|
0
|
1 @c Copyright (C) 1988, 1989, 1992, 1993, 1994, 1995, 1996, 1997, 1998,
|
|
|
2 @c 1999, 2000, 2001, 2003, 2004, 2005, 2008 Free Software Foundation, Inc.
|
|
|
3 @c This is part of the GCC manual.
|
|
|
4 @c For copying conditions, see the file gcc.texi.
|
|
|
5
|
|
|
6 @node Fragments
|
|
|
7 @chapter Makefile Fragments
|
|
|
8 @cindex makefile fragment
|
|
|
9
|
|
|
10 When you configure GCC using the @file{configure} script, it will
|
|
|
11 construct the file @file{Makefile} from the template file
|
|
|
12 @file{Makefile.in}. When it does this, it can incorporate makefile
|
|
|
13 fragments from the @file{config} directory. These are used to set
|
|
|
14 Makefile parameters that are not amenable to being calculated by
|
|
|
15 autoconf. The list of fragments to incorporate is set by
|
|
|
16 @file{config.gcc} (and occasionally @file{config.build}
|
|
|
17 and @file{config.host}); @xref{System Config}.
|
|
|
18
|
|
|
19 Fragments are named either @file{t-@var{target}} or @file{x-@var{host}},
|
|
|
20 depending on whether they are relevant to configuring GCC to produce
|
|
|
21 code for a particular target, or to configuring GCC to run on a
|
|
|
22 particular host. Here @var{target} and @var{host} are mnemonics
|
|
|
23 which usually have some relationship to the canonical system name, but
|
|
|
24 no formal connection.
|
|
|
25
|
|
|
26 If these files do not exist, it means nothing needs to be added for a
|
|
|
27 given target or host. Most targets need a few @file{t-@var{target}}
|
|
|
28 fragments, but needing @file{x-@var{host}} fragments is rare.
|
|
|
29
|
|
|
30 @menu
|
|
|
31 * Target Fragment:: Writing @file{t-@var{target}} files.
|
|
|
32 * Host Fragment:: Writing @file{x-@var{host}} files.
|
|
|
33 @end menu
|
|
|
34
|
|
|
35 @node Target Fragment
|
|
|
36 @section Target Makefile Fragments
|
|
|
37 @cindex target makefile fragment
|
|
|
38 @cindex @file{t-@var{target}}
|
|
|
39
|
|
|
40 Target makefile fragments can set these Makefile variables.
|
|
|
41
|
|
|
42 @table @code
|
|
|
43 @findex LIBGCC2_CFLAGS
|
|
|
44 @item LIBGCC2_CFLAGS
|
|
|
45 Compiler flags to use when compiling @file{libgcc2.c}.
|
|
|
46
|
|
|
47 @findex LIB2FUNCS_EXTRA
|
|
|
48 @item LIB2FUNCS_EXTRA
|
|
|
49 A list of source file names to be compiled or assembled and inserted
|
|
|
50 into @file{libgcc.a}.
|
|
|
51
|
|
|
52 @findex Floating Point Emulation
|
|
|
53 @item Floating Point Emulation
|
|
|
54 To have GCC include software floating point libraries in @file{libgcc.a}
|
|
|
55 define @code{FPBIT} and @code{DPBIT} along with a few rules as follows:
|
|
|
56 @smallexample
|
|
|
57 # We want fine grained libraries, so use the new code
|
|
|
58 # to build the floating point emulation libraries.
|
|
|
59 FPBIT = fp-bit.c
|
|
|
60 DPBIT = dp-bit.c
|
|
|
61
|
|
|
62
|
|
|
63 fp-bit.c: $(srcdir)/config/fp-bit.c
|
|
|
64 echo '#define FLOAT' > fp-bit.c
|
|
|
65 cat $(srcdir)/config/fp-bit.c >> fp-bit.c
|
|
|
66
|
|
|
67 dp-bit.c: $(srcdir)/config/fp-bit.c
|
|
|
68 cat $(srcdir)/config/fp-bit.c > dp-bit.c
|
|
|
69 @end smallexample
|
|
|
70
|
|
|
71 You may need to provide additional #defines at the beginning of @file{fp-bit.c}
|
|
|
72 and @file{dp-bit.c} to control target endianness and other options.
