OpenCores
URL https://opencores.org/ocsvn/openrisc/openrisc/trunk

Subversion Repositories openrisc

[/] [openrisc/] [trunk/] [gnu-old/] [gcc-4.2.2/] [gcc/] [doc/] [tm.texi] - Blame information for rev 816

Details | Compare with Previous | View Log

Line No. Rev Author Line
1 38 julius
@c Copyright (C) 1988,1989,1992,1993,1994,1995,1996,1997,1998,1999,2000,2001,
2
@c 2002, 2003, 2004, 2005, 2006 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 Target Macros
7
@chapter Target Description Macros and Functions
8
@cindex machine description macros
9
@cindex target description macros
10
@cindex macros, target description
11
@cindex @file{tm.h} macros
12
 
13
In addition to the file @file{@var{machine}.md}, a machine description
14
includes a C header file conventionally given the name
15
@file{@var{machine}.h} and a C source file named @file{@var{machine}.c}.
16
The header file defines numerous macros that convey the information
17
about the target machine that does not fit into the scheme of the
18
@file{.md} file.  The file @file{tm.h} should be a link to
19
@file{@var{machine}.h}.  The header file @file{config.h} includes
20
@file{tm.h} and most compiler source files include @file{config.h}.  The
21
source file defines a variable @code{targetm}, which is a structure
22
containing pointers to functions and data relating to the target
23
machine.  @file{@var{machine}.c} should also contain their definitions,
24
if they are not defined elsewhere in GCC, and other functions called
25
through the macros defined in the @file{.h} file.
26
 
27
@menu
28
* Target Structure::    The @code{targetm} variable.
29
* Driver::              Controlling how the driver runs the compilation passes.
30
* Run-time Target::     Defining @samp{-m} options like @option{-m68000} and @option{-m68020}.
31
* Per-Function Data::   Defining data structures for per-function information.
32
* Storage Layout::      Defining sizes and alignments of data.
33
* Type Layout::         Defining sizes and properties of basic user data types.
34
* Registers::           Naming and describing the hardware registers.
35
* Register Classes::    Defining the classes of hardware registers.
36
* Old Constraints::     The old way to define machine-specific constraints.
37
* Stack and Calling::   Defining which way the stack grows and by how much.
38
* Varargs::             Defining the varargs macros.
39
* Trampolines::         Code set up at run time to enter a nested function.
40
* Library Calls::       Controlling how library routines are implicitly called.
41
* Addressing Modes::    Defining addressing modes valid for memory operands.
42
* Anchored Addresses::  Defining how @option{-fsection-anchors} should work.
43
* Condition Code::      Defining how insns update the condition code.
44
* Costs::               Defining relative costs of different operations.
45
* Scheduling::          Adjusting the behavior of the instruction scheduler.
46
* Sections::            Dividing storage into text, data, and other sections.
47
* PIC::                 Macros for position independent code.
48
* Assembler Format::    Defining how to write insns and pseudo-ops to output.
49
* Debugging Info::      Defining the format of debugging output.
50
* Floating Point::      Handling floating point for cross-compilers.
51
* Mode Switching::      Insertion of mode-switching instructions.
52
* Target Attributes::   Defining target-specific uses of @code{__attribute__}.
53
* MIPS Coprocessors::   MIPS coprocessor support and how to customize it.
54
* PCH Target::          Validity checking for precompiled headers.
55
* C++ ABI::             Controlling C++ ABI changes.
56
* Misc::                Everything else.
57
@end menu
58
 
59
@node Target Structure
60
@section The Global @code{targetm} Variable
61
@cindex target hooks
62
@cindex target functions
63
 
64
@deftypevar {struct gcc_target} targetm
65
The target @file{.c} file must define the global @code{targetm} variable
66
which contains pointers to functions and data relating to the target
67
machine.  The variable is declared in @file{target.h};
68
@file{target-def.h} defines the macro @code{TARGET_INITIALIZER} which is
69
used to initialize the variable, and macros for the default initializers
70
for elements of the structure.  The @file{.c} file should override those
71
macros for which the default definition is inappropriate.  For example:
72
@smallexample
73
#include "target.h"
74
#include "target-def.h"
75
 
76
/* @r{Initialize the GCC target structure.}  */
77
 
78
#undef TARGET_COMP_TYPE_ATTRIBUTES
79
#define TARGET_COMP_TYPE_ATTRIBUTES @var{machine}_comp_type_attributes
80
 
81
struct gcc_target targetm = TARGET_INITIALIZER;
82
@end smallexample
83
@end deftypevar
84
 
85
Where a macro should be defined in the @file{.c} file in this manner to
86
form part of the @code{targetm} structure, it is documented below as a
87
``Target Hook'' with a prototype.  Many macros will change in future
88
from being defined in the @file{.h} file to being part of the
89
@code{targetm} structure.
90
 
91
@node Driver
92
@section Controlling the Compilation Driver, @file{gcc}
93
@cindex driver
94
@cindex controlling the compilation driver
95
 
96
@c prevent bad page break with this line
97
You can control the compilation driver.
98
 
99
@defmac SWITCH_TAKES_ARG (@var{char})
100
A C expression which determines whether the option @option{-@var{char}}
101
takes arguments.  The value should be the number of arguments that
102
option takes--zero, for many options.
103
 
104
By default, this macro is defined as
105
@code{DEFAULT_SWITCH_TAKES_ARG}, which handles the standard options
106
properly.  You need not define @code{SWITCH_TAKES_ARG} unless you
107
wish to add additional options which take arguments.  Any redefinition
108
should call @code{DEFAULT_SWITCH_TAKES_ARG} and then check for
109
additional options.
110
@end defmac
111
 
112
@defmac WORD_SWITCH_TAKES_ARG (@var{name})
113
A C expression which determines whether the option @option{-@var{name}}
114
takes arguments.  The value should be the number of arguments that
115
option takes--zero, for many options.  This macro rather than
116
@code{SWITCH_TAKES_ARG} is used for multi-character option names.
117
 
118
By default, this macro is defined as
119
@code{DEFAULT_WORD_SWITCH_TAKES_ARG}, which handles the standard options
120
properly.  You need not define @code{WORD_SWITCH_TAKES_ARG} unless you
121
wish to add additional options which take arguments.  Any redefinition
122
should call @code{DEFAULT_WORD_SWITCH_TAKES_ARG} and then check for
123
additional options.
124
@end defmac
125
 
126
@defmac SWITCH_CURTAILS_COMPILATION (@var{char})
127
A C expression which determines whether the option @option{-@var{char}}
128
stops compilation before the generation of an executable.  The value is
129
boolean, nonzero if the option does stop an executable from being
130
generated, zero otherwise.
131
 
132
By default, this macro is defined as
133
@code{DEFAULT_SWITCH_CURTAILS_COMPILATION}, which handles the standard
134
options properly.  You need not define
135
@code{SWITCH_CURTAILS_COMPILATION} unless you wish to add additional
136
options which affect the generation of an executable.  Any redefinition
137
should call @code{DEFAULT_SWITCH_CURTAILS_COMPILATION} and then check
138
for additional options.
139
@end defmac
140
 
141
@defmac SWITCHES_NEED_SPACES
142
A string-valued C expression which enumerates the options for which
143
the linker needs a space between the option and its argument.
144
 
145
If this macro is not defined, the default value is @code{""}.
146
@end defmac
147
 
148
@defmac TARGET_OPTION_TRANSLATE_TABLE
149
If defined, a list of pairs of strings, the first of which is a
150
potential command line target to the @file{gcc} driver program, and the
151
second of which is a space-separated (tabs and other whitespace are not
152
supported) list of options with which to replace the first option.  The
153
target defining this list is responsible for assuring that the results
154
are valid.  Replacement options may not be the @code{--opt} style, they
155
must be the @code{-opt} style.  It is the intention of this macro to
156
provide a mechanism for substitution that affects the multilibs chosen,
157
such as one option that enables many options, some of which select
158
multilibs.  Example nonsensical definition, where @option{-malt-abi},
159
@option{-EB}, and @option{-mspoo} cause different multilibs to be chosen:
160
 
161
@smallexample
162
#define TARGET_OPTION_TRANSLATE_TABLE \
163
@{ "-fast",   "-march=fast-foo -malt-abi -I/usr/fast-foo" @}, \
164
@{ "-compat", "-EB -malign=4 -mspoo" @}
165
@end smallexample
166
@end defmac
167
 
168
@defmac DRIVER_SELF_SPECS
169
A list of specs for the driver itself.  It should be a suitable
170
initializer for an array of strings, with no surrounding braces.
171
 
172
The driver applies these specs to its own command line between loading
173
default @file{specs} files (but not command-line specified ones) and
174
choosing the multilib directory or running any subcommands.  It
175
applies them in the order given, so each spec can depend on the
176
options added by earlier ones.  It is also possible to remove options
177
using @samp{%<@var{option}} in the usual way.
178
 
179
This macro can be useful when a port has several interdependent target
180
options.  It provides a way of standardizing the command line so
181
that the other specs are easier to write.
182
 
183
Do not define this macro if it does not need to do anything.
184
@end defmac
185
 
186
@defmac OPTION_DEFAULT_SPECS
187
A list of specs used to support configure-time default options (i.e.@:
188
@option{--with} options) in the driver.  It should be a suitable initializer
189
for an array of structures, each containing two strings, without the
190
outermost pair of surrounding braces.
191
 
192
The first item in the pair is the name of the default.  This must match
193
the code in @file{config.gcc} for the target.  The second item is a spec
194
to apply if a default with this name was specified.  The string
195
@samp{%(VALUE)} in the spec will be replaced by the value of the default
196
everywhere it occurs.
197
 
198
The driver will apply these specs to its own command line between loading
199
default @file{specs} files and processing @code{DRIVER_SELF_SPECS}, using
200
the same mechanism as @code{DRIVER_SELF_SPECS}.
201
 
202
Do not define this macro if it does not need to do anything.
203
@end defmac
204
 
205
@defmac CPP_SPEC
206
A C string constant that tells the GCC driver program options to
207
pass to CPP@.  It can also specify how to translate options you
208
give to GCC into options for GCC to pass to the CPP@.
209
 
210
Do not define this macro if it does not need to do anything.
211
@end defmac
212
 
213
@defmac CPLUSPLUS_CPP_SPEC
214
This macro is just like @code{CPP_SPEC}, but is used for C++, rather
215
than C@.  If you do not define this macro, then the value of
216
@code{CPP_SPEC} (if any) will be used instead.
217
@end defmac
218
 
219
@defmac CC1_SPEC
220
A C string constant that tells the GCC driver program options to
221
pass to @code{cc1}, @code{cc1plus}, @code{f771}, and the other language
222
front ends.
223
It can also specify how to translate options you give to GCC into options
224
for GCC to pass to front ends.
225
 
226
Do not define this macro if it does not need to do anything.
227
@end defmac
228
 
229
@defmac CC1PLUS_SPEC
230
A C string constant that tells the GCC driver program options to
231
pass to @code{cc1plus}.  It can also specify how to translate options you
232
give to GCC into options for GCC to pass to the @code{cc1plus}.
233
 
234
Do not define this macro if it does not need to do anything.
235
Note that everything defined in CC1_SPEC is already passed to
236
@code{cc1plus} so there is no need to duplicate the contents of
237
CC1_SPEC in CC1PLUS_SPEC@.
238
@end defmac
239
 
240
@defmac ASM_SPEC
241
A C string constant that tells the GCC driver program options to
242
pass to the assembler.  It can also specify how to translate options
243
you give to GCC into options for GCC to pass to the assembler.
244
See the file @file{sun3.h} for an example of this.
245
 
246
Do not define this macro if it does not need to do anything.
247
@end defmac
248
 
249
@defmac ASM_FINAL_SPEC
250
A C string constant that tells the GCC driver program how to
251
run any programs which cleanup after the normal assembler.
252
Normally, this is not needed.  See the file @file{mips.h} for
253
an example of this.
254
 
255
Do not define this macro if it does not need to do anything.
256
@end defmac
257
 
258
@defmac AS_NEEDS_DASH_FOR_PIPED_INPUT
259
Define this macro, with no value, if the driver should give the assembler
260
an argument consisting of a single dash, @option{-}, to instruct it to
261
read from its standard input (which will be a pipe connected to the
262
output of the compiler proper).  This argument is given after any
263
@option{-o} option specifying the name of the output file.
264
 
265
If you do not define this macro, the assembler is assumed to read its
266
standard input if given no non-option arguments.  If your assembler
267
cannot read standard input at all, use a @samp{%@{pipe:%e@}} construct;
268
see @file{mips.h} for instance.
269
@end defmac
270
 
271
@defmac LINK_SPEC
272
A C string constant that tells the GCC driver program options to
273
pass to the linker.  It can also specify how to translate options you
274
give to GCC into options for GCC to pass to the linker.
275
 
276
Do not define this macro if it does not need to do anything.
277
@end defmac
278
 
279
@defmac LIB_SPEC
280
Another C string constant used much like @code{LINK_SPEC}.  The difference
281
between the two is that @code{LIB_SPEC} is used at the end of the
282
command given to the linker.
283
 
284
If this macro is not defined, a default is provided that
285
loads the standard C library from the usual place.  See @file{gcc.c}.
286
@end defmac
287
 
288
@defmac LIBGCC_SPEC
289
Another C string constant that tells the GCC driver program
290
how and when to place a reference to @file{libgcc.a} into the
291
linker command line.  This constant is placed both before and after
292
the value of @code{LIB_SPEC}.
293
 
294
If this macro is not defined, the GCC driver provides a default that
295
passes the string @option{-lgcc} to the linker.
296
@end defmac
297
 
298
@defmac REAL_LIBGCC_SPEC
299
By default, if @code{ENABLE_SHARED_LIBGCC} is defined, the
300
@code{LIBGCC_SPEC} is not directly used by the driver program but is
301
instead modified to refer to different versions of @file{libgcc.a}
302
depending on the values of the command line flags @option{-static},
303
@option{-shared}, @option{-static-libgcc}, and @option{-shared-libgcc}.  On
304
targets where these modifications are inappropriate, define
305
@code{REAL_LIBGCC_SPEC} instead.  @code{REAL_LIBGCC_SPEC} tells the
306
driver how to place a reference to @file{libgcc} on the link command
307
line, but, unlike @code{LIBGCC_SPEC}, it is used unmodified.
308
@end defmac
309
 
310
@defmac USE_LD_AS_NEEDED
311
A macro that controls the modifications to @code{LIBGCC_SPEC}
312
mentioned in @code{REAL_LIBGCC_SPEC}.  If nonzero, a spec will be
313
generated that uses --as-needed and the shared libgcc in place of the
314
static exception handler library, when linking without any of
315
@code{-static}, @code{-static-libgcc}, or @code{-shared-libgcc}.
316
@end defmac
317
 
318
@defmac LINK_EH_SPEC
319
If defined, this C string constant is added to @code{LINK_SPEC}.
320
When @code{USE_LD_AS_NEEDED} is zero or undefined, it also affects
321
the modifications to @code{LIBGCC_SPEC} mentioned in
322
@code{REAL_LIBGCC_SPEC}.
323
@end defmac
324
 
325
@defmac STARTFILE_SPEC
326
Another C string constant used much like @code{LINK_SPEC}.  The
327
difference between the two is that @code{STARTFILE_SPEC} is used at
328
the very beginning of the command given to the linker.
329
 
330
If this macro is not defined, a default is provided that loads the
331
standard C startup file from the usual place.  See @file{gcc.c}.
332
@end defmac
333
 
334
@defmac ENDFILE_SPEC
335
Another C string constant used much like @code{LINK_SPEC}.  The
336
difference between the two is that @code{ENDFILE_SPEC} is used at
337
the very end of the command given to the linker.
338
 
339
Do not define this macro if it does not need to do anything.
340
@end defmac
341
 
342
@defmac THREAD_MODEL_SPEC
343
GCC @code{-v} will print the thread model GCC was configured to use.
344
However, this doesn't work on platforms that are multilibbed on thread
345
models, such as AIX 4.3.  On such platforms, define
346
@code{THREAD_MODEL_SPEC} such that it evaluates to a string without
347
blanks that names one of the recognized thread models.  @code{%*}, the
348
default value of this macro, will expand to the value of
349
@code{thread_file} set in @file{config.gcc}.
350
@end defmac
351
 
352
@defmac SYSROOT_SUFFIX_SPEC
353
Define this macro to add a suffix to the target sysroot when GCC is
354
configured with a sysroot.  This will cause GCC to search for usr/lib,
355
et al, within sysroot+suffix.
356
@end defmac
357
 
358
@defmac SYSROOT_HEADERS_SUFFIX_SPEC
359
Define this macro to add a headers_suffix to the target sysroot when
360
GCC is configured with a sysroot.  This will cause GCC to pass the
361
updated sysroot+headers_suffix to CPP, causing it to search for
362
usr/include, et al, within sysroot+headers_suffix.
363
@end defmac
364
 
365
@defmac EXTRA_SPECS
366
Define this macro to provide additional specifications to put in the
367
@file{specs} file that can be used in various specifications like
368
@code{CC1_SPEC}.
369
 
370
The definition should be an initializer for an array of structures,
371
containing a string constant, that defines the specification name, and a
372
string constant that provides the specification.
373
 
374
Do not define this macro if it does not need to do anything.
375
 
376
@code{EXTRA_SPECS} is useful when an architecture contains several
377
related targets, which have various @code{@dots{}_SPECS} which are similar
378
to each other, and the maintainer would like one central place to keep
379
these definitions.
380
 
381
For example, the PowerPC System V.4 targets use @code{EXTRA_SPECS} to
382
define either @code{_CALL_SYSV} when the System V calling sequence is
383
used or @code{_CALL_AIX} when the older AIX-based calling sequence is
384
used.
385
 
386
The @file{config/rs6000/rs6000.h} target file defines:
387
 
388
@smallexample
389
#define EXTRA_SPECS \
390
  @{ "cpp_sysv_default", CPP_SYSV_DEFAULT @},
391
 
392
#define CPP_SYS_DEFAULT ""
393
@end smallexample
394
 
395
The @file{config/rs6000/sysv.h} target file defines:
396
@smallexample
397
#undef CPP_SPEC
398
#define CPP_SPEC \
399
"%@{posix: -D_POSIX_SOURCE @} \
400
%@{mcall-sysv: -D_CALL_SYSV @} \
401
%@{!mcall-sysv: %(cpp_sysv_default) @} \
402
%@{msoft-float: -D_SOFT_FLOAT@} %@{mcpu=403: -D_SOFT_FLOAT@}"
403
 
404
#undef CPP_SYSV_DEFAULT
405
#define CPP_SYSV_DEFAULT "-D_CALL_SYSV"
406
@end smallexample
407
 
408
while the @file{config/rs6000/eabiaix.h} target file defines
409
@code{CPP_SYSV_DEFAULT} as:
410
 
411
@smallexample
412
#undef CPP_SYSV_DEFAULT
413
#define CPP_SYSV_DEFAULT "-D_CALL_AIX"
414
@end smallexample
415
@end defmac
416
 
417
@defmac LINK_LIBGCC_SPECIAL_1
418
Define this macro if the driver program should find the library
419
@file{libgcc.a}.  If you do not define this macro, the driver program will pass
420
the argument @option{-lgcc} to tell the linker to do the search.
421
@end defmac
422
 
423
@defmac LINK_GCC_C_SEQUENCE_SPEC
424
The sequence in which libgcc and libc are specified to the linker.
425
By default this is @code{%G %L %G}.
426
@end defmac
427
 
428
@defmac LINK_COMMAND_SPEC
429
A C string constant giving the complete command line need to execute the
430
linker.  When you do this, you will need to update your port each time a
431
change is made to the link command line within @file{gcc.c}.  Therefore,
432
define this macro only if you need to completely redefine the command
433
line for invoking the linker and there is no other way to accomplish
434
the effect you need.  Overriding this macro may be avoidable by overriding
435
@code{LINK_GCC_C_SEQUENCE_SPEC} instead.
436
@end defmac
437
 
438
@defmac LINK_ELIMINATE_DUPLICATE_LDIRECTORIES
439
A nonzero value causes @command{collect2} to remove duplicate @option{-L@var{directory}} search
440
directories from linking commands.  Do not give it a nonzero value if
441
removing duplicate search directories changes the linker's semantics.
442
@end defmac
443
 
444
@defmac MULTILIB_DEFAULTS
445
Define this macro as a C expression for the initializer of an array of
446
string to tell the driver program which options are defaults for this
447
target and thus do not need to be handled specially when using
448
@code{MULTILIB_OPTIONS}.
449
 
450
Do not define this macro if @code{MULTILIB_OPTIONS} is not defined in
451
the target makefile fragment or if none of the options listed in
452
@code{MULTILIB_OPTIONS} are set by default.
453
@xref{Target Fragment}.
454
@end defmac
455
 
456
@defmac RELATIVE_PREFIX_NOT_LINKDIR
457
Define this macro to tell @command{gcc} that it should only translate
458
a @option{-B} prefix into a @option{-L} linker option if the prefix
459
indicates an absolute file name.
460
@end defmac
461
 
462
@defmac MD_EXEC_PREFIX
463
If defined, this macro is an additional prefix to try after
464
@code{STANDARD_EXEC_PREFIX}.  @code{MD_EXEC_PREFIX} is not searched
465
when the @option{-b} option is used, or the compiler is built as a cross
466
compiler.  If you define @code{MD_EXEC_PREFIX}, then be sure to add it
467
to the list of directories used to find the assembler in @file{configure.in}.
468
@end defmac
469
 
470
@defmac STANDARD_STARTFILE_PREFIX
471
Define this macro as a C string constant if you wish to override the
472
standard choice of @code{libdir} as the default prefix to
473
try when searching for startup files such as @file{crt0.o}.
474
@code{STANDARD_STARTFILE_PREFIX} is not searched when the compiler
475
is built as a cross compiler.
476
@end defmac
477
 
478
@defmac STANDARD_STARTFILE_PREFIX_1
479
Define this macro as a C string constant if you wish to override the
480
standard choice of @code{/lib} as a prefix to try after the default prefix
481
when searching for startup files such as @file{crt0.o}.
482
@code{STANDARD_STARTFILE_PREFIX_1} is not searched when the compiler
483
is built as a cross compiler.
484
@end defmac
485
 
486
@defmac STANDARD_STARTFILE_PREFIX_2
487
Define this macro as a C string constant if you wish to override the
488
standard choice of @code{/lib} as yet another prefix to try after the
489
default prefix when searching for startup files such as @file{crt0.o}.
490
@code{STANDARD_STARTFILE_PREFIX_2} is not searched when the compiler
491
is built as a cross compiler.
492
@end defmac
493
 
494
@defmac MD_STARTFILE_PREFIX
495
If defined, this macro supplies an additional prefix to try after the
496
standard prefixes.  @code{MD_EXEC_PREFIX} is not searched when the
497
@option{-b} option is used, or when the compiler is built as a cross
498
compiler.
499
@end defmac
500
 
501
@defmac MD_STARTFILE_PREFIX_1
502
If defined, this macro supplies yet another prefix to try after the
503
standard prefixes.  It is not searched when the @option{-b} option is
504
used, or when the compiler is built as a cross compiler.
505
@end defmac
506
 
507
@defmac INIT_ENVIRONMENT
508
Define this macro as a C string constant if you wish to set environment
509
variables for programs called by the driver, such as the assembler and
510
loader.  The driver passes the value of this macro to @code{putenv} to
511
initialize the necessary environment variables.
512
@end defmac
513
 
514
@defmac LOCAL_INCLUDE_DIR
515
Define this macro as a C string constant if you wish to override the
516
standard choice of @file{/usr/local/include} as the default prefix to
517
try when searching for local header files.  @code{LOCAL_INCLUDE_DIR}
518
comes before @code{SYSTEM_INCLUDE_DIR} in the search order.
519
 
520
Cross compilers do not search either @file{/usr/local/include} or its
521
replacement.
522
@end defmac
523
 
524
@defmac MODIFY_TARGET_NAME
525
Define this macro if you wish to define command-line switches that
526
modify the default target name.
527
 
528
For each switch, you can include a string to be appended to the first
529
part of the configuration name or a string to be deleted from the
530
configuration name, if present.  The definition should be an initializer
531
for an array of structures.  Each array element should have three
532
elements: the switch name (a string constant, including the initial
533
dash), one of the enumeration codes @code{ADD} or @code{DELETE} to
534
indicate whether the string should be inserted or deleted, and the string
535
to be inserted or deleted (a string constant).
536
 
537
For example, on a machine where @samp{64} at the end of the
538
configuration name denotes a 64-bit target and you want the @option{-32}
539
and @option{-64} switches to select between 32- and 64-bit targets, you would
540
code
541
 
542
@smallexample
543
#define MODIFY_TARGET_NAME \
544
  @{ @{ "-32", DELETE, "64"@}, \
545
     @{"-64", ADD, "64"@}@}
546
@end smallexample
547
@end defmac
548
 
549
@defmac SYSTEM_INCLUDE_DIR
550
Define this macro as a C string constant if you wish to specify a
551
system-specific directory to search for header files before the standard
552
directory.  @code{SYSTEM_INCLUDE_DIR} comes before
553
@code{STANDARD_INCLUDE_DIR} in the search order.
554
 
555
Cross compilers do not use this macro and do not search the directory
556
specified.
557
@end defmac
558
 
559
@defmac STANDARD_INCLUDE_DIR
560
Define this macro as a C string constant if you wish to override the
561
standard choice of @file{/usr/include} as the default prefix to
562
try when searching for header files.
563
 
564
Cross compilers ignore this macro and do not search either
565
@file{/usr/include} or its replacement.
566
@end defmac
567
 
568
@defmac STANDARD_INCLUDE_COMPONENT
569
The ``component'' corresponding to @code{STANDARD_INCLUDE_DIR}.
570
See @code{INCLUDE_DEFAULTS}, below, for the description of components.
571
If you do not define this macro, no component is used.
572
@end defmac
573
 
574
@defmac INCLUDE_DEFAULTS
575
Define this macro if you wish to override the entire default search path
576
for include files.  For a native compiler, the default search path
577
usually consists of @code{GCC_INCLUDE_DIR}, @code{LOCAL_INCLUDE_DIR},
578
@code{SYSTEM_INCLUDE_DIR}, @code{GPLUSPLUS_INCLUDE_DIR}, and
579
@code{STANDARD_INCLUDE_DIR}.  In addition, @code{GPLUSPLUS_INCLUDE_DIR}
580
and @code{GCC_INCLUDE_DIR} are defined automatically by @file{Makefile},
581
and specify private search areas for GCC@.  The directory
582
@code{GPLUSPLUS_INCLUDE_DIR} is used only for C++ programs.
583
 
584
The definition should be an initializer for an array of structures.
585
Each array element should have four elements: the directory name (a
586
string constant), the component name (also a string constant), a flag
587
for C++-only directories,
588
and a flag showing that the includes in the directory don't need to be
589
wrapped in @code{extern @samp{C}} when compiling C++.  Mark the end of
590
the array with a null element.
591
 
592
The component name denotes what GNU package the include file is part of,
593
if any, in all uppercase letters.  For example, it might be @samp{GCC}
594
or @samp{BINUTILS}.  If the package is part of a vendor-supplied
595
operating system, code the component name as @samp{0}.
596
 
597
For example, here is the definition used for VAX/VMS:
598
 
599
@smallexample
600
#define INCLUDE_DEFAULTS \
601
@{                                       \
602
  @{ "GNU_GXX_INCLUDE:", "G++", 1, 1@},   \
603
  @{ "GNU_CC_INCLUDE:", "GCC", 0, 0@},    \
604
  @{ "SYS$SYSROOT:[SYSLIB.]", 0, 0, 0@},  \
605
  @{ ".", 0, 0, 0@},                      \
606
  @{ 0, 0, 0, 0@}                         \
607
@}
608
@end smallexample
609
@end defmac
610
 
611
Here is the order of prefixes tried for exec files:
612
 
613
@enumerate
614
@item
615
Any prefixes specified by the user with @option{-B}.
616
 
617
@item
618
The environment variable @code{GCC_EXEC_PREFIX}, if any.
619
 
620
@item
621
The directories specified by the environment variable @code{COMPILER_PATH}.
622
 
623
@item
624
The macro @code{STANDARD_EXEC_PREFIX}.
625
 
626
@item
627
@file{/usr/lib/gcc/}.
628
 
629
@item
630
The macro @code{MD_EXEC_PREFIX}, if any.
631
@end enumerate
632
 
633
Here is the order of prefixes tried for startfiles:
634
 
635
@enumerate
636
@item
637
Any prefixes specified by the user with @option{-B}.
638
 
639
@item
640
The environment variable @code{GCC_EXEC_PREFIX}, if any.
641
 
642
@item
643
The directories specified by the environment variable @code{LIBRARY_PATH}
644
(or port-specific name; native only, cross compilers do not use this).
645
 
646
@item
647
The macro @code{STANDARD_EXEC_PREFIX}.
648
 
649
@item
650
@file{/usr/lib/gcc/}.
651
 
652
@item
653
The macro @code{MD_EXEC_PREFIX}, if any.
654
 
655
@item
656
The macro @code{MD_STARTFILE_PREFIX}, if any.
657
 
658
@item
659
The macro @code{STANDARD_STARTFILE_PREFIX}.
660
 
661
@item
662
@file{/lib/}.
663
 
664
@item
665
@file{/usr/lib/}.
666
@end enumerate
667
 
668
@node Run-time Target
669
@section Run-time Target Specification
670
@cindex run-time target specification
671
@cindex predefined macros
672
@cindex target specifications
673
 
674
@c prevent bad page break with this line
675
Here are run-time target specifications.
676
 
677
@defmac TARGET_CPU_CPP_BUILTINS ()
678
This function-like macro expands to a block of code that defines
679
built-in preprocessor macros and assertions for the target cpu, using
680
the functions @code{builtin_define}, @code{builtin_define_std} and
681
@code{builtin_assert}.  When the front end
682
calls this macro it provides a trailing semicolon, and since it has
683
finished command line option processing your code can use those
684
results freely.
685
 
686
@code{builtin_assert} takes a string in the form you pass to the
687
command-line option @option{-A}, such as @code{cpu=mips}, and creates
688
the assertion.  @code{builtin_define} takes a string in the form
689
accepted by option @option{-D} and unconditionally defines the macro.
690
 
691
@code{builtin_define_std} takes a string representing the name of an
692
object-like macro.  If it doesn't lie in the user's namespace,
693
@code{builtin_define_std} defines it unconditionally.  Otherwise, it
694
defines a version with two leading underscores, and another version
695
with two leading and trailing underscores, and defines the original
696
only if an ISO standard was not requested on the command line.  For
697
example, passing @code{unix} defines @code{__unix}, @code{__unix__}
698
and possibly @code{unix}; passing @code{_mips} defines @code{__mips},
699
@code{__mips__} and possibly @code{_mips}, and passing @code{_ABI64}
700
defines only @code{_ABI64}.
701
 
702
You can also test for the C dialect being compiled.  The variable
703
@code{c_language} is set to one of @code{clk_c}, @code{clk_cplusplus}
704
or @code{clk_objective_c}.  Note that if we are preprocessing
705
assembler, this variable will be @code{clk_c} but the function-like
706
macro @code{preprocessing_asm_p()} will return true, so you might want
707
to check for that first.  If you need to check for strict ANSI, the
708
variable @code{flag_iso} can be used.  The function-like macro
709
@code{preprocessing_trad_p()} can be used to check for traditional
710
preprocessing.
711
@end defmac
712
 
713
@defmac TARGET_OS_CPP_BUILTINS ()
714
Similarly to @code{TARGET_CPU_CPP_BUILTINS} but this macro is optional
715
and is used for the target operating system instead.
716
@end defmac
717
 
718
@defmac TARGET_OBJFMT_CPP_BUILTINS ()
719
Similarly to @code{TARGET_CPU_CPP_BUILTINS} but this macro is optional
720
and is used for the target object format.  @file{elfos.h} uses this
721
macro to define @code{__ELF__}, so you probably do not need to define
722
it yourself.
723
@end defmac
724
 
725
@deftypevar {extern int} target_flags
726
This variable is declared in @file{options.h}, which is included before
727
any target-specific headers.
728
@end deftypevar
729
 
730
@deftypevar {Target Hook} int TARGET_DEFAULT_TARGET_FLAGS
731
This variable specifies the initial value of @code{target_flags}.
732
Its default setting is 0.
733
@end deftypevar
734
 
735
@cindex optional hardware or system features
736
@cindex features, optional, in system conventions
737
 
738
@deftypefn {Target Hook} bool TARGET_HANDLE_OPTION (size_t @var{code}, const char *@var{arg}, int @var{value})
739
This hook is called whenever the user specifies one of the
740
target-specific options described by the @file{.opt} definition files
741
(@pxref{Options}).  It has the opportunity to do some option-specific
742
processing and should return true if the option is valid.  The default
743
definition does nothing but return true.
744
 
745
@var{code} specifies the @code{OPT_@var{name}} enumeration value
746
associated with the selected option; @var{name} is just a rendering of
747
the option name in which non-alphanumeric characters are replaced by
748
underscores.  @var{arg} specifies the string argument and is null if
749
no argument was given.  If the option is flagged as a @code{UInteger}
750
(@pxref{Option properties}), @var{value} is the numeric value of the
751
argument.  Otherwise @var{value} is 1 if the positive form of the
752
option was used and 0 if the ``no-'' form was.
753
@end deftypefn
754
 
755
@defmac TARGET_VERSION
756
This macro is a C statement to print on @code{stderr} a string
757
describing the particular machine description choice.  Every machine
758
description should define @code{TARGET_VERSION}.  For example:
759
 
760
@smallexample
761
#ifdef MOTOROLA
762
#define TARGET_VERSION \
763
  fprintf (stderr, " (68k, Motorola syntax)");
764
#else
765
#define TARGET_VERSION \
766
  fprintf (stderr, " (68k, MIT syntax)");
767
#endif
768
@end smallexample
769
@end defmac
770
 
771
@defmac OVERRIDE_OPTIONS
772
Sometimes certain combinations of command options do not make sense on
773
a particular target machine.  You can define a macro
774
@code{OVERRIDE_OPTIONS} to take account of this.  This macro, if
775
defined, is executed once just after all the command options have been
776
parsed.
777
 
778
Don't use this macro to turn on various extra optimizations for
779
@option{-O}.  That is what @code{OPTIMIZATION_OPTIONS} is for.
780
@end defmac
781
 
782
@defmac C_COMMON_OVERRIDE_OPTIONS
783
This is similar to @code{OVERRIDE_OPTIONS} but is only used in the C
784
language frontends (C, Objective-C, C++, Objective-C++) and so can be
785
used to alter option flag variables which only exist in those
786
frontends.
787
@end defmac
788
 
789
@defmac OPTIMIZATION_OPTIONS (@var{level}, @var{size})
790
Some machines may desire to change what optimizations are performed for
791
various optimization levels.   This macro, if defined, is executed once
792
just after the optimization level is determined and before the remainder
793
of the command options have been parsed.  Values set in this macro are
794
used as the default values for the other command line options.
795
 
796
@var{level} is the optimization level specified; 2 if @option{-O2} is
797
specified, 1 if @option{-O} is specified, and 0 if neither is specified.
798
 
799
@var{size} is nonzero if @option{-Os} is specified and zero otherwise.
800
 
801
You should not use this macro to change options that are not
802
machine-specific.  These should uniformly selected by the same
803
optimization level on all supported machines.  Use this macro to enable
804
machine-specific optimizations.
805
 
806
@strong{Do not examine @code{write_symbols} in
807
this macro!} The debugging options are not supposed to alter the
808
generated code.
809
@end defmac
810
 
811
@defmac CAN_DEBUG_WITHOUT_FP
812
Define this macro if debugging can be performed even without a frame
813
pointer.  If this macro is defined, GCC will turn on the
814
@option{-fomit-frame-pointer} option whenever @option{-O} is specified.
815
@end defmac
816
 
817
@node Per-Function Data
818
@section Defining data structures for per-function information.
819
@cindex per-function data
820
@cindex data structures
821
 
822
If the target needs to store information on a per-function basis, GCC
823
provides a macro and a couple of variables to allow this.  Note, just
824
using statics to store the information is a bad idea, since GCC supports
825
nested functions, so you can be halfway through encoding one function
826
when another one comes along.
827
 
828
GCC defines a data structure called @code{struct function} which
829
contains all of the data specific to an individual function.  This
830
structure contains a field called @code{machine} whose type is
831
@code{struct machine_function *}, which can be used by targets to point
832
to their own specific data.
833
 
834
If a target needs per-function specific data it should define the type
835
@code{struct machine_function} and also the macro @code{INIT_EXPANDERS}.
836
This macro should be used to initialize the function pointer
837
@code{init_machine_status}.  This pointer is explained below.
838
 
839
One typical use of per-function, target specific data is to create an
840
RTX to hold the register containing the function's return address.  This
841
RTX can then be used to implement the @code{__builtin_return_address}
842
function, for level 0.
843
 
844
Note---earlier implementations of GCC used a single data area to hold
845
all of the per-function information.  Thus when processing of a nested
846
function began the old per-function data had to be pushed onto a
847
stack, and when the processing was finished, it had to be popped off the
848
stack.  GCC used to provide function pointers called
849
@code{save_machine_status} and @code{restore_machine_status} to handle
850
the saving and restoring of the target specific information.  Since the
851
single data area approach is no longer used, these pointers are no
852
longer supported.
853
 
854
@defmac INIT_EXPANDERS
855
Macro called to initialize any target specific information.  This macro
856
is called once per function, before generation of any RTL has begun.
857
The intention of this macro is to allow the initialization of the
858
function pointer @code{init_machine_status}.
859
@end defmac
860
 
861
@deftypevar {void (*)(struct function *)} init_machine_status
862
If this function pointer is non-@code{NULL} it will be called once per
863
function, before function compilation starts, in order to allow the
864
target to perform any target specific initialization of the
865
@code{struct function} structure.  It is intended that this would be
866
used to initialize the @code{machine} of that structure.
867
 
868
@code{struct machine_function} structures are expected to be freed by GC@.
869
Generally, any memory that they reference must be allocated by using
870
@code{ggc_alloc}, including the structure itself.
871
@end deftypevar
872
 
873
@node Storage Layout
874
@section Storage Layout
875
@cindex storage layout
876
 
877
Note that the definitions of the macros in this table which are sizes or
878
alignments measured in bits do not need to be constant.  They can be C
879
expressions that refer to static variables, such as the @code{target_flags}.
880
@xref{Run-time Target}.
881
 
882
@defmac BITS_BIG_ENDIAN
883
Define this macro to have the value 1 if the most significant bit in a
884
byte has the lowest number; otherwise define it to have the value zero.
885
This means that bit-field instructions count from the most significant
886
bit.  If the machine has no bit-field instructions, then this must still
887
be defined, but it doesn't matter which value it is defined to.  This
888
macro need not be a constant.
889
 
890
This macro does not affect the way structure fields are packed into
891
bytes or words; that is controlled by @code{BYTES_BIG_ENDIAN}.
892
@end defmac
893
 
894
@defmac BYTES_BIG_ENDIAN
895
Define this macro to have the value 1 if the most significant byte in a
896
word has the lowest number.  This macro need not be a constant.
897
@end defmac
898
 
899
@defmac WORDS_BIG_ENDIAN
900
Define this macro to have the value 1 if, in a multiword object, the
901
most significant word has the lowest number.  This applies to both
902
memory locations and registers; GCC fundamentally assumes that the
903
order of words in memory is the same as the order in registers.  This
904
macro need not be a constant.
905
@end defmac
906
 
907
@defmac LIBGCC2_WORDS_BIG_ENDIAN
908
Define this macro if @code{WORDS_BIG_ENDIAN} is not constant.  This must be a
909
constant value with the same meaning as @code{WORDS_BIG_ENDIAN}, which will be
910
used only when compiling @file{libgcc2.c}.  Typically the value will be set
911
based on preprocessor defines.
912
@end defmac
913
 
914
@defmac FLOAT_WORDS_BIG_ENDIAN
915
Define this macro to have the value 1 if @code{DFmode}, @code{XFmode} or
916
@code{TFmode} floating point numbers are stored in memory with the word
917
containing the sign bit at the lowest address; otherwise define it to
918
have the value 0.  This macro need not be a constant.
919
 
920
You need not define this macro if the ordering is the same as for
921
multi-word integers.
922
@end defmac
923
 
924
@defmac BITS_PER_UNIT
925
Define this macro to be the number of bits in an addressable storage
926
unit (byte).  If you do not define this macro the default is 8.
927
@end defmac
928
 
929
@defmac BITS_PER_WORD
930
Number of bits in a word.  If you do not define this macro, the default
931
is @code{BITS_PER_UNIT * UNITS_PER_WORD}.
932
@end defmac
933
 
934
@defmac MAX_BITS_PER_WORD
935
Maximum number of bits in a word.  If this is undefined, the default is
936
@code{BITS_PER_WORD}.  Otherwise, it is the constant value that is the
937
largest value that @code{BITS_PER_WORD} can have at run-time.
938
@end defmac
939
 
940
@defmac UNITS_PER_WORD
941
Number of storage units in a word; normally the size of a general-purpose
942
register, a power of two from 1 or 8.
943
@end defmac
944
 
945
@defmac MIN_UNITS_PER_WORD
946
Minimum number of units in a word.  If this is undefined, the default is
947
@code{UNITS_PER_WORD}.  Otherwise, it is the constant value that is the
948
smallest value that @code{UNITS_PER_WORD} can have at run-time.
949
@end defmac
950
 
951
@defmac UNITS_PER_SIMD_WORD
952
Number of units in the vectors that the vectorizer can produce.
953
The default is equal to @code{UNITS_PER_WORD}, because the vectorizer
954
can do some transformations even in absence of specialized @acronym{SIMD}
955
hardware.
956
@end defmac
957
 
958
@defmac POINTER_SIZE
959
Width of a pointer, in bits.  You must specify a value no wider than the
960
width of @code{Pmode}.  If it is not equal to the width of @code{Pmode},
961
you must define @code{POINTERS_EXTEND_UNSIGNED}.  If you do not specify
962
a value the default is @code{BITS_PER_WORD}.
963
@end defmac
964
 
965
@defmac POINTERS_EXTEND_UNSIGNED
966
A C expression whose value is greater than zero if pointers that need to be
967
extended from being @code{POINTER_SIZE} bits wide to @code{Pmode} are to
968
be zero-extended and zero if they are to be sign-extended.  If the value
969
is less then zero then there must be an "ptr_extend" instruction that
970
extends a pointer from @code{POINTER_SIZE} to @code{Pmode}.
971
 
972
You need not define this macro if the @code{POINTER_SIZE} is equal
973
to the width of @code{Pmode}.
974
@end defmac
975
 
976
@defmac PROMOTE_MODE (@var{m}, @var{unsignedp}, @var{type})
977
A macro to update @var{m} and @var{unsignedp} when an object whose type
978
is @var{type} and which has the specified mode and signedness is to be
979
stored in a register.  This macro is only called when @var{type} is a
980
scalar type.
981
 
982
On most RISC machines, which only have operations that operate on a full
983
register, define this macro to set @var{m} to @code{word_mode} if
984
@var{m} is an integer mode narrower than @code{BITS_PER_WORD}.  In most
985
cases, only integer modes should be widened because wider-precision
986
floating-point operations are usually more expensive than their narrower
987
counterparts.
988
 
989
For most machines, the macro definition does not change @var{unsignedp}.
990
However, some machines, have instructions that preferentially handle
991
either signed or unsigned quantities of certain modes.  For example, on
992
the DEC Alpha, 32-bit loads from memory and 32-bit add instructions
993
sign-extend the result to 64 bits.  On such machines, set
994
@var{unsignedp} according to which kind of extension is more efficient.
995
 
996
Do not define this macro if it would never modify @var{m}.
997
@end defmac
998
 
999
@defmac PROMOTE_FUNCTION_MODE
1000
Like @code{PROMOTE_MODE}, but is applied to outgoing function arguments or
1001
function return values, as specified by @code{TARGET_PROMOTE_FUNCTION_ARGS}
1002
and @code{TARGET_PROMOTE_FUNCTION_RETURN}, respectively.
1003
 
1004
The default is @code{PROMOTE_MODE}.
1005
@end defmac
1006
 
1007
@deftypefn {Target Hook} bool TARGET_PROMOTE_FUNCTION_ARGS (tree @var{fntype})
1008
This target hook should return @code{true} if the promotion described by
1009
@code{PROMOTE_FUNCTION_MODE} should be done for outgoing function
1010
arguments.
1011
@end deftypefn
1012
 
1013
@deftypefn {Target Hook} bool TARGET_PROMOTE_FUNCTION_RETURN (tree @var{fntype})
1014
This target hook should return @code{true} if the promotion described by
1015
@code{PROMOTE_FUNCTION_MODE} should be done for the return value of
1016
functions.
1017
 
1018
If this target hook returns @code{true}, @code{TARGET_FUNCTION_VALUE}
1019
must perform the same promotions done by @code{PROMOTE_FUNCTION_MODE}.
1020
@end deftypefn
1021
 
1022
@defmac PARM_BOUNDARY
1023
Normal alignment required for function parameters on the stack, in
1024
bits.  All stack parameters receive at least this much alignment
1025
regardless of data type.  On most machines, this is the same as the
1026
size of an integer.
1027
@end defmac
1028
 
1029
@defmac STACK_BOUNDARY
1030
Define this macro to the minimum alignment enforced by hardware for the
1031
stack pointer on this machine.  The definition is a C expression for the
1032
desired alignment (measured in bits).  This value is used as a default
1033
if @code{PREFERRED_STACK_BOUNDARY} is not defined.  On most machines,
1034
this should be the same as @code{PARM_BOUNDARY}.
1035
@end defmac
1036
 
1037
@defmac PREFERRED_STACK_BOUNDARY
1038
Define this macro if you wish to preserve a certain alignment for the
1039
stack pointer, greater than what the hardware enforces.  The definition
1040
is a C expression for the desired alignment (measured in bits).  This
1041
macro must evaluate to a value equal to or larger than
1042
@code{STACK_BOUNDARY}.
1043
@end defmac
1044
 
1045
@defmac FUNCTION_BOUNDARY
1046
Alignment required for a function entry point, in bits.
1047
@end defmac
1048
 
1049
@defmac BIGGEST_ALIGNMENT
1050
Biggest alignment that any data type can require on this machine, in bits.
1051
@end defmac
1052
 
1053
@defmac MINIMUM_ATOMIC_ALIGNMENT
1054
If defined, the smallest alignment, in bits, that can be given to an
1055
object that can be referenced in one operation, without disturbing any
1056
nearby object.  Normally, this is @code{BITS_PER_UNIT}, but may be larger
1057
on machines that don't have byte or half-word store operations.
1058
@end defmac
1059
 
1060
@defmac BIGGEST_FIELD_ALIGNMENT
1061
Biggest alignment that any structure or union field can require on this
1062
machine, in bits.  If defined, this overrides @code{BIGGEST_ALIGNMENT} for
1063
structure and union fields only, unless the field alignment has been set
1064
by the @code{__attribute__ ((aligned (@var{n})))} construct.
1065
@end defmac
1066
 
1067
@defmac ADJUST_FIELD_ALIGN (@var{field}, @var{computed})
1068
An expression for the alignment of a structure field @var{field} if the
1069
alignment computed in the usual way (including applying of
1070
@code{BIGGEST_ALIGNMENT} and @code{BIGGEST_FIELD_ALIGNMENT} to the
1071
alignment) is @var{computed}.  It overrides alignment only if the
1072
field alignment has not been set by the
1073
@code{__attribute__ ((aligned (@var{n})))} construct.
1074
@end defmac
1075
 
1076
@defmac MAX_OFILE_ALIGNMENT
1077
Biggest alignment supported by the object file format of this machine.
1078
Use this macro to limit the alignment which can be specified using the
1079
@code{__attribute__ ((aligned (@var{n})))} construct.  If not defined,
1080
the default value is @code{BIGGEST_ALIGNMENT}.
1081
@end defmac
1082
 
1083
@defmac DATA_ALIGNMENT (@var{type}, @var{basic-align})
1084
If defined, a C expression to compute the alignment for a variable in
1085
the static store.  @var{type} is the data type, and @var{basic-align} is
1086
the alignment that the object would ordinarily have.  The value of this
1087
macro is used instead of that alignment to align the object.
1088
 
1089
If this macro is not defined, then @var{basic-align} is used.
1090
 
1091
@findex strcpy
1092
One use of this macro is to increase alignment of medium-size data to
1093
make it all fit in fewer cache lines.  Another is to cause character
1094
arrays to be word-aligned so that @code{strcpy} calls that copy
1095
constants to character arrays can be done inline.
1096
@end defmac
1097
 
1098
@defmac CONSTANT_ALIGNMENT (@var{constant}, @var{basic-align})
1099
If defined, a C expression to compute the alignment given to a constant
1100
that is being placed in memory.  @var{constant} is the constant and
1101
@var{basic-align} is the alignment that the object would ordinarily
1102
have.  The value of this macro is used instead of that alignment to
1103
align the object.
1104
 
1105
If this macro is not defined, then @var{basic-align} is used.
1106
 
1107
The typical use of this macro is to increase alignment for string
1108
constants to be word aligned so that @code{strcpy} calls that copy
1109
constants can be done inline.
1110
@end defmac
1111
 
1112
@defmac LOCAL_ALIGNMENT (@var{type}, @var{basic-align})
1113
If defined, a C expression to compute the alignment for a variable in
1114
the local store.  @var{type} is the data type, and @var{basic-align} is
1115
the alignment that the object would ordinarily have.  The value of this
1116
macro is used instead of that alignment to align the object.
1117
 
1118
If this macro is not defined, then @var{basic-align} is used.
1119
 
1120
One use of this macro is to increase alignment of medium-size data to
1121
make it all fit in fewer cache lines.
1122
@end defmac
1123
 
1124
@defmac EMPTY_FIELD_BOUNDARY
1125
Alignment in bits to be given to a structure bit-field that follows an
1126
empty field such as @code{int : 0;}.
1127
 
1128
If @code{PCC_BITFIELD_TYPE_MATTERS} is true, it overrides this macro.
1129
@end defmac
1130
 
1131
@defmac STRUCTURE_SIZE_BOUNDARY
1132
Number of bits which any structure or union's size must be a multiple of.
1133
Each structure or union's size is rounded up to a multiple of this.
1134
 
1135
If you do not define this macro, the default is the same as
1136
@code{BITS_PER_UNIT}.
1137
@end defmac
1138
 
1139
@defmac STRICT_ALIGNMENT
1140
Define this macro to be the value 1 if instructions will fail to work
1141
if given data not on the nominal alignment.  If instructions will merely
1142
go slower in that case, define this macro as 0.
1143
@end defmac
1144
 
1145
@defmac PCC_BITFIELD_TYPE_MATTERS
1146
Define this if you wish to imitate the way many other C compilers handle
1147
alignment of bit-fields and the structures that contain them.
1148
 
1149
The behavior is that the type written for a named bit-field (@code{int},
1150
@code{short}, or other integer type) imposes an alignment for the entire
1151
structure, as if the structure really did contain an ordinary field of
1152
that type.  In addition, the bit-field is placed within the structure so
1153
that it would fit within such a field, not crossing a boundary for it.
1154
 
1155
Thus, on most machines, a named bit-field whose type is written as
1156
@code{int} would not cross a four-byte boundary, and would force
1157
four-byte alignment for the whole structure.  (The alignment used may
1158
not be four bytes; it is controlled by the other alignment parameters.)
1159
 
1160
An unnamed bit-field will not affect the alignment of the containing
1161
structure.
1162
 
1163
If the macro is defined, its definition should be a C expression;
1164
a nonzero value for the expression enables this behavior.
1165
 
1166
Note that if this macro is not defined, or its value is zero, some
1167
bit-fields may cross more than one alignment boundary.  The compiler can
1168
support such references if there are @samp{insv}, @samp{extv}, and
1169
@samp{extzv} insns that can directly reference memory.
1170
 
1171
The other known way of making bit-fields work is to define
1172
@code{STRUCTURE_SIZE_BOUNDARY} as large as @code{BIGGEST_ALIGNMENT}.
1173
Then every structure can be accessed with fullwords.
1174
 
1175
Unless the machine has bit-field instructions or you define
1176
@code{STRUCTURE_SIZE_BOUNDARY} that way, you must define
1177
@code{PCC_BITFIELD_TYPE_MATTERS} to have a nonzero value.
1178
 
1179
If your aim is to make GCC use the same conventions for laying out
1180
bit-fields as are used by another compiler, here is how to investigate
1181
what the other compiler does.  Compile and run this program:
1182
 
1183
@smallexample
1184
struct foo1
1185
@{
1186
  char x;
1187
  char :0;
1188
  char y;
1189
@};
1190
 
1191
struct foo2
1192
@{
1193
  char x;
1194
  int :0;
1195
  char y;
1196
@};
1197
 
1198
main ()
1199
@{
1200
  printf ("Size of foo1 is %d\n",
1201
          sizeof (struct foo1));
1202
  printf ("Size of foo2 is %d\n",
1203
          sizeof (struct foo2));
1204
  exit (0);
1205
@}
1206
@end smallexample
1207
 
1208
If this prints 2 and 5, then the compiler's behavior is what you would
1209
get from @code{PCC_BITFIELD_TYPE_MATTERS}.
1210
@end defmac
1211
 
1212
@defmac BITFIELD_NBYTES_LIMITED
1213
Like @code{PCC_BITFIELD_TYPE_MATTERS} except that its effect is limited
1214
to aligning a bit-field within the structure.
1215
@end defmac
1216
 
1217
@deftypefn {Target Hook} bool TARGET_ALIGN_ANON_BITFIELDS (void)
1218
When @code{PCC_BITFIELD_TYPE_MATTERS} is true this hook will determine
1219
whether unnamed bitfields affect the alignment of the containing
1220
structure.  The hook should return true if the structure should inherit
1221
the alignment requirements of an unnamed bitfield's type.
1222
@end deftypefn
1223
 
1224
@deftypefn {Target Hook} bool TARGET_NARROW_VOLATILE_BITFIELDS (void)
1225
This target hook should return @code{true} if accesses to volatile bitfields
1226
should use the narrowest mode possible.  It should return @code{false} if
1227
these accesses should use the bitfield container type.
1228
 
1229
The default is @code{!TARGET_STRICT_ALIGN}.
1230
@end deftypefn
1231
 
1232
@defmac MEMBER_TYPE_FORCES_BLK (@var{field}, @var{mode})
1233
Return 1 if a structure or array containing @var{field} should be accessed using
1234
@code{BLKMODE}.
1235
 
1236
If @var{field} is the only field in the structure, @var{mode} is its
1237
mode, otherwise @var{mode} is VOIDmode.  @var{mode} is provided in the
1238
case where structures of one field would require the structure's mode to
1239
retain the field's mode.
1240
 
1241
Normally, this is not needed.  See the file @file{c4x.h} for an example
1242
of how to use this macro to prevent a structure having a floating point
1243
field from being accessed in an integer mode.
1244
@end defmac
1245
 
1246
@defmac ROUND_TYPE_ALIGN (@var{type}, @var{computed}, @var{specified})
1247
Define this macro as an expression for the alignment of a type (given
1248
by @var{type} as a tree node) if the alignment computed in the usual
1249
way is @var{computed} and the alignment explicitly specified was
1250
@var{specified}.
1251
 
1252
The default is to use @var{specified} if it is larger; otherwise, use
1253
the smaller of @var{computed} and @code{BIGGEST_ALIGNMENT}
1254
@end defmac
1255
 
1256
@defmac MAX_FIXED_MODE_SIZE
1257
An integer expression for the size in bits of the largest integer
1258
machine mode that should actually be used.  All integer machine modes of
1259
this size or smaller can be used for structures and unions with the
1260
appropriate sizes.  If this macro is undefined, @code{GET_MODE_BITSIZE
1261
(DImode)} is assumed.
1262
@end defmac
1263
 
1264
@defmac STACK_SAVEAREA_MODE (@var{save_level})
1265
If defined, an expression of type @code{enum machine_mode} that
1266
specifies the mode of the save area operand of a
1267
@code{save_stack_@var{level}} named pattern (@pxref{Standard Names}).
1268
@var{save_level} is one of @code{SAVE_BLOCK}, @code{SAVE_FUNCTION}, or
1269
@code{SAVE_NONLOCAL} and selects which of the three named patterns is
1270
having its mode specified.
1271
 
1272
You need not define this macro if it always returns @code{Pmode}.  You
1273
would most commonly define this macro if the
1274
@code{save_stack_@var{level}} patterns need to support both a 32- and a
1275
64-bit mode.
1276
@end defmac
1277
 
1278
@defmac STACK_SIZE_MODE
1279
If defined, an expression of type @code{enum machine_mode} that
1280
specifies the mode of the size increment operand of an
1281
@code{allocate_stack} named pattern (@pxref{Standard Names}).
1282
 
1283
You need not define this macro if it always returns @code{word_mode}.
1284
You would most commonly define this macro if the @code{allocate_stack}
1285
pattern needs to support both a 32- and a 64-bit mode.
1286
@end defmac
1287
 
1288
@defmac TARGET_FLOAT_FORMAT
1289
A code distinguishing the floating point format of the target machine.
1290
There are four defined values:
1291
 
1292
@ftable @code
1293
@item IEEE_FLOAT_FORMAT
1294
This code indicates IEEE floating point.  It is the default; there is no
1295
need to define @code{TARGET_FLOAT_FORMAT} when the format is IEEE@.
1296
 
1297
@item VAX_FLOAT_FORMAT
1298
This code indicates the ``F float'' (for @code{float}) and ``D float''
1299
or ``G float'' formats (for @code{double}) used on the VAX and PDP-11@.
1300
 
1301
@item IBM_FLOAT_FORMAT
1302
This code indicates the format used on the IBM System/370.
1303
 
1304
@item C4X_FLOAT_FORMAT
1305
This code indicates the format used on the TMS320C3x/C4x.
1306
@end ftable
1307
 
1308
If your target uses a floating point format other than these, you must
1309
define a new @var{name}_FLOAT_FORMAT code for it, and add support for
1310
it to @file{real.c}.
1311
 
1312
The ordering of the component words of floating point values stored in
1313
memory is controlled by @code{FLOAT_WORDS_BIG_ENDIAN}.
1314
@end defmac
1315
 
1316
@defmac MODE_HAS_NANS (@var{mode})
1317
When defined, this macro should be true if @var{mode} has a NaN
1318
representation.  The compiler assumes that NaNs are not equal to
1319
anything (including themselves) and that addition, subtraction,
1320
multiplication and division all return NaNs when one operand is
1321
NaN@.
1322
 
1323
By default, this macro is true if @var{mode} is a floating-point
1324
mode and the target floating-point format is IEEE@.
1325
@end defmac
1326
 
1327
@defmac MODE_HAS_INFINITIES (@var{mode})
1328
This macro should be true if @var{mode} can represent infinity.  At
1329
present, the compiler uses this macro to decide whether @samp{x - x}
1330
is always defined.  By default, the macro is true when @var{mode}
1331
is a floating-point mode and the target format is IEEE@.
1332
@end defmac
1333
 
1334
@defmac MODE_HAS_SIGNED_ZEROS (@var{mode})
1335
True if @var{mode} distinguishes between positive and negative zero.
1336
The rules are expected to follow the IEEE standard:
1337
 
1338
@itemize @bullet
1339
@item
1340
@samp{x + x} has the same sign as @samp{x}.
1341
 
1342
@item
1343
If the sum of two values with opposite sign is zero, the result is
1344
positive for all rounding modes expect towards @minus{}infinity, for
1345
which it is negative.
1346
 
1347
@item
1348
The sign of a product or quotient is negative when exactly one
1349
of the operands is negative.
1350
@end itemize
1351
 
1352
The default definition is true if @var{mode} is a floating-point
1353
mode and the target format is IEEE@.
1354
@end defmac
1355
 
1356
@defmac MODE_HAS_SIGN_DEPENDENT_ROUNDING (@var{mode})
1357
If defined, this macro should be true for @var{mode} if it has at
1358
least one rounding mode in which @samp{x} and @samp{-x} can be
1359
rounded to numbers of different magnitude.  Two such modes are
1360
towards @minus{}infinity and towards +infinity.
1361
 
1362
The default definition of this macro is true if @var{mode} is
1363
a floating-point mode and the target format is IEEE@.
1364
@end defmac
1365
 
