same options when generating and using the
30101 precompiled header. The following are known to be safe:
30103 @gccoptlist{-fmessage-length= -fpreprocessed -fsched-interblock @gol
30104 -fsched-spec -fsched-spec-load -fsched-spec-load-dangerous @gol
30105 -fsched-verbose=@var{number} -fschedule-insns -fvisibility= @gol
30110 For all of these except the last, the compiler automatically
30111 ignores the precompiled header if the conditions aren't met. If you
30112 find an option combination that doesn't work and doesn't cause the
30113 precompiled header to be ignored, please consider filing a bug report,
30116 If you do use differing options when generating and using the
30117 precompiled header, the actual behavior is a mixture of the
30118 behavior for the options. For instance, if you use @option{-g} to
30119 generate the precompiled header but not when using it, you may or may
30120 not get debugging information for routines in the precompiled header.