]>
Commit | Line | Data |
---|---|---|
d0a5eb32 | 1 | @c Copyright (C) 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, |
79cf5994 | 2 | @c 2000, 2001, 2002, 2003, 2004, 2005, 2006 Free Software Foundation, Inc. |
74291a4b MM |
3 | @c This is part of the GCC manual. |
4 | @c For copying conditions, see the file gcc.texi. | |
5 | ||
9d86bffc | 6 | @ignore |
9d530538 MM |
7 | @c man begin INCLUDE |
8 | @include gcc-vers.texi | |
9 | @c man end | |
10 | ||
9d86bffc | 11 | @c man begin COPYRIGHT |
2b6dd222 JM |
12 | Copyright @copyright{} 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998, |
13 | 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006 Free Software Foundation, Inc. | |
9d86bffc | 14 | |
77bd67cb | 15 | Permission is granted to copy, distribute and/or modify this document |
b3a8389d | 16 | under the terms of the GNU Free Documentation License, Version 1.2 or |
77bd67cb JM |
17 | any later version published by the Free Software Foundation; with the |
18 | Invariant Sections being ``GNU General Public License'' and ``Funding | |
19 | Free Software'', the Front-Cover texts being (a) (see below), and with | |
20 | the Back-Cover Texts being (b) (see below). A copy of the license is | |
21 | included in the gfdl(7) man page. | |
9d86bffc | 22 | |
77bd67cb | 23 | (a) The FSF's Front-Cover Text is: |
9d86bffc | 24 | |
77bd67cb JM |
25 | A GNU Manual |
26 | ||
27 | (b) The FSF's Back-Cover Text is: | |
28 | ||
29 | You have freedom to copy and modify this GNU Manual, like GNU | |
30 | software. Copies published by the Free Software Foundation raise | |
31 | funds for GNU development. | |
9d86bffc JM |
32 | @c man end |
33 | @c Set file name and title for the man page. | |
34 | @setfilename gcc | |
35 | @settitle GNU project C and C++ compiler | |
36 | @c man begin SYNOPSIS | |
630d3d5a JM |
37 | gcc [@option{-c}|@option{-S}|@option{-E}] [@option{-std=}@var{standard}] |
38 | [@option{-g}] [@option{-pg}] [@option{-O}@var{level}] | |
39 | [@option{-W}@var{warn}@dots{}] [@option{-pedantic}] | |
40 | [@option{-I}@var{dir}@dots{}] [@option{-L}@var{dir}@dots{}] | |
41 | [@option{-D}@var{macro}[=@var{defn}]@dots{}] [@option{-U}@var{macro}] | |
42 | [@option{-f}@var{option}@dots{}] [@option{-m}@var{machine-option}@dots{}] | |
9d530538 | 43 | [@option{-o} @var{outfile}] [@@@var{file}] @var{infile}@dots{} |
9d86bffc JM |
44 | |
45 | Only the most useful options are listed here; see below for the | |
46 | remainder. @samp{g++} accepts mostly the same options as @samp{gcc}. | |
47 | @c man end | |
48 | @c man begin SEEALSO | |
77bd67cb | 49 | gpl(7), gfdl(7), fsf-funding(7), |
b4117c30 ZW |
50 | cpp(1), gcov(1), as(1), ld(1), gdb(1), adb(1), dbx(1), sdb(1) |
51 | and the Info entries for @file{gcc}, @file{cpp}, @file{as}, | |
9d86bffc JM |
52 | @file{ld}, @file{binutils} and @file{gdb}. |
53 | @c man end | |
2642624b JM |
54 | @c man begin BUGS |
55 | For instructions on reporting bugs, see | |
ee86feaf | 56 | @w{@uref{http://gcc.gnu.org/bugs.html}}. |
2642624b JM |
57 | @c man end |
58 | @c man begin AUTHOR | |
24dbb440 GP |
59 | See the Info entry for @command{gcc}, or |
60 | @w{@uref{http://gcc.gnu.org/onlinedocs/gcc/Contributors.html}}, | |
61 | for contributors to GCC@. | |
2642624b | 62 | @c man end |
9d86bffc JM |
63 | @end ignore |
64 | ||
74291a4b | 65 | @node Invoking GCC |
0c2d1a2a JB |
66 | @chapter GCC Command Options |
67 | @cindex GCC command options | |
74291a4b | 68 | @cindex command options |
0c2d1a2a | 69 | @cindex options, GCC command |
74291a4b | 70 | |
9d86bffc | 71 | @c man begin DESCRIPTION |
0c2d1a2a | 72 | When you invoke GCC, it normally does preprocessing, compilation, |
74291a4b | 73 | assembly and linking. The ``overall options'' allow you to stop this |
630d3d5a | 74 | process at an intermediate stage. For example, the @option{-c} option |
74291a4b MM |
75 | says not to run the linker. Then the output consists of object files |
76 | output by the assembler. | |
77 | ||
78 | Other options are passed on to one stage of processing. Some options | |
79 | control the preprocessor and others the compiler itself. Yet other | |
80 | options control the assembler and linker; most of these are not | |
81 | documented here, since you rarely need to use any of them. | |
82 | ||
83 | @cindex C compilation options | |
0c2d1a2a | 84 | Most of the command line options that you can use with GCC are useful |
74291a4b MM |
85 | for C programs; when an option is only useful with another language |
86 | (usually C++), the explanation says so explicitly. If the description | |
87 | for a particular option does not mention a source language, you can use | |
88 | that option with all supported languages. | |
89 | ||
90 | @cindex C++ compilation options | |
91 | @xref{Invoking G++,,Compiling C++ Programs}, for a summary of special | |
92 | options for compiling C++ programs. | |
93 | ||
94 | @cindex grouping options | |
95 | @cindex options, grouping | |
bedc7537 | 96 | The @command{gcc} program accepts options and file names as operands. Many |
b192711e | 97 | options have multi-letter names; therefore multiple single-letter options |
630d3d5a | 98 | may @emph{not} be grouped: @option{-dr} is very different from @w{@samp{-d |
74291a4b MM |
99 | -r}}. |
100 | ||
101 | @cindex order of options | |
102 | @cindex options, order | |
103 | You can mix options and other arguments. For the most part, the order | |
104 | you use doesn't matter. Order does matter when you use several options | |
630d3d5a | 105 | of the same kind; for example, if you specify @option{-L} more than once, |
74291a4b MM |
106 | the directories are searched in the order specified. |
107 | ||
108 | Many options have long names starting with @samp{-f} or with | |
1d8eeb63 | 109 | @samp{-W}---for example, |
630d3d5a | 110 | @option{-fstrength-reduce}, @option{-Wformat} and so on. Most of |
74291a4b | 111 | these have both positive and negative forms; the negative form of |
630d3d5a | 112 | @option{-ffoo} would be @option{-fno-foo}. This manual documents |
74291a4b MM |
113 | only one of these two forms, whichever one is not the default. |
114 | ||
9d86bffc JM |
115 | @c man end |
116 | ||
cd3bb277 JM |
117 | @xref{Option Index}, for an index to GCC's options. |
118 | ||
74291a4b MM |
119 | @menu |
120 | * Option Summary:: Brief list of all options, without explanations. | |
121 | * Overall Options:: Controlling the kind of output: | |
122 | an executable, object files, assembler files, | |
123 | or preprocessed source. | |
124 | * Invoking G++:: Compiling C++ programs. | |
125 | * C Dialect Options:: Controlling the variant of C language compiled. | |
126 | * C++ Dialect Options:: Variations on C++. | |
46e34f96 ZL |
127 | * Objective-C and Objective-C++ Dialect Options:: Variations on Objective-C |
128 | and Objective-C++. | |
764dbbf2 | 129 | * Language Independent Options:: Controlling how diagnostics should be |
02f52e19 | 130 | formatted. |
74291a4b MM |
131 | * Warning Options:: How picky should the compiler be? |
132 | * Debugging Options:: Symbol tables, measurements, and debugging dumps. | |
133 | * Optimize Options:: How much optimization? | |
134 | * Preprocessor Options:: Controlling header files and macro definitions. | |
135 | Also, getting dependency information for Make. | |
136 | * Assembler Options:: Passing options to the assembler. | |
137 | * Link Options:: Specifying libraries and so on. | |
138 | * Directory Options:: Where to find header files and libraries. | |
139 | Where to find the compiler executable files. | |
a743d340 | 140 | * Spec Files:: How to pass switches to sub-processes. |
0c2d1a2a | 141 | * Target Options:: Running a cross-compiler, or an old version of GCC. |
74291a4b MM |
142 | * Submodel Options:: Specifying minor hardware or convention variations, |
143 | such as 68010 vs 68020. | |
144 | * Code Gen Options:: Specifying conventions for function calls, data layout | |
145 | and register usage. | |
0c2d1a2a | 146 | * Environment Variables:: Env vars that affect GCC. |
17211ab5 | 147 | * Precompiled Headers:: Compiling a header once, and using it many times. |
74291a4b MM |
148 | * Running Protoize:: Automatically adding or removing function prototypes. |
149 | @end menu | |
150 | ||
4bc1997b JM |
151 | @c man begin OPTIONS |
152 | ||
74291a4b MM |
153 | @node Option Summary |
154 | @section Option Summary | |
155 | ||
156 | Here is a summary of all the options, grouped by type. Explanations are | |
157 | in the following sections. | |
158 | ||
159 | @table @emph | |
160 | @item Overall Options | |
161 | @xref{Overall Options,,Options Controlling the Kind of Output}. | |
0855eab7 | 162 | @gccoptlist{-c -S -E -o @var{file} -combine -pipe -pass-exit-codes @gol |
9d530538 | 163 | -x @var{language} -v -### --help --target-help --version @@@var{file}} |
74291a4b MM |
164 | |
165 | @item C Language Options | |
166 | @xref{C Dialect Options,,Options Controlling C Dialect}. | |
9a94f7f3 JM |
167 | @gccoptlist{-ansi -std=@var{standard} -aux-info @var{filename} @gol |
168 | -fno-asm -fno-builtin -fno-builtin-@var{function} @gol | |
750491fc | 169 | -fhosted -ffreestanding -fms-extensions @gol |
8a035a6b | 170 | -trigraphs -no-integrated-cpp -traditional -traditional-cpp @gol |
4bc1997b JM |
171 | -fallow-single-precision -fcond-mismatch @gol |
172 | -fsigned-bitfields -fsigned-char @gol | |
3521b33c | 173 | -funsigned-bitfields -funsigned-char} |
74291a4b MM |
174 | |
175 | @item C++ Language Options | |
176 | @xref{C++ Dialect Options,,Options Controlling C++ Dialect}. | |
9a94f7f3 | 177 | @gccoptlist{-fabi-version=@var{n} -fno-access-control -fcheck-new @gol |
d63d5d0c | 178 | -fconserve-space -ffriend-injection -fno-const-strings @gol |
aa0cc562 | 179 | -fno-elide-constructors @gol |
7813d14c | 180 | -fno-enforce-eh-specs @gol |
1dbb6023 | 181 | -ffor-scope -fno-for-scope -fno-gnu-keywords @gol |
90ecce3e | 182 | -fno-implicit-templates @gol |
4bc1997b JM |
183 | -fno-implicit-inline-templates @gol |
184 | -fno-implement-inlines -fms-extensions @gol | |
aa0cc562 | 185 | -fno-nonansi-builtins -fno-operator-names @gol |
4bc1997b | 186 | -fno-optional-diags -fpermissive @gol |
aa0cc562 | 187 | -frepo -fno-rtti -fstats -ftemplate-depth-@var{n} @gol |
40aac948 | 188 | -fno-threadsafe-statics -fuse-cxa-atexit -fno-weak -nostdinc++ @gol |
d7afec4b ND |
189 | -fno-default-inline -fvisibility-inlines-hidden @gol |
190 | -Wabi -Wctor-dtor-privacy @gol | |
4bc1997b | 191 | -Wnon-virtual-dtor -Wreorder @gol |
b2f97e4a | 192 | -Weffc++ -Wno-deprecated -Wstrict-null-sentinel @gol |
4bc1997b JM |
193 | -Wno-non-template-friend -Wold-style-cast @gol |
194 | -Woverloaded-virtual -Wno-pmf-conversions @gol | |
4e2db7a8 | 195 | -Wsign-promo} |
74291a4b | 196 | |
46e34f96 ZL |
197 | @item Objective-C and Objective-C++ Language Options |
198 | @xref{Objective-C and Objective-C++ Dialect Options,,Options Controlling | |
199 | Objective-C and Objective-C++ Dialects}. | |
264fa2db ZL |
200 | @gccoptlist{ |
201 | -fconstant-string-class=@var{class-name} @gol | |
202 | -fgnu-runtime -fnext-runtime @gol | |
203 | -fno-nil-receivers @gol | |
6e955430 ZL |
204 | -fobjc-call-cxx-cdtors @gol |
205 | -fobjc-direct-dispatch @gol | |
264fa2db | 206 | -fobjc-exceptions @gol |
6e955430 | 207 | -fobjc-gc @gol |
264fa2db ZL |
208 | -freplace-objc-classes @gol |
209 | -fzero-link @gol | |
210 | -gen-decls @gol | |
6e955430 ZL |
211 | -Wassign-intercept @gol |
212 | -Wno-protocol -Wselector @gol | |
213 | -Wstrict-selector-match @gol | |
214 | -Wundeclared-selector} | |
60de6385 | 215 | |
764dbbf2 GDR |
216 | @item Language Independent Options |
217 | @xref{Language Independent Options,,Options to Control Diagnostic Messages Formatting}. | |
9a94f7f3 | 218 | @gccoptlist{-fmessage-length=@var{n} @gol |
ccf08a6e DD |
219 | -fdiagnostics-show-location=@r{[}once@r{|}every-line@r{]}} @gol |
220 | -fdiagnostics-show-options | |
764dbbf2 | 221 | |
74291a4b MM |
222 | @item Warning Options |
223 | @xref{Warning Options,,Options to Request or Suppress Warnings}. | |
9a94f7f3 | 224 | @gccoptlist{-fsyntax-only -pedantic -pedantic-errors @gol |
690a704a | 225 | -w -Wextra -Wall -Waggregate-return -Walways-true -Wno-attributes @gol |
b7e20b53 | 226 | -Wc++-compat -Wcast-align -Wcast-qual -Wchar-subscripts -Wcomment @gol |
e23bd218 | 227 | -Wconversion -Wno-deprecated-declarations @gol |
90689ae1 | 228 | -Wdisabled-optimization -Wno-div-by-zero -Wno-endif-labels @gol |
79cf5994 | 229 | -Werror -Werror-* -Werror-implicit-function-declaration @gol |
c65a01af | 230 | -Wfatal-errors -Wfloat-equal -Wformat -Wformat=2 @gol |
fb0317c6 | 231 | -Wno-format-extra-args -Wformat-nonliteral @gol |
c76f4e8e | 232 | -Wformat-security -Wformat-y2k @gol |
fb0317c6 VR |
233 | -Wimplicit -Wimplicit-function-declaration -Wimplicit-int @gol |
234 | -Wimport -Wno-import -Winit-self -Winline @gol | |
53a2494e | 235 | -Wno-int-to-pointer-cast @gol |
fb0317c6 | 236 | -Wno-invalid-offsetof -Winvalid-pch @gol |
f9cc1a70 | 237 | -Wlarger-than-@var{len} -Wunsafe-loop-optimizations -Wlong-long @gol |
eaac4679 | 238 | -Wmain -Wmissing-braces -Wmissing-field-initializers @gol |
b02398bd BE |
239 | -Wmissing-format-attribute -Wmissing-include-dirs @gol |
240 | -Wmissing-noreturn @gol | |
fb0317c6 | 241 | -Wno-multichar -Wnonnull -Wpacked -Wpadded @gol |
53a2494e JM |
242 | -Wparentheses -Wpointer-arith -Wno-pointer-to-int-cast @gol |
243 | -Wredundant-decls @gol | |
4bc1997b | 244 | -Wreturn-type -Wsequence-point -Wshadow @gol |
0aca9021 JW |
245 | -Wsign-compare -Wstack-protector @gol |
246 | -Wstrict-aliasing -Wstrict-aliasing=2 @gol | |
d77314ec | 247 | -Wstring-literal-comparison @gol |
9a94f7f3 JM |
248 | -Wswitch -Wswitch-default -Wswitch-enum @gol |
249 | -Wsystem-headers -Wtrigraphs -Wundef -Wuninitialized @gol | |
b9b8dde3 | 250 | -Wunknown-pragmas -Wno-pragmas -Wunreachable-code @gol |
4bc1997b | 251 | -Wunused -Wunused-function -Wunused-label -Wunused-parameter @gol |
0aca9021 | 252 | -Wunused-value -Wunused-variable -Wvariadic-macros @gol |
d35a40fc | 253 | -Wvolatile-register-var -Wwrite-strings} |
74291a4b | 254 | |
fe50c0eb | 255 | @item C-only Warning Options |
9a94f7f3 | 256 | @gccoptlist{-Wbad-function-cast -Wmissing-declarations @gol |
fb0317c6 | 257 | -Wmissing-prototypes -Wnested-externs -Wold-style-definition @gol |
85617eba | 258 | -Wstrict-prototypes -Wtraditional @gol |
f4e9414e | 259 | -Wdeclaration-after-statement -Wpointer-sign} |
fe50c0eb | 260 | |
74291a4b MM |
261 | @item Debugging Options |
262 | @xref{Debugging Options,,Options for Debugging Your Program or GCC}. | |
9a94f7f3 JM |
263 | @gccoptlist{-d@var{letters} -dumpspecs -dumpmachine -dumpversion @gol |
264 | -fdump-unnumbered -fdump-translation-unit@r{[}-@var{n}@r{]} @gol | |
aee96fe9 | 265 | -fdump-class-hierarchy@r{[}-@var{n}@r{]} @gol |
9b3e897d | 266 | -fdump-ipa-all -fdump-ipa-cgraph @gol |
6de9cd9a | 267 | -fdump-tree-all @gol |
9a94f7f3 JM |
268 | -fdump-tree-original@r{[}-@var{n}@r{]} @gol |
269 | -fdump-tree-optimized@r{[}-@var{n}@r{]} @gol | |
22367161 | 270 | -fdump-tree-inlined@r{[}-@var{n}@r{]} @gol |
6de9cd9a DN |
271 | -fdump-tree-cfg -fdump-tree-vcg -fdump-tree-alias @gol |
272 | -fdump-tree-ch @gol | |
273 | -fdump-tree-ssa@r{[}-@var{n}@r{]} -fdump-tree-pre@r{[}-@var{n}@r{]} @gol | |
274 | -fdump-tree-ccp@r{[}-@var{n}@r{]} -fdump-tree-dce@r{[}-@var{n}@r{]} @gol | |
275 | -fdump-tree-gimple@r{[}-raw@r{]} -fdump-tree-mudflap@r{[}-@var{n}@r{]} @gol | |
276 | -fdump-tree-dom@r{[}-@var{n}@r{]} @gol | |
277 | -fdump-tree-dse@r{[}-@var{n}@r{]} @gol | |
278 | -fdump-tree-phiopt@r{[}-@var{n}@r{]} @gol | |
279 | -fdump-tree-forwprop@r{[}-@var{n}@r{]} @gol | |
280 | -fdump-tree-copyrename@r{[}-@var{n}@r{]} @gol | |
79fe1b3b | 281 | -fdump-tree-nrv -fdump-tree-vect @gol |
fa555252 | 282 | -fdump-tree-sink @gol |
6de9cd9a | 283 | -fdump-tree-sra@r{[}-@var{n}@r{]} @gol |
c75ab022 | 284 | -fdump-tree-salias @gol |
ff2ad0f7 | 285 | -fdump-tree-fre@r{[}-@var{n}@r{]} @gol |
08873e96 | 286 | -fdump-tree-vrp@r{[}-@var{n}@r{]} @gol |
c866976a | 287 | -ftree-vectorizer-verbose=@var{n} @gol |
0bca51f0 | 288 | -fdump-tree-storeccp@r{[}-@var{n}@r{]} @gol |
a37db56b | 289 | -feliminate-dwarf2-dups -feliminate-unused-debug-types @gol |
8a76829c | 290 | -feliminate-unused-debug-symbols -fmem-report -fprofile-arcs @gol |
a37db56b | 291 | -frandom-seed=@var{string} -fsched-verbose=@var{n} @gol |
014a1138 | 292 | -ftest-coverage -ftime-report -fvar-tracking @gol |
def66b10 | 293 | -g -g@var{level} -gcoff -gdwarf-2 @gol |
5f98259a | 294 | -ggdb -gstabs -gstabs+ -gvms -gxcoff -gxcoff+ @gol |
4bc1997b | 295 | -p -pg -print-file-name=@var{library} -print-libgcc-file-name @gol |
b1018de6 | 296 | -print-multi-directory -print-multi-lib @gol |
4bc1997b JM |
297 | -print-prog-name=@var{program} -print-search-dirs -Q @gol |
298 | -save-temps -time} | |
74291a4b MM |
299 | |
300 | @item Optimization Options | |
301 | @xref{Optimize Options,,Options that Control Optimization}. | |
9a94f7f3 | 302 | @gccoptlist{-falign-functions=@var{n} -falign-jumps=@var{n} @gol |
4bc1997b | 303 | -falign-labels=@var{n} -falign-loops=@var{n} @gol |
6de9cd9a | 304 | -fbounds-check -fmudflap -fmudflapth -fmudflapir @gol |
fca9dc00 | 305 | -fbranch-probabilities -fprofile-values -fvpt -fbranch-target-load-optimize @gol |
1194fc79 | 306 | -fbranch-target-load-optimize2 -fbtr-bb-exclusive @gol |
c7463669 | 307 | -fcaller-saves -fcprop-registers -fcse-follow-jumps @gol |
3450cbc4 | 308 | -fcse-skip-blocks -fcx-limited-range -fdata-sections @gol |
d63db217 | 309 | -fdelayed-branch -fdelete-null-pointer-checks -fearly-inlining @gol |
4bc1997b | 310 | -fexpensive-optimizations -ffast-math -ffloat-store @gol |
1d8eeb63 | 311 | -fforce-addr -ffunction-sections @gol |
db643b91 SH |
312 | -fgcse -fgcse-lm -fgcse-sm -fgcse-las -fgcse-after-reload @gol |
313 | -floop-optimize -fcrossjumping -fif-conversion -fif-conversion2 @gol | |
355866de RG |
314 | -finline-functions -finline-functions-called-once @gol |
315 | -finline-limit=@var{n} -fkeep-inline-functions @gol | |
201556f0 | 316 | -fkeep-static-consts -fmerge-constants -fmerge-all-constants @gol |
cd280abb | 317 | -fmodulo-sched -fno-branch-count-reg @gol |
5e962776 | 318 | -fno-default-inline -fno-defer-pop -floop-optimize2 -fmove-loop-invariants @gol |
feb48bde | 319 | -fno-function-cse -fno-guess-branch-probability @gol |
6cfc0341 | 320 | -fno-inline -fno-math-errno -fno-peephole -fno-peephole2 @gol |
f9cc1a70 | 321 | -funsafe-math-optimizations -funsafe-loop-optimizations -ffinite-math-only @gol |
474eccc6 | 322 | -fno-toplevel-reorder -fno-trapping-math -fno-zero-initialized-in-bss @gol |
4bc1997b | 323 | -fomit-frame-pointer -foptimize-register-move @gol |
5d22c1a5 | 324 | -foptimize-sibling-calls -fprefetch-loop-arrays @gol |
a8a5f53a | 325 | -fprofile-generate -fprofile-use @gol |
a924fe34 | 326 | -fregmove -frename-registers @gol |
750054a2 | 327 | -freorder-blocks -freorder-blocks-and-partition -freorder-functions @gol |
4bc1997b | 328 | -frerun-cse-after-loop -frerun-loop-opt @gol |
38109dab GL |
329 | -frounding-math -frtl-abstract-sequences @gol |
330 | -fschedule-insns -fschedule-insns2 @gol | |
e03b7153 | 331 | -fno-sched-interblock -fno-sched-spec -fsched-spec-load @gol |
569fa502 | 332 | -fsched-spec-load-dangerous @gol |
0aca9021 | 333 | -fsched-stalled-insns=@var{n} -fsched-stalled-insns-dep=@var{n} @gol |
569fa502 | 334 | -fsched2-use-superblocks @gol |
d72372e4 | 335 | -fsched2-use-traces -freschedule-modulo-scheduled-loops @gol |
8a76829c | 336 | -fsignaling-nans -fsingle-precision-constant @gol |
0aca9021 | 337 | -fstack-protector -fstack-protector-all @gol |
9a94f7f3 JM |
338 | -fstrength-reduce -fstrict-aliasing -ftracer -fthread-jumps @gol |
339 | -funroll-all-loops -funroll-loops -fpeel-loops @gol | |
113d659a | 340 | -fsplit-ivs-in-unroller -funswitch-loops @gol |
f37a4f14 | 341 | -fvariable-expansion-in-unroller @gol |
c66b6c66 | 342 | -ftree-pre -ftree-ccp -ftree-dce -ftree-loop-optimize @gol |
599eabdb | 343 | -ftree-loop-linear -ftree-loop-im -ftree-loop-ivcanon -fivopts @gol |
fa555252 | 344 | -ftree-dominator-opts -ftree-dse -ftree-copyrename -ftree-sink @gol |
79fe1b3b | 345 | -ftree-ch -ftree-sra -ftree-ter -ftree-lrs -ftree-fre -ftree-vectorize @gol |
c12cc930 | 346 | -ftree-vect-loop-version -ftree-salias -fweb @gol |
ce91e74c | 347 | -ftree-copy-prop -ftree-store-ccp -ftree-store-copy-prop -fwhole-program @gol |
3af64fd6 | 348 | --param @var{name}=@var{value} |
4bc1997b | 349 | -O -O0 -O1 -O2 -O3 -Os} |
74291a4b MM |
350 | |
351 | @item Preprocessor Options | |
352 | @xref{Preprocessor Options,,Options Controlling the Preprocessor}. | |
c2d635bc | 353 | @gccoptlist{-A@var{question}=@var{answer} @gol |
9a94f7f3 | 354 | -A-@var{question}@r{[}=@var{answer}@r{]} @gol |
4bc1997b JM |
355 | -C -dD -dI -dM -dN @gol |
356 | -D@var{macro}@r{[}=@var{defn}@r{]} -E -H @gol | |
357 | -idirafter @var{dir} @gol | |
358 | -include @var{file} -imacros @var{file} @gol | |
359 | -iprefix @var{file} -iwithprefix @var{dir} @gol | |
bdd42dd9 | 360 | -iwithprefixbefore @var{dir} -isystem @var{dir} @gol |
2b6dd222 | 361 | -imultilib @var{dir} -isysroot @var{dir} @gol |
b20d9f0c AO |
362 | -M -MM -MF -MG -MP -MQ -MT -nostdinc @gol |
363 | -P -fworking-directory -remap @gol | |
9a94f7f3 JM |
364 | -trigraphs -undef -U@var{macro} -Wp,@var{option} @gol |
365 | -Xpreprocessor @var{option}} | |
74291a4b MM |
366 | |
367 | @item Assembler Option | |
368 | @xref{Assembler Options,,Passing Options to the Assembler}. | |
9a94f7f3 | 369 | @gccoptlist{-Wa,@var{option} -Xassembler @var{option}} |
74291a4b MM |
370 | |
371 | @item Linker Options | |
372 | @xref{Link Options,,Options for Linking}. | |
9a94f7f3 | 373 | @gccoptlist{@var{object-file-name} -l@var{library} @gol |
0cbc4d77 | 374 | -nostartfiles -nodefaultlibs -nostdlib -pie -rdynamic @gol |
4bc1997b | 375 | -s -static -static-libgcc -shared -shared-libgcc -symbolic @gol |
aee96fe9 | 376 | -Wl,@var{option} -Xlinker @var{option} @gol |
4bc1997b | 377 | -u @var{symbol}} |
74291a4b MM |
378 | |
379 | @item Directory Options | |
380 | @xref{Directory Options,,Options for Directory Search}. | |
160633c6 MM |
381 | @gccoptlist{-B@var{prefix} -I@var{dir} -iquote@var{dir} -L@var{dir} |
382 | -specs=@var{file} -I- --sysroot=@var{dir}} | |
74291a4b MM |
383 | |
384 | @item Target Options | |
385 | @c I wrote this xref this way to avoid overfull hbox. -- rms | |
386 | @xref{Target Options}. | |
9a94f7f3 | 387 | @gccoptlist{-V @var{version} -b @var{machine}} |
74291a4b MM |
388 | |
389 | @item Machine Dependent Options | |
390 | @xref{Submodel Options,,Hardware Models and Configurations}. | |
39bc1876 NS |
391 | @c This list is ordered alphanumerically by subsection name. |
392 | @c Try and put the significant identifier (CPU or system) first, | |
393 | @c so users have a clue at guessing where the ones they want will be. | |
5d22c1a5 | 394 | |
39bc1876 NS |
395 | @emph{ARC Options} |
396 | @gccoptlist{-EB -EL @gol | |
397 | -mmangle-cpu -mcpu=@var{cpu} -mtext=@var{text-section} @gol | |
398 | -mdata=@var{data-section} -mrodata=@var{readonly-data-section}} | |
74291a4b | 399 | |
74291a4b | 400 | @emph{ARM Options} |
9a94f7f3 | 401 | @gccoptlist{-mapcs-frame -mno-apcs-frame @gol |
5848830f | 402 | -mabi=@var{name} @gol |
310668e8 JM |
403 | -mapcs-stack-check -mno-apcs-stack-check @gol |
404 | -mapcs-float -mno-apcs-float @gol | |
405 | -mapcs-reentrant -mno-apcs-reentrant @gol | |
406 | -msched-prolog -mno-sched-prolog @gol | |
407 | -mlittle-endian -mbig-endian -mwords-little-endian @gol | |
34a86306 | 408 | -mfloat-abi=@var{name} -msoft-float -mhard-float -mfpe @gol |
310668e8 | 409 | -mthumb-interwork -mno-thumb-interwork @gol |
9b66ebb1 | 410 | -mcpu=@var{name} -march=@var{name} -mfpu=@var{name} @gol |
247f8561 | 411 | -mstructure-size-boundary=@var{n} @gol |
4bc1997b | 412 | -mabort-on-noreturn @gol |
310668e8 JM |
413 | -mlong-calls -mno-long-calls @gol |
414 | -msingle-pic-base -mno-single-pic-base @gol | |
247f8561 PB |
415 | -mpic-register=@var{reg} @gol |
416 | -mnop-fun-dllimport @gol | |
9b6b54e2 | 417 | -mcirrus-fix-invalid-insns -mno-cirrus-fix-invalid-insns @gol |
247f8561 | 418 | -mpoke-function-name @gol |
310668e8 JM |
419 | -mthumb -marm @gol |
420 | -mtpcs-frame -mtpcs-leaf-frame @gol | |
d3585b76 DJ |
421 | -mcaller-super-interworking -mcallee-super-interworking @gol |
422 | -mtp=@var{name}} | |
74291a4b | 423 | |
39bc1876 NS |
424 | @emph{AVR Options} |
425 | @gccoptlist{-mmcu=@var{mcu} -msize -minit-stack=@var{n} -mno-interrupts @gol | |
426 | -mcall-prologues -mno-tablejump -mtiny-stack -mint8} | |
861bb6c1 | 427 | |
0d4a78eb | 428 | @emph{Blackfin Options} |
3fb192d2 BS |
429 | @gccoptlist{-momit-leaf-frame-pointer -mno-omit-leaf-frame-pointer @gol |
430 | -mspecld-anomaly -mno-specld-anomaly -mcsync-anomaly -mno-csync-anomaly @gol | |
431 | -mlow-64k -mno-low64k -mid-shared-library @gol | |
b6877196 BS |
432 | -mno-id-shared-library -mshared-library-id=@var{n} @gol |
433 | -mlong-calls -mno-long-calls} | |
0d4a78eb | 434 | |
39bc1876 NS |
435 | @emph{CRIS Options} |
436 | @gccoptlist{-mcpu=@var{cpu} -march=@var{cpu} -mtune=@var{cpu} @gol | |
437 | -mmax-stack-frame=@var{n} -melinux-stacksize=@var{n} @gol | |
438 | -metrax4 -metrax100 -mpdebug -mcc-init -mno-side-effects @gol | |
439 | -mstack-align -mdata-align -mconst-align @gol | |
440 | -m32-bit -m16-bit -m8-bit -mno-prologue-epilogue -mno-gotplt @gol | |
441 | -melf -maout -melinux -mlinux -sim -sim2 @gol | |
442 | -mmul-bug-workaround -mno-mul-bug-workaround} | |
74291a4b | 443 | |
53054e77 PW |
444 | @emph{CRX Options} |
445 | @gccoptlist{-mmac -mpush-args} | |
446 | ||
48aec0bc | 447 | @emph{Darwin Options} |
6d2f9dd3 JM |
448 | @gccoptlist{-all_load -allowable_client -arch -arch_errors_fatal @gol |
449 | -arch_only -bind_at_load -bundle -bundle_loader @gol | |
450 | -client_name -compatibility_version -current_version @gol | |
5079843a | 451 | -dead_strip @gol |
6d2f9dd3 JM |
452 | -dependency-file -dylib_file -dylinker_install_name @gol |
453 | -dynamic -dynamiclib -exported_symbols_list @gol | |
454 | -filelist -flat_namespace -force_cpusubtype_ALL @gol | |
455 | -force_flat_namespace -headerpad_max_install_names @gol | |
456 | -image_base -init -install_name -keep_private_externs @gol | |
457 | -multi_module -multiply_defined -multiply_defined_unused @gol | |
5079843a DP |
458 | -noall_load -no_dead_strip_inits_and_terms @gol |
459 | -nofixprebinding -nomultidefs -noprebind -noseglinkedit @gol | |
6d2f9dd3 JM |
460 | -pagezero_size -prebind -prebind_all_twolevel_modules @gol |
461 | -private_bundle -read_only_relocs -sectalign @gol | |
462 | -sectobjectsymbols -whyload -seg1addr @gol | |
463 | -sectcreate -sectobjectsymbols -sectorder @gol | |
5826770c | 464 | -segaddr -segs_read_only_addr -segs_read_write_addr @gol |
6d2f9dd3 JM |
465 | -seg_addr_table -seg_addr_table_filename -seglinkedit @gol |
466 | -segprot -segs_read_only_addr -segs_read_write_addr @gol | |
467 | -single_module -static -sub_library -sub_umbrella @gol | |
468 | -twolevel_namespace -umbrella -undefined @gol | |
469 | -unexported_symbols_list -weak_reference_mismatches @gol | |
337f2a22 | 470 | -whatsloaded -F -gused -gfull -mmacosx-version-min=@var{version} @gol |
ed5b9f96 | 471 | -mone-byte-bool} |
48aec0bc | 472 | |
74291a4b | 473 | @emph{DEC Alpha Options} |
9a94f7f3 | 474 | @gccoptlist{-mno-fp-regs -msoft-float -malpha-as -mgas @gol |
4bc1997b JM |
475 | -mieee -mieee-with-inexact -mieee-conformant @gol |
476 | -mfp-trap-mode=@var{mode} -mfp-rounding-mode=@var{mode} @gol | |
477 | -mtrap-precision=@var{mode} -mbuild-constants @gol | |
58605ba0 RH |
478 | -mcpu=@var{cpu-type} -mtune=@var{cpu-type} @gol |
479 | -mbwx -mmax -mfix -mcix @gol | |
480 | -mfloat-vax -mfloat-ieee @gol | |
9a94f7f3 JM |
481 | -mexplicit-relocs -msmall-data -mlarge-data @gol |
482 | -msmall-text -mlarge-text @gol | |
4bc1997b | 483 | -mmemory-latency=@var{time}} |
74291a4b | 484 | |
d7c23cdc | 485 | @emph{DEC Alpha/VMS Options} |
9a94f7f3 | 486 | @gccoptlist{-mvms-return-codes} |
d7c23cdc | 487 | |
39bc1876 NS |
488 | @emph{FRV Options} |
489 | @gccoptlist{-mgpr-32 -mgpr-64 -mfpr-32 -mfpr-64 @gol | |
490 | -mhard-float -msoft-float @gol | |
491 | -malloc-cc -mfixed-cc -mdword -mno-dword @gol | |
492 | -mdouble -mno-double @gol | |
493 | -mmedia -mno-media -mmuladd -mno-muladd @gol | |
c557edf4 RS |
494 | -mfdpic -minline-plt -mgprel-ro -multilib-library-pic @gol |
495 | -mlinked-fp -mlong-calls -malign-labels @gol | |
496 | -mlibrary-pic -macc-4 -macc-8 @gol | |
39bc1876 | 497 | -mpack -mno-pack -mno-eflags -mcond-move -mno-cond-move @gol |
38c28a25 | 498 | -moptimize-membar -mno-optimize-membar @gol |
39bc1876 NS |
499 | -mscc -mno-scc -mcond-exec -mno-cond-exec @gol |
500 | -mvliw-branch -mno-vliw-branch @gol | |
501 | -mmulti-cond-exec -mno-multi-cond-exec -mnested-cond-exec @gol | |
502 | -mno-nested-cond-exec -mtomcat-stats @gol | |
e4dd71de | 503 | -mTLS -mtls @gol |
39bc1876 NS |
504 | -mcpu=@var{cpu}} |
505 | ||
74291a4b | 506 | @emph{H8/300 Options} |
9a94f7f3 | 507 | @gccoptlist{-mrelax -mh -ms -mn -mint32 -malign-300} |
74291a4b | 508 | |
39bc1876 NS |
509 | @emph{HPPA Options} |
510 | @gccoptlist{-march=@var{architecture-type} @gol | |
511 | -mbig-switch -mdisable-fpregs -mdisable-indexing @gol | |
512 | -mfast-indirect-calls -mgas -mgnu-ld -mhp-ld @gol | |
a2017852 | 513 | -mfixed-range=@var{register-range} @gol |
39bc1876 NS |
514 | -mjump-in-delay -mlinker-opt -mlong-calls @gol |
515 | -mlong-load-store -mno-big-switch -mno-disable-fpregs @gol | |
516 | -mno-disable-indexing -mno-fast-indirect-calls -mno-gas @gol | |
517 | -mno-jump-in-delay -mno-long-load-store @gol | |
518 | -mno-portable-runtime -mno-soft-float @gol | |
519 | -mno-space-regs -msoft-float -mpa-risc-1-0 @gol | |
520 | -mpa-risc-1-1 -mpa-risc-2-0 -mportable-runtime @gol | |
521 | -mschedule=@var{cpu-type} -mspace-regs -msio -mwsio @gol | |
d711cf67 | 522 | -munix=@var{unix-std} -nolibdld -static -threads} |
74291a4b | 523 | |
39bc1876 NS |
524 | @emph{i386 and x86-64 Options} |
525 | @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol | |
526 | -mfpmath=@var{unit} @gol | |
527 | -masm=@var{dialect} -mno-fancy-math-387 @gol | |
528 | -mno-fp-ret-in-387 -msoft-float -msvr3-shlib @gol | |
529 | -mno-wide-multiply -mrtd -malign-double @gol | |
530 | -mpreferred-stack-boundary=@var{num} @gol | |
531 | -mmmx -msse -msse2 -msse3 -m3dnow @gol | |
532 | -mthreads -mno-align-stringops -minline-all-stringops @gol | |
533 | -mpush-args -maccumulate-outgoing-args -m128bit-long-double @gol | |
1f97667f RG |
534 | -m96bit-long-double -mregparm=@var{num} -msseregparm @gol |
535 | -momit-leaf-frame-pointer -mno-red-zone -mno-tls-direct-seg-refs @gol | |
39bc1876 | 536 | -mcmodel=@var{code-model} @gol |
7dcbf659 | 537 | -m32 -m64 -mlarge-data-threshold=@var{num}} |
56b2d7a7 | 538 | |
39bc1876 NS |
539 | @emph{IA-64 Options} |
540 | @gccoptlist{-mbig-endian -mlittle-endian -mgnu-as -mgnu-ld -mno-pic @gol | |
4c334b37 | 541 | -mvolatile-asm-stop -mregister-names -mno-sdata @gol |
39bc1876 NS |
542 | -mconstant-gp -mauto-pic -minline-float-divide-min-latency @gol |
543 | -minline-float-divide-max-throughput @gol | |
544 | -minline-int-divide-min-latency @gol | |
28b43def SE |
545 | -minline-int-divide-max-throughput @gol |
546 | -minline-sqrt-min-latency -minline-sqrt-max-throughput @gol | |
547 | -mno-dwarf2-asm -mearly-stop-bits @gol | |
548 | -mfixed-range=@var{register-range} -mtls-size=@var{tls-size} @gol | |
549 | -mtune=@var{cpu-type} -mt -pthread -milp32 -mlp64} | |
282a61e6 | 550 | |
39bc1876 NS |
551 | @emph{M32R/D Options} |
552 | @gccoptlist{-m32r2 -m32rx -m32r @gol | |
553 | -mdebug @gol | |
554 | -malign-loops -mno-align-loops @gol | |
555 | -missue-rate=@var{number} @gol | |
556 | -mbranch-cost=@var{number} @gol | |
557 | -mmodel=@var{code-size-model-type} @gol | |
558 | -msdata=@var{sdata-type} @gol | |
559 | -mno-flush-func -mflush-func=@var{name} @gol | |
560 | -mno-flush-trap -mflush-trap=@var{number} @gol | |
561 | -G @var{num}} | |
83575957 | 562 | |
38b2d076 DD |
563 | @emph{M32C Options} |
564 | @gccoptlist{-mcpu=@var{cpu} -msim -memregs=@var{number}} | |
565 | ||
39bc1876 NS |
566 | @emph{M680x0 Options} |
567 | @gccoptlist{-m68000 -m68020 -m68020-40 -m68020-60 -m68030 -m68040 @gol | |
dcc21c4c PB |
568 | -m68060 -mcpu32 -m5200 -mcfv4e -m68881 -mbitfield @gol |
569 | -mc68000 -mc68020 @gol | |
39bc1876 NS |
570 | -mnobitfield -mrtd -mshort -msoft-float -mpcrel @gol |
571 | -malign-int -mstrict-align -msep-data -mno-sep-data @gol | |
572 | -mshared-library-id=n -mid-shared-library -mno-id-shared-library} | |
789a3090 | 573 | |
39bc1876 NS |
574 | @emph{M68hc1x Options} |
575 | @gccoptlist{-m6811 -m6812 -m68hc11 -m68hc12 -m68hcs12 @gol | |
576 | -mauto-incdec -minmax -mlong-calls -mshort @gol | |
577 | -msoft-reg-count=@var{count}} | |
052a4b28 | 578 | |
789a3090 | 579 | @emph{MCore Options} |
9a94f7f3 | 580 | @gccoptlist{-mhardlit -mno-hardlit -mdiv -mno-div -mrelax-immediates @gol |
310668e8 JM |
581 | -mno-relax-immediates -mwide-bitfields -mno-wide-bitfields @gol |
582 | -m4byte-functions -mno-4byte-functions -mcallgraph-data @gol | |
583 | -mno-callgraph-data -mslow-bytes -mno-slow-bytes -mno-lsim @gol | |
584 | -mlittle-endian -mbig-endian -m210 -m340 -mstack-increment} | |
f84271d9 | 585 | |
39bc1876 NS |
586 | @emph{MIPS Options} |
587 | @gccoptlist{-EL -EB -march=@var{arch} -mtune=@var{arch} @gol | |
588 | -mips1 -mips2 -mips3 -mips4 -mips32 -mips32r2 -mips64 @gol | |
589 | -mips16 -mno-mips16 -mabi=@var{abi} -mabicalls -mno-abicalls @gol | |
590 | -mxgot -mno-xgot -mgp32 -mgp64 -mfp32 -mfp64 @gol | |
591 | -mhard-float -msoft-float -msingle-float -mdouble-float @gol | |
118ea793 | 592 | -mdsp -mpaired-single -mips3d @gol |
fb8136b2 | 593 | -mlong64 -mlong32 -msym32 -mno-sym32 @gol |
39bc1876 NS |
594 | -G@var{num} -membedded-data -mno-embedded-data @gol |
595 | -muninit-const-in-rodata -mno-uninit-const-in-rodata @gol | |
596 | -msplit-addresses -mno-split-addresses @gol | |
597 | -mexplicit-relocs -mno-explicit-relocs @gol | |
39bc1876 | 598 | -mcheck-zero-division -mno-check-zero-division @gol |
9f0df97a | 599 | -mdivide-traps -mdivide-breaks @gol |
39bc1876 NS |
600 | -mmemcpy -mno-memcpy -mlong-calls -mno-long-calls @gol |
601 | -mmad -mno-mad -mfused-madd -mno-fused-madd -nocpp @gol | |
602 | -mfix-r4000 -mno-fix-r4000 -mfix-r4400 -mno-fix-r4400 @gol | |
0ac40e7a RS |
603 | -mfix-vr4120 -mno-fix-vr4120 -mfix-vr4130 @gol |
604 | -mfix-sb1 -mno-fix-sb1 @gol | |
39bc1876 NS |
605 | -mflush-func=@var{func} -mno-flush-func @gol |
606 | -mbranch-likely -mno-branch-likely @gol | |
607 | -mfp-exceptions -mno-fp-exceptions @gol | |
608 | -mvr4130-align -mno-vr4130-align} | |
bcf684c7 | 609 | |
39bc1876 NS |
610 | @emph{MMIX Options} |
611 | @gccoptlist{-mlibfuncs -mno-libfuncs -mepsilon -mno-epsilon -mabi=gnu @gol | |
612 | -mabi=mmixware -mzero-extend -mknuthdiv -mtoplevel-symbols @gol | |
613 | -melf -mbranch-predict -mno-branch-predict -mbase-addresses @gol | |
614 | -mno-base-addresses -msingle-exit -mno-single-exit} | |
df6194d4 | 615 | |
39bc1876 NS |
616 | @emph{MN10300 Options} |
617 | @gccoptlist{-mmult-bug -mno-mult-bug @gol | |
618 | -mam33 -mno-am33 @gol | |
619 | -mam33-2 -mno-am33-2 @gol | |
b1eb8119 | 620 | -mreturn-pointer-on-d0 @gol |
39bc1876 | 621 | -mno-crt0 -mrelax} |
91abf72d | 622 | |
6c9ac67a NS |
623 | @emph{MT Options} |
624 | @gccoptlist{-mno-crt0 -mbacc -msim @gol | |
c85ba4fb AH |
625 | -march=@var{cpu-type} } |
626 | ||
9f85bca7 | 627 | @emph{PDP-11 Options} |
9a94f7f3 | 628 | @gccoptlist{-mfpu -msoft-float -mac0 -mno-ac0 -m40 -m45 -m10 @gol |
9f85bca7 JM |
629 | -mbcopy -mbcopy-builtin -mint32 -mno-int16 @gol |
630 | -mint16 -mno-int32 -mfloat32 -mno-float64 @gol | |
631 | -mfloat64 -mno-float32 -mabshi -mno-abshi @gol | |
632 | -mbranch-expensive -mbranch-cheap @gol | |
633 | -msplit -mno-split -munix-asm -mdec-asm} | |
634 | ||
39bc1876 NS |
635 | @emph{PowerPC Options} |
636 | See RS/6000 and PowerPC Options. | |
637 | ||
638 | @emph{RS/6000 and PowerPC Options} | |
639 | @gccoptlist{-mcpu=@var{cpu-type} @gol | |
640 | -mtune=@var{cpu-type} @gol | |
641 | -mpower -mno-power -mpower2 -mno-power2 @gol | |
642 | -mpowerpc -mpowerpc64 -mno-powerpc @gol | |
643 | -maltivec -mno-altivec @gol | |
644 | -mpowerpc-gpopt -mno-powerpc-gpopt @gol | |
645 | -mpowerpc-gfxopt -mno-powerpc-gfxopt @gol | |
9719f3b7 | 646 | -mmfcrf -mno-mfcrf -mpopcntb -mno-popcntb -mfprnd -mno-fprnd @gol |
39bc1876 NS |
647 | -mnew-mnemonics -mold-mnemonics @gol |
648 | -mfull-toc -mminimal-toc -mno-fp-in-toc -mno-sum-in-toc @gol | |
de17c25f | 649 | -m64 -m32 -mxl-compat -mno-xl-compat -mpe @gol |
39bc1876 NS |
650 | -malign-power -malign-natural @gol |
651 | -msoft-float -mhard-float -mmultiple -mno-multiple @gol | |
652 | -mstring -mno-string -mupdate -mno-update @gol | |
653 | -mfused-madd -mno-fused-madd -mbit-align -mno-bit-align @gol | |
654 | -mstrict-align -mno-strict-align -mrelocatable @gol | |
655 | -mno-relocatable -mrelocatable-lib -mno-relocatable-lib @gol | |
656 | -mtoc -mno-toc -mlittle -mlittle-endian -mbig -mbig-endian @gol | |
ef765ea9 | 657 | -mdynamic-no-pic -maltivec -mswdiv @gol |
39bc1876 NS |
658 | -mprioritize-restricted-insns=@var{priority} @gol |
659 | -msched-costly-dep=@var{dependence_type} @gol | |
660 | -minsert-sched-nops=@var{scheme} @gol | |
661 | -mcall-sysv -mcall-netbsd @gol | |
662 | -maix-struct-return -msvr4-struct-return @gol | |
7f970b70 | 663 | -mabi=@var{abi-type} -msecure-plt -mbss-plt @gol |
78f5898b | 664 | -misel -mno-isel @gol |
39bc1876 | 665 | -misel=yes -misel=no @gol |
78f5898b | 666 | -mspe -mno-spe @gol |
39bc1876 | 667 | -mspe=yes -mspe=no @gol |
78f5898b | 668 | -mvrsave -mno-vrsave @gol |
131aeb82 | 669 | -mmulhw -mno-mulhw @gol |
4d4cbc0e | 670 | -mfloat-gprs=yes -mfloat-gprs=no -mfloat-gprs=single -mfloat-gprs=double @gol |
39bc1876 NS |
671 | -mprototype -mno-prototype @gol |
672 | -msim -mmvme -mads -myellowknife -memb -msdata @gol | |
673 | -msdata=@var{opt} -mvxworks -mwindiss -G @var{num} -pthread} | |
674 | ||
675 | @emph{S/390 and zSeries Options} | |
676 | @gccoptlist{-mtune=@var{cpu-type} -march=@var{cpu-type} @gol | |
b3d31392 AK |
677 | -mhard-float -msoft-float -mbackchain -mno-backchain @gol |
678 | -mpacked-stack -mno-packed-stack @gol | |
39bc1876 NS |
679 | -msmall-exec -mno-small-exec -mmvcle -mno-mvcle @gol |
680 | -m64 -m31 -mdebug -mno-debug -mesa -mzarch @gol | |
d75f90f1 AK |
681 | -mtpf-trace -mno-tpf-trace -mfused-madd -mno-fused-madd @gol |
682 | -mwarn-framesize -mwarn-dynamicstack -mstack-size -mstack-guard} | |
39bc1876 NS |
683 | |
684 | @emph{SH Options} | |
685 | @gccoptlist{-m1 -m2 -m2e -m3 -m3e @gol | |
686 | -m4-nofpu -m4-single-only -m4-single -m4 @gol | |
312209c6 | 687 | -m4a-nofpu -m4a-single-only -m4a-single -m4a -m4al @gol |
39bc1876 NS |
688 | -m5-64media -m5-64media-nofpu @gol |
689 | -m5-32media -m5-32media-nofpu @gol | |
690 | -m5-compact -m5-compact-nofpu @gol | |
691 | -mb -ml -mdalign -mrelax @gol | |
2acc29bd | 692 | -mbigtable -mfmovd -mhitachi -mrenesas -mno-renesas -mnomacsave @gol |
39bc1876 | 693 | -mieee -misize -mpadstruct -mspace @gol |
73a4d10b R |
694 | -mprefergot -musermode -multcost=@var{number} -mdiv=@var{strategy} @gol |
695 | -mdivsi3_libfunc=@var{name} @gol | |
696 | -madjust-unroll -mindexed-addressing -mgettrcost=@var{number} -mpt-fixed @gol | |
697 | -minvalid-symbols} | |
39bc1876 NS |
698 | |
699 | @emph{SPARC Options} | |
700 | @gccoptlist{-mcpu=@var{cpu-type} @gol | |
701 | -mtune=@var{cpu-type} @gol | |
702 | -mcmodel=@var{code-model} @gol | |
703 | -m32 -m64 -mapp-regs -mno-app-regs @gol | |
704 | -mfaster-structs -mno-faster-structs @gol | |
705 | -mfpu -mno-fpu -mhard-float -msoft-float @gol | |
706 | -mhard-quad-float -msoft-quad-float @gol | |
707 | -mimpure-text -mno-impure-text -mlittle-endian @gol | |
708 | -mstack-bias -mno-stack-bias @gol | |
709 | -munaligned-doubles -mno-unaligned-doubles @gol | |
6bfb2f93 EB |
710 | -mv8plus -mno-v8plus -mvis -mno-vis |
711 | -threads -pthreads} | |
39bc1876 NS |
712 | |
713 | @emph{System V Options} | |
714 | @gccoptlist{-Qy -Qn -YP,@var{paths} -Ym,@var{dir}} | |
715 | ||
716 | @emph{TMS320C3x/C4x Options} | |
717 | @gccoptlist{-mcpu=@var{cpu} -mbig -msmall -mregparm -mmemparm @gol | |
718 | -mfast-fix -mmpyi -mbk -mti -mdp-isr-reload @gol | |
719 | -mrpts=@var{count} -mrptb -mdb -mloop-unsigned @gol | |
720 | -mparallel-insns -mparallel-mpy -mpreserve-float} | |
721 | ||
722 | @emph{V850 Options} | |
723 | @gccoptlist{-mlong-calls -mno-long-calls -mep -mno-ep @gol | |
724 | -mprolog-function -mno-prolog-function -mspace @gol | |
725 | -mtda=@var{n} -msda=@var{n} -mzda=@var{n} @gol | |
726 | -mapp-regs -mno-app-regs @gol | |
727 | -mdisable-callt -mno-disable-callt @gol | |
728 | -mv850e1 @gol | |
729 | -mv850e @gol | |
730 | -mv850 -mbig-switch} | |
731 | ||
732 | @emph{VAX Options} | |
733 | @gccoptlist{-mg -mgnu -munix} | |
734 | ||
735 | @emph{x86-64 Options} | |
736 | See i386 and x86-64 Options. | |
737 | ||
69a0611f | 738 | @emph{Xstormy16 Options} |
9a94f7f3 | 739 | @gccoptlist{-msim} |
69a0611f | 740 | |
03984308 | 741 | @emph{Xtensa Options} |
6cedbe44 | 742 | @gccoptlist{-mconst16 -mno-const16 @gol |
9a94f7f3 | 743 | -mfused-madd -mno-fused-madd @gol |
9a94f7f3 JM |
744 | -mtext-section-literals -mno-text-section-literals @gol |
745 | -mtarget-align -mno-target-align @gol | |
746 | -mlongcalls -mno-longcalls} | |
03984308 | 747 | |
39bc1876 NS |
748 | @emph{zSeries Options} |
749 | See S/390 and zSeries Options. | |
70899148 | 750 | |
74291a4b MM |
751 | @item Code Generation Options |
752 | @xref{Code Gen Options,,Options for Code Generation Conventions}. | |
9a94f7f3 JM |
753 | @gccoptlist{-fcall-saved-@var{reg} -fcall-used-@var{reg} @gol |
754 | -ffixed-@var{reg} -fexceptions @gol | |
5d22c1a5 | 755 | -fnon-call-exceptions -funwind-tables @gol |
a944ceb9 | 756 | -fasynchronous-unwind-tables @gol |
4bc1997b | 757 | -finhibit-size-directive -finstrument-functions @gol |
dc170a87 | 758 | -fno-common -fno-ident @gol |
24a4dd31 | 759 | -fpcc-struct-return -fpic -fPIC -fpie -fPIE @gol |
82c0180d | 760 | -fno-jump-tables @gol |
4bc1997b | 761 | -freg-struct-return -fshared-data -fshort-enums @gol |
271bd540 | 762 | -fshort-double -fshort-wchar @gol |
467cecf3 | 763 | -fverbose-asm -fpack-struct[=@var{n}] -fstack-check @gol |
4bc1997b JM |
764 | -fstack-limit-register=@var{reg} -fstack-limit-symbol=@var{sym} @gol |
765 | -fargument-alias -fargument-noalias @gol | |
478c9e72 | 766 | -fargument-noalias-global -fleading-underscore @gol |
d4463dfc | 767 | -ftls-model=@var{model} @gol |
d7afec4b | 768 | -ftrapv -fwrapv -fbounds-check @gol |
953ff289 | 769 | -fvisibility -fopenmp} |
74291a4b MM |
770 | @end table |
771 | ||
772 | @menu | |
773 | * Overall Options:: Controlling the kind of output: | |
774 | an executable, object files, assembler files, | |
775 | or preprocessed source. | |
776 | * C Dialect Options:: Controlling the variant of C language compiled. | |
777 | * C++ Dialect Options:: Variations on C++. | |
46e34f96 ZL |
778 | * Objective-C and Objective-C++ Dialect Options:: Variations on Objective-C |
779 | and Objective-C++. | |
764dbbf2 | 780 | * Language Independent Options:: Controlling how diagnostics should be |
02f52e19 | 781 | formatted. |
74291a4b MM |
782 | * Warning Options:: How picky should the compiler be? |
783 | * Debugging Options:: Symbol tables, measurements, and debugging dumps. | |
784 | * Optimize Options:: How much optimization? | |
785 | * Preprocessor Options:: Controlling header files and macro definitions. | |
786 | Also, getting dependency information for Make. | |
787 | * Assembler Options:: Passing options to the assembler. | |
788 | * Link Options:: Specifying libraries and so on. | |
789 | * Directory Options:: Where to find header files and libraries. | |
790 | Where to find the compiler executable files. | |
a743d340 | 791 | * Spec Files:: How to pass switches to sub-processes. |
0c2d1a2a | 792 | * Target Options:: Running a cross-compiler, or an old version of GCC. |
74291a4b MM |
793 | @end menu |
794 | ||
795 | @node Overall Options | |
796 | @section Options Controlling the Kind of Output | |
797 | ||
798 | Compilation can involve up to four stages: preprocessing, compilation | |
d1bd0ded GK |
799 | proper, assembly and linking, always in that order. GCC is capable of |
800 | preprocessing and compiling several files either into several | |
801 | assembler input files, or into one assembler input file; then each | |
802 | assembler input file produces an object file, and linking combines all | |
803 | the object files (those newly compiled, and those specified as input) | |
804 | into an executable file. | |
74291a4b MM |
805 | |
806 | @cindex file name suffix | |
807 | For any given input file, the file name suffix determines what kind of | |
808 | compilation is done: | |
809 | ||
2642624b | 810 | @table @gcctabopt |
74291a4b MM |
811 | @item @var{file}.c |
812 | C source code which must be preprocessed. | |
813 | ||
814 | @item @var{file}.i | |
815 | C source code which should not be preprocessed. | |
816 | ||
817 | @item @var{file}.ii | |
818 | C++ source code which should not be preprocessed. | |
819 | ||
820 | @item @var{file}.m | |
46e34f96 ZL |
821 | Objective-C source code. Note that you must link with the @file{libobjc} |
822 | library to make an Objective-C program work. | |
74291a4b | 823 | |
b9265ec1 JM |
824 | @item @var{file}.mi |
825 | Objective-C source code which should not be preprocessed. | |
826 | ||
46e34f96 ZL |
827 | @item @var{file}.mm |
828 | @itemx @var{file}.M | |
829 | Objective-C++ source code. Note that you must link with the @file{libobjc} | |
830 | library to make an Objective-C++ program work. Note that @samp{.M} refers | |
831 | to a literal capital M@. | |
832 | ||
833 | @item @var{file}.mii | |
834 | Objective-C++ source code which should not be preprocessed. | |
835 | ||
74291a4b | 836 | @item @var{file}.h |
46e34f96 ZL |
837 | C, C++, Objective-C or Objective-C++ header file to be turned into a |
838 | precompiled header. | |
74291a4b MM |
839 | |
840 | @item @var{file}.cc | |
b9265ec1 | 841 | @itemx @var{file}.cp |
74291a4b MM |
842 | @itemx @var{file}.cxx |
843 | @itemx @var{file}.cpp | |
ee8acf89 | 844 | @itemx @var{file}.CPP |
b9265ec1 | 845 | @itemx @var{file}.c++ |
74291a4b MM |
846 | @itemx @var{file}.C |
847 | C++ source code which must be preprocessed. Note that in @samp{.cxx}, | |
848 | the last two letters must both be literally @samp{x}. Likewise, | |
161d7b59 | 849 | @samp{.C} refers to a literal capital C@. |
74291a4b | 850 | |
6e955430 ZL |
851 | @item @var{file}.mm |
852 | @itemx @var{file}.M | |
853 | Objective-C++ source code which must be preprocessed. | |
854 | ||
855 | @item @var{file}.mii | |
856 | Objective-C++ source code which should not be preprocessed. | |
857 | ||
17211ab5 GK |
858 | @item @var{file}.hh |
859 | @itemx @var{file}.H | |
860 | C++ header file to be turned into a precompiled header. | |
861 | ||
b9265ec1 JM |
862 | @item @var{file}.f |
863 | @itemx @var{file}.for | |
864 | @itemx @var{file}.FOR | |
80a0c50a | 865 | Fixed form Fortran source code which should not be preprocessed. |
b9265ec1 JM |
866 | |
867 | @item @var{file}.F | |
868 | @itemx @var{file}.fpp | |
869 | @itemx @var{file}.FPP | |
80a0c50a | 870 | Fixed form Fortran source code which must be preprocessed (with the traditional |
b9265ec1 JM |
871 | preprocessor). |
872 | ||
6de9cd9a DN |
873 | @item @var{file}.f90 |
874 | @itemx @var{file}.f95 | |
80a0c50a | 875 | Free form Fortran source code which should not be preprocessed. |
6de9cd9a | 876 | |
5a006700 MR |
877 | @item @var{file}.F90 |
878 | @itemx @var{file}.F95 | |
80a0c50a | 879 | Free form Fortran source code which must be preprocessed (with the |
5a006700 MR |
880 | traditional preprocessor). |
881 | ||
b9265ec1 JM |
882 | @c FIXME: Descriptions of Java file types. |
883 | @c @var{file}.java | |
884 | @c @var{file}.class | |
885 | @c @var{file}.zip | |
886 | @c @var{file}.jar | |
887 | ||
e23381df GB |
888 | @item @var{file}.ads |
889 | Ada source code file which contains a library unit declaration (a | |
890 | declaration of a package, subprogram, or generic, or a generic | |
891 | instantiation), or a library unit renaming declaration (a package, | |
892 | generic, or subprogram renaming declaration). Such files are also | |
893 | called @dfn{specs}. | |
894 | ||
895 | @itemx @var{file}.adb | |
896 | Ada source code file containing a library unit body (a subprogram or | |
897 | package body). Such files are also called @dfn{bodies}. | |
898 | ||
b9265ec1 | 899 | @c GCC also knows about some suffixes for languages not yet included: |
b9265ec1 JM |
900 | @c Pascal: |
901 | @c @var{file}.p | |
902 | @c @var{file}.pas | |
80a0c50a TS |
903 | @c Ratfor: |
904 | @c @var{file}.r | |
b9265ec1 | 905 | |
74291a4b MM |
906 | @item @var{file}.s |
907 | Assembler code. | |
908 | ||
909 | @item @var{file}.S | |
910 | Assembler code which must be preprocessed. | |
911 | ||
912 | @item @var{other} | |
913 | An object file to be fed straight into linking. | |
914 | Any file name with no recognized suffix is treated this way. | |
915 | @end table | |
916 | ||
cd3bb277 | 917 | @opindex x |
630d3d5a | 918 | You can specify the input language explicitly with the @option{-x} option: |
74291a4b | 919 | |
2642624b | 920 | @table @gcctabopt |
74291a4b MM |
921 | @item -x @var{language} |
922 | Specify explicitly the @var{language} for the following input files | |
923 | (rather than letting the compiler choose a default based on the file | |
924 | name suffix). This option applies to all following input files until | |
630d3d5a | 925 | the next @option{-x} option. Possible values for @var{language} are: |
3ab51846 | 926 | @smallexample |
46e34f96 | 927 | c c-header c-cpp-output |
17211ab5 | 928 | c++ c++-header c++-cpp-output |
46e34f96 ZL |
929 | objective-c objective-c-header objective-c-cpp-output |
930 | objective-c++ objective-c++-header objective-c++-cpp-output | |
74291a4b | 931 | assembler assembler-with-cpp |
e23381df | 932 | ada |
80a0c50a | 933 | f77 f77-cpp-input |
acd1a829 | 934 | f95 f95-cpp-input |
e23381df | 935 | java |
b38b97c4 | 936 | treelang |
3ab51846 | 937 | @end smallexample |
74291a4b MM |
938 | |
939 | @item -x none | |
940 | Turn off any specification of a language, so that subsequent files are | |
630d3d5a | 941 | handled according to their file name suffixes (as they are if @option{-x} |
74291a4b | 942 | has not been used at all). |
14a774a9 RK |
943 | |
944 | @item -pass-exit-codes | |
cd3bb277 | 945 | @opindex pass-exit-codes |
bedc7537 | 946 | Normally the @command{gcc} program will exit with the code of 1 if any |
14a774a9 | 947 | phase of the compiler returns a non-success return code. If you specify |
630d3d5a | 948 | @option{-pass-exit-codes}, the @command{gcc} program will instead return with |
14a774a9 RK |
949 | numerically highest error produced by any phase that returned an error |
950 | indication. | |
74291a4b MM |
951 | @end table |
952 | ||
953 | If you only want some of the stages of compilation, you can use | |
630d3d5a JM |
954 | @option{-x} (or filename suffixes) to tell @command{gcc} where to start, and |
955 | one of the options @option{-c}, @option{-S}, or @option{-E} to say where | |
bedc7537 NC |
956 | @command{gcc} is to stop. Note that some combinations (for example, |
957 | @samp{-x cpp-output -E}) instruct @command{gcc} to do nothing at all. | |
74291a4b | 958 | |
2642624b | 959 | @table @gcctabopt |
74291a4b | 960 | @item -c |
cd3bb277 | 961 | @opindex c |
74291a4b MM |
962 | Compile or assemble the source files, but do not link. The linking |
963 | stage simply is not done. The ultimate output is in the form of an | |
964 | object file for each source file. | |
965 | ||
966 | By default, the object file name for a source file is made by replacing | |
967 | the suffix @samp{.c}, @samp{.i}, @samp{.s}, etc., with @samp{.o}. | |
968 | ||
969 | Unrecognized input files, not requiring compilation or assembly, are | |
970 | ignored. | |
971 | ||
972 | @item -S | |
cd3bb277 | 973 | @opindex S |
74291a4b MM |
974 | Stop after the stage of compilation proper; do not assemble. The output |
975 | is in the form of an assembler code file for each non-assembler input | |
976 | file specified. | |
977 | ||
978 | By default, the assembler file name for a source file is made by | |
979 | replacing the suffix @samp{.c}, @samp{.i}, etc., with @samp{.s}. | |
980 | ||
981 | Input files that don't require compilation are ignored. | |
982 | ||
983 | @item -E | |
cd3bb277 | 984 | @opindex E |
74291a4b MM |
985 | Stop after the preprocessing stage; do not run the compiler proper. The |
986 | output is in the form of preprocessed source code, which is sent to the | |
987 | standard output. | |
988 | ||
989 | Input files which don't require preprocessing are ignored. | |
990 | ||
991 | @cindex output file option | |
992 | @item -o @var{file} | |
cd3bb277 | 993 | @opindex o |
74291a4b MM |
994 | Place output in file @var{file}. This applies regardless to whatever |
995 | sort of output is being produced, whether it be an executable file, | |
996 | an object file, an assembler file or preprocessed C code. | |
997 | ||
488061c8 GK |
998 | If @option{-o} is not specified, the default is to put an executable |
999 | file in @file{a.out}, the object file for | |
1000 | @file{@var{source}.@var{suffix}} in @file{@var{source}.o}, its | |
1001 | assembler file in @file{@var{source}.s}, a precompiled header file in | |
1002 | @file{@var{source}.@var{suffix}.gch}, and all preprocessed C source on | |
1003 | standard output. | |
74291a4b MM |
1004 | |
1005 | @item -v | |
cd3bb277 | 1006 | @opindex v |
74291a4b MM |
1007 | Print (on standard error output) the commands executed to run the stages |
1008 | of compilation. Also print the version number of the compiler driver | |
1009 | program and of the preprocessor and the compiler proper. | |
1010 | ||
e8b3c8ac IR |
1011 | @item -### |
1012 | @opindex ### | |
1013 | Like @option{-v} except the commands are not executed and all command | |
1014 | arguments are quoted. This is useful for shell scripts to capture the | |
1015 | driver-generated command lines. | |
1016 | ||
74291a4b | 1017 | @item -pipe |
cd3bb277 | 1018 | @opindex pipe |
74291a4b MM |
1019 | Use pipes rather than temporary files for communication between the |
1020 | various stages of compilation. This fails to work on some systems where | |
1021 | the assembler is unable to read from a pipe; but the GNU assembler has | |
1022 | no trouble. | |
844642e6 | 1023 | |
0855eab7 CT |
1024 | @item -combine |
1025 | @opindex combine | |
1026 | If you are compiling multiple source files, this option tells the driver | |
f26c1794 | 1027 | to pass all the source files to the compiler at once (for those |
0855eab7 CT |
1028 | languages for which the compiler can handle this). This will allow |
1029 | intermodule analysis (IMA) to be performed by the compiler. Currently the only | |
78466c0e | 1030 | language for which this is supported is C@. If you pass source files for |
0855eab7 CT |
1031 | multiple languages to the driver, using this option, the driver will invoke |
1032 | the compiler(s) that support IMA once each, passing each compiler all the | |
1033 | source files appropriate for it. For those languages that do not support | |
1034 | IMA this option will be ignored, and the compiler will be invoked once for | |
1035 | each source file in that language. If you use this option in conjunction | |
78466c0e JM |
1036 | with @option{-save-temps}, the compiler will generate multiple |
1037 | pre-processed files | |
1038 | (one for each source file), but only one (combined) @file{.o} or | |
1039 | @file{.s} file. | |
0855eab7 | 1040 | |
844642e6 | 1041 | @item --help |
cd3bb277 | 1042 | @opindex help |
844642e6 | 1043 | Print (on the standard output) a description of the command line options |
bedc7537 NC |
1044 | understood by @command{gcc}. If the @option{-v} option is also specified |
1045 | then @option{--help} will also be passed on to the various processes | |
1046 | invoked by @command{gcc}, so that they can display the command line options | |
65ca2d60 | 1047 | they accept. If the @option{-Wextra} option is also specified then command |
844642e6 NC |
1048 | line options which have no documentation associated with them will also |
1049 | be displayed. | |
10501d8f CC |
1050 | |
1051 | @item --target-help | |
cd3bb277 | 1052 | @opindex target-help |
10501d8f CC |
1053 | Print (on the standard output) a description of target specific command |
1054 | line options for each tool. | |
e03b7153 RS |
1055 | |
1056 | @item --version | |
1057 | @opindex version | |
8a36672b | 1058 | Display the version number and copyrights of the invoked GCC@. |
9d530538 MM |
1059 | |
1060 | @include @value{srcdir}/../libiberty/at-file.texi | |
74291a4b MM |
1061 | @end table |
1062 | ||
1063 | @node Invoking G++ | |
1064 | @section Compiling C++ Programs | |
1065 | ||
1066 | @cindex suffixes for C++ source | |
1067 | @cindex C++ source file suffixes | |
1068 | C++ source files conventionally use one of the suffixes @samp{.C}, | |
17211ab5 GK |
1069 | @samp{.cc}, @samp{.cpp}, @samp{.CPP}, @samp{.c++}, @samp{.cp}, or |
1070 | @samp{.cxx}; C++ header files often use @samp{.hh} or @samp{.H}; and | |
0c2d1a2a | 1071 | preprocessed C++ files use the suffix @samp{.ii}. GCC recognizes |
bba975d4 | 1072 | files with these names and compiles them as C++ programs even if you |
17211ab5 GK |
1073 | call the compiler the same way as for compiling C programs (usually |
1074 | with the name @command{gcc}). | |
74291a4b MM |
1075 | |
1076 | @findex g++ | |
1077 | @findex c++ | |
1078 | However, C++ programs often require class libraries as well as a | |
1079 | compiler that understands the C++ language---and under some | |
17211ab5 GK |
1080 | circumstances, you might want to compile programs or header files from |
1081 | standard input, or otherwise without a suffix that flags them as C++ | |
1082 | programs. You might also like to precompile a C header file with a | |
1083 | @samp{.h} extension to be used in C++ compilations. @command{g++} is a | |
1084 | program that calls GCC with the default language set to C++, and | |
1085 | automatically specifies linking against the C++ library. On many | |
1086 | systems, @command{g++} is also installed with the name @command{c++}. | |
74291a4b | 1087 | |
bedc7537 | 1088 | @cindex invoking @command{g++} |
74291a4b MM |
1089 | When you compile C++ programs, you may specify many of the same |
1090 | command-line options that you use for compiling programs in any | |
1091 | language; or command-line options meaningful for C and related | |
1092 | languages; or options that are meaningful only for C++ programs. | |
1093 | @xref{C Dialect Options,,Options Controlling C Dialect}, for | |
161d7b59 | 1094 | explanations of options for languages related to C@. |
74291a4b MM |
1095 | @xref{C++ Dialect Options,,Options Controlling C++ Dialect}, for |
1096 | explanations of options that are meaningful only for C++ programs. | |
1097 | ||
1098 | @node C Dialect Options | |
1099 | @section Options Controlling C Dialect | |
1100 | @cindex dialect options | |
1101 | @cindex language dialect options | |
1102 | @cindex options, dialect | |
1103 | ||
1104 | The following options control the dialect of C (or languages derived | |
46e34f96 ZL |
1105 | from C, such as C++, Objective-C and Objective-C++) that the compiler |
1106 | accepts: | |
74291a4b | 1107 | |
2642624b | 1108 | @table @gcctabopt |
74291a4b | 1109 | @cindex ANSI support |
c1030c7c | 1110 | @cindex ISO support |
74291a4b | 1111 | @item -ansi |
cd3bb277 | 1112 | @opindex ansi |
3764f879 | 1113 | In C mode, support all ISO C90 programs. In C++ mode, |
775afb25 | 1114 | remove GNU extensions that conflict with ISO C++. |
74291a4b | 1115 | |
c1030c7c | 1116 | This turns off certain features of GCC that are incompatible with ISO |
3764f879 | 1117 | C90 (when compiling C code), or of standard C++ (when compiling C++ code), |
0c2d1a2a | 1118 | such as the @code{asm} and @code{typeof} keywords, and |
74291a4b MM |
1119 | predefined macros such as @code{unix} and @code{vax} that identify the |
1120 | type of system you are using. It also enables the undesirable and | |
02f52e19 | 1121 | rarely used ISO trigraph feature. For the C compiler, |
0c2d1a2a | 1122 | it disables recognition of C++ style @samp{//} comments as well as |
775afb25 | 1123 | the @code{inline} keyword. |
74291a4b MM |
1124 | |
1125 | The alternate keywords @code{__asm__}, @code{__extension__}, | |
1126 | @code{__inline__} and @code{__typeof__} continue to work despite | |
630d3d5a | 1127 | @option{-ansi}. You would not want to use them in an ISO C program, of |
74291a4b | 1128 | course, but it is useful to put them in header files that might be included |
630d3d5a | 1129 | in compilations done with @option{-ansi}. Alternate predefined macros |
74291a4b | 1130 | such as @code{__unix__} and @code{__vax__} are also available, with or |
630d3d5a | 1131 | without @option{-ansi}. |
74291a4b | 1132 | |
630d3d5a JM |
1133 | The @option{-ansi} option does not cause non-ISO programs to be |
1134 | rejected gratuitously. For that, @option{-pedantic} is required in | |
1135 | addition to @option{-ansi}. @xref{Warning Options}. | |
74291a4b | 1136 | |
630d3d5a | 1137 | The macro @code{__STRICT_ANSI__} is predefined when the @option{-ansi} |
74291a4b MM |
1138 | option is used. Some header files may notice this macro and refrain |
1139 | from declaring certain functions or defining certain macros that the | |
c1030c7c | 1140 | ISO standard doesn't call for; this is to avoid interfering with any |
74291a4b MM |
1141 | programs that might use these names for other things. |
1142 | ||
c771326b JM |
1143 | Functions which would normally be built in but do not have semantics |
1144 | defined by ISO C (such as @code{alloca} and @code{ffs}) are not built-in | |
630d3d5a | 1145 | functions with @option{-ansi} is used. @xref{Other Builtins,,Other |
f0523f02 | 1146 | built-in functions provided by GCC}, for details of the functions |
01702459 | 1147 | affected. |
74291a4b | 1148 | |
49419c8f | 1149 | @item -std= |
cd3bb277 | 1150 | @opindex std |
aee96fe9 | 1151 | Determine the language standard. This option is currently only |
f749a36b NB |
1152 | supported when compiling C or C++. A value for this option must be |
1153 | provided; possible values are | |
3932261a | 1154 | |
ee457005 | 1155 | @table @samp |
aee96fe9 JM |
1156 | @item c89 |
1157 | @itemx iso9899:1990 | |
3764f879 | 1158 | ISO C90 (same as @option{-ansi}). |
3043b30e ML |
1159 | |
1160 | @item iso9899:199409 | |
3764f879 | 1161 | ISO C90 as modified in amendment 1. |
3043b30e | 1162 | |
49419c8f | 1163 | @item c99 |
aee96fe9 JM |
1164 | @itemx c9x |
1165 | @itemx iso9899:1999 | |
1166 | @itemx iso9899:199x | |
1167 | ISO C99. Note that this standard is not yet fully supported; see | |
1168 | @w{@uref{http://gcc.gnu.org/c99status.html}} for more information. The | |
1169 | names @samp{c9x} and @samp{iso9899:199x} are deprecated. | |
3043b30e ML |
1170 | |
1171 | @item gnu89 | |
3764f879 | 1172 | Default, ISO C90 plus GNU extensions (including some C99 features). |
3043b30e | 1173 | |
49419c8f | 1174 | @item gnu99 |
31775d31 | 1175 | @itemx gnu9x |
d15a05b3 EC |
1176 | ISO C99 plus GNU extensions. When ISO C99 is fully implemented in GCC, |
1177 | this will become the default. The name @samp{gnu9x} is deprecated. | |
49419c8f | 1178 | |
f749a36b NB |
1179 | @item c++98 |
1180 | The 1998 ISO C++ standard plus amendments. | |
1181 | ||
1182 | @item gnu++98 | |
1183 | The same as @option{-std=c++98} plus GNU extensions. This is the | |
1184 | default for C++ code. | |
ee457005 | 1185 | @end table |
3043b30e ML |
1186 | |
1187 | Even when this option is not specified, you can still use some of the | |
1188 | features of newer standards in so far as they do not conflict with | |
1189 | previous C standards. For example, you may use @code{__restrict__} even | |
bedc7537 | 1190 | when @option{-std=c99} is not specified. |
3932261a | 1191 | |
5490d604 | 1192 | The @option{-std} options specifying some version of ISO C have the same |
3764f879 | 1193 | effects as @option{-ansi}, except that features that were not in ISO C90 |
5490d604 JM |
1194 | but are in the specified version (for example, @samp{//} comments and |
1195 | the @code{inline} keyword in ISO C99) are not disabled. | |
1196 | ||
c1030c7c JM |
1197 | @xref{Standards,,Language Standards Supported by GCC}, for details of |
1198 | these standard versions. | |
1199 | ||
b1018de6 AO |
1200 | @item -aux-info @var{filename} |
1201 | @opindex aux-info | |
1202 | Output to the given filename prototyped declarations for all functions | |
1203 | declared and/or defined in a translation unit, including those in header | |
161d7b59 | 1204 | files. This option is silently ignored in any language other than C@. |
b1018de6 AO |
1205 | |
1206 | Besides declarations, the file indicates, in comments, the origin of | |
1207 | each declaration (source file and line), whether the declaration was | |
1208 | implicit, prototyped or unprototyped (@samp{I}, @samp{N} for new or | |
1209 | @samp{O} for old, respectively, in the first character after the line | |
1210 | number and the colon), and whether it came from a declaration or a | |
1211 | definition (@samp{C} or @samp{F}, respectively, in the following | |
1212 | character). In the case of function definitions, a K&R-style list of | |
1213 | arguments followed by their declarations is also provided, inside | |
1214 | comments, after the declaration. | |
1215 | ||
74291a4b | 1216 | @item -fno-asm |
cd3bb277 | 1217 | @opindex fno-asm |
74291a4b MM |
1218 | Do not recognize @code{asm}, @code{inline} or @code{typeof} as a |
1219 | keyword, so that code can use these words as identifiers. You can use | |
1220 | the keywords @code{__asm__}, @code{__inline__} and @code{__typeof__} | |
630d3d5a | 1221 | instead. @option{-ansi} implies @option{-fno-asm}. |
74291a4b MM |
1222 | |
1223 | In C++, this switch only affects the @code{typeof} keyword, since | |
1224 | @code{asm} and @code{inline} are standard keywords. You may want to | |
630d3d5a | 1225 | use the @option{-fno-gnu-keywords} flag instead, which has the same |
5490d604 JM |
1226 | effect. In C99 mode (@option{-std=c99} or @option{-std=gnu99}), this |
1227 | switch only affects the @code{asm} and @code{typeof} keywords, since | |
1228 | @code{inline} is a standard keyword in ISO C99. | |
74291a4b MM |
1229 | |
1230 | @item -fno-builtin | |
a3926fe1 | 1231 | @itemx -fno-builtin-@var{function} |
cd3bb277 | 1232 | @opindex fno-builtin |
c771326b JM |
1233 | @cindex built-in functions |
1234 | Don't recognize built-in functions that do not begin with | |
01702459 | 1235 | @samp{__builtin_} as prefix. @xref{Other Builtins,,Other built-in |
f0523f02 | 1236 | functions provided by GCC}, for details of the functions affected, |
c771326b | 1237 | including those which are not built-in functions when @option{-ansi} or |
5490d604 JM |
1238 | @option{-std} options for strict ISO C conformance are used because they |
1239 | do not have an ISO standard meaning. | |
74291a4b | 1240 | |
c771326b | 1241 | GCC normally generates special code to handle certain built-in functions |
74291a4b MM |
1242 | more efficiently; for instance, calls to @code{alloca} may become single |
1243 | instructions that adjust the stack directly, and calls to @code{memcpy} | |
1244 | may become inline copy loops. The resulting code is often both smaller | |
1245 | and faster, but since the function calls no longer appear as such, you | |
1246 | cannot set a breakpoint on those calls, nor can you change the behavior | |
e6e931b7 JM |
1247 | of the functions by linking with a different library. In addition, |
1248 | when a function is recognized as a built-in function, GCC may use | |
1249 | information about that function to warn about problems with calls to | |
1250 | that function, or to generate more efficient code, even if the | |
1251 | resulting code still contains calls to that function. For example, | |
1252 | warnings are given with @option{-Wformat} for bad calls to | |
1253 | @code{printf}, when @code{printf} is built in, and @code{strlen} is | |
1254 | known not to modify global memory. | |
74291a4b | 1255 | |
a3926fe1 RS |
1256 | With the @option{-fno-builtin-@var{function}} option |
1257 | only the built-in function @var{function} is | |
7d14c755 JM |
1258 | disabled. @var{function} must not begin with @samp{__builtin_}. If a |
1259 | function is named this is not built-in in this version of GCC, this | |
1260 | option is ignored. There is no corresponding | |
1261 | @option{-fbuiltin-@var{function}} option; if you wish to enable | |
1262 | built-in functions selectively when using @option{-fno-builtin} or | |
1263 | @option{-ffreestanding}, you may define macros such as: | |
1264 | ||
1265 | @smallexample | |
1266 | #define abs(n) __builtin_abs ((n)) | |
1267 | #define strcpy(d, s) __builtin_strcpy ((d), (s)) | |
1268 | @end smallexample | |
1269 | ||
861bb6c1 | 1270 | @item -fhosted |
cd3bb277 | 1271 | @opindex fhosted |
861bb6c1 JL |
1272 | @cindex hosted environment |
1273 | ||
1274 | Assert that compilation takes place in a hosted environment. This implies | |
630d3d5a | 1275 | @option{-fbuiltin}. A hosted environment is one in which the |
861bb6c1 JL |
1276 | entire standard library is available, and in which @code{main} has a return |
1277 | type of @code{int}. Examples are nearly everything except a kernel. | |
630d3d5a | 1278 | This is equivalent to @option{-fno-freestanding}. |
861bb6c1 JL |
1279 | |
1280 | @item -ffreestanding | |
cd3bb277 | 1281 | @opindex ffreestanding |
861bb6c1 JL |
1282 | @cindex hosted environment |
1283 | ||
1284 | Assert that compilation takes place in a freestanding environment. This | |
630d3d5a | 1285 | implies @option{-fno-builtin}. A freestanding environment |
861bb6c1 JL |
1286 | is one in which the standard library may not exist, and program startup may |
1287 | not necessarily be at @code{main}. The most obvious example is an OS kernel. | |
630d3d5a | 1288 | This is equivalent to @option{-fno-hosted}. |
861bb6c1 | 1289 | |
c1030c7c JM |
1290 | @xref{Standards,,Language Standards Supported by GCC}, for details of |
1291 | freestanding and hosted environments. | |
1292 | ||
750491fc RH |
1293 | @item -fms-extensions |
1294 | @opindex fms-extensions | |
1295 | Accept some non-standard constructs used in Microsoft header files. | |
1296 | ||
2fbebc71 JM |
1297 | Some cases of unnamed fields in structures and unions are only |
1298 | accepted with this option. @xref{Unnamed Fields,,Unnamed struct/union | |
1299 | fields within structs/unions}, for details. | |
1300 | ||
74291a4b | 1301 | @item -trigraphs |
cd3bb277 | 1302 | @opindex trigraphs |
3bce8a01 NB |
1303 | Support ISO C trigraphs. The @option{-ansi} option (and @option{-std} |
1304 | options for strict ISO C conformance) implies @option{-trigraphs}. | |
74291a4b | 1305 | |
8a035a6b AH |
1306 | @item -no-integrated-cpp |
1307 | @opindex no-integrated-cpp | |
1308 | Performs a compilation in two passes: preprocessing and compiling. This | |
1309 | option allows a user supplied "cc1", "cc1plus", or "cc1obj" via the | |
8a36672b | 1310 | @option{-B} option. The user supplied compilation step can then add in |
8a035a6b | 1311 | an additional preprocessing step after normal preprocessing but before |
8a36672b | 1312 | compiling. The default is to use the integrated cpp (internal cpp) |
8a035a6b AH |
1313 | |
1314 | The semantics of this option will change if "cc1", "cc1plus", and | |
1315 | "cc1obj" are merged. | |
1316 | ||
74291a4b MM |
1317 | @cindex traditional C language |
1318 | @cindex C language, traditional | |
1319 | @item -traditional | |
f458d1d5 | 1320 | @itemx -traditional-cpp |
cd3bb277 | 1321 | @opindex traditional-cpp |
f458d1d5 ZW |
1322 | @opindex traditional |
1323 | Formerly, these options caused GCC to attempt to emulate a pre-standard | |
1324 | C compiler. They are now only supported with the @option{-E} switch. | |
1325 | The preprocessor continues to support a pre-standard mode. See the GNU | |
1326 | CPP manual for details. | |
74291a4b MM |
1327 | |
1328 | @item -fcond-mismatch | |
cd3bb277 | 1329 | @opindex fcond-mismatch |
74291a4b | 1330 | Allow conditional expressions with mismatched types in the second and |
a7537031 JM |
1331 | third arguments. The value of such an expression is void. This option |
1332 | is not supported for C++. | |
74291a4b MM |
1333 | |
1334 | @item -funsigned-char | |
cd3bb277 | 1335 | @opindex funsigned-char |
74291a4b MM |
1336 | Let the type @code{char} be unsigned, like @code{unsigned char}. |
1337 | ||
1338 | Each kind of machine has a default for what @code{char} should | |
1339 | be. It is either like @code{unsigned char} by default or like | |
1340 | @code{signed char} by default. | |
1341 | ||
1342 | Ideally, a portable program should always use @code{signed char} or | |
1343 | @code{unsigned char} when it depends on the signedness of an object. | |
1344 | But many programs have been written to use plain @code{char} and | |
1345 | expect it to be signed, or expect it to be unsigned, depending on the | |
1346 | machines they were written for. This option, and its inverse, let you | |
1347 | make such a program work with the opposite default. | |
1348 | ||
1349 | The type @code{char} is always a distinct type from each of | |
1350 | @code{signed char} or @code{unsigned char}, even though its behavior | |
1351 | is always just like one of those two. | |
1352 | ||
1353 | @item -fsigned-char | |
cd3bb277 | 1354 | @opindex fsigned-char |
74291a4b MM |
1355 | Let the type @code{char} be signed, like @code{signed char}. |
1356 | ||
630d3d5a JM |
1357 | Note that this is equivalent to @option{-fno-unsigned-char}, which is |
1358 | the negative form of @option{-funsigned-char}. Likewise, the option | |
1359 | @option{-fno-signed-char} is equivalent to @option{-funsigned-char}. | |
74291a4b | 1360 | |
74291a4b MM |
1361 | @item -fsigned-bitfields |
1362 | @itemx -funsigned-bitfields | |
1363 | @itemx -fno-signed-bitfields | |
1364 | @itemx -fno-unsigned-bitfields | |
cd3bb277 JM |
1365 | @opindex fsigned-bitfields |
1366 | @opindex funsigned-bitfields | |
1367 | @opindex fno-signed-bitfields | |
1368 | @opindex fno-unsigned-bitfields | |
c771326b | 1369 | These options control whether a bit-field is signed or unsigned, when the |
74291a4b | 1370 | declaration does not use either @code{signed} or @code{unsigned}. By |
c771326b | 1371 | default, such a bit-field is signed, because this is consistent: the |
74291a4b | 1372 | basic integer types such as @code{int} are signed types. |
74291a4b MM |
1373 | @end table |
1374 | ||
1375 | @node C++ Dialect Options | |
1376 | @section Options Controlling C++ Dialect | |
1377 | ||
1378 | @cindex compiler options, C++ | |
1379 | @cindex C++ options, command line | |
1380 | @cindex options, C++ | |
1381 | This section describes the command-line options that are only meaningful | |
1382 | for C++ programs; but you can also use most of the GNU compiler options | |
1383 | regardless of what language your program is in. For example, you | |
1384 | might compile a file @code{firstClass.C} like this: | |
1385 | ||
3ab51846 | 1386 | @smallexample |
1dc5fc4b | 1387 | g++ -g -frepo -O -c firstClass.C |
3ab51846 | 1388 | @end smallexample |
74291a4b MM |
1389 | |
1390 | @noindent | |
630d3d5a | 1391 | In this example, only @option{-frepo} is an option meant |
74291a4b | 1392 | only for C++ programs; you can use the other options with any |
161d7b59 | 1393 | language supported by GCC@. |
74291a4b MM |
1394 | |
1395 | Here is a list of options that are @emph{only} for compiling C++ programs: | |
1396 | ||
2642624b | 1397 | @table @gcctabopt |
2d3e278d MM |
1398 | |
1399 | @item -fabi-version=@var{n} | |
1400 | @opindex fabi-version | |
8a36672b | 1401 | Use version @var{n} of the C++ ABI@. Version 2 is the version of the |
57702a80 MM |
1402 | C++ ABI that first appeared in G++ 3.4. Version 1 is the version of |
1403 | the C++ ABI that first appeared in G++ 3.2. Version 0 will always be | |
1404 | the version that conforms most closely to the C++ ABI specification. | |
1405 | Therefore, the ABI obtained using version 0 will change as ABI bugs | |
1406 | are fixed. | |
2d3e278d | 1407 | |
d150ccef | 1408 | The default is version 2. |
46c83bce | 1409 | |
74291a4b | 1410 | @item -fno-access-control |
cd3bb277 | 1411 | @opindex fno-access-control |
74291a4b MM |
1412 | Turn off all access checking. This switch is mainly useful for working |
1413 | around bugs in the access control code. | |
1414 | ||
74291a4b | 1415 | @item -fcheck-new |
cd3bb277 | 1416 | @opindex fcheck-new |
74291a4b | 1417 | Check that the pointer returned by @code{operator new} is non-null |
6d9c4c83 JW |
1418 | before attempting to modify the storage allocated. This check is |
1419 | normally unnecessary because the C++ standard specifies that | |
1420 | @code{operator new} will only return @code{0} if it is declared | |
1421 | @samp{throw()}, in which case the compiler will always check the | |
1422 | return value even without this option. In all other cases, when | |
1423 | @code{operator new} has a non-empty exception specification, memory | |
1424 | exhaustion is signalled by throwing @code{std::bad_alloc}. See also | |
1425 | @samp{new (nothrow)}. | |
1dc5fc4b | 1426 | |
74291a4b | 1427 | @item -fconserve-space |
cd3bb277 | 1428 | @opindex fconserve-space |
74291a4b MM |
1429 | Put uninitialized or runtime-initialized global variables into the |
1430 | common segment, as C does. This saves space in the executable at the | |
1431 | cost of not diagnosing duplicate definitions. If you compile with this | |
1432 | flag and your program mysteriously crashes after @code{main()} has | |
1433 | completed, you may have an object that is being destroyed twice because | |
1434 | two definitions were merged. | |
1435 | ||
1dc5fc4b JM |
1436 | This option is no longer useful on most targets, now that support has |
1437 | been added for putting variables into BSS without making them common. | |
1438 | ||
d63d5d0c ILT |
1439 | @item -ffriend-injection |
1440 | @opindex ffriend-injection | |
1441 | Inject friend functions into the enclosing namespace, so that they are | |
1442 | visible outside the scope of the class in which they are declared. | |
1443 | Friend functions were documented to work this way in the old Annotated | |
1444 | C++ Reference Manual, and versions of G++ before 4.1 always worked | |
1445 | that way. However, in ISO C++ a friend function which is not declared | |
1446 | in an enclosing scope can only be found using argument dependent | |
1447 | lookup. This option causes friends to be injected as they were in | |
1448 | earlier releases. | |
1449 | ||
1450 | This option is for compatibility, and may be removed in a future | |
1451 | release of G++. | |
1452 | ||
02f52e19 | 1453 | @item -fno-const-strings |
cd3bb277 | 1454 | @opindex fno-const-strings |
fcca588c MM |
1455 | Give string constants type @code{char *} instead of type @code{const |
1456 | char *}. By default, G++ uses type @code{const char *} as required by | |
630d3d5a | 1457 | the standard. Even if you use @option{-fno-const-strings}, you cannot |
3521b33c | 1458 | actually modify the value of a string constant. |
fcca588c MM |
1459 | |
1460 | This option might be removed in a future release of G++. For maximum | |
1461 | portability, you should structure your code so that it works with | |
1462 | string constants that have type @code{const char *}. | |
1463 | ||
1dc5fc4b | 1464 | @item -fno-elide-constructors |
cd3bb277 | 1465 | @opindex fno-elide-constructors |
1dc5fc4b JM |
1466 | The C++ standard allows an implementation to omit creating a temporary |
1467 | which is only used to initialize another object of the same type. | |
aee96fe9 | 1468 | Specifying this option disables that optimization, and forces G++ to |
1dc5fc4b | 1469 | call the copy constructor in all cases. |
74291a4b | 1470 | |
dd1ba632 | 1471 | @item -fno-enforce-eh-specs |
cd3bb277 | 1472 | @opindex fno-enforce-eh-specs |
4381020e JM |
1473 | Don't generate code to check for violation of exception specifications |
1474 | at runtime. This option violates the C++ standard, but may be useful | |
1475 | for reducing code size in production builds, much like defining | |
1476 | @samp{NDEBUG}. This does not give user code permission to throw | |
1477 | exceptions in violation of the exception specifications; the compiler | |
1478 | will still optimize based on the specifications, so throwing an | |
1479 | unexpected exception will result in undefined behavior. | |
dd1ba632 | 1480 | |
74291a4b | 1481 | @item -ffor-scope |
8c81598d | 1482 | @itemx -fno-for-scope |
cd3bb277 JM |
1483 | @opindex ffor-scope |
1484 | @opindex fno-for-scope | |
695ac33f | 1485 | If @option{-ffor-scope} is specified, the scope of variables declared in |
74291a4b | 1486 | a @i{for-init-statement} is limited to the @samp{for} loop itself, |
34527c47 | 1487 | as specified by the C++ standard. |
695ac33f | 1488 | If @option{-fno-for-scope} is specified, the scope of variables declared in |
74291a4b | 1489 | a @i{for-init-statement} extends to the end of the enclosing scope, |
aee96fe9 | 1490 | as was the case in old versions of G++, and other (traditional) |
74291a4b MM |
1491 | implementations of C++. |
1492 | ||
1493 | The default if neither flag is given to follow the standard, | |
1494 | but to allow and give a warning for old-style code that would | |
1495 | otherwise be invalid, or have different behavior. | |
1496 | ||
1497 | @item -fno-gnu-keywords | |
cd3bb277 | 1498 | @opindex fno-gnu-keywords |
9762e8a4 | 1499 | Do not recognize @code{typeof} as a keyword, so that code can use this |
767094dd | 1500 | word as an identifier. You can use the keyword @code{__typeof__} instead. |
630d3d5a | 1501 | @option{-ansi} implies @option{-fno-gnu-keywords}. |
74291a4b | 1502 | |
1dc5fc4b | 1503 | @item -fno-implicit-templates |
cd3bb277 | 1504 | @opindex fno-implicit-templates |
bba975d4 | 1505 | Never emit code for non-inline templates which are instantiated |
e979f9e8 | 1506 | implicitly (i.e.@: by use); only emit code for explicit instantiations. |
bba975d4 JM |
1507 | @xref{Template Instantiation}, for more information. |
1508 | ||
1509 | @item -fno-implicit-inline-templates | |
cd3bb277 | 1510 | @opindex fno-implicit-inline-templates |
bba975d4 JM |
1511 | Don't emit code for implicit instantiations of inline templates, either. |
1512 | The default is to handle inlines differently so that compiles with and | |
1513 | without optimization will need the same set of explicit instantiations. | |
1dc5fc4b | 1514 | |
74291a4b | 1515 | @item -fno-implement-inlines |
cd3bb277 | 1516 | @opindex fno-implement-inlines |
74291a4b MM |
1517 | To save space, do not emit out-of-line copies of inline functions |
1518 | controlled by @samp{#pragma implementation}. This will cause linker | |
1519 | errors if these functions are not inlined everywhere they are called. | |
1520 | ||
631cf95d | 1521 | @item -fms-extensions |
cd3bb277 | 1522 | @opindex fms-extensions |
32fb1fb2 PE |
1523 | Disable pedantic warnings about constructs used in MFC, such as implicit |
1524 | int and getting a pointer to member function via non-standard syntax. | |
631cf95d | 1525 | |
fcca588c | 1526 | @item -fno-nonansi-builtins |
cd3bb277 | 1527 | @opindex fno-nonansi-builtins |
c771326b | 1528 | Disable built-in declarations of functions that are not mandated by |
161d7b59 | 1529 | ANSI/ISO C@. These include @code{ffs}, @code{alloca}, @code{_exit}, |
fcca588c MM |
1530 | @code{index}, @code{bzero}, @code{conjf}, and other related functions. |
1531 | ||
775afb25 | 1532 | @item -fno-operator-names |
cd3bb277 | 1533 | @opindex fno-operator-names |
775afb25 | 1534 | Do not treat the operator name keywords @code{and}, @code{bitand}, |
74291a4b | 1535 | @code{bitor}, @code{compl}, @code{not}, @code{or} and @code{xor} as |
775afb25 | 1536 | synonyms as keywords. |
74291a4b | 1537 | |
4f8b4fd9 | 1538 | @item -fno-optional-diags |
cd3bb277 | 1539 | @opindex fno-optional-diags |
4f8b4fd9 | 1540 | Disable diagnostics that the standard says a compiler does not need to |
aee96fe9 | 1541 | issue. Currently, the only such diagnostic issued by G++ is the one for |
bba975d4 | 1542 | a name having multiple meanings within a class. |
4f8b4fd9 | 1543 | |
8c7707b0 | 1544 | @item -fpermissive |
cd3bb277 | 1545 | @opindex fpermissive |
4a386498 MM |
1546 | Downgrade some diagnostics about nonconformant code from errors to |
1547 | warnings. Thus, using @option{-fpermissive} will allow some | |
1548 | nonconforming code to compile. | |
8c7707b0 | 1549 | |
8c81598d | 1550 | @item -frepo |
cd3bb277 | 1551 | @opindex frepo |
9c34dbbf ZW |
1552 | Enable automatic template instantiation at link time. This option also |
1553 | implies @option{-fno-implicit-templates}. @xref{Template | |
1554 | Instantiation}, for more information. | |
8c81598d | 1555 | |
8c7707b0 | 1556 | @item -fno-rtti |
cd3bb277 | 1557 | @opindex fno-rtti |
a7fbfcf9 JM |
1558 | Disable generation of information about every class with virtual |
1559 | functions for use by the C++ runtime type identification features | |
1560 | (@samp{dynamic_cast} and @samp{typeid}). If you don't use those parts | |
1561 | of the language, you can save some space by using this flag. Note that | |
1562 | exception handling uses the same information, but it will generate it as | |
1563 | needed. | |
8c7707b0 | 1564 | |
fcca588c | 1565 | @item -fstats |
cd3bb277 | 1566 | @opindex fstats |
fcca588c MM |
1567 | Emit statistics about front-end processing at the end of the compilation. |
1568 | This information is generally only useful to the G++ development team. | |
1569 | ||
1dc5fc4b | 1570 | @item -ftemplate-depth-@var{n} |
cd3bb277 | 1571 | @opindex ftemplate-depth |
1dc5fc4b JM |
1572 | Set the maximum instantiation depth for template classes to @var{n}. |
1573 | A limit on the template instantiation depth is needed to detect | |
767094dd | 1574 | endless recursions during template class instantiation. ANSI/ISO C++ |
1dc5fc4b JM |
1575 | conforming programs must not rely on a maximum depth greater than 17. |
1576 | ||
40aac948 JM |
1577 | @item -fno-threadsafe-statics |
1578 | @opindex fno-threadsafe-statics | |
1579 | Do not emit the extra code to use the routines specified in the C++ | |
1580 | ABI for thread-safe initialization of local statics. You can use this | |
1581 | option to reduce code size slightly in code that doesn't need to be | |
1582 | thread-safe. | |
1583 | ||
fc693822 | 1584 | @item -fuse-cxa-atexit |
cd3bb277 | 1585 | @opindex fuse-cxa-atexit |
fc693822 MM |
1586 | Register destructors for objects with static storage duration with the |
1587 | @code{__cxa_atexit} function rather than the @code{atexit} function. | |
1588 | This option is required for fully standards-compliant handling of static | |
1589 | destructors, but will only work if your C library supports | |
1590 | @code{__cxa_atexit}. | |
1591 | ||
d7afec4b ND |
1592 | @item -fvisibility-inlines-hidden |
1593 | @opindex fvisibility-inlines-hidden | |
1594 | Causes all inlined methods to be marked with | |
1595 | @code{__attribute__ ((visibility ("hidden")))} so that they do not | |
1596 | appear in the export table of a DSO and do not require a PLT indirection | |
8a36672b | 1597 | when used within the DSO@. Enabling this option can have a dramatic effect |
d7afec4b | 1598 | on load and link times of a DSO as it massively reduces the size of the |
8a36672b | 1599 | dynamic export table when the library makes heavy use of templates. While |
d7afec4b ND |
1600 | it can cause bloating through duplication of code within each DSO where |
1601 | it is used, often the wastage is less than the considerable space occupied | |
1602 | by a long symbol name in the export table which is typical when using | |
8a36672b | 1603 | templates and namespaces. For even more savings, combine with the |
4ec7afd7 | 1604 | @option{-fvisibility=hidden} switch. |
d7afec4b | 1605 | |
02f52e19 | 1606 | @item -fno-weak |
cd3bb277 | 1607 | @opindex fno-weak |
90ecce3e | 1608 | Do not use weak symbol support, even if it is provided by the linker. |
fcca588c MM |
1609 | By default, G++ will use weak symbols if they are available. This |
1610 | option exists only for testing, and should not be used by end-users; | |
1611 | it will result in inferior code and has no benefits. This option may | |
1612 | be removed in a future release of G++. | |
1613 | ||
74291a4b | 1614 | @item -nostdinc++ |
cd3bb277 | 1615 | @opindex nostdinc++ |
74291a4b MM |
1616 | Do not search for header files in the standard directories specific to |
1617 | C++, but do still search the other standard directories. (This option | |
e5e809f4 | 1618 | is used when building the C++ library.) |
74291a4b MM |
1619 | @end table |
1620 | ||
1621 | In addition, these optimization, warning, and code generation options | |
1622 | have meanings only for C++ programs: | |
1623 | ||
2642624b | 1624 | @table @gcctabopt |
74291a4b | 1625 | @item -fno-default-inline |
cd3bb277 | 1626 | @opindex fno-default-inline |
74291a4b | 1627 | Do not assume @samp{inline} for functions defined inside a class scope. |
1dc5fc4b JM |
1628 | @xref{Optimize Options,,Options That Control Optimization}. Note that these |
1629 | functions will have linkage like inline functions; they just won't be | |
1630 | inlined by default. | |
74291a4b | 1631 | |
eca7f13c MM |
1632 | @item -Wabi @r{(C++ only)} |
1633 | @opindex Wabi | |
1634 | Warn when G++ generates code that is probably not compatible with the | |
8a36672b | 1635 | vendor-neutral C++ ABI@. Although an effort has been made to warn about |
daf2f129 | 1636 | all such cases, there are probably some cases that are not warned about, |
eca7f13c MM |
1637 | even though G++ is generating incompatible code. There may also be |
1638 | cases where warnings are emitted even though the code that is generated | |