|
|
|
73
|
|
|
74
|
|
|
75 @findex CRTSTUFF_T_CFLAGS
|
|
|
76 @item CRTSTUFF_T_CFLAGS
|
|
|
77 Special flags used when compiling @file{crtstuff.c}.
|
|
|
78 @xref{Initialization}.
|
|
|
79
|
|
|
80 @findex CRTSTUFF_T_CFLAGS_S
|
|
|
81 @item CRTSTUFF_T_CFLAGS_S
|
|
|
82 Special flags used when compiling @file{crtstuff.c} for shared
|
|
|
83 linking. Used if you use @file{crtbeginS.o} and @file{crtendS.o}
|
|
|
84 in @code{EXTRA-PARTS}.
|
|
|
85 @xref{Initialization}.
|
|
|
86
|
|
|
87 @findex MULTILIB_OPTIONS
|
|
|
88 @item MULTILIB_OPTIONS
|
|
|
89 For some targets, invoking GCC in different ways produces objects
|
|
|
90 that can not be linked together. For example, for some targets GCC
|
|
|
91 produces both big and little endian code. For these targets, you must
|
|
|
92 arrange for multiple versions of @file{libgcc.a} to be compiled, one for
|
|
|
93 each set of incompatible options. When GCC invokes the linker, it
|
|
|
94 arranges to link in the right version of @file{libgcc.a}, based on
|
|
|
95 the command line options used.
|
|
|
96
|
|
|
97 The @code{MULTILIB_OPTIONS} macro lists the set of options for which
|
|
|
98 special versions of @file{libgcc.a} must be built. Write options that
|
|
|
99 are mutually incompatible side by side, separated by a slash. Write
|
|
|
100 options that may be used together separated by a space. The build
|
|
|
101 procedure will build all combinations of compatible options.
|
|
|
102
|
|
|
103 For example, if you set @code{MULTILIB_OPTIONS} to @samp{m68000/m68020
|
|
|
104 msoft-float}, @file{Makefile} will build special versions of
|
|
|
105 @file{libgcc.a} using the following sets of options: @option{-m68000},
|
|
|
106 @option{-m68020}, @option{-msoft-float}, @samp{-m68000 -msoft-float}, and
|
|
|
107 @samp{-m68020 -msoft-float}.
|
|
|
108
|
|
|
109 @findex MULTILIB_DIRNAMES
|
|
|
110 @item MULTILIB_DIRNAMES
|
|
|
111 If @code{MULTILIB_OPTIONS} is used, this variable specifies the
|
|
|
112 directory names that should be used to hold the various libraries.
|
|
|
113 Write one element in @code{MULTILIB_DIRNAMES} for each element in
|
|
|
114 @code{MULTILIB_OPTIONS}. If @code{MULTILIB_DIRNAMES} is not used, the
|
|
|
115 default value will be @code{MULTILIB_OPTIONS}, with all slashes treated
|
|
|
116 as spaces.
|
|
|
117
|
|
|
118 For example, if @code{MULTILIB_OPTIONS} is set to @samp{m68000/m68020
|
|
|
119 msoft-float}, then the default value of @code{MULTILIB_DIRNAMES} is
|
|
|
120 @samp{m68000 m68020 msoft-float}. You may specify a different value if
|
|
|
121 you desire a different set of directory names.
|
|
|
122
|
|
|
123 @findex MULTILIB_MATCHES
|
|
|
124 @item MULTILIB_MATCHES
|
|
|
125 Sometimes the same option may be written in two different ways. If an
|
|
|
126 option is listed in @code{MULTILIB_OPTIONS}, GCC needs to know about
|
|
|
127 any synonyms. In that case, set @code{MULTILIB_MATCHES} to a list of
|
|
|
128 items of the form @samp{option=option} to describe all relevant
|
|
|
129 synonyms. For example, @samp{m68000=mc68000 m68020=mc68020}.