1366
@defmac ROUND_TOWARDS_ZERO
1367
If defined, this macro should be true if the prevailing rounding
1368
mode is towards zero.  A true value has the following effects:
1369
 
1370
@itemize @bullet
1371
@item
1372
@code{MODE_HAS_SIGN_DEPENDENT_ROUNDING} will be false for all modes.
1373
 
1374
@item
1375
@file{libgcc.a}'s floating-point emulator will round towards zero
1376
rather than towards nearest.
1377
 
1378
@item
1379
The compiler's floating-point emulator will round towards zero after
1380
doing arithmetic, and when converting from the internal float format to
1381
the target format.
1382
@end itemize
1383
 
1384
The macro does not affect the parsing of string literals.  When the
1385
primary rounding mode is towards zero, library functions like
1386
@code{strtod} might still round towards nearest, and the compiler's
1387
parser should behave like the target's @code{strtod} where possible.
1388
 
1389
Not defining this macro is equivalent to returning zero.
1390
@end defmac
1391
 
1392
@defmac LARGEST_EXPONENT_IS_NORMAL (@var{size})
1393
This macro should return true if floats with @var{size}
1394
bits do not have a NaN or infinity representation, but use the largest
1395
exponent for normal numbers instead.
1396
 
1397
Defining this macro to true for @var{size} causes @code{MODE_HAS_NANS}
1398
and @code{MODE_HAS_INFINITIES} to be false for @var{size}-bit modes.
1399
It also affects the way @file{libgcc.a} and @file{real.c} emulate
1400
floating-point arithmetic.
1401
 
1402
The default definition of this macro returns false for all sizes.
1403
@end defmac
1404
 
1405
@deftypefn {Target Hook} bool TARGET_VECTOR_OPAQUE_P (tree @var{type})
1406
This target hook should return @code{true} a vector is opaque.  That
1407
is, if no cast is needed when copying a vector value of type
1408
@var{type} into another vector lvalue of the same size.  Vector opaque
1409
types cannot be initialized.  The default is that there are no such
1410
types.
1411
@end deftypefn
1412
 
1413
@deftypefn {Target Hook} bool TARGET_MS_BITFIELD_LAYOUT_P (tree @var{record_type})
1414
This target hook returns @code{true} if bit-fields in the given
1415
@var{record_type} are to be laid out following the rules of Microsoft
1416
Visual C/C++, namely: (i) a bit-field won't share the same storage
1417
unit with the previous bit-field if their underlying types have
1418
different sizes, and the bit-field will be aligned to the highest
1419
alignment of the underlying types of itself and of the previous
1420
bit-field; (ii) a zero-sized bit-field will affect the alignment of
1421
the whole enclosing structure, even if it is unnamed; except that
1422
(iii) a zero-sized bit-field will be disregarded unless it follows
1423
another bit-field of nonzero size.  If this hook returns @code{true},
1424
other macros that control bit-field layout are ignored.
1425
 
1426
When a bit-field is inserted into a packed record, the whole size
1427
of the underlying type is used by one or more same-size adjacent
1428
bit-fields (that is, if its long:3, 32 bits is used in the record,
1429
and any additional adjacent long bit-fields are packed into the same
1430
chunk of 32 bits.  However, if the size changes, a new field of that
1431
size is allocated).  In an unpacked record, this is the same as using
1432
alignment, but not equivalent when packing.
1433
 
1434
If both MS bit-fields and @samp{__attribute__((packed))} are used,
1435
the latter will take precedence.  If @samp{__attribute__((packed))} is
1436
used on a single field when MS bit-fields are in use, it will take
1437
precedence for that field, but the alignment of the rest of the structure
1438
may affect its placement.
1439
@end deftypefn
1440
 
1441
@deftypefn {Target Hook} {bool} TARGET_DECIMAL_FLOAT_SUPPORTED_P (void)
1442
Returns true if the target supports decimal floating point.
1443
@end deftypefn
1444
 
1445
@deftypefn {Target Hook} {const char *} TARGET_MANGLE_FUNDAMENTAL_TYPE (tree @var{type})
1446
If your target defines any fundamental types, define this hook to
1447
return the appropriate encoding for these types as part of a C++
1448
mangled name.  The @var{type} argument is the tree structure
1449
representing the type to be mangled.  The hook may be applied to trees
1450
which are not target-specific fundamental types; it should return
1451
@code{NULL} for all such types, as well as arguments it does not
1452
recognize.  If the return value is not @code{NULL}, it must point to
1453
a statically-allocated string constant.
1454
 
1455
Target-specific fundamental types might be new fundamental types or
1456
qualified versions of ordinary fundamental types.  Encode new
1457
fundamental types as @samp{@w{u @var{n} @var{name}}}, where @var{name}
1458
is the name used for the type in source code, and @var{n} is the
1459
length of @var{name} in decimal.  Encode qualified versions of
1460
ordinary types as @samp{@w{U @var{n} @var{name} @var{code}}}, where
1461
@var{name} is the name used for the type qualifier in source code,
1462
@var{n} is the length of @var{name} as above, and @var{code} is the
1463
code used to represent the unqualified version of this type.  (See
1464
@code{write_builtin_type} in @file{cp/mangle.c} for the list of
1465
codes.)  In both cases the spaces are for clarity; do not include any
1466
spaces in your string.
1467
 
1468
The default version of this hook always returns @code{NULL}, which is
1469
appropriate for a target that does not define any new fundamental
1470
types.
1471
@end deftypefn
1472
 
1473
@node Type Layout
1474
@section Layout of Source Language Data Types
1475
 
1476
These macros define the sizes and other characteristics of the standard
1477
basic data types used in programs being compiled.  Unlike the macros in
1478
the previous section, these apply to specific features of C and related
1479
languages, rather than to fundamental aspects of storage layout.
1480
 
1481
@defmac INT_TYPE_SIZE
1482
A C expression for the size in bits of the type @code{int} on the
1483
target machine.  If you don't define this, the default is one word.
1484
@end defmac
1485
 
1486
@defmac SHORT_TYPE_SIZE
1487
A C expression for the size in bits of the type @code{short} on the
1488
target machine.  If you don't define this, the default is half a word.
1489
(If this would be less than one storage unit, it is rounded up to one
1490
unit.)
1491
@end defmac
1492
 
1493
@defmac LONG_TYPE_SIZE
1494
A C expression for the size in bits of the type @code{long} on the
1495
target machine.  If you don't define this, the default is one word.
1496
@end defmac
1497
 
1498
@defmac ADA_LONG_TYPE_SIZE
1499
On some machines, the size used for the Ada equivalent of the type
1500
@code{long} by a native Ada compiler differs from that used by C@.  In
1501
that situation, define this macro to be a C expression to be used for
1502
the size of that type.  If you don't define this, the default is the
1503
value of @code{LONG_TYPE_SIZE}.
1504
@end defmac
1505
 
1506
@defmac LONG_LONG_TYPE_SIZE
1507
A C expression for the size in bits of the type @code{long long} on the
1508
target machine.  If you don't define this, the default is two
1509
words.  If you want to support GNU Ada on your machine, the value of this
1510
macro must be at least 64.
1511
@end defmac
1512
 
1513
@defmac CHAR_TYPE_SIZE
1514
A C expression for the size in bits of the type @code{char} on the
1515
target machine.  If you don't define this, the default is
1516
@code{BITS_PER_UNIT}.
1517
@end defmac
1518
 
1519
@defmac BOOL_TYPE_SIZE
1520
A C expression for the size in bits of the C++ type @code{bool} and
1521
C99 type @code{_Bool} on the target machine.  If you don't define
1522
this, and you probably shouldn't, the default is @code{CHAR_TYPE_SIZE}.
1523
@end defmac
1524
 
1525
@defmac FLOAT_TYPE_SIZE
1526
A C expression for the size in bits of the type @code{float} on the
1527
target machine.  If you don't define this, the default is one word.
1528
@end defmac
1529
 
1530
@defmac DOUBLE_TYPE_SIZE
1531
A C expression for the size in bits of the type @code{double} on the
1532
target machine.  If you don't define this, the default is two
1533
words.
1534
@end defmac
1535
 
1536
@defmac LONG_DOUBLE_TYPE_SIZE
1537
A C expression for the size in bits of the type @code{long double} on
1538
the target machine.  If you don't define this, the default is two
1539
words.
1540
@end defmac
1541
 
1542
@defmac LIBGCC2_LONG_DOUBLE_TYPE_SIZE
1543
Define this macro if @code{LONG_DOUBLE_TYPE_SIZE} is not constant or
1544
if you want routines in @file{libgcc2.a} for a size other than
1545
@code{LONG_DOUBLE_TYPE_SIZE}.  If you don't define this, the
1546
default is @code{LONG_DOUBLE_TYPE_SIZE}.
1547
@end defmac
1548
 
1549
@defmac LIBGCC2_HAS_DF_MODE
1550
Define this macro if neither @code{LIBGCC2_DOUBLE_TYPE_SIZE} nor
1551
@code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE} is
1552
@code{DFmode} but you want @code{DFmode} routines in @file{libgcc2.a}
1553
anyway.  If you don't define this and either @code{LIBGCC2_DOUBLE_TYPE_SIZE}
1554
or @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE} is 64 then the default is 1,
1555
otherwise it is 0.
1556
@end defmac
1557
 
1558
@defmac LIBGCC2_HAS_XF_MODE
1559
Define this macro if @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE} is not
1560
@code{XFmode} but you want @code{XFmode} routines in @file{libgcc2.a}
1561
anyway.  If you don't define this and @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE}
1562
is 80 then the default is 1, otherwise it is 0.
1563
@end defmac
1564
 
1565
@defmac LIBGCC2_HAS_TF_MODE
1566
Define this macro if @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE} is not
1567
@code{TFmode} but you want @code{TFmode} routines in @file{libgcc2.a}
1568
anyway.  If you don't define this and @code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE}
1569
is 128 then the default is 1, otherwise it is 0.
1570
@end defmac
1571
 
1572
@defmac SF_SIZE
1573
@defmacx DF_SIZE
1574
@defmacx XF_SIZE
1575
@defmacx TF_SIZE
1576
Define these macros to be the size in bits of the mantissa of
1577
@code{SFmode}, @code{DFmode}, @code{XFmode} and @code{TFmode} values,
1578
if the defaults in @file{libgcc2.h} are inappropriate.  By default,
1579
@code{FLT_MANT_DIG} is used for @code{SF_SIZE}, @code{LDBL_MANT_DIG}
1580
for @code{XF_SIZE} and @code{TF_SIZE}, and @code{DBL_MANT_DIG} or
1581
@code{LDBL_MANT_DIG} for @code{DF_SIZE} according to whether
1582
@code{LIBGCC2_DOUBLE_TYPE_SIZE} or
1583
@code{LIBGCC2_LONG_DOUBLE_TYPE_SIZE} is 64.
1584
@end defmac
1585
 
1586
@defmac TARGET_FLT_EVAL_METHOD
1587
A C expression for the value for @code{FLT_EVAL_METHOD} in @file{float.h},
1588
assuming, if applicable, that the floating-point control word is in its
1589
default state.  If you do not define this macro the value of
1590
@code{FLT_EVAL_METHOD} will be zero.
1591
@end defmac
1592
 
1593
@defmac WIDEST_HARDWARE_FP_SIZE
1594
A C expression for the size in bits of the widest floating-point format
1595
supported by the hardware.  If you define this macro, you must specify a
1596
value less than or equal to the value of @code{LONG_DOUBLE_TYPE_SIZE}.
1597
If you do not define this macro, the value of @code{LONG_DOUBLE_TYPE_SIZE}
1598
is the default.
1599
@end defmac
1600
 
1601
@defmac DEFAULT_SIGNED_CHAR
1602
An expression whose value is 1 or 0, according to whether the type
1603
@code{char} should be signed or unsigned by default.  The user can
1604
always override this default with the options @option{-fsigned-char}
1605
and @option{-funsigned-char}.
1606
@end defmac
1607
 
1608
@deftypefn {Target Hook} bool TARGET_DEFAULT_SHORT_ENUMS (void)
1609
This target hook should return true if the compiler should give an
1610
@code{enum} type only as many bytes as it takes to represent the range
1611
of possible values of that type.  It should return false if all
1612
@code{enum} types should be allocated like @code{int}.
1613
 
1614
The default is to return false.
1615
@end deftypefn
1616
 
1617
@defmac SIZE_TYPE
1618
A C expression for a string describing the name of the data type to use
1619
for size values.  The typedef name @code{size_t} is defined using the
1620
contents of the string.
1621
 
1622
The string can contain more than one keyword.  If so, separate them with
1623
spaces, and write first any length keyword, then @code{unsigned} if
1624
appropriate, and finally @code{int}.  The string must exactly match one
1625
of the data type names defined in the function
1626
@code{init_decl_processing} in the file @file{c-decl.c}.  You may not
1627
omit @code{int} or change the order---that would cause the compiler to
1628
crash on startup.
1629
 
1630
If you don't define this macro, the default is @code{"long unsigned
1631
int"}.
1632
@end defmac
1633
 
1634
@defmac PTRDIFF_TYPE
1635
A C expression for a string describing the name of the data type to use
1636
for the result of subtracting two pointers.  The typedef name
1637
@code{ptrdiff_t} is defined using the contents of the string.  See
1638
@code{SIZE_TYPE} above for more information.
1639
 
1640
If you don't define this macro, the default is @code{"long int"}.
1641
@end defmac
1642
 
1643
@defmac WCHAR_TYPE
1644
A C expression for a string describing the name of the data type to use
1645
for wide characters.  The typedef name @code{wchar_t} is defined using
1646
the contents of the string.  See @code{SIZE_TYPE} above for more
1647
information.
1648
 
1649
If you don't define this macro, the default is @code{"int"}.
1650
@end defmac
1651
 
1652
@defmac WCHAR_TYPE_SIZE
1653
A C expression for the size in bits of the data type for wide
1654
characters.  This is used in @code{cpp}, which cannot make use of
1655
@code{WCHAR_TYPE}.
1656
@end defmac
1657
 
1658
@defmac WINT_TYPE
1659
A C expression for a string describing the name of the data type to
1660
use for wide characters passed to @code{printf} and returned from
1661
@code{getwc}.  The typedef name @code{wint_t} is defined using the
1662
contents of the string.  See @code{SIZE_TYPE} above for more
1663
information.
1664
 
1665
If you don't define this macro, the default is @code{"unsigned int"}.
1666
@end defmac
1667
 
1668
@defmac INTMAX_TYPE
1669
A C expression for a string describing the name of the data type that
1670
can represent any value of any standard or extended signed integer type.
1671
The typedef name @code{intmax_t} is defined using the contents of the
1672
string.  See @code{SIZE_TYPE} above for more information.
1673
 
1674
If you don't define this macro, the default is the first of
1675
@code{"int"}, @code{"long int"}, or @code{"long long int"} that has as
1676
much precision as @code{long long int}.
1677
@end defmac
1678
 
1679
@defmac UINTMAX_TYPE
1680
A C expression for a string describing the name of the data type that
1681
can represent any value of any standard or extended unsigned integer
1682
type.  The typedef name @code{uintmax_t} is defined using the contents
1683
of the string.  See @code{SIZE_TYPE} above for more information.
1684
 
1685
If you don't define this macro, the default is the first of
1686
@code{"unsigned int"}, @code{"long unsigned int"}, or @code{"long long
1687
unsigned int"} that has as much precision as @code{long long unsigned
1688
int}.
1689
@end defmac
1690
 
1691
@defmac TARGET_PTRMEMFUNC_VBIT_LOCATION
1692
The C++ compiler represents a pointer-to-member-function with a struct
1693
that looks like:
1694
 
1695
@smallexample
1696
  struct @{
1697
    union @{
1698
      void (*fn)();
1699
      ptrdiff_t vtable_index;
1700
    @};
1701
    ptrdiff_t delta;
1702
  @};
1703
@end smallexample
1704
 
1705
@noindent
1706
The C++ compiler must use one bit to indicate whether the function that
1707
will be called through a pointer-to-member-function is virtual.
1708
Normally, we assume that the low-order bit of a function pointer must
1709
always be zero.  Then, by ensuring that the vtable_index is odd, we can
1710
distinguish which variant of the union is in use.  But, on some
1711
platforms function pointers can be odd, and so this doesn't work.  In
1712
that case, we use the low-order bit of the @code{delta} field, and shift
1713
the remainder of the @code{delta} field to the left.
1714
 
1715
GCC will automatically make the right selection about where to store
1716
this bit using the @code{FUNCTION_BOUNDARY} setting for your platform.
1717
However, some platforms such as ARM/Thumb have @code{FUNCTION_BOUNDARY}
1718
set such that functions always start at even addresses, but the lowest
1719
bit of pointers to functions indicate whether the function at that
1720
address is in ARM or Thumb mode.  If this is the case of your
1721
architecture, you should define this macro to
1722
@code{ptrmemfunc_vbit_in_delta}.
1723
 
1724
In general, you should not have to define this macro.  On architectures
1725
in which function addresses are always even, according to
1726
@code{FUNCTION_BOUNDARY}, GCC will automatically define this macro to
1727
@code{ptrmemfunc_vbit_in_pfn}.
1728
@end defmac
1729
 
1730
@defmac TARGET_VTABLE_USES_DESCRIPTORS
1731
Normally, the C++ compiler uses function pointers in vtables.  This
1732
macro allows the target to change to use ``function descriptors''
1733
instead.  Function descriptors are found on targets for whom a
1734
function pointer is actually a small data structure.  Normally the
1735
data structure consists of the actual code address plus a data
1736
pointer to which the function's data is relative.
1737
 
1738
If vtables are used, the value of this macro should be the number
1739
of words that the function descriptor occupies.
1740
@end defmac
1741
 
1742
@defmac TARGET_VTABLE_ENTRY_ALIGN
1743
By default, the vtable entries are void pointers, the so the alignment
1744
is the same as pointer alignment.  The value of this macro specifies
1745
the alignment of the vtable entry in bits.  It should be defined only
1746
when special alignment is necessary. */
1747
@end defmac
1748
 
1749
@defmac TARGET_VTABLE_DATA_ENTRY_DISTANCE
1750
There are a few non-descriptor entries in the vtable at offsets below
1751
zero.  If these entries must be padded (say, to preserve the alignment
1752
specified by @code{TARGET_VTABLE_ENTRY_ALIGN}), set this to the number
1753
of words in each data entry.
1754
@end defmac
1755
 
1756
@node Registers
1757
@section Register Usage
1758
@cindex register usage
1759
 
1760
This section explains how to describe what registers the target machine
1761
has, and how (in general) they can be used.
1762
 
1763
The description of which registers a specific instruction can use is
1764
done with register classes; see @ref{Register Classes}.  For information
1765
on using registers to access a stack frame, see @ref{Frame Registers}.
1766
For passing values in registers, see @ref{Register Arguments}.
1767
For returning values in registers, see @ref{Scalar Return}.
1768
 
1769
@menu
1770
* Register Basics::             Number and kinds of registers.
1771
* Allocation Order::            Order in which registers are allocated.
1772
* Values in Registers::         What kinds of values each reg can hold.
1773
* Leaf Functions::              Renumbering registers for leaf functions.
1774
* Stack Registers::             Handling a register stack such as 80387.
1775
@end menu
1776
 
1777
@node Register Basics
1778
@subsection Basic Characteristics of Registers
1779
 
1780
@c prevent bad page break with this line
1781
Registers have various characteristics.
1782
 
1783
@defmac FIRST_PSEUDO_REGISTER
1784
Number of hardware registers known to the compiler.  They receive
1785
numbers 0 through @code{FIRST_PSEUDO_REGISTER-1}; thus, the first
1786
pseudo register's number really is assigned the number
1787
@code{FIRST_PSEUDO_REGISTER}.
1788
@end defmac
1789
 
1790
@defmac FIXED_REGISTERS
1791
@cindex fixed register
1792
An initializer that says which registers are used for fixed purposes
1793
all throughout the compiled code and are therefore not available for
1794
general allocation.  These would include the stack pointer, the frame
1795
pointer (except on machines where that can be used as a general
1796
register when no frame pointer is needed), the program counter on
1797
machines where that is considered one of the addressable registers,
1798
and any other numbered register with a standard use.
1799
 
1800
This information is expressed as a sequence of numbers, separated by
1801
commas and surrounded by braces.  The @var{n}th number is 1 if
1802
register @var{n} is fixed, 0 otherwise.
1803
 
1804
The table initialized from this macro, and the table initialized by
1805
the following one, may be overridden at run time either automatically,
1806
by the actions of the macro @code{CONDITIONAL_REGISTER_USAGE}, or by
1807
the user with the command options @option{-ffixed-@var{reg}},
1808
@option{-fcall-used-@var{reg}} and @option{-fcall-saved-@var{reg}}.
1809
@end defmac
1810
 
1811
@defmac CALL_USED_REGISTERS
1812
@cindex call-used register
1813
@cindex call-clobbered register
1814
@cindex call-saved register
1815
Like @code{FIXED_REGISTERS} but has 1 for each register that is
1816
clobbered (in general) by function calls as well as for fixed
1817
registers.  This macro therefore identifies the registers that are not
1818
available for general allocation of values that must live across
1819
function calls.
1820
 
1821
If a register has 0 in @code{CALL_USED_REGISTERS}, the compiler
1822
automatically saves it on function entry and restores it on function
1823
exit, if the register is used within the function.
1824
@end defmac
1825
 
1826
@defmac CALL_REALLY_USED_REGISTERS
1827
@cindex call-used register
1828
@cindex call-clobbered register
1829
@cindex call-saved register
1830
Like @code{CALL_USED_REGISTERS} except this macro doesn't require
1831
that the entire set of @code{FIXED_REGISTERS} be included.
1832
(@code{CALL_USED_REGISTERS} must be a superset of @code{FIXED_REGISTERS}).
1833
This macro is optional.  If not specified, it defaults to the value
1834
of @code{CALL_USED_REGISTERS}.
1835
@end defmac
1836
 
1837
@defmac HARD_REGNO_CALL_PART_CLOBBERED (@var{regno}, @var{mode})
1838
@cindex call-used register
1839
@cindex call-clobbered register
1840
@cindex call-saved register
1841
A C expression that is nonzero if it is not permissible to store a
1842
value of mode @var{mode} in hard register number @var{regno} across a
1843
call without some part of it being clobbered.  For most machines this
1844
macro need not be defined.  It is only required for machines that do not
1845
preserve the entire contents of a register across a call.
1846
@end defmac
1847
 
1848
@findex fixed_regs
1849
@findex call_used_regs
1850
@findex global_regs
1851
@findex reg_names
1852
@findex reg_class_contents
1853
@defmac CONDITIONAL_REGISTER_USAGE
1854
Zero or more C statements that may conditionally modify five variables
1855
@code{fixed_regs}, @code{call_used_regs}, @code{global_regs},
1856
@code{reg_names}, and @code{reg_class_contents}, to take into account
1857
any dependence of these register sets on target flags.  The first three
1858
of these are of type @code{char []} (interpreted as Boolean vectors).
1859
@code{global_regs} is a @code{const char *[]}, and
1860
@code{reg_class_contents} is a @code{HARD_REG_SET}.  Before the macro is
1861
called, @code{fixed_regs}, @code{call_used_regs},
1862
@code{reg_class_contents}, and @code{reg_names} have been initialized
1863
from @code{FIXED_REGISTERS}, @code{CALL_USED_REGISTERS},
1864
@code{REG_CLASS_CONTENTS}, and @code{REGISTER_NAMES}, respectively.
1865
@code{global_regs} has been cleared, and any @option{-ffixed-@var{reg}},
1866
@option{-fcall-used-@var{reg}} and @option{-fcall-saved-@var{reg}}
1867
command options have been applied.
1868
 
1869
You need not define this macro if it has no work to do.
1870
 
1871
@cindex disabling certain registers
1872
@cindex controlling register usage
1873
If the usage of an entire class of registers depends on the target
1874
flags, you may indicate this to GCC by using this macro to modify
1875
@code{fixed_regs} and @code{call_used_regs} to 1 for each of the
1876
registers in the classes which should not be used by GCC@.  Also define
1877
the macro @code{REG_CLASS_FROM_LETTER} / @code{REG_CLASS_FROM_CONSTRAINT}
1878
to return @code{NO_REGS} if it
1879
is called with a letter for a class that shouldn't be used.
1880
 
1881
(However, if this class is not included in @code{GENERAL_REGS} and all
1882
of the insn patterns whose constraints permit this class are
1883
controlled by target switches, then GCC will automatically avoid using
1884
these registers when the target switches are opposed to them.)
1885
@end defmac
1886
 
1887
@defmac INCOMING_REGNO (@var{out})
1888
Define this macro if the target machine has register windows.  This C
1889
expression returns the register number as seen by the called function
1890
corresponding to the register number @var{out} as seen by the calling
1891
function.  Return @var{out} if register number @var{out} is not an
1892
outbound register.
1893
@end defmac
1894
 
1895
@defmac OUTGOING_REGNO (@var{in})
1896
Define this macro if the target machine has register windows.  This C
1897
expression returns the register number as seen by the calling function
1898
corresponding to the register number @var{in} as seen by the called
1899
function.  Return @var{in} if register number @var{in} is not an inbound
1900
register.
1901
@end defmac
1902
 
1903
@defmac LOCAL_REGNO (@var{regno})
1904
Define this macro if the target machine has register windows.  This C
1905
expression returns true if the register is call-saved but is in the
1906
register window.  Unlike most call-saved registers, such registers
1907
need not be explicitly restored on function exit or during non-local
1908
gotos.
1909
@end defmac
1910
 
1911
@defmac PC_REGNUM
1912
If the program counter has a register number, define this as that
1913
register number.  Otherwise, do not define it.
1914
@end defmac
1915
 
1916
@node Allocation Order
1917
@subsection Order of Allocation of Registers
1918
@cindex order of register allocation
1919
@cindex register allocation order
1920
 
1921
@c prevent bad page break with this line
1922
Registers are allocated in order.
1923
 
1924
@defmac REG_ALLOC_ORDER
1925
If defined, an initializer for a vector of integers, containing the
1926
numbers of hard registers in the order in which GCC should prefer
1927
to use them (from most preferred to least).
1928
 
1929
If this macro is not defined, registers are used lowest numbered first
1930
(all else being equal).
1931
 
1932
One use of this macro is on machines where the highest numbered
1933
registers must always be saved and the save-multiple-registers
1934
instruction supports only sequences of consecutive registers.  On such
1935
machines, define @code{REG_ALLOC_ORDER} to be an initializer that lists
1936
the highest numbered allocable register first.
1937
@end defmac
1938
 
1939
@defmac ORDER_REGS_FOR_LOCAL_ALLOC
1940
A C statement (sans semicolon) to choose the order in which to allocate
1941
hard registers for pseudo-registers local to a basic block.
1942
 
1943
Store the desired register order in the array @code{reg_alloc_order}.
1944
Element 0 should be the register to allocate first; element 1, the next
1945
register; and so on.
1946
 
1947
The macro body should not assume anything about the contents of
1948
@code{reg_alloc_order} before execution of the macro.
1949
 
1950
On most machines, it is not necessary to define this macro.
1951
@end defmac
1952
 
1953
@node Values in Registers
1954
@subsection How Values Fit in Registers
1955
 
1956
This section discusses the macros that describe which kinds of values
1957
(specifically, which machine modes) each register can hold, and how many
1958
consecutive registers are needed for a given mode.
1959
 
1960
@defmac HARD_REGNO_NREGS (@var{regno}, @var{mode})
1961
A C expression for the number of consecutive hard registers, starting
1962
at register number @var{regno}, required to hold a value of mode
1963
@var{mode}.
1964
 
1965
On a machine where all registers are exactly one word, a suitable
1966
definition of this macro is
1967
 
1968
@smallexample
1969
#define HARD_REGNO_NREGS(REGNO, MODE)            \
1970
   ((GET_MODE_SIZE (MODE) + UNITS_PER_WORD - 1)  \
1971
    / UNITS_PER_WORD)
1972
@end smallexample
1973
@end defmac
1974
 
1975
@defmac HARD_REGNO_NREGS_HAS_PADDING (@var{regno}, @var{mode})
1976
A C expression that is nonzero if a value of mode @var{mode}, stored
1977
in memory, ends with padding that causes it to take up more space than
1978
in registers starting at register number @var{regno} (as determined by
1979
multiplying GCC's notion of the size of the register when containing
1980
this mode by the number of registers returned by
1981
@code{HARD_REGNO_NREGS}).  By default this is zero.
1982
 
1983
For example, if a floating-point value is stored in three 32-bit
1984
registers but takes up 128 bits in memory, then this would be
1985
nonzero.
1986
 
1987
This macros only needs to be defined if there are cases where
1988
@code{subreg_regno_offset} and @code{subreg_offset_representable_p}
1989
would otherwise wrongly determine that a @code{subreg} can be
1990
represented by an offset to the register number, when in fact such a
1991
@code{subreg} would contain some of the padding not stored in
1992
registers and so not be representable.
1993
@end defmac
1994
 
1995
@defmac HARD_REGNO_NREGS_WITH_PADDING (@var{regno}, @var{mode})
1996
For values of @var{regno} and @var{mode} for which
1997
@code{HARD_REGNO_NREGS_HAS_PADDING} returns nonzero, a C expression
1998
returning the greater number of registers required to hold the value
1999
including any padding.  In the example above, the value would be four.
2000
@end defmac
2001
 
2002
@defmac REGMODE_NATURAL_SIZE (@var{mode})
2003
Define this macro if the natural size of registers that hold values
2004
of mode @var{mode} is not the word size.  It is a C expression that
2005
should give the natural size in bytes for the specified mode.  It is
2006
used by the register allocator to try to optimize its results.  This
2007
happens for example on SPARC 64-bit where the natural size of
2008
floating-point registers is still 32-bit.
2009
@end defmac
2010
 
2011
@defmac HARD_REGNO_MODE_OK (@var{regno}, @var{mode})
2012
A C expression that is nonzero if it is permissible to store a value
2013
of mode @var{mode} in hard register number @var{regno} (or in several
2014
registers starting with that one).  For a machine where all registers
2015
are equivalent, a suitable definition is
2016
 
2017
@smallexample
2018
#define HARD_REGNO_MODE_OK(REGNO, MODE) 1
2019
@end smallexample
2020
 
2021
You need not include code to check for the numbers of fixed registers,
2022
because the allocation mechanism considers them to be always occupied.
2023
 
2024
@cindex register pairs
2025
On some machines, double-precision values must be kept in even/odd
2026
register pairs.  You can implement that by defining this macro to reject
2027
odd register numbers for such modes.
2028
 
2029
The minimum requirement for a mode to be OK in a register is that the
2030
@samp{mov@var{mode}} instruction pattern support moves between the
2031
register and other hard register in the same class and that moving a
2032
value into the register and back out not alter it.
2033
 
2034
Since the same instruction used to move @code{word_mode} will work for
2035
all narrower integer modes, it is not necessary on any machine for
2036
@code{HARD_REGNO_MODE_OK} to distinguish between these modes, provided
2037
you define patterns @samp{movhi}, etc., to take advantage of this.  This
2038
is useful because of the interaction between @code{HARD_REGNO_MODE_OK}
2039
and @code{MODES_TIEABLE_P}; it is very desirable for all integer modes
2040
to be tieable.
2041
 
2042
Many machines have special registers for floating point arithmetic.
2043
Often people assume that floating point machine modes are allowed only
2044
in floating point registers.  This is not true.  Any registers that
2045
can hold integers can safely @emph{hold} a floating point machine
2046
mode, whether or not floating arithmetic can be done on it in those
2047
registers.  Integer move instructions can be used to move the values.
2048
 
2049
On some machines, though, the converse is true: fixed-point machine
2050
modes may not go in floating registers.  This is true if the floating
2051
registers normalize any value stored in them, because storing a
2052
non-floating value there would garble it.  In this case,
2053
@code{HARD_REGNO_MODE_OK} should reject fixed-point machine modes in
2054
floating registers.  But if the floating registers do not automatically
2055
normalize, if you can store any bit pattern in one and retrieve it
2056
unchanged without a trap, then any machine mode may go in a floating
2057
register, so you can define this macro to say so.
2058
 
2059
The primary significance of special floating registers is rather that
2060
they are the registers acceptable in floating point arithmetic
2061
instructions.  However, this is of no concern to
2062
@code{HARD_REGNO_MODE_OK}.  You handle it by writing the proper
2063
constraints for those instructions.
2064
 
2065
On some machines, the floating registers are especially slow to access,
2066
so that it is better to store a value in a stack frame than in such a
2067
register if floating point arithmetic is not being done.  As long as the
2068
floating registers are not in class @code{GENERAL_REGS}, they will not
2069
be used unless some pattern's constraint asks for one.
2070
@end defmac
2071
 
2072
@defmac HARD_REGNO_RENAME_OK (@var{from}, @var{to})
2073
A C expression that is nonzero if it is OK to rename a hard register
2074
@var{from} to another hard register @var{to}.
2075
 
2076
One common use of this macro is to prevent renaming of a register to
2077
another register that is not saved by a prologue in an interrupt
2078
handler.
2079
 
2080
The default is always nonzero.
2081
@end defmac
2082
 
2083
@defmac MODES_TIEABLE_P (@var{mode1}, @var{mode2})
2084
A C expression that is nonzero if a value of mode
2085
@var{mode1} is accessible in mode @var{mode2} without copying.
2086
 
2087
If @code{HARD_REGNO_MODE_OK (@var{r}, @var{mode1})} and
2088
@code{HARD_REGNO_MODE_OK (@var{r}, @var{mode2})} are always the same for
2089
any @var{r}, then @code{MODES_TIEABLE_P (@var{mode1}, @var{mode2})}
2090
should be nonzero.  If they differ for any @var{r}, you should define
2091
this macro to return zero unless some other mechanism ensures the
2092
accessibility of the value in a narrower mode.
2093
 
2094
You should define this macro to return nonzero in as many cases as
2095
possible since doing so will allow GCC to perform better register
2096
allocation.
2097
@end defmac
2098
 
2099
@defmac AVOID_CCMODE_COPIES
2100
Define this macro if the compiler should avoid copies to/from @code{CCmode}
2101
registers.  You should only define this macro if support for copying to/from
2102
@code{CCmode} is incomplete.
2103
@end defmac
2104
 
2105
@node Leaf Functions
2106
@subsection Handling Leaf Functions
2107
 
2108
@cindex leaf functions
2109
@cindex functions, leaf
2110
On some machines, a leaf function (i.e., one which makes no calls) can run
2111
more efficiently if it does not make its own register window.  Often this
2112
means it is required to receive its arguments in the registers where they
2113
are passed by the caller, instead of the registers where they would
2114
normally arrive.
2115
 
2116
The special treatment for leaf functions generally applies only when
2117
other conditions are met; for example, often they may use only those
2118
registers for its own variables and temporaries.  We use the term ``leaf
2119
function'' to mean a function that is suitable for this special
2120
handling, so that functions with no calls are not necessarily ``leaf
2121
functions''.
2122
 
2123
GCC assigns register numbers before it knows whether the function is
2124
suitable for leaf function treatment.  So it needs to renumber the
2125
registers in order to output a leaf function.  The following macros
2126
accomplish this.
2127
 
2128
@defmac LEAF_REGISTERS
2129
Name of a char vector, indexed by hard register number, which
2130
contains 1 for a register that is allowable in a candidate for leaf
2131
function treatment.
2132
 
2133
If leaf function treatment involves renumbering the registers, then the
2134
registers marked here should be the ones before renumbering---those that
2135
GCC would ordinarily allocate.  The registers which will actually be
2136
used in the assembler code, after renumbering, should not be marked with 1
2137
in this vector.
2138
 
2139
Define this macro only if the target machine offers a way to optimize
2140
the treatment of leaf functions.
2141
@end defmac
2142
 
2143
@defmac LEAF_REG_REMAP (@var{regno})
2144
A C expression whose value is the register number to which @var{regno}
2145
should be renumbered, when a function is treated as a leaf function.
2146
 
2147
If @var{regno} is a register number which should not appear in a leaf
2148
function before renumbering, then the expression should yield @minus{}1, which
2149
will cause the compiler to abort.
2150
 
2151
Define this macro only if the target machine offers a way to optimize the
2152
treatment of leaf functions, and registers need to be renumbered to do
2153
this.
2154
@end defmac
2155
 
2156
@findex current_function_is_leaf
2157
@findex current_function_uses_only_leaf_regs
2158
@code{TARGET_ASM_FUNCTION_PROLOGUE} and
2159
@code{TARGET_ASM_FUNCTION_EPILOGUE} must usually treat leaf functions
2160
specially.  They can test the C variable @code{current_function_is_leaf}
2161
which is nonzero for leaf functions.  @code{current_function_is_leaf} is
2162
set prior to local register allocation and is valid for the remaining
2163
compiler passes.  They can also test the C variable
2164
@code{current_function_uses_only_leaf_regs} which is nonzero for leaf
2165
functions which only use leaf registers.
2166
@code{current_function_uses_only_leaf_regs} is valid after all passes
2167
that modify the instructions have been run and is only useful if
2168
@code{LEAF_REGISTERS} is defined.
2169
@c changed this to fix overfull.  ALSO:  why the "it" at the beginning
2170
@c of the next paragraph?!  --mew 2feb93
2171
 
2172
@node Stack Registers
2173
@subsection Registers That Form a Stack
2174
 
2175
There are special features to handle computers where some of the
2176
``registers'' form a stack.  Stack registers are normally written by
2177
pushing onto the stack, and are numbered relative to the top of the
2178
stack.
2179
 
2180
Currently, GCC can only handle one group of stack-like registers, and
2181
they must be consecutively numbered.  Furthermore, the existing
2182
support for stack-like registers is specific to the 80387 floating
2183
point coprocessor.  If you have a new architecture that uses
2184
stack-like registers, you will need to do substantial work on
2185
@file{reg-stack.c} and write your machine description to cooperate
2186
with it, as well as defining these macros.
2187
 
2188
@defmac STACK_REGS
2189
Define this if the machine has any stack-like registers.
2190
@end defmac
2191
 
2192
@defmac FIRST_STACK_REG
2193
The number of the first stack-like register.  This one is the top
2194
of the stack.
2195
@end defmac
2196
 
2197
@defmac LAST_STACK_REG
2198
The number of the last stack-like register.  This one is the bottom of
2199
the stack.
2200
@end defmac
2201
 
2202
@node Register Classes
2203
@section Register Classes
2204
@cindex register class definitions
2205
@cindex class definitions, register
2206
 
2207
On many machines, the numbered registers are not all equivalent.
2208
For example, certain registers may not be allowed for indexed addressing;
2209
certain registers may not be allowed in some instructions.  These machine
2210
restrictions are described to the compiler using @dfn{register classes}.
2211
 
2212
You define a number of register classes, giving each one a name and saying
2213
which of the registers belong to it.  Then you can specify register classes
2214
that are allowed as operands to particular instruction patterns.
2215
 
2216
@findex ALL_REGS
2217
@findex NO_REGS
2218
In general, each register will belong to several classes.  In fact, one
2219
class must be named @code{ALL_REGS} and contain all the registers.  Another
2220
class must be named @code{NO_REGS} and contain no registers.  Often the
2221
union of two classes will be another class; however, this is not required.
2222
 
2223
@findex GENERAL_REGS
2224
One of the classes must be named @code{GENERAL_REGS}.  There is nothing
2225
terribly special about the name, but the operand constraint letters
2226
@samp{r} and @samp{g} specify this class.  If @code{GENERAL_REGS} is
2227
the same as @code{ALL_REGS}, just define it as a macro which expands
2228
to @code{ALL_REGS}.
2229
 
2230
Order the classes so that if class @var{x} is contained in class @var{y}
2231
then @var{x} has a lower class number than @var{y}.
2232
 
2233
The way classes other than @code{GENERAL_REGS} are specified in operand
2234
constraints is through machine-dependent operand constraint letters.
2235
You can define such letters to correspond to various classes, then use
2236
them in operand constraints.
2237
 
2238
You should define a class for the union of two classes whenever some
2239
instruction allows both classes.  For example, if an instruction allows
2240
either a floating point (coprocessor) register or a general register for a
2241
certain operand, you should define a class @code{FLOAT_OR_GENERAL_REGS}
2242
which includes both of them.  Otherwise you will get suboptimal code.
2243
 
2244
You must also specify certain redundant information about the register
2245
classes: for each class, which classes contain it and which ones are
2246
contained in it; for each pair of classes, the largest class contained
2247
in their union.
2248
 
2249
When a value occupying several consecutive registers is expected in a
2250
certain class, all the registers used must belong to that class.
2251
Therefore, register classes cannot be used to enforce a requirement for
2252
a register pair to start with an even-numbered register.  The way to
2253
specify this requirement is with @code{HARD_REGNO_MODE_OK}.
2254
 
2255
Register classes used for input-operands of bitwise-and or shift
2256
instructions have a special requirement: each such class must have, for
2257
each fixed-point machine mode, a subclass whose registers can transfer that
2258
mode to or from memory.  For example, on some machines, the operations for
2259
single-byte values (@code{QImode}) are limited to certain registers.  When
2260
this is so, each register class that is used in a bitwise-and or shift
2261
instruction must have a subclass consisting of registers from which
2262
single-byte values can be loaded or stored.  This is so that
2263
@code{PREFERRED_RELOAD_CLASS} can always have a possible value to return.
2264
 
2265
@deftp {Data type} {enum reg_class}
2266
An enumerated type that must be defined with all the register class names
2267
as enumerated values.  @code{NO_REGS} must be first.  @code{ALL_REGS}
2268
must be the last register class, followed by one more enumerated value,
2269
@code{LIM_REG_CLASSES}, which is not a register class but rather
2270
tells how many classes there are.
2271
 
2272
Each register class has a number, which is the value of casting
2273
the class name to type @code{int}.  The number serves as an index
2274
in many of the tables described below.
2275
@end deftp
2276
 
2277
@defmac N_REG_CLASSES
2278
The number of distinct register classes, defined as follows:
2279
 
2280
@smallexample
2281
#define N_REG_CLASSES (int) LIM_REG_CLASSES
2282
@end smallexample
2283
@end defmac
2284
 
2285
@defmac REG_CLASS_NAMES
2286
An initializer containing the names of the register classes as C string
2287
constants.  These names are used in writing some of the debugging dumps.
2288
@end defmac
2289
 
2290
@defmac REG_CLASS_CONTENTS
2291
An initializer containing the contents of the register classes, as integers
2292
which are bit masks.  The @var{n}th integer specifies the contents of class
2293
@var{n}.  The way the integer @var{mask} is interpreted is that
2294
register @var{r} is in the class if @code{@var{mask} & (1 << @var{r})} is 1.
2295
 
2296
When the machine has more than 32 registers, an integer does not suffice.
2297
Then the integers are replaced by sub-initializers, braced groupings containing
2298
several integers.  Each sub-initializer must be suitable as an initializer
2299
for the type @code{HARD_REG_SET} which is defined in @file{hard-reg-set.h}.
2300
In this situation, the first integer in each sub-initializer corresponds to
2301
registers 0 through 31, the second integer to registers 32 through 63, and
2302
so on.
2303
@end defmac
2304
 
2305
@defmac REGNO_REG_CLASS (@var{regno})
2306
A C expression whose value is a register class containing hard register
2307
@var{regno}.  In general there is more than one such class; choose a class
2308
which is @dfn{minimal}, meaning that no smaller class also contains the
2309
register.
2310
@end defmac
2311
 
2312
@defmac BASE_REG_CLASS
2313
A macro whose definition is the name of the class to which a valid
2314
base register must belong.  A base register is one used in an address
2315
which is the register value plus a displacement.
2316
@end defmac
2317
 
2318
@defmac MODE_BASE_REG_CLASS (@var{mode})
2319
This is a variation of the @code{BASE_REG_CLASS} macro which allows
2320
the selection of a base register in a mode dependent manner.  If
2321
@var{mode} is VOIDmode then it should return the same value as
2322
@code{BASE_REG_CLASS}.
2323
@end defmac
2324
 
2325
@defmac MODE_BASE_REG_REG_CLASS (@var{mode})
2326
A C expression whose value is the register class to which a valid
2327
base register must belong in order to be used in a base plus index
2328
register address.  You should define this macro if base plus index
2329
addresses have different requirements than other base register uses.
2330
@end defmac
2331
 
2332
@defmac MODE_CODE_BASE_REG_CLASS (@var{mode}, @var{outer_code}, @var{index_code})
2333
A C expression whose value is the register class to which a valid
2334
base register must belong.  @var{outer_code} and @var{index_code} define the
2335
context in which the base register occurs.  @var{outer_code} is the code of
2336
the immediately enclosing expression (@code{MEM} for the top level of an
2337
address, @code{ADDRESS} for something that occurs in an
2338
@code{address_operand}).  @var{index_code} is the code of the corresponding
2339
index expression if @var{outer_code} is @code{PLUS}; @code{SCRATCH} otherwise.
2340
@end defmac
2341
 
2342
@defmac INDEX_REG_CLASS
2343
A macro whose definition is the name of the class to which a valid
2344
index register must belong.  An index register is one used in an
2345
address where its value is either multiplied by a scale factor or
2346
added to another register (as well as added to a displacement).
2347
@end defmac
2348
 
2349
@defmac REGNO_OK_FOR_BASE_P (@var{num})
2350
A C expression which is nonzero if register number @var{num} is
2351
suitable for use as a base register in operand addresses.  It may be
2352
either a suitable hard register or a pseudo register that has been
2353
allocated such a hard register.
2354
@end defmac
2355
 
2356
@defmac REGNO_MODE_OK_FOR_BASE_P (@var{num}, @var{mode})
2357
A C expression that is just like @code{REGNO_OK_FOR_BASE_P}, except that
2358
that expression may examine the mode of the memory reference in
2359
@var{mode}.  You should define this macro if the mode of the memory
2360
reference affects whether a register may be used as a base register.  If
2361
you define this macro, the compiler will use it instead of
2362
@code{REGNO_OK_FOR_BASE_P}.  The mode may be @code{VOIDmode} for addresses
2363
that appear outside a @code{MEM}, i.e. as an @code{address_operand}.
2364
 
2365
@end defmac
2366
 
2367
@defmac REGNO_MODE_OK_FOR_REG_BASE_P (@var{num}, @var{mode})
2368
A C expression which is nonzero if register number @var{num} is suitable for
2369
use as a base register in base plus index operand addresses, accessing
2370
memory in mode @var{mode}.  It may be either a suitable hard register or a
2371
pseudo register that has been allocated such a hard register.  You should
2372
define this macro if base plus index addresses have different requirements
2373
than other base register uses.
2374
 
2375
Use of this macro is deprecated; please use the more general
2376
@code{REGNO_MODE_CODE_OK_FOR_BASE_P}.
2377
@end defmac
2378
 
2379
@defmac REGNO_MODE_CODE_OK_FOR_BASE_P (@var{num}, @var{mode}, @var{outer_code}, @var{index_code})
2380
A C expression that is just like @code{REGNO_MODE_OK_FOR_BASE_P}, except that
2381
that expression may examine the context in which the register appears in the
2382
memory reference.  @var{outer_code} is the code of the immediately enclosing
2383
expression (@code{MEM} if at the top level of the address, @code{ADDRESS} for
2384
something that occurs in an @code{address_operand}).  @var{index_code} is the
2385
code of the corresponding index expression if @var{outer_code} is @code{PLUS};
2386
@code{SCRATCH} otherwise.  The mode may be @code{VOIDmode} for addresses
2387
that appear outside a @code{MEM}, i.e. as an @code{address_operand}.
2388
@end defmac
2389
 
2390
@defmac REGNO_OK_FOR_INDEX_P (@var{num})
2391
A C expression which is nonzero if register number @var{num} is
2392
suitable for use as an index register in operand addresses.  It may be
2393
either a suitable hard register or a pseudo register that has been
2394
allocated such a hard register.
2395
 
2396
The difference between an index register and a base register is that
2397
the index register may be scaled.  If an address involves the sum of
2398
two registers, neither one of them scaled, then either one may be
2399
labeled the ``base'' and the other the ``index''; but whichever
2400
labeling is used must fit the machine's constraints of which registers
2401
may serve in each capacity.  The compiler will try both labelings,
2402
looking for one that is valid, and will reload one or both registers
2403
only if neither labeling works.
2404
@end defmac
2405
 
2406
@defmac PREFERRED_RELOAD_CLASS (@var{x}, @var{class})
2407
A C expression that places additional restrictions on the register class
2408
to use when it is necessary to copy value @var{x} into a register in class
2409
@var{class}.  The value is a register class; perhaps @var{class}, or perhaps
2410
another, smaller class.  On many machines, the following definition is
2411
safe:
2412
 
2413
@smallexample
2414
#define PREFERRED_RELOAD_CLASS(X,CLASS) CLASS
2415
@end smallexample
2416
 
2417
Sometimes returning a more restrictive class makes better code.  For
2418
example, on the 68000, when @var{x} is an integer constant that is in range
2419
for a @samp{moveq} instruction, the value of this macro is always
2420
@code{DATA_REGS} as long as @var{class} includes the data registers.
2421
Requiring a data register guarantees that a @samp{moveq} will be used.
2422
 
2423
One case where @code{PREFERRED_RELOAD_CLASS} must not return
2424
@var{class} is if @var{x} is a legitimate constant which cannot be
2425
loaded into some register class.  By returning @code{NO_REGS} you can
2426
force @var{x} into a memory location.  For example, rs6000 can load
2427
immediate values into general-purpose registers, but does not have an
2428
instruction for loading an immediate value into a floating-point
2429
register, so @code{PREFERRED_RELOAD_CLASS} returns @code{NO_REGS} when
2430
@var{x} is a floating-point constant.  If the constant can't be loaded
2431
into any kind of register, code generation will be better if
2432
@code{LEGITIMATE_CONSTANT_P} makes the constant illegitimate instead
2433
of using @code{PREFERRED_RELOAD_CLASS}.
2434
 
2435
If an insn has pseudos in it after register allocation, reload will go
2436
through the alternatives and call repeatedly @code{PREFERRED_RELOAD_CLASS}
2437
to find the best one.  Returning @code{NO_REGS}, in this case, makes
2438
reload add a @code{!} in front of the constraint: the x86 back-end uses
2439
this feature to discourage usage of 387 registers when math is done in
2440
the SSE registers (and vice versa).
2441
@end defmac
2442
 
2443
@defmac PREFERRED_OUTPUT_RELOAD_CLASS (@var{x}, @var{class})
2444
Like @code{PREFERRED_RELOAD_CLASS}, but for output reloads instead of
2445
input reloads.  If you don't define this macro, the default is to use
2446
@var{class}, unchanged.
2447
 
2448
You can also use @code{PREFERRED_OUTPUT_RELOAD_CLASS} to discourage
2449
reload from using some alternatives, like @code{PREFERRED_RELOAD_CLASS}.
2450
@end defmac
2451
 
2452
@defmac LIMIT_RELOAD_CLASS (@var{mode}, @var{class})
2453
A C expression that places additional restrictions on the register class
2454
to use when it is necessary to be able to hold a value of mode
2455
@var{mode} in a reload register for which class @var{class} would
2456
ordinarily be used.
2457
 
2458
Unlike @code{PREFERRED_RELOAD_CLASS}, this macro should be used when
2459
there are certain modes that simply can't go in certain reload classes.
2460
 
2461
The value is a register class; perhaps @var{class}, or perhaps another,
2462
smaller class.
2463
 
2464
Don't define this macro unless the target machine has limitations which
2465
require the macro to do something nontrivial.
2466
@end defmac
2467
 
2468
@deftypefn {Target Hook} enum reg_class TARGET_SECONDARY_RELOAD (bool @var{in_p}, rtx @var{x}, enum reg_class @var{reload_class}, enum machine_mode @var{reload_mode}, secondary_reload_info *@var{sri})
2469
Many machines have some registers that cannot be copied directly to or
2470
from memory or even from other types of registers.  An example is the
2471
@samp{MQ} register, which on most machines, can only be copied to or
2472
from general registers, but not memory.  Below, we shall be using the
2473
term 'intermediate register' when a move operation cannot be performed
2474
directly, but has to be done by copying the source into the intermediate
2475
register first, and then copying the intermediate register to the
2476
destination.  An intermediate register always has the same mode as
2477
source and destination.  Since it holds the actual value being copied,
2478
reload might apply optimizations to re-use an intermediate register
2479
and eliding the copy from the source when it can determine that the
2480
intermediate register still holds the required value.
2481
 
2482
Another kind of secondary reload is required on some machines which
2483
allow copying all registers to and from memory, but require a scratch
2484
register for stores to some memory locations (e.g., those with symbolic
2485
address on the RT, and those with certain symbolic address on the SPARC
2486
when compiling PIC)@.  Scratch registers need not have the same mode
2487
as the value being copied, and usually hold a different value that
2488
that being copied.  Special patterns in the md file are needed to
2489
describe how the copy is performed with the help of the scratch register;
2490
these patterns also describe the number, register class(es) and mode(s)
2491
of the scratch register(s).
2492
 
2493
In some cases, both an intermediate and a scratch register are required.
2494
 
2495
For input reloads, this target hook is called with nonzero @var{in_p},
2496
and @var{x} is an rtx that needs to be copied to a register in of class
2497
@var{reload_class} in @var{reload_mode}.  For output reloads, this target
2498
hook is called with zero @var{in_p}, and a register of class @var{reload_mode}
2499
needs to be copied to rtx @var{x} in @var{reload_mode}.
2500
 
2501
If copying a register of @var{reload_class} from/to @var{x} requires
2502
an intermediate register, the hook @code{secondary_reload} should
2503
return the register class required for this intermediate register.
2504
If no intermediate register is required, it should return NO_REGS.
2505
If more than one intermediate register is required, describe the one
2506
that is closest in the copy chain to the reload register.
2507
 
2508
If scratch registers are needed, you also have to describe how to
2509
perform the copy from/to the reload register to/from this
2510
closest intermediate register.  Or if no intermediate register is
2511
required, but still a scratch register is needed, describe the
2512
copy  from/to the reload register to/from the reload operand @var{x}.
2513
 
2514
You do this by setting @code{sri->icode} to the instruction code of a pattern
2515
in the md file which performs the move.  Operands 0 and 1 are the output
2516
and input of this copy, respectively.  Operands from operand 2 onward are
2517
for scratch operands.  These scratch operands must have a mode, and a
2518
single-register-class
2519
@c [later: or memory]
2520
output constraint.
2521
 
2522
When an intermediate register is used, the @code{secondary_reload}
2523
hook will be called again to determine how to copy the intermediate
2524
register to/from the reload operand @var{x}, so your hook must also
2525
have code to handle the register class of the intermediate operand.
2526
 
2527
@c [For later: maybe we'll allow multi-alternative reload patterns -
2528
@c   the port maintainer could name a mov<mode> pattern that has clobbers -
2529
@c   and match the constraints of input and output to determine the required
2530
@c   alternative.  A restriction would be that constraints used to match
2531
@c   against reloads registers would have to be written as register class
2532
@c   constraints, or we need a new target macro / hook that tells us if an
2533
@c   arbitrary constraint can match an unknown register of a given class.
2534
@c   Such a macro / hook would also be useful in other places.]
2535
 
2536
 
2537
@var{x} might be a pseudo-register or a @code{subreg} of a
2538
pseudo-register, which could either be in a hard register or in memory.
2539
Use @code{true_regnum} to find out; it will return @minus{}1 if the pseudo is
2540
in memory and the hard register number if it is in a register.
2541
 
2542
Scratch operands in memory (constraint @code{"=m"} / @code{"=&m"}) are
2543
currently not supported.  For the time being, you will have to continue
2544
to use @code{SECONDARY_MEMORY_NEEDED} for that purpose.
2545
 
2546
@code{copy_cost} also uses this target hook to find out how values are
2547
copied.  If you want it to include some extra cost for the need to allocate
2548
(a) scratch register(s), set @code{sri->extra_cost} to the additional cost.
2549
Or if two dependent moves are supposed to have a lower cost than the sum
2550
of the individual moves due to expected fortuitous scheduling and/or special
2551
forwarding logic, you can set @code{sri->extra_cost} to a negative amount.
2552
@end deftypefn
2553
 
2554
@defmac SECONDARY_RELOAD_CLASS (@var{class}, @var{mode}, @var{x})
2555
@defmacx SECONDARY_INPUT_RELOAD_CLASS (@var{class}, @var{mode}, @var{x})
2556
@defmacx SECONDARY_OUTPUT_RELOAD_CLASS (@var{class}, @var{mode}, @var{x})
2557
These macros are obsolete, new ports should use the target hook
2558
@code{TARGET_SECONDARY_RELOAD} instead.
2559
 
2560
These are obsolete macros, replaced by the @code{TARGET_SECONDARY_RELOAD}
2561
target hook.  Older ports still define these macros to indicate to the
2562
reload phase that it may
2563
need to allocate at least one register for a reload in addition to the
2564
register to contain the data.  Specifically, if copying @var{x} to a
2565
register @var{class} in @var{mode} requires an intermediate register,
2566
you were supposed to define @code{SECONDARY_INPUT_RELOAD_CLASS} to return the
2567
largest register class all of whose registers can be used as
2568
intermediate registers or scratch registers.
2569
 
2570
If copying a register @var{class} in @var{mode} to @var{x} requires an
2571
intermediate or scratch register, @code{SECONDARY_OUTPUT_RELOAD_CLASS}
2572
was supposed to be defined be defined to return the largest register
2573
class required.  If the
2574
requirements for input and output reloads were the same, the macro
2575
@code{SECONDARY_RELOAD_CLASS} should have been used instead of defining both
2576
macros identically.
2577
 
2578
The values returned by these macros are often @code{GENERAL_REGS}.
2579
Return @code{NO_REGS} if no spare register is needed; i.e., if @var{x}
2580
can be directly copied to or from a register of @var{class} in
2581
@var{mode} without requiring a scratch register.  Do not define this
2582
macro if it would always return @code{NO_REGS}.
2583
 
2584
If a scratch register is required (either with or without an
2585
intermediate register), you were supposed to define patterns for
2586
@samp{reload_in@var{m}} or @samp{reload_out@var{m}}, as required
2587
(@pxref{Standard Names}.  These patterns, which were normally
2588
implemented with a @code{define_expand}, should be similar to the
2589
@samp{mov@var{m}} patterns, except that operand 2 is the scratch
2590
register.
2591
 
2592
These patterns need constraints for the reload register and scratch
2593
register that
2594
contain a single register class.  If the original reload register (whose
2595
class is @var{class}) can meet the constraint given in the pattern, the
2596
value returned by these macros is used for the class of the scratch
2597
register.  Otherwise, two additional reload registers are required.
2598
Their classes are obtained from the constraints in the insn pattern.
2599
 
2600
@var{x} might be a pseudo-register or a @code{subreg} of a
2601
pseudo-register, which could either be in a hard register or in memory.
2602
Use @code{true_regnum} to find out; it will return @minus{}1 if the pseudo is
2603
in memory and the hard register number if it is in a register.
2604
 
2605
These macros should not be used in the case where a particular class of
2606
registers can only be copied to memory and not to another class of
2607
registers.  In that case, secondary reload registers are not needed and
2608
would not be helpful.  Instead, a stack location must be used to perform
2609
the copy and the @code{mov@var{m}} pattern should use memory as an
2610
intermediate storage.  This case often occurs between floating-point and
2611
general registers.
2612
@end defmac
2613
 
2614
@defmac SECONDARY_MEMORY_NEEDED (@var{class1}, @var{class2}, @var{m})
2615
Certain machines have the property that some registers cannot be copied
2616
to some other registers without using memory.  Define this macro on
2617
those machines to be a C expression that is nonzero if objects of mode
2618
@var{m} in registers of @var{class1} can only be copied to registers of
2619
class @var{class2} by storing a register of @var{class1} into memory
2620
and loading that memory location into a register of @var{class2}.
2621
 
2622
Do not define this macro if its value would always be zero.
2623
@end defmac
2624
 
2625
@defmac SECONDARY_MEMORY_NEEDED_RTX (@var{mode})
2626
Normally when @code{SECONDARY_MEMORY_NEEDED} is defined, the compiler
2627
allocates a stack slot for a memory location needed for register copies.
2628
If this macro is defined, the compiler instead uses the memory location
2629
defined by this macro.
2630
 