1639 | will be compatible. | |
1640 | ||
1641 | You should rewrite your code to avoid these warnings if you are | |
1642 | concerned about the fact that code generated by G++ may not be binary | |
1643 | compatible with code generated by other compilers. | |
1644 | ||
3364c33b | 1645 | The known incompatibilities at this point include: |
eca7f13c MM |
1646 | |
1647 | @itemize @bullet | |
1648 | ||
1649 | @item | |
1650 | Incorrect handling of tail-padding for bit-fields. G++ may attempt to | |
1651 | pack data into the same byte as a base class. For example: | |
1652 | ||
1653 | @smallexample | |
1654 | struct A @{ virtual void f(); int f1 : 1; @}; | |
1655 | struct B : public A @{ int f2 : 1; @}; | |
1656 | @end smallexample | |
1657 | ||
1658 | @noindent | |
1659 | In this case, G++ will place @code{B::f2} into the same byte | |
daf2f129 | 1660 | as@code{A::f1}; other compilers will not. You can avoid this problem |
eca7f13c MM |
1661 | by explicitly padding @code{A} so that its size is a multiple of the |
1662 | byte size on your platform; that will cause G++ and other compilers to | |
1663 | layout @code{B} identically. | |
1664 | ||
1665 | @item | |
1666 | Incorrect handling of tail-padding for virtual bases. G++ does not use | |
1667 | tail padding when laying out virtual bases. For example: | |
1668 | ||
1669 | @smallexample | |
1670 | struct A @{ virtual void f(); char c1; @}; | |
1671 | struct B @{ B(); char c2; @}; | |
1672 | struct C : public A, public virtual B @{@}; | |
1673 | @end smallexample | |
1674 | ||
1675 | @noindent | |
1676 | In this case, G++ will not place @code{B} into the tail-padding for | |
1677 | @code{A}; other compilers will. You can avoid this problem by | |
1678 | explicitly padding @code{A} so that its size is a multiple of its | |
1679 | alignment (ignoring virtual base classes); that will cause G++ and other | |
1680 | compilers to layout @code{C} identically. | |
1681 | ||
2d3e278d MM |
1682 | @item |
1683 | Incorrect handling of bit-fields with declared widths greater than that | |
1684 | of their underlying types, when the bit-fields appear in a union. For | |
1685 | example: | |
1686 | ||
1687 | @smallexample | |
1688 | union U @{ int i : 4096; @}; | |
1689 | @end smallexample | |
1690 | ||
1691 | @noindent | |
1692 | Assuming that an @code{int} does not have 4096 bits, G++ will make the | |
1693 | union too small by the number of bits in an @code{int}. | |
1694 | ||
956d9305 MM |
1695 | @item |
1696 | Empty classes can be placed at incorrect offsets. For example: | |
daf2f129 | 1697 | |
956d9305 MM |
1698 | @smallexample |
1699 | struct A @{@}; | |
1700 | ||
1701 | struct B @{ | |
1702 | A a; | |
1703 | virtual void f (); | |
1704 | @}; | |
1705 | ||
1706 | struct C : public B, public A @{@}; | |
1707 | @end smallexample | |
1708 | ||
1709 | @noindent | |
c0478a66 | 1710 | G++ will place the @code{A} base class of @code{C} at a nonzero offset; |
956d9305 MM |
1711 | it should be placed at offset zero. G++ mistakenly believes that the |
1712 | @code{A} data member of @code{B} is already at offset zero. | |
1713 | ||
6397d80b MM |
1714 | @item |
1715 | Names of template functions whose types involve @code{typename} or | |
1716 | template template parameters can be mangled incorrectly. | |
1717 | ||
1718 | @smallexample | |
1719 | template <typename Q> | |
1720 | void f(typename Q::X) @{@} | |
1721 | ||
1722 | template <template <typename> class Q> | |
1723 | void f(typename Q<int>::X) @{@} | |
1724 | @end smallexample | |
1725 | ||
1726 | @noindent | |
3364c33b | 1727 | Instantiations of these templates may be mangled incorrectly. |
6397d80b | 1728 | |
eca7f13c MM |
1729 | @end itemize |
1730 | ||
aee96fe9 | 1731 | @item -Wctor-dtor-privacy @r{(C++ only)} |
cd3bb277 | 1732 | @opindex Wctor-dtor-privacy |
9eff22bc LG |
1733 | Warn when a class seems unusable because all the constructors or |
1734 | destructors in that class are private, and it has neither friends nor | |
78d0a54d | 1735 | public static member functions. |
bba975d4 | 1736 | |
aee96fe9 | 1737 | @item -Wnon-virtual-dtor @r{(C++ only)} |
cd3bb277 | 1738 | @opindex Wnon-virtual-dtor |
9eff22bc | 1739 | Warn when a class appears to be polymorphic, thereby requiring a virtual |
efee9ded GDR |
1740 | destructor, yet it declares a non-virtual one. This warning is also |
1741 | enabled if -Weffc++ is specified. | |
bba975d4 | 1742 | |
aee96fe9 | 1743 | @item -Wreorder @r{(C++ only)} |
cd3bb277 | 1744 | @opindex Wreorder |
bba975d4 JM |
1745 | @cindex reordering, warning |
1746 | @cindex warning for reordering of member initializers | |
1747 | Warn when the order of member initializers given in the code does not | |
1748 | match the order in which they must be executed. For instance: | |
1749 | ||
1750 | @smallexample | |
1751 | struct A @{ | |
1752 | int i; | |
1753 | int j; | |
1754 | A(): j (0), i (1) @{ @} | |
1755 | @}; | |
1756 | @end smallexample | |
1757 | ||
9eff22bc LG |
1758 | The compiler will rearrange the member initializers for @samp{i} |
1759 | and @samp{j} to match the declaration order of the members, emitting | |
1760 | a warning to that effect. This warning is enabled by @option{-Wall}. | |
bba975d4 JM |
1761 | @end table |
1762 | ||
630d3d5a | 1763 | The following @option{-W@dots{}} options are not affected by @option{-Wall}. |
bba975d4 | 1764 | |
2642624b | 1765 | @table @gcctabopt |
aee96fe9 | 1766 | @item -Weffc++ @r{(C++ only)} |
cd3bb277 | 1767 | @opindex Weffc++ |
77f6c1eb RS |
1768 | Warn about violations of the following style guidelines from Scott Meyers' |
1769 | @cite{Effective C++} book: | |
1770 | ||
1771 | @itemize @bullet | |
1772 | @item | |
1773 | Item 11: Define a copy constructor and an assignment operator for classes | |
1774 | with dynamically allocated memory. | |
1775 | ||
1776 | @item | |
1777 | Item 12: Prefer initialization to assignment in constructors. | |
1778 | ||
1779 | @item | |
1780 | Item 14: Make destructors virtual in base classes. | |
1781 | ||
1782 | @item | |
1783 | Item 15: Have @code{operator=} return a reference to @code{*this}. | |
1784 | ||
1785 | @item | |
1786 | Item 23: Don't try to return a reference when you must return an object. | |
1787 | ||
1788 | @end itemize | |
1789 | ||
daf2f129 | 1790 | Also warn about violations of the following style guidelines from |
9eff22bc | 1791 | Scott Meyers' @cite{More Effective C++} book: |
77f6c1eb RS |
1792 | |
1793 | @itemize @bullet | |
1794 | @item | |
1795 | Item 6: Distinguish between prefix and postfix forms of increment and | |
1796 | decrement operators. | |
1797 | ||
1798 | @item | |
1799 | Item 7: Never overload @code{&&}, @code{||}, or @code{,}. | |
1800 | ||
1801 | @end itemize | |
1802 | ||
9eff22bc LG |
1803 | When selecting this option, be aware that the standard library |
1804 | headers do not obey all of these guidelines; use @samp{grep -v} | |
77f6c1eb | 1805 | to filter out those warnings. |
bba975d4 | 1806 | |
aee96fe9 | 1807 | @item -Wno-deprecated @r{(C++ only)} |
cd3bb277 | 1808 | @opindex Wno-deprecated |
767094dd | 1809 | Do not warn about usage of deprecated features. @xref{Deprecated Features}. |
2de45c06 | 1810 | |
b2f97e4a MM |
1811 | @item -Wstrict-null-sentinel @r{(C++ only)} |
1812 | @opindex Wstrict-null-sentinel | |
1813 | Warn also about the use of an uncasted @code{NULL} as sentinel. When | |
1814 | compiling only with GCC this is a valid sentinel, as @code{NULL} is defined | |
1815 | to @code{__null}. Although it is a null pointer constant not a null pointer, | |
1816 | it is guaranteed to of the same size as a pointer. But this use is | |
1817 | not portable across different compilers. | |
1818 | ||
aee96fe9 | 1819 | @item -Wno-non-template-friend @r{(C++ only)} |
cd3bb277 | 1820 | @opindex Wno-non-template-friend |
bba975d4 | 1821 | Disable warnings when non-templatized friend functions are declared |
9eff22bc | 1822 | within a template. Since the advent of explicit template specification |
aee96fe9 | 1823 | support in G++, if the name of the friend is an unqualified-id (i.e., |
bba975d4 | 1824 | @samp{friend foo(int)}), the C++ language specification demands that the |
767094dd | 1825 | friend declare or define an ordinary, nontemplate function. (Section |
aee96fe9 | 1826 | 14.5.3). Before G++ implemented explicit specification, unqualified-ids |
bba975d4 | 1827 | could be interpreted as a particular specialization of a templatized |
767094dd | 1828 | function. Because this non-conforming behavior is no longer the default |
aee96fe9 | 1829 | behavior for G++, @option{-Wnon-template-friend} allows the compiler to |
9eff22bc | 1830 | check existing code for potential trouble spots and is on by default. |
2228d450 | 1831 | This new compiler behavior can be turned off with |
630d3d5a | 1832 | @option{-Wno-non-template-friend} which keeps the conformant compiler code |
2228d450 | 1833 | but disables the helpful warning. |
bba975d4 | 1834 | |
aee96fe9 | 1835 | @item -Wold-style-cast @r{(C++ only)} |
cd3bb277 | 1836 | @opindex Wold-style-cast |
323728aa | 1837 | Warn if an old-style (C-style) cast to a non-void type is used within |
7cd5c075 GP |
1838 | a C++ program. The new-style casts (@samp{dynamic_cast}, |
1839 | @samp{static_cast}, @samp{reinterpret_cast}, and @samp{const_cast}) are | |
1840 | less vulnerable to unintended effects and much easier to search for. | |
bba975d4 | 1841 | |
aee96fe9 | 1842 | @item -Woverloaded-virtual @r{(C++ only)} |
cd3bb277 | 1843 | @opindex Woverloaded-virtual |
bba975d4 JM |
1844 | @cindex overloaded virtual fn, warning |
1845 | @cindex warning for overloaded virtual fn | |
3747f3dc MM |
1846 | Warn when a function declaration hides virtual functions from a |
1847 | base class. For example, in: | |
1848 | ||
1849 | @smallexample | |
1850 | struct A @{ | |
1851 | virtual void f(); | |
1852 | @}; | |
1853 | ||
1854 | struct B: public A @{ | |
1855 | void f(int); | |
1856 | @}; | |
1857 | @end smallexample | |
1858 | ||
1859 | the @code{A} class version of @code{f} is hidden in @code{B}, and code | |
9eff22bc | 1860 | like: |
3747f3dc MM |
1861 | |
1862 | @smallexample | |
1863 | B* b; | |
1864 | b->f(); | |
1865 | @end smallexample | |
1866 | ||
1867 | will fail to compile. | |
bba975d4 | 1868 | |
aee96fe9 | 1869 | @item -Wno-pmf-conversions @r{(C++ only)} |
cd3bb277 | 1870 | @opindex Wno-pmf-conversions |
bba975d4 JM |
1871 | Disable the diagnostic for converting a bound pointer to member function |
1872 | to a plain pointer. | |
1873 | ||
aee96fe9 | 1874 | @item -Wsign-promo @r{(C++ only)} |
cd3bb277 | 1875 | @opindex Wsign-promo |
bba975d4 | 1876 | Warn when overload resolution chooses a promotion from unsigned or |
2eac577f | 1877 | enumerated type to a signed type, over a conversion to an unsigned type of |
aee96fe9 | 1878 | the same size. Previous versions of G++ would try to preserve |
bba975d4 JM |
1879 | unsignedness, but the standard mandates the current behavior. |
1880 | ||
bba975d4 JM |
1881 | @smallexample |
1882 | struct A @{ | |
1883 | operator int (); | |
1884 | A& operator = (int); | |
1885 | @}; | |
1886 | ||
1887 | main () | |
1888 | @{ | |
1889 | A a,b; | |
1890 | a = b; | |
1891 | @} | |
1892 | @end smallexample | |
74291a4b | 1893 | |
aee96fe9 | 1894 | In this example, G++ will synthesize a default @samp{A& operator = |
bba975d4 | 1895 | (const A&);}, while cfront will use the user-defined @samp{operator =}. |
74291a4b MM |
1896 | @end table |
1897 | ||
46e34f96 ZL |
1898 | @node Objective-C and Objective-C++ Dialect Options |
1899 | @section Options Controlling Objective-C and Objective-C++ Dialects | |
60de6385 | 1900 | |
46e34f96 ZL |
1901 | @cindex compiler options, Objective-C and Objective-C++ |
1902 | @cindex Objective-C and Objective-C++ options, command line | |
1903 | @cindex options, Objective-C and Objective-C++ | |
1904 | (NOTE: This manual does not describe the Objective-C and Objective-C++ | |
1905 | languages themselves. See @xref{Standards,,Language Standards | |
1906 | Supported by GCC}, for references.) | |
264fa2db | 1907 | |
60de6385 | 1908 | This section describes the command-line options that are only meaningful |
46e34f96 ZL |
1909 | for Objective-C and Objective-C++ programs, but you can also use most of |
1910 | the language-independent GNU compiler options. | |
1911 | For example, you might compile a file @code{some_class.m} like this: | |
60de6385 | 1912 | |
3ab51846 | 1913 | @smallexample |
60de6385 | 1914 | gcc -g -fgnu-runtime -O -c some_class.m |
3ab51846 | 1915 | @end smallexample |
60de6385 SS |
1916 | |
1917 | @noindent | |
9eff22bc | 1918 | In this example, @option{-fgnu-runtime} is an option meant only for |
46e34f96 ZL |
1919 | Objective-C and Objective-C++ programs; you can use the other options with |
1920 | any language supported by GCC@. | |
1921 | ||
1922 | Note that since Objective-C is an extension of the C language, Objective-C | |
f0eb93a8 | 1923 | compilations may also use options specific to the C front-end (e.g., |
46e34f96 ZL |
1924 | @option{-Wtraditional}). Similarly, Objective-C++ compilations may use |
1925 | C++-specific options (e.g., @option{-Wabi}). | |
60de6385 SS |
1926 | |
1927 | Here is a list of options that are @emph{only} for compiling Objective-C | |
46e34f96 | 1928 | and Objective-C++ programs: |
60de6385 SS |
1929 | |
1930 | @table @gcctabopt | |
630d3d5a | 1931 | @item -fconstant-string-class=@var{class-name} |
cd3bb277 | 1932 | @opindex fconstant-string-class |
630d3d5a | 1933 | Use @var{class-name} as the name of the class to instantiate for each |
695ac33f | 1934 | literal string specified with the syntax @code{@@"@dots{}"}. The default |
264fa2db ZL |
1935 | class name is @code{NXConstantString} if the GNU runtime is being used, and |
1936 | @code{NSConstantString} if the NeXT runtime is being used (see below). The | |
1937 | @option{-fconstant-cfstrings} option, if also present, will override the | |
1938 | @option{-fconstant-string-class} setting and cause @code{@@"@dots{}"} literals | |
1939 | to be laid out as constant CoreFoundation strings. | |
60de6385 SS |
1940 | |
1941 | @item -fgnu-runtime | |
cd3bb277 | 1942 | @opindex fgnu-runtime |
60de6385 SS |
1943 | Generate object code compatible with the standard GNU Objective-C |
1944 | runtime. This is the default for most types of systems. | |
1945 | ||
1946 | @item -fnext-runtime | |
cd3bb277 | 1947 | @opindex fnext-runtime |
60de6385 | 1948 | Generate output compatible with the NeXT runtime. This is the default |
1f676100 NP |
1949 | for NeXT-based systems, including Darwin and Mac OS X@. The macro |
1950 | @code{__NEXT_RUNTIME__} is predefined if (and only if) this option is | |
1951 | used. | |
60de6385 | 1952 | |
264fa2db | 1953 | @item -fno-nil-receivers |
5ad7ae7f | 1954 | @opindex fno-nil-receivers |
daf2f129 JM |
1955 | Assume that all Objective-C message dispatches (e.g., |
1956 | @code{[receiver message:arg]}) in this translation unit ensure that the receiver | |
46e34f96 ZL |
1957 | is not @code{nil}. This allows for more efficient entry points in the runtime |
1958 | to be used. Currently, this option is only available in conjunction with | |
264fa2db ZL |
1959 | the NeXT runtime on Mac OS X 10.3 and later. |
1960 | ||
6e955430 ZL |
1961 | @item -fobjc-call-cxx-cdtors |
1962 | @opindex fobjc-call-cxx-cdtors | |
1963 | For each Objective-C class, check if any of its instance variables is a | |
1964 | C++ object with a non-trivial default constructor. If so, synthesize a | |
1965 | special @code{- (id) .cxx_construct} instance method that will run | |
1966 | non-trivial default constructors on any such instance variables, in order, | |
1967 | and then return @code{self}. Similarly, check if any instance variable | |
1968 | is a C++ object with a non-trivial destructor, and if so, synthesize a | |
1969 | special @code{- (void) .cxx_destruct} method that will run | |
1970 | all such default destructors, in reverse order. | |
1971 | ||
1972 | The @code{- (id) .cxx_construct} and/or @code{- (void) .cxx_destruct} methods | |
1973 | thusly generated will only operate on instance variables declared in the | |
1974 | current Objective-C class, and not those inherited from superclasses. It | |
1975 | is the responsibility of the Objective-C runtime to invoke all such methods | |
1976 | in an object's inheritance hierarchy. The @code{- (id) .cxx_construct} methods | |
1977 | will be invoked by the runtime immediately after a new object | |
1978 | instance is allocated; the @code{- (void) .cxx_destruct} methods will | |
1979 | be invoked immediately before the runtime deallocates an object instance. | |
1980 | ||
1981 | As of this writing, only the NeXT runtime on Mac OS X 10.4 and later has | |
1982 | support for invoking the @code{- (id) .cxx_construct} and | |
1983 | @code{- (void) .cxx_destruct} methods. | |
1984 | ||
1985 | @item -fobjc-direct-dispatch | |
1986 | @opindex fobjc-direct-dispatch | |
1987 | Allow fast jumps to the message dispatcher. On Darwin this is | |
128a79fb | 1988 | accomplished via the comm page. |
6e955430 | 1989 | |
264fa2db | 1990 | @item -fobjc-exceptions |
5ad7ae7f | 1991 | @opindex fobjc-exceptions |
daf2f129 | 1992 | Enable syntactic support for structured exception handling in Objective-C, |
3dd9b65f MS |
1993 | similar to what is offered by C++ and Java. This option is |
1994 | unavailable in conjunction with the NeXT runtime on Mac OS X 10.2 and | |
1995 | earlier. | |
264fa2db ZL |
1996 | |
1997 | @smallexample | |
1998 | @@try @{ | |
1999 | @dots{} | |
2000 | @@throw expr; | |
2001 | @dots{} | |
2002 | @} | |
2003 | @@catch (AnObjCClass *exc) @{ | |
2004 | @dots{} | |
2005 | @@throw expr; | |
2006 | @dots{} | |
2007 | @@throw; | |
2008 | @dots{} | |
2009 | @} | |
2010 | @@catch (AnotherClass *exc) @{ | |
2011 | @dots{} | |
2012 | @} | |
2013 | @@catch (id allOthers) @{ | |
2014 | @dots{} | |
2015 | @} | |
2016 | @@finally @{ | |
2017 | @dots{} | |
2018 | @@throw expr; | |
2019 | @dots{} | |
2020 | @} | |
2021 | @end smallexample | |
2022 | ||
2023 | The @code{@@throw} statement may appear anywhere in an Objective-C or | |
daf2f129 JM |
2024 | Objective-C++ program; when used inside of a @code{@@catch} block, the |
2025 | @code{@@throw} may appear without an argument (as shown above), in which case | |
264fa2db ZL |
2026 | the object caught by the @code{@@catch} will be rethrown. |
2027 | ||
2028 | Note that only (pointers to) Objective-C objects may be thrown and | |
2029 | caught using this scheme. When an object is thrown, it will be caught | |
2030 | by the nearest @code{@@catch} clause capable of handling objects of that type, | |
daf2f129 JM |
2031 | analogously to how @code{catch} blocks work in C++ and Java. A |
2032 | @code{@@catch(id @dots{})} clause (as shown above) may also be provided to catch | |
264fa2db ZL |
2033 | any and all Objective-C exceptions not caught by previous @code{@@catch} |
2034 | clauses (if any). | |
2035 | ||
2036 | The @code{@@finally} clause, if present, will be executed upon exit from the | |
2037 | immediately preceding @code{@@try @dots{} @@catch} section. This will happen | |
2038 | regardless of whether any exceptions are thrown, caught or rethrown | |
2039 | inside the @code{@@try @dots{} @@catch} section, analogously to the behavior | |
2040 | of the @code{finally} clause in Java. | |
2041 | ||
2042 | There are several caveats to using the new exception mechanism: | |
2043 | ||
2044 | @itemize @bullet | |
2045 | @item | |
daf2f129 | 2046 | Although currently designed to be binary compatible with @code{NS_HANDLER}-style |
264fa2db ZL |
2047 | idioms provided by the @code{NSException} class, the new |
2048 | exceptions can only be used on Mac OS X 10.3 (Panther) and later | |
2049 | systems, due to additional functionality needed in the (NeXT) Objective-C | |
2050 | runtime. | |
2051 | ||
2052 | @item | |
2053 | As mentioned above, the new exceptions do not support handling | |
daf2f129 | 2054 | types other than Objective-C objects. Furthermore, when used from |
264fa2db ZL |
2055 | Objective-C++, the Objective-C exception model does not interoperate with C++ |
2056 | exceptions at this time. This means you cannot @code{@@throw} an exception | |
daf2f129 | 2057 | from Objective-C and @code{catch} it in C++, or vice versa |
264fa2db ZL |
2058 | (i.e., @code{throw @dots{} @@catch}). |
2059 | @end itemize | |
daf2f129 | 2060 | |
264fa2db ZL |
2061 | The @option{-fobjc-exceptions} switch also enables the use of synchronization |
2062 | blocks for thread-safe execution: | |
2063 | ||
2064 | @smallexample | |
2065 | @@synchronized (ObjCClass *guard) @{ | |
2066 | @dots{} | |
2067 | @} | |
2068 | @end smallexample | |
2069 | ||
2070 | Upon entering the @code{@@synchronized} block, a thread of execution shall | |
2071 | first check whether a lock has been placed on the corresponding @code{guard} | |
2072 | object by another thread. If it has, the current thread shall wait until | |
daf2f129 | 2073 | the other thread relinquishes its lock. Once @code{guard} becomes available, |
264fa2db ZL |
2074 | the current thread will place its own lock on it, execute the code contained in |
2075 | the @code{@@synchronized} block, and finally relinquish the lock (thereby | |
2076 | making @code{guard} available to other threads). | |
2077 | ||
2078 | Unlike Java, Objective-C does not allow for entire methods to be marked | |
2079 | @code{@@synchronized}. Note that throwing exceptions out of | |
2080 | @code{@@synchronized} blocks is allowed, and will cause the guarding object | |
2081 | to be unlocked properly. | |
2082 | ||
6e955430 ZL |
2083 | @item -fobjc-gc |
2084 | @opindex fobjc-gc | |
2085 | Enable garbage collection (GC) in Objective-C and Objective-C++ programs. | |
2086 | ||
264fa2db | 2087 | @item -freplace-objc-classes |
5ad7ae7f | 2088 | @opindex freplace-objc-classes |
264fa2db ZL |
2089 | Emit a special marker instructing @command{ld(1)} not to statically link in |
2090 | the resulting object file, and allow @command{dyld(1)} to load it in at | |
2091 | run time instead. This is used in conjunction with the Fix-and-Continue | |
daf2f129 | 2092 | debugging mode, where the object file in question may be recompiled and |
264fa2db ZL |
2093 | dynamically reloaded in the course of program execution, without the need |
2094 | to restart the program itself. Currently, Fix-and-Continue functionality | |
daf2f129 | 2095 | is only available in conjunction with the NeXT runtime on Mac OS X 10.3 |
264fa2db ZL |
2096 | and later. |
2097 | ||
2098 | @item -fzero-link | |
5ad7ae7f | 2099 | @opindex fzero-link |
264fa2db ZL |
2100 | When compiling for the NeXT runtime, the compiler ordinarily replaces calls |
2101 | to @code{objc_getClass("@dots{}")} (when the name of the class is known at | |
2102 | compile time) with static class references that get initialized at load time, | |
2103 | which improves run-time performance. Specifying the @option{-fzero-link} flag | |
2104 | suppresses this behavior and causes calls to @code{objc_getClass("@dots{}")} | |
daf2f129 | 2105 | to be retained. This is useful in Zero-Link debugging mode, since it allows |
264fa2db ZL |
2106 | for individual class implementations to be modified during program execution. |
2107 | ||
60de6385 | 2108 | @item -gen-decls |
cd3bb277 | 2109 | @opindex gen-decls |
60de6385 SS |
2110 | Dump interface declarations for all classes seen in the source file to a |
2111 | file named @file{@var{sourcename}.decl}. | |
2112 | ||
6e955430 ZL |
2113 | @item -Wassign-intercept |
2114 | @opindex Wassign-intercept | |
2115 | Warn whenever an Objective-C assignment is being intercepted by the | |
2116 | garbage collector. | |
2117 | ||
60de6385 | 2118 | @item -Wno-protocol |
cd3bb277 | 2119 | @opindex Wno-protocol |
1f676100 NP |
2120 | If a class is declared to implement a protocol, a warning is issued for |
2121 | every method in the protocol that is not implemented by the class. The | |
6335b0aa | 2122 | default behavior is to issue a warning for every method not explicitly |
1f676100 | 2123 | implemented in the class, even if a method implementation is inherited |
4ec7afd7 | 2124 | from the superclass. If you use the @option{-Wno-protocol} option, then |
1f676100 NP |
2125 | methods inherited from the superclass are considered to be implemented, |
2126 | and no warning is issued for them. | |
60de6385 SS |
2127 | |
2128 | @item -Wselector | |
cd3bb277 | 2129 | @opindex Wselector |
1f676100 NP |
2130 | Warn if multiple methods of different types for the same selector are |
2131 | found during compilation. The check is performed on the list of methods | |
2132 | in the final stage of compilation. Additionally, a check is performed | |
9eff22bc LG |
2133 | for each selector appearing in a @code{@@selector(@dots{})} |
2134 | expression, and a corresponding method for that selector has been found | |
1f676100 NP |
2135 | during compilation. Because these checks scan the method table only at |
2136 | the end of compilation, these warnings are not produced if the final | |
2137 | stage of compilation is not reached, for example because an error is | |
4ec7afd7 | 2138 | found during compilation, or because the @option{-fsyntax-only} option is |
1f676100 NP |
2139 | being used. |
2140 | ||
6e955430 ZL |
2141 | @item -Wstrict-selector-match |
2142 | @opindex Wstrict-selector-match | |
2143 | Warn if multiple methods with differing argument and/or return types are | |
2144 | found for a given selector when attempting to send a message using this | |
2145 | selector to a receiver of type @code{id} or @code{Class}. When this flag | |
2146 | is off (which is the default behavior), the compiler will omit such warnings | |
2147 | if any differences found are confined to types which share the same size | |
2148 | and alignment. | |
2149 | ||
1f676100 NP |
2150 | @item -Wundeclared-selector |
2151 | @opindex Wundeclared-selector | |
2152 | Warn if a @code{@@selector(@dots{})} expression referring to an | |
2153 | undeclared selector is found. A selector is considered undeclared if no | |
daf2f129 | 2154 | method with that name has been declared before the |
9eff22bc LG |
2155 | @code{@@selector(@dots{})} expression, either explicitly in an |
2156 | @code{@@interface} or @code{@@protocol} declaration, or implicitly in | |
2157 | an @code{@@implementation} section. This option always performs its | |
2158 | checks as soon as a @code{@@selector(@dots{})} expression is found, | |
4ec7afd7 | 2159 | while @option{-Wselector} only performs its checks in the final stage of |
9eff22bc | 2160 | compilation. This also enforces the coding style convention |
1f676100 | 2161 | that methods and selectors must be declared before being used. |
60de6385 | 2162 | |
7989e4dc | 2163 | @item -print-objc-runtime-info |
5ad7ae7f | 2164 | @opindex print-objc-runtime-info |
7989e4dc RO |
2165 | Generate C header describing the largest structure that is passed by |
2166 | value, if any. | |
60de6385 SS |
2167 | |
2168 | @end table | |
2169 | ||
764dbbf2 GDR |
2170 | @node Language Independent Options |
2171 | @section Options to Control Diagnostic Messages Formatting | |
2172 | @cindex options to control diagnostics formatting | |
2173 | @cindex diagnostic messages | |
2174 | @cindex message formatting | |
2175 | ||
b192711e | 2176 | Traditionally, diagnostic messages have been formatted irrespective of |
e979f9e8 | 2177 | the output device's aspect (e.g.@: its width, @dots{}). The options described |
764dbbf2 | 2178 | below can be used to control the diagnostic messages formatting |
e979f9e8 | 2179 | algorithm, e.g.@: how many characters per line, how often source location |
6c0a4eab | 2180 | information should be reported. Right now, only the C++ front end can |
764dbbf2 | 2181 | honor these options. However it is expected, in the near future, that |
6c0a4eab | 2182 | the remaining front ends would be able to digest them correctly. |
764dbbf2 | 2183 | |
2642624b | 2184 | @table @gcctabopt |
764dbbf2 | 2185 | @item -fmessage-length=@var{n} |
cd3bb277 | 2186 | @opindex fmessage-length |
764dbbf2 | 2187 | Try to format error messages so that they fit on lines of about @var{n} |
aee96fe9 | 2188 | characters. The default is 72 characters for @command{g++} and 0 for the rest of |
161d7b59 | 2189 | the front ends supported by GCC@. If @var{n} is zero, then no |
02f52e19 | 2190 | line-wrapping will be done; each error message will appear on a single |
764dbbf2 GDR |
2191 | line. |
2192 | ||
cd3bb277 | 2193 | @opindex fdiagnostics-show-location |
764dbbf2 | 2194 | @item -fdiagnostics-show-location=once |
b192711e | 2195 | Only meaningful in line-wrapping mode. Instructs the diagnostic messages |
764dbbf2 GDR |
2196 | reporter to emit @emph{once} source location information; that is, in |
2197 | case the message is too long to fit on a single physical line and has to | |
2198 | be wrapped, the source location won't be emitted (as prefix) again, | |
2199 | over and over, in subsequent continuation lines. This is the default | |
c21cd8b1 | 2200 | behavior. |
764dbbf2 GDR |
2201 | |
2202 | @item -fdiagnostics-show-location=every-line | |
2203 | Only meaningful in line-wrapping mode. Instructs the diagnostic | |
2204 | messages reporter to emit the same source location information (as | |
4fe9b91c | 2205 | prefix) for physical lines that result from the process of breaking |
b192711e | 2206 | a message which is too long to fit on a single line. |
764dbbf2 | 2207 | |
ccf08a6e DD |
2208 | @item -fdiagnostics-show-options |
2209 | @opindex fdiagnostics-show-options | |
2210 | This option instructs the diagnostic machinery to add text to each | |
2211 | diagnostic emitted, which indicates which command line option directly | |
2212 | controls that diagnostic, when such an option is known to the | |
2213 | diagnostic machinery. | |
2214 | ||
764dbbf2 GDR |
2215 | @end table |
2216 | ||
74291a4b MM |
2217 | @node Warning Options |
2218 | @section Options to Request or Suppress Warnings | |
2219 | @cindex options to control warnings | |
2220 | @cindex warning messages | |
2221 | @cindex messages, warning | |
2222 | @cindex suppressing warnings | |
2223 | ||
2224 | Warnings are diagnostic messages that report constructions which | |
2225 | are not inherently erroneous but which are risky or suggest there | |
2226 | may have been an error. | |
2227 | ||
2228 | You can request many specific warnings with options beginning @samp{-W}, | |
630d3d5a | 2229 | for example @option{-Wimplicit} to request warnings on implicit |
74291a4b MM |
2230 | declarations. Each of these specific warning options also has a |
2231 | negative form beginning @samp{-Wno-} to turn off warnings; | |
630d3d5a | 2232 | for example, @option{-Wno-implicit}. This manual lists only one of the |
74291a4b MM |
2233 | two forms, whichever is not the default. |
2234 | ||
62aaa62c GP |
2235 | The following options control the amount and kinds of warnings produced |
2236 | by GCC; for further, language-specific options also refer to | |
46e34f96 ZL |
2237 | @ref{C++ Dialect Options} and @ref{Objective-C and Objective-C++ Dialect |
2238 | Options}. | |
74291a4b | 2239 | |
2642624b | 2240 | @table @gcctabopt |
74291a4b MM |
2241 | @cindex syntax checking |
2242 | @item -fsyntax-only | |
cd3bb277 | 2243 | @opindex fsyntax-only |
74291a4b MM |
2244 | Check the code for syntax errors, but don't do anything beyond that. |
2245 | ||
2246 | @item -pedantic | |
cd3bb277 | 2247 | @opindex pedantic |
074e95e3 JM |
2248 | Issue all the warnings demanded by strict ISO C and ISO C++; |
2249 | reject all programs that use forbidden extensions, and some other | |
2250 | programs that do not follow ISO C and ISO C++. For ISO C, follows the | |
630d3d5a | 2251 | version of the ISO C standard specified by any @option{-std} option used. |
74291a4b | 2252 | |
074e95e3 | 2253 | Valid ISO C and ISO C++ programs should compile properly with or without |
5490d604 | 2254 | this option (though a rare few will require @option{-ansi} or a |
161d7b59 | 2255 | @option{-std} option specifying the required version of ISO C)@. However, |
b1d16193 JL |
2256 | without this option, certain GNU extensions and traditional C and C++ |
2257 | features are supported as well. With this option, they are rejected. | |
74291a4b | 2258 | |
630d3d5a | 2259 | @option{-pedantic} does not cause warning messages for use of the |
74291a4b MM |
2260 | alternate keywords whose names begin and end with @samp{__}. Pedantic |
2261 | warnings are also disabled in the expression that follows | |
2262 | @code{__extension__}. However, only system header files should use | |
2263 | these escape routes; application programs should avoid them. | |
2264 | @xref{Alternate Keywords}. | |
2265 | ||
630d3d5a | 2266 | Some users try to use @option{-pedantic} to check programs for strict ISO |
74291a4b | 2267 | C conformance. They soon find that it does not do quite what they want: |
c1030c7c | 2268 | it finds some non-ISO practices, but not all---only those for which |
074e95e3 JM |
2269 | ISO C @emph{requires} a diagnostic, and some others for which |
2270 | diagnostics have been added. | |
74291a4b | 2271 | |
074e95e3 | 2272 | A feature to report any failure to conform to ISO C might be useful in |
74291a4b | 2273 | some instances, but would require considerable additional work and would |
630d3d5a | 2274 | be quite different from @option{-pedantic}. We don't have plans to |
892d0a6d | 2275 | support such a feature in the near future. |
74291a4b | 2276 | |
91ea548a JM |
2277 | Where the standard specified with @option{-std} represents a GNU |
2278 | extended dialect of C, such as @samp{gnu89} or @samp{gnu99}, there is a | |
2279 | corresponding @dfn{base standard}, the version of ISO C on which the GNU | |
2280 | extended dialect is based. Warnings from @option{-pedantic} are given | |
2281 | where they are required by the base standard. (It would not make sense | |
2282 | for such warnings to be given only for features not in the specified GNU | |
2283 | C dialect, since by definition the GNU dialects of C include all | |
2284 | features the compiler supports with the given option, and there would be | |
2285 | nothing to warn about.) | |
2286 | ||
74291a4b | 2287 | @item -pedantic-errors |
cd3bb277 | 2288 | @opindex pedantic-errors |
630d3d5a | 2289 | Like @option{-pedantic}, except that errors are produced rather than |
74291a4b MM |
2290 | warnings. |
2291 | ||
2292 | @item -w | |
cd3bb277 | 2293 | @opindex w |
74291a4b MM |
2294 | Inhibit all warning messages. |
2295 | ||
2296 | @item -Wno-import | |
cd3bb277 | 2297 | @opindex Wno-import |
74291a4b MM |
2298 | Inhibit warning messages about the use of @samp{#import}. |
2299 | ||
2300 | @item -Wchar-subscripts | |
cd3bb277 | 2301 | @opindex Wchar-subscripts |
74291a4b MM |
2302 | Warn if an array subscript has type @code{char}. This is a common cause |
2303 | of error, as programmers often forget that this type is signed on some | |
2304 | machines. | |
69cdf050 | 2305 | This warning is enabled by @option{-Wall}. |
74291a4b MM |
2306 | |
2307 | @item -Wcomment | |
cd3bb277 | 2308 | @opindex Wcomment |
74291a4b MM |
2309 | Warn whenever a comment-start sequence @samp{/*} appears in a @samp{/*} |
2310 | comment, or whenever a Backslash-Newline appears in a @samp{//} comment. | |
69cdf050 | 2311 | This warning is enabled by @option{-Wall}. |
74291a4b | 2312 | |
c65a01af RG |
2313 | @item -Wfatal-errors |
2314 | @opindex Wfatal-errors | |
2315 | This option causes the compiler to abort compilation on the first error | |
2316 | occurred rather than trying to keep going and printing further error | |
2317 | messages. | |
2318 | ||
74291a4b | 2319 | @item -Wformat |
cd3bb277 | 2320 | @opindex Wformat |
e6e931b7 JM |
2321 | @opindex ffreestanding |
2322 | @opindex fno-builtin | |
74291a4b MM |
2323 | Check calls to @code{printf} and @code{scanf}, etc., to make sure that |
2324 | the arguments supplied have types appropriate to the format string | |
26f6672d JM |
2325 | specified, and that the conversions specified in the format string make |
2326 | sense. This includes standard functions, and others specified by format | |
2327 | attributes (@pxref{Function Attributes}), in the @code{printf}, | |
2328 | @code{scanf}, @code{strftime} and @code{strfmon} (an X/Open extension, | |
a2bec818 | 2329 | not in the C standard) families (or other target-specific families). |
e6e931b7 JM |
2330 | Which functions are checked without format attributes having been |
2331 | specified depends on the standard version selected, and such checks of | |
2332 | functions without the attribute specified are disabled by | |
2333 | @option{-ffreestanding} or @option{-fno-builtin}. | |
74291a4b | 2334 | |
8308e0b7 | 2335 | The formats are checked against the format features supported by GNU |
3764f879 | 2336 | libc version 2.2. These include all ISO C90 and C99 features, as well |
8308e0b7 JM |
2337 | as features from the Single Unix Specification and some BSD and GNU |
2338 | extensions. Other library implementations may not support all these | |
2339 | features; GCC does not support warning about features that go beyond a | |
630d3d5a JM |
2340 | particular library's limitations. However, if @option{-pedantic} is used |
2341 | with @option{-Wformat}, warnings will be given about format features not | |
26f6672d JM |
2342 | in the selected standard version (but not for @code{strfmon} formats, |
2343 | since those are not in any version of the C standard). @xref{C Dialect | |
2344 | Options,,Options Controlling C Dialect}. | |
8308e0b7 | 2345 | |
b34c7881 JT |
2346 | Since @option{-Wformat} also checks for null format arguments for |
2347 | several functions, @option{-Wformat} also implies @option{-Wnonnull}. | |
2348 | ||
630d3d5a | 2349 | @option{-Wformat} is included in @option{-Wall}. For more control over some |
c76f4e8e | 2350 | aspects of format checking, the options @option{-Wformat-y2k}, |
e964a556 JT |
2351 | @option{-Wno-format-extra-args}, @option{-Wno-format-zero-length}, |
2352 | @option{-Wformat-nonliteral}, @option{-Wformat-security}, and | |
2353 | @option{-Wformat=2} are available, but are not included in @option{-Wall}. | |
4d808927 | 2354 | |
c76f4e8e JM |
2355 | @item -Wformat-y2k |
2356 | @opindex Wformat-y2k | |
2357 | If @option{-Wformat} is specified, also warn about @code{strftime} | |
4d808927 JM |
2358 | formats which may yield only a two-digit year. |
2359 | ||
2360 | @item -Wno-format-extra-args | |
cd3bb277 | 2361 | @opindex Wno-format-extra-args |
630d3d5a | 2362 | If @option{-Wformat} is specified, do not warn about excess arguments to a |
4d808927 JM |
2363 | @code{printf} or @code{scanf} format function. The C standard specifies |
2364 | that such arguments are ignored. | |
2365 | ||
7e5fb12f JM |
2366 | Where the unused arguments lie between used arguments that are |
2367 | specified with @samp{$} operand number specifications, normally | |
2368 | warnings are still given, since the implementation could not know what | |
2369 | type to pass to @code{va_arg} to skip the unused arguments. However, | |
2370 | in the case of @code{scanf} formats, this option will suppress the | |
2371 | warning if the unused arguments are all pointers, since the Single | |
2372 | Unix Specification says that such unused arguments are allowed. | |
2373 | ||
e964a556 JT |
2374 | @item -Wno-format-zero-length |
2375 | @opindex Wno-format-zero-length | |
2376 | If @option{-Wformat} is specified, do not warn about zero-length formats. | |
2377 | The C standard specifies that zero-length formats are allowed. | |
2378 | ||
4d808927 | 2379 | @item -Wformat-nonliteral |
cd3bb277 | 2380 | @opindex Wformat-nonliteral |
630d3d5a | 2381 | If @option{-Wformat} is specified, also warn if the format string is not a |
4d808927 JM |
2382 | string literal and so cannot be checked, unless the format function |
2383 | takes its format arguments as a @code{va_list}. | |
2384 | ||
c907e684 | 2385 | @item -Wformat-security |
cd3bb277 | 2386 | @opindex Wformat-security |
630d3d5a | 2387 | If @option{-Wformat} is specified, also warn about uses of format |
c907e684 JM |
2388 | functions that represent possible security problems. At present, this |
2389 | warns about calls to @code{printf} and @code{scanf} functions where the | |
2390 | format string is not a string literal and there are no format arguments, | |
2391 | as in @code{printf (foo);}. This may be a security hole if the format | |
2392 | string came from untrusted input and contains @samp{%n}. (This is | |
630d3d5a JM |
2393 | currently a subset of what @option{-Wformat-nonliteral} warns about, but |
2394 | in future warnings may be added to @option{-Wformat-security} that are not | |
2395 | included in @option{-Wformat-nonliteral}.) | |
c907e684 | 2396 | |
4d808927 | 2397 | @item -Wformat=2 |
cd3bb277 | 2398 | @opindex Wformat=2 |
630d3d5a JM |
2399 | Enable @option{-Wformat} plus format checks not included in |
2400 | @option{-Wformat}. Currently equivalent to @samp{-Wformat | |
c76f4e8e | 2401 | -Wformat-nonliteral -Wformat-security -Wformat-y2k}. |
4d808927 | 2402 | |
b34c7881 JT |
2403 | @item -Wnonnull |
2404 | @opindex Wnonnull | |
f6d9224f | 2405 | Warn about passing a null pointer for arguments marked as |
b34c7881 JT |
2406 | requiring a non-null value by the @code{nonnull} function attribute. |
2407 | ||
2408 | @option{-Wnonnull} is included in @option{-Wall} and @option{-Wformat}. It | |
2409 | can be disabled with the @option{-Wno-nonnull} option. | |
2410 | ||
46e34f96 | 2411 | @item -Winit-self @r{(C, C++, Objective-C and Objective-C++ only)} |
3390f9c9 | 2412 | @opindex Winit-self |
f6d9224f GP |
2413 | Warn about uninitialized variables which are initialized with themselves. |
2414 | Note this option can only be used with the @option{-Wuninitialized} option, | |
2415 | which in turn only works with @option{-O1} and above. | |
3390f9c9 | 2416 | |
f6d9224f GP |
2417 | For example, GCC will warn about @code{i} being uninitialized in the |
2418 | following snippet only when @option{-Winit-self} has been specified: | |
3390f9c9 AP |
2419 | @smallexample |
2420 | @group | |
2421 | int f() | |
2422 | @{ | |
2423 | int i = i; | |
2424 | return i; | |
2425 | @} | |
2426 | @end group | |
2427 | @end smallexample | |
2428 | ||
e9a25f70 | 2429 | @item -Wimplicit-int |
cd3bb277 | 2430 | @opindex Wimplicit-int |
e9a25f70 | 2431 | Warn when a declaration does not specify a type. |
69cdf050 | 2432 | This warning is enabled by @option{-Wall}. |
e9a25f70 | 2433 | |
f5963e61 JL |
2434 | @item -Wimplicit-function-declaration |
2435 | @itemx -Werror-implicit-function-declaration | |
cd3bb277 JM |
2436 | @opindex Wimplicit-function-declaration |
2437 | @opindex Werror-implicit-function-declaration | |
f5963e61 | 2438 | Give a warning (or error) whenever a function is used before being |
b99cfc22 JM |
2439 | declared. The form @option{-Wno-error-implicit-function-declaration} |
2440 | is not supported. | |
69cdf050 | 2441 | This warning is enabled by @option{-Wall} (as a warning, not an error). |
e9a25f70 | 2442 | |
74291a4b | 2443 | @item -Wimplicit |
cd3bb277 | 2444 | @opindex Wimplicit |
630d3d5a | 2445 | Same as @option{-Wimplicit-int} and @option{-Wimplicit-function-declaration}. |
69cdf050 | 2446 | This warning is enabled by @option{-Wall}. |
861bb6c1 JL |
2447 | |
2448 | @item -Wmain | |
cd3bb277 | 2449 | @opindex Wmain |
861bb6c1 JL |
2450 | Warn if the type of @samp{main} is suspicious. @samp{main} should be a |
2451 | function with external linkage, returning int, taking either zero | |
2452 | arguments, two, or three arguments of appropriate types. | |
69cdf050 | 2453 | This warning is enabled by @option{-Wall}. |
4a870dba | 2454 | |
1f0c3120 | 2455 | @item -Wmissing-braces |
cd3bb277 | 2456 | @opindex Wmissing-braces |
1f0c3120 JM |
2457 | Warn if an aggregate or union initializer is not fully bracketed. In |
2458 | the following example, the initializer for @samp{a} is not fully | |
2459 | bracketed, but that for @samp{b} is fully bracketed. | |
2460 | ||
2461 | @smallexample | |
2462 | int a[2][2] = @{ 0, 1, 2, 3 @}; | |
2463 | int b[2][2] = @{ @{ 0, 1 @}, @{ 2, 3 @} @}; | |
2464 | @end smallexample | |
2465 | ||
69cdf050 JM |
2466 | This warning is enabled by @option{-Wall}. |
2467 | ||
46e34f96 | 2468 | @item -Wmissing-include-dirs @r{(C, C++, Objective-C and Objective-C++ only)} |
b02398bd BE |
2469 | @opindex Wmissing-include-dirs |
2470 | Warn if a user-supplied include directory does not exist. | |
2471 | ||
74291a4b | 2472 | @item -Wparentheses |
cd3bb277 | 2473 | @opindex Wparentheses |
74291a4b MM |
2474 | Warn if parentheses are omitted in certain contexts, such |
2475 | as when there is an assignment in a context where a truth value | |
2476 | is expected, or when operators are nested whose precedence people | |
3e3970a2 JM |
2477 | often get confused about. Only the warning for an assignment used as |
2478 | a truth value is supported when compiling C++; the other warnings are | |
2479 | only supported when compiling C@. | |
2480 | ||
2481 | Also warn if a comparison like @samp{x<=y<=z} appears; this is | |
2482 | equivalent to @samp{(x<=y ? 1 : 0) <= z}, which is a different | |
2483 | interpretation from that of ordinary mathematical notation. | |
74291a4b | 2484 | |
e9a25f70 JL |
2485 | Also warn about constructions where there may be confusion to which |
2486 | @code{if} statement an @code{else} branch belongs. Here is an example of | |
2487 | such a case: | |
2488 | ||
2489 | @smallexample | |
aee96fe9 | 2490 | @group |
e9a25f70 JL |
2491 | @{ |
2492 | if (a) | |
2493 | if (b) | |
2494 | foo (); | |
2495 | else | |
2496 | bar (); | |
2497 | @} | |
aee96fe9 | 2498 | @end group |
e9a25f70 JL |
2499 | @end smallexample |
2500 | ||
2501 | In C, every @code{else} branch belongs to the innermost possible @code{if} | |
2502 | statement, which in this example is @code{if (b)}. This is often not | |
2503 | what the programmer expected, as illustrated in the above example by | |
2504 | indentation the programmer chose. When there is the potential for this | |
f0523f02 | 2505 | confusion, GCC will issue a warning when this flag is specified. |
e9a25f70 JL |
2506 | To eliminate the warning, add explicit braces around the innermost |
2507 | @code{if} statement so there is no way the @code{else} could belong to | |
2508 | the enclosing @code{if}. The resulting code would look like this: | |
2509 | ||
2510 | @smallexample | |
aee96fe9 | 2511 | @group |
e9a25f70 JL |
2512 | @{ |
2513 | if (a) | |
2514 | @{ | |
2515 | if (b) | |
2516 | foo (); | |
2517 | else | |
2518 | bar (); | |
2519 | @} | |
2520 | @} | |
aee96fe9 | 2521 | @end group |
e9a25f70 JL |
2522 | @end smallexample |
2523 | ||
69cdf050 JM |
2524 | This warning is enabled by @option{-Wall}. |
2525 | ||
bb58bec5 | 2526 | @item -Wsequence-point |
cd3bb277 | 2527 | @opindex Wsequence-point |
bb58bec5 JM |
2528 | Warn about code that may have undefined semantics because of violations |
2529 | of sequence point rules in the C standard. | |
2530 | ||
2531 | The C standard defines the order in which expressions in a C program are | |
2532 | evaluated in terms of @dfn{sequence points}, which represent a partial | |
2533 | ordering between the execution of parts of the program: those executed | |
2534 | before the sequence point, and those executed after it. These occur | |
2535 | after the evaluation of a full expression (one which is not part of a | |
2536 | larger expression), after the evaluation of the first operand of a | |
2537 | @code{&&}, @code{||}, @code{? :} or @code{,} (comma) operator, before a | |
2538 | function is called (but after the evaluation of its arguments and the | |
2539 | expression denoting the called function), and in certain other places. | |
2540 | Other than as expressed by the sequence point rules, the order of | |
2541 | evaluation of subexpressions of an expression is not specified. All | |
2542 | these rules describe only a partial order rather than a total order, | |
2543 | since, for example, if two functions are called within one expression | |
2544 | with no sequence point between them, the order in which the functions | |
2545 | are called is not specified. However, the standards committee have | |
2546 | ruled that function calls do not overlap. | |
2547 | ||
2548 | It is not specified when between sequence points modifications to the | |
2549 | values of objects take effect. Programs whose behavior depends on this | |
2550 | have undefined behavior; the C standard specifies that ``Between the | |
2551 | previous and next sequence point an object shall have its stored value | |
2552 | modified at most once by the evaluation of an expression. Furthermore, | |
2553 | the prior value shall be read only to determine the value to be | |
2554 | stored.''. If a program breaks these rules, the results on any | |
2555 | particular implementation are entirely unpredictable. | |
2556 | ||
2557 | Examples of code with undefined behavior are @code{a = a++;}, @code{a[n] | |
2558 | = b[n++]} and @code{a[i++] = i;}. Some more complicated cases are not | |
2559 | diagnosed by this option, and it may give an occasional false positive | |
2560 | result, but in general it has been found fairly effective at detecting | |
2561 | this sort of problem in programs. | |
2562 | ||
2563 | The present implementation of this option only works for C programs. A | |
2564 | future implementation may also work for C++ programs. | |
2565 | ||
9c34dbbf ZW |
2566 | The C standard is worded confusingly, therefore there is some debate |
2567 | over the precise meaning of the sequence point rules in subtle cases. | |
2568 | Links to discussions of the problem, including proposed formal | |
962e6e00 | 2569 | definitions, may be found on the GCC readings page, at |
c5122d75 | 2570 | @w{@uref{http://gcc.gnu.org/readings.html}}. |
bb58bec5 | 2571 | |
69cdf050 JM |
2572 | This warning is enabled by @option{-Wall}. |
2573 | ||
74291a4b | 2574 | @item -Wreturn-type |
cd3bb277 | 2575 | @opindex Wreturn-type |
32c4c36c ML |
2576 | Warn whenever a function is defined with a return-type that defaults to |
2577 | @code{int}. Also warn about any @code{return} statement with no | |
02f52e19 | 2578 | return-value in a function whose return-type is not @code{void}. |
32c4c36c | 2579 | |
e508a019 JM |
2580 | For C, also warn if the return type of a function has a type qualifier |
2581 | such as @code{const}. Such a type qualifier has no effect, since the | |
2582 | value returned by a function is not an lvalue. ISO C prohibits | |
2583 | qualified @code{void} return types on function definitions, so such | |
2584 | return types always receive a warning even without this option. | |
2585 | ||
32c4c36c | 2586 | For C++, a function without return type always produces a diagnostic |
767094dd | 2587 | message, even when @option{-Wno-return-type} is specified. The only |
32c4c36c | 2588 | exceptions are @samp{main} and functions defined in system headers. |
74291a4b | 2589 | |
69cdf050 JM |
2590 | This warning is enabled by @option{-Wall}. |
2591 | ||
74291a4b | 2592 | @item -Wswitch |
cd3bb277 | 2593 | @opindex Wswitch |
2eac577f | 2594 | Warn whenever a @code{switch} statement has an index of enumerated type |
74291a4b MM |
2595 | and lacks a @code{case} for one or more of the named codes of that |
2596 | enumeration. (The presence of a @code{default} label prevents this | |
2597 | warning.) @code{case} labels outside the enumeration range also | |
2598 | provoke warnings when this option is used. | |
69cdf050 | 2599 | This warning is enabled by @option{-Wall}. |
74291a4b | 2600 | |
d6961341 AC |
2601 | @item -Wswitch-default |
2602 | @opindex Wswitch-switch | |
2603 | Warn whenever a @code{switch} statement does not have a @code{default} | |
2604 | case. | |
2605 | ||
173028e5 AC |
2606 | @item -Wswitch-enum |
2607 | @opindex Wswitch-enum | |
2eac577f | 2608 | Warn whenever a @code{switch} statement has an index of enumerated type |
173028e5 AC |
2609 | and lacks a @code{case} for one or more of the named codes of that |
2610 | enumeration. @code{case} labels outside the enumeration range also | |
2611 | provoke warnings when this option is used. | |
2612 | ||
74291a4b | 2613 | @item -Wtrigraphs |
cd3bb277 | 2614 | @opindex Wtrigraphs |
f2ecb02d JM |
2615 | Warn if any trigraphs are encountered that might change the meaning of |
2616 | the program (trigraphs within comments are not warned about). | |
69cdf050 | 2617 | This warning is enabled by @option{-Wall}. |
74291a4b | 2618 | |
078721e1 | 2619 | @item -Wunused-function |
cd3bb277 | 2620 | @opindex Wunused-function |
078721e1 | 2621 | Warn whenever a static function is declared but not defined or a |
aa58883c | 2622 | non-inline static function is unused. |
69cdf050 | 2623 | This warning is enabled by @option{-Wall}. |
74291a4b | 2624 | |
078721e1 | 2625 | @item -Wunused-label |
cd3bb277 | 2626 | @opindex Wunused-label |
078721e1 | 2627 | Warn whenever a label is declared but not used. |
69cdf050 | 2628 | This warning is enabled by @option{-Wall}. |
078721e1 AC |
2629 | |
2630 | To suppress this warning use the @samp{unused} attribute | |
2631 | (@pxref{Variable Attributes}). | |
2632 | ||
2633 | @item -Wunused-parameter | |
cd3bb277 | 2634 | @opindex Wunused-parameter |
078721e1 AC |
2635 | Warn whenever a function parameter is unused aside from its declaration. |
2636 | ||
2637 | To suppress this warning use the @samp{unused} attribute | |
2638 | (@pxref{Variable Attributes}). | |
956d6950 | 2639 | |
078721e1 | 2640 | @item -Wunused-variable |
cd3bb277 | 2641 | @opindex Wunused-variable |
078721e1 AC |
2642 | Warn whenever a local variable or non-constant static variable is unused |
2643 | aside from its declaration | |
69cdf050 | 2644 | This warning is enabled by @option{-Wall}. |
078721e1 AC |
2645 | |
2646 | To suppress this warning use the @samp{unused} attribute | |
74291a4b MM |
2647 | (@pxref{Variable Attributes}). |
2648 | ||
078721e1 | 2649 | @item -Wunused-value |
cd3bb277 | 2650 | @opindex Wunused-value |
078721e1 | 2651 | Warn whenever a statement computes a result that is explicitly not used. |
69cdf050 | 2652 | This warning is enabled by @option{-Wall}. |
078721e1 AC |
2653 | |
2654 | To suppress this warning cast the expression to @samp{void}. | |
2655 | ||
2656 | @item -Wunused | |
cd3bb277 | 2657 | @opindex Wunused |
d3075b6c | 2658 | All the above @option{-Wunused} options combined. |
078721e1 AC |
2659 | |
2660 | In order to get a warning about an unused function parameter, you must | |
65ca2d60 PE |
2661 | either specify @samp{-Wextra -Wunused} (note that @samp{-Wall} implies |
2662 | @samp{-Wunused}), or separately specify @option{-Wunused-parameter}. | |
078721e1 | 2663 | |
74291a4b | 2664 | @item -Wuninitialized |
cd3bb277 | 2665 | @opindex Wuninitialized |
c5c76735 JL |
2666 | Warn if an automatic variable is used without first being initialized or |
2667 | if a variable may be clobbered by a @code{setjmp} call. | |
74291a4b MM |
2668 | |
2669 | These warnings are possible only in optimizing compilation, | |
2670 | because they require data flow information that is computed only | |
630d3d5a | 2671 | when optimizing. If you don't specify @option{-O}, you simply won't |
74291a4b MM |
2672 | get these warnings. |
2673 | ||
3390f9c9 AP |
2674 | If you want to warn about code which uses the uninitialized value of the |
2675 | variable in its own initializer, use the @option{-Winit-self} option. | |
2676 | ||
8ceac9f8 JM |
2677 | These warnings occur for individual uninitialized or clobbered |
2678 | elements of structure, union or array variables as well as for | |
2679 | variables which are uninitialized or clobbered as a whole. They do | |
2680 | not occur for variables or elements declared @code{volatile}. Because | |
2681 | these warnings depend on optimization, the exact variables or elements | |
2682 | for which there are warnings will depend on the precise optimization | |
2683 | options and version of GCC used. | |
74291a4b MM |
2684 | |
2685 | Note that there may be no warning about a variable that is used only | |
2686 | to compute a value that itself is never used, because such | |
2687 | computations may be deleted by data flow analysis before the warnings | |
2688 | are printed. | |
2689 | ||
0c2d1a2a | 2690 | These warnings are made optional because GCC is not smart |
74291a4b MM |
2691 | enough to see all the reasons why the code might be correct |
2692 | despite appearing to have an error. Here is one example of how | |
2693 | this can happen: | |
2694 | ||
2695 | @smallexample | |
aee96fe9 | 2696 | @group |
74291a4b MM |
2697 | @{ |
2698 | int x; | |
2699 | switch (y) | |
2700 | @{ | |
2701 | case 1: x = 1; | |
2702 | break; | |
2703 | case 2: x = 4; | |
2704 | break; | |
2705 | case 3: x = 5; | |
2706 | @} | |
2707 | foo (x); | |
2708 | @} | |
aee96fe9 | 2709 | @end group |
74291a4b MM |
2710 | @end smallexample |
2711 | ||
2712 | @noindent | |
2713 | If the value of @code{y} is always 1, 2 or 3, then @code{x} is | |
0c2d1a2a | 2714 | always initialized, but GCC doesn't know this. Here is |
74291a4b MM |
2715 | another common case: |
2716 | ||
2717 | @smallexample | |
2718 | @{ | |
2719 | int save_y; | |
2720 | if (change_y) save_y = y, y = new_y; | |
2721 | @dots{} | |
2722 | if (change_y) y = save_y; | |
2723 | @} | |
2724 | @end smallexample | |
2725 | ||
2726 | @noindent | |
2727 | This has no bug because @code{save_y} is used only if it is set. | |
2728 | ||
20300b05 | 2729 | @cindex @code{longjmp} warnings |
b192711e | 2730 | This option also warns when a non-volatile automatic variable might be |
c5c76735 JL |
2731 | changed by a call to @code{longjmp}. These warnings as well are possible |
2732 | only in optimizing compilation. | |
20300b05 GK |
2733 | |
2734 | The compiler sees only the calls to @code{setjmp}. It cannot know | |
2735 | where @code{longjmp} will be called; in fact, a signal handler could | |
2736 | call it at any point in the code. As a result, you may get a warning | |
2737 | even when there is in fact no problem because @code{longjmp} cannot | |
2738 | in fact be called at the place which would cause a problem. | |
2739 | ||
74291a4b MM |
2740 | Some spurious warnings can be avoided if you declare all the functions |
2741 | you use that never return as @code{noreturn}. @xref{Function | |
2742 | Attributes}. | |
2743 | ||
69cdf050 JM |
2744 | This warning is enabled by @option{-Wall}. |
2745 | ||
d300e551 | 2746 | @item -Wunknown-pragmas |
cd3bb277 | 2747 | @opindex Wunknown-pragmas |
d300e551 NC |
2748 | @cindex warning for unknown pragmas |
2749 | @cindex unknown pragmas, warning | |
2750 | @cindex pragmas, warning of unknown | |
2751 | Warn when a #pragma directive is encountered which is not understood by | |
161d7b59 | 2752 | GCC@. If this command line option is used, warnings will even be issued |
d300e551 | 2753 | for unknown pragmas in system header files. This is not the case if |
630d3d5a | 2754 | the warnings were only enabled by the @option{-Wall} command line option. |
d300e551 | 2755 | |
b9b8dde3 DD |
2756 | @item -Wno-pragmas |
2757 | @opindex Wno-pragmas | |
2758 | @opindex Wpragmas | |
2759 | Do not warn about misuses of pragmas, such as incorrect parameters, | |
2760 | invalid syntax, or conflicts between pragmas. See also | |
2761 | @samp{-Wunknown-pragmas}. | |
2762 | ||
bf52f899 NS |
2763 | @item -Wstrict-aliasing |
2764 | @opindex Wstrict-aliasing | |
2765 | This option is only active when @option{-fstrict-aliasing} is active. | |
2766 | It warns about code which might break the strict aliasing rules that the | |
8a36672b JM |
2767 | compiler is using for optimization. The warning does not catch all |
2768 | cases, but does attempt to catch the more common pitfalls. It is | |
bf52f899 NS |
2769 | included in @option{-Wall}. |
2770 | ||
5399d643 JW |
2771 | @item -Wstrict-aliasing=2 |
2772 | @opindex Wstrict-aliasing=2 | |
2773 | This option is only active when @option{-fstrict-aliasing} is active. | |
66f33c03 JW |
2774 | It warns about code which might break the strict aliasing rules that the |
2775 | compiler is using for optimization. This warning catches more cases than | |
2776 | @option{-Wstrict-aliasing}, but it will also give a warning for some ambiguous | |
2777 | cases that are safe. | |
5399d643 | 2778 | |
74291a4b | 2779 | @item -Wall |
cd3bb277 | 2780 | @opindex Wall |
74291a4b MM |
2781 | All of the above @samp{-W} options combined. This enables all the |
2782 | warnings about constructions that some users consider questionable, and | |
2783 | that are easy to avoid (or modify to prevent the warning), even in | |
bd8f9aec SP |
2784 | conjunction with macros. This also enables some language-specific |
2785 | warnings described in @ref{C++ Dialect Options} and | |
46e34f96 | 2786 | @ref{Objective-C and Objective-C++ Dialect Options}. |
74291a4b MM |
2787 | @end table |
2788 | ||
630d3d5a | 2789 | The following @option{-W@dots{}} options are not implied by @option{-Wall}. |
74291a4b MM |
2790 | Some of them warn about constructions that users generally do not |
2791 | consider questionable, but which occasionally you might wish to check | |
2792 | for; others warn about constructions that are necessary or hard to avoid | |
2793 | in some cases, and there is no simple way to modify the code to suppress | |
2794 | the warning. | |
2795 | ||
2642624b | 2796 | @table @gcctabopt |
65ca2d60 | 2797 | @item -Wextra |
cd3bb277 | 2798 | @opindex W |
65ca2d60 PE |
2799 | @opindex Wextra |
2800 | (This option used to be called @option{-W}. The older name is still | |
2801 | supported, but the newer name is more descriptive.) Print extra warning | |
2802 | messages for these events: | |
74291a4b MM |
2803 | |
2804 | @itemize @bullet | |
74291a4b MM |
2805 | @item |
2806 | A function can return either with or without a value. (Falling | |
2807 | off the end of the function body is considered returning without | |
2808 | a value.) For example, this function would evoke such a | |
2809 | warning: | |
2810 | ||
2811 | @smallexample | |
2812 | @group | |
2813 | foo (a) | |
2814 | @{ | |
2815 | if (a > 0) | |
2816 | return a; | |
2817 | @} | |
2818 | @end group | |
2819 | @end smallexample | |
2820 | ||
2821 | @item | |
2822 | An expression-statement or the left-hand side of a comma expression | |
2823 | contains no side effects. | |
2824 | To suppress the warning, cast the unused expression to void. | |
2825 | For example, an expression such as @samp{x[i,j]} will cause a warning, | |
2826 | but @samp{x[(void)i,j]} will not. | |
2827 | ||
2828 | @item | |
65ca2d60 | 2829 | An unsigned value is compared against zero with @samp{<} or @samp{>=}. |
74291a4b | 2830 | |
74291a4b MM |
2831 | @item |
2832 | Storage-class specifiers like @code{static} are not the first things in | |
2833 | a declaration. According to the C Standard, this usage is obsolescent. | |
2834 | ||
2835 | @item | |
630d3d5a | 2836 | If @option{-Wall} or @option{-Wunused} is also specified, warn about unused |
74291a4b MM |
2837 | arguments. |
2838 | ||
e9a25f70 JL |
2839 | @item |
2840 | A comparison between signed and unsigned values could produce an | |
2841 | incorrect result when the signed value is converted to unsigned. | |
630d3d5a | 2842 | (But don't warn if @option{-Wno-sign-compare} is also specified.) |
e9a25f70 | 2843 | |
dbde0d5d BH |
2844 | @item |
2845 | An aggregate has an initializer which does not initialize all members. | |
eaac4679 RS |
2846 | This warning can be independently controlled by |
2847 | @option{-Wmissing-field-initializers}. | |
65ca2d60 PE |
2848 | |
2849 | @item | |
2850 | A function parameter is declared without a type specifier in K&R-style | |
2851 | functions: | |
2852 | ||
2853 | @smallexample | |
2854 | void foo(bar) @{ @} | |
2855 | @end smallexample | |
2856 | ||
2857 | @item | |
2858 | An empty body occurs in an @samp{if} or @samp{else} statement. | |
2859 | ||
2860 | @item | |
2861 | A pointer is compared against integer zero with @samp{<}, @samp{<=}, | |
2862 | @samp{>}, or @samp{>=}. | |
2863 | ||
2864 | @item | |
2865 | A variable might be changed by @samp{longjmp} or @samp{vfork}. | |
2866 | ||
2867 | @item | |
2868 | Any of several floating-point events that often indicate errors, such as | |
2869 | overflow, underflow, loss of precision, etc. | |
2870 | ||
2871 | @item @r{(C++ only)} | |
2872 | An enumerator and a non-enumerator both appear in a conditional expression. | |
2873 | ||
2874 | @item @r{(C++ only)} | |
2875 | A non-static reference or non-static @samp{const} member appears in a | |
2876 | class without constructors. | |
2877 | ||
2878 | @item @r{(C++ only)} | |
2879 | Ambiguous virtual bases. | |
2880 | ||
2881 | @item @r{(C++ only)} | |
2882 | Subscripting an array which has been declared @samp{register}. | |
2883 | ||
2884 | @item @r{(C++ only)} | |
2885 | Taking the address of a variable which has been declared @samp{register}. | |
2886 | ||
2887 | @item @r{(C++ only)} | |
62b9c42c | 2888 | A base class is not initialized in a derived class' copy constructor. |
74291a4b MM |
2889 | @end itemize |
2890 | ||
75227a33 GK |
2891 | @item -Wno-div-by-zero |
2892 | @opindex Wno-div-by-zero | |
2893 | @opindex Wdiv-by-zero | |
2894 | Do not warn about compile-time integer division by zero. Floating point | |
2895 | division by zero is not warned about, as it can be a legitimate way of | |
2896 | obtaining infinities and NaNs. | |
2897 | ||
2898 | @item -Wsystem-headers | |
2899 | @opindex Wsystem-headers | |
2900 | @cindex warnings from system headers | |
2901 | @cindex system headers, warnings from | |
2902 | Print warning messages for constructs found in system header files. | |
2903 | Warnings from system headers are normally suppressed, on the assumption | |
2904 | that they usually do not indicate real problems and would only make the | |
2905 | compiler output harder to read. Using this command line option tells | |
2906 | GCC to emit warnings from system headers as if they occurred in user | |
2907 | code. However, note that using @option{-Wall} in conjunction with this | |
2908 | option will @emph{not} warn about unknown pragmas in system | |
2909 | headers---for that, @option{-Wunknown-pragmas} must also be used. | |
2910 | ||
f793a95e | 2911 | @item -Wfloat-equal |
cd3bb277 | 2912 | @opindex Wfloat-equal |
f793a95e JL |
2913 | Warn if floating point values are used in equality comparisons. |
2914 | ||
488d3985 GK |
2915 | The idea behind this is that sometimes it is convenient (for the |
2916 | programmer) to consider floating-point values as approximations to | |
2917 | infinitely precise real numbers. If you are doing this, then you need | |
c0478a66 | 2918 | to compute (by analyzing the code, or in some other way) the maximum or |
488d3985 GK |
2919 | likely maximum error that the computation introduces, and allow for it |
2920 | when performing comparisons (and when producing output, but that's a | |
2921 | different problem). In particular, instead of testing for equality, you | |
2922 | would check to see whether the two values have ranges that overlap; and | |
2923 | this is done with the relational operators, so equality comparisons are | |
2924 | probably mistaken. | |
2925 | ||
aee96fe9 | 2926 | @item -Wtraditional @r{(C only)} |
cd3bb277 | 2927 | @opindex Wtraditional |
74291a4b | 2928 | Warn about certain constructs that behave differently in traditional and |
161d7b59 | 2929 | ISO C@. Also warn about ISO C constructs that have no traditional C |
c8abc684 | 2930 | equivalent, and/or problematic constructs which should be avoided. |
74291a4b MM |
2931 | |
2932 | @itemize @bullet | |
2933 | @item | |
da312b55 NB |
2934 | Macro parameters that appear within string literals in the macro body. |
2935 | In traditional C macro replacement takes place within string literals, | |
161d7b59 | 2936 | but does not in ISO C@. |
da312b55 NB |
2937 | |
2938 | @item | |
2939 | In traditional C, some preprocessor directives did not exist. | |
2940 | Traditional preprocessors would only consider a line to be a directive | |
2941 | if the @samp{#} appeared in column 1 on the line. Therefore | |
630d3d5a | 2942 | @option{-Wtraditional} warns about directives that traditional C |
da312b55 NB |
2943 | understands but would ignore because the @samp{#} does not appear as the |
2944 | first character on the line. It also suggests you hide directives like | |
2945 | @samp{#pragma} not understood by traditional C by indenting them. Some | |
c21cd8b1 | 2946 | traditional implementations would not recognize @samp{#elif}, so it |
da312b55 NB |
2947 | suggests avoiding it altogether. |
2948 | ||
2949 | @item | |
2950 | A function-like macro that appears without arguments. | |
2951 | ||
2952 | @item | |
2953 | The unary plus operator. | |
2954 | ||
2955 | @item | |
c771326b JM |
2956 | The @samp{U} integer constant suffix, or the @samp{F} or @samp{L} floating point |
2957 | constant suffixes. (Traditional C does support the @samp{L} suffix on integer | |
da312b55 | 2958 | constants.) Note, these suffixes appear in macros defined in the system |
e979f9e8 | 2959 | headers of most modern systems, e.g.@: the @samp{_MIN}/@samp{_MAX} macros in @code{<limits.h>}. |
c8abc684 | 2960 | Use of these macros in user code might normally lead to spurious |
2dd76960 | 2961 | warnings, however GCC's integrated preprocessor has enough context to |
c8abc684 | 2962 | avoid warning in these cases. |
74291a4b MM |
2963 | |
2964 | @item | |
2965 | A function declared external in one block and then used after the end of | |
2966 | the block. | |
2967 | ||
2968 | @item | |
2969 | A @code{switch} statement has an operand of type @code{long}. | |
db838bb8 KG |
2970 | |
2971 | @item | |
2972 | A non-@code{static} function declaration follows a @code{static} one. | |
2973 | This construct is not accepted by some traditional C compilers. | |
48776cde KG |
2974 | |
2975 | @item | |
c1030c7c | 2976 | The ISO type of an integer constant has a different width or |
48776cde | 2977 | signedness from its traditional type. This warning is only issued if |
e979f9e8 | 2978 | the base of the constant is ten. I.e.@: hexadecimal or octal values, which |
48776cde | 2979 | typically represent bit patterns, are not warned about. |
bb66adca KG |
2980 | |
2981 | @item | |
c1030c7c | 2982 | Usage of ISO string concatenation is detected. |
7f094a94 | 2983 | |
895ea614 KG |
2984 | @item |
2985 | Initialization of automatic aggregates. | |
2986 | ||
2987 | @item | |
2988 | Identifier conflicts with labels. Traditional C lacks a separate | |
2989 | namespace for labels. | |
253b6b82 KG |
2990 | |
2991 | @item | |
2992 | Initialization of unions. If the initializer is zero, the warning is | |
2993 | omitted. This is done under the assumption that the zero initializer in | |
e979f9e8 | 2994 | user code appears conditioned on e.g.@: @code{__STDC__} to avoid missing |
253b6b82 KG |
2995 | initializer warnings and relies on default initialization to zero in the |
2996 | traditional C case. | |
03829ad2 KG |
2997 | |
2998 | @item | |
3ed56f8a KG |
2999 | Conversions by prototypes between fixed/floating point values and vice |
3000 | versa. The absence of these prototypes when compiling with traditional | |
3001 | C would cause serious problems. This is a subset of the possible | |
630d3d5a | 3002 | conversion warnings, for the full set use @option{-Wconversion}. |
622d3731 KG |
3003 | |
3004 | @item | |
3005 | Use of ISO C style function definitions. This warning intentionally is | |
3006 | @emph{not} issued for prototype declarations or variadic functions | |
3007 | because these ISO C features will appear in your code when using | |
3008 | libiberty's traditional C compatibility macros, @code{PARAMS} and | |
3009 | @code{VPARAMS}. This warning is also bypassed for nested functions | |
2dd76960 | 3010 | because that feature is already a GCC extension and thus not relevant to |
622d3731 | 3011 | traditional C compatibility. |
74291a4b MM |
3012 | @end itemize |
3013 | ||
85617eba HPN |
3014 | @item -Wdeclaration-after-statement @r{(C only)} |
3015 | @opindex Wdeclaration-after-statement | |
3016 | Warn when a declaration is found after a statement in a block. This | |
3017 | construct, known from C++, was introduced with ISO C99 and is by default | |
3018 | allowed in GCC@. It is not supported by ISO C90 and was not supported by | |
3019 | GCC versions before GCC 3.0. @xref{Mixed Declarations}. | |
3020 | ||
861bb6c1 | 3021 | @item -Wundef |
cd3bb277 | 3022 | @opindex Wundef |
861bb6c1 JL |
3023 | Warn if an undefined identifier is evaluated in an @samp{#if} directive. |
3024 | ||
90689ae1 JM |
3025 | @item -Wno-endif-labels |
3026 | @opindex Wno-endif-labels | |
909de5da | 3027 | @opindex Wendif-labels |
90689ae1 | 3028 | Do not warn whenever an @samp{#else} or an @samp{#endif} are followed by text. |
909de5da | 3029 | |
74291a4b | 3030 | @item -Wshadow |
cd3bb277 | 3031 | @opindex Wshadow |
d773df5a DB |
3032 | Warn whenever a local variable shadows another local variable, parameter or |
3033 | global variable or whenever a built-in function is shadowed. | |
74291a4b | 3034 | |
74291a4b | 3035 | @item -Wlarger-than-@var{len} |
cd3bb277 | 3036 | @opindex Wlarger-than |
74291a4b MM |
3037 | Warn whenever an object of larger than @var{len} bytes is defined. |
3038 | ||
f9cc1a70 PB |
3039 | @item -Wunsafe-loop-optimizations |
3040 | @opindex Wunsafe-loop-optimizations | |
3041 | Warn if the loop cannot be optimized because the compiler could not | |
3042 | assume anything on the bounds of the loop indices. With | |
3043 | @option{-funsafe-loop-optimizations} warn if the compiler made | |
3044 | such assumptions. | |
3045 | ||
74291a4b | 3046 | @item -Wpointer-arith |
cd3bb277 | 3047 | @opindex Wpointer-arith |
74291a4b MM |
3048 | Warn about anything that depends on the ``size of'' a function type or |
3049 | of @code{void}. GNU C assigns these types a size of 1, for | |
3050 | convenience in calculations with @code{void *} pointers and pointers | |
3051 | to functions. | |
3052 | ||
aee96fe9 | 3053 | @item -Wbad-function-cast @r{(C only)} |
cd3bb277 | 3054 | @opindex Wbad-function-cast |
74291a4b MM |
3055 | Warn whenever a function call is cast to a non-matching type. |
3056 | For example, warn if @code{int malloc()} is cast to @code{anything *}. | |
3057 | ||
b7e20b53 GDR |
3058 | @item -Wc++-compat |
3059 | Warn about ISO C constructs that are outside of the common subset of | |
3060 | ISO C and ISO C++, e.g.@: request for implicit conversion from | |
3061 | @code{void *} to a pointer to non-@code{void} type. | |
3062 | ||
74291a4b | 3063 | @item -Wcast-qual |
cd3bb277 | 3064 | @opindex Wcast-qual |
74291a4b MM |
3065 | Warn whenever a pointer is cast so as to remove a type qualifier from |
3066 | the target type. For example, warn if a @code{const char *} is cast | |
3067 | to an ordinary @code{char *}. | |
3068 | ||
3069 | @item -Wcast-align | |
cd3bb277 | 3070 | @opindex Wcast-align |
74291a4b MM |
3071 | Warn whenever a pointer is cast such that the required alignment of the |
3072 | target is increased. For example, warn if a @code{char *} is cast to | |
3073 | an @code{int *} on machines where integers can only be accessed at | |
3074 | two- or four-byte boundaries. | |
3075 | ||
3076 | @item -Wwrite-strings | |
cd3bb277 | 3077 | @opindex Wwrite-strings |
aee96fe9 JM |
3078 | When compiling C, give string constants the type @code{const |
3079 | char[@var{length}]} so that | |
74291a4b | 3080 | copying the address of one into a non-@code{const} @code{char *} |
aee96fe9 JM |
3081 | pointer will get a warning; when compiling C++, warn about the |
3082 | deprecated conversion from string constants to @code{char *}. | |
3083 | These warnings will help you find at | |
74291a4b MM |
3084 | compile time code that can try to write into a string constant, but |
3085 | only if you have been very careful about using @code{const} in | |
3086 | declarations and prototypes. Otherwise, it will just be a nuisance; | |
630d3d5a | 3087 | this is why we did not make @option{-Wall} request these warnings. |
74291a4b MM |
3088 | |
3089 | @item -Wconversion | |
cd3bb277 | 3090 | @opindex Wconversion |
74291a4b MM |
3091 | Warn if a prototype causes a type conversion that is different from what |
3092 | would happen to the same argument in the absence of a prototype. This | |
3093 | includes conversions of fixed point to floating and vice versa, and | |
3094 | conversions changing the width or signedness of a fixed point argument | |
3095 | except when the same as the default promotion. | |
3096 | ||
3097 | Also, warn if a negative integer constant expression is implicitly | |
3098 | converted to an unsigned type. For example, warn about the assignment | |
3099 | @code{x = -1} if @code{x} is unsigned. But do not warn about explicit | |
3100 | casts like @code{(unsigned) -1}. | |
3101 | ||
e9a25f70 | 3102 | @item -Wsign-compare |
cd3bb277 | 3103 | @opindex Wsign-compare |
e9a25f70 JL |
3104 | @cindex warning for comparison of signed and unsigned values |
3105 | @cindex comparison of signed and unsigned values, warning | |
3106 | @cindex signed and unsigned values, comparison warning | |
3107 | Warn when a comparison between signed and unsigned values could produce | |
3108 | an incorrect result when the signed value is converted to unsigned. | |
65ca2d60 PE |
3109 | This warning is also enabled by @option{-Wextra}; to get the other warnings |
3110 | of @option{-Wextra} without this warning, use @samp{-Wextra -Wno-sign-compare}. | |
e9a25f70 | 3111 | |
74291a4b | 3112 | @item -Waggregate-return |
cd3bb277 | 3113 | @opindex Waggregate-return |
74291a4b MM |
3114 | Warn if any functions that return structures or unions are defined or |
3115 | called. (In languages where you can return an array, this also elicits | |
3116 | a warning.) | |
3117 | ||
690a704a BE |
3118 | @item -Walways-true |
3119 | @opindex Walways-true | |
03237161 BE |
3120 | Warn about comparisons which are always true such as testing if |
3121 | unsigned values are greater than or equal to zero. This warning is | |
3122 | enabled by @option{-Wall}. | |
690a704a | 3123 | |
5c498b10 DD |
3124 | @item -Wno-attributes |
3125 | @opindex Wno-attributes | |
3126 | @opindex Wattributes | |
3127 | Do not warn if an unexpected @code{__attribute__} is used, such as | |
3128 | unrecognized attributes, function attributes applied to variables, | |
3129 | etc. This will not stop errors for incorrect use of supported | |
3130 | attributes. | |
3131 | ||
aee96fe9 | 3132 | @item -Wstrict-prototypes @r{(C only)} |
cd3bb277 | 3133 | @opindex Wstrict-prototypes |
74291a4b MM |
3134 | Warn if a function is declared or defined without specifying the |
3135 | argument types. (An old-style function definition is permitted without | |
3136 | a warning if preceded by a declaration which specifies the argument | |
3137 | types.) | |
3138 | ||
c034f121 AJ |
3139 | @item -Wold-style-definition @r{(C only)} |
3140 | @opindex Wold-style-definition | |
3141 | Warn if an old-style function definition is used. A warning is given | |
3142 | even if there is a previous prototype. | |
3143 | ||
aee96fe9 | 3144 | @item -Wmissing-prototypes @r{(C only)} |
cd3bb277 | 3145 | @opindex Wmissing-prototypes |
74291a4b MM |
3146 | Warn if a global function is defined without a previous prototype |
3147 | declaration. This warning is issued even if the definition itself | |
3148 | provides a prototype. The aim is to detect global functions that fail | |
3149 | to be declared in header files. | |
3150 | ||
da635858 | 3151 | @item -Wmissing-declarations @r{(C only)} |
cd3bb277 | 3152 | @opindex Wmissing-declarations |
74291a4b MM |
3153 | Warn if a global function is defined without a previous declaration. |
3154 | Do so even if the definition itself provides a prototype. | |
3155 | Use this option to detect global functions that are not declared in | |
3156 | header files. | |
3157 | ||
eaac4679 RS |
3158 | @item -Wmissing-field-initializers |
3159 | @opindex Wmissing-field-initializers | |
3160 | @opindex W | |
3161 | @opindex Wextra | |
3162 | Warn if a structure's initializer has some fields missing. For | |
3163 | example, the following code would cause such a warning, because | |
3164 | @code{x.h} is implicitly zero: | |
3165 | ||
3166 | @smallexample | |
3167 | struct s @{ int f, g, h; @}; | |
3168 | struct s x = @{ 3, 4 @}; | |
3169 | @end smallexample | |
3170 | ||
3171 | This option does not warn about designated initializers, so the following | |
3172 | modification would not trigger a warning: | |
3173 | ||
3174 | @smallexample | |
3175 | struct s @{ int f, g, h; @}; | |
3176 | struct s x = @{ .f = 3, .g = 4 @}; | |
3177 | @end smallexample | |
3178 | ||
3179 | This warning is included in @option{-Wextra}. To get other @option{-Wextra} | |
3180 | warnings without this one, use @samp{-Wextra -Wno-missing-field-initializers}. | |
3181 | ||
0ca3fb0a | 3182 | @item -Wmissing-noreturn |
cd3bb277 | 3183 | @opindex Wmissing-noreturn |
0ca3fb0a KG |
3184 | Warn about functions which might be candidates for attribute @code{noreturn}. |
3185 | Note these are only possible candidates, not absolute ones. Care should | |
3186 | be taken to manually verify functions actually do not ever return before | |
3187 | adding the @code{noreturn} attribute, otherwise subtle code generation | |
21c7361e AJ |
3188 | bugs could be introduced. You will not get a warning for @code{main} in |
3189 | hosted C environments. | |
0ca3fb0a | 3190 | |
74ff4629 | 3191 | @item -Wmissing-format-attribute |
cd3bb277 JM |
3192 | @opindex Wmissing-format-attribute |
3193 | @opindex Wformat | |
7876a414 KG |
3194 | Warn about function pointers which might be candidates for @code{format} |
3195 | attributes. Note these are only possible candidates, not absolute ones. | |
3196 | GCC will guess that function pointers with @code{format} attributes that | |
3197 | are used in assignment, initialization, parameter passing or return | |
3198 | statements should have a corresponding @code{format} attribute in the | |
3199 | resulting type. I.e.@: the left-hand side of the assignment or | |
3200 | initialization, the type of the parameter variable, or the return type | |
3201 | of the containing function respectively should also have a @code{format} | |
3202 | attribute to avoid the warning. | |
3203 | ||
3204 | GCC will also warn about function definitions which might be | |
3205 | candidates for @code{format} attributes. Again, these are only | |
3206 | possible candidates. GCC will guess that @code{format} attributes | |
3207 | might be appropriate for any function that calls a function like | |
3208 | @code{vprintf} or @code{vscanf}, but this might not always be the | |
74ff4629 | 3209 | case, and some functions for which @code{format} attributes are |
7876a414 | 3210 | appropriate may not be detected. |
74ff4629 | 3211 | |
75227a33 GK |
3212 | @item -Wno-multichar |
3213 | @opindex Wno-multichar | |
3214 | @opindex Wmultichar | |
3215 | Do not warn if a multicharacter constant (@samp{'FOOF'}) is used. | |
3216 | Usually they indicate a typo in the user's code, as they have | |
3217 | implementation-defined values, and should not be used in portable code. | |
3218 | ||
50668cf6 GK |
3219 | @item -Wnormalized=<none|id|nfc|nfkc> |
3220 | @opindex Wnormalized | |
3221 | @cindex NFC | |
3222 | @cindex NFKC | |
3223 | @cindex character set, input normalization | |
3224 | In ISO C and ISO C++, two identifiers are different if they are | |
3225 | different sequences of characters. However, sometimes when characters | |
3226 | outside the basic ASCII character set are used, you can have two | |
3227 | different character sequences that look the same. To avoid confusion, | |
3228 | the ISO 10646 standard sets out some @dfn{normalization rules} which | |
3229 | when applied ensure that two sequences that look the same are turned into | |
3230 | the same sequence. GCC can warn you if you are using identifiers which | |
3231 | have not been normalized; this option controls that warning. | |
3232 | ||
3233 | There are four levels of warning that GCC supports. The default is | |
3234 | @option{-Wnormalized=nfc}, which warns about any identifier which is | |
3235 | not in the ISO 10646 ``C'' normalized form, @dfn{NFC}. NFC is the | |
3236 | recommended form for most uses. | |
3237 | ||
3238 | Unfortunately, there are some characters which ISO C and ISO C++ allow | |
3239 | in identifiers that when turned into NFC aren't allowable as | |
3240 | identifiers. That is, there's no way to use these symbols in portable | |
3241 | ISO C or C++ and have all your identifiers in NFC. | |
3242 | @option{-Wnormalized=id} suppresses the warning for these characters. | |
3243 | It is hoped that future versions of the standards involved will correct | |
3244 | this, which is why this option is not the default. | |
3245 | ||
3246 | You can switch the warning off for all characters by writing | |
3247 | @option{-Wnormalized=none}. You would only want to do this if you | |
3248 | were using some other normalization scheme (like ``D''), because | |
3249 | otherwise you can easily create bugs that are literally impossible to see. | |
3250 | ||
3251 | Some characters in ISO 10646 have distinct meanings but look identical | |
3252 | in some fonts or display methodologies, especially once formatting has | |
3253 | been applied. For instance @code{\u207F}, ``SUPERSCRIPT LATIN SMALL | |
3254 | LETTER N'', will display just like a regular @code{n} which has been | |
3255 | placed in a superscript. ISO 10646 defines the @dfn{NFKC} | |
3256 | normalisation scheme to convert all these into a standard form as | |
3257 | well, and GCC will warn if your code is not in NFKC if you use | |
3258 | @option{-Wnormalized=nfkc}. This warning is comparable to warning | |
3259 | about every identifier that contains the letter O because it might be | |
3260 | confused with the digit 0, and so is not the default, but may be | |
3261 | useful as a local coding convention if the programming environment is | |
3262 | unable to be fixed to display these characters distinctly. | |
3263 | ||
e23bd218 IR |
3264 | @item -Wno-deprecated-declarations |
3265 | @opindex Wno-deprecated-declarations | |
3266 | Do not warn about uses of functions, variables, and types marked as | |
f282ffb3 | 3267 | deprecated by using the @code{deprecated} attribute. |
e23bd218 IR |
3268 | (@pxref{Function Attributes}, @pxref{Variable Attributes}, |
3269 | @pxref{Type Attributes}.) | |
3270 | ||
3c12fcc2 | 3271 | @item -Wpacked |
cd3bb277 | 3272 | @opindex Wpacked |
3c12fcc2 | 3273 | Warn if a structure is given the packed attribute, but the packed |
02f52e19 | 3274 | attribute has no effect on the layout or size of the structure. |
3c12fcc2 GM |
3275 | Such structures may be mis-aligned for little benefit. For |
3276 | instance, in this code, the variable @code{f.x} in @code{struct bar} | |
3277 | will be misaligned even though @code{struct bar} does not itself | |
3278 | have the packed attribute: | |
3279 | ||
3280 | @smallexample | |
3281 | @group | |
3282 | struct foo @{ | |
3283 | int x; | |
3284 | char a, b, c, d; | |
3285 | @} __attribute__((packed)); | |
3286 | struct bar @{ | |
3287 | char z; | |
3288 | struct foo f; | |
3289 | @}; | |
3290 | @end group | |
3291 | @end smallexample | |
3292 | ||
3293 | @item -Wpadded | |
cd3bb277 | 3294 | @opindex Wpadded |
3c12fcc2 GM |
3295 | Warn if padding is included in a structure, either to align an element |
3296 | of the structure or to align the whole structure. Sometimes when this | |
3297 | happens it is possible to rearrange the fields of the structure to | |
3298 | reduce the padding and so make the structure smaller. | |
3299 | ||
74291a4b | 3300 | @item -Wredundant-decls |
cd3bb277 | 3301 | @opindex Wredundant-decls |
74291a4b MM |
3302 | Warn if anything is declared more than once in the same scope, even in |
3303 | cases where multiple declaration is valid and changes nothing. | |
3304 | ||
aee96fe9 | 3305 | @item -Wnested-externs @r{(C only)} |
cd3bb277 | 3306 | @opindex Wnested-externs |
252215a7 | 3307 | Warn if an @code{extern} declaration is encountered within a function. |
74291a4b | 3308 | |
312f6255 | 3309 | @item -Wunreachable-code |
cd3bb277 | 3310 | @opindex Wunreachable-code |
312f6255 GK |
3311 | Warn if the compiler detects that code will never be executed. |
3312 | ||
3313 | This option is intended to warn when the compiler detects that at | |
3314 | least a whole line of source code will never be executed, because | |
3315 | some condition is never satisfied or because it is after a | |
3316 | procedure that never returns. | |
3317 | ||
3318 | It is possible for this option to produce a warning even though there | |
3319 | are circumstances under which part of the affected line can be executed, | |
3320 | so care should be taken when removing apparently-unreachable code. | |
3321 | ||
3322 | For instance, when a function is inlined, a warning may mean that the | |
02f52e19 | 3323 | line is unreachable in only one inlined copy of the function. |
312f6255 | 3324 | |
630d3d5a | 3325 | This option is not made part of @option{-Wall} because in a debugging |
312f6255 GK |
3326 | version of a program there is often substantial code which checks |
3327 | correct functioning of the program and is, hopefully, unreachable | |
3328 | because the program does work. Another common use of unreachable | |
c21cd8b1 | 3329 | code is to provide behavior which is selectable at compile-time. |
312f6255 | 3330 | |
74291a4b | 3331 | @item -Winline |
cd3bb277 | 3332 | @opindex Winline |
c5c76735 | 3333 | Warn if a function can not be inlined and it was declared as inline. |
ae4a7155 | 3334 | Even with this option, the compiler will not warn about failures to |
daf2f129 | 3335 | inline functions declared in system headers. |
ae4a7155 MM |
3336 | |
3337 | The compiler uses a variety of heuristics to determine whether or not | |
3338 | to inline a function. For example, the compiler takes into account | |
0bdcd332 | 3339 | the size of the function being inlined and the amount of inlining |
ae4a7155 MM |
3340 | that has already been done in the current function. Therefore, |
3341 | seemingly insignificant changes in the source program can cause the | |