|
|
|
130
|
|
|
131 @findex MULTILIB_EXCEPTIONS
|
|
|
132 @item MULTILIB_EXCEPTIONS
|
|
|
133 Sometimes when there are multiple sets of @code{MULTILIB_OPTIONS} being
|
|
|
134 specified, there are combinations that should not be built. In that
|
|
|
135 case, set @code{MULTILIB_EXCEPTIONS} to be all of the switch exceptions
|
|
|
136 in shell case syntax that should not be built.
|
|
|
137
|
|
|
138 For example the ARM processor cannot execute both hardware floating
|
|
|
139 point instructions and the reduced size THUMB instructions at the same
|
|
|
140 time, so there is no need to build libraries with both of these
|
|
|
141 options enabled. Therefore @code{MULTILIB_EXCEPTIONS} is set to:
|
|
|
142 @smallexample
|
|
|
143 *mthumb/*mhard-float*
|
|
|
144 @end smallexample
|
|
|
145
|
|
|
146 @findex MULTILIB_EXTRA_OPTS
|
|
|
147 @item MULTILIB_EXTRA_OPTS
|
|
|
148 Sometimes it is desirable that when building multiple versions of
|
|
|
149 @file{libgcc.a} certain options should always be passed on to the
|
|
|
150 compiler. In that case, set @code{MULTILIB_EXTRA_OPTS} to be the list
|
|
|
151 of options to be used for all builds. If you set this, you should
|
|
|
152 probably set @code{CRTSTUFF_T_CFLAGS} to a dash followed by it.
|
|
|
153
|
|
|
154 @findex NATIVE_SYSTEM_HEADER_DIR
|
|
|
155 @item NATIVE_SYSTEM_HEADER_DIR
|
|
|
156 If the default location for system headers is not @file{/usr/include},
|
|
|
157 you must set this to the directory containing the headers. This value
|
|
|
158 should match the value of the @code{SYSTEM_INCLUDE_DIR} macro.
|
|
|
159
|
|
|
160 @findex SPECS
|
|
|
161 @item SPECS
|
|
|
162 Unfortunately, setting @code{MULTILIB_EXTRA_OPTS} is not enough, since
|
|
|
163 it does not affect the build of target libraries, at least not the
|
|
|
164 build of the default multilib. One possible work-around is to use
|
|
|
165 @code{DRIVER_SELF_SPECS} to bring options from the @file{specs} file
|
|
|
166 as if they had been passed in the compiler driver command line.
|
|
|
167 However, you don't want to be adding these options after the toolchain
|
|
|
168 is installed, so you can instead tweak the @file{specs} file that will
|
|
|
169 be used during the toolchain build, while you still install the
|
|
|
170 original, built-in @file{specs}. The trick is to set @code{SPECS} to
|
|
|
171 some other filename (say @file{specs.install}), that will then be
|
|
|
172 created out of the built-in specs, and introduce a @file{Makefile}
|
|
|
173 rule to generate the @file{specs} file that's going to be used at
|
|
|
174 build time out of your @file{specs.install}.
|
|
|
175
|
|
|
176 @item T_CFLAGS
|
|
|
177 These are extra flags to pass to the C compiler. They are used both
|
|
|
178 when building GCC, and when compiling things with the just-built GCC@.
|
|
|
179 This variable is deprecated and should not be used.
|
|
|
180 @end table
|
|
|
181
|
|
|
182 @node Host Fragment
|
|
|
183 @section Host Makefile Fragments
|
|
|
184 @cindex host makefile fragment
|
|
|
185 @cindex @file{x-@var{host}}
|
|
|
186
|
|
|
187 The use of @file{x-@var{host}} fragments is discouraged. You should only
|
|
|
188 use it for makefile dependencies.
|