2631
Do not define this macro if you do not define
2632
@code{SECONDARY_MEMORY_NEEDED}.
2633
@end defmac
2634
 
2635
@defmac SECONDARY_MEMORY_NEEDED_MODE (@var{mode})
2636
When the compiler needs a secondary memory location to copy between two
2637
registers of mode @var{mode}, it normally allocates sufficient memory to
2638
hold a quantity of @code{BITS_PER_WORD} bits and performs the store and
2639
load operations in a mode that many bits wide and whose class is the
2640
same as that of @var{mode}.
2641
 
2642
This is right thing to do on most machines because it ensures that all
2643
bits of the register are copied and prevents accesses to the registers
2644
in a narrower mode, which some machines prohibit for floating-point
2645
registers.
2646
 
2647
However, this default behavior is not correct on some machines, such as
2648
the DEC Alpha, that store short integers in floating-point registers
2649
differently than in integer registers.  On those machines, the default
2650
widening will not work correctly and you must define this macro to
2651
suppress that widening in some cases.  See the file @file{alpha.h} for
2652
details.
2653
 
2654
Do not define this macro if you do not define
2655
@code{SECONDARY_MEMORY_NEEDED} or if widening @var{mode} to a mode that
2656
is @code{BITS_PER_WORD} bits wide is correct for your machine.
2657
@end defmac
2658
 
2659
@defmac SMALL_REGISTER_CLASSES
2660
On some machines, it is risky to let hard registers live across arbitrary
2661
insns.  Typically, these machines have instructions that require values
2662
to be in specific registers (like an accumulator), and reload will fail
2663
if the required hard register is used for another purpose across such an
2664
insn.
2665
 
2666
Define @code{SMALL_REGISTER_CLASSES} to be an expression with a nonzero
2667
value on these machines.  When this macro has a nonzero value, the
2668
compiler will try to minimize the lifetime of hard registers.
2669
 
2670
It is always safe to define this macro with a nonzero value, but if you
2671
unnecessarily define it, you will reduce the amount of optimizations
2672
that can be performed in some cases.  If you do not define this macro
2673
with a nonzero value when it is required, the compiler will run out of
2674
spill registers and print a fatal error message.  For most machines, you
2675
should not define this macro at all.
2676
@end defmac
2677
 
2678
@defmac CLASS_LIKELY_SPILLED_P (@var{class})
2679
A C expression whose value is nonzero if pseudos that have been assigned
2680
to registers of class @var{class} would likely be spilled because
2681
registers of @var{class} are needed for spill registers.
2682
 
2683
The default value of this macro returns 1 if @var{class} has exactly one
2684
register and zero otherwise.  On most machines, this default should be
2685
used.  Only define this macro to some other expression if pseudos
2686
allocated by @file{local-alloc.c} end up in memory because their hard
2687
registers were needed for spill registers.  If this macro returns nonzero
2688
for those classes, those pseudos will only be allocated by
2689
@file{global.c}, which knows how to reallocate the pseudo to another
2690
register.  If there would not be another register available for
2691
reallocation, you should not change the definition of this macro since
2692
the only effect of such a definition would be to slow down register
2693
allocation.
2694
@end defmac
2695
 
2696
@defmac CLASS_MAX_NREGS (@var{class}, @var{mode})
2697
A C expression for the maximum number of consecutive registers
2698
of class @var{class} needed to hold a value of mode @var{mode}.
2699
 
2700
This is closely related to the macro @code{HARD_REGNO_NREGS}.  In fact,
2701
the value of the macro @code{CLASS_MAX_NREGS (@var{class}, @var{mode})}
2702
should be the maximum value of @code{HARD_REGNO_NREGS (@var{regno},
2703
@var{mode})} for all @var{regno} values in the class @var{class}.
2704
 
2705
This macro helps control the handling of multiple-word values
2706
in the reload pass.
2707
@end defmac
2708
 
2709
@defmac CANNOT_CHANGE_MODE_CLASS (@var{from}, @var{to}, @var{class})
2710
If defined, a C expression that returns nonzero for a @var{class} for which
2711
a change from mode @var{from} to mode @var{to} is invalid.
2712
 
2713
For the example, loading 32-bit integer or floating-point objects into
2714
floating-point registers on the Alpha extends them to 64 bits.
2715
Therefore loading a 64-bit object and then storing it as a 32-bit object
2716
does not store the low-order 32 bits, as would be the case for a normal
2717
register.  Therefore, @file{alpha.h} defines @code{CANNOT_CHANGE_MODE_CLASS}
2718
as below:
2719
 
2720
@smallexample
2721
#define CANNOT_CHANGE_MODE_CLASS(FROM, TO, CLASS) \
2722
  (GET_MODE_SIZE (FROM) != GET_MODE_SIZE (TO) \
2723
   ? reg_classes_intersect_p (FLOAT_REGS, (CLASS)) : 0)
2724
@end smallexample
2725
@end defmac
2726
 
2727
@node Old Constraints
2728
@section Obsolete Macros for Defining Constraints
2729
@cindex defining constraints, obsolete method
2730
@cindex constraints, defining, obsolete method
2731
 
2732
Machine-specific constraints can be defined with these macros instead
2733
of the machine description constructs described in @ref{Define
2734
Constraints}.  This mechanism is obsolete.  New ports should not use
2735
it; old ports should convert to the new mechanism.
2736
 
2737
@defmac CONSTRAINT_LEN (@var{char}, @var{str})
2738
For the constraint at the start of @var{str}, which starts with the letter
2739
@var{c}, return the length.  This allows you to have register class /
2740
constant / extra constraints that are longer than a single letter;
2741
you don't need to define this macro if you can do with single-letter
2742
constraints only.  The definition of this macro should use
2743
DEFAULT_CONSTRAINT_LEN for all the characters that you don't want
2744
to handle specially.
2745
There are some sanity checks in genoutput.c that check the constraint lengths
2746
for the md file, so you can also use this macro to help you while you are
2747
transitioning from a byzantine single-letter-constraint scheme: when you
2748
return a negative length for a constraint you want to re-use, genoutput
2749
will complain about every instance where it is used in the md file.
2750
@end defmac
2751
 
2752
@defmac REG_CLASS_FROM_LETTER (@var{char})
2753
A C expression which defines the machine-dependent operand constraint
2754
letters for register classes.  If @var{char} is such a letter, the
2755
value should be the register class corresponding to it.  Otherwise,
2756
the value should be @code{NO_REGS}.  The register letter @samp{r},
2757
corresponding to class @code{GENERAL_REGS}, will not be passed
2758
to this macro; you do not need to handle it.
2759
@end defmac
2760
 
2761
@defmac REG_CLASS_FROM_CONSTRAINT (@var{char}, @var{str})
2762
Like @code{REG_CLASS_FROM_LETTER}, but you also get the constraint string
2763
passed in @var{str}, so that you can use suffixes to distinguish between
2764
different variants.
2765
@end defmac
2766
 
2767
@defmac CONST_OK_FOR_LETTER_P (@var{value}, @var{c})
2768
A C expression that defines the machine-dependent operand constraint
2769
letters (@samp{I}, @samp{J}, @samp{K}, @dots{} @samp{P}) that specify
2770
particular ranges of integer values.  If @var{c} is one of those
2771
letters, the expression should check that @var{value}, an integer, is in
2772
the appropriate range and return 1 if so, 0 otherwise.  If @var{c} is
2773
not one of those letters, the value should be 0 regardless of
2774
@var{value}.
2775
@end defmac
2776
 
2777
@defmac CONST_OK_FOR_CONSTRAINT_P (@var{value}, @var{c}, @var{str})
2778
Like @code{CONST_OK_FOR_LETTER_P}, but you also get the constraint
2779
string passed in @var{str}, so that you can use suffixes to distinguish
2780
between different variants.
2781
@end defmac
2782
 
2783
@defmac CONST_DOUBLE_OK_FOR_LETTER_P (@var{value}, @var{c})
2784
A C expression that defines the machine-dependent operand constraint
2785
letters that specify particular ranges of @code{const_double} values
2786
(@samp{G} or @samp{H}).
2787
 
2788
If @var{c} is one of those letters, the expression should check that
2789
@var{value}, an RTX of code @code{const_double}, is in the appropriate
2790
range and return 1 if so, 0 otherwise.  If @var{c} is not one of those
2791
letters, the value should be 0 regardless of @var{value}.
2792
 
2793
@code{const_double} is used for all floating-point constants and for
2794
@code{DImode} fixed-point constants.  A given letter can accept either
2795
or both kinds of values.  It can use @code{GET_MODE} to distinguish
2796
between these kinds.
2797
@end defmac
2798
 
2799
@defmac CONST_DOUBLE_OK_FOR_CONSTRAINT_P (@var{value}, @var{c}, @var{str})
2800
Like @code{CONST_DOUBLE_OK_FOR_LETTER_P}, but you also get the constraint
2801
string passed in @var{str}, so that you can use suffixes to distinguish
2802
between different variants.
2803
@end defmac
2804
 
2805
@defmac EXTRA_CONSTRAINT (@var{value}, @var{c})
2806
A C expression that defines the optional machine-dependent constraint
2807
letters that can be used to segregate specific types of operands, usually
2808
memory references, for the target machine.  Any letter that is not
2809
elsewhere defined and not matched by @code{REG_CLASS_FROM_LETTER} /
2810
@code{REG_CLASS_FROM_CONSTRAINT}
2811
may be used.  Normally this macro will not be defined.
2812
 
2813
If it is required for a particular target machine, it should return 1
2814
if @var{value} corresponds to the operand type represented by the
2815
constraint letter @var{c}.  If @var{c} is not defined as an extra
2816
constraint, the value returned should be 0 regardless of @var{value}.
2817
 
2818
For example, on the ROMP, load instructions cannot have their output
2819
in r0 if the memory reference contains a symbolic address.  Constraint
2820
letter @samp{Q} is defined as representing a memory address that does
2821
@emph{not} contain a symbolic address.  An alternative is specified with
2822
a @samp{Q} constraint on the input and @samp{r} on the output.  The next
2823
alternative specifies @samp{m} on the input and a register class that
2824
does not include r0 on the output.
2825
@end defmac
2826
 
2827
@defmac EXTRA_CONSTRAINT_STR (@var{value}, @var{c}, @var{str})
2828
Like @code{EXTRA_CONSTRAINT}, but you also get the constraint string passed
2829
in @var{str}, so that you can use suffixes to distinguish between different
2830
variants.
2831
@end defmac
2832
 
2833
@defmac EXTRA_MEMORY_CONSTRAINT (@var{c}, @var{str})
2834
A C expression that defines the optional machine-dependent constraint
2835
letters, amongst those accepted by @code{EXTRA_CONSTRAINT}, that should
2836
be treated like memory constraints by the reload pass.
2837
 
2838
It should return 1 if the operand type represented by the constraint
2839
at the start of @var{str}, the first letter of which is the letter @var{c},
2840
 comprises a subset of all memory references including
2841
all those whose address is simply a base register.  This allows the reload
2842
pass to reload an operand, if it does not directly correspond to the operand
2843
type of @var{c}, by copying its address into a base register.
2844
 
2845
For example, on the S/390, some instructions do not accept arbitrary
2846
memory references, but only those that do not make use of an index
2847
register.  The constraint letter @samp{Q} is defined via
2848
@code{EXTRA_CONSTRAINT} as representing a memory address of this type.
2849
If the letter @samp{Q} is marked as @code{EXTRA_MEMORY_CONSTRAINT},
2850
a @samp{Q} constraint can handle any memory operand, because the
2851
reload pass knows it can be reloaded by copying the memory address
2852
into a base register if required.  This is analogous to the way
2853
a @samp{o} constraint can handle any memory operand.
2854
@end defmac
2855
 
2856
@defmac EXTRA_ADDRESS_CONSTRAINT (@var{c}, @var{str})
2857
A C expression that defines the optional machine-dependent constraint
2858
letters, amongst those accepted by @code{EXTRA_CONSTRAINT} /
2859
@code{EXTRA_CONSTRAINT_STR}, that should
2860
be treated like address constraints by the reload pass.
2861
 
2862
It should return 1 if the operand type represented by the constraint
2863
at the start of @var{str}, which starts with the letter @var{c}, comprises
2864
a subset of all memory addresses including
2865
all those that consist of just a base register.  This allows the reload
2866
pass to reload an operand, if it does not directly correspond to the operand
2867
type of @var{str}, by copying it into a base register.
2868
 
2869
Any constraint marked as @code{EXTRA_ADDRESS_CONSTRAINT} can only
2870
be used with the @code{address_operand} predicate.  It is treated
2871
analogously to the @samp{p} constraint.
2872
@end defmac
2873
 
2874
@node Stack and Calling
2875
@section Stack Layout and Calling Conventions
2876
@cindex calling conventions
2877
 
2878
@c prevent bad page break with this line
2879
This describes the stack layout and calling conventions.
2880
 
2881
@menu
2882
* Frame Layout::
2883
* Exception Handling::
2884
* Stack Checking::
2885
* Frame Registers::
2886
* Elimination::
2887
* Stack Arguments::
2888
* Register Arguments::
2889
* Scalar Return::
2890
* Aggregate Return::
2891
* Caller Saves::
2892
* Function Entry::
2893
* Profiling::
2894
* Tail Calls::
2895
* Stack Smashing Protection::
2896
@end menu
2897
 
2898
@node Frame Layout
2899
@subsection Basic Stack Layout
2900
@cindex stack frame layout
2901
@cindex frame layout
2902
 
2903
@c prevent bad page break with this line
2904
Here is the basic stack layout.
2905
 
2906
@defmac STACK_GROWS_DOWNWARD
2907
Define this macro if pushing a word onto the stack moves the stack
2908
pointer to a smaller address.
2909
 
2910
When we say, ``define this macro if @dots{}'', it means that the
2911
compiler checks this macro only with @code{#ifdef} so the precise
2912
definition used does not matter.
2913
@end defmac
2914
 
2915
@defmac STACK_PUSH_CODE
2916
This macro defines the operation used when something is pushed
2917
on the stack.  In RTL, a push operation will be
2918
@code{(set (mem (STACK_PUSH_CODE (reg sp))) @dots{})}
2919
 
2920
The choices are @code{PRE_DEC}, @code{POST_DEC}, @code{PRE_INC},
2921
and @code{POST_INC}.  Which of these is correct depends on
2922
the stack direction and on whether the stack pointer points
2923
to the last item on the stack or whether it points to the
2924
space for the next item on the stack.
2925
 
2926
The default is @code{PRE_DEC} when @code{STACK_GROWS_DOWNWARD} is
2927
defined, which is almost always right, and @code{PRE_INC} otherwise,
2928
which is often wrong.
2929
@end defmac
2930
 
2931
@defmac FRAME_GROWS_DOWNWARD
2932
Define this macro to nonzero value if the addresses of local variable slots
2933
are at negative offsets from the frame pointer.
2934
@end defmac
2935
 
2936
@defmac ARGS_GROW_DOWNWARD
2937
Define this macro if successive arguments to a function occupy decreasing
2938
addresses on the stack.
2939
@end defmac
2940
 
2941
@defmac STARTING_FRAME_OFFSET
2942
Offset from the frame pointer to the first local variable slot to be allocated.
2943
 
2944
If @code{FRAME_GROWS_DOWNWARD}, find the next slot's offset by
2945
subtracting the first slot's length from @code{STARTING_FRAME_OFFSET}.
2946
Otherwise, it is found by adding the length of the first slot to the
2947
value @code{STARTING_FRAME_OFFSET}.
2948
@c i'm not sure if the above is still correct.. had to change it to get
2949
@c rid of an overfull.  --mew 2feb93
2950
@end defmac
2951
 
2952
@defmac STACK_ALIGNMENT_NEEDED
2953
Define to zero to disable final alignment of the stack during reload.
2954
The nonzero default for this macro is suitable for most ports.
2955
 
2956
On ports where @code{STARTING_FRAME_OFFSET} is nonzero or where there
2957
is a register save block following the local block that doesn't require
2958
alignment to @code{STACK_BOUNDARY}, it may be beneficial to disable
2959
stack alignment and do it in the backend.
2960
@end defmac
2961
 
2962
@defmac STACK_POINTER_OFFSET
2963
Offset from the stack pointer register to the first location at which
2964
outgoing arguments are placed.  If not specified, the default value of
2965
zero is used.  This is the proper value for most machines.
2966
 
2967
If @code{ARGS_GROW_DOWNWARD}, this is the offset to the location above
2968
the first location at which outgoing arguments are placed.
2969
@end defmac
2970
 
2971
@defmac FIRST_PARM_OFFSET (@var{fundecl})
2972
Offset from the argument pointer register to the first argument's
2973
address.  On some machines it may depend on the data type of the
2974
function.
2975
 
2976
If @code{ARGS_GROW_DOWNWARD}, this is the offset to the location above
2977
the first argument's address.
2978
@end defmac
2979
 
2980
@defmac STACK_DYNAMIC_OFFSET (@var{fundecl})
2981
Offset from the stack pointer register to an item dynamically allocated
2982
on the stack, e.g., by @code{alloca}.
2983
 
2984
The default value for this macro is @code{STACK_POINTER_OFFSET} plus the
2985
length of the outgoing arguments.  The default is correct for most
2986
machines.  See @file{function.c} for details.
2987
@end defmac
2988
 
2989
@defmac INITIAL_FRAME_ADDRESS_RTX
2990
A C expression whose value is RTL representing the address of the initial
2991
stack frame. This address is passed to @code{RETURN_ADDR_RTX} and
2992
@code{DYNAMIC_CHAIN_ADDRESS}.  If you don't define this macro, a reasonable
2993
default value will be used.  Define this macro in order to make frame pointer
2994
elimination work in the presence of @code{__builtin_frame_address (count)} and
2995
@code{__builtin_return_address (count)} for @code{count} not equal to zero.
2996
@end defmac
2997
 
2998
@defmac DYNAMIC_CHAIN_ADDRESS (@var{frameaddr})
2999
A C expression whose value is RTL representing the address in a stack
3000
frame where the pointer to the caller's frame is stored.  Assume that
3001
@var{frameaddr} is an RTL expression for the address of the stack frame
3002
itself.
3003
 
3004
If you don't define this macro, the default is to return the value
3005
of @var{frameaddr}---that is, the stack frame address is also the
3006
address of the stack word that points to the previous frame.
3007
@end defmac
3008
 
3009
@defmac SETUP_FRAME_ADDRESSES
3010
If defined, a C expression that produces the machine-specific code to
3011
setup the stack so that arbitrary frames can be accessed.  For example,
3012
on the SPARC, we must flush all of the register windows to the stack
3013
before we can access arbitrary stack frames.  You will seldom need to
3014
define this macro.
3015
@end defmac
3016
 
3017
@deftypefn {Target Hook} bool TARGET_BUILTIN_SETJMP_FRAME_VALUE ()
3018
This target hook should return an rtx that is used to store
3019
the address of the current frame into the built in @code{setjmp} buffer.
3020
The default value, @code{virtual_stack_vars_rtx}, is correct for most
3021
machines.  One reason you may need to define this target hook is if
3022
@code{hard_frame_pointer_rtx} is the appropriate value on your machine.
3023
@end deftypefn
3024
 
3025
@defmac FRAME_ADDR_RTX (@var{frameaddr})
3026
A C expression whose value is RTL representing the value of the frame
3027
address for the current frame.  @var{frameaddr} is the frame pointer
3028
of the current frame.  This is used for __builtin_frame_address.
3029
You need only define this macro if the frame address is not the same
3030
as the frame pointer.  Most machines do not need to define it.
3031
@end defmac
3032
 
3033
@defmac RETURN_ADDR_RTX (@var{count}, @var{frameaddr})
3034
A C expression whose value is RTL representing the value of the return
3035
address for the frame @var{count} steps up from the current frame, after
3036
the prologue.  @var{frameaddr} is the frame pointer of the @var{count}
3037
frame, or the frame pointer of the @var{count} @minus{} 1 frame if
3038
@code{RETURN_ADDR_IN_PREVIOUS_FRAME} is defined.
3039
 
3040
The value of the expression must always be the correct address when
3041
@var{count} is zero, but may be @code{NULL_RTX} if there is not way to
3042
determine the return address of other frames.
3043
@end defmac
3044
 
3045
@defmac RETURN_ADDR_IN_PREVIOUS_FRAME
3046
Define this if the return address of a particular stack frame is accessed
3047
from the frame pointer of the previous stack frame.
3048
@end defmac
3049
 
3050
@defmac INCOMING_RETURN_ADDR_RTX
3051
A C expression whose value is RTL representing the location of the
3052
incoming return address at the beginning of any function, before the
3053
prologue.  This RTL is either a @code{REG}, indicating that the return
3054
value is saved in @samp{REG}, or a @code{MEM} representing a location in
3055
the stack.
3056
 
3057
You only need to define this macro if you want to support call frame
3058
debugging information like that provided by DWARF 2.
3059
 
3060
If this RTL is a @code{REG}, you should also define
3061
@code{DWARF_FRAME_RETURN_COLUMN} to @code{DWARF_FRAME_REGNUM (REGNO)}.
3062
@end defmac
3063
 
3064
@defmac DWARF_ALT_FRAME_RETURN_COLUMN
3065
A C expression whose value is an integer giving a DWARF 2 column
3066
number that may be used as an alternate return column.  This should
3067
be defined only if @code{DWARF_FRAME_RETURN_COLUMN} is set to a
3068
general register, but an alternate column needs to be used for
3069
signal frames.
3070
@end defmac
3071
 
3072
@defmac DWARF_ZERO_REG
3073
A C expression whose value is an integer giving a DWARF 2 register
3074
number that is considered to always have the value zero.  This should
3075
only be defined if the target has an architected zero register, and
3076
someone decided it was a good idea to use that register number to
3077
terminate the stack backtrace.  New ports should avoid this.
3078
@end defmac
3079
 
3080
@deftypefn {Target Hook} void TARGET_DWARF_HANDLE_FRAME_UNSPEC (const char *@var{label}, rtx @var{pattern}, int @var{index})
3081
This target hook allows the backend to emit frame-related insns that
3082
contain UNSPECs or UNSPEC_VOLATILEs.  The DWARF 2 call frame debugging
3083
info engine will invoke it on insns of the form
3084
@smallexample
3085
(set (reg) (unspec [...] UNSPEC_INDEX))
3086
@end smallexample
3087
and
3088
@smallexample
3089
(set (reg) (unspec_volatile [...] UNSPECV_INDEX)).
3090
@end smallexample
3091
to let the backend emit the call frame instructions.  @var{label} is
3092
the CFI label attached to the insn, @var{pattern} is the pattern of
3093
the insn and @var{index} is @code{UNSPEC_INDEX} or @code{UNSPECV_INDEX}.
3094
@end deftypefn
3095
 
3096
@defmac INCOMING_FRAME_SP_OFFSET
3097
A C expression whose value is an integer giving the offset, in bytes,
3098
from the value of the stack pointer register to the top of the stack
3099
frame at the beginning of any function, before the prologue.  The top of
3100
the frame is defined to be the value of the stack pointer in the
3101
previous frame, just before the call instruction.
3102
 
3103
You only need to define this macro if you want to support call frame
3104
debugging information like that provided by DWARF 2.
3105
@end defmac
3106
 
3107
@defmac ARG_POINTER_CFA_OFFSET (@var{fundecl})
3108
A C expression whose value is an integer giving the offset, in bytes,
3109
from the argument pointer to the canonical frame address (cfa).  The
3110
final value should coincide with that calculated by
3111
@code{INCOMING_FRAME_SP_OFFSET}.  Which is unfortunately not usable
3112
during virtual register instantiation.
3113
 
3114
The default value for this macro is @code{FIRST_PARM_OFFSET (fundecl)},
3115
which is correct for most machines; in general, the arguments are found
3116
immediately before the stack frame.  Note that this is not the case on
3117
some targets that save registers into the caller's frame, such as SPARC
3118
and rs6000, and so such targets need to define this macro.
3119
 
3120
You only need to define this macro if the default is incorrect, and you
3121
want to support call frame debugging information like that provided by
3122
DWARF 2.
3123
@end defmac
3124
 
3125
@defmac FRAME_POINTER_CFA_OFFSET (@var{fundecl})
3126
If defined, a C expression whose value is an integer giving the offset
3127
in bytes from the frame pointer to the canonical frame address (cfa).
3128
The final value should coincide with that calculated by
3129
@code{INCOMING_FRAME_SP_OFFSET}.
3130
 
3131
Normally the CFA is calculated as an offset from the argument pointer,
3132
via @code{ARG_POINTER_CFA_OFFSET}, but if the argument pointer is
3133
variable due to the ABI, this may not be possible.  If this macro is
3134
defined, it implies that the virtual register instantiation should be
3135
based on the frame pointer instead of the argument pointer.  Only one
3136
of @code{FRAME_POINTER_CFA_OFFSET} and @code{ARG_POINTER_CFA_OFFSET}
3137
should be defined.
3138
@end defmac
3139
 
3140
@defmac CFA_FRAME_BASE_OFFSET (@var{fundecl})
3141
If defined, a C expression whose value is an integer giving the offset
3142
in bytes from the canonical frame address (cfa) to the frame base used
3143
in DWARF 2 debug information.  The default is zero.  A different value
3144
may reduce the size of debug information on some ports.
3145
@end defmac
3146
 
3147
@node Exception Handling
3148
@subsection Exception Handling Support
3149
@cindex exception handling
3150
 
3151
@defmac EH_RETURN_DATA_REGNO (@var{N})
3152
A C expression whose value is the @var{N}th register number used for
3153
data by exception handlers, or @code{INVALID_REGNUM} if fewer than
3154
@var{N} registers are usable.
3155
 
3156
The exception handling library routines communicate with the exception
3157
handlers via a set of agreed upon registers.  Ideally these registers
3158
should be call-clobbered; it is possible to use call-saved registers,
3159
but may negatively impact code size.  The target must support at least
3160
2 data registers, but should define 4 if there are enough free registers.
3161
 
3162
You must define this macro if you want to support call frame exception
3163
handling like that provided by DWARF 2.
3164
@end defmac
3165
 
3166
@defmac EH_RETURN_STACKADJ_RTX
3167
A C expression whose value is RTL representing a location in which
3168
to store a stack adjustment to be applied before function return.
3169
This is used to unwind the stack to an exception handler's call frame.
3170
It will be assigned zero on code paths that return normally.
3171
 
3172
Typically this is a call-clobbered hard register that is otherwise
3173
untouched by the epilogue, but could also be a stack slot.
3174
 
3175
Do not define this macro if the stack pointer is saved and restored
3176
by the regular prolog and epilog code in the call frame itself; in
3177
this case, the exception handling library routines will update the
3178
stack location to be restored in place.  Otherwise, you must define
3179
this macro if you want to support call frame exception handling like
3180
that provided by DWARF 2.
3181
@end defmac
3182
 
3183
@defmac EH_RETURN_HANDLER_RTX
3184
A C expression whose value is RTL representing a location in which
3185
to store the address of an exception handler to which we should
3186
return.  It will not be assigned on code paths that return normally.
3187
 
3188
Typically this is the location in the call frame at which the normal
3189
return address is stored.  For targets that return by popping an
3190
address off the stack, this might be a memory address just below
3191
the @emph{target} call frame rather than inside the current call
3192
frame.  If defined, @code{EH_RETURN_STACKADJ_RTX} will have already
3193
been assigned, so it may be used to calculate the location of the
3194
target call frame.
3195
 
3196
Some targets have more complex requirements than storing to an
3197
address calculable during initial code generation.  In that case
3198
the @code{eh_return} instruction pattern should be used instead.
3199
 
3200
If you want to support call frame exception handling, you must
3201
define either this macro or the @code{eh_return} instruction pattern.
3202
@end defmac
3203
 
3204
@defmac RETURN_ADDR_OFFSET
3205
If defined, an integer-valued C expression for which rtl will be generated
3206
to add it to the exception handler address before it is searched in the
3207
exception handling tables, and to subtract it again from the address before
3208
using it to return to the exception handler.
3209
@end defmac
3210
 
3211
@defmac ASM_PREFERRED_EH_DATA_FORMAT (@var{code}, @var{global})
3212
This macro chooses the encoding of pointers embedded in the exception
3213
handling sections.  If at all possible, this should be defined such
3214
that the exception handling section will not require dynamic relocations,
3215
and so may be read-only.
3216
 
3217
@var{code} is 0 for data, 1 for code labels, 2 for function pointers.
3218
@var{global} is true if the symbol may be affected by dynamic relocations.
3219
The macro should return a combination of the @code{DW_EH_PE_*} defines
3220
as found in @file{dwarf2.h}.
3221
 
3222
If this macro is not defined, pointers will not be encoded but
3223
represented directly.
3224
@end defmac
3225
 
3226
@defmac ASM_MAYBE_OUTPUT_ENCODED_ADDR_RTX (@var{file}, @var{encoding}, @var{size}, @var{addr}, @var{done})
3227
This macro allows the target to emit whatever special magic is required
3228
to represent the encoding chosen by @code{ASM_PREFERRED_EH_DATA_FORMAT}.
3229
Generic code takes care of pc-relative and indirect encodings; this must
3230
be defined if the target uses text-relative or data-relative encodings.
3231
 
3232
This is a C statement that branches to @var{done} if the format was
3233
handled.  @var{encoding} is the format chosen, @var{size} is the number
3234
of bytes that the format occupies, @var{addr} is the @code{SYMBOL_REF}
3235
to be emitted.
3236
@end defmac
3237
 
3238
@defmac MD_UNWIND_SUPPORT
3239
A string specifying a file to be #include'd in unwind-dw2.c.  The file
3240
so included typically defines @code{MD_FALLBACK_FRAME_STATE_FOR}.
3241
@end defmac
3242
 
3243
@defmac MD_FALLBACK_FRAME_STATE_FOR (@var{context}, @var{fs})
3244
This macro allows the target to add cpu and operating system specific
3245
code to the call-frame unwinder for use when there is no unwind data
3246
available.  The most common reason to implement this macro is to unwind
3247
through signal frames.
3248
 
3249
This macro is called from @code{uw_frame_state_for} in @file{unwind-dw2.c}
3250
and @file{unwind-ia64.c}.  @var{context} is an @code{_Unwind_Context};
3251
@var{fs} is an @code{_Unwind_FrameState}.  Examine @code{context->ra}
3252
for the address of the code being executed and @code{context->cfa} for
3253
the stack pointer value.  If the frame can be decoded, the register save
3254
addresses should be updated in @var{fs} and the macro should evaluate to
3255
@code{_URC_NO_REASON}.  If the frame cannot be decoded, the macro should
3256
evaluate to @code{_URC_END_OF_STACK}.
3257
 
3258
For proper signal handling in Java this macro is accompanied by
3259
@code{MAKE_THROW_FRAME}, defined in @file{libjava/include/*-signal.h} headers.
3260
@end defmac
3261
 
3262
@defmac MD_HANDLE_UNWABI (@var{context}, @var{fs})
3263
This macro allows the target to add operating system specific code to the
3264
call-frame unwinder to handle the IA-64 @code{.unwabi} unwinding directive,
3265
usually used for signal or interrupt frames.
3266
 
3267
This macro is called from @code{uw_update_context} in @file{unwind-ia64.c}.
3268
@var{context} is an @code{_Unwind_Context};
3269
@var{fs} is an @code{_Unwind_FrameState}.  Examine @code{fs->unwabi}
3270
for the abi and context in the @code{.unwabi} directive.  If the
3271
@code{.unwabi} directive can be handled, the register save addresses should
3272
be updated in @var{fs}.
3273
@end defmac
3274
 
3275
@defmac TARGET_USES_WEAK_UNWIND_INFO
3276
A C expression that evaluates to true if the target requires unwind
3277
info to be given comdat linkage.  Define it to be @code{1} if comdat
3278
linkage is necessary.  The default is @code{0}.
3279
@end defmac
3280
 
3281
@node Stack Checking
3282
@subsection Specifying How Stack Checking is Done
3283
 
3284
GCC will check that stack references are within the boundaries of
3285
the stack, if the @option{-fstack-check} is specified, in one of three ways:
3286
 
3287
@enumerate
3288
@item
3289
If the value of the @code{STACK_CHECK_BUILTIN} macro is nonzero, GCC
3290
will assume that you have arranged for stack checking to be done at
3291
appropriate places in the configuration files, e.g., in
3292
@code{TARGET_ASM_FUNCTION_PROLOGUE}.  GCC will do not other special
3293
processing.
3294
 
3295
@item
3296
If @code{STACK_CHECK_BUILTIN} is zero and you defined a named pattern
3297
called @code{check_stack} in your @file{md} file, GCC will call that
3298
pattern with one argument which is the address to compare the stack
3299
value against.  You must arrange for this pattern to report an error if
3300
the stack pointer is out of range.
3301
 
3302
@item
3303
If neither of the above are true, GCC will generate code to periodically
3304
``probe'' the stack pointer using the values of the macros defined below.
3305
@end enumerate
3306
 
3307
Normally, you will use the default values of these macros, so GCC
3308
will use the third approach.
3309
 
3310
@defmac STACK_CHECK_BUILTIN
3311
A nonzero value if stack checking is done by the configuration files in a
3312
machine-dependent manner.  You should define this macro if stack checking
3313
is require by the ABI of your machine or if you would like to have to stack
3314
checking in some more efficient way than GCC's portable approach.
3315
The default value of this macro is zero.
3316
@end defmac
3317
 
3318
@defmac STACK_CHECK_PROBE_INTERVAL
3319
An integer representing the interval at which GCC must generate stack
3320
probe instructions.  You will normally define this macro to be no larger
3321
than the size of the ``guard pages'' at the end of a stack area.  The
3322
default value of 4096 is suitable for most systems.
3323
@end defmac
3324
 
3325
@defmac STACK_CHECK_PROBE_LOAD
3326
A integer which is nonzero if GCC should perform the stack probe
3327
as a load instruction and zero if GCC should use a store instruction.
3328
The default is zero, which is the most efficient choice on most systems.
3329
@end defmac
3330
 
3331
@defmac STACK_CHECK_PROTECT
3332
The number of bytes of stack needed to recover from a stack overflow,
3333
for languages where such a recovery is supported.  The default value of
3334
75 words should be adequate for most machines.
3335
@end defmac
3336
 
3337
@defmac STACK_CHECK_MAX_FRAME_SIZE
3338
The maximum size of a stack frame, in bytes.  GCC will generate probe
3339
instructions in non-leaf functions to ensure at least this many bytes of
3340
stack are available.  If a stack frame is larger than this size, stack
3341
checking will not be reliable and GCC will issue a warning.  The
3342
default is chosen so that GCC only generates one instruction on most
3343
systems.  You should normally not change the default value of this macro.
3344
@end defmac
3345
 
3346
@defmac STACK_CHECK_FIXED_FRAME_SIZE
3347
GCC uses this value to generate the above warning message.  It
3348
represents the amount of fixed frame used by a function, not including
3349
space for any callee-saved registers, temporaries and user variables.
3350
You need only specify an upper bound for this amount and will normally
3351
use the default of four words.
3352
@end defmac
3353
 
3354
@defmac STACK_CHECK_MAX_VAR_SIZE
3355
The maximum size, in bytes, of an object that GCC will place in the
3356
fixed area of the stack frame when the user specifies
3357
@option{-fstack-check}.
3358
GCC computed the default from the values of the above macros and you will
3359
normally not need to override that default.
3360
@end defmac
3361
 
3362
@need 2000
3363
@node Frame Registers
3364
@subsection Registers That Address the Stack Frame
3365
 
3366
@c prevent bad page break with this line
3367
This discusses registers that address the stack frame.
3368
 
3369
@defmac STACK_POINTER_REGNUM
3370
The register number of the stack pointer register, which must also be a
3371
fixed register according to @code{FIXED_REGISTERS}.  On most machines,
3372
the hardware determines which register this is.
3373
@end defmac
3374
 
3375
@defmac FRAME_POINTER_REGNUM
3376
The register number of the frame pointer register, which is used to
3377
access automatic variables in the stack frame.  On some machines, the
3378
hardware determines which register this is.  On other machines, you can
3379
choose any register you wish for this purpose.
3380
@end defmac
3381
 
3382
@defmac HARD_FRAME_POINTER_REGNUM
3383
On some machines the offset between the frame pointer and starting
3384
offset of the automatic variables is not known until after register
3385
allocation has been done (for example, because the saved registers are
3386
between these two locations).  On those machines, define
3387
@code{FRAME_POINTER_REGNUM} the number of a special, fixed register to
3388
be used internally until the offset is known, and define
3389
@code{HARD_FRAME_POINTER_REGNUM} to be the actual hard register number
3390
used for the frame pointer.
3391
 
3392
You should define this macro only in the very rare circumstances when it
3393
is not possible to calculate the offset between the frame pointer and
3394
the automatic variables until after register allocation has been
3395
completed.  When this macro is defined, you must also indicate in your
3396
definition of @code{ELIMINABLE_REGS} how to eliminate
3397
@code{FRAME_POINTER_REGNUM} into either @code{HARD_FRAME_POINTER_REGNUM}
3398
or @code{STACK_POINTER_REGNUM}.
3399
 
3400
Do not define this macro if it would be the same as
3401
@code{FRAME_POINTER_REGNUM}.
3402
@end defmac
3403
 
3404
@defmac ARG_POINTER_REGNUM
3405
The register number of the arg pointer register, which is used to access
3406
the function's argument list.  On some machines, this is the same as the
3407
frame pointer register.  On some machines, the hardware determines which
3408
register this is.  On other machines, you can choose any register you
3409
wish for this purpose.  If this is not the same register as the frame
3410
pointer register, then you must mark it as a fixed register according to
3411
@code{FIXED_REGISTERS}, or arrange to be able to eliminate it
3412
(@pxref{Elimination}).
3413
@end defmac
3414
 
3415
@defmac RETURN_ADDRESS_POINTER_REGNUM
3416
The register number of the return address pointer register, which is used to
3417
access the current function's return address from the stack.  On some
3418
machines, the return address is not at a fixed offset from the frame
3419
pointer or stack pointer or argument pointer.  This register can be defined
3420
to point to the return address on the stack, and then be converted by
3421
@code{ELIMINABLE_REGS} into either the frame pointer or stack pointer.
3422
 
3423
Do not define this macro unless there is no other way to get the return
3424
address from the stack.
3425
@end defmac
3426
 
3427
@defmac STATIC_CHAIN_REGNUM
3428
@defmacx STATIC_CHAIN_INCOMING_REGNUM
3429
Register numbers used for passing a function's static chain pointer.  If
3430
register windows are used, the register number as seen by the called
3431
function is @code{STATIC_CHAIN_INCOMING_REGNUM}, while the register
3432
number as seen by the calling function is @code{STATIC_CHAIN_REGNUM}.  If
3433
these registers are the same, @code{STATIC_CHAIN_INCOMING_REGNUM} need
3434
not be defined.
3435
 
3436
The static chain register need not be a fixed register.
3437
 
3438
If the static chain is passed in memory, these macros should not be
3439
defined; instead, the next two macros should be defined.
3440
@end defmac
3441
 
3442
@defmac STATIC_CHAIN
3443
@defmacx STATIC_CHAIN_INCOMING
3444
If the static chain is passed in memory, these macros provide rtx giving
3445
@code{mem} expressions that denote where they are stored.
3446
@code{STATIC_CHAIN} and @code{STATIC_CHAIN_INCOMING} give the locations
3447
as seen by the calling and called functions, respectively.  Often the former
3448
will be at an offset from the stack pointer and the latter at an offset from
3449
the frame pointer.
3450
 
3451
@findex stack_pointer_rtx
3452
@findex frame_pointer_rtx
3453
@findex arg_pointer_rtx
3454
The variables @code{stack_pointer_rtx}, @code{frame_pointer_rtx}, and
3455
@code{arg_pointer_rtx} will have been initialized prior to the use of these
3456
macros and should be used to refer to those items.
3457
 
3458
If the static chain is passed in a register, the two previous macros should
3459
be defined instead.
3460
@end defmac
3461
 
3462
@defmac DWARF_FRAME_REGISTERS
3463
This macro specifies the maximum number of hard registers that can be
3464
saved in a call frame.  This is used to size data structures used in
3465
DWARF2 exception handling.
3466
 
3467
Prior to GCC 3.0, this macro was needed in order to establish a stable
3468
exception handling ABI in the face of adding new hard registers for ISA
3469
extensions.  In GCC 3.0 and later, the EH ABI is insulated from changes
3470
in the number of hard registers.  Nevertheless, this macro can still be
3471
used to reduce the runtime memory requirements of the exception handling
3472
routines, which can be substantial if the ISA contains a lot of
3473
registers that are not call-saved.
3474
 
3475
If this macro is not defined, it defaults to
3476
@code{FIRST_PSEUDO_REGISTER}.
3477
@end defmac
3478
 
3479
@defmac PRE_GCC3_DWARF_FRAME_REGISTERS
3480
 
3481
This macro is similar to @code{DWARF_FRAME_REGISTERS}, but is provided
3482
for backward compatibility in pre GCC 3.0 compiled code.
3483
 
3484
If this macro is not defined, it defaults to
3485
@code{DWARF_FRAME_REGISTERS}.
3486
@end defmac
3487
 
3488
@defmac DWARF_REG_TO_UNWIND_COLUMN (@var{regno})
3489
 
3490
Define this macro if the target's representation for dwarf registers
3491
is different than the internal representation for unwind column.
3492
Given a dwarf register, this macro should return the internal unwind
3493
column number to use instead.
3494
 
3495
See the PowerPC's SPE target for an example.
3496
@end defmac
3497
 
3498
@defmac DWARF_FRAME_REGNUM (@var{regno})
3499
 
3500
Define this macro if the target's representation for dwarf registers
3501
used in .eh_frame or .debug_frame is different from that used in other
3502
debug info sections.  Given a GCC hard register number, this macro
3503
should return the .eh_frame register number.  The default is
3504
@code{DBX_REGISTER_NUMBER (@var{regno})}.
3505
 
3506
@end defmac
3507
 
3508
@defmac DWARF2_FRAME_REG_OUT (@var{regno}, @var{for_eh})
3509
 
3510
Define this macro to map register numbers held in the call frame info
3511
that GCC has collected using @code{DWARF_FRAME_REGNUM} to those that
3512
should be output in .debug_frame (@code{@var{for_eh}} is zero) and
3513
.eh_frame (@code{@var{for_eh}} is nonzero).  The default is to
3514
return @code{@var{regno}}.
3515
 
3516
@end defmac
3517
 
3518
@node Elimination
3519
@subsection Eliminating Frame Pointer and Arg Pointer
3520
 
3521
@c prevent bad page break with this line
3522
This is about eliminating the frame pointer and arg pointer.
3523
 
3524
@defmac FRAME_POINTER_REQUIRED
3525
A C expression which is nonzero if a function must have and use a frame
3526
pointer.  This expression is evaluated  in the reload pass.  If its value is
3527
nonzero the function will have a frame pointer.
3528
 
3529
The expression can in principle examine the current function and decide
3530
according to the facts, but on most machines the constant 0 or the
3531
constant 1 suffices.  Use 0 when the machine allows code to be generated
3532
with no frame pointer, and doing so saves some time or space.  Use 1
3533
when there is no possible advantage to avoiding a frame pointer.
3534
 
3535
In certain cases, the compiler does not know how to produce valid code
3536
without a frame pointer.  The compiler recognizes those cases and
3537
automatically gives the function a frame pointer regardless of what
3538
@code{FRAME_POINTER_REQUIRED} says.  You don't need to worry about
3539
them.
3540
 
3541
In a function that does not require a frame pointer, the frame pointer
3542
register can be allocated for ordinary usage, unless you mark it as a
3543
fixed register.  See @code{FIXED_REGISTERS} for more information.
3544
@end defmac
3545
 
3546
@findex get_frame_size
3547
@defmac INITIAL_FRAME_POINTER_OFFSET (@var{depth-var})
3548
A C statement to store in the variable @var{depth-var} the difference
3549
between the frame pointer and the stack pointer values immediately after
3550
the function prologue.  The value would be computed from information
3551
such as the result of @code{get_frame_size ()} and the tables of
3552
registers @code{regs_ever_live} and @code{call_used_regs}.
3553
 
3554
If @code{ELIMINABLE_REGS} is defined, this macro will be not be used and
3555
need not be defined.  Otherwise, it must be defined even if
3556
@code{FRAME_POINTER_REQUIRED} is defined to always be true; in that
3557
case, you may set @var{depth-var} to anything.
3558
@end defmac
3559
 
3560
@defmac ELIMINABLE_REGS
3561
If defined, this macro specifies a table of register pairs used to
3562
eliminate unneeded registers that point into the stack frame.  If it is not
3563
defined, the only elimination attempted by the compiler is to replace
3564
references to the frame pointer with references to the stack pointer.
3565
 
3566
The definition of this macro is a list of structure initializations, each
3567
of which specifies an original and replacement register.
3568
 
3569
On some machines, the position of the argument pointer is not known until
3570
the compilation is completed.  In such a case, a separate hard register
3571
must be used for the argument pointer.  This register can be eliminated by
3572
replacing it with either the frame pointer or the argument pointer,
3573
depending on whether or not the frame pointer has been eliminated.
3574
 
3575
In this case, you might specify:
3576
@smallexample
3577
#define ELIMINABLE_REGS  \
3578
@{@{ARG_POINTER_REGNUM, STACK_POINTER_REGNUM@}, \
3579
 @{ARG_POINTER_REGNUM, FRAME_POINTER_REGNUM@}, \
3580
 @{FRAME_POINTER_REGNUM, STACK_POINTER_REGNUM@}@}
3581
@end smallexample
3582
 
3583
Note that the elimination of the argument pointer with the stack pointer is
3584
specified first since that is the preferred elimination.
3585
@end defmac
3586
 
3587
@defmac CAN_ELIMINATE (@var{from-reg}, @var{to-reg})
3588
A C expression that returns nonzero if the compiler is allowed to try
3589
to replace register number @var{from-reg} with register number
3590
@var{to-reg}.  This macro need only be defined if @code{ELIMINABLE_REGS}
3591
is defined, and will usually be the constant 1, since most of the cases
3592
preventing register elimination are things that the compiler already
3593
knows about.
3594
@end defmac
3595
 
3596
@defmac INITIAL_ELIMINATION_OFFSET (@var{from-reg}, @var{to-reg}, @var{offset-var})
3597
This macro is similar to @code{INITIAL_FRAME_POINTER_OFFSET}.  It
3598
specifies the initial difference between the specified pair of
3599
registers.  This macro must be defined if @code{ELIMINABLE_REGS} is
3600
defined.
3601
@end defmac
3602
 
3603
@node Stack Arguments
3604
@subsection Passing Function Arguments on the Stack
3605
@cindex arguments on stack
3606
@cindex stack arguments
3607
 
3608
The macros in this section control how arguments are passed
3609
on the stack.  See the following section for other macros that
3610
control passing certain arguments in registers.
3611
 
3612
@deftypefn {Target Hook} bool TARGET_PROMOTE_PROTOTYPES (tree @var{fntype})
3613
This target hook returns @code{true} if an argument declared in a
3614
prototype as an integral type smaller than @code{int} should actually be
3615
passed as an @code{int}.  In addition to avoiding errors in certain
3616
cases of mismatch, it also makes for better code on certain machines.
3617
The default is to not promote prototypes.
3618
@end deftypefn
3619
 
3620
@defmac PUSH_ARGS
3621
A C expression.  If nonzero, push insns will be used to pass
3622
outgoing arguments.
3623
If the target machine does not have a push instruction, set it to zero.
3624
That directs GCC to use an alternate strategy: to
3625
allocate the entire argument block and then store the arguments into
3626
it.  When @code{PUSH_ARGS} is nonzero, @code{PUSH_ROUNDING} must be defined too.
3627
@end defmac
3628
 
3629
@defmac PUSH_ARGS_REVERSED
3630
A C expression.  If nonzero, function arguments will be evaluated from
3631
last to first, rather than from first to last.  If this macro is not
3632
defined, it defaults to @code{PUSH_ARGS} on targets where the stack
3633
and args grow in opposite directions, and 0 otherwise.
3634
@end defmac
3635
 
3636
@defmac PUSH_ROUNDING (@var{npushed})
3637
A C expression that is the number of bytes actually pushed onto the
3638
stack when an instruction attempts to push @var{npushed} bytes.
3639
 
3640
On some machines, the definition
3641
 
3642
@smallexample
3643
#define PUSH_ROUNDING(BYTES) (BYTES)
3644
@end smallexample
3645
 
3646
@noindent
3647
will suffice.  But on other machines, instructions that appear
3648
to push one byte actually push two bytes in an attempt to maintain
3649
alignment.  Then the definition should be
3650
 
3651
@smallexample
3652
#define PUSH_ROUNDING(BYTES) (((BYTES) + 1) & ~1)
3653
@end smallexample
3654
@end defmac
3655
 
3656
@findex current_function_outgoing_args_size
3657
@defmac ACCUMULATE_OUTGOING_ARGS
3658
A C expression.  If nonzero, the maximum amount of space required for outgoing arguments
3659
will be computed and placed into the variable
3660
@code{current_function_outgoing_args_size}.  No space will be pushed
3661
onto the stack for each call; instead, the function prologue should
3662
increase the stack frame size by this amount.
3663
 
3664
Setting both @code{PUSH_ARGS} and @code{ACCUMULATE_OUTGOING_ARGS}
3665
is not proper.
3666
@end defmac
3667
 
3668
@defmac REG_PARM_STACK_SPACE (@var{fndecl})
3669
Define this macro if functions should assume that stack space has been
3670
allocated for arguments even when their values are passed in
3671
registers.
3672
 
3673
The value of this macro is the size, in bytes, of the area reserved for
3674
arguments passed in registers for the function represented by @var{fndecl},
3675
which can be zero if GCC is calling a library function.
3676
 
3677
This space can be allocated by the caller, or be a part of the
3678
machine-dependent stack frame: @code{OUTGOING_REG_PARM_STACK_SPACE} says
3679
which.
3680
@end defmac
3681
@c above is overfull.  not sure what to do.  --mew 5feb93  did
3682
@c something, not sure if it looks good.  --mew 10feb93
3683
 
3684
@defmac OUTGOING_REG_PARM_STACK_SPACE
3685
Define this if it is the responsibility of the caller to allocate the area
3686
reserved for arguments passed in registers.
3687
 
3688
If @code{ACCUMULATE_OUTGOING_ARGS} is defined, this macro controls
3689
whether the space for these arguments counts in the value of
3690
@code{current_function_outgoing_args_size}.
3691
@end defmac
3692
 
3693
@defmac STACK_PARMS_IN_REG_PARM_AREA
3694
Define this macro if @code{REG_PARM_STACK_SPACE} is defined, but the
3695
stack parameters don't skip the area specified by it.
3696
@c i changed this, makes more sens and it should have taken care of the
3697
@c overfull.. not as specific, tho.  --mew 5feb93
3698
 
3699
Normally, when a parameter is not passed in registers, it is placed on the
3700
stack beyond the @code{REG_PARM_STACK_SPACE} area.  Defining this macro
3701
suppresses this behavior and causes the parameter to be passed on the
3702
stack in its natural location.
3703
@end defmac
3704
 
3705
@defmac RETURN_POPS_ARGS (@var{fundecl}, @var{funtype}, @var{stack-size})
3706
A C expression that should indicate the number of bytes of its own
3707
arguments that a function pops on returning, or 0 if the
3708
function pops no arguments and the caller must therefore pop them all
3709
after the function returns.
3710
 
3711
@var{fundecl} is a C variable whose value is a tree node that describes
3712
the function in question.  Normally it is a node of type
3713
@code{FUNCTION_DECL} that describes the declaration of the function.
3714
From this you can obtain the @code{DECL_ATTRIBUTES} of the function.
3715
 
3716
@var{funtype} is a C variable whose value is a tree node that
3717
describes the function in question.  Normally it is a node of type
3718
@code{FUNCTION_TYPE} that describes the data type of the function.
3719
From this it is possible to obtain the data types of the value and
3720
arguments (if known).
3721
 
3722
When a call to a library function is being considered, @var{fundecl}
3723
will contain an identifier node for the library function.  Thus, if
3724
you need to distinguish among various library functions, you can do so
3725
by their names.  Note that ``library function'' in this context means
3726
a function used to perform arithmetic, whose name is known specially
3727
in the compiler and was not mentioned in the C code being compiled.
3728
 
3729
@var{stack-size} is the number of bytes of arguments passed on the
3730
stack.  If a variable number of bytes is passed, it is zero, and
3731
argument popping will always be the responsibility of the calling function.
3732
 
3733
On the VAX, all functions always pop their arguments, so the definition
3734
of this macro is @var{stack-size}.  On the 68000, using the standard
3735
calling convention, no functions pop their arguments, so the value of
3736
the macro is always 0 in this case.  But an alternative calling
3737
convention is available in which functions that take a fixed number of
3738
arguments pop them but other functions (such as @code{printf}) pop
3739
nothing (the caller pops all).  When this convention is in use,
3740
@var{funtype} is examined to determine whether a function takes a fixed
3741
number of arguments.
3742
@end defmac
3743
 
3744
@defmac CALL_POPS_ARGS (@var{cum})
3745
A C expression that should indicate the number of bytes a call sequence
3746
pops off the stack.  It is added to the value of @code{RETURN_POPS_ARGS}
3747
when compiling a function call.
3748
 
3749
@var{cum} is the variable in which all arguments to the called function
3750
have been accumulated.
3751
 
3752
On certain architectures, such as the SH5, a call trampoline is used
3753
that pops certain registers off the stack, depending on the arguments
3754
that have been passed to the function.  Since this is a property of the
3755
call site, not of the called function, @code{RETURN_POPS_ARGS} is not
3756
appropriate.
3757
@end defmac
3758
 
3759
@node Register Arguments
3760
@subsection Passing Arguments in Registers
3761
@cindex arguments in registers
3762
@cindex registers arguments
3763
 
3764
This section describes the macros which let you control how various
3765
types of arguments are passed in registers or how they are arranged in
3766
the stack.
3767
 
3768
@defmac FUNCTION_ARG (@var{cum}, @var{mode}, @var{type}, @var{named})
3769
A C expression that controls whether a function argument is passed
3770
in a register, and which register.
3771
 
3772
The arguments are @var{cum}, which summarizes all the previous
3773
arguments; @var{mode}, the machine mode of the argument; @var{type},
3774
the data type of the argument as a tree node or 0 if that is not known
3775
(which happens for C support library functions); and @var{named},
3776
which is 1 for an ordinary argument and 0 for nameless arguments that
3777
correspond to @samp{@dots{}} in the called function's prototype.
3778
@var{type} can be an incomplete type if a syntax error has previously
3779
occurred.
3780
 
3781
The value of the expression is usually either a @code{reg} RTX for the
3782
hard register in which to pass the argument, or zero to pass the
3783
argument on the stack.
3784
 
3785
For machines like the VAX and 68000, where normally all arguments are
3786
pushed, zero suffices as a definition.
3787
 
3788
The value of the expression can also be a @code{parallel} RTX@.  This is
3789
used when an argument is passed in multiple locations.  The mode of the
3790
@code{parallel} should be the mode of the entire argument.  The
3791
@code{parallel} holds any number of @code{expr_list} pairs; each one
3792
describes where part of the argument is passed.  In each
3793
@code{expr_list} the first operand must be a @code{reg} RTX for the hard
3794
register in which to pass this part of the argument, and the mode of the
3795
register RTX indicates how large this part of the argument is.  The
3796
second operand of the @code{expr_list} is a @code{const_int} which gives
3797
the offset in bytes into the entire argument of where this part starts.
3798
As a special exception the first @code{expr_list} in the @code{parallel}
3799
RTX may have a first operand of zero.  This indicates that the entire
3800
argument is also stored on the stack.
3801
 
3802
The last time this macro is called, it is called with @code{MODE ==
3803
VOIDmode}, and its result is passed to the @code{call} or @code{call_value}
3804
pattern as operands 2 and 3 respectively.
3805
 
3806
@cindex @file{stdarg.h} and register arguments
3807
The usual way to make the ISO library @file{stdarg.h} work on a machine
3808
where some arguments are usually passed in registers, is to cause
3809
nameless arguments to be passed on the stack instead.  This is done
3810
by making @code{FUNCTION_ARG} return 0 whenever @var{named} is 0.
3811
 
3812
@cindex @code{TARGET_MUST_PASS_IN_STACK}, and @code{FUNCTION_ARG}
3813
@cindex @code{REG_PARM_STACK_SPACE}, and @code{FUNCTION_ARG}
3814
You may use the hook @code{targetm.calls.must_pass_in_stack}
3815
in the definition of this macro to determine if this argument is of a
3816
type that must be passed in the stack.  If @code{REG_PARM_STACK_SPACE}
3817
is not defined and @code{FUNCTION_ARG} returns nonzero for such an
3818
argument, the compiler will abort.  If @code{REG_PARM_STACK_SPACE} is
3819
defined, the argument will be computed in the stack and then loaded into
3820
a register.
3821
@end defmac
3822
 
3823
@deftypefn {Target Hook} bool TARGET_MUST_PASS_IN_STACK (enum machine_mode @var{mode}, tree @var{type})
3824
This target hook should return @code{true} if we should not pass @var{type}
3825
solely in registers.  The file @file{expr.h} defines a
3826
definition that is usually appropriate, refer to @file{expr.h} for additional
3827
documentation.
3828
@end deftypefn
3829
 
3830
@defmac FUNCTION_INCOMING_ARG (@var{cum}, @var{mode}, @var{type}, @var{named})
3831
Define this macro if the target machine has ``register windows'', so
3832
that the register in which a function sees an arguments is not
3833
necessarily the same as the one in which the caller passed the
3834
argument.
3835
 
3836
For such machines, @code{FUNCTION_ARG} computes the register in which
3837
the caller passes the value, and @code{FUNCTION_INCOMING_ARG} should
3838
be defined in a similar fashion to tell the function being called
3839
where the arguments will arrive.
3840
 
3841
If @code{FUNCTION_INCOMING_ARG} is not defined, @code{FUNCTION_ARG}
3842
serves both purposes.
3843
@end defmac
3844
 
3845
@deftypefn {Target Hook} int TARGET_ARG_PARTIAL_BYTES (CUMULATIVE_ARGS *@var{cum}, enum machine_mode @var{mode}, tree @var{type}, bool @var{named})
3846
This target hook returns the number of bytes at the beginning of an
3847
argument that must be put in registers.  The value must be zero for
3848
arguments that are passed entirely in registers or that are entirely
3849
pushed on the stack.
3850
 
3851
On some machines, certain arguments must be passed partially in
3852
registers and partially in memory.  On these machines, typically the
3853
first few words of arguments are passed in registers, and the rest
3854
on the stack.  If a multi-word argument (a @code{double} or a
3855
structure) crosses that boundary, its first few words must be passed
3856
in registers and the rest must be pushed.  This macro tells the
3857
compiler when this occurs, and how many bytes should go in registers.
3858
 
3859
@code{FUNCTION_ARG} for these arguments should return the first
3860
register to be used by the caller for this argument; likewise
3861
@code{FUNCTION_INCOMING_ARG}, for the called function.
3862
@end deftypefn
3863
 
3864
@deftypefn {Target Hook} bool TARGET_PASS_BY_REFERENCE (CUMULATIVE_ARGS *@var{cum}, enum machine_mode @var{mode}, tree @var{type}, bool @var{named})
3865
This target hook should return @code{true} if an argument at the
3866
position indicated by @var{cum} should be passed by reference.  This
3867
predicate is queried after target independent reasons for being
3868
passed by reference, such as @code{TREE_ADDRESSABLE (type)}.
3869
 
3870
If the hook returns true, a copy of that argument is made in memory and a
3871
pointer to the argument is passed instead of the argument itself.
3872
The pointer is passed in whatever way is appropriate for passing a pointer
3873
to that type.
3874
@end deftypefn
3875
 
3876
@deftypefn {Target Hook} bool TARGET_CALLEE_COPIES (CUMULATIVE_ARGS *@var{cum}, enum machine_mode @var{mode}, tree @var{type}, bool @var{named})
3877
The function argument described by the parameters to this hook is
3878
known to be passed by reference.  The hook should return true if the
3879
function argument should be copied by the callee instead of copied
3880
by the caller.
3881
 
3882
For any argument for which the hook returns true, if it can be
3883
determined that the argument is not modified, then a copy need
3884
not be generated.
3885
 
3886
The default version of this hook always returns false.
3887
@end deftypefn
3888
 
3889
@defmac CUMULATIVE_ARGS
3890
A C type for declaring a variable that is used as the first argument of
3891
@code{FUNCTION_ARG} and other related values.  For some target machines,
3892
the type @code{int} suffices and can hold the number of bytes of
3893
argument so far.
3894
 