3342 | warnings produced by @option{-Winline} to appear or disappear. | |
74291a4b | 3343 | |
a01fff59 MA |
3344 | @item -Wno-invalid-offsetof @r{(C++ only)} |
3345 | @opindex Wno-invalid-offsetof | |
3346 | Suppress warnings from applying the @samp{offsetof} macro to a non-POD | |
3347 | type. According to the 1998 ISO C++ standard, applying @samp{offsetof} | |
3348 | to a non-POD type is undefined. In existing C++ implementations, | |
3349 | however, @samp{offsetof} typically gives meaningful results even when | |
3350 | applied to certain kinds of non-POD types. (Such as a simple | |
3351 | @samp{struct} that fails to be a POD type only by virtue of having a | |
3352 | constructor.) This flag is for users who are aware that they are | |
3353 | writing nonportable code and who have deliberately chosen to ignore the | |
3354 | warning about it. | |
3355 | ||
3356 | The restrictions on @samp{offsetof} may be relaxed in a future version | |
3357 | of the C++ standard. | |
3358 | ||
53a2494e JM |
3359 | @item -Wno-int-to-pointer-cast @r{(C only)} |
3360 | @opindex Wno-int-to-pointer-cast | |
3361 | Suppress warnings from casts to pointer type of an integer of a | |
3362 | different size. | |
3363 | ||
3364 | @item -Wno-pointer-to-int-cast @r{(C only)} | |
3365 | @opindex Wno-pointer-to-int-cast | |
3366 | Suppress warnings from casts from a pointer to an integer type of a | |
3367 | different size. | |
3368 | ||
17211ab5 GK |
3369 | @item -Winvalid-pch |
3370 | @opindex Winvalid-pch | |
3371 | Warn if a precompiled header (@pxref{Precompiled Headers}) is found in | |
3372 | the search path but can't be used. | |
3373 | ||
795add94 | 3374 | @item -Wlong-long |
cd3bb277 JM |
3375 | @opindex Wlong-long |
3376 | @opindex Wno-long-long | |
795add94 | 3377 | Warn if @samp{long long} type is used. This is default. To inhibit |
630d3d5a JM |
3378 | the warning messages, use @option{-Wno-long-long}. Flags |
3379 | @option{-Wlong-long} and @option{-Wno-long-long} are taken into account | |
3380 | only when @option{-pedantic} flag is used. | |
795add94 | 3381 | |
7c4d376d RH |
3382 | @item -Wvariadic-macros |
3383 | @opindex Wvariadic-macros | |
3384 | @opindex Wno-variadic-macros | |
3385 | Warn if variadic macros are used in pedantic ISO C90 mode, or the GNU | |
3386 | alternate syntax when in pedantic ISO C99 mode. This is default. | |
3387 | To inhibit the warning messages, use @option{-Wno-variadic-macros}. | |
3388 | ||
d35a40fc DE |
3389 | @item -Wvolatile-register-var |
3390 | @opindex Wvolatile-register-var | |
3391 | @opindex Wno-volatile-register-var | |
3392 | Warn if a register variable is declared volatile. The volatile | |
3393 | modifier does not inhibit all optimizations that may eliminate reads | |
3394 | and/or writes to register variables. | |
3395 | ||
18424ae1 | 3396 | @item -Wdisabled-optimization |
cd3bb277 | 3397 | @opindex Wdisabled-optimization |
18424ae1 BL |
3398 | Warn if a requested optimization pass is disabled. This warning does |
3399 | not generally indicate that there is anything wrong with your code; it | |
3400 | merely indicates that GCC's optimizers were unable to handle the code | |
3401 | effectively. Often, the problem is that your code is too big or too | |
3402 | complex; GCC will refuse to optimize programs when the optimization | |
3403 | itself is likely to take inordinate amounts of time. | |
3404 | ||
f4e9414e AO |
3405 | @item -Wpointer-sign |
3406 | @opindex Wpointer-sign | |
f2fd3821 | 3407 | @opindex Wno-pointer-sign |
f4e9414e AO |
3408 | Warn for pointer argument passing or assignment with different signedness. |
3409 | This option is only supported for C and Objective-C@. It is implied by | |
3410 | @option{-Wall} and by @option{-pedantic}, which can be disabled with | |
3411 | @option{-Wno-pointer-sign}. | |
f2fd3821 | 3412 | |
74291a4b | 3413 | @item -Werror |
cd3bb277 | 3414 | @opindex Werror |
74291a4b | 3415 | Make all warnings into errors. |
0aca9021 | 3416 | |
79cf5994 DD |
3417 | @item -Werror= |
3418 | @opindex Werror= | |
3419 | Make the specified warning into an errors. The specifier for a | |
3420 | warning is appended, for example @option{-Werror=switch} turns the | |
3421 | warnings controlled by @option{-Wswitch} into errors. This switch | |
3422 | takes a negative form, to be used to negate @option{-Werror} for | |
3423 | specific warnings, for example @option{-Wno-error=switch} makes | |
3424 | @option{-Wswitch} warnings not be errors, even when @option{-Werror} | |
3425 | is in effect. You can use the @option{-fdiagnostics-show-option} | |
3426 | option to have each controllable warning amended with the option which | |
3427 | controls it, to determine what to use with this option. | |
3428 | ||
3429 | Note that specifying @option{-Werror=}@var{foo} automatically implies | |
3430 | @option{-W}@var{foo}. However, @option{-Wno-error=}@var{foo} does not | |
3431 | imply anything. | |
3432 | ||
0aca9021 | 3433 | @item -Wstack-protector |
d77314ec | 3434 | @opindex Wstack-protector |
0aca9021 JW |
3435 | This option is only active when @option{-fstack-protector} is active. It |
3436 | warns about functions that will not be protected against stack smashing. | |
3437 | ||
d77314ec RS |
3438 | @item -Wstring-literal-comparison |
3439 | @opindex Wstring-literal-comparison | |
3440 | Warn about suspicious comparisons to string literal constants. In C, | |
3441 | direct comparisons against the memory address of a string literal, such | |
3442 | as @code{if (x == "abc")}, typically indicate a programmer error, and | |
3443 | even when intentional, result in unspecified behavior and are not portable. | |
3444 | Usually these warnings alert that the programmer intended to use | |
3445 | @code{strcmp}. This warning is enabled by @option{-Wall}. | |
3446 | ||
74291a4b MM |
3447 | @end table |
3448 | ||
3449 | @node Debugging Options | |
0c2d1a2a | 3450 | @section Options for Debugging Your Program or GCC |
74291a4b MM |
3451 | @cindex options, debugging |
3452 | @cindex debugging information options | |
3453 | ||
0c2d1a2a | 3454 | GCC has various special options that are used for debugging |
74291a4b MM |
3455 | either your program or GCC: |
3456 | ||
2642624b | 3457 | @table @gcctabopt |
74291a4b | 3458 | @item -g |
cd3bb277 | 3459 | @opindex g |
74291a4b | 3460 | Produce debugging information in the operating system's native format |
f8ca7e49 | 3461 | (stabs, COFF, XCOFF, or DWARF 2)@. GDB can work with this debugging |
74291a4b MM |
3462 | information. |
3463 | ||
630d3d5a | 3464 | On most systems that use stabs format, @option{-g} enables use of extra |
74291a4b MM |
3465 | debugging information that only GDB can use; this extra information |
3466 | makes debugging work better in GDB but will probably make other debuggers | |
3467 | crash or | |
3468 | refuse to read the program. If you want to control for certain whether | |
630d3d5a | 3469 | to generate the extra information, use @option{-gstabs+}, @option{-gstabs}, |
def66b10 | 3470 | @option{-gxcoff+}, @option{-gxcoff}, or @option{-gvms} (see below). |
74291a4b | 3471 | |
f8ca7e49 | 3472 | GCC allows you to use @option{-g} with |
630d3d5a | 3473 | @option{-O}. The shortcuts taken by optimized code may occasionally |
74291a4b MM |
3474 | produce surprising results: some variables you declared may not exist |
3475 | at all; flow of control may briefly move where you did not expect it; | |
3476 | some statements may not be executed because they compute constant | |
3477 | results or their values were already at hand; some statements may | |
3478 | execute in different places because they were moved out of loops. | |
3479 | ||
3480 | Nevertheless it proves possible to debug optimized output. This makes | |
3481 | it reasonable to use the optimizer for programs that might have bugs. | |
3482 | ||
0c2d1a2a | 3483 | The following options are useful when GCC is generated with the |
74291a4b MM |
3484 | capability for more than one debugging format. |
3485 | ||
3486 | @item -ggdb | |
cd3bb277 | 3487 | @opindex ggdb |
161d7b59 | 3488 | Produce debugging information for use by GDB@. This means to use the |
861bb6c1 JL |
3489 | most expressive format available (DWARF 2, stabs, or the native format |
3490 | if neither of those are supported), including GDB extensions if at all | |
3491 | possible. | |
74291a4b MM |
3492 | |
3493 | @item -gstabs | |
cd3bb277 | 3494 | @opindex gstabs |
74291a4b MM |
3495 | Produce debugging information in stabs format (if that is supported), |
3496 | without GDB extensions. This is the format used by DBX on most BSD | |
3497 | systems. On MIPS, Alpha and System V Release 4 systems this option | |
161d7b59 | 3498 | produces stabs debugging output which is not understood by DBX or SDB@. |
74291a4b MM |
3499 | On System V Release 4 systems this option requires the GNU assembler. |
3500 | ||
6a08f7b3 DP |
3501 | @item -feliminate-unused-debug-symbols |
3502 | @opindex feliminate-unused-debug-symbols | |
3503 | Produce debugging information in stabs format (if that is supported), | |
c0cbdbd9 | 3504 | for only symbols that are actually used. |
6a08f7b3 | 3505 | |
74291a4b | 3506 | @item -gstabs+ |
cd3bb277 | 3507 | @opindex gstabs+ |
74291a4b | 3508 | Produce debugging information in stabs format (if that is supported), |
161d7b59 | 3509 | using GNU extensions understood only by the GNU debugger (GDB)@. The |
74291a4b MM |
3510 | use of these extensions is likely to make other debuggers crash or |
3511 | refuse to read the program. | |
3512 | ||
3513 | @item -gcoff | |
cd3bb277 | 3514 | @opindex gcoff |
74291a4b MM |
3515 | Produce debugging information in COFF format (if that is supported). |
3516 | This is the format used by SDB on most System V systems prior to | |
3517 | System V Release 4. | |
3518 | ||
3519 | @item -gxcoff | |
cd3bb277 | 3520 | @opindex gxcoff |
74291a4b MM |
3521 | Produce debugging information in XCOFF format (if that is supported). |
3522 | This is the format used by the DBX debugger on IBM RS/6000 systems. | |
3523 | ||
3524 | @item -gxcoff+ | |
cd3bb277 | 3525 | @opindex gxcoff+ |
74291a4b | 3526 | Produce debugging information in XCOFF format (if that is supported), |
161d7b59 | 3527 | using GNU extensions understood only by the GNU debugger (GDB)@. The |
74291a4b MM |
3528 | use of these extensions is likely to make other debuggers crash or |
3529 | refuse to read the program, and may cause assemblers other than the GNU | |
3530 | assembler (GAS) to fail with an error. | |
3531 | ||
861bb6c1 | 3532 | @item -gdwarf-2 |
cd3bb277 | 3533 | @opindex gdwarf-2 |
861bb6c1 | 3534 | Produce debugging information in DWARF version 2 format (if that is |
f8ca7e49 ZW |
3535 | supported). This is the format used by DBX on IRIX 6. With this |
3536 | option, GCC uses features of DWARF version 3 when they are useful; | |
3537 | version 3 is upward compatible with version 2, but may still cause | |
3538 | problems for older debuggers. | |
74291a4b | 3539 | |
5f98259a RK |
3540 | @item -gvms |
3541 | @opindex gvms | |
3542 | Produce debugging information in VMS debug format (if that is | |
3543 | supported). This is the format used by DEBUG on VMS systems. | |
3544 | ||
74291a4b MM |
3545 | @item -g@var{level} |
3546 | @itemx -ggdb@var{level} | |
3547 | @itemx -gstabs@var{level} | |
3548 | @itemx -gcoff@var{level} | |
3549 | @itemx -gxcoff@var{level} | |
5f98259a | 3550 | @itemx -gvms@var{level} |
74291a4b MM |
3551 | Request debugging information and also use @var{level} to specify how |
3552 | much information. The default level is 2. | |
3553 | ||
3554 | Level 1 produces minimal information, enough for making backtraces in | |
3555 | parts of the program that you don't plan to debug. This includes | |
3556 | descriptions of functions and external variables, but no information | |
3557 | about local variables and no line numbers. | |
3558 | ||
3559 | Level 3 includes extra information, such as all the macro definitions | |
3560 | present in the program. Some debuggers support macro expansion when | |
630d3d5a | 3561 | you use @option{-g3}. |
74291a4b | 3562 | |
f8ca7e49 ZW |
3563 | @option{-gdwarf-2} does not accept a concatenated debug level, because |
3564 | GCC used to support an option @option{-gdwarf} that meant to generate | |
3565 | debug information in version 1 of the DWARF format (which is very | |
3566 | different from version 2), and it would have been too confusing. That | |
3567 | debug format is long obsolete, but the option cannot be changed now. | |
3568 | Instead use an additional @option{-g@var{level}} option to change the | |
3569 | debug level for DWARF2. | |
eb7715a4 | 3570 | |
e03b7153 RS |
3571 | @item -feliminate-dwarf2-dups |
3572 | @opindex feliminate-dwarf2-dups | |
3573 | Compress DWARF2 debugging information by eliminating duplicated | |
3574 | information about each symbol. This option only makes sense when | |
3575 | generating DWARF2 debugging information with @option{-gdwarf-2}. | |
3576 | ||
05739753 | 3577 | @cindex @command{prof} |
74291a4b | 3578 | @item -p |
cd3bb277 | 3579 | @opindex p |
74291a4b | 3580 | Generate extra code to write profile information suitable for the |
05739753 | 3581 | analysis program @command{prof}. You must use this option when compiling |
74291a4b MM |
3582 | the source files you want data about, and you must also use it when |
3583 | linking. | |
3584 | ||
05739753 | 3585 | @cindex @command{gprof} |
74291a4b | 3586 | @item -pg |
cd3bb277 | 3587 | @opindex pg |
74291a4b | 3588 | Generate extra code to write profile information suitable for the |
05739753 | 3589 | analysis program @command{gprof}. You must use this option when compiling |
74291a4b MM |
3590 | the source files you want data about, and you must also use it when |
3591 | linking. | |
3592 | ||
898f531b | 3593 | @item -Q |
cd3bb277 | 3594 | @opindex Q |
898f531b JL |
3595 | Makes the compiler print out each function name as it is compiled, and |
3596 | print some statistics about each pass when it finishes. | |
3597 | ||
1f0c3120 | 3598 | @item -ftime-report |
cd3bb277 | 3599 | @opindex ftime-report |
1f0c3120 JM |
3600 | Makes the compiler print some statistics about the time consumed by each |
3601 | pass when it finishes. | |
3602 | ||
3603 | @item -fmem-report | |
cd3bb277 | 3604 | @opindex fmem-report |
1f0c3120 JM |
3605 | Makes the compiler print some statistics about permanent memory |
3606 | allocation when it finishes. | |
3607 | ||
861bb6c1 | 3608 | @item -fprofile-arcs |
cd3bb277 | 3609 | @opindex fprofile-arcs |
23af32e6 NS |
3610 | Add code so that program flow @dfn{arcs} are instrumented. During |
3611 | execution the program records how many times each branch and call is | |
3612 | executed and how many times it is taken or returns. When the compiled | |
3613 | program exits it saves this data to a file called | |
8a36672b | 3614 | @file{@var{auxname}.gcda} for each source file. The data may be used for |
23af32e6 | 3615 | profile-directed optimizations (@option{-fbranch-probabilities}), or for |
8a36672b | 3616 | test coverage analysis (@option{-ftest-coverage}). Each object file's |
23af32e6 NS |
3617 | @var{auxname} is generated from the name of the output file, if |
3618 | explicitly specified and it is not the final executable, otherwise it is | |
8a36672b | 3619 | the basename of the source file. In both cases any suffix is removed |
431ae0bf | 3620 | (e.g.@: @file{foo.gcda} for input file @file{dir/foo.c}, or |
a4878735 | 3621 | @file{dir/foo.gcda} for output file specified as @option{-o dir/foo.o}). |
992f396f | 3622 | @xref{Cross-profiling}. |
23af32e6 | 3623 | |
ee4c708e BE |
3624 | @cindex @command{gcov} |
3625 | @item --coverage | |
3626 | @opindex coverage | |
3627 | ||
3628 | This option is used to compile and link code instrumented for coverage | |
3629 | analysis. The option is a synonym for @option{-fprofile-arcs} | |
3630 | @option{-ftest-coverage} (when compiling) and @option{-lgcov} (when | |
3631 | linking). See the documentation for those options for more details. | |
3632 | ||
23af32e6 NS |
3633 | @itemize |
3634 | ||
3635 | @item | |
3636 | Compile the source files with @option{-fprofile-arcs} plus optimization | |
8a36672b JM |
3637 | and code generation options. For test coverage analysis, use the |
3638 | additional @option{-ftest-coverage} option. You do not need to profile | |
23af32e6 NS |
3639 | every source file in a program. |
3640 | ||
3641 | @item | |
8555daff NS |
3642 | Link your object files with @option{-lgcov} or @option{-fprofile-arcs} |
3643 | (the latter implies the former). | |
23af32e6 NS |
3644 | |
3645 | @item | |
3646 | Run the program on a representative workload to generate the arc profile | |
8a36672b | 3647 | information. This may be repeated any number of times. You can run |
8555daff | 3648 | concurrent instances of your program, and provided that the file system |
8a36672b | 3649 | supports locking, the data files will be correctly updated. Also |
8555daff NS |
3650 | @code{fork} calls are detected and correctly handled (double counting |
3651 | will not happen). | |
23af32e6 NS |
3652 | |
3653 | @item | |
3654 | For profile-directed optimizations, compile the source files again with | |
3655 | the same optimization and code generation options plus | |
630d3d5a | 3656 | @option{-fbranch-probabilities} (@pxref{Optimize Options,,Options that |
3de87bf2 JJ |
3657 | Control Optimization}). |
3658 | ||
23af32e6 NS |
3659 | @item |
3660 | For test coverage analysis, use @command{gcov} to produce human readable | |
8a36672b | 3661 | information from the @file{.gcno} and @file{.gcda} files. Refer to the |
23af32e6 NS |
3662 | @command{gcov} documentation for further information. |
3663 | ||
3664 | @end itemize | |
3de87bf2 JJ |
3665 | |
3666 | With @option{-fprofile-arcs}, for each function of your program GCC | |
3667 | creates a program flow graph, then finds a spanning tree for the graph. | |
3668 | Only arcs that are not on the spanning tree have to be instrumented: the | |
3669 | compiler adds code to count the number of times that these arcs are | |
3670 | executed. When an arc is the only exit or only entrance to a block, the | |
3671 | instrumentation code can be added to the block; otherwise, a new basic | |
3672 | block must be created to hold the instrumentation code. | |
3673 | ||
861bb6c1 JL |
3674 | @need 2000 |
3675 | @item -ftest-coverage | |
cd3bb277 | 3676 | @opindex ftest-coverage |
a4878735 | 3677 | Produce a notes file that the @command{gcov} code-coverage utility |
23af32e6 | 3678 | (@pxref{Gcov,, @command{gcov}---a Test Coverage Program}) can use to |
8a36672b JM |
3679 | show program coverage. Each source file's note file is called |
3680 | @file{@var{auxname}.gcno}. Refer to the @option{-fprofile-arcs} option | |
23af32e6 | 3681 | above for a description of @var{auxname} and instructions on how to |
8a36672b | 3682 | generate test coverage data. Coverage data will match the source files |
23af32e6 | 3683 | more closely, if you do not optimize. |
3de87bf2 | 3684 | |
74291a4b | 3685 | @item -d@var{letters} |
9f8628ba | 3686 | @item -fdump-rtl-@var{pass} |
cd3bb277 | 3687 | @opindex d |
74291a4b | 3688 | Says to make debugging dumps during compilation at times specified by |
9f8628ba PB |
3689 | @var{letters}. This is used for debugging the RTL-based passes of the |
3690 | compiler. The file names for most of the dumps are made by appending a | |
3691 | pass number and a word to the @var{dumpname}. @var{dumpname} is generated | |
3692 | from the name of the output file, if explicitly specified and it is not | |
3693 | an executable, otherwise it is the basename of the source file. | |
74291a4b | 3694 | |
9f8628ba PB |
3695 | Most debug dumps can be enabled either passing a letter to the @option{-d} |
3696 | option, or with a long @option{-fdump-rtl} switch; here are the possible | |
3697 | letters for use in @var{letters} and @var{pass}, and their meanings: | |
3698 | ||
3699 | @table @gcctabopt | |
3700 | @item -dA | |
cd3bb277 | 3701 | @opindex dA |
375e2d5c | 3702 | Annotate the assembler output with miscellaneous debugging information. |
9f8628ba PB |
3703 | |
3704 | @item -db | |
3705 | @itemx -fdump-rtl-bp | |
cd3bb277 | 3706 | @opindex db |
9f8628ba PB |
3707 | @opindex fdump-rtl-bp |
3708 | Dump after computing branch probabilities, to @file{@var{file}.09.bp}. | |
3709 | ||
3710 | @item -dB | |
3711 | @itemx -fdump-rtl-bbro | |
cd3bb277 | 3712 | @opindex dB |
9f8628ba PB |
3713 | @opindex fdump-rtl-bbro |
3714 | Dump after block reordering, to @file{@var{file}.30.bbro}. | |
3715 | ||
3716 | @item -dc | |
3717 | @itemx -fdump-rtl-combine | |
cd3bb277 | 3718 | @opindex dc |
9f8628ba PB |
3719 | @opindex fdump-rtl-combine |
3720 | Dump after instruction combination, to the file @file{@var{file}.17.combine}. | |
3721 | ||
3722 | @item -dC | |
3723 | @itemx -fdump-rtl-ce1 | |
3724 | @itemx -fdump-rtl-ce2 | |
cd3bb277 | 3725 | @opindex dC |
9f8628ba PB |
3726 | @opindex fdump-rtl-ce1 |
3727 | @opindex fdump-rtl-ce2 | |
3728 | @option{-dC} and @option{-fdump-rtl-ce1} enable dumping after the | |
3729 | first if conversion, to the file @file{@var{file}.11.ce1}. @option{-dC} | |
3730 | and @option{-fdump-rtl-ce2} enable dumping after the second if | |
3731 | conversion, to the file @file{@var{file}.18.ce2}. | |
3732 | ||
3733 | @item -dd | |
3734 | @itemx -fdump-rtl-btl | |
3735 | @itemx -fdump-rtl-dbr | |
cd3bb277 | 3736 | @opindex dd |
9f8628ba PB |
3737 | @opindex fdump-rtl-btl |
3738 | @opindex fdump-rtl-dbr | |
3739 | @option{-dd} and @option{-fdump-rtl-btl} enable dumping after branch | |
0bdcd332 | 3740 | target load optimization, to @file{@var{file}.31.btl}. @option{-dd} |
9f8628ba PB |
3741 | and @option{-fdump-rtl-dbr} enable dumping after delayed branch |
3742 | scheduling, to @file{@var{file}.36.dbr}. | |
3743 | ||
3744 | @item -dD | |
cd3bb277 | 3745 | @opindex dD |
f5963e61 JL |
3746 | Dump all macro definitions, at the end of preprocessing, in addition to |
3747 | normal output. | |
9f8628ba PB |
3748 | |
3749 | @item -dE | |
3750 | @itemx -fdump-rtl-ce3 | |
cd3bb277 | 3751 | @opindex dE |
9f8628ba PB |
3752 | @opindex fdump-rtl-ce3 |
3753 | Dump after the third if conversion, to @file{@var{file}.28.ce3}. | |
3754 | ||
3755 | @item -df | |
3756 | @itemx -fdump-rtl-cfg | |
3757 | @itemx -fdump-rtl-life | |
cd3bb277 | 3758 | @opindex df |
9f8628ba PB |
3759 | @opindex fdump-rtl-cfg |
3760 | @opindex fdump-rtl-life | |
3761 | @option{-df} and @option{-fdump-rtl-cfg} enable dumping after control | |
3762 | and data flow analysis, to @file{@var{file}.08.cfg}. @option{-df} | |
3763 | and @option{-fdump-rtl-cfg} enable dumping dump after life analysis, | |
3764 | to @file{@var{file}.16.life}. | |
3765 | ||
3766 | @item -dg | |
3767 | @itemx -fdump-rtl-greg | |
cd3bb277 | 3768 | @opindex dg |
9f8628ba PB |
3769 | @opindex fdump-rtl-greg |
3770 | Dump after global register allocation, to @file{@var{file}.23.greg}. | |
3771 | ||
3772 | @item -dG | |
3773 | @itemx -fdump-rtl-gcse | |
3774 | @itemx -fdump-rtl-bypass | |
cd3bb277 | 3775 | @opindex dG |
9f8628ba PB |
3776 | @opindex fdump-rtl-gcse |
3777 | @opindex fdump-rtl-bypass | |
3778 | @option{-dG} and @option{-fdump-rtl-gcse} enable dumping after GCSE, to | |
3779 | @file{@var{file}.05.gcse}. @option{-dG} and @option{-fdump-rtl-bypass} | |
3780 | enable dumping after jump bypassing and control flow optimizations, to | |
3781 | @file{@var{file}.07.bypass}. | |
3782 | ||
3783 | @item -dh | |
3784 | @itemx -fdump-rtl-eh | |
7fedea11 | 3785 | @opindex dh |
9f8628ba PB |
3786 | @opindex fdump-rtl-eh |
3787 | Dump after finalization of EH handling code, to @file{@var{file}.02.eh}. | |
3788 | ||
3789 | @item -di | |
3790 | @itemx -fdump-rtl-sibling | |
cd3bb277 | 3791 | @opindex di |
9f8628ba PB |
3792 | @opindex fdump-rtl-sibling |
3793 | Dump after sibling call optimizations, to @file{@var{file}.01.sibling}. | |
3794 | ||
3795 | @item -dj | |
3796 | @itemx -fdump-rtl-jump | |
cd3bb277 | 3797 | @opindex dj |
9f8628ba PB |
3798 | @opindex fdump-rtl-jump |
3799 | Dump after the first jump optimization, to @file{@var{file}.03.jump}. | |
3800 | ||
3801 | @item -dk | |
3802 | @itemx -fdump-rtl-stack | |
cd3bb277 | 3803 | @opindex dk |
9f8628ba PB |
3804 | @opindex fdump-rtl-stack |
3805 | Dump after conversion from registers to stack, to @file{@var{file}.33.stack}. | |
3806 | ||
3807 | @item -dl | |
3808 | @itemx -fdump-rtl-lreg | |
cd3bb277 | 3809 | @opindex dl |
9f8628ba PB |
3810 | @opindex fdump-rtl-lreg |
3811 | Dump after local register allocation, to @file{@var{file}.22.lreg}. | |
3812 | ||
3813 | @item -dL | |
3814 | @itemx -fdump-rtl-loop | |
3815 | @itemx -fdump-rtl-loop2 | |
cd3bb277 | 3816 | @opindex dL |
9f8628ba PB |
3817 | @opindex fdump-rtl-loop |
3818 | @opindex fdump-rtl-loop2 | |
3819 | @option{-dL} and @option{-fdump-rtl-loop} enable dumping after the first | |
3820 | loop optimization pass, to @file{@var{file}.06.loop}. @option{-dL} and | |
3821 | @option{-fdump-rtl-loop2} enable dumping after the second pass, to | |
3822 | @file{@var{file}.13.loop2}. | |
3823 | ||
3824 | @item -dm | |
3825 | @itemx -fdump-rtl-sms | |
e5626198 | 3826 | @opindex dm |
9f8628ba | 3827 | @opindex fdump-rtl-sms |
f0eb93a8 | 3828 | Dump after modulo scheduling, to @file{@var{file}.20.sms}. |
9f8628ba PB |
3829 | |
3830 | @item -dM | |
3831 | @itemx -fdump-rtl-mach | |
cd3bb277 | 3832 | @opindex dM |
9f8628ba | 3833 | @opindex fdump-rtl-mach |
c0478a66 | 3834 | Dump after performing the machine dependent reorganization pass, to |
9f8628ba PB |
3835 | @file{@var{file}.35.mach}. |
3836 | ||
3837 | @item -dn | |
3838 | @itemx -fdump-rtl-rnreg | |
cd3bb277 | 3839 | @opindex dn |
9f8628ba PB |
3840 | @opindex fdump-rtl-rnreg |
3841 | Dump after register renumbering, to @file{@var{file}.29.rnreg}. | |
3842 | ||
3843 | @item -dN | |
3844 | @itemx -fdump-rtl-regmove | |
cd3bb277 | 3845 | @opindex dN |
9f8628ba PB |
3846 | @opindex fdump-rtl-regmove |
3847 | Dump after the register move pass, to @file{@var{file}.19.regmove}. | |
3848 | ||
3849 | @item -do | |
3850 | @itemx -fdump-rtl-postreload | |
7fedea11 | 3851 | @opindex do |
9f8628ba PB |
3852 | @opindex fdump-rtl-postreload |
3853 | Dump after post-reload optimizations, to @file{@var{file}.24.postreload}. | |
3854 | ||
3855 | @item -dr | |
3856 | @itemx -fdump-rtl-expand | |
cd3bb277 | 3857 | @opindex dr |
9f8628ba PB |
3858 | @opindex fdump-rtl-expand |
3859 | Dump after RTL generation, to @file{@var{file}.00.expand}. | |
3860 | ||
3861 | @item -dR | |
3862 | @itemx -fdump-rtl-sched2 | |
cd3bb277 | 3863 | @opindex dR |
9f8628ba PB |
3864 | @opindex fdump-rtl-sched2 |
3865 | Dump after the second scheduling pass, to @file{@var{file}.32.sched2}. | |
3866 | ||
3867 | @item -ds | |
3868 | @itemx -fdump-rtl-cse | |
cd3bb277 | 3869 | @opindex ds |
9f8628ba | 3870 | @opindex fdump-rtl-cse |
032713aa | 3871 | Dump after CSE (including the jump optimization that sometimes follows |
9f8628ba PB |
3872 | CSE), to @file{@var{file}.04.cse}. |
3873 | ||
3874 | @item -dS | |
3875 | @itemx -fdump-rtl-sched | |
cd3bb277 | 3876 | @opindex dS |
9f8628ba PB |
3877 | @opindex fdump-rtl-sched |
3878 | Dump after the first scheduling pass, to @file{@var{file}.21.sched}. | |
3879 | ||
3880 | @item -dt | |
3881 | @itemx -fdump-rtl-cse2 | |
cd3bb277 | 3882 | @opindex dt |
9f8628ba | 3883 | @opindex fdump-rtl-cse2 |
032713aa | 3884 | Dump after the second CSE pass (including the jump optimization that |
9f8628ba PB |
3885 | sometimes follows CSE), to @file{@var{file}.15.cse2}. |
3886 | ||
3887 | @item -dT | |
3888 | @itemx -fdump-rtl-tracer | |
4319ef2a | 3889 | @opindex dT |
9f8628ba PB |
3890 | @opindex fdump-rtl-tracer |
3891 | Dump after running tracer, to @file{@var{file}.12.tracer}. | |
3892 | ||
3893 | @item -dV | |
3894 | @itemx -fdump-rtl-vpt | |
3895 | @itemx -fdump-rtl-vartrack | |
9313cfdd | 3896 | @opindex dV |
9f8628ba PB |
3897 | @opindex fdump-rtl-vpt |
3898 | @opindex fdump-rtl-vartrack | |
3899 | @option{-dV} and @option{-fdump-rtl-vpt} enable dumping after the value | |
3900 | profile transformations, to @file{@var{file}.10.vpt}. @option{-dV} | |
3901 | and @option{-fdump-rtl-vartrack} enable dumping after variable tracking, | |
3902 | to @file{@var{file}.34.vartrack}. | |
3903 | ||
3904 | @item -dw | |
3905 | @itemx -fdump-rtl-flow2 | |
cd3bb277 | 3906 | @opindex dw |
9f8628ba PB |
3907 | @opindex fdump-rtl-flow2 |
3908 | Dump after the second flow pass, to @file{@var{file}.26.flow2}. | |
3909 | ||
3910 | @item -dz | |
3911 | @itemx -fdump-rtl-peephole2 | |
cd3bb277 | 3912 | @opindex dz |
9f8628ba PB |
3913 | @opindex fdump-rtl-peephole2 |
3914 | Dump after the peephole pass, to @file{@var{file}.27.peephole2}. | |
3915 | ||
3916 | @item -dZ | |
3917 | @itemx -fdump-rtl-web | |
9313cfdd | 3918 | @opindex dZ |
9f8628ba PB |
3919 | @opindex fdump-rtl-web |
3920 | Dump after live range splitting, to @file{@var{file}.14.web}. | |
3921 | ||
3922 | @item -da | |
3923 | @itemx -fdump-rtl-all | |
cd3bb277 | 3924 | @opindex da |
9f8628ba | 3925 | @opindex fdump-rtl-all |
74291a4b | 3926 | Produce all the dumps listed above. |
9f8628ba PB |
3927 | |
3928 | @item -dH | |
886e0865 GK |
3929 | @opindex dH |
3930 | Produce a core dump whenever an error occurs. | |
9f8628ba PB |
3931 | |
3932 | @item -dm | |
cd3bb277 | 3933 | @opindex dm |
74291a4b MM |
3934 | Print statistics on memory usage, at the end of the run, to |
3935 | standard error. | |
9f8628ba PB |
3936 | |
3937 | @item -dp | |
cd3bb277 | 3938 | @opindex dp |
74291a4b | 3939 | Annotate the assembler output with a comment indicating which |
f20b5577 MM |
3940 | pattern and alternative was used. The length of each instruction is |
3941 | also printed. | |
9f8628ba PB |
3942 | |
3943 | @item -dP | |
cd3bb277 | 3944 | @opindex dP |
2856c3e3 | 3945 | Dump the RTL in the assembler output as a comment before each instruction. |
630d3d5a | 3946 | Also turns on @option{-dp} annotation. |
9f8628ba PB |
3947 | |
3948 | @item -dv | |
cd3bb277 | 3949 | @opindex dv |
9f8628ba PB |
3950 | For each of the other indicated dump files (either with @option{-d} or |
3951 | @option{-fdump-rtl-@var{pass}}), dump a representation of the control flow | |
3952 | graph suitable for viewing with VCG to @file{@var{file}.@var{pass}.vcg}. | |
3953 | ||
3954 | @item -dx | |
cd3bb277 | 3955 | @opindex dx |
62a1403d | 3956 | Just generate RTL for a function instead of compiling it. Usually used |
9f8628ba PB |
3957 | with @samp{r} (@option{-fdump-rtl-expand}). |
3958 | ||
3959 | @item -dy | |
cd3bb277 | 3960 | @opindex dy |
032713aa | 3961 | Dump debugging information during parsing, to standard error. |
74291a4b MM |
3962 | @end table |
3963 | ||
b707b450 | 3964 | @item -fdump-unnumbered |
cd3bb277 | 3965 | @opindex fdump-unnumbered |
695ac33f | 3966 | When doing debugging dumps (see @option{-d} option above), suppress instruction |
b707b450 | 3967 | numbers and line number note output. This makes it more feasible to |
b192711e | 3968 | use diff on debugging dumps for compiler invocations with different |
695ac33f | 3969 | options, in particular with and without @option{-g}. |
b707b450 | 3970 | |
223dcf1c AN |
3971 | @item -fdump-translation-unit @r{(C++ only)} |
3972 | @itemx -fdump-translation-unit-@var{options} @r{(C++ only)} | |
f70a54cb CR |
3973 | @opindex fdump-translation-unit |
3974 | Dump a representation of the tree structure for the entire translation | |
3975 | unit to a file. The file name is made by appending @file{.tu} to the | |
3976 | source file name. If the @samp{-@var{options}} form is used, @var{options} | |
3977 | controls the details of the dump as described for the | |
3978 | @option{-fdump-tree} options. | |
3979 | ||
aee96fe9 | 3980 | @item -fdump-class-hierarchy @r{(C++ only)} |
22367161 | 3981 | @itemx -fdump-class-hierarchy-@var{options} @r{(C++ only)} |
cd3bb277 | 3982 | @opindex fdump-class-hierarchy |
e76b4820 | 3983 | Dump a representation of each class's hierarchy and virtual function |
767094dd | 3984 | table layout to a file. The file name is made by appending @file{.class} |
22367161 NS |
3985 | to the source file name. If the @samp{-@var{options}} form is used, |
3986 | @var{options} controls the details of the dump as described for the | |
3987 | @option{-fdump-tree} options. | |
3988 | ||
9b3e897d PB |
3989 | @item -fdump-ipa-@var{switch} |
3990 | @opindex fdump-ipa | |
83c99486 | 3991 | Control the dumping at various stages of inter-procedural analysis |
9b3e897d | 3992 | language tree to a file. The file name is generated by appending a switch |
8a36672b | 3993 | specific suffix to the source file name. The following dumps are possible: |
9b3e897d PB |
3994 | |
3995 | @table @samp | |
3996 | @item all | |
3997 | Enables all inter-procedural analysis dumps; currently the only produced | |
3998 | dump is the @samp{cgraph} dump. | |
3999 | ||
4000 | @item cgraph | |
4001 | Dumps information about call-graph optimization, unused function removal, | |
4002 | and inlining decisions. | |
4003 | @end table | |
4004 | ||
0c8c236b JJ |
4005 | @item -fdump-tree-@var{switch} |
4006 | @itemx -fdump-tree-@var{switch}-@var{options} | |
22367161 NS |
4007 | @opindex fdump-tree |
4008 | Control the dumping at various stages of processing the intermediate | |
4009 | language tree to a file. The file name is generated by appending a switch | |
4010 | specific suffix to the source file name. If the @samp{-@var{options}} | |
4011 | form is used, @var{options} is a list of @samp{-} separated options that | |
8a36672b JM |
4012 | control the details of the dump. Not all options are applicable to all |
4013 | dumps, those which are not meaningful will be ignored. The following | |
22367161 | 4014 | options are available |
f71f87f9 | 4015 | |
e76b4820 | 4016 | @table @samp |
22367161 | 4017 | @item address |
767094dd | 4018 | Print the address of each node. Usually this is not meaningful as it |
8a36672b | 4019 | changes according to the environment and source file. Its primary use |
22367161 NS |
4020 | is for tying up a dump file with a debug environment. |
4021 | @item slim | |
4022 | Inhibit dumping of members of a scope or body of a function merely | |
6de9cd9a DN |
4023 | because that scope has been reached. Only dump such items when they |
4024 | are directly reachable by some other path. When dumping pretty-printed | |
4025 | trees, this option inhibits dumping the bodies of control structures. | |
4026 | @item raw | |
4027 | Print a raw representation of the tree. By default, trees are | |
4028 | pretty-printed into a C-like representation. | |
4029 | @item details | |
4030 | Enable more detailed dumps (not honored by every dump option). | |
4031 | @item stats | |
4032 | Enable dumping various statistics about the pass (not honored by every dump | |
4033 | option). | |
4034 | @item blocks | |
4035 | Enable showing basic block boundaries (disabled in raw dumps). | |
4036 | @item vops | |
4037 | Enable showing virtual operands for every statement. | |
4038 | @item lineno | |
4039 | Enable showing line numbers for statements. | |
4040 | @item uid | |
4041 | Enable showing the unique ID (@code{DECL_UID}) for each variable. | |
22367161 | 4042 | @item all |
6de9cd9a | 4043 | Turn on all options, except @option{raw}, @option{slim} and @option{lineno}. |
e76b4820 NS |
4044 | @end table |
4045 | ||
4046 | The following tree dumps are possible: | |
4047 | @table @samp | |
6de9cd9a | 4048 | |
e76b4820 NS |
4049 | @item original |
4050 | Dump before any tree based optimization, to @file{@var{file}.original}. | |
6de9cd9a | 4051 | |
e76b4820 NS |
4052 | @item optimized |
4053 | Dump after all tree based optimization, to @file{@var{file}.optimized}. | |
6de9cd9a | 4054 | |
6be77748 | 4055 | @item inlined |
9c34dbbf | 4056 | Dump after function inlining, to @file{@var{file}.inlined}. |
6de9cd9a DN |
4057 | |
4058 | @item gimple | |
4059 | @opindex fdump-tree-gimple | |
4060 | Dump each function before and after the gimplification pass to a file. The | |
4061 | file name is made by appending @file{.gimple} to the source file name. | |
4062 | ||
4063 | @item cfg | |
4064 | @opindex fdump-tree-cfg | |
4065 | Dump the control flow graph of each function to a file. The file name is | |
4066 | made by appending @file{.cfg} to the source file name. | |
4067 | ||
4068 | @item vcg | |
4069 | @opindex fdump-tree-vcg | |
4070 | Dump the control flow graph of each function to a file in VCG format. The | |
4071 | file name is made by appending @file{.vcg} to the source file name. Note | |
4072 | that if the file contains more than one function, the generated file cannot | |
8a36672b | 4073 | be used directly by VCG@. You will need to cut and paste each function's |
6de9cd9a DN |
4074 | graph into its own separate file first. |
4075 | ||
4076 | @item ch | |
4077 | @opindex fdump-tree-ch | |
4078 | Dump each function after copying loop headers. The file name is made by | |
4079 | appending @file{.ch} to the source file name. | |
4080 | ||
4081 | @item ssa | |
4082 | @opindex fdump-tree-ssa | |
4083 | Dump SSA related information to a file. The file name is made by appending | |
4084 | @file{.ssa} to the source file name. | |
4085 | ||
c75ab022 DB |
4086 | @item salias |
4087 | @opindex fdump-tree-salias | |
4088 | Dump structure aliasing variable information to a file. This file name | |
4089 | is made by appending @file{.salias} to the source file name. | |
4090 | ||
6de9cd9a DN |
4091 | @item alias |
4092 | @opindex fdump-tree-alias | |
4093 | Dump aliasing information for each function. The file name is made by | |
4094 | appending @file{.alias} to the source file name. | |
4095 | ||
4096 | @item ccp | |
4097 | @opindex fdump-tree-ccp | |
8a36672b | 4098 | Dump each function after CCP@. The file name is made by appending |
6de9cd9a DN |
4099 | @file{.ccp} to the source file name. |
4100 | ||
0bca51f0 DN |
4101 | @item storeccp |
4102 | @opindex fdump-tree-storeccp | |
4103 | Dump each function after STORE-CCP. The file name is made by appending | |
4104 | @file{.storeccp} to the source file name. | |
4105 | ||
6de9cd9a DN |
4106 | @item pre |
4107 | @opindex fdump-tree-pre | |
4108 | Dump trees after partial redundancy elimination. The file name is made | |
4109 | by appending @file{.pre} to the source file name. | |
4110 | ||
ff2ad0f7 DN |
4111 | @item fre |
4112 | @opindex fdump-tree-fre | |
4113 | Dump trees after full redundancy elimination. The file name is made | |
4114 | by appending @file{.fre} to the source file name. | |
4115 | ||
0bca51f0 DN |
4116 | @item copyprop |
4117 | @opindex fdump-tree-copyprop | |
4118 | Dump trees after copy propagation. The file name is made | |
4119 | by appending @file{.copyprop} to the source file name. | |
4120 | ||
4121 | @item store_copyprop | |
4122 | @opindex fdump-tree-store_copyprop | |
4123 | Dump trees after store copy-propagation. The file name is made | |
4124 | by appending @file{.store_copyprop} to the source file name. | |
4125 | ||
6de9cd9a DN |
4126 | @item dce |
4127 | @opindex fdump-tree-dce | |
4128 | Dump each function after dead code elimination. The file name is made by | |
4129 | appending @file{.dce} to the source file name. | |
4130 | ||
4131 | @item mudflap | |
4132 | @opindex fdump-tree-mudflap | |
4133 | Dump each function after adding mudflap instrumentation. The file name is | |
4134 | made by appending @file{.mudflap} to the source file name. | |
4135 | ||
4136 | @item sra | |
4137 | @opindex fdump-tree-sra | |
4138 | Dump each function after performing scalar replacement of aggregates. The | |
4139 | file name is made by appending @file{.sra} to the source file name. | |
4140 | ||
fa555252 DB |
4141 | @item sink |
4142 | @opindex fdump-tree-sink | |
4143 | Dump each function after performing code sinking. The file name is made | |
4144 | by appending @file{.sink} to the source file name. | |
4145 | ||
6de9cd9a DN |
4146 | @item dom |
4147 | @opindex fdump-tree-dom | |
4148 | Dump each function after applying dominator tree optimizations. The file | |
4149 | name is made by appending @file{.dom} to the source file name. | |
4150 | ||
4151 | @item dse | |
4152 | @opindex fdump-tree-dse | |
4153 | Dump each function after applying dead store elimination. The file | |
4154 | name is made by appending @file{.dse} to the source file name. | |
4155 | ||
4156 | @item phiopt | |
4157 | @opindex fdump-tree-phiopt | |
4158 | Dump each function after optimizing PHI nodes into straightline code. The file | |
4159 | name is made by appending @file{.phiopt} to the source file name. | |
4160 | ||
4161 | @item forwprop | |
4162 | @opindex fdump-tree-forwprop | |
4163 | Dump each function after forward propagating single use variables. The file | |
4164 | name is made by appending @file{.forwprop} to the source file name. | |
4165 | ||
4166 | @item copyrename | |
4167 | @opindex fdump-tree-copyrename | |
4168 | Dump each function after applying the copy rename optimization. The file | |
4169 | name is made by appending @file{.copyrename} to the source file name. | |
4170 | ||
4171 | @item nrv | |
4172 | @opindex fdump-tree-nrv | |
4173 | Dump each function after applying the named return value optimization on | |
4174 | generic trees. The file name is made by appending @file{.nrv} to the source | |
4175 | file name. | |
4176 | ||
79fe1b3b DN |
4177 | @item vect |
4178 | @opindex fdump-tree-vect | |
f0eb93a8 | 4179 | Dump each function after applying vectorization of loops. The file name is |
79fe1b3b DN |
4180 | made by appending @file{.vect} to the source file name. |
4181 | ||
08873e96 DN |
4182 | @item vrp |
4183 | @opindex fdump-tree-vrp | |
4184 | Dump each function after Value Range Propagation (VRP). The file name | |
4185 | is made by appending @file{.vrp} to the source file name. | |
4186 | ||
6de9cd9a DN |
4187 | @item all |
4188 | @opindex fdump-tree-all | |
4189 | Enable all the available tree dumps with the flags provided in this option. | |
e76b4820 | 4190 | @end table |
9965d119 | 4191 | |
c866976a LB |
4192 | @item -ftree-vectorizer-verbose=@var{n} |
4193 | @opindex ftree-vectorizer-verbose | |
4194 | This option controls the amount of debugging output the vectorizer prints. | |
4195 | This information is written to standard error, unless @option{-fdump-tree-all} | |
4196 | or @option{-fdump-tree-vect} is specified, in which case it is output to the | |
4197 | usual dump listing file, @file{.vect}. | |
4198 | ||
a37db56b GK |
4199 | @item -frandom-seed=@var{string} |
4200 | @opindex frandom-string | |
4201 | This option provides a seed that GCC uses when it would otherwise use | |
e61a2eb7 | 4202 | random numbers. It is used to generate certain symbol names |
8a36672b | 4203 | that have to be different in every compiled file. It is also used to |
e61a2eb7 | 4204 | place unique stamps in coverage data files and the object files that |
8a36672b | 4205 | produce them. You can use the @option{-frandom-seed} option to produce |
e61a2eb7 | 4206 | reproducibly identical object files. |
a37db56b GK |
4207 | |
4208 | The @var{string} should be different for every file you compile. | |
4209 | ||
e03b7153 RS |
4210 | @item -fsched-verbose=@var{n} |
4211 | @opindex fsched-verbose | |
4212 | On targets that use instruction scheduling, this option controls the | |
4213 | amount of debugging output the scheduler prints. This information is | |
4214 | written to standard error, unless @option{-dS} or @option{-dR} is | |
4215 | specified, in which case it is output to the usual dump | |
4216 | listing file, @file{.sched} or @file{.sched2} respectively. However | |
4217 | for @var{n} greater than nine, the output is always printed to standard | |
4218 | error. | |
4219 | ||
4220 | For @var{n} greater than zero, @option{-fsched-verbose} outputs the | |
4221 | same information as @option{-dRS}. For @var{n} greater than one, it | |
4222 | also output basic block probabilities, detailed ready list information | |
4223 | and unit/insn info. For @var{n} greater than two, it includes RTL | |
4224 | at abort point, control-flow and regions info. And for @var{n} over | |
4225 | four, @option{-fsched-verbose} also includes dependence info. | |
4226 | ||
74291a4b | 4227 | @item -save-temps |
cd3bb277 | 4228 | @opindex save-temps |
74291a4b MM |
4229 | Store the usual ``temporary'' intermediate files permanently; place them |
4230 | in the current directory and name them based on the source file. Thus, | |
4231 | compiling @file{foo.c} with @samp{-c -save-temps} would produce files | |
f2ecb02d JM |
4232 | @file{foo.i} and @file{foo.s}, as well as @file{foo.o}. This creates a |
4233 | preprocessed @file{foo.i} output file even though the compiler now | |
4234 | normally uses an integrated preprocessor. | |
74291a4b | 4235 | |
1f7edb8b RS |
4236 | When used in combination with the @option{-x} command line option, |
4237 | @option{-save-temps} is sensible enough to avoid over writing an | |
4238 | input source file with the same extension as an intermediate file. | |
4239 | The corresponding intermediate file may be obtained by renaming the | |
4240 | source file before using @option{-save-temps}. | |
4241 | ||
03c41c05 | 4242 | @item -time |
cd3bb277 | 4243 | @opindex time |
03c41c05 | 4244 | Report the CPU time taken by each subprocess in the compilation |
f2ecb02d JM |
4245 | sequence. For C source files, this is the compiler proper and assembler |
4246 | (plus the linker if linking is done). The output looks like this: | |
03c41c05 ZW |
4247 | |
4248 | @smallexample | |
03c41c05 ZW |
4249 | # cc1 0.12 0.01 |
4250 | # as 0.00 0.01 | |
4251 | @end smallexample | |
4252 | ||
d78aa55c JM |
4253 | The first number on each line is the ``user time'', that is time spent |
4254 | executing the program itself. The second number is ``system time'', | |
03c41c05 ZW |
4255 | time spent executing operating system routines on behalf of the program. |
4256 | Both numbers are in seconds. | |
4257 | ||
014a1138 JZ |
4258 | @item -fvar-tracking |
4259 | @opindex fvar-tracking | |
8a36672b JM |
4260 | Run variable tracking pass. It computes where variables are stored at each |
4261 | position in code. Better debugging information is then generated | |
014a1138 JZ |
4262 | (if the debugging information format supports this information). |
4263 | ||
4264 | It is enabled by default when compiling with optimization (@option{-Os}, | |
4265 | @option{-O}, @option{-O2}, ...), debugging information (@option{-g}) and | |
4266 | the debug info format supports it. | |
4267 | ||
74291a4b | 4268 | @item -print-file-name=@var{library} |
cd3bb277 | 4269 | @opindex print-file-name |
74291a4b MM |
4270 | Print the full absolute name of the library file @var{library} that |
4271 | would be used when linking---and don't do anything else. With this | |
0c2d1a2a | 4272 | option, GCC does not compile or link anything; it just prints the |
74291a4b MM |
4273 | file name. |
4274 | ||
b1018de6 AO |
4275 | @item -print-multi-directory |
4276 | @opindex print-multi-directory | |
4277 | Print the directory name corresponding to the multilib selected by any | |
4278 | other switches present in the command line. This directory is supposed | |
4279 | to exist in @env{GCC_EXEC_PREFIX}. | |
4280 | ||
4281 | @item -print-multi-lib | |
4282 | @opindex print-multi-lib | |
4283 | Print the mapping from multilib directory names to compiler switches | |
4284 | that enable them. The directory name is separated from the switches by | |
4285 | @samp{;}, and each switch starts with an @samp{@@} instead of the | |
4286 | @samp{-}, without spaces between multiple switches. This is supposed to | |
4287 | ease shell-processing. | |
4288 | ||
74291a4b | 4289 | @item -print-prog-name=@var{program} |
cd3bb277 | 4290 | @opindex print-prog-name |
630d3d5a | 4291 | Like @option{-print-file-name}, but searches for a program such as @samp{cpp}. |
74291a4b MM |
4292 | |
4293 | @item -print-libgcc-file-name | |
cd3bb277 | 4294 | @opindex print-libgcc-file-name |
630d3d5a | 4295 | Same as @option{-print-file-name=libgcc.a}. |
74291a4b | 4296 | |
630d3d5a | 4297 | This is useful when you use @option{-nostdlib} or @option{-nodefaultlibs} |
74291a4b MM |
4298 | but you do want to link with @file{libgcc.a}. You can do |
4299 | ||
3ab51846 | 4300 | @smallexample |
74291a4b | 4301 | gcc -nostdlib @var{files}@dots{} `gcc -print-libgcc-file-name` |
3ab51846 | 4302 | @end smallexample |
74291a4b MM |
4303 | |
4304 | @item -print-search-dirs | |
cd3bb277 | 4305 | @opindex print-search-dirs |
74291a4b | 4306 | Print the name of the configured installation directory and a list of |
2dd76960 | 4307 | program and library directories @command{gcc} will search---and don't do anything else. |
74291a4b | 4308 | |
2dd76960 | 4309 | This is useful when @command{gcc} prints the error message |
3c0b7970 JM |
4310 | @samp{installation problem, cannot exec cpp0: No such file or directory}. |
4311 | To resolve this you either need to put @file{cpp0} and the other compiler | |
2dd76960 | 4312 | components where @command{gcc} expects to find them, or you can set the environment |
bedc7537 | 4313 | variable @env{GCC_EXEC_PREFIX} to the directory where you installed them. |
78466c0e | 4314 | Don't forget the trailing @samp{/}. |
74291a4b | 4315 | @xref{Environment Variables}. |
1f0c3120 JM |
4316 | |
4317 | @item -dumpmachine | |
cd3bb277 | 4318 | @opindex dumpmachine |
1f0c3120 JM |
4319 | Print the compiler's target machine (for example, |
4320 | @samp{i686-pc-linux-gnu})---and don't do anything else. | |
4321 | ||
4322 | @item -dumpversion | |
cd3bb277 | 4323 | @opindex dumpversion |
1f0c3120 JM |
4324 | Print the compiler version (for example, @samp{3.0})---and don't do |
4325 | anything else. | |
4326 | ||
4327 | @item -dumpspecs | |
cd3bb277 | 4328 | @opindex dumpspecs |
1f0c3120 JM |
4329 | Print the compiler's built-in specs---and don't do anything else. (This |
4330 | is used when GCC itself is being built.) @xref{Spec Files}. | |
73c68f61 SS |
4331 | |
4332 | @item -feliminate-unused-debug-types | |
4333 | @opindex feliminate-unused-debug-types | |
4334 | Normally, when producing DWARF2 output, GCC will emit debugging | |
4335 | information for all types declared in a compilation | |
4336 | unit, regardless of whether or not they are actually used | |
4337 | in that compilation unit. Sometimes this is useful, such as | |
4338 | if, in the debugger, you want to cast a value to a type that is | |
4339 | not actually used in your program (but is declared). More often, | |
4340 | however, this results in a significant amount of wasted space. | |
4341 | With this option, GCC will avoid producing debug symbol output | |
4342 | for types that are nowhere used in the source file being compiled. | |
74291a4b MM |
4343 | @end table |
4344 | ||
4345 | @node Optimize Options | |
4346 | @section Options That Control Optimization | |
4347 | @cindex optimize options | |
4348 | @cindex options, optimization | |
4349 | ||
147d1cd3 JQ |
4350 | These options control various sorts of optimizations. |
4351 | ||
4352 | Without any optimization option, the compiler's goal is to reduce the | |
4353 | cost of compilation and to make debugging produce the expected | |
4354 | results. Statements are independent: if you stop the program with a | |
4355 | breakpoint between statements, you can then assign a new value to any | |
4356 | variable or change the program counter to any other statement in the | |
4357 | function and get exactly the results you would expect from the source | |
4358 | code. | |
4359 | ||
4360 | Turning on optimization flags makes the compiler attempt to improve | |
4361 | the performance and/or code size at the expense of compilation time | |
4362 | and possibly the ability to debug the program. | |
4363 | ||
a451b0bd | 4364 | The compiler performs optimization based on the knowledge it has of |
7797ff53 PB |
4365 | the program. Optimization levels @option{-O2} and above, in |
4366 | particular, enable @emph{unit-at-a-time} mode, which allows the | |
4367 | compiler to consider information gained from later functions in | |
4368 | the file when compiling a function. Compiling multiple files at | |
4369 | once to a single output file in @emph{unit-at-a-time} mode allows | |
d1bd0ded GK |
4370 | the compiler to use information gained from all of the files when |
4371 | compiling each of them. | |
4372 | ||
147d1cd3 JQ |
4373 | Not all optimizations are controlled directly by a flag. Only |
4374 | optimizations that have a flag are listed. | |
74291a4b | 4375 | |
2642624b | 4376 | @table @gcctabopt |
74291a4b MM |
4377 | @item -O |
4378 | @itemx -O1 | |
cd3bb277 JM |
4379 | @opindex O |
4380 | @opindex O1 | |
74291a4b MM |
4381 | Optimize. Optimizing compilation takes somewhat more time, and a lot |
4382 | more memory for a large function. | |
4383 | ||
630d3d5a | 4384 | With @option{-O}, the compiler tries to reduce code size and execution |
9c34dbbf ZW |
4385 | time, without performing any optimizations that take a great deal of |
4386 | compilation time. | |
74291a4b | 4387 | |
daf2f129 | 4388 | @option{-O} turns on the following optimization flags: |
9a94f7f3 | 4389 | @gccoptlist{-fdefer-pop @gol |
08711fdf SB |
4390 | -fdelayed-branch @gol |
4391 | -fguess-branch-probability @gol | |
4392 | -fcprop-registers @gol | |
9a94f7f3 | 4393 | -floop-optimize @gol |
9a94f7f3 JM |
4394 | -fif-conversion @gol |
4395 | -fif-conversion2 @gol | |
08711fdf SB |
4396 | -ftree-ccp @gol |
4397 | -ftree-dce @gol | |
ec7dea0a | 4398 | -ftree-dominator-opts @gol |
08711fdf SB |
4399 | -ftree-dse @gol |
4400 | -ftree-ter @gol | |
3c632d89 | 4401 | -ftree-lrs @gol |
08711fdf SB |
4402 | -ftree-sra @gol |
4403 | -ftree-copyrename @gol | |
4404 | -ftree-fre @gol | |
4405 | -ftree-ch @gol | |
4406 | -fmerge-constants} | |
fad893da JQ |
4407 | |
4408 | @option{-O} also turns on @option{-fomit-frame-pointer} on machines | |
4409 | where doing so does not interfere with debugging. | |
4410 | ||
74291a4b | 4411 | @item -O2 |
cd3bb277 | 4412 | @opindex O2 |
0c2d1a2a | 4413 | Optimize even more. GCC performs nearly all supported optimizations |
74291a4b | 4414 | that do not involve a space-speed tradeoff. The compiler does not |
630d3d5a JM |
4415 | perform loop unrolling or function inlining when you specify @option{-O2}. |
4416 | As compared to @option{-O}, this option increases both compilation time | |
74291a4b MM |
4417 | and the performance of the generated code. |
4418 | ||
fad893da JQ |
4419 | @option{-O2} turns on all optimization flags specified by @option{-O}. It |
4420 | also turns on the following optimization flags: | |
08711fdf SB |
4421 | @gccoptlist{-fthread-jumps @gol |
4422 | -fcrossjumping @gol | |
9a94f7f3 | 4423 | -foptimize-sibling-calls @gol |
9a94f7f3 | 4424 | -fcse-follow-jumps -fcse-skip-blocks @gol |
08711fdf | 4425 | -fgcse -fgcse-lm @gol |
9a94f7f3 | 4426 | -fexpensive-optimizations @gol |
08711fdf SB |
4427 | -fstrength-reduce @gol |
4428 | -frerun-cse-after-loop -frerun-loop-opt @gol | |
9a94f7f3 JM |
4429 | -fcaller-saves @gol |
4430 | -fpeephole2 @gol | |
08711fdf SB |
4431 | -fschedule-insns -fschedule-insns2 @gol |
4432 | -fsched-interblock -fsched-spec @gol | |
4433 | -fregmove @gol | |
9a94f7f3 | 4434 | -fstrict-aliasing @gol |
08711fdf SB |
4435 | -fdelete-null-pointer-checks @gol |
4436 | -freorder-blocks -freorder-functions @gol | |
23a44080 | 4437 | -funit-at-a-time @gol |
9a94f7f3 | 4438 | -falign-functions -falign-jumps @gol |
b684a3df | 4439 | -falign-loops -falign-labels @gol |
08873e96 | 4440 | -ftree-vrp @gol |
08711fdf | 4441 | -ftree-pre} |
74291a4b | 4442 | |
081ca317 BL |
4443 | Please note the warning under @option{-fgcse} about |
4444 | invoking @option{-O2} on programs that use computed gotos. | |
4445 | ||
74291a4b | 4446 | @item -O3 |
cd3bb277 | 4447 | @opindex O3 |
630d3d5a | 4448 | Optimize yet more. @option{-O3} turns on all optimizations specified by |
dafc5b82 | 4449 | @option{-O2} and also turns on the @option{-finline-functions}, |
08711fdf | 4450 | @option{-funswitch-loops} and @option{-fgcse-after-reload} options. |
74291a4b MM |
4451 | |
4452 | @item -O0 | |
cd3bb277 | 4453 | @opindex O0 |
fad893da | 4454 | Do not optimize. This is the default. |
74291a4b | 4455 | |
c6aded7c | 4456 | @item -Os |
cd3bb277 | 4457 | @opindex Os |
630d3d5a | 4458 | Optimize for size. @option{-Os} enables all @option{-O2} optimizations that |
c6aded7c AG |
4459 | do not typically increase code size. It also performs further |
4460 | optimizations designed to reduce code size. | |
4461 | ||
fad893da | 4462 | @option{-Os} disables the following optimization flags: |
9a94f7f3 | 4463 | @gccoptlist{-falign-functions -falign-jumps -falign-loops @gol |
c12cc930 KB |
4464 | -falign-labels -freorder-blocks -freorder-blocks-and-partition @gol |
4465 | -fprefetch-loop-arrays -ftree-vect-loop-version} | |
fad893da | 4466 | |
630d3d5a | 4467 | If you use multiple @option{-O} options, with or without level numbers, |
74291a4b MM |
4468 | the last such option is the one that is effective. |
4469 | @end table | |
4470 | ||
630d3d5a | 4471 | Options of the form @option{-f@var{flag}} specify machine-independent |
74291a4b | 4472 | flags. Most flags have both positive and negative forms; the negative |
147d1cd3 JQ |
4473 | form of @option{-ffoo} would be @option{-fno-foo}. In the table |
4474 | below, only one of the forms is listed---the one you typically will | |
4475 | use. You can figure out the other form by either removing @samp{no-} | |
4476 | or adding it. | |
4477 | ||
4478 | The following options control specific optimizations. They are either | |
4479 | activated by @option{-O} options or are related to ones that are. You | |
4480 | can use the following flags in the rare cases when ``fine-tuning'' of | |
4481 | optimizations to be performed is desired. | |
74291a4b | 4482 | |
2642624b | 4483 | @table @gcctabopt |
74291a4b | 4484 | @item -fno-default-inline |
cd3bb277 | 4485 | @opindex fno-default-inline |
74291a4b MM |
4486 | Do not make member functions inline by default merely because they are |
4487 | defined inside the class scope (C++ only). Otherwise, when you specify | |
630d3d5a | 4488 | @w{@option{-O}}, member functions defined inside class scope are compiled |
74291a4b MM |
4489 | inline by default; i.e., you don't need to add @samp{inline} in front of |
4490 | the member function name. | |
4491 | ||
4492 | @item -fno-defer-pop | |
cd3bb277 | 4493 | @opindex fno-defer-pop |
74291a4b MM |
4494 | Always pop the arguments to each function call as soon as that function |
4495 | returns. For machines which must pop arguments after a function call, | |
4496 | the compiler normally lets arguments accumulate on the stack for several | |
4497 | function calls and pops them all at once. | |
4498 | ||
38df970e JQ |
4499 | Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}. |
4500 | ||
74291a4b | 4501 | @item -fforce-mem |
cd3bb277 | 4502 | @opindex fforce-mem |
74291a4b MM |
4503 | Force memory operands to be copied into registers before doing |
4504 | arithmetic on them. This produces better code by making all memory | |
4505 | references potential common subexpressions. When they are not common | |
4506 | subexpressions, instruction combination should eliminate the separate | |
1d8eeb63 | 4507 | register-load. This option is now a nop and will be removed in 4.2. |
74291a4b MM |
4508 | |
4509 | @item -fforce-addr | |
cd3bb277 | 4510 | @opindex fforce-addr |
74291a4b | 4511 | Force memory address constants to be copied into registers before |
1d8eeb63 | 4512 | doing arithmetic on them. |
74291a4b MM |
4513 | |
4514 | @item -fomit-frame-pointer | |
cd3bb277 | 4515 | @opindex fomit-frame-pointer |
74291a4b MM |
4516 | Don't keep the frame pointer in a register for functions that |
4517 | don't need one. This avoids the instructions to save, set up and | |
4518 | restore frame pointers; it also makes an extra register available | |
4519 | in many functions. @strong{It also makes debugging impossible on | |
4520 | some machines.} | |
4521 | ||
8aeea6e6 | 4522 | On some machines, such as the VAX, this flag has no effect, because |
74291a4b MM |
4523 | the standard calling sequence automatically handles the frame pointer |
4524 | and nothing is saved by pretending it doesn't exist. The | |
4525 | machine-description macro @code{FRAME_POINTER_REQUIRED} controls | |
4526 | whether a target machine supports this flag. @xref{Registers,,Register | |
b11cc610 | 4527 | Usage, gccint, GNU Compiler Collection (GCC) Internals}. |
74291a4b | 4528 | |
38df970e JQ |
4529 | Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}. |
4530 | ||
1aaef9c1 | 4531 | @item -foptimize-sibling-calls |
cd3bb277 | 4532 | @opindex foptimize-sibling-calls |
1aaef9c1 JH |
4533 | Optimize sibling and tail recursive calls. |
4534 | ||
38df970e JQ |
4535 | Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}. |
4536 | ||
74291a4b | 4537 | @item -fno-inline |
cd3bb277 | 4538 | @opindex fno-inline |
74291a4b MM |
4539 | Don't pay attention to the @code{inline} keyword. Normally this option |
4540 | is used to keep the compiler from expanding any functions inline. | |
4541 | Note that if you are not optimizing, no functions can be expanded inline. | |
4542 | ||
4543 | @item -finline-functions | |
cd3bb277 | 4544 | @opindex finline-functions |
74291a4b MM |
4545 | Integrate all simple functions into their callers. The compiler |
4546 | heuristically decides which functions are simple enough to be worth | |
4547 | integrating in this way. | |
4548 | ||
4549 | If all calls to a given function are integrated, and the function is | |
4550 | declared @code{static}, then the function is normally not output as | |
4551 | assembler code in its own right. | |
4552 | ||
38df970e | 4553 | Enabled at level @option{-O3}. |
355866de RG |
4554 | |
4555 | @item -finline-functions-called-once | |
4556 | @opindex finline-functions-called-once | |
4557 | Consider all @code{static} functions called once for inlining into their | |
4558 | caller even if they are not marked @code{inline}. If a call to a given | |
4559 | function is integrated, then the function is not output as assembler code | |
4560 | in its own right. | |
4561 | ||
4562 | Enabled if @option{-funit-at-a-time} is enabled. | |
38df970e | 4563 | |
d63db217 JH |
4564 | @item -fearly-inlining |
4565 | @opindex fearly-inlining | |
4566 | Inline functions marked by @code{always_inline} and functions whose body seems | |
4567 | smaller than the function call overhead early before doing | |
4568 | @option{-fprofile-generate} instrumentation and real inlining pass. Doing so | |
4569 | makes profiling significantly cheaper and usually inlining faster on programs | |
4570 | having large chains of nested wrapper functions. | |
4571 | ||
4572 | Enabled by default. | |
4573 | ||
efa3896a | 4574 | @item -finline-limit=@var{n} |
cd3bb277 | 4575 | @opindex finline-limit |
2dd76960 | 4576 | By default, GCC limits the size of functions that can be inlined. This flag |
f9e814f1 | 4577 | allows the control of this limit for functions that are explicitly marked as |
3364c33b | 4578 | inline (i.e., marked with the inline keyword or defined within the class |
02f52e19 | 4579 | definition in c++). @var{n} is the size of functions that can be inlined in |
f9e814f1 | 4580 | number of pseudo instructions (not counting parameter handling). The default |
93ee12c4 GP |
4581 | value of @var{n} is 600. |
4582 | Increasing this value can result in more inlined code at | |
f9e814f1 | 4583 | the cost of compilation time and memory consumption. Decreasing usually makes |
02f52e19 AJ |
4584 | the compilation faster and less code will be inlined (which presumably |
4585 | means slower programs). This option is particularly useful for programs that | |
aee96fe9 | 4586 | use inlining heavily such as those based on recursive templates with C++. |
f9e814f1 | 4587 | |
bc522472 KG |
4588 | Inlining is actually controlled by a number of parameters, which may be |
4589 | specified individually by using @option{--param @var{name}=@var{value}}. | |
daf2f129 | 4590 | The @option{-finline-limit=@var{n}} option sets some of these parameters |
bc522472 KG |
4591 | as follows: |
4592 | ||
4593 | @table @gcctabopt | |
bc522472 KG |
4594 | @item max-inline-insns-single |
4595 | is set to @var{n}/2. | |
6d7fe8b3 | 4596 | @item max-inline-insns-auto |
bc522472 KG |
4597 | is set to @var{n}/2. |
4598 | @item min-inline-insns | |
4599 | is set to 130 or @var{n}/4, whichever is smaller. | |
4600 | @item max-inline-insns-rtl | |
4601 | is set to @var{n}. | |
4602 | @end table | |
4603 | ||
f7a01847 | 4604 | See below for a documentation of the individual |
bc522472 KG |
4605 | parameters controlling inlining. |
4606 | ||
f9e814f1 | 4607 | @emph{Note:} pseudo instruction represents, in this particular context, an |
0fa2e4df | 4608 | abstract measurement of function's size. In no way does it represent a count |
f9e814f1 TP |
4609 | of assembly instructions and as such its exact meaning might change from one |
4610 | release to an another. | |
4611 | ||
74291a4b | 4612 | @item -fkeep-inline-functions |
cd3bb277 | 4613 | @opindex fkeep-inline-functions |
1a10290c MM |
4614 | In C, emit @code{static} functions that are declared @code{inline} |
4615 | into the object file, even if the function has been inlined into all | |
4616 | of its callers. This switch does not affect functions using the | |
8a36672b | 4617 | @code{extern inline} extension in GNU C@. In C++, emit any and all |
1a10290c | 4618 | inline functions into the object file. |
74291a4b MM |
4619 | |
4620 | @item -fkeep-static-consts | |
cd3bb277 | 4621 | @opindex fkeep-static-consts |
74291a4b MM |
4622 | Emit variables declared @code{static const} when optimization isn't turned |
4623 | on, even if the variables aren't referenced. | |
4624 | ||
0c2d1a2a | 4625 | GCC enables this option by default. If you want to force the compiler to |
74291a4b | 4626 | check if the variable was referenced, regardless of whether or not |
630d3d5a | 4627 | optimization is turned on, use the @option{-fno-keep-static-consts} option. |
74291a4b | 4628 | |
201556f0 JJ |
4629 | @item -fmerge-constants |
4630 | Attempt to merge identical constants (string constants and floating point | |
3364c33b | 4631 | constants) across compilation units. |
201556f0 | 4632 | |
3364c33b JQ |
4633 | This option is the default for optimized compilation if the assembler and |
4634 | linker support it. Use @option{-fno-merge-constants} to inhibit this | |
4635 | behavior. | |
201556f0 | 4636 | |
38df970e JQ |
4637 | Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}. |
4638 | ||
201556f0 JJ |
4639 | @item -fmerge-all-constants |
4640 | Attempt to merge identical constants and identical variables. | |
4641 | ||
4642 | This option implies @option{-fmerge-constants}. In addition to | |
431ae0bf | 4643 | @option{-fmerge-constants} this considers e.g.@: even constant initialized |
201556f0 JJ |
4644 | arrays or initialized constant variables with integral or floating point |
4645 | types. Languages like C or C++ require each non-automatic variable to | |
4646 | have distinct location, so using this option will result in non-conforming | |
c21cd8b1 | 4647 | behavior. |
201556f0 | 4648 | |
e5626198 AZ |
4649 | @item -fmodulo-sched |
4650 | @opindex fmodulo-sched | |
4651 | Perform swing modulo scheduling immediately before the first scheduling | |
4652 | pass. This pass looks at innermost loops and reorders their | |
4653 | instructions by overlapping different iterations. | |
4654 | ||
e03b7153 RS |
4655 | @item -fno-branch-count-reg |
4656 | @opindex fno-branch-count-reg | |
4657 | Do not use ``decrement and branch'' instructions on a count register, | |
4658 | but instead generate a sequence of instructions that decrement a | |
4659 | register, compare it against zero, then branch based upon the result. | |
4660 | This option is only meaningful on architectures that support such | |
4661 | instructions, which include x86, PowerPC, IA-64 and S/390. | |
4662 | ||
38df970e JQ |
4663 | The default is @option{-fbranch-count-reg}, enabled when |
4664 | @option{-fstrength-reduce} is enabled. | |
4665 | ||
74291a4b | 4666 | @item -fno-function-cse |
cd3bb277 | 4667 | @opindex fno-function-cse |
74291a4b MM |
4668 | Do not put function addresses in registers; make each instruction that |
4669 | calls a constant function contain the function's address explicitly. | |
4670 | ||
4671 | This option results in less efficient code, but some strange hacks | |
4672 | that alter the assembler output may be confused by the optimizations | |
4673 | performed when this option is not used. | |
4674 | ||
38df970e JQ |
4675 | The default is @option{-ffunction-cse} |
4676 | ||
27b41650 KG |
4677 | @item -fno-zero-initialized-in-bss |
4678 | @opindex fno-zero-initialized-in-bss | |
4679 | If the target supports a BSS section, GCC by default puts variables that | |
4680 | are initialized to zero into BSS@. This can save space in the resulting | |
4681 | code. | |
4682 | ||
4683 | This option turns off this behavior because some programs explicitly | |
4684 | rely on variables going to the data section. E.g., so that the | |
4685 | resulting executable can find the beginning of that section and/or make | |
4686 | assumptions based on that. | |
4687 | ||
4688 | The default is @option{-fzero-initialized-in-bss}. | |
e03b7153 | 4689 | |
6de9cd9a DN |
4690 | @item -fbounds-check |
4691 | @opindex fbounds-check | |
4692 | For front-ends that support it, generate additional code to check that | |
4693 | indices used to access arrays are within the declared range. This is | |
4694 | currently only supported by the Java and Fortran front-ends, where | |
4695 | this option defaults to true and false respectively. | |
4696 | ||
4697 | @item -fmudflap -fmudflapth -fmudflapir | |
4698 | @opindex fmudflap | |
4699 | @opindex fmudflapth | |
4700 | @opindex fmudflapir | |
4701 | @cindex bounds checking | |
4702 | @cindex mudflap | |
4703 | For front-ends that support it (C and C++), instrument all risky | |
4704 | pointer/array dereferencing operations, some standard library | |
4705 | string/heap functions, and some other associated constructs with | |
4706 | range/validity tests. Modules so instrumented should be immune to | |
4707 | buffer overflows, invalid heap use, and some other classes of C/C++ | |
4708 | programming errors. The instrumentation relies on a separate runtime | |
4709 | library (@file{libmudflap}), which will be linked into a program if | |
4710 | @option{-fmudflap} is given at link time. Run-time behavior of the | |
4711 | instrumented program is controlled by the @env{MUDFLAP_OPTIONS} | |
4712 | environment variable. See @code{env MUDFLAP_OPTIONS=-help a.out} | |
4713 | for its options. | |
4714 | ||
4715 | Use @option{-fmudflapth} instead of @option{-fmudflap} to compile and to | |
4716 | link if your program is multi-threaded. Use @option{-fmudflapir}, in | |
4717 | addition to @option{-fmudflap} or @option{-fmudflapth}, if | |
4718 | instrumentation should ignore pointer reads. This produces less | |
4719 | instrumentation (and therefore faster execution) and still provides | |
4720 | some protection against outright memory corrupting writes, but allows | |
4721 | erroneously read data to propagate within a program. | |
4722 | ||
953ff289 DN |
4723 | @item -fopenmp |
4724 | @opindex fopenmp | |
4725 | @cindex openmp parallel | |
4726 | Enable handling of OpenMP directives @code{#pragma omp} in C/C++ and | |
4727 | @code{!$omp} in Fortran. When @option{-fopenmp} is specified, the | |
4728 | compiler generates parallel code according to the OpenMP Application | |
4729 | Program Interface v2.5. To generate the final exectuable, the runtime | |
4730 | library @code{libgomp} must be linked in using @option{-lgomp}. | |
4731 | ||
74291a4b | 4732 | @item -fstrength-reduce |
cd3bb277 | 4733 | @opindex fstrength-reduce |
74291a4b MM |
4734 | Perform the optimizations of loop strength reduction and |
4735 | elimination of iteration variables. | |
4736 | ||
38df970e JQ |
4737 | Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}. |
4738 | ||
74291a4b | 4739 | @item -fthread-jumps |
cd3bb277 | 4740 | @opindex fthread-jumps |
74291a4b MM |
4741 | Perform optimizations where we check to see if a jump branches to a |
4742 | location where another comparison subsumed by the first is found. If | |
4743 | so, the first branch is redirected to either the destination of the | |
4744 | second branch or a point immediately following it, depending on whether | |
4745 | the condition is known to be true or false. | |
4746 | ||
08711fdf | 4747 | Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}. |
38df970e | 4748 | |
74291a4b | 4749 | @item -fcse-follow-jumps |
cd3bb277 | 4750 | @opindex fcse-follow-jumps |
74291a4b MM |
4751 | In common subexpression elimination, scan through jump instructions |
4752 | when the target of the jump is not reached by any other path. For | |
4753 | example, when CSE encounters an @code{if} statement with an | |
4754 | @code{else} clause, CSE will follow the jump when the condition | |
4755 | tested is false. | |
4756 | ||
38df970e JQ |
4757 | Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}. |
4758 | ||
74291a4b | 4759 | @item -fcse-skip-blocks |
cd3bb277 | 4760 | @opindex fcse-skip-blocks |
630d3d5a | 4761 | This is similar to @option{-fcse-follow-jumps}, but causes CSE to |
74291a4b MM |
4762 | follow jumps which conditionally skip over blocks. When CSE |
4763 | encounters a simple @code{if} statement with no else clause, | |
630d3d5a | 4764 | @option{-fcse-skip-blocks} causes CSE to follow the jump around the |
74291a4b MM |
4765 | body of the @code{if}. |
4766 | ||
38df970e JQ |
4767 | Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}. |
4768 | ||
74291a4b | 4769 | @item -frerun-cse-after-loop |
cd3bb277 | 4770 | @opindex frerun-cse-after-loop |
74291a4b MM |
4771 | Re-run common subexpression elimination after loop optimizations has been |
4772 | performed. | |
4773 | ||
38df970e JQ |
4774 | Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}. |
4775 | ||
6d6d0fa0 | 4776 | @item -frerun-loop-opt |
cd3bb277 | 4777 | @opindex frerun-loop-opt |
6d6d0fa0 JL |
4778 | Run the loop optimizer twice. |
4779 | ||
38df970e JQ |
4780 | Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}. |
4781 | ||
7506f491 | 4782 | @item -fgcse |
cd3bb277 | 4783 | @opindex fgcse |
7506f491 DE |
4784 | Perform a global common subexpression elimination pass. |
4785 | This pass also performs global constant and copy propagation. | |
4786 | ||
081ca317 BL |
4787 | @emph{Note:} When compiling a program using computed gotos, a GCC |
4788 | extension, you may get better runtime performance if you disable | |
3364c33b | 4789 | the global common subexpression elimination pass by adding |
081ca317 BL |
4790 | @option{-fno-gcse} to the command line. |
4791 | ||
38df970e JQ |
4792 | Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}. |
4793 | ||
a13d4ebf | 4794 | @item -fgcse-lm |
cd3bb277 | 4795 | @opindex fgcse-lm |
695ac33f | 4796 | When @option{-fgcse-lm} is enabled, global common subexpression elimination will |
767094dd | 4797 | attempt to move loads which are only killed by stores into themselves. This |
a13d4ebf | 4798 | allows a loop containing a load/store sequence to be changed to a load outside |
02f52e19 | 4799 | the loop, and a copy/store within the loop. |
a13d4ebf | 4800 | |
38df970e JQ |
4801 | Enabled by default when gcse is enabled. |
4802 | ||
a13d4ebf | 4803 | @item -fgcse-sm |
cd3bb277 | 4804 | @opindex fgcse-sm |
f5f2e3cd MH |
4805 | When @option{-fgcse-sm} is enabled, a store motion pass is run after |
4806 | global common subexpression elimination. This pass will attempt to move | |
4807 | stores out of loops. When used in conjunction with @option{-fgcse-lm}, | |
4808 | loops containing a load/store sequence can be changed to a load before | |
4809 | the loop and a store after the loop. | |
4810 | ||
08711fdf | 4811 | Not enabled at any optimization level. |
f5f2e3cd MH |
4812 | |
4813 | @item -fgcse-las | |
4814 | @opindex fgcse-las | |
4815 | When @option{-fgcse-las} is enabled, the global common subexpression | |
4816 | elimination pass eliminates redundant loads that come after stores to the | |
2206e783 | 4817 | same memory location (both partial and full redundancies). |
a13d4ebf | 4818 | |
08711fdf | 4819 | Not enabled at any optimization level. |
38df970e | 4820 | |
db643b91 SH |
4821 | @item -fgcse-after-reload |
4822 | @opindex fgcse-after-reload | |
4823 | When @option{-fgcse-after-reload} is enabled, a redundant load elimination | |
8a36672b | 4824 | pass is performed after reload. The purpose of this pass is to cleanup |
db643b91 SH |
4825 | redundant spilling. |
4826 | ||
96327cdc JH |
4827 | @item -floop-optimize |
4828 | @opindex floop-optimize | |
4829 | Perform loop optimizations: move constant expressions out of loops, simplify | |
c94583fe | 4830 | exit test conditions and optionally do strength-reduction as well. |
96327cdc | 4831 | |
38df970e JQ |
4832 | Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}. |
4833 | ||
5e962776 ZD |
4834 | @item -floop-optimize2 |
4835 | @opindex floop-optimize2 | |
4836 | Perform loop optimizations using the new loop optimizer. The optimizations | |
4837 | (loop unrolling, peeling and unswitching, loop invariant motion) are enabled | |
4838 | by separate flags. | |
4839 | ||
f9cc1a70 PB |
4840 | @item -funsafe-loop-optimizations |
4841 | @opindex funsafe-loop-optimizations | |
4842 | If given, the loop optimizer will assume that loop indices do not | |
4843 | overflow, and that the loops with nontrivial exit condition are not | |
4844 | infinite. This enables a wider range of loop optimizations even if | |
4845 | the loop optimizer itself cannot prove that these assumptions are valid. | |
4846 | Using @option{-Wunsafe-loop-optimizations}, the compiler will warn you | |
4847 | if it finds this kind of loop. | |
4848 | ||
96327cdc JH |
4849 | @item -fcrossjumping |
4850 | @opindex crossjumping | |
8a36672b | 4851 | Perform cross-jumping transformation. This transformation unifies equivalent code and save code size. The |
96327cdc JH |
4852 | resulting code may or may not perform better than without cross-jumping. |
4853 | ||
53071270 | 4854 | Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}. |
38df970e | 4855 | |
2c4b77f3 JH |
4856 | @item -fif-conversion |
4857 | @opindex if-conversion | |
4858 | Attempt to transform conditional jumps into branch-less equivalents. This | |
4859 | include use of conditional moves, min, max, set flags and abs instructions, and | |
4860 | some tricks doable by standard arithmetics. The use of conditional execution | |
4861 | on chips where it is available is controlled by @code{if-conversion2}. | |
4862 | ||
38df970e JQ |
4863 | Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}. |
4864 | ||
2c4b77f3 JH |
4865 | @item -fif-conversion2 |
4866 | @opindex if-conversion2 | |
4867 | Use conditional execution (where available) to transform conditional jumps into | |
4868 | branch-less equivalents. | |
4869 | ||
38df970e JQ |
4870 | Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}. |
4871 | ||
b6d24183 | 4872 | @item -fdelete-null-pointer-checks |
cd3bb277 | 4873 | @opindex fdelete-null-pointer-checks |
9c34dbbf ZW |
4874 | Use global dataflow analysis to identify and eliminate useless checks |
4875 | for null pointers. The compiler assumes that dereferencing a null | |
4876 | pointer would have halted the program. If a pointer is checked after | |
4877 | it has already been dereferenced, it cannot be null. | |
4878 | ||
4879 | In some environments, this assumption is not true, and programs can | |
4880 | safely dereference null pointers. Use | |
4881 | @option{-fno-delete-null-pointer-checks} to disable this optimization | |
4882 | for programs which depend on that behavior. | |
b6d24183 | 4883 | |
38df970e JQ |
4884 | Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}. |
4885 | ||
74291a4b | 4886 | @item -fexpensive-optimizations |
cd3bb277 | 4887 | @opindex fexpensive-optimizations |
74291a4b MM |
4888 | Perform a number of minor optimizations that are relatively expensive. |
4889 | ||
38df970e JQ |
4890 | Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}. |
4891 | ||
639726ba | 4892 | @item -foptimize-register-move |
59d40964 | 4893 | @itemx -fregmove |
cd3bb277 JM |
4894 | @opindex foptimize-register-move |
4895 | @opindex fregmove | |
9ec36da5 JL |
4896 | Attempt to reassign register numbers in move instructions and as |
4897 | operands of other simple instructions in order to maximize the amount of | |
56159047 | 4898 | register tying. This is especially helpful on machines with two-operand |
38df970e | 4899 | instructions. |
9ec36da5 | 4900 | |
bedc7537 | 4901 | Note @option{-fregmove} and @option{-foptimize-register-move} are the same |
9ec36da5 JL |
4902 | optimization. |
4903 | ||
38df970e JQ |
4904 | Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}. |
4905 | ||
74291a4b | 4906 | @item -fdelayed-branch |
cd3bb277 | 4907 | @opindex fdelayed-branch |
74291a4b MM |
4908 | If supported for the target machine, attempt to reorder instructions |
4909 | to exploit instruction slots available after delayed branch | |
4910 | instructions. | |
4911 | ||
38df970e JQ |
4912 | Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}. |
4913 | ||
74291a4b | 4914 | @item -fschedule-insns |
cd3bb277 | 4915 | @opindex fschedule-insns |
74291a4b MM |
4916 | If supported for the target machine, attempt to reorder instructions to |
4917 | eliminate execution stalls due to required data being unavailable. This | |
4918 | helps machines that have slow floating point or memory load instructions | |
4919 | by allowing other instructions to be issued until the result of the load | |
4920 | or floating point instruction is required. | |
4921 | ||
38df970e JQ |
4922 | Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}. |
4923 | ||
74291a4b | 4924 | @item -fschedule-insns2 |
cd3bb277 | 4925 | @opindex fschedule-insns2 |
630d3d5a | 4926 | Similar to @option{-fschedule-insns}, but requests an additional pass of |
74291a4b MM |
4927 | instruction scheduling after register allocation has been done. This is |
4928 | especially useful on machines with a relatively small number of | |
4929 | registers and where memory load instructions take more than one cycle. | |
4930 | ||
38df970e JQ |
4931 | Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}. |
4932 | ||
e03b7153 RS |
4933 | @item -fno-sched-interblock |
4934 | @opindex fno-sched-interblock | |
4935 | Don't schedule instructions across basic blocks. This is normally | |
4936 | enabled by default when scheduling before register allocation, i.e.@: | |
4937 | with @option{-fschedule-insns} or at @option{-O2} or higher. | |
4938 | ||
4939 | @item -fno-sched-spec | |
4940 | @opindex fno-sched-spec | |
4941 | Don't allow speculative motion of non-load instructions. This is normally | |
4942 | enabled by default when scheduling before register allocation, i.e.@: | |
4943 | with @option{-fschedule-insns} or at @option{-O2} or higher. | |
4944 | ||
4945 | @item -fsched-spec-load | |
4946 | @opindex fsched-spec-load | |
4947 | Allow speculative motion of some load instructions. This only makes | |
4948 | sense when scheduling before register allocation, i.e.@: with | |
4949 | @option{-fschedule-insns} or at @option{-O2} or higher. | |
4950 | ||
4951 | @item -fsched-spec-load-dangerous | |
4952 | @opindex fsched-spec-load-dangerous | |
4953 | Allow speculative motion of more load instructions. This only makes | |
4954 | sense when scheduling before register allocation, i.e.@: with | |
4955 | @option{-fschedule-insns} or at @option{-O2} or higher. | |
4956 | ||
569fa502 DN |
4957 | @item -fsched-stalled-insns=@var{n} |
4958 | @opindex fsched-stalled-insns | |
4959 | Define how many insns (if any) can be moved prematurely from the queue | |
4960 | of stalled insns into the ready list, during the second scheduling pass. | |
4961 | ||
4962 | @item -fsched-stalled-insns-dep=@var{n} | |
4963 | @opindex fsched-stalled-insns-dep | |
daf2f129 JM |
4964 | Define how many insn groups (cycles) will be examined for a dependency |
4965 | on a stalled insn that is candidate for premature removal from the queue | |
4966 | of stalled insns. Has an effect only during the second scheduling pass, | |
569fa502 DN |
4967 | and only if @option{-fsched-stalled-insns} is used and its value is not zero. |
4968 | ||
b9422b69 JH |
4969 | @item -fsched2-use-superblocks |
4970 | @opindex fsched2-use-superblocks | |
61aeb06f | 4971 | When scheduling after register allocation, do use superblock scheduling |
62b9c42c | 4972 | algorithm. Superblock scheduling allows motion across basic block boundaries |
b9422b69 | 4973 | resulting on faster schedules. This option is experimental, as not all machine |
62b9c42c | 4974 | descriptions used by GCC model the CPU closely enough to avoid unreliable |
daf2f129 | 4975 | results from the algorithm. |
b9422b69 JH |
4976 | |
4977 | This only makes sense when scheduling after register allocation, i.e.@: with | |
4978 | @option{-fschedule-insns2} or at @option{-O2} or higher. | |
4979 | ||
4980 | @item -fsched2-use-traces | |
4981 | @opindex fsched2-use-traces | |
4982 | Use @option{-fsched2-use-superblocks} algorithm when scheduling after register | |
4983 | allocation and additionally perform code duplication in order to increase the | |
4984 | size of superblocks using tracer pass. See @option{-ftracer} for details on | |
4985 | trace formation. | |
4986 | ||
62b9c42c | 4987 | This mode should produce faster but significantly longer programs. Also |
4ec7afd7 KH |
4988 | without @option{-fbranch-probabilities} the traces constructed may not |
4989 | match the reality and hurt the performance. This only makes | |
b9422b69 JH |
4990 | sense when scheduling after register allocation, i.e.@: with |
4991 | @option{-fschedule-insns2} or at @option{-O2} or higher. | |
4992 | ||
d72372e4 MH |
4993 | @item -freschedule-modulo-scheduled-loops |
4994 | @opindex fscheduling-in-modulo-scheduled-loops | |
4995 | The modulo scheduling comes before the traditional scheduling, if a loop was modulo scheduled | |
4996 | we may want to prevent the later scheduling passes from changing its schedule, we use this | |
4997 | option to control that. | |
4998 | ||
74291a4b | 4999 | @item -fcaller-saves |
cd3bb277 | 5000 | @opindex fcaller-saves |
74291a4b MM |
5001 | Enable values to be allocated in registers that will be clobbered by |
5002 | function calls, by emitting extra instructions to save and restore the | |
5003 | registers around such calls. Such allocation is done only when it | |
5004 | seems to result in better code than would otherwise be produced. | |
5005 | ||
81610a0d HPN |
5006 | This option is always enabled by default on certain machines, usually |
5007 | those which have no call-preserved registers to use instead. | |
5008 | ||
38df970e | 5009 | Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}. |
74291a4b | 5010 | |
6de9cd9a DN |
5011 | @item -ftree-pre |
5012 | Perform Partial Redundancy Elimination (PRE) on trees. This flag is | |
08711fdf | 5013 | enabled by default at @option{-O2} and @option{-O3}. |
ff2ad0f7 DN |
5014 | |
5015 | @item -ftree-fre | |
5016 | Perform Full Redundancy Elimination (FRE) on trees. The difference | |
5017 | between FRE and PRE is that FRE only considers expressions | |
5018 | that are computed on all paths leading to the redundant computation. | |
5019 | This analysis faster than PRE, though it exposes fewer redundancies. | |
4ec7afd7 | 5020 | This flag is enabled by default at @option{-O} and higher. |
6de9cd9a | 5021 | |
0bca51f0 DN |
5022 | @item -ftree-copy-prop |
5023 | Perform copy propagation on trees. This pass eliminates unnecessary | |
5024 | copy operations. This flag is enabled by default at @option{-O} and | |
5025 | higher. | |
5026 | ||
5027 | @item -ftree-store-copy-prop | |
5028 | Perform copy propagation of memory loads and stores. This pass | |
5029 | eliminates unnecessary copy operations in memory references | |
5030 | (structures, global variables, arrays, etc). This flag is enabled by | |
5031 | default at @option{-O2} and higher. | |
5032 | ||
c75ab022 DB |
5033 | @item -ftree-salias |
5034 | Perform structural alias analysis on trees. This flag | |
5035 | is enabled by default at @option{-O} and higher. | |
5036 | ||
fa555252 DB |
5037 | @item -ftree-sink |
5038 | Perform forward store motion on trees. This flag is | |
5039 | enabled by default at @option{-O} and higher. | |
5040 | ||
6de9cd9a | 5041 | @item -ftree-ccp |
0bca51f0 DN |
5042 | Perform sparse conditional constant propagation (CCP) on trees. This |
5043 | pass only operates on local scalar variables and is enabled by default | |
5044 | at @option{-O} and higher. | |
5045 | ||
5046 | @item -ftree-store-ccp | |
5047 | Perform sparse conditional constant propagation (CCP) on trees. This | |
5048 | pass operates on both local scalar variables and memory stores and | |
5049 | loads (global variables, structures, arrays, etc). This flag is | |
5050 | enabled by default at @option{-O2} and higher. | |
6de9cd9a DN |
5051 | |
5052 | @item -ftree-dce | |
5053 | Perform dead code elimination (DCE) on trees. This flag is enabled by | |
4ec7afd7 | 5054 | default at @option{-O} and higher. |
6de9cd9a DN |
5055 | |
5056 | @item -ftree-dominator-opts | |
5c1c631e DN |
5057 | Perform a variety of simple scalar cleanups (constant/copy |
5058 | propagation, redundancy elimination, range propagation and expression | |
5059 | simplification) based on a dominator tree traversal. This also | |
5060 | performs jump threading (to reduce jumps to jumps). This flag is | |
5061 | enabled by default at @option{-O} and higher. | |
6de9cd9a DN |
5062 | |
5063 | @item -ftree-ch | |
5064 | Perform loop header copying on trees. This is beneficial since it increases | |
83c99486 | 5065 | effectiveness of code motion optimizations. It also saves one jump. This flag |
4ec7afd7 KH |
5066 | is enabled by default at @option{-O} and higher. It is not enabled |
5067 | for @option{-Os}, since it usually increases code size. | |
6de9cd9a | 5068 | |
c66b6c66 | 5069 | @item -ftree-loop-optimize |
4ec7afd7 KH |
5070 | Perform loop optimizations on trees. This flag is enabled by default |
5071 | at @option{-O} and higher. | |
c66b6c66 | 5072 | |
599eabdb | 5073 | @item -ftree-loop-linear |
8a36672b | 5074 | Perform linear loop transformations on tree. This flag can improve cache |
599eabdb DB |
5075 | performance and allow further loop optimizations to take place. |
5076 | ||
3450cbc4 DJ |
5077 | @item -ftree-loop-im |
5078 | Perform loop invariant motion on trees. This pass moves only invariants that | |
ec7dea0a | 5079 | would be hard to handle at RTL level (function calls, operations that expand to |
a7e5372d ZD |
5080 | nontrivial sequences of insns). With @option{-funswitch-loops} it also moves |
5081 | operands of conditions that are invariant out of the loop, so that we can use | |
5082 | just trivial invariantness analysis in loop unswitching. The pass also includes | |
5083 | store motion. | |
5084 | ||
3450cbc4 | 5085 | @item -ftree-loop-ivcanon |
82b85a85 ZD |
5086 | Create a canonical counter for number of iterations in the loop for that |
5087 | determining number of iterations requires complicated analysis. Later | |
5088 | optimizations then may determine the number easily. Useful especially | |
5089 | in connection with unrolling. | |
5090 | ||
8b11a64c ZD |
5091 | @item -fivopts |
5092 | Perform induction variable optimizations (strength reduction, induction | |
5093 | variable merging and induction variable elimination) on trees. | |
5094 | ||
6de9cd9a DN |
5095 | @item -ftree-sra |
5096 | Perform scalar replacement of aggregates. This pass replaces structure | |
5097 | references with scalars to prevent committing structures to memory too | |
4ec7afd7 | 5098 | early. This flag is enabled by default at @option{-O} and higher. |
6de9cd9a DN |
5099 | |
5100 | @item -ftree-copyrename | |
f26c1794 EC |
5101 | Perform copy renaming on trees. This pass attempts to rename compiler |
5102 | temporaries to other variables at copy locations, usually resulting in | |
5103 | variable names which more closely resemble the original variables. This flag | |
4ec7afd7 | 5104 | is enabled by default at @option{-O} and higher. |
6de9cd9a DN |
5105 | |
5106 | @item -ftree-ter | |
5107 | Perform temporary expression replacement during the SSA->normal phase. Single | |
f26c1794 EC |
5108 | use/single def temporaries are replaced at their use location with their |
5109 | defining expression. This results in non-GIMPLE code, but gives the expanders | |
6de9cd9a | 5110 | much more complex trees to work on resulting in better RTL generation. This is |
4ec7afd7 | 5111 | enabled by default at @option{-O} and higher. |
6de9cd9a DN |
5112 | |
5113 | @item -ftree-lrs | |
f26c1794 EC |
5114 | Perform live range splitting during the SSA->normal phase. Distinct live |
5115 | ranges of a variable are split into unique variables, allowing for better | |
4ec7afd7 | 5116 | optimization later. This is enabled by default at @option{-O} and higher. |
6de9cd9a | 5117 | |
79fe1b3b DN |
5118 | @item -ftree-vectorize |
5119 | Perform loop vectorization on trees. | |
5120 | ||
c12cc930 KB |
5121 | @item -ftree-vect-loop-version |
5122 | @opindex ftree-vect-loop-version | |
5123 | Perform loop versioning when doing loop vectorization on trees. When a loop | |
5124 | appears to be vectorizable except that data alignment or data dependence cannot | |
5125 | be determined at compile time then vectorized and non-vectorized versions of | |
5126 | the loop are generated along with runtime checks for alignment or dependence | |
5127 | to control which version is executed. This option is enabled by default | |
5128 | except at level @option{-Os} where it is disabled. | |
5129 | ||
08873e96 DN |
5130 | @item -ftree-vrp |
5131 | Perform Value Range Propagation on trees. This is similar to the | |
5132 | constant propagation pass, but instead of values, ranges of values are | |
5133 | propagated. This allows the optimizers to remove unnecessary range | |
5134 | checks like array bound checks and null pointer checks. This is | |
5135 | enabled by default at @option{-O2} and higher. Null pointer check | |
5136 | elimination is only done if @option{-fdelete-null-pointer-checks} is | |
5137 | enabled. | |
5138 | ||
6de9cd9a DN |
5139 | @item -ftracer |
5140 | @opindex ftracer | |
8a36672b | 5141 | Perform tail duplication to enlarge superblock size. This transformation |
6de9cd9a DN |
5142 | simplifies the control flow of the function allowing other optimizations to do |
5143 | better job. | |
5144 | ||
5145 | @item -funroll-loops | |
5146 | @opindex funroll-loops | |
5147 | Unroll loops whose number of iterations can be determined at compile | |
5148 | time or upon entry to the loop. @option{-funroll-loops} implies both | |
5149 | @option{-fstrength-reduce} and @option{-frerun-cse-after-loop}. This | |
5150 | option makes code larger, and may or may not make it run faster. | |
5151 | ||
5152 | @item -funroll-all-loops | |
5153 | @opindex funroll-all-loops | |
5154 | Unroll all loops, even if their number of iterations is uncertain when | |
5155 | the loop is entered. This usually makes programs run more slowly. | |
5156 | @option{-funroll-all-loops} implies the same options as | |
5157 | @option{-funroll-loops}, | |
5158 | ||
113d659a ZD |
5159 | @item -fsplit-ivs-in-unroller |
5160 | @opindex -fsplit-ivs-in-unroller | |
5161 | Enables expressing of values of induction variables in later iterations | |
5162 | of the unrolled loop using the value in the first iteration. This breaks | |
7869fe47 | 5163 | long dependency chains, thus improving efficiency of the scheduling passes. |
113d659a ZD |
5164 | |
5165 | Combination of @option{-fweb} and CSE is often sufficient to obtain the | |
8a36672b | 5166 | same effect. However in cases the loop body is more complicated than |
113d659a ZD |
5167 | a single basic block, this is not reliable. It also does not work at all |
5168 | on some of the architectures due to restrictions in the CSE pass. | |
5169 | ||
5170 | This optimization is enabled by default. | |
5171 | ||
f37a4f14 RE |
5172 | @item -fvariable-expansion-in-unroller |
5173 | @opindex -fvariable-expansion-in-unroller | |
f0eb93a8 | 5174 | With this option, the compiler will create multiple copies of some |
f37a4f14 RE |
5175 | local variables when unrolling a loop which can result in superior code. |
5176 | ||
6de9cd9a DN |
5177 | @item -fprefetch-loop-arrays |
5178 | @opindex fprefetch-loop-arrays | |
5179 | If supported by the target machine, generate instructions to prefetch | |
5180 | memory to improve the performance of loops that access large arrays. | |
5181 | ||
e5eb27e5 JL |
5182 | These options may generate better or worse code; results are highly |
5183 | dependent on the structure of loops within the source code. | |
5184 | ||
74291a4b | 5185 | @item -fno-peephole |
6cfc0341 | 5186 | @itemx -fno-peephole2 |
cd3bb277 | 5187 | @opindex fno-peephole |
6cfc0341 RH |
5188 | @opindex fno-peephole2 |
5189 | Disable any machine-specific peephole optimizations. The difference | |
630d3d5a | 5190 | between @option{-fno-peephole} and @option{-fno-peephole2} is in how they |
6cfc0341 RH |
5191 | are implemented in the compiler; some targets use one, some use the |
5192 | other, a few use both. | |
861bb6c1 | 5193 | |
38df970e JQ |
5194 | @option{-fpeephole} is enabled by default. |
5195 | @option{-fpeephole2} enabled at levels @option{-O2}, @option{-O3}, @option{-Os}. | |
5196 | ||
454d0cc7 | 5197 | @item -fno-guess-branch-probability |
cd3bb277 | 5198 | @opindex fno-guess-branch-probability |
ddd8e3ca JW |
5199 | Do not guess branch probabilities using heuristics. |
5200 | ||
5201 | GCC will use heuristics to guess branch probabilities if they are | |
5202 | not provided by profiling feedback (@option{-fprofile-arcs}). These | |
5203 | heuristics are based on the control flow graph. If some branch probabilities | |
5204 | are specified by @samp{__builtin_expect}, then the heuristics will be | |
5205 | used to guess branch probabilities for the rest of the control flow graph, | |
5206 | taking the @samp{__builtin_expect} info into account. The interactions | |
5207 | between the heuristics and @samp{__builtin_expect} can be complex, and in | |
5208 | some cases, it may be useful to disable the heuristics so that the effects | |
5209 | of @samp{__builtin_expect} are easier to understand. | |
454d0cc7 | 5210 | |
38df970e JQ |
5211 | The default is @option{-fguess-branch-probability} at levels |
5212 | @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}. | |
5213 | ||
194734e9 JH |
5214 | @item -freorder-blocks |
5215 | @opindex freorder-blocks | |
5216 | Reorder basic blocks in the compiled function in order to reduce number of | |
5217 | taken branches and improve code locality. | |
5218 | ||
3f8b659d | 5219 | Enabled at levels @option{-O2}, @option{-O3}. |
38df970e | 5220 | |
750054a2 CT |
5221 | @item -freorder-blocks-and-partition |
5222 | @opindex freorder-blocks-and-partition | |
5223 | In addition to reordering basic blocks in the compiled function, in order | |
5224 | to reduce number of taken branches, partitions hot and cold basic blocks | |
5225 | into separate sections of the assembly and .o files, to improve | |
5226 | paging and cache locality performance. | |
5227 | ||
8e8d5162 CT |
5228 | This optimization is automatically turned off in the presence of |
5229 | exception handling, for linkonce sections, for functions with a user-defined | |
5230 | section attribute and on any architecture that does not support named | |
5231 | sections. | |
5232 | ||
194734e9 JH |
5233 | @item -freorder-functions |
5234 | @opindex freorder-functions | |
3e8cb558 JM |
5235 | Reorder functions in the object file in order to |
5236 | improve code locality. This is implemented by using special | |
3a4bdd05 RH |
5237 | subsections @code{.text.hot} for most frequently executed functions and |
5238 | @code{.text.unlikely} for unlikely executed functions. Reordering is done by | |
194734e9 | 5239 | the linker so object file format must support named sections and linker must |
3364c33b | 5240 | place them in a reasonable way. |
194734e9 JH |
5241 | |
5242 | Also profile feedback must be available in to make this option effective. See | |
5243 | @option{-fprofile-arcs} for details. | |
5244 | ||
38df970e JQ |
5245 | Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}. |
5246 | ||
41472af8 | 5247 | @item -fstrict-aliasing |
cd3bb277 | 5248 | @opindex fstrict-aliasing |
41472af8 MM |
5249 | Allows the compiler to assume the strictest aliasing rules applicable to |
5250 | the language being compiled. For C (and C++), this activates | |
5251 | optimizations based on the type of expressions. In particular, an | |
5252 | object of one type is assumed never to reside at the same address as an | |
5253 | object of a different type, unless the types are almost the same. For | |
5254 | example, an @code{unsigned int} can alias an @code{int}, but not a | |
5255 | @code{void*} or a @code{double}. A character type may alias any other | |
02f52e19 | 5256 | type. |
41472af8 MM |
5257 | |
5258 | Pay special attention to code like this: | |
3ab51846 | 5259 | @smallexample |
02f52e19 | 5260 | union a_union @{ |
41472af8 MM |
5261 | int i; |
5262 | double d; | |
5263 | @}; | |
5264 | ||
5265 | int f() @{ | |
5266 | a_union t; | |
5267 | t.d = 3.0; | |
5268 | return t.i; | |
5269 | @} | |
3ab51846 | 5270 | @end smallexample |
41472af8 MM |
5271 | The practice of reading from a different union member than the one most |
5272 | recently written to (called ``type-punning'') is common. Even with | |
630d3d5a | 5273 | @option{-fstrict-aliasing}, type-punning is allowed, provided the memory |
41472af8 MM |
5274 | is accessed through the union type. So, the code above will work as |
5275 | expected. However, this code might not: | |
3ab51846 | 5276 | @smallexample |
02f52e19 | 5277 | int f() @{ |
41472af8 MM |
5278 | a_union t; |
5279 | int* ip; | |
5280 | t.d = 3.0; | |
5281 | ip = &t.i; | |
5282 | return *ip; | |
5283 | @} | |
3ab51846 | 5284 | @end smallexample |
41472af8 | 5285 | |
41472af8 MM |
5286 | Every language that wishes to perform language-specific alias analysis |
5287 | should define a function that computes, given an @code{tree} | |
5288 | node, an alias set for the node. Nodes in different alias sets are not | |
5289 | allowed to alias. For an example, see the C front-end function | |
5290 | @code{c_get_alias_set}. | |
41472af8 | 5291 | |
38df970e | 5292 | Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}. |
852b81bb | 5293 | |
efa3896a GK |
5294 | @item -falign-functions |
5295 | @itemx -falign-functions=@var{n} | |
cd3bb277 | 5296 | @opindex falign-functions |
efa3896a GK |
5297 | Align the start of functions to the next power-of-two greater than |
5298 | @var{n}, skipping up to @var{n} bytes. For instance, | |
630d3d5a JM |
5299 | @option{-falign-functions=32} aligns functions to the next 32-byte |
5300 | boundary, but @option{-falign-functions=24} would align to the next | |
efa3896a GK |
5301 | 32-byte boundary only if this can be done by skipping 23 bytes or less. |
5302 | ||
630d3d5a | 5303 | @option{-fno-align-functions} and @option{-falign-functions=1} are |
efa3896a GK |
5304 | equivalent and mean that functions will not be aligned. |
5305 | ||
5306 | Some assemblers only support this flag when @var{n} is a power of two; | |
5307 | in that case, it is rounded up. | |
5308 | ||
561913cb | 5309 | If @var{n} is not specified or is zero, use a machine-dependent default. |
efa3896a | 5310 | |
38df970e JQ |
5311 | Enabled at levels @option{-O2}, @option{-O3}. |
5312 | ||
efa3896a GK |
5313 | @item -falign-labels |
5314 | @itemx -falign-labels=@var{n} | |
cd3bb277 | 5315 | @opindex falign-labels |
efa3896a | 5316 | Align all branch targets to a power-of-two boundary, skipping up to |
630d3d5a | 5317 | @var{n} bytes like @option{-falign-functions}. This option can easily |
efa3896a GK |
5318 | make code slower, because it must insert dummy operations for when the |
5319 | branch target is reached in the usual flow of the code. | |
5320 | ||
561913cb AP |
5321 | @option{-fno-align-labels} and @option{-falign-labels=1} are |
5322 | equivalent and mean that labels will not be aligned. | |
5323 | ||
630d3d5a | 5324 | If @option{-falign-loops} or @option{-falign-jumps} are applicable and |
efa3896a GK |
5325 | are greater than this value, then their values are used instead. |
5326 | ||
561913cb AP |
5327 | If @var{n} is not specified or is zero, use a machine-dependent default |
5328 | which is very likely to be @samp{1}, meaning no alignment. | |
efa3896a | 5329 | |
38df970e JQ |
5330 | Enabled at levels @option{-O2}, @option{-O3}. |
5331 | ||
efa3896a GK |
5332 | @item -falign-loops |
5333 | @itemx -falign-loops=@var{n} | |
cd3bb277 | 5334 | @opindex falign-loops |
efa3896a | 5335 | Align loops to a power-of-two boundary, skipping up to @var{n} bytes |
630d3d5a | 5336 | like @option{-falign-functions}. The hope is that the loop will be |
efa3896a GK |
5337 | executed many times, which will make up for any execution of the dummy |
5338 | operations. | |
5339 | ||
561913cb AP |
5340 | @option{-fno-align-loops} and @option{-falign-loops=1} are |
5341 | equivalent and mean that loops will not be aligned. | |
5342 | ||
5343 | If @var{n} is not specified or is zero, use a machine-dependent default. | |
efa3896a | 5344 | |
38df970e JQ |
5345 | Enabled at levels @option{-O2}, @option{-O3}. |
5346 | ||
efa3896a GK |
5347 | @item -falign-jumps |
5348 | @itemx -falign-jumps=@var{n} | |
cd3bb277 | 5349 | @opindex falign-jumps |
efa3896a GK |
5350 | Align branch targets to a power-of-two boundary, for branch targets |
5351 | where the targets can only be reached by jumping, skipping up to @var{n} | |
630d3d5a | 5352 | bytes like @option{-falign-functions}. In this case, no dummy operations |
efa3896a GK |
5353 | need be executed. |
5354 | ||
561913cb AP |
5355 | @option{-fno-align-jumps} and @option{-falign-jumps=1} are |
5356 | equivalent and mean that loops will not be aligned. | |
5357 | ||
5358 | If @var{n} is not specified or is zero, use a machine-dependent default. | |
efa3896a | 5359 | |
38df970e JQ |
5360 | Enabled at levels @option{-O2}, @option{-O3}. |
5361 | ||
7797ff53 PB |
5362 | @item -funit-at-a-time |
5363 | @opindex funit-at-a-time | |
5364 | Parse the whole compilation unit before starting to produce code. | |
5365 | This allows some extra optimizations to take place but consumes | |
5366 | more memory (in general). There are some compatibility issues | |
474eccc6 | 5367 | with @emph{unit-at-a-time} mode: |
7797ff53 PB |
5368 | @itemize @bullet |
5369 | @item | |
5370 | enabling @emph{unit-at-a-time} mode may change the order | |
5371 | in which functions, variables, and top-level @code{asm} statements | |
5372 | are emitted, and will likely break code relying on some particular | |
5373 | ordering. The majority of such top-level @code{asm} statements, | |
474eccc6 ILT |
5374 | though, can be replaced by @code{section} attributes. The |
5375 | @option{fno-toplevel-reorder} option may be used to keep the ordering | |
5376 | used in the input file, at the cost of some optimizations. | |
7797ff53 PB |
5377 | |
5378 | @item | |
5379 | @emph{unit-at-a-time} mode removes unreferenced static variables | |
60d436c9 | 5380 | and functions. This may result in undefined references |
7797ff53 PB |
5381 | when an @code{asm} statement refers directly to variables or functions |
5382 | that are otherwise unused. In that case either the variable/function | |
5383 | shall be listed as an operand of the @code{asm} statement operand or, | |
5384 | in the case of top-level @code{asm} statements the attribute @code{used} | |
5385 | shall be used on the declaration. | |
5386 | ||
5387 | @item | |
5388 | Static functions now can use non-standard passing conventions that | |
8a36672b | 5389 | may break @code{asm} statements calling functions directly. Again, |
7797ff53 PB |
5390 | attribute @code{used} will prevent this behavior. |
5391 | @end itemize | |
5392 | ||
5393 | As a temporary workaround, @option{-fno-unit-at-a-time} can be used, | |
8a36672b | 5394 | but this scheme may not be supported by future releases of GCC@. |
7797ff53 PB |
5395 | |
5396 | Enabled at levels @option{-O2}, @option{-O3}. | |
5397 | ||
474eccc6 ILT |
5398 | @item -fno-toplevel-reorder |
5399 | Do not reorder top-level functions, variables, and @code{asm} | |
5400 | statements. Output them in the same order that they appear in the | |
5401 | input file. When this option is used, unreferenced static variables | |
5402 | will not be removed. This option is intended to support existing code | |
5403 | which relies on a particular ordering. For new code, it is better to | |
5404 | use attributes. | |
5405 | ||
7260e9a0 JH |
5406 | @item -fweb |
5407 | @opindex fweb | |
5408 | Constructs webs as commonly used for register allocation purposes and assign | |
962e6e00 | 5409 | each web individual pseudo register. This allows the register allocation pass |
7260e9a0 JH |
5410 | to operate on pseudos directly, but also strengthens several other optimization |
5411 | passes, such as CSE, loop optimizer and trivial dead code remover. It can, | |
5412 | however, make debugging impossible, since variables will no longer stay in a | |
5413 | ``home register''. | |
5414 | ||
7869fe47 | 5415 | Enabled by default with @option{-funroll-loops}. |
d4463dfc | 5416 | |
ce91e74c JH |
5417 | @item -fwhole-program |
5418 | @opindex fwhole-program | |
5419 | Assume that the current compilation unit represents whole program being | |
5420 | compiled. All public functions and variables with the exception of @code{main} | |
f341de7b KH |
5421 | and those merged by attribute @code{externally_visible} become static functions |
5422 | and in a affect gets more aggressively optimized by interprocedural optimizers. | |
ce91e74c | 5423 | While this option is equivalent to proper use of @code{static} keyword for |
f341de7b | 5424 | programs consisting of single file, in combination with option |
ce91e74c JH |
5425 | @option{--combine} this flag can be used to compile most of smaller scale C |
5426 | programs since the functions and variables become local for the whole combined | |
5427 | compilation unit, not for the single source file itself. | |
5428 | ||
5429 | ||
d4463dfc JQ |
5430 | @item -fno-cprop-registers |
5431 | @opindex fno-cprop-registers | |
5432 | After register allocation and post-register allocation instruction splitting, | |
5433 | we perform a copy-propagation pass to try to reduce scheduling dependencies | |
5434 | and occasionally eliminate the copy. | |
5435 | ||
5436 | Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}. | |
5437 | ||
a8a5f53a JH |
5438 | @item -fprofile-generate |
5439 | @opindex fprofile-generate | |
a8a5f53a | 5440 | |
33adcb6c MM |
5441 | Enable options usually used for instrumenting application to produce |
5442 | profile useful for later recompilation with profile feedback based | |
4ec7afd7 | 5443 | optimization. You must use @option{-fprofile-generate} both when |
33adcb6c MM |
5444 | compiling and when linking your program. |
5445 | ||
5446 | The following options are enabled: @code{-fprofile-arcs}, @code{-fprofile-values}, @code{-fvpt}. | |
a8a5f53a JH |
5447 | |
5448 | @item -fprofile-use | |
5449 | @opindex fprofile-use | |
5450 | Enable profile feedback directed optimizations, and optimizations | |
5451 | generally profitable only with profile feedback available. | |
5452 | ||
8a76829c JH |
5453 | The following options are enabled: @code{-fbranch-probabilities}, @code{-fvpt}, |
5454 | @code{-funroll-loops}, @code{-fpeel-loops}, @code{-ftracer}, | |
5455 | @code{-fno-loop-optimize}. | |
a8a5f53a | 5456 | |
d4463dfc JQ |
5457 | @end table |
5458 | ||
5459 | The following options control compiler behavior regarding floating | |
5460 | point arithmetic. These options trade off between speed and | |
5461 | correctness. All must be specifically enabled. | |
5462 | ||
5463 | @table @gcctabopt | |
5464 | @item -ffloat-store | |
5465 | @opindex ffloat-store | |
5466 | Do not store floating point variables in registers, and inhibit other | |
5467 | options that might change whether a floating point value is taken from a | |
5468 | register or memory. | |
5469 | ||
5470 | @cindex floating point precision | |
5471 | This option prevents undesirable excess precision on machines such as | |
5472 | the 68000 where the floating registers (of the 68881) keep more | |
5473 | precision than a @code{double} is supposed to have. Similarly for the | |
5474 | x86 architecture. For most programs, the excess precision does only | |
5475 | good, but a few programs rely on the precise definition of IEEE floating | |
5476 | point. Use @option{-ffloat-store} for such programs, after modifying | |
5477 | them to store all pertinent intermediate computations into variables. | |
5478 | ||
5479 | @item -ffast-math | |
5480 | @opindex ffast-math | |
5481 | Sets @option{-fno-math-errno}, @option{-funsafe-math-optimizations}, @* | |
039c3d42 | 5482 | @option{-fno-trapping-math}, @option{-ffinite-math-only}, |
c7463669 RH |
5483 | @option{-fno-rounding-math}, @option{-fno-signaling-nans} |
5484 | and @option{fcx-limited-range}. | |
d4463dfc JQ |
5485 | |
5486 | This option causes the preprocessor macro @code{__FAST_MATH__} to be defined. | |
5487 | ||
5488 | This option should never be turned on by any @option{-O} option since | |
5489 | it can result in incorrect output for programs which depend on | |
5490 | an exact implementation of IEEE or ISO rules/specifications for | |
5491 | math functions. | |
5492 | ||
5493 | @item -fno-math-errno | |
5494 | @opindex fno-math-errno | |
5495 | Do not set ERRNO after calling math functions that are executed | |
5496 | with a single instruction, e.g., sqrt. A program that relies on | |
5497 | IEEE exceptions for math error handling may want to use this flag | |
5498 | for speed while maintaining IEEE arithmetic compatibility. | |
5499 | ||
5500 | This option should never be turned on by any @option{-O} option since | |
5501 | it can result in incorrect output for programs which depend on | |
5502 | an exact implementation of IEEE or ISO rules/specifications for | |
5503 | math functions. | |
5504 | ||
5505 | The default is @option{-fmath-errno}. | |
5506 | ||
2e3f0db6 DJ |
5507 | On Darwin systems, the math library never sets @code{errno}. There is therefore |
5508 | no reason for the compiler to consider the possibility that it might, | |
5509 | and @option{-fno-math-errno} is the default. | |
5510 | ||
d4463dfc JQ |
5511 | @item -funsafe-math-optimizations |
5512 | @opindex funsafe-math-optimizations | |
5513 | Allow optimizations for floating-point arithmetic that (a) assume | |
5514 | that arguments and results are valid and (b) may violate IEEE or | |
5515 | ANSI standards. When used at link-time, it may include libraries | |
5516 | or startup files that change the default FPU control word or other | |
5517 | similar optimizations. | |
5518 | ||
5519 | This option should never be turned on by any @option{-O} option since | |
5520 | it can result in incorrect output for programs which depend on | |
5521 | an exact implementation of IEEE or ISO rules/specifications for | |
5522 | math functions. | |
5523 | ||
5524 | The default is @option{-fno-unsafe-math-optimizations}. | |
5525 | ||
5526 | @item -ffinite-math-only | |
5527 | @opindex ffinite-math-only | |
5528 | Allow optimizations for floating-point arithmetic that assume | |
5529 | that arguments and results are not NaNs or +-Infs. | |
5530 | ||
5531 | This option should never be turned on by any @option{-O} option since | |
5532 | it can result in incorrect output for programs which depend on | |
5533 | an exact implementation of IEEE or ISO rules/specifications. | |
5534 | ||
5535 | The default is @option{-fno-finite-math-only}. | |
5536 | ||
5537 | @item -fno-trapping-math | |
5538 | @opindex fno-trapping-math | |
5539 | Compile code assuming that floating-point operations cannot generate | |
5540 | user-visible traps. These traps include division by zero, overflow, | |
5541 | underflow, inexact result and invalid operation. This option implies | |
5542 | @option{-fno-signaling-nans}. Setting this option may allow faster | |
5543 | code if one relies on ``non-stop'' IEEE arithmetic, for example. | |
5544 | ||
5545 | This option should never be turned on by any @option{-O} option since | |
5546 | it can result in incorrect output for programs which depend on | |
5547 | an exact implementation of IEEE or ISO rules/specifications for | |
5548 | math functions. | |
5549 | ||
5550 | The default is @option{-ftrapping-math}. | |
5551 | ||
039c3d42 RS |
5552 | @item -frounding-math |
5553 | @opindex frounding-math | |
5554 | Disable transformations and optimizations that assume default floating | |
5555 | point rounding behavior. This is round-to-zero for all floating point | |
5556 | to integer conversions, and round-to-nearest for all other arithmetic | |
5557 | truncations. This option should be specified for programs that change | |
5558 | the FP rounding mode dynamically, or that may be executed with a | |
5559 | non-default rounding mode. This option disables constant folding of | |
5560 | floating point expressions at compile-time (which may be affected by | |
5561 | rounding mode) and arithmetic transformations that are unsafe in the | |
5562 | presence of sign-dependent rounding modes. | |
5563 | ||
5564 | The default is @option{-fno-rounding-math}. | |
5565 | ||
5566 | This option is experimental and does not currently guarantee to | |
5567 | disable all GCC optimizations that are affected by rounding mode. | |
2dd76960 | 5568 | Future versions of GCC may provide finer control of this setting |
039c3d42 RS |
5569 | using C99's @code{FENV_ACCESS} pragma. This command line option |
5570 | will be used to specify the default state for @code{FENV_ACCESS}. | |
5571 | ||
38109dab GL |
5572 | @item -frtl-abstract-sequences |
5573 | @opindex frtl-abstract-sequences | |
5574 | It is a size optimization method. This option is to find identical | |
5575 | sequences of code, which can be turned into pseudo-procedures and | |
5576 | then replace all occurrences with calls to the newly created | |
5577 | subroutine. It is kind of an opposite of @option{-finline-functions}. | |
5578 | This optimization runs at RTL level. | |
5579 | ||
d4463dfc JQ |
5580 | @item -fsignaling-nans |
5581 | @opindex fsignaling-nans | |
5582 | Compile code assuming that IEEE signaling NaNs may generate user-visible | |
5583 | traps during floating-point operations. Setting this option disables | |
5584 | optimizations that may change the number of exceptions visible with | |
5585 | signaling NaNs. This option implies @option{-ftrapping-math}. | |
5586 | ||
5587 | This option causes the preprocessor macro @code{__SUPPORT_SNAN__} to | |
5588 | be defined. | |
5589 | ||
5590 | The default is @option{-fno-signaling-nans}. | |
5591 | ||
5592 | This option is experimental and does not currently guarantee to | |
5593 | disable all GCC optimizations that affect signaling NaN behavior. | |
5594 | ||
5595 | @item -fsingle-precision-constant | |
5596 | @opindex fsingle-precision-constant | |
5597 | Treat floating point constant as single precision constant instead of | |
5598 | implicitly converting it to double precision constant. | |
5599 | ||
c7463669 RH |
5600 | @item -fcx-limited-range |
5601 | @itemx -fno-cx-limited-range | |
5602 | @opindex fcx-limited-range | |
5603 | @opindex fno-cx-limited-range | |
5604 | When enabled, this option states that a range reduction step is not | |
5605 | needed when performing complex division. The default is | |
5606 | @option{-fno-cx-limited-range}, but is enabled by @option{-ffast-math}. | |
5607 | ||
5608 | This option controls the default setting of the ISO C99 | |
5609 | @code{CX_LIMITED_RANGE} pragma. Nevertheless, the option applies to | |
5610 | all languages. | |
d4463dfc JQ |
5611 | |
5612 | @end table | |
5613 | ||
5614 | The following options control optimizations that may improve | |
5615 | performance, but are not enabled by any @option{-O} options. This | |
5616 | section includes experimental options that may produce broken code. | |
5617 | ||
5618 | @table @gcctabopt | |
5619 | @item -fbranch-probabilities | |
5620 | @opindex fbranch-probabilities | |
5621 | After running a program compiled with @option{-fprofile-arcs} | |
5622 | (@pxref{Debugging Options,, Options for Debugging Your Program or | |
5623 | @command{gcc}}), you can compile it a second time using | |
5624 | @option{-fbranch-probabilities}, to improve optimizations based on | |
5625 | the number of times each branch was taken. When the program | |
5626 | compiled with @option{-fprofile-arcs} exits it saves arc execution | |
a4878735 | 5627 | counts to a file called @file{@var{sourcename}.gcda} for each source |
d4463dfc JQ |
5628 | file The information in this data file is very dependent on the |
5629 | structure of the generated code, so you must use the same source code | |
5630 | and the same optimization options for both compilations. | |
5631 | ||
daf2f129 | 5632 | With @option{-fbranch-probabilities}, GCC puts a |
d4463dfc JQ |
5633 | @samp{REG_BR_PROB} note on each @samp{JUMP_INSN} and @samp{CALL_INSN}. |
5634 | These can be used to improve optimization. Currently, they are only | |
5635 | used in one place: in @file{reorg.c}, instead of guessing which path a | |
5636 | branch is mostly to take, the @samp{REG_BR_PROB} values are used to | |
5637 | exactly determine which path is taken more often. | |
5638 | ||
fc5eb4a1 ZD |
5639 | @item -fprofile-values |
5640 | @opindex fprofile-values | |
5641 | If combined with @option{-fprofile-arcs}, it adds code so that some | |
5642 | data about values of expressions in the program is gathered. | |
5643 | ||
6e885ee3 ZD |
5644 | With @option{-fbranch-probabilities}, it reads back the data gathered |
5645 | from profiling values of expressions and adds @samp{REG_VALUE_PROFILE} | |
5646 | notes to instructions for their later usage in optimizations. | |
5647 | ||
7797ff53 | 5648 | Enabled with @option{-fprofile-generate} and @option{-fprofile-use}. |
8d3b3fb7 | 5649 | |
fca9dc00 ZD |
5650 | @item -fvpt |
5651 | @opindex fvpt | |
5652 | If combined with @option{-fprofile-arcs}, it instructs the compiler to add | |
5653 | a code to gather information about values of expressions. | |
5654 | ||
5655 | With @option{-fbranch-probabilities}, it reads back the data gathered | |
5656 | and actually performs the optimizations based on them. | |
5657 | Currently the optimizations include specialization of division operation | |
5658 | using the knowledge about the value of the denominator. | |
5659 | ||
a7b1dc36 PB |
5660 | @item -frename-registers |
5661 | @opindex frename-registers | |
5662 | Attempt to avoid false dependencies in scheduled code by making use | |
5663 | of registers left over after register allocation. This optimization | |
5664 | will most benefit processors with lots of registers. Depending on the | |
5665 | debug information format adopted by the target, however, it can | |
5666 | make debugging impossible, since variables will no longer stay in | |
5667 | a ``home register''. | |
5668 | ||
7869fe47 | 5669 | Enabled by default with @option{-funroll-loops}. |
a7b1dc36 | 5670 | |
d4463dfc JQ |
5671 | @item -ftracer |
5672 | @opindex ftracer | |
8a36672b | 5673 | Perform tail duplication to enlarge superblock size. This transformation |
d4463dfc JQ |
5674 | simplifies the control flow of the function allowing other optimizations to do |
5675 | better job. | |
5676 | ||
7797ff53 | 5677 | Enabled with @option{-fprofile-use}. |
8d3b3fb7 | 5678 | |
d4463dfc JQ |
5679 | @item -funroll-loops |
5680 | @opindex funroll-loops | |
b17d5d7c ZD |
5681 | Unroll loops whose number of iterations can be determined at compile time or |
5682 | upon entry to the loop. @option{-funroll-loops} implies | |
7869fe47 ZD |
5683 | @option{-frerun-cse-after-loop}, @option{-fweb} and @option{-frename-registers}. |
5684 | It also turns on complete loop peeling (i.e.@: complete removal of loops with | |
5685 | small constant number of iterations). This option makes code larger, and may | |
5686 | or may not make it run faster. | |
d4463dfc | 5687 | |
7797ff53 | 5688 | Enabled with @option{-fprofile-use}. |
8d3b3fb7 | 5689 | |
d4463dfc JQ |
5690 | @item -funroll-all-loops |
5691 | @opindex funroll-all-loops | |
5692 | Unroll all loops, even if their number of iterations is uncertain when | |
5693 | the loop is entered. This usually makes programs run more slowly. | |
5694 | @option{-funroll-all-loops} implies the same options as | |
b17d5d7c ZD |
5695 | @option{-funroll-loops}. |
5696 | ||
5697 | @item -fpeel-loops | |
5698 | @opindex fpeel-loops | |
5699 | Peels the loops for that there is enough information that they do not | |
5700 | roll much (from profile feedback). It also turns on complete loop peeling | |
431ae0bf | 5701 | (i.e.@: complete removal of loops with small constant number of iterations). |
b17d5d7c | 5702 | |
7797ff53 | 5703 | Enabled with @option{-fprofile-use}. |
8d3b3fb7 | 5704 | |
5e962776 ZD |
5705 | @item -fmove-loop-invariants |
5706 | @opindex fmove-loop-invariants | |
5707 | Enables the loop invariant motion pass in the new loop optimizer. Enabled | |
5708 | at level @option{-O1} | |
5709 | ||
b17d5d7c ZD |
5710 | @item -funswitch-loops |
5711 | @opindex funswitch-loops | |
5712 | Move branches with loop invariant conditions out of the loop, with duplicates | |
5713 | of the loop on both branches (modified according to result of the condition). | |
5714 | ||
d4463dfc JQ |
5715 | @item -fprefetch-loop-arrays |
5716 | @opindex fprefetch-loop-arrays | |
5717 | If supported by the target machine, generate instructions to prefetch | |
5718 | memory to improve the performance of loops that access large arrays. | |
5719 | ||
5720 | Disabled at level @option{-Os}. | |
5721 | ||
5722 | @item -ffunction-sections | |
5723 | @itemx -fdata-sections | |
5724 | @opindex ffunction-sections | |
5725 | @opindex fdata-sections | |
5726 | Place each function or data item into its own section in the output | |
5727 | file if the target supports arbitrary sections. The name of the | |
5728 | function or the name of the data item determines the section's name | |
5729 | in the output file. | |
5730 | ||
5731 | Use these options on systems where the linker can perform optimizations | |
f401d0f5 JDA |
5732 | to improve locality of reference in the instruction space. Most systems |
5733 | using the ELF object format and SPARC processors running Solaris 2 have | |
5734 | linkers with such optimizations. AIX may have these optimizations in | |
5735 | the future. | |
d4463dfc JQ |
5736 | |
5737 | Only use these options when there are significant benefits from doing | |
5738 | so. When you specify these options, the assembler and linker will | |
5739 | create larger object and executable files and will also be slower. | |
5740 | You will not be able to use @code{gprof} on all systems if you | |
5741 | specify this option and you may have problems with debugging if | |
5742 | you specify both this option and @option{-g}. | |
5743 | ||
fe3ad572 SC |
5744 | @item -fbranch-target-load-optimize |
5745 | @opindex fbranch-target-load-optimize | |
5746 | Perform branch target register load optimization before prologue / epilogue | |
5747 | threading. | |
5748 | The use of target registers can typically be exposed only during reload, | |
5749 | thus hoisting loads out of loops and doing inter-block scheduling needs | |
5750 | a separate optimization pass. | |
5751 | ||
5752 | @item -fbranch-target-load-optimize2 | |
5753 | @opindex fbranch-target-load-optimize2 | |
5754 | Perform branch target register load optimization after prologue / epilogue | |
5755 | threading. | |
5756 | ||
1194fc79 R |
5757 | @item -fbtr-bb-exclusive |
5758 | @opindex fbtr-bb-exclusive | |
8d3b3fb7 | 5759 | When performing branch target register load optimization, don't reuse |
1194fc79 R |
5760 | branch target registers in within any basic block. |
5761 | ||
0aca9021 JW |
5762 | @item -fstack-protector |
5763 | Emit extra code to check for buffer overflows, such as stack smashing | |
5764 | attacks. This is done by adding a guard variable to functions with | |
5765 | vulnerable objects. This includes functions that call alloca, and | |
5766 | functions with buffers larger than 8 bytes. The guards are initialized | |
5767 | when a function is entered and then checked when the function exits. | |
5768 | If a guard check fails, an error message is printed and the program exits. | |
5769 | ||
5770 | @item -fstack-protector-all | |
5771 | Like @option{-fstack-protector} except that all functions are protected. | |
5772 | ||
3af64fd6 | 5773 | @item --param @var{name}=@var{value} |
cd3bb277 | 5774 | @opindex param |
3af64fd6 MM |
5775 | In some places, GCC uses various constants to control the amount of |
5776 | optimization that is done. For example, GCC will not inline functions | |
5777 | that contain more that a certain number of instructions. You can | |
5778 | control some of these constants on the command-line using the | |
630d3d5a | 5779 | @option{--param} option. |
3af64fd6 | 5780 | |
b00567b0 ILT |
5781 | The names of specific parameters, and the meaning of the values, are |
5782 | tied to the internals of the compiler, and are subject to change | |
5783 | without notice in future releases. | |
5784 | ||
4fe9b91c | 5785 | In each case, the @var{value} is an integer. The allowable choices for |
3af64fd6 MM |
5786 | @var{name} are given in the following table: |
5787 | ||
5788 | @table @gcctabopt | |
31617ef1 DB |
5789 | @item salias-max-implicit-fields |
5790 | The maximum number of fields in a variable without direct | |
5791 | structure accesses for which structure aliasing will consider trying | |
5792 | to track each field. The default is 5 | |
5793 | ||
a916f21d RG |
5794 | @item salias-max-array-elements |
5795 | The maximum number of elements an array can have and its elements | |
5796 | still be tracked individually by structure aliasing. The default is 4 | |
5797 | ||
61b58001 BE |
5798 | @item sra-max-structure-size |
5799 | The maximum structure size, in bytes, at which the scalar replacement | |
5800 | of aggregates (SRA) optimization will perform block copies. The | |
5801 | default value, 0, implies that GCC will select the most appropriate | |
5802 | size itself. | |
5803 | ||
5804 | @item sra-field-structure-ratio | |
1c170b5e | 5805 | The threshold ratio (as a percentage) between instantiated fields and |
61b58001 BE |
5806 | the complete structure size. We say that if the ratio of the number |
5807 | of bytes in instantiated fields to the number of bytes in the complete | |
5808 | structure exceeds this parameter, then block copies are not used. The | |
5809 | default is 75. | |
5810 | ||
5f24e0dc RH |
5811 | @item max-crossjump-edges |
5812 | The maximum number of incoming edges to consider for crossjumping. | |
d203738b | 5813 | The algorithm used by @option{-fcrossjumping} is @math{O(N^2)} in |
5f24e0dc RH |
5814 | the number of edges incoming to each block. Increasing values mean |
5815 | more aggressive optimization, making the compile time increase with | |
5816 | probably small improvement in executable size. | |
5817 | ||
12183e0f PH |
5818 | @item min-crossjump-insns |
5819 | The minimum number of instructions which must be matched at the end | |
5820 | of two blocks before crossjumping will be performed on them. This | |
5821 | value is ignored in the case where all instructions in the block being | |
5822 | crossjumped from are matched. The default value is 5. | |
5823 | ||
f935b9e0 DE |
5824 | @item max-grow-copy-bb-insns |
5825 | The maximum code size expansion factor when copying basic blocks | |
5826 | instead of jumping. The expansion is relative to a jump instruction. | |
5827 | The default value is 8. | |
5828 | ||
bbcb0c05 SB |
5829 | @item max-goto-duplication-insns |
5830 | The maximum number of instructions to duplicate to a block that jumps | |
5831 | to a computed goto. To avoid @math{O(N^2)} behavior in a number of | |
5832 | passes, GCC factors computed gotos early in the compilation process, | |
5833 | and unfactors them as late as possible. Only computed jumps at the | |
5834 | end of a basic blocks with no more than max-goto-duplication-insns are | |
5835 | unfactored. The default value is 8. | |
5836 | ||
1c4c47db JO |
5837 | @item max-delay-slot-insn-search |
5838 | The maximum number of instructions to consider when looking for an | |
5839 | instruction to fill a delay slot. If more than this arbitrary number of | |
5840 | instructions is searched, the time savings from filling the delay slot | |
5841 | will be minimal so stop searching. Increasing values mean more | |
5842 | aggressive optimization, making the compile time increase with probably | |
5843 | small improvement in executable run time. | |
5844 | ||
5845 | @item max-delay-slot-live-search | |
5846 | When trying to fill delay slots, the maximum number of instructions to | |
5847 | consider when searching for a block with valid live register | |
5848 | information. Increasing this arbitrarily chosen value means more | |
5849 | aggressive optimization, increasing the compile time. This parameter | |
5850 | should be removed when the delay slot code is rewritten to maintain the | |
5851 | control-flow graph. | |
33d3b05b MM |
5852 | |
5853 | @item max-gcse-memory | |
5854 | The approximate maximum amount of memory that will be allocated in | |
5855 | order to perform the global common subexpression elimination | |
5856 | optimization. If more memory than specified is required, the | |
5857 | optimization will not be done. | |
3af64fd6 | 5858 | |
740f35a0 | 5859 | @item max-gcse-passes |
8d3b3fb7 | 5860 | The maximum number of passes of GCSE to run. The default is 1. |
740f35a0 | 5861 | |
4a121cc3 | 5862 | @item max-pending-list-length |
0c688a7d | 5863 | The maximum number of pending dependencies scheduling will allow |
4a121cc3 AM |
5864 | before flushing the current state and starting over. Large functions |
5865 | with few branches or calls can create excessively large lists which | |
5866 | needlessly consume memory and resources. | |
5867 | ||
bc522472 KG |
5868 | @item max-inline-insns-single |
5869 | Several parameters control the tree inliner used in gcc. | |
2dd76960 | 5870 | This number sets the maximum number of instructions (counted in GCC's |
daf2f129 | 5871 | internal representation) in a single function that the tree inliner |
bc522472 KG |
5872 | will consider for inlining. This only affects functions declared |
5873 | inline and methods implemented in a class declaration (C++). | |
e5c4f28a | 5874 | The default value is 450. |
bc522472 KG |
5875 | |
5876 | @item max-inline-insns-auto | |
5877 | When you use @option{-finline-functions} (included in @option{-O3}), | |
5878 | a lot of functions that would otherwise not be considered for inlining | |
5879 | by the compiler will be investigated. To those functions, a different | |
5880 | (more restrictive) limit compared to functions declared inline can | |
5881 | be applied. | |
e5c4f28a | 5882 | The default value is 90. |
bc522472 | 5883 | |
b58b1157 | 5884 | @item large-function-insns |
74515a0a RG |
5885 | The limit specifying really large functions. For functions larger than this |
5886 | limit after inlining inlining is constrained by | |
5887 | @option{--param large-function-growth}. This parameter is useful primarily | |
5888 | to avoid extreme compilation time caused by non-linear algorithms used by the | |
5889 | backend. | |
b58b1157 | 5890 | This parameter is ignored when @option{-funit-at-a-time} is not used. |
e5c4f28a | 5891 | The default value is 2700. |
b58b1157 JH |
5892 | |
5893 | @item large-function-growth | |
e53e0c56 | 5894 | Specifies maximal growth of large function caused by inlining in percents. |
b58b1157 | 5895 | This parameter is ignored when @option{-funit-at-a-time} is not used. |
74515a0a RG |
5896 | The default value is 100 which limits large function growth to 2.0 times |
5897 | the original size. | |
b58b1157 | 5898 | |
b4f32d07 JH |
5899 | @item large-unit-insns |
5900 | The limit specifying large translation unit. Growth caused by inlining of | |
5901 | units larger than this limit is limited by @option{--param inline-unit-growth}. | |
5902 | For small units this might be too tight (consider unit consisting of function A | |
5903 | that is inline and B that just calls A three time. If B is small relative to | |
5904 | A, the growth of unit is 300\% and yet such inlining is very sane. For very | |
5905 | large units consisting of small inlininable functions however the overall unit | |
5906 | growth limit is needed to avoid exponential explosion of code size. Thus for | |
5907 | smaller units, the size is increased to @option{--param large-unit-insns} | |
6416ae7f | 5908 | before applying @option{--param inline-unit-growth}. The default is 10000 |
b4f32d07 | 5909 | |
b58b1157 JH |
5910 | @item inline-unit-growth |
5911 | Specifies maximal overall growth of the compilation unit caused by inlining. | |
5912 | This parameter is ignored when @option{-funit-at-a-time} is not used. | |
74515a0a RG |
5913 | The default value is 50 which limits unit growth to 1.5 times the original |
5914 | size. | |
b58b1157 | 5915 | |
6de9cd9a DN |
5916 | @item max-inline-insns-recursive |
5917 | @itemx max-inline-insns-recursive-auto | |
5918 | Specifies maximum number of instructions out-of-line copy of self recursive inline | |
5919 | function can grow into by performing recursive inlining. | |
5920 | ||
5921 | For functions declared inline @option{--param max-inline-insns-recursive} is | |
5922 | taken into acount. For function not declared inline, recursive inlining | |
5923 | happens only when @option{-finline-functions} (included in @option{-O3}) is | |
5924 | enabled and @option{--param max-inline-insns-recursive-auto} is used. The | |
e5c4f28a | 5925 | default value is 450. |
6de9cd9a DN |
5926 | |
5927 | @item max-inline-recursive-depth | |
5928 | @itemx max-inline-recursive-depth-auto | |
5929 | Specifies maximum recursion depth used by the recursive inlining. | |
5930 | ||
5931 | For functions declared inline @option{--param max-inline-recursive-depth} is | |
5932 | taken into acount. For function not declared inline, recursive inlining | |
5933 | happens only when @option{-finline-functions} (included in @option{-O3}) is | |
5934 | enabled and @option{--param max-inline-recursive-depth-auto} is used. The | |
e5c4f28a RG |
5935 | default value is 450. |
5936 | ||
c5a4444c JH |
5937 | @item min-inline-recursive-probability |
5938 | Recursive inlining is profitable only for function having deep recursion | |
5939 | in average and can hurt for function having little recursion depth by | |
5940 | increasing the prologue size or complexity of function body to other | |
5941 | optimizers. | |
5942 | ||
5943 | When profile feedback is available (see @option{-fprofile-generate}) the actual | |
5944 | recursion depth can be guessed from probability that function will recurse via | |
5945 | given call expression. This parameter limits inlining only to call expression | |
5946 | whose probability exceeds given threshold (in percents). The default value is | |
5947 | 10. | |
5948 | ||
e5c4f28a RG |
5949 | @item inline-call-cost |
5950 | Specify cost of call instruction relative to simple arithmetics operations | |
2e0ea515 KH |
5951 | (having cost of 1). Increasing this cost disqualifies inlining of non-leaf |
5952 | functions and at the same time increases size of leaf function that is believed to | |
5953 | reduce function size by being inlined. In effect it increases amount of | |
e5c4f28a | 5954 | inlining for code having large abstraction penalty (many functions that just |
2e0ea515 KH |
5955 | pass the arguments to other functions) and decrease inlining for code with low |
5956 | abstraction penalty. The default value is 16. | |
6de9cd9a | 5957 | |
03e9dbc9 MM |
5958 | @item max-unrolled-insns |
5959 | The maximum number of instructions that a loop should have if that loop | |
5960 | is unrolled, and if the loop is unrolled, it determines how many times | |
5961 | the loop code is unrolled. | |
5962 | ||
b17d5d7c ZD |
5963 | @item max-average-unrolled-insns |
5964 | The maximum number of instructions biased by probabilities of their execution | |
5965 | that a loop should have if that loop is unrolled, and if the loop is unrolled, | |
5966 | it determines how many times the loop code is unrolled. | |
5967 | ||
5968 | @item max-unroll-times | |
5969 | The maximum number of unrollings of a single loop. | |
5970 | ||
5971 | @item max-peeled-insns | |
5972 | The maximum number of instructions that a loop should have if that loop | |
5973 | is peeled, and if the loop is peeled, it determines how many times | |
5974 | the loop code is peeled. | |
5975 | ||
5976 | @item max-peel-times | |
5977 | The maximum number of peelings of a single loop. | |
5978 | ||
5979 | @item max-completely-peeled-insns | |
5980 | The maximum number of insns of a completely peeled loop. | |
5981 | ||
5982 | @item max-completely-peel-times | |
5983 | The maximum number of iterations of a loop to be suitable for complete peeling. | |
5984 | ||
01a132bb ZD |
5985 | @item max-unswitch-insns |
5986 | The maximum number of insns of an unswitched loop. | |
5987 | ||
5988 | @item max-unswitch-level | |
5989 | The maximum number of branches unswitched in a single loop. | |
5990 | ||
a7e5372d ZD |
5991 | @item lim-expensive |
5992 | The minimum cost of an expensive expression in the loop invariant motion. | |
5993 | ||
8b11a64c ZD |
5994 | @item iv-consider-all-candidates-bound |
5995 | Bound on number of candidates for induction variables below that | |
5996 | all candidates are considered for each use in induction variable | |
5997 | optimizations. Only the most relevant candidates are considered | |
5998 | if there are more candidates, to avoid quadratic time complexity. | |
5999 | ||
6000 | @item iv-max-considered-uses | |
6001 | The induction variable optimizations give up on loops that contain more | |
6002 | induction variable uses. | |
6003 | ||
36f5ada1 ZD |
6004 | @item iv-always-prune-cand-set-bound |
6005 | If number of candidates in the set is smaller than this value, | |
6006 | we always try to remove unnecessary ivs from the set during its | |
6007 | optimization when a new iv is added to the set. | |
6008 | ||
2412d35c SP |
6009 | @item scev-max-expr-size |
6010 | Bound on size of expressions used in the scalar evolutions analyzer. | |
6011 | Large expressions slow the analyzer. | |
6012 | ||
c12cc930 | 6013 | @item vect-max-version-checks |
5116a5d2 | 6014 | The maximum number of runtime checks that can be performed when doing |
c12cc930 KB |
6015 | loop versioning in the vectorizer. See option ftree-vect-loop-version |
6016 | for more information. | |
6017 | ||
e9eb809d ZD |
6018 | @item max-iterations-to-track |
6019 | ||
6020 | The maximum number of iterations of a loop the brute force algorithm | |
6021 | for analysis of # of iterations of the loop tries to evaluate. | |
6022 | ||
194734e9 JH |
6023 | @item hot-bb-count-fraction |
6024 | Select fraction of the maximal count of repetitions of basic block in program | |
6025 | given basic block needs to have to be considered hot. | |
6026 | ||
6027 | @item hot-bb-frequency-fraction | |
6028 | Select fraction of the maximal frequency of executions of basic block in | |
6029 | function given basic block needs to have to be considered hot | |
5c856b23 | 6030 | |
95b9a3a5 JH |
6031 | @item max-predicted-iterations |
6032 | The maximum number of loop iterations we predict statically. This is useful | |
6033 | in cases where function contain single loop with known bound and other loop | |
6034 | with unknown. We predict the known number of iterations correctly, while | |
c83eecad KH |
6035 | the unknown number of iterations average to roughly 10. This means that the |
6036 | loop without bounds would appear artificially cold relative to the other one. | |
95b9a3a5 | 6037 | |
5c856b23 JH |
6038 | @item tracer-dynamic-coverage |
6039 | @itemx tracer-dynamic-coverage-feedback | |
6040 | ||
3364c33b JQ |
6041 | This value is used to limit superblock formation once the given percentage of |
6042 | executed instructions is covered. This limits unnecessary code size | |
6043 | expansion. | |
5c856b23 JH |
6044 | |
6045 | The @option{tracer-dynamic-coverage-feedback} is used only when profile | |
6046 | feedback is available. The real profiles (as opposed to statically estimated | |
6047 | ones) are much less balanced allowing the threshold to be larger value. | |
6048 | ||
6049 | @item tracer-max-code-growth | |
6050 | Stop tail duplication once code growth has reached given percentage. This is | |
3364c33b | 6051 | rather hokey argument, as most of the duplicates will be eliminated later in |
5c856b23 JH |
6052 | cross jumping, so it may be set to much higher values than is the desired code |
6053 | growth. | |
6054 | ||
6055 | @item tracer-min-branch-ratio | |
6056 | ||
6057 | Stop reverse growth when the reverse probability of best edge is less than this | |
6058 | threshold (in percent). | |
6059 | ||
6060 | @item tracer-min-branch-ratio | |
6061 | @itemx tracer-min-branch-ratio-feedback | |
6062 | ||
6063 | Stop forward growth if the best edge do have probability lower than this | |
6064 | threshold. | |
6065 | ||
3364c33b | 6066 | Similarly to @option{tracer-dynamic-coverage} two values are present, one for |
5c856b23 JH |
6067 | compilation for profile feedback and one for compilation without. The value |
6068 | for compilation with profile feedback needs to be more conservative (higher) in | |
6069 | order to make tracer effective. | |
6070 | ||
c48ec590 ZD |
6071 | @item max-cse-path-length |
6072 | ||
8d3b3fb7 | 6073 | Maximum number of basic blocks on path that cse considers. The default is 10. |
c48ec590 | 6074 | |
95b9a3a5 JH |
6075 | @item max-cse-insns |
6076 | The maximum instructions CSE process before flushing. The default is 1000. | |
6077 | ||
6de9cd9a DN |
6078 | @item global-var-threshold |
6079 | ||
8a36672b JM |
6080 | Counts the number of function calls (@var{n}) and the number of |
6081 | call-clobbered variables (@var{v}). If @var{n}x@var{v} is larger than this limit, a | |
6de9cd9a DN |
6082 | single artificial variable will be created to represent all the |
6083 | call-clobbered variables at function call sites. This artificial | |
6084 | variable will then be made to alias every call-clobbered variable. | |
8a36672b | 6085 | (done as @code{int * size_t} on the host machine; beware overflow). |
6de9cd9a DN |
6086 | |
6087 | @item max-aliased-vops | |
6088 | ||
27ef2cdd | 6089 | Maximum number of virtual operands allowed to represent aliases |
6de9cd9a DN |
6090 | before triggering the alias grouping heuristic. Alias grouping |
6091 | reduces compile times and memory consumption needed for aliasing at | |
6092 | the expense of precision loss in alias information. | |
6093 | ||
3788cc17 ZW |
6094 | @item ggc-min-expand |
6095 | ||
6096 | GCC uses a garbage collector to manage its own memory allocation. This | |
6097 | parameter specifies the minimum percentage by which the garbage | |
6098 | collector's heap should be allowed to expand between collections. | |
6099 | Tuning this may improve compilation speed; it has no effect on code | |
6100 | generation. | |
6101 | ||
9ac121af | 6102 | The default is 30% + 70% * (RAM/1GB) with an upper bound of 100% when |
8a36672b JM |
6103 | RAM >= 1GB@. If @code{getrlimit} is available, the notion of "RAM" is |
6104 | the smallest of actual RAM and @code{RLIMIT_DATA} or @code{RLIMIT_AS}. If | |
16226f1e KG |
6105 | GCC is not able to calculate RAM on a particular platform, the lower |
6106 | bound of 30% is used. Setting this parameter and | |
737c38d1 GK |
6107 | @option{ggc-min-heapsize} to zero causes a full collection to occur at |
6108 | every opportunity. This is extremely slow, but can be useful for | |
6109 | debugging. | |
3788cc17 ZW |
6110 | |
6111 | @item ggc-min-heapsize | |
6112 | ||
6113 | Minimum size of the garbage collector's heap before it begins bothering | |
6114 | to collect garbage. The first collection occurs after the heap expands | |
6115 | by @option{ggc-min-expand}% beyond @option{ggc-min-heapsize}. Again, | |
6116 | tuning this may improve compilation speed, and has no effect on code | |
6117 | generation. | |
6118 | ||
d37e6b50 GK |
6119 | The default is the smaller of RAM/8, RLIMIT_RSS, or a limit which |
6120 | tries to ensure that RLIMIT_DATA or RLIMIT_AS are not exceeded, but | |
6121 | with a lower bound of 4096 (four megabytes) and an upper bound of | |
6122 | 131072 (128 megabytes). If GCC is not able to calculate RAM on a | |
6123 | particular platform, the lower bound is used. Setting this parameter | |
6124 | very large effectively disables garbage collection. Setting this | |
6125 | parameter and @option{ggc-min-expand} to zero causes a full collection | |
6126 | to occur at every opportunity. | |
3788cc17 | 6127 | |
0bcf8261 JH |
6128 | @item max-reload-search-insns |
6129 | The maximum number of instruction reload should look backward for equivalent | |
6130 | register. Increasing values mean more aggressive optimization, making the | |
6131 | compile time increase with probably slightly better performance. The default | |
6132 | value is 100. | |
6133 | ||
c65ecebc JH |
6134 | @item max-cselib-memory-location |
6135 | The maximum number of memory locations cselib should take into acount. | |
6136 | Increasing values mean more aggressive optimization, making the compile time | |
6137 | increase with probably slightly better performance. The default value is 500. | |
6138 | ||
95b9a3a5 JH |
6139 | @item max-flow-memory-location |
6140 | Similar as @option{max-cselib-memory-location} but for dataflow liveness. | |
6141 | The default value is 100. | |
6142 | ||
26f74aa3 JH |
6143 | @item reorder-blocks-duplicate |
6144 | @itemx reorder-blocks-duplicate-feedback | |
6145 | ||
65d2d764 | 6146 | Used by basic block reordering pass to decide whether to use unconditional |
b222f49a | 6147 | branch or duplicate the code on its destination. Code is duplicated when its |
26f74aa3 | 6148 | estimated size is smaller than this value multiplied by the estimated size of |
65d2d764 | 6149 | unconditional jump in the hot spots of the program. |
26f74aa3 JH |
6150 | |
6151 | The @option{reorder-block-duplicate-feedback} is used only when profile | |
6152 | feedback is available and may be set to higher values than | |
6153 | @option{reorder-block-duplicate} since information about the hot spots is more | |
6154 | accurate. | |
f72c6b56 DE |
6155 | |
6156 | @item max-sched-region-blocks | |
6157 | The maximum number of blocks in a region to be considered for | |
6158 | interblock scheduling. The default value is 10. | |
6159 | ||
fbce7685 | 6160 | @item max-sched-region-insns |
f72c6b56 DE |
6161 | The maximum number of insns in a region to be considered for |
6162 | interblock scheduling. The default value is 100. | |
31ebd7c8 | 6163 | |
6f48c21a PS |
6164 | @item min-sched-prob |
6165 | The minimum probability of reaching a source block for interblock | |
6166 | speculative scheduling. The default value is 40. | |
6167 | ||
49c3b9a8 JJ |
6168 | @item max-last-value-rtl |
6169 | ||
6170 | The maximum size measured as number of RTLs that can be recorded in an expression | |
6171 | in combiner for a pseudo register as last known value of that register. The default | |
6172 | is 10000. | |
6173 | ||
31ebd7c8 NS |
6174 | @item integer-share-limit |
6175 | Small integer constants can use a shared data structure, reducing the | |
6176 | compiler's memory usage and increasing its speed. This sets the maximum | |
6177 | value of a shared integer constant's. The default value is 256. | |
6178 | ||
84d65814 DN |
6179 | @item min-virtual-mappings |
6180 | Specifies the minimum number of virtual mappings in the incremental | |
6181 | SSA updater that should be registered to trigger the virtual mappings | |
6182 | heuristic defined by virtual-mappings-ratio. The default value is | |
6183 | 100. | |
6184 | ||
6185 | @item virtual-mappings-ratio | |
6186 | If the number of virtual mappings is virtual-mappings-ratio bigger | |
6187 | than the number of virtual symbols to be updated, then the incremental | |
6188 | SSA updater switches to a full update for those symbols. The default | |
6189 | ratio is 3. | |
6190 | ||
0aca9021 JW |
6191 | @item ssp-buffer-size |
6192 | The minimum size of buffers (i.e. arrays) that will receive stack smashing | |
6193 | protection when @option{-fstack-protection} is used. | |
6194 | ||
43f31be5 JL |
6195 | @item max-jump-thread-duplication-stmts |
6196 | Maximum number of statements allowed in a block that needs to be | |
6197 | duplicated when threading jumps. | |
1c4c47db | 6198 | @end table |
74291a4b MM |
6199 | @end table |
6200 | ||
6201 | @node Preprocessor Options | |
6202 | @section Options Controlling the Preprocessor | |
6203 | @cindex preprocessor options | |
6204 | @cindex options, preprocessor | |
6205 | ||
6206 | These options control the C preprocessor, which is run on each C source | |
6207 | file before actual compilation. | |
6208 | ||
630d3d5a JM |
6209 | If you use the @option{-E} option, nothing is done except preprocessing. |
6210 | Some of these options make sense only together with @option{-E} because | |
74291a4b MM |
6211 | they cause the preprocessor output to be unsuitable for actual |
6212 | compilation. | |
6213 | ||
4977bab6 | 6214 | @table @gcctabopt |
cd3bb277 | 6215 | @opindex Wp |
40adaa27 NB |
6216 | You can use @option{-Wp,@var{option}} to bypass the compiler driver |
6217 | and pass @var{option} directly through to the preprocessor. If | |
6218 | @var{option} contains commas, it is split into multiple options at the | |
6219 | commas. However, many options are modified, translated or interpreted | |
6220 | by the compiler driver before being passed to the preprocessor, and | |
6221 | @option{-Wp} forcibly bypasses this phase. The preprocessor's direct | |
6222 | interface is undocumented and subject to change, so whenever possible | |
6223 | you should avoid using @option{-Wp} and let the driver handle the | |
6224 | options instead. | |
6225 | ||
4977bab6 ZW |
6226 | @item -Xpreprocessor @var{option} |
6227 | @opindex preprocessor | |
6228 | Pass @var{option} as an option to the preprocessor. You can use this to | |
6229 | supply system-specific preprocessor options which GCC does not know how to | |
6230 | recognize. | |
6231 | ||
6232 | If you want to pass an option that takes an argument, you must use | |
6233 | @option{-Xpreprocessor} twice, once for the option and once for the argument. | |
6234 | @end table | |
6235 | ||
40adaa27 | 6236 | @include cppopts.texi |
74291a4b MM |
6237 | |
6238 | @node Assembler Options | |
6239 | @section Passing Options to the Assembler | |
6240 | ||
6241 | @c prevent bad page break with this line | |
6242 | You can pass options to the assembler. | |
6243 | ||
2642624b | 6244 | @table @gcctabopt |
aee96fe9 | 6245 | @item -Wa,@var{option} |
cd3bb277 | 6246 | @opindex Wa |
74291a4b MM |
6247 | Pass @var{option} as an option to the assembler. If @var{option} |
6248 | contains commas, it is split into multiple options at the commas. | |
4977bab6 ZW |
6249 | |
6250 | @item -Xassembler @var{option} | |
6251 | @opindex Xassembler | |
6252 | Pass @var{option} as an option to the assembler. You can use this to | |
6253 | supply system-specific assembler options which GCC does not know how to | |
6254 | recognize. | |
6255 | ||
6256 | If you want to pass an option that takes an argument, you must use | |
6257 | @option{-Xassembler} twice, once for the option and once for the argument. | |
6258 | ||
74291a4b MM |
6259 | @end table |
6260 | ||
6261 | @node Link Options | |
6262 | @section Options for Linking | |
6263 | @cindex link options | |
6264 | @cindex options, linking | |
6265 | ||
6266 | These options come into play when the compiler links object files into | |
6267 | an executable output file. They are meaningless if the compiler is | |
6268 | not doing a link step. | |
6269 | ||
2642624b | 6270 | @table @gcctabopt |
74291a4b MM |
6271 | @cindex file names |
6272 | @item @var{object-file-name} | |
6273 | A file name that does not end in a special recognized suffix is | |
6274 | considered to name an object file or library. (Object files are | |
6275 | distinguished from libraries by the linker according to the file | |
6276 | contents.) If linking is done, these object files are used as input | |
6277 | to the linker. | |
6278 | ||
6279 | @item -c | |
6280 | @itemx -S | |
6281 | @itemx -E | |
cd3bb277 JM |
6282 | @opindex c |
6283 | @opindex S | |
6284 | @opindex E | |
74291a4b MM |
6285 | If any of these options is used, then the linker is not run, and |
6286 | object file names should not be used as arguments. @xref{Overall | |
6287 | Options}. | |
6288 | ||
6289 | @cindex Libraries | |
6290 | @item -l@var{library} | |
4275c4c4 | 6291 | @itemx -l @var{library} |
cd3bb277 | 6292 | @opindex l |
4275c4c4 JS |
6293 | Search the library named @var{library} when linking. (The second |
6294 | alternative with the library as a separate argument is only for | |
6295 | POSIX compliance and is not recommended.) | |
74291a4b MM |
6296 | |
6297 | It makes a difference where in the command you write this option; the | |
4275c4c4 | 6298 | linker searches and processes libraries and object files in the order they |
74291a4b MM |
6299 | are specified. Thus, @samp{foo.o -lz bar.o} searches library @samp{z} |
6300 | after file @file{foo.o} but before @file{bar.o}. If @file{bar.o} refers | |
6301 | to functions in @samp{z}, those functions may not be loaded. | |
6302 | ||
6303 | The linker searches a standard list of directories for the library, | |
6304 | which is actually a file named @file{lib@var{library}.a}. The linker | |
6305 | then uses this file as if it had been specified precisely by name. | |
6306 | ||
6307 | The directories searched include several standard system directories | |
630d3d5a | 6308 | plus any that you specify with @option{-L}. |
74291a4b MM |
6309 | |
6310 | Normally the files found this way are library files---archive files | |
6311 | whose members are object files. The linker handles an archive file by | |
6312 | scanning through it for members which define symbols that have so far | |
6313 | been referenced but not defined. But if the file that is found is an | |
6314 | ordinary object file, it is linked in the usual fashion. The only | |
630d3d5a JM |
6315 | difference between using an @option{-l} option and specifying a file name |
6316 | is that @option{-l} surrounds @var{library} with @samp{lib} and @samp{.a} | |
74291a4b MM |
6317 | and searches several directories. |
6318 | ||
6319 | @item -lobjc | |
cd3bb277 | 6320 | @opindex lobjc |
630d3d5a | 6321 | You need this special case of the @option{-l} option in order to |
46e34f96 | 6322 | link an Objective-C or Objective-C++ program. |
74291a4b MM |
6323 | |
6324 | @item -nostartfiles | |
cd3bb277 | 6325 | @opindex nostartfiles |
74291a4b | 6326 | Do not use the standard system startup files when linking. |
bedc7537 NC |
6327 | The standard system libraries are used normally, unless @option{-nostdlib} |
6328 | or @option{-nodefaultlibs} is used. | |
74291a4b MM |
6329 | |
6330 | @item -nodefaultlibs | |
cd3bb277 | 6331 | @opindex nodefaultlibs |
74291a4b MM |
6332 | Do not use the standard system libraries when linking. |
6333 | Only the libraries you specify will be passed to the linker. | |
bedc7537 | 6334 | The standard startup files are used normally, unless @option{-nostartfiles} |
8f99553f JM |
6335 | is used. The compiler may generate calls to @code{memcmp}, |
6336 | @code{memset}, @code{memcpy} and @code{memmove}. | |
6337 | These entries are usually resolved by entries in | |
4754172c CM |
6338 | libc. These entry points should be supplied through some other |
6339 | mechanism when this option is specified. | |
74291a4b MM |
6340 | |
6341 | @item -nostdlib | |
cd3bb277 | 6342 | @opindex nostdlib |
74291a4b MM |
6343 | Do not use the standard system startup files or libraries when linking. |
6344 | No startup files and only the libraries you specify will be passed to | |
8f99553f JM |
6345 | the linker. The compiler may generate calls to @code{memcmp}, @code{memset}, |
6346 | @code{memcpy} and @code{memmove}. | |
6347 | These entries are usually resolved by entries in | |
4754172c CM |
6348 | libc. These entry points should be supplied through some other |
6349 | mechanism when this option is specified. | |
74291a4b | 6350 | |
630d3d5a JM |
6351 | @cindex @option{-lgcc}, use with @option{-nostdlib} |
6352 | @cindex @option{-nostdlib} and unresolved references | |
6353 | @cindex unresolved references and @option{-nostdlib} | |
6354 | @cindex @option{-lgcc}, use with @option{-nodefaultlibs} | |
6355 | @cindex @option{-nodefaultlibs} and unresolved references | |
6356 | @cindex unresolved references and @option{-nodefaultlibs} | |
6357 | One of the standard libraries bypassed by @option{-nostdlib} and | |
6358 | @option{-nodefaultlibs} is @file{libgcc.a}, a library of internal subroutines | |
0c2d1a2a | 6359 | that GCC uses to overcome shortcomings of particular machines, or special |
74291a4b | 6360 | needs for some languages. |
b11cc610 JM |
6361 | (@xref{Interface,,Interfacing to GCC Output,gccint,GNU Compiler |
6362 | Collection (GCC) Internals}, | |
74291a4b | 6363 | for more discussion of @file{libgcc.a}.) |
74291a4b | 6364 | In most cases, you need @file{libgcc.a} even when you want to avoid |
630d3d5a JM |
6365 | other standard libraries. In other words, when you specify @option{-nostdlib} |
6366 | or @option{-nodefaultlibs} you should usually specify @option{-lgcc} as well. | |
0c2d1a2a | 6367 | This ensures that you have no unresolved references to internal GCC |
74291a4b | 6368 | library subroutines. (For example, @samp{__main}, used to ensure C++ |
b11cc610 JM |
6369 | constructors will be called; @pxref{Collect2,,@code{collect2}, gccint, |
6370 | GNU Compiler Collection (GCC) Internals}.) | |
74291a4b | 6371 | |
24a4dd31 JJ |
6372 | @item -pie |
6373 | @opindex pie | |
6374 | Produce a position independent executable on targets which support it. | |
6375 | For predictable results, you must also specify the same set of options | |
6376 | that were used to generate code (@option{-fpie}, @option{-fPIE}, | |
6377 | or model suboptions) when you specify this option. | |
6378 | ||
0cbc4d77 WB |
6379 | @item -rdynamic |
6380 | @opindex rdynamic | |
6381 | Pass the flag @option{-export-dynamic} to the ELF linker, on targets | |
6382 | that support it. This instructs the linker to add all symbols, not | |
6383 | only used ones, to the dynamic symbol table. This option is needed | |
6384 | for some uses of @code{dlopen} or to allow obtaining backtraces | |
6385 | from within a program. | |
6386 | ||
74291a4b | 6387 | @item -s |
cd3bb277 | 6388 | @opindex s |
74291a4b MM |
6389 | Remove all symbol table and relocation information from the executable. |
6390 | ||
6391 | @item -static | |
cd3bb277 | 6392 | @opindex static |
74291a4b MM |
6393 | On systems that support dynamic linking, this prevents linking with the shared |
6394 | libraries. On other systems, this option has no effect. | |
6395 | ||
6396 | @item -shared | |
cd3bb277 | 6397 | @opindex shared |
74291a4b | 6398 | Produce a shared object which can then be linked with other objects to |
1d3b0e2c | 6399 | form an executable. Not all systems support this option. For predictable |
02f52e19 | 6400 | results, you must also specify the same set of options that were used to |
630d3d5a | 6401 | generate code (@option{-fpic}, @option{-fPIC}, or model suboptions) |
2642624b | 6402 | when you specify this option.@footnote{On some systems, @samp{gcc -shared} |
767094dd | 6403 | needs to build supplementary stub code for constructors to work. On |
2642624b | 6404 | multi-libbed systems, @samp{gcc -shared} must select the correct support |
1d3b0e2c | 6405 | libraries to link against. Failing to supply the correct flags may lead |
767094dd | 6406 | to subtle defects. Supplying them in cases where they are not necessary |
1d3b0e2c | 6407 | is innocuous.} |
74291a4b | 6408 | |
9db0819e RH |
6409 | @item -shared-libgcc |
6410 | @itemx -static-libgcc | |
cd3bb277 JM |
6411 | @opindex shared-libgcc |
6412 | @opindex static-libgcc | |
9db0819e RH |
6413 | On systems that provide @file{libgcc} as a shared library, these options |
6414 | force the use of either the shared or static version respectively. | |
6415 | If no shared version of @file{libgcc} was built when the compiler was | |
6416 | configured, these options have no effect. | |
6417 | ||
6418 | There are several situations in which an application should use the | |
6419 | shared @file{libgcc} instead of the static version. The most common | |
6420 | of these is when the application wishes to throw and catch exceptions | |
6421 | across different shared libraries. In that case, each of the libraries | |
6422 | as well as the application itself should use the shared @file{libgcc}. | |
6423 | ||
5c181756 AO |
6424 | Therefore, the G++ and GCJ drivers automatically add |
6425 | @option{-shared-libgcc} whenever you build a shared library or a main | |
6426 | executable, because C++ and Java programs typically use exceptions, so | |
6427 | this is the right thing to do. | |
6428 | ||
6429 | If, instead, you use the GCC driver to create shared libraries, you may | |
6430 | find that they will not always be linked with the shared @file{libgcc}. | |
d3144423 EB |
6431 | If GCC finds, at its configuration time, that you have a non-GNU linker |
6432 | or a GNU linker that does not support option @option{--eh-frame-hdr}, | |
6433 | it will link the shared version of @file{libgcc} into shared libraries | |
6434 | by default. Otherwise, it will take advantage of the linker and optimize | |
6435 | away the linking with the shared version of @file{libgcc}, linking with | |
6436 | the static version of libgcc by default. This allows exceptions to | |
6437 | propagate through such shared libraries, without incurring relocation | |
6438 | costs at library load time. | |
5c181756 AO |
6439 | |
6440 | However, if a library or main executable is supposed to throw or catch | |
6441 | exceptions, you must link it using the G++ or GCJ driver, as appropriate | |
6442 | for the languages used in the program, or using the option | |
6443 | @option{-shared-libgcc}, such that it is linked with the shared | |
049f6ec9 | 6444 | @file{libgcc}. |
9db0819e | 6445 | |
74291a4b | 6446 | @item -symbolic |
cd3bb277 | 6447 | @opindex symbolic |
74291a4b MM |
6448 | Bind references to global symbols when building a shared object. Warn |
6449 | about any unresolved references (unless overridden by the link editor | |
6450 | option @samp{-Xlinker -z -Xlinker defs}). Only a few systems support | |
6451 | this option. | |
6452 | ||
6453 | @item -Xlinker @var{option} | |
cd3bb277 | 6454 | @opindex Xlinker |
74291a4b | 6455 | Pass @var{option} as an option to the linker. You can use this to |
0c2d1a2a | 6456 | supply system-specific linker options which GCC does not know how to |
74291a4b MM |
6457 | recognize. |
6458 | ||
6459 | If you want to pass an option that takes an argument, you must use | |
630d3d5a JM |
6460 | @option{-Xlinker} twice, once for the option and once for the argument. |
6461 | For example, to pass @option{-assert definitions}, you must write | |
74291a4b | 6462 | @samp{-Xlinker -assert -Xlinker definitions}. It does not work to write |
630d3d5a | 6463 | @option{-Xlinker "-assert definitions"}, because this passes the entire |
74291a4b MM |
6464 | string as a single argument, which is not what the linker expects. |
6465 | ||
aee96fe9 | 6466 | @item -Wl,@var{option} |
cd3bb277 | 6467 | @opindex Wl |
74291a4b MM |
6468 | Pass @var{option} as an option to the linker. If @var{option} contains |
6469 | commas, it is split into multiple options at the commas. | |
6470 | ||
6471 | @item -u @var{symbol} | |
cd3bb277 | 6472 | @opindex u |
74291a4b | 6473 | Pretend the symbol @var{symbol} is undefined, to force linking of |
630d3d5a | 6474 | library modules to define it. You can use @option{-u} multiple times with |
74291a4b MM |
6475 | different symbols to force loading of additional library modules. |
6476 | @end table | |
6477 | ||
6478 | @node Directory Options | |
6479 | @section Options for Directory Search | |
6480 | @cindex directory options | |
6481 | @cindex options, directory search | |
6482 | @cindex search path | |
6483 | ||
6484 | These options specify directories to search for header files, for | |
6485 | libraries and for parts of the compiler: | |
6486 | ||
2642624b | 6487 | @table @gcctabopt |
74291a4b | 6488 | @item -I@var{dir} |
cd3bb277 | 6489 | @opindex I |
861bb6c1 JL |
6490 | Add the directory @var{dir} to the head of the list of directories to be |
6491 | searched for header files. This can be used to override a system header | |
6492 | file, substituting your own version, since these directories are | |
d0a5eb32 RK |
6493 | searched before the system header file directories. However, you should |
6494 | not use this option to add directories that contain vendor-supplied | |
767094dd | 6495 | system header files (use @option{-isystem} for that). If you use more than |
630d3d5a | 6496 | one @option{-I} option, the directories are scanned in left-to-right |
74291a4b MM |
6497 | order; the standard system directories come after. |
6498 | ||
dbead49c | 6499 | If a standard system include directory, or a directory specified with |
48209ce5 JDA |
6500 | @option{-isystem}, is also specified with @option{-I}, the @option{-I} |
6501 | option will be ignored. The directory will still be searched but as a | |
6502 | system directory at its normal position in the system include chain. | |
6503 | This is to ensure that GCC's procedure to fix buggy system headers and | |
3364c33b | 6504 | the ordering for the include_next directive are not inadvertently changed. |
48209ce5 JDA |
6505 | If you really need to change the search order for system directories, |
6506 | use the @option{-nostdinc} and/or @option{-isystem} options. | |
dbead49c | 6507 | |
4bed3787 MS |
6508 | @item -iquote@var{dir} |
6509 | @opindex iquote | |
6510 | Add the directory @var{dir} to the head of the list of directories to | |
6511 | be searched for header files only for the case of @samp{#include | |
6512 | "@var{file}"}; they are not searched for @samp{#include <@var{file}>}, | |
6513 | otherwise just like @option{-I}. | |
74291a4b MM |
6514 | |
6515 | @item -L@var{dir} | |
cd3bb277 | 6516 | @opindex L |
74291a4b | 6517 | Add directory @var{dir} to the list of directories to be searched |
630d3d5a | 6518 | for @option{-l}. |
74291a4b MM |
6519 | |
6520 | @item -B@var{prefix} | |
cd3bb277 | 6521 | @opindex B |
74291a4b MM |
6522 | This option specifies where to find the executables, libraries, |
6523 | include files, and data files of the compiler itself. | |
6524 | ||
6525 | The compiler driver program runs one or more of the subprograms | |
6526 | @file{cpp}, @file{cc1}, @file{as} and @file{ld}. It tries | |
6527 | @var{prefix} as a prefix for each program it tries to run, both with and | |
6528 | without @samp{@var{machine}/@var{version}/} (@pxref{Target Options}). | |
6529 | ||
6530 | For each subprogram to be run, the compiler driver first tries the | |
630d3d5a | 6531 | @option{-B} prefix, if any. If that name is not found, or if @option{-B} |
74291a4b | 6532 | was not specified, the driver tries two standard prefixes, which are |
8e5f33ff | 6533 | @file{/usr/lib/gcc/} and @file{/usr/local/lib/gcc/}. If neither of |
74291a4b MM |
6534 | those results in a file name that is found, the unmodified program |
6535 | name is searched for using the directories specified in your | |
bedc7537 | 6536 | @env{PATH} environment variable. |
74291a4b | 6537 | |
07804c3b NC |
6538 | The compiler will check to see if the path provided by the @option{-B} |
6539 | refers to a directory, and if necessary it will add a directory | |
6540 | separator character at the end of the path. | |
6541 | ||
630d3d5a | 6542 | @option{-B} prefixes that effectively specify directory names also apply |
74291a4b | 6543 | to libraries in the linker, because the compiler translates these |
630d3d5a | 6544 | options into @option{-L} options for the linker. They also apply to |
74291a4b | 6545 | includes files in the preprocessor, because the compiler translates these |
630d3d5a | 6546 | options into @option{-isystem} options for the preprocessor. In this case, |
74291a4b MM |
6547 | the compiler appends @samp{include} to the prefix. |
6548 | ||
6549 | The run-time support file @file{libgcc.a} can also be searched for using | |
630d3d5a | 6550 | the @option{-B} prefix, if needed. If it is not found there, the two |
74291a4b MM |
6551 | standard prefixes above are tried, and that is all. The file is left |
6552 | out of the link if it is not found by those means. | |
6553 | ||
630d3d5a | 6554 | Another way to specify a prefix much like the @option{-B} prefix is to use |
bedc7537 | 6555 | the environment variable @env{GCC_EXEC_PREFIX}. @xref{Environment |
74291a4b | 6556 | Variables}. |
861bb6c1 | 6557 | |
07804c3b | 6558 | As a special kludge, if the path provided by @option{-B} is |
bf4eebe0 NC |
6559 | @file{[dir/]stage@var{N}/}, where @var{N} is a number in the range 0 to |
6560 | 9, then it will be replaced by @file{[dir/]include}. This is to help | |
7dac2f89 | 6561 | with boot-strapping the compiler. |
07804c3b | 6562 | |
861bb6c1 | 6563 | @item -specs=@var{file} |
cd3bb277 | 6564 | @opindex specs |
861bb6c1 JL |
6565 | Process @var{file} after the compiler reads in the standard @file{specs} |
6566 | file, in order to override the defaults that the @file{gcc} driver | |
6567 | program uses when determining what switches to pass to @file{cc1}, | |
6568 | @file{cc1plus}, @file{as}, @file{ld}, etc. More than one | |
630d3d5a | 6569 | @option{-specs=@var{file}} can be specified on the command line, and they |
861bb6c1 | 6570 | are processed in order, from left to right. |
4bed3787 | 6571 | |
160633c6 MM |
6572 | @item --sysroot=@var{dir} |
6573 | @opindex sysroot | |
6574 | Use @var{dir} as the logical root directory for headers and libraries. | |
6575 | For example, if the compiler would normally search for headers in | |
6576 | @file{/usr/include} and libraries in @file{/usr/lib}, it will instead | |
6577 | search @file{@var{dir}/usr/include} and @file{@var{dir}/usr/lib}. | |
6578 | ||
6579 | If you use both this option and the @option{-isysroot} option, then | |
6580 | the @option{--sysroot} option will apply to libraries, but the | |
6581 | @option{-isysroot} option will apply to header files. | |
6582 | ||
6583 | The GNU linker (beginning with version 2.16) has the necessary support | |
6584 | for this option. If your linker does not support this option, the | |
6585 | header file aspect of @option{--sysroot} will still work, but the | |
6586 | library aspect will not. | |
6587 | ||
4bed3787 MS |
6588 | @item -I- |
6589 | @opindex I- | |
6590 | This option has been deprecated. Please use @option{-iquote} instead for | |
6591 | @option{-I} directories before the @option{-I-} and remove the @option{-I-}. | |
6592 | Any directories you specify with @option{-I} options before the @option{-I-} | |
6593 | option are searched only for the case of @samp{#include "@var{file}"}; | |
6594 | they are not searched for @samp{#include <@var{file}>}. | |
6595 | ||
6596 | If additional directories are specified with @option{-I} options after | |
6597 | the @option{-I-}, these directories are searched for all @samp{#include} | |
6598 | directives. (Ordinarily @emph{all} @option{-I} directories are used | |
6599 | this way.) | |
6600 | ||
6601 | In addition, the @option{-I-} option inhibits the use of the current | |
6602 | directory (where the current input file came from) as the first search | |
6603 | directory for @samp{#include "@var{file}"}. There is no way to | |
6604 | override this effect of @option{-I-}. With @option{-I.} you can specify | |
6605 | searching the directory which was current when the compiler was | |
6606 | invoked. That is not exactly the same as what the preprocessor does | |
6607 | by default, but it is often satisfactory. | |
6608 | ||
6609 | @option{-I-} does not inhibit the use of the standard system directories | |
6610 | for header files. Thus, @option{-I-} and @option{-nostdinc} are | |
6611 | independent. | |
74291a4b MM |
6612 | @end table |
6613 | ||
ee457005 JM |
6614 | @c man end |
6615 | ||
a743d340 NC |
6616 | @node Spec Files |
6617 | @section Specifying subprocesses and the switches to pass to them | |
6618 | @cindex Spec Files | |
d2d42a91 | 6619 | |
bedc7537 | 6620 | @command{gcc} is a driver program. It performs its job by invoking a |
a743d340 NC |
6621 | sequence of other programs to do the work of compiling, assembling and |
6622 | linking. GCC interprets its command-line parameters and uses these to | |
6623 | deduce which programs it should invoke, and which command-line options | |
c21cd8b1 | 6624 | it ought to place on their command lines. This behavior is controlled |
a743d340 NC |
6625 | by @dfn{spec strings}. In most cases there is one spec string for each |
6626 | program that GCC can invoke, but a few programs have multiple spec | |
c21cd8b1 | 6627 | strings to control their behavior. The spec strings built into GCC can |
630d3d5a | 6628 | be overridden by using the @option{-specs=} command-line switch to specify |
a743d340 NC |
6629 | a spec file. |
6630 | ||
6631 | @dfn{Spec files} are plaintext files that are used to construct spec | |
6632 | strings. They consist of a sequence of directives separated by blank | |
6633 | lines. The type of directive is determined by the first non-whitespace | |
6634 | character on the line and it can be one of the following: | |
6635 | ||
6636 | @table @code | |
6637 | @item %@var{command} | |
6638 | Issues a @var{command} to the spec file processor. The commands that can | |
02f52e19 | 6639 | appear here are: |
a743d340 NC |
6640 | |
6641 | @table @code | |
6642 | @item %include <@var{file}> | |
6643 | @cindex %include | |
6644 | Search for @var{file} and insert its text at the current point in the | |
6645 | specs file. | |
6646 | ||
6647 | @item %include_noerr <@var{file}> | |
6648 | @cindex %include_noerr | |
6649 | Just like @samp{%include}, but do not generate an error message if the include | |
6650 | file cannot be found. | |
6651 | ||
6652 | @item %rename @var{old_name} @var{new_name} | |
6653 | @cindex %rename | |
6654 | Rename the spec string @var{old_name} to @var{new_name}. | |
6655 | ||
6656 | @end table | |
6657 | ||
6658 | @item *[@var{spec_name}]: | |
6659 | This tells the compiler to create, override or delete the named spec | |
6660 | string. All lines after this directive up to the next directive or | |
6661 | blank line are considered to be the text for the spec string. If this | |
6662 | results in an empty string then the spec will be deleted. (Or, if the | |
6663 | spec did not exist, then nothing will happened.) Otherwise, if the spec | |
6664 | does not currently exist a new spec will be created. If the spec does | |
6665 | exist then its contents will be overridden by the text of this | |
6666 | directive, unless the first character of that text is the @samp{+} | |
6667 | character, in which case the text will be appended to the spec. | |
6668 | ||
6669 | @item [@var{suffix}]: | |
6670 | Creates a new @samp{[@var{suffix}] spec} pair. All lines after this directive | |
6671 | and up to the next directive or blank line are considered to make up the | |
02f52e19 | 6672 | spec string for the indicated suffix. When the compiler encounters an |
a743d340 NC |
6673 | input file with the named suffix, it will processes the spec string in |