3895
There is no need to record in @code{CUMULATIVE_ARGS} anything about the
3896
arguments that have been passed on the stack.  The compiler has other
3897
variables to keep track of that.  For target machines on which all
3898
arguments are passed on the stack, there is no need to store anything in
3899
@code{CUMULATIVE_ARGS}; however, the data structure must exist and
3900
should not be empty, so use @code{int}.
3901
@end defmac
3902
 
3903
@defmac INIT_CUMULATIVE_ARGS (@var{cum}, @var{fntype}, @var{libname}, @var{fndecl}, @var{n_named_args})
3904
A C statement (sans semicolon) for initializing the variable
3905
@var{cum} for the state at the beginning of the argument list.  The
3906
variable has type @code{CUMULATIVE_ARGS}.  The value of @var{fntype}
3907
is the tree node for the data type of the function which will receive
3908
the args, or 0 if the args are to a compiler support library function.
3909
For direct calls that are not libcalls, @var{fndecl} contain the
3910
declaration node of the function.  @var{fndecl} is also set when
3911
@code{INIT_CUMULATIVE_ARGS} is used to find arguments for the function
3912
being compiled.  @var{n_named_args} is set to the number of named
3913
arguments, including a structure return address if it is passed as a
3914
parameter, when making a call.  When processing incoming arguments,
3915
@var{n_named_args} is set to @minus{}1.
3916
 
3917
When processing a call to a compiler support library function,
3918
@var{libname} identifies which one.  It is a @code{symbol_ref} rtx which
3919
contains the name of the function, as a string.  @var{libname} is 0 when
3920
an ordinary C function call is being processed.  Thus, each time this
3921
macro is called, either @var{libname} or @var{fntype} is nonzero, but
3922
never both of them at once.
3923
@end defmac
3924
 
3925
@defmac INIT_CUMULATIVE_LIBCALL_ARGS (@var{cum}, @var{mode}, @var{libname})
3926
Like @code{INIT_CUMULATIVE_ARGS} but only used for outgoing libcalls,
3927
it gets a @code{MODE} argument instead of @var{fntype}, that would be
3928
@code{NULL}.  @var{indirect} would always be zero, too.  If this macro
3929
is not defined, @code{INIT_CUMULATIVE_ARGS (cum, NULL_RTX, libname,
3930
0)} is used instead.
3931
@end defmac
3932
 
3933
@defmac INIT_CUMULATIVE_INCOMING_ARGS (@var{cum}, @var{fntype}, @var{libname})
3934
Like @code{INIT_CUMULATIVE_ARGS} but overrides it for the purposes of
3935
finding the arguments for the function being compiled.  If this macro is
3936
undefined, @code{INIT_CUMULATIVE_ARGS} is used instead.
3937
 
3938
The value passed for @var{libname} is always 0, since library routines
3939
with special calling conventions are never compiled with GCC@.  The
3940
argument @var{libname} exists for symmetry with
3941
@code{INIT_CUMULATIVE_ARGS}.
3942
@c could use "this macro" in place of @code{INIT_CUMULATIVE_ARGS}, maybe.
3943
@c --mew 5feb93   i switched the order of the sentences.  --mew 10feb93
3944
@end defmac
3945
 
3946
@defmac FUNCTION_ARG_ADVANCE (@var{cum}, @var{mode}, @var{type}, @var{named})
3947
A C statement (sans semicolon) to update the summarizer variable
3948
@var{cum} to advance past an argument in the argument list.  The
3949
values @var{mode}, @var{type} and @var{named} describe that argument.
3950
Once this is done, the variable @var{cum} is suitable for analyzing
3951
the @emph{following} argument with @code{FUNCTION_ARG}, etc.
3952
 
3953
This macro need not do anything if the argument in question was passed
3954
on the stack.  The compiler knows how to track the amount of stack space
3955
used for arguments without any special help.
3956
@end defmac
3957
 
3958
@defmac FUNCTION_ARG_PADDING (@var{mode}, @var{type})
3959
If defined, a C expression which determines whether, and in which direction,
3960
to pad out an argument with extra space.  The value should be of type
3961
@code{enum direction}: either @code{upward} to pad above the argument,
3962
@code{downward} to pad below, or @code{none} to inhibit padding.
3963
 
3964
The @emph{amount} of padding is always just enough to reach the next
3965
multiple of @code{FUNCTION_ARG_BOUNDARY}; this macro does not control
3966
it.
3967
 
3968
This macro has a default definition which is right for most systems.
3969
For little-endian machines, the default is to pad upward.  For
3970
big-endian machines, the default is to pad downward for an argument of
3971
constant size shorter than an @code{int}, and upward otherwise.
3972
@end defmac
3973
 
3974
@defmac PAD_VARARGS_DOWN
3975
If defined, a C expression which determines whether the default
3976
implementation of va_arg will attempt to pad down before reading the
3977
next argument, if that argument is smaller than its aligned space as
3978
controlled by @code{PARM_BOUNDARY}.  If this macro is not defined, all such
3979
arguments are padded down if @code{BYTES_BIG_ENDIAN} is true.
3980
@end defmac
3981
 
3982
@defmac BLOCK_REG_PADDING (@var{mode}, @var{type}, @var{first})
3983
Specify padding for the last element of a block move between registers and
3984
memory.  @var{first} is nonzero if this is the only element.  Defining this
3985
macro allows better control of register function parameters on big-endian
3986
machines, without using @code{PARALLEL} rtl.  In particular,
3987
@code{MUST_PASS_IN_STACK} need not test padding and mode of types in
3988
registers, as there is no longer a "wrong" part of a register;  For example,
3989
a three byte aggregate may be passed in the high part of a register if so
3990
required.
3991
@end defmac
3992
 
3993
@defmac FUNCTION_ARG_BOUNDARY (@var{mode}, @var{type})
3994
If defined, a C expression that gives the alignment boundary, in bits,
3995
of an argument with the specified mode and type.  If it is not defined,
3996
@code{PARM_BOUNDARY} is used for all arguments.
3997
@end defmac
3998
 
3999
@defmac FUNCTION_ARG_REGNO_P (@var{regno})
4000
A C expression that is nonzero if @var{regno} is the number of a hard
4001
register in which function arguments are sometimes passed.  This does
4002
@emph{not} include implicit arguments such as the static chain and
4003
the structure-value address.  On many machines, no registers can be
4004
used for this purpose since all function arguments are pushed on the
4005
stack.
4006
@end defmac
4007
 
4008
@deftypefn {Target Hook} bool TARGET_SPLIT_COMPLEX_ARG (tree @var{type})
4009
This hook should return true if parameter of type @var{type} are passed
4010
as two scalar parameters.  By default, GCC will attempt to pack complex
4011
arguments into the target's word size.  Some ABIs require complex arguments
4012
to be split and treated as their individual components.  For example, on
4013
AIX64, complex floats should be passed in a pair of floating point
4014
registers, even though a complex float would fit in one 64-bit floating
4015
point register.
4016
 
4017
The default value of this hook is @code{NULL}, which is treated as always
4018
false.
4019
@end deftypefn
4020
 
4021
@deftypefn {Target Hook} tree TARGET_BUILD_BUILTIN_VA_LIST (void)
4022
This hook returns a type node for @code{va_list} for the target.
4023
The default version of the hook returns @code{void*}.
4024
@end deftypefn
4025
 
4026
@deftypefn {Target Hook} tree TARGET_GIMPLIFY_VA_ARG_EXPR (tree @var{valist}, tree @var{type}, tree *@var{pre_p}, tree *@var{post_p})
4027
This hook performs target-specific gimplification of
4028
@code{VA_ARG_EXPR}.  The first two parameters correspond to the
4029
arguments to @code{va_arg}; the latter two are as in
4030
@code{gimplify.c:gimplify_expr}.
4031
@end deftypefn
4032
 
4033
@deftypefn {Target Hook} bool TARGET_VALID_POINTER_MODE (enum machine_mode @var{mode})
4034
Define this to return nonzero if the port can handle pointers
4035
with machine mode @var{mode}.  The default version of this
4036
hook returns true for both @code{ptr_mode} and @code{Pmode}.
4037
@end deftypefn
4038
 
4039
@deftypefn {Target Hook} bool TARGET_SCALAR_MODE_SUPPORTED_P (enum machine_mode @var{mode})
4040
Define this to return nonzero if the port is prepared to handle
4041
insns involving scalar mode @var{mode}.  For a scalar mode to be
4042
considered supported, all the basic arithmetic and comparisons
4043
must work.
4044
 
4045
The default version of this hook returns true for any mode
4046
required to handle the basic C types (as defined by the port).
4047
Included here are the double-word arithmetic supported by the
4048
code in @file{optabs.c}.
4049
@end deftypefn
4050
 
4051
@deftypefn {Target Hook} bool TARGET_VECTOR_MODE_SUPPORTED_P (enum machine_mode @var{mode})
4052
Define this to return nonzero if the port is prepared to handle
4053
insns involving vector mode @var{mode}.  At the very least, it
4054
must have move patterns for this mode.
4055
@end deftypefn
4056
 
4057
@node Scalar Return
4058
@subsection How Scalar Function Values Are Returned
4059
@cindex return values in registers
4060
@cindex values, returned by functions
4061
@cindex scalars, returned as values
4062
 
4063
This section discusses the macros that control returning scalars as
4064
values---values that can fit in registers.
4065
 
4066
@deftypefn {Target Hook} rtx TARGET_FUNCTION_VALUE (tree @var{ret_type}, tree @var{fn_decl_or_type}, bool @var{outgoing})
4067
 
4068
Define this to return an RTX representing the place where a function
4069
returns or receives a value of data type @var{ret_type}, a tree node
4070
node representing a data type.  @var{fn_decl_or_type} is a tree node
4071
representing @code{FUNCTION_DECL} or @code{FUNCTION_TYPE} of a
4072
function being called.  If @var{outgoing} is false, the hook should
4073
compute the register in which the caller will see the return value.
4074
Otherwise, the hook should return an RTX representing the place where
4075
a function returns a value.
4076
 
4077
On many machines, only @code{TYPE_MODE (@var{ret_type})} is relevant.
4078
(Actually, on most machines, scalar values are returned in the same
4079
place regardless of mode.)  The value of the expression is usually a
4080
@code{reg} RTX for the hard register where the return value is stored.
4081
The value can also be a @code{parallel} RTX, if the return value is in
4082
multiple places.  See @code{FUNCTION_ARG} for an explanation of the
4083
@code{parallel} form.
4084
 
4085
If @code{TARGET_PROMOTE_FUNCTION_RETURN} returns true, you must apply
4086
the same promotion rules specified in @code{PROMOTE_MODE} if
4087
@var{valtype} is a scalar type.
4088
 
4089
If the precise function being called is known, @var{func} is a tree
4090
node (@code{FUNCTION_DECL}) for it; otherwise, @var{func} is a null
4091
pointer.  This makes it possible to use a different value-returning
4092
convention for specific functions when all their calls are
4093
known.
4094
 
4095
Some target machines have ``register windows'' so that the register in
4096
which a function returns its value is not the same as the one in which
4097
the caller sees the value.  For such machines, you should return
4098
different RTX depending on @var{outgoing}.
4099
 
4100
@code{TARGET_FUNCTION_VALUE} is not used for return values with
4101
aggregate data types, because these are returned in another way.  See
4102
@code{TARGET_STRUCT_VALUE_RTX} and related macros, below.
4103
@end deftypefn
4104
 
4105
@defmac FUNCTION_VALUE (@var{valtype}, @var{func})
4106
This macro has been deprecated.  Use @code{TARGET_FUNCTION_VALUE} for
4107
a new target instead.
4108
@end defmac
4109
 
4110
@defmac FUNCTION_OUTGOING_VALUE (@var{valtype}, @var{func})
4111
This macro has been deprecated.  Use @code{TARGET_FUNCTION_VALUE} for
4112
a new target instead.
4113
@end defmac
4114
 
4115
@defmac LIBCALL_VALUE (@var{mode})
4116
A C expression to create an RTX representing the place where a library
4117
function returns a value of mode @var{mode}.  If the precise function
4118
being called is known, @var{func} is a tree node
4119
(@code{FUNCTION_DECL}) for it; otherwise, @var{func} is a null
4120
pointer.  This makes it possible to use a different value-returning
4121
convention for specific functions when all their calls are
4122
known.
4123
 
4124
Note that ``library function'' in this context means a compiler
4125
support routine, used to perform arithmetic, whose name is known
4126
specially by the compiler and was not mentioned in the C code being
4127
compiled.
4128
 
4129
The definition of @code{LIBRARY_VALUE} need not be concerned aggregate
4130
data types, because none of the library functions returns such types.
4131
@end defmac
4132
 
4133
@defmac FUNCTION_VALUE_REGNO_P (@var{regno})
4134
A C expression that is nonzero if @var{regno} is the number of a hard
4135
register in which the values of called function may come back.
4136
 
4137
A register whose use for returning values is limited to serving as the
4138
second of a pair (for a value of type @code{double}, say) need not be
4139
recognized by this macro.  So for most machines, this definition
4140
suffices:
4141
 
4142
@smallexample
4143
#define FUNCTION_VALUE_REGNO_P(N) ((N) == 0)
4144
@end smallexample
4145
 
4146
If the machine has register windows, so that the caller and the called
4147
function use different registers for the return value, this macro
4148
should recognize only the caller's register numbers.
4149
@end defmac
4150
 
4151
@defmac APPLY_RESULT_SIZE
4152
Define this macro if @samp{untyped_call} and @samp{untyped_return}
4153
need more space than is implied by @code{FUNCTION_VALUE_REGNO_P} for
4154
saving and restoring an arbitrary return value.
4155
@end defmac
4156
 
4157
@deftypefn {Target Hook} bool TARGET_RETURN_IN_MSB (tree @var{type})
4158
This hook should return true if values of type @var{type} are returned
4159
at the most significant end of a register (in other words, if they are
4160
padded at the least significant end).  You can assume that @var{type}
4161
is returned in a register; the caller is required to check this.
4162
 
4163
Note that the register provided by @code{TARGET_FUNCTION_VALUE} must
4164
be able to hold the complete return value.  For example, if a 1-, 2-
4165
or 3-byte structure is returned at the most significant end of a
4166
4-byte register, @code{TARGET_FUNCTION_VALUE} should provide an
4167
@code{SImode} rtx.
4168
@end deftypefn
4169
 
4170
@node Aggregate Return
4171
@subsection How Large Values Are Returned
4172
@cindex aggregates as return values
4173
@cindex large return values
4174
@cindex returning aggregate values
4175
@cindex structure value address
4176
 
4177
When a function value's mode is @code{BLKmode} (and in some other
4178
cases), the value is not returned according to
4179
@code{TARGET_FUNCTION_VALUE} (@pxref{Scalar Return}).  Instead, the
4180
caller passes the address of a block of memory in which the value
4181
should be stored.  This address is called the @dfn{structure value
4182
address}.
4183
 
4184
This section describes how to control returning structure values in
4185
memory.
4186
 
4187
@deftypefn {Target Hook} bool TARGET_RETURN_IN_MEMORY (tree @var{type}, tree @var{fntype})
4188
This target hook should return a nonzero value to say to return the
4189
function value in memory, just as large structures are always returned.
4190
Here @var{type} will be the data type of the value, and @var{fntype}
4191
will be the type of the function doing the returning, or @code{NULL} for
4192
libcalls.
4193
 
4194
Note that values of mode @code{BLKmode} must be explicitly handled
4195
by this function.  Also, the option @option{-fpcc-struct-return}
4196
takes effect regardless of this macro.  On most systems, it is
4197
possible to leave the hook undefined; this causes a default
4198
definition to be used, whose value is the constant 1 for @code{BLKmode}
4199
values, and 0 otherwise.
4200
 
4201
Do not use this hook to indicate that structures and unions should always
4202
be returned in memory.  You should instead use @code{DEFAULT_PCC_STRUCT_RETURN}
4203
to indicate this.
4204
@end deftypefn
4205
 
4206
@defmac DEFAULT_PCC_STRUCT_RETURN
4207
Define this macro to be 1 if all structure and union return values must be
4208
in memory.  Since this results in slower code, this should be defined
4209
only if needed for compatibility with other compilers or with an ABI@.
4210
If you define this macro to be 0, then the conventions used for structure
4211
and union return values are decided by the @code{TARGET_RETURN_IN_MEMORY}
4212
target hook.
4213
 
4214
If not defined, this defaults to the value 1.
4215
@end defmac
4216
 
4217
@deftypefn {Target Hook} rtx TARGET_STRUCT_VALUE_RTX (tree @var{fndecl}, int @var{incoming})
4218
This target hook should return the location of the structure value
4219
address (normally a @code{mem} or @code{reg}), or 0 if the address is
4220
passed as an ``invisible'' first argument.  Note that @var{fndecl} may
4221
be @code{NULL}, for libcalls.  You do not need to define this target
4222
hook if the address is always passed as an ``invisible'' first
4223
argument.
4224
 
4225
On some architectures the place where the structure value address
4226
is found by the called function is not the same place that the
4227
caller put it.  This can be due to register windows, or it could
4228
be because the function prologue moves it to a different place.
4229
@var{incoming} is @code{1} or @code{2} when the location is needed in
4230
the context of the called function, and @code{0} in the context of
4231
the caller.
4232
 
4233
If @var{incoming} is nonzero and the address is to be found on the
4234
stack, return a @code{mem} which refers to the frame pointer. If
4235
@var{incoming} is @code{2}, the result is being used to fetch the
4236
structure value address at the beginning of a function.  If you need
4237
to emit adjusting code, you should do it at this point.
4238
@end deftypefn
4239
 
4240
@defmac PCC_STATIC_STRUCT_RETURN
4241
Define this macro if the usual system convention on the target machine
4242
for returning structures and unions is for the called function to return
4243
the address of a static variable containing the value.
4244
 
4245
Do not define this if the usual system convention is for the caller to
4246
pass an address to the subroutine.
4247
 
4248
This macro has effect in @option{-fpcc-struct-return} mode, but it does
4249
nothing when you use @option{-freg-struct-return} mode.
4250
@end defmac
4251
 
4252
@node Caller Saves
4253
@subsection Caller-Saves Register Allocation
4254
 
4255
If you enable it, GCC can save registers around function calls.  This
4256
makes it possible to use call-clobbered registers to hold variables that
4257
must live across calls.
4258
 
4259
@defmac CALLER_SAVE_PROFITABLE (@var{refs}, @var{calls})
4260
A C expression to determine whether it is worthwhile to consider placing
4261
a pseudo-register in a call-clobbered hard register and saving and
4262
restoring it around each function call.  The expression should be 1 when
4263
this is worth doing, and 0 otherwise.
4264
 
4265
If you don't define this macro, a default is used which is good on most
4266
machines: @code{4 * @var{calls} < @var{refs}}.
4267
@end defmac
4268
 
4269
@defmac HARD_REGNO_CALLER_SAVE_MODE (@var{regno}, @var{nregs})
4270
A C expression specifying which mode is required for saving @var{nregs}
4271
of a pseudo-register in call-clobbered hard register @var{regno}.  If
4272
@var{regno} is unsuitable for caller save, @code{VOIDmode} should be
4273
returned.  For most machines this macro need not be defined since GCC
4274
will select the smallest suitable mode.
4275
@end defmac
4276
 
4277
@node Function Entry
4278
@subsection Function Entry and Exit
4279
@cindex function entry and exit
4280
@cindex prologue
4281
@cindex epilogue
4282
 
4283
This section describes the macros that output function entry
4284
(@dfn{prologue}) and exit (@dfn{epilogue}) code.
4285
 
4286
@deftypefn {Target Hook} void TARGET_ASM_FUNCTION_PROLOGUE (FILE *@var{file}, HOST_WIDE_INT @var{size})
4287
If defined, a function that outputs the assembler code for entry to a
4288
function.  The prologue is responsible for setting up the stack frame,
4289
initializing the frame pointer register, saving registers that must be
4290
saved, and allocating @var{size} additional bytes of storage for the
4291
local variables.  @var{size} is an integer.  @var{file} is a stdio
4292
stream to which the assembler code should be output.
4293
 
4294
The label for the beginning of the function need not be output by this
4295
macro.  That has already been done when the macro is run.
4296
 
4297
@findex regs_ever_live
4298
To determine which registers to save, the macro can refer to the array
4299
@code{regs_ever_live}: element @var{r} is nonzero if hard register
4300
@var{r} is used anywhere within the function.  This implies the function
4301
prologue should save register @var{r}, provided it is not one of the
4302
call-used registers.  (@code{TARGET_ASM_FUNCTION_EPILOGUE} must likewise use
4303
@code{regs_ever_live}.)
4304
 
4305
On machines that have ``register windows'', the function entry code does
4306
not save on the stack the registers that are in the windows, even if
4307
they are supposed to be preserved by function calls; instead it takes
4308
appropriate steps to ``push'' the register stack, if any non-call-used
4309
registers are used in the function.
4310
 
4311
@findex frame_pointer_needed
4312
On machines where functions may or may not have frame-pointers, the
4313
function entry code must vary accordingly; it must set up the frame
4314
pointer if one is wanted, and not otherwise.  To determine whether a
4315
frame pointer is in wanted, the macro can refer to the variable
4316
@code{frame_pointer_needed}.  The variable's value will be 1 at run
4317
time in a function that needs a frame pointer.  @xref{Elimination}.
4318
 
4319
The function entry code is responsible for allocating any stack space
4320
required for the function.  This stack space consists of the regions
4321
listed below.  In most cases, these regions are allocated in the
4322
order listed, with the last listed region closest to the top of the
4323
stack (the lowest address if @code{STACK_GROWS_DOWNWARD} is defined, and
4324
the highest address if it is not defined).  You can use a different order
4325
for a machine if doing so is more convenient or required for
4326
compatibility reasons.  Except in cases where required by standard
4327
or by a debugger, there is no reason why the stack layout used by GCC
4328
need agree with that used by other compilers for a machine.
4329
@end deftypefn
4330
 
4331
@deftypefn {Target Hook} void TARGET_ASM_FUNCTION_END_PROLOGUE (FILE *@var{file})
4332
If defined, a function that outputs assembler code at the end of a
4333
prologue.  This should be used when the function prologue is being
4334
emitted as RTL, and you have some extra assembler that needs to be
4335
emitted.  @xref{prologue instruction pattern}.
4336
@end deftypefn
4337
 
4338
@deftypefn {Target Hook} void TARGET_ASM_FUNCTION_BEGIN_EPILOGUE (FILE *@var{file})
4339
If defined, a function that outputs assembler code at the start of an
4340
epilogue.  This should be used when the function epilogue is being
4341
emitted as RTL, and you have some extra assembler that needs to be
4342
emitted.  @xref{epilogue instruction pattern}.
4343
@end deftypefn
4344
 
4345
@deftypefn {Target Hook} void TARGET_ASM_FUNCTION_EPILOGUE (FILE *@var{file}, HOST_WIDE_INT @var{size})
4346
If defined, a function that outputs the assembler code for exit from a
4347
function.  The epilogue is responsible for restoring the saved
4348
registers and stack pointer to their values when the function was
4349
called, and returning control to the caller.  This macro takes the
4350
same arguments as the macro @code{TARGET_ASM_FUNCTION_PROLOGUE}, and the
4351
registers to restore are determined from @code{regs_ever_live} and
4352
@code{CALL_USED_REGISTERS} in the same way.
4353
 
4354
On some machines, there is a single instruction that does all the work
4355
of returning from the function.  On these machines, give that
4356
instruction the name @samp{return} and do not define the macro
4357
@code{TARGET_ASM_FUNCTION_EPILOGUE} at all.
4358
 
4359
Do not define a pattern named @samp{return} if you want the
4360
@code{TARGET_ASM_FUNCTION_EPILOGUE} to be used.  If you want the target
4361
switches to control whether return instructions or epilogues are used,
4362
define a @samp{return} pattern with a validity condition that tests the
4363
target switches appropriately.  If the @samp{return} pattern's validity
4364
condition is false, epilogues will be used.
4365
 
4366
On machines where functions may or may not have frame-pointers, the
4367
function exit code must vary accordingly.  Sometimes the code for these
4368
two cases is completely different.  To determine whether a frame pointer
4369
is wanted, the macro can refer to the variable
4370
@code{frame_pointer_needed}.  The variable's value will be 1 when compiling
4371
a function that needs a frame pointer.
4372
 
4373
Normally, @code{TARGET_ASM_FUNCTION_PROLOGUE} and
4374
@code{TARGET_ASM_FUNCTION_EPILOGUE} must treat leaf functions specially.
4375
The C variable @code{current_function_is_leaf} is nonzero for such a
4376
function.  @xref{Leaf Functions}.
4377
 
4378
On some machines, some functions pop their arguments on exit while
4379
others leave that for the caller to do.  For example, the 68020 when
4380
given @option{-mrtd} pops arguments in functions that take a fixed
4381
number of arguments.
4382
 
4383
@findex current_function_pops_args
4384
Your definition of the macro @code{RETURN_POPS_ARGS} decides which
4385
functions pop their own arguments.  @code{TARGET_ASM_FUNCTION_EPILOGUE}
4386
needs to know what was decided.  The variable that is called
4387
@code{current_function_pops_args} is the number of bytes of its
4388
arguments that a function should pop.  @xref{Scalar Return}.
4389
@c what is the "its arguments" in the above sentence referring to, pray
4390
@c tell?  --mew 5feb93
4391
@end deftypefn
4392
 
4393
@itemize @bullet
4394
@item
4395
@findex current_function_pretend_args_size
4396
A region of @code{current_function_pretend_args_size} bytes of
4397
uninitialized space just underneath the first argument arriving on the
4398
stack.  (This may not be at the very start of the allocated stack region
4399
if the calling sequence has pushed anything else since pushing the stack
4400
arguments.  But usually, on such machines, nothing else has been pushed
4401
yet, because the function prologue itself does all the pushing.)  This
4402
region is used on machines where an argument may be passed partly in
4403
registers and partly in memory, and, in some cases to support the
4404
features in @code{<stdarg.h>}.
4405
 
4406
@item
4407
An area of memory used to save certain registers used by the function.
4408
The size of this area, which may also include space for such things as
4409
the return address and pointers to previous stack frames, is
4410
machine-specific and usually depends on which registers have been used
4411
in the function.  Machines with register windows often do not require
4412
a save area.
4413
 
4414
@item
4415
A region of at least @var{size} bytes, possibly rounded up to an allocation
4416
boundary, to contain the local variables of the function.  On some machines,
4417
this region and the save area may occur in the opposite order, with the
4418
save area closer to the top of the stack.
4419
 
4420
@item
4421
@cindex @code{ACCUMULATE_OUTGOING_ARGS} and stack frames
4422
Optionally, when @code{ACCUMULATE_OUTGOING_ARGS} is defined, a region of
4423
@code{current_function_outgoing_args_size} bytes to be used for outgoing
4424
argument lists of the function.  @xref{Stack Arguments}.
4425
@end itemize
4426
 
4427
@defmac EXIT_IGNORE_STACK
4428
Define this macro as a C expression that is nonzero if the return
4429
instruction or the function epilogue ignores the value of the stack
4430
pointer; in other words, if it is safe to delete an instruction to
4431
adjust the stack pointer before a return from the function.  The
4432
default is 0.
4433
 
4434
Note that this macro's value is relevant only for functions for which
4435
frame pointers are maintained.  It is never safe to delete a final
4436
stack adjustment in a function that has no frame pointer, and the
4437
compiler knows this regardless of @code{EXIT_IGNORE_STACK}.
4438
@end defmac
4439
 
4440
@defmac EPILOGUE_USES (@var{regno})
4441
Define this macro as a C expression that is nonzero for registers that are
4442
used by the epilogue or the @samp{return} pattern.  The stack and frame
4443
pointer registers are already assumed to be used as needed.
4444
@end defmac
4445
 
4446
@defmac EH_USES (@var{regno})
4447
Define this macro as a C expression that is nonzero for registers that are
4448
used by the exception handling mechanism, and so should be considered live
4449
on entry to an exception edge.
4450
@end defmac
4451
 
4452
@defmac DELAY_SLOTS_FOR_EPILOGUE
4453
Define this macro if the function epilogue contains delay slots to which
4454
instructions from the rest of the function can be ``moved''.  The
4455
definition should be a C expression whose value is an integer
4456
representing the number of delay slots there.
4457
@end defmac
4458
 
4459
@defmac ELIGIBLE_FOR_EPILOGUE_DELAY (@var{insn}, @var{n})
4460
A C expression that returns 1 if @var{insn} can be placed in delay
4461
slot number @var{n} of the epilogue.
4462
 
4463
The argument @var{n} is an integer which identifies the delay slot now
4464
being considered (since different slots may have different rules of
4465
eligibility).  It is never negative and is always less than the number
4466
of epilogue delay slots (what @code{DELAY_SLOTS_FOR_EPILOGUE} returns).
4467
If you reject a particular insn for a given delay slot, in principle, it
4468
may be reconsidered for a subsequent delay slot.  Also, other insns may
4469
(at least in principle) be considered for the so far unfilled delay
4470
slot.
4471
 
4472
@findex current_function_epilogue_delay_list
4473
@findex final_scan_insn
4474
The insns accepted to fill the epilogue delay slots are put in an RTL
4475
list made with @code{insn_list} objects, stored in the variable
4476
@code{current_function_epilogue_delay_list}.  The insn for the first
4477
delay slot comes first in the list.  Your definition of the macro
4478
@code{TARGET_ASM_FUNCTION_EPILOGUE} should fill the delay slots by
4479
outputting the insns in this list, usually by calling
4480
@code{final_scan_insn}.
4481
 
4482
You need not define this macro if you did not define
4483
@code{DELAY_SLOTS_FOR_EPILOGUE}.
4484
@end defmac
4485
 
4486
@deftypefn {Target Hook} void TARGET_ASM_OUTPUT_MI_THUNK (FILE *@var{file}, tree @var{thunk_fndecl}, HOST_WIDE_INT @var{delta}, HOST_WIDE_INT @var{vcall_offset}, tree @var{function})
4487
A function that outputs the assembler code for a thunk
4488
function, used to implement C++ virtual function calls with multiple
4489
inheritance.  The thunk acts as a wrapper around a virtual function,
4490
adjusting the implicit object parameter before handing control off to
4491
the real function.
4492
 
4493
First, emit code to add the integer @var{delta} to the location that
4494
contains the incoming first argument.  Assume that this argument
4495
contains a pointer, and is the one used to pass the @code{this} pointer
4496
in C++.  This is the incoming argument @emph{before} the function prologue,
4497
e.g.@: @samp{%o0} on a sparc.  The addition must preserve the values of
4498
all other incoming arguments.
4499
 
4500
Then, if @var{vcall_offset} is nonzero, an additional adjustment should be
4501
made after adding @code{delta}.  In particular, if @var{p} is the
4502
adjusted pointer, the following adjustment should be made:
4503
 
4504
@smallexample
4505
p += (*((ptrdiff_t **)p))[vcall_offset/sizeof(ptrdiff_t)]
4506
@end smallexample
4507
 
4508
After the additions, emit code to jump to @var{function}, which is a
4509
@code{FUNCTION_DECL}.  This is a direct pure jump, not a call, and does
4510
not touch the return address.  Hence returning from @var{FUNCTION} will
4511
return to whoever called the current @samp{thunk}.
4512
 
4513
The effect must be as if @var{function} had been called directly with
4514
the adjusted first argument.  This macro is responsible for emitting all
4515
of the code for a thunk function; @code{TARGET_ASM_FUNCTION_PROLOGUE}
4516
and @code{TARGET_ASM_FUNCTION_EPILOGUE} are not invoked.
4517
 
4518
The @var{thunk_fndecl} is redundant.  (@var{delta} and @var{function}
4519
have already been extracted from it.)  It might possibly be useful on
4520
some targets, but probably not.
4521
 
4522
If you do not define this macro, the target-independent code in the C++
4523
front end will generate a less efficient heavyweight thunk that calls
4524
@var{function} instead of jumping to it.  The generic approach does
4525
not support varargs.
4526
@end deftypefn
4527
 
4528
@deftypefn {Target Hook} bool TARGET_ASM_CAN_OUTPUT_MI_THUNK (tree @var{thunk_fndecl}, HOST_WIDE_INT @var{delta}, HOST_WIDE_INT @var{vcall_offset}, tree @var{function})
4529
A function that returns true if TARGET_ASM_OUTPUT_MI_THUNK would be able
4530
to output the assembler code for the thunk function specified by the
4531
arguments it is passed, and false otherwise.  In the latter case, the
4532
generic approach will be used by the C++ front end, with the limitations
4533
previously exposed.
4534
@end deftypefn
4535
 
4536
@node Profiling
4537
@subsection Generating Code for Profiling
4538
@cindex profiling, code generation
4539
 
4540
These macros will help you generate code for profiling.
4541
 
4542
@defmac FUNCTION_PROFILER (@var{file}, @var{labelno})
4543
A C statement or compound statement to output to @var{file} some
4544
assembler code to call the profiling subroutine @code{mcount}.
4545
 
4546
@findex mcount
4547
The details of how @code{mcount} expects to be called are determined by
4548
your operating system environment, not by GCC@.  To figure them out,
4549
compile a small program for profiling using the system's installed C
4550
compiler and look at the assembler code that results.
4551
 
4552
Older implementations of @code{mcount} expect the address of a counter
4553
variable to be loaded into some register.  The name of this variable is
4554
@samp{LP} followed by the number @var{labelno}, so you would generate
4555
the name using @samp{LP%d} in a @code{fprintf}.
4556
@end defmac
4557
 
4558
@defmac PROFILE_HOOK
4559
A C statement or compound statement to output to @var{file} some assembly
4560
code to call the profiling subroutine @code{mcount} even the target does
4561
not support profiling.
4562
@end defmac
4563
 
4564
@defmac NO_PROFILE_COUNTERS
4565
Define this macro to be an expression with a nonzero value if the
4566
@code{mcount} subroutine on your system does not need a counter variable
4567
allocated for each function.  This is true for almost all modern
4568
implementations.  If you define this macro, you must not use the
4569
@var{labelno} argument to @code{FUNCTION_PROFILER}.
4570
@end defmac
4571
 
4572
@defmac PROFILE_BEFORE_PROLOGUE
4573
Define this macro if the code for function profiling should come before
4574
the function prologue.  Normally, the profiling code comes after.
4575
@end defmac
4576
 
4577
@node Tail Calls
4578
@subsection Permitting tail calls
4579
@cindex tail calls
4580
 
4581
@deftypefn {Target Hook} bool TARGET_FUNCTION_OK_FOR_SIBCALL (tree @var{decl}, tree @var{exp})
4582
True if it is ok to do sibling call optimization for the specified
4583
call expression @var{exp}.  @var{decl} will be the called function,
4584
or @code{NULL} if this is an indirect call.
4585
 
4586
It is not uncommon for limitations of calling conventions to prevent
4587
tail calls to functions outside the current unit of translation, or
4588
during PIC compilation.  The hook is used to enforce these restrictions,
4589
as the @code{sibcall} md pattern can not fail, or fall over to a
4590
``normal'' call.  The criteria for successful sibling call optimization
4591
may vary greatly between different architectures.
4592
@end deftypefn
4593
 
4594
@deftypefn {Target Hook} void TARGET_EXTRA_LIVE_ON_ENTRY (bitmap *@var{regs})
4595
Add any hard registers to @var{regs} that are live on entry to the
4596
function.  This hook only needs to be defined to provide registers that
4597
cannot be found by examination of FUNCTION_ARG_REGNO_P, the callee saved
4598
registers, STATIC_CHAIN_INCOMING_REGNUM, STATIC_CHAIN_REGNUM,
4599
TARGET_STRUCT_VALUE_RTX, FRAME_POINTER_REGNUM, EH_USES,
4600
FRAME_POINTER_REGNUM, ARG_POINTER_REGNUM, and the PIC_OFFSET_TABLE_REGNUM.
4601
@end deftypefn
4602
 
4603
@node Stack Smashing Protection
4604
@subsection Stack smashing protection
4605
@cindex stack smashing protection
4606
 
4607
@deftypefn {Target Hook} tree TARGET_STACK_PROTECT_GUARD (void)
4608
This hook returns a @code{DECL} node for the external variable to use
4609
for the stack protection guard.  This variable is initialized by the
4610
runtime to some random value and is used to initialize the guard value
4611
that is placed at the top of the local stack frame.  The type of this
4612
variable must be @code{ptr_type_node}.
4613
 
4614
The default version of this hook creates a variable called
4615
@samp{__stack_chk_guard}, which is normally defined in @file{libgcc2.c}.
4616
@end deftypefn
4617
 
4618
@deftypefn {Target Hook} tree TARGET_STACK_PROTECT_FAIL (void)
4619
This hook returns a tree expression that alerts the runtime that the
4620
stack protect guard variable has been modified.  This expression should
4621
involve a call to a @code{noreturn} function.
4622
 
4623
The default version of this hook invokes a function called
4624
@samp{__stack_chk_fail}, taking no arguments.  This function is
4625
normally defined in @file{libgcc2.c}.
4626
@end deftypefn
4627
 
4628
@node Varargs
4629
@section Implementing the Varargs Macros
4630
@cindex varargs implementation
4631
 
4632
GCC comes with an implementation of @code{<varargs.h>} and
4633
@code{<stdarg.h>} that work without change on machines that pass arguments
4634
on the stack.  Other machines require their own implementations of
4635
varargs, and the two machine independent header files must have
4636
conditionals to include it.
4637
 
4638
ISO @code{<stdarg.h>} differs from traditional @code{<varargs.h>} mainly in
4639
the calling convention for @code{va_start}.  The traditional
4640
implementation takes just one argument, which is the variable in which
4641
to store the argument pointer.  The ISO implementation of
4642
@code{va_start} takes an additional second argument.  The user is
4643
supposed to write the last named argument of the function here.
4644
 
4645
However, @code{va_start} should not use this argument.  The way to find
4646
the end of the named arguments is with the built-in functions described
4647
below.
4648
 
4649
@defmac __builtin_saveregs ()
4650
Use this built-in function to save the argument registers in memory so
4651
that the varargs mechanism can access them.  Both ISO and traditional
4652
versions of @code{va_start} must use @code{__builtin_saveregs}, unless
4653
you use @code{TARGET_SETUP_INCOMING_VARARGS} (see below) instead.
4654
 
4655
On some machines, @code{__builtin_saveregs} is open-coded under the
4656
control of the target hook @code{TARGET_EXPAND_BUILTIN_SAVEREGS}.  On
4657
other machines, it calls a routine written in assembler language,
4658
found in @file{libgcc2.c}.
4659
 
4660
Code generated for the call to @code{__builtin_saveregs} appears at the
4661
beginning of the function, as opposed to where the call to
4662
@code{__builtin_saveregs} is written, regardless of what the code is.
4663
This is because the registers must be saved before the function starts
4664
to use them for its own purposes.
4665
@c i rewrote the first sentence above to fix an overfull hbox. --mew
4666
@c 10feb93
4667
@end defmac
4668
 
4669
@defmac __builtin_args_info (@var{category})
4670
Use this built-in function to find the first anonymous arguments in
4671
registers.
4672
 
4673
In general, a machine may have several categories of registers used for
4674
arguments, each for a particular category of data types.  (For example,
4675
on some machines, floating-point registers are used for floating-point
4676
arguments while other arguments are passed in the general registers.)
4677
To make non-varargs functions use the proper calling convention, you
4678
have defined the @code{CUMULATIVE_ARGS} data type to record how many
4679
registers in each category have been used so far
4680
 
4681
@code{__builtin_args_info} accesses the same data structure of type
4682
@code{CUMULATIVE_ARGS} after the ordinary argument layout is finished
4683
with it, with @var{category} specifying which word to access.  Thus, the
4684
value indicates the first unused register in a given category.
4685
 
4686
Normally, you would use @code{__builtin_args_info} in the implementation
4687
of @code{va_start}, accessing each category just once and storing the
4688
value in the @code{va_list} object.  This is because @code{va_list} will
4689
have to update the values, and there is no way to alter the
4690
values accessed by @code{__builtin_args_info}.
4691
@end defmac
4692
 
4693
@defmac __builtin_next_arg (@var{lastarg})
4694
This is the equivalent of @code{__builtin_args_info}, for stack
4695
arguments.  It returns the address of the first anonymous stack
4696
argument, as type @code{void *}.  If @code{ARGS_GROW_DOWNWARD}, it
4697
returns the address of the location above the first anonymous stack
4698
argument.  Use it in @code{va_start} to initialize the pointer for
4699
fetching arguments from the stack.  Also use it in @code{va_start} to
4700
verify that the second parameter @var{lastarg} is the last named argument
4701
of the current function.
4702
@end defmac
4703
 
4704
@defmac __builtin_classify_type (@var{object})
4705
Since each machine has its own conventions for which data types are
4706
passed in which kind of register, your implementation of @code{va_arg}
4707
has to embody these conventions.  The easiest way to categorize the
4708
specified data type is to use @code{__builtin_classify_type} together
4709
with @code{sizeof} and @code{__alignof__}.
4710
 
4711
@code{__builtin_classify_type} ignores the value of @var{object},
4712
considering only its data type.  It returns an integer describing what
4713
kind of type that is---integer, floating, pointer, structure, and so on.
4714
 
4715
The file @file{typeclass.h} defines an enumeration that you can use to
4716
interpret the values of @code{__builtin_classify_type}.
4717
@end defmac
4718
 
4719
These machine description macros help implement varargs:
4720
 
4721
@deftypefn {Target Hook} rtx TARGET_EXPAND_BUILTIN_SAVEREGS (void)
4722
If defined, this hook produces the machine-specific code for a call to
4723
@code{__builtin_saveregs}.  This code will be moved to the very
4724
beginning of the function, before any parameter access are made.  The
4725
return value of this function should be an RTX that contains the value
4726
to use as the return of @code{__builtin_saveregs}.
4727
@end deftypefn
4728
 
4729
@deftypefn {Target Hook} void TARGET_SETUP_INCOMING_VARARGS (CUMULATIVE_ARGS *@var{args_so_far}, enum machine_mode @var{mode}, tree @var{type}, int *@var{pretend_args_size}, int @var{second_time})
4730
This target hook offers an alternative to using
4731
@code{__builtin_saveregs} and defining the hook
4732
@code{TARGET_EXPAND_BUILTIN_SAVEREGS}.  Use it to store the anonymous
4733
register arguments into the stack so that all the arguments appear to
4734
have been passed consecutively on the stack.  Once this is done, you can
4735
use the standard implementation of varargs that works for machines that
4736
pass all their arguments on the stack.
4737
 
4738
The argument @var{args_so_far} points to the @code{CUMULATIVE_ARGS} data
4739
structure, containing the values that are obtained after processing the
4740
named arguments.  The arguments @var{mode} and @var{type} describe the
4741
last named argument---its machine mode and its data type as a tree node.
4742
 
4743
The target hook should do two things: first, push onto the stack all the
4744
argument registers @emph{not} used for the named arguments, and second,
4745
store the size of the data thus pushed into the @code{int}-valued
4746
variable pointed to by @var{pretend_args_size}.  The value that you
4747
store here will serve as additional offset for setting up the stack
4748
frame.
4749
 
4750
Because you must generate code to push the anonymous arguments at
4751
compile time without knowing their data types,
4752
@code{TARGET_SETUP_INCOMING_VARARGS} is only useful on machines that
4753
have just a single category of argument register and use it uniformly
4754
for all data types.
4755
 
4756
If the argument @var{second_time} is nonzero, it means that the
4757
arguments of the function are being analyzed for the second time.  This
4758
happens for an inline function, which is not actually compiled until the
4759
end of the source file.  The hook @code{TARGET_SETUP_INCOMING_VARARGS} should
4760
not generate any instructions in this case.
4761
@end deftypefn
4762
 
4763
@deftypefn {Target Hook} bool TARGET_STRICT_ARGUMENT_NAMING (CUMULATIVE_ARGS *@var{ca})
4764
Define this hook to return @code{true} if the location where a function
4765
argument is passed depends on whether or not it is a named argument.
4766
 
4767
This hook controls how the @var{named} argument to @code{FUNCTION_ARG}
4768
is set for varargs and stdarg functions.  If this hook returns
4769
@code{true}, the @var{named} argument is always true for named
4770
arguments, and false for unnamed arguments.  If it returns @code{false},
4771
but @code{TARGET_PRETEND_OUTGOING_VARARGS_NAMED} returns @code{true},
4772
then all arguments are treated as named.  Otherwise, all named arguments
4773
except the last are treated as named.
4774
 
4775
You need not define this hook if it always returns zero.
4776
@end deftypefn
4777
 
4778
@deftypefn {Target Hook} bool TARGET_PRETEND_OUTGOING_VARARGS_NAMED
4779
If you need to conditionally change ABIs so that one works with
4780
@code{TARGET_SETUP_INCOMING_VARARGS}, but the other works like neither
4781
@code{TARGET_SETUP_INCOMING_VARARGS} nor @code{TARGET_STRICT_ARGUMENT_NAMING} was
4782
defined, then define this hook to return @code{true} if
4783
@code{TARGET_SETUP_INCOMING_VARARGS} is used, @code{false} otherwise.
4784
Otherwise, you should not define this hook.
4785
@end deftypefn
4786
 
4787
@node Trampolines
4788
@section Trampolines for Nested Functions
4789
@cindex trampolines for nested functions
4790
@cindex nested functions, trampolines for
4791
 
4792
A @dfn{trampoline} is a small piece of code that is created at run time
4793
when the address of a nested function is taken.  It normally resides on
4794
the stack, in the stack frame of the containing function.  These macros
4795
tell GCC how to generate code to allocate and initialize a
4796
trampoline.
4797
 
4798
The instructions in the trampoline must do two things: load a constant
4799
address into the static chain register, and jump to the real address of
4800
the nested function.  On CISC machines such as the m68k, this requires
4801
two instructions, a move immediate and a jump.  Then the two addresses
4802
exist in the trampoline as word-long immediate operands.  On RISC
4803
machines, it is often necessary to load each address into a register in
4804
two parts.  Then pieces of each address form separate immediate
4805
operands.
4806
 
4807
The code generated to initialize the trampoline must store the variable
4808
parts---the static chain value and the function address---into the
4809
immediate operands of the instructions.  On a CISC machine, this is
4810
simply a matter of copying each address to a memory reference at the
4811
proper offset from the start of the trampoline.  On a RISC machine, it
4812
may be necessary to take out pieces of the address and store them
4813
separately.
4814
 
4815
@defmac TRAMPOLINE_TEMPLATE (@var{file})
4816
A C statement to output, on the stream @var{file}, assembler code for a
4817
block of data that contains the constant parts of a trampoline.  This
4818
code should not include a label---the label is taken care of
4819
automatically.
4820
 
4821
If you do not define this macro, it means no template is needed
4822
for the target.  Do not define this macro on systems where the block move
4823
code to copy the trampoline into place would be larger than the code
4824
to generate it on the spot.
4825
@end defmac
4826
 
4827
@defmac TRAMPOLINE_SECTION
4828
Return the section into which the trampoline template is to be placed
4829
(@pxref{Sections}).  The default value is @code{readonly_data_section}.
4830
@end defmac
4831
 
4832
@defmac TRAMPOLINE_SIZE
4833
A C expression for the size in bytes of the trampoline, as an integer.
4834
@end defmac
4835
 
4836
@defmac TRAMPOLINE_ALIGNMENT
4837
Alignment required for trampolines, in bits.
4838
 
4839
If you don't define this macro, the value of @code{BIGGEST_ALIGNMENT}
4840
is used for aligning trampolines.
4841
@end defmac
4842
 
4843
@defmac INITIALIZE_TRAMPOLINE (@var{addr}, @var{fnaddr}, @var{static_chain})
4844
A C statement to initialize the variable parts of a trampoline.
4845
@var{addr} is an RTX for the address of the trampoline; @var{fnaddr} is
4846
an RTX for the address of the nested function; @var{static_chain} is an
4847
RTX for the static chain value that should be passed to the function
4848
when it is called.
4849
@end defmac
4850
 
4851
@defmac TRAMPOLINE_ADJUST_ADDRESS (@var{addr})
4852
A C statement that should perform any machine-specific adjustment in
4853
the address of the trampoline.  Its argument contains the address that
4854
was passed to @code{INITIALIZE_TRAMPOLINE}.  In case the address to be
4855
used for a function call should be different from the address in which
4856
the template was stored, the different address should be assigned to
4857
@var{addr}.  If this macro is not defined, @var{addr} will be used for
4858
function calls.
4859
 
4860
@cindex @code{TARGET_ASM_FUNCTION_EPILOGUE} and trampolines
4861
@cindex @code{TARGET_ASM_FUNCTION_PROLOGUE} and trampolines
4862
If this macro is not defined, by default the trampoline is allocated as
4863
a stack slot.  This default is right for most machines.  The exceptions
4864
are machines where it is impossible to execute instructions in the stack
4865
area.  On such machines, you may have to implement a separate stack,
4866
using this macro in conjunction with @code{TARGET_ASM_FUNCTION_PROLOGUE}
4867
and @code{TARGET_ASM_FUNCTION_EPILOGUE}.
4868
 
4869
@var{fp} points to a data structure, a @code{struct function}, which
4870
describes the compilation status of the immediate containing function of
4871
the function which the trampoline is for.  The stack slot for the
4872
trampoline is in the stack frame of this containing function.  Other
4873
allocation strategies probably must do something analogous with this
4874
information.
4875
@end defmac
4876
 
4877
Implementing trampolines is difficult on many machines because they have
4878
separate instruction and data caches.  Writing into a stack location
4879
fails to clear the memory in the instruction cache, so when the program
4880
jumps to that location, it executes the old contents.
4881
 
4882
Here are two possible solutions.  One is to clear the relevant parts of
4883
the instruction cache whenever a trampoline is set up.  The other is to
4884
make all trampolines identical, by having them jump to a standard
4885
subroutine.  The former technique makes trampoline execution faster; the
4886
latter makes initialization faster.
4887
 
4888
To clear the instruction cache when a trampoline is initialized, define
4889
the following macro.
4890
 
4891
@defmac CLEAR_INSN_CACHE (@var{beg}, @var{end})
4892
If defined, expands to a C expression clearing the @emph{instruction
4893
cache} in the specified interval.  The definition of this macro would
4894
typically be a series of @code{asm} statements.  Both @var{beg} and
4895
@var{end} are both pointer expressions.
4896
@end defmac
4897
 
4898
The operating system may also require the stack to be made executable
4899
before calling the trampoline.  To implement this requirement, define
4900
the following macro.
4901
 
4902
@defmac ENABLE_EXECUTE_STACK
4903
Define this macro if certain operations must be performed before executing
4904
code located on the stack.  The macro should expand to a series of C
4905
file-scope constructs (e.g.@: functions) and provide a unique entry point
4906
named @code{__enable_execute_stack}.  The target is responsible for
4907
emitting calls to the entry point in the code, for example from the
4908
@code{INITIALIZE_TRAMPOLINE} macro.
4909
@end defmac
4910
 
4911
To use a standard subroutine, define the following macro.  In addition,
4912
you must make sure that the instructions in a trampoline fill an entire
4913
cache line with identical instructions, or else ensure that the
4914
beginning of the trampoline code is always aligned at the same point in
4915
its cache line.  Look in @file{m68k.h} as a guide.
4916
 
4917
@defmac TRANSFER_FROM_TRAMPOLINE
4918
Define this macro if trampolines need a special subroutine to do their
4919
work.  The macro should expand to a series of @code{asm} statements
4920
which will be compiled with GCC@.  They go in a library function named
4921
@code{__transfer_from_trampoline}.
4922
 
4923
If you need to avoid executing the ordinary prologue code of a compiled
4924
C function when you jump to the subroutine, you can do so by placing a
4925
special label of your own in the assembler code.  Use one @code{asm}
4926
statement to generate an assembler label, and another to make the label
4927
global.  Then trampolines can use that label to jump directly to your
4928
special assembler code.
4929
@end defmac
4930
 
4931
@node Library Calls
4932
@section Implicit Calls to Library Routines
4933
@cindex library subroutine names
4934
@cindex @file{libgcc.a}
4935
 
4936
@c prevent bad page break with this line
4937
Here is an explanation of implicit calls to library routines.
4938
 
4939
@defmac DECLARE_LIBRARY_RENAMES
4940
This macro, if defined, should expand to a piece of C code that will get
4941
expanded when compiling functions for libgcc.a.  It can be used to
4942
provide alternate names for GCC's internal library functions if there
4943
are ABI-mandated names that the compiler should provide.
4944
@end defmac
4945
 
4946
@findex init_one_libfunc
4947
@findex set_optab_libfunc
4948
@deftypefn {Target Hook} void TARGET_INIT_LIBFUNCS (void)
4949
This hook should declare additional library routines or rename
4950
existing ones, using the functions @code{set_optab_libfunc} and
4951
@code{init_one_libfunc} defined in @file{optabs.c}.
4952
@code{init_optabs} calls this macro after initializing all the normal
4953
library routines.
4954
 
4955
The default is to do nothing.  Most ports don't need to define this hook.
4956
@end deftypefn
4957
 
4958
@defmac FLOAT_LIB_COMPARE_RETURNS_BOOL (@var{mode}, @var{comparison})
4959
This macro should return @code{true} if the library routine that
4960
implements the floating point comparison operator @var{comparison} in
4961
mode @var{mode} will return a boolean, and @var{false} if it will
4962
return a tristate.
4963
 
4964
GCC's own floating point libraries return tristates from the
4965
comparison operators, so the default returns false always.  Most ports
4966
don't need to define this macro.
4967
@end defmac
4968
 
4969
@defmac TARGET_LIB_INT_CMP_BIASED
4970
This macro should evaluate to @code{true} if the integer comparison
4971
functions (like @code{__cmpdi2}) return 0 to indicate that the first
4972
operand is smaller than the second, 1 to indicate that they are equal,
4973
and 2 to indicate that the first operand is greater than the second.
4974
If this macro evaluates to @code{false} the comparison functions return
4975
@minus{}1, 0, and 1 instead of 0, 1, and 2.  If the target uses the routines
4976
in @file{libgcc.a}, you do not need to define this macro.
4977
@end defmac
4978
 
4979
@cindex US Software GOFAST, floating point emulation library
4980
@cindex floating point emulation library, US Software GOFAST
4981
@cindex GOFAST, floating point emulation library
4982
@findex gofast_maybe_init_libfuncs
4983
@defmac US_SOFTWARE_GOFAST
4984
Define this macro if your system C library uses the US Software GOFAST
4985
library to provide floating point emulation.
4986
 
4987
In addition to defining this macro, your architecture must set
4988
@code{TARGET_INIT_LIBFUNCS} to @code{gofast_maybe_init_libfuncs}, or
4989
else call that function from its version of that hook.  It is defined
4990
in @file{config/gofast.h}, which must be included by your
4991
architecture's @file{@var{cpu}.c} file.  See @file{sparc/sparc.c} for
4992
an example.
4993
 
4994
If this macro is defined, the
4995
@code{TARGET_FLOAT_LIB_COMPARE_RETURNS_BOOL} target hook must return
4996
false for @code{SFmode} and @code{DFmode} comparisons.
4997
@end defmac
4998
 
4999
@cindex @code{EDOM}, implicit usage
5000
@findex matherr
5001
@defmac TARGET_EDOM
5002
The value of @code{EDOM} on the target machine, as a C integer constant
5003
expression.  If you don't define this macro, GCC does not attempt to
5004
deposit the value of @code{EDOM} into @code{errno} directly.  Look in
5005
@file{/usr/include/errno.h} to find the value of @code{EDOM} on your
5006
system.
5007
 
5008
If you do not define @code{TARGET_EDOM}, then compiled code reports
5009
domain errors by calling the library function and letting it report the
5010
error.  If mathematical functions on your system use @code{matherr} when
5011
there is an error, then you should leave @code{TARGET_EDOM} undefined so
5012
that @code{matherr} is used normally.
5013
@end defmac
5014
 
5015
@cindex @code{errno}, implicit usage
5016
@defmac GEN_ERRNO_RTX
5017
Define this macro as a C expression to create an rtl expression that
5018
refers to the global ``variable'' @code{errno}.  (On certain systems,
5019
@code{errno} may not actually be a variable.)  If you don't define this
5020
macro, a reasonable default is used.
5021
@end defmac
5022
 
5023
@cindex C99 math functions, implicit usage
5024
@defmac TARGET_C99_FUNCTIONS
5025
When this macro is nonzero, GCC will implicitly optimize @code{sin} calls into
5026
@code{sinf} and similarly for other functions defined by C99 standard.  The
5027
default is nonzero that should be proper value for most modern systems, however
5028
number of existing systems lacks support for these functions in the runtime so
5029
they needs this macro to be redefined to 0.
5030
@end defmac
5031
 
5032
@defmac NEXT_OBJC_RUNTIME
5033
Define this macro to generate code for Objective-C message sending using
5034
the calling convention of the NeXT system.  This calling convention
5035
involves passing the object, the selector and the method arguments all
5036
at once to the method-lookup library function.
5037
 
5038
The default calling convention passes just the object and the selector
5039
to the lookup function, which returns a pointer to the method.
5040
@end defmac
5041
 
5042
@node Addressing Modes
5043
@section Addressing Modes
5044
@cindex addressing modes
5045
 
5046
@c prevent bad page break with this line
5047
This is about addressing modes.
5048
 
5049
@defmac HAVE_PRE_INCREMENT
5050
@defmacx HAVE_PRE_DECREMENT
5051
@defmacx HAVE_POST_INCREMENT
5052
@defmacx HAVE_POST_DECREMENT
5053
A C expression that is nonzero if the machine supports pre-increment,
5054
pre-decrement, post-increment, or post-decrement addressing respectively.
5055
@end defmac
5056
 
5057
@defmac HAVE_PRE_MODIFY_DISP
5058
@defmacx HAVE_POST_MODIFY_DISP
5059
A C expression that is nonzero if the machine supports pre- or
5060
post-address side-effect generation involving constants other than
5061
the size of the memory operand.
5062
@end defmac
5063
 
5064
@defmac HAVE_PRE_MODIFY_REG
5065
@defmacx HAVE_POST_MODIFY_REG
5066
A C expression that is nonzero if the machine supports pre- or
5067
post-address side-effect generation involving a register displacement.
5068
@end defmac
5069
 
5070
@defmac CONSTANT_ADDRESS_P (@var{x})
5071
A C expression that is 1 if the RTX @var{x} is a constant which
5072
is a valid address.  On most machines, this can be defined as
5073
@code{CONSTANT_P (@var{x})}, but a few machines are more restrictive
5074
in which constant addresses are supported.
5075
@end defmac
5076
 
5077
@defmac CONSTANT_P (@var{x})
5078
@code{CONSTANT_P}, which is defined by target-independent code,
5079
accepts integer-values expressions whose values are not explicitly
5080
known, such as @code{symbol_ref}, @code{label_ref}, and @code{high}
5081
expressions and @code{const} arithmetic expressions, in addition to
5082
@code{const_int} and @code{const_double} expressions.
5083
@end defmac
5084
 
5085
@defmac MAX_REGS_PER_ADDRESS
5086
A number, the maximum number of registers that can appear in a valid
5087
memory address.  Note that it is up to you to specify a value equal to
5088
the maximum number that @code{GO_IF_LEGITIMATE_ADDRESS} would ever
5089
accept.
5090
@end defmac
5091
 
5092
@defmac GO_IF_LEGITIMATE_ADDRESS (@var{mode}, @var{x}, @var{label})
5093
A C compound statement with a conditional @code{goto @var{label};}
5094
executed if @var{x} (an RTX) is a legitimate memory address on the
5095
target machine for a memory operand of mode @var{mode}.
5096
 
5097
It usually pays to define several simpler macros to serve as
5098
subroutines for this one.  Otherwise it may be too complicated to
5099
understand.
5100
 
5101
This macro must exist in two variants: a strict variant and a
5102
non-strict one.  The strict variant is used in the reload pass.  It
5103
must be defined so that any pseudo-register that has not been
5104
allocated a hard register is considered a memory reference.  In
5105
contexts where some kind of register is required, a pseudo-register
5106
with no hard register must be rejected.
5107
 
5108
The non-strict variant is used in other passes.  It must be defined to
5109
accept all pseudo-registers in every context where some kind of
5110
register is required.
5111
 