6674 | order to work out how to compile that file. For example: | |
6675 | ||
6676 | @smallexample | |
6677 | .ZZ: | |
6678 | z-compile -input %i | |
6679 | @end smallexample | |
6680 | ||
6681 | This says that any input file whose name ends in @samp{.ZZ} should be | |
6682 | passed to the program @samp{z-compile}, which should be invoked with the | |
630d3d5a | 6683 | command-line switch @option{-input} and with the result of performing the |
a743d340 NC |
6684 | @samp{%i} substitution. (See below.) |
6685 | ||
6686 | As an alternative to providing a spec string, the text that follows a | |
6687 | suffix directive can be one of the following: | |
6688 | ||
6689 | @table @code | |
6690 | @item @@@var{language} | |
6691 | This says that the suffix is an alias for a known @var{language}. This is | |
bedc7537 | 6692 | similar to using the @option{-x} command-line switch to GCC to specify a |
a743d340 NC |
6693 | language explicitly. For example: |
6694 | ||
6695 | @smallexample | |
6696 | .ZZ: | |
6697 | @@c++ | |
6698 | @end smallexample | |
6699 | ||
6700 | Says that .ZZ files are, in fact, C++ source files. | |
6701 | ||
6702 | @item #@var{name} | |
6703 | This causes an error messages saying: | |
6704 | ||
6705 | @smallexample | |
6706 | @var{name} compiler not installed on this system. | |
6707 | @end smallexample | |
6708 | @end table | |
6709 | ||
6710 | GCC already has an extensive list of suffixes built into it. | |
6711 | This directive will add an entry to the end of the list of suffixes, but | |
6712 | since the list is searched from the end backwards, it is effectively | |
6713 | possible to override earlier entries using this technique. | |
6714 | ||
6715 | @end table | |
6716 | ||
6717 | GCC has the following spec strings built into it. Spec files can | |
6718 | override these strings or create their own. Note that individual | |
02f52e19 | 6719 | targets can also add their own spec strings to this list. |
a743d340 NC |
6720 | |
6721 | @smallexample | |
6722 | asm Options to pass to the assembler | |
6723 | asm_final Options to pass to the assembler post-processor | |
6724 | cpp Options to pass to the C preprocessor | |
6725 | cc1 Options to pass to the C compiler | |
6726 | cc1plus Options to pass to the C++ compiler | |
6727 | endfile Object files to include at the end of the link | |
6728 | link Options to pass to the linker | |
6729 | lib Libraries to include on the command line to the linker | |
6730 | libgcc Decides which GCC support library to pass to the linker | |
6731 | linker Sets the name of the linker | |
6732 | predefines Defines to be passed to the C preprocessor | |
310668e8 JM |
6733 | signed_char Defines to pass to CPP to say whether @code{char} is signed |
6734 | by default | |
a743d340 NC |
6735 | startfile Object files to include at the start of the link |
6736 | @end smallexample | |
6737 | ||
6738 | Here is a small example of a spec file: | |
6739 | ||
6740 | @smallexample | |
6741 | %rename lib old_lib | |
6742 | ||
6743 | *lib: | |
6744 | --start-group -lgcc -lc -leval1 --end-group %(old_lib) | |
6745 | @end smallexample | |
6746 | ||
6747 | This example renames the spec called @samp{lib} to @samp{old_lib} and | |
6748 | then overrides the previous definition of @samp{lib} with a new one. | |
6749 | The new definition adds in some extra command-line options before | |
6750 | including the text of the old definition. | |
6751 | ||
6752 | @dfn{Spec strings} are a list of command-line options to be passed to their | |
6753 | corresponding program. In addition, the spec strings can contain | |
6754 | @samp{%}-prefixed sequences to substitute variable text or to | |
6755 | conditionally insert text into the command line. Using these constructs | |
6756 | it is possible to generate quite complex command lines. | |
6757 | ||
6758 | Here is a table of all defined @samp{%}-sequences for spec | |
6759 | strings. Note that spaces are not generated automatically around the | |
6760 | results of expanding these sequences. Therefore you can concatenate them | |
02f52e19 | 6761 | together or combine them with constant text in a single argument. |
a743d340 NC |
6762 | |
6763 | @table @code | |
6764 | @item %% | |
6765 | Substitute one @samp{%} into the program name or argument. | |
6766 | ||
6767 | @item %i | |
6768 | Substitute the name of the input file being processed. | |
6769 | ||
6770 | @item %b | |
6771 | Substitute the basename of the input file being processed. | |
6772 | This is the substring up to (and not including) the last period | |
6773 | and not including the directory. | |
6774 | ||
371e300b NC |
6775 | @item %B |
6776 | This is the same as @samp{%b}, but include the file suffix (text after | |
6777 | the last period). | |
6778 | ||
a743d340 NC |
6779 | @item %d |
6780 | Marks the argument containing or following the @samp{%d} as a | |
6781 | temporary file name, so that that file will be deleted if GCC exits | |
6782 | successfully. Unlike @samp{%g}, this contributes no text to the | |
02f52e19 | 6783 | argument. |
a743d340 NC |
6784 | |
6785 | @item %g@var{suffix} | |
6786 | Substitute a file name that has suffix @var{suffix} and is chosen | |
6787 | once per compilation, and mark the argument in the same way as | |
6788 | @samp{%d}. To reduce exposure to denial-of-service attacks, the file | |
02f52e19 | 6789 | name is now chosen in a way that is hard to predict even when previously |
695ac33f | 6790 | chosen file names are known. For example, @samp{%g.s @dots{} %g.o @dots{} %g.s} |
a743d340 NC |
6791 | might turn into @samp{ccUVUUAU.s ccXYAXZ12.o ccUVUUAU.s}. @var{suffix} matches |
6792 | the regexp @samp{[.A-Za-z]*} or the special string @samp{%O}, which is | |
6793 | treated exactly as if @samp{%O} had been preprocessed. Previously, @samp{%g} | |
6794 | was simply substituted with a file name chosen once per compilation, | |
6795 | without regard to any appended suffix (which was therefore treated | |
6796 | just like ordinary text), making such attacks more likely to succeed. | |
6797 | ||
6798 | @item %u@var{suffix} | |
6799 | Like @samp{%g}, but generates a new temporary file name even if | |
6800 | @samp{%u@var{suffix}} was already seen. | |
6801 | ||
6802 | @item %U@var{suffix} | |
6803 | Substitutes the last file name generated with @samp{%u@var{suffix}}, generating a | |
6804 | new one if there is no such last file name. In the absence of any | |
6805 | @samp{%u@var{suffix}}, this is just like @samp{%g@var{suffix}}, except they don't share | |
695ac33f | 6806 | the same suffix @emph{space}, so @samp{%g.s @dots{} %U.s @dots{} %g.s @dots{} %U.s} |
a743d340 NC |
6807 | would involve the generation of two distinct file names, one |
6808 | for each @samp{%g.s} and another for each @samp{%U.s}. Previously, @samp{%U} was | |
6809 | simply substituted with a file name chosen for the previous @samp{%u}, | |
6810 | without regard to any appended suffix. | |
6811 | ||
4977bab6 | 6812 | @item %j@var{suffix} |
aee96fe9 | 6813 | Substitutes the name of the @code{HOST_BIT_BUCKET}, if any, and if it is |
371e300b NC |
6814 | writable, and if save-temps is off; otherwise, substitute the name |
6815 | of a temporary file, just like @samp{%u}. This temporary file is not | |
6816 | meant for communication between processes, but rather as a junk | |
6817 | disposal mechanism. | |
6818 | ||
4977bab6 ZW |
6819 | @item %|@var{suffix} |
6820 | @itemx %m@var{suffix} | |
6821 | Like @samp{%g}, except if @option{-pipe} is in effect. In that case | |
6822 | @samp{%|} substitutes a single dash and @samp{%m} substitutes nothing at | |
6823 | all. These are the two most common ways to instruct a program that it | |
6824 | should read from standard input or write to standard output. If you | |
6825 | need something more elaborate you can use an @samp{%@{pipe:@code{X}@}} | |
6826 | construct: see for example @file{f/lang-specs.h}. | |
6827 | ||
371e300b NC |
6828 | @item %.@var{SUFFIX} |
6829 | Substitutes @var{.SUFFIX} for the suffixes of a matched switch's args | |
767094dd | 6830 | when it is subsequently output with @samp{%*}. @var{SUFFIX} is |
371e300b NC |
6831 | terminated by the next space or %. |
6832 | ||
a743d340 NC |
6833 | @item %w |
6834 | Marks the argument containing or following the @samp{%w} as the | |
6835 | designated output file of this compilation. This puts the argument | |
6836 | into the sequence of arguments that @samp{%o} will substitute later. | |
6837 | ||
6838 | @item %o | |
6839 | Substitutes the names of all the output files, with spaces | |
6840 | automatically placed around them. You should write spaces | |
6841 | around the @samp{%o} as well or the results are undefined. | |
6842 | @samp{%o} is for use in the specs for running the linker. | |
6843 | Input files whose names have no recognized suffix are not compiled | |
6844 | at all, but they are included among the output files, so they will | |
6845 | be linked. | |
6846 | ||
6847 | @item %O | |
6848 | Substitutes the suffix for object files. Note that this is | |
6849 | handled specially when it immediately follows @samp{%g, %u, or %U}, | |
6850 | because of the need for those to form complete file names. The | |
6851 | handling is such that @samp{%O} is treated exactly as if it had already | |
6852 | been substituted, except that @samp{%g, %u, and %U} do not currently | |
6853 | support additional @var{suffix} characters following @samp{%O} as they would | |
6854 | following, for example, @samp{.o}. | |
6855 | ||
6856 | @item %p | |
6857 | Substitutes the standard macro predefinitions for the | |
6858 | current target machine. Use this when running @code{cpp}. | |
6859 | ||
6860 | @item %P | |
6861 | Like @samp{%p}, but puts @samp{__} before and after the name of each | |
6862 | predefined macro, except for macros that start with @samp{__} or with | |
c1030c7c | 6863 | @samp{_@var{L}}, where @var{L} is an uppercase letter. This is for ISO |
161d7b59 | 6864 | C@. |
a743d340 NC |
6865 | |
6866 | @item %I | |
047d636f | 6867 | Substitute any of @option{-iprefix} (made from @env{GCC_EXEC_PREFIX}), |
2b6dd222 | 6868 | @option{-isysroot} (made from @env{TARGET_SYSTEM_ROOT}), |
047d636f | 6869 | @option{-isystem} (made from @env{COMPILER_PATH} and @option{-B} options) |
2b6dd222 | 6870 | and @option{-imultilib} as necessary. |
a743d340 NC |
6871 | |
6872 | @item %s | |
6873 | Current argument is the name of a library or startup file of some sort. | |
6874 | Search for that file in a standard list of directories and substitute | |
02f52e19 | 6875 | the full name found. |
a743d340 NC |
6876 | |
6877 | @item %e@var{str} | |
6878 | Print @var{str} as an error message. @var{str} is terminated by a newline. | |
6879 | Use this when inconsistent options are detected. | |
6880 | ||
a743d340 NC |
6881 | @item %(@var{name}) |
6882 | Substitute the contents of spec string @var{name} at this point. | |
6883 | ||
6884 | @item %[@var{name}] | |
630d3d5a | 6885 | Like @samp{%(@dots{})} but put @samp{__} around @option{-D} arguments. |
a743d340 NC |
6886 | |
6887 | @item %x@{@var{option}@} | |
6888 | Accumulate an option for @samp{%X}. | |
6889 | ||
6890 | @item %X | |
630d3d5a | 6891 | Output the accumulated linker options specified by @option{-Wl} or a @samp{%x} |
a743d340 NC |
6892 | spec string. |
6893 | ||
6894 | @item %Y | |
630d3d5a | 6895 | Output the accumulated assembler options specified by @option{-Wa}. |
a743d340 NC |
6896 | |
6897 | @item %Z | |
630d3d5a | 6898 | Output the accumulated preprocessor options specified by @option{-Wp}. |
a743d340 | 6899 | |
a743d340 NC |
6900 | @item %a |
6901 | Process the @code{asm} spec. This is used to compute the | |
6902 | switches to be passed to the assembler. | |
6903 | ||
6904 | @item %A | |
6905 | Process the @code{asm_final} spec. This is a spec string for | |
6906 | passing switches to an assembler post-processor, if such a program is | |
6907 | needed. | |
6908 | ||
6909 | @item %l | |
6910 | Process the @code{link} spec. This is the spec for computing the | |
6911 | command line passed to the linker. Typically it will make use of the | |
6912 | @samp{%L %G %S %D and %E} sequences. | |
6913 | ||
6914 | @item %D | |
630d3d5a | 6915 | Dump out a @option{-L} option for each directory that GCC believes might |
a743d340 | 6916 | contain startup files. If the target supports multilibs then the |
02f52e19 | 6917 | current multilib directory will be prepended to each of these paths. |
a743d340 NC |
6918 | |
6919 | @item %L | |
6920 | Process the @code{lib} spec. This is a spec string for deciding which | |
02f52e19 | 6921 | libraries should be included on the command line to the linker. |
a743d340 NC |
6922 | |
6923 | @item %G | |
6924 | Process the @code{libgcc} spec. This is a spec string for deciding | |
02f52e19 | 6925 | which GCC support library should be included on the command line to the linker. |
a743d340 NC |
6926 | |
6927 | @item %S | |
6928 | Process the @code{startfile} spec. This is a spec for deciding which | |
6929 | object files should be the first ones passed to the linker. Typically | |
02f52e19 | 6930 | this might be a file named @file{crt0.o}. |
a743d340 NC |
6931 | |
6932 | @item %E | |
6933 | Process the @code{endfile} spec. This is a spec string that specifies | |
02f52e19 | 6934 | the last object files that will be passed to the linker. |
a743d340 NC |
6935 | |
6936 | @item %C | |
6937 | Process the @code{cpp} spec. This is used to construct the arguments | |
6938 | to be passed to the C preprocessor. | |
6939 | ||
a743d340 NC |
6940 | @item %1 |
6941 | Process the @code{cc1} spec. This is used to construct the options to be | |
6942 | passed to the actual C compiler (@samp{cc1}). | |
6943 | ||
6944 | @item %2 | |
6945 | Process the @code{cc1plus} spec. This is used to construct the options to be | |
6946 | passed to the actual C++ compiler (@samp{cc1plus}). | |
6947 | ||
6948 | @item %* | |
6949 | Substitute the variable part of a matched option. See below. | |
6950 | Note that each comma in the substituted string is replaced by | |
6951 | a single space. | |
6952 | ||
4977bab6 ZW |
6953 | @item %<@code{S} |
6954 | Remove all occurrences of @code{-S} from the command line. Note---this | |
6955 | command is position dependent. @samp{%} commands in the spec string | |
6956 | before this one will see @code{-S}, @samp{%} commands in the spec string | |
6957 | after this one will not. | |
6958 | ||
f3226a90 JT |
6959 | @item %:@var{function}(@var{args}) |
6960 | Call the named function @var{function}, passing it @var{args}. | |
6961 | @var{args} is first processed as a nested spec string, then split | |
6962 | into an argument vector in the usual fashion. The function returns | |
6963 | a string which is processed as if it had appeared literally as part | |
6964 | of the current spec. | |
6965 | ||
6966 | The following built-in spec functions are provided: | |
6967 | ||
6968 | @table @code | |
6969 | @item @code{if-exists} | |
6970 | The @code{if-exists} spec function takes one argument, an absolute | |
6971 | pathname to a file. If the file exists, @code{if-exists} returns the | |
6972 | pathname. Here is a small example of its usage: | |
6973 | ||
6974 | @smallexample | |
6975 | *startfile: | |
6976 | crt0%O%s %:if-exists(crti%O%s) crtbegin%O%s | |
6977 | @end smallexample | |
152a5a9c JT |
6978 | |
6979 | @item @code{if-exists-else} | |
6980 | The @code{if-exists-else} spec function is similar to the @code{if-exists} | |
6981 | spec function, except that it takes two arguments. The first argument is | |
6982 | an absolute pathname to a file. If the file exists, @code{if-exists-else} | |
6983 | returns the pathname. If it does not exist, it returns the second argument. | |
6984 | This way, @code{if-exists-else} can be used to select one file or another, | |
6985 | based on the existence of the first. Here is a small example of its usage: | |
6986 | ||
daf2f129 | 6987 | @smallexample |
152a5a9c | 6988 | *startfile: |
f5034c5e JM |
6989 | crt0%O%s %:if-exists(crti%O%s) \ |
6990 | %:if-exists-else(crtbeginT%O%s crtbegin%O%s) | |
152a5a9c | 6991 | @end smallexample |
3dd53121 AP |
6992 | |
6993 | @item @code{replace-outfile} | |
6994 | The @code{replace-outfile} spec function takes two arguments. It looks for the | |
6995 | first argument in the outfiles array and replaces it with the second argument. Here | |
6996 | is a small example of its usage: | |
6997 | ||
6998 | @smallexample | |
6999 | %@{fgnu-runtime:%:replace-outfile(-lobjc -lobjc-gnu)@} | |
7000 | @end smallexample | |
7001 | ||
daf2f129 | 7002 | @end table |
f3226a90 | 7003 | |
a743d340 | 7004 | @item %@{@code{S}@} |
161d7b59 | 7005 | Substitutes the @code{-S} switch, if that switch was given to GCC@. |
a743d340 NC |
7006 | If that switch was not specified, this substitutes nothing. Note that |
7007 | the leading dash is omitted when specifying this option, and it is | |
7008 | automatically inserted if the substitution is performed. Thus the spec | |
630d3d5a JM |
7009 | string @samp{%@{foo@}} would match the command-line option @option{-foo} |
7010 | and would output the command line option @option{-foo}. | |
a743d340 NC |
7011 | |
7012 | @item %W@{@code{S}@} | |
7013 | Like %@{@code{S}@} but mark last argument supplied within as a file to be | |
02f52e19 | 7014 | deleted on failure. |
a743d340 NC |
7015 | |
7016 | @item %@{@code{S}*@} | |
7017 | Substitutes all the switches specified to GCC whose names start | |
7018 | with @code{-S}, but which also take an argument. This is used for | |
695ac33f | 7019 | switches like @option{-o}, @option{-D}, @option{-I}, etc. |
630d3d5a | 7020 | GCC considers @option{-o foo} as being |
a743d340 | 7021 | one switch whose names starts with @samp{o}. %@{o*@} would substitute this |
02f52e19 | 7022 | text, including the space. Thus two arguments would be generated. |
a743d340 | 7023 | |
371e300b NC |
7024 | @item %@{@code{S}*&@code{T}*@} |
7025 | Like %@{@code{S}*@}, but preserve order of @code{S} and @code{T} options | |
7026 | (the order of @code{S} and @code{T} in the spec is not significant). | |
7027 | There can be any number of ampersand-separated variables; for each the | |
7028 | wild card is optional. Useful for CPP as @samp{%@{D*&U*&A*@}}. | |
7029 | ||
a743d340 | 7030 | @item %@{@code{S}:@code{X}@} |
4977bab6 | 7031 | Substitutes @code{X}, if the @samp{-S} switch was given to GCC@. |
a743d340 NC |
7032 | |
7033 | @item %@{!@code{S}:@code{X}@} | |
4977bab6 | 7034 | Substitutes @code{X}, if the @samp{-S} switch was @emph{not} given to GCC@. |
a743d340 | 7035 | |
4977bab6 ZW |
7036 | @item %@{@code{S}*:@code{X}@} |
7037 | Substitutes @code{X} if one or more switches whose names start with | |
7038 | @code{-S} are specified to GCC@. Normally @code{X} is substituted only | |
7039 | once, no matter how many such switches appeared. However, if @code{%*} | |
7040 | appears somewhere in @code{X}, then @code{X} will be substituted once | |
7041 | for each matching switch, with the @code{%*} replaced by the part of | |
7042 | that switch that matched the @code{*}. | |
a743d340 NC |
7043 | |
7044 | @item %@{.@code{S}:@code{X}@} | |
4977bab6 | 7045 | Substitutes @code{X}, if processing a file with suffix @code{S}. |
a743d340 NC |
7046 | |
7047 | @item %@{!.@code{S}:@code{X}@} | |
4977bab6 | 7048 | Substitutes @code{X}, if @emph{not} processing a file with suffix @code{S}. |
a743d340 NC |
7049 | |
7050 | @item %@{@code{S}|@code{P}:@code{X}@} | |
4977bab6 ZW |
7051 | Substitutes @code{X} if either @code{-S} or @code{-P} was given to GCC@. |
7052 | This may be combined with @samp{!}, @samp{.}, and @code{*} sequences as well, | |
7053 | although they have a stronger binding than the @samp{|}. If @code{%*} | |
7054 | appears in @code{X}, all of the alternatives must be starred, and only | |
7055 | the first matching alternative is substituted. | |
7056 | ||
7057 | For example, a spec string like this: | |
a743d340 NC |
7058 | |
7059 | @smallexample | |
7060 | %@{.c:-foo@} %@{!.c:-bar@} %@{.c|d:-baz@} %@{!.c|d:-boggle@} | |
7061 | @end smallexample | |
7062 | ||
7063 | will output the following command-line options from the following input | |
7064 | command-line options: | |
7065 | ||
7066 | @smallexample | |
7067 | fred.c -foo -baz | |
7068 | jim.d -bar -boggle | |
7069 | -d fred.c -foo -baz -boggle | |
7070 | -d jim.d -bar -baz -boggle | |
7071 | @end smallexample | |
7072 | ||
4977bab6 ZW |
7073 | @item %@{S:X; T:Y; :D@} |
7074 | ||
c0cbdbd9 KH |
7075 | If @code{S} was given to GCC, substitutes @code{X}; else if @code{T} was |
7076 | given to GCC, substitutes @code{Y}; else substitutes @code{D}. There can | |
daf2f129 | 7077 | be as many clauses as you need. This may be combined with @code{.}, |
4977bab6 ZW |
7078 | @code{!}, @code{|}, and @code{*} as needed. |
7079 | ||
7080 | ||
a743d340 NC |
7081 | @end table |
7082 | ||
4977bab6 ZW |
7083 | The conditional text @code{X} in a %@{@code{S}:@code{X}@} or similar |
7084 | construct may contain other nested @samp{%} constructs or spaces, or | |
7085 | even newlines. They are processed as usual, as described above. | |
7086 | Trailing white space in @code{X} is ignored. White space may also | |
7087 | appear anywhere on the left side of the colon in these constructs, | |
7088 | except between @code{.} or @code{*} and the corresponding word. | |
a743d340 | 7089 | |
4977bab6 ZW |
7090 | The @option{-O}, @option{-f}, @option{-m}, and @option{-W} switches are |
7091 | handled specifically in these constructs. If another value of | |
7092 | @option{-O} or the negated form of a @option{-f}, @option{-m}, or | |
7093 | @option{-W} switch is found later in the command line, the earlier | |
7094 | switch value is ignored, except with @{@code{S}*@} where @code{S} is | |
7095 | just one letter, which passes all matching options. | |
a743d340 | 7096 | |
4977bab6 ZW |
7097 | The character @samp{|} at the beginning of the predicate text is used to |
7098 | indicate that a command should be piped to the following command, but | |
7099 | only if @option{-pipe} is specified. | |
a743d340 NC |
7100 | |
7101 | It is built into GCC which switches take arguments and which do not. | |
7102 | (You might think it would be useful to generalize this to allow each | |
7103 | compiler's spec to say which switches take arguments. But this cannot | |
7104 | be done in a consistent fashion. GCC cannot even decide which input | |
7105 | files have been specified without knowing which switches take arguments, | |
7106 | and it must know which input files to compile in order to tell which | |
02f52e19 | 7107 | compilers to run). |
a743d340 | 7108 | |
630d3d5a | 7109 | GCC also knows implicitly that arguments starting in @option{-l} are to be |
a743d340 NC |
7110 | treated as compiler output files, and passed to the linker in their |
7111 | proper position among the other output files. | |
7112 | ||
ee457005 JM |
7113 | @c man begin OPTIONS |
7114 | ||
74291a4b MM |
7115 | @node Target Options |
7116 | @section Specifying Target Machine and Compiler Version | |
7117 | @cindex target options | |
7118 | @cindex cross compiling | |
7119 | @cindex specifying machine version | |
7120 | @cindex specifying compiler version and target machine | |
7121 | @cindex compiler version, specifying | |
7122 | @cindex target machine, specifying | |
7123 | ||
37a4aa31 GK |
7124 | The usual way to run GCC is to run the executable called @file{gcc}, or |
7125 | @file{<machine>-gcc} when cross-compiling, or | |
7126 | @file{<machine>-gcc-<version>} to run a version other than the one that | |
7127 | was installed last. Sometimes this is inconvenient, so GCC provides | |
7128 | options that will switch to another cross-compiler or version. | |
74291a4b | 7129 | |
2642624b | 7130 | @table @gcctabopt |
74291a4b | 7131 | @item -b @var{machine} |
cd3bb277 | 7132 | @opindex b |
74291a4b | 7133 | The argument @var{machine} specifies the target machine for compilation. |
74291a4b MM |
7134 | |
7135 | The value to use for @var{machine} is the same as was specified as the | |
0c2d1a2a | 7136 | machine type when configuring GCC as a cross-compiler. For |
74291a4b | 7137 | example, if a cross-compiler was configured with @samp{configure |
3300bf07 PG |
7138 | arm-elf}, meaning to compile for an arm processor with elf binaries, |
7139 | then you would specify @option{-b arm-elf} to run that cross compiler. | |
7140 | Because there are other options beginning with @option{-b}, the | |
7141 | configuration must contain a hyphen. | |
74291a4b | 7142 | |
37a4aa31 GK |
7143 | @item -V @var{version} |
7144 | @opindex V | |
7145 | The argument @var{version} specifies which version of GCC to run. | |
7146 | This is useful when multiple versions are installed. For example, | |
3300bf07 | 7147 | @var{version} might be @samp{4.0}, meaning to run GCC version 4.0. |
74291a4b MM |
7148 | @end table |
7149 | ||
37a4aa31 GK |
7150 | The @option{-V} and @option{-b} options work by running the |
7151 | @file{<machine>-gcc-<version>} executable, so there's no real reason to | |
7152 | use them if you can just run that directly. | |
74291a4b MM |
7153 | |
7154 | @node Submodel Options | |
7155 | @section Hardware Models and Configurations | |
7156 | @cindex submodel options | |
7157 | @cindex specifying hardware config | |
7158 | @cindex hardware models and configurations, specifying | |
7159 | @cindex machine dependent options | |
7160 | ||
630d3d5a | 7161 | Earlier we discussed the standard option @option{-b} which chooses among |
74291a4b | 7162 | different installed compilers for completely different target |
8aeea6e6 | 7163 | machines, such as VAX vs.@: 68000 vs.@: 80386. |
74291a4b MM |
7164 | |
7165 | In addition, each of these target machine types can have its own | |
7166 | special options, starting with @samp{-m}, to choose among various | |
7167 | hardware models or configurations---for example, 68010 vs 68020, | |
7168 | floating coprocessor or none. A single installed version of the | |
7169 | compiler can compile for any model or configuration, according to the | |
7170 | options specified. | |
7171 | ||
7172 | Some configurations of the compiler also support additional special | |
7173 | options, usually for compatibility with other compilers on the same | |
7174 | platform. | |
7175 | ||
39bc1876 NS |
7176 | @c This list is ordered alphanumerically by subsection name. |
7177 | @c It should be the same order and spelling as these options are listed | |
7178 | @c in Machine Dependent Options | |
7179 | ||
74291a4b | 7180 | @menu |
39bc1876 | 7181 | * ARC Options:: |
74291a4b | 7182 | * ARM Options:: |
39bc1876 | 7183 | * AVR Options:: |
0d4a78eb | 7184 | * Blackfin Options:: |
39bc1876 | 7185 | * CRIS Options:: |
53054e77 | 7186 | * CRX Options:: |
48aec0bc | 7187 | * Darwin Options:: |
74291a4b | 7188 | * DEC Alpha Options:: |
d7c23cdc | 7189 | * DEC Alpha/VMS Options:: |
39bc1876 | 7190 | * FRV Options:: |
74291a4b | 7191 | * H8/300 Options:: |
39bc1876 NS |
7192 | * HPPA Options:: |
7193 | * i386 and x86-64 Options:: | |
7194 | * IA-64 Options:: | |
38b2d076 | 7195 | * M32C Options:: |
39bc1876 NS |
7196 | * M32R/D Options:: |
7197 | * M680x0 Options:: | |
7198 | * M68hc1x Options:: | |
7199 | * MCore Options:: | |
7200 | * MIPS Options:: | |
7201 | * MMIX Options:: | |
7202 | * MN10300 Options:: | |
6c9ac67a | 7203 | * MT Options:: |
39bc1876 NS |
7204 | * PDP-11 Options:: |
7205 | * PowerPC Options:: | |
7206 | * RS/6000 and PowerPC Options:: | |
7207 | * S/390 and zSeries Options:: | |
74291a4b | 7208 | * SH Options:: |
39bc1876 | 7209 | * SPARC Options:: |
74291a4b | 7210 | * System V Options:: |
282a61e6 | 7211 | * TMS320C3x/C4x Options:: |
f84271d9 | 7212 | * V850 Options:: |
39bc1876 NS |
7213 | * VAX Options:: |
7214 | * x86-64 Options:: | |
69a0611f | 7215 | * Xstormy16 Options:: |
03984308 | 7216 | * Xtensa Options:: |
39bc1876 | 7217 | * zSeries Options:: |
74291a4b MM |
7218 | @end menu |
7219 | ||
39bc1876 NS |
7220 | @node ARC Options |
7221 | @subsection ARC Options | |
7222 | @cindex ARC Options | |
74291a4b | 7223 | |
39bc1876 | 7224 | These options are defined for ARC implementations: |
74291a4b | 7225 | |
2642624b | 7226 | @table @gcctabopt |
39bc1876 NS |
7227 | @item -EL |
7228 | @opindex EL | |
7229 | Compile code for little endian mode. This is the default. | |
74cf1c6d | 7230 | |
39bc1876 NS |
7231 | @item -EB |
7232 | @opindex EB | |
7233 | Compile code for big endian mode. | |
74291a4b | 7234 | |
39bc1876 NS |
7235 | @item -mmangle-cpu |
7236 | @opindex mmangle-cpu | |
7237 | Prepend the name of the cpu to all public symbol names. | |
7238 | In multiple-processor systems, there are many ARC variants with different | |
7239 | instruction and register set characteristics. This flag prevents code | |
7240 | compiled for one cpu to be linked with code compiled for another. | |
7241 | No facility exists for handling variants that are ``almost identical''. | |
7242 | This is an all or nothing option. | |
74291a4b | 7243 | |
39bc1876 NS |
7244 | @item -mcpu=@var{cpu} |
7245 | @opindex mcpu | |
7246 | Compile code for ARC variant @var{cpu}. | |
7247 | Which variants are supported depend on the configuration. | |
7248 | All variants support @option{-mcpu=base}, this is the default. | |
74291a4b | 7249 | |
39bc1876 NS |
7250 | @item -mtext=@var{text-section} |
7251 | @itemx -mdata=@var{data-section} | |
7252 | @itemx -mrodata=@var{readonly-data-section} | |
7253 | @opindex mtext | |
7254 | @opindex mdata | |
7255 | @opindex mrodata | |
7256 | Put functions, data, and readonly data in @var{text-section}, | |
7257 | @var{data-section}, and @var{readonly-data-section} respectively | |
7258 | by default. This can be overridden with the @code{section} attribute. | |
7259 | @xref{Variable Attributes}. | |
74291a4b | 7260 | |
39bc1876 | 7261 | @end table |
74291a4b | 7262 | |
39bc1876 NS |
7263 | @node ARM Options |
7264 | @subsection ARM Options | |
7265 | @cindex ARM options | |
74291a4b | 7266 | |
39bc1876 NS |
7267 | These @samp{-m} options are defined for Advanced RISC Machines (ARM) |
7268 | architectures: | |
74cf1c6d | 7269 | |
39bc1876 NS |
7270 | @table @gcctabopt |
7271 | @item -mabi=@var{name} | |
7272 | @opindex mabi | |
8a36672b | 7273 | Generate code for the specified ABI@. Permissible values are: @samp{apcs-gnu}, |
077fc835 | 7274 | @samp{atpcs}, @samp{aapcs}, @samp{aapcs-linux} and @samp{iwmmxt}. |
74cf1c6d | 7275 | |
39bc1876 NS |
7276 | @item -mapcs-frame |
7277 | @opindex mapcs-frame | |
7278 | Generate a stack frame that is compliant with the ARM Procedure Call | |
7279 | Standard for all functions, even if this is not strictly necessary for | |
7280 | correct execution of the code. Specifying @option{-fomit-frame-pointer} | |
7281 | with this option will cause the stack frames not to be generated for | |
7282 | leaf functions. The default is @option{-mno-apcs-frame}. | |
74291a4b | 7283 | |
39bc1876 NS |
7284 | @item -mapcs |
7285 | @opindex mapcs | |
7286 | This is a synonym for @option{-mapcs-frame}. | |
74291a4b | 7287 | |
39bc1876 NS |
7288 | @ignore |
7289 | @c not currently implemented | |
7290 | @item -mapcs-stack-check | |
7291 | @opindex mapcs-stack-check | |
7292 | Generate code to check the amount of stack space available upon entry to | |
7293 | every function (that actually uses some stack space). If there is | |
7294 | insufficient space available then either the function | |
7295 | @samp{__rt_stkovf_split_small} or @samp{__rt_stkovf_split_big} will be | |
7296 | called, depending upon the amount of stack space required. The run time | |
7297 | system is required to provide these functions. The default is | |
7298 | @option{-mno-apcs-stack-check}, since this produces smaller code. | |
74cf1c6d | 7299 | |
39bc1876 NS |
7300 | @c not currently implemented |
7301 | @item -mapcs-float | |
7302 | @opindex mapcs-float | |
7303 | Pass floating point arguments using the float point registers. This is | |
7304 | one of the variants of the APCS@. This option is recommended if the | |
7305 | target hardware has a floating point unit or if a lot of floating point | |
7306 | arithmetic is going to be performed by the code. The default is | |
7307 | @option{-mno-apcs-float}, since integer only code is slightly increased in | |
7308 | size if @option{-mapcs-float} is used. | |
74291a4b | 7309 | |
39bc1876 NS |
7310 | @c not currently implemented |
7311 | @item -mapcs-reentrant | |
7312 | @opindex mapcs-reentrant | |
7313 | Generate reentrant, position independent code. The default is | |
7314 | @option{-mno-apcs-reentrant}. | |
7315 | @end ignore | |
74291a4b | 7316 | |
39bc1876 NS |
7317 | @item -mthumb-interwork |
7318 | @opindex mthumb-interwork | |
7319 | Generate code which supports calling between the ARM and Thumb | |
7320 | instruction sets. Without this option the two instruction sets cannot | |
7321 | be reliably used inside one program. The default is | |
7322 | @option{-mno-thumb-interwork}, since slightly larger code is generated | |
7323 | when @option{-mthumb-interwork} is specified. | |
7324 | ||
7325 | @item -mno-sched-prolog | |
7326 | @opindex mno-sched-prolog | |
7327 | Prevent the reordering of instructions in the function prolog, or the | |
7328 | merging of those instruction with the instructions in the function's | |
7329 | body. This means that all functions will start with a recognizable set | |
7330 | of instructions (or in fact one of a choice from a small set of | |
7331 | different function prologues), and this information can be used to | |
7332 | locate the start if functions inside an executable piece of code. The | |
7333 | default is @option{-msched-prolog}. | |
7334 | ||
7335 | @item -mhard-float | |
7336 | @opindex mhard-float | |
7337 | Generate output containing floating point instructions. This is the | |
7338 | default. | |
861bb6c1 | 7339 | |
74291a4b | 7340 | @item -msoft-float |
cd3bb277 | 7341 | @opindex msoft-float |
74291a4b | 7342 | Generate output containing library calls for floating point. |
39bc1876 | 7343 | @strong{Warning:} the requisite libraries are not available for all ARM |
74291a4b | 7344 | targets. Normally the facilities of the machine's usual C compiler are |
39bc1876 NS |
7345 | used, but this cannot be done directly in cross-compilation. You must make |
7346 | your own arrangements to provide suitable library functions for | |
7347 | cross-compilation. | |
74291a4b | 7348 | |
39bc1876 NS |
7349 | @option{-msoft-float} changes the calling convention in the output file; |
7350 | therefore, it is only useful if you compile @emph{all} of a program with | |
7351 | this option. In particular, you need to compile @file{libgcc.a}, the | |
7352 | library that comes with GCC, with @option{-msoft-float} in order for | |
7353 | this to work. | |
74291a4b | 7354 | |
39bc1876 NS |
7355 | @item -mfloat-abi=@var{name} |
7356 | @opindex mfloat-abi | |
7357 | Specifies which ABI to use for floating point values. Permissible values | |
7358 | are: @samp{soft}, @samp{softfp} and @samp{hard}. | |
74291a4b | 7359 | |
39bc1876 NS |
7360 | @samp{soft} and @samp{hard} are equivalent to @option{-msoft-float} |
7361 | and @option{-mhard-float} respectively. @samp{softfp} allows the generation | |
7362 | of floating point instructions, but still uses the soft-float calling | |
7363 | conventions. | |
74291a4b | 7364 | |
39bc1876 NS |
7365 | @item -mlittle-endian |
7366 | @opindex mlittle-endian | |
7367 | Generate code for a processor running in little-endian mode. This is | |
7368 | the default for all standard configurations. | |
74291a4b | 7369 | |
39bc1876 NS |
7370 | @item -mbig-endian |
7371 | @opindex mbig-endian | |
7372 | Generate code for a processor running in big-endian mode; the default is | |
7373 | to compile code for a little-endian processor. | |
74291a4b | 7374 | |
39bc1876 NS |
7375 | @item -mwords-little-endian |
7376 | @opindex mwords-little-endian | |
7377 | This option only applies when generating code for big-endian processors. | |
7378 | Generate code for a little-endian word order but a big-endian byte | |
7379 | order. That is, a byte order of the form @samp{32107654}. Note: this | |
7380 | option should only be used if you require compatibility with code for | |
7381 | big-endian ARM processors generated by versions of the compiler prior to | |
7382 | 2.8. | |
74291a4b | 7383 | |
39bc1876 NS |
7384 | @item -mcpu=@var{name} |
7385 | @opindex mcpu | |
7386 | This specifies the name of the target ARM processor. GCC uses this name | |
7387 | to determine what kind of instructions it can emit when generating | |
7388 | assembly code. Permissible names are: @samp{arm2}, @samp{arm250}, | |
7389 | @samp{arm3}, @samp{arm6}, @samp{arm60}, @samp{arm600}, @samp{arm610}, | |
7390 | @samp{arm620}, @samp{arm7}, @samp{arm7m}, @samp{arm7d}, @samp{arm7dm}, | |
7391 | @samp{arm7di}, @samp{arm7dmi}, @samp{arm70}, @samp{arm700}, | |
7392 | @samp{arm700i}, @samp{arm710}, @samp{arm710c}, @samp{arm7100}, | |
d98a72fd RE |
7393 | @samp{arm7500}, @samp{arm7500fe}, @samp{arm7tdmi}, @samp{arm7tdmi-s}, |
7394 | @samp{arm8}, @samp{strongarm}, @samp{strongarm110}, @samp{strongarm1100}, | |
39bc1876 | 7395 | @samp{arm8}, @samp{arm810}, @samp{arm9}, @samp{arm9e}, @samp{arm920}, |
d98a72fd RE |
7396 | @samp{arm920t}, @samp{arm922t}, @samp{arm946e-s}, @samp{arm966e-s}, |
7397 | @samp{arm968e-s}, @samp{arm926ej-s}, @samp{arm940t}, @samp{arm9tdmi}, | |
7398 | @samp{arm10tdmi}, @samp{arm1020t}, @samp{arm1026ej-s}, | |
f9e8581a | 7399 | @samp{arm10e}, @samp{arm1020e}, @samp{arm1022e}, |
fa91adc6 PB |
7400 | @samp{arm1136j-s}, @samp{arm1136jf-s}, @samp{mpcore}, @samp{mpcorenovfp}, |
7401 | @samp{arm1176jz-s}, @samp{arm1176jzf-s}, @samp{xscale}, @samp{iwmmxt}, | |
39bc1876 | 7402 | @samp{ep9312}. |
74291a4b | 7403 | |
39bc1876 NS |
7404 | @itemx -mtune=@var{name} |
7405 | @opindex mtune | |
7406 | This option is very similar to the @option{-mcpu=} option, except that | |
7407 | instead of specifying the actual target processor type, and hence | |
7408 | restricting which instructions can be used, it specifies that GCC should | |
7409 | tune the performance of the code as if the target were of the type | |
7410 | specified in this option, but still choosing the instructions that it | |
7411 | will generate based on the cpu specified by a @option{-mcpu=} option. | |
7412 | For some ARM implementations better performance can be obtained by using | |
7413 | this option. | |
861bb6c1 | 7414 | |
39bc1876 NS |
7415 | @item -march=@var{name} |
7416 | @opindex march | |
7417 | This specifies the name of the target ARM architecture. GCC uses this | |
7418 | name to determine what kind of instructions it can emit when generating | |
7419 | assembly code. This option can be used in conjunction with or instead | |
7420 | of the @option{-mcpu=} option. Permissible names are: @samp{armv2}, | |
7421 | @samp{armv2a}, @samp{armv3}, @samp{armv3m}, @samp{armv4}, @samp{armv4t}, | |
7422 | @samp{armv5}, @samp{armv5t}, @samp{armv5te}, @samp{armv6}, @samp{armv6j}, | |
7423 | @samp{iwmmxt}, @samp{ep9312}. | |
861bb6c1 | 7424 | |
39bc1876 NS |
7425 | @item -mfpu=@var{name} |
7426 | @itemx -mfpe=@var{number} | |
7427 | @itemx -mfp=@var{number} | |
7428 | @opindex mfpu | |
7429 | @opindex mfpe | |
7430 | @opindex mfp | |
7431 | This specifies what floating point hardware (or hardware emulation) is | |
7432 | available on the target. Permissible names are: @samp{fpa}, @samp{fpe2}, | |
7433 | @samp{fpe3}, @samp{maverick}, @samp{vfp}. @option{-mfp} and @option{-mfpe} | |
7434 | are synonyms for @option{-mfpu}=@samp{fpe}@var{number}, for compatibility | |
7435 | with older versions of GCC@. | |
861bb6c1 | 7436 | |
39bc1876 NS |
7437 | If @option{-msoft-float} is specified this specifies the format of |
7438 | floating point values. | |
fb868474 | 7439 | |
39bc1876 NS |
7440 | @item -mstructure-size-boundary=@var{n} |
7441 | @opindex mstructure-size-boundary | |
7442 | The size of all structures and unions will be rounded up to a multiple | |
7443 | of the number of bits set by this option. Permissible values are 8, 32 | |
7444 | and 64. The default value varies for different toolchains. For the COFF | |
7445 | targeted toolchain the default value is 8. A value of 64 is only allowed | |
7446 | if the underlying ABI supports it. | |
b71733d5 | 7447 | |
39bc1876 NS |
7448 | Specifying the larger number can produce faster, more efficient code, but |
7449 | can also increase the size of the program. Different values are potentially | |
7450 | incompatible. Code compiled with one value cannot necessarily expect to | |
7451 | work with code or libraries compiled with another value, if they exchange | |
7452 | information using structures or unions. | |
24f9c4df | 7453 | |
39bc1876 NS |
7454 | @item -mabort-on-noreturn |
7455 | @opindex mabort-on-noreturn | |
7456 | Generate a call to the function @code{abort} at the end of a | |
7457 | @code{noreturn} function. It will be executed if the function tries to | |
7458 | return. | |
24f9c4df | 7459 | |
39bc1876 NS |
7460 | @item -mlong-calls |
7461 | @itemx -mno-long-calls | |
7462 | @opindex mlong-calls | |
7463 | @opindex mno-long-calls | |
7464 | Tells the compiler to perform function calls by first loading the | |
7465 | address of the function into a register and then performing a subroutine | |
7466 | call on this register. This switch is needed if the target function | |
7467 | will lie outside of the 64 megabyte addressing range of the offset based | |
7468 | version of subroutine call instruction. | |
24f9c4df | 7469 | |
39bc1876 NS |
7470 | Even if this switch is enabled, not all function calls will be turned |
7471 | into long calls. The heuristic is that static functions, functions | |
7472 | which have the @samp{short-call} attribute, functions that are inside | |
7473 | the scope of a @samp{#pragma no_long_calls} directive and functions whose | |
7474 | definitions have already been compiled within the current compilation | |
7475 | unit, will not be turned into long calls. The exception to this rule is | |
7476 | that weak function definitions, functions with the @samp{long-call} | |
7477 | attribute or the @samp{section} attribute, and functions that are within | |
7478 | the scope of a @samp{#pragma long_calls} directive, will always be | |
7479 | turned into long calls. | |
24f9c4df | 7480 | |
39bc1876 NS |
7481 | This feature is not enabled by default. Specifying |
7482 | @option{-mno-long-calls} will restore the default behavior, as will | |
7483 | placing the function calls within the scope of a @samp{#pragma | |
7484 | long_calls_off} directive. Note these switches have no effect on how | |
7485 | the compiler generates code to handle function calls via function | |
7486 | pointers. | |
24f9c4df | 7487 | |
39bc1876 NS |
7488 | @item -mnop-fun-dllimport |
7489 | @opindex mnop-fun-dllimport | |
7490 | Disable support for the @code{dllimport} attribute. | |
74291a4b | 7491 | |
39bc1876 NS |
7492 | @item -msingle-pic-base |
7493 | @opindex msingle-pic-base | |
7494 | Treat the register used for PIC addressing as read-only, rather than | |
7495 | loading it in the prologue for each function. The run-time system is | |
7496 | responsible for initializing this register with an appropriate value | |
7497 | before execution begins. | |
2856c3e3 | 7498 | |
39bc1876 NS |
7499 | @item -mpic-register=@var{reg} |
7500 | @opindex mpic-register | |
7501 | Specify the register to be used for PIC addressing. The default is R10 | |
7502 | unless stack-checking is enabled, when R9 is used. | |
2856c3e3 | 7503 | |
39bc1876 NS |
7504 | @item -mcirrus-fix-invalid-insns |
7505 | @opindex mcirrus-fix-invalid-insns | |
7506 | @opindex mno-cirrus-fix-invalid-insns | |
7507 | Insert NOPs into the instruction stream to in order to work around | |
7508 | problems with invalid Maverick instruction combinations. This option | |
7509 | is only valid if the @option{-mcpu=ep9312} option has been used to | |
7510 | enable generation of instructions for the Cirrus Maverick floating | |
7511 | point co-processor. This option is not enabled by default, since the | |
7512 | problem is only present in older Maverick implementations. The default | |
7513 | can be re-enabled by use of the @option{-mno-cirrus-fix-invalid-insns} | |
7514 | switch. | |
2856c3e3 | 7515 | |
39bc1876 NS |
7516 | @item -mpoke-function-name |
7517 | @opindex mpoke-function-name | |
7518 | Write the name of each function into the text section, directly | |
7519 | preceding the function prologue. The generated code is similar to this: | |
2856c3e3 | 7520 | |
39bc1876 NS |
7521 | @smallexample |
7522 | t0 | |
7523 | .ascii "arm_poke_function_name", 0 | |
7524 | .align | |
7525 | t1 | |
7526 | .word 0xff000000 + (t1 - t0) | |
7527 | arm_poke_function_name | |
7528 | mov ip, sp | |
7529 | stmfd sp!, @{fp, ip, lr, pc@} | |
7530 | sub fp, ip, #4 | |
7531 | @end smallexample | |
f077f169 | 7532 | |
39bc1876 NS |
7533 | When performing a stack backtrace, code can inspect the value of |
7534 | @code{pc} stored at @code{fp + 0}. If the trace function then looks at | |
7535 | location @code{pc - 12} and the top 8 bits are set, then we know that | |
7536 | there is a function name embedded immediately preceding this location | |
7537 | and has length @code{((pc[-3]) & 0xff000000)}. | |
2856c3e3 | 7538 | |
39bc1876 NS |
7539 | @item -mthumb |
7540 | @opindex mthumb | |
7541 | Generate code for the 16-bit Thumb instruction set. The default is to | |
7542 | use the 32-bit ARM instruction set. | |
8a0b86f5 | 7543 | |
39bc1876 NS |
7544 | @item -mtpcs-frame |
7545 | @opindex mtpcs-frame | |
7546 | Generate a stack frame that is compliant with the Thumb Procedure Call | |
7547 | Standard for all non-leaf functions. (A leaf function is one that does | |
7548 | not call any other functions.) The default is @option{-mno-tpcs-frame}. | |
058edcdb | 7549 | |
39bc1876 NS |
7550 | @item -mtpcs-leaf-frame |
7551 | @opindex mtpcs-leaf-frame | |
7552 | Generate a stack frame that is compliant with the Thumb Procedure Call | |
7553 | Standard for all leaf functions. (A leaf function is one that does | |
7554 | not call any other functions.) The default is @option{-mno-apcs-leaf-frame}. | |
2856c3e3 | 7555 | |
39bc1876 NS |
7556 | @item -mcallee-super-interworking |
7557 | @opindex mcallee-super-interworking | |
7558 | Gives all externally visible functions in the file being compiled an ARM | |
7559 | instruction set header which switches to Thumb mode before executing the | |
7560 | rest of the function. This allows these functions to be called from | |
7561 | non-interworking code. | |
7562 | ||
7563 | @item -mcaller-super-interworking | |
7564 | @opindex mcaller-super-interworking | |
7565 | Allows calls via function pointers (including virtual functions) to | |
7566 | execute correctly regardless of whether the target code has been | |
7567 | compiled for interworking or not. There is a small overhead in the cost | |
7568 | of executing a function pointer if this option is enabled. | |
2856c3e3 | 7569 | |
d3585b76 DJ |
7570 | @item -mtp=@var{name} |
7571 | @opindex mtp | |
7572 | Specify the access model for the thread local storage pointer. The valid | |
7573 | models are @option{soft}, which generates calls to @code{__aeabi_read_tp}, | |
7574 | @option{cp15}, which fetches the thread pointer from @code{cp15} directly | |
7575 | (supported in the arm6k architecture), and @option{auto}, which uses the | |
7576 | best available method for the selected processor. The default setting is | |
7577 | @option{auto}. | |
7578 | ||
2856c3e3 SC |
7579 | @end table |
7580 | ||
39bc1876 NS |
7581 | @node AVR Options |
7582 | @subsection AVR Options | |
7583 | @cindex AVR Options | |
74291a4b | 7584 | |
39bc1876 | 7585 | These options are defined for AVR implementations: |
74291a4b | 7586 | |
2642624b | 7587 | @table @gcctabopt |
39bc1876 NS |
7588 | @item -mmcu=@var{mcu} |
7589 | @opindex mmcu | |
7590 | Specify ATMEL AVR instruction set or MCU type. | |
74291a4b | 7591 | |
39bc1876 NS |
7592 | Instruction set avr1 is for the minimal AVR core, not supported by the C |
7593 | compiler, only for assembler programs (MCU types: at90s1200, attiny10, | |
7594 | attiny11, attiny12, attiny15, attiny28). | |
74291a4b | 7595 | |
39bc1876 NS |
7596 | Instruction set avr2 (default) is for the classic AVR core with up to |
7597 | 8K program memory space (MCU types: at90s2313, at90s2323, attiny22, | |
7598 | at90s2333, at90s2343, at90s4414, at90s4433, at90s4434, at90s8515, | |
7599 | at90c8534, at90s8535). | |
74291a4b | 7600 | |
39bc1876 NS |
7601 | Instruction set avr3 is for the classic AVR core with up to 128K program |
7602 | memory space (MCU types: atmega103, atmega603, at43usb320, at76c711). | |
74291a4b | 7603 | |
39bc1876 NS |
7604 | Instruction set avr4 is for the enhanced AVR core with up to 8K program |
7605 | memory space (MCU types: atmega8, atmega83, atmega85). | |
74291a4b | 7606 | |
39bc1876 NS |
7607 | Instruction set avr5 is for the enhanced AVR core with up to 128K program |
7608 | memory space (MCU types: atmega16, atmega161, atmega163, atmega32, atmega323, | |
7609 | atmega64, atmega128, at43usb355, at94k). | |
74291a4b | 7610 | |
39bc1876 NS |
7611 | @item -msize |
7612 | @opindex msize | |
7613 | Output instruction sizes to the asm file. | |
74291a4b | 7614 | |
39bc1876 NS |
7615 | @item -minit-stack=@var{N} |
7616 | @opindex minit-stack | |
7617 | Specify the initial stack address, which may be a symbol or numeric value, | |
7618 | @samp{__stack} is the default. | |
74291a4b | 7619 | |
39bc1876 NS |
7620 | @item -mno-interrupts |
7621 | @opindex mno-interrupts | |
7622 | Generated code is not compatible with hardware interrupts. | |
7623 | Code size will be smaller. | |
74291a4b | 7624 | |
39bc1876 NS |
7625 | @item -mcall-prologues |
7626 | @opindex mcall-prologues | |
7627 | Functions prologues/epilogues expanded as call to appropriate | |
7628 | subroutines. Code size will be smaller. | |
74291a4b | 7629 | |
39bc1876 NS |
7630 | @item -mno-tablejump |
7631 | @opindex mno-tablejump | |
7632 | Do not generate tablejump insns which sometimes increase code size. | |
74291a4b | 7633 | |
39bc1876 NS |
7634 | @item -mtiny-stack |
7635 | @opindex mtiny-stack | |
7636 | Change only the low 8 bits of the stack pointer. | |
74291a4b | 7637 | |
39bc1876 NS |
7638 | @item -mint8 |
7639 | @opindex mint8 | |
8a36672b | 7640 | Assume int to be 8 bit integer. This affects the sizes of all types: A |
39bc1876 | 7641 | char will be 1 byte, an int will be 1 byte, an long will be 2 bytes |
8a36672b | 7642 | and long long will be 4 bytes. Please note that this option does not |
39bc1876 NS |
7643 | comply to the C standards, but it will provide you with smaller code |
7644 | size. | |
7645 | @end table | |
74291a4b | 7646 | |
0d4a78eb BS |
7647 | @node Blackfin Options |
7648 | @subsection Blackfin Options | |
7649 | @cindex Blackfin Options | |
7650 | ||
7651 | @table @gcctabopt | |
7652 | @item -momit-leaf-frame-pointer | |
7653 | @opindex momit-leaf-frame-pointer | |
7654 | Don't keep the frame pointer in a register for leaf functions. This | |
7655 | avoids the instructions to save, set up and restore frame pointers and | |
7656 | makes an extra register available in leaf functions. The option | |
7657 | @option{-fomit-frame-pointer} removes the frame pointer for all functions | |
7658 | which might make debugging harder. | |
7659 | ||
3fb192d2 BS |
7660 | @item -mspecld-anomaly |
7661 | @opindex mspecld-anomaly | |
0d4a78eb BS |
7662 | When enabled, the compiler will ensure that the generated code does not |
7663 | contain speculative loads after jump instructions. This option is enabled | |
7664 | by default. | |
7665 | ||
3fb192d2 BS |
7666 | @item -mno-specld-anomaly |
7667 | @opindex mno-specld-anomaly | |
0d4a78eb BS |
7668 | Don't generate extra code to prevent speculative loads from occurring. |
7669 | ||
3fb192d2 | 7670 | @item -mcsync-anomaly |
161c21b6 | 7671 | @opindex mcsync-anomaly |
3fb192d2 BS |
7672 | When enabled, the compiler will ensure that the generated code does not |
7673 | contain CSYNC or SSYNC instructions too soon after conditional branches. | |
7674 | This option is enabled by default. | |
7675 | ||
7676 | @item -mno-csync-anomaly | |
161c21b6 | 7677 | @opindex mno-csync-anomaly |
3fb192d2 BS |
7678 | Don't generate extra code to prevent CSYNC or SSYNC instructions from |
7679 | occurring too soon after a conditional branch. | |
7680 | ||
0d4a78eb | 7681 | @item -mlow-64k |
9821b257 | 7682 | @opindex mlow-64k |
0d4a78eb BS |
7683 | When enabled, the compiler is free to take advantage of the knowledge that |
7684 | the entire program fits into the low 64k of memory. | |
7685 | ||
7686 | @item -mno-low-64k | |
7687 | @opindex mno-low-64k | |
7688 | Assume that the program is arbitrarily large. This is the default. | |
7689 | ||
7690 | @item -mid-shared-library | |
7691 | @opindex mid-shared-library | |
7692 | Generate code that supports shared libraries via the library ID method. | |
7693 | This allows for execute in place and shared libraries in an environment | |
7694 | without virtual memory management. This option implies @option{-fPIC}. | |
7695 | ||
7696 | @item -mno-id-shared-library | |
7697 | @opindex mno-id-shared-library | |
7698 | Generate code that doesn't assume ID based shared libraries are being used. | |
7699 | This is the default. | |
7700 | ||
7701 | @item -mshared-library-id=n | |
7702 | @opindex mshared-library-id | |
7703 | Specified the identification number of the ID based shared library being | |
7704 | compiled. Specifying a value of 0 will generate more compact code, specifying | |
7705 | other values will force the allocation of that number to the current | |
7706 | library but is no more space or time efficient than omitting this option. | |
b6877196 BS |
7707 | |
7708 | @item -mlong-calls | |
7709 | @itemx -mno-long-calls | |
7710 | @opindex mlong-calls | |
7711 | @opindex mno-long-calls | |
7712 | Tells the compiler to perform function calls by first loading the | |
7713 | address of the function into a register and then performing a subroutine | |
7714 | call on this register. This switch is needed if the target function | |
7715 | will lie outside of the 24 bit addressing range of the offset based | |
7716 | version of subroutine call instruction. | |
7717 | ||
7718 | This feature is not enabled by default. Specifying | |
7719 | @option{-mno-long-calls} will restore the default behavior. Note these | |
7720 | switches have no effect on how the compiler generates code to handle | |
7721 | function calls via function pointers. | |
0d4a78eb BS |
7722 | @end table |
7723 | ||
39bc1876 NS |
7724 | @node CRIS Options |
7725 | @subsection CRIS Options | |
7726 | @cindex CRIS Options | |
74291a4b | 7727 | |
39bc1876 | 7728 | These options are defined specifically for the CRIS ports. |
74291a4b | 7729 | |
39bc1876 NS |
7730 | @table @gcctabopt |
7731 | @item -march=@var{architecture-type} | |
7732 | @itemx -mcpu=@var{architecture-type} | |
7733 | @opindex march | |
7734 | @opindex mcpu | |
7735 | Generate code for the specified architecture. The choices for | |
7736 | @var{architecture-type} are @samp{v3}, @samp{v8} and @samp{v10} for | |
8a36672b | 7737 | respectively ETRAX@w{ }4, ETRAX@w{ }100, and ETRAX@w{ }100@w{ }LX@. |
39bc1876 NS |
7738 | Default is @samp{v0} except for cris-axis-linux-gnu, where the default is |
7739 | @samp{v10}. | |
c219ddf7 | 7740 | |
39bc1876 NS |
7741 | @item -mtune=@var{architecture-type} |
7742 | @opindex mtune | |
7743 | Tune to @var{architecture-type} everything applicable about the generated | |
7744 | code, except for the ABI and the set of available instructions. The | |
7745 | choices for @var{architecture-type} are the same as for | |
7746 | @option{-march=@var{architecture-type}}. | |
54284728 | 7747 | |
39bc1876 NS |
7748 | @item -mmax-stack-frame=@var{n} |
7749 | @opindex mmax-stack-frame | |
7750 | Warn when the stack frame of a function exceeds @var{n} bytes. | |
54284728 | 7751 | |
39bc1876 NS |
7752 | @item -melinux-stacksize=@var{n} |
7753 | @opindex melinux-stacksize | |
7754 | Only available with the @samp{cris-axis-aout} target. Arranges for | |
7755 | indications in the program to the kernel loader that the stack of the | |
7756 | program should be set to @var{n} bytes. | |
54284728 | 7757 | |
39bc1876 NS |
7758 | @item -metrax4 |
7759 | @itemx -metrax100 | |
7760 | @opindex metrax4 | |
7761 | @opindex metrax100 | |
7762 | The options @option{-metrax4} and @option{-metrax100} are synonyms for | |
7763 | @option{-march=v3} and @option{-march=v8} respectively. | |
c0498f43 | 7764 | |
39bc1876 NS |
7765 | @item -mmul-bug-workaround |
7766 | @itemx -mno-mul-bug-workaround | |
7767 | @opindex mmul-bug-workaround | |
7768 | @opindex mno-mul-bug-workaround | |
7769 | Work around a bug in the @code{muls} and @code{mulu} instructions for CPU | |
7770 | models where it applies. This option is active by default. | |
c0498f43 | 7771 | |
39bc1876 NS |
7772 | @item -mpdebug |
7773 | @opindex mpdebug | |
7774 | Enable CRIS-specific verbose debug-related information in the assembly | |
7775 | code. This option also has the effect to turn off the @samp{#NO_APP} | |
7776 | formatted-code indicator to the assembler at the beginning of the | |
7777 | assembly file. | |
c0498f43 | 7778 | |
39bc1876 NS |
7779 | @item -mcc-init |
7780 | @opindex mcc-init | |
7781 | Do not use condition-code results from previous instruction; always emit | |
7782 | compare and test instructions before use of condition codes. | |
74291a4b | 7783 | |
39bc1876 NS |
7784 | @item -mno-side-effects |
7785 | @opindex mno-side-effects | |
7786 | Do not emit instructions with side-effects in addressing modes other than | |
7787 | post-increment. | |
238b11b5 | 7788 | |
39bc1876 NS |
7789 | @item -mstack-align |
7790 | @itemx -mno-stack-align | |
7791 | @itemx -mdata-align | |
7792 | @itemx -mno-data-align | |
7793 | @itemx -mconst-align | |
7794 | @itemx -mno-const-align | |
7795 | @opindex mstack-align | |
7796 | @opindex mno-stack-align | |
7797 | @opindex mdata-align | |
7798 | @opindex mno-data-align | |
7799 | @opindex mconst-align | |
7800 | @opindex mno-const-align | |
7801 | These options (no-options) arranges (eliminate arrangements) for the | |
7802 | stack-frame, individual data and constants to be aligned for the maximum | |
7803 | single data access size for the chosen CPU model. The default is to | |
7804 | arrange for 32-bit alignment. ABI details such as structure layout are | |
7805 | not affected by these options. | |
238b11b5 | 7806 | |
39bc1876 NS |
7807 | @item -m32-bit |
7808 | @itemx -m16-bit | |
7809 | @itemx -m8-bit | |
7810 | @opindex m32-bit | |
7811 | @opindex m16-bit | |
7812 | @opindex m8-bit | |
7813 | Similar to the stack- data- and const-align options above, these options | |
7814 | arrange for stack-frame, writable data and constants to all be 32-bit, | |
7815 | 16-bit or 8-bit aligned. The default is 32-bit alignment. | |
238b11b5 | 7816 | |
39bc1876 NS |
7817 | @item -mno-prologue-epilogue |
7818 | @itemx -mprologue-epilogue | |
7819 | @opindex mno-prologue-epilogue | |
7820 | @opindex mprologue-epilogue | |
7821 | With @option{-mno-prologue-epilogue}, the normal function prologue and | |
7822 | epilogue that sets up the stack-frame are omitted and no return | |
7823 | instructions or return sequences are generated in the code. Use this | |
7824 | option only together with visual inspection of the compiled code: no | |
7825 | warnings or errors are generated when call-saved registers must be saved, | |
7826 | or storage for local variable needs to be allocated. | |
238b11b5 | 7827 | |
39bc1876 NS |
7828 | @item -mno-gotplt |
7829 | @itemx -mgotplt | |
7830 | @opindex mno-gotplt | |
7831 | @opindex mgotplt | |
7832 | With @option{-fpic} and @option{-fPIC}, don't generate (do generate) | |
7833 | instruction sequences that load addresses for functions from the PLT part | |
7834 | of the GOT rather than (traditional on other architectures) calls to the | |
8a36672b | 7835 | PLT@. The default is @option{-mgotplt}. |
238b11b5 | 7836 | |
39bc1876 NS |
7837 | @item -maout |
7838 | @opindex maout | |
7839 | Legacy no-op option only recognized with the cris-axis-aout target. | |
c0498f43 | 7840 | |
39bc1876 NS |
7841 | @item -melf |
7842 | @opindex melf | |
7843 | Legacy no-op option only recognized with the cris-axis-elf and | |
7844 | cris-axis-linux-gnu targets. | |
74291a4b | 7845 | |
39bc1876 NS |
7846 | @item -melinux |
7847 | @opindex melinux | |
7848 | Only recognized with the cris-axis-aout target, where it selects a | |
7849 | GNU/linux-like multilib, include files and instruction set for | |
7850 | @option{-march=v8}. | |
ded17aad | 7851 | |
39bc1876 NS |
7852 | @item -mlinux |
7853 | @opindex mlinux | |
7854 | Legacy no-op option only recognized with the cris-axis-linux-gnu target. | |
ded17aad | 7855 | |
39bc1876 NS |
7856 | @item -sim |
7857 | @opindex sim | |
7858 | This option, recognized for the cris-axis-aout and cris-axis-elf arranges | |
7859 | to link with input-output functions from a simulator library. Code, | |
7860 | initialized data and zero-initialized data are allocated consecutively. | |
74291a4b | 7861 | |
39bc1876 NS |
7862 | @item -sim2 |
7863 | @opindex sim2 | |
7864 | Like @option{-sim}, but pass linker options to locate initialized data at | |
7865 | 0x40000000 and zero-initialized data at 0x80000000. | |
74291a4b MM |
7866 | @end table |
7867 | ||
53054e77 PW |
7868 | @node CRX Options |
7869 | @subsection CRX Options | |
7870 | @cindex CRX Options | |
7871 | ||
7872 | These options are defined specifically for the CRX ports. | |
7873 | ||
7874 | @table @gcctabopt | |
7875 | ||
7876 | @item -mmac | |
7877 | @opindex mmac | |
7878 | Enable the use of multiply-accumulate instructions. Disabled by default. | |
7879 | ||
7880 | @item -mpush-args | |
7881 | @opindex mpush-args | |
7882 | Push instructions will be used to pass outgoing arguments when functions | |
7883 | are called. Enabled by default. | |
7884 | @end table | |
7885 | ||
39bc1876 NS |
7886 | @node Darwin Options |
7887 | @subsection Darwin Options | |
7888 | @cindex Darwin options | |
74291a4b | 7889 | |
39bc1876 | 7890 | These options are defined for all architectures running the Darwin operating |
965a7e90 GK |
7891 | system. |
7892 | ||
7893 | FSF GCC on Darwin does not create ``fat'' object files; it will create | |
7894 | an object file for the single architecture that it was built to | |
7895 | target. Apple's GCC on Darwin does create ``fat'' files if multiple | |
46bfe5e3 GK |
7896 | @option{-arch} options are used; it does so by running the compiler or |
7897 | linker multiple times and joining the results together with | |
7898 | @file{lipo}. | |
965a7e90 | 7899 | |
46bfe5e3 GK |
7900 | The subtype of the file created (like @samp{ppc7400} or @samp{ppc970} or |
7901 | @samp{i686}) is determined by the flags that specify the ISA | |
7902 | that GCC is targetting, like @option{-mcpu} or @option{-march}. The | |
965a7e90 | 7903 | @option{-force_cpusubtype_ALL} option can be used to override this. |
74291a4b | 7904 | |
38b974a6 | 7905 | The Darwin tools vary in their behavior when presented with an ISA |
46bfe5e3 GK |
7906 | mismatch. The assembler, @file{as}, will only permit instructions to |
7907 | be used that are valid for the subtype of the file it is generating, | |
7908 | so you cannot put 64-bit instructions in an @samp{ppc750} object file. | |
7909 | The linker for shared libraries, @file{/usr/bin/libtool}, will fail | |
7910 | and print an error if asked to create a shared library with a less | |
7911 | restrictive subtype than its input files (for instance, trying to put | |
7912 | a @samp{ppc970} object file in a @samp{ppc7400} library). The linker | |
7913 | for executables, @file{ld}, will quietly give the executable the most | |
7914 | restrictive subtype of any of its input files. | |
7915 | ||
2642624b | 7916 | @table @gcctabopt |
39bc1876 NS |
7917 | @item -F@var{dir} |
7918 | @opindex F | |
7919 | Add the framework directory @var{dir} to the head of the list of | |
7920 | directories to be searched for header files. These directories are | |
7921 | interleaved with those specified by @option{-I} options and are | |
7922 | scanned in a left-to-right order. | |
5848830f | 7923 | |
39bc1876 NS |
7924 | A framework directory is a directory with frameworks in it. A |
7925 | framework is a directory with a @samp{"Headers"} and/or | |
7926 | @samp{"PrivateHeaders"} directory contained directly in it that ends | |
7927 | in @samp{".framework"}. The name of a framework is the name of this | |
7928 | directory excluding the @samp{".framework"}. Headers associated with | |
7929 | the framework are found in one of those two directories, with | |
7930 | @samp{"Headers"} being searched first. A subframework is a framework | |
7931 | directory that is in a framework's @samp{"Frameworks"} directory. | |
7932 | Includes of subframework headers can only appear in a header of a | |
7933 | framework that contains the subframework, or in a sibling subframework | |
7934 | header. Two subframeworks are siblings if they occur in the same | |
7935 | framework. A subframework should not have the same name as a | |
7936 | framework, a warning will be issued if this is violated. Currently a | |
7937 | subframework cannot have subframeworks, in the future, the mechanism | |
7938 | may be extended to support this. The standard frameworks can be found | |
3e558e80 MS |
7939 | in @samp{"/System/Library/Frameworks"} and |
7940 | @samp{"/Library/Frameworks"}. An example include looks like | |
39bc1876 NS |
7941 | @code{#include <Framework/header.h>}, where @samp{Framework} denotes |
7942 | the name of the framework and header.h is found in the | |
7943 | @samp{"PrivateHeaders"} or @samp{"Headers"} directory. | |
157a620e | 7944 | |
7aded944 DP |
7945 | @item -gused |
7946 | @opindex -gused | |
8a36672b | 7947 | Emit debugging information for symbols that are used. For STABS |
7aded944 | 7948 | debugging format, this enables @option{-feliminate-unused-debug-symbols}. |
8a36672b | 7949 | This is by default ON@. |
7aded944 DP |
7950 | |
7951 | @item -gfull | |
7952 | @opindex -gfull | |
7953 | Emit debugging information for all symbols and types. | |
7954 | ||
ed5b9f96 GK |
7955 | @item -mmacosx-version-min=@var{version} |
7956 | The earliest version of MacOS X that this executable will run on | |
7957 | is @var{version}. Typical values of @var{version} include @code{10.1}, | |
7958 | @code{10.2}, and @code{10.3.9}. | |
7959 | ||
7960 | The default for this option is to make choices that seem to be most | |
7961 | useful. | |
7962 | ||
8f4220dc MA |
7963 | @item -mone-byte-bool |
7964 | @opindex -mone-byte-bool | |
7965 | Override the defaults for @samp{bool} so that @samp{sizeof(bool)==1}. | |
f0eb93a8 | 7966 | By default @samp{sizeof(bool)} is @samp{4} when compiling for |
8f4220dc MA |
7967 | Darwin/PowerPC and @samp{1} when compiling for Darwin/x86, so this |
7968 | option has no effect on x86. | |
7969 | ||
7970 | @strong{Warning:} The @option{-mone-byte-bool} switch causes GCC | |
7971 | to generate code that is not binary compatible with code generated | |
7972 | without that switch. Using this switch may require recompiling all | |
f0eb93a8 | 7973 | other modules in a program, including system libraries. Use this |
8f4220dc MA |
7974 | switch to conform to a non-default data model. |
7975 | ||
699c914a MS |
7976 | @item -mfix-and-continue |
7977 | @itemx -ffix-and-continue | |
7978 | @itemx -findirect-data | |
7979 | @opindex mfix-and-continue | |
7980 | @opindex ffix-and-continue | |
7981 | @opindex findirect-data | |
7982 | Generate code suitable for fast turn around development. Needed to | |
7983 | enable gdb to dynamically load @code{.o} files into already running | |
7984 | programs. @option{-findirect-data} and @option{-ffix-and-continue} | |
7985 | are provided for backwards compatibility. | |
7986 | ||
39bc1876 NS |
7987 | @item -all_load |
7988 | @opindex all_load | |
7989 | Loads all members of static archive libraries. | |
7990 | See man ld(1) for more information. | |
74291a4b | 7991 | |
39bc1876 NS |
7992 | @item -arch_errors_fatal |
7993 | @opindex arch_errors_fatal | |
7994 | Cause the errors having to do with files that have the wrong architecture | |
7995 | to be fatal. | |
157a620e | 7996 | |
39bc1876 NS |
7997 | @item -bind_at_load |
7998 | @opindex bind_at_load | |
7999 | Causes the output file to be marked such that the dynamic linker will | |
8000 | bind all undefined references when the file is loaded or launched. | |
157a620e | 8001 | |
39bc1876 NS |
8002 | @item -bundle |
8003 | @opindex bundle | |
8004 | Produce a Mach-o bundle format file. | |
8005 | See man ld(1) for more information. | |
157a620e | 8006 | |
39bc1876 NS |
8007 | @item -bundle_loader @var{executable} |
8008 | @opindex bundle_loader | |
965a7e90 | 8009 | This option specifies the @var{executable} that will be loading the build |
8a36672b | 8010 | output file being linked. See man ld(1) for more information. |
157a620e | 8011 | |
965a7e90 GK |
8012 | @item -dynamiclib |
8013 | @opindex -dynamiclib | |
8014 | When passed this option, GCC will produce a dynamic library instead of | |
8015 | an executable when linking, using the Darwin @file{libtool} command. | |
157a620e | 8016 | |
965a7e90 GK |
8017 | @item -force_cpusubtype_ALL |
8018 | @opindex -force_cpusubtype_ALL | |
8019 | This causes GCC's output file to have the @var{ALL} subtype, instead of | |
8020 | one controlled by the @option{-mcpu} or @option{-march} option. | |
8021 | ||
8022 | @item -allowable_client @var{client_name} | |
39bc1876 NS |
8023 | @itemx -client_name |
8024 | @itemx -compatibility_version | |
8025 | @itemx -current_version | |
5079843a | 8026 | @itemx -dead_strip |
39bc1876 NS |
8027 | @itemx -dependency-file |
8028 | @itemx -dylib_file | |
8029 | @itemx -dylinker_install_name | |
8030 | @itemx -dynamic | |
39bc1876 NS |
8031 | @itemx -exported_symbols_list |
8032 | @itemx -filelist | |
8033 | @itemx -flat_namespace | |
39bc1876 NS |
8034 | @itemx -force_flat_namespace |
8035 | @itemx -headerpad_max_install_names | |
8036 | @itemx -image_base | |
8037 | @itemx -init | |
8038 | @itemx -install_name | |
8039 | @itemx -keep_private_externs | |
8040 | @itemx -multi_module | |
8041 | @itemx -multiply_defined | |
8042 | @itemx -multiply_defined_unused | |
8043 | @itemx -noall_load | |
89aa5a20 | 8044 | @itemx -no_dead_strip_inits_and_terms |
39bc1876 NS |
8045 | @itemx -nofixprebinding |
8046 | @itemx -nomultidefs | |
8047 | @itemx -noprebind | |
8048 | @itemx -noseglinkedit | |
8049 | @itemx -pagezero_size | |
8050 | @itemx -prebind | |
8051 | @itemx -prebind_all_twolevel_modules | |
8052 | @itemx -private_bundle | |
8053 | @itemx -read_only_relocs | |
8054 | @itemx -sectalign | |
8055 | @itemx -sectobjectsymbols | |
8056 | @itemx -whyload | |
8057 | @itemx -seg1addr | |
8058 | @itemx -sectcreate | |
8059 | @itemx -sectobjectsymbols | |
8060 | @itemx -sectorder | |
5826770c DP |
8061 | @itemx -segaddr |
8062 | @itemx -segs_read_only_addr | |
8063 | @itemx -segs_read_write_addr | |
39bc1876 NS |
8064 | @itemx -seg_addr_table |
8065 | @itemx -seg_addr_table_filename | |
8066 | @itemx -seglinkedit | |
8067 | @itemx -segprot | |
8068 | @itemx -segs_read_only_addr | |
8069 | @itemx -segs_read_write_addr | |
8070 | @itemx -single_module | |
8071 | @itemx -static | |
8072 | @itemx -sub_library | |
8073 | @itemx -sub_umbrella | |
8074 | @itemx -twolevel_namespace | |
8075 | @itemx -umbrella | |
8076 | @itemx -undefined | |
8077 | @itemx -unexported_symbols_list | |
8078 | @itemx -weak_reference_mismatches | |
8079 | @itemx -whatsloaded | |
74291a4b | 8080 | |
39bc1876 | 8081 | @opindex allowable_client |
39bc1876 NS |
8082 | @opindex client_name |
8083 | @opindex compatibility_version | |
8084 | @opindex current_version | |
5079843a | 8085 | @opindex dead_strip |
39bc1876 NS |
8086 | @opindex dependency-file |
8087 | @opindex dylib_file | |
8088 | @opindex dylinker_install_name | |
8089 | @opindex dynamic | |
39bc1876 NS |
8090 | @opindex exported_symbols_list |
8091 | @opindex filelist | |
8092 | @opindex flat_namespace | |
39bc1876 NS |
8093 | @opindex force_flat_namespace |
8094 | @opindex headerpad_max_install_names | |
8095 | @opindex image_base | |
8096 | @opindex init | |
8097 | @opindex install_name | |
8098 | @opindex keep_private_externs | |
8099 | @opindex multi_module | |
8100 | @opindex multiply_defined | |
8101 | @opindex multiply_defined_unused | |
8102 | @opindex noall_load | |
5079843a | 8103 | @opindex no_dead_strip_inits_and_terms |
39bc1876 NS |
8104 | @opindex nofixprebinding |
8105 | @opindex nomultidefs | |
8106 | @opindex noprebind | |
8107 | @opindex noseglinkedit | |
8108 | @opindex pagezero_size | |
8109 | @opindex prebind | |
8110 | @opindex prebind_all_twolevel_modules | |
8111 | @opindex private_bundle | |
8112 | @opindex read_only_relocs | |
8113 | @opindex sectalign | |
8114 | @opindex sectobjectsymbols | |
8115 | @opindex whyload | |
8116 | @opindex seg1addr | |
8117 | @opindex sectcreate | |
8118 | @opindex sectobjectsymbols | |
8119 | @opindex sectorder | |
5826770c DP |
8120 | @opindex segaddr |
8121 | @opindex segs_read_only_addr | |
8122 | @opindex segs_read_write_addr | |
39bc1876 NS |
8123 | @opindex seg_addr_table |
8124 | @opindex seg_addr_table_filename | |
8125 | @opindex seglinkedit | |
8126 | @opindex segprot | |
8127 | @opindex segs_read_only_addr | |
8128 | @opindex segs_read_write_addr | |
8129 | @opindex single_module | |
8130 | @opindex static | |
8131 | @opindex sub_library | |
8132 | @opindex sub_umbrella | |
8133 | @opindex twolevel_namespace | |
8134 | @opindex umbrella | |
8135 | @opindex undefined | |
8136 | @opindex unexported_symbols_list | |
8137 | @opindex weak_reference_mismatches | |
8138 | @opindex whatsloaded | |
8139 | ||
965a7e90 | 8140 | These options are passed to the Darwin linker. The Darwin linker man page |
39bc1876 NS |
8141 | describes them in detail. |
8142 | @end table | |
8143 | ||
8144 | @node DEC Alpha Options | |
8145 | @subsection DEC Alpha Options | |
8146 | ||
8147 | These @samp{-m} options are defined for the DEC Alpha implementations: | |
8148 | ||
8149 | @table @gcctabopt | |
8150 | @item -mno-soft-float | |
8151 | @itemx -msoft-float | |
8152 | @opindex mno-soft-float | |
cd3bb277 | 8153 | @opindex msoft-float |
39bc1876 NS |
8154 | Use (do not use) the hardware floating-point instructions for |
8155 | floating-point operations. When @option{-msoft-float} is specified, | |
8156 | functions in @file{libgcc.a} will be used to perform floating-point | |
8157 | operations. Unless they are replaced by routines that emulate the | |
8158 | floating-point operations, or compiled in such a way as to call such | |
8159 | emulations routines, these routines will issue floating-point | |
8160 | operations. If you are compiling for an Alpha without floating-point | |
8161 | operations, you must ensure that the library is built so as not to call | |
8162 | them. | |
74291a4b | 8163 | |
39bc1876 NS |
8164 | Note that Alpha implementations without floating-point operations are |
8165 | required to have floating-point registers. | |
74291a4b | 8166 | |
39bc1876 NS |
8167 | @item -mfp-reg |
8168 | @itemx -mno-fp-regs | |
8169 | @opindex mfp-reg | |
8170 | @opindex mno-fp-regs | |
8171 | Generate code that uses (does not use) the floating-point register set. | |
8172 | @option{-mno-fp-regs} implies @option{-msoft-float}. If the floating-point | |
8173 | register set is not used, floating point operands are passed in integer | |
8174 | registers as if they were integers and floating-point results are passed | |
8175 | in @code{$0} instead of @code{$f0}. This is a non-standard calling sequence, | |
8176 | so any function with a floating-point argument or return value called by code | |
8177 | compiled with @option{-mno-fp-regs} must also be compiled with that | |
8178 | option. | |
9b66ebb1 | 8179 | |
39bc1876 NS |
8180 | A typical use of this option is building a kernel that does not use, |
8181 | and hence need not save and restore, any floating-point registers. | |
9b66ebb1 | 8182 | |
39bc1876 NS |
8183 | @item -mieee |
8184 | @opindex mieee | |
8185 | The Alpha architecture implements floating-point hardware optimized for | |
8186 | maximum performance. It is mostly compliant with the IEEE floating | |
8187 | point standard. However, for full compliance, software assistance is | |
8188 | required. This option generates code fully IEEE compliant code | |
8189 | @emph{except} that the @var{inexact-flag} is not maintained (see below). | |
8190 | If this option is turned on, the preprocessor macro @code{_IEEE_FP} is | |
8191 | defined during compilation. The resulting code is less efficient but is | |
8192 | able to correctly support denormalized numbers and exceptional IEEE | |
8193 | values such as not-a-number and plus/minus infinity. Other Alpha | |
8194 | compilers call this option @option{-ieee_with_no_inexact}. | |
74291a4b | 8195 | |
39bc1876 NS |
8196 | @item -mieee-with-inexact |
8197 | @opindex mieee-with-inexact | |
8198 | This is like @option{-mieee} except the generated code also maintains | |
8199 | the IEEE @var{inexact-flag}. Turning on this option causes the | |
8200 | generated code to implement fully-compliant IEEE math. In addition to | |
8201 | @code{_IEEE_FP}, @code{_IEEE_FP_EXACT} is defined as a preprocessor | |
8202 | macro. On some Alpha implementations the resulting code may execute | |
8203 | significantly slower than the code generated by default. Since there is | |
8204 | very little code that depends on the @var{inexact-flag}, you should | |
8205 | normally not specify this option. Other Alpha compilers call this | |
8206 | option @option{-ieee_with_inexact}. | |
74291a4b | 8207 | |
39bc1876 NS |
8208 | @item -mfp-trap-mode=@var{trap-mode} |
8209 | @opindex mfp-trap-mode | |
8210 | This option controls what floating-point related traps are enabled. | |
8211 | Other Alpha compilers call this option @option{-fptm @var{trap-mode}}. | |
8212 | The trap mode can be set to one of four values: | |
74291a4b | 8213 | |
39bc1876 NS |
8214 | @table @samp |
8215 | @item n | |
8216 | This is the default (normal) setting. The only traps that are enabled | |
8217 | are the ones that cannot be disabled in software (e.g., division by zero | |
8218 | trap). | |
62b10bbc | 8219 | |
39bc1876 NS |
8220 | @item u |
8221 | In addition to the traps enabled by @samp{n}, underflow traps are enabled | |
8222 | as well. | |
157a620e | 8223 | |
39bc1876 NS |
8224 | @item su |
8225 | Like @samp{su}, but the instructions are marked to be safe for software | |
8226 | completion (see Alpha architecture manual for details). | |
157a620e | 8227 | |
39bc1876 NS |
8228 | @item sui |
8229 | Like @samp{su}, but inexact traps are enabled as well. | |
8230 | @end table | |
9b66ebb1 | 8231 | |
39bc1876 NS |
8232 | @item -mfp-rounding-mode=@var{rounding-mode} |
8233 | @opindex mfp-rounding-mode | |
8234 | Selects the IEEE rounding mode. Other Alpha compilers call this option | |
8235 | @option{-fprm @var{rounding-mode}}. The @var{rounding-mode} can be one | |
8236 | of: | |
157a620e | 8237 | |
39bc1876 NS |
8238 | @table @samp |
8239 | @item n | |
8240 | Normal IEEE rounding mode. Floating point numbers are rounded towards | |
8241 | the nearest machine number or towards the even machine number in case | |
8242 | of a tie. | |
5848830f | 8243 | |
39bc1876 NS |
8244 | @item m |
8245 | Round towards minus infinity. | |
157a620e | 8246 | |
39bc1876 NS |
8247 | @item c |
8248 | Chopped rounding mode. Floating point numbers are rounded towards zero. | |
f5a1b0d2 | 8249 | |
39bc1876 NS |
8250 | @item d |
8251 | Dynamic rounding mode. A field in the floating point control register | |
8252 | (@var{fpcr}, see Alpha architecture reference manual) controls the | |
8253 | rounding mode in effect. The C library initializes this register for | |
8254 | rounding towards plus infinity. Thus, unless your program modifies the | |
8255 | @var{fpcr}, @samp{d} corresponds to round towards plus infinity. | |
8256 | @end table | |
c27ba912 | 8257 | |
39bc1876 NS |
8258 | @item -mtrap-precision=@var{trap-precision} |
8259 | @opindex mtrap-precision | |
8260 | In the Alpha architecture, floating point traps are imprecise. This | |
8261 | means without software assistance it is impossible to recover from a | |
8262 | floating trap and program execution normally needs to be terminated. | |
8263 | GCC can generate code that can assist operating system trap handlers | |
8264 | in determining the exact location that caused a floating point trap. | |
8265 | Depending on the requirements of an application, different levels of | |
8266 | precisions can be selected: | |
c27ba912 | 8267 | |
39bc1876 NS |
8268 | @table @samp |
8269 | @item p | |
8270 | Program precision. This option is the default and means a trap handler | |
8271 | can only identify which program caused a floating point exception. | |
c27ba912 | 8272 | |
39bc1876 NS |
8273 | @item f |
8274 | Function precision. The trap handler can determine the function that | |
8275 | caused a floating point exception. | |
62b10bbc | 8276 | |
39bc1876 NS |
8277 | @item i |
8278 | Instruction precision. The trap handler can determine the exact | |
8279 | instruction that caused a floating point exception. | |
8280 | @end table | |
ed0e6530 | 8281 | |
39bc1876 NS |
8282 | Other Alpha compilers provide the equivalent options called |
8283 | @option{-scope_safe} and @option{-resumption_safe}. | |
ed0e6530 | 8284 | |
39bc1876 NS |
8285 | @item -mieee-conformant |
8286 | @opindex mieee-conformant | |
8287 | This option marks the generated code as IEEE conformant. You must not | |
8288 | use this option unless you also specify @option{-mtrap-precision=i} and either | |
8289 | @option{-mfp-trap-mode=su} or @option{-mfp-trap-mode=sui}. Its only effect | |
8290 | is to emit the line @samp{.eflag 48} in the function prologue of the | |
8291 | generated assembly file. Under DEC Unix, this has the effect that | |
8292 | IEEE-conformant math library routines will be linked in. | |
9b6b54e2 | 8293 | |
39bc1876 NS |
8294 | @item -mbuild-constants |
8295 | @opindex mbuild-constants | |
8296 | Normally GCC examines a 32- or 64-bit integer constant to | |
8297 | see if it can construct it from smaller constants in two or three | |
8298 | instructions. If it cannot, it will output the constant as a literal and | |
8299 | generate code to load it from the data segment at runtime. | |
74291a4b | 8300 | |
39bc1876 NS |
8301 | Use this option to require GCC to construct @emph{all} integer constants |
8302 | using code, even if it takes more instructions (the maximum is six). | |
157a620e | 8303 | |
39bc1876 NS |
8304 | You would typically use this option to build a shared library dynamic |
8305 | loader. Itself a shared library, it must relocate itself in memory | |
8306 | before it can find the variables and constants in its own data segment. | |
8307 | ||
8308 | @item -malpha-as | |
8309 | @itemx -mgas | |
8310 | @opindex malpha-as | |
8311 | @opindex mgas | |
8312 | Select whether to generate code to be assembled by the vendor-supplied | |
8313 | assembler (@option{-malpha-as}) or by the GNU assembler @option{-mgas}. | |
8314 | ||
8315 | @item -mbwx | |
8316 | @itemx -mno-bwx | |
8317 | @itemx -mcix | |
8318 | @itemx -mno-cix | |
8319 | @itemx -mfix | |
8320 | @itemx -mno-fix | |
8321 | @itemx -mmax | |
8322 | @itemx -mno-max | |
8323 | @opindex mbwx | |
8324 | @opindex mno-bwx | |
8325 | @opindex mcix | |
8326 | @opindex mno-cix | |
8327 | @opindex mfix | |
8328 | @opindex mno-fix | |
8329 | @opindex mmax | |
8330 | @opindex mno-max | |
8331 | Indicate whether GCC should generate code to use the optional BWX, | |
8332 | CIX, FIX and MAX instruction sets. The default is to use the instruction | |
8333 | sets supported by the CPU type specified via @option{-mcpu=} option or that | |
8334 | of the CPU on which GCC was built if none was specified. | |
157a620e | 8335 | |
39bc1876 NS |
8336 | @item -mfloat-vax |
8337 | @itemx -mfloat-ieee | |
8338 | @opindex mfloat-vax | |
8339 | @opindex mfloat-ieee | |
8340 | Generate code that uses (does not use) VAX F and G floating point | |
8341 | arithmetic instead of IEEE single and double precision. | |
157a620e | 8342 | |
39bc1876 NS |
8343 | @item -mexplicit-relocs |
8344 | @itemx -mno-explicit-relocs | |
8345 | @opindex mexplicit-relocs | |
8346 | @opindex mno-explicit-relocs | |
8347 | Older Alpha assemblers provided no way to generate symbol relocations | |
8348 | except via assembler macros. Use of these macros does not allow | |
8349 | optimal instruction scheduling. GNU binutils as of version 2.12 | |
8350 | supports a new syntax that allows the compiler to explicitly mark | |
8351 | which relocations should apply to which instructions. This option | |
8352 | is mostly useful for debugging, as GCC detects the capabilities of | |
8353 | the assembler when it is built and sets the default accordingly. | |
157a620e | 8354 | |
39bc1876 NS |
8355 | @item -msmall-data |
8356 | @itemx -mlarge-data | |
8357 | @opindex msmall-data | |
8358 | @opindex mlarge-data | |
8359 | When @option{-mexplicit-relocs} is in effect, static data is | |
8360 | accessed via @dfn{gp-relative} relocations. When @option{-msmall-data} | |
8361 | is used, objects 8 bytes long or smaller are placed in a @dfn{small data area} | |
8362 | (the @code{.sdata} and @code{.sbss} sections) and are accessed via | |
8363 | 16-bit relocations off of the @code{$gp} register. This limits the | |
8364 | size of the small data area to 64KB, but allows the variables to be | |
8365 | directly accessed via a single instruction. | |
62b10bbc | 8366 | |
39bc1876 | 8367 | The default is @option{-mlarge-data}. With this option the data area |
8a36672b | 8368 | is limited to just below 2GB@. Programs that require more than 2GB of |
39bc1876 NS |
8369 | data must use @code{malloc} or @code{mmap} to allocate the data in the |
8370 | heap instead of in the program's data segment. | |
62b10bbc | 8371 | |
39bc1876 NS |
8372 | When generating code for shared libraries, @option{-fpic} implies |
8373 | @option{-msmall-data} and @option{-fPIC} implies @option{-mlarge-data}. | |
4bdc1ac7 | 8374 | |
39bc1876 NS |
8375 | @item -msmall-text |
8376 | @itemx -mlarge-text | |
8377 | @opindex msmall-text | |
8378 | @opindex mlarge-text | |
8379 | When @option{-msmall-text} is used, the compiler assumes that the | |
8380 | code of the entire program (or shared library) fits in 4MB, and is | |
8381 | thus reachable with a branch instruction. When @option{-msmall-data} | |
8382 | is used, the compiler can assume that all local symbols share the | |
8383 | same @code{$gp} value, and thus reduce the number of instructions | |
8384 | required for a function call from 4 to 1. | |
157a620e | 8385 | |
39bc1876 | 8386 | The default is @option{-mlarge-text}. |
d2d42a91 | 8387 | |
39bc1876 NS |
8388 | @item -mcpu=@var{cpu_type} |
8389 | @opindex mcpu | |
8390 | Set the instruction set and instruction scheduling parameters for | |
8391 | machine type @var{cpu_type}. You can specify either the @samp{EV} | |
8392 | style name or the corresponding chip number. GCC supports scheduling | |
8393 | parameters for the EV4, EV5 and EV6 family of processors and will | |
8394 | choose the default values for the instruction set from the processor | |
8395 | you specify. If you do not specify a processor type, GCC will default | |
8396 | to the processor on which the compiler was built. | |
6d6d0fa0 | 8397 | |
39bc1876 | 8398 | Supported values for @var{cpu_type} are |
6d6d0fa0 | 8399 | |
39bc1876 NS |
8400 | @table @samp |
8401 | @item ev4 | |
8402 | @itemx ev45 | |
8403 | @itemx 21064 | |
8404 | Schedules as an EV4 and has no instruction set extensions. | |
ecff22ab | 8405 | |
39bc1876 NS |
8406 | @item ev5 |
8407 | @itemx 21164 | |
8408 | Schedules as an EV5 and has no instruction set extensions. | |
705ac34f | 8409 | |
39bc1876 NS |
8410 | @item ev56 |
8411 | @itemx 21164a | |
8412 | Schedules as an EV5 and supports the BWX extension. | |
705ac34f | 8413 | |
39bc1876 NS |
8414 | @item pca56 |
8415 | @itemx 21164pc | |
8416 | @itemx 21164PC | |
8417 | Schedules as an EV5 and supports the BWX and MAX extensions. | |
c474f76b | 8418 | |
39bc1876 NS |
8419 | @item ev6 |
8420 | @itemx 21264 | |
8421 | Schedules as an EV6 and supports the BWX, FIX, and MAX extensions. | |
ecff22ab | 8422 | |
39bc1876 NS |
8423 | @item ev67 |
8424 | @itemx 21264a | |
8425 | Schedules as an EV6 and supports the BWX, CIX, FIX, and MAX extensions. | |
6d6d0fa0 JL |
8426 | @end table |
8427 | ||
39bc1876 NS |
8428 | @item -mtune=@var{cpu_type} |
8429 | @opindex mtune | |
8430 | Set only the instruction scheduling parameters for machine type | |
8431 | @var{cpu_type}. The instruction set is not changed. | |
ecff22ab | 8432 | |
39bc1876 NS |
8433 | @item -mmemory-latency=@var{time} |
8434 | @opindex mmemory-latency | |
8435 | Sets the latency the scheduler should assume for typical memory | |
8436 | references as seen by the application. This number is highly | |
8437 | dependent on the memory access patterns used by the application | |
8438 | and the size of the external cache on the machine. | |
861bb6c1 | 8439 | |
39bc1876 | 8440 | Valid options for @var{time} are |
6975bd2c | 8441 | |
39bc1876 NS |
8442 | @table @samp |
8443 | @item @var{number} | |
8444 | A decimal number representing clock cycles. | |
98180123 | 8445 | |
39bc1876 NS |
8446 | @item L1 |
8447 | @itemx L2 | |
8448 | @itemx L3 | |
8449 | @itemx main | |
8450 | The compiler contains estimates of the number of clock cycles for | |
8451 | ``typical'' EV4 & EV5 hardware for the Level 1, 2 & 3 caches | |
8452 | (also called Dcache, Scache, and Bcache), as well as to main memory. | |
8453 | Note that L3 is only valid for EV5. | |
de41e41c | 8454 | |
39bc1876 NS |
8455 | @end table |
8456 | @end table | |
861bb6c1 | 8457 | |
39bc1876 NS |
8458 | @node DEC Alpha/VMS Options |
8459 | @subsection DEC Alpha/VMS Options | |
861bb6c1 | 8460 | |
39bc1876 | 8461 | These @samp{-m} options are defined for the DEC Alpha/VMS implementations: |
861bb6c1 | 8462 | |
39bc1876 NS |
8463 | @table @gcctabopt |
8464 | @item -mvms-return-codes | |
8465 | @opindex mvms-return-codes | |
8466 | Return VMS condition codes from main. The default is to return POSIX | |
8467 | style condition (e.g.@ error) codes. | |
8468 | @end table | |
861bb6c1 | 8469 | |
39bc1876 NS |
8470 | @node FRV Options |
8471 | @subsection FRV Options | |
8472 | @cindex FRV Options | |
861bb6c1 | 8473 | |
39bc1876 NS |
8474 | @table @gcctabopt |
8475 | @item -mgpr-32 | |
8476 | @opindex mgpr-32 | |
861bb6c1 | 8477 | |
39bc1876 | 8478 | Only use the first 32 general purpose registers. |
861bb6c1 | 8479 | |
39bc1876 NS |
8480 | @item -mgpr-64 |
8481 | @opindex mgpr-64 | |
861bb6c1 | 8482 | |
39bc1876 | 8483 | Use all 64 general purpose registers. |
861bb6c1 | 8484 | |
39bc1876 NS |
8485 | @item -mfpr-32 |
8486 | @opindex mfpr-32 | |
861bb6c1 | 8487 | |
39bc1876 | 8488 | Use only the first 32 floating point registers. |
ad126521 | 8489 | |
39bc1876 NS |
8490 | @item -mfpr-64 |
8491 | @opindex mfpr-64 | |
ad126521 | 8492 | |
39bc1876 | 8493 | Use all 64 floating point registers |
ad126521 | 8494 | |
39bc1876 NS |
8495 | @item -mhard-float |
8496 | @opindex mhard-float | |
ad126521 | 8497 | |
39bc1876 | 8498 | Use hardware instructions for floating point operations. |
ad126521 | 8499 | |
39bc1876 NS |
8500 | @item -msoft-float |
8501 | @opindex msoft-float | |
ad126521 | 8502 | |
39bc1876 | 8503 | Use library routines for floating point operations. |
ad126521 | 8504 | |
39bc1876 NS |
8505 | @item -malloc-cc |
8506 | @opindex malloc-cc | |
ad126521 | 8507 | |
39bc1876 | 8508 | Dynamically allocate condition code registers. |
ad126521 | 8509 | |
39bc1876 NS |
8510 | @item -mfixed-cc |
8511 | @opindex mfixed-cc | |
861bb6c1 | 8512 | |
39bc1876 NS |
8513 | Do not try to dynamically allocate condition code registers, only |
8514 | use @code{icc0} and @code{fcc0}. | |
74291a4b | 8515 | |
39bc1876 NS |
8516 | @item -mdword |
8517 | @opindex mdword | |
74291a4b | 8518 | |
39bc1876 | 8519 | Change ABI to use double word insns. |
74291a4b | 8520 | |
39bc1876 NS |
8521 | @item -mno-dword |
8522 | @opindex mno-dword | |
74291a4b | 8523 | |
39bc1876 | 8524 | Do not use double word instructions. |
74291a4b | 8525 | |
39bc1876 NS |
8526 | @item -mdouble |
8527 | @opindex mdouble | |
74291a4b | 8528 | |
39bc1876 | 8529 | Use floating point double instructions. |
7fe90e7b | 8530 | |
39bc1876 NS |
8531 | @item -mno-double |
8532 | @opindex mno-double | |
74291a4b | 8533 | |
39bc1876 | 8534 | Do not use floating point double instructions. |
74291a4b | 8535 | |
39bc1876 NS |
8536 | @item -mmedia |
8537 | @opindex mmedia | |
74291a4b | 8538 | |
39bc1876 | 8539 | Use media instructions. |
9c34dbbf | 8540 | |
39bc1876 NS |
8541 | @item -mno-media |
8542 | @opindex mno-media | |
74291a4b | 8543 | |
39bc1876 | 8544 | Do not use media instructions. |
74291a4b | 8545 | |
39bc1876 NS |
8546 | @item -mmuladd |
8547 | @opindex mmuladd | |
9c34dbbf | 8548 | |
39bc1876 | 8549 | Use multiply and add/subtract instructions. |
66188a7e | 8550 | |
39bc1876 NS |
8551 | @item -mno-muladd |
8552 | @opindex mno-muladd | |
74291a4b | 8553 | |
39bc1876 | 8554 | Do not use multiply and add/subtract instructions. |
74291a4b | 8555 | |
d711cf67 JDA |
8556 | @item -mfdpic |
8557 | @opindex mfdpic | |
8558 | ||
8559 | Select the FDPIC ABI, that uses function descriptors to represent | |
8560 | pointers to functions. Without any PIC/PIE-related options, it | |
8561 | implies @option{-fPIE}. With @option{-fpic} or @option{-fpie}, it | |
8562 | assumes GOT entries and small data are within a 12-bit range from the | |
8563 | GOT base address; with @option{-fPIC} or @option{-fPIE}, GOT offsets | |
8564 | are computed with 32 bits. | |
8565 | ||
8566 | @item -minline-plt | |
8567 | @opindex minline-plt | |
8568 | ||
8569 | Enable inlining of PLT entries in function calls to functions that are | |
8570 | not known to bind locally. It has no effect without @option{-mfdpic}. | |
8571 | It's enabled by default if optimizing for speed and compiling for | |
8572 | shared libraries (i.e., @option{-fPIC} or @option{-fpic}), or when an | |
8573 | optimization option such as @option{-O3} or above is present in the | |
8574 | command line. | |
8575 | ||
e4dd71de AH |
8576 | @item -mTLS |
8577 | @opindex TLS | |
8578 | ||
8579 | Assume a large TLS segment when generating thread-local code. | |
8580 | ||
8581 | @item -mtls | |
8582 | @opindex tls | |
8583 | ||
8584 | Do not assume a large TLS segment when generating thread-local code. | |
8585 | ||
d711cf67 JDA |
8586 | @item -mgprel-ro |
8587 | @opindex mgprel-ro | |
8588 | ||
8589 | Enable the use of @code{GPREL} relocations in the FDPIC ABI for data | |
8590 | that is known to be in read-only sections. It's enabled by default, | |
8591 | except for @option{-fpic} or @option{-fpie}: even though it may help | |
8592 | make the global offset table smaller, it trades 1 instruction for 4. | |
8593 | With @option{-fPIC} or @option{-fPIE}, it trades 3 instructions for 4, | |
8594 | one of which may be shared by multiple symbols, and it avoids the need | |
8595 | for a GOT entry for the referenced symbol, so it's more likely to be a | |
8596 | win. If it is not, @option{-mno-gprel-ro} can be used to disable it. | |
8597 | ||
8598 | @item -multilib-library-pic | |
8599 | @opindex multilib-library-pic | |
8600 | ||
8601 | Link with the (library, not FD) pic libraries. It's implied by | |
8602 | @option{-mlibrary-pic}, as well as by @option{-fPIC} and | |
8603 | @option{-fpic} without @option{-mfdpic}. You should never have to use | |
8604 | it explicitly. | |
8605 | ||
8606 | @item -mlinked-fp | |
8607 | @opindex mlinked-fp | |
8608 | ||
8609 | Follow the EABI requirement of always creating a frame pointer whenever | |
8610 | a stack frame is allocated. This option is enabled by default and can | |
8611 | be disabled with @option{-mno-linked-fp}. | |
8612 | ||
c557edf4 RS |
8613 | @item -mlong-calls |
8614 | @opindex mlong-calls | |
8615 | ||
8616 | Use indirect addressing to call functions outside the current | |
8617 | compilation unit. This allows the functions to be placed anywhere | |
8618 | within the 32-bit address space. | |
8619 | ||
8620 | @item -malign-labels | |
8621 | @opindex malign-labels | |
8622 | ||
8623 | Try to align labels to an 8-byte boundary by inserting nops into the | |
8624 | previous packet. This option only has an effect when VLIW packing | |
8625 | is enabled. It doesn't create new packets; it merely adds nops to | |
8626 | existing ones. | |
8627 | ||
39bc1876 NS |
8628 | @item -mlibrary-pic |
8629 | @opindex mlibrary-pic | |
0ac081f6 | 8630 | |
39bc1876 | 8631 | Generate position-independent EABI code. |
6c8875e5 | 8632 | |
39bc1876 NS |
8633 | @item -macc-4 |
8634 | @opindex macc-4 | |
6c8875e5 | 8635 | |
39bc1876 | 8636 | Use only the first four media accumulator registers. |
6c8875e5 | 8637 | |
39bc1876 NS |
8638 | @item -macc-8 |
8639 | @opindex macc-8 | |
993f19a8 | 8640 | |
39bc1876 | 8641 | Use all eight media accumulator registers. |
5da702b1 | 8642 | |
39bc1876 NS |
8643 | @item -mpack |
8644 | @opindex mpack | |
74291a4b | 8645 | |
39bc1876 | 8646 | Pack VLIW instructions. |
74291a4b | 8647 | |
39bc1876 NS |
8648 | @item -mno-pack |
8649 | @opindex mno-pack | |
74291a4b | 8650 | |
39bc1876 | 8651 | Do not pack VLIW instructions. |
7fe90e7b | 8652 | |
39bc1876 NS |
8653 | @item -mno-eflags |
8654 | @opindex mno-eflags | |
74291a4b | 8655 | |
39bc1876 | 8656 | Do not mark ABI switches in e_flags. |
861bb6c1 | 8657 | |
39bc1876 NS |
8658 | @item -mcond-move |
8659 | @opindex mcond-move | |
6184e8a4 | 8660 | |
39bc1876 | 8661 | Enable the use of conditional-move instructions (default). |
74291a4b | 8662 | |
39bc1876 NS |
8663 | This switch is mainly for debugging the compiler and will likely be removed |
8664 | in a future version. | |
74291a4b | 8665 | |
39bc1876 NS |
8666 | @item -mno-cond-move |
8667 | @opindex mno-cond-move | |
74291a4b | 8668 | |
39bc1876 | 8669 | Disable the use of conditional-move instructions. |
861bb6c1 | 8670 | |
39bc1876 NS |
8671 | This switch is mainly for debugging the compiler and will likely be removed |
8672 | in a future version. | |
861bb6c1 | 8673 | |
39bc1876 NS |
8674 | @item -mscc |
8675 | @opindex mscc | |
74291a4b | 8676 | |
39bc1876 | 8677 | Enable the use of conditional set instructions (default). |
74291a4b | 8678 | |
39bc1876 NS |
8679 | This switch is mainly for debugging the compiler and will likely be removed |
8680 | in a future version. | |
74291a4b | 8681 | |
39bc1876 NS |
8682 | @item -mno-scc |
8683 | @opindex mno-scc | |
74291a4b | 8684 | |
39bc1876 | 8685 | Disable the use of conditional set instructions. |
74291a4b | 8686 | |
39bc1876 NS |
8687 | This switch is mainly for debugging the compiler and will likely be removed |
8688 | in a future version. | |
74291a4b | 8689 | |
39bc1876 NS |
8690 | @item -mcond-exec |
8691 | @opindex mcond-exec | |
74291a4b | 8692 | |
39bc1876 | 8693 | Enable the use of conditional execution (default). |
74291a4b | 8694 | |
39bc1876 NS |
8695 | This switch is mainly for debugging the compiler and will likely be removed |
8696 | in a future version. | |
ab82a49f | 8697 | |
39bc1876 NS |
8698 | @item -mno-cond-exec |
8699 | @opindex mno-cond-exec | |
79ae11c4 | 8700 | |
39bc1876 | 8701 | Disable the use of conditional execution. |
daf2f129 | 8702 | |
39bc1876 NS |
8703 | This switch is mainly for debugging the compiler and will likely be removed |
8704 | in a future version. | |
cbe26ab8 | 8705 | |
39bc1876 NS |
8706 | @item -mvliw-branch |
8707 | @opindex mvliw-branch | |
8708 | ||
8709 | Run a pass to pack branches into VLIW instructions (default). | |
8710 | ||
8711 | This switch is mainly for debugging the compiler and will likely be removed | |
8712 | in a future version. | |
8713 | ||
8714 | @item -mno-vliw-branch | |
8715 | @opindex mno-vliw-branch | |
8716 | ||
8717 | Do not run a pass to pack branches into VLIW instructions. | |
8718 | ||
8719 | This switch is mainly for debugging the compiler and will likely be removed | |
8720 | in a future version. | |
74291a4b | 8721 | |
39bc1876 NS |
8722 | @item -mmulti-cond-exec |
8723 | @opindex mmulti-cond-exec | |
74291a4b | 8724 | |
39bc1876 NS |
8725 | Enable optimization of @code{&&} and @code{||} in conditional execution |
8726 | (default). | |
74291a4b | 8727 | |
39bc1876 NS |
8728 | This switch is mainly for debugging the compiler and will likely be removed |
8729 | in a future version. | |
74291a4b | 8730 | |
39bc1876 NS |
8731 | @item -mno-multi-cond-exec |
8732 | @opindex mno-multi-cond-exec | |
74291a4b | 8733 | |
39bc1876 | 8734 | Disable optimization of @code{&&} and @code{||} in conditional execution. |
48180d68 | 8735 | |
39bc1876 NS |
8736 | This switch is mainly for debugging the compiler and will likely be removed |
8737 | in a future version. | |
edf1b3f3 | 8738 | |
39bc1876 NS |
8739 | @item -mnested-cond-exec |
8740 | @opindex mnested-cond-exec | |
9904592e | 8741 | |
39bc1876 | 8742 | Enable nested conditional execution optimizations (default). |
9904592e | 8743 | |
39bc1876 NS |
8744 | This switch is mainly for debugging the compiler and will likely be removed |
8745 | in a future version. | |
0ac081f6 | 8746 | |
39bc1876 NS |
8747 | @item -mno-nested-cond-exec |
8748 | @opindex mno-nested-cond-exec | |
76a773f3 | 8749 | |
39bc1876 | 8750 | Disable nested conditional execution optimizations. |
74291a4b | 8751 | |
39bc1876 NS |
8752 | This switch is mainly for debugging the compiler and will likely be removed |
8753 | in a future version. | |
74291a4b | 8754 | |
38c28a25 AH |
8755 | @item -moptimize-membar |
8756 | @opindex moptimize-membar | |
8757 | ||
8758 | This switch removes redundant @code{membar} instructions from the | |
8759 | compiler generated code. It is enabled by default. | |
8760 | ||
8761 | @item -mno-optimize-membar | |
8762 | @opindex mno-optimize-membar | |
8763 | ||
8764 | This switch disables the automatic removal of redundant @code{membar} | |
8765 | instructions from the generated code. | |
8766 | ||
39bc1876 NS |
8767 | @item -mtomcat-stats |
8768 | @opindex mtomcat-stats | |
e9a25f70 | 8769 | |
39bc1876 | 8770 | Cause gas to print out tomcat statistics. |
e9a25f70 | 8771 | |
39bc1876 NS |
8772 | @item -mcpu=@var{cpu} |
8773 | @opindex mcpu | |
74291a4b | 8774 | |
39bc1876 | 8775 | Select the processor type for which to generate code. Possible values are |
c557edf4 RS |
8776 | @samp{frv}, @samp{fr550}, @samp{tomcat}, @samp{fr500}, @samp{fr450}, |
8777 | @samp{fr405}, @samp{fr400}, @samp{fr300} and @samp{simple}. | |
bff46771 | 8778 | |
39bc1876 | 8779 | @end table |
8d8269fa | 8780 | |
39bc1876 NS |
8781 | @node H8/300 Options |
8782 | @subsection H8/300 Options | |
74291a4b | 8783 | |
39bc1876 | 8784 | These @samp{-m} options are defined for the H8/300 implementations: |
74291a4b | 8785 | |
39bc1876 NS |
8786 | @table @gcctabopt |
8787 | @item -mrelax | |
8788 | @opindex mrelax | |
8789 | Shorten some address references at link time, when possible; uses the | |
8790 | linker option @option{-relax}. @xref{H8/300,, @code{ld} and the H8/300, | |
8791 | ld, Using ld}, for a fuller description. | |
74291a4b | 8792 | |
39bc1876 NS |
8793 | @item -mh |
8794 | @opindex mh | |
8795 | Generate code for the H8/300H@. | |
74291a4b | 8796 | |
39bc1876 NS |
8797 | @item -ms |
8798 | @opindex ms | |
8799 | Generate code for the H8S@. | |
74291a4b | 8800 | |
39bc1876 NS |
8801 | @item -mn |
8802 | @opindex mn | |
8803 | Generate code for the H8S and H8/300H in the normal mode. This switch | |
4ec7afd7 | 8804 | must be used either with @option{-mh} or @option{-ms}. |
74291a4b | 8805 | |
39bc1876 NS |
8806 | @item -ms2600 |
8807 | @opindex ms2600 | |
8808 | Generate code for the H8S/2600. This switch must be used with @option{-ms}. | |
74291a4b | 8809 | |
39bc1876 NS |
8810 | @item -mint32 |
8811 | @opindex mint32 | |
8812 | Make @code{int} data 32 bits by default. | |
74291a4b | 8813 | |
39bc1876 NS |
8814 | @item -malign-300 |
8815 | @opindex malign-300 | |
8816 | On the H8/300H and H8S, use the same alignment rules as for the H8/300. | |
8817 | The default for the H8/300H and H8S is to align longs and floats on 4 | |
8818 | byte boundaries. | |
8819 | @option{-malign-300} causes them to be aligned on 2 byte boundaries. | |
8820 | This option has no effect on the H8/300. | |
8821 | @end table | |
f5a1b0d2 | 8822 | |
39bc1876 NS |
8823 | @node HPPA Options |
8824 | @subsection HPPA Options | |
8825 | @cindex HPPA Options | |
a5c76ee6 | 8826 | |
39bc1876 | 8827 | These @samp{-m} options are defined for the HPPA family of computers: |
a5c76ee6 | 8828 | |
39bc1876 NS |
8829 | @table @gcctabopt |
8830 | @item -march=@var{architecture-type} | |
8831 | @opindex march | |
8832 | Generate code for the specified architecture. The choices for | |
8833 | @var{architecture-type} are @samp{1.0} for PA 1.0, @samp{1.1} for PA | |
8834 | 1.1, and @samp{2.0} for PA 2.0 processors. Refer to | |
8835 | @file{/usr/lib/sched.models} on an HP-UX system to determine the proper | |
8836 | architecture option for your machine. Code compiled for lower numbered | |
8837 | architectures will run on higher numbered architectures, but not the | |
8838 | other way around. | |
8839 | ||
39bc1876 NS |
8840 | @item -mpa-risc-1-0 |
8841 | @itemx -mpa-risc-1-1 | |
8842 | @itemx -mpa-risc-2-0 | |
8843 | @opindex mpa-risc-1-0 | |
8844 | @opindex mpa-risc-1-1 | |
8845 | @opindex mpa-risc-2-0 | |
8846 | Synonyms for @option{-march=1.0}, @option{-march=1.1}, and @option{-march=2.0} respectively. | |
5a26b329 | 8847 | |
39bc1876 NS |
8848 | @item -mbig-switch |
8849 | @opindex mbig-switch | |
8850 | Generate code suitable for big switch tables. Use this option only if | |
8851 | the assembler/linker complain about out of range branches within a switch | |
8852 | table. | |
efdba735 | 8853 | |
39bc1876 NS |
8854 | @item -mjump-in-delay |
8855 | @opindex mjump-in-delay | |
8856 | Fill delay slots of function calls with unconditional jump instructions | |
8857 | by modifying the return pointer for the function call to be the target | |
8858 | of the conditional jump. | |
a5c76ee6 | 8859 | |
39bc1876 NS |
8860 | @item -mdisable-fpregs |
8861 | @opindex mdisable-fpregs | |
8862 | Prevent floating point registers from being used in any manner. This is | |
8863 | necessary for compiling kernels which perform lazy context switching of | |
8864 | floating point registers. If you use this option and attempt to perform | |
8865 | floating point operations, the compiler will abort. | |
0a379b7a | 8866 | |
39bc1876 NS |
8867 | @item -mdisable-indexing |
8868 | @opindex mdisable-indexing | |
8869 | Prevent the compiler from using indexing address modes. This avoids some | |
8870 | rather obscure problems when compiling MIG generated code under MACH@. | |
f5a1b0d2 | 8871 | |
39bc1876 NS |
8872 | @item -mno-space-regs |
8873 | @opindex mno-space-regs | |
8874 | Generate code that assumes the target has no space registers. This allows | |
8875 | GCC to generate faster indirect calls and use unscaled index address modes. | |
d2d42a91 | 8876 | |
39bc1876 | 8877 | Such code is suitable for level 0 PA systems and kernels. |
d2d42a91 | 8878 | |
39bc1876 NS |
8879 | @item -mfast-indirect-calls |
8880 | @opindex mfast-indirect-calls | |
8881 | Generate code that assumes calls never cross space boundaries. This | |
8882 | allows GCC to emit code which performs faster indirect calls. | |
f08a3544 | 8883 | |
39bc1876 NS |
8884 | This option will not work in the presence of shared libraries or nested |
8885 | functions. | |
f08a3544 | 8886 | |
a2017852 JDA |
8887 | @item -mfixed-range=@var{register-range} |
8888 | @opindex mfixed-range | |
8889 | Generate code treating the given register range as fixed registers. | |
8890 | A fixed register is one that the register allocator can not use. This is | |
8891 | useful when compiling kernel code. A register range is specified as | |
8892 | two registers separated by a dash. Multiple register ranges can be | |
8893 | specified separated by a comma. | |
8894 | ||
39bc1876 NS |
8895 | @item -mlong-load-store |
8896 | @opindex mlong-load-store | |
8897 | Generate 3-instruction load and store sequences as sometimes required by | |
8898 | the HP-UX 10 linker. This is equivalent to the @samp{+k} option to | |
8899 | the HP compilers. | |
61c85ff1 | 8900 | |
39bc1876 NS |
8901 | @item -mportable-runtime |
8902 | @opindex mportable-runtime | |
8903 | Use the portable calling conventions proposed by HP for ELF systems. | |
61c85ff1 | 8904 | |
39bc1876 NS |
8905 | @item -mgas |
8906 | @opindex mgas | |
8907 | Enable the use of assembler directives only GAS understands. | |
61c85ff1 | 8908 | |
39bc1876 NS |
8909 | @item -mschedule=@var{cpu-type} |
8910 | @opindex mschedule | |
8911 | Schedule code according to the constraints for the machine type | |
8912 | @var{cpu-type}. The choices for @var{cpu-type} are @samp{700} | |
8913 | @samp{7100}, @samp{7100LC}, @samp{7200}, @samp{7300} and @samp{8000}. Refer | |
8914 | to @file{/usr/lib/sched.models} on an HP-UX system to determine the | |
8915 | proper scheduling option for your machine. The default scheduling is | |
8916 | @samp{8000}. | |
61c85ff1 | 8917 | |
39bc1876 NS |
8918 | @item -mlinker-opt |
8919 | @opindex mlinker-opt | |
8920 | Enable the optimization pass in the HP-UX linker. Note this makes symbolic | |
8921 | debugging impossible. It also triggers a bug in the HP-UX 8 and HP-UX 9 | |
8922 | linkers in which they give bogus error messages when linking some programs. | |
61c85ff1 | 8923 | |
39bc1876 NS |
8924 | @item -msoft-float |
8925 | @opindex msoft-float | |
8926 | Generate output containing library calls for floating point. | |
8927 | @strong{Warning:} the requisite libraries are not available for all HPPA | |
8928 | targets. Normally the facilities of the machine's usual C compiler are | |
8929 | used, but this cannot be done directly in cross-compilation. You must make | |
8930 | your own arrangements to provide suitable library functions for | |
8931 | cross-compilation. The embedded target @samp{hppa1.1-*-pro} | |
8932 | does provide software floating point support. | |
31775d31 | 8933 | |
39bc1876 NS |
8934 | @option{-msoft-float} changes the calling convention in the output file; |
8935 | therefore, it is only useful if you compile @emph{all} of a program with | |
8936 | this option. In particular, you need to compile @file{libgcc.a}, the | |
8937 | library that comes with GCC, with @option{-msoft-float} in order for | |
8938 | this to work. | |
61c85ff1 | 8939 | |
39bc1876 NS |
8940 | @item -msio |
8941 | @opindex msio | |
8a36672b | 8942 | Generate the predefine, @code{_SIO}, for server IO@. The default is |
39bc1876 | 8943 | @option{-mwsio}. This generates the predefines, @code{__hp9000s700}, |
8a36672b JM |
8944 | @code{__hp9000s700__} and @code{_WSIO}, for workstation IO@. These |
8945 | options are available under HP-UX and HI-UX@. | |
39bc1876 NS |
8946 | |
8947 | @item -mgnu-ld | |
8948 | @opindex gnu-ld | |
8949 | Use GNU ld specific options. This passes @option{-shared} to ld when | |
8950 | building a shared library. It is the default when GCC is configured, | |
8951 | explicitly or implicitly, with the GNU linker. This option does not | |
8952 | have any affect on which ld is called, it only changes what parameters | |
8953 | are passed to that ld. The ld that is called is determined by the | |
8954 | @option{--with-ld} configure option, GCC's program search path, and | |
8955 | finally by the user's @env{PATH}. The linker used by GCC can be printed | |
a73035ae SE |
8956 | using @samp{which `gcc -print-prog-name=ld`}. This option is only available |
8957 | on the 64 bit HP-UX GCC, i.e. configured with @samp{hppa*64*-*-hpux*}. | |
48aec0bc | 8958 | |
39bc1876 NS |
8959 | @item -mhp-ld |
8960 | @opindex hp-ld | |
8961 | Use HP ld specific options. This passes @option{-b} to ld when building | |
8962 | a shared library and passes @option{+Accept TypeMismatch} to ld on all | |
8963 | links. It is the default when GCC is configured, explicitly or | |
8964 | implicitly, with the HP linker. This option does not have any affect on | |
8965 | which ld is called, it only changes what parameters are passed to that | |
8966 | ld. The ld that is called is determined by the @option{--with-ld} | |
8967 | configure option, GCC's program search path, and finally by the user's | |
8968 | @env{PATH}. The linker used by GCC can be printed using @samp{which | |
a73035ae SE |
8969 | `gcc -print-prog-name=ld`}. This option is only available on the 64 bit |
8970 | HP-UX GCC, i.e. configured with @samp{hppa*64*-*-hpux*}. | |
48aec0bc | 8971 | |
39bc1876 NS |
8972 | @item -mlong-calls |
8973 | @opindex mno-long-calls | |
8974 | Generate code that uses long call sequences. This ensures that a call | |
8975 | is always able to reach linker generated stubs. The default is to generate | |
8976 | long calls only when the distance from the call site to the beginning | |
8977 | of the function or translation unit, as the case may be, exceeds a | |
8978 | predefined limit set by the branch type being used. The limits for | |
8979 | normal calls are 7,600,000 and 240,000 bytes, respectively for the | |
8980 | PA 2.0 and PA 1.X architectures. Sibcalls are always limited at | |
8981 | 240,000 bytes. | |
a27fb29b | 8982 | |
39bc1876 NS |
8983 | Distances are measured from the beginning of functions when using the |
8984 | @option{-ffunction-sections} option, or when using the @option{-mgas} | |
8985 | and @option{-mno-portable-runtime} options together under HP-UX with | |
8986 | the SOM linker. | |
7dac2f89 | 8987 | |
39bc1876 NS |
8988 | It is normally not desirable to use this option as it will degrade |
8989 | performance. However, it may be useful in large applications, | |
8990 | particularly when partial linking is used to build the application. | |
74291a4b | 8991 | |
39bc1876 NS |
8992 | The types of long calls used depends on the capabilities of the |
8993 | assembler and linker, and the type of code being generated. The | |
8994 | impact on systems that support long absolute calls, and long pic | |
8995 | symbol-difference or pc-relative calls should be relatively small. | |
8996 | However, an indirect call is used on 32-bit ELF systems in pic code | |
8997 | and it is quite long. | |
74291a4b | 8998 | |
d711cf67 JDA |
8999 | @item -munix=@var{unix-std} |
9000 | @opindex march | |
9001 | Generate compiler predefines and select a startfile for the specified | |
9002 | UNIX standard. The choices for @var{unix-std} are @samp{93}, @samp{95} | |
9003 | and @samp{98}. @samp{93} is supported on all HP-UX versions. @samp{95} | |
9004 | is available on HP-UX 10.10 and later. @samp{98} is available on HP-UX | |
9005 | 11.11 and later. The default values are @samp{93} for HP-UX 10.00, | |
9006 | @samp{95} for HP-UX 10.10 though to 11.00, and @samp{98} for HP-UX 11.11 | |
9007 | and later. | |
9008 | ||
9009 | @option{-munix=93} provides the same predefines as GCC 3.3 and 3.4. | |
9010 | @option{-munix=95} provides additional predefines for @code{XOPEN_UNIX} | |
9011 | and @code{_XOPEN_SOURCE_EXTENDED}, and the startfile @file{unix95.o}. | |
9012 | @option{-munix=98} provides additional predefines for @code{_XOPEN_UNIX}, | |
9013 | @code{_XOPEN_SOURCE_EXTENDED}, @code{_INCLUDE__STDC_A1_SOURCE} and | |
9014 | @code{_INCLUDE_XOPEN_SOURCE_500}, and the startfile @file{unix98.o}. | |
9015 | ||
9016 | It is @emph{important} to note that this option changes the interfaces | |
9017 | for various library routines. It also affects the operational behavior | |
9018 | of the C library. Thus, @emph{extreme} care is needed in using this | |
9019 | option. | |
9020 | ||
9021 | Library code that is intended to operate with more than one UNIX | |
9022 | standard must test, set and restore the variable @var{__xpg4_extended_mask} | |
9023 | as appropriate. Most GNU software doesn't provide this capability. | |
9024 | ||
39bc1876 NS |
9025 | @item -nolibdld |
9026 | @opindex nolibdld | |
9027 | Suppress the generation of link options to search libdld.sl when the | |
9028 | @option{-static} option is specified on HP-UX 10 and later. | |
9029 | ||
9030 | @item -static | |
9031 | @opindex static | |
9032 | The HP-UX implementation of setlocale in libc has a dependency on | |
9033 | libdld.sl. There isn't an archive version of libdld.sl. Thus, | |
9034 | when the @option{-static} option is specified, special link options | |
9035 | are needed to resolve this dependency. | |
9036 | ||
9037 | On HP-UX 10 and later, the GCC driver adds the necessary options to | |
9038 | link with libdld.sl when the @option{-static} option is specified. | |
9039 | This causes the resulting binary to be dynamic. On the 64-bit port, | |
9040 | the linkers generate dynamic binaries by default in any case. The | |
9041 | @option{-nolibdld} option can be used to prevent the GCC driver from | |
9042 | adding these link options. | |
9043 | ||
9044 | @item -threads | |
9045 | @opindex threads | |
9046 | Add support for multithreading with the @dfn{dce thread} library | |
8a36672b | 9047 | under HP-UX@. This option sets flags for both the preprocessor and |
39bc1876 NS |
9048 | linker. |
9049 | @end table | |
9050 | ||
9051 | @node i386 and x86-64 Options | |
9052 | @subsection Intel 386 and AMD x86-64 Options | |
9053 | @cindex i386 Options | |
9054 | @cindex x86-64 Options | |
9055 | @cindex Intel 386 Options | |
9056 | @cindex AMD x86-64 Options | |
9057 | ||
9058 | These @samp{-m} options are defined for the i386 and x86-64 family of | |
9059 | computers: | |
9060 | ||
9061 | @table @gcctabopt | |
9062 | @item -mtune=@var{cpu-type} | |
9063 | @opindex mtune | |
9064 | Tune to @var{cpu-type} everything applicable about the generated code, except | |
9065 | for the ABI and the set of available instructions. The choices for | |
9066 | @var{cpu-type} are: | |
9067 | @table @emph | |
786872fd JH |
9068 | @item generic |
9069 | Produce code optimized for the most common IA32/AMD64/EM64T processors. | |
9070 | If you know the CPU on which your code will run, then you should use | |
9071 | the corresponding @option{-mtune} option instead of | |
9072 | @option{-mtune=generic}. But, if you do not know exactly what CPU users | |
9073 | of your application will have, then you should use this option. | |
9074 | ||
9075 | As new processors are deployed in the marketplace, the behavior of this | |
9076 | option will change. Therefore, if you upgrade to a newer version of | |
9077 | GCC, the code generated option will change to reflect the processors | |
9078 | that were most common when that version of GCC was released. | |
9079 | ||
9080 | There is no @option{-march=generic} option because @option{-march} | |
9081 | indicates the instruction set the compiler can use, and there is no | |
9082 | generic instruction set applicable to all processors. In contrast, | |
9083 | @option{-mtune} indicates the processor (or, in this case, collection of | |
9084 | processors) for which the code is optimized. | |
39bc1876 | 9085 | @item i386 |
8a36672b | 9086 | Original Intel's i386 CPU@. |
39bc1876 | 9087 | @item i486 |
8a36672b | 9088 | Intel's i486 CPU@. (No scheduling is implemented for this chip.) |
39bc1876 NS |
9089 | @item i586, pentium |
9090 | Intel Pentium CPU with no MMX support. | |
9091 | @item pentium-mmx | |
9092 | Intel PentiumMMX CPU based on Pentium core with MMX instruction set support. | |
786872fd | 9093 | @item pentiumpro |
8a36672b | 9094 | Intel PentiumPro CPU@. |
786872fd JH |
9095 | @item i686 |
9096 | Same as @code{generic}, but when used as @code{march} option, PentiumPro | |
9097 | instruction set will be used, so the code will run on all i686 familly chips. | |
39bc1876 NS |
9098 | @item pentium2 |
9099 | Intel Pentium2 CPU based on PentiumPro core with MMX instruction set support. | |
9100 | @item pentium3, pentium3m | |
9101 | Intel Pentium3 CPU based on PentiumPro core with MMX and SSE instruction set | |
9102 | support. | |
9103 | @item pentium-m | |
9104 | Low power version of Intel Pentium3 CPU with MMX, SSE and SSE2 instruction set | |
9105 | support. Used by Centrino notebooks. | |
9106 | @item pentium4, pentium4m | |
9107 | Intel Pentium4 CPU with MMX, SSE and SSE2 instruction set support. | |
9108 | @item prescott | |
9109 | Improved version of Intel Pentium4 CPU with MMX, SSE, SSE2 and SSE3 instruction | |
9110 | set support. | |
9111 | @item nocona | |
9112 | Improved version of Intel Pentium4 CPU with 64-bit extensions, MMX, SSE, | |
9113 | SSE2 and SSE3 instruction set support. | |
9114 | @item k6 | |
9115 | AMD K6 CPU with MMX instruction set support. | |
9116 | @item k6-2, k6-3 | |
9117 | Improved versions of AMD K6 CPU with MMX and 3dNOW! instruction set support. | |
9118 | @item athlon, athlon-tbird | |
9119 | AMD Athlon CPU with MMX, 3dNOW!, enhanced 3dNOW! and SSE prefetch instructions | |
9120 | support. | |
9121 | @item athlon-4, athlon-xp, athlon-mp | |
9122 | Improved AMD Athlon CPU with MMX, 3dNOW!, enhanced 3dNOW! and full SSE | |
9123 | instruction set support. | |
9124 | @item k8, opteron, athlon64, athlon-fx | |
9125 | AMD K8 core based CPUs with x86-64 instruction set support. (This supersets | |
9126 | MMX, SSE, SSE2, 3dNOW!, enhanced 3dNOW! and 64-bit instruction set extensions.) | |
9127 | @item winchip-c6 | |
9128 | IDT Winchip C6 CPU, dealt in same way as i486 with additional MMX instruction | |
9129 | set support. | |
9130 | @item winchip2 | |
9131 | IDT Winchip2 CPU, dealt in same way as i486 with additional MMX and 3dNOW! | |
9132 | instruction set support. | |
9133 | @item c3 | |
8a36672b | 9134 | Via C3 CPU with MMX and 3dNOW! instruction set support. (No scheduling is |
39bc1876 NS |
9135 | implemented for this chip.) |
9136 | @item c3-2 | |
9137 | Via C3-2 CPU with MMX and SSE instruction set support. (No scheduling is | |
9138 | implemented for this chip.) | |
9139 | @end table | |
9140 | ||
9141 | While picking a specific @var{cpu-type} will schedule things appropriately | |
9142 | for that particular chip, the compiler will not generate any code that | |
9143 | does not run on the i386 without the @option{-march=@var{cpu-type}} option | |
9144 | being used. | |
9145 | ||
9146 | @item -march=@var{cpu-type} | |
9147 | @opindex march | |
9148 | Generate instructions for the machine type @var{cpu-type}. The choices | |
9149 | for @var{cpu-type} are the same as for @option{-mtune}. Moreover, | |
9150 | specifying @option{-march=@var{cpu-type}} implies @option{-mtune=@var{cpu-type}}. | |
74291a4b | 9151 | |
39bc1876 NS |
9152 | @item -mcpu=@var{cpu-type} |
9153 | @opindex mcpu | |
9154 | A deprecated synonym for @option{-mtune}. | |
3398f47f | 9155 | |
39bc1876 NS |
9156 | @item -m386 |
9157 | @itemx -m486 | |
9158 | @itemx -mpentium | |
9159 | @itemx -mpentiumpro | |
9160 | @opindex m386 | |
9161 | @opindex m486 | |
9162 | @opindex mpentium | |
9163 | @opindex mpentiumpro | |
9164 | These options are synonyms for @option{-mtune=i386}, @option{-mtune=i486}, | |
9165 | @option{-mtune=pentium}, and @option{-mtune=pentiumpro} respectively. | |
9166 | These synonyms are deprecated. | |
74291a4b | 9167 | |
39bc1876 NS |
9168 | @item -mfpmath=@var{unit} |
9169 | @opindex march | |
9170 | Generate floating point arithmetics for selected unit @var{unit}. The choices | |
9171 | for @var{unit} are: | |
2d2a50c3 | 9172 | |
39bc1876 NS |
9173 | @table @samp |
9174 | @item 387 | |
9175 | Use the standard 387 floating point coprocessor present majority of chips and | |
9176 | emulated otherwise. Code compiled with this option will run almost everywhere. | |
9177 | The temporary results are computed in 80bit precision instead of precision | |
9178 | specified by the type resulting in slightly different results compared to most | |
8a36672b | 9179 | of other chips. See @option{-ffloat-store} for more detailed description. |
74291a4b | 9180 | |
39bc1876 | 9181 | This is the default choice for i386 compiler. |
74291a4b | 9182 | |
39bc1876 NS |
9183 | @item sse |
9184 | Use scalar floating point instructions present in the SSE instruction set. | |
9185 | This instruction set is supported by Pentium3 and newer chips, in the AMD line | |
9186 | by Athlon-4, Athlon-xp and Athlon-mp chips. The earlier version of SSE | |
9187 | instruction set supports only single precision arithmetics, thus the double and | |
9188 | extended precision arithmetics is still done using 387. Later version, present | |
9189 | only in Pentium4 and the future AMD x86-64 chips supports double precision | |
9190 | arithmetics too. | |
a27fb29b | 9191 | |
c2b43d7a UB |
9192 | For the i386 compiler, you need to use @option{-march=@var{cpu-type}}, @option{-msse} |
9193 | or @option{-msse2} switches to enable SSE extensions and make this option | |
9194 | effective. For the x86-64 compiler, these extensions are enabled by default. | |
3398f47f | 9195 | |
39bc1876 NS |
9196 | The resulting code should be considerably faster in the majority of cases and avoid |
9197 | the numerical instability problems of 387 code, but may break some existing | |
9198 | code that expects temporaries to be 80bit. | |
cd3f11a6 | 9199 | |
39bc1876 | 9200 | This is the default choice for the x86-64 compiler. |
74291a4b | 9201 | |
39bc1876 NS |
9202 | @item sse,387 |
9203 | Attempt to utilize both instruction sets at once. This effectively double the | |
9204 | amount of available registers and on chips with separate execution units for | |
9205 | 387 and SSE the execution resources too. Use this option with care, as it is | |
9206 | still experimental, because the GCC register allocator does not model separate | |
9207 | functional units well resulting in instable performance. | |
9208 | @end table | |
ee692410 | 9209 | |
39bc1876 NS |
9210 | @item -masm=@var{dialect} |
9211 | @opindex masm=@var{dialect} | |
1f4c2c57 MS |
9212 | Output asm instructions using selected @var{dialect}. Supported |
9213 | choices are @samp{intel} or @samp{att} (the default one). Darwin does | |
13b3c362 | 9214 | not support @samp{intel}. |
ee692410 | 9215 | |
39bc1876 NS |
9216 | @item -mieee-fp |
9217 | @itemx -mno-ieee-fp | |
9218 | @opindex mieee-fp | |
9219 | @opindex mno-ieee-fp | |
9220 | Control whether or not the compiler uses IEEE floating point | |
9221 | comparisons. These handle correctly the case where the result of a | |
9222 | comparison is unordered. | |
ee692410 | 9223 | |
39bc1876 NS |
9224 | @item -msoft-float |
9225 | @opindex msoft-float | |
9226 | Generate output containing library calls for floating point. | |
9227 | @strong{Warning:} the requisite libraries are not part of GCC@. | |
9228 | Normally the facilities of the machine's usual C compiler are used, but | |
9229 | this can't be done directly in cross-compilation. You must make your | |
9230 | own arrangements to provide suitable library functions for | |
9231 | cross-compilation. | |
ee692410 | 9232 | |
39bc1876 NS |
9233 | On machines where a function returns floating point results in the 80387 |
9234 | register stack, some floating point opcodes may be emitted even if | |
9235 | @option{-msoft-float} is used. | |
ee692410 | 9236 | |
39bc1876 NS |
9237 | @item -mno-fp-ret-in-387 |
9238 | @opindex mno-fp-ret-in-387 | |
9239 | Do not use the FPU registers for return values of functions. | |
ee692410 | 9240 | |
39bc1876 NS |
9241 | The usual calling convention has functions return values of types |
9242 | @code{float} and @code{double} in an FPU register, even if there | |
9243 | is no FPU@. The idea is that the operating system should emulate | |
9244 | an FPU@. | |
5a4b3afd | 9245 | |
39bc1876 NS |
9246 | The option @option{-mno-fp-ret-in-387} causes such values to be returned |
9247 | in ordinary CPU registers instead. | |
5a4b3afd | 9248 | |
39bc1876 NS |
9249 | @item -mno-fancy-math-387 |
9250 | @opindex mno-fancy-math-387 | |
9251 | Some 387 emulators do not support the @code{sin}, @code{cos} and | |
9252 | @code{sqrt} instructions for the 387. Specify this option to avoid | |
9253 | generating those instructions. This option is the default on FreeBSD, | |
9254 | OpenBSD and NetBSD@. This option is overridden when @option{-march} | |
9255 | indicates that the target cpu will always have an FPU and so the | |
9256 | instruction will not need emulation. As of revision 2.6.1, these | |
9257 | instructions are not generated unless you also use the | |
9258 | @option{-funsafe-math-optimizations} switch. | |
5a4b3afd | 9259 | |
39bc1876 NS |
9260 | @item -malign-double |
9261 | @itemx -mno-align-double | |
9262 | @opindex malign-double | |
9263 | @opindex mno-align-double | |
9264 | Control whether GCC aligns @code{double}, @code{long double}, and | |
9265 | @code{long long} variables on a two word boundary or a one word | |
9266 | boundary. Aligning @code{double} variables on a two word boundary will | |
9267 | produce code that runs somewhat faster on a @samp{Pentium} at the | |
9268 | expense of more memory. | |
5a4b3afd | 9269 | |
39bc1876 NS |
9270 | @strong{Warning:} if you use the @option{-malign-double} switch, |
9271 | structures containing the above types will be aligned differently than | |
9272 | the published application binary interface specifications for the 386 | |
9273 | and will not be binary compatible with structures in code compiled | |
9274 | without that switch. | |
5a4b3afd | 9275 | |
39bc1876 NS |
9276 | @item -m96bit-long-double |
9277 | @itemx -m128bit-long-double | |
9278 | @opindex m96bit-long-double | |
9279 | @opindex m128bit-long-double | |
8a36672b | 9280 | These switches control the size of @code{long double} type. The i386 |
39bc1876 NS |
9281 | application binary interface specifies the size to be 96 bits, |
9282 | so @option{-m96bit-long-double} is the default in 32 bit mode. | |
5a4b3afd | 9283 | |
39bc1876 NS |
9284 | Modern architectures (Pentium and newer) would prefer @code{long double} |
9285 | to be aligned to an 8 or 16 byte boundary. In arrays or structures | |
9286 | conforming to the ABI, this would not be possible. So specifying a | |
9287 | @option{-m128bit-long-double} will align @code{long double} | |
9288 | to a 16 byte boundary by padding the @code{long double} with an additional | |
9289 | 32 bit zero. | |
5a4b3afd | 9290 | |
39bc1876 NS |
9291 | In the x86-64 compiler, @option{-m128bit-long-double} is the default choice as |
9292 | its ABI specifies that @code{long double} is to be aligned on 16 byte boundary. | |
5a4b3afd | 9293 | |
39bc1876 NS |
9294 | Notice that neither of these options enable any extra precision over the x87 |
9295 | standard of 80 bits for a @code{long double}. | |
5a4b3afd | 9296 | |
39bc1876 NS |
9297 | @strong{Warning:} if you override the default value for your target ABI, the |
9298 | structures and arrays containing @code{long double} variables will change | |
9299 | their size as well as function calling convention for function taking | |
9300 | @code{long double} will be modified. Hence they will not be binary | |
9301 | compatible with arrays or structures in code compiled without that switch. | |
5a4b3afd | 9302 | |
7dcbf659 JH |
9303 | @item -mmlarge-data-threshold=@var{number} |
9304 | @opindex mlarge-data-threshold=@var{number} | |
9305 | When @option{-mcmodel=medium} is specified, the data greater than | |
9306 | @var{threshold} are placed in large data section. This value must be the | |
a4d05547 | 9307 | same across all object linked into the binary and defaults to 65535. |
5a4b3afd | 9308 | |
39bc1876 NS |
9309 | @item -msvr3-shlib |
9310 | @itemx -mno-svr3-shlib | |
9311 | @opindex msvr3-shlib | |
9312 | @opindex mno-svr3-shlib | |
9313 | Control whether GCC places uninitialized local variables into the | |
9314 | @code{bss} or @code{data} segments. @option{-msvr3-shlib} places them | |
9315 | into @code{bss}. These options are meaningful only on System V Release 3. | |
5a4b3afd | 9316 | |
39bc1876 NS |
9317 | @item -mrtd |
9318 | @opindex mrtd | |
9319 | Use a different function-calling convention, in which functions that | |
9320 | take a fixed number of arguments return with the @code{ret} @var{num} | |
9321 | instruction, which pops their arguments while returning. This saves one | |
9322 | instruction in the caller since there is no need to pop the arguments | |
9323 | there. | |
5a4b3afd | 9324 | |
39bc1876 NS |
9325 | You can specify that an individual function is called with this calling |
9326 | sequence with the function attribute @samp{stdcall}. You can also | |
9327 | override the @option{-mrtd} option by using the function attribute | |
9328 | @samp{cdecl}. @xref{Function Attributes}. | |
74291a4b | 9329 | |
39bc1876 NS |
9330 | @strong{Warning:} this calling convention is incompatible with the one |
9331 | normally used on Unix, so you cannot use it if you need to call | |
9332 | libraries compiled with the Unix compiler. | |
74291a4b | 9333 | |
39bc1876 NS |
9334 | Also, you must provide function prototypes for all functions that |
9335 | take variable numbers of arguments (including @code{printf}); | |
9336 | otherwise incorrect code will be generated for calls to those | |
9337 | functions. | |
02f52e19 | 9338 | |
39bc1876 NS |
9339 | In addition, seriously incorrect code will result if you call a |
9340 | function with too many arguments. (Normally, extra arguments are | |
9341 | harmlessly ignored.) | |
1cf959cb | 9342 | |
39bc1876 NS |
9343 | @item -mregparm=@var{num} |
9344 | @opindex mregparm | |
9345 | Control how many registers are used to pass integer arguments. By | |
9346 | default, no registers are used to pass arguments, and at most 3 | |
9347 | registers can be used. You can control this behavior for a specific | |
9348 | function by using the function attribute @samp{regparm}. | |
9349 | @xref{Function Attributes}. | |
1cf959cb | 9350 | |
39bc1876 NS |
9351 | @strong{Warning:} if you use this switch, and |
9352 | @var{num} is nonzero, then you must build all modules with the same | |
9353 | value, including any libraries. This includes the system libraries and | |
9354 | startup modules. | |
5a4b3afd | 9355 | |
1f97667f RG |
9356 | @item -msseregparm |
9357 | @opindex msseregparm | |
9358 | Use SSE register passing conventions for float and double arguments | |
9359 | and return values. You can control this behavior for a specific | |
9360 | function by using the function attribute @samp{sseregparm}. | |
9361 | @xref{Function Attributes}. | |
9362 | ||
9363 | @strong{Warning:} if you use this switch then you must build all | |
9364 | modules with the same value, including any libraries. This includes | |
9365 | the system libraries and startup modules. | |
9366 | ||
39bc1876 NS |
9367 | @item -mpreferred-stack-boundary=@var{num} |
9368 | @opindex mpreferred-stack-boundary | |
9369 | Attempt to keep the stack boundary aligned to a 2 raised to @var{num} | |
9370 | byte boundary. If @option{-mpreferred-stack-boundary} is not specified, | |
9371 | the default is 4 (16 bytes or 128 bits), except when optimizing for code | |
9372 | size (@option{-Os}), in which case the default is the minimum correct | |
9373 | alignment (4 bytes for x86, and 8 bytes for x86-64). | |
5a4b3afd | 9374 | |
39bc1876 NS |
9375 | On Pentium and PentiumPro, @code{double} and @code{long double} values |
9376 | should be aligned to an 8 byte boundary (see @option{-malign-double}) or | |
9377 | suffer significant run time performance penalties. On Pentium III, the | |
9378 | Streaming SIMD Extension (SSE) data type @code{__m128} suffers similar | |
9379 | penalties if it is not 16 byte aligned. | |
5a4b3afd | 9380 | |
39bc1876 NS |
9381 | To ensure proper alignment of this values on the stack, the stack boundary |
9382 | must be as aligned as that required by any value stored on the stack. | |
9383 | Further, every function must be generated such that it keeps the stack | |
9384 | aligned. Thus calling a function compiled with a higher preferred | |
9385 | stack boundary from a function compiled with a lower preferred stack | |
9386 | boundary will most likely misalign the stack. It is recommended that | |
9387 | libraries that use callbacks always use the default setting. | |
5a4b3afd | 9388 | |
39bc1876 NS |
9389 | This extra alignment does consume extra stack space, and generally |
9390 | increases code size. Code that is sensitive to stack space usage, such | |
9391 | as embedded systems and operating system kernels, may want to reduce the | |
9392 | preferred alignment to @option{-mpreferred-stack-boundary=2}. | |
5a4b3afd | 9393 | |
39bc1876 NS |
9394 | @item -mmmx |
9395 | @itemx -mno-mmx | |
9396 | @item -msse | |
9397 | @itemx -mno-sse | |
9398 | @item -msse2 | |
9399 | @itemx -mno-sse2 | |
9400 | @item -msse3 | |
9401 | @itemx -mno-sse3 | |
9402 | @item -m3dnow | |
9403 | @itemx -mno-3dnow | |
9404 | @opindex mmmx | |
9405 | @opindex mno-mmx | |
9406 | @opindex msse | |
9407 | @opindex mno-sse | |
9408 | @opindex m3dnow | |
9409 | @opindex mno-3dnow | |
75576871 BB |
9410 | These switches enable or disable the use of instructions in the MMX, |
9411 | SSE, SSE2 or 3DNow! extended instruction sets. These extensions are | |
9412 | also available as built-in functions: see @ref{X86 Built-in Functions}, | |
9413 | for details of the functions enabled and disabled by these switches. | |
74291a4b | 9414 | |
39bc1876 | 9415 | To have SSE/SSE2 instructions generated automatically from floating-point |
75576871 BB |
9416 | code (as opposed to 387 instructions), see @option{-mfpmath=sse}. |
9417 | ||
9418 | These options will enable GCC to use these extended instructions in | |
9419 | generated code, even without @option{-mfpmath=sse}. Applications which | |
9420 | perform runtime CPU detection must compile separate files for each | |
9421 | supported architecture, using the appropriate flags. In particular, | |
9422 | the file containing the CPU detection code should be compiled without | |
9423 | these options. | |
74291a4b | 9424 | |
39bc1876 NS |
9425 | @item -mpush-args |
9426 | @itemx -mno-push-args | |
9427 | @opindex mpush-args | |
9428 | @opindex mno-push-args | |
9429 | Use PUSH operations to store outgoing parameters. This method is shorter | |
9430 | and usually equally fast as method using SUB/MOV operations and is enabled | |
9431 | by default. In some cases disabling it may improve performance because of | |
9432 | improved scheduling and reduced dependencies. | |
74291a4b | 9433 | |
39bc1876 NS |
9434 | @item -maccumulate-outgoing-args |
9435 | @opindex maccumulate-outgoing-args | |
9436 | If enabled, the maximum amount of space required for outgoing arguments will be | |
9437 | computed in the function prologue. This is faster on most modern CPUs | |
9438 | because of reduced dependencies, improved scheduling and reduced stack usage | |
9439 | when preferred stack boundary is not equal to 2. The drawback is a notable | |
9440 | increase in code size. This switch implies @option{-mno-push-args}. | |
63357d93 | 9441 | |
39bc1876 NS |
9442 | @item -mthreads |
9443 | @opindex mthreads | |
9444 | Support thread-safe exception handling on @samp{Mingw32}. Code that relies | |
9445 | on thread-safe exception handling must compile and link all code with the | |
9446 | @option{-mthreads} option. When compiling, @option{-mthreads} defines | |
9447 | @option{-D_MT}; when linking, it links in a special thread helper library | |
9448 | @option{-lmingwthrd} which cleans up per thread exception handling data. | |
5ef1a99d | 9449 | |
39bc1876 NS |
9450 | @item -mno-align-stringops |
9451 | @opindex mno-align-stringops | |
9452 | Do not align destination of inlined string operations. This switch reduces | |
9453 | code size and improves performance in case the destination is already aligned, | |
9454 | but GCC doesn't know about it. | |
46490403 | 9455 | |
39bc1876 NS |
9456 | @item -minline-all-stringops |
9457 | @opindex minline-all-stringops | |
9458 | By default GCC inlines string operations only when destination is known to be | |
9459 | aligned at least to 4 byte boundary. This enables more inlining, increase code | |
9460 | size, but may improve performance of code that depends on fast memcpy, strlen | |
9461 | and memset for short lengths. | |
c235ddf2 | 9462 | |
39bc1876 NS |
9463 | @item -momit-leaf-frame-pointer |
9464 | @opindex momit-leaf-frame-pointer | |
9465 | Don't keep the frame pointer in a register for leaf functions. This | |
9466 | avoids the instructions to save, set up and restore frame pointers and | |
9467 | makes an extra register available in leaf functions. The option | |
9468 | @option{-fomit-frame-pointer} removes the frame pointer for all functions | |
9469 | which might make debugging harder. | |
c235ddf2 | 9470 | |
39bc1876 NS |
9471 | @item -mtls-direct-seg-refs |
9472 | @itemx -mno-tls-direct-seg-refs | |
9473 | @opindex mtls-direct-seg-refs | |
9474 | Controls whether TLS variables may be accessed with offsets from the | |
9475 | TLS segment register (@code{%gs} for 32-bit, @code{%fs} for 64-bit), | |
9476 | or whether the thread base pointer must be added. Whether or not this | |
9477 | is legal depends on the operating system, and whether it maps the | |
9478 | segment to cover the entire TLS area. | |
beadc644 | 9479 | |
39bc1876 NS |
9480 | For systems that use GNU libc, the default is on. |
9481 | @end table | |
af34e51e | 9482 | |
39bc1876 NS |
9483 | These @samp{-m} switches are supported in addition to the above |
9484 | on AMD x86-64 processors in 64-bit environments. | |
50d32cf6 | 9485 | |
39bc1876 NS |
9486 | @table @gcctabopt |
9487 | @item -m32 | |
9488 | @itemx -m64 | |
9489 | @opindex m32 | |
9490 | @opindex m64 | |
9491 | Generate code for a 32-bit or 64-bit environment. | |
9492 | The 32-bit environment sets int, long and pointer to 32 bits and | |
9493 | generates code that runs on any i386 system. | |
9494 | The 64-bit environment sets int to 32 bits and long and pointer | |
9495 | to 64 bits and generates code for AMD's x86-64 architecture. | |
50d32cf6 | 9496 | |
39bc1876 NS |
9497 | @item -mno-red-zone |
9498 | @opindex no-red-zone | |
9499 | Do not use a so called red zone for x86-64 code. The red zone is mandated | |
9500 | by the x86-64 ABI, it is a 128-byte area beyond the location of the | |
9501 | stack pointer that will not be modified by signal or interrupt handlers | |
9502 | and therefore can be used for temporary data without adjusting the stack | |
9503 | pointer. The flag @option{-mno-red-zone} disables this red zone. | |
dc884a86 | 9504 | |
39bc1876 NS |
9505 | @item -mcmodel=small |
9506 | @opindex mcmodel=small | |
9507 | Generate code for the small code model: the program and its symbols must | |
9508 | be linked in the lower 2 GB of the address space. Pointers are 64 bits. | |
9509 | Programs can be statically or dynamically linked. This is the default | |
9510 | code model. | |
dc884a86 | 9511 | |
39bc1876 NS |
9512 | @item -mcmodel=kernel |
9513 | @opindex mcmodel=kernel | |
9514 | Generate code for the kernel code model. The kernel runs in the | |
9515 | negative 2 GB of the address space. | |
9516 | This model has to be used for Linux kernel code. | |
9517 | ||
9518 | @item -mcmodel=medium | |
9519 | @opindex mcmodel=medium | |
9520 | Generate code for the medium model: The program is linked in the lower 2 | |
9521 | GB of the address space but symbols can be located anywhere in the | |
9522 | address space. Programs can be statically or dynamically linked, but | |
9523 | building of shared libraries are not supported with the medium model. | |
9524 | ||
9525 | @item -mcmodel=large | |
9526 | @opindex mcmodel=large | |
9527 | Generate code for the large model: This model makes no assumptions | |
9528 | about addresses and sizes of sections. Currently GCC does not implement | |
9529 | this model. | |
74291a4b MM |
9530 | @end table |
9531 | ||
39bc1876 NS |
9532 | @node IA-64 Options |
9533 | @subsection IA-64 Options | |
9534 | @cindex IA-64 Options | |
74291a4b | 9535 | |
39bc1876 | 9536 | These are the @samp{-m} options defined for the Intel IA-64 architecture. |
74291a4b | 9537 | |
2642624b | 9538 | @table @gcctabopt |
39bc1876 NS |
9539 | @item -mbig-endian |
9540 | @opindex mbig-endian | |
9541 | Generate code for a big endian target. This is the default for HP-UX@. | |
6f670fde | 9542 | |
39bc1876 NS |
9543 | @item -mlittle-endian |
9544 | @opindex mlittle-endian | |
9545 | Generate code for a little endian target. This is the default for AIX5 | |
9546 | and GNU/Linux. | |
a9f3e1a4 | 9547 | |
39bc1876 NS |
9548 | @item -mgnu-as |
9549 | @itemx -mno-gnu-as | |
9550 | @opindex mgnu-as | |
9551 | @opindex mno-gnu-as | |
9552 | Generate (or don't) code for the GNU assembler. This is the default. | |
9553 | @c Also, this is the default if the configure option @option{--with-gnu-as} | |
9554 | @c is used. | |
9d913bbf | 9555 | |
39bc1876 NS |
9556 | @item -mgnu-ld |
9557 | @itemx -mno-gnu-ld | |
9558 | @opindex mgnu-ld | |
9559 | @opindex mno-gnu-ld | |
9560 | Generate (or don't) code for the GNU linker. This is the default. | |
9561 | @c Also, this is the default if the configure option @option{--with-gnu-ld} | |
9562 | @c is used. | |
a9f3e1a4 | 9563 | |
39bc1876 NS |
9564 | @item -mno-pic |
9565 | @opindex mno-pic | |
9566 | Generate code that does not use a global pointer register. The result | |
9567 | is not position independent code, and violates the IA-64 ABI@. | |
74291a4b | 9568 | |
39bc1876 NS |
9569 | @item -mvolatile-asm-stop |
9570 | @itemx -mno-volatile-asm-stop | |
9571 | @opindex mvolatile-asm-stop | |
9572 | @opindex mno-volatile-asm-stop | |
9573 | Generate (or don't) a stop bit immediately before and after volatile asm | |
9574 | statements. | |
965f5423 | 9575 | |
39bc1876 NS |
9576 | @item -mregister-names |
9577 | @itemx -mno-register-names | |
9578 | @opindex mregister-names | |
9579 | @opindex mno-register-names | |
9580 | Generate (or don't) @samp{in}, @samp{loc}, and @samp{out} register names for | |
9581 | the stacked registers. This may make assembler output more readable. | |
965f5423 | 9582 | |
39bc1876 NS |
9583 | @item -mno-sdata |
9584 | @itemx -msdata | |
9585 | @opindex mno-sdata | |
9586 | @opindex msdata | |
9587 | Disable (or enable) optimizations that use the small data section. This may | |
9588 | be useful for working around optimizer bugs. | |
9589 | ||
9590 | @item -mconstant-gp | |
9591 | @opindex mconstant-gp | |
9592 | Generate code that uses a single constant global pointer value. This is | |
9593 | useful when compiling kernel code. | |
9594 | ||
9595 | @item -mauto-pic | |
9596 | @opindex mauto-pic | |
9597 | Generate code that is self-relocatable. This implies @option{-mconstant-gp}. | |
9598 | This is useful when compiling firmware code. | |
9599 | ||
9600 | @item -minline-float-divide-min-latency | |
9601 | @opindex minline-float-divide-min-latency | |
9602 | Generate code for inline divides of floating point values | |
9603 | using the minimum latency algorithm. | |
965f5423 | 9604 | |
39bc1876 NS |
9605 | @item -minline-float-divide-max-throughput |
9606 | @opindex minline-float-divide-max-throughput | |
9607 | Generate code for inline divides of floating point values | |
9608 | using the maximum throughput algorithm. | |
965f5423 | 9609 | |
39bc1876 NS |
9610 | @item -minline-int-divide-min-latency |
9611 | @opindex minline-int-divide-min-latency | |
9612 | Generate code for inline divides of integer values | |
9613 | using the minimum latency algorithm. | |
965f5423 | 9614 | |
39bc1876 NS |
9615 | @item -minline-int-divide-max-throughput |
9616 | @opindex minline-int-divide-max-throughput | |
9617 | Generate code for inline divides of integer values | |
9618 | using the maximum throughput algorithm. | |
965f5423 | 9619 | |
28b43def SE |
9620 | @item -minline-sqrt-min-latency |
9621 | @opindex minline-sqrt-min-latency | |
9622 | Generate code for inline square roots | |
9623 | using the minimum latency algorithm. | |
9624 | ||
9625 | @item -minline-sqrt-max-throughput | |
9626 | @opindex minline-sqrt-max-throughput | |
9627 | Generate code for inline square roots | |
9628 | using the maximum throughput algorithm. | |
9629 | ||
39bc1876 NS |
9630 | @item -mno-dwarf2-asm |
9631 | @itemx -mdwarf2-asm | |
9632 | @opindex mno-dwarf2-asm | |
9633 | @opindex mdwarf2-asm | |
9634 | Don't (or do) generate assembler code for the DWARF2 line number debugging | |
9635 | info. This may be useful when not using the GNU assembler. | |
965f5423 | 9636 | |
28b43def SE |
9637 | @item -mearly-stop-bits |
9638 | @itemx -mno-early-stop-bits | |
9639 | @opindex mearly-stop-bits | |
9640 | @opindex mno-early-stop-bits | |
9641 | Allow stop bits to be placed earlier than immediately preceding the | |
9642 | instruction that triggered the stop bit. This can improve instruction | |
9643 | scheduling, but does not always do so. | |
9644 | ||
39bc1876 NS |
9645 | @item -mfixed-range=@var{register-range} |
9646 | @opindex mfixed-range | |
9647 | Generate code treating the given register range as fixed registers. | |
9648 | A fixed register is one that the register allocator can not use. This is | |
9649 | useful when compiling kernel code. A register range is specified as | |
9650 | two registers separated by a dash. Multiple register ranges can be | |
9651 | specified separated by a comma. | |
04e149ab | 9652 | |
28b43def SE |
9653 | @item -mtls-size=@var{tls-size} |
9654 | @opindex mtls-size | |
9655 | Specify bit size of immediate TLS offsets. Valid values are 14, 22, and | |
9656 | 64. | |
9657 | ||
7f9dc823 JW |
9658 | @item -mtune=@var{cpu-type} |
9659 | @opindex mtune | |
28b43def SE |
9660 | Tune the instruction scheduling for a particular CPU, Valid values are |
9661 | itanium, itanium1, merced, itanium2, and mckinley. | |
9662 | ||
9663 | @item -mt | |
9664 | @itemx -pthread | |
9665 | @opindex mt | |
9666 | @opindex pthread | |
9667 | Add support for multithreading using the POSIX threads library. This | |
9668 | option sets flags for both the preprocessor and linker. It does | |
9669 | not affect the thread safety of object code produced by the compiler or | |
9670 | that of libraries supplied with it. These are HP-UX specific flags. | |
9671 | ||
9672 | @item -milp32 | |
9673 | @itemx -mlp64 | |
9674 | @opindex milp32 | |
9675 | @opindex mlp64 | |
9676 | Generate code for a 32-bit or 64-bit environment. | |
9677 | The 32-bit environment sets int, long and pointer to 32 bits. | |
9678 | The 64-bit environment sets int to 32 bits and long and pointer | |
9679 | to 64 bits. These are HP-UX specific flags. | |
9680 | ||
39bc1876 | 9681 | @end table |
74291a4b | 9682 | |
38b2d076 DD |
9683 | @node M32C Options |
9684 | @subsection M32C Options | |
9685 | @cindex M32C options | |
9686 | ||
9687 | @table @gcctabopt | |
9688 | @item -mcpu=@var{name} | |
9689 | @opindex mcpu= | |
9690 | Select the CPU for which code is generated. @var{name} may be one of | |
9691 | @samp{r8c} for the R8C/Tiny series, @samp{m16c} for the M16C (up to | |
9692 | /60) series, @samp{m32cm} for the M16C/80 series, or @samp{m32c} for | |
9693 | the M32C/80 series. | |
9694 | ||
9695 | @item -msim | |
9696 | @opindex msim | |
9697 | Specifies that the program will be run on the simulator. This causes | |
9698 | an alternate runtime library to be linked in which supports, for | |
9699 | example, file I/O. You must not use this option when generating | |
9700 | programs that will run on real hardware; you must provide your own | |
9701 | runtime library for whatever I/O functions are needed. | |
9702 | ||
9703 | @item -memregs=@var{number} | |
9704 | @opindex memregs= | |
9705 | Specifies the number of memory-based pseudo-registers GCC will use | |
9706 | during code generation. These pseudo-registers will be used like real | |
9707 | registers, so there is a tradeoff between GCC's ability to fit the | |
9708 | code into available registers, and the performance penalty of using | |
9709 | memory instead of registers. Note that all modules in a program must | |
9710 | be compiled with the same value for this option. Because of that, you | |
9711 | must not use this option with the default runtime libraries gcc | |
9712 | builds. | |
9713 | ||
9714 | @end table | |
9715 | ||
39bc1876 NS |
9716 | @node M32R/D Options |
9717 | @subsection M32R/D Options | |
9718 | @cindex M32R/D options | |
74291a4b | 9719 | |
39bc1876 | 9720 | These @option{-m} options are defined for Renesas M32R/D architectures: |
74291a4b | 9721 | |
39bc1876 NS |
9722 | @table @gcctabopt |
9723 | @item -m32r2 | |
9724 | @opindex m32r2 | |
9725 | Generate code for the M32R/2@. | |
74291a4b | 9726 | |
39bc1876 NS |
9727 | @item -m32rx |
9728 | @opindex m32rx | |
9729 | Generate code for the M32R/X@. | |
74291a4b | 9730 | |
39bc1876 NS |
9731 | @item -m32r |
9732 | @opindex m32r | |
9733 | Generate code for the M32R@. This is the default. | |
74291a4b | 9734 | |
39bc1876 NS |
9735 | @item -mmodel=small |
9736 | @opindex mmodel=small | |
9737 | Assume all objects live in the lower 16MB of memory (so that their addresses | |
9738 | can be loaded with the @code{ld24} instruction), and assume all subroutines | |
9739 | are reachable with the @code{bl} instruction. | |
9740 | This is the default. | |
74291a4b | 9741 | |
39bc1876 NS |
9742 | The addressability of a particular object can be set with the |
9743 | @code{model} attribute. | |
74291a4b | 9744 | |
39bc1876 NS |
9745 | @item -mmodel=medium |
9746 | @opindex mmodel=medium | |
9747 | Assume objects may be anywhere in the 32-bit address space (the compiler | |
9748 | will generate @code{seth/add3} instructions to load their addresses), and | |
9749 | assume all subroutines are reachable with the @code{bl} instruction. | |
194734e9 | 9750 | |
39bc1876 NS |
9751 | @item -mmodel=large |
9752 | @opindex mmodel=large | |
9753 | Assume objects may be anywhere in the 32-bit address space (the compiler | |
9754 | will generate @code{seth/add3} instructions to load their addresses), and | |
9755 | assume subroutines may not be reachable with the @code{bl} instruction | |
9756 | (the compiler will generate the much slower @code{seth/add3/jl} | |
9757 | instruction sequence). | |
a5f3dd66 | 9758 | |
39bc1876 NS |
9759 | @item -msdata=none |
9760 | @opindex msdata=none | |
9761 | Disable use of the small data area. Variables will be put into | |
9762 | one of @samp{.data}, @samp{bss}, or @samp{.rodata} (unless the | |
9763 | @code{section} attribute has been specified). | |
9764 | This is the default. | |
a5f3dd66 | 9765 | |
39bc1876 NS |
9766 | The small data area consists of sections @samp{.sdata} and @samp{.sbss}. |
9767 | Objects may be explicitly put in the small data area with the | |
9768 | @code{section} attribute using one of these sections. | |
daf2f129 | 9769 | |
39bc1876 NS |
9770 | @item -msdata=sdata |
9771 | @opindex msdata=sdata | |
9772 | Put small global and static data in the small data area, but do not | |
9773 | generate special code to reference them. | |
2b589241 | 9774 | |
39bc1876 NS |
9775 | @item -msdata=use |
9776 | @opindex msdata=use | |
9777 | Put small global and static data in the small data area, and generate | |
9778 | special instructions to reference them. | |
2b589241 | 9779 | |
39bc1876 NS |
9780 | @item -G @var{num} |
9781 | @opindex G | |
9782 | @cindex smaller data references | |
9783 | Put global and static objects less than or equal to @var{num} bytes | |
9784 | into the small data or bss sections instead of the normal data or bss | |
9785 | sections. The default value of @var{num} is 8. | |
9786 | The @option{-msdata} option must be set to one of @samp{sdata} or @samp{use} | |
9787 | for this option to have any effect. | |
74291a4b | 9788 | |
39bc1876 NS |
9789 | All modules should be compiled with the same @option{-G @var{num}} value. |
9790 | Compiling with different values of @var{num} may or may not work; if it | |
9791 | doesn't the linker will give an error message---incorrect code will not be | |
9792 | generated. | |
74291a4b | 9793 | |
39bc1876 NS |
9794 | @item -mdebug |
9795 | @opindex mdebug | |
9796 | Makes the M32R specific code in the compiler display some statistics | |
9797 | that might help in debugging programs. | |
74291a4b | 9798 | |
39bc1876 NS |
9799 | @item -malign-loops |
9800 | @opindex malign-loops | |
9801 | Align all loops to a 32-byte boundary. | |
74291a4b | 9802 | |
39bc1876 NS |
9803 | @item -mno-align-loops |
9804 | @opindex mno-align-loops | |
9805 | Do not enforce a 32-byte alignment for loops. This is the default. | |
74291a4b | 9806 | |
39bc1876 NS |
9807 | @item -missue-rate=@var{number} |
9808 | @opindex missue-rate=@var{number} | |
9809 | Issue @var{number} instructions per cycle. @var{number} can only be 1 | |
9810 | or 2. | |
74291a4b | 9811 | |
39bc1876 NS |
9812 | @item -mbranch-cost=@var{number} |
9813 | @opindex mbranch-cost=@var{number} | |
9814 | @var{number} can only be 1 or 2. If it is 1 then branches will be | |
9815 | preferred over conditional code, if it is 2, then the opposite will | |
9816 | apply. | |
74291a4b | 9817 | |
39bc1876 NS |
9818 | @item -mflush-trap=@var{number} |
9819 | @opindex mflush-trap=@var{number} | |
9820 | Specifies the trap number to use to flush the cache. The default is | |
9821 | 12. Valid numbers are between 0 and 15 inclusive. | |
74291a4b | 9822 | |
39bc1876 NS |
9823 | @item -mno-flush-trap |
9824 | @opindex mno-flush-trap | |
9825 | Specifies that the cache cannot be flushed by using a trap. | |
74291a4b | 9826 | |
39bc1876 NS |
9827 | @item -mflush-func=@var{name} |
9828 | @opindex mflush-func=@var{name} | |
9829 | Specifies the name of the operating system function to call to flush | |
9830 | the cache. The default is @emph{_flush_cache}, but a function call | |
9831 | will only be used if a trap is not available. | |
3af4bd89 | 9832 | |
39bc1876 NS |
9833 | @item -mno-flush-func |
9834 | @opindex mno-flush-func | |
9835 | Indicates that there is no OS function for flushing the cache. | |
3af4bd89 | 9836 | |
39bc1876 | 9837 | @end table |
3af4bd89 | 9838 | |
39bc1876 NS |
9839 | @node M680x0 Options |
9840 | @subsection M680x0 Options | |
9841 | @cindex M680x0 options | |
f22a97d2 | 9842 | |
39bc1876 NS |
9843 | These are the @samp{-m} options defined for the 68000 series. The default |
9844 | values for these options depends on which style of 68000 was selected when | |
9845 | the compiler was configured; the defaults for the most common choices are | |
9846 | given below. | |
1255c85c | 9847 | |
39bc1876 NS |
9848 | @table @gcctabopt |
9849 | @item -m68000 | |
9850 | @itemx -mc68000 | |
9851 | @opindex m68000 | |
9852 | @opindex mc68000 | |
9853 | Generate output for a 68000. This is the default | |
9854 | when the compiler is configured for 68000-based systems. | |
1255c85c | 9855 | |
39bc1876 NS |
9856 | Use this option for microcontrollers with a 68000 or EC000 core, |
9857 | including the 68008, 68302, 68306, 68307, 68322, 68328 and 68356. | |
7eafc329 | 9858 | |
39bc1876 NS |
9859 | @item -m68020 |
9860 | @itemx -mc68020 | |
9861 | @opindex m68020 | |
9862 | @opindex mc68020 | |
9863 | Generate output for a 68020. This is the default | |
9864 | when the compiler is configured for 68020-based systems. | |
f73ad30e | 9865 | |
39bc1876 NS |
9866 | @item -m68881 |
9867 | @opindex m68881 | |
9868 | Generate output containing 68881 instructions for floating point. | |
9869 | This is the default for most 68020 systems unless @option{--nfp} was | |
9870 | specified when the compiler was configured. | |
f73ad30e | 9871 | |
39bc1876 NS |
9872 | @item -m68030 |
9873 | @opindex m68030 | |
9874 | Generate output for a 68030. This is the default when the compiler is | |
9875 | configured for 68030-based systems. | |
79f05c19 | 9876 | |
39bc1876 NS |
9877 | @item -m68040 |
9878 | @opindex m68040 | |
9879 | Generate output for a 68040. This is the default when the compiler is | |
9880 | configured for 68040-based systems. | |
79f05c19 | 9881 | |
39bc1876 NS |
9882 | This option inhibits the use of 68881/68882 instructions that have to be |
9883 | emulated by software on the 68040. Use this option if your 68040 does not | |
9884 | have code to emulate those instructions. | |
762e166b | 9885 | |
39bc1876 NS |
9886 | @item -m68060 |
9887 | @opindex m68060 | |
9888 | Generate output for a 68060. This is the default when the compiler is | |
9889 | configured for 68060-based systems. | |
74dc3e94 | 9890 | |
39bc1876 NS |
9891 | This option inhibits the use of 68020 and 68881/68882 instructions that |
9892 | have to be emulated by software on the 68060. Use this option if your 68060 | |
9893 | does not have code to emulate those instructions. | |
74dc3e94 | 9894 | |
39bc1876 NS |
9895 | @item -mcpu32 |
9896 | @opindex mcpu32 | |
9897 | Generate output for a CPU32. This is the default | |
9898 | when the compiler is configured for CPU32-based systems. | |
3af4bd89 | 9899 | |
39bc1876 NS |
9900 | Use this option for microcontrollers with a |
9901 | CPU32 or CPU32+ core, including the 68330, 68331, 68332, 68333, 68334, | |
9902 | 68336, 68340, 68341, 68349 and 68360. | |
14f73b5a | 9903 | |
39bc1876 NS |
9904 | @item -m5200 |
9905 | @opindex m5200 | |
9906 | Generate output for a 520X ``coldfire'' family cpu. This is the default | |
9907 | when the compiler is configured for 520X-based systems. | |
14f73b5a | 9908 | |
39bc1876 NS |
9909 | Use this option for microcontroller with a 5200 core, including |
9910 | the MCF5202, MCF5203, MCF5204 and MCF5202. | |
a7701995 | 9911 | |
dcc21c4c PB |
9912 | @item -mcfv4e |
9913 | @opindex mcfv4e | |
9914 | Generate output for a ColdFire V4e family cpu (e.g.@: 547x/548x). | |
9915 | This includes use of hardware floating point instructions. | |
a7701995 | 9916 | |
39bc1876 NS |
9917 | @item -m68020-40 |
9918 | @opindex m68020-40 | |
9919 | Generate output for a 68040, without using any of the new instructions. | |
9920 | This results in code which can run relatively efficiently on either a | |
9921 | 68020/68881 or a 68030 or a 68040. The generated code does use the | |
9922 | 68881 instructions that are emulated on the 68040. | |
a7701995 | 9923 | |
39bc1876 NS |
9924 | @item -m68020-60 |
9925 | @opindex m68020-60 | |
9926 | Generate output for a 68060, without using any of the new instructions. | |
9927 | This results in code which can run relatively efficiently on either a | |
9928 | 68020/68881 or a 68030 or a 68040. The generated code does use the | |
9929 | 68881 instructions that are emulated on the 68060. | |
a7701995 | 9930 | |
39bc1876 NS |
9931 | @item -msoft-float |
9932 | @opindex msoft-float | |
9933 | Generate output containing library calls for floating point. | |
9934 | @strong{Warning:} the requisite libraries are not available for all m68k | |
9935 | targets. Normally the facilities of the machine's usual C compiler are | |
9936 | used, but this can't be done directly in cross-compilation. You must | |
9937 | make your own arrangements to provide suitable library functions for | |
9938 | cross-compilation. The embedded targets @samp{m68k-*-aout} and | |
9939 | @samp{m68k-*-coff} do provide software floating point support. | |
14f73b5a | 9940 | |
39bc1876 NS |
9941 | @item -mshort |
9942 | @opindex mshort | |
9943 | Consider type @code{int} to be 16 bits wide, like @code{short int}. | |
9944 | Additionally, parameters passed on the stack are also aligned to a | |
9945 | 16-bit boundary even on targets whose API mandates promotion to 32-bit. | |
74291a4b | 9946 | |
39bc1876 NS |
9947 | @item -mnobitfield |
9948 | @opindex mnobitfield | |
9949 | Do not use the bit-field instructions. The @option{-m68000}, @option{-mcpu32} | |
9950 | and @option{-m5200} options imply @w{@option{-mnobitfield}}. | |
74291a4b | 9951 | |
39bc1876 NS |
9952 | @item -mbitfield |
9953 | @opindex mbitfield | |
9954 | Do use the bit-field instructions. The @option{-m68020} option implies | |
9955 | @option{-mbitfield}. This is the default if you use a configuration | |
9956 | designed for a 68020. | |
9957 | ||
9958 | @item -mrtd | |
9959 | @opindex mrtd | |
9960 | Use a different function-calling convention, in which functions | |
9961 | that take a fixed number of arguments return with the @code{rtd} | |
9962 | instruction, which pops their arguments while returning. This | |
9963 | saves one instruction in the caller since there is no need to pop | |
9964 | the arguments there. | |
ea3bfbfe | 9965 | |
39bc1876 NS |
9966 | This calling convention is incompatible with the one normally |
9967 | used on Unix, so you cannot use it if you need to call libraries | |
9968 | compiled with the Unix compiler. | |
74291a4b | 9969 | |
39bc1876 NS |
9970 | Also, you must provide function prototypes for all functions that |
9971 | take variable numbers of arguments (including @code{printf}); | |
9972 | otherwise incorrect code will be generated for calls to those | |
9973 | functions. | |
861bb6c1 | 9974 | |
39bc1876 NS |
9975 | In addition, seriously incorrect code will result if you call a |
9976 | function with too many arguments. (Normally, extra arguments are | |
9977 | harmlessly ignored.) | |
74291a4b | 9978 | |
39bc1876 NS |
9979 | The @code{rtd} instruction is supported by the 68010, 68020, 68030, |
9980 | 68040, 68060 and CPU32 processors, but not by the 68000 or 5200. | |
74291a4b | 9981 | |
39bc1876 NS |
9982 | @item -malign-int |
9983 | @itemx -mno-align-int | |
9984 | @opindex malign-int | |
9985 | @opindex mno-align-int | |
9986 | Control whether GCC aligns @code{int}, @code{long}, @code{long long}, | |
9987 | @code{float}, @code{double}, and @code{long double} variables on a 32-bit | |
9988 | boundary (@option{-malign-int}) or a 16-bit boundary (@option{-mno-align-int}). | |
9989 | Aligning variables on 32-bit boundaries produces code that runs somewhat | |
9990 | faster on processors with 32-bit busses at the expense of more memory. | |
74291a4b | 9991 | |
39bc1876 NS |
9992 | @strong{Warning:} if you use the @option{-malign-int} switch, GCC will |
9993 | align structures containing the above types differently than | |
9994 | most published application binary interface specifications for the m68k. | |
74291a4b | 9995 | |
39bc1876 NS |
9996 | @item -mpcrel |
9997 | @opindex mpcrel | |
9998 | Use the pc-relative addressing mode of the 68000 directly, instead of | |
9999 | using a global offset table. At present, this option implies @option{-fpic}, | |
10000 | allowing at most a 16-bit offset for pc-relative addressing. @option{-fPIC} is | |
10001 | not presently supported with @option{-mpcrel}, though this could be supported for | |
10002 | 68020 and higher processors. | |
74291a4b | 10003 | |
39bc1876 NS |
10004 | @item -mno-strict-align |
10005 | @itemx -mstrict-align | |
10006 | @opindex mno-strict-align | |
10007 | @opindex mstrict-align | |
10008 | Do not (do) assume that unaligned memory references will be handled by | |
10009 | the system. | |
74291a4b | 10010 | |
39bc1876 NS |
10011 | @item -msep-data |
10012 | Generate code that allows the data segment to be located in a different | |
10013 | area of memory from the text segment. This allows for execute in place in | |
4ec7afd7 KH |
10014 | an environment without virtual memory management. This option implies |
10015 | @option{-fPIC}. | |
74291a4b | 10016 | |
39bc1876 NS |
10017 | @item -mno-sep-data |
10018 | Generate code that assumes that the data segment follows the text segment. | |
10019 | This is the default. | |
74291a4b | 10020 | |
39bc1876 NS |
10021 | @item -mid-shared-library |
10022 | Generate code that supports shared libraries via the library ID method. | |
10023 | This allows for execute in place and shared libraries in an environment | |
4ec7afd7 | 10024 | without virtual memory management. This option implies @option{-fPIC}. |
74291a4b | 10025 | |
39bc1876 NS |
10026 | @item -mno-id-shared-library |
10027 | Generate code that doesn't assume ID based shared libraries are being used. | |
10028 | This is the default. | |
74291a4b | 10029 | |
39bc1876 NS |
10030 | @item -mshared-library-id=n |
10031 | Specified the identification number of the ID based shared library being | |
10032 | compiled. Specifying a value of 0 will generate more compact code, specifying | |
10033 | other values will force the allocation of that number to the current | |
10034 | library but is no more space or time efficient than omitting this option. | |
74291a4b | 10035 | |
39bc1876 | 10036 | @end table |
74291a4b | 10037 | |
39bc1876 NS |
10038 | @node M68hc1x Options |
10039 | @subsection M68hc1x Options | |
10040 | @cindex M68hc1x options | |
74291a4b | 10041 | |
39bc1876 NS |
10042 | These are the @samp{-m} options defined for the 68hc11 and 68hc12 |
10043 | microcontrollers. The default values for these options depends on | |
10044 | which style of microcontroller was selected when the compiler was configured; | |
10045 | the defaults for the most common choices are given below. | |
c219e1da | 10046 | |
39bc1876 NS |
10047 | @table @gcctabopt |
10048 | @item -m6811 | |
10049 | @itemx -m68hc11 | |
10050 | @opindex m6811 | |
10051 | @opindex m68hc11 | |
10052 | Generate output for a 68HC11. This is the default | |
10053 | when the compiler is configured for 68HC11-based systems. | |
c5d3d49b | 10054 | |
39bc1876 NS |
10055 | @item -m6812 |
10056 | @itemx -m68hc12 | |
10057 | @opindex m6812 | |
10058 | @opindex m68hc12 | |
10059 | Generate output for a 68HC12. This is the default | |
10060 | when the compiler is configured for 68HC12-based systems. | |
c5d3d49b | 10061 | |
39bc1876 NS |
10062 | @item -m68S12 |
10063 | @itemx -m68hcs12 | |
10064 | @opindex m68S12 | |
10065 | @opindex m68hcs12 | |
10066 | Generate output for a 68HCS12. | |
c5d3d49b | 10067 | |
39bc1876 NS |
10068 | @item -mauto-incdec |
10069 | @opindex mauto-incdec | |
10070 | Enable the use of 68HC12 pre and post auto-increment and auto-decrement | |
10071 | addressing modes. | |
34208acf | 10072 | |
39bc1876 NS |
10073 | @item -minmax |
10074 | @itemx -nominmax | |
10075 | @opindex minmax | |
10076 | @opindex mnominmax | |
10077 | Enable the use of 68HC12 min and max instructions. | |
34208acf | 10078 | |
39bc1876 NS |
10079 | @item -mlong-calls |
10080 | @itemx -mno-long-calls | |
10081 | @opindex mlong-calls | |
10082 | @opindex mno-long-calls | |
10083 | Treat all calls as being far away (near). If calls are assumed to be | |
10084 | far away, the compiler will use the @code{call} instruction to | |
10085 | call a function and the @code{rtc} instruction for returning. | |
34208acf | 10086 | |
39bc1876 NS |
10087 | @item -mshort |
10088 | @opindex mshort | |
10089 | Consider type @code{int} to be 16 bits wide, like @code{short int}. | |
34208acf | 10090 | |
39bc1876 NS |
10091 | @item -msoft-reg-count=@var{count} |
10092 | @opindex msoft-reg-count | |
10093 | Specify the number of pseudo-soft registers which are used for the | |
10094 | code generation. The maximum number is 32. Using more pseudo-soft | |
10095 | register may or may not result in better code depending on the program. | |
10096 | The default is 4 for 68HC11 and 2 for 68HC12. | |
34208acf | 10097 | |
39bc1876 | 10098 | @end table |
34208acf | 10099 | |
39bc1876 NS |
10100 | @node MCore Options |
10101 | @subsection MCore Options | |
10102 | @cindex MCore options | |
34208acf | 10103 | |
39bc1876 NS |
10104 | These are the @samp{-m} options defined for the Motorola M*Core |
10105 | processors. | |
34208acf | 10106 | |
39bc1876 | 10107 | @table @gcctabopt |
34208acf | 10108 | |
39bc1876 NS |
10109 | @item -mhardlit |
10110 | @itemx -mno-hardlit | |
10111 | @opindex mhardlit | |
10112 | @opindex mno-hardlit | |
10113 | Inline constants into the code stream if it can be done in two | |
10114 | instructions or less. | |
34208acf | 10115 | |
39bc1876 NS |
10116 | @item -mdiv |
10117 | @itemx -mno-div | |
10118 | @opindex mdiv | |
10119 | @opindex mno-div | |
10120 | Use the divide instruction. (Enabled by default). | |
a02aa5b0 | 10121 | |
39bc1876 NS |
10122 | @item -mrelax-immediate |
10123 | @itemx -mno-relax-immediate | |
10124 | @opindex mrelax-immediate | |
10125 | @opindex mno-relax-immediate | |
10126 | Allow arbitrary sized immediates in bit operations. | |
a02aa5b0 | 10127 | |
39bc1876 NS |
10128 | @item -mwide-bitfields |
10129 | @itemx -mno-wide-bitfields | |
10130 | @opindex mwide-bitfields | |
10131 | @opindex mno-wide-bitfields | |
10132 | Always treat bit-fields as int-sized. | |
a02aa5b0 | 10133 | |
39bc1876 NS |
10134 | @item -m4byte-functions |
10135 | @itemx -mno-4byte-functions | |
10136 | @opindex m4byte-functions | |
10137 | @opindex mno-4byte-functions | |
10138 | Force all functions to be aligned to a four byte boundary. | |
a02aa5b0 | 10139 | |
39bc1876 NS |
10140 | @item -mcallgraph-data |
10141 | @itemx -mno-callgraph-data | |
10142 | @opindex mcallgraph-data | |
10143 | @opindex mno-callgraph-data | |
10144 | Emit callgraph information. | |
f401d0f5 | 10145 | |
39bc1876 NS |
10146 | @item -mslow-bytes |
10147 | @itemx -mno-slow-bytes | |
10148 | @opindex mslow-bytes | |
10149 | @opindex mno-slow-bytes | |
10150 | Prefer word access when reading byte quantities. | |
f401d0f5 | 10151 | |
39bc1876 NS |
10152 | @item -mlittle-endian |
10153 | @itemx -mbig-endian | |
10154 | @opindex mlittle-endian | |
10155 | @opindex mbig-endian | |
10156 | Generate code for a little endian target. | |
f401d0f5 | 10157 | |
39bc1876 NS |
10158 | @item -m210 |
10159 | @itemx -m340 | |
10160 | @opindex m210 | |
10161 | @opindex m340 | |
10162 | Generate code for the 210 processor. | |
74291a4b MM |
10163 | @end table |
10164 | ||
39bc1876 NS |
10165 | @node MIPS Options |
10166 | @subsection MIPS Options | |
10167 | @cindex MIPS options | |
74291a4b | 10168 | |
2642624b | 10169 | @table @gcctabopt |
74291a4b | 10170 | |
39bc1876 NS |
10171 | @item -EB |
10172 | @opindex EB | |
10173 | Generate big-endian code. | |
74291a4b | 10174 | |
39bc1876 NS |
10175 | @item -EL |
10176 | @opindex EL | |
10177 | Generate little-endian code. This is the default for @samp{mips*el-*-*} | |
10178 | configurations. | |
74291a4b | 10179 | |
39bc1876 NS |
10180 | @item -march=@var{arch} |
10181 | @opindex march | |
10182 | Generate code that will run on @var{arch}, which can be the name of a | |
10183 | generic MIPS ISA, or the name of a particular processor. | |
10184 | The ISA names are: | |
10185 | @samp{mips1}, @samp{mips2}, @samp{mips3}, @samp{mips4}, | |
10186 | @samp{mips32}, @samp{mips32r2}, and @samp{mips64}. | |
10187 | The processor names are: | |
79dad081 DU |
10188 | @samp{4kc}, @samp{4km}, @samp{4kp}, |
10189 | @samp{5kc}, @samp{5kf}, | |
10190 | @samp{20kc}, | |
10191 | @samp{24k}, @samp{24kc}, @samp{24kf}, @samp{24kx}, | |
39bc1876 | 10192 | @samp{m4k}, |
39bc1876 | 10193 | @samp{orion}, |
79dad081 DU |
10194 | @samp{r2000}, @samp{r3000}, @samp{r3900}, @samp{r4000}, @samp{r4400}, |
10195 | @samp{r4600}, @samp{r4650}, @samp{r6000}, @samp{r8000}, | |
10196 | @samp{rm7000}, @samp{rm9000}, | |
39bc1876 | 10197 | @samp{sb1}, |
79dad081 | 10198 | @samp{sr71000}, |
39bc1876 NS |
10199 | @samp{vr4100}, @samp{vr4111}, @samp{vr4120}, @samp{vr4130}, @samp{vr4300}, |
10200 | @samp{vr5000}, @samp{vr5400} and @samp{vr5500}. | |
10201 | The special value @samp{from-abi} selects the | |
10202 | most compatible architecture for the selected ABI (that is, | |
10203 | @samp{mips1} for 32-bit ABIs and @samp{mips3} for 64-bit ABIs)@. | |
74291a4b | 10204 | |
39bc1876 NS |
10205 | In processor names, a final @samp{000} can be abbreviated as @samp{k} |
10206 | (for example, @samp{-march=r2k}). Prefixes are optional, and | |
10207 | @samp{vr} may be written @samp{r}. | |
74291a4b | 10208 | |
39bc1876 NS |
10209 | GCC defines two macros based on the value of this option. The first |
10210 | is @samp{_MIPS_ARCH}, which gives the name of target architecture, as | |
10211 | a string. The second has the form @samp{_MIPS_ARCH_@var{foo}}, | |
10212 | where @var{foo} is the capitalized value of @samp{_MIPS_ARCH}@. | |
10213 | For example, @samp{-march=r2000} will set @samp{_MIPS_ARCH} | |
10214 | to @samp{"r2000"} and define the macro @samp{_MIPS_ARCH_R2000}. | |
74291a4b | 10215 | |
39bc1876 NS |
10216 | Note that the @samp{_MIPS_ARCH} macro uses the processor names given |
10217 | above. In other words, it will have the full prefix and will not | |
10218 | abbreviate @samp{000} as @samp{k}. In the case of @samp{from-abi}, | |
10219 | the macro names the resolved architecture (either @samp{"mips1"} or | |
10220 | @samp{"mips3"}). It names the default architecture when no | |
10221 | @option{-march} option is given. | |
74291a4b | 10222 | |
39bc1876 NS |
10223 | @item -mtune=@var{arch} |
10224 | @opindex mtune | |
10225 | Optimize for @var{arch}. Among other things, this option controls | |
10226 | the way instructions are scheduled, and the perceived cost of arithmetic | |
10227 | operations. The list of @var{arch} values is the same as for | |
10228 | @option{-march}. | |
74291a4b | 10229 | |
39bc1876 NS |
10230 | When this option is not used, GCC will optimize for the processor |
10231 | specified by @option{-march}. By using @option{-march} and | |
10232 | @option{-mtune} together, it is possible to generate code that will | |
10233 | run on a family of processors, but optimize the code for one | |
10234 | particular member of that family. | |
74291a4b | 10235 | |
39bc1876 NS |
10236 | @samp{-mtune} defines the macros @samp{_MIPS_TUNE} and |
10237 | @samp{_MIPS_TUNE_@var{foo}}, which work in the same way as the | |
10238 | @samp{-march} ones described above. | |
74291a4b | 10239 | |
39bc1876 NS |
10240 | @item -mips1 |
10241 | @opindex mips1 | |
10242 | Equivalent to @samp{-march=mips1}. | |
74291a4b | 10243 | |
39bc1876 NS |
10244 | @item -mips2 |
10245 | @opindex mips2 | |
10246 | Equivalent to @samp{-march=mips2}. | |
74291a4b | 10247 | |
39bc1876 NS |
10248 | @item -mips3 |
10249 | @opindex mips3 | |
10250 | Equivalent to @samp{-march=mips3}. | |
74291a4b | 10251 | |
39bc1876 NS |
10252 | @item -mips4 |
10253 | @opindex mips4 | |
10254 | Equivalent to @samp{-march=mips4}. | |
10255 | ||
10256 | @item -mips32 | |
10257 | @opindex mips32 | |
10258 | Equivalent to @samp{-march=mips32}. | |
74291a4b | 10259 | |
39bc1876 NS |
10260 | @item -mips32r2 |
10261 | @opindex mips32r2 | |
10262 | Equivalent to @samp{-march=mips32r2}. | |
74291a4b | 10263 | |
39bc1876 NS |
10264 | @item -mips64 |
10265 | @opindex mips64 | |
10266 | Equivalent to @samp{-march=mips64}. | |
74291a4b | 10267 | |
39bc1876 NS |
10268 | @item -mips16 |
10269 | @itemx -mno-mips16 | |
10270 | @opindex mips16 | |
10271 | @opindex mno-mips16 | |
7cc63a88 | 10272 | Generate (do not generate) MIPS16 code. If GCC is targetting a |
71cb39e6 | 10273 | MIPS32 or MIPS64 architecture, it will make use of the MIPS16e ASE@. |
74291a4b | 10274 | |
39bc1876 NS |
10275 | @item -mabi=32 |
10276 | @itemx -mabi=o64 | |
10277 | @itemx -mabi=n32 | |
10278 | @itemx -mabi=64 | |
10279 | @itemx -mabi=eabi | |
10280 | @opindex mabi=32 | |
10281 | @opindex mabi=o64 | |
10282 | @opindex mabi=n32 | |
10283 | @opindex mabi=64 | |
10284 | @opindex mabi=eabi | |
10285 | Generate code for the given ABI@. | |
74291a4b | 10286 | |
39bc1876 NS |
10287 | Note that the EABI has a 32-bit and a 64-bit variant. GCC normally |
10288 | generates 64-bit code when you select a 64-bit architecture, but you | |
10289 | can use @option{-mgp32} to get 32-bit code instead. | |
74291a4b | 10290 | |
39bc1876 NS |
10291 | For information about the O64 ABI, see |
10292 | @w{@uref{http://gcc.gnu.org/projects/mipso64-abi.html}}. | |
74291a4b | 10293 | |
39bc1876 NS |
10294 | @item -mabicalls |
10295 | @itemx -mno-abicalls | |
10296 | @opindex mabicalls | |
10297 | @opindex mno-abicalls | |
10298 | Generate (do not generate) SVR4-style position-independent code. | |
10299 | @option{-mabicalls} is the default for SVR4-based systems. | |
74291a4b | 10300 | |
39bc1876 NS |
10301 | @item -mxgot |
10302 | @itemx -mno-xgot | |
10303 | @opindex mxgot | |
10304 | @opindex mno-xgot | |
10305 | Lift (do not lift) the usual restrictions on the size of the global | |
10306 | offset table. | |
74291a4b | 10307 | |
8a36672b | 10308 | GCC normally uses a single instruction to load values from the GOT@. |
39bc1876 NS |
10309 | While this is relatively efficient, it will only work if the GOT |