5112
@findex REG_OK_STRICT
5113
Compiler source files that want to use the strict variant of this
5114
macro define the macro @code{REG_OK_STRICT}.  You should use an
5115
@code{#ifdef REG_OK_STRICT} conditional to define the strict variant
5116
in that case and the non-strict variant otherwise.
5117
 
5118
Subroutines to check for acceptable registers for various purposes (one
5119
for base registers, one for index registers, and so on) are typically
5120
among the subroutines used to define @code{GO_IF_LEGITIMATE_ADDRESS}.
5121
Then only these subroutine macros need have two variants; the higher
5122
levels of macros may be the same whether strict or not.
5123
 
5124
Normally, constant addresses which are the sum of a @code{symbol_ref}
5125
and an integer are stored inside a @code{const} RTX to mark them as
5126
constant.  Therefore, there is no need to recognize such sums
5127
specifically as legitimate addresses.  Normally you would simply
5128
recognize any @code{const} as legitimate.
5129
 
5130
Usually @code{PRINT_OPERAND_ADDRESS} is not prepared to handle constant
5131
sums that are not marked with  @code{const}.  It assumes that a naked
5132
@code{plus} indicates indexing.  If so, then you @emph{must} reject such
5133
naked constant sums as illegitimate addresses, so that none of them will
5134
be given to @code{PRINT_OPERAND_ADDRESS}.
5135
 
5136
@cindex @code{TARGET_ENCODE_SECTION_INFO} and address validation
5137
On some machines, whether a symbolic address is legitimate depends on
5138
the section that the address refers to.  On these machines, define the
5139
target hook @code{TARGET_ENCODE_SECTION_INFO} to store the information
5140
into the @code{symbol_ref}, and then check for it here.  When you see a
5141
@code{const}, you will have to look inside it to find the
5142
@code{symbol_ref} in order to determine the section.  @xref{Assembler
5143
Format}.
5144
@end defmac
5145
 
5146
@defmac FIND_BASE_TERM (@var{x})
5147
A C expression to determine the base term of address @var{x}.
5148
This macro is used in only one place: `find_base_term' in alias.c.
5149
 
5150
It is always safe for this macro to not be defined.  It exists so
5151
that alias analysis can understand machine-dependent addresses.
5152
 
5153
The typical use of this macro is to handle addresses containing
5154
a label_ref or symbol_ref within an UNSPEC@.
5155
@end defmac
5156
 
5157
@defmac LEGITIMIZE_ADDRESS (@var{x}, @var{oldx}, @var{mode}, @var{win})
5158
A C compound statement that attempts to replace @var{x} with a valid
5159
memory address for an operand of mode @var{mode}.  @var{win} will be a
5160
C statement label elsewhere in the code; the macro definition may use
5161
 
5162
@smallexample
5163
GO_IF_LEGITIMATE_ADDRESS (@var{mode}, @var{x}, @var{win});
5164
@end smallexample
5165
 
5166
@noindent
5167
to avoid further processing if the address has become legitimate.
5168
 
5169
@findex break_out_memory_refs
5170
@var{x} will always be the result of a call to @code{break_out_memory_refs},
5171
and @var{oldx} will be the operand that was given to that function to produce
5172
@var{x}.
5173
 
5174
The code generated by this macro should not alter the substructure of
5175
@var{x}.  If it transforms @var{x} into a more legitimate form, it
5176
should assign @var{x} (which will always be a C variable) a new value.
5177
 
5178
It is not necessary for this macro to come up with a legitimate
5179
address.  The compiler has standard ways of doing so in all cases.  In
5180
fact, it is safe to omit this macro.  But often a
5181
machine-dependent strategy can generate better code.
5182
@end defmac
5183
 
5184
@defmac LEGITIMIZE_RELOAD_ADDRESS (@var{x}, @var{mode}, @var{opnum}, @var{type}, @var{ind_levels}, @var{win})
5185
A C compound statement that attempts to replace @var{x}, which is an address
5186
that needs reloading, with a valid memory address for an operand of mode
5187
@var{mode}.  @var{win} will be a C statement label elsewhere in the code.
5188
It is not necessary to define this macro, but it might be useful for
5189
performance reasons.
5190
 
5191
For example, on the i386, it is sometimes possible to use a single
5192
reload register instead of two by reloading a sum of two pseudo
5193
registers into a register.  On the other hand, for number of RISC
5194
processors offsets are limited so that often an intermediate address
5195
needs to be generated in order to address a stack slot.  By defining
5196
@code{LEGITIMIZE_RELOAD_ADDRESS} appropriately, the intermediate addresses
5197
generated for adjacent some stack slots can be made identical, and thus
5198
be shared.
5199
 
5200
@emph{Note}: This macro should be used with caution.  It is necessary
5201
to know something of how reload works in order to effectively use this,
5202
and it is quite easy to produce macros that build in too much knowledge
5203
of reload internals.
5204
 
5205
@emph{Note}: This macro must be able to reload an address created by a
5206
previous invocation of this macro.  If it fails to handle such addresses
5207
then the compiler may generate incorrect code or abort.
5208
 
5209
@findex push_reload
5210
The macro definition should use @code{push_reload} to indicate parts that
5211
need reloading; @var{opnum}, @var{type} and @var{ind_levels} are usually
5212
suitable to be passed unaltered to @code{push_reload}.
5213
 
5214
The code generated by this macro must not alter the substructure of
5215
@var{x}.  If it transforms @var{x} into a more legitimate form, it
5216
should assign @var{x} (which will always be a C variable) a new value.
5217
This also applies to parts that you change indirectly by calling
5218
@code{push_reload}.
5219
 
5220
@findex strict_memory_address_p
5221
The macro definition may use @code{strict_memory_address_p} to test if
5222
the address has become legitimate.
5223
 
5224
@findex copy_rtx
5225
If you want to change only a part of @var{x}, one standard way of doing
5226
this is to use @code{copy_rtx}.  Note, however, that is unshares only a
5227
single level of rtl.  Thus, if the part to be changed is not at the
5228
top level, you'll need to replace first the top level.
5229
It is not necessary for this macro to come up with a legitimate
5230
address;  but often a machine-dependent strategy can generate better code.
5231
@end defmac
5232
 
5233
@defmac GO_IF_MODE_DEPENDENT_ADDRESS (@var{addr}, @var{label})
5234
A C statement or compound statement with a conditional @code{goto
5235
@var{label};} executed if memory address @var{x} (an RTX) can have
5236
different meanings depending on the machine mode of the memory
5237
reference it is used for or if the address is valid for some modes
5238
but not others.
5239
 
5240
Autoincrement and autodecrement addresses typically have mode-dependent
5241
effects because the amount of the increment or decrement is the size
5242
of the operand being addressed.  Some machines have other mode-dependent
5243
addresses.  Many RISC machines have no mode-dependent addresses.
5244
 
5245
You may assume that @var{addr} is a valid address for the machine.
5246
@end defmac
5247
 
5248
@defmac LEGITIMATE_CONSTANT_P (@var{x})
5249
A C expression that is nonzero if @var{x} is a legitimate constant for
5250
an immediate operand on the target machine.  You can assume that
5251
@var{x} satisfies @code{CONSTANT_P}, so you need not check this.  In fact,
5252
@samp{1} is a suitable definition for this macro on machines where
5253
anything @code{CONSTANT_P} is valid.
5254
@end defmac
5255
 
5256
@deftypefn {Target Hook} rtx TARGET_DELEGITIMIZE_ADDRESS (rtx @var{x})
5257
This hook is used to undo the possibly obfuscating effects of the
5258
@code{LEGITIMIZE_ADDRESS} and @code{LEGITIMIZE_RELOAD_ADDRESS} target
5259
macros.  Some backend implementations of these macros wrap symbol
5260
references inside an @code{UNSPEC} rtx to represent PIC or similar
5261
addressing modes.  This target hook allows GCC's optimizers to understand
5262
the semantics of these opaque @code{UNSPEC}s by converting them back
5263
into their original form.
5264
@end deftypefn
5265
 
5266
@deftypefn {Target Hook} bool TARGET_CANNOT_FORCE_CONST_MEM (rtx @var{x})
5267
This hook should return true if @var{x} is of a form that cannot (or
5268
should not) be spilled to the constant pool.  The default version of
5269
this hook returns false.
5270
 
5271
The primary reason to define this hook is to prevent reload from
5272
deciding that a non-legitimate constant would be better reloaded
5273
from the constant pool instead of spilling and reloading a register
5274
holding the constant.  This restriction is often true of addresses
5275
of TLS symbols for various targets.
5276
@end deftypefn
5277
 
5278
@deftypefn {Target Hook} bool TARGET_USE_BLOCKS_FOR_CONSTANT_P (enum machine_mode @var{mode}, rtx @var{x})
5279
This hook should return true if pool entries for constant @var{x} can
5280
be placed in an @code{object_block} structure.  @var{mode} is the mode
5281
of @var{x}.
5282
 
5283
The default version returns false for all constants.
5284
@end deftypefn
5285
 
5286
@deftypefn {Target Hook} tree TARGET_VECTORIZE_BUILTIN_MASK_FOR_LOAD (void)
5287
This hook should return the DECL of a function @var{f} that given an
5288
address @var{addr} as an argument returns a mask @var{m} that can be
5289
used to extract from two vectors the relevant data that resides in
5290
@var{addr} in case @var{addr} is not properly aligned.
5291
 
5292
The autovectrizer, when vectorizing a load operation from an address
5293
@var{addr} that may be unaligned, will generate two vector loads from
5294
the two aligned addresses around @var{addr}. It then generates a
5295
@code{REALIGN_LOAD} operation to extract the relevant data from the
5296
two loaded vectors. The first two arguments to @code{REALIGN_LOAD},
5297
@var{v1} and @var{v2}, are the two vectors, each of size @var{VS}, and
5298
the third argument, @var{OFF}, defines how the data will be extracted
5299
from these two vectors: if @var{OFF} is 0, then the returned vector is
5300
@var{v2}; otherwise, the returned vector is composed from the last
5301
@var{VS}-@var{OFF} elements of @var{v1} concatenated to the first
5302
@var{OFF} elements of @var{v2}.
5303
 
5304
If this hook is defined, the autovectorizer will generate a call
5305
to @var{f} (using the DECL tree that this hook returns) and will
5306
use the return value of @var{f} as the argument @var{OFF} to
5307
@code{REALIGN_LOAD}. Therefore, the mask @var{m} returned by @var{f}
5308
should comply with the semantics expected by @code{REALIGN_LOAD}
5309
described above.
5310
If this hook is not defined, then @var{addr} will be used as
5311
the argument @var{OFF} to @code{REALIGN_LOAD}, in which case the low
5312
log2(@var{VS})-1 bits of @var{addr} will be considered.
5313
@end deftypefn
5314
 
5315
@node Anchored Addresses
5316
@section Anchored Addresses
5317
@cindex anchored addresses
5318
@cindex @option{-fsection-anchors}
5319
 
5320
GCC usually addresses every static object as a separate entity.
5321
For example, if we have:
5322
 
5323
@smallexample
5324
static int a, b, c;
5325
int foo (void) @{ return a + b + c; @}
5326
@end smallexample
5327
 
5328
the code for @code{foo} will usually calculate three separate symbolic
5329
addresses: those of @code{a}, @code{b} and @code{c}.  On some targets,
5330
it would be better to calculate just one symbolic address and access
5331
the three variables relative to it.  The equivalent pseudocode would
5332
be something like:
5333
 
5334
@smallexample
5335
int foo (void)
5336
@{
5337
  register int *xr = &x;
5338
  return xr[&a - &x] + xr[&b - &x] + xr[&c - &x];
5339
@}
5340
@end smallexample
5341
 
5342
(which isn't valid C).  We refer to shared addresses like @code{x} as
5343
``section anchors''.  Their use is controlled by @option{-fsection-anchors}.
5344
 
5345
The hooks below describe the target properties that GCC needs to know
5346
in order to make effective use of section anchors.  It won't use
5347
section anchors at all unless either @code{TARGET_MIN_ANCHOR_OFFSET}
5348
or @code{TARGET_MAX_ANCHOR_OFFSET} is set to a nonzero value.
5349
 
5350
@deftypevar {Target Hook} HOST_WIDE_INT TARGET_MIN_ANCHOR_OFFSET
5351
The minimum offset that should be applied to a section anchor.
5352
On most targets, it should be the smallest offset that can be
5353
applied to a base register while still giving a legitimate address
5354
for every mode.  The default value is 0.
5355
@end deftypevar
5356
 
5357
@deftypevar {Target Hook} HOST_WIDE_INT TARGET_MAX_ANCHOR_OFFSET
5358
Like @code{TARGET_MIN_ANCHOR_OFFSET}, but the maximum (inclusive)
5359
offset that should be applied to section anchors.  The default
5360
value is 0.
5361
@end deftypevar
5362
 
5363
@deftypefn {Target Hook} void TARGET_ASM_OUTPUT_ANCHOR (rtx @var{x})
5364
Write the assembly code to define section anchor @var{x}, which is a
5365
@code{SYMBOL_REF} for which @samp{SYMBOL_REF_ANCHOR_P (@var{x})} is true.
5366
The hook is called with the assembly output position set to the beginning
5367
of @code{SYMBOL_REF_BLOCK (@var{x})}.
5368
 
5369
If @code{ASM_OUTPUT_DEF} is available, the hook's default definition uses
5370
it to define the symbol as @samp{. + SYMBOL_REF_BLOCK_OFFSET (@var{x})}.
5371
If @code{ASM_OUTPUT_DEF} is not available, the hook's default definition
5372
is @code{NULL}, which disables the use of section anchors altogether.
5373
@end deftypefn
5374
 
5375
@deftypefn {Target Hook} bool TARGET_USE_ANCHORS_FOR_SYMBOL_P (rtx @var{x})
5376
Return true if GCC should attempt to use anchors to access @code{SYMBOL_REF}
5377
@var{x}.  You can assume @samp{SYMBOL_REF_HAS_BLOCK_INFO_P (@var{x})} and
5378
@samp{!SYMBOL_REF_ANCHOR_P (@var{x})}.
5379
 
5380
The default version is correct for most targets, but you might need to
5381
intercept this hook to handle things like target-specific attributes
5382
or target-specific sections.
5383
@end deftypefn
5384
 
5385
@node Condition Code
5386
@section Condition Code Status
5387
@cindex condition code status
5388
 
5389
@c prevent bad page break with this line
5390
This describes the condition code status.
5391
 
5392
@findex cc_status
5393
The file @file{conditions.h} defines a variable @code{cc_status} to
5394
describe how the condition code was computed (in case the interpretation of
5395
the condition code depends on the instruction that it was set by).  This
5396
variable contains the RTL expressions on which the condition code is
5397
currently based, and several standard flags.
5398
 
5399
Sometimes additional machine-specific flags must be defined in the machine
5400
description header file.  It can also add additional machine-specific
5401
information by defining @code{CC_STATUS_MDEP}.
5402
 
5403
@defmac CC_STATUS_MDEP
5404
C code for a data type which is used for declaring the @code{mdep}
5405
component of @code{cc_status}.  It defaults to @code{int}.
5406
 
5407
This macro is not used on machines that do not use @code{cc0}.
5408
@end defmac
5409
 
5410
@defmac CC_STATUS_MDEP_INIT
5411
A C expression to initialize the @code{mdep} field to ``empty''.
5412
The default definition does nothing, since most machines don't use
5413
the field anyway.  If you want to use the field, you should probably
5414
define this macro to initialize it.
5415
 
5416
This macro is not used on machines that do not use @code{cc0}.
5417
@end defmac
5418
 
5419
@defmac NOTICE_UPDATE_CC (@var{exp}, @var{insn})
5420
A C compound statement to set the components of @code{cc_status}
5421
appropriately for an insn @var{insn} whose body is @var{exp}.  It is
5422
this macro's responsibility to recognize insns that set the condition
5423
code as a byproduct of other activity as well as those that explicitly
5424
set @code{(cc0)}.
5425
 
5426
This macro is not used on machines that do not use @code{cc0}.
5427
 
5428
If there are insns that do not set the condition code but do alter
5429
other machine registers, this macro must check to see whether they
5430
invalidate the expressions that the condition code is recorded as
5431
reflecting.  For example, on the 68000, insns that store in address
5432
registers do not set the condition code, which means that usually
5433
@code{NOTICE_UPDATE_CC} can leave @code{cc_status} unaltered for such
5434
insns.  But suppose that the previous insn set the condition code
5435
based on location @samp{a4@@(102)} and the current insn stores a new
5436
value in @samp{a4}.  Although the condition code is not changed by
5437
this, it will no longer be true that it reflects the contents of
5438
@samp{a4@@(102)}.  Therefore, @code{NOTICE_UPDATE_CC} must alter
5439
@code{cc_status} in this case to say that nothing is known about the
5440
condition code value.
5441
 
5442
The definition of @code{NOTICE_UPDATE_CC} must be prepared to deal
5443
with the results of peephole optimization: insns whose patterns are
5444
@code{parallel} RTXs containing various @code{reg}, @code{mem} or
5445
constants which are just the operands.  The RTL structure of these
5446
insns is not sufficient to indicate what the insns actually do.  What
5447
@code{NOTICE_UPDATE_CC} should do when it sees one is just to run
5448
@code{CC_STATUS_INIT}.
5449
 
5450
A possible definition of @code{NOTICE_UPDATE_CC} is to call a function
5451
that looks at an attribute (@pxref{Insn Attributes}) named, for example,
5452
@samp{cc}.  This avoids having detailed information about patterns in
5453
two places, the @file{md} file and in @code{NOTICE_UPDATE_CC}.
5454
@end defmac
5455
 
5456
@defmac SELECT_CC_MODE (@var{op}, @var{x}, @var{y})
5457
Returns a mode from class @code{MODE_CC} to be used when comparison
5458
operation code @var{op} is applied to rtx @var{x} and @var{y}.  For
5459
example, on the SPARC, @code{SELECT_CC_MODE} is defined as (see
5460
@pxref{Jump Patterns} for a description of the reason for this
5461
definition)
5462
 
5463
@smallexample
5464
#define SELECT_CC_MODE(OP,X,Y) \
5465
  (GET_MODE_CLASS (GET_MODE (X)) == MODE_FLOAT          \
5466
   ? ((OP == EQ || OP == NE) ? CCFPmode : CCFPEmode)    \
5467
   : ((GET_CODE (X) == PLUS || GET_CODE (X) == MINUS    \
5468
       || GET_CODE (X) == NEG) \
5469
      ? CC_NOOVmode : CCmode))
5470
@end smallexample
5471
 
5472
You should define this macro if and only if you define extra CC modes
5473
in @file{@var{machine}-modes.def}.
5474
@end defmac
5475
 
5476
@defmac CANONICALIZE_COMPARISON (@var{code}, @var{op0}, @var{op1})
5477
On some machines not all possible comparisons are defined, but you can
5478
convert an invalid comparison into a valid one.  For example, the Alpha
5479
does not have a @code{GT} comparison, but you can use an @code{LT}
5480
comparison instead and swap the order of the operands.
5481
 
5482
On such machines, define this macro to be a C statement to do any
5483
required conversions.  @var{code} is the initial comparison code
5484
and @var{op0} and @var{op1} are the left and right operands of the
5485
comparison, respectively.  You should modify @var{code}, @var{op0}, and
5486
@var{op1} as required.
5487
 
5488
GCC will not assume that the comparison resulting from this macro is
5489
valid but will see if the resulting insn matches a pattern in the
5490
@file{md} file.
5491
 
5492
You need not define this macro if it would never change the comparison
5493
code or operands.
5494
@end defmac
5495
 
5496
@defmac REVERSIBLE_CC_MODE (@var{mode})
5497
A C expression whose value is one if it is always safe to reverse a
5498
comparison whose mode is @var{mode}.  If @code{SELECT_CC_MODE}
5499
can ever return @var{mode} for a floating-point inequality comparison,
5500
then @code{REVERSIBLE_CC_MODE (@var{mode})} must be zero.
5501
 
5502
You need not define this macro if it would always returns zero or if the
5503
floating-point format is anything other than @code{IEEE_FLOAT_FORMAT}.
5504
For example, here is the definition used on the SPARC, where floating-point
5505
inequality comparisons are always given @code{CCFPEmode}:
5506
 
5507
@smallexample
5508
#define REVERSIBLE_CC_MODE(MODE)  ((MODE) != CCFPEmode)
5509
@end smallexample
5510
@end defmac
5511
 
5512
@defmac REVERSE_CONDITION (@var{code}, @var{mode})
5513
A C expression whose value is reversed condition code of the @var{code} for
5514
comparison done in CC_MODE @var{mode}.  The macro is used only in case
5515
@code{REVERSIBLE_CC_MODE (@var{mode})} is nonzero.  Define this macro in case
5516
machine has some non-standard way how to reverse certain conditionals.  For
5517
instance in case all floating point conditions are non-trapping, compiler may
5518
freely convert unordered compares to ordered one.  Then definition may look
5519
like:
5520
 
5521
@smallexample
5522
#define REVERSE_CONDITION(CODE, MODE) \
5523
   ((MODE) != CCFPmode ? reverse_condition (CODE) \
5524
    : reverse_condition_maybe_unordered (CODE))
5525
@end smallexample
5526
@end defmac
5527
 
5528
@defmac REVERSE_CONDEXEC_PREDICATES_P (@var{op1}, @var{op2})
5529
A C expression that returns true if the conditional execution predicate
5530
@var{op1}, a comparison operation, is the inverse of @var{op2} and vice
5531
versa.  Define this to return 0 if the target has conditional execution
5532
predicates that cannot be reversed safely.  There is no need to validate
5533
that the arguments of op1 and op2 are the same, this is done separately.
5534
If no expansion is specified, this macro is defined as follows:
5535
 
5536
@smallexample
5537
#define REVERSE_CONDEXEC_PREDICATES_P (x, y) \
5538
   (GET_CODE ((x)) == reversed_comparison_code ((y), NULL))
5539
@end smallexample
5540
@end defmac
5541
 
5542
@deftypefn {Target Hook} bool TARGET_FIXED_CONDITION_CODE_REGS (unsigned int *, unsigned int *)
5543
On targets which do not use @code{(cc0)}, and which use a hard
5544
register rather than a pseudo-register to hold condition codes, the
5545
regular CSE passes are often not able to identify cases in which the
5546
hard register is set to a common value.  Use this hook to enable a
5547
small pass which optimizes such cases.  This hook should return true
5548
to enable this pass, and it should set the integers to which its
5549
arguments point to the hard register numbers used for condition codes.
5550
When there is only one such register, as is true on most systems, the
5551
integer pointed to by the second argument should be set to
5552
@code{INVALID_REGNUM}.
5553
 
5554
The default version of this hook returns false.
5555
@end deftypefn
5556
 
5557
@deftypefn {Target Hook} enum machine_mode TARGET_CC_MODES_COMPATIBLE (enum machine_mode, enum machine_mode)
5558
On targets which use multiple condition code modes in class
5559
@code{MODE_CC}, it is sometimes the case that a comparison can be
5560
validly done in more than one mode.  On such a system, define this
5561
target hook to take two mode arguments and to return a mode in which
5562
both comparisons may be validly done.  If there is no such mode,
5563
return @code{VOIDmode}.
5564
 
5565
The default version of this hook checks whether the modes are the
5566
same.  If they are, it returns that mode.  If they are different, it
5567
returns @code{VOIDmode}.
5568
@end deftypefn
5569
 
5570
@node Costs
5571
@section Describing Relative Costs of Operations
5572
@cindex costs of instructions
5573
@cindex relative costs
5574
@cindex speed of instructions
5575
 
5576
These macros let you describe the relative speed of various operations
5577
on the target machine.
5578
 
5579
@defmac REGISTER_MOVE_COST (@var{mode}, @var{from}, @var{to})
5580
A C expression for the cost of moving data of mode @var{mode} from a
5581
register in class @var{from} to one in class @var{to}.  The classes are
5582
expressed using the enumeration values such as @code{GENERAL_REGS}.  A
5583
value of 2 is the default; other values are interpreted relative to
5584
that.
5585
 
5586
It is not required that the cost always equal 2 when @var{from} is the
5587
same as @var{to}; on some machines it is expensive to move between
5588
registers if they are not general registers.
5589
 
5590
If reload sees an insn consisting of a single @code{set} between two
5591
hard registers, and if @code{REGISTER_MOVE_COST} applied to their
5592
classes returns a value of 2, reload does not check to ensure that the
5593
constraints of the insn are met.  Setting a cost of other than 2 will
5594
allow reload to verify that the constraints are met.  You should do this
5595
if the @samp{mov@var{m}} pattern's constraints do not allow such copying.
5596
@end defmac
5597
 
5598
@defmac MEMORY_MOVE_COST (@var{mode}, @var{class}, @var{in})
5599
A C expression for the cost of moving data of mode @var{mode} between a
5600
register of class @var{class} and memory; @var{in} is zero if the value
5601
is to be written to memory, nonzero if it is to be read in.  This cost
5602
is relative to those in @code{REGISTER_MOVE_COST}.  If moving between
5603
registers and memory is more expensive than between two registers, you
5604
should define this macro to express the relative cost.
5605
 
5606
If you do not define this macro, GCC uses a default cost of 4 plus
5607
the cost of copying via a secondary reload register, if one is
5608
needed.  If your machine requires a secondary reload register to copy
5609
between memory and a register of @var{class} but the reload mechanism is
5610
more complex than copying via an intermediate, define this macro to
5611
reflect the actual cost of the move.
5612
 
5613
GCC defines the function @code{memory_move_secondary_cost} if
5614
secondary reloads are needed.  It computes the costs due to copying via
5615
a secondary register.  If your machine copies from memory using a
5616
secondary register in the conventional way but the default base value of
5617
4 is not correct for your machine, define this macro to add some other
5618
value to the result of that function.  The arguments to that function
5619
are the same as to this macro.
5620
@end defmac
5621
 
5622
@defmac BRANCH_COST
5623
A C expression for the cost of a branch instruction.  A value of 1 is
5624
the default; other values are interpreted relative to that.
5625
@end defmac
5626
 
5627
Here are additional macros which do not specify precise relative costs,
5628
but only that certain actions are more expensive than GCC would
5629
ordinarily expect.
5630
 
5631
@defmac SLOW_BYTE_ACCESS
5632
Define this macro as a C expression which is nonzero if accessing less
5633
than a word of memory (i.e.@: a @code{char} or a @code{short}) is no
5634
faster than accessing a word of memory, i.e., if such access
5635
require more than one instruction or if there is no difference in cost
5636
between byte and (aligned) word loads.
5637
 
5638
When this macro is not defined, the compiler will access a field by
5639
finding the smallest containing object; when it is defined, a fullword
5640
load will be used if alignment permits.  Unless bytes accesses are
5641
faster than word accesses, using word accesses is preferable since it
5642
may eliminate subsequent memory access if subsequent accesses occur to
5643
other fields in the same word of the structure, but to different bytes.
5644
@end defmac
5645
 
5646
@defmac SLOW_UNALIGNED_ACCESS (@var{mode}, @var{alignment})
5647
Define this macro to be the value 1 if memory accesses described by the
5648
@var{mode} and @var{alignment} parameters have a cost many times greater
5649
than aligned accesses, for example if they are emulated in a trap
5650
handler.
5651
 
5652
When this macro is nonzero, the compiler will act as if
5653
@code{STRICT_ALIGNMENT} were nonzero when generating code for block
5654
moves.  This can cause significantly more instructions to be produced.
5655
Therefore, do not set this macro nonzero if unaligned accesses only add a
5656
cycle or two to the time for a memory access.
5657
 
5658
If the value of this macro is always zero, it need not be defined.  If
5659
this macro is defined, it should produce a nonzero value when
5660
@code{STRICT_ALIGNMENT} is nonzero.
5661
@end defmac
5662
 
5663
@defmac MOVE_RATIO
5664
The threshold of number of scalar memory-to-memory move insns, @emph{below}
5665
which a sequence of insns should be generated instead of a
5666
string move insn or a library call.  Increasing the value will always
5667
make code faster, but eventually incurs high cost in increased code size.
5668
 
5669
Note that on machines where the corresponding move insn is a
5670
@code{define_expand} that emits a sequence of insns, this macro counts
5671
the number of such sequences.
5672
 
5673
If you don't define this, a reasonable default is used.
5674
@end defmac
5675
 
5676
@defmac MOVE_BY_PIECES_P (@var{size}, @var{alignment})
5677
A C expression used to determine whether @code{move_by_pieces} will be used to
5678
copy a chunk of memory, or whether some other block move mechanism
5679
will be used.  Defaults to 1 if @code{move_by_pieces_ninsns} returns less
5680
than @code{MOVE_RATIO}.
5681
@end defmac
5682
 
5683
@defmac MOVE_MAX_PIECES
5684
A C expression used by @code{move_by_pieces} to determine the largest unit
5685
a load or store used to copy memory is.  Defaults to @code{MOVE_MAX}.
5686
@end defmac
5687
 
5688
@defmac CLEAR_RATIO
5689
The threshold of number of scalar move insns, @emph{below} which a sequence
5690
of insns should be generated to clear memory instead of a string clear insn
5691
or a library call.  Increasing the value will always make code faster, but
5692
eventually incurs high cost in increased code size.
5693
 
5694
If you don't define this, a reasonable default is used.
5695
@end defmac
5696
 
5697
@defmac CLEAR_BY_PIECES_P (@var{size}, @var{alignment})
5698
A C expression used to determine whether @code{clear_by_pieces} will be used
5699
to clear a chunk of memory, or whether some other block clear mechanism
5700
will be used.  Defaults to 1 if @code{move_by_pieces_ninsns} returns less
5701
than @code{CLEAR_RATIO}.
5702
@end defmac
5703
 
5704
@defmac STORE_BY_PIECES_P (@var{size}, @var{alignment})
5705
A C expression used to determine whether @code{store_by_pieces} will be
5706
used to set a chunk of memory to a constant value, or whether some other
5707
mechanism will be used.  Used by @code{__builtin_memset} when storing
5708
values other than constant zero and by @code{__builtin_strcpy} when
5709
when called with a constant source string.
5710
Defaults to 1 if @code{move_by_pieces_ninsns} returns less
5711
than @code{MOVE_RATIO}.
5712
@end defmac
5713
 
5714
@defmac USE_LOAD_POST_INCREMENT (@var{mode})
5715
A C expression used to determine whether a load postincrement is a good
5716
thing to use for a given mode.  Defaults to the value of
5717
@code{HAVE_POST_INCREMENT}.
5718
@end defmac
5719
 
5720
@defmac USE_LOAD_POST_DECREMENT (@var{mode})
5721
A C expression used to determine whether a load postdecrement is a good
5722
thing to use for a given mode.  Defaults to the value of
5723
@code{HAVE_POST_DECREMENT}.
5724
@end defmac
5725
 
5726
@defmac USE_LOAD_PRE_INCREMENT (@var{mode})
5727
A C expression used to determine whether a load preincrement is a good
5728
thing to use for a given mode.  Defaults to the value of
5729
@code{HAVE_PRE_INCREMENT}.
5730
@end defmac
5731
 
5732
@defmac USE_LOAD_PRE_DECREMENT (@var{mode})
5733
A C expression used to determine whether a load predecrement is a good
5734
thing to use for a given mode.  Defaults to the value of
5735
@code{HAVE_PRE_DECREMENT}.
5736
@end defmac
5737
 
5738
@defmac USE_STORE_POST_INCREMENT (@var{mode})
5739
A C expression used to determine whether a store postincrement is a good
5740
thing to use for a given mode.  Defaults to the value of
5741
@code{HAVE_POST_INCREMENT}.
5742
@end defmac
5743
 
5744
@defmac USE_STORE_POST_DECREMENT (@var{mode})
5745
A C expression used to determine whether a store postdecrement is a good
5746
thing to use for a given mode.  Defaults to the value of
5747
@code{HAVE_POST_DECREMENT}.
5748
@end defmac
5749
 
5750
@defmac USE_STORE_PRE_INCREMENT (@var{mode})
5751
This macro is used to determine whether a store preincrement is a good
5752
thing to use for a given mode.  Defaults to the value of
5753
@code{HAVE_PRE_INCREMENT}.
5754
@end defmac
5755
 
5756
@defmac USE_STORE_PRE_DECREMENT (@var{mode})
5757
This macro is used to determine whether a store predecrement is a good
5758
thing to use for a given mode.  Defaults to the value of
5759
@code{HAVE_PRE_DECREMENT}.
5760
@end defmac
5761
 
5762
@defmac NO_FUNCTION_CSE
5763
Define this macro if it is as good or better to call a constant
5764
function address than to call an address kept in a register.
5765
@end defmac
5766
 
5767
@defmac RANGE_TEST_NON_SHORT_CIRCUIT
5768
Define this macro if a non-short-circuit operation produced by
5769
@samp{fold_range_test ()} is optimal.  This macro defaults to true if
5770
@code{BRANCH_COST} is greater than or equal to the value 2.
5771
@end defmac
5772
 
5773
@deftypefn {Target Hook} bool TARGET_RTX_COSTS (rtx @var{x}, int @var{code}, int @var{outer_code}, int *@var{total})
5774
This target hook describes the relative costs of RTL expressions.
5775
 
5776
The cost may depend on the precise form of the expression, which is
5777
available for examination in @var{x}, and the rtx code of the expression
5778
in which it is contained, found in @var{outer_code}.  @var{code} is the
5779
expression code---redundant, since it can be obtained with
5780
@code{GET_CODE (@var{x})}.
5781
 
5782
In implementing this hook, you can use the construct
5783
@code{COSTS_N_INSNS (@var{n})} to specify a cost equal to @var{n} fast
5784
instructions.
5785
 
5786
On entry to the hook, @code{*@var{total}} contains a default estimate
5787
for the cost of the expression.  The hook should modify this value as
5788
necessary.  Traditionally, the default costs are @code{COSTS_N_INSNS (5)}
5789
for multiplications, @code{COSTS_N_INSNS (7)} for division and modulus
5790
operations, and @code{COSTS_N_INSNS (1)} for all other operations.
5791
 
5792
When optimizing for code size, i.e.@: when @code{optimize_size} is
5793
nonzero, this target hook should be used to estimate the relative
5794
size cost of an expression, again relative to @code{COSTS_N_INSNS}.
5795
 
5796
The hook returns true when all subexpressions of @var{x} have been
5797
processed, and false when @code{rtx_cost} should recurse.
5798
@end deftypefn
5799
 
5800
@deftypefn {Target Hook} int TARGET_ADDRESS_COST (rtx @var{address})
5801
This hook computes the cost of an addressing mode that contains
5802
@var{address}.  If not defined, the cost is computed from
5803
the @var{address} expression and the @code{TARGET_RTX_COST} hook.
5804
 
5805
For most CISC machines, the default cost is a good approximation of the
5806
true cost of the addressing mode.  However, on RISC machines, all
5807
instructions normally have the same length and execution time.  Hence
5808
all addresses will have equal costs.
5809
 
5810
In cases where more than one form of an address is known, the form with
5811
the lowest cost will be used.  If multiple forms have the same, lowest,
5812
cost, the one that is the most complex will be used.
5813
 
5814
For example, suppose an address that is equal to the sum of a register
5815
and a constant is used twice in the same basic block.  When this macro
5816
is not defined, the address will be computed in a register and memory
5817
references will be indirect through that register.  On machines where
5818
the cost of the addressing mode containing the sum is no higher than
5819
that of a simple indirect reference, this will produce an additional
5820
instruction and possibly require an additional register.  Proper
5821
specification of this macro eliminates this overhead for such machines.
5822
 
5823
This hook is never called with an invalid address.
5824
 
5825
On machines where an address involving more than one register is as
5826
cheap as an address computation involving only one register, defining
5827
@code{TARGET_ADDRESS_COST} to reflect this can cause two registers to
5828
be live over a region of code where only one would have been if
5829
@code{TARGET_ADDRESS_COST} were not defined in that manner.  This effect
5830
should be considered in the definition of this macro.  Equivalent costs
5831
should probably only be given to addresses with different numbers of
5832
registers on machines with lots of registers.
5833
@end deftypefn
5834
 
5835
@node Scheduling
5836
@section Adjusting the Instruction Scheduler
5837
 
5838
The instruction scheduler may need a fair amount of machine-specific
5839
adjustment in order to produce good code.  GCC provides several target
5840
hooks for this purpose.  It is usually enough to define just a few of
5841
them: try the first ones in this list first.
5842
 
5843
@deftypefn {Target Hook} int TARGET_SCHED_ISSUE_RATE (void)
5844
This hook returns the maximum number of instructions that can ever
5845
issue at the same time on the target machine.  The default is one.
5846
Although the insn scheduler can define itself the possibility of issue
5847
an insn on the same cycle, the value can serve as an additional
5848
constraint to issue insns on the same simulated processor cycle (see
5849
hooks @samp{TARGET_SCHED_REORDER} and @samp{TARGET_SCHED_REORDER2}).
5850
This value must be constant over the entire compilation.  If you need
5851
it to vary depending on what the instructions are, you must use
5852
@samp{TARGET_SCHED_VARIABLE_ISSUE}.
5853
@end deftypefn
5854
 
5855
@deftypefn {Target Hook} int TARGET_SCHED_VARIABLE_ISSUE (FILE *@var{file}, int @var{verbose}, rtx @var{insn}, int @var{more})
5856
This hook is executed by the scheduler after it has scheduled an insn
5857
from the ready list.  It should return the number of insns which can
5858
still be issued in the current cycle.  The default is
5859
@samp{@w{@var{more} - 1}} for insns other than @code{CLOBBER} and
5860
@code{USE}, which normally are not counted against the issue rate.
5861
You should define this hook if some insns take more machine resources
5862
than others, so that fewer insns can follow them in the same cycle.
5863
@var{file} is either a null pointer, or a stdio stream to write any
5864
debug output to.  @var{verbose} is the verbose level provided by
5865
@option{-fsched-verbose-@var{n}}.  @var{insn} is the instruction that
5866
was scheduled.
5867
@end deftypefn
5868
 
5869
@deftypefn {Target Hook} int TARGET_SCHED_ADJUST_COST (rtx @var{insn}, rtx @var{link}, rtx @var{dep_insn}, int @var{cost})
5870
This function corrects the value of @var{cost} based on the
5871
relationship between @var{insn} and @var{dep_insn} through the
5872
dependence @var{link}.  It should return the new value.  The default
5873
is to make no adjustment to @var{cost}.  This can be used for example
5874
to specify to the scheduler using the traditional pipeline description
5875
that an output- or anti-dependence does not incur the same cost as a
5876
data-dependence.  If the scheduler using the automaton based pipeline
5877
description, the cost of anti-dependence is zero and the cost of
5878
output-dependence is maximum of one and the difference of latency
5879
times of the first and the second insns.  If these values are not
5880
acceptable, you could use the hook to modify them too.  See also
5881
@pxref{Processor pipeline description}.
5882
@end deftypefn
5883
 
5884
@deftypefn {Target Hook} int TARGET_SCHED_ADJUST_PRIORITY (rtx @var{insn}, int @var{priority})
5885
This hook adjusts the integer scheduling priority @var{priority} of
5886
@var{insn}.  It should return the new priority.  Increase the priority to
5887
execute @var{insn} earlier, reduce the priority to execute @var{insn}
5888
later.  Do not define this hook if you do not need to adjust the
5889
scheduling priorities of insns.
5890
@end deftypefn
5891
 
5892
@deftypefn {Target Hook} int TARGET_SCHED_REORDER (FILE *@var{file}, int @var{verbose}, rtx *@var{ready}, int *@var{n_readyp}, int @var{clock})
5893
This hook is executed by the scheduler after it has scheduled the ready
5894
list, to allow the machine description to reorder it (for example to
5895
combine two small instructions together on @samp{VLIW} machines).
5896
@var{file} is either a null pointer, or a stdio stream to write any
5897
debug output to.  @var{verbose} is the verbose level provided by
5898
@option{-fsched-verbose-@var{n}}.  @var{ready} is a pointer to the ready
5899
list of instructions that are ready to be scheduled.  @var{n_readyp} is
5900
a pointer to the number of elements in the ready list.  The scheduler
5901
reads the ready list in reverse order, starting with
5902
@var{ready}[@var{*n_readyp}-1] and going to @var{ready}[0].  @var{clock}
5903
is the timer tick of the scheduler.  You may modify the ready list and
5904
the number of ready insns.  The return value is the number of insns that
5905
can issue this cycle; normally this is just @code{issue_rate}.  See also
5906
@samp{TARGET_SCHED_REORDER2}.
5907
@end deftypefn
5908
 
5909
@deftypefn {Target Hook} int TARGET_SCHED_REORDER2 (FILE *@var{file}, int @var{verbose}, rtx *@var{ready}, int *@var{n_ready}, @var{clock})
5910
Like @samp{TARGET_SCHED_REORDER}, but called at a different time.  That
5911
function is called whenever the scheduler starts a new cycle.  This one
5912
is called once per iteration over a cycle, immediately after
5913
@samp{TARGET_SCHED_VARIABLE_ISSUE}; it can reorder the ready list and
5914
return the number of insns to be scheduled in the same cycle.  Defining
5915
this hook can be useful if there are frequent situations where
5916
scheduling one insn causes other insns to become ready in the same
5917
cycle.  These other insns can then be taken into account properly.
5918
@end deftypefn
5919
 
5920
@deftypefn {Target Hook} void TARGET_SCHED_DEPENDENCIES_EVALUATION_HOOK (rtx @var{head}, rtx @var{tail})
5921
This hook is called after evaluation forward dependencies of insns in
5922
chain given by two parameter values (@var{head} and @var{tail}
5923
correspondingly) but before insns scheduling of the insn chain.  For
5924
example, it can be used for better insn classification if it requires
5925
analysis of dependencies.  This hook can use backward and forward
5926
dependencies of the insn scheduler because they are already
5927
calculated.
5928
@end deftypefn
5929
 
5930
@deftypefn {Target Hook} void TARGET_SCHED_INIT (FILE *@var{file}, int @var{verbose}, int @var{max_ready})
5931
This hook is executed by the scheduler at the beginning of each block of
5932
instructions that are to be scheduled.  @var{file} is either a null
5933
pointer, or a stdio stream to write any debug output to.  @var{verbose}
5934
is the verbose level provided by @option{-fsched-verbose-@var{n}}.
5935
@var{max_ready} is the maximum number of insns in the current scheduling
5936
region that can be live at the same time.  This can be used to allocate
5937
scratch space if it is needed, e.g.@: by @samp{TARGET_SCHED_REORDER}.
5938
@end deftypefn
5939
 
5940
@deftypefn {Target Hook} void TARGET_SCHED_FINISH (FILE *@var{file}, int @var{verbose})
5941
This hook is executed by the scheduler at the end of each block of
5942
instructions that are to be scheduled.  It can be used to perform
5943
cleanup of any actions done by the other scheduling hooks.  @var{file}
5944
is either a null pointer, or a stdio stream to write any debug output
5945
to.  @var{verbose} is the verbose level provided by
5946
@option{-fsched-verbose-@var{n}}.
5947
@end deftypefn
5948
 
5949
@deftypefn {Target Hook} void TARGET_SCHED_INIT_GLOBAL (FILE *@var{file}, int @var{verbose}, int @var{old_max_uid})
5950
This hook is executed by the scheduler after function level initializations.
5951
@var{file} is either a null pointer, or a stdio stream to write any debug output to.
5952
@var{verbose} is the verbose level provided by @option{-fsched-verbose-@var{n}}.
5953
@var{old_max_uid} is the maximum insn uid when scheduling begins.
5954
@end deftypefn
5955
 
5956
@deftypefn {Target Hook} void TARGET_SCHED_FINISH_GLOBAL (FILE *@var{file}, int @var{verbose})
5957
This is the cleanup hook corresponding to @code{TARGET_SCHED_INIT_GLOBAL}.
5958
@var{file} is either a null pointer, or a stdio stream to write any debug output to.
5959
@var{verbose} is the verbose level provided by @option{-fsched-verbose-@var{n}}.
5960
@end deftypefn
5961
 
5962
@deftypefn {Target Hook} int TARGET_SCHED_DFA_PRE_CYCLE_INSN (void)
5963
The hook returns an RTL insn.  The automaton state used in the
5964
pipeline hazard recognizer is changed as if the insn were scheduled
5965
when the new simulated processor cycle starts.  Usage of the hook may
5966
simplify the automaton pipeline description for some @acronym{VLIW}
5967
processors.  If the hook is defined, it is used only for the automaton
5968
based pipeline description.  The default is not to change the state
5969
when the new simulated processor cycle starts.
5970
@end deftypefn
5971
 
5972
@deftypefn {Target Hook} void TARGET_SCHED_INIT_DFA_PRE_CYCLE_INSN (void)
5973
The hook can be used to initialize data used by the previous hook.
5974
@end deftypefn
5975
 
5976
@deftypefn {Target Hook} int TARGET_SCHED_DFA_POST_CYCLE_INSN (void)
5977
The hook is analogous to @samp{TARGET_SCHED_DFA_PRE_CYCLE_INSN} but used
5978
to changed the state as if the insn were scheduled when the new
5979
simulated processor cycle finishes.
5980
@end deftypefn
5981
 
5982
@deftypefn {Target Hook} void TARGET_SCHED_INIT_DFA_POST_CYCLE_INSN (void)
5983
The hook is analogous to @samp{TARGET_SCHED_INIT_DFA_PRE_CYCLE_INSN} but
5984
used to initialize data used by the previous hook.
5985
@end deftypefn
5986
 
5987
@deftypefn {Target Hook} int TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD (void)
5988
This hook controls better choosing an insn from the ready insn queue
5989
for the @acronym{DFA}-based insn scheduler.  Usually the scheduler
5990
chooses the first insn from the queue.  If the hook returns a positive
5991
value, an additional scheduler code tries all permutations of
5992
@samp{TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD ()}
5993
subsequent ready insns to choose an insn whose issue will result in
5994
maximal number of issued insns on the same cycle.  For the
5995
@acronym{VLIW} processor, the code could actually solve the problem of
5996
packing simple insns into the @acronym{VLIW} insn.  Of course, if the
5997
rules of @acronym{VLIW} packing are described in the automaton.
5998
 
5999
This code also could be used for superscalar @acronym{RISC}
6000
processors.  Let us consider a superscalar @acronym{RISC} processor
6001
with 3 pipelines.  Some insns can be executed in pipelines @var{A} or
6002
@var{B}, some insns can be executed only in pipelines @var{B} or
6003
@var{C}, and one insn can be executed in pipeline @var{B}.  The
6004
processor may issue the 1st insn into @var{A} and the 2nd one into
6005
@var{B}.  In this case, the 3rd insn will wait for freeing @var{B}
6006
until the next cycle.  If the scheduler issues the 3rd insn the first,
6007
the processor could issue all 3 insns per cycle.
6008
 
6009
Actually this code demonstrates advantages of the automaton based
6010
pipeline hazard recognizer.  We try quickly and easy many insn
6011
schedules to choose the best one.
6012
 
6013
The default is no multipass scheduling.
6014
@end deftypefn
6015
 
6016
@deftypefn {Target Hook} int TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD (rtx)
6017
 
6018
This hook controls what insns from the ready insn queue will be
6019
considered for the multipass insn scheduling.  If the hook returns
6020
zero for insn passed as the parameter, the insn will be not chosen to
6021
be issued.
6022
 
6023
The default is that any ready insns can be chosen to be issued.
6024
@end deftypefn
6025
 
6026
@deftypefn {Target Hook} int TARGET_SCHED_DFA_NEW_CYCLE (FILE *, int, rtx, int, int, int *)
6027
 
6028
This hook is called by the insn scheduler before issuing insn passed
6029
as the third parameter on given cycle.  If the hook returns nonzero,
6030
the insn is not issued on given processors cycle.  Instead of that,
6031
the processor cycle is advanced.  If the value passed through the last
6032
parameter is zero, the insn ready queue is not sorted on the new cycle
6033
start as usually.  The first parameter passes file for debugging
6034
output.  The second one passes the scheduler verbose level of the
6035
debugging output.  The forth and the fifth parameter values are
6036
correspondingly processor cycle on which the previous insn has been
6037
issued and the current processor cycle.
6038
@end deftypefn
6039
 
6040
@deftypefn {Target Hook} bool TARGET_SCHED_IS_COSTLY_DEPENDENCE (rtx @var{insn1}, rtx @var{insn2}, rtx @var{dep_link}, int @var{dep_cost}, int @var{distance})
6041
This hook is used to define which dependences are considered costly by
6042
the target, so costly that it is not advisable to schedule the insns that
6043
are involved in the dependence too close to one another.  The parameters
6044
to this hook are as follows:  The second parameter @var{insn2} is dependent
6045
upon the first parameter @var{insn1}.  The dependence between @var{insn1}
6046
and @var{insn2} is represented by the third parameter @var{dep_link}.  The
6047
fourth parameter @var{cost} is the cost of the dependence, and the fifth
6048
parameter @var{distance} is the distance in cycles between the two insns.
6049
The hook returns @code{true} if considering the distance between the two
6050
insns the dependence between them is considered costly by the target,
6051
and @code{false} otherwise.
6052
 
6053
Defining this hook can be useful in multiple-issue out-of-order machines,
6054
where (a) it's practically hopeless to predict the actual data/resource
6055
delays, however: (b) there's a better chance to predict the actual grouping
6056
that will be formed, and (c) correctly emulating the grouping can be very
6057
important.  In such targets one may want to allow issuing dependent insns
6058
closer to one another---i.e., closer than the dependence distance;  however,
6059
not in cases of "costly dependences", which this hooks allows to define.
6060
@end deftypefn
6061
 
6062
@deftypefn {Target Hook} int TARGET_SCHED_ADJUST_COST_2 (rtx @var{insn}, int @var{dep_type}, rtx @var{dep_insn}, int @var{cost})
6063
This hook is a modified version of @samp{TARGET_SCHED_ADJUST_COST}.  Instead
6064
of passing dependence as a second parameter, it passes a type of that
6065
dependence.  This is useful to calculate cost of dependence between insns
6066
not having the corresponding link.  If @samp{TARGET_SCHED_ADJUST_COST_2} is
6067
defined it is used instead of @samp{TARGET_SCHED_ADJUST_COST}.
6068
@end deftypefn
6069
 
6070
@deftypefn {Target Hook} void TARGET_SCHED_H_I_D_EXTENDED (void)
6071
This hook is called by the insn scheduler after emitting a new instruction to
6072
the instruction stream.  The hook notifies a target backend to extend its
6073
per instruction data structures.
6074
@end deftypefn
6075
 
6076
@deftypefn {Target Hook} int TARGET_SCHED_SPECULATE_INSN (rtx @var{insn}, int @var{request}, rtx *@var{new_pat})
6077
This hook is called by the insn scheduler when @var{insn} has only
6078
speculative dependencies and therefore can be scheduled speculatively.
6079
The hook is used to check if the pattern of @var{insn} has a speculative
6080
version and, in case of successful check, to generate that speculative
6081
pattern.  The hook should return 1, if the instruction has a speculative form,
6082
or -1, if it doesn't.  @var{request} describes the type of requested
6083
speculation.  If the return value equals 1 then @var{new_pat} is assigned
6084
the generated speculative pattern.
6085
@end deftypefn
6086
 
6087
@deftypefn {Target Hook} int TARGET_SCHED_NEEDS_BLOCK_P (rtx @var{insn})
6088
This hook is called by the insn scheduler during generation of recovery code
6089
for @var{insn}.  It should return nonzero, if the corresponding check
6090
instruction should branch to recovery code, or zero otherwise.
6091
@end deftypefn
6092
 
6093
@deftypefn {Target Hook} rtx TARGET_SCHED_GEN_CHECK (rtx @var{insn}, rtx @var{label}, int @var{mutate_p})
6094
This hook is called by the insn scheduler to generate a pattern for recovery
6095
check instruction.  If @var{mutate_p} is zero, then @var{insn} is a
6096
speculative instruction for which the check should be generated.
6097
@var{label} is either a label of a basic block, where recovery code should
6098
be emitted, or a null pointer, when requested check doesn't branch to
6099
recovery code (a simple check).  If @var{mutate_p} is nonzero, then
6100
a pattern for a branchy check corresponding to a simple check denoted by
6101
@var{insn} should be generated.  In this case @var{label} can't be null.
6102
@end deftypefn
6103
 
6104
@deftypefn {Target Hook} int TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC (rtx @var{insn})
6105
This hook is used as a workaround for
6106
@samp{TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD} not being
6107
called on the first instruction of the ready list.  The hook is used to
6108
discard speculative instruction that stand first in the ready list from
6109
being scheduled on the current cycle.  For non-speculative instructions,
6110
the hook should always return nonzero.  For example, in the ia64 backend
6111
the hook is used to cancel data speculative insns when the ALAT table
6112
is nearly full.
6113
@end deftypefn
6114
 
6115
@deftypefn {Target Hook} void TARGET_SCHED_SET_SCHED_FLAGS (unsigned int *@var{flags}, spec_info_t @var{spec_info})
6116
This hook is used by the insn scheduler to find out what features should be
6117
enabled/used.  @var{flags} initially may have either the SCHED_RGN or SCHED_EBB
6118
bit set.  This denotes the scheduler pass for which the data should be
6119
provided.  The target backend should modify @var{flags} by modifying
6120
the bits corresponding to the following features: USE_DEPS_LIST, USE_GLAT,
6121
DETACH_LIFE_INFO, and DO_SPECULATION.  For the DO_SPECULATION feature
6122
an additional structure @var{spec_info} should be filled by the target.
6123
The structure describes speculation types that can be used in the scheduler.
6124
@end deftypefn
6125
 
6126
@node Sections
6127
@section Dividing the Output into Sections (Texts, Data, @dots{})
6128
@c the above section title is WAY too long.  maybe cut the part between
6129
@c the (...)?  --mew 10feb93
6130
 
6131
An object file is divided into sections containing different types of
6132
data.  In the most common case, there are three sections: the @dfn{text
6133
section}, which holds instructions and read-only data; the @dfn{data
6134
section}, which holds initialized writable data; and the @dfn{bss
6135
section}, which holds uninitialized data.  Some systems have other kinds
6136
of sections.
6137
 
6138
@file{varasm.c} provides several well-known sections, such as
6139
@code{text_section}, @code{data_section} and @code{bss_section}.
6140
The normal way of controlling a @code{@var{foo}_section} variable
6141
is to define the associated @code{@var{FOO}_SECTION_ASM_OP} macro,
6142
as described below.  The macros are only read once, when @file{varasm.c}
6143
initializes itself, so their values must be run-time constants.
6144
They may however depend on command-line flags.
6145
 
6146
@emph{Note:} Some run-time files, such @file{crtstuff.c}, also make
6147
use of the @code{@var{FOO}_SECTION_ASM_OP} macros, and expect them
6148
to be string literals.
6149
 
6150
Some assemblers require a different string to be written every time a
6151
section is selected.  If your assembler falls into this category, you
6152
should define the @code{TARGET_ASM_INIT_SECTIONS} hook and use
6153
@code{get_unnamed_section} to set up the sections.
6154
 
6155
You must always create a @code{text_section}, either by defining
6156
@code{TEXT_SECTION_ASM_OP} or by initializing @code{text_section}
6157
in @code{TARGET_ASM_INIT_SECTIONS}.  The same is true of
6158
@code{data_section} and @code{DATA_SECTION_ASM_OP}.  If you do not
6159
create a distinct @code{readonly_data_section}, the default is to
6160
reuse @code{text_section}.
6161
 
6162
All the other @file{varasm.c} sections are optional, and are null
6163
if the target does not provide them.
6164
 
6165
@defmac TEXT_SECTION_ASM_OP
6166
A C expression whose value is a string, including spacing, containing the
6167
assembler operation that should precede instructions and read-only data.
6168
Normally @code{"\t.text"} is right.
6169
@end defmac
6170
 
6171
@defmac HOT_TEXT_SECTION_NAME
6172
If defined, a C string constant for the name of the section containing most
6173
frequently executed functions of the program.  If not defined, GCC will provide
6174
a default definition if the target supports named sections.
6175
@end defmac
6176
 
6177
@defmac UNLIKELY_EXECUTED_TEXT_SECTION_NAME
6178
If defined, a C string constant for the name of the section containing unlikely
6179
executed functions in the program.
6180
@end defmac
6181
 
6182
@defmac DATA_SECTION_ASM_OP
6183
A C expression whose value is a string, including spacing, containing the
6184
assembler operation to identify the following data as writable initialized
6185
data.  Normally @code{"\t.data"} is right.
6186
@end defmac
6187
 
6188
@defmac SDATA_SECTION_ASM_OP
6189
If defined, a C expression whose value is a string, including spacing,
6190
containing the assembler operation to identify the following data as
6191
initialized, writable small data.
6192
@end defmac
6193
 
6194
@defmac READONLY_DATA_SECTION_ASM_OP
6195
A C expression whose value is a string, including spacing, containing the
6196
assembler operation to identify the following data as read-only initialized
6197
data.
6198
@end defmac
6199
 
6200
@defmac BSS_SECTION_ASM_OP
6201
If defined, a C expression whose value is a string, including spacing,
6202
containing the assembler operation to identify the following data as
6203
uninitialized global data.  If not defined, and neither
6204
@code{ASM_OUTPUT_BSS} nor @code{ASM_OUTPUT_ALIGNED_BSS} are defined,
6205
uninitialized global data will be output in the data section if
6206
@option{-fno-common} is passed, otherwise @code{ASM_OUTPUT_COMMON} will be
6207
used.
6208
@end defmac
6209
 
6210
@defmac SBSS_SECTION_ASM_OP
6211
If defined, a C expression whose value is a string, including spacing,
6212
containing the assembler operation to identify the following data as
6213
uninitialized, writable small data.
6214
@end defmac
6215
 
6216
@defmac INIT_SECTION_ASM_OP
6217
If defined, a C expression whose value is a string, including spacing,
6218
containing the assembler operation to identify the following data as
6219
initialization code.  If not defined, GCC will assume such a section does
6220
not exist.  This section has no corresponding @code{init_section}
6221
variable; it is used entirely in runtime code.
6222
@end defmac
6223
 
6224
@defmac FINI_SECTION_ASM_OP
6225
If defined, a C expression whose value is a string, including spacing,
6226
containing the assembler operation to identify the following data as
6227
finalization code.  If not defined, GCC will assume such a section does
6228
not exist.  This section has no corresponding @code{fini_section}
6229
variable; it is used entirely in runtime code.
6230
@end defmac
6231
 
6232
@defmac INIT_ARRAY_SECTION_ASM_OP
6233
If defined, a C expression whose value is a string, including spacing,
6234
containing the assembler operation to identify the following data as
6235
part of the @code{.init_array} (or equivalent) section.  If not
6236
defined, GCC will assume such a section does not exist.  Do not define
6237
both this macro and @code{INIT_SECTION_ASM_OP}.
6238
@end defmac
6239
 
6240
@defmac FINI_ARRAY_SECTION_ASM_OP
6241
If defined, a C expression whose value is a string, including spacing,
6242
containing the assembler operation to identify the following data as
6243
part of the @code{.fini_array} (or equivalent) section.  If not
6244
defined, GCC will assume such a section does not exist.  Do not define
6245
both this macro and @code{FINI_SECTION_ASM_OP}.
6246
@end defmac
6247
 
6248
@defmac CRT_CALL_STATIC_FUNCTION (@var{section_op}, @var{function})
6249
If defined, an ASM statement that switches to a different section
6250
via @var{section_op}, calls @var{function}, and switches back to
6251
the text section.  This is used in @file{crtstuff.c} if
6252
@code{INIT_SECTION_ASM_OP} or @code{FINI_SECTION_ASM_OP} to calls
6253
to initialization and finalization functions from the init and fini
6254
sections.  By default, this macro uses a simple function call.  Some
6255
ports need hand-crafted assembly code to avoid dependencies on
6256
registers initialized in the function prologue or to ensure that
6257
constant pools don't end up too far way in the text section.
6258
@end defmac
6259
 
6260
@defmac TARGET_LIBGCC_SDATA_SECTION
6261
If defined, a string which names the section into which small
6262
variables defined in crtstuff and libgcc should go.  This is useful
6263
when the target has options for optimizing access to small data, and
6264
you want the crtstuff and libgcc routines to be conservative in what
6265
they expect of your application yet liberal in what your application
6266
expects.  For example, for targets with a @code{.sdata} section (like
6267
MIPS), you could compile crtstuff with @code{-G 0} so that it doesn't
6268
require small data support from your application, but use this macro
6269
to put small data into @code{.sdata} so that your application can
6270
access these variables whether it uses small data or not.
6271
@end defmac
6272
 