10310 | is smaller than about 64k. Anything larger will cause the linker | |
10311 | to report an error such as: | |
74291a4b | 10312 | |
39bc1876 NS |
10313 | @cindex relocation truncated to fit (MIPS) |
10314 | @smallexample | |
10315 | relocation truncated to fit: R_MIPS_GOT16 foobar | |
10316 | @end smallexample | |
74291a4b | 10317 | |
39bc1876 NS |
10318 | If this happens, you should recompile your code with @option{-mxgot}. |
10319 | It should then work with very large GOTs, although it will also be | |
10320 | less efficient, since it will take three instructions to fetch the | |
10321 | value of a global symbol. | |
956d6950 | 10322 | |
39bc1876 NS |
10323 | Note that some linkers can create multiple GOTs. If you have such a |
10324 | linker, you should only need to use @option{-mxgot} when a single object | |
10325 | file accesses more than 64k's worth of GOT entries. Very few do. | |
956d6950 | 10326 | |
39bc1876 NS |
10327 | These options have no effect unless GCC is generating position |
10328 | independent code. | |
956d6950 | 10329 | |
39bc1876 NS |
10330 | @item -mgp32 |
10331 | @opindex mgp32 | |
10332 | Assume that general-purpose registers are 32 bits wide. | |
58605ba0 | 10333 | |
39bc1876 NS |
10334 | @item -mgp64 |
10335 | @opindex mgp64 | |
10336 | Assume that general-purpose registers are 64 bits wide. | |
58605ba0 | 10337 | |
39bc1876 NS |
10338 | @item -mfp32 |
10339 | @opindex mfp32 | |
10340 | Assume that floating-point registers are 32 bits wide. | |
58605ba0 | 10341 | |
39bc1876 NS |
10342 | @item -mfp64 |
10343 | @opindex mfp64 | |
10344 | Assume that floating-point registers are 64 bits wide. | |
58605ba0 | 10345 | |
39bc1876 NS |
10346 | @item -mhard-float |
10347 | @opindex mhard-float | |
10348 | Use floating-point coprocessor instructions. | |
58605ba0 | 10349 | |
39bc1876 NS |
10350 | @item -msoft-float |
10351 | @opindex msoft-float | |
10352 | Do not use floating-point coprocessor instructions. Implement | |
10353 | floating-point calculations using library calls instead. | |
3094247f | 10354 | |
39bc1876 NS |
10355 | @item -msingle-float |
10356 | @opindex msingle-float | |
10357 | Assume that the floating-point coprocessor only supports single-precision | |
10358 | operations. | |
3094247f | 10359 | |
39bc1876 NS |
10360 | @itemx -mdouble-float |
10361 | @opindex mdouble-float | |
10362 | Assume that the floating-point coprocessor supports double-precision | |
10363 | operations. This is the default. | |
956d6950 | 10364 | |
118ea793 CF |
10365 | @itemx -mdsp |
10366 | @itemx -mno-dsp | |
10367 | @opindex mdsp | |
10368 | @opindex mno-dsp | |
10369 | Use (do not use) the MIPS DSP ASE. @xref{MIPS DSP Built-in Functions}. | |
10370 | ||
06a4ab70 CF |
10371 | @itemx -mpaired-single |
10372 | @itemx -mno-paired-single | |
10373 | @opindex mpaired-single | |
10374 | @opindex mno-paired-single | |
d840bfd3 CF |
10375 | Use (do not use) paired-single floating-point instructions. |
10376 | @xref{MIPS Paired-Single Support}. This option can only be used | |
10377 | when generating 64-bit code and requires hardware floating-point | |
10378 | support to be enabled. | |
06a4ab70 CF |
10379 | |
10380 | @itemx -mips3d | |
10381 | @itemx -mno-mips3d | |
10382 | @opindex mips3d | |
10383 | @opindex mno-mips3d | |
8a36672b | 10384 | Use (do not use) the MIPS-3D ASE@. @xref{MIPS-3D Built-in Functions}. |
d840bfd3 | 10385 | The option @option{-mips3d} implies @option{-mpaired-single}. |
06a4ab70 | 10386 | |
39bc1876 NS |
10387 | @item -mlong64 |
10388 | @opindex mlong64 | |
10389 | Force @code{long} types to be 64 bits wide. See @option{-mlong32} for | |
10390 | an explanation of the default and the way that the pointer size is | |
10391 | determined. | |
956d6950 | 10392 | |
39bc1876 NS |
10393 | @item -mlong32 |
10394 | @opindex mlong32 | |
10395 | Force @code{long}, @code{int}, and pointer types to be 32 bits wide. | |
956d6950 | 10396 | |
39bc1876 NS |
10397 | The default size of @code{int}s, @code{long}s and pointers depends on |
10398 | the ABI@. All the supported ABIs use 32-bit @code{int}s. The n64 ABI | |
10399 | uses 64-bit @code{long}s, as does the 64-bit EABI; the others use | |
10400 | 32-bit @code{long}s. Pointers are the same size as @code{long}s, | |
10401 | or the same size as integer registers, whichever is smaller. | |
956d6950 | 10402 | |
acda0e26 RS |
10403 | @item -msym32 |
10404 | @itemx -mno-sym32 | |
10405 | @opindex msym32 | |
10406 | @opindex mno-sym32 | |
10407 | Assume (do not assume) that all symbols have 32-bit values, regardless | |
10408 | of the selected ABI@. This option is useful in combination with | |
10409 | @option{-mabi=64} and @option{-mno-abicalls} because it allows GCC | |
10410 | to generate shorter and faster references to symbolic addresses. | |
10411 | ||
39bc1876 NS |
10412 | @item -G @var{num} |
10413 | @opindex G | |
10414 | @cindex smaller data references (MIPS) | |
10415 | @cindex gp-relative references (MIPS) | |
10416 | Put global and static items less than or equal to @var{num} bytes into | |
10417 | the small data or bss section instead of the normal data or bss section. | |
10418 | This allows the data to be accessed using a single instruction. | |
10419 | ||
10420 | All modules should be compiled with the same @option{-G @var{num}} | |
10421 | value. | |
956d6950 | 10422 | |
39bc1876 NS |
10423 | @item -membedded-data |
10424 | @itemx -mno-embedded-data | |
10425 | @opindex membedded-data | |
10426 | @opindex mno-embedded-data | |
10427 | Allocate variables to the read-only data section first if possible, then | |
10428 | next in the small data section if possible, otherwise in data. This gives | |
10429 | slightly slower code than the default, but reduces the amount of RAM required | |
10430 | when executing, and thus may be preferred for some embedded systems. | |
58605ba0 | 10431 | |
39bc1876 NS |
10432 | @item -muninit-const-in-rodata |
10433 | @itemx -mno-uninit-const-in-rodata | |
10434 | @opindex muninit-const-in-rodata | |
10435 | @opindex mno-uninit-const-in-rodata | |
10436 | Put uninitialized @code{const} variables in the read-only data section. | |
10437 | This option is only meaningful in conjunction with @option{-membedded-data}. | |
4f69985c | 10438 | |
39bc1876 NS |
10439 | @item -msplit-addresses |
10440 | @itemx -mno-split-addresses | |
10441 | @opindex msplit-addresses | |
10442 | @opindex mno-split-addresses | |
10443 | Enable (disable) use of the @code{%hi()} and @code{%lo()} assembler | |
38b974a6 | 10444 | relocation operators. This option has been superseded by |
39bc1876 | 10445 | @option{-mexplicit-relocs} but is retained for backwards compatibility. |
58605ba0 | 10446 | |
39bc1876 NS |
10447 | @item -mexplicit-relocs |
10448 | @itemx -mno-explicit-relocs | |
10449 | @opindex mexplicit-relocs | |
10450 | @opindex mno-explicit-relocs | |
10451 | Use (do not use) assembler relocation operators when dealing with symbolic | |
10452 | addresses. The alternative, selected by @option{-mno-explicit-relocs}, | |
10453 | is to use assembler macros instead. | |
4f69985c | 10454 | |
12e4afe4 RS |
10455 | @option{-mexplicit-relocs} is the default if GCC was configured |
10456 | to use an assembler that supports relocation operators. | |
4f69985c | 10457 | |
39bc1876 NS |
10458 | @item -mcheck-zero-division |
10459 | @itemx -mno-check-zero-division | |
10460 | @opindex mcheck-zero-division | |
10461 | @opindex mno-check-zero-division | |
10462 | Trap (do not trap) on integer division by zero. The default is | |
10463 | @option{-mcheck-zero-division}. | |
4f69985c | 10464 | |
9f0df97a DD |
10465 | @item -mdivide-traps |
10466 | @itemx -mdivide-breaks | |
10467 | @opindex mdivide-traps | |
10468 | @opindex mdivide-breaks | |
10469 | MIPS systems check for division by zero by generating either a | |
10470 | conditional trap or a break instruction. Using traps results in | |
10471 | smaller code, but is only supported on MIPS II and later. Also, some | |
10472 | versions of the Linux kernel have a bug that prevents trap from | |
8a36672b | 10473 | generating the proper signal (@code{SIGFPE}). Use @option{-mdivide-traps} to |
9f0df97a DD |
10474 | allow conditional traps on architectures that support them and |
10475 | @option{-mdivide-breaks} to force the use of breaks. | |
10476 | ||
10477 | The default is usually @option{-mdivide-traps}, but this can be | |
10478 | overridden at configure time using @option{--with-divide=breaks}. | |
10479 | Divide-by-zero checks can be completely disabled using | |
10480 | @option{-mno-check-zero-division}. | |
10481 | ||
39bc1876 NS |
10482 | @item -mmemcpy |
10483 | @itemx -mno-memcpy | |
10484 | @opindex mmemcpy | |
10485 | @opindex mno-memcpy | |
10486 | Force (do not force) the use of @code{memcpy()} for non-trivial block | |
10487 | moves. The default is @option{-mno-memcpy}, which allows GCC to inline | |
10488 | most constant-sized copies. | |
74291a4b | 10489 | |
39bc1876 NS |
10490 | @item -mlong-calls |
10491 | @itemx -mno-long-calls | |
10492 | @opindex mlong-calls | |
10493 | @opindex mno-long-calls | |
10494 | Disable (do not disable) use of the @code{jal} instruction. Calling | |
10495 | functions using @code{jal} is more efficient but requires the caller | |
10496 | and callee to be in the same 256 megabyte segment. | |
d7c23cdc | 10497 | |
39bc1876 NS |
10498 | This option has no effect on abicalls code. The default is |
10499 | @option{-mno-long-calls}. | |
d7c23cdc | 10500 | |
39bc1876 NS |
10501 | @item -mmad |
10502 | @itemx -mno-mad | |
10503 | @opindex mmad | |
10504 | @opindex mno-mad | |
10505 | Enable (disable) use of the @code{mad}, @code{madu} and @code{mul} | |
8a36672b | 10506 | instructions, as provided by the R4650 ISA@. |
d7c23cdc | 10507 | |
39bc1876 NS |
10508 | @item -mfused-madd |
10509 | @itemx -mno-fused-madd | |
10510 | @opindex mfused-madd | |
10511 | @opindex mno-fused-madd | |
10512 | Enable (disable) use of the floating point multiply-accumulate | |
10513 | instructions, when they are available. The default is | |
10514 | @option{-mfused-madd}. | |
74291a4b | 10515 | |
39bc1876 NS |
10516 | When multiply-accumulate instructions are used, the intermediate |
10517 | product is calculated to infinite precision and is not subject to | |
10518 | the FCSR Flush to Zero bit. This may be undesirable in some | |
10519 | circumstances. | |
74291a4b | 10520 | |
39bc1876 NS |
10521 | @item -nocpp |
10522 | @opindex nocpp | |
10523 | Tell the MIPS assembler to not run its preprocessor over user | |
10524 | assembler files (with a @samp{.s} suffix) when assembling them. | |
74291a4b | 10525 | |
39bc1876 NS |
10526 | @item -mfix-r4000 |
10527 | @itemx -mno-fix-r4000 | |
10528 | @opindex mfix-r4000 | |
10529 | @opindex mno-fix-r4000 | |
10530 | Work around certain R4000 CPU errata: | |
10531 | @itemize @minus | |
10532 | @item | |
10533 | A double-word or a variable shift may give an incorrect result if executed | |
10534 | immediately after starting an integer division. | |
10535 | @item | |
10536 | A double-word or a variable shift may give an incorrect result if executed | |
10537 | while an integer multiplication is in progress. | |
10538 | @item | |
10539 | An integer division may give an incorrect result if started in a delay slot | |
10540 | of a taken branch or a jump. | |
10541 | @end itemize | |
74291a4b | 10542 | |
39bc1876 NS |
10543 | @item -mfix-r4400 |
10544 | @itemx -mno-fix-r4400 | |
10545 | @opindex mfix-r4400 | |
10546 | @opindex mno-fix-r4400 | |
10547 | Work around certain R4400 CPU errata: | |
10548 | @itemize @minus | |
10549 | @item | |
10550 | A double-word or a variable shift may give an incorrect result if executed | |
10551 | immediately after starting an integer division. | |
10552 | @end itemize | |
dcb9d1f0 | 10553 | |
39bc1876 NS |
10554 | @item -mfix-vr4120 |
10555 | @itemx -mno-fix-vr4120 | |
10556 | @opindex mfix-vr4120 | |
10557 | Work around certain VR4120 errata: | |
10558 | @itemize @minus | |
10559 | @item | |
10560 | @code{dmultu} does not always produce the correct result. | |
10561 | @item | |
10562 | @code{div} and @code{ddiv} do not always produce the correct result if one | |
10563 | of the operands is negative. | |
10564 | @end itemize | |
10565 | The workarounds for the division errata rely on special functions in | |
10566 | @file{libgcc.a}. At present, these functions are only provided by | |
10567 | the @code{mips64vr*-elf} configurations. | |
39ba95b5 | 10568 | |
39bc1876 NS |
10569 | Other VR4120 errata require a nop to be inserted between certain pairs of |
10570 | instructions. These errata are handled by the assembler, not by GCC itself. | |
17f0f8fa | 10571 | |
0ac40e7a RS |
10572 | @item -mfix-vr4130 |
10573 | @opindex mfix-vr4130 | |
10574 | Work around the VR4130 @code{mflo}/@code{mfhi} errata. The | |
10575 | workarounds are implemented by the assembler rather than by GCC, | |
10576 | although GCC will avoid using @code{mflo} and @code{mfhi} if the | |
10577 | VR4130 @code{macc}, @code{macchi}, @code{dmacc} and @code{dmacchi} | |
10578 | instructions are available instead. | |
10579 | ||
39bc1876 NS |
10580 | @item -mfix-sb1 |
10581 | @itemx -mno-fix-sb1 | |
10582 | @opindex mfix-sb1 | |
10583 | Work around certain SB-1 CPU core errata. | |
10584 | (This flag currently works around the SB-1 revision 2 | |
10585 | ``F1'' and ``F2'' floating point errata.) | |
74291a4b | 10586 | |
39bc1876 NS |
10587 | @item -mflush-func=@var{func} |
10588 | @itemx -mno-flush-func | |
10589 | @opindex mflush-func | |
10590 | Specifies the function to call to flush the I and D caches, or to not | |
10591 | call any such function. If called, the function must take the same | |
10592 | arguments as the common @code{_flush_func()}, that is, the address of the | |
10593 | memory range for which the cache is being flushed, the size of the | |
10594 | memory range, and the number 3 (to flush both caches). The default | |
10595 | depends on the target GCC was configured for, but commonly is either | |
10596 | @samp{_flush_func} or @samp{__cpu_flush}. | |
74291a4b | 10597 | |
39bc1876 NS |
10598 | @item -mbranch-likely |
10599 | @itemx -mno-branch-likely | |
10600 | @opindex mbranch-likely | |
10601 | @opindex mno-branch-likely | |
10602 | Enable or disable use of Branch Likely instructions, regardless of the | |
10603 | default for the selected architecture. By default, Branch Likely | |
10604 | instructions may be generated if they are supported by the selected | |
10605 | architecture. An exception is for the MIPS32 and MIPS64 architectures | |
10606 | and processors which implement those architectures; for those, Branch | |
10607 | Likely instructions will not be generated by default because the MIPS32 | |
10608 | and MIPS64 architectures specifically deprecate their use. | |
74291a4b | 10609 | |
39bc1876 NS |
10610 | @item -mfp-exceptions |
10611 | @itemx -mno-fp-exceptions | |
10612 | @opindex mfp-exceptions | |
10613 | Specifies whether FP exceptions are enabled. This affects how we schedule | |
10614 | FP instructions for some processors. The default is that FP exceptions are | |
10615 | enabled. | |
74291a4b | 10616 | |
39bc1876 NS |
10617 | For instance, on the SB-1, if FP exceptions are disabled, and we are emitting |
10618 | 64-bit code, then we can use both FP pipes. Otherwise, we can only use one | |
10619 | FP pipe. | |
74291a4b | 10620 | |
39bc1876 NS |
10621 | @item -mvr4130-align |
10622 | @itemx -mno-vr4130-align | |
10623 | @opindex mvr4130-align | |
10624 | The VR4130 pipeline is two-way superscalar, but can only issue two | |
10625 | instructions together if the first one is 8-byte aligned. When this | |
10626 | option is enabled, GCC will align pairs of instructions that it | |
10627 | thinks should execute in parallel. | |
74291a4b | 10628 | |
39bc1876 NS |
10629 | This option only has an effect when optimizing for the VR4130. |
10630 | It normally makes code faster, but at the expense of making it bigger. | |
10631 | It is enabled by default at optimization level @option{-O3}. | |
10632 | @end table | |
3a8699c7 | 10633 | |
39bc1876 NS |
10634 | @node MMIX Options |
10635 | @subsection MMIX Options | |
10636 | @cindex MMIX Options | |
74291a4b | 10637 | |
39bc1876 | 10638 | These options are defined for the MMIX: |
74291a4b | 10639 | |
39bc1876 NS |
10640 | @table @gcctabopt |
10641 | @item -mlibfuncs | |
10642 | @itemx -mno-libfuncs | |
10643 | @opindex mlibfuncs | |
10644 | @opindex mno-libfuncs | |
10645 | Specify that intrinsic library functions are being compiled, passing all | |
10646 | values in registers, no matter the size. | |
3cadd778 | 10647 | |
39bc1876 NS |
10648 | @item -mepsilon |
10649 | @itemx -mno-epsilon | |
10650 | @opindex mepsilon | |
10651 | @opindex mno-epsilon | |
10652 | Generate floating-point comparison instructions that compare with respect | |
10653 | to the @code{rE} epsilon register. | |
3cadd778 | 10654 | |
39bc1876 NS |
10655 | @item -mabi=mmixware |
10656 | @itemx -mabi=gnu | |
10657 | @opindex mabi-mmixware | |
10658 | @opindex mabi=gnu | |
10659 | Generate code that passes function parameters and return values that (in | |
10660 | the called function) are seen as registers @code{$0} and up, as opposed to | |
10661 | the GNU ABI which uses global registers @code{$231} and up. | |
3cadd778 | 10662 | |
39bc1876 NS |
10663 | @item -mzero-extend |
10664 | @itemx -mno-zero-extend | |
10665 | @opindex mzero-extend | |
10666 | @opindex mno-zero-extend | |
10667 | When reading data from memory in sizes shorter than 64 bits, use (do not | |
10668 | use) zero-extending load instructions by default, rather than | |
10669 | sign-extending ones. | |
3cadd778 | 10670 | |
39bc1876 NS |
10671 | @item -mknuthdiv |
10672 | @itemx -mno-knuthdiv | |
10673 | @opindex mknuthdiv | |
10674 | @opindex mno-knuthdiv | |
10675 | Make the result of a division yielding a remainder have the same sign as | |
10676 | the divisor. With the default, @option{-mno-knuthdiv}, the sign of the | |
10677 | remainder follows the sign of the dividend. Both methods are | |
10678 | arithmetically valid, the latter being almost exclusively used. | |
74291a4b | 10679 | |
39bc1876 NS |
10680 | @item -mtoplevel-symbols |
10681 | @itemx -mno-toplevel-symbols | |
10682 | @opindex mtoplevel-symbols | |
10683 | @opindex mno-toplevel-symbols | |
10684 | Prepend (do not prepend) a @samp{:} to all global symbols, so the assembly | |
10685 | code can be used with the @code{PREFIX} assembly directive. | |
74291a4b | 10686 | |
39bc1876 NS |
10687 | @item -melf |
10688 | @opindex melf | |
10689 | Generate an executable in the ELF format, rather than the default | |
10690 | @samp{mmo} format used by the @command{mmix} simulator. | |
3d5a0820 | 10691 | |
39bc1876 NS |
10692 | @item -mbranch-predict |
10693 | @itemx -mno-branch-predict | |
10694 | @opindex mbranch-predict | |
10695 | @opindex mno-branch-predict | |
10696 | Use (do not use) the probable-branch instructions, when static branch | |
10697 | prediction indicates a probable branch. | |
3cadd778 | 10698 | |
39bc1876 NS |
10699 | @item -mbase-addresses |
10700 | @itemx -mno-base-addresses | |
10701 | @opindex mbase-addresses | |
10702 | @opindex mno-base-addresses | |
10703 | Generate (do not generate) code that uses @emph{base addresses}. Using a | |
10704 | base address automatically generates a request (handled by the assembler | |
10705 | and the linker) for a constant to be set up in a global register. The | |
10706 | register is used for one or more base address requests within the range 0 | |
10707 | to 255 from the value held in the register. The generally leads to short | |
10708 | and fast code, but the number of different data items that can be | |
10709 | addressed is limited. This means that a program that uses lots of static | |
10710 | data may require @option{-mno-base-addresses}. | |
3cadd778 | 10711 | |
39bc1876 NS |
10712 | @item -msingle-exit |
10713 | @itemx -mno-single-exit | |
10714 | @opindex msingle-exit | |
10715 | @opindex mno-single-exit | |
10716 | Force (do not force) generated code to have a single exit point in each | |
10717 | function. | |
10718 | @end table | |
3cadd778 | 10719 | |
39bc1876 NS |
10720 | @node MN10300 Options |
10721 | @subsection MN10300 Options | |
10722 | @cindex MN10300 options | |
3cadd778 | 10723 | |
39bc1876 | 10724 | These @option{-m} options are defined for Matsushita MN10300 architectures: |
3cadd778 | 10725 | |
39bc1876 NS |
10726 | @table @gcctabopt |
10727 | @item -mmult-bug | |
10728 | @opindex mmult-bug | |
10729 | Generate code to avoid bugs in the multiply instructions for the MN10300 | |
10730 | processors. This is the default. | |
c474f76b | 10731 | |
39bc1876 NS |
10732 | @item -mno-mult-bug |
10733 | @opindex mno-mult-bug | |
10734 | Do not generate code to avoid bugs in the multiply instructions for the | |
10735 | MN10300 processors. | |
3cadd778 | 10736 | |
39bc1876 NS |
10737 | @item -mam33 |
10738 | @opindex mam33 | |
10739 | Generate code which uses features specific to the AM33 processor. | |
3cadd778 | 10740 | |
39bc1876 NS |
10741 | @item -mno-am33 |
10742 | @opindex mno-am33 | |
10743 | Do not generate code which uses features specific to the AM33 processor. This | |
10744 | is the default. | |
1a66cd67 | 10745 | |
b1eb8119 DD |
10746 | @item -mreturn-pointer-on-d0 |
10747 | @opindex mreturn-pointer-on-d0 | |
10748 | When generating a function which returns a pointer, return the pointer | |
10749 | in both @code{a0} and @code{d0}. Otherwise, the pointer is returned | |
10750 | only in a0, and attempts to call such functions without a prototype | |
10751 | would result in errors. Note that this option is on by default; use | |
10752 | @option{-mno-return-pointer-on-d0} to disable it. | |
10753 | ||
39bc1876 NS |
10754 | @item -mno-crt0 |
10755 | @opindex mno-crt0 | |
10756 | Do not link in the C run-time initialization object file. | |
93ca1662 | 10757 | |
39bc1876 NS |
10758 | @item -mrelax |
10759 | @opindex mrelax | |
10760 | Indicate to the linker that it should perform a relaxation optimization pass | |
10761 | to shorten branches, calls and absolute memory addresses. This option only | |
10762 | has an effect when used on the command line for the final link step. | |
10763 | ||
10764 | This option makes symbolic debugging impossible. | |
74291a4b MM |
10765 | @end table |
10766 | ||
6c9ac67a NS |
10767 | @node MT Options |
10768 | @subsection MT Options | |
10769 | @cindex MT options | |
c85ba4fb | 10770 | |
6c9ac67a | 10771 | These @option{-m} options are defined for Morpho MT architectures: |
c85ba4fb AH |
10772 | |
10773 | @table @gcctabopt | |
10774 | ||
10775 | @item -march=@var{cpu-type} | |
10776 | @opindex march | |
10777 | Generate code that will run on @var{cpu-type}, which is the name of a system | |
10778 | representing a certain processor type. Possible values for | |
6c9ac67a NS |
10779 | @var{cpu-type} are @samp{ms1-64-001}, @samp{ms1-16-002}, |
10780 | @samp{ms1-16-003} and @samp{ms2}. | |
c85ba4fb | 10781 | |
6c9ac67a | 10782 | When this option is not used, the default is @option{-march=ms1-16-002}. |
c85ba4fb AH |
10783 | |
10784 | @item -mbacc | |
10785 | @opindex mbacc | |
10786 | Use byte loads and stores when generating code. | |
10787 | ||
10788 | @item -mno-bacc | |
10789 | @opindex mno-bacc | |
10790 | Do not use byte loads and stores when generating code. | |
10791 | ||
10792 | @item -msim | |
10793 | @opindex msim | |
10794 | Use simulator runtime | |
10795 | ||
10796 | @item -mno-crt0 | |
10797 | @opindex mno-crt0 | |
10798 | Do not link in the C run-time initialization object file | |
57b51d4d | 10799 | @file{crti.o}. Other run-time initialization and termination files |
c85ba4fb AH |
10800 | such as @file{startup.o} and @file{exit.o} are still included on the |
10801 | linker command line. | |
10802 | ||
10803 | @end table | |
10804 | ||
39bc1876 NS |
10805 | @node PDP-11 Options |
10806 | @subsection PDP-11 Options | |
10807 | @cindex PDP-11 Options | |
f84271d9 | 10808 | |
39bc1876 | 10809 | These options are defined for the PDP-11: |
f84271d9 | 10810 | |
2642624b | 10811 | @table @gcctabopt |
39bc1876 NS |
10812 | @item -mfpu |
10813 | @opindex mfpu | |
10814 | Use hardware FPP floating point. This is the default. (FIS floating | |
10815 | point on the PDP-11/40 is not supported.) | |
f84271d9 | 10816 | |
39bc1876 NS |
10817 | @item -msoft-float |
10818 | @opindex msoft-float | |
10819 | Do not use hardware floating point. | |
f84271d9 | 10820 | |
39bc1876 NS |
10821 | @item -mac0 |
10822 | @opindex mac0 | |
10823 | Return floating-point results in ac0 (fr0 in Unix assembler syntax). | |
f84271d9 | 10824 | |
39bc1876 NS |
10825 | @item -mno-ac0 |
10826 | @opindex mno-ac0 | |
10827 | Return floating-point results in memory. This is the default. | |
10828 | ||
10829 | @item -m40 | |
10830 | @opindex m40 | |
10831 | Generate code for a PDP-11/40. | |
10832 | ||
10833 | @item -m45 | |
10834 | @opindex m45 | |
10835 | Generate code for a PDP-11/45. This is the default. | |
f84271d9 | 10836 | |
39bc1876 NS |
10837 | @item -m10 |
10838 | @opindex m10 | |
10839 | Generate code for a PDP-11/10. | |
f84271d9 | 10840 | |
39bc1876 NS |
10841 | @item -mbcopy-builtin |
10842 | @opindex bcopy-builtin | |
70128ad9 | 10843 | Use inline @code{movmemhi} patterns for copying memory. This is the |
39bc1876 | 10844 | default. |
f84271d9 | 10845 | |
39bc1876 NS |
10846 | @item -mbcopy |
10847 | @opindex mbcopy | |
70128ad9 | 10848 | Do not use inline @code{movmemhi} patterns for copying memory. |
02f52e19 | 10849 | |
39bc1876 NS |
10850 | @item -mint16 |
10851 | @itemx -mno-int32 | |
10852 | @opindex mint16 | |
10853 | @opindex mno-int32 | |
10854 | Use 16-bit @code{int}. This is the default. | |
48f0be1b | 10855 | |
39bc1876 NS |
10856 | @item -mint32 |
10857 | @itemx -mno-int16 | |
10858 | @opindex mint32 | |
10859 | @opindex mno-int16 | |
10860 | Use 32-bit @code{int}. | |
b4378319 | 10861 | |
39bc1876 NS |
10862 | @item -mfloat64 |
10863 | @itemx -mno-float32 | |
10864 | @opindex mfloat64 | |
10865 | @opindex mno-float32 | |
10866 | Use 64-bit @code{float}. This is the default. | |
b4378319 | 10867 | |
39bc1876 NS |
10868 | @item -mfloat32 |
10869 | @itemx -mno-float64 | |
10870 | @opindex mfloat32 | |
10871 | @opindex mno-float64 | |
10872 | Use 32-bit @code{float}. | |
daf2f129 | 10873 | |
39bc1876 NS |
10874 | @item -mabshi |
10875 | @opindex mabshi | |
10876 | Use @code{abshi2} pattern. This is the default. | |
232830b7 | 10877 | |
39bc1876 NS |
10878 | @item -mno-abshi |
10879 | @opindex mno-abshi | |
10880 | Do not use @code{abshi2} pattern. | |
b4378319 | 10881 | |
39bc1876 NS |
10882 | @item -mbranch-expensive |
10883 | @opindex mbranch-expensive | |
10884 | Pretend that branches are expensive. This is for experimenting with | |
10885 | code generation only. | |
b4378319 | 10886 | |
39bc1876 NS |
10887 | @item -mbranch-cheap |
10888 | @opindex mbranch-cheap | |
10889 | Do not pretend that branches are expensive. This is the default. | |
b4378319 | 10890 | |
39bc1876 NS |
10891 | @item -msplit |
10892 | @opindex msplit | |
8a36672b | 10893 | Generate code for a system with split I&D@. |
b4378319 | 10894 | |
39bc1876 NS |
10895 | @item -mno-split |
10896 | @opindex mno-split | |
8a36672b | 10897 | Generate code for a system without split I&D@. This is the default. |
f84271d9 | 10898 | |
39bc1876 NS |
10899 | @item -munix-asm |
10900 | @opindex munix-asm | |
10901 | Use Unix assembler syntax. This is the default when configured for | |
10902 | @samp{pdp11-*-bsd}. | |
56b2d7a7 | 10903 | |
39bc1876 NS |
10904 | @item -mdec-asm |
10905 | @opindex mdec-asm | |
10906 | Use DEC assembler syntax. This is the default when configured for any | |
10907 | PDP-11 target other than @samp{pdp11-*-bsd}. | |
10908 | @end table | |
56b2d7a7 | 10909 | |
39bc1876 NS |
10910 | @node PowerPC Options |
10911 | @subsection PowerPC Options | |
10912 | @cindex PowerPC options | |
56b2d7a7 | 10913 | |
39bc1876 | 10914 | These are listed under @xref{RS/6000 and PowerPC Options}. |
56b2d7a7 | 10915 | |
39bc1876 NS |
10916 | @node RS/6000 and PowerPC Options |
10917 | @subsection IBM RS/6000 and PowerPC Options | |
10918 | @cindex RS/6000 and PowerPC Options | |
10919 | @cindex IBM RS/6000 and PowerPC Options | |
56b2d7a7 | 10920 | |
39bc1876 NS |
10921 | These @samp{-m} options are defined for the IBM RS/6000 and PowerPC: |
10922 | @table @gcctabopt | |
10923 | @item -mpower | |
10924 | @itemx -mno-power | |
10925 | @itemx -mpower2 | |
10926 | @itemx -mno-power2 | |
10927 | @itemx -mpowerpc | |
10928 | @itemx -mno-powerpc | |
10929 | @itemx -mpowerpc-gpopt | |
10930 | @itemx -mno-powerpc-gpopt | |
10931 | @itemx -mpowerpc-gfxopt | |
10932 | @itemx -mno-powerpc-gfxopt | |
10933 | @itemx -mpowerpc64 | |
10934 | @itemx -mno-powerpc64 | |
9719f3b7 DE |
10935 | @itemx -mmfcrf |
10936 | @itemx -mno-mfcrf | |
10937 | @itemx -mpopcntb | |
10938 | @itemx -mno-popcntb | |
10939 | @itemx -mfprnd | |
10940 | @itemx -mno-fprnd | |
39bc1876 NS |
10941 | @opindex mpower |
10942 | @opindex mno-power | |
10943 | @opindex mpower2 | |
10944 | @opindex mno-power2 | |
10945 | @opindex mpowerpc | |
10946 | @opindex mno-powerpc | |
10947 | @opindex mpowerpc-gpopt | |
10948 | @opindex mno-powerpc-gpopt | |
10949 | @opindex mpowerpc-gfxopt | |
10950 | @opindex mno-powerpc-gfxopt | |
10951 | @opindex mpowerpc64 | |
10952 | @opindex mno-powerpc64 | |
9719f3b7 DE |
10953 | @opindex mmfcrf |
10954 | @opindex mno-mfcrf | |
10955 | @opindex mpopcntb | |
10956 | @opindex mno-popcntb | |
10957 | @opindex mfprnd | |
10958 | @opindex mno-fprnd | |
39bc1876 NS |
10959 | GCC supports two related instruction set architectures for the |
10960 | RS/6000 and PowerPC@. The @dfn{POWER} instruction set are those | |
10961 | instructions supported by the @samp{rios} chip set used in the original | |
10962 | RS/6000 systems and the @dfn{PowerPC} instruction set is the | |
9719f3b7 DE |
10963 | architecture of the Freescale MPC5xx, MPC6xx, MPC8xx microprocessors, and |
10964 | the IBM 4xx, 6xx, and follow-on microprocessors. | |
56b2d7a7 | 10965 | |
39bc1876 NS |
10966 | Neither architecture is a subset of the other. However there is a |
10967 | large common subset of instructions supported by both. An MQ | |
10968 | register is included in processors supporting the POWER architecture. | |
56b2d7a7 | 10969 | |
39bc1876 NS |
10970 | You use these options to specify which instructions are available on the |
10971 | processor you are using. The default value of these options is | |
10972 | determined when configuring GCC@. Specifying the | |
10973 | @option{-mcpu=@var{cpu_type}} overrides the specification of these | |
10974 | options. We recommend you use the @option{-mcpu=@var{cpu_type}} option | |
10975 | rather than the options listed above. | |
56b2d7a7 | 10976 | |
39bc1876 NS |
10977 | The @option{-mpower} option allows GCC to generate instructions that |
10978 | are found only in the POWER architecture and to use the MQ register. | |
10979 | Specifying @option{-mpower2} implies @option{-power} and also allows GCC | |
10980 | to generate instructions that are present in the POWER2 architecture but | |
10981 | not the original POWER architecture. | |
83575957 | 10982 | |
39bc1876 NS |
10983 | The @option{-mpowerpc} option allows GCC to generate instructions that |
10984 | are found only in the 32-bit subset of the PowerPC architecture. | |
10985 | Specifying @option{-mpowerpc-gpopt} implies @option{-mpowerpc} and also allows | |
10986 | GCC to use the optional PowerPC architecture instructions in the | |
10987 | General Purpose group, including floating-point square root. Specifying | |
10988 | @option{-mpowerpc-gfxopt} implies @option{-mpowerpc} and also allows GCC to | |
10989 | use the optional PowerPC architecture instructions in the Graphics | |
10990 | group, including floating-point select. | |
83575957 | 10991 | |
9719f3b7 DE |
10992 | The @option{-mmfcrf} option allows GCC to generate the move from |
10993 | condition register field instruction implemented on the POWER4 | |
10994 | processor and other processors that support the PowerPC V2.01 | |
10995 | architecture. | |
10996 | The @option{-mpopcntb} option allows GCC to generate the popcount and | |
10997 | double precision FP reciprocal estimate instruction implemented on the | |
10998 | POWER5 processor and other processors that support the PowerPC V2.02 | |
10999 | architecture. | |
11000 | The @option{-mfprnd} option allows GCC to generate the FP round to | |
11001 | integer instructions implemented on the POWER5+ processor and other | |
11002 | processors that support the PowerPC V2.03 architecture. | |
11003 | ||
39bc1876 NS |
11004 | The @option{-mpowerpc64} option allows GCC to generate the additional |
11005 | 64-bit instructions that are found in the full PowerPC64 architecture | |
11006 | and to treat GPRs as 64-bit, doubleword quantities. GCC defaults to | |
11007 | @option{-mno-powerpc64}. | |
83575957 | 11008 | |
39bc1876 NS |
11009 | If you specify both @option{-mno-power} and @option{-mno-powerpc}, GCC |
11010 | will use only the instructions in the common subset of both | |
11011 | architectures plus some special AIX common-mode calls, and will not use | |
11012 | the MQ register. Specifying both @option{-mpower} and @option{-mpowerpc} | |
11013 | permits GCC to use any instruction from either architecture and to | |
11014 | allow use of the MQ register; specify this for the Motorola MPC601. | |
83575957 | 11015 | |
39bc1876 NS |
11016 | @item -mnew-mnemonics |
11017 | @itemx -mold-mnemonics | |
11018 | @opindex mnew-mnemonics | |
11019 | @opindex mold-mnemonics | |
11020 | Select which mnemonics to use in the generated assembler code. With | |
11021 | @option{-mnew-mnemonics}, GCC uses the assembler mnemonics defined for | |
11022 | the PowerPC architecture. With @option{-mold-mnemonics} it uses the | |
11023 | assembler mnemonics defined for the POWER architecture. Instructions | |
11024 | defined in only one architecture have only one mnemonic; GCC uses that | |
11025 | mnemonic irrespective of which of these options is specified. | |
83575957 | 11026 | |
39bc1876 NS |
11027 | GCC defaults to the mnemonics appropriate for the architecture in |
11028 | use. Specifying @option{-mcpu=@var{cpu_type}} sometimes overrides the | |
11029 | value of these option. Unless you are building a cross-compiler, you | |
11030 | should normally not specify either @option{-mnew-mnemonics} or | |
11031 | @option{-mold-mnemonics}, but should instead accept the default. | |
83575957 | 11032 | |
39bc1876 NS |
11033 | @item -mcpu=@var{cpu_type} |
11034 | @opindex mcpu | |
11035 | Set architecture type, register usage, choice of mnemonics, and | |
11036 | instruction scheduling parameters for machine type @var{cpu_type}. | |
11037 | Supported values for @var{cpu_type} are @samp{401}, @samp{403}, | |
11038 | @samp{405}, @samp{405fp}, @samp{440}, @samp{440fp}, @samp{505}, | |
11039 | @samp{601}, @samp{602}, @samp{603}, @samp{603e}, @samp{604}, | |
11040 | @samp{604e}, @samp{620}, @samp{630}, @samp{740}, @samp{7400}, | |
11041 | @samp{7450}, @samp{750}, @samp{801}, @samp{821}, @samp{823}, | |
9719f3b7 | 11042 | @samp{860}, @samp{970}, @samp{8540}, @samp{ec603e}, @samp{G3}, |
39bc1876 | 11043 | @samp{G4}, @samp{G5}, @samp{power}, @samp{power2}, @samp{power3}, |
9719f3b7 DE |
11044 | @samp{power4}, @samp{power5}, @samp{power5+}, |
11045 | @samp{common}, @samp{powerpc}, @samp{powerpc64}, | |
98c41d98 | 11046 | @samp{rios}, @samp{rios1}, @samp{rios2}, @samp{rsc}, and @samp{rs64}. |
83575957 | 11047 | |
39bc1876 NS |
11048 | @option{-mcpu=common} selects a completely generic processor. Code |
11049 | generated under this option will run on any POWER or PowerPC processor. | |
11050 | GCC will use only the instructions in the common subset of both | |
11051 | architectures, and will not use the MQ register. GCC assumes a generic | |
11052 | processor model for scheduling purposes. | |
83575957 | 11053 | |
39bc1876 NS |
11054 | @option{-mcpu=power}, @option{-mcpu=power2}, @option{-mcpu=powerpc}, and |
11055 | @option{-mcpu=powerpc64} specify generic POWER, POWER2, pure 32-bit | |
11056 | PowerPC (i.e., not MPC601), and 64-bit PowerPC architecture machine | |
11057 | types, with an appropriate, generic processor model assumed for | |
11058 | scheduling purposes. | |
83575957 | 11059 | |
39bc1876 NS |
11060 | The other options specify a specific processor. Code generated under |
11061 | those options will run best on that processor, and may not run at all on | |
11062 | others. | |
83575957 | 11063 | |
39bc1876 | 11064 | The @option{-mcpu} options automatically enable or disable the |
9719f3b7 DE |
11065 | following options: @option{-maltivec}, @option{-mfprnd}, |
11066 | @option{-mhard-float}, @option{-mmfcrf}, @option{-mmultiple}, | |
11067 | @option{-mnew-mnemonics}, @option{-mpopcntb}, @option{-mpower}, | |
11068 | @option{-mpower2}, @option{-mpowerpc64}, @option{-mpowerpc-gpopt}, | |
131aeb82 JM |
11069 | @option{-mpowerpc-gfxopt}, @option{-mstring}, @option{-mmulhw}. |
11070 | The particular options | |
9719f3b7 DE |
11071 | set for any particular CPU will vary between compiler versions, |
11072 | depending on what setting seems to produce optimal code for that CPU; | |
11073 | it doesn't necessarily reflect the actual hardware's capabilities. If | |
11074 | you wish to set an individual option to a particular value, you may | |
11075 | specify it after the @option{-mcpu} option, like @samp{-mcpu=970 | |
11076 | -mno-altivec}. | |
5d7c2819 | 11077 | |
39bc1876 | 11078 | On AIX, the @option{-maltivec} and @option{-mpowerpc64} options are |
9719f3b7 | 11079 | not enabled or disabled by the @option{-mcpu} option at present because |
39bc1876 NS |
11080 | AIX does not have full support for these options. You may still |
11081 | enable or disable them individually if you're sure it'll work in your | |
11082 | environment. | |
83575957 | 11083 | |
39bc1876 NS |
11084 | @item -mtune=@var{cpu_type} |
11085 | @opindex mtune | |
11086 | Set the instruction scheduling parameters for machine type | |
11087 | @var{cpu_type}, but do not set the architecture type, register usage, or | |
11088 | choice of mnemonics, as @option{-mcpu=@var{cpu_type}} would. The same | |
11089 | values for @var{cpu_type} are used for @option{-mtune} as for | |
11090 | @option{-mcpu}. If both are specified, the code generated will use the | |
11091 | architecture, registers, and mnemonics set by @option{-mcpu}, but the | |
11092 | scheduling parameters set by @option{-mtune}. | |
83575957 | 11093 | |
ef765ea9 DE |
11094 | @item -mswdiv |
11095 | @itemx -mno-swdiv | |
11096 | @opindex mswdiv | |
11097 | @opindex mno-swdiv | |
11098 | Generate code to compute division as reciprocal estimate and iterative | |
11099 | refinement, creating opportunities for increased throughput. This | |
11100 | feature requires: optional PowerPC Graphics instruction set for single | |
11101 | precision and FRE instruction for double precision, assuming divides | |
11102 | cannot generate user-visible traps, and the domain values not include | |
11103 | Infinities, denormals or zero denominator. | |
11104 | ||
39bc1876 NS |
11105 | @item -maltivec |
11106 | @itemx -mno-altivec | |
11107 | @opindex maltivec | |
11108 | @opindex mno-altivec | |
7088c6e4 GK |
11109 | Generate code that uses (does not use) AltiVec instructions, and also |
11110 | enable the use of built-in functions that allow more direct access to | |
11111 | the AltiVec instruction set. You may also need to set | |
39bc1876 NS |
11112 | @option{-mabi=altivec} to adjust the current ABI with AltiVec ABI |
11113 | enhancements. | |
83575957 | 11114 | |
78f5898b AH |
11115 | @item -mvrsave |
11116 | @item -mno-vrsave | |
11117 | @opindex mvrsave | |
11118 | @opindex mno-vrsave | |
11119 | Generate VRSAVE instructions when generating AltiVec code. | |
11120 | ||
7f970b70 AM |
11121 | @item -msecure-plt |
11122 | @opindex msecure-plt | |
11123 | Generate code that allows ld and ld.so to build executables and shared | |
11124 | libraries with non-exec .plt and .got sections. This is a PowerPC | |
11125 | 32-bit SYSV ABI option. | |
11126 | ||
11127 | @item -mbss-plt | |
11128 | @opindex mbss-plt | |
11129 | Generate code that uses a BSS .plt section that ld.so fills in, and | |
11130 | requires .plt and .got sections that are both writable and executable. | |
11131 | This is a PowerPC 32-bit SYSV ABI option. | |
11132 | ||
78f5898b AH |
11133 | @item -misel |
11134 | @itemx -mno-isel | |
39bc1876 | 11135 | @opindex misel |
78f5898b | 11136 | @opindex mno-isel |
39bc1876 | 11137 | This switch enables or disables the generation of ISEL instructions. |
83575957 | 11138 | |
78f5898b AH |
11139 | @item -misel=@var{yes/no} |
11140 | This switch has been deprecated. Use @option{-misel} and | |
11141 | @option{-mno-isel} instead. | |
11142 | ||
11143 | @item -mspe | |
11144 | @itemx -mno-isel | |
39bc1876 | 11145 | @opindex mspe |
78f5898b | 11146 | @opindex mno-spe |
39bc1876 NS |
11147 | This switch enables or disables the generation of SPE simd |
11148 | instructions. | |
83575957 | 11149 | |
78f5898b AH |
11150 | @item -mspe=@var{yes/no} |
11151 | This option has been deprecated. Use @option{-mspe} and | |
11152 | @option{-mno-spe} instead. | |
11153 | ||
4d4cbc0e | 11154 | @item -mfloat-gprs=@var{yes/single/double/no} |
39bc1876 NS |
11155 | @itemx -mfloat-gprs |
11156 | @opindex mfloat-gprs | |
11157 | This switch enables or disables the generation of floating point | |
11158 | operations on the general purpose registers for architectures that | |
4d4cbc0e AH |
11159 | support it. |
11160 | ||
11161 | The argument @var{yes} or @var{single} enables the use of | |
11162 | single-precision floating point operations. | |
11163 | ||
11164 | The argument @var{double} enables the use of single and | |
11165 | double-precision floating point operations. | |
11166 | ||
11167 | The argument @var{no} disables floating point operations on the | |
11168 | general purpose registers. | |
11169 | ||
11170 | This option is currently only available on the MPC854x. | |
83575957 | 11171 | |
49bd1d27 SS |
11172 | @item -m32 |
11173 | @itemx -m64 | |
11174 | @opindex m32 | |
11175 | @opindex m64 | |
11176 | Generate code for 32-bit or 64-bit environments of Darwin and SVR4 | |
11177 | targets (including GNU/Linux). The 32-bit environment sets int, long | |
11178 | and pointer to 32 bits and generates code that runs on any PowerPC | |
11179 | variant. The 64-bit environment sets int to 32 bits and long and | |
11180 | pointer to 64 bits, and generates code for PowerPC64, as for | |
11181 | @option{-mpowerpc64}. | |
11182 | ||
39bc1876 NS |
11183 | @item -mfull-toc |
11184 | @itemx -mno-fp-in-toc | |
11185 | @itemx -mno-sum-in-toc | |
11186 | @itemx -mminimal-toc | |
11187 | @opindex mfull-toc | |
11188 | @opindex mno-fp-in-toc | |
11189 | @opindex mno-sum-in-toc | |
11190 | @opindex mminimal-toc | |
11191 | Modify generation of the TOC (Table Of Contents), which is created for | |
11192 | every executable file. The @option{-mfull-toc} option is selected by | |
11193 | default. In that case, GCC will allocate at least one TOC entry for | |
11194 | each unique non-automatic variable reference in your program. GCC | |
11195 | will also place floating-point constants in the TOC@. However, only | |
11196 | 16,384 entries are available in the TOC@. | |
83575957 | 11197 | |
39bc1876 NS |
11198 | If you receive a linker error message that saying you have overflowed |
11199 | the available TOC space, you can reduce the amount of TOC space used | |
11200 | with the @option{-mno-fp-in-toc} and @option{-mno-sum-in-toc} options. | |
11201 | @option{-mno-fp-in-toc} prevents GCC from putting floating-point | |
11202 | constants in the TOC and @option{-mno-sum-in-toc} forces GCC to | |
11203 | generate code to calculate the sum of an address and a constant at | |
11204 | run-time instead of putting that sum into the TOC@. You may specify one | |
11205 | or both of these options. Each causes GCC to produce very slightly | |
11206 | slower and larger code at the expense of conserving TOC space. | |
83575957 | 11207 | |
39bc1876 NS |
11208 | If you still run out of space in the TOC even when you specify both of |
11209 | these options, specify @option{-mminimal-toc} instead. This option causes | |
11210 | GCC to make only one TOC entry for every file. When you specify this | |
11211 | option, GCC will produce code that is slower and larger but which | |
11212 | uses extremely little TOC space. You may wish to use this option | |
11213 | only on files that contain less frequently executed code. | |
83575957 | 11214 | |
39bc1876 NS |
11215 | @item -maix64 |
11216 | @itemx -maix32 | |
11217 | @opindex maix64 | |
11218 | @opindex maix32 | |
11219 | Enable 64-bit AIX ABI and calling convention: 64-bit pointers, 64-bit | |
11220 | @code{long} type, and the infrastructure needed to support them. | |
11221 | Specifying @option{-maix64} implies @option{-mpowerpc64} and | |
11222 | @option{-mpowerpc}, while @option{-maix32} disables the 64-bit ABI and | |
11223 | implies @option{-mno-powerpc64}. GCC defaults to @option{-maix32}. | |
83575957 | 11224 | |
de17c25f DE |
11225 | @item -mxl-compat |
11226 | @itemx -mno-xl-compat | |
11227 | @opindex mxl-compat | |
11228 | @opindex mno-xl-compat | |
11229 | Produce code that conforms more closely to IBM XLC semantics when using | |
11230 | AIX-compatible ABI. Pass floating-point arguments to prototyped | |
11231 | functions beyond the register save area (RSA) on the stack in addition | |
11232 | to argument FPRs. Do not assume that most significant double in 128 | |
11233 | bit long double value is properly rounded when comparing values. | |
11234 | ||
11235 | The AIX calling convention was extended but not initially documented to | |
39bc1876 NS |
11236 | handle an obscure K&R C case of calling a function that takes the |
11237 | address of its arguments with fewer arguments than declared. AIX XL | |
11238 | compilers access floating point arguments which do not fit in the | |
11239 | RSA from the stack when a subroutine is compiled without | |
11240 | optimization. Because always storing floating-point arguments on the | |
11241 | stack is inefficient and rarely needed, this option is not enabled by | |
11242 | default and only is necessary when calling subroutines compiled by AIX | |
11243 | XL compilers without optimization. | |
83575957 | 11244 | |
39bc1876 NS |
11245 | @item -mpe |
11246 | @opindex mpe | |
11247 | Support @dfn{IBM RS/6000 SP} @dfn{Parallel Environment} (PE)@. Link an | |
11248 | application written to use message passing with special startup code to | |
11249 | enable the application to run. The system must have PE installed in the | |
11250 | standard location (@file{/usr/lpp/ppe.poe/}), or the @file{specs} file | |
11251 | must be overridden with the @option{-specs=} option to specify the | |
11252 | appropriate directory location. The Parallel Environment does not | |
11253 | support threads, so the @option{-mpe} option and the @option{-pthread} | |
11254 | option are incompatible. | |
83575957 | 11255 | |
39bc1876 NS |
11256 | @item -malign-natural |
11257 | @itemx -malign-power | |
11258 | @opindex malign-natural | |
11259 | @opindex malign-power | |
d9168963 | 11260 | On AIX, 32-bit Darwin, and 64-bit PowerPC GNU/Linux, the option |
39bc1876 NS |
11261 | @option{-malign-natural} overrides the ABI-defined alignment of larger |
11262 | types, such as floating-point doubles, on their natural size-based boundary. | |
11263 | The option @option{-malign-power} instructs GCC to follow the ABI-specified | |
8a36672b | 11264 | alignment rules. GCC defaults to the standard alignment defined in the ABI@. |
83575957 | 11265 | |
d9168963 SS |
11266 | On 64-bit Darwin, natural alignment is the default, and @option{-malign-power} |
11267 | is not supported. | |
11268 | ||
39bc1876 NS |
11269 | @item -msoft-float |
11270 | @itemx -mhard-float | |
11271 | @opindex msoft-float | |
11272 | @opindex mhard-float | |
11273 | Generate code that does not use (uses) the floating-point register set. | |
11274 | Software floating point emulation is provided if you use the | |
11275 | @option{-msoft-float} option, and pass the option to GCC when linking. | |
83575957 | 11276 | |
39bc1876 NS |
11277 | @item -mmultiple |
11278 | @itemx -mno-multiple | |
11279 | @opindex mmultiple | |
11280 | @opindex mno-multiple | |
11281 | Generate code that uses (does not use) the load multiple word | |
11282 | instructions and the store multiple word instructions. These | |
11283 | instructions are generated by default on POWER systems, and not | |
11284 | generated on PowerPC systems. Do not use @option{-mmultiple} on little | |
11285 | endian PowerPC systems, since those instructions do not work when the | |
11286 | processor is in little endian mode. The exceptions are PPC740 and | |
11287 | PPC750 which permit the instructions usage in little endian mode. | |
83575957 | 11288 | |
39bc1876 NS |
11289 | @item -mstring |
11290 | @itemx -mno-string | |
11291 | @opindex mstring | |
11292 | @opindex mno-string | |
11293 | Generate code that uses (does not use) the load string instructions | |
11294 | and the store string word instructions to save multiple registers and | |
11295 | do small block moves. These instructions are generated by default on | |
11296 | POWER systems, and not generated on PowerPC systems. Do not use | |
11297 | @option{-mstring} on little endian PowerPC systems, since those | |
11298 | instructions do not work when the processor is in little endian mode. | |
11299 | The exceptions are PPC740 and PPC750 which permit the instructions | |
11300 | usage in little endian mode. | |
052a4b28 | 11301 | |
39bc1876 NS |
11302 | @item -mupdate |
11303 | @itemx -mno-update | |
11304 | @opindex mupdate | |
11305 | @opindex mno-update | |
11306 | Generate code that uses (does not use) the load or store instructions | |
11307 | that update the base register to the address of the calculated memory | |
11308 | location. These instructions are generated by default. If you use | |
11309 | @option{-mno-update}, there is a small window between the time that the | |
11310 | stack pointer is updated and the address of the previous frame is | |
11311 | stored, which means code that walks the stack frame across interrupts or | |
11312 | signals may get corrupted data. | |
052a4b28 | 11313 | |
39bc1876 NS |
11314 | @item -mfused-madd |
11315 | @itemx -mno-fused-madd | |
11316 | @opindex mfused-madd | |
11317 | @opindex mno-fused-madd | |
11318 | Generate code that uses (does not use) the floating point multiply and | |
11319 | accumulate instructions. These instructions are generated by default if | |
11320 | hardware floating is used. | |
3a69a7d5 | 11321 | |
131aeb82 JM |
11322 | @item -mmulhw |
11323 | @itemx -mno-mulhw | |
11324 | @opindex mmulhw | |
11325 | @opindex mno-mulhw | |
11326 | Generate code that uses (does not use) the half-word multiply and | |
11327 | multiply-accumulate instructions on the IBM 405 and 440 processors. | |
11328 | These instructions are generated by default when targetting those | |
11329 | processors. | |
11330 | ||
39bc1876 NS |
11331 | @item -mno-bit-align |
11332 | @itemx -mbit-align | |
11333 | @opindex mno-bit-align | |
11334 | @opindex mbit-align | |
11335 | On System V.4 and embedded PowerPC systems do not (do) force structures | |
11336 | and unions that contain bit-fields to be aligned to the base type of the | |
11337 | bit-field. | |
3a69a7d5 | 11338 | |
39bc1876 NS |
11339 | For example, by default a structure containing nothing but 8 |
11340 | @code{unsigned} bit-fields of length 1 would be aligned to a 4 byte | |
11341 | boundary and have a size of 4 bytes. By using @option{-mno-bit-align}, | |
11342 | the structure would be aligned to a 1 byte boundary and be one byte in | |
11343 | size. | |
3a69a7d5 | 11344 | |
39bc1876 NS |
11345 | @item -mno-strict-align |
11346 | @itemx -mstrict-align | |
11347 | @opindex mno-strict-align | |
11348 | @opindex mstrict-align | |
11349 | On System V.4 and embedded PowerPC systems do not (do) assume that | |
11350 | unaligned memory references will be handled by the system. | |
3a69a7d5 | 11351 | |
39bc1876 NS |
11352 | @item -mrelocatable |
11353 | @itemx -mno-relocatable | |
11354 | @opindex mrelocatable | |
11355 | @opindex mno-relocatable | |
11356 | On embedded PowerPC systems generate code that allows (does not allow) | |
11357 | the program to be relocated to a different address at runtime. If you | |
11358 | use @option{-mrelocatable} on any module, all objects linked together must | |
11359 | be compiled with @option{-mrelocatable} or @option{-mrelocatable-lib}. | |
3a69a7d5 | 11360 | |
39bc1876 NS |
11361 | @item -mrelocatable-lib |
11362 | @itemx -mno-relocatable-lib | |
11363 | @opindex mrelocatable-lib | |
11364 | @opindex mno-relocatable-lib | |
11365 | On embedded PowerPC systems generate code that allows (does not allow) | |
11366 | the program to be relocated to a different address at runtime. Modules | |
11367 | compiled with @option{-mrelocatable-lib} can be linked with either modules | |
11368 | compiled without @option{-mrelocatable} and @option{-mrelocatable-lib} or | |
11369 | with modules compiled with the @option{-mrelocatable} options. | |
052a4b28 | 11370 | |
39bc1876 NS |
11371 | @item -mno-toc |
11372 | @itemx -mtoc | |
11373 | @opindex mno-toc | |
11374 | @opindex mtoc | |
11375 | On System V.4 and embedded PowerPC systems do not (do) assume that | |
11376 | register 2 contains a pointer to a global area pointing to the addresses | |
11377 | used in the program. | |
052a4b28 | 11378 | |
39bc1876 NS |
11379 | @item -mlittle |
11380 | @itemx -mlittle-endian | |
11381 | @opindex mlittle | |
11382 | @opindex mlittle-endian | |
11383 | On System V.4 and embedded PowerPC systems compile code for the | |
11384 | processor in little endian mode. The @option{-mlittle-endian} option is | |
11385 | the same as @option{-mlittle}. | |
052a4b28 | 11386 | |
39bc1876 NS |
11387 | @item -mbig |
11388 | @itemx -mbig-endian | |
11389 | @opindex mbig | |
11390 | @opindex mbig-endian | |
11391 | On System V.4 and embedded PowerPC systems compile code for the | |
11392 | processor in big endian mode. The @option{-mbig-endian} option is | |
11393 | the same as @option{-mbig}. | |
052a4b28 | 11394 | |
39bc1876 NS |
11395 | @item -mdynamic-no-pic |
11396 | @opindex mdynamic-no-pic | |
11397 | On Darwin and Mac OS X systems, compile code so that it is not | |
11398 | relocatable, but that its external references are relocatable. The | |
11399 | resulting code is suitable for applications, but not shared | |
11400 | libraries. | |
3a69a7d5 | 11401 | |
39bc1876 NS |
11402 | @item -mprioritize-restricted-insns=@var{priority} |
11403 | @opindex mprioritize-restricted-insns | |
11404 | This option controls the priority that is assigned to | |
11405 | dispatch-slot restricted instructions during the second scheduling | |
11406 | pass. The argument @var{priority} takes the value @var{0/1/2} to assign | |
11407 | @var{no/highest/second-highest} priority to dispatch slot restricted | |
11408 | instructions. | |
3a69a7d5 | 11409 | |
39bc1876 NS |
11410 | @item -msched-costly-dep=@var{dependence_type} |
11411 | @opindex msched-costly-dep | |
11412 | This option controls which dependences are considered costly | |
11413 | by the target during instruction scheduling. The argument | |
11414 | @var{dependence_type} takes one of the following values: | |
11415 | @var{no}: no dependence is costly, | |
11416 | @var{all}: all dependences are costly, | |
11417 | @var{true_store_to_load}: a true dependence from store to load is costly, | |
11418 | @var{store_to_load}: any dependence from store to load is costly, | |
11419 | @var{number}: any dependence which latency >= @var{number} is costly. | |
11338cda | 11420 | |
39bc1876 NS |
11421 | @item -minsert-sched-nops=@var{scheme} |
11422 | @opindex minsert-sched-nops | |
11423 | This option controls which nop insertion scheme will be used during | |
8a36672b | 11424 | the second scheduling pass. The argument @var{scheme} takes one of the |
39bc1876 NS |
11425 | following values: |
11426 | @var{no}: Don't insert nops. | |
11427 | @var{pad}: Pad with nops any dispatch group which has vacant issue slots, | |
11428 | according to the scheduler's grouping. | |
11429 | @var{regroup_exact}: Insert nops to force costly dependent insns into | |
11430 | separate groups. Insert exactly as many nops as needed to force an insn | |
11431 | to a new group, according to the estimated processor grouping. | |
11432 | @var{number}: Insert nops to force costly dependent insns into | |
11433 | separate groups. Insert @var{number} nops to force an insn to a new group. | |
052a4b28 | 11434 | |
39bc1876 NS |
11435 | @item -mcall-sysv |
11436 | @opindex mcall-sysv | |
11437 | On System V.4 and embedded PowerPC systems compile code using calling | |
11438 | conventions that adheres to the March 1995 draft of the System V | |
11439 | Application Binary Interface, PowerPC processor supplement. This is the | |
11440 | default unless you configured GCC using @samp{powerpc-*-eabiaix}. | |
789a3090 | 11441 | |
39bc1876 NS |
11442 | @item -mcall-sysv-eabi |
11443 | @opindex mcall-sysv-eabi | |
11444 | Specify both @option{-mcall-sysv} and @option{-meabi} options. | |
789a3090 | 11445 | |
39bc1876 NS |
11446 | @item -mcall-sysv-noeabi |
11447 | @opindex mcall-sysv-noeabi | |
11448 | Specify both @option{-mcall-sysv} and @option{-mno-eabi} options. | |
789a3090 | 11449 | |
39bc1876 NS |
11450 | @item -mcall-solaris |
11451 | @opindex mcall-solaris | |
11452 | On System V.4 and embedded PowerPC systems compile code for the Solaris | |
11453 | operating system. | |
789a3090 | 11454 | |
39bc1876 NS |
11455 | @item -mcall-linux |
11456 | @opindex mcall-linux | |
11457 | On System V.4 and embedded PowerPC systems compile code for the | |
11458 | Linux-based GNU system. | |
789a3090 | 11459 | |
39bc1876 NS |
11460 | @item -mcall-gnu |
11461 | @opindex mcall-gnu | |
11462 | On System V.4 and embedded PowerPC systems compile code for the | |
11463 | Hurd-based GNU system. | |
789a3090 | 11464 | |
39bc1876 NS |
11465 | @item -mcall-netbsd |
11466 | @opindex mcall-netbsd | |
11467 | On System V.4 and embedded PowerPC systems compile code for the | |
11468 | NetBSD operating system. | |
789a3090 | 11469 | |
39bc1876 NS |
11470 | @item -maix-struct-return |
11471 | @opindex maix-struct-return | |
11472 | Return all structures in memory (as specified by the AIX ABI)@. | |
789a3090 | 11473 | |
39bc1876 NS |
11474 | @item -msvr4-struct-return |
11475 | @opindex msvr4-struct-return | |
11476 | Return structures smaller than 8 bytes in registers (as specified by the | |
11477 | SVR4 ABI)@. | |
789a3090 | 11478 | |
78f5898b AH |
11479 | @item -mabi=@var{abi-type} |
11480 | @opindex mabi | |
11481 | Extend the current ABI with a particular extension, or remove such extension. | |
11482 | Valid values are @var{altivec}, @var{no-altivec}, @var{spe}, | |
602ea4d3 JJ |
11483 | @var{no-spe}, @var{ibmlongdouble}, @var{ieeelongdouble}@. |
11484 | ||
11485 | @item -mabi=spe | |
11486 | @opindex mabi=spe | |
11487 | Extend the current ABI with SPE ABI extensions. This does not change | |
11488 | the default ABI, instead it adds the SPE ABI extensions to the current | |
11489 | ABI@. | |
11490 | ||
11491 | @item -mabi=no-spe | |
11492 | @opindex mabi=no-spe | |
11493 | Disable Booke SPE ABI extensions for the current ABI@. | |
11494 | ||
11495 | @item -mabi=ibmlongdouble | |
11496 | @opindex mabi=ibmlongdouble | |
11497 | Change the current ABI to use IBM extended precision long double. | |
11498 | This is a PowerPC 32-bit SYSV ABI option. | |
11499 | ||
11500 | @item -mabi=ieeelongdouble | |
11501 | @opindex mabi=ieeelongdouble | |
11502 | Change the current ABI to use IEEE extended precision long double. | |
11503 | This is a PowerPC 32-bit Linux ABI option. | |
789a3090 | 11504 | |
39bc1876 NS |
11505 | @item -mprototype |
11506 | @itemx -mno-prototype | |
11507 | @opindex mprototype | |
11508 | @opindex mno-prototype | |
11509 | On System V.4 and embedded PowerPC systems assume that all calls to | |
11510 | variable argument functions are properly prototyped. Otherwise, the | |
11511 | compiler must insert an instruction before every non prototyped call to | |
11512 | set or clear bit 6 of the condition code register (@var{CR}) to | |
11513 | indicate whether floating point values were passed in the floating point | |
11514 | registers in case the function takes a variable arguments. With | |
11515 | @option{-mprototype}, only calls to prototyped variable argument functions | |
11516 | will set or clear the bit. | |
83575957 | 11517 | |
39bc1876 NS |
11518 | @item -msim |
11519 | @opindex msim | |
11520 | On embedded PowerPC systems, assume that the startup module is called | |
11521 | @file{sim-crt0.o} and that the standard C libraries are @file{libsim.a} and | |
11522 | @file{libc.a}. This is the default for @samp{powerpc-*-eabisim}. | |
11523 | configurations. | |
df6194d4 | 11524 | |
39bc1876 NS |
11525 | @item -mmvme |
11526 | @opindex mmvme | |
11527 | On embedded PowerPC systems, assume that the startup module is called | |
11528 | @file{crt0.o} and the standard C libraries are @file{libmvme.a} and | |
11529 | @file{libc.a}. | |
df6194d4 | 11530 | |
39bc1876 NS |
11531 | @item -mads |
11532 | @opindex mads | |
11533 | On embedded PowerPC systems, assume that the startup module is called | |
11534 | @file{crt0.o} and the standard C libraries are @file{libads.a} and | |
11535 | @file{libc.a}. | |
df6194d4 | 11536 | |
39bc1876 NS |
11537 | @item -myellowknife |
11538 | @opindex myellowknife | |
11539 | On embedded PowerPC systems, assume that the startup module is called | |
11540 | @file{crt0.o} and the standard C libraries are @file{libyk.a} and | |
11541 | @file{libc.a}. | |
df6194d4 | 11542 | |
39bc1876 NS |
11543 | @item -mvxworks |
11544 | @opindex mvxworks | |
11545 | On System V.4 and embedded PowerPC systems, specify that you are | |
11546 | compiling for a VxWorks system. | |
df6194d4 | 11547 | |
39bc1876 NS |
11548 | @item -mwindiss |
11549 | @opindex mwindiss | |
11550 | Specify that you are compiling for the WindISS simulation environment. | |
df6194d4 | 11551 | |
39bc1876 NS |
11552 | @item -memb |
11553 | @opindex memb | |
11554 | On embedded PowerPC systems, set the @var{PPC_EMB} bit in the ELF flags | |
11555 | header to indicate that @samp{eabi} extended relocations are used. | |
df6194d4 | 11556 | |
39bc1876 NS |
11557 | @item -meabi |
11558 | @itemx -mno-eabi | |
11559 | @opindex meabi | |
11560 | @opindex mno-eabi | |
11561 | On System V.4 and embedded PowerPC systems do (do not) adhere to the | |
11562 | Embedded Applications Binary Interface (eabi) which is a set of | |
11563 | modifications to the System V.4 specifications. Selecting @option{-meabi} | |
11564 | means that the stack is aligned to an 8 byte boundary, a function | |
11565 | @code{__eabi} is called to from @code{main} to set up the eabi | |
11566 | environment, and the @option{-msdata} option can use both @code{r2} and | |
11567 | @code{r13} to point to two separate small data areas. Selecting | |
11568 | @option{-mno-eabi} means that the stack is aligned to a 16 byte boundary, | |
11569 | do not call an initialization function from @code{main}, and the | |
11570 | @option{-msdata} option will only use @code{r13} to point to a single | |
11571 | small data area. The @option{-meabi} option is on by default if you | |
11572 | configured GCC using one of the @samp{powerpc*-*-eabi*} options. | |
df6194d4 | 11573 | |
39bc1876 NS |
11574 | @item -msdata=eabi |
11575 | @opindex msdata=eabi | |
11576 | On System V.4 and embedded PowerPC systems, put small initialized | |
11577 | @code{const} global and static data in the @samp{.sdata2} section, which | |
11578 | is pointed to by register @code{r2}. Put small initialized | |
11579 | non-@code{const} global and static data in the @samp{.sdata} section, | |
11580 | which is pointed to by register @code{r13}. Put small uninitialized | |
11581 | global and static data in the @samp{.sbss} section, which is adjacent to | |
11582 | the @samp{.sdata} section. The @option{-msdata=eabi} option is | |
11583 | incompatible with the @option{-mrelocatable} option. The | |
11584 | @option{-msdata=eabi} option also sets the @option{-memb} option. | |
df6194d4 | 11585 | |
39bc1876 NS |
11586 | @item -msdata=sysv |
11587 | @opindex msdata=sysv | |
11588 | On System V.4 and embedded PowerPC systems, put small global and static | |
11589 | data in the @samp{.sdata} section, which is pointed to by register | |
11590 | @code{r13}. Put small uninitialized global and static data in the | |
11591 | @samp{.sbss} section, which is adjacent to the @samp{.sdata} section. | |
11592 | The @option{-msdata=sysv} option is incompatible with the | |
11593 | @option{-mrelocatable} option. | |
df6194d4 | 11594 | |
39bc1876 | 11595 | @item -msdata=default |
df6194d4 | 11596 | @itemx -msdata |
39bc1876 | 11597 | @opindex msdata=default |
cd3bb277 | 11598 | @opindex msdata |
39bc1876 NS |
11599 | On System V.4 and embedded PowerPC systems, if @option{-meabi} is used, |
11600 | compile code the same as @option{-msdata=eabi}, otherwise compile code the | |
11601 | same as @option{-msdata=sysv}. | |
df6194d4 | 11602 | |
39bc1876 NS |
11603 | @item -msdata-data |
11604 | @opindex msdata-data | |
05f25017 AM |
11605 | On System V.4 and embedded PowerPC systems, put small global |
11606 | data in the @samp{.sdata} section. Put small uninitialized global | |
11607 | data in the @samp{.sbss} section. Do not use register @code{r13} | |
39bc1876 NS |
11608 | to address small data however. This is the default behavior unless |
11609 | other @option{-msdata} options are used. | |
df6194d4 | 11610 | |
39bc1876 NS |
11611 | @item -msdata=none |
11612 | @itemx -mno-sdata | |
11613 | @opindex msdata=none | |
11614 | @opindex mno-sdata | |
11615 | On embedded PowerPC systems, put all initialized global and static data | |
11616 | in the @samp{.data} section, and all uninitialized data in the | |
11617 | @samp{.bss} section. | |
df6194d4 | 11618 | |
39bc1876 NS |
11619 | @item -G @var{num} |
11620 | @opindex G | |
11621 | @cindex smaller data references (PowerPC) | |
11622 | @cindex .sdata/.sdata2 references (PowerPC) | |
11623 | On embedded PowerPC systems, put global and static items less than or | |
11624 | equal to @var{num} bytes into the small data or bss sections instead of | |
11625 | the normal data or bss section. By default, @var{num} is 8. The | |
11626 | @option{-G @var{num}} switch is also passed to the linker. | |
11627 | All modules should be compiled with the same @option{-G @var{num}} value. | |
dcffbade | 11628 | |
39bc1876 NS |
11629 | @item -mregnames |
11630 | @itemx -mno-regnames | |
11631 | @opindex mregnames | |
11632 | @opindex mno-regnames | |
11633 | On System V.4 and embedded PowerPC systems do (do not) emit register | |
11634 | names in the assembly language output using symbolic forms. | |
dcffbade | 11635 | |
39bc1876 NS |
11636 | @item -mlongcall |
11637 | @itemx -mno-longcall | |
11638 | @opindex mlongcall | |
11639 | @opindex mno-longcall | |
11640 | Default to making all function calls indirectly, using a register, so | |
11641 | that functions which reside further than 32 megabytes (33,554,432 | |
11642 | bytes) from the current location can be called. This setting can be | |
11643 | overridden by the @code{shortcall} function attribute, or by | |
11644 | @code{#pragma longcall(0)}. | |
dcffbade | 11645 | |
39bc1876 NS |
11646 | Some linkers are capable of detecting out-of-range calls and generating |
11647 | glue code on the fly. On these systems, long calls are unnecessary and | |
11648 | generate slower code. As of this writing, the AIX linker can do this, | |
11649 | as can the GNU linker for PowerPC/64. It is planned to add this feature | |
11650 | to the GNU linker for 32-bit PowerPC systems as well. | |
df6194d4 | 11651 | |
39bc1876 NS |
11652 | On Darwin/PPC systems, @code{#pragma longcall} will generate ``jbsr |
11653 | callee, L42'', plus a ``branch island'' (glue code). The two target | |
d78aa55c | 11654 | addresses represent the callee and the ``branch island''. The |
39bc1876 NS |
11655 | Darwin/PPC linker will prefer the first address and generate a ``bl |
11656 | callee'' if the PPC ``bl'' instruction will reach the callee directly; | |
11657 | otherwise, the linker will generate ``bl L42'' to call the ``branch | |
d78aa55c | 11658 | island''. The ``branch island'' is appended to the body of the |
39bc1876 NS |
11659 | calling function; it computes the full 32-bit address of the callee |
11660 | and jumps to it. | |
df6194d4 | 11661 | |
39bc1876 NS |
11662 | On Mach-O (Darwin) systems, this option directs the compiler emit to |
11663 | the glue for every direct call, and the Darwin linker decides whether | |
11664 | to use or discard it. | |
11665 | ||
11666 | In the future, we may cause GCC to ignore all longcall specifications | |
11667 | when the linker is known to generate glue. | |
11668 | ||
11669 | @item -pthread | |
11670 | @opindex pthread | |
11671 | Adds support for multithreading with the @dfn{pthreads} library. | |
11672 | This option sets flags for both the preprocessor and linker. | |
30028c85 | 11673 | |
df6194d4 JW |
11674 | @end table |
11675 | ||
91abf72d HP |
11676 | @node S/390 and zSeries Options |
11677 | @subsection S/390 and zSeries Options | |
11678 | @cindex S/390 and zSeries Options | |
11679 | ||
11680 | These are the @samp{-m} options defined for the S/390 and zSeries architecture. | |
11681 | ||
11682 | @table @gcctabopt | |
11683 | @item -mhard-float | |
11684 | @itemx -msoft-float | |
11685 | @opindex mhard-float | |
11686 | @opindex msoft-float | |
11687 | Use (do not use) the hardware floating-point instructions and registers | |
11688 | for floating-point operations. When @option{-msoft-float} is specified, | |
11689 | functions in @file{libgcc.a} will be used to perform floating-point | |
11690 | operations. When @option{-mhard-float} is specified, the compiler | |
11691 | generates IEEE floating-point instructions. This is the default. | |
11692 | ||
11693 | @item -mbackchain | |
11694 | @itemx -mno-backchain | |
11695 | @opindex mbackchain | |
11696 | @opindex mno-backchain | |
b3d31392 AK |
11697 | Store (do not store) the address of the caller's frame as backchain pointer |
11698 | into the callee's stack frame. | |
adf39f8f AK |
11699 | A backchain may be needed to allow debugging using tools that do not understand |
11700 | DWARF-2 call frame information. | |
b3d31392 AK |
11701 | When @option{-mno-packed-stack} is in effect, the backchain pointer is stored |
11702 | at the bottom of the stack frame; when @option{-mpacked-stack} is in effect, | |
11703 | the backchain is placed into the topmost word of the 96/160 byte register | |
11704 | save area. | |
11705 | ||
11706 | In general, code compiled with @option{-mbackchain} is call-compatible with | |
11707 | code compiled with @option{-mmo-backchain}; however, use of the backchain | |
11708 | for debugging purposes usually requires that the whole binary is built with | |
6b78f6be AK |
11709 | @option{-mbackchain}. Note that the combination of @option{-mbackchain}, |
11710 | @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order | |
11711 | to build a linux kernel use @option{-msoft-float}. | |
b3d31392 AK |
11712 | |
11713 | The default is to not maintain the backchain. | |
11714 | ||
11715 | @item -mpacked-stack | |
11716 | @item -mno-packed-stack | |
11717 | @opindex mpacked-stack | |
11718 | @opindex mno-packed-stack | |
11719 | Use (do not use) the packed stack layout. When @option{-mno-packed-stack} is | |
11720 | specified, the compiler uses the all fields of the 96/160 byte register save | |
11721 | area only for their default purpose; unused fields still take up stack space. | |
11722 | When @option{-mpacked-stack} is specified, register save slots are densely | |
11723 | packed at the top of the register save area; unused space is reused for other | |
11724 | purposes, allowing for more efficient use of the available stack space. | |
11725 | However, when @option{-mbackchain} is also in effect, the topmost word of | |
11726 | the save area is always used to store the backchain, and the return address | |
11727 | register is always saved two words below the backchain. | |
11728 | ||
f2fd3821 | 11729 | As long as the stack frame backchain is not used, code generated with |
b3d31392 AK |
11730 | @option{-mpacked-stack} is call-compatible with code generated with |
11731 | @option{-mno-packed-stack}. Note that some non-FSF releases of GCC 2.95 for | |
11732 | S/390 or zSeries generated code that uses the stack frame backchain at run | |
11733 | time, not just for debugging purposes. Such code is not call-compatible | |
11734 | with code compiled with @option{-mpacked-stack}. Also, note that the | |
6b78f6be AK |
11735 | combination of @option{-mbackchain}, |
11736 | @option{-mpacked-stack} and @option{-mhard-float} is not supported. In order | |
11737 | to build a linux kernel use @option{-msoft-float}. | |
b3d31392 AK |
11738 | |
11739 | The default is to not use the packed stack layout. | |
91abf72d HP |
11740 | |
11741 | @item -msmall-exec | |
11742 | @itemx -mno-small-exec | |
11743 | @opindex msmall-exec | |
11744 | @opindex mno-small-exec | |
f282ffb3 JM |
11745 | Generate (or do not generate) code using the @code{bras} instruction |
11746 | to do subroutine calls. | |
91abf72d HP |
11747 | This only works reliably if the total executable size does not |
11748 | exceed 64k. The default is to use the @code{basr} instruction instead, | |
11749 | which does not have this limitation. | |
11750 | ||
11751 | @item -m64 | |
11752 | @itemx -m31 | |
11753 | @opindex m64 | |
11754 | @opindex m31 | |
11755 | When @option{-m31} is specified, generate code compliant to the | |
95fef11f JM |
11756 | GNU/Linux for S/390 ABI@. When @option{-m64} is specified, generate |
11757 | code compliant to the GNU/Linux for zSeries ABI@. This allows GCC in | |
91abf72d | 11758 | particular to generate 64-bit instructions. For the @samp{s390} |
f282ffb3 | 11759 | targets, the default is @option{-m31}, while the @samp{s390x} |
91abf72d HP |
11760 | targets default to @option{-m64}. |
11761 | ||
1fec52be HP |
11762 | @item -mzarch |
11763 | @itemx -mesa | |
11764 | @opindex mzarch | |
11765 | @opindex mesa | |
daf2f129 JM |
11766 | When @option{-mzarch} is specified, generate code using the |
11767 | instructions available on z/Architecture. | |
11768 | When @option{-mesa} is specified, generate code using the | |
8a36672b | 11769 | instructions available on ESA/390. Note that @option{-mesa} is |
1fec52be | 11770 | not possible with @option{-m64}. |
95fef11f | 11771 | When generating code compliant to the GNU/Linux for S/390 ABI, |
f13e0d4e | 11772 | the default is @option{-mesa}. When generating code compliant |
95fef11f | 11773 | to the GNU/Linux for zSeries ABI, the default is @option{-mzarch}. |
1fec52be | 11774 | |
91abf72d HP |
11775 | @item -mmvcle |
11776 | @itemx -mno-mvcle | |
11777 | @opindex mmvcle | |
11778 | @opindex mno-mvcle | |
f282ffb3 | 11779 | Generate (or do not generate) code using the @code{mvcle} instruction |
3364c33b | 11780 | to perform block moves. When @option{-mno-mvcle} is specified, |
8daf098e AS |
11781 | use a @code{mvc} loop instead. This is the default unless optimizing for |
11782 | size. | |
91abf72d HP |
11783 | |
11784 | @item -mdebug | |
11785 | @itemx -mno-debug | |
11786 | @opindex mdebug | |
11787 | @opindex mno-debug | |
11788 | Print (or do not print) additional debug information when compiling. | |
11789 | The default is to not print debug information. | |
11790 | ||
f13e0d4e | 11791 | @item -march=@var{cpu-type} |
1fec52be | 11792 | @opindex march |
f13e0d4e | 11793 | Generate code that will run on @var{cpu-type}, which is the name of a system |
8a36672b | 11794 | representing a certain processor type. Possible values for |
f13e0d4e UW |
11795 | @var{cpu-type} are @samp{g5}, @samp{g6}, @samp{z900}, and @samp{z990}. |
11796 | When generating code using the instructions available on z/Architecture, | |
11797 | the default is @option{-march=z900}. Otherwise, the default is | |
11798 | @option{-march=g5}. | |
1fec52be | 11799 | |
f13e0d4e | 11800 | @item -mtune=@var{cpu-type} |
35082351 | 11801 | @opindex mtune |
1fec52be | 11802 | Tune to @var{cpu-type} everything applicable about the generated code, |
f13e0d4e UW |
11803 | except for the ABI and the set of available instructions. |
11804 | The list of @var{cpu-type} values is the same as for @option{-march}. | |
11805 | The default is the value used for @option{-march}. | |
1fec52be | 11806 | |
f26c1794 EC |
11807 | @item -mtpf-trace |
11808 | @itemx -mno-tpf-trace | |
11809 | @opindex mtpf-trace | |
11810 | @opindex mno-tpf-trace | |
11811 | Generate code that adds (does not add) in TPF OS specific branches to trace | |
11812 | routines in the operating system. This option is off by default, even | |
8a36672b | 11813 | when compiling for the TPF OS@. |
f26c1794 | 11814 | |
f2d226e1 AK |
11815 | @item -mfused-madd |
11816 | @itemx -mno-fused-madd | |
11817 | @opindex mfused-madd | |
11818 | @opindex mno-fused-madd | |
11819 | Generate code that uses (does not use) the floating point multiply and | |
11820 | accumulate instructions. These instructions are generated by default if | |
11821 | hardware floating point is used. | |
d75f90f1 AK |
11822 | |
11823 | @item -mwarn-framesize=@var{framesize} | |
11824 | @opindex mwarn-framesize | |
11825 | Emit a warning if the current function exceeds the given frame size. Because | |
11826 | this is a compile time check it doesn't need to be a real problem when the program | |
f0eb93a8 JM |
11827 | runs. It is intended to identify functions which most probably cause |
11828 | a stack overflow. It is useful to be used in an environment with limited stack | |
431ae0bf | 11829 | size e.g.@: the linux kernel. |
d75f90f1 AK |
11830 | |
11831 | @item -mwarn-dynamicstack | |
11832 | @opindex mwarn-dynamicstack | |
11833 | Emit a warning if the function calls alloca or uses dynamically | |
11834 | sized arrays. This is generally a bad idea with a limited stack size. | |
11835 | ||
11836 | @item -mstack-guard=@var{stack-guard} | |
11837 | @item -mstack-size=@var{stack-size} | |
11838 | @opindex mstack-guard | |
11839 | @opindex mstack-size | |
11840 | These arguments always have to be used in conjunction. If they are present the s390 | |
11841 | back end emits additional instructions in the function prologue which trigger a trap | |
f0eb93a8 JM |
11842 | if the stack size is @var{stack-guard} bytes above the @var{stack-size} |
11843 | (remember that the stack on s390 grows downward). These options are intended to | |
11844 | be used to help debugging stack overflow problems. The additionally emitted code | |
f695eccf | 11845 | causes only little overhead and hence can also be used in production like systems |
f0eb93a8 | 11846 | without greater performance degradation. The given values have to be exact |
f695eccf AK |
11847 | powers of 2 and @var{stack-size} has to be greater than @var{stack-guard} without |
11848 | exceeding 64k. | |
d75f90f1 AK |
11849 | In order to be efficient the extra code makes the assumption that the stack starts |
11850 | at an address aligned to the value given by @var{stack-size}. | |
91abf72d HP |
11851 | @end table |
11852 | ||
39bc1876 NS |
11853 | @node SH Options |
11854 | @subsection SH Options | |
bcf684c7 | 11855 | |
39bc1876 | 11856 | These @samp{-m} options are defined for the SH implementations: |
bcf684c7 | 11857 | |
5d22c1a5 | 11858 | @table @gcctabopt |
39bc1876 NS |
11859 | @item -m1 |
11860 | @opindex m1 | |
11861 | Generate code for the SH1. | |
9f85bca7 | 11862 | |
39bc1876 NS |
11863 | @item -m2 |
11864 | @opindex m2 | |
11865 | Generate code for the SH2. | |
9f85bca7 | 11866 | |
39bc1876 NS |
11867 | @item -m2e |
11868 | Generate code for the SH2e. | |
9f85bca7 | 11869 | |
39bc1876 NS |
11870 | @item -m3 |
11871 | @opindex m3 | |
11872 | Generate code for the SH3. | |
9f85bca7 | 11873 | |
39bc1876 NS |
11874 | @item -m3e |
11875 | @opindex m3e | |
11876 | Generate code for the SH3e. | |
9f85bca7 | 11877 | |
39bc1876 NS |
11878 | @item -m4-nofpu |
11879 | @opindex m4-nofpu | |
11880 | Generate code for the SH4 without a floating-point unit. | |
9f85bca7 | 11881 | |
39bc1876 NS |
11882 | @item -m4-single-only |
11883 | @opindex m4-single-only | |
11884 | Generate code for the SH4 with a floating-point unit that only | |
11885 | supports single-precision arithmetic. | |
9f85bca7 | 11886 | |
39bc1876 NS |
11887 | @item -m4-single |
11888 | @opindex m4-single | |
11889 | Generate code for the SH4 assuming the floating-point unit is in | |
11890 | single-precision mode by default. | |
9f85bca7 | 11891 | |
39bc1876 NS |
11892 | @item -m4 |
11893 | @opindex m4 | |
11894 | Generate code for the SH4. | |
9f85bca7 | 11895 | |
312209c6 AO |
11896 | @item -m4a-nofpu |
11897 | @opindex m4a-nofpu | |
11898 | Generate code for the SH4al-dsp, or for a SH4a in such a way that the | |
11899 | floating-point unit is not used. | |
11900 | ||
11901 | @item -m4a-single-only | |
11902 | @opindex m4a-single-only | |
11903 | Generate code for the SH4a, in such a way that no double-precision | |
11904 | floating point operations are used. | |
11905 | ||
11906 | @item -m4a-single | |
11907 | @opindex m4a-single | |
11908 | Generate code for the SH4a assuming the floating-point unit is in | |
11909 | single-precision mode by default. | |
11910 | ||
11911 | @item -m4a | |
11912 | @opindex m4a | |
11913 | Generate code for the SH4a. | |
11914 | ||
11915 | @item -m4al | |
11916 | @opindex m4al | |
11917 | Same as @option{-m4a-nofpu}, except that it implicitly passes | |
11918 | @option{-dsp} to the assembler. GCC doesn't generate any DSP | |
11919 | instructions at the moment. | |
11920 | ||
39bc1876 NS |
11921 | @item -mb |
11922 | @opindex mb | |
11923 | Compile code for the processor in big endian mode. | |
9f85bca7 | 11924 | |
39bc1876 NS |
11925 | @item -ml |
11926 | @opindex ml | |
11927 | Compile code for the processor in little endian mode. | |
9f85bca7 | 11928 | |
39bc1876 NS |
11929 | @item -mdalign |
11930 | @opindex mdalign | |
11931 | Align doubles at 64-bit boundaries. Note that this changes the calling | |
11932 | conventions, and thus some functions from the standard C library will | |
11933 | not work unless you recompile it first with @option{-mdalign}. | |
9f85bca7 | 11934 | |
39bc1876 NS |
11935 | @item -mrelax |
11936 | @opindex mrelax | |
11937 | Shorten some address references at link time, when possible; uses the | |
11938 | linker option @option{-relax}. | |
9f85bca7 | 11939 | |
39bc1876 NS |
11940 | @item -mbigtable |
11941 | @opindex mbigtable | |
11942 | Use 32-bit offsets in @code{switch} tables. The default is to use | |
11943 | 16-bit offsets. | |
9f85bca7 | 11944 | |
39bc1876 NS |
11945 | @item -mfmovd |
11946 | @opindex mfmovd | |
11947 | Enable the use of the instruction @code{fmovd}. | |
9f85bca7 | 11948 | |
39bc1876 NS |
11949 | @item -mhitachi |
11950 | @opindex mhitachi | |
11951 | Comply with the calling conventions defined by Renesas. | |
9f85bca7 | 11952 | |
2acc29bd NC |
11953 | @item -mrenesas |
11954 | @opindex mhitachi | |
11955 | Comply with the calling conventions defined by Renesas. | |
11956 | ||
11957 | @item -mno-renesas | |
11958 | @opindex mhitachi | |
11959 | Comply with the calling conventions defined for GCC before the Renesas | |
11960 | conventions were available. This option is the default for all | |
11961 | targets of the SH toolchain except for @samp{sh-symbianelf}. | |
11962 | ||
39bc1876 NS |
11963 | @item -mnomacsave |
11964 | @opindex mnomacsave | |
11965 | Mark the @code{MAC} register as call-clobbered, even if | |
11966 | @option{-mhitachi} is given. | |
9f85bca7 | 11967 | |
39bc1876 NS |
11968 | @item -mieee |
11969 | @opindex mieee | |
11970 | Increase IEEE-compliance of floating-point code. | |
73a4d10b R |
11971 | At the moment, this is equivalent to @option{-fno-finite-math-only}. |
11972 | When generating 16 bit SH opcodes, getting IEEE-conforming results for | |
11973 | comparisons of NANs / infinities incurs extra overhead in every | |
11974 | floating point comparison, therefore the default is set to | |
11975 | @option{-ffinite-math-only}. | |
9f85bca7 | 11976 | |
39bc1876 NS |
11977 | @item -misize |
11978 | @opindex misize | |
11979 | Dump instruction size and location in the assembly code. | |
9f85bca7 | 11980 | |
39bc1876 NS |
11981 | @item -mpadstruct |
11982 | @opindex mpadstruct | |
11983 | This option is deprecated. It pads structures to multiple of 4 bytes, | |
11984 | which is incompatible with the SH ABI@. | |
9f85bca7 | 11985 | |
39bc1876 NS |
11986 | @item -mspace |
11987 | @opindex mspace | |
11988 | Optimize for space instead of speed. Implied by @option{-Os}. | |
9f85bca7 | 11989 | |
39bc1876 NS |
11990 | @item -mprefergot |
11991 | @opindex mprefergot | |
11992 | When generating position-independent code, emit function calls using | |
11993 | the Global Offset Table instead of the Procedure Linkage Table. | |
9f85bca7 | 11994 | |
39bc1876 NS |
11995 | @item -musermode |
11996 | @opindex musermode | |
11997 | Generate a library function call to invalidate instruction cache | |
11998 | entries, after fixing up a trampoline. This library function call | |
11999 | doesn't assume it can write to the whole memory address space. This | |
12000 | is the default when the target is @code{sh-*-linux*}. | |
73a4d10b R |
12001 | |
12002 | @item -multcost=@var{number} | |
12003 | @opindex multcost=@var{number} | |
12004 | Set the cost to assume for a multiply insn. | |
12005 | ||
12006 | @item -mdiv=@var{strategy} | |
12007 | @opindex mdiv=@var{strategy} | |
12008 | Set the division strategy to use for SHmedia code. @var{strategy} must be | |
12009 | one of: call, call2, fp, inv, inv:minlat, inv20u, inv20l, inv:call, | |
12010 | inv:call2, inv:fp . | |
12011 | "fp" performs the operation in floating point. This has a very high latency, | |
12012 | but needs only a few instructions, so it might be a good choice if | |
30dc60c7 | 12013 | your code has enough easily exploitable ILP to allow the compiler to |
73a4d10b R |
12014 | schedule the floating point instructions together with other instructions. |
12015 | Division by zero causes a floating point exception. | |
12016 | "inv" uses integer operations to calculate the inverse of the divisor, | |
30dc60c7 | 12017 | and then multiplies the dividend with the inverse. This strategy allows |
73a4d10b R |
12018 | cse and hoisting of the inverse calculation. Division by zero calculates |
12019 | an unspecified result, but does not trap. | |
12020 | "inv:minlat" is a variant of "inv" where if no cse / hoisting opportunities | |
12021 | have been found, or if the entire operation has been hoisted to the same | |
12022 | place, the last stages of the inverse calculation are intertwined with the | |
12023 | final multiply to reduce the overall latency, at the expense of using a few | |
12024 | more instructions, and thus offering fewer scheduling opportunities with | |
12025 | other code. | |
12026 | "call" calls a library function that usually implements the inv:minlat | |
12027 | strategy. | |
12028 | This gives high code density for m5-*media-nofpu compilations. | |
12029 | "call2" uses a different entry point of the same library function, where it | |
12030 | assumes that a pointer to a lookup table has already been set up, which | |
12031 | exposes the pointer load to cse / code hoisting optimizations. | |
12032 | "inv:call", "inv:call2" and "inv:fp" all use the "inv" algorithm for initial | |
12033 | code generation, but if the code stays unoptimized, revert to the "call", | |
30dc60c7 | 12034 | "call2", or "fp" strategies, respectively. Note that the |
73a4d10b R |
12035 | potentially-trapping side effect of division by zero is carried by a |
12036 | separate instruction, so it is possible that all the integer instructions | |
12037 | are hoisted out, but the marker for the side effect stays where it is. | |
12038 | A recombination to fp operations or a call is not possible in that case. | |
12039 | "inv20u" and "inv20l" are variants of the "inv:minlat" strategy. In the case | |
12040 | that the inverse calculation was nor separated from the multiply, they speed | |
12041 | up division where the dividend fits into 20 bits (plus sign where applicable), | |
12042 | by inserting a test to skip a number of operations in this case; this test | |
30dc60c7 | 12043 | slows down the case of larger dividends. inv20u assumes the case of a such |
73a4d10b R |
12044 | a small dividend to be unlikely, and inv20l assumes it to be likely. |
12045 | ||
12046 | @item -mdivsi3_libfunc=@var{name} | |
12047 | @opindex mdivsi3_libfunc=@var{name} | |
12048 | Set the name of the library function used for 32 bit signed division to | |
12049 | @var{name}. This only affect the name used in the call and inv:call | |
12050 | division strategies, and the compiler will still expect the same | |
12051 | sets of input/output/clobbered registers as if this option was not present. | |
12052 | ||
12053 | @item -madjust-unroll | |
12054 | @opindex madjust-unroll | |
12055 | Throttle unrolling to avoid thrashing target registers. | |
12056 | This option only has an effect if the gcc code base supports the | |
12057 | TARGET_ADJUST_UNROLL_MAX target hook. | |
12058 | ||
12059 | @item -mindexed-addressing | |
12060 | @opindex mindexed-addressing | |
12061 | Enable the use of the indexed addressing mode for SHmedia32/SHcompact. | |
12062 | This is only safe if the hardware and/or OS implement 32 bit wrap-around | |
12063 | semantics for the indexed addressing mode. The architecture allows the | |
12064 | implementation of processors with 64 bit MMU, which the OS could use to | |
30dc60c7 | 12065 | get 32 bit addressing, but since no current hardware implementation supports |
73a4d10b R |