6273
@defmac FORCE_CODE_SECTION_ALIGN
6274
If defined, an ASM statement that aligns a code section to some
6275
arbitrary boundary.  This is used to force all fragments of the
6276
@code{.init} and @code{.fini} sections to have to same alignment
6277
and thus prevent the linker from having to add any padding.
6278
@end defmac
6279
 
6280
@defmac JUMP_TABLES_IN_TEXT_SECTION
6281
Define this macro to be an expression with a nonzero value if jump
6282
tables (for @code{tablejump} insns) should be output in the text
6283
section, along with the assembler instructions.  Otherwise, the
6284
readonly data section is used.
6285
 
6286
This macro is irrelevant if there is no separate readonly data section.
6287
@end defmac
6288
 
6289
@deftypefn {Target Hook} void TARGET_ASM_INIT_SECTIONS (void)
6290
Define this hook if you need to do something special to set up the
6291
@file{varasm.c} sections, or if your target has some special sections
6292
of its own that you need to create.
6293
 
6294
GCC calls this hook after processing the command line, but before writing
6295
any assembly code, and before calling any of the section-returning hooks
6296
described below.
6297
@end deftypefn
6298
 
6299
@deftypefn {Target Hook} TARGET_ASM_RELOC_RW_MASK (void)
6300
Return a mask describing how relocations should be treated when
6301
selecting sections.  Bit 1 should be set if global relocations
6302
should be placed in a read-write section; bit 0 should be set if
6303
local relocations should be placed in a read-write section.
6304
 
6305
The default version of this function returns 3 when @option{-fpic}
6306
is in effect, and 0 otherwise.  The hook is typically redefined
6307
when the target cannot support (some kinds of) dynamic relocations
6308
in read-only sections even in executables.
6309
@end deftypefn
6310
 
6311
@deftypefn {Target Hook} {section *} TARGET_ASM_SELECT_SECTION (tree @var{exp}, int @var{reloc}, unsigned HOST_WIDE_INT @var{align})
6312
Return the section into which @var{exp} should be placed.  You can
6313
assume that @var{exp} is either a @code{VAR_DECL} node or a constant of
6314
some sort.  @var{reloc} indicates whether the initial value of @var{exp}
6315
requires link-time relocations.  Bit 0 is set when variable contains
6316
local relocations only, while bit 1 is set for global relocations.
6317
@var{align} is the constant alignment in bits.
6318
 
6319
The default version of this function takes care of putting read-only
6320
variables in @code{readonly_data_section}.
6321
 
6322
See also @var{USE_SELECT_SECTION_FOR_FUNCTIONS}.
6323
@end deftypefn
6324
 
6325
@defmac USE_SELECT_SECTION_FOR_FUNCTIONS
6326
Define this macro if you wish TARGET_ASM_SELECT_SECTION to be called
6327
for @code{FUNCTION_DECL}s as well as for variables and constants.
6328
 
6329
In the case of a @code{FUNCTION_DECL}, @var{reloc} will be zero if the
6330
function has been determined to be likely to be called, and nonzero if
6331
it is unlikely to be called.
6332
@end defmac
6333
 
6334
@deftypefn {Target Hook} void TARGET_ASM_UNIQUE_SECTION (tree @var{decl}, int @var{reloc})
6335
Build up a unique section name, expressed as a @code{STRING_CST} node,
6336
and assign it to @samp{DECL_SECTION_NAME (@var{decl})}.
6337
As with @code{TARGET_ASM_SELECT_SECTION}, @var{reloc} indicates whether
6338
the initial value of @var{exp} requires link-time relocations.
6339
 
6340
The default version of this function appends the symbol name to the
6341
ELF section name that would normally be used for the symbol.  For
6342
example, the function @code{foo} would be placed in @code{.text.foo}.
6343
Whatever the actual target object format, this is often good enough.
6344
@end deftypefn
6345
 
6346
@deftypefn {Target Hook} {section *} TARGET_ASM_FUNCTION_RODATA_SECTION (tree @var{decl})
6347
Return the readonly data section associated with
6348
@samp{DECL_SECTION_NAME (@var{decl})}.
6349
The default version of this function selects @code{.gnu.linkonce.r.name} if
6350
the function's section is @code{.gnu.linkonce.t.name}, @code{.rodata.name}
6351
if function is in @code{.text.name}, and the normal readonly-data section
6352
otherwise.
6353
@end deftypefn
6354
 
6355
@deftypefn {Target Hook} {section *} TARGET_ASM_SELECT_RTX_SECTION (enum machine_mode @var{mode}, rtx @var{x}, unsigned HOST_WIDE_INT @var{align})
6356
Return the section into which a constant @var{x}, of mode @var{mode},
6357
should be placed.  You can assume that @var{x} is some kind of
6358
constant in RTL@.  The argument @var{mode} is redundant except in the
6359
case of a @code{const_int} rtx.  @var{align} is the constant alignment
6360
in bits.
6361
 
6362
The default version of this function takes care of putting symbolic
6363
constants in @code{flag_pic} mode in @code{data_section} and everything
6364
else in @code{readonly_data_section}.
6365
@end deftypefn
6366
 
6367
@deftypefn {Target Hook} void TARGET_ENCODE_SECTION_INFO (tree @var{decl}, rtx @var{rtl}, int @var{new_decl_p})
6368
Define this hook if references to a symbol or a constant must be
6369
treated differently depending on something about the variable or
6370
function named by the symbol (such as what section it is in).
6371
 
6372
The hook is executed immediately after rtl has been created for
6373
@var{decl}, which may be a variable or function declaration or
6374
an entry in the constant pool.  In either case, @var{rtl} is the
6375
rtl in question.  Do @emph{not} use @code{DECL_RTL (@var{decl})}
6376
in this hook; that field may not have been initialized yet.
6377
 
6378
In the case of a constant, it is safe to assume that the rtl is
6379
a @code{mem} whose address is a @code{symbol_ref}.  Most decls
6380
will also have this form, but that is not guaranteed.  Global
6381
register variables, for instance, will have a @code{reg} for their
6382
rtl.  (Normally the right thing to do with such unusual rtl is
6383
leave it alone.)
6384
 
6385
The @var{new_decl_p} argument will be true if this is the first time
6386
that @code{TARGET_ENCODE_SECTION_INFO} has been invoked on this decl.  It will
6387
be false for subsequent invocations, which will happen for duplicate
6388
declarations.  Whether or not anything must be done for the duplicate
6389
declaration depends on whether the hook examines @code{DECL_ATTRIBUTES}.
6390
@var{new_decl_p} is always true when the hook is called for a constant.
6391
 
6392
@cindex @code{SYMBOL_REF_FLAG}, in @code{TARGET_ENCODE_SECTION_INFO}
6393
The usual thing for this hook to do is to record flags in the
6394
@code{symbol_ref}, using @code{SYMBOL_REF_FLAG} or @code{SYMBOL_REF_FLAGS}.
6395
Historically, the name string was modified if it was necessary to
6396
encode more than one bit of information, but this practice is now
6397
discouraged; use @code{SYMBOL_REF_FLAGS}.
6398
 
6399
The default definition of this hook, @code{default_encode_section_info}
6400
in @file{varasm.c}, sets a number of commonly-useful bits in
6401
@code{SYMBOL_REF_FLAGS}.  Check whether the default does what you need
6402
before overriding it.
6403
@end deftypefn
6404
 
6405
@deftypefn {Target Hook} const char *TARGET_STRIP_NAME_ENCODING (const char *name)
6406
Decode @var{name} and return the real name part, sans
6407
the characters that @code{TARGET_ENCODE_SECTION_INFO}
6408
may have added.
6409
@end deftypefn
6410
 
6411
@deftypefn {Target Hook} bool TARGET_IN_SMALL_DATA_P (tree @var{exp})
6412
Returns true if @var{exp} should be placed into a ``small data'' section.
6413
The default version of this hook always returns false.
6414
@end deftypefn
6415
 
6416
@deftypevar {Target Hook} bool TARGET_HAVE_SRODATA_SECTION
6417
Contains the value true if the target places read-only
6418
``small data'' into a separate section.  The default value is false.
6419
@end deftypevar
6420
 
6421
@deftypefn {Target Hook} bool TARGET_BINDS_LOCAL_P (tree @var{exp})
6422
Returns true if @var{exp} names an object for which name resolution
6423
rules must resolve to the current ``module'' (dynamic shared library
6424
or executable image).
6425
 
6426
The default version of this hook implements the name resolution rules
6427
for ELF, which has a looser model of global name binding than other
6428
currently supported object file formats.
6429
@end deftypefn
6430
 
6431
@deftypevar {Target Hook} bool TARGET_HAVE_TLS
6432
Contains the value true if the target supports thread-local storage.
6433
The default value is false.
6434
@end deftypevar
6435
 
6436
 
6437
@node PIC
6438
@section Position Independent Code
6439
@cindex position independent code
6440
@cindex PIC
6441
 
6442
This section describes macros that help implement generation of position
6443
independent code.  Simply defining these macros is not enough to
6444
generate valid PIC; you must also add support to the macros
6445
@code{GO_IF_LEGITIMATE_ADDRESS} and @code{PRINT_OPERAND_ADDRESS}, as
6446
well as @code{LEGITIMIZE_ADDRESS}.  You must modify the definition of
6447
@samp{movsi} to do something appropriate when the source operand
6448
contains a symbolic address.  You may also need to alter the handling of
6449
switch statements so that they use relative addresses.
6450
@c i rearranged the order of the macros above to try to force one of
6451
@c them to the next line, to eliminate an overfull hbox. --mew 10feb93
6452
 
6453
@defmac PIC_OFFSET_TABLE_REGNUM
6454
The register number of the register used to address a table of static
6455
data addresses in memory.  In some cases this register is defined by a
6456
processor's ``application binary interface'' (ABI)@.  When this macro
6457
is defined, RTL is generated for this register once, as with the stack
6458
pointer and frame pointer registers.  If this macro is not defined, it
6459
is up to the machine-dependent files to allocate such a register (if
6460
necessary).  Note that this register must be fixed when in use (e.g.@:
6461
when @code{flag_pic} is true).
6462
@end defmac
6463
 
6464
@defmac PIC_OFFSET_TABLE_REG_CALL_CLOBBERED
6465
Define this macro if the register defined by
6466
@code{PIC_OFFSET_TABLE_REGNUM} is clobbered by calls.  Do not define
6467
this macro if @code{PIC_OFFSET_TABLE_REGNUM} is not defined.
6468
@end defmac
6469
 
6470
@defmac LEGITIMATE_PIC_OPERAND_P (@var{x})
6471
A C expression that is nonzero if @var{x} is a legitimate immediate
6472
operand on the target machine when generating position independent code.
6473
You can assume that @var{x} satisfies @code{CONSTANT_P}, so you need not
6474
check this.  You can also assume @var{flag_pic} is true, so you need not
6475
check it either.  You need not define this macro if all constants
6476
(including @code{SYMBOL_REF}) can be immediate operands when generating
6477
position independent code.
6478
@end defmac
6479
 
6480
@node Assembler Format
6481
@section Defining the Output Assembler Language
6482
 
6483
This section describes macros whose principal purpose is to describe how
6484
to write instructions in assembler language---rather than what the
6485
instructions do.
6486
 
6487
@menu
6488
* File Framework::       Structural information for the assembler file.
6489
* Data Output::          Output of constants (numbers, strings, addresses).
6490
* Uninitialized Data::   Output of uninitialized variables.
6491
* Label Output::         Output and generation of labels.
6492
* Initialization::       General principles of initialization
6493
                           and termination routines.
6494
* Macros for Initialization::
6495
                         Specific macros that control the handling of
6496
                           initialization and termination routines.
6497
* Instruction Output::   Output of actual instructions.
6498
* Dispatch Tables::      Output of jump tables.
6499
* Exception Region Output:: Output of exception region code.
6500
* Alignment Output::     Pseudo ops for alignment and skipping data.
6501
@end menu
6502
 
6503
@node File Framework
6504
@subsection The Overall Framework of an Assembler File
6505
@cindex assembler format
6506
@cindex output of assembler code
6507
 
6508
@c prevent bad page break with this line
6509
This describes the overall framework of an assembly file.
6510
 
6511
@deftypefn {Target Hook} void TARGET_ASM_FILE_START ()
6512
@findex default_file_start
6513
Output to @code{asm_out_file} any text which the assembler expects to
6514
find at the beginning of a file.  The default behavior is controlled
6515
by two flags, documented below.  Unless your target's assembler is
6516
quite unusual, if you override the default, you should call
6517
@code{default_file_start} at some point in your target hook.  This
6518
lets other target files rely on these variables.
6519
@end deftypefn
6520
 
6521
@deftypevr {Target Hook} bool TARGET_ASM_FILE_START_APP_OFF
6522
If this flag is true, the text of the macro @code{ASM_APP_OFF} will be
6523
printed as the very first line in the assembly file, unless
6524
@option{-fverbose-asm} is in effect.  (If that macro has been defined
6525
to the empty string, this variable has no effect.)  With the normal
6526
definition of @code{ASM_APP_OFF}, the effect is to notify the GNU
6527
assembler that it need not bother stripping comments or extra
6528
whitespace from its input.  This allows it to work a bit faster.
6529
 
6530
The default is false.  You should not set it to true unless you have
6531
verified that your port does not generate any extra whitespace or
6532
comments that will cause GAS to issue errors in NO_APP mode.
6533
@end deftypevr
6534
 
6535
@deftypevr {Target Hook} bool TARGET_ASM_FILE_START_FILE_DIRECTIVE
6536
If this flag is true, @code{output_file_directive} will be called
6537
for the primary source file, immediately after printing
6538
@code{ASM_APP_OFF} (if that is enabled).  Most ELF assemblers expect
6539
this to be done.  The default is false.
6540
@end deftypevr
6541
 
6542
@deftypefn {Target Hook} void TARGET_ASM_FILE_END ()
6543
Output to @code{asm_out_file} any text which the assembler expects
6544
to find at the end of a file.  The default is to output nothing.
6545
@end deftypefn
6546
 
6547
@deftypefun void file_end_indicate_exec_stack ()
6548
Some systems use a common convention, the @samp{.note.GNU-stack}
6549
special section, to indicate whether or not an object file relies on
6550
the stack being executable.  If your system uses this convention, you
6551
should define @code{TARGET_ASM_FILE_END} to this function.  If you
6552
need to do other things in that hook, have your hook function call
6553
this function.
6554
@end deftypefun
6555
 
6556
@defmac ASM_COMMENT_START
6557
A C string constant describing how to begin a comment in the target
6558
assembler language.  The compiler assumes that the comment will end at
6559
the end of the line.
6560
@end defmac
6561
 
6562
@defmac ASM_APP_ON
6563
A C string constant for text to be output before each @code{asm}
6564
statement or group of consecutive ones.  Normally this is
6565
@code{"#APP"}, which is a comment that has no effect on most
6566
assemblers but tells the GNU assembler that it must check the lines
6567
that follow for all valid assembler constructs.
6568
@end defmac
6569
 
6570
@defmac ASM_APP_OFF
6571
A C string constant for text to be output after each @code{asm}
6572
statement or group of consecutive ones.  Normally this is
6573
@code{"#NO_APP"}, which tells the GNU assembler to resume making the
6574
time-saving assumptions that are valid for ordinary compiler output.
6575
@end defmac
6576
 
6577
@defmac ASM_OUTPUT_SOURCE_FILENAME (@var{stream}, @var{name})
6578
A C statement to output COFF information or DWARF debugging information
6579
which indicates that filename @var{name} is the current source file to
6580
the stdio stream @var{stream}.
6581
 
6582
This macro need not be defined if the standard form of output
6583
for the file format in use is appropriate.
6584
@end defmac
6585
 
6586
@defmac OUTPUT_QUOTED_STRING (@var{stream}, @var{string})
6587
A C statement to output the string @var{string} to the stdio stream
6588
@var{stream}.  If you do not call the function @code{output_quoted_string}
6589
in your config files, GCC will only call it to output filenames to
6590
the assembler source.  So you can use it to canonicalize the format
6591
of the filename using this macro.
6592
@end defmac
6593
 
6594
@defmac ASM_OUTPUT_IDENT (@var{stream}, @var{string})
6595
A C statement to output something to the assembler file to handle a
6596
@samp{#ident} directive containing the text @var{string}.  If this
6597
macro is not defined, nothing is output for a @samp{#ident} directive.
6598
@end defmac
6599
 
6600
@deftypefn {Target Hook} void TARGET_ASM_NAMED_SECTION (const char *@var{name}, unsigned int @var{flags}, unsigned int @var{align})
6601
Output assembly directives to switch to section @var{name}.  The section
6602
should have attributes as specified by @var{flags}, which is a bit mask
6603
of the @code{SECTION_*} flags defined in @file{output.h}.  If @var{align}
6604
is nonzero, it contains an alignment in bytes to be used for the section,
6605
otherwise some target default should be used.  Only targets that must
6606
specify an alignment within the section directive need pay attention to
6607
@var{align} -- we will still use @code{ASM_OUTPUT_ALIGN}.
6608
@end deftypefn
6609
 
6610
@deftypefn {Target Hook} bool TARGET_HAVE_NAMED_SECTIONS
6611
This flag is true if the target supports @code{TARGET_ASM_NAMED_SECTION}.
6612
@end deftypefn
6613
 
6614
@anchor{TARGET_HAVE_SWITCHABLE_BSS_SECTIONS}
6615
@deftypefn {Target Hook} bool TARGET_HAVE_SWITCHABLE_BSS_SECTIONS
6616
This flag is true if we can create zeroed data by switching to a BSS
6617
section and then using @code{ASM_OUTPUT_SKIP} to allocate the space.
6618
This is true on most ELF targets.
6619
@end deftypefn
6620
 
6621
@deftypefn {Target Hook} {unsigned int} TARGET_SECTION_TYPE_FLAGS (tree @var{decl}, const char *@var{name}, int @var{reloc})
6622
Choose a set of section attributes for use by @code{TARGET_ASM_NAMED_SECTION}
6623
based on a variable or function decl, a section name, and whether or not the
6624
declaration's initializer may contain runtime relocations.  @var{decl} may be
6625
 null, in which case read-write data should be assumed.
6626
 
6627
The default version of this function handles choosing code vs data,
6628
read-only vs read-write data, and @code{flag_pic}.  You should only
6629
need to override this if your target has special flags that might be
6630
set via @code{__attribute__}.
6631
@end deftypefn
6632
 
6633
@need 2000
6634
@node Data Output
6635
@subsection Output of Data
6636
 
6637
 
6638
@deftypevr {Target Hook} {const char *} TARGET_ASM_BYTE_OP
6639
@deftypevrx {Target Hook} {const char *} TARGET_ASM_ALIGNED_HI_OP
6640
@deftypevrx {Target Hook} {const char *} TARGET_ASM_ALIGNED_SI_OP
6641
@deftypevrx {Target Hook} {const char *} TARGET_ASM_ALIGNED_DI_OP
6642
@deftypevrx {Target Hook} {const char *} TARGET_ASM_ALIGNED_TI_OP
6643
@deftypevrx {Target Hook} {const char *} TARGET_ASM_UNALIGNED_HI_OP
6644
@deftypevrx {Target Hook} {const char *} TARGET_ASM_UNALIGNED_SI_OP
6645
@deftypevrx {Target Hook} {const char *} TARGET_ASM_UNALIGNED_DI_OP
6646
@deftypevrx {Target Hook} {const char *} TARGET_ASM_UNALIGNED_TI_OP
6647
These hooks specify assembly directives for creating certain kinds
6648
of integer object.  The @code{TARGET_ASM_BYTE_OP} directive creates a
6649
byte-sized object, the @code{TARGET_ASM_ALIGNED_HI_OP} one creates an
6650
aligned two-byte object, and so on.  Any of the hooks may be
6651
@code{NULL}, indicating that no suitable directive is available.
6652
 
6653
The compiler will print these strings at the start of a new line,
6654
followed immediately by the object's initial value.  In most cases,
6655
the string should contain a tab, a pseudo-op, and then another tab.
6656
@end deftypevr
6657
 
6658
@deftypefn {Target Hook} bool TARGET_ASM_INTEGER (rtx @var{x}, unsigned int @var{size}, int @var{aligned_p})
6659
The @code{assemble_integer} function uses this hook to output an
6660
integer object.  @var{x} is the object's value, @var{size} is its size
6661
in bytes and @var{aligned_p} indicates whether it is aligned.  The
6662
function should return @code{true} if it was able to output the
6663
object.  If it returns false, @code{assemble_integer} will try to
6664
split the object into smaller parts.
6665
 
6666
The default implementation of this hook will use the
6667
@code{TARGET_ASM_BYTE_OP} family of strings, returning @code{false}
6668
when the relevant string is @code{NULL}.
6669
@end deftypefn
6670
 
6671
@defmac OUTPUT_ADDR_CONST_EXTRA (@var{stream}, @var{x}, @var{fail})
6672
A C statement to recognize @var{rtx} patterns that
6673
@code{output_addr_const} can't deal with, and output assembly code to
6674
@var{stream} corresponding to the pattern @var{x}.  This may be used to
6675
allow machine-dependent @code{UNSPEC}s to appear within constants.
6676
 
6677
If @code{OUTPUT_ADDR_CONST_EXTRA} fails to recognize a pattern, it must
6678
@code{goto fail}, so that a standard error message is printed.  If it
6679
prints an error message itself, by calling, for example,
6680
@code{output_operand_lossage}, it may just complete normally.
6681
@end defmac
6682
 
6683
@defmac ASM_OUTPUT_ASCII (@var{stream}, @var{ptr}, @var{len})
6684
A C statement to output to the stdio stream @var{stream} an assembler
6685
instruction to assemble a string constant containing the @var{len}
6686
bytes at @var{ptr}.  @var{ptr} will be a C expression of type
6687
@code{char *} and @var{len} a C expression of type @code{int}.
6688
 
6689
If the assembler has a @code{.ascii} pseudo-op as found in the
6690
Berkeley Unix assembler, do not define the macro
6691
@code{ASM_OUTPUT_ASCII}.
6692
@end defmac
6693
 
6694
@defmac ASM_OUTPUT_FDESC (@var{stream}, @var{decl}, @var{n})
6695
A C statement to output word @var{n} of a function descriptor for
6696
@var{decl}.  This must be defined if @code{TARGET_VTABLE_USES_DESCRIPTORS}
6697
is defined, and is otherwise unused.
6698
@end defmac
6699
 
6700
@defmac CONSTANT_POOL_BEFORE_FUNCTION
6701
You may define this macro as a C expression.  You should define the
6702
expression to have a nonzero value if GCC should output the constant
6703
pool for a function before the code for the function, or a zero value if
6704
GCC should output the constant pool after the function.  If you do
6705
not define this macro, the usual case, GCC will output the constant
6706
pool before the function.
6707
@end defmac
6708
 
6709
@defmac ASM_OUTPUT_POOL_PROLOGUE (@var{file}, @var{funname}, @var{fundecl}, @var{size})
6710
A C statement to output assembler commands to define the start of the
6711
constant pool for a function.  @var{funname} is a string giving
6712
the name of the function.  Should the return type of the function
6713
be required, it can be obtained via @var{fundecl}.  @var{size}
6714
is the size, in bytes, of the constant pool that will be written
6715
immediately after this call.
6716
 
6717
If no constant-pool prefix is required, the usual case, this macro need
6718
not be defined.
6719
@end defmac
6720
 
6721
@defmac ASM_OUTPUT_SPECIAL_POOL_ENTRY (@var{file}, @var{x}, @var{mode}, @var{align}, @var{labelno}, @var{jumpto})
6722
A C statement (with or without semicolon) to output a constant in the
6723
constant pool, if it needs special treatment.  (This macro need not do
6724
anything for RTL expressions that can be output normally.)
6725
 
6726
The argument @var{file} is the standard I/O stream to output the
6727
assembler code on.  @var{x} is the RTL expression for the constant to
6728
output, and @var{mode} is the machine mode (in case @var{x} is a
6729
@samp{const_int}).  @var{align} is the required alignment for the value
6730
@var{x}; you should output an assembler directive to force this much
6731
alignment.
6732
 
6733
The argument @var{labelno} is a number to use in an internal label for
6734
the address of this pool entry.  The definition of this macro is
6735
responsible for outputting the label definition at the proper place.
6736
Here is how to do this:
6737
 
6738
@smallexample
6739
@code{(*targetm.asm_out.internal_label)} (@var{file}, "LC", @var{labelno});
6740
@end smallexample
6741
 
6742
When you output a pool entry specially, you should end with a
6743
@code{goto} to the label @var{jumpto}.  This will prevent the same pool
6744
entry from being output a second time in the usual manner.
6745
 
6746
You need not define this macro if it would do nothing.
6747
@end defmac
6748
 
6749
@defmac ASM_OUTPUT_POOL_EPILOGUE (@var{file} @var{funname} @var{fundecl} @var{size})
6750
A C statement to output assembler commands to at the end of the constant
6751
pool for a function.  @var{funname} is a string giving the name of the
6752
function.  Should the return type of the function be required, you can
6753
obtain it via @var{fundecl}.  @var{size} is the size, in bytes, of the
6754
constant pool that GCC wrote immediately before this call.
6755
 
6756
If no constant-pool epilogue is required, the usual case, you need not
6757
define this macro.
6758
@end defmac
6759
 
6760
@defmac IS_ASM_LOGICAL_LINE_SEPARATOR (@var{C})
6761
Define this macro as a C expression which is nonzero if @var{C} is
6762
used as a logical line separator by the assembler.
6763
 
6764
If you do not define this macro, the default is that only
6765
the character @samp{;} is treated as a logical line separator.
6766
@end defmac
6767
 
6768
@deftypevr {Target Hook} {const char *} TARGET_ASM_OPEN_PAREN
6769
@deftypevrx {Target Hook} {const char *} TARGET_ASM_CLOSE_PAREN
6770
These target hooks are C string constants, describing the syntax in the
6771
assembler for grouping arithmetic expressions.  If not overridden, they
6772
default to normal parentheses, which is correct for most assemblers.
6773
@end deftypevr
6774
 
6775
  These macros are provided by @file{real.h} for writing the definitions
6776
of @code{ASM_OUTPUT_DOUBLE} and the like:
6777
 
6778
@defmac REAL_VALUE_TO_TARGET_SINGLE (@var{x}, @var{l})
6779
@defmacx REAL_VALUE_TO_TARGET_DOUBLE (@var{x}, @var{l})
6780
@defmacx REAL_VALUE_TO_TARGET_LONG_DOUBLE (@var{x}, @var{l})
6781
@defmacx REAL_VALUE_TO_TARGET_DECIMAL32 (@var{x}, @var{l})
6782
@defmacx REAL_VALUE_TO_TARGET_DECIMAL64 (@var{x}, @var{l})
6783
@defmacx REAL_VALUE_TO_TARGET_DECIMAL128 (@var{x}, @var{l})
6784
These translate @var{x}, of type @code{REAL_VALUE_TYPE}, to the
6785
target's floating point representation, and store its bit pattern in
6786
the variable @var{l}.  For @code{REAL_VALUE_TO_TARGET_SINGLE} and
6787
@code{REAL_VALUE_TO_TARGET_DECIMAL32}, this variable should be a
6788
simple @code{long int}.  For the others, it should be an array of
6789
@code{long int}.  The number of elements in this array is determined
6790
by the size of the desired target floating point data type: 32 bits of
6791
it go in each @code{long int} array element.  Each array element holds
6792
32 bits of the result, even if @code{long int} is wider than 32 bits
6793
on the host machine.
6794
 
6795
The array element values are designed so that you can print them out
6796
using @code{fprintf} in the order they should appear in the target
6797
machine's memory.
6798
@end defmac
6799
 
6800
@node Uninitialized Data
6801
@subsection Output of Uninitialized Variables
6802
 
6803
Each of the macros in this section is used to do the whole job of
6804
outputting a single uninitialized variable.
6805
 
6806
@defmac ASM_OUTPUT_COMMON (@var{stream}, @var{name}, @var{size}, @var{rounded})
6807
A C statement (sans semicolon) to output to the stdio stream
6808
@var{stream} the assembler definition of a common-label named
6809
@var{name} whose size is @var{size} bytes.  The variable @var{rounded}
6810
is the size rounded up to whatever alignment the caller wants.
6811
 
6812
Use the expression @code{assemble_name (@var{stream}, @var{name})} to
6813
output the name itself; before and after that, output the additional
6814
assembler syntax for defining the name, and a newline.
6815
 
6816
This macro controls how the assembler definitions of uninitialized
6817
common global variables are output.
6818
@end defmac
6819
 
6820
@defmac ASM_OUTPUT_ALIGNED_COMMON (@var{stream}, @var{name}, @var{size}, @var{alignment})
6821
Like @code{ASM_OUTPUT_COMMON} except takes the required alignment as a
6822
separate, explicit argument.  If you define this macro, it is used in
6823
place of @code{ASM_OUTPUT_COMMON}, and gives you more flexibility in
6824
handling the required alignment of the variable.  The alignment is specified
6825
as the number of bits.
6826
@end defmac
6827
 
6828
@defmac ASM_OUTPUT_ALIGNED_DECL_COMMON (@var{stream}, @var{decl}, @var{name}, @var{size}, @var{alignment})
6829
Like @code{ASM_OUTPUT_ALIGNED_COMMON} except that @var{decl} of the
6830
variable to be output, if there is one, or @code{NULL_TREE} if there
6831
is no corresponding variable.  If you define this macro, GCC will use it
6832
in place of both @code{ASM_OUTPUT_COMMON} and
6833
@code{ASM_OUTPUT_ALIGNED_COMMON}.  Define this macro when you need to see
6834
the variable's decl in order to chose what to output.
6835
@end defmac
6836
 
6837
@defmac ASM_OUTPUT_BSS (@var{stream}, @var{decl}, @var{name}, @var{size}, @var{rounded})
6838
A C statement (sans semicolon) to output to the stdio stream
6839
@var{stream} the assembler definition of uninitialized global @var{decl} named
6840
@var{name} whose size is @var{size} bytes.  The variable @var{rounded}
6841
is the size rounded up to whatever alignment the caller wants.
6842
 
6843
Try to use function @code{asm_output_bss} defined in @file{varasm.c} when
6844
defining this macro.  If unable, use the expression
6845
@code{assemble_name (@var{stream}, @var{name})} to output the name itself;
6846
before and after that, output the additional assembler syntax for defining
6847
the name, and a newline.
6848
 
6849
There are two ways of handling global BSS.  One is to define either
6850
this macro or its aligned counterpart, @code{ASM_OUTPUT_ALIGNED_BSS}.
6851
The other is to have @code{TARGET_ASM_SELECT_SECTION} return a
6852
switchable BSS section (@pxref{TARGET_HAVE_SWITCHABLE_BSS_SECTIONS}).
6853
You do not need to do both.
6854
 
6855
Some languages do not have @code{common} data, and require a
6856
non-common form of global BSS in order to handle uninitialized globals
6857
efficiently.  C++ is one example of this.  However, if the target does
6858
not support global BSS, the front end may choose to make globals
6859
common in order to save space in the object file.
6860
@end defmac
6861
 
6862
@defmac ASM_OUTPUT_ALIGNED_BSS (@var{stream}, @var{decl}, @var{name}, @var{size}, @var{alignment})
6863
Like @code{ASM_OUTPUT_BSS} except takes the required alignment as a
6864
separate, explicit argument.  If you define this macro, it is used in
6865
place of @code{ASM_OUTPUT_BSS}, and gives you more flexibility in
6866
handling the required alignment of the variable.  The alignment is specified
6867
as the number of bits.
6868
 
6869
Try to use function @code{asm_output_aligned_bss} defined in file
6870
@file{varasm.c} when defining this macro.
6871
@end defmac
6872
 
6873
@defmac ASM_OUTPUT_LOCAL (@var{stream}, @var{name}, @var{size}, @var{rounded})
6874
A C statement (sans semicolon) to output to the stdio stream
6875
@var{stream} the assembler definition of a local-common-label named
6876
@var{name} whose size is @var{size} bytes.  The variable @var{rounded}
6877
is the size rounded up to whatever alignment the caller wants.
6878
 
6879
Use the expression @code{assemble_name (@var{stream}, @var{name})} to
6880
output the name itself; before and after that, output the additional
6881
assembler syntax for defining the name, and a newline.
6882
 
6883
This macro controls how the assembler definitions of uninitialized
6884
static variables are output.
6885
@end defmac
6886
 
6887
@defmac ASM_OUTPUT_ALIGNED_LOCAL (@var{stream}, @var{name}, @var{size}, @var{alignment})
6888
Like @code{ASM_OUTPUT_LOCAL} except takes the required alignment as a
6889
separate, explicit argument.  If you define this macro, it is used in
6890
place of @code{ASM_OUTPUT_LOCAL}, and gives you more flexibility in
6891
handling the required alignment of the variable.  The alignment is specified
6892
as the number of bits.
6893
@end defmac
6894
 
6895
@defmac ASM_OUTPUT_ALIGNED_DECL_LOCAL (@var{stream}, @var{decl}, @var{name}, @var{size}, @var{alignment})
6896
Like @code{ASM_OUTPUT_ALIGNED_DECL} except that @var{decl} of the
6897
variable to be output, if there is one, or @code{NULL_TREE} if there
6898
is no corresponding variable.  If you define this macro, GCC will use it
6899
in place of both @code{ASM_OUTPUT_DECL} and
6900
@code{ASM_OUTPUT_ALIGNED_DECL}.  Define this macro when you need to see
6901
the variable's decl in order to chose what to output.
6902
@end defmac
6903
 
6904
@node Label Output
6905
@subsection Output and Generation of Labels
6906
 
6907
@c prevent bad page break with this line
6908
This is about outputting labels.
6909
 
6910
@findex assemble_name
6911
@defmac ASM_OUTPUT_LABEL (@var{stream}, @var{name})
6912
A C statement (sans semicolon) to output to the stdio stream
6913
@var{stream} the assembler definition of a label named @var{name}.
6914
Use the expression @code{assemble_name (@var{stream}, @var{name})} to
6915
output the name itself; before and after that, output the additional
6916
assembler syntax for defining the name, and a newline.  A default
6917
definition of this macro is provided which is correct for most systems.
6918
@end defmac
6919
 
6920
@findex assemble_name_raw
6921
@defmac ASM_OUTPUT_INTERNAL_LABEL (@var{stream}, @var{name})
6922
Identical to @code{ASM_OUTPUT_LABEL}, except that @var{name} is known
6923
to refer to a compiler-generated label.  The default definition uses
6924
@code{assemble_name_raw}, which is like @code{assemble_name} except
6925
that it is more efficient.
6926
@end defmac
6927
 
6928
@defmac SIZE_ASM_OP
6929
A C string containing the appropriate assembler directive to specify the
6930
size of a symbol, without any arguments.  On systems that use ELF, the
6931
default (in @file{config/elfos.h}) is @samp{"\t.size\t"}; on other
6932
systems, the default is not to define this macro.
6933
 
6934
Define this macro only if it is correct to use the default definitions
6935
of @code{ASM_OUTPUT_SIZE_DIRECTIVE} and @code{ASM_OUTPUT_MEASURED_SIZE}
6936
for your system.  If you need your own custom definitions of those
6937
macros, or if you do not need explicit symbol sizes at all, do not
6938
define this macro.
6939
@end defmac
6940
 
6941
@defmac ASM_OUTPUT_SIZE_DIRECTIVE (@var{stream}, @var{name}, @var{size})
6942
A C statement (sans semicolon) to output to the stdio stream
6943
@var{stream} a directive telling the assembler that the size of the
6944
symbol @var{name} is @var{size}.  @var{size} is a @code{HOST_WIDE_INT}.
6945
If you define @code{SIZE_ASM_OP}, a default definition of this macro is
6946
provided.
6947
@end defmac
6948
 
6949
@defmac ASM_OUTPUT_MEASURED_SIZE (@var{stream}, @var{name})
6950
A C statement (sans semicolon) to output to the stdio stream
6951
@var{stream} a directive telling the assembler to calculate the size of
6952
the symbol @var{name} by subtracting its address from the current
6953
address.
6954
 
6955
If you define @code{SIZE_ASM_OP}, a default definition of this macro is
6956
provided.  The default assumes that the assembler recognizes a special
6957
@samp{.} symbol as referring to the current address, and can calculate
6958
the difference between this and another symbol.  If your assembler does
6959
not recognize @samp{.} or cannot do calculations with it, you will need
6960
to redefine @code{ASM_OUTPUT_MEASURED_SIZE} to use some other technique.
6961
@end defmac
6962
 
6963
@defmac TYPE_ASM_OP
6964
A C string containing the appropriate assembler directive to specify the
6965
type of a symbol, without any arguments.  On systems that use ELF, the
6966
default (in @file{config/elfos.h}) is @samp{"\t.type\t"}; on other
6967
systems, the default is not to define this macro.
6968
 
6969
Define this macro only if it is correct to use the default definition of
6970
@code{ASM_OUTPUT_TYPE_DIRECTIVE} for your system.  If you need your own
6971
custom definition of this macro, or if you do not need explicit symbol
6972
types at all, do not define this macro.
6973
@end defmac
6974
 
6975
@defmac TYPE_OPERAND_FMT
6976
A C string which specifies (using @code{printf} syntax) the format of
6977
the second operand to @code{TYPE_ASM_OP}.  On systems that use ELF, the
6978
default (in @file{config/elfos.h}) is @samp{"@@%s"}; on other systems,
6979
the default is not to define this macro.
6980
 
6981
Define this macro only if it is correct to use the default definition of
6982
@code{ASM_OUTPUT_TYPE_DIRECTIVE} for your system.  If you need your own
6983
custom definition of this macro, or if you do not need explicit symbol
6984
types at all, do not define this macro.
6985
@end defmac
6986
 
6987
@defmac ASM_OUTPUT_TYPE_DIRECTIVE (@var{stream}, @var{type})
6988
A C statement (sans semicolon) to output to the stdio stream
6989
@var{stream} a directive telling the assembler that the type of the
6990
symbol @var{name} is @var{type}.  @var{type} is a C string; currently,
6991
that string is always either @samp{"function"} or @samp{"object"}, but
6992
you should not count on this.
6993
 
6994
If you define @code{TYPE_ASM_OP} and @code{TYPE_OPERAND_FMT}, a default
6995
definition of this macro is provided.
6996
@end defmac
6997
 
6998
@defmac ASM_DECLARE_FUNCTION_NAME (@var{stream}, @var{name}, @var{decl})
6999
A C statement (sans semicolon) to output to the stdio stream
7000
@var{stream} any text necessary for declaring the name @var{name} of a
7001
function which is being defined.  This macro is responsible for
7002
outputting the label definition (perhaps using
7003
@code{ASM_OUTPUT_LABEL}).  The argument @var{decl} is the
7004
@code{FUNCTION_DECL} tree node representing the function.
7005
 
7006
If this macro is not defined, then the function name is defined in the
7007
usual manner as a label (by means of @code{ASM_OUTPUT_LABEL}).
7008
 
7009
You may wish to use @code{ASM_OUTPUT_TYPE_DIRECTIVE} in the definition
7010
of this macro.
7011
@end defmac
7012
 
7013
@defmac ASM_DECLARE_FUNCTION_SIZE (@var{stream}, @var{name}, @var{decl})
7014
A C statement (sans semicolon) to output to the stdio stream
7015
@var{stream} any text necessary for declaring the size of a function
7016
which is being defined.  The argument @var{name} is the name of the
7017
function.  The argument @var{decl} is the @code{FUNCTION_DECL} tree node
7018
representing the function.
7019
 
7020
If this macro is not defined, then the function size is not defined.
7021
 
7022
You may wish to use @code{ASM_OUTPUT_MEASURED_SIZE} in the definition
7023
of this macro.
7024
@end defmac
7025
 
7026
@defmac ASM_DECLARE_OBJECT_NAME (@var{stream}, @var{name}, @var{decl})
7027
A C statement (sans semicolon) to output to the stdio stream
7028
@var{stream} any text necessary for declaring the name @var{name} of an
7029
initialized variable which is being defined.  This macro must output the
7030
label definition (perhaps using @code{ASM_OUTPUT_LABEL}).  The argument
7031
@var{decl} is the @code{VAR_DECL} tree node representing the variable.
7032
 
7033
If this macro is not defined, then the variable name is defined in the
7034
usual manner as a label (by means of @code{ASM_OUTPUT_LABEL}).
7035
 
7036
You may wish to use @code{ASM_OUTPUT_TYPE_DIRECTIVE} and/or
7037
@code{ASM_OUTPUT_SIZE_DIRECTIVE} in the definition of this macro.
7038
@end defmac
7039
 
7040
@defmac ASM_DECLARE_CONSTANT_NAME (@var{stream}, @var{name}, @var{exp}, @var{size})
7041
A C statement (sans semicolon) to output to the stdio stream
7042
@var{stream} any text necessary for declaring the name @var{name} of a
7043
constant which is being defined.  This macro is responsible for
7044
outputting the label definition (perhaps using
7045
@code{ASM_OUTPUT_LABEL}).  The argument @var{exp} is the
7046
value of the constant, and @var{size} is the size of the constant
7047
in bytes.  @var{name} will be an internal label.
7048
 
7049
If this macro is not defined, then the @var{name} is defined in the
7050
usual manner as a label (by means of @code{ASM_OUTPUT_LABEL}).
7051
 
7052
You may wish to use @code{ASM_OUTPUT_TYPE_DIRECTIVE} in the definition
7053
of this macro.
7054
@end defmac
7055
 
7056
@defmac ASM_DECLARE_REGISTER_GLOBAL (@var{stream}, @var{decl}, @var{regno}, @var{name})
7057
A C statement (sans semicolon) to output to the stdio stream
7058
@var{stream} any text necessary for claiming a register @var{regno}
7059
for a global variable @var{decl} with name @var{name}.
7060
 
7061
If you don't define this macro, that is equivalent to defining it to do
7062
nothing.
7063
@end defmac
7064
 
7065
@defmac ASM_FINISH_DECLARE_OBJECT (@var{stream}, @var{decl}, @var{toplevel}, @var{atend})
7066
A C statement (sans semicolon) to finish up declaring a variable name
7067
once the compiler has processed its initializer fully and thus has had a
7068
chance to determine the size of an array when controlled by an
7069
initializer.  This is used on systems where it's necessary to declare
7070
something about the size of the object.
7071
 
7072
If you don't define this macro, that is equivalent to defining it to do
7073
nothing.
7074
 
7075
You may wish to use @code{ASM_OUTPUT_SIZE_DIRECTIVE} and/or
7076
@code{ASM_OUTPUT_MEASURED_SIZE} in the definition of this macro.
7077
@end defmac
7078
 
7079
@deftypefn {Target Hook} void TARGET_ASM_GLOBALIZE_LABEL (FILE *@var{stream}, const char *@var{name})
7080
This target hook is a function to output to the stdio stream
7081
@var{stream} some commands that will make the label @var{name} global;
7082
that is, available for reference from other files.
7083
 
7084
The default implementation relies on a proper definition of
7085
@code{GLOBAL_ASM_OP}.
7086
@end deftypefn
7087
 
7088
@defmac ASM_WEAKEN_LABEL (@var{stream}, @var{name})
7089
A C statement (sans semicolon) to output to the stdio stream
7090
@var{stream} some commands that will make the label @var{name} weak;
7091
that is, available for reference from other files but only used if
7092
no other definition is available.  Use the expression
7093
@code{assemble_name (@var{stream}, @var{name})} to output the name
7094
itself; before and after that, output the additional assembler syntax
7095
for making that name weak, and a newline.
7096
 
7097
If you don't define this macro or @code{ASM_WEAKEN_DECL}, GCC will not
7098
support weak symbols and you should not define the @code{SUPPORTS_WEAK}
7099
macro.
7100
@end defmac
7101
 
7102
@defmac ASM_WEAKEN_DECL (@var{stream}, @var{decl}, @var{name}, @var{value})
7103
Combines (and replaces) the function of @code{ASM_WEAKEN_LABEL} and
7104
@code{ASM_OUTPUT_WEAK_ALIAS}, allowing access to the associated function
7105
or variable decl.  If @var{value} is not @code{NULL}, this C statement
7106
should output to the stdio stream @var{stream} assembler code which
7107
defines (equates) the weak symbol @var{name} to have the value
7108
@var{value}.  If @var{value} is @code{NULL}, it should output commands
7109
to make @var{name} weak.
7110
@end defmac
7111
 
7112
@defmac ASM_OUTPUT_WEAKREF (@var{stream}, @var{decl}, @var{name}, @var{value})
7113
Outputs a directive that enables @var{name} to be used to refer to
7114
symbol @var{value} with weak-symbol semantics.  @code{decl} is the
7115
declaration of @code{name}.
7116
@end defmac
7117
 
7118
@defmac SUPPORTS_WEAK
7119
A C expression which evaluates to true if the target supports weak symbols.
7120
 
7121
If you don't define this macro, @file{defaults.h} provides a default
7122
definition.  If either @code{ASM_WEAKEN_LABEL} or @code{ASM_WEAKEN_DECL}
7123
is defined, the default definition is @samp{1}; otherwise, it is
7124
@samp{0}.  Define this macro if you want to control weak symbol support
7125
with a compiler flag such as @option{-melf}.
7126
@end defmac
7127
 
7128
@defmac MAKE_DECL_ONE_ONLY (@var{decl})
7129
A C statement (sans semicolon) to mark @var{decl} to be emitted as a
7130
public symbol such that extra copies in multiple translation units will
7131
be discarded by the linker.  Define this macro if your object file
7132
format provides support for this concept, such as the @samp{COMDAT}
7133
section flags in the Microsoft Windows PE/COFF format, and this support
7134
requires changes to @var{decl}, such as putting it in a separate section.
7135
@end defmac
7136
 
7137
@defmac SUPPORTS_ONE_ONLY
7138
A C expression which evaluates to true if the target supports one-only
7139
semantics.
7140
 
7141
If you don't define this macro, @file{varasm.c} provides a default
7142
definition.  If @code{MAKE_DECL_ONE_ONLY} is defined, the default
7143
definition is @samp{1}; otherwise, it is @samp{0}.  Define this macro if
7144
you want to control one-only symbol support with a compiler flag, or if
7145
setting the @code{DECL_ONE_ONLY} flag is enough to mark a declaration to
7146
be emitted as one-only.
7147
@end defmac
7148
 
7149
@deftypefn {Target Hook} void TARGET_ASM_ASSEMBLE_VISIBILITY (tree @var{decl}, const char *@var{visibility})
7150
This target hook is a function to output to @var{asm_out_file} some
7151
commands that will make the symbol(s) associated with @var{decl} have
7152
hidden, protected or internal visibility as specified by @var{visibility}.
7153
@end deftypefn
7154
 
7155
@defmac TARGET_WEAK_NOT_IN_ARCHIVE_TOC
7156
A C expression that evaluates to true if the target's linker expects
7157
that weak symbols do not appear in a static archive's table of contents.
7158
The default is @code{0}.
7159
 
7160
Leaving weak symbols out of an archive's table of contents means that,
7161
if a symbol will only have a definition in one translation unit and
7162
will have undefined references from other translation units, that
7163
symbol should not be weak.  Defining this macro to be nonzero will
7164
thus have the effect that certain symbols that would normally be weak
7165
(explicit template instantiations, and vtables for polymorphic classes
7166
with noninline key methods) will instead be nonweak.
7167
 
7168
The C++ ABI requires this macro to be zero.  Define this macro for
7169
targets where full C++ ABI compliance is impossible and where linker
7170
restrictions require weak symbols to be left out of a static archive's
7171
table of contents.
7172
@end defmac
7173
 
7174
@defmac ASM_OUTPUT_EXTERNAL (@var{stream}, @var{decl}, @var{name})
7175
A C statement (sans semicolon) to output to the stdio stream
7176
@var{stream} any text necessary for declaring the name of an external
7177
symbol named @var{name} which is referenced in this compilation but
7178
not defined.  The value of @var{decl} is the tree node for the
7179
declaration.
7180
 
7181
This macro need not be defined if it does not need to output anything.
7182
The GNU assembler and most Unix assemblers don't require anything.
7183
@end defmac
7184
 
7185
@deftypefn {Target Hook} void TARGET_ASM_EXTERNAL_LIBCALL (rtx @var{symref})
7186
This target hook is a function to output to @var{asm_out_file} an assembler
7187
pseudo-op to declare a library function name external.  The name of the
7188
library function is given by @var{symref}, which is a @code{symbol_ref}.
7189
@end deftypefn
7190
 
7191
@deftypefn {Target Hook} void TARGET_ASM_MARK_DECL_PRESERVED (tree @var{decl})
7192
This target hook is a function to output to @var{asm_out_file} an assembler
7193
directive to annotate used symbol.  Darwin target use .no_dead_code_strip
7194
directive.
7195
@end deftypefn
7196
 
7197
@defmac ASM_OUTPUT_LABELREF (@var{stream}, @var{name})
7198
A C statement (sans semicolon) to output to the stdio stream
7199
@var{stream} a reference in assembler syntax to a label named
7200
@var{name}.  This should add @samp{_} to the front of the name, if that
7201
is customary on your operating system, as it is in most Berkeley Unix
7202
systems.  This macro is used in @code{assemble_name}.
7203
@end defmac
7204
 
7205
@defmac ASM_OUTPUT_SYMBOL_REF (@var{stream}, @var{sym})
7206
A C statement (sans semicolon) to output a reference to
7207
@code{SYMBOL_REF} @var{sym}.  If not defined, @code{assemble_name}
7208
will be used to output the name of the symbol.  This macro may be used
7209
to modify the way a symbol is referenced depending on information
7210
encoded by @code{TARGET_ENCODE_SECTION_INFO}.
7211
@end defmac
7212
 
7213
@defmac ASM_OUTPUT_LABEL_REF (@var{stream}, @var{buf})
7214
A C statement (sans semicolon) to output a reference to @var{buf}, the
7215
result of @code{ASM_GENERATE_INTERNAL_LABEL}.  If not defined,
7216
@code{assemble_name} will be used to output the name of the symbol.
7217
This macro is not used by @code{output_asm_label}, or the @code{%l}
7218
specifier that calls it; the intention is that this macro should be set
7219
when it is necessary to output a label differently when its address is
7220
being taken.
7221
@end defmac
7222
 
7223
@deftypefn {Target Hook} void TARGET_ASM_INTERNAL_LABEL (FILE *@var{stream}, const char *@var{prefix}, unsigned long @var{labelno})
7224
A function to output to the stdio stream @var{stream} a label whose
7225
name is made from the string @var{prefix} and the number @var{labelno}.
7226
 
7227
It is absolutely essential that these labels be distinct from the labels
7228
used for user-level functions and variables.  Otherwise, certain programs
7229
will have name conflicts with internal labels.
7230
 
7231
It is desirable to exclude internal labels from the symbol table of the
7232
object file.  Most assemblers have a naming convention for labels that
7233
should be excluded; on many systems, the letter @samp{L} at the
7234
beginning of a label has this effect.  You should find out what
7235
convention your system uses, and follow it.
7236
 
7237
The default version of this function utilizes @code{ASM_GENERATE_INTERNAL_LABEL}.
7238
@end deftypefn
7239
 
7240
@defmac ASM_OUTPUT_DEBUG_LABEL (@var{stream}, @var{prefix}, @var{num})
7241
A C statement to output to the stdio stream @var{stream} a debug info
7242
label whose name is made from the string @var{prefix} and the number
7243
@var{num}.  This is useful for VLIW targets, where debug info labels
7244
may need to be treated differently than branch target labels.  On some
7245
systems, branch target labels must be at the beginning of instruction
7246
bundles, but debug info labels can occur in the middle of instruction
7247
bundles.
7248
 
7249
If this macro is not defined, then @code{(*targetm.asm_out.internal_label)} will be
7250
used.
7251
@end defmac
7252
 
7253
@defmac ASM_GENERATE_INTERNAL_LABEL (@var{string}, @var{prefix}, @var{num})
7254
A C statement to store into the string @var{string} a label whose name
7255
is made from the string @var{prefix} and the number @var{num}.
7256
 
7257
This string, when output subsequently by @code{assemble_name}, should
7258
produce the output that @code{(*targetm.asm_out.internal_label)} would produce
7259
with the same @var{prefix} and @var{num}.
7260
 
7261
If the string begins with @samp{*}, then @code{assemble_name} will
7262
output the rest of the string unchanged.  It is often convenient for
7263
@code{ASM_GENERATE_INTERNAL_LABEL} to use @samp{*} in this way.  If the
7264
string doesn't start with @samp{*}, then @code{ASM_OUTPUT_LABELREF} gets
7265
to output the string, and may change it.  (Of course,
7266
@code{ASM_OUTPUT_LABELREF} is also part of your machine description, so
7267
you should know what it does on your machine.)
7268
@end defmac
7269
 
7270
@defmac ASM_FORMAT_PRIVATE_NAME (@var{outvar}, @var{name}, @var{number})
7271
A C expression to assign to @var{outvar} (which is a variable of type
7272
@code{char *}) a newly allocated string made from the string
7273
@var{name} and the number @var{number}, with some suitable punctuation
7274
added.  Use @code{alloca} to get space for the string.
7275
 
7276
The string will be used as an argument to @code{ASM_OUTPUT_LABELREF} to
7277
produce an assembler label for an internal static variable whose name is
7278
@var{name}.  Therefore, the string must be such as to result in valid
7279
assembler code.  The argument @var{number} is different each time this
7280
macro is executed; it prevents conflicts between similarly-named
7281
internal static variables in different scopes.
7282
 
7283
Ideally this string should not be a valid C identifier, to prevent any
7284
conflict with the user's own symbols.  Most assemblers allow periods
7285
or percent signs in assembler symbols; putting at least one of these
7286
between the name and the number will suffice.
7287
 
7288
If this macro is not defined, a default definition will be provided
7289
which is correct for most systems.
7290
@end defmac
7291
 
7292
@defmac ASM_OUTPUT_DEF (@var{stream}, @var{name}, @var{value})
7293
A C statement to output to the stdio stream @var{stream} assembler code
7294
which defines (equates) the symbol @var{name} to have the value @var{value}.
7295
 
7296
@findex SET_ASM_OP
7297
If @code{SET_ASM_OP} is defined, a default definition is provided which is
7298
correct for most systems.
7299
@end defmac
7300
 
7301
@defmac ASM_OUTPUT_DEF_FROM_DECLS (@var{stream}, @var{decl_of_name}, @var{decl_of_value})
7302
A C statement to output to the stdio stream @var{stream} assembler code
7303
which defines (equates) the symbol whose tree node is @var{decl_of_name}
7304
to have the value of the tree node @var{decl_of_value}.  This macro will
7305
be used in preference to @samp{ASM_OUTPUT_DEF} if it is defined and if
7306
the tree nodes are available.
7307
 
7308
@findex SET_ASM_OP
7309
If @code{SET_ASM_OP} is defined, a default definition is provided which is
7310
correct for most systems.
7311
@end defmac
7312
 
7313
@defmac TARGET_DEFERRED_OUTPUT_DEFS (@var{decl_of_name}, @var{decl_of_value})
7314
A C statement that evaluates to true if the assembler code which defines
7315
(equates) the symbol whose tree node is @var{decl_of_name} to have the value
7316
of the tree node @var{decl_of_value} should be emitted near the end of the
7317
current compilation unit.  The default is to not defer output of defines.
7318
This macro affects defines output by @samp{ASM_OUTPUT_DEF} and
7319
@samp{ASM_OUTPUT_DEF_FROM_DECLS}.
7320
@end defmac
7321
 
7322
@defmac ASM_OUTPUT_WEAK_ALIAS (@var{stream}, @var{name}, @var{value})
7323
A C statement to output to the stdio stream @var{stream} assembler code
7324
which defines (equates) the weak symbol @var{name} to have the value
7325
@var{value}.  If @var{value} is @code{NULL}, it defines @var{name} as
7326
an undefined weak symbol.
7327
 
7328
Define this macro if the target only supports weak aliases; define
7329
@code{ASM_OUTPUT_DEF} instead if possible.
7330
@end defmac
7331
 
7332
@defmac OBJC_GEN_METHOD_LABEL (@var{buf}, @var{is_inst}, @var{class_name}, @var{cat_name}, @var{sel_name})
7333
Define this macro to override the default assembler names used for
7334
Objective-C methods.
7335
 
7336
The default name is a unique method number followed by the name of the
7337
class (e.g.@: @samp{_1_Foo}).  For methods in categories, the name of
7338
the category is also included in the assembler name (e.g.@:
7339
@samp{_1_Foo_Bar}).
7340
 
7341
These names are safe on most systems, but make debugging difficult since
7342
the method's selector is not present in the name.  Therefore, particular
7343
systems define other ways of computing names.
7344
 
7345
@var{buf} is an expression of type @code{char *} which gives you a
7346
buffer in which to store the name; its length is as long as
7347
@var{class_name}, @var{cat_name} and @var{sel_name} put together, plus
7348
50 characters extra.
7349
 
7350
The argument @var{is_inst} specifies whether the method is an instance
7351
method or a class method; @var{class_name} is the name of the class;
7352
@var{cat_name} is the name of the category (or @code{NULL} if the method is not
7353
in a category); and @var{sel_name} is the name of the selector.
7354
 
7355
On systems where the assembler can handle quoted names, you can use this
7356
macro to provide more human-readable names.
7357
@end defmac
7358
 
7359
@defmac ASM_DECLARE_CLASS_REFERENCE (@var{stream}, @var{name})
7360
A C statement (sans semicolon) to output to the stdio stream
7361
@var{stream} commands to declare that the label @var{name} is an
7362
Objective-C class reference.  This is only needed for targets whose
7363
linkers have special support for NeXT-style runtimes.
7364
@end defmac
7365
 
7366
@defmac ASM_DECLARE_UNRESOLVED_REFERENCE (@var{stream}, @var{name})
7367
A C statement (sans semicolon) to output to the stdio stream
7368
@var{stream} commands to declare that the label @var{name} is an
7369
unresolved Objective-C class reference.  This is only needed for targets
7370
whose linkers have special support for NeXT-style runtimes.
7371
@end defmac
7372
 
7373
@node Initialization
7374
@subsection How Initialization Functions Are Handled
7375
@cindex initialization routines
7376
@cindex termination routines
7377
@cindex constructors, output of
7378
@cindex destructors, output of
7379
 
7380
The compiled code for certain languages includes @dfn{constructors}
7381
(also called @dfn{initialization routines})---functions to initialize
7382
data in the program when the program is started.  These functions need
7383
to be called before the program is ``started''---that is to say, before
7384
@code{main} is called.
7385
 
7386
Compiling some languages generates @dfn{destructors} (also called
7387
@dfn{termination routines}) that should be called when the program
7388
terminates.
7389
 
7390
To make the initialization and termination functions work, the compiler
7391
must output something in the assembler code to cause those functions to
7392
be called at the appropriate time.  When you port the compiler to a new
7393
system, you need to specify how to do this.
7394
 
7395
There are two major ways that GCC currently supports the execution of
7396
initialization and termination functions.  Each way has two variants.
7397
Much of the structure is common to all four variations.
7398
 
7399
@findex __CTOR_LIST__
7400
@findex __DTOR_LIST__
7401
The linker must build two lists of these functions---a list of
7402
initialization functions, called @code{__CTOR_LIST__}, and a list of
7403
termination functions, called @code{__DTOR_LIST__}.
7404
 
7405
Each list always begins with an ignored function pointer (which may hold
7406
0, @minus{}1, or a count of the function pointers after it, depending on
7407
the environment).  This is followed by a series of zero or more function
7408
pointers to constructors (or destructors), followed by a function
7409
pointer containing zero.
7410
 
7411
Depending on the operating system and its executable file format, either
7412
@file{crtstuff.c} or @file{libgcc2.c} traverses these lists at startup
7413
time and exit time.  Constructors are called in reverse order of the
7414
list; destructors in forward order.
7415
 
7416
The best way to handle static constructors works only for object file
7417
formats which provide arbitrarily-named sections.  A section is set
7418
aside for a list of constructors, and another for a list of destructors.
7419
Traditionally these are called @samp{.ctors} and @samp{.dtors}.  Each
7420
object file that defines an initialization function also puts a word in
7421
the constructor section to point to that function.  The linker
7422
accumulates all these words into one contiguous @samp{.ctors} section.
7423
Termination functions are handled similarly.
7424
 