12066 | this or any other way to make the indexed addressing mode safe to use in |
12067 | the 32 bit ABI, the default is -mno-indexed-addressing. | |
12068 | ||
12069 | @item -mgettrcost=@var{number} | |
12070 | @opindex mgettrcost=@var{number} | |
12071 | Set the cost assumed for the gettr instruction to @var{number}. | |
12072 | The default is 2 if @option{-mpt-fixed} is in effect, 100 otherwise. | |
12073 | ||
12074 | @item -mpt-fixed | |
12075 | @opindex mpt-fixed | |
12076 | Assume pt* instructions won't trap. This will generally generate better | |
12077 | scheduled code, but is unsafe on current hardware. The current architecture | |
12078 | definition says that ptabs and ptrel trap when the target anded with 3 is 3. | |
12079 | This has the unintentional effect of making it unsafe to schedule ptabs / | |
12080 | ptrel before a branch, or hoist it out of a loop. For example, | |
12081 | __do_global_ctors, a part of libgcc that runs constructors at program | |
12082 | startup, calls functions in a list which is delimited by -1. With the | |
12083 | -mpt-fixed option, the ptabs will be done before testing against -1. | |
12084 | That means that all the constructors will be run a bit quicker, but when | |
30dc60c7 | 12085 | the loop comes to the end of the list, the program crashes because ptabs |
73a4d10b R |
12086 | loads -1 into a target register. Since this option is unsafe for any |
12087 | hardware implementing the current architecture specification, the default | |
12088 | is -mno-pt-fixed. Unless the user specifies a specific cost with | |
12089 | @option{-mgettrcost}, -mno-pt-fixed also implies @option{-mgettrcost=100}; | |
12090 | this deters register allocation using target registers for storing | |
12091 | ordinary integers. | |
12092 | ||
12093 | @item -minvalid-symbols | |
12094 | @opindex minvalid-symbols | |
12095 | Assume symbols might be invalid. Ordinary function symbols generated by | |
12096 | the compiler will always be valid to load with movi/shori/ptabs or | |
12097 | movi/shori/ptrel, but with assembler and/or linker tricks it is possible | |
12098 | to generate symbols that will cause ptabs / ptrel to trap. | |
12099 | This option is only meaningful when @option{-mno-pt-fixed} is in effect. | |
12100 | It will then prevent cross-basic-block cse, hoisting and most scheduling | |
12101 | of symbol loads. The default is @option{-mno-invalid-symbols}. | |
9f85bca7 JM |
12102 | @end table |
12103 | ||
39bc1876 NS |
12104 | @node SPARC Options |
12105 | @subsection SPARC Options | |
12106 | @cindex SPARC options | |
69a0611f | 12107 | |
39bc1876 | 12108 | These @samp{-m} options are supported on the SPARC: |
69a0611f GK |
12109 | |
12110 | @table @gcctabopt | |
39bc1876 NS |
12111 | @item -mno-app-regs |
12112 | @itemx -mapp-regs | |
12113 | @opindex mno-app-regs | |
12114 | @opindex mapp-regs | |
12115 | Specify @option{-mapp-regs} to generate output using the global registers | |
12116 | 2 through 4, which the SPARC SVR4 ABI reserves for applications. This | |
d14f5ecb | 12117 | is the default. |
69a0611f | 12118 | |
39bc1876 NS |
12119 | To be fully SVR4 ABI compliant at the cost of some performance loss, |
12120 | specify @option{-mno-app-regs}. You should compile libraries and system | |
12121 | software with this option. | |
70899148 | 12122 | |
39bc1876 NS |
12123 | @item -mfpu |
12124 | @itemx -mhard-float | |
12125 | @opindex mfpu | |
12126 | @opindex mhard-float | |
12127 | Generate output containing floating point instructions. This is the | |
12128 | default. | |
70899148 | 12129 | |
39bc1876 NS |
12130 | @item -mno-fpu |
12131 | @itemx -msoft-float | |
12132 | @opindex mno-fpu | |
12133 | @opindex msoft-float | |
12134 | Generate output containing library calls for floating point. | |
12135 | @strong{Warning:} the requisite libraries are not available for all SPARC | |
12136 | targets. Normally the facilities of the machine's usual C compiler are | |
12137 | used, but this cannot be done directly in cross-compilation. You must make | |
12138 | your own arrangements to provide suitable library functions for | |
12139 | cross-compilation. The embedded targets @samp{sparc-*-aout} and | |
12140 | @samp{sparclite-*-*} do provide software floating point support. | |
70899148 | 12141 | |
39bc1876 NS |
12142 | @option{-msoft-float} changes the calling convention in the output file; |
12143 | therefore, it is only useful if you compile @emph{all} of a program with | |
12144 | this option. In particular, you need to compile @file{libgcc.a}, the | |
12145 | library that comes with GCC, with @option{-msoft-float} in order for | |
12146 | this to work. | |
70899148 | 12147 | |
39bc1876 NS |
12148 | @item -mhard-quad-float |
12149 | @opindex mhard-quad-float | |
12150 | Generate output containing quad-word (long double) floating point | |
12151 | instructions. | |
70899148 | 12152 | |
39bc1876 NS |
12153 | @item -msoft-quad-float |
12154 | @opindex msoft-quad-float | |
12155 | Generate output containing library calls for quad-word (long double) | |
12156 | floating point instructions. The functions called are those specified | |
12157 | in the SPARC ABI@. This is the default. | |
70899148 | 12158 | |
39bc1876 NS |
12159 | As of this writing, there are no SPARC implementations that have hardware |
12160 | support for the quad-word floating point instructions. They all invoke | |
12161 | a trap handler for one of these instructions, and then the trap handler | |
12162 | emulates the effect of the instruction. Because of the trap handler overhead, | |
12163 | this is much slower than calling the ABI library routines. Thus the | |
12164 | @option{-msoft-quad-float} option is the default. | |
70899148 | 12165 | |
39bc1876 NS |
12166 | @item -mno-unaligned-doubles |
12167 | @itemx -munaligned-doubles | |
12168 | @opindex mno-unaligned-doubles | |
12169 | @opindex munaligned-doubles | |
12170 | Assume that doubles have 8 byte alignment. This is the default. | |
70899148 | 12171 | |
39bc1876 NS |
12172 | With @option{-munaligned-doubles}, GCC assumes that doubles have 8 byte |
12173 | alignment only if they are contained in another type, or if they have an | |
12174 | absolute address. Otherwise, it assumes they have 4 byte alignment. | |
12175 | Specifying this option avoids some rare compatibility problems with code | |
12176 | generated by other compilers. It is not the default because it results | |
12177 | in a performance loss, especially for floating point code. | |
70899148 | 12178 | |
39bc1876 NS |
12179 | @item -mno-faster-structs |
12180 | @itemx -mfaster-structs | |
12181 | @opindex mno-faster-structs | |
12182 | @opindex mfaster-structs | |
12183 | With @option{-mfaster-structs}, the compiler assumes that structures | |
12184 | should have 8 byte alignment. This enables the use of pairs of | |
12185 | @code{ldd} and @code{std} instructions for copies in structure | |
12186 | assignment, in place of twice as many @code{ld} and @code{st} pairs. | |
12187 | However, the use of this changed alignment directly violates the SPARC | |
12188 | ABI@. Thus, it's intended only for use on targets where the developer | |
12189 | acknowledges that their resulting code will not be directly in line with | |
12190 | the rules of the ABI@. | |
70899148 | 12191 | |
39bc1876 NS |
12192 | @item -mimpure-text |
12193 | @opindex mimpure-text | |
12194 | @option{-mimpure-text}, used in addition to @option{-shared}, tells | |
12195 | the compiler to not pass @option{-z text} to the linker when linking a | |
12196 | shared object. Using this option, you can link position-dependent | |
12197 | code into a shared object. | |
70899148 | 12198 | |
39bc1876 NS |
12199 | @option{-mimpure-text} suppresses the ``relocations remain against |
12200 | allocatable but non-writable sections'' linker error message. | |
12201 | However, the necessary relocations will trigger copy-on-write, and the | |
12202 | shared object is not actually shared across processes. Instead of | |
12203 | using @option{-mimpure-text}, you should compile all source code with | |
12204 | @option{-fpic} or @option{-fPIC}. | |
12205 | ||
12206 | This option is only available on SunOS and Solaris. | |
12207 | ||
12208 | @item -mcpu=@var{cpu_type} | |
12209 | @opindex mcpu | |
12210 | Set the instruction set, register set, and instruction scheduling parameters | |
12211 | for machine type @var{cpu_type}. Supported values for @var{cpu_type} are | |
12212 | @samp{v7}, @samp{cypress}, @samp{v8}, @samp{supersparc}, @samp{sparclite}, | |
12213 | @samp{f930}, @samp{f934}, @samp{hypersparc}, @samp{sparclite86x}, | |
12214 | @samp{sparclet}, @samp{tsc701}, @samp{v9}, @samp{ultrasparc}, and | |
12215 | @samp{ultrasparc3}. | |
70899148 | 12216 | |
39bc1876 NS |
12217 | Default instruction scheduling parameters are used for values that select |
12218 | an architecture and not an implementation. These are @samp{v7}, @samp{v8}, | |
12219 | @samp{sparclite}, @samp{sparclet}, @samp{v9}. | |
70899148 | 12220 | |
39bc1876 NS |
12221 | Here is a list of each supported architecture and their supported |
12222 | implementations. | |
70899148 | 12223 | |
39bc1876 NS |
12224 | @smallexample |
12225 | v7: cypress | |
12226 | v8: supersparc, hypersparc | |
12227 | sparclite: f930, f934, sparclite86x | |
12228 | sparclet: tsc701 | |
12229 | v9: ultrasparc, ultrasparc3 | |
12230 | @end smallexample | |
70899148 | 12231 | |
39bc1876 NS |
12232 | By default (unless configured otherwise), GCC generates code for the V7 |
12233 | variant of the SPARC architecture. With @option{-mcpu=cypress}, the compiler | |
12234 | additionally optimizes it for the Cypress CY7C602 chip, as used in the | |
12235 | SPARCStation/SPARCServer 3xx series. This is also appropriate for the older | |
12236 | SPARCStation 1, 2, IPX etc. | |
70899148 | 12237 | |
39bc1876 NS |
12238 | With @option{-mcpu=v8}, GCC generates code for the V8 variant of the SPARC |
12239 | architecture. The only difference from V7 code is that the compiler emits | |
12240 | the integer multiply and integer divide instructions which exist in SPARC-V8 | |
12241 | but not in SPARC-V7. With @option{-mcpu=supersparc}, the compiler additionally | |
12242 | optimizes it for the SuperSPARC chip, as used in the SPARCStation 10, 1000 and | |
12243 | 2000 series. | |
70899148 | 12244 | |
39bc1876 NS |
12245 | With @option{-mcpu=sparclite}, GCC generates code for the SPARClite variant of |
12246 | the SPARC architecture. This adds the integer multiply, integer divide step | |
12247 | and scan (@code{ffs}) instructions which exist in SPARClite but not in SPARC-V7. | |
12248 | With @option{-mcpu=f930}, the compiler additionally optimizes it for the | |
8a36672b | 12249 | Fujitsu MB86930 chip, which is the original SPARClite, with no FPU@. With |
39bc1876 | 12250 | @option{-mcpu=f934}, the compiler additionally optimizes it for the Fujitsu |
8a36672b | 12251 | MB86934 chip, which is the more recent SPARClite with FPU@. |
70899148 | 12252 | |
39bc1876 NS |
12253 | With @option{-mcpu=sparclet}, GCC generates code for the SPARClet variant of |
12254 | the SPARC architecture. This adds the integer multiply, multiply/accumulate, | |
12255 | integer divide step and scan (@code{ffs}) instructions which exist in SPARClet | |
12256 | but not in SPARC-V7. With @option{-mcpu=tsc701}, the compiler additionally | |
12257 | optimizes it for the TEMIC SPARClet chip. | |
70899148 | 12258 | |
39bc1876 NS |
12259 | With @option{-mcpu=v9}, GCC generates code for the V9 variant of the SPARC |
12260 | architecture. This adds 64-bit integer and floating-point move instructions, | |
12261 | 3 additional floating-point condition code registers and conditional move | |
12262 | instructions. With @option{-mcpu=ultrasparc}, the compiler additionally | |
12263 | optimizes it for the Sun UltraSPARC I/II chips. With | |
12264 | @option{-mcpu=ultrasparc3}, the compiler additionally optimizes it for the | |
12265 | Sun UltraSPARC III chip. | |
70899148 | 12266 | |
39bc1876 NS |
12267 | @item -mtune=@var{cpu_type} |
12268 | @opindex mtune | |
12269 | Set the instruction scheduling parameters for machine type | |
12270 | @var{cpu_type}, but do not set the instruction set or register set that the | |
12271 | option @option{-mcpu=@var{cpu_type}} would. | |
70899148 | 12272 | |
39bc1876 NS |
12273 | The same values for @option{-mcpu=@var{cpu_type}} can be used for |
12274 | @option{-mtune=@var{cpu_type}}, but the only useful values are those | |
12275 | that select a particular cpu implementation. Those are @samp{cypress}, | |
12276 | @samp{supersparc}, @samp{hypersparc}, @samp{f930}, @samp{f934}, | |
12277 | @samp{sparclite86x}, @samp{tsc701}, @samp{ultrasparc}, and | |
12278 | @samp{ultrasparc3}. | |
70899148 | 12279 | |
39bc1876 NS |
12280 | @item -mv8plus |
12281 | @itemx -mno-v8plus | |
12282 | @opindex mv8plus | |
12283 | @opindex mno-v8plus | |
8a36672b | 12284 | With @option{-mv8plus}, GCC generates code for the SPARC-V8+ ABI@. The |
39bc1876 NS |
12285 | difference from the V8 ABI is that the global and out registers are |
12286 | considered 64-bit wide. This is enabled by default on Solaris in 32-bit | |
12287 | mode for all SPARC-V9 processors. | |
70899148 | 12288 | |
39bc1876 NS |
12289 | @item -mvis |
12290 | @itemx -mno-vis | |
12291 | @opindex mvis | |
12292 | @opindex mno-vis | |
12293 | With @option{-mvis}, GCC generates code that takes advantage of the UltraSPARC | |
12294 | Visual Instruction Set extensions. The default is @option{-mno-vis}. | |
12295 | @end table | |
70899148 | 12296 | |
39bc1876 NS |
12297 | These @samp{-m} options are supported in addition to the above |
12298 | on SPARC-V9 processors in 64-bit environments: | |
70899148 | 12299 | |
39bc1876 NS |
12300 | @table @gcctabopt |
12301 | @item -mlittle-endian | |
12302 | @opindex mlittle-endian | |
8a36672b | 12303 | Generate code for a processor running in little-endian mode. It is only |
d4e1591f | 12304 | available for a few configurations and most notably not on Solaris and Linux. |
70899148 | 12305 | |
39bc1876 NS |
12306 | @item -m32 |
12307 | @itemx -m64 | |
12308 | @opindex m32 | |
12309 | @opindex m64 | |
12310 | Generate code for a 32-bit or 64-bit environment. | |
12311 | The 32-bit environment sets int, long and pointer to 32 bits. | |
12312 | The 64-bit environment sets int to 32 bits and long and pointer | |
12313 | to 64 bits. | |
70899148 | 12314 | |
39bc1876 NS |
12315 | @item -mcmodel=medlow |
12316 | @opindex mcmodel=medlow | |
12317 | Generate code for the Medium/Low code model: 64-bit addresses, programs | |
12318 | must be linked in the low 32 bits of memory. Programs can be statically | |
12319 | or dynamically linked. | |
70899148 | 12320 | |
39bc1876 NS |
12321 | @item -mcmodel=medmid |
12322 | @opindex mcmodel=medmid | |
12323 | Generate code for the Medium/Middle code model: 64-bit addresses, programs | |
12324 | must be linked in the low 44 bits of memory, the text and data segments must | |
12325 | be less than 2GB in size and the data segment must be located within 2GB of | |
12326 | the text segment. | |
70899148 | 12327 | |
39bc1876 NS |
12328 | @item -mcmodel=medany |
12329 | @opindex mcmodel=medany | |
12330 | Generate code for the Medium/Anywhere code model: 64-bit addresses, programs | |
12331 | may be linked anywhere in memory, the text and data segments must be less | |
12332 | than 2GB in size and the data segment must be located within 2GB of the | |
12333 | text segment. | |
70899148 | 12334 | |
39bc1876 NS |
12335 | @item -mcmodel=embmedany |
12336 | @opindex mcmodel=embmedany | |
12337 | Generate code for the Medium/Anywhere code model for embedded systems: | |
12338 | 64-bit addresses, the text and data segments must be less than 2GB in | |
12339 | size, both starting anywhere in memory (determined at link time). The | |
12340 | global register %g4 points to the base of the data segment. Programs | |
12341 | are statically linked and PIC is not supported. | |
70899148 | 12342 | |
39bc1876 NS |
12343 | @item -mstack-bias |
12344 | @itemx -mno-stack-bias | |
12345 | @opindex mstack-bias | |
12346 | @opindex mno-stack-bias | |
12347 | With @option{-mstack-bias}, GCC assumes that the stack pointer, and | |
12348 | frame pointer if present, are offset by @minus{}2047 which must be added back | |
12349 | when making stack frame references. This is the default in 64-bit mode. | |
12350 | Otherwise, assume no such offset is present. | |
12351 | @end table | |
70899148 | 12352 | |
6bfb2f93 EB |
12353 | These switches are supported in addition to the above on Solaris: |
12354 | ||
12355 | @table @gcctabopt | |
12356 | @item -threads | |
12357 | @opindex threads | |
12358 | Add support for multithreading using the Solaris threads library. This | |
12359 | option sets flags for both the preprocessor and linker. This option does | |
12360 | not affect the thread safety of object code produced by the compiler or | |
12361 | that of libraries supplied with it. | |
12362 | ||
12363 | @item -pthreads | |
12364 | @opindex pthreads | |
12365 | Add support for multithreading using the POSIX threads library. This | |
12366 | option sets flags for both the preprocessor and linker. This option does | |
12367 | not affect the thread safety of object code produced by the compiler or | |
12368 | that of libraries supplied with it. | |
12369 | @end table | |
12370 | ||
39bc1876 NS |
12371 | @node System V Options |
12372 | @subsection Options for System V | |
70899148 | 12373 | |
39bc1876 NS |
12374 | These additional options are available on System V Release 4 for |
12375 | compatibility with other compilers on those systems: | |
70899148 | 12376 | |
39bc1876 NS |
12377 | @table @gcctabopt |
12378 | @item -G | |
12379 | @opindex G | |
12380 | Create a shared object. | |
12381 | It is recommended that @option{-symbolic} or @option{-shared} be used instead. | |
70899148 | 12382 | |
39bc1876 NS |
12383 | @item -Qy |
12384 | @opindex Qy | |
12385 | Identify the versions of each tool used by the compiler, in a | |
12386 | @code{.ident} assembler directive in the output. | |
70899148 | 12387 | |
39bc1876 NS |
12388 | @item -Qn |
12389 | @opindex Qn | |
12390 | Refrain from adding @code{.ident} directives to the output file (this is | |
12391 | the default). | |
70899148 | 12392 | |
39bc1876 NS |
12393 | @item -YP,@var{dirs} |
12394 | @opindex YP | |
12395 | Search the directories @var{dirs}, and no others, for libraries | |
12396 | specified with @option{-l}. | |
70899148 | 12397 | |
39bc1876 NS |
12398 | @item -Ym,@var{dir} |
12399 | @opindex Ym | |
12400 | Look in the directory @var{dir} to find the M4 preprocessor. | |
12401 | The assembler uses this option. | |
12402 | @c This is supposed to go with a -Yd for predefined M4 macro files, but | |
12403 | @c the generic assembler that comes with Solaris takes just -Ym. | |
12404 | @end table | |
70899148 | 12405 | |
39bc1876 NS |
12406 | @node TMS320C3x/C4x Options |
12407 | @subsection TMS320C3x/C4x Options | |
12408 | @cindex TMS320C3x/C4x Options | |
70899148 | 12409 | |
39bc1876 | 12410 | These @samp{-m} options are defined for TMS320C3x/C4x implementations: |
70899148 | 12411 | |
39bc1876 | 12412 | @table @gcctabopt |
70899148 | 12413 | |
39bc1876 NS |
12414 | @item -mcpu=@var{cpu_type} |
12415 | @opindex mcpu | |
12416 | Set the instruction set, register set, and instruction scheduling | |
12417 | parameters for machine type @var{cpu_type}. Supported values for | |
12418 | @var{cpu_type} are @samp{c30}, @samp{c31}, @samp{c32}, @samp{c40}, and | |
12419 | @samp{c44}. The default is @samp{c40} to generate code for the | |
12420 | TMS320C40. | |
70899148 | 12421 | |
39bc1876 NS |
12422 | @item -mbig-memory |
12423 | @itemx -mbig | |
12424 | @itemx -msmall-memory | |
12425 | @itemx -msmall | |
12426 | @opindex mbig-memory | |
12427 | @opindex mbig | |
12428 | @opindex msmall-memory | |
12429 | @opindex msmall | |
12430 | Generates code for the big or small memory model. The small memory | |
12431 | model assumed that all data fits into one 64K word page. At run-time | |
12432 | the data page (DP) register must be set to point to the 64K page | |
12433 | containing the .bss and .data program sections. The big memory model is | |
12434 | the default and requires reloading of the DP register for every direct | |
12435 | memory access. | |
70899148 | 12436 | |
39bc1876 NS |
12437 | @item -mbk |
12438 | @itemx -mno-bk | |
12439 | @opindex mbk | |
12440 | @opindex mno-bk | |
12441 | Allow (disallow) allocation of general integer operands into the block | |
12442 | count register BK@. | |
70899148 | 12443 | |
39bc1876 NS |
12444 | @item -mdb |
12445 | @itemx -mno-db | |
12446 | @opindex mdb | |
12447 | @opindex mno-db | |
12448 | Enable (disable) generation of code using decrement and branch, | |
12449 | DBcond(D), instructions. This is enabled by default for the C4x. To be | |
12450 | on the safe side, this is disabled for the C3x, since the maximum | |
12451 | iteration count on the C3x is @math{2^{23} + 1} (but who iterates loops more than | |
12452 | @math{2^{23}} times on the C3x?). Note that GCC will try to reverse a loop so | |
12453 | that it can utilize the decrement and branch instruction, but will give | |
12454 | up if there is more than one memory reference in the loop. Thus a loop | |
12455 | where the loop counter is decremented can generate slightly more | |
12456 | efficient code, in cases where the RPTB instruction cannot be utilized. | |
70899148 | 12457 | |
39bc1876 NS |
12458 | @item -mdp-isr-reload |
12459 | @itemx -mparanoid | |
12460 | @opindex mdp-isr-reload | |
12461 | @opindex mparanoid | |
12462 | Force the DP register to be saved on entry to an interrupt service | |
12463 | routine (ISR), reloaded to point to the data section, and restored on | |
12464 | exit from the ISR@. This should not be required unless someone has | |
12465 | violated the small memory model by modifying the DP register, say within | |
12466 | an object library. | |
70899148 | 12467 | |
39bc1876 NS |
12468 | @item -mmpyi |
12469 | @itemx -mno-mpyi | |
12470 | @opindex mmpyi | |
12471 | @opindex mno-mpyi | |
12472 | For the C3x use the 24-bit MPYI instruction for integer multiplies | |
12473 | instead of a library call to guarantee 32-bit results. Note that if one | |
12474 | of the operands is a constant, then the multiplication will be performed | |
12475 | using shifts and adds. If the @option{-mmpyi} option is not specified for the C3x, | |
12476 | then squaring operations are performed inline instead of a library call. | |
70899148 | 12477 | |
39bc1876 NS |
12478 | @item -mfast-fix |
12479 | @itemx -mno-fast-fix | |
12480 | @opindex mfast-fix | |
12481 | @opindex mno-fast-fix | |
12482 | The C3x/C4x FIX instruction to convert a floating point value to an | |
12483 | integer value chooses the nearest integer less than or equal to the | |
12484 | floating point value rather than to the nearest integer. Thus if the | |
12485 | floating point number is negative, the result will be incorrectly | |
12486 | truncated an additional code is necessary to detect and correct this | |
12487 | case. This option can be used to disable generation of the additional | |
12488 | code required to correct the result. | |
70899148 | 12489 | |
39bc1876 NS |
12490 | @item -mrptb |
12491 | @itemx -mno-rptb | |
12492 | @opindex mrptb | |
12493 | @opindex mno-rptb | |
12494 | Enable (disable) generation of repeat block sequences using the RPTB | |
12495 | instruction for zero overhead looping. The RPTB construct is only used | |
12496 | for innermost loops that do not call functions or jump across the loop | |
12497 | boundaries. There is no advantage having nested RPTB loops due to the | |
12498 | overhead required to save and restore the RC, RS, and RE registers. | |
12499 | This is enabled by default with @option{-O2}. | |
70899148 | 12500 | |
39bc1876 NS |
12501 | @item -mrpts=@var{count} |
12502 | @itemx -mno-rpts | |
12503 | @opindex mrpts | |
12504 | @opindex mno-rpts | |
12505 | Enable (disable) the use of the single instruction repeat instruction | |
12506 | RPTS@. If a repeat block contains a single instruction, and the loop | |
12507 | count can be guaranteed to be less than the value @var{count}, GCC will | |
12508 | emit a RPTS instruction instead of a RPTB@. If no value is specified, | |
12509 | then a RPTS will be emitted even if the loop count cannot be determined | |
12510 | at compile time. Note that the repeated instruction following RPTS does | |
12511 | not have to be reloaded from memory each iteration, thus freeing up the | |
12512 | CPU buses for operands. However, since interrupts are blocked by this | |
12513 | instruction, it is disabled by default. | |
70899148 | 12514 | |
39bc1876 NS |
12515 | @item -mloop-unsigned |
12516 | @itemx -mno-loop-unsigned | |
12517 | @opindex mloop-unsigned | |
12518 | @opindex mno-loop-unsigned | |
12519 | The maximum iteration count when using RPTS and RPTB (and DB on the C40) | |
12520 | is @math{2^{31} + 1} since these instructions test if the iteration count is | |
12521 | negative to terminate the loop. If the iteration count is unsigned | |
12522 | there is a possibility than the @math{2^{31} + 1} maximum iteration count may be | |
12523 | exceeded. This switch allows an unsigned iteration count. | |
70899148 | 12524 | |
39bc1876 NS |
12525 | @item -mti |
12526 | @opindex mti | |
12527 | Try to emit an assembler syntax that the TI assembler (asm30) is happy | |
12528 | with. This also enforces compatibility with the API employed by the TI | |
12529 | C3x C compiler. For example, long doubles are passed as structures | |
12530 | rather than in floating point registers. | |
70899148 | 12531 | |
39bc1876 NS |
12532 | @item -mregparm |
12533 | @itemx -mmemparm | |
12534 | @opindex mregparm | |
12535 | @opindex mmemparm | |
12536 | Generate code that uses registers (stack) for passing arguments to functions. | |
12537 | By default, arguments are passed in registers where possible rather | |
12538 | than by pushing arguments on to the stack. | |
70899148 | 12539 | |
39bc1876 NS |
12540 | @item -mparallel-insns |
12541 | @itemx -mno-parallel-insns | |
12542 | @opindex mparallel-insns | |
12543 | @opindex mno-parallel-insns | |
12544 | Allow the generation of parallel instructions. This is enabled by | |
12545 | default with @option{-O2}. | |
70899148 | 12546 | |
39bc1876 NS |
12547 | @item -mparallel-mpy |
12548 | @itemx -mno-parallel-mpy | |
12549 | @opindex mparallel-mpy | |
12550 | @opindex mno-parallel-mpy | |
12551 | Allow the generation of MPY||ADD and MPY||SUB parallel instructions, | |
12552 | provided @option{-mparallel-insns} is also specified. These instructions have | |
12553 | tight register constraints which can pessimize the code generation | |
12554 | of large functions. | |
70899148 | 12555 | |
39bc1876 | 12556 | @end table |
70899148 | 12557 | |
39bc1876 NS |
12558 | @node V850 Options |
12559 | @subsection V850 Options | |
12560 | @cindex V850 Options | |
70899148 | 12561 | |
39bc1876 | 12562 | These @samp{-m} options are defined for V850 implementations: |
70899148 | 12563 | |
39bc1876 NS |
12564 | @table @gcctabopt |
12565 | @item -mlong-calls | |
12566 | @itemx -mno-long-calls | |
12567 | @opindex mlong-calls | |
12568 | @opindex mno-long-calls | |
12569 | Treat all calls as being far away (near). If calls are assumed to be | |
12570 | far away, the compiler will always load the functions address up into a | |
12571 | register, and call indirect through the pointer. | |
70899148 | 12572 | |
39bc1876 NS |
12573 | @item -mno-ep |
12574 | @itemx -mep | |
12575 | @opindex mno-ep | |
12576 | @opindex mep | |
12577 | Do not optimize (do optimize) basic blocks that use the same index | |
12578 | pointer 4 or more times to copy pointer into the @code{ep} register, and | |
12579 | use the shorter @code{sld} and @code{sst} instructions. The @option{-mep} | |
12580 | option is on by default if you optimize. | |
70899148 | 12581 | |
39bc1876 NS |
12582 | @item -mno-prolog-function |
12583 | @itemx -mprolog-function | |
12584 | @opindex mno-prolog-function | |
12585 | @opindex mprolog-function | |
12586 | Do not use (do use) external functions to save and restore registers | |
12587 | at the prologue and epilogue of a function. The external functions | |
12588 | are slower, but use less code space if more than one function saves | |
12589 | the same number of registers. The @option{-mprolog-function} option | |
12590 | is on by default if you optimize. | |
70899148 | 12591 | |
39bc1876 NS |
12592 | @item -mspace |
12593 | @opindex mspace | |
12594 | Try to make the code as small as possible. At present, this just turns | |
12595 | on the @option{-mep} and @option{-mprolog-function} options. | |
70899148 | 12596 | |
39bc1876 NS |
12597 | @item -mtda=@var{n} |
12598 | @opindex mtda | |
12599 | Put static or global variables whose size is @var{n} bytes or less into | |
12600 | the tiny data area that register @code{ep} points to. The tiny data | |
12601 | area can hold up to 256 bytes in total (128 bytes for byte references). | |
70899148 | 12602 | |
39bc1876 NS |
12603 | @item -msda=@var{n} |
12604 | @opindex msda | |
12605 | Put static or global variables whose size is @var{n} bytes or less into | |
12606 | the small data area that register @code{gp} points to. The small data | |
12607 | area can hold up to 64 kilobytes. | |
70899148 | 12608 | |
39bc1876 NS |
12609 | @item -mzda=@var{n} |
12610 | @opindex mzda | |
12611 | Put static or global variables whose size is @var{n} bytes or less into | |
12612 | the first 32 kilobytes of memory. | |
70899148 | 12613 | |
39bc1876 NS |
12614 | @item -mv850 |
12615 | @opindex mv850 | |
12616 | Specify that the target processor is the V850. | |
70899148 | 12617 | |
39bc1876 NS |
12618 | @item -mbig-switch |
12619 | @opindex mbig-switch | |
12620 | Generate code suitable for big switch tables. Use this option only if | |
12621 | the assembler/linker complain about out of range branches within a switch | |
12622 | table. | |
70899148 | 12623 | |
39bc1876 NS |
12624 | @item -mapp-regs |
12625 | @opindex mapp-regs | |
12626 | This option will cause r2 and r5 to be used in the code generated by | |
12627 | the compiler. This setting is the default. | |
70899148 | 12628 | |
39bc1876 NS |
12629 | @item -mno-app-regs |
12630 | @opindex mno-app-regs | |
12631 | This option will cause r2 and r5 to be treated as fixed registers. | |
70899148 | 12632 | |
39bc1876 NS |
12633 | @item -mv850e1 |
12634 | @opindex mv850e1 | |
12635 | Specify that the target processor is the V850E1. The preprocessor | |
12636 | constants @samp{__v850e1__} and @samp{__v850e__} will be defined if | |
12637 | this option is used. | |
70899148 | 12638 | |
39bc1876 NS |
12639 | @item -mv850e |
12640 | @opindex mv850e | |
8a36672b | 12641 | Specify that the target processor is the V850E@. The preprocessor |
39bc1876 | 12642 | constant @samp{__v850e__} will be defined if this option is used. |
70899148 | 12643 | |
39bc1876 NS |
12644 | If neither @option{-mv850} nor @option{-mv850e} nor @option{-mv850e1} |
12645 | are defined then a default target processor will be chosen and the | |
12646 | relevant @samp{__v850*__} preprocessor constant will be defined. | |
70899148 | 12647 | |
39bc1876 NS |
12648 | The preprocessor constants @samp{__v850} and @samp{__v851__} are always |
12649 | defined, regardless of which processor variant is the target. | |
70899148 | 12650 | |
39bc1876 NS |
12651 | @item -mdisable-callt |
12652 | @opindex mdisable-callt | |
12653 | This option will suppress generation of the CALLT instruction for the | |
12654 | v850e and v850e1 flavors of the v850 architecture. The default is | |
12655 | @option{-mno-disable-callt} which allows the CALLT instruction to be used. | |
70899148 | 12656 | |
39bc1876 | 12657 | @end table |
70899148 | 12658 | |
39bc1876 NS |
12659 | @node VAX Options |
12660 | @subsection VAX Options | |
12661 | @cindex VAX options | |
70899148 | 12662 | |
39bc1876 | 12663 | These @samp{-m} options are defined for the VAX: |
70899148 | 12664 | |
39bc1876 NS |
12665 | @table @gcctabopt |
12666 | @item -munix | |
12667 | @opindex munix | |
12668 | Do not output certain jump instructions (@code{aobleq} and so on) | |
12669 | that the Unix assembler for the VAX cannot handle across long | |
12670 | ranges. | |
70899148 | 12671 | |
39bc1876 NS |
12672 | @item -mgnu |
12673 | @opindex mgnu | |
12674 | Do output those jump instructions, on the assumption that you | |
12675 | will assemble with the GNU assembler. | |
70899148 | 12676 | |
39bc1876 NS |
12677 | @item -mg |
12678 | @opindex mg | |
12679 | Output code for g-format floating point numbers instead of d-format. | |
12680 | @end table | |
70899148 | 12681 | |
39bc1876 NS |
12682 | @node x86-64 Options |
12683 | @subsection x86-64 Options | |
12684 | @cindex x86-64 options | |
70899148 | 12685 | |
39bc1876 | 12686 | These are listed under @xref{i386 and x86-64 Options}. |
70899148 | 12687 | |
39bc1876 NS |
12688 | @node Xstormy16 Options |
12689 | @subsection Xstormy16 Options | |
12690 | @cindex Xstormy16 Options | |
70899148 | 12691 | |
39bc1876 | 12692 | These options are defined for Xstormy16: |
70899148 | 12693 | |
39bc1876 NS |
12694 | @table @gcctabopt |
12695 | @item -msim | |
12696 | @opindex msim | |
12697 | Choose startup files and linker script suitable for the simulator. | |
70899148 BS |
12698 | @end table |
12699 | ||
03984308 BW |
12700 | @node Xtensa Options |
12701 | @subsection Xtensa Options | |
12702 | @cindex Xtensa Options | |
12703 | ||
6cedbe44 | 12704 | These options are supported for Xtensa targets: |
03984308 BW |
12705 | |
12706 | @table @gcctabopt | |
f42f5a1b BW |
12707 | @item -mconst16 |
12708 | @itemx -mno-const16 | |
12709 | @opindex mconst16 | |
12710 | @opindex mno-const16 | |
6c2e8d1c BW |
12711 | Enable or disable use of @code{CONST16} instructions for loading |
12712 | constant values. The @code{CONST16} instruction is currently not a | |
12713 | standard option from Tensilica. When enabled, @code{CONST16} | |
12714 | instructions are always used in place of the standard @code{L32R} | |
12715 | instructions. The use of @code{CONST16} is enabled by default only if | |
12716 | the @code{L32R} instruction is not available. | |
12717 | ||
03984308 BW |
12718 | @item -mfused-madd |
12719 | @itemx -mno-fused-madd | |
12720 | @opindex mfused-madd | |
12721 | @opindex mno-fused-madd | |
12722 | Enable or disable use of fused multiply/add and multiply/subtract | |
12723 | instructions in the floating-point option. This has no effect if the | |
12724 | floating-point option is not also enabled. Disabling fused multiply/add | |
12725 | and multiply/subtract instructions forces the compiler to use separate | |
12726 | instructions for the multiply and add/subtract operations. This may be | |
12727 | desirable in some cases where strict IEEE 754-compliant results are | |
12728 | required: the fused multiply add/subtract instructions do not round the | |
12729 | intermediate result, thereby producing results with @emph{more} bits of | |
12730 | precision than specified by the IEEE standard. Disabling fused multiply | |
12731 | add/subtract instructions also ensures that the program output is not | |
12732 | sensitive to the compiler's ability to combine multiply and add/subtract | |
12733 | operations. | |
12734 | ||
03984308 BW |
12735 | @item -mtext-section-literals |
12736 | @itemx -mno-text-section-literals | |
12737 | @opindex mtext-section-literals | |
12738 | @opindex mno-text-section-literals | |
12739 | Control the treatment of literal pools. The default is | |
12740 | @option{-mno-text-section-literals}, which places literals in a separate | |
12741 | section in the output file. This allows the literal pool to be placed | |
12742 | in a data RAM/ROM, and it also allows the linker to combine literal | |
12743 | pools from separate object files to remove redundant literals and | |
12744 | improve code size. With @option{-mtext-section-literals}, the literals | |
12745 | are interspersed in the text section in order to keep them as close as | |
12746 | possible to their references. This may be necessary for large assembly | |
12747 | files. | |
12748 | ||
12749 | @item -mtarget-align | |
12750 | @itemx -mno-target-align | |
12751 | @opindex mtarget-align | |
12752 | @opindex mno-target-align | |
12753 | When this option is enabled, GCC instructs the assembler to | |
12754 | automatically align instructions to reduce branch penalties at the | |
12755 | expense of some code density. The assembler attempts to widen density | |
12756 | instructions to align branch targets and the instructions following call | |
12757 | instructions. If there are not enough preceding safe density | |
12758 | instructions to align a target, no widening will be performed. The | |
12759 | default is @option{-mtarget-align}. These options do not affect the | |
12760 | treatment of auto-aligned instructions like @code{LOOP}, which the | |
12761 | assembler will always align, either by widening density instructions or | |
12762 | by inserting no-op instructions. | |
12763 | ||
12764 | @item -mlongcalls | |
12765 | @itemx -mno-longcalls | |
12766 | @opindex mlongcalls | |
12767 | @opindex mno-longcalls | |
12768 | When this option is enabled, GCC instructs the assembler to translate | |
12769 | direct calls to indirect calls unless it can determine that the target | |
12770 | of a direct call is in the range allowed by the call instruction. This | |
12771 | translation typically occurs for calls to functions in other source | |
12772 | files. Specifically, the assembler translates a direct @code{CALL} | |
12773 | instruction into an @code{L32R} followed by a @code{CALLX} instruction. | |
12774 | The default is @option{-mno-longcalls}. This option should be used in | |
12775 | programs where the call target can potentially be out of range. This | |
12776 | option is implemented in the assembler, not the compiler, so the | |
12777 | assembly code generated by GCC will still show direct call | |
12778 | instructions---look at the disassembled object code to see the actual | |
12779 | instructions. Note that the assembler will use an indirect call for | |
12780 | every cross-file call, not just those that really will be out of range. | |
12781 | @end table | |
12782 | ||
39bc1876 NS |
12783 | @node zSeries Options |
12784 | @subsection zSeries Options | |
12785 | @cindex zSeries options | |
12786 | ||
12787 | These are listed under @xref{S/390 and zSeries Options}. | |
12788 | ||
74291a4b MM |
12789 | @node Code Gen Options |
12790 | @section Options for Code Generation Conventions | |
12791 | @cindex code generation conventions | |
12792 | @cindex options, code generation | |
12793 | @cindex run-time options | |
12794 | ||
12795 | These machine-independent options control the interface conventions | |
12796 | used in code generation. | |
12797 | ||
12798 | Most of them have both positive and negative forms; the negative form | |
630d3d5a | 12799 | of @option{-ffoo} would be @option{-fno-foo}. In the table below, only |
74291a4b MM |
12800 | one of the forms is listed---the one which is not the default. You |
12801 | can figure out the other form by either removing @samp{no-} or adding | |
12802 | it. | |
12803 | ||
2642624b | 12804 | @table @gcctabopt |
d4463dfc JQ |
12805 | @item -fbounds-check |
12806 | @opindex fbounds-check | |
12807 | For front-ends that support it, generate additional code to check that | |
12808 | indices used to access arrays are within the declared range. This is | |
12809 | currently only supported by the Java and Fortran 77 front-ends, where | |
12810 | this option defaults to true and false respectively. | |
12811 | ||
12812 | @item -ftrapv | |
12813 | @opindex ftrapv | |
12814 | This option generates traps for signed overflow on addition, subtraction, | |
12815 | multiplication operations. | |
12816 | ||
4fa26a60 RS |
12817 | @item -fwrapv |
12818 | @opindex fwrapv | |
12819 | This option instructs the compiler to assume that signed arithmetic | |
12820 | overflow of addition, subtraction and multiplication wraps around | |
c0cbdbd9 | 12821 | using twos-complement representation. This flag enables some optimizations |
aa58883c | 12822 | and disables others. This option is enabled by default for the Java |
4fa26a60 RS |
12823 | front-end, as required by the Java language specification. |
12824 | ||
956d6950 | 12825 | @item -fexceptions |
cd3bb277 | 12826 | @opindex fexceptions |
767094dd | 12827 | Enable exception handling. Generates extra code needed to propagate |
f0523f02 | 12828 | exceptions. For some targets, this implies GCC will generate frame |
c5c76735 JL |
12829 | unwind information for all functions, which can produce significant data |
12830 | size overhead, although it does not affect execution. If you do not | |
f0523f02 | 12831 | specify this option, GCC will enable it by default for languages like |
90ecce3e | 12832 | C++ which normally require exception handling, and disable it for |
c5c76735 JL |
12833 | languages like C that do not normally require it. However, you may need |
12834 | to enable this option when compiling C code that needs to interoperate | |
12835 | properly with exception handlers written in C++. You may also wish to | |
12836 | disable this option if you are compiling older C++ programs that don't | |
12837 | use exception handling. | |
956d6950 | 12838 | |
6cfc0341 RH |
12839 | @item -fnon-call-exceptions |
12840 | @opindex fnon-call-exceptions | |
12841 | Generate code that allows trapping instructions to throw exceptions. | |
12842 | Note that this requires platform-specific runtime support that does | |
12843 | not exist everywhere. Moreover, it only allows @emph{trapping} | |
e979f9e8 | 12844 | instructions to throw exceptions, i.e.@: memory references or floating |
6cfc0341 RH |
12845 | point instructions. It does not allow exceptions to be thrown from |
12846 | arbitrary signal handlers such as @code{SIGALRM}. | |
12847 | ||
14a774a9 | 12848 | @item -funwind-tables |
cd3bb277 | 12849 | @opindex funwind-tables |
bedc7537 | 12850 | Similar to @option{-fexceptions}, except that it will just generate any needed |
14a774a9 RK |
12851 | static data, but will not affect the generated code in any other way. |
12852 | You will normally not enable this option; instead, a language processor | |
12853 | that needs this handling would enable it on your behalf. | |
12854 | ||
b932f770 | 12855 | @item -fasynchronous-unwind-tables |
2a9dc917 | 12856 | @opindex fasynchronous-unwind-tables |
b932f770 JH |
12857 | Generate unwind table in dwarf2 format, if supported by target machine. The |
12858 | table is exact at each instruction boundary, so it can be used for stack | |
12859 | unwinding from asynchronous events (such as debugger or garbage collector). | |
12860 | ||
74291a4b | 12861 | @item -fpcc-struct-return |
cd3bb277 | 12862 | @opindex fpcc-struct-return |
74291a4b MM |
12863 | Return ``short'' @code{struct} and @code{union} values in memory like |
12864 | longer ones, rather than in registers. This convention is less | |
12865 | efficient, but it has the advantage of allowing intercallability between | |
a9c60612 JJ |
12866 | GCC-compiled files and files compiled with other compilers, particularly |
12867 | the Portable C Compiler (pcc). | |
74291a4b MM |
12868 | |
12869 | The precise convention for returning structures in memory depends | |
12870 | on the target configuration macros. | |
12871 | ||
12872 | Short structures and unions are those whose size and alignment match | |
12873 | that of some integer type. | |
12874 | ||
a9c60612 JJ |
12875 | @strong{Warning:} code compiled with the @option{-fpcc-struct-return} |
12876 | switch is not binary compatible with code compiled with the | |
12877 | @option{-freg-struct-return} switch. | |
12878 | Use it to conform to a non-default application binary interface. | |
12879 | ||
74291a4b | 12880 | @item -freg-struct-return |
cd3bb277 | 12881 | @opindex freg-struct-return |
9c34dbbf ZW |
12882 | Return @code{struct} and @code{union} values in registers when possible. |
12883 | This is more efficient for small structures than | |
12884 | @option{-fpcc-struct-return}. | |
74291a4b | 12885 | |
9c34dbbf | 12886 | If you specify neither @option{-fpcc-struct-return} nor |
630d3d5a | 12887 | @option{-freg-struct-return}, GCC defaults to whichever convention is |
0c2d1a2a | 12888 | standard for the target. If there is no standard convention, GCC |
9c34dbbf ZW |
12889 | defaults to @option{-fpcc-struct-return}, except on targets where GCC is |
12890 | the principal compiler. In those cases, we can choose the standard, and | |
12891 | we chose the more efficient register return alternative. | |
74291a4b | 12892 | |
a9c60612 JJ |
12893 | @strong{Warning:} code compiled with the @option{-freg-struct-return} |
12894 | switch is not binary compatible with code compiled with the | |
12895 | @option{-fpcc-struct-return} switch. | |
12896 | Use it to conform to a non-default application binary interface. | |
12897 | ||
74291a4b | 12898 | @item -fshort-enums |
cd3bb277 | 12899 | @opindex fshort-enums |
74291a4b MM |
12900 | Allocate to an @code{enum} type only as many bytes as it needs for the |
12901 | declared range of possible values. Specifically, the @code{enum} type | |
12902 | will be equivalent to the smallest integer type which has enough room. | |
12903 | ||
a9c60612 JJ |
12904 | @strong{Warning:} the @option{-fshort-enums} switch causes GCC to generate |
12905 | code that is not binary compatible with code generated without that switch. | |
12906 | Use it to conform to a non-default application binary interface. | |
12907 | ||
74291a4b | 12908 | @item -fshort-double |
cd3bb277 | 12909 | @opindex fshort-double |
74291a4b MM |
12910 | Use the same size for @code{double} as for @code{float}. |
12911 | ||
a9c60612 JJ |
12912 | @strong{Warning:} the @option{-fshort-double} switch causes GCC to generate |
12913 | code that is not binary compatible with code generated without that switch. | |
12914 | Use it to conform to a non-default application binary interface. | |
12915 | ||
12916 | @item -fshort-wchar | |
12917 | @opindex fshort-wchar | |
12918 | Override the underlying type for @samp{wchar_t} to be @samp{short | |
12919 | unsigned int} instead of the default for the target. This option is | |
12920 | useful for building programs to run under WINE@. | |
12921 | ||
12922 | @strong{Warning:} the @option{-fshort-wchar} switch causes GCC to generate | |
12923 | code that is not binary compatible with code generated without that switch. | |
12924 | Use it to conform to a non-default application binary interface. | |
12925 | ||
74291a4b | 12926 | @item -fshared-data |
cd3bb277 | 12927 | @opindex fshared-data |
74291a4b MM |
12928 | Requests that the data and non-@code{const} variables of this |
12929 | compilation be shared data rather than private data. The distinction | |
12930 | makes sense only on certain operating systems, where shared data is | |
12931 | shared between processes running the same program, while private data | |
12932 | exists in one copy per process. | |
12933 | ||
12934 | @item -fno-common | |
cd3bb277 | 12935 | @opindex fno-common |
90ecce3e | 12936 | In C, allocate even uninitialized global variables in the data section of the |
74291a4b MM |
12937 | object file, rather than generating them as common blocks. This has the |
12938 | effect that if the same variable is declared (without @code{extern}) in | |
12939 | two different compilations, you will get an error when you link them. | |
12940 | The only reason this might be useful is if you wish to verify that the | |
12941 | program will work on other systems which always work this way. | |
12942 | ||
12943 | @item -fno-ident | |
cd3bb277 | 12944 | @opindex fno-ident |
74291a4b MM |
12945 | Ignore the @samp{#ident} directive. |
12946 | ||
74291a4b | 12947 | @item -finhibit-size-directive |
cd3bb277 | 12948 | @opindex finhibit-size-directive |
74291a4b MM |
12949 | Don't output a @code{.size} assembler directive, or anything else that |
12950 | would cause trouble if the function is split in the middle, and the | |
12951 | two halves are placed at locations far apart in memory. This option is | |
12952 | used when compiling @file{crtstuff.c}; you should not need to use it | |
12953 | for anything else. | |
12954 | ||
12955 | @item -fverbose-asm | |
cd3bb277 | 12956 | @opindex fverbose-asm |
74291a4b MM |
12957 | Put extra commentary information in the generated assembly code to |
12958 | make it more readable. This option is generally only of use to those | |
12959 | who actually need to read the generated assembly code (perhaps while | |
12960 | debugging the compiler itself). | |
12961 | ||
630d3d5a | 12962 | @option{-fno-verbose-asm}, the default, causes the |
74291a4b MM |
12963 | extra information to be omitted and is useful when comparing two assembler |
12964 | files. | |
12965 | ||
74291a4b | 12966 | @item -fpic |
cd3bb277 | 12967 | @opindex fpic |
74291a4b MM |
12968 | @cindex global offset table |
12969 | @cindex PIC | |
12970 | Generate position-independent code (PIC) suitable for use in a shared | |
12971 | library, if supported for the target machine. Such code accesses all | |
161d7b59 | 12972 | constant addresses through a global offset table (GOT)@. The dynamic |
861bb6c1 | 12973 | loader resolves the GOT entries when the program starts (the dynamic |
0c2d1a2a | 12974 | loader is not part of GCC; it is part of the operating system). If |
861bb6c1 JL |
12975 | the GOT size for the linked executable exceeds a machine-specific |
12976 | maximum size, you get an error message from the linker indicating that | |
630d3d5a | 12977 | @option{-fpic} does not work; in that case, recompile with @option{-fPIC} |
b6e69d94 | 12978 | instead. (These maximums are 8k on the SPARC and 32k |
861bb6c1 | 12979 | on the m68k and RS/6000. The 386 has no such limit.) |
74291a4b MM |
12980 | |
12981 | Position-independent code requires special support, and therefore works | |
0c2d1a2a | 12982 | only on certain machines. For the 386, GCC supports PIC for System V |
74291a4b MM |
12983 | but not for the Sun 386i. Code generated for the IBM RS/6000 is always |
12984 | position-independent. | |
12985 | ||
3d119f8f KG |
12986 | When this flag is set, the macros @code{__pic__} and @code{__PIC__} |
12987 | are defined to 1. | |
12988 | ||
74291a4b | 12989 | @item -fPIC |
cd3bb277 | 12990 | @opindex fPIC |
74291a4b MM |
12991 | If supported for the target machine, emit position-independent code, |
12992 | suitable for dynamic linking and avoiding any limit on the size of the | |
068d4f38 BE |
12993 | global offset table. This option makes a difference on the m68k, |
12994 | PowerPC and SPARC@. | |
74291a4b MM |
12995 | |
12996 | Position-independent code requires special support, and therefore works | |
12997 | only on certain machines. | |
12998 | ||
3d119f8f KG |
12999 | When this flag is set, the macros @code{__pic__} and @code{__PIC__} |
13000 | are defined to 2. | |
13001 | ||
24a4dd31 JJ |
13002 | @item -fpie |
13003 | @itemx -fPIE | |
13004 | @opindex fpie | |
13005 | @opindex fPIE | |
13006 | These options are similar to @option{-fpic} and @option{-fPIC}, but | |
13007 | generated position independent code can be only linked into executables. | |
13008 | Usually these options are used when @option{-pie} GCC option will be | |
13009 | used during linking. | |
13010 | ||
82c0180d JM |
13011 | @item -fno-jump-tables |
13012 | @opindex fno-jump-tables | |
13013 | Do not use jump tables for switch statements even where it would be | |
13014 | more efficient than other code generation strategies. This option is | |
13015 | of use in conjunction with @option{-fpic} or @option{-fPIC} for | |
13016 | building code which forms part of a dynamic linker and cannot | |
13017 | reference the address of a jump table. On some targets, jump tables | |
13018 | do not require a GOT and this option is not needed. | |
13019 | ||
74291a4b | 13020 | @item -ffixed-@var{reg} |
cd3bb277 | 13021 | @opindex ffixed |
74291a4b MM |
13022 | Treat the register named @var{reg} as a fixed register; generated code |
13023 | should never refer to it (except perhaps as a stack pointer, frame | |
13024 | pointer or in some other fixed role). | |
13025 | ||
13026 | @var{reg} must be the name of a register. The register names accepted | |
13027 | are machine-specific and are defined in the @code{REGISTER_NAMES} | |
13028 | macro in the machine description macro file. | |
13029 | ||
13030 | This flag does not have a negative form, because it specifies a | |
13031 | three-way choice. | |
13032 | ||
13033 | @item -fcall-used-@var{reg} | |
cd3bb277 | 13034 | @opindex fcall-used |
956d6950 | 13035 | Treat the register named @var{reg} as an allocable register that is |
74291a4b MM |
13036 | clobbered by function calls. It may be allocated for temporaries or |
13037 | variables that do not live across a call. Functions compiled this way | |
13038 | will not save and restore the register @var{reg}. | |
13039 | ||
cb2fdc84 GRK |
13040 | It is an error to used this flag with the frame pointer or stack pointer. |
13041 | Use of this flag for other registers that have fixed pervasive roles in | |
13042 | the machine's execution model will produce disastrous results. | |
74291a4b MM |
13043 | |
13044 | This flag does not have a negative form, because it specifies a | |
13045 | three-way choice. | |
13046 | ||
13047 | @item -fcall-saved-@var{reg} | |
cd3bb277 | 13048 | @opindex fcall-saved |
956d6950 | 13049 | Treat the register named @var{reg} as an allocable register saved by |
74291a4b MM |
13050 | functions. It may be allocated even for temporaries or variables that |
13051 | live across a call. Functions compiled this way will save and restore | |
13052 | the register @var{reg} if they use it. | |
13053 | ||
cb2fdc84 GRK |
13054 | It is an error to used this flag with the frame pointer or stack pointer. |
13055 | Use of this flag for other registers that have fixed pervasive roles in | |
13056 | the machine's execution model will produce disastrous results. | |
74291a4b MM |
13057 | |
13058 | A different sort of disaster will result from the use of this flag for | |
13059 | a register in which function values may be returned. | |
13060 | ||
13061 | This flag does not have a negative form, because it specifies a | |
13062 | three-way choice. | |
13063 | ||
467cecf3 | 13064 | @item -fpack-struct[=@var{n}] |
cd3bb277 | 13065 | @opindex fpack-struct |
467cecf3 | 13066 | Without a value specified, pack all structure members together without |
8a36672b | 13067 | holes. When a value is specified (which must be a small power of two), pack |
467cecf3 JB |
13068 | structure members according to this value, representing the maximum |
13069 | alignment (that is, objects with default alignment requirements larger than | |
13070 | this will be output potentially unaligned at the next fitting location. | |
a9c60612 JJ |
13071 | |
13072 | @strong{Warning:} the @option{-fpack-struct} switch causes GCC to generate | |
13073 | code that is not binary compatible with code generated without that switch. | |
3364c33b | 13074 | Additionally, it makes the code suboptimal. |
a9c60612 | 13075 | Use it to conform to a non-default application binary interface. |
74291a4b | 13076 | |
07417085 | 13077 | @item -finstrument-functions |
cd3bb277 | 13078 | @opindex finstrument-functions |
07417085 KR |
13079 | Generate instrumentation calls for entry and exit to functions. Just |
13080 | after function entry and just before function exit, the following | |
13081 | profiling functions will be called with the address of the current | |
13082 | function and its call site. (On some platforms, | |
13083 | @code{__builtin_return_address} does not work beyond the current | |
13084 | function, so the call site information may not be available to the | |
13085 | profiling functions otherwise.) | |
13086 | ||
3ab51846 | 13087 | @smallexample |
310668e8 JM |
13088 | void __cyg_profile_func_enter (void *this_fn, |
13089 | void *call_site); | |
13090 | void __cyg_profile_func_exit (void *this_fn, | |
13091 | void *call_site); | |
3ab51846 | 13092 | @end smallexample |
07417085 KR |
13093 | |
13094 | The first argument is the address of the start of the current function, | |
13095 | which may be looked up exactly in the symbol table. | |
13096 | ||
13097 | This instrumentation is also done for functions expanded inline in other | |
13098 | functions. The profiling calls will indicate where, conceptually, the | |
13099 | inline function is entered and exited. This means that addressable | |
13100 | versions of such functions must be available. If all your uses of a | |
13101 | function are expanded inline, this may mean an additional expansion of | |
13102 | code size. If you use @samp{extern inline} in your C code, an | |
13103 | addressable version of such functions must be provided. (This is | |
13104 | normally the case anyways, but if you get lucky and the optimizer always | |
13105 | expands the functions inline, you might have gotten away without | |
13106 | providing static copies.) | |
13107 | ||
13108 | A function may be given the attribute @code{no_instrument_function}, in | |
13109 | which case this instrumentation will not be done. This can be used, for | |
13110 | example, for the profiling functions listed above, high-priority | |
13111 | interrupt routines, and any functions from which the profiling functions | |
13112 | cannot safely be called (perhaps signal handlers, if the profiling | |
13113 | routines generate output or allocate memory). | |
13114 | ||
861bb6c1 | 13115 | @item -fstack-check |
cd3bb277 | 13116 | @opindex fstack-check |
861bb6c1 JL |
13117 | Generate code to verify that you do not go beyond the boundary of the |
13118 | stack. You should specify this flag if you are running in an | |
13119 | environment with multiple threads, but only rarely need to specify it in | |
13120 | a single-threaded environment since stack overflow is automatically | |
13121 | detected on nearly all systems if there is only one stack. | |
13122 | ||
a157febd GK |
13123 | Note that this switch does not actually cause checking to be done; the |
13124 | operating system must do that. The switch causes generation of code | |
13125 | to ensure that the operating system sees the stack being extended. | |
13126 | ||
13127 | @item -fstack-limit-register=@var{reg} | |
13128 | @itemx -fstack-limit-symbol=@var{sym} | |
13129 | @itemx -fno-stack-limit | |
cd3bb277 JM |
13130 | @opindex fstack-limit-register |
13131 | @opindex fstack-limit-symbol | |
13132 | @opindex fno-stack-limit | |
a157febd GK |
13133 | Generate code to ensure that the stack does not grow beyond a certain value, |
13134 | either the value of a register or the address of a symbol. If the stack | |
13135 | would grow beyond the value, a signal is raised. For most targets, | |
13136 | the signal is raised before the stack overruns the boundary, so | |
13137 | it is possible to catch the signal without taking special precautions. | |
13138 | ||
9c34dbbf ZW |
13139 | For instance, if the stack starts at absolute address @samp{0x80000000} |
13140 | and grows downwards, you can use the flags | |
13141 | @option{-fstack-limit-symbol=__stack_limit} and | |
13142 | @option{-Wl,--defsym,__stack_limit=0x7ffe0000} to enforce a stack limit | |
13143 | of 128KB@. Note that this may only work with the GNU linker. | |
a157febd | 13144 | |
e5eb27e5 JL |
13145 | @cindex aliasing of parameters |
13146 | @cindex parameters, aliased | |
13147 | @item -fargument-alias | |
04afd9d6 JL |
13148 | @itemx -fargument-noalias |
13149 | @itemx -fargument-noalias-global | |
cd3bb277 JM |
13150 | @opindex fargument-alias |
13151 | @opindex fargument-noalias | |
13152 | @opindex fargument-noalias-global | |
e5eb27e5 JL |
13153 | Specify the possible relationships among parameters and between |
13154 | parameters and global data. | |
13155 | ||
630d3d5a | 13156 | @option{-fargument-alias} specifies that arguments (parameters) may |
9c34dbbf | 13157 | alias each other and may alias global storage.@* |
630d3d5a | 13158 | @option{-fargument-noalias} specifies that arguments do not alias |
9c34dbbf | 13159 | each other, but may alias global storage.@* |
630d3d5a | 13160 | @option{-fargument-noalias-global} specifies that arguments do not |
e5eb27e5 JL |
13161 | alias each other and do not alias global storage. |
13162 | ||
13163 | Each language will automatically use whatever option is required by | |
13164 | the language standard. You should not need to use these options yourself. | |
19283265 RH |
13165 | |
13166 | @item -fleading-underscore | |
cd3bb277 | 13167 | @opindex fleading-underscore |
695ac33f | 13168 | This option and its counterpart, @option{-fno-leading-underscore}, forcibly |
19283265 RH |
13169 | change the way C symbols are represented in the object file. One use |
13170 | is to help link with legacy assembly code. | |
13171 | ||
a9c60612 JJ |
13172 | @strong{Warning:} the @option{-fleading-underscore} switch causes GCC to |
13173 | generate code that is not binary compatible with code generated without that | |
13174 | switch. Use it to conform to a non-default application binary interface. | |
13175 | Not all targets provide complete support for this switch. | |
3d78f2e9 RH |
13176 | |
13177 | @item -ftls-model=@var{model} | |
13178 | Alter the thread-local storage model to be used (@pxref{Thread-Local}). | |
13179 | The @var{model} argument should be one of @code{global-dynamic}, | |
13180 | @code{local-dynamic}, @code{initial-exec} or @code{local-exec}. | |
13181 | ||
13182 | The default without @option{-fpic} is @code{initial-exec}; with | |
13183 | @option{-fpic} the default is @code{global-dynamic}. | |
d7afec4b ND |
13184 | |
13185 | @item -fvisibility=@var{default|internal|hidden|protected} | |
13186 | @opindex fvisibility | |
78466c0e | 13187 | Set the default ELF image symbol visibility to the specified option---all |
27ef2cdd | 13188 | symbols will be marked with this unless overridden within the code. |
d7afec4b | 13189 | Using this feature can very substantially improve linking and |
27ef2cdd | 13190 | load times of shared object libraries, produce more optimized |
d7afec4b ND |
13191 | code, provide near-perfect API export and prevent symbol clashes. |
13192 | It is @strong{strongly} recommended that you use this in any shared objects | |
13193 | you distribute. | |
f0eb93a8 | 13194 | |
d7afec4b ND |
13195 | Despite the nomenclature, @code{default} always means public ie; |
13196 | available to be linked against from outside the shared object. | |
13197 | @code{protected} and @code{internal} are pretty useless in real-world | |
13198 | usage so the only other commonly used option will be @code{hidden}. | |
78466c0e JM |
13199 | The default if @option{-fvisibility} isn't specified is |
13200 | @code{default}, i.e., make every | |
13201 | symbol public---this causes the same behavior as previous versions of | |
8a36672b | 13202 | GCC@. |
f0eb93a8 | 13203 | |
d7afec4b ND |
13204 | A good explanation of the benefits offered by ensuring ELF |
13205 | symbols have the correct visibility is given by ``How To Write | |
13206 | Shared Libraries'' by Ulrich Drepper (which can be found at | |
78466c0e | 13207 | @w{@uref{http://people.redhat.com/~drepper/}})---however a superior |
d7afec4b ND |
13208 | solution made possible by this option to marking things hidden when |
13209 | the default is public is to make the default hidden and mark things | |
8a36672b | 13210 | public. This is the norm with DLL's on Windows and with @option{-fvisibility=hidden} |
d7afec4b ND |
13211 | and @code{__attribute__ ((visibility("default")))} instead of |
13212 | @code{__declspec(dllexport)} you get almost identical semantics with | |
8a36672b | 13213 | identical syntax. This is a great boon to those working with |
d7afec4b ND |
13214 | cross-platform projects. |
13215 | ||
13216 | For those adding visibility support to existing code, you may find | |
8a36672b | 13217 | @samp{#pragma GCC visibility} of use. This works by you enclosing |
d7afec4b ND |
13218 | the declarations you wish to set visibility for with (for example) |
13219 | @samp{#pragma GCC visibility push(hidden)} and | |
be1b1c9b L |
13220 | @samp{#pragma GCC visibility pop}. |
13221 | Bear in mind that symbol visibility should be viewed @strong{as | |
d7afec4b ND |
13222 | part of the API interface contract} and thus all new code should |
13223 | always specify visibility when it is not the default ie; declarations | |
13224 | only for use within the local DSO should @strong{always} be marked explicitly | |
78466c0e | 13225 | as hidden as so to avoid PLT indirection overheads---making this |
d7afec4b ND |
13226 | abundantly clear also aids readability and self-documentation of the code. |
13227 | Note that due to ISO C++ specification requirements, operator new and | |
13228 | operator delete must always be of default visibility. | |
13229 | ||
ce84fa41 DP |
13230 | An overview of these techniques, their benefits and how to use them |
13231 | is at @w{@uref{http://gcc.gnu.org/wiki/Visibility}}. | |
13232 | ||
74291a4b MM |
13233 | @end table |
13234 | ||
ee457005 JM |
13235 | @c man end |
13236 | ||
74291a4b | 13237 | @node Environment Variables |
0c2d1a2a | 13238 | @section Environment Variables Affecting GCC |
74291a4b MM |
13239 | @cindex environment variables |
13240 | ||
ee457005 | 13241 | @c man begin ENVIRONMENT |
0c2d1a2a JB |
13242 | This section describes several environment variables that affect how GCC |
13243 | operates. Some of them work by specifying directories or prefixes to use | |
767094dd | 13244 | when searching for various kinds of files. Some are used to specify other |
46103ab4 | 13245 | aspects of the compilation environment. |
74291a4b | 13246 | |
74291a4b | 13247 | Note that you can also specify places to search using options such as |
630d3d5a | 13248 | @option{-B}, @option{-I} and @option{-L} (@pxref{Directory Options}). These |
74291a4b | 13249 | take precedence over places specified using environment variables, which |
161d7b59 | 13250 | in turn take precedence over those specified by the configuration of GCC@. |
b11cc610 JM |
13251 | @xref{Driver,, Controlling the Compilation Driver @file{gcc}, gccint, |
13252 | GNU Compiler Collection (GCC) Internals}. | |
74291a4b | 13253 | |
bedc7537 | 13254 | @table @env |
ab87f8c8 JL |
13255 | @item LANG |
13256 | @itemx LC_CTYPE | |
13257 | @c @itemx LC_COLLATE | |
13258 | @itemx LC_MESSAGES | |
13259 | @c @itemx LC_MONETARY | |
13260 | @c @itemx LC_NUMERIC | |
13261 | @c @itemx LC_TIME | |
13262 | @itemx LC_ALL | |
13263 | @findex LANG | |
13264 | @findex LC_CTYPE | |
13265 | @c @findex LC_COLLATE | |
13266 | @findex LC_MESSAGES | |
13267 | @c @findex LC_MONETARY | |
13268 | @c @findex LC_NUMERIC | |
13269 | @c @findex LC_TIME | |
13270 | @findex LC_ALL | |
13271 | @cindex locale | |
0c2d1a2a JB |
13272 | These environment variables control the way that GCC uses |
13273 | localization information that allow GCC to work with different | |
13274 | national conventions. GCC inspects the locale categories | |
bedc7537 | 13275 | @env{LC_CTYPE} and @env{LC_MESSAGES} if it has been configured to do |
ab87f8c8 | 13276 | so. These locale categories can be set to any value supported by your |
0e6d31fb JM |
13277 | installation. A typical value is @samp{en_GB.UTF-8} for English in the United |
13278 | Kingdom encoded in UTF-8. | |
ab87f8c8 | 13279 | |
bedc7537 | 13280 | The @env{LC_CTYPE} environment variable specifies character |
0c2d1a2a | 13281 | classification. GCC uses it to determine the character boundaries in |
ab87f8c8 JL |
13282 | a string; this is needed for some multibyte encodings that contain quote |
13283 | and escape characters that would otherwise be interpreted as a string | |
13284 | end or escape. | |
13285 | ||
bedc7537 | 13286 | The @env{LC_MESSAGES} environment variable specifies the language to |
ab87f8c8 JL |
13287 | use in diagnostic messages. |
13288 | ||
bedc7537 NC |
13289 | If the @env{LC_ALL} environment variable is set, it overrides the value |
13290 | of @env{LC_CTYPE} and @env{LC_MESSAGES}; otherwise, @env{LC_CTYPE} | |
13291 | and @env{LC_MESSAGES} default to the value of the @env{LANG} | |
0c2d1a2a | 13292 | environment variable. If none of these variables are set, GCC |
ab87f8c8 JL |
13293 | defaults to traditional C English behavior. |
13294 | ||
74291a4b MM |
13295 | @item TMPDIR |
13296 | @findex TMPDIR | |
bedc7537 | 13297 | If @env{TMPDIR} is set, it specifies the directory to use for temporary |
0c2d1a2a | 13298 | files. GCC uses temporary files to hold the output of one stage of |
74291a4b MM |
13299 | compilation which is to be used as input to the next stage: for example, |
13300 | the output of the preprocessor, which is the input to the compiler | |
13301 | proper. | |
13302 | ||
13303 | @item GCC_EXEC_PREFIX | |
13304 | @findex GCC_EXEC_PREFIX | |
bedc7537 | 13305 | If @env{GCC_EXEC_PREFIX} is set, it specifies a prefix to use in the |
74291a4b MM |
13306 | names of the subprograms executed by the compiler. No slash is added |
13307 | when this prefix is combined with the name of a subprogram, but you can | |
13308 | specify a prefix that ends with a slash if you wish. | |
13309 | ||
f0523f02 | 13310 | If @env{GCC_EXEC_PREFIX} is not set, GCC will attempt to figure out |
0deb20df TT |
13311 | an appropriate prefix to use based on the pathname it was invoked with. |
13312 | ||
0c2d1a2a | 13313 | If GCC cannot find the subprogram using the specified prefix, it |
74291a4b MM |
13314 | tries looking in the usual places for the subprogram. |
13315 | ||
bedc7537 | 13316 | The default value of @env{GCC_EXEC_PREFIX} is |
8e5f33ff | 13317 | @file{@var{prefix}/lib/gcc/} where @var{prefix} is the value |
74291a4b MM |
13318 | of @code{prefix} when you ran the @file{configure} script. |
13319 | ||
630d3d5a | 13320 | Other prefixes specified with @option{-B} take precedence over this prefix. |
74291a4b MM |
13321 | |
13322 | This prefix is also used for finding files such as @file{crt0.o} that are | |
13323 | used for linking. | |
13324 | ||
13325 | In addition, the prefix is used in an unusual way in finding the | |
13326 | directories to search for header files. For each of the standard | |
8e5f33ff | 13327 | directories whose name normally begins with @samp{/usr/local/lib/gcc} |
bedc7537 | 13328 | (more precisely, with the value of @env{GCC_INCLUDE_DIR}), GCC tries |
74291a4b | 13329 | replacing that beginning with the specified prefix to produce an |
630d3d5a | 13330 | alternate directory name. Thus, with @option{-Bfoo/}, GCC will search |
74291a4b MM |
13331 | @file{foo/bar} where it would normally search @file{/usr/local/lib/bar}. |
13332 | These alternate directories are searched first; the standard directories | |
13333 | come next. | |
13334 | ||
13335 | @item COMPILER_PATH | |
13336 | @findex COMPILER_PATH | |
bedc7537 NC |
13337 | The value of @env{COMPILER_PATH} is a colon-separated list of |
13338 | directories, much like @env{PATH}. GCC tries the directories thus | |
74291a4b | 13339 | specified when searching for subprograms, if it can't find the |
bedc7537 | 13340 | subprograms using @env{GCC_EXEC_PREFIX}. |
74291a4b MM |
13341 | |
13342 | @item LIBRARY_PATH | |
13343 | @findex LIBRARY_PATH | |
bedc7537 NC |
13344 | The value of @env{LIBRARY_PATH} is a colon-separated list of |
13345 | directories, much like @env{PATH}. When configured as a native compiler, | |
0c2d1a2a | 13346 | GCC tries the directories thus specified when searching for special |
bedc7537 | 13347 | linker files, if it can't find them using @env{GCC_EXEC_PREFIX}. Linking |
0c2d1a2a | 13348 | using GCC also uses these directories when searching for ordinary |
630d3d5a JM |
13349 | libraries for the @option{-l} option (but directories specified with |
13350 | @option{-L} come first). | |
74291a4b | 13351 | |
56f48ce9 DB |
13352 | @item LANG |
13353 | @findex LANG | |
13354 | @cindex locale definition | |
767094dd | 13355 | This variable is used to pass locale information to the compiler. One way in |
56f48ce9 DB |
13356 | which this information is used is to determine the character set to be used |
13357 | when character literals, string literals and comments are parsed in C and C++. | |
13358 | When the compiler is configured to allow multibyte characters, | |
bedc7537 | 13359 | the following values for @env{LANG} are recognized: |
56f48ce9 | 13360 | |
2642624b | 13361 | @table @samp |
56f48ce9 DB |
13362 | @item C-JIS |
13363 | Recognize JIS characters. | |
13364 | @item C-SJIS | |
13365 | Recognize SJIS characters. | |
13366 | @item C-EUCJP | |
13367 | Recognize EUCJP characters. | |
13368 | @end table | |
13369 | ||
bedc7537 | 13370 | If @env{LANG} is not defined, or if it has some other value, then the |
56f48ce9 DB |
13371 | compiler will use mblen and mbtowc as defined by the default locale to |
13372 | recognize and translate multibyte characters. | |
74291a4b MM |
13373 | @end table |
13374 | ||
40adaa27 NB |
13375 | @noindent |
13376 | Some additional environments variables affect the behavior of the | |
13377 | preprocessor. | |
13378 | ||
13379 | @include cppenv.texi | |
13380 | ||
9d86bffc JM |
13381 | @c man end |
13382 | ||
17211ab5 GK |
13383 | @node Precompiled Headers |
13384 | @section Using Precompiled Headers | |
13385 | @cindex precompiled headers | |
13386 | @cindex speed of compilation | |
13387 | ||
13388 | Often large projects have many header files that are included in every | |
13389 | source file. The time the compiler takes to process these header files | |
13390 | over and over again can account for nearly all of the time required to | |
13391 | build the project. To make builds faster, GCC allows users to | |
13392 | `precompile' a header file; then, if builds can use the precompiled | |
13393 | header file they will be much faster. | |
13394 | ||
13395 | To create a precompiled header file, simply compile it as you would any | |
13396 | other file, if necessary using the @option{-x} option to make the driver | |
13397 | treat it as a C or C++ header file. You will probably want to use a | |
13398 | tool like @command{make} to keep the precompiled header up-to-date when | |
13399 | the headers it contains change. | |
13400 | ||
13401 | A precompiled header file will be searched for when @code{#include} is | |
13402 | seen in the compilation. As it searches for the included file | |
24726b96 | 13403 | (@pxref{Search Path,,Search Path,cpp,The C Preprocessor}) the |
17211ab5 GK |
13404 | compiler looks for a precompiled header in each directory just before it |
13405 | looks for the include file in that directory. The name searched for is | |
d8fad4ea | 13406 | the name specified in the @code{#include} with @samp{.gch} appended. If |
17211ab5 GK |
13407 | the precompiled header file can't be used, it is ignored. |
13408 | ||
13409 | For instance, if you have @code{#include "all.h"}, and you have | |
d8fad4ea | 13410 | @file{all.h.gch} in the same directory as @file{all.h}, then the |
17211ab5 GK |
13411 | precompiled header file will be used if possible, and the original |
13412 | header will be used otherwise. | |
13413 | ||
13414 | Alternatively, you might decide to put the precompiled header file in a | |
13415 | directory and use @option{-I} to ensure that directory is searched | |
13416 | before (or instead of) the directory containing the original header. | |
13417 | Then, if you want to check that the precompiled header file is always | |
13418 | used, you can put a file of the same name as the original header in this | |
13419 | directory containing an @code{#error} command. | |
13420 | ||
13421 | This also works with @option{-include}. So yet another way to use | |
13422 | precompiled headers, good for projects not designed with precompiled | |
13423 | header files in mind, is to simply take most of the header files used by | |
13424 | a project, include them from another header file, precompile that header | |
13425 | file, and @option{-include} the precompiled header. If the header files | |
13426 | have guards against multiple inclusion, they will be skipped because | |
13427 | they've already been included (in the precompiled header). | |
13428 | ||
13429 | If you need to precompile the same header file for different | |
13430 | languages, targets, or compiler options, you can instead make a | |
d8fad4ea | 13431 | @emph{directory} named like @file{all.h.gch}, and put each precompiled |
54e109ed GK |
13432 | header in the directory, perhaps using @option{-o}. It doesn't matter |
13433 | what you call the files in the directory, every precompiled header in | |
13434 | the directory will be considered. The first precompiled header | |
13435 | encountered in the directory that is valid for this compilation will | |
13436 | be used; they're searched in no particular order. | |
17211ab5 GK |
13437 | |
13438 | There are many other possibilities, limited only by your imagination, | |
13439 | good sense, and the constraints of your build system. | |
13440 | ||
13441 | A precompiled header file can be used only when these conditions apply: | |
13442 | ||
13443 | @itemize | |
13444 | @item | |
13445 | Only one precompiled header can be used in a particular compilation. | |
54e109ed | 13446 | |
17211ab5 GK |
13447 | @item |
13448 | A precompiled header can't be used once the first C token is seen. You | |
13449 | can have preprocessor directives before a precompiled header; you can | |
13450 | even include a precompiled header from inside another header, so long as | |
13451 | there are no C tokens before the @code{#include}. | |
54e109ed | 13452 | |
17211ab5 GK |
13453 | @item |
13454 | The precompiled header file must be produced for the same language as | |
13455 | the current compilation. You can't use a C precompiled header for a C++ | |
13456 | compilation. | |
54e109ed | 13457 | |
17211ab5 | 13458 | @item |
3fd30b88 GK |
13459 | The precompiled header file must have been produced by the same compiler |
13460 | binary as the current compilation is using. | |
54e109ed | 13461 | |
17211ab5 | 13462 | @item |
54e109ed GK |
13463 | Any macros defined before the precompiled header is included must |
13464 | either be defined in the same way as when the precompiled header was | |
13465 | generated, or must not affect the precompiled header, which usually | |
0bdcd332 | 13466 | means that they don't appear in the precompiled header at all. |
54e109ed GK |
13467 | |
13468 | The @option{-D} option is one way to define a macro before a | |
13469 | precompiled header is included; using a @code{#define} can also do it. | |
13470 | There are also some options that define macros implicitly, like | |
13471 | @option{-O} and @option{-Wdeprecated}; the same rule applies to macros | |
13472 | defined this way. | |
13473 | ||
13474 | @item If debugging information is output when using the precompiled | |
13475 | header, using @option{-g} or similar, the same kind of debugging information | |
13476 | must have been output when building the precompiled header. However, | |
13477 | a precompiled header built using @option{-g} can be used in a compilation | |
13478 | when no debugging information is being output. | |
13479 | ||
13480 | @item The same @option{-m} options must generally be used when building | |
13481 | and using the precompiled header. @xref{Submodel Options}, | |
13482 | for any cases where this rule is relaxed. | |
13483 | ||
13484 | @item Each of the following options must be the same when building and using | |
13485 | the precompiled header: | |
13486 | ||
13487 | @gccoptlist{-fexceptions -funit-at-a-time} | |
13488 | ||
17211ab5 | 13489 | @item |
54e109ed GK |
13490 | Some other command-line options starting with @option{-f}, |
13491 | @option{-p}, or @option{-O} must be defined in the same way as when | |
13492 | the precompiled header was generated. At present, it's not clear | |
13493 | which options are safe to change and which are not; the safest choice | |
13494 | is to use exactly the same options when generating and using the | |
13495 | precompiled header. The following are known to be safe: | |
13496 | ||
17ccdd2c | 13497 | @gccoptlist{-fmessage-length= -fpreprocessed |
3fd30b88 | 13498 | -fsched-interblock -fsched-spec -fsched-spec-load -fsched-spec-load-dangerous |
17ccdd2c | 13499 | -fsched-verbose=<number> -fschedule-insns -fvisibility= |
3fd30b88 | 13500 | -pedantic-errors} |
54e109ed | 13501 | |
17211ab5 GK |
13502 | @end itemize |
13503 | ||
54e109ed GK |
13504 | For all of these except the last, the compiler will automatically |
13505 | ignore the precompiled header if the conditions aren't met. If you | |
13506 | find an option combination that doesn't work and doesn't cause the | |
13507 | precompiled header to be ignored, please consider filing a bug report, | |
13508 | see @ref{Bugs}. | |
17211ab5 | 13509 | |
c0d578e6 | 13510 | If you do use differing options when generating and using the |
83c99486 KH |
13511 | precompiled header, the actual behavior will be a mixture of the |
13512 | behavior for the options. For instance, if you use @option{-g} to | |
c0d578e6 GK |
13513 | generate the precompiled header but not when using it, you may or may |
13514 | not get debugging information for routines in the precompiled header. | |
13515 | ||
74291a4b MM |
13516 | @node Running Protoize |
13517 | @section Running Protoize | |
13518 | ||
161d7b59 | 13519 | The program @code{protoize} is an optional part of GCC@. You can use |
c1030c7c | 13520 | it to add prototypes to a program, thus converting the program to ISO |
74291a4b MM |
13521 | C in one respect. The companion program @code{unprotoize} does the |
13522 | reverse: it removes argument types from any prototypes that are found. | |
13523 | ||
13524 | When you run these programs, you must specify a set of source files as | |
13525 | command line arguments. The conversion programs start out by compiling | |
13526 | these files to see what functions they define. The information gathered | |
13527 | about a file @var{foo} is saved in a file named @file{@var{foo}.X}. | |
13528 | ||
13529 | After scanning comes actual conversion. The specified files are all | |
13530 | eligible to be converted; any files they include (whether sources or | |
13531 | just headers) are eligible as well. | |
13532 | ||
13533 | But not all the eligible files are converted. By default, | |
13534 | @code{protoize} and @code{unprotoize} convert only source and header | |
13535 | files in the current directory. You can specify additional directories | |
630d3d5a | 13536 | whose files should be converted with the @option{-d @var{directory}} |
74291a4b | 13537 | option. You can also specify particular files to exclude with the |
630d3d5a | 13538 | @option{-x @var{file}} option. A file is converted if it is eligible, its |
74291a4b MM |
13539 | directory name matches one of the specified directory names, and its |
13540 | name within the directory has not been excluded. | |
13541 | ||
13542 | Basic conversion with @code{protoize} consists of rewriting most | |
13543 | function definitions and function declarations to specify the types of | |
13544 | the arguments. The only ones not rewritten are those for varargs | |
13545 | functions. | |
13546 | ||
13547 | @code{protoize} optionally inserts prototype declarations at the | |
13548 | beginning of the source file, to make them available for any calls that | |
13549 | precede the function's definition. Or it can insert prototype | |
13550 | declarations with block scope in the blocks where undeclared functions | |
13551 | are called. | |
13552 | ||
13553 | Basic conversion with @code{unprotoize} consists of rewriting most | |
13554 | function declarations to remove any argument types, and rewriting | |
c1030c7c | 13555 | function definitions to the old-style pre-ISO form. |
74291a4b MM |
13556 | |
13557 | Both conversion programs print a warning for any function declaration or | |
13558 | definition that they can't convert. You can suppress these warnings | |
630d3d5a | 13559 | with @option{-q}. |
74291a4b MM |
13560 | |
13561 | The output from @code{protoize} or @code{unprotoize} replaces the | |
13562 | original source file. The original file is renamed to a name ending | |
02f52e19 | 13563 | with @samp{.save} (for DOS, the saved filename ends in @samp{.sav} |
a7db8bbb MK |
13564 | without the original @samp{.c} suffix). If the @samp{.save} (@samp{.sav} |
13565 | for DOS) file already exists, then the source file is simply discarded. | |
74291a4b | 13566 | |
0c2d1a2a | 13567 | @code{protoize} and @code{unprotoize} both depend on GCC itself to |
74291a4b | 13568 | scan the program and collect information about the functions it uses. |
0c2d1a2a | 13569 | So neither of these programs will work until GCC is installed. |
74291a4b MM |
13570 | |
13571 | Here is a table of the options you can use with @code{protoize} and | |
13572 | @code{unprotoize}. Each option works with both programs unless | |
13573 | otherwise stated. | |
13574 | ||
13575 | @table @code | |
13576 | @item -B @var{directory} | |
13577 | Look for the file @file{SYSCALLS.c.X} in @var{directory}, instead of the | |
13578 | usual directory (normally @file{/usr/local/lib}). This file contains | |
13579 | prototype information about standard system functions. This option | |
13580 | applies only to @code{protoize}. | |
13581 | ||
13582 | @item -c @var{compilation-options} | |
05739753 | 13583 | Use @var{compilation-options} as the options when running @command{gcc} to |
630d3d5a | 13584 | produce the @samp{.X} files. The special option @option{-aux-info} is |
05739753 | 13585 | always passed in addition, to tell @command{gcc} to write a @samp{.X} file. |
74291a4b MM |
13586 | |
13587 | Note that the compilation options must be given as a single argument to | |
13588 | @code{protoize} or @code{unprotoize}. If you want to specify several | |
05739753 | 13589 | @command{gcc} options, you must quote the entire set of compilation options |
74291a4b MM |
13590 | to make them a single word in the shell. |
13591 | ||
05739753 | 13592 | There are certain @command{gcc} arguments that you cannot use, because they |
630d3d5a JM |
13593 | would produce the wrong kind of output. These include @option{-g}, |
13594 | @option{-O}, @option{-c}, @option{-S}, and @option{-o} If you include these in | |
74291a4b MM |
13595 | the @var{compilation-options}, they are ignored. |
13596 | ||
13597 | @item -C | |
a7db8bbb | 13598 | Rename files to end in @samp{.C} (@samp{.cc} for DOS-based file |
02f52e19 | 13599 | systems) instead of @samp{.c}. This is convenient if you are converting |
ee77eda5 | 13600 | a C program to C++. This option applies only to @code{protoize}. |
74291a4b MM |
13601 | |
13602 | @item -g | |
13603 | Add explicit global declarations. This means inserting explicit | |
13604 | declarations at the beginning of each source file for each function | |
13605 | that is called in the file and was not declared. These declarations | |
13606 | precede the first function definition that contains a call to an | |
13607 | undeclared function. This option applies only to @code{protoize}. | |
13608 | ||
13609 | @item -i @var{string} | |
13610 | Indent old-style parameter declarations with the string @var{string}. | |
13611 | This option applies only to @code{protoize}. | |
13612 | ||
13613 | @code{unprotoize} converts prototyped function definitions to old-style | |
13614 | function definitions, where the arguments are declared between the | |
13615 | argument list and the initial @samp{@{}. By default, @code{unprotoize} | |
13616 | uses five spaces as the indentation. If you want to indent with just | |
630d3d5a | 13617 | one space instead, use @option{-i " "}. |
74291a4b MM |
13618 | |
13619 | @item -k | |
13620 | Keep the @samp{.X} files. Normally, they are deleted after conversion | |
13621 | is finished. | |
13622 | ||
13623 | @item -l | |
630d3d5a | 13624 | Add explicit local declarations. @code{protoize} with @option{-l} inserts |
74291a4b MM |
13625 | a prototype declaration for each function in each block which calls the |
13626 | function without any declaration. This option applies only to | |
13627 | @code{protoize}. | |
13628 | ||
13629 | @item -n | |
13630 | Make no real changes. This mode just prints information about the conversions | |
630d3d5a | 13631 | that would have been done without @option{-n}. |
74291a4b MM |
13632 | |
13633 | @item -N | |
13634 | Make no @samp{.save} files. The original files are simply deleted. | |
13635 | Use this option with caution. | |
13636 | ||
13637 | @item -p @var{program} | |
13638 | Use the program @var{program} as the compiler. Normally, the name | |
13639 | @file{gcc} is used. | |
13640 | ||
13641 | @item -q | |
13642 | Work quietly. Most warnings are suppressed. | |
13643 | ||
13644 | @item -v | |
05739753 | 13645 | Print the version number, just like @option{-v} for @command{gcc}. |
74291a4b MM |
13646 | @end table |
13647 | ||
13648 | If you need special compiler options to compile one of your program's | |
13649 | source files, then you should generate that file's @samp{.X} file | |
05739753 | 13650 | specially, by running @command{gcc} on that source file with the |
630d3d5a | 13651 | appropriate options and the option @option{-aux-info}. Then run |
74291a4b MM |
13652 | @code{protoize} on the entire set of files. @code{protoize} will use |
13653 | the existing @samp{.X} file because it is newer than the source file. | |
13654 | For example: | |
13655 | ||
3ab51846 | 13656 | @smallexample |
b1018de6 | 13657 | gcc -Dfoo=bar file1.c -aux-info file1.X |
74291a4b | 13658 | protoize *.c |
3ab51846 | 13659 | @end smallexample |
74291a4b MM |
13660 | |
13661 | @noindent | |
13662 | You need to include the special files along with the rest in the | |
13663 | @code{protoize} command, even though their @samp{.X} files already | |
13664 | exist, because otherwise they won't get converted. | |
13665 | ||
13666 | @xref{Protoize Caveats}, for more information on how to use | |
13667 | @code{protoize} successfully. |