7425
This method will be chosen as the default by @file{target-def.h} if
7426
@code{TARGET_ASM_NAMED_SECTION} is defined.  A target that does not
7427
support arbitrary sections, but does support special designated
7428
constructor and destructor sections may define @code{CTORS_SECTION_ASM_OP}
7429
and @code{DTORS_SECTION_ASM_OP} to achieve the same effect.
7430
 
7431
When arbitrary sections are available, there are two variants, depending
7432
upon how the code in @file{crtstuff.c} is called.  On systems that
7433
support a @dfn{.init} section which is executed at program startup,
7434
parts of @file{crtstuff.c} are compiled into that section.  The
7435
program is linked by the @command{gcc} driver like this:
7436
 
7437
@smallexample
7438
ld -o @var{output_file} crti.o crtbegin.o @dots{} -lgcc crtend.o crtn.o
7439
@end smallexample
7440
 
7441
The prologue of a function (@code{__init}) appears in the @code{.init}
7442
section of @file{crti.o}; the epilogue appears in @file{crtn.o}.  Likewise
7443
for the function @code{__fini} in the @dfn{.fini} section.  Normally these
7444
files are provided by the operating system or by the GNU C library, but
7445
are provided by GCC for a few targets.
7446
 
7447
The objects @file{crtbegin.o} and @file{crtend.o} are (for most targets)
7448
compiled from @file{crtstuff.c}.  They contain, among other things, code
7449
fragments within the @code{.init} and @code{.fini} sections that branch
7450
to routines in the @code{.text} section.  The linker will pull all parts
7451
of a section together, which results in a complete @code{__init} function
7452
that invokes the routines we need at startup.
7453
 
7454
To use this variant, you must define the @code{INIT_SECTION_ASM_OP}
7455
macro properly.
7456
 
7457
If no init section is available, when GCC compiles any function called
7458
@code{main} (or more accurately, any function designated as a program
7459
entry point by the language front end calling @code{expand_main_function}),
7460
it inserts a procedure call to @code{__main} as the first executable code
7461
after the function prologue.  The @code{__main} function is defined
7462
in @file{libgcc2.c} and runs the global constructors.
7463
 
7464
In file formats that don't support arbitrary sections, there are again
7465
two variants.  In the simplest variant, the GNU linker (GNU @code{ld})
7466
and an `a.out' format must be used.  In this case,
7467
@code{TARGET_ASM_CONSTRUCTOR} is defined to produce a @code{.stabs}
7468
entry of type @samp{N_SETT}, referencing the name @code{__CTOR_LIST__},
7469
and with the address of the void function containing the initialization
7470
code as its value.  The GNU linker recognizes this as a request to add
7471
the value to a @dfn{set}; the values are accumulated, and are eventually
7472
placed in the executable as a vector in the format described above, with
7473
a leading (ignored) count and a trailing zero element.
7474
@code{TARGET_ASM_DESTRUCTOR} is handled similarly.  Since no init
7475
section is available, the absence of @code{INIT_SECTION_ASM_OP} causes
7476
the compilation of @code{main} to call @code{__main} as above, starting
7477
the initialization process.
7478
 
7479
The last variant uses neither arbitrary sections nor the GNU linker.
7480
This is preferable when you want to do dynamic linking and when using
7481
file formats which the GNU linker does not support, such as `ECOFF'@.  In
7482
this case, @code{TARGET_HAVE_CTORS_DTORS} is false, initialization and
7483
termination functions are recognized simply by their names.  This requires
7484
an extra program in the linkage step, called @command{collect2}.  This program
7485
pretends to be the linker, for use with GCC; it does its job by running
7486
the ordinary linker, but also arranges to include the vectors of
7487
initialization and termination functions.  These functions are called
7488
via @code{__main} as described above.  In order to use this method,
7489
@code{use_collect2} must be defined in the target in @file{config.gcc}.
7490
 
7491
@ifinfo
7492
The following section describes the specific macros that control and
7493
customize the handling of initialization and termination functions.
7494
@end ifinfo
7495
 
7496
@node Macros for Initialization
7497
@subsection Macros Controlling Initialization Routines
7498
 
7499
Here are the macros that control how the compiler handles initialization
7500
and termination functions:
7501
 
7502
@defmac INIT_SECTION_ASM_OP
7503
If defined, a C string constant, including spacing, for the assembler
7504
operation to identify the following data as initialization code.  If not
7505
defined, GCC will assume such a section does not exist.  When you are
7506
using special sections for initialization and termination functions, this
7507
macro also controls how @file{crtstuff.c} and @file{libgcc2.c} arrange to
7508
run the initialization functions.
7509
@end defmac
7510
 
7511
@defmac HAS_INIT_SECTION
7512
If defined, @code{main} will not call @code{__main} as described above.
7513
This macro should be defined for systems that control start-up code
7514
on a symbol-by-symbol basis, such as OSF/1, and should not
7515
be defined explicitly for systems that support @code{INIT_SECTION_ASM_OP}.
7516
@end defmac
7517
 
7518
@defmac LD_INIT_SWITCH
7519
If defined, a C string constant for a switch that tells the linker that
7520
the following symbol is an initialization routine.
7521
@end defmac
7522
 
7523
@defmac LD_FINI_SWITCH
7524
If defined, a C string constant for a switch that tells the linker that
7525
the following symbol is a finalization routine.
7526
@end defmac
7527
 
7528
@defmac COLLECT_SHARED_INIT_FUNC (@var{stream}, @var{func})
7529
If defined, a C statement that will write a function that can be
7530
automatically called when a shared library is loaded.  The function
7531
should call @var{func}, which takes no arguments.  If not defined, and
7532
the object format requires an explicit initialization function, then a
7533
function called @code{_GLOBAL__DI} will be generated.
7534
 
7535
This function and the following one are used by collect2 when linking a
7536
shared library that needs constructors or destructors, or has DWARF2
7537
exception tables embedded in the code.
7538
@end defmac
7539
 
7540
@defmac COLLECT_SHARED_FINI_FUNC (@var{stream}, @var{func})
7541
If defined, a C statement that will write a function that can be
7542
automatically called when a shared library is unloaded.  The function
7543
should call @var{func}, which takes no arguments.  If not defined, and
7544
the object format requires an explicit finalization function, then a
7545
function called @code{_GLOBAL__DD} will be generated.
7546
@end defmac
7547
 
7548
@defmac INVOKE__main
7549
If defined, @code{main} will call @code{__main} despite the presence of
7550
@code{INIT_SECTION_ASM_OP}.  This macro should be defined for systems
7551
where the init section is not actually run automatically, but is still
7552
useful for collecting the lists of constructors and destructors.
7553
@end defmac
7554
 
7555
@defmac SUPPORTS_INIT_PRIORITY
7556
If nonzero, the C++ @code{init_priority} attribute is supported and the
7557
compiler should emit instructions to control the order of initialization
7558
of objects.  If zero, the compiler will issue an error message upon
7559
encountering an @code{init_priority} attribute.
7560
@end defmac
7561
 
7562
@deftypefn {Target Hook} bool TARGET_HAVE_CTORS_DTORS
7563
This value is true if the target supports some ``native'' method of
7564
collecting constructors and destructors to be run at startup and exit.
7565
It is false if we must use @command{collect2}.
7566
@end deftypefn
7567
 
7568
@deftypefn {Target Hook} void TARGET_ASM_CONSTRUCTOR (rtx @var{symbol}, int @var{priority})
7569
If defined, a function that outputs assembler code to arrange to call
7570
the function referenced by @var{symbol} at initialization time.
7571
 
7572
Assume that @var{symbol} is a @code{SYMBOL_REF} for a function taking
7573
no arguments and with no return value.  If the target supports initialization
7574
priorities, @var{priority} is a value between 0 and @code{MAX_INIT_PRIORITY};
7575
otherwise it must be @code{DEFAULT_INIT_PRIORITY}.
7576
 
7577
If this macro is not defined by the target, a suitable default will
7578
be chosen if (1) the target supports arbitrary section names, (2) the
7579
target defines @code{CTORS_SECTION_ASM_OP}, or (3) @code{USE_COLLECT2}
7580
is not defined.
7581
@end deftypefn
7582
 
7583
@deftypefn {Target Hook} void TARGET_ASM_DESTRUCTOR (rtx @var{symbol}, int @var{priority})
7584
This is like @code{TARGET_ASM_CONSTRUCTOR} but used for termination
7585
functions rather than initialization functions.
7586
@end deftypefn
7587
 
7588
If @code{TARGET_HAVE_CTORS_DTORS} is true, the initialization routine
7589
generated for the generated object file will have static linkage.
7590
 
7591
If your system uses @command{collect2} as the means of processing
7592
constructors, then that program normally uses @command{nm} to scan
7593
an object file for constructor functions to be called.
7594
 
7595
On certain kinds of systems, you can define this macro to make
7596
@command{collect2} work faster (and, in some cases, make it work at all):
7597
 
7598
@defmac OBJECT_FORMAT_COFF
7599
Define this macro if the system uses COFF (Common Object File Format)
7600
object files, so that @command{collect2} can assume this format and scan
7601
object files directly for dynamic constructor/destructor functions.
7602
 
7603
This macro is effective only in a native compiler; @command{collect2} as
7604
part of a cross compiler always uses @command{nm} for the target machine.
7605
@end defmac
7606
 
7607
@defmac REAL_NM_FILE_NAME
7608
Define this macro as a C string constant containing the file name to use
7609
to execute @command{nm}.  The default is to search the path normally for
7610
@command{nm}.
7611
 
7612
If your system supports shared libraries and has a program to list the
7613
dynamic dependencies of a given library or executable, you can define
7614
these macros to enable support for running initialization and
7615
termination functions in shared libraries:
7616
@end defmac
7617
 
7618
@defmac LDD_SUFFIX
7619
Define this macro to a C string constant containing the name of the program
7620
which lists dynamic dependencies, like @command{"ldd"} under SunOS 4.
7621
@end defmac
7622
 
7623
@defmac PARSE_LDD_OUTPUT (@var{ptr})
7624
Define this macro to be C code that extracts filenames from the output
7625
of the program denoted by @code{LDD_SUFFIX}.  @var{ptr} is a variable
7626
of type @code{char *} that points to the beginning of a line of output
7627
from @code{LDD_SUFFIX}.  If the line lists a dynamic dependency, the
7628
code must advance @var{ptr} to the beginning of the filename on that
7629
line.  Otherwise, it must set @var{ptr} to @code{NULL}.
7630
@end defmac
7631
 
7632
@node Instruction Output
7633
@subsection Output of Assembler Instructions
7634
 
7635
@c prevent bad page break with this line
7636
This describes assembler instruction output.
7637
 
7638
@defmac REGISTER_NAMES
7639
A C initializer containing the assembler's names for the machine
7640
registers, each one as a C string constant.  This is what translates
7641
register numbers in the compiler into assembler language.
7642
@end defmac
7643
 
7644
@defmac ADDITIONAL_REGISTER_NAMES
7645
If defined, a C initializer for an array of structures containing a name
7646
and a register number.  This macro defines additional names for hard
7647
registers, thus allowing the @code{asm} option in declarations to refer
7648
to registers using alternate names.
7649
@end defmac
7650
 
7651
@defmac ASM_OUTPUT_OPCODE (@var{stream}, @var{ptr})
7652
Define this macro if you are using an unusual assembler that
7653
requires different names for the machine instructions.
7654
 
7655
The definition is a C statement or statements which output an
7656
assembler instruction opcode to the stdio stream @var{stream}.  The
7657
macro-operand @var{ptr} is a variable of type @code{char *} which
7658
points to the opcode name in its ``internal'' form---the form that is
7659
written in the machine description.  The definition should output the
7660
opcode name to @var{stream}, performing any translation you desire, and
7661
increment the variable @var{ptr} to point at the end of the opcode
7662
so that it will not be output twice.
7663
 
7664
In fact, your macro definition may process less than the entire opcode
7665
name, or more than the opcode name; but if you want to process text
7666
that includes @samp{%}-sequences to substitute operands, you must take
7667
care of the substitution yourself.  Just be sure to increment
7668
@var{ptr} over whatever text should not be output normally.
7669
 
7670
@findex recog_data.operand
7671
If you need to look at the operand values, they can be found as the
7672
elements of @code{recog_data.operand}.
7673
 
7674
If the macro definition does nothing, the instruction is output
7675
in the usual way.
7676
@end defmac
7677
 
7678
@defmac FINAL_PRESCAN_INSN (@var{insn}, @var{opvec}, @var{noperands})
7679
If defined, a C statement to be executed just prior to the output of
7680
assembler code for @var{insn}, to modify the extracted operands so
7681
they will be output differently.
7682
 
7683
Here the argument @var{opvec} is the vector containing the operands
7684
extracted from @var{insn}, and @var{noperands} is the number of
7685
elements of the vector which contain meaningful data for this insn.
7686
The contents of this vector are what will be used to convert the insn
7687
template into assembler code, so you can change the assembler output
7688
by changing the contents of the vector.
7689
 
7690
This macro is useful when various assembler syntaxes share a single
7691
file of instruction patterns; by defining this macro differently, you
7692
can cause a large class of instructions to be output differently (such
7693
as with rearranged operands).  Naturally, variations in assembler
7694
syntax affecting individual insn patterns ought to be handled by
7695
writing conditional output routines in those patterns.
7696
 
7697
If this macro is not defined, it is equivalent to a null statement.
7698
@end defmac
7699
 
7700
@defmac PRINT_OPERAND (@var{stream}, @var{x}, @var{code})
7701
A C compound statement to output to stdio stream @var{stream} the
7702
assembler syntax for an instruction operand @var{x}.  @var{x} is an
7703
RTL expression.
7704
 
7705
@var{code} is a value that can be used to specify one of several ways
7706
of printing the operand.  It is used when identical operands must be
7707
printed differently depending on the context.  @var{code} comes from
7708
the @samp{%} specification that was used to request printing of the
7709
operand.  If the specification was just @samp{%@var{digit}} then
7710
@var{code} is 0; if the specification was @samp{%@var{ltr}
7711
@var{digit}} then @var{code} is the ASCII code for @var{ltr}.
7712
 
7713
@findex reg_names
7714
If @var{x} is a register, this macro should print the register's name.
7715
The names can be found in an array @code{reg_names} whose type is
7716
@code{char *[]}.  @code{reg_names} is initialized from
7717
@code{REGISTER_NAMES}.
7718
 
7719
When the machine description has a specification @samp{%@var{punct}}
7720
(a @samp{%} followed by a punctuation character), this macro is called
7721
with a null pointer for @var{x} and the punctuation character for
7722
@var{code}.
7723
@end defmac
7724
 
7725
@defmac PRINT_OPERAND_PUNCT_VALID_P (@var{code})
7726
A C expression which evaluates to true if @var{code} is a valid
7727
punctuation character for use in the @code{PRINT_OPERAND} macro.  If
7728
@code{PRINT_OPERAND_PUNCT_VALID_P} is not defined, it means that no
7729
punctuation characters (except for the standard one, @samp{%}) are used
7730
in this way.
7731
@end defmac
7732
 
7733
@defmac PRINT_OPERAND_ADDRESS (@var{stream}, @var{x})
7734
A C compound statement to output to stdio stream @var{stream} the
7735
assembler syntax for an instruction operand that is a memory reference
7736
whose address is @var{x}.  @var{x} is an RTL expression.
7737
 
7738
@cindex @code{TARGET_ENCODE_SECTION_INFO} usage
7739
On some machines, the syntax for a symbolic address depends on the
7740
section that the address refers to.  On these machines, define the hook
7741
@code{TARGET_ENCODE_SECTION_INFO} to store the information into the
7742
@code{symbol_ref}, and then check for it here.  @xref{Assembler
7743
Format}.
7744
@end defmac
7745
 
7746
@findex dbr_sequence_length
7747
@defmac DBR_OUTPUT_SEQEND (@var{file})
7748
A C statement, to be executed after all slot-filler instructions have
7749
been output.  If necessary, call @code{dbr_sequence_length} to
7750
determine the number of slots filled in a sequence (zero if not
7751
currently outputting a sequence), to decide how many no-ops to output,
7752
or whatever.
7753
 
7754
Don't define this macro if it has nothing to do, but it is helpful in
7755
reading assembly output if the extent of the delay sequence is made
7756
explicit (e.g.@: with white space).
7757
@end defmac
7758
 
7759
@findex final_sequence
7760
Note that output routines for instructions with delay slots must be
7761
prepared to deal with not being output as part of a sequence
7762
(i.e.@: when the scheduling pass is not run, or when no slot fillers could be
7763
found.)  The variable @code{final_sequence} is null when not
7764
processing a sequence, otherwise it contains the @code{sequence} rtx
7765
being output.
7766
 
7767
@findex asm_fprintf
7768
@defmac REGISTER_PREFIX
7769
@defmacx LOCAL_LABEL_PREFIX
7770
@defmacx USER_LABEL_PREFIX
7771
@defmacx IMMEDIATE_PREFIX
7772
If defined, C string expressions to be used for the @samp{%R}, @samp{%L},
7773
@samp{%U}, and @samp{%I} options of @code{asm_fprintf} (see
7774
@file{final.c}).  These are useful when a single @file{md} file must
7775
support multiple assembler formats.  In that case, the various @file{tm.h}
7776
files can define these macros differently.
7777
@end defmac
7778
 
7779
@defmac ASM_FPRINTF_EXTENSIONS (@var{file}, @var{argptr}, @var{format})
7780
If defined this macro should expand to a series of @code{case}
7781
statements which will be parsed inside the @code{switch} statement of
7782
the @code{asm_fprintf} function.  This allows targets to define extra
7783
printf formats which may useful when generating their assembler
7784
statements.  Note that uppercase letters are reserved for future
7785
generic extensions to asm_fprintf, and so are not available to target
7786
specific code.  The output file is given by the parameter @var{file}.
7787
The varargs input pointer is @var{argptr} and the rest of the format
7788
string, starting the character after the one that is being switched
7789
upon, is pointed to by @var{format}.
7790
@end defmac
7791
 
7792
@defmac ASSEMBLER_DIALECT
7793
If your target supports multiple dialects of assembler language (such as
7794
different opcodes), define this macro as a C expression that gives the
7795
numeric index of the assembler language dialect to use, with zero as the
7796
first variant.
7797
 
7798
If this macro is defined, you may use constructs of the form
7799
@smallexample
7800
@samp{@{option0|option1|option2@dots{}@}}
7801
@end smallexample
7802
@noindent
7803
in the output templates of patterns (@pxref{Output Template}) or in the
7804
first argument of @code{asm_fprintf}.  This construct outputs
7805
@samp{option0}, @samp{option1}, @samp{option2}, etc., if the value of
7806
@code{ASSEMBLER_DIALECT} is zero, one, two, etc.  Any special characters
7807
within these strings retain their usual meaning.  If there are fewer
7808
alternatives within the braces than the value of
7809
@code{ASSEMBLER_DIALECT}, the construct outputs nothing.
7810
 
7811
If you do not define this macro, the characters @samp{@{}, @samp{|} and
7812
@samp{@}} do not have any special meaning when used in templates or
7813
operands to @code{asm_fprintf}.
7814
 
7815
Define the macros @code{REGISTER_PREFIX}, @code{LOCAL_LABEL_PREFIX},
7816
@code{USER_LABEL_PREFIX} and @code{IMMEDIATE_PREFIX} if you can express
7817
the variations in assembler language syntax with that mechanism.  Define
7818
@code{ASSEMBLER_DIALECT} and use the @samp{@{option0|option1@}} syntax
7819
if the syntax variant are larger and involve such things as different
7820
opcodes or operand order.
7821
@end defmac
7822
 
7823
@defmac ASM_OUTPUT_REG_PUSH (@var{stream}, @var{regno})
7824
A C expression to output to @var{stream} some assembler code
7825
which will push hard register number @var{regno} onto the stack.
7826
The code need not be optimal, since this macro is used only when
7827
profiling.
7828
@end defmac
7829
 
7830
@defmac ASM_OUTPUT_REG_POP (@var{stream}, @var{regno})
7831
A C expression to output to @var{stream} some assembler code
7832
which will pop hard register number @var{regno} off of the stack.
7833
The code need not be optimal, since this macro is used only when
7834
profiling.
7835
@end defmac
7836
 
7837
@node Dispatch Tables
7838
@subsection Output of Dispatch Tables
7839
 
7840
@c prevent bad page break with this line
7841
This concerns dispatch tables.
7842
 
7843
@cindex dispatch table
7844
@defmac ASM_OUTPUT_ADDR_DIFF_ELT (@var{stream}, @var{body}, @var{value}, @var{rel})
7845
A C statement to output to the stdio stream @var{stream} an assembler
7846
pseudo-instruction to generate a difference between two labels.
7847
@var{value} and @var{rel} are the numbers of two internal labels.  The
7848
definitions of these labels are output using
7849
@code{(*targetm.asm_out.internal_label)}, and they must be printed in the same
7850
way here.  For example,
7851
 
7852
@smallexample
7853
fprintf (@var{stream}, "\t.word L%d-L%d\n",
7854
         @var{value}, @var{rel})
7855
@end smallexample
7856
 
7857
You must provide this macro on machines where the addresses in a
7858
dispatch table are relative to the table's own address.  If defined, GCC
7859
will also use this macro on all machines when producing PIC@.
7860
@var{body} is the body of the @code{ADDR_DIFF_VEC}; it is provided so that the
7861
mode and flags can be read.
7862
@end defmac
7863
 
7864
@defmac ASM_OUTPUT_ADDR_VEC_ELT (@var{stream}, @var{value})
7865
This macro should be provided on machines where the addresses
7866
in a dispatch table are absolute.
7867
 
7868
The definition should be a C statement to output to the stdio stream
7869
@var{stream} an assembler pseudo-instruction to generate a reference to
7870
a label.  @var{value} is the number of an internal label whose
7871
definition is output using @code{(*targetm.asm_out.internal_label)}.
7872
For example,
7873
 
7874
@smallexample
7875
fprintf (@var{stream}, "\t.word L%d\n", @var{value})
7876
@end smallexample
7877
@end defmac
7878
 
7879
@defmac ASM_OUTPUT_CASE_LABEL (@var{stream}, @var{prefix}, @var{num}, @var{table})
7880
Define this if the label before a jump-table needs to be output
7881
specially.  The first three arguments are the same as for
7882
@code{(*targetm.asm_out.internal_label)}; the fourth argument is the
7883
jump-table which follows (a @code{jump_insn} containing an
7884
@code{addr_vec} or @code{addr_diff_vec}).
7885
 
7886
This feature is used on system V to output a @code{swbeg} statement
7887
for the table.
7888
 
7889
If this macro is not defined, these labels are output with
7890
@code{(*targetm.asm_out.internal_label)}.
7891
@end defmac
7892
 
7893
@defmac ASM_OUTPUT_CASE_END (@var{stream}, @var{num}, @var{table})
7894
Define this if something special must be output at the end of a
7895
jump-table.  The definition should be a C statement to be executed
7896
after the assembler code for the table is written.  It should write
7897
the appropriate code to stdio stream @var{stream}.  The argument
7898
@var{table} is the jump-table insn, and @var{num} is the label-number
7899
of the preceding label.
7900
 
7901
If this macro is not defined, nothing special is output at the end of
7902
the jump-table.
7903
@end defmac
7904
 
7905
@deftypefn {Target Hook} void TARGET_ASM_EMIT_UNWIND_LABEL (@var{stream}, @var{decl}, @var{for_eh}, @var{empty})
7906
This target hook emits a label at the beginning of each FDE@.  It
7907
should be defined on targets where FDEs need special labels, and it
7908
should write the appropriate label, for the FDE associated with the
7909
function declaration @var{decl}, to the stdio stream @var{stream}.
7910
The third argument, @var{for_eh}, is a boolean: true if this is for an
7911
exception table.  The fourth argument, @var{empty}, is a boolean:
7912
true if this is a placeholder label for an omitted FDE@.
7913
 
7914
The default is that FDEs are not given nonlocal labels.
7915
@end deftypefn
7916
 
7917
@deftypefn {Target Hook} void TARGET_ASM_EMIT_EXCEPT_TABLE_LABEL (@var{stream})
7918
This target hook emits a label at the beginning of the exception table.
7919
It should be defined on targets where it is desirable for the table
7920
to be broken up according to function.
7921
 
7922
The default is that no label is emitted.
7923
@end deftypefn
7924
 
7925
@deftypefn {Target Hook} void TARGET_UNWIND_EMIT (FILE * @var{stream}, rtx @var{insn})
7926
This target hook emits and assembly directives required to unwind the
7927
given instruction.  This is only used when TARGET_UNWIND_INFO is set.
7928
@end deftypefn
7929
 
7930
@node Exception Region Output
7931
@subsection Assembler Commands for Exception Regions
7932
 
7933
@c prevent bad page break with this line
7934
 
7935
This describes commands marking the start and the end of an exception
7936
region.
7937
 
7938
@defmac EH_FRAME_SECTION_NAME
7939
If defined, a C string constant for the name of the section containing
7940
exception handling frame unwind information.  If not defined, GCC will
7941
provide a default definition if the target supports named sections.
7942
@file{crtstuff.c} uses this macro to switch to the appropriate section.
7943
 
7944
You should define this symbol if your target supports DWARF 2 frame
7945
unwind information and the default definition does not work.
7946
@end defmac
7947
 
7948
@defmac EH_FRAME_IN_DATA_SECTION
7949
If defined, DWARF 2 frame unwind information will be placed in the
7950
data section even though the target supports named sections.  This
7951
might be necessary, for instance, if the system linker does garbage
7952
collection and sections cannot be marked as not to be collected.
7953
 
7954
Do not define this macro unless @code{TARGET_ASM_NAMED_SECTION} is
7955
also defined.
7956
@end defmac
7957
 
7958
@defmac EH_TABLES_CAN_BE_READ_ONLY
7959
Define this macro to 1 if your target is such that no frame unwind
7960
information encoding used with non-PIC code will ever require a
7961
runtime relocation, but the linker may not support merging read-only
7962
and read-write sections into a single read-write section.
7963
@end defmac
7964
 
7965
@defmac MASK_RETURN_ADDR
7966
An rtx used to mask the return address found via @code{RETURN_ADDR_RTX}, so
7967
that it does not contain any extraneous set bits in it.
7968
@end defmac
7969
 
7970
@defmac DWARF2_UNWIND_INFO
7971
Define this macro to 0 if your target supports DWARF 2 frame unwind
7972
information, but it does not yet work with exception handling.
7973
Otherwise, if your target supports this information (if it defines
7974
@samp{INCOMING_RETURN_ADDR_RTX} and either @samp{UNALIGNED_INT_ASM_OP}
7975
or @samp{OBJECT_FORMAT_ELF}), GCC will provide a default definition of 1.
7976
 
7977
If @code{TARGET_UNWIND_INFO} is defined, the target specific unwinder
7978
will be used in all cases.  Defining this macro will enable the generation
7979
of DWARF 2 frame debugging information.
7980
 
7981
If @code{TARGET_UNWIND_INFO} is not defined, and this macro is defined to 1,
7982
the DWARF 2 unwinder will be the default exception handling mechanism;
7983
otherwise, the @code{setjmp}/@code{longjmp}-based scheme will be used by
7984
default.
7985
@end defmac
7986
 
7987
@defmac TARGET_UNWIND_INFO
7988
Define this macro if your target has ABI specified unwind tables.  Usually
7989
these will be output by @code{TARGET_UNWIND_EMIT}.
7990
@end defmac
7991
 
7992
@deftypevar {Target Hook} bool TARGET_UNWIND_TABLES_DEFAULT
7993
This variable should be set to @code{true} if the target ABI requires unwinding
7994
tables even when exceptions are not used.
7995
@end deftypevar
7996
 
7997
@defmac MUST_USE_SJLJ_EXCEPTIONS
7998
This macro need only be defined if @code{DWARF2_UNWIND_INFO} is
7999
runtime-variable.  In that case, @file{except.h} cannot correctly
8000
determine the corresponding definition of @code{MUST_USE_SJLJ_EXCEPTIONS},
8001
so the target must provide it directly.
8002
@end defmac
8003
 
8004
@defmac DONT_USE_BUILTIN_SETJMP
8005
Define this macro to 1 if the @code{setjmp}/@code{longjmp}-based scheme
8006
should use the @code{setjmp}/@code{longjmp} functions from the C library
8007
instead of the @code{__builtin_setjmp}/@code{__builtin_longjmp} machinery.
8008
@end defmac
8009
 
8010
@defmac DWARF_CIE_DATA_ALIGNMENT
8011
This macro need only be defined if the target might save registers in the
8012
function prologue at an offset to the stack pointer that is not aligned to
8013
@code{UNITS_PER_WORD}.  The definition should be the negative minimum
8014
alignment if @code{STACK_GROWS_DOWNWARD} is defined, and the positive
8015
minimum alignment otherwise.  @xref{SDB and DWARF}.  Only applicable if
8016
the target supports DWARF 2 frame unwind information.
8017
@end defmac
8018
 
8019
@deftypevar {Target Hook} bool TARGET_TERMINATE_DW2_EH_FRAME_INFO
8020
Contains the value true if the target should add a zero word onto the
8021
end of a Dwarf-2 frame info section when used for exception handling.
8022
Default value is false if @code{EH_FRAME_SECTION_NAME} is defined, and
8023
true otherwise.
8024
@end deftypevar
8025
 
8026
@deftypefn {Target Hook} rtx TARGET_DWARF_REGISTER_SPAN (rtx @var{reg})
8027
Given a register, this hook should return a parallel of registers to
8028
represent where to find the register pieces.  Define this hook if the
8029
register and its mode are represented in Dwarf in non-contiguous
8030
locations, or if the register should be represented in more than one
8031
register in Dwarf.  Otherwise, this hook should return @code{NULL_RTX}.
8032
If not defined, the default is to return @code{NULL_RTX}.
8033
@end deftypefn
8034
 
8035
@deftypefn {Target Hook} bool TARGET_ASM_TTYPE (rtx @var{sym})
8036
This hook is used to output a reference from a frame unwinding table to
8037
the type_info object identified by @var{sym}.  It should return @code{true}
8038
if the reference was output.  Returning @code{false} will cause the
8039
reference to be output using the normal Dwarf2 routines.
8040
@end deftypefn
8041
 
8042
@deftypefn {Target Hook} bool TARGET_ARM_EABI_UNWINDER
8043
This hook should be set to @code{true} on targets that use an ARM EABI
8044
based unwinding library, and @code{false} on other targets.  This effects
8045
the format of unwinding tables, and how the unwinder in entered after
8046
running a cleanup.  The default is @code{false}.
8047
@end deftypefn
8048
 
8049
@node Alignment Output
8050
@subsection Assembler Commands for Alignment
8051
 
8052
@c prevent bad page break with this line
8053
This describes commands for alignment.
8054
 
8055
@defmac JUMP_ALIGN (@var{label})
8056
The alignment (log base 2) to put in front of @var{label}, which is
8057
a common destination of jumps and has no fallthru incoming edge.
8058
 
8059
This macro need not be defined if you don't want any special alignment
8060
to be done at such a time.  Most machine descriptions do not currently
8061
define the macro.
8062
 
8063
Unless it's necessary to inspect the @var{label} parameter, it is better
8064
to set the variable @var{align_jumps} in the target's
8065
@code{OVERRIDE_OPTIONS}.  Otherwise, you should try to honor the user's
8066
selection in @var{align_jumps} in a @code{JUMP_ALIGN} implementation.
8067
@end defmac
8068
 
8069
@defmac LABEL_ALIGN_AFTER_BARRIER (@var{label})
8070
The alignment (log base 2) to put in front of @var{label}, which follows
8071
a @code{BARRIER}.
8072
 
8073
This macro need not be defined if you don't want any special alignment
8074
to be done at such a time.  Most machine descriptions do not currently
8075
define the macro.
8076
@end defmac
8077
 
8078
@defmac LABEL_ALIGN_AFTER_BARRIER_MAX_SKIP
8079
The maximum number of bytes to skip when applying
8080
@code{LABEL_ALIGN_AFTER_BARRIER}.  This works only if
8081
@code{ASM_OUTPUT_MAX_SKIP_ALIGN} is defined.
8082
@end defmac
8083
 
8084
@defmac LOOP_ALIGN (@var{label})
8085
The alignment (log base 2) to put in front of @var{label}, which follows
8086
a @code{NOTE_INSN_LOOP_BEG} note.
8087
 
8088
This macro need not be defined if you don't want any special alignment
8089
to be done at such a time.  Most machine descriptions do not currently
8090
define the macro.
8091
 
8092
Unless it's necessary to inspect the @var{label} parameter, it is better
8093
to set the variable @code{align_loops} in the target's
8094
@code{OVERRIDE_OPTIONS}.  Otherwise, you should try to honor the user's
8095
selection in @code{align_loops} in a @code{LOOP_ALIGN} implementation.
8096
@end defmac
8097
 
8098
@defmac LOOP_ALIGN_MAX_SKIP
8099
The maximum number of bytes to skip when applying @code{LOOP_ALIGN}.
8100
This works only if @code{ASM_OUTPUT_MAX_SKIP_ALIGN} is defined.
8101
@end defmac
8102
 
8103
@defmac LABEL_ALIGN (@var{label})
8104
The alignment (log base 2) to put in front of @var{label}.
8105
If @code{LABEL_ALIGN_AFTER_BARRIER} / @code{LOOP_ALIGN} specify a different alignment,
8106
the maximum of the specified values is used.
8107
 
8108
Unless it's necessary to inspect the @var{label} parameter, it is better
8109
to set the variable @code{align_labels} in the target's
8110
@code{OVERRIDE_OPTIONS}.  Otherwise, you should try to honor the user's
8111
selection in @code{align_labels} in a @code{LABEL_ALIGN} implementation.
8112
@end defmac
8113
 
8114
@defmac LABEL_ALIGN_MAX_SKIP
8115
The maximum number of bytes to skip when applying @code{LABEL_ALIGN}.
8116
This works only if @code{ASM_OUTPUT_MAX_SKIP_ALIGN} is defined.
8117
@end defmac
8118
 
8119
@defmac ASM_OUTPUT_SKIP (@var{stream}, @var{nbytes})
8120
A C statement to output to the stdio stream @var{stream} an assembler
8121
instruction to advance the location counter by @var{nbytes} bytes.
8122
Those bytes should be zero when loaded.  @var{nbytes} will be a C
8123
expression of type @code{int}.
8124
@end defmac
8125
 
8126
@defmac ASM_NO_SKIP_IN_TEXT
8127
Define this macro if @code{ASM_OUTPUT_SKIP} should not be used in the
8128
text section because it fails to put zeros in the bytes that are skipped.
8129
This is true on many Unix systems, where the pseudo--op to skip bytes
8130
produces no-op instructions rather than zeros when used in the text
8131
section.
8132
@end defmac
8133
 
8134
@defmac ASM_OUTPUT_ALIGN (@var{stream}, @var{power})
8135
A C statement to output to the stdio stream @var{stream} an assembler
8136
command to advance the location counter to a multiple of 2 to the
8137
@var{power} bytes.  @var{power} will be a C expression of type @code{int}.
8138
@end defmac
8139
 
8140
@defmac ASM_OUTPUT_ALIGN_WITH_NOP (@var{stream}, @var{power})
8141
Like @code{ASM_OUTPUT_ALIGN}, except that the ``nop'' instruction is used
8142
for padding, if necessary.
8143
@end defmac
8144
 
8145
@defmac ASM_OUTPUT_MAX_SKIP_ALIGN (@var{stream}, @var{power}, @var{max_skip})
8146
A C statement to output to the stdio stream @var{stream} an assembler
8147
command to advance the location counter to a multiple of 2 to the
8148
@var{power} bytes, but only if @var{max_skip} or fewer bytes are needed to
8149
satisfy the alignment request.  @var{power} and @var{max_skip} will be
8150
a C expression of type @code{int}.
8151
@end defmac
8152
 
8153
@need 3000
8154
@node Debugging Info
8155
@section Controlling Debugging Information Format
8156
 
8157
@c prevent bad page break with this line
8158
This describes how to specify debugging information.
8159
 
8160
@menu
8161
* All Debuggers::      Macros that affect all debugging formats uniformly.
8162
* DBX Options::        Macros enabling specific options in DBX format.
8163
* DBX Hooks::          Hook macros for varying DBX format.
8164
* File Names and DBX:: Macros controlling output of file names in DBX format.
8165
* SDB and DWARF::      Macros for SDB (COFF) and DWARF formats.
8166
* VMS Debug::          Macros for VMS debug format.
8167
@end menu
8168
 
8169
@node All Debuggers
8170
@subsection Macros Affecting All Debugging Formats
8171
 
8172
@c prevent bad page break with this line
8173
These macros affect all debugging formats.
8174
 
8175
@defmac DBX_REGISTER_NUMBER (@var{regno})
8176
A C expression that returns the DBX register number for the compiler
8177
register number @var{regno}.  In the default macro provided, the value
8178
of this expression will be @var{regno} itself.  But sometimes there are
8179
some registers that the compiler knows about and DBX does not, or vice
8180
versa.  In such cases, some register may need to have one number in the
8181
compiler and another for DBX@.
8182
 
8183
If two registers have consecutive numbers inside GCC, and they can be
8184
used as a pair to hold a multiword value, then they @emph{must} have
8185
consecutive numbers after renumbering with @code{DBX_REGISTER_NUMBER}.
8186
Otherwise, debuggers will be unable to access such a pair, because they
8187
expect register pairs to be consecutive in their own numbering scheme.
8188
 
8189
If you find yourself defining @code{DBX_REGISTER_NUMBER} in way that
8190
does not preserve register pairs, then what you must do instead is
8191
redefine the actual register numbering scheme.
8192
@end defmac
8193
 
8194
@defmac DEBUGGER_AUTO_OFFSET (@var{x})
8195
A C expression that returns the integer offset value for an automatic
8196
variable having address @var{x} (an RTL expression).  The default
8197
computation assumes that @var{x} is based on the frame-pointer and
8198
gives the offset from the frame-pointer.  This is required for targets
8199
that produce debugging output for DBX or COFF-style debugging output
8200
for SDB and allow the frame-pointer to be eliminated when the
8201
@option{-g} options is used.
8202
@end defmac
8203
 
8204
@defmac DEBUGGER_ARG_OFFSET (@var{offset}, @var{x})
8205
A C expression that returns the integer offset value for an argument
8206
having address @var{x} (an RTL expression).  The nominal offset is
8207
@var{offset}.
8208
@end defmac
8209
 
8210
@defmac PREFERRED_DEBUGGING_TYPE
8211
A C expression that returns the type of debugging output GCC should
8212
produce when the user specifies just @option{-g}.  Define
8213
this if you have arranged for GCC to support more than one format of
8214
debugging output.  Currently, the allowable values are @code{DBX_DEBUG},
8215
@code{SDB_DEBUG}, @code{DWARF_DEBUG}, @code{DWARF2_DEBUG},
8216
@code{XCOFF_DEBUG}, @code{VMS_DEBUG}, and @code{VMS_AND_DWARF2_DEBUG}.
8217
 
8218
When the user specifies @option{-ggdb}, GCC normally also uses the
8219
value of this macro to select the debugging output format, but with two
8220
exceptions.  If @code{DWARF2_DEBUGGING_INFO} is defined, GCC uses the
8221
value @code{DWARF2_DEBUG}.  Otherwise, if @code{DBX_DEBUGGING_INFO} is
8222
defined, GCC uses @code{DBX_DEBUG}.
8223
 
8224
The value of this macro only affects the default debugging output; the
8225
user can always get a specific type of output by using @option{-gstabs},
8226
@option{-gcoff}, @option{-gdwarf-2}, @option{-gxcoff}, or @option{-gvms}.
8227
@end defmac
8228
 
8229
@node DBX Options
8230
@subsection Specific Options for DBX Output
8231
 
8232
@c prevent bad page break with this line
8233
These are specific options for DBX output.
8234
 
8235
@defmac DBX_DEBUGGING_INFO
8236
Define this macro if GCC should produce debugging output for DBX
8237
in response to the @option{-g} option.
8238
@end defmac
8239
 
8240
@defmac XCOFF_DEBUGGING_INFO
8241
Define this macro if GCC should produce XCOFF format debugging output
8242
in response to the @option{-g} option.  This is a variant of DBX format.
8243
@end defmac
8244
 
8245
@defmac DEFAULT_GDB_EXTENSIONS
8246
Define this macro to control whether GCC should by default generate
8247
GDB's extended version of DBX debugging information (assuming DBX-format
8248
debugging information is enabled at all).  If you don't define the
8249
macro, the default is 1: always generate the extended information
8250
if there is any occasion to.
8251
@end defmac
8252
 
8253
@defmac DEBUG_SYMS_TEXT
8254
Define this macro if all @code{.stabs} commands should be output while
8255
in the text section.
8256
@end defmac
8257
 
8258
@defmac ASM_STABS_OP
8259
A C string constant, including spacing, naming the assembler pseudo op to
8260
use instead of @code{"\t.stabs\t"} to define an ordinary debugging symbol.
8261
If you don't define this macro, @code{"\t.stabs\t"} is used.  This macro
8262
applies only to DBX debugging information format.
8263
@end defmac
8264
 
8265
@defmac ASM_STABD_OP
8266
A C string constant, including spacing, naming the assembler pseudo op to
8267
use instead of @code{"\t.stabd\t"} to define a debugging symbol whose
8268
value is the current location.  If you don't define this macro,
8269
@code{"\t.stabd\t"} is used.  This macro applies only to DBX debugging
8270
information format.
8271
@end defmac
8272
 
8273
@defmac ASM_STABN_OP
8274
A C string constant, including spacing, naming the assembler pseudo op to
8275
use instead of @code{"\t.stabn\t"} to define a debugging symbol with no
8276
name.  If you don't define this macro, @code{"\t.stabn\t"} is used.  This
8277
macro applies only to DBX debugging information format.
8278
@end defmac
8279
 
8280
@defmac DBX_NO_XREFS
8281
Define this macro if DBX on your system does not support the construct
8282
@samp{xs@var{tagname}}.  On some systems, this construct is used to
8283
describe a forward reference to a structure named @var{tagname}.
8284
On other systems, this construct is not supported at all.
8285
@end defmac
8286
 
8287
@defmac DBX_CONTIN_LENGTH
8288
A symbol name in DBX-format debugging information is normally
8289
continued (split into two separate @code{.stabs} directives) when it
8290
exceeds a certain length (by default, 80 characters).  On some
8291
operating systems, DBX requires this splitting; on others, splitting
8292
must not be done.  You can inhibit splitting by defining this macro
8293
with the value zero.  You can override the default splitting-length by
8294
defining this macro as an expression for the length you desire.
8295
@end defmac
8296
 
8297
@defmac DBX_CONTIN_CHAR
8298
Normally continuation is indicated by adding a @samp{\} character to
8299
the end of a @code{.stabs} string when a continuation follows.  To use
8300
a different character instead, define this macro as a character
8301
constant for the character you want to use.  Do not define this macro
8302
if backslash is correct for your system.
8303
@end defmac
8304
 
8305
@defmac DBX_STATIC_STAB_DATA_SECTION
8306
Define this macro if it is necessary to go to the data section before
8307
outputting the @samp{.stabs} pseudo-op for a non-global static
8308
variable.
8309
@end defmac
8310
 
8311
@defmac DBX_TYPE_DECL_STABS_CODE
8312
The value to use in the ``code'' field of the @code{.stabs} directive
8313
for a typedef.  The default is @code{N_LSYM}.
8314
@end defmac
8315
 
8316
@defmac DBX_STATIC_CONST_VAR_CODE
8317
The value to use in the ``code'' field of the @code{.stabs} directive
8318
for a static variable located in the text section.  DBX format does not
8319
provide any ``right'' way to do this.  The default is @code{N_FUN}.
8320
@end defmac
8321
 
8322
@defmac DBX_REGPARM_STABS_CODE
8323
The value to use in the ``code'' field of the @code{.stabs} directive
8324
for a parameter passed in registers.  DBX format does not provide any
8325
``right'' way to do this.  The default is @code{N_RSYM}.
8326
@end defmac
8327
 
8328
@defmac DBX_REGPARM_STABS_LETTER
8329
The letter to use in DBX symbol data to identify a symbol as a parameter
8330
passed in registers.  DBX format does not customarily provide any way to
8331
do this.  The default is @code{'P'}.
8332
@end defmac
8333
 
8334
@defmac DBX_FUNCTION_FIRST
8335
Define this macro if the DBX information for a function and its
8336
arguments should precede the assembler code for the function.  Normally,
8337
in DBX format, the debugging information entirely follows the assembler
8338
code.
8339
@end defmac
8340
 
8341
@defmac DBX_BLOCKS_FUNCTION_RELATIVE
8342
Define this macro, with value 1, if the value of a symbol describing
8343
the scope of a block (@code{N_LBRAC} or @code{N_RBRAC}) should be
8344
relative to the start of the enclosing function.  Normally, GCC uses
8345
an absolute address.
8346
@end defmac
8347
 
8348
@defmac DBX_LINES_FUNCTION_RELATIVE
8349
Define this macro, with value 1, if the value of a symbol indicating
8350
the current line number (@code{N_SLINE}) should be relative to the
8351
start of the enclosing function.  Normally, GCC uses an absolute address.
8352
@end defmac
8353
 
8354
@defmac DBX_USE_BINCL
8355
Define this macro if GCC should generate @code{N_BINCL} and
8356
@code{N_EINCL} stabs for included header files, as on Sun systems.  This
8357
macro also directs GCC to output a type number as a pair of a file
8358
number and a type number within the file.  Normally, GCC does not
8359
generate @code{N_BINCL} or @code{N_EINCL} stabs, and it outputs a single
8360
number for a type number.
8361
@end defmac
8362
 
8363
@node DBX Hooks
8364
@subsection Open-Ended Hooks for DBX Format
8365
 
8366
@c prevent bad page break with this line
8367
These are hooks for DBX format.
8368
 
8369
@defmac DBX_OUTPUT_LBRAC (@var{stream}, @var{name})
8370
Define this macro to say how to output to @var{stream} the debugging
8371
information for the start of a scope level for variable names.  The
8372
argument @var{name} is the name of an assembler symbol (for use with
8373
@code{assemble_name}) whose value is the address where the scope begins.
8374
@end defmac
8375
 
8376
@defmac DBX_OUTPUT_RBRAC (@var{stream}, @var{name})
8377
Like @code{DBX_OUTPUT_LBRAC}, but for the end of a scope level.
8378
@end defmac
8379
 
8380
@defmac DBX_OUTPUT_NFUN (@var{stream}, @var{lscope_label}, @var{decl})
8381
Define this macro if the target machine requires special handling to
8382
output an @code{N_FUN} entry for the function @var{decl}.
8383
@end defmac
8384
 
8385
@defmac DBX_OUTPUT_SOURCE_LINE (@var{stream}, @var{line}, @var{counter})
8386
A C statement to output DBX debugging information before code for line
8387
number @var{line} of the current source file to the stdio stream
8388
@var{stream}.  @var{counter} is the number of time the macro was
8389
invoked, including the current invocation; it is intended to generate
8390
unique labels in the assembly output.
8391
 
8392
This macro should not be defined if the default output is correct, or
8393
if it can be made correct by defining @code{DBX_LINES_FUNCTION_RELATIVE}.
8394
@end defmac
8395
 
8396
@defmac NO_DBX_FUNCTION_END
8397
Some stabs encapsulation formats (in particular ECOFF), cannot handle the
8398
@code{.stabs "",N_FUN,,0,0,Lscope-function-1} gdb dbx extension construct.
8399
On those machines, define this macro to turn this feature off without
8400
disturbing the rest of the gdb extensions.
8401
@end defmac
8402
 
8403
@defmac NO_DBX_BNSYM_ENSYM
8404
Some assemblers cannot handle the @code{.stabd BNSYM/ENSYM,0,0} gdb dbx
8405
extension construct.  On those machines, define this macro to turn this
8406
feature off without disturbing the rest of the gdb extensions.
8407
@end defmac
8408
 
8409
@node File Names and DBX
8410
@subsection File Names in DBX Format
8411
 
8412
@c prevent bad page break with this line
8413
This describes file names in DBX format.
8414
 
8415
@defmac DBX_OUTPUT_MAIN_SOURCE_FILENAME (@var{stream}, @var{name})
8416
A C statement to output DBX debugging information to the stdio stream
8417
@var{stream}, which indicates that file @var{name} is the main source
8418
file---the file specified as the input file for compilation.
8419
This macro is called only once, at the beginning of compilation.
8420
 
8421
This macro need not be defined if the standard form of output
8422
for DBX debugging information is appropriate.
8423
 
8424
It may be necessary to refer to a label equal to the beginning of the
8425
text section.  You can use @samp{assemble_name (stream, ltext_label_name)}
8426
to do so.  If you do this, you must also set the variable
8427
@var{used_ltext_label_name} to @code{true}.
8428
@end defmac
8429
 
8430
@defmac NO_DBX_MAIN_SOURCE_DIRECTORY
8431
Define this macro, with value 1, if GCC should not emit an indication
8432
of the current directory for compilation and current source language at
8433
the beginning of the file.
8434
@end defmac
8435
 
8436
@defmac NO_DBX_GCC_MARKER
8437
Define this macro, with value 1, if GCC should not emit an indication
8438
that this object file was compiled by GCC@.  The default is to emit
8439
an @code{N_OPT} stab at the beginning of every source file, with
8440
@samp{gcc2_compiled.} for the string and value 0.
8441
@end defmac
8442
 
8443
@defmac DBX_OUTPUT_MAIN_SOURCE_FILE_END (@var{stream}, @var{name})
8444
A C statement to output DBX debugging information at the end of
8445
compilation of the main source file @var{name}.  Output should be
8446
written to the stdio stream @var{stream}.
8447
 
8448
If you don't define this macro, nothing special is output at the end
8449
of compilation, which is correct for most machines.
8450
@end defmac
8451
 
8452
@defmac DBX_OUTPUT_NULL_N_SO_AT_MAIN_SOURCE_FILE_END
8453
Define this macro @emph{instead of} defining
8454
@code{DBX_OUTPUT_MAIN_SOURCE_FILE_END}, if what needs to be output at
8455
the end of compilation is a @code{N_SO} stab with an empty string,
8456
whose value is the highest absolute text address in the file.
8457
@end defmac
8458
 
8459
@need 2000
8460
@node SDB and DWARF
8461
@subsection Macros for SDB and DWARF Output
8462
 
8463
@c prevent bad page break with this line
8464
Here are macros for SDB and DWARF output.
8465
 
8466
@defmac SDB_DEBUGGING_INFO
8467
Define this macro if GCC should produce COFF-style debugging output
8468
for SDB in response to the @option{-g} option.
8469
@end defmac
8470
 
8471
@defmac DWARF2_DEBUGGING_INFO
8472
Define this macro if GCC should produce dwarf version 2 format
8473
debugging output in response to the @option{-g} option.
8474
 
8475
@deftypefn {Target Hook} int TARGET_DWARF_CALLING_CONVENTION (tree @var{function})
8476
Define this to enable the dwarf attribute @code{DW_AT_calling_convention} to
8477
be emitted for each function.  Instead of an integer return the enum
8478
value for the @code{DW_CC_} tag.
8479
@end deftypefn
8480
 
8481
To support optional call frame debugging information, you must also
8482
define @code{INCOMING_RETURN_ADDR_RTX} and either set
8483
@code{RTX_FRAME_RELATED_P} on the prologue insns if you use RTL for the
8484
prologue, or call @code{dwarf2out_def_cfa} and @code{dwarf2out_reg_save}
8485
as appropriate from @code{TARGET_ASM_FUNCTION_PROLOGUE} if you don't.
8486
@end defmac
8487
 
8488
@defmac DWARF2_FRAME_INFO
8489
Define this macro to a nonzero value if GCC should always output
8490
Dwarf 2 frame information.  If @code{DWARF2_UNWIND_INFO}
8491
(@pxref{Exception Region Output} is nonzero, GCC will output this
8492
information not matter how you define @code{DWARF2_FRAME_INFO}.
8493
@end defmac
8494
 
8495
@defmac DWARF2_ASM_LINE_DEBUG_INFO
8496
Define this macro to be a nonzero value if the assembler can generate Dwarf 2
8497
line debug info sections.  This will result in much more compact line number
8498
tables, and hence is desirable if it works.
8499
@end defmac
8500
 
8501
@defmac ASM_OUTPUT_DWARF_DELTA (@var{stream}, @var{size}, @var{label1}, @var{label2})
8502
A C statement to issue assembly directives that create a difference
8503
@var{lab1} minus @var{lab2}, using an integer of the given @var{size}.
8504
@end defmac
8505
 
8506
@defmac ASM_OUTPUT_DWARF_OFFSET (@var{stream}, @var{size}, @var{label}, @var{section})
8507
A C statement to issue assembly directives that create a
8508
section-relative reference to the given @var{label}, using an integer of the
8509
given @var{size}.  The label is known to be defined in the given @var{section}.
8510
@end defmac
8511
 
8512
@defmac ASM_OUTPUT_DWARF_PCREL (@var{stream}, @var{size}, @var{label})
8513
A C statement to issue assembly directives that create a self-relative
8514
reference to the given @var{label}, using an integer of the given @var{size}.
8515
@end defmac
8516
 
8517
@deftypefn {Target Hook} void TARGET_ASM_OUTPUT_DWARF_DTPREL (FILE *@var{FILE}, int @var{size}, rtx @var{x})
8518
If defined, this target hook is a function which outputs a DTP-relative
8519
reference to the given TLS symbol of the specified size.
8520
@end deftypefn
8521
 
8522
@defmac PUT_SDB_@dots{}
8523
Define these macros to override the assembler syntax for the special
8524
SDB assembler directives.  See @file{sdbout.c} for a list of these
8525
macros and their arguments.  If the standard syntax is used, you need
8526
not define them yourself.
8527
@end defmac
8528
 
8529
@defmac SDB_DELIM
8530
Some assemblers do not support a semicolon as a delimiter, even between
8531
SDB assembler directives.  In that case, define this macro to be the
8532
delimiter to use (usually @samp{\n}).  It is not necessary to define
8533
a new set of @code{PUT_SDB_@var{op}} macros if this is the only change
8534
required.
8535
@end defmac
8536
 
8537
@defmac SDB_ALLOW_UNKNOWN_REFERENCES
8538
Define this macro to allow references to unknown structure,
8539
union, or enumeration tags to be emitted.  Standard COFF does not
8540
allow handling of unknown references, MIPS ECOFF has support for
8541
it.
8542
@end defmac
8543
 
8544
@defmac SDB_ALLOW_FORWARD_REFERENCES
8545
Define this macro to allow references to structure, union, or
8546
enumeration tags that have not yet been seen to be handled.  Some
8547
assemblers choke if forward tags are used, while some require it.
8548
@end defmac
8549
 
8550
@defmac SDB_OUTPUT_SOURCE_LINE (@var{stream}, @var{line})
8551
A C statement to output SDB debugging information before code for line
8552
number @var{line} of the current source file to the stdio stream
8553
@var{stream}.  The default is to emit an @code{.ln} directive.
8554
@end defmac
8555
 
8556
@need 2000
8557
@node VMS Debug
8558
@subsection Macros for VMS Debug Format
8559
 
8560
@c prevent bad page break with this line
8561
Here are macros for VMS debug format.
8562
 
8563
@defmac VMS_DEBUGGING_INFO
8564
Define this macro if GCC should produce debugging output for VMS
8565
in response to the @option{-g} option.  The default behavior for VMS
8566
is to generate minimal debug info for a traceback in the absence of
8567
@option{-g} unless explicitly overridden with @option{-g0}.  This
8568
behavior is controlled by @code{OPTIMIZATION_OPTIONS} and
8569
@code{OVERRIDE_OPTIONS}.
8570
@end defmac
8571
 
8572
@node Floating Point
8573
@section Cross Compilation and Floating Point
8574
@cindex cross compilation and floating point
8575
@cindex floating point and cross compilation
8576
 
8577
While all modern machines use twos-complement representation for integers,
8578
there are a variety of representations for floating point numbers.  This
8579
means that in a cross-compiler the representation of floating point numbers
8580
in the compiled program may be different from that used in the machine
8581
doing the compilation.
8582
 
8583
Because different representation systems may offer different amounts of
8584
range and precision, all floating point constants must be represented in
8585
the target machine's format.  Therefore, the cross compiler cannot
8586
safely use the host machine's floating point arithmetic; it must emulate
8587
the target's arithmetic.  To ensure consistency, GCC always uses
8588
emulation to work with floating point values, even when the host and
8589
target floating point formats are identical.
8590
 
8591
The following macros are provided by @file{real.h} for the compiler to
8592
use.  All parts of the compiler which generate or optimize
8593
floating-point calculations must use these macros.  They may evaluate
8594
their operands more than once, so operands must not have side effects.
8595
 
8596
@defmac REAL_VALUE_TYPE
8597
The C data type to be used to hold a floating point value in the target
8598
machine's format.  Typically this is a @code{struct} containing an
8599
array of @code{HOST_WIDE_INT}, but all code should treat it as an opaque
8600
quantity.
8601
@end defmac
8602
 
8603
@deftypefn Macro int REAL_VALUES_EQUAL (REAL_VALUE_TYPE @var{x}, REAL_VALUE_TYPE @var{y})
8604
Compares for equality the two values, @var{x} and @var{y}.  If the target
8605
floating point format supports negative zeroes and/or NaNs,
8606
@samp{REAL_VALUES_EQUAL (-0.0, 0.0)} is true, and
8607
@samp{REAL_VALUES_EQUAL (NaN, NaN)} is false.
8608
@end deftypefn
8609
 
8610
@deftypefn Macro int REAL_VALUES_LESS (REAL_VALUE_TYPE @var{x}, REAL_VALUE_TYPE @var{y})
8611
Tests whether @var{x} is less than @var{y}.
8612
@end deftypefn
8613
 
8614
@deftypefn Macro HOST_WIDE_INT REAL_VALUE_FIX (REAL_VALUE_TYPE @var{x})
8615
Truncates @var{x} to a signed integer, rounding toward zero.
8616
@end deftypefn
8617
 
8618
@deftypefn Macro {unsigned HOST_WIDE_INT} REAL_VALUE_UNSIGNED_FIX (REAL_VALUE_TYPE @var{x})
8619
Truncates @var{x} to an unsigned integer, rounding toward zero.  If
8620
@var{x} is negative, returns zero.
8621
@end deftypefn
8622
 
8623
@deftypefn Macro REAL_VALUE_TYPE REAL_VALUE_ATOF (const char *@var{string}, enum machine_mode @var{mode})
8624
Converts @var{string} into a floating point number in the target machine's
8625
representation for mode @var{mode}.  This routine can handle both
8626
decimal and hexadecimal floating point constants, using the syntax
8627
defined by the C language for both.
8628
@end deftypefn
8629
 
8630
@deftypefn Macro int REAL_VALUE_NEGATIVE (REAL_VALUE_TYPE @var{x})
8631
Returns 1 if @var{x} is negative (including negative zero), 0 otherwise.
8632
@end deftypefn
8633
 
8634
@deftypefn Macro int REAL_VALUE_ISINF (REAL_VALUE_TYPE @var{x})
8635
Determines whether @var{x} represents infinity (positive or negative).
8636
@end deftypefn
8637
 
8638
@deftypefn Macro int REAL_VALUE_ISNAN (REAL_VALUE_TYPE @var{x})
8639
Determines whether @var{x} represents a ``NaN'' (not-a-number).
8640
@end deftypefn
8641
 
8642
@deftypefn Macro void REAL_ARITHMETIC (REAL_VALUE_TYPE @var{output}, enum tree_code @var{code}, REAL_VALUE_TYPE @var{x}, REAL_VALUE_TYPE @var{y})
8643
Calculates an arithmetic operation on the two floating point values
8644
@var{x} and @var{y}, storing the result in @var{output} (which must be a
8645
variable).
8646
 
8647
The operation to be performed is specified by @var{code}.  Only the
8648
following codes are supported: @code{PLUS_EXPR}, @code{MINUS_EXPR},
8649
@code{MULT_EXPR}, @code{RDIV_EXPR}, @code{MAX_EXPR}, @code{MIN_EXPR}.
8650
 
8651
If @code{REAL_ARITHMETIC} is asked to evaluate division by zero and the
8652
target's floating point format cannot represent infinity, it will call
8653
@code{abort}.  Callers should check for this situation first, using
8654
@code{MODE_HAS_INFINITIES}.  @xref{Storage Layout}.
8655
@end deftypefn
8656
 
8657
@deftypefn Macro REAL_VALUE_TYPE REAL_VALUE_NEGATE (REAL_VALUE_TYPE @var{x})
8658
Returns the negative of the floating point value @var{x}.
8659
@end deftypefn
8660
 
8661
@deftypefn Macro REAL_VALUE_TYPE REAL_VALUE_ABS (REAL_VALUE_TYPE @var{x})
8662
Returns the absolute value of @var{x}.
8663
@end deftypefn
8664
 
8665
@deftypefn Macro REAL_VALUE_TYPE REAL_VALUE_TRUNCATE (REAL_VALUE_TYPE @var{mode}, enum machine_mode @var{x})
8666
Truncates the floating point value @var{x} to fit in @var{mode}.  The
8667
return value is still a full-size @code{REAL_VALUE_TYPE}, but it has an
8668
appropriate bit pattern to be output asa floating constant whose
8669
precision accords with mode @var{mode}.
8670
@end deftypefn
8671
 
8672
@deftypefn Macro void REAL_VALUE_TO_INT (HOST_WIDE_INT @var{low}, HOST_WIDE_INT @var{high}, REAL_VALUE_TYPE @var{x})
8673
Converts a floating point value @var{x} into a double-precision integer
8674
which is then stored into @var{low} and @var{high}.  If the value is not
8675
integral, it is truncated.
8676
@end deftypefn
8677
 
8678
@deftypefn Macro void REAL_VALUE_FROM_INT (REAL_VALUE_TYPE @var{x}, HOST_WIDE_INT @var{low}, HOST_WIDE_INT @var{high}, enum machine_mode @var{mode})
8679
Converts a double-precision integer found in @var{low} and @var{high},
8680
into a floating point value which is then stored into @var{x}.  The
8681
value is truncated to fit in mode @var{mode}.
8682
@end deftypefn
8683
 
8684
@node Mode Switching
8685
@section Mode Switching Instructions
8686
@cindex mode switching
8687
The following macros control mode switching optimizations:
8688
 
8689
@defmac OPTIMIZE_MODE_SWITCHING (@var{entity})
8690
Define this macro if the port needs extra instructions inserted for mode
8691
switching in an optimizing compilation.
8692
 
8693
For an example, the SH4 can perform both single and double precision
8694
floating point operations, but to perform a single precision operation,
8695
the FPSCR PR bit has to be cleared, while for a double precision
8696
operation, this bit has to be set.  Changing the PR bit requires a general
8697
purpose register as a scratch register, hence these FPSCR sets have to
8698
be inserted before reload, i.e.@: you can't put this into instruction emitting
8699
or @code{TARGET_MACHINE_DEPENDENT_REORG}.
8700
 
8701
You can have multiple entities that are mode-switched, and select at run time
8702
which entities actually need it.  @code{OPTIMIZE_MODE_SWITCHING} should
8703
return nonzero for any @var{entity} that needs mode-switching.
8704
If you define this macro, you also have to define
8705
@code{NUM_MODES_FOR_MODE_SWITCHING}, @code{MODE_NEEDED},
8706
@code{MODE_PRIORITY_TO_MODE} and @code{EMIT_MODE_SET}.
8707
@code{MODE_AFTER}, @code{MODE_ENTRY}, and @code{MODE_EXIT}
8708
are optional.
8709
@end defmac
8710
 
8711
@defmac NUM_MODES_FOR_MODE_SWITCHING
8712
If you define @code{OPTIMIZE_MODE_SWITCHING}, you have to define this as
8713
initializer for an array of integers.  Each initializer element
8714
N refers to an entity that needs mode switching, and specifies the number
8715
of different modes that might need to be set for this entity.
8716
The position of the initializer in the initializer---starting counting at
8717
zero---determines the integer that is used to refer to the mode-switched
8718
entity in question.
8719
In macros that take mode arguments / yield a mode result, modes are
8720
represented as numbers 0 @dots{} N @minus{} 1.  N is used to specify that no mode
8721
switch is needed / supplied.
8722
@end defmac
8723
 
8724
@defmac MODE_NEEDED (@var{entity}, @var{insn})
8725
@var{entity} is an integer specifying a mode-switched entity.  If
8726
@code{OPTIMIZE_MODE_SWITCHING} is defined, you must define this macro to
8727
return an integer value not larger than the corresponding element in
8728
@code{NUM_MODES_FOR_MODE_SWITCHING}, to denote the mode that @var{entity} must
8729
be switched into prior to the execution of @var{insn}.
8730
@end defmac
8731
 
8732
@defmac MODE_AFTER (@var{mode}, @var{insn})
8733
If this macro is defined, it is evaluated for every @var{insn} during
8734
mode switching.  It determines the mode that an insn results in (if
8735
different from the incoming mode).
8736
@end defmac
8737
 
8738
@defmac MODE_ENTRY (@var{entity})
8739
If this macro is defined, it is evaluated for every @var{entity} that needs
8740
mode switching.  It should evaluate to an integer, which is a mode that
8741
@var{entity} is assumed to be switched to at function entry.  If @code{MODE_ENTRY}
8742
is defined then @code{MODE_EXIT} must be defined.
8743
@end defmac
8744
 
8745
@defmac MODE_EXIT (@var{entity})
8746
If this macro is defined, it is evaluated for every @var{entity} that needs
8747
mode switching.  It should evaluate to an integer, which is a mode that
8748
@var{entity} is assumed to be switched to at function exit.  If @code{MODE_EXIT}
8749
is defined then @code{MODE_ENTRY} must be defined.
8750
@end defmac
8751
 
8752
@defmac MODE_PRIORITY_TO_MODE (@var{entity}, @var{n})
8753
This macro specifies the order in which modes for @var{entity} are processed.
8754
 
8755
lowest.  The value of the macro should be an integer designating a mode
8756
for @var{entity}.  For any fixed @var{entity}, @code{mode_priority_to_mode}
8757
(@var{entity}, @var{n}) shall be a bijection in 0 @dots{}
8758
@code{num_modes_for_mode_switching[@var{entity}] - 1}.
8759
@end defmac
8760
 
8761
@defmac EMIT_MODE_SET (@var{entity}, @var{mode}, @var{hard_regs_live})
8762
Generate one or more insns to set @var{entity} to @var{mode}.
8763
@var{hard_reg_live} is the set of hard registers live at the point where
8764
the insn(s) are to be inserted.
8765
@end defmac
8766
 
8767
@node Target Attributes
8768
@section Defining target-specific uses of @code{__attribute__}
8769
@cindex target attributes
8770
@cindex machine attributes
8771
@cindex attributes, target-specific
8772
 
8773
Target-specific attributes may be defined for functions, data and types.
8774
These are described using the following target hooks; they also need to
8775
be documented in @file{extend.texi}.
8776
 
8777
@deftypevr {Target Hook} {const struct attribute_spec *} TARGET_ATTRIBUTE_TABLE
8778
If defined, this target hook points to an array of @samp{struct
8779
attribute_spec} (defined in @file{tree.h}) specifying the machine
8780
specific attributes for this target and some of the restrictions on the
8781
entities to which these attributes are applied and the arguments they
8782
take.
8783
@end deftypevr
8784
 
8785
@deftypefn {Target Hook} int TARGET_COMP_TYPE_ATTRIBUTES (tree @var{type1}, tree @var{type2})
8786
If defined, this target hook is a function which returns zero if the attributes on
8787
@var{type1} and @var{type2} are incompatible, one if they are compatible,
8788
and two if they are nearly compatible (which causes a warning to be
8789
generated).  If this is not defined, machine-specific attributes are
8790
supposed always to be compatible.
8791
@end deftypefn
8792
 
8793
@deftypefn {Target Hook} void TARGET_SET_DEFAULT_TYPE_ATTRIBUTES (tree @var{type})
8794
If defined, this target hook is a function which assigns default attributes to
8795
newly defined @var{type}.
8796
@end deftypefn
8797
 
8798
@deftypefn {Target Hook} tree TARGET_MERGE_TYPE_ATTRIBUTES (tree @var{type1}, tree @var{type2})
8799
Define this target hook if the merging of type attributes needs special
8800
handling.  If defined, the result is a list of the combined
8801
@code{TYPE_ATTRIBUTES} of @var{type1} and @var{type2}.  It is assumed
8802
that @code{comptypes} has already been called and returned 1.  This
8803
function may call @code{merge_attributes} to handle machine-independent
8804
merging.
8805
@end deftypefn
8806
 
8807
@deftypefn {Target Hook} tree TARGET_MERGE_DECL_ATTRIBUTES (tree @var{olddecl}, tree @var{newdecl})
8808
Define this target hook if the merging of decl attributes needs special
8809
handling.  If defined, the result is a list of the combined
8810
@code{DECL_ATTRIBUTES} of @var{olddecl} and @var{newdecl}.
8811
@var{newdecl} is a duplicate declaration of @var{olddecl}.  Examples of
8812
when this is needed are when one attribute overrides another, or when an
8813
attribute is nullified by a subsequent definition.  This function may
8814
call @code{merge_attributes} to handle machine-independent merging.
8815
 
8816
@findex TARGET_DLLIMPORT_DECL_ATTRIBUTES
8817
If the only target-specific handling you require is @samp{dllimport}
8818
for Microsoft Windows targets, you should define the macro
8819
@code{TARGET_DLLIMPORT_DECL_ATTRIBUTES} to @code{1}.  The compiler
8820
will then define a function called
8821
@code{merge_dllimport_decl_attributes} which can then be defined as
8822
the expansion of @code{TARGET_MERGE_DECL_ATTRIBUTES}.  You can also
8823
add @code{handle_dll_attribute} in the attribute table for your port
8824
to perform initial processing of the @samp{dllimport} and
8825
@samp{dllexport} attributes.  This is done in @file{i386/cygwin.h} and
8826
@file{i386/i386.c}, for example.
8827
@end deftypefn
8828
 
8829
@deftypefn {Target Hook} bool TARGET_VALID_DLLIMPORT_ATTRIBUTE_P (tree @var{decl})
8830
@var{decl} is a variable or function with @code{__attribute__((dllimport))}
8831
specified. Use this hook if the target needs to add extra validation
8832
checks to @code{handle_dll_attribute}.
8833
@end deftypefn
8834
 
8835
@defmac TARGET_DECLSPEC
8836
Define this macro to a nonzero value if you want to treat
8837
@code{__declspec(X)} as equivalent to @code{__attribute((X))}.  By
8838
default, this behavior is enabled only for targets that define
8839
@code{TARGET_DLLIMPORT_DECL_ATTRIBUTES}.  The current implementation
8840
of @code{__declspec} is via a built-in macro, but you should not rely
8841
on this implementation detail.
8842
@end defmac
8843
 
8844
@deftypefn {Target Hook} void TARGET_INSERT_ATTRIBUTES (tree @var{node}, tree *@var{attr_ptr})
8845
Define this target hook if you want to be able to add attributes to a decl
8846
when it is being created.  This is normally useful for back ends which
8847
wish to implement a pragma by using the attributes which correspond to
8848
the pragma's effect.  The @var{node} argument is the decl which is being
8849
created.  The @var{attr_ptr} argument is a pointer to the attribute list
8850
for this decl.  The list itself should not be modified, since it may be
8851
shared with other decls, but attributes may be chained on the head of
8852
the list and @code{*@var{attr_ptr}} modified to point to the new
8853
attributes, or a copy of the list may be made if further changes are
8854
needed.
8855
@end deftypefn
8856
 
8857
@deftypefn {Target Hook} bool TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P (tree @var{fndecl})
8858
@cindex inlining
8859
This target hook returns @code{true} if it is ok to inline @var{fndecl}
8860
into the current function, despite its having target-specific
8861
attributes, @code{false} otherwise.  By default, if a function has a
8862
target specific attribute attached to it, it will not be inlined.
8863
@end deftypefn
8864
 
8865
@node MIPS Coprocessors
8866
@section Defining coprocessor specifics for MIPS targets.
8867
@cindex MIPS coprocessor-definition macros
8868
 
8869
The MIPS specification allows MIPS implementations to have as many as 4
8870
coprocessors, each with as many as 32 private registers.  GCC supports
8871
accessing these registers and transferring values between the registers
8872
and memory using asm-ized variables.  For example:
8873
 
8874
@smallexample
8875
  register unsigned int cp0count asm ("c0r1");
8876
  unsigned int d;
8877
 
8878
  d = cp0count + 3;
8879
@end smallexample
8880
 
8881
(``c0r1'' is the default name of register 1 in coprocessor 0; alternate
8882
names may be added as described below, or the default names may be
8883
overridden entirely in @code{SUBTARGET_CONDITIONAL_REGISTER_USAGE}.)
8884
 
8885
Coprocessor registers are assumed to be epilogue-used; sets to them will
8886
be preserved even if it does not appear that the register is used again
8887
later in the function.
8888
 
8889
Another note: according to the MIPS spec, coprocessor 1 (if present) is
8890
the FPU@.  One accesses COP1 registers through standard mips
8891
floating-point support; they are not included in this mechanism.
8892
 
8893
There is one macro used in defining the MIPS coprocessor interface which
8894
you may want to override in subtargets; it is described below.
8895
 
8896
@defmac ALL_COP_ADDITIONAL_REGISTER_NAMES
8897
A comma-separated list (with leading comma) of pairs describing the
8898
alternate names of coprocessor registers.  The format of each entry should be
8899
@smallexample
8900
@{ @var{alternatename}, @var{register_number}@}
8901
@end smallexample
8902
Default: empty.
8903
@end defmac
8904
 
8905
@node PCH Target
8906
@section Parameters for Precompiled Header Validity Checking
8907
@cindex parameters, precompiled headers
8908
 
8909
@deftypefn {Target Hook} void *TARGET_GET_PCH_VALIDITY (size_t *@var{sz})
8910
This hook returns the data needed by @code{TARGET_PCH_VALID_P} and sets
8911
@samp{*@var{sz}} to the size of the data in bytes.
8912
@end deftypefn
8913
 
8914
@deftypefn {Target Hook} const char *TARGET_PCH_VALID_P (const void *@var{data}, size_t @var{sz})
8915
This hook checks whether the options used to create a PCH file are
8916
compatible with the current settings.  It returns @code{NULL}
8917
if so and a suitable error message if not.  Error messages will
8918
be presented to the user and must be localized using @samp{_(@var{msg})}.
8919
 
8920
@var{data} is the data that was returned by @code{TARGET_GET_PCH_VALIDITY}
8921
when the PCH file was created and @var{sz} is the size of that data in bytes.
8922
It's safe to assume that the data was created by the same version of the
8923
compiler, so no format checking is needed.
8924
 
8925
The default definition of @code{default_pch_valid_p} should be
8926
suitable for most targets.
8927
@end deftypefn
8928
 
8929
@deftypefn {Target Hook} const char *TARGET_CHECK_PCH_TARGET_FLAGS (int @var{pch_flags})
8930
If this hook is nonnull, the default implementation of
8931
@code{TARGET_PCH_VALID_P} will use it to check for compatible values
8932
of @code{target_flags}.  @var{pch_flags} specifies the value that
8933
@code{target_flags} had when the PCH file was created.  The return
8934
value is the same as for @code{TARGET_PCH_VALID_P}.
8935
@end deftypefn
8936
 
8937
@node C++ ABI
8938
@section C++ ABI parameters
8939
@cindex parameters, c++ abi
8940
 
8941
@deftypefn {Target Hook} tree TARGET_CXX_GUARD_TYPE (void)
8942
Define this hook to override the integer type used for guard variables.
8943
These are used to implement one-time construction of static objects.  The
8944
default is long_long_integer_type_node.
8945
@end deftypefn
8946
 
8947
@deftypefn {Target Hook} bool TARGET_CXX_GUARD_MASK_BIT (void)
8948
This hook determines how guard variables are used.  It should return
8949
@code{false} (the default) if first byte should be used.  A return value of
8950
@code{true} indicates the least significant bit should be used.
8951
@end deftypefn
8952
 
8953
@deftypefn {Target Hook} tree TARGET_CXX_GET_COOKIE_SIZE (tree @var{type})
8954
This hook returns the size of the cookie to use when allocating an array
8955
whose elements have the indicated @var{type}.  Assumes that it is already
8956
known that a cookie is needed.  The default is
8957
@code{max(sizeof (size_t), alignof(type))}, as defined in section 2.7 of the
8958
IA64/Generic C++ ABI@.
8959
@end deftypefn
8960
 
8961
@deftypefn {Target Hook} bool TARGET_CXX_COOKIE_HAS_SIZE (void)
8962
This hook should return @code{true} if the element size should be stored in
8963
array cookies.  The default is to return @code{false}.
8964
@end deftypefn
8965
 
8966
@deftypefn {Target Hook} int TARGET_CXX_IMPORT_EXPORT_CLASS (tree  @var{type}, int @var{import_export})
8967
If defined by a backend this hook allows the decision made to export
8968
class @var{type} to be overruled.  Upon entry @var{import_export}
8969
will contain 1 if the class is going to be exported, @minus{}1 if it is going
8970
to be imported and 0 otherwise.  This function should return the
8971
modified value and perform any other actions necessary to support the
8972
backend's targeted operating system.
8973
@end deftypefn
8974
 
8975
@deftypefn {Target Hook} bool TARGET_CXX_CDTOR_RETURNS_THIS (void)
8976
This hook should return @code{true} if constructors and destructors return
8977
the address of the object created/destroyed.  The default is to return
8978
@code{false}.
8979
@end deftypefn
8980
 
8981
@deftypefn {Target Hook} bool TARGET_CXX_KEY_METHOD_MAY_BE_INLINE (void)
8982
This hook returns true if the key method for a class (i.e., the method
8983
which, if defined in the current translation unit, causes the virtual
8984
table to be emitted) may be an inline function.  Under the standard
8985
Itanium C++ ABI the key method may be an inline function so long as
8986
the function is not declared inline in the class definition.  Under
8987
some variants of the ABI, an inline function can never be the key
8988
method.  The default is to return @code{true}.
8989
@end deftypefn
8990
 
8991
@deftypefn {Target Hook} void TARGET_CXX_DETERMINE_CLASS_DATA_VISIBILITY (tree @var{decl})
8992
@var{decl} is a virtual table, virtual table table, typeinfo object,
8993
or other similar implicit class data object that will be emitted with
8994
external linkage in this translation unit.  No ELF visibility has been
8995
explicitly specified.  If the target needs to specify a visibility
8996
other than that of the containing class, use this hook to set
8997
@code{DECL_VISIBILITY} and @code{DECL_VISIBILITY_SPECIFIED}.
8998
@end deftypefn
8999
 
9000
@deftypefn {Target Hook} bool TARGET_CXX_CLASS_DATA_ALWAYS_COMDAT (void)
9001
This hook returns true (the default) if virtual tables and other
9002
similar implicit class data objects are always COMDAT if they have
9003
external linkage.  If this hook returns false, then class data for
9004
classes whose virtual table will be emitted in only one translation
9005
unit will not be COMDAT.
9006
@end deftypefn
9007
 
9008
@deftypefn {Target Hook} bool TARGET_CXX_USE_AEABI_ATEXIT (void)
9009
This hook returns true if @code{__aeabi_atexit} (as defined by the ARM EABI)
9010
should be used to register static destructors when @option{-fuse-cxa-atexit}
9011
is in effect.  The default is to return false to use @code{__cxa_atexit}.
9012
@end deftypefn
9013
 
9014
@deftypefn {Target Hook} void TARGET_CXX_ADJUST_CLASS_AT_DEFINITION (tree @var{type})
9015
@var{type} is a C++ class (i.e., RECORD_TYPE or UNION_TYPE) that has just been
9016
defined.  Use this hook to make adjustments to the class (eg, tweak
9017
visibility or perform any other required target modifications).
9018
@end deftypefn
9019
 
9020
@node Misc
9021
@section Miscellaneous Parameters
9022
@cindex parameters, miscellaneous
9023
 
9024
@c prevent bad page break with this line
9025
Here are several miscellaneous parameters.
9026
 
9027
@defmac HAS_LONG_COND_BRANCH
9028
Define this boolean macro to indicate whether or not your architecture
9029
has conditional branches that can span all of memory.  It is used in
9030
conjunction with an optimization that partitions hot and cold basic
9031
blocks into separate sections of the executable.  If this macro is
9032
set to false, gcc will convert any conditional branches that attempt
9033
to cross between sections into unconditional branches or indirect jumps.
9034
@end defmac
9035
 
9036
@defmac HAS_LONG_UNCOND_BRANCH
9037
Define this boolean macro to indicate whether or not your architecture
9038
has unconditional branches that can span all of memory.  It is used in
9039
conjunction with an optimization that partitions hot and cold basic
9040
blocks into separate sections of the executable.  If this macro is
9041
set to false, gcc will convert any unconditional branches that attempt
9042
to cross between sections into indirect jumps.
9043
@end defmac
9044
 
9045
@defmac CASE_VECTOR_MODE
9046
An alias for a machine mode name.  This is the machine mode that
9047
elements of a jump-table should have.
9048
@end defmac
9049
 
9050
@defmac CASE_VECTOR_SHORTEN_MODE (@var{min_offset}, @var{max_offset}, @var{body})
9051
Optional: return the preferred mode for an @code{addr_diff_vec}
9052
when the minimum and maximum offset are known.  If you define this,
9053
it enables extra code in branch shortening to deal with @code{addr_diff_vec}.
9054
To make this work, you also have to define @code{INSN_ALIGN} and
9055
make the alignment for @code{addr_diff_vec} explicit.
9056
The @var{body} argument is provided so that the offset_unsigned and scale
9057
flags can be updated.
9058
@end defmac
9059
 
9060
@defmac CASE_VECTOR_PC_RELATIVE
9061
Define this macro to be a C expression to indicate when jump-tables
9062
should contain relative addresses.  You need not define this macro if
9063
jump-tables never contain relative addresses, or jump-tables should
9064
contain relative addresses only when @option{-fPIC} or @option{-fPIC}
9065
is in effect.
9066
@end defmac
9067
 
9068
@defmac CASE_VALUES_THRESHOLD
9069
Define this to be the smallest number of different values for which it
9070
is best to use a jump-table instead of a tree of conditional branches.
9071
The default is four for machines with a @code{casesi} instruction and
9072
five otherwise.  This is best for most machines.
9073
@end defmac
9074
 
9075
@defmac CASE_USE_BIT_TESTS
9076
Define this macro to be a C expression to indicate whether C switch
9077
statements may be implemented by a sequence of bit tests.  This is
9078
advantageous on processors that can efficiently implement left shift
9079
of 1 by the number of bits held in a register, but inappropriate on
9080
targets that would require a loop.  By default, this macro returns
9081
@code{true} if the target defines an @code{ashlsi3} pattern, and
9082
@code{false} otherwise.
9083
@end defmac
9084
 
9085
@defmac WORD_REGISTER_OPERATIONS
9086
Define this macro if operations between registers with integral mode
9087
smaller than a word are always performed on the entire register.
9088
Most RISC machines have this property and most CISC machines do not.
9089
@end defmac
9090
 
9091
@defmac LOAD_EXTEND_OP (@var{mem_mode})
9092
Define this macro to be a C expression indicating when insns that read
9093
memory in @var{mem_mode}, an integral mode narrower than a word, set the
9094
bits outside of @var{mem_mode} to be either the sign-extension or the
9095
zero-extension of the data read.  Return @code{SIGN_EXTEND} for values
9096
of @var{mem_mode} for which the
9097
insn sign-extends, @code{ZERO_EXTEND} for which it zero-extends, and
9098
@code{UNKNOWN} for other modes.
9099
 
9100
This macro is not called with @var{mem_mode} non-integral or with a width
9101
greater than or equal to @code{BITS_PER_WORD}, so you may return any
9102
value in this case.  Do not define this macro if it would always return
9103
@code{UNKNOWN}.  On machines where this macro is defined, you will normally
9104
define it as the constant @code{SIGN_EXTEND} or @code{ZERO_EXTEND}.
9105
 
9106
You may return a non-@code{UNKNOWN} value even if for some hard registers
9107
the sign extension is not performed, if for the @code{REGNO_REG_CLASS}
9108
of these hard registers @code{CANNOT_CHANGE_MODE_CLASS} returns nonzero
9109
when the @var{from} mode is @var{mem_mode} and the @var{to} mode is any
9110
integral mode larger than this but not larger than @code{word_mode}.
9111
 
9112
You must return @code{UNKNOWN} if for some hard registers that allow this
9113
mode, @code{CANNOT_CHANGE_MODE_CLASS} says that they cannot change to
9114
@code{word_mode}, but that they can change to another integral mode that
9115
is larger then @var{mem_mode} but still smaller than @code{word_mode}.
9116
@end defmac
9117
 
9118
@defmac SHORT_IMMEDIATES_SIGN_EXTEND
9119
Define this macro if loading short immediate values into registers sign
9120
extends.
9121
@end defmac
9122
 
9123
@defmac FIXUNS_TRUNC_LIKE_FIX_TRUNC
9124
Define this macro if the same instructions that convert a floating
9125
point number to a signed fixed point number also convert validly to an
9126
unsigned one.
9127
@end defmac
9128
 
9129
@deftypefn {Target Hook} int TARGET_MIN_DIVISIONS_FOR_RECIP_MUL (enum machine_mode @var{mode})
9130
When @option{-ffast-math} is in effect, GCC tries to optimize
9131
divisions by the same divisor, by turning them into multiplications by
9132
the reciprocal.  This target hook specifies the minimum number of divisions
9133
that should be there for GCC to perform the optimization for a variable
9134
of mode @var{mode}.  The default implementation returns 3 if the machine
9135
has an instruction for the division, and 2 if it does not.
9136
@end deftypefn
9137
 
9138
@defmac MOVE_MAX
9139
The maximum number of bytes that a single instruction can move quickly
9140
between memory and registers or between two memory locations.
9141
@end defmac
9142
 
9143
@defmac MAX_MOVE_MAX
9144
The maximum number of bytes that a single instruction can move quickly
9145
between memory and registers or between two memory locations.  If this
9146
is undefined, the default is @code{MOVE_MAX}.  Otherwise, it is the
9147
constant value that is the largest value that @code{MOVE_MAX} can have
9148
at run-time.
9149
@end defmac
9150
 
9151
@defmac SHIFT_COUNT_TRUNCATED
9152
A C expression that is nonzero if on this machine the number of bits
9153
actually used for the count of a shift operation is equal to the number
9154
of bits needed to represent the size of the object being shifted.  When
9155
this macro is nonzero, the compiler will assume that it is safe to omit
9156
a sign-extend, zero-extend, and certain bitwise `and' instructions that
9157
truncates the count of a shift operation.  On machines that have
9158
instructions that act on bit-fields at variable positions, which may
9159
include `bit test' instructions, a nonzero @code{SHIFT_COUNT_TRUNCATED}
9160
also enables deletion of truncations of the values that serve as
9161
arguments to bit-field instructions.
9162
 
9163
If both types of instructions truncate the count (for shifts) and
9164
position (for bit-field operations), or if no variable-position bit-field
9165
instructions exist, you should define this macro.
9166
 
9167
However, on some machines, such as the 80386 and the 680x0, truncation
9168
only applies to shift operations and not the (real or pretended)
9169
bit-field operations.  Define @code{SHIFT_COUNT_TRUNCATED} to be zero on
9170
such machines.  Instead, add patterns to the @file{md} file that include
9171
the implied truncation of the shift instructions.
9172
 
9173
You need not define this macro if it would always have the value of zero.
9174
@end defmac
9175
 
9176
@anchor{TARGET_SHIFT_TRUNCATION_MASK}
9177
@deftypefn {Target Hook} int TARGET_SHIFT_TRUNCATION_MASK (enum machine_mode @var{mode})
9178
This function describes how the standard shift patterns for @var{mode}
9179
deal with shifts by negative amounts or by more than the width of the mode.
9180
@xref{shift patterns}.
9181
 
9182
On many machines, the shift patterns will apply a mask @var{m} to the
9183
shift count, meaning that a fixed-width shift of @var{x} by @var{y} is
9184
equivalent to an arbitrary-width shift of @var{x} by @var{y & m}.  If
9185
this is true for mode @var{mode}, the function should return @var{m},
9186
otherwise it should return 0.  A return value of 0 indicates that no
9187
particular behavior is guaranteed.
9188
 
9189
Note that, unlike @code{SHIFT_COUNT_TRUNCATED}, this function does
9190
@emph{not} apply to general shift rtxes; it applies only to instructions
9191
that are generated by the named shift patterns.
9192
 
9193
The default implementation of this function returns
9194
@code{GET_MODE_BITSIZE (@var{mode}) - 1} if @code{SHIFT_COUNT_TRUNCATED}
9195
and 0 otherwise.  This definition is always safe, but if
9196
@code{SHIFT_COUNT_TRUNCATED} is false, and some shift patterns
9197
nevertheless truncate the shift count, you may get better code
9198
by overriding it.
9199
@end deftypefn
9200
 
9201
@defmac TRULY_NOOP_TRUNCATION (@var{outprec}, @var{inprec})
9202
A C expression which is nonzero if on this machine it is safe to
9203
``convert'' an integer of @var{inprec} bits to one of @var{outprec}
9204
bits (where @var{outprec} is smaller than @var{inprec}) by merely
9205
operating on it as if it had only @var{outprec} bits.
9206
 
9207
On many machines, this expression can be 1.
9208
 
9209
@c rearranged this, removed the phrase "it is reported that".  this was
9210
@c to fix an overfull hbox.  --mew 10feb93
9211
When @code{TRULY_NOOP_TRUNCATION} returns 1 for a pair of sizes for
9212
modes for which @code{MODES_TIEABLE_P} is 0, suboptimal code can result.
9213
If this is the case, making @code{TRULY_NOOP_TRUNCATION} return 0 in
9214
such cases may improve things.
9215
@end defmac
9216
 
9217
@deftypefn {Target Hook} int TARGET_MODE_REP_EXTENDED (enum machine_mode @var{mode}, enum machine_mode @var{rep_mode})
9218
The representation of an integral mode can be such that the values
9219
are always extended to a wider integral mode.  Return
9220
@code{SIGN_EXTEND} if values of @var{mode} are represented in
9221
sign-extended form to @var{rep_mode}.  Return @code{UNKNOWN}
9222
otherwise.  (Currently, none of the targets use zero-extended
9223
representation this way so unlike @code{LOAD_EXTEND_OP},
9224
@code{TARGET_MODE_REP_EXTENDED} is expected to return either
9225
@code{SIGN_EXTEND} or @code{UNKNOWN}.  Also no target extends
9226
@var{mode} to @var{mode_rep} so that @var{mode_rep} is not the next
9227
widest integral mode and currently we take advantage of this fact.)
9228
 
9229
Similarly to @code{LOAD_EXTEND_OP} you may return a non-@code{UNKNOWN}
9230
value even if the extension is not performed on certain hard registers
9231
as long as for the @code{REGNO_REG_CLASS} of these hard registers
9232
@code{CANNOT_CHANGE_MODE_CLASS} returns nonzero.
9233
 
9234
Note that @code{TARGET_MODE_REP_EXTENDED} and @code{LOAD_EXTEND_OP}
9235
describe two related properties.  If you define
9236
@code{TARGET_MODE_REP_EXTENDED (mode, word_mode)} you probably also want
9237
to define @code{LOAD_EXTEND_OP (mode)} to return the same type of
9238
extension.
9239
 
9240
In order to enforce the representation of @code{mode},
9241
@code{TRULY_NOOP_TRUNCATION} should return false when truncating to
9242
@code{mode}.
9243
@end deftypefn
9244
 
9245
@defmac STORE_FLAG_VALUE
9246
A C expression describing the value returned by a comparison operator
9247
with an integral mode and stored by a store-flag instruction
9248
(@samp{s@var{cond}}) when the condition is true.  This description must
9249
apply to @emph{all} the @samp{s@var{cond}} patterns and all the
9250
comparison operators whose results have a @code{MODE_INT} mode.
9251
 
9252
A value of 1 or @minus{}1 means that the instruction implementing the
9253
comparison operator returns exactly 1 or @minus{}1 when the comparison is true
9254
and 0 when the comparison is false.  Otherwise, the value indicates
9255
which bits of the result are guaranteed to be 1 when the comparison is
9256
true.  This value is interpreted in the mode of the comparison
9257
operation, which is given by the mode of the first operand in the
9258
@samp{s@var{cond}} pattern.  Either the low bit or the sign bit of
9259
@code{STORE_FLAG_VALUE} be on.  Presently, only those bits are used by
9260
the compiler.
9261
 
9262
If @code{STORE_FLAG_VALUE} is neither 1 or @minus{}1, the compiler will
9263
generate code that depends only on the specified bits.  It can also
9264
replace comparison operators with equivalent operations if they cause
9265
the required bits to be set, even if the remaining bits are undefined.
9266
For example, on a machine whose comparison operators return an
9267
@code{SImode} value and where @code{STORE_FLAG_VALUE} is defined as
9268
@samp{0x80000000}, saying that just the sign bit is relevant, the
9269
expression
9270
 
9271
@smallexample
9272
(ne:SI (and:SI @var{x} (const_int @var{power-of-2})) (const_int 0))
9273
@end smallexample
9274
 
9275
@noindent
9276
can be converted to
9277
 
9278
@smallexample
9279
(ashift:SI @var{x} (const_int @var{n}))
9280
@end smallexample
9281
 
9282
@noindent
9283
where @var{n} is the appropriate shift count to move the bit being
9284
tested into the sign bit.
9285
 
9286
There is no way to describe a machine that always sets the low-order bit
9287
for a true value, but does not guarantee the value of any other bits,
9288
but we do not know of any machine that has such an instruction.  If you
9289
are trying to port GCC to such a machine, include an instruction to
9290
perform a logical-and of the result with 1 in the pattern for the
9291
comparison operators and let us know at @email{gcc@@gcc.gnu.org}.
9292
 
9293
Often, a machine will have multiple instructions that obtain a value
9294
from a comparison (or the condition codes).  Here are rules to guide the
9295
choice of value for @code{STORE_FLAG_VALUE}, and hence the instructions
9296
to be used:
9297
 
9298
@itemize @bullet
9299
@item
9300
Use the shortest sequence that yields a valid definition for
9301
@code{STORE_FLAG_VALUE}.  It is more efficient for the compiler to
9302
``normalize'' the value (convert it to, e.g., 1 or 0) than for the
9303
comparison operators to do so because there may be opportunities to
9304
combine the normalization with other operations.
9305
 
9306
@item
9307
For equal-length sequences, use a value of 1 or @minus{}1, with @minus{}1 being
9308
slightly preferred on machines with expensive jumps and 1 preferred on
9309
other machines.
9310
 
9311
@item
9312
As a second choice, choose a value of @samp{0x80000001} if instructions
9313
exist that set both the sign and low-order bits but do not define the
9314
others.
9315
 
9316
@item
9317
Otherwise, use a value of @samp{0x80000000}.
9318
@end itemize
9319
 
9320
Many machines can produce both the value chosen for
9321
@code{STORE_FLAG_VALUE} and its negation in the same number of
9322
instructions.  On those machines, you should also define a pattern for
9323
those cases, e.g., one matching
9324
 
9325
@smallexample
9326
(set @var{A} (neg:@var{m} (ne:@var{m} @var{B} @var{C})))
9327
@end smallexample
9328
 
9329
Some machines can also perform @code{and} or @code{plus} operations on
9330
condition code values with less instructions than the corresponding
9331
@samp{s@var{cond}} insn followed by @code{and} or @code{plus}.  On those
9332
machines, define the appropriate patterns.  Use the names @code{incscc}
9333
and @code{decscc}, respectively, for the patterns which perform
9334
@code{plus} or @code{minus} operations on condition code values.  See
9335
@file{rs6000.md} for some examples.  The GNU Superoptizer can be used to
9336
find such instruction sequences on other machines.
9337
 
9338
If this macro is not defined, the default value, 1, is used.  You need
9339
not define @code{STORE_FLAG_VALUE} if the machine has no store-flag
9340
instructions, or if the value generated by these instructions is 1.
9341
@end defmac
9342
 
9343
@defmac FLOAT_STORE_FLAG_VALUE (@var{mode})
9344
A C expression that gives a nonzero @code{REAL_VALUE_TYPE} value that is
9345
returned when comparison operators with floating-point results are true.
9346
Define this macro on machines that have comparison operations that return
9347
floating-point values.  If there are no such operations, do not define
9348
this macro.
9349
@end defmac
9350
 
9351
@defmac VECTOR_STORE_FLAG_VALUE (@var{mode})
9352
A C expression that gives a rtx representing the nonzero true element
9353
for vector comparisons.  The returned rtx should be valid for the inner
9354
mode of @var{mode} which is guaranteed to be a vector mode.  Define
9355
this macro on machines that have vector comparison operations that
9356
return a vector result.  If there are no such operations, do not define
9357
this macro.  Typically, this macro is defined as @code{const1_rtx} or
9358
@code{constm1_rtx}.  This macro may return @code{NULL_RTX} to prevent
9359
the compiler optimizing such vector comparison operations for the
9360
given mode.
9361
@end defmac
9362
 
9363
@defmac CLZ_DEFINED_VALUE_AT_ZERO (@var{mode}, @var{value})
9364
@defmacx CTZ_DEFINED_VALUE_AT_ZERO (@var{mode}, @var{value})
9365
A C expression that evaluates to true if the architecture defines a value
9366
for @code{clz} or @code{ctz} with a zero operand.  If so, @var{value}
9367
should be set to this value.  If this macro is not defined, the value of
9368
@code{clz} or @code{ctz} is assumed to be undefined.
9369
 
9370
This macro must be defined if the target's expansion for @code{ffs}
9371
relies on a particular value to get correct results.  Otherwise it
9372
is not necessary, though it may be used to optimize some corner cases.
9373
 
9374
Note that regardless of this macro the ``definedness'' of @code{clz}
9375
and @code{ctz} at zero do @emph{not} extend to the builtin functions
9376
visible to the user.  Thus one may be free to adjust the value at will
9377
to match the target expansion of these operations without fear of
9378
breaking the API@.
9379
@end defmac
9380
 
9381
@defmac Pmode
9382
An alias for the machine mode for pointers.  On most machines, define
9383
this to be the integer mode corresponding to the width of a hardware
9384
pointer; @code{SImode} on 32-bit machine or @code{DImode} on 64-bit machines.
9385
On some machines you must define this to be one of the partial integer
9386
modes, such as @code{PSImode}.
9387
 
9388
The width of @code{Pmode} must be at least as large as the value of
9389
@code{POINTER_SIZE}.  If it is not equal, you must define the macro
9390
@code{POINTERS_EXTEND_UNSIGNED} to specify how pointers are extended
9391
to @code{Pmode}.
9392
@end defmac
9393
 
9394
@defmac FUNCTION_MODE
9395
An alias for the machine mode used for memory references to functions
9396
being called, in @code{call} RTL expressions.  On most machines this
9397
should be @code{QImode}.
9398
@end defmac
9399
 
9400
@defmac STDC_0_IN_SYSTEM_HEADERS
9401
In normal operation, the preprocessor expands @code{__STDC__} to the
9402
constant 1, to signify that GCC conforms to ISO Standard C@.  On some
9403
hosts, like Solaris, the system compiler uses a different convention,
9404
where @code{__STDC__} is normally 0, but is 1 if the user specifies
9405
strict conformance to the C Standard.
9406
 
9407
Defining @code{STDC_0_IN_SYSTEM_HEADERS} makes GNU CPP follows the host
9408
convention when processing system header files, but when processing user
9409
files @code{__STDC__} will always expand to 1.
9410
@end defmac
9411
 
9412
@defmac NO_IMPLICIT_EXTERN_C
9413
Define this macro if the system header files support C++ as well as C@.
9414
This macro inhibits the usual method of using system header files in
9415
C++, which is to pretend that the file's contents are enclosed in
9416
@samp{extern "C" @{@dots{}@}}.
9417
@end defmac
9418
 
9419
@findex #pragma
9420
@findex pragma
9421
@defmac REGISTER_TARGET_PRAGMAS ()
9422
Define this macro if you want to implement any target-specific pragmas.
9423
If defined, it is a C expression which makes a series of calls to
9424
@code{c_register_pragma} or @code{c_register_pragma_with_expansion}
9425
for each pragma.  The macro may also do any
9426
setup required for the pragmas.
9427
 
9428
The primary reason to define this macro is to provide compatibility with
9429
other compilers for the same target.  In general, we discourage
9430
definition of target-specific pragmas for GCC@.
9431
 
9432
If the pragma can be implemented by attributes then you should consider
9433
defining the target hook @samp{TARGET_INSERT_ATTRIBUTES} as well.
9434
 
9435
Preprocessor macros that appear on pragma lines are not expanded.  All
9436
@samp{#pragma} directives that do not match any registered pragma are
9437
silently ignored, unless the user specifies @option{-Wunknown-pragmas}.
9438
@end defmac
9439
 
9440
@deftypefun void c_register_pragma (const char *@var{space}, const char *@var{name}, void (*@var{callback}) (struct cpp_reader *))
9441
@deftypefunx void c_register_pragma_with_expansion (const char *@var{space}, const char *@var{name}, void (*@var{callback}) (struct cpp_reader *))
9442
 
9443
Each call to @code{c_register_pragma} or
9444
@code{c_register_pragma_with_expansion} establishes one pragma.  The
9445
@var{callback} routine will be called when the preprocessor encounters a
9446
pragma of the form
9447
 
9448
@smallexample
9449
#pragma [@var{space}] @var{name} @dots{}
9450
@end smallexample
9451
 
9452
@var{space} is the case-sensitive namespace of the pragma, or
9453
@code{NULL} to put the pragma in the global namespace.  The callback
9454
routine receives @var{pfile} as its first argument, which can be passed
9455
on to cpplib's functions if necessary.  You can lex tokens after the
9456
@var{name} by calling @code{pragma_lex}.  Tokens that are not read by the
9457
callback will be silently ignored.  The end of the line is indicated by
9458
a token of type @code{CPP_EOF}.  Macro expansion occurs on the
9459
arguments of pragmas registered with
9460
@code{c_register_pragma_with_expansion} but not on the arguments of
9461
pragmas registered with @code{c_register_pragma}.
9462
 
9463
For an example use of this routine, see @file{c4x.h} and the callback
9464
routines defined in @file{c4x-c.c}.
9465
 
9466
Note that the use of @code{pragma_lex} is specific to the C and C++
9467
compilers.  It will not work in the Java or Fortran compilers, or any
9468
other language compilers for that matter.  Thus if @code{pragma_lex} is going
9469
to be called from target-specific code, it must only be done so when
9470
building the C and C++ compilers.  This can be done by defining the
9471
variables @code{c_target_objs} and @code{cxx_target_objs} in the
9472
target entry in the @file{config.gcc} file.  These variables should name
9473
the target-specific, language-specific object file which contains the
9474
code that uses @code{pragma_lex}.  Note it will also be necessary to add a
9475
rule to the makefile fragment pointed to by @code{tmake_file} that shows
9476
how to build this object file.
9477
@end deftypefun
9478
 
9479
@findex #pragma
9480
@findex pragma
9481
@defmac HANDLE_SYSV_PRAGMA
9482
Define this macro (to a value of 1) if you want the System V style
9483
pragmas @samp{#pragma pack(<n>)} and @samp{#pragma weak <name>
9484
[=<value>]} to be supported by gcc.
9485
 
9486
The pack pragma specifies the maximum alignment (in bytes) of fields
9487
within a structure, in much the same way as the @samp{__aligned__} and
9488
@samp{__packed__} @code{__attribute__}s do.  A pack value of zero resets
9489
the behavior to the default.
9490
 
9491
A subtlety for Microsoft Visual C/C++ style bit-field packing
9492
(e.g.@: -mms-bitfields) for targets that support it:
9493
When a bit-field is inserted into a packed record, the whole size
9494
of the underlying type is used by one or more same-size adjacent
9495
bit-fields (that is, if its long:3, 32 bits is used in the record,
9496
and any additional adjacent long bit-fields are packed into the same
9497
chunk of 32 bits.  However, if the size changes, a new field of that
9498
size is allocated).
9499
 
9500
If both MS bit-fields and @samp{__attribute__((packed))} are used,
9501
the latter will take precedence.  If @samp{__attribute__((packed))} is
9502
used on a single field when MS bit-fields are in use, it will take
9503
precedence for that field, but the alignment of the rest of the structure
9504
may affect its placement.
9505
 
9506
The weak pragma only works if @code{SUPPORTS_WEAK} and
9507
@code{ASM_WEAKEN_LABEL} are defined.  If enabled it allows the creation
9508
of specifically named weak labels, optionally with a value.
9509
@end defmac
9510
 
9511
@findex #pragma
9512
@findex pragma
9513
@defmac HANDLE_PRAGMA_PACK_PUSH_POP
9514
Define this macro (to a value of 1) if you want to support the Win32
9515
style pragmas @samp{#pragma pack(push[,@var{n}])} and @samp{#pragma
9516
pack(pop)}.  The @samp{pack(push,[@var{n}])} pragma specifies the maximum
9517
alignment (in bytes) of fields within a structure, in much the same way as
9518
the @samp{__aligned__} and @samp{__packed__} @code{__attribute__}s do.  A
9519
pack value of zero resets the behavior to the default.  Successive
9520
invocations of this pragma cause the previous values to be stacked, so
9521
that invocations of @samp{#pragma pack(pop)} will return to the previous
9522
value.
9523
@end defmac
9524
 
9525
@defmac HANDLE_PRAGMA_PACK_WITH_EXPANSION
9526
Define this macro, as well as
9527
@code{HANDLE_SYSV_PRAGMA}, if macros should be expanded in the
9528
arguments of @samp{#pragma pack}.
9529
@end defmac
9530
 
9531
@defmac TARGET_DEFAULT_PACK_STRUCT
9532
If your target requires a structure packing default other than 0 (meaning
9533
the machine default), define this macro to the necessary value (in bytes).
9534
This must be a value that would also be valid to use with
9535
@samp{#pragma pack()} (that is, a small power of two).
9536
@end defmac
9537
 
9538
@defmac DOLLARS_IN_IDENTIFIERS
9539
Define this macro to control use of the character @samp{$} in
9540
identifier names for the C family of languages.  0 means @samp{$} is
9541
not allowed by default; 1 means it is allowed.  1 is the default;
9542
there is no need to define this macro in that case.
9543
@end defmac
9544
 
9545
@defmac NO_DOLLAR_IN_LABEL
9546
Define this macro if the assembler does not accept the character
9547
@samp{$} in label names.  By default constructors and destructors in
9548
G++ have @samp{$} in the identifiers.  If this macro is defined,
9549
@samp{.} is used instead.
9550
@end defmac
9551
 
9552
@defmac NO_DOT_IN_LABEL
9553
Define this macro if the assembler does not accept the character
9554
@samp{.} in label names.  By default constructors and destructors in G++
9555
have names that use @samp{.}.  If this macro is defined, these names
9556
are rewritten to avoid @samp{.}.
9557
@end defmac
9558
 
9559
@defmac INSN_SETS_ARE_DELAYED (@var{insn})
9560
Define this macro as a C expression that is nonzero if it is safe for the
9561
delay slot scheduler to place instructions in the delay slot of @var{insn},
9562
even if they appear to use a resource set or clobbered in @var{insn}.
9563
@var{insn} is always a @code{jump_insn} or an @code{insn}; GCC knows that
9564
every @code{call_insn} has this behavior.  On machines where some @code{insn}
9565
or @code{jump_insn} is really a function call and hence has this behavior,
9566
you should define this macro.
9567
 
9568
You need not define this macro if it would always return zero.
9569
@end defmac
9570
 
9571
@defmac INSN_REFERENCES_ARE_DELAYED (@var{insn})
9572
Define this macro as a C expression that is nonzero if it is safe for the
9573
delay slot scheduler to place instructions in the delay slot of @var{insn},
9574
even if they appear to set or clobber a resource referenced in @var{insn}.
9575
@var{insn} is always a @code{jump_insn} or an @code{insn}.  On machines where
9576
some @code{insn} or @code{jump_insn} is really a function call and its operands
9577
are registers whose use is actually in the subroutine it calls, you should
9578
define this macro.  Doing so allows the delay slot scheduler to move
9579
instructions which copy arguments into the argument registers into the delay
9580
slot of @var{insn}.
9581
 
9582
You need not define this macro if it would always return zero.
9583
@end defmac
9584
 
9585
@defmac MULTIPLE_SYMBOL_SPACES
9586
Define this macro as a C expression that is nonzero if, in some cases,
9587
global symbols from one translation unit may not be bound to undefined
9588
symbols in another translation unit without user intervention.  For
9589
instance, under Microsoft Windows symbols must be explicitly imported
9590
from shared libraries (DLLs).
9591
 
9592
You need not define this macro if it would always evaluate to zero.
9593
@end defmac
9594
 
9595
@deftypefn {Target Hook} tree TARGET_MD_ASM_CLOBBERS (tree @var{outputs}, tree @var{inputs}, tree @var{clobbers})
9596
This target hook should add to @var{clobbers} @code{STRING_CST} trees for
9597
any hard regs the port wishes to automatically clobber for an asm.
9598
It should return the result of the last @code{tree_cons} used to add a
9599
clobber.  The @var{outputs}, @var{inputs} and @var{clobber} lists are the
9600
corresponding parameters to the asm and may be inspected to avoid
9601
clobbering a register that is an input or output of the asm.  You can use
9602
@code{tree_overlaps_hard_reg_set}, declared in @file{tree.h}, to test
9603
for overlap with regards to asm-declared registers.
9604
@end deftypefn
9605
 
9606
@defmac MATH_LIBRARY
9607
Define this macro as a C string constant for the linker argument to link
9608
in the system math library, or @samp{""} if the target does not have a
9609
separate math library.
9610
 
9611
You need only define this macro if the default of @samp{"-lm"} is wrong.
9612
@end defmac
9613
 
9614
@defmac LIBRARY_PATH_ENV
9615
Define this macro as a C string constant for the environment variable that
9616
specifies where the linker should look for libraries.
9617
 
9618
You need only define this macro if the default of @samp{"LIBRARY_PATH"}
9619
is wrong.
9620
@end defmac
9621
 
9622
@defmac TARGET_POSIX_IO
9623
Define this macro if the target supports the following POSIX@ file
9624
functions, access, mkdir and  file locking with fcntl / F_SETLKW@.
9625
Defining @code{TARGET_POSIX_IO} will enable the test coverage code
9626
to use file locking when exiting a program, which avoids race conditions
9627
if the program has forked. It will also create directories at run-time
9628
for cross-profiling.
9629
@end defmac
9630
 
9631
@defmac MAX_CONDITIONAL_EXECUTE
9632
 
9633
A C expression for the maximum number of instructions to execute via
9634
conditional execution instructions instead of a branch.  A value of
9635
@code{BRANCH_COST}+1 is the default if the machine does not use cc0, and
9636
1 if it does use cc0.
9637
@end defmac
9638
 
9639
@defmac IFCVT_MODIFY_TESTS (@var{ce_info}, @var{true_expr}, @var{false_expr})
9640
Used if the target needs to perform machine-dependent modifications on the
9641
conditionals used for turning basic blocks into conditionally executed code.
9642
@var{ce_info} points to a data structure, @code{struct ce_if_block}, which
9643
contains information about the currently processed blocks.  @var{true_expr}
9644
and @var{false_expr} are the tests that are used for converting the
9645
then-block and the else-block, respectively.  Set either @var{true_expr} or
9646
@var{false_expr} to a null pointer if the tests cannot be converted.
9647
@end defmac
9648
 
9649
@defmac IFCVT_MODIFY_MULTIPLE_TESTS (@var{ce_info}, @var{bb}, @var{true_expr}, @var{false_expr})
9650
Like @code{IFCVT_MODIFY_TESTS}, but used when converting more complicated
9651
if-statements into conditions combined by @code{and} and @code{or} operations.
9652
@var{bb} contains the basic block that contains the test that is currently
9653
being processed and about to be turned into a condition.
9654
@end defmac
9655
 
9656
@defmac IFCVT_MODIFY_INSN (@var{ce_info}, @var{pattern}, @var{insn})
9657
A C expression to modify the @var{PATTERN} of an @var{INSN} that is to
9658
be converted to conditional execution format.  @var{ce_info} points to
9659
a data structure, @code{struct ce_if_block}, which contains information
9660
about the currently processed blocks.
9661
@end defmac
9662
 
9663
@defmac IFCVT_MODIFY_FINAL (@var{ce_info})
9664
A C expression to perform any final machine dependent modifications in
9665
converting code to conditional execution.  The involved basic blocks
9666
can be found in the @code{struct ce_if_block} structure that is pointed
9667
to by @var{ce_info}.
9668
@end defmac
9669
 
9670
@defmac IFCVT_MODIFY_CANCEL (@var{ce_info})
9671
A C expression to cancel any machine dependent modifications in
9672
converting code to conditional execution.  The involved basic blocks
9673
can be found in the @code{struct ce_if_block} structure that is pointed
9674
to by @var{ce_info}.
9675
@end defmac
9676
 
9677
@defmac IFCVT_INIT_EXTRA_FIELDS (@var{ce_info})
9678
A C expression to initialize any extra fields in a @code{struct ce_if_block}
9679
structure, which are defined by the @code{IFCVT_EXTRA_FIELDS} macro.
9680
@end defmac
9681
 
9682
@defmac IFCVT_EXTRA_FIELDS
9683
If defined, it should expand to a set of field declarations that will be
9684
added to the @code{struct ce_if_block} structure.  These should be initialized
9685
by the @code{IFCVT_INIT_EXTRA_FIELDS} macro.
9686
@end defmac
9687
 
9688
@deftypefn {Target Hook} void TARGET_MACHINE_DEPENDENT_REORG ()
9689
If non-null, this hook performs a target-specific pass over the
9690
instruction stream.  The compiler will run it at all optimization levels,
9691
just before the point at which it normally does delayed-branch scheduling.
9692
 
9693
The exact purpose of the hook varies from target to target.  Some use
9694
it to do transformations that are necessary for correctness, such as
9695
laying out in-function constant pools or avoiding hardware hazards.
9696
Others use it as an opportunity to do some machine-dependent optimizations.
9697
 
9698
You need not implement the hook if it has nothing to do.  The default
9699
definition is null.
9700
@end deftypefn
9701
 
9702
@deftypefn {Target Hook} void TARGET_INIT_BUILTINS ()
9703
Define this hook if you have any machine-specific built-in functions
9704
that need to be defined.  It should be a function that performs the
9705
necessary setup.
9706
 
9707
Machine specific built-in functions can be useful to expand special machine
9708
instructions that would otherwise not normally be generated because
9709
they have no equivalent in the source language (for example, SIMD vector
9710
instructions or prefetch instructions).
9711
 
9712
To create a built-in function, call the function
9713
@code{lang_hooks.builtin_function}
9714
which is defined by the language front end.  You can use any type nodes set
9715
up by @code{build_common_tree_nodes} and @code{build_common_tree_nodes_2};
9716
only language front ends that use those two functions will call
9717
@samp{TARGET_INIT_BUILTINS}.
9718
@end deftypefn
9719
 
9720
@deftypefn {Target Hook} rtx TARGET_EXPAND_BUILTIN (tree @var{exp}, rtx @var{target}, rtx @var{subtarget}, enum machine_mode @var{mode}, int @var{ignore})
9721
 
9722
Expand a call to a machine specific built-in function that was set up by
9723
@samp{TARGET_INIT_BUILTINS}.  @var{exp} is the expression for the
9724
function call; the result should go to @var{target} if that is
9725
convenient, and have mode @var{mode} if that is convenient.
9726
@var{subtarget} may be used as the target for computing one of
9727
@var{exp}'s operands.  @var{ignore} is nonzero if the value is to be
9728
ignored.  This function should return the result of the call to the
9729
built-in function.
9730
@end deftypefn
9731
 
9732
@deftypefn {Target Hook} tree TARGET_RESOLVE_OVERLOADED_BUILTIN (tree @var{fndecl}, tree @var{arglist})
9733
 
9734
Select a replacement for a machine specific built-in function that
9735
was set up by @samp{TARGET_INIT_BUILTINS}.  This is done
9736
@emph{before} regular type checking, and so allows the target to
9737
implement a crude form of function overloading.  @var{fndecl} is the
9738
declaration of the built-in function.  @var{arglist} is the list of
9739
arguments passed to the built-in function.  The result is a
9740
complete expression that implements the operation, usually
9741
another @code{CALL_EXPR}.
9742
@end deftypefn
9743
 
9744
@deftypefn {Target Hook} tree TARGET_FOLD_BUILTIN (tree @var{fndecl}, tree @var{arglist}, bool @var{ignore})
9745
 
9746
Fold a call to a machine specific built-in function that was set up by
9747
@samp{TARGET_INIT_BUILTINS}.  @var{fndecl} is the declaration of the
9748
built-in function.  @var{arglist} is the list of arguments passed to
9749
the built-in function.  The result is another tree containing a
9750
simplified expression for the call's result.  If @var{ignore} is true
9751
the value will be ignored.
9752
@end deftypefn
9753
 
9754
@deftypefn {Target Hook} const char * TARGET_INVALID_WITHIN_DOLOOP (rtx @var{insn})
9755
 
9756
Take an instruction in @var{insn} and return NULL if it is valid within a
9757
low-overhead loop, otherwise return a string why doloop could not be applied.
9758
 
9759
Many targets use special registers for low-overhead looping. For any
9760
instruction that clobbers these this function should return a string indicating
9761
the reason why the doloop could not be applied.
9762
By default, the RTL loop optimizer does not use a present doloop pattern for
9763
loops containing function calls or branch on table instructions.
9764
@end deftypefn
9765
 
9766
@defmac MD_CAN_REDIRECT_BRANCH (@var{branch1}, @var{branch2})
9767
 
9768
Take a branch insn in @var{branch1} and another in @var{branch2}.
9769
Return true if redirecting @var{branch1} to the destination of
9770
@var{branch2} is possible.
9771
 
9772
On some targets, branches may have a limited range.  Optimizing the
9773
filling of delay slots can result in branches being redirected, and this
9774
may in turn cause a branch offset to overflow.
9775
@end defmac
9776
 
9777
@deftypefn {Target Hook} bool TARGET_COMMUTATIVE_P (rtx @var{x}, @var{outer_code})
9778
This target hook returns @code{true} if @var{x} is considered to be commutative.
9779
Usually, this is just COMMUTATIVE_P (@var{x}), but the HP PA doesn't consider
9780
PLUS to be commutative inside a MEM.  @var{outer_code} is the rtx code
9781
of the enclosing rtl, if known, otherwise it is UNKNOWN.
9782
@end deftypefn
9783
 
9784
@deftypefn {Target Hook} rtx TARGET_ALLOCATE_INITIAL_VALUE (rtx @var{hard_reg})
9785
 
9786
When the initial value of a hard register has been copied in a pseudo
9787
register, it is often not necessary to actually allocate another register
9788
to this pseudo register, because the original hard register or a stack slot
9789
it has been saved into can be used.  @code{TARGET_ALLOCATE_INITIAL_VALUE}
9790
is called at the start of register allocation once for each hard register
9791
that had its initial value copied by using
9792
@code{get_func_hard_reg_initial_val} or @code{get_hard_reg_initial_val}.
9793
Possible values are @code{NULL_RTX}, if you don't want
9794
to do any special allocation, a @code{REG} rtx---that would typically be
9795
the hard register itself, if it is known not to be clobbered---or a
9796
@code{MEM}.
9797
If you are returning a @code{MEM}, this is only a hint for the allocator;
9798
it might decide to use another register anyways.
9799
You may use @code{current_function_leaf_function} in the hook, functions
9800
that use @code{REG_N_SETS}, to determine if the hard
9801
register in question will not be clobbered.
9802
The default value of this hook is @code{NULL}, which disables any special
9803
allocation.
9804
@end deftypefn
9805
 
9806
@defmac TARGET_OBJECT_SUFFIX
9807
Define this macro to be a C string representing the suffix for object
9808
files on your target machine.  If you do not define this macro, GCC will
9809
use @samp{.o} as the suffix for object files.
9810
@end defmac
9811
 
9812
@defmac TARGET_EXECUTABLE_SUFFIX
9813
Define this macro to be a C string representing the suffix to be
9814
automatically added to executable files on your target machine.  If you
9815
do not define this macro, GCC will use the null string as the suffix for
9816
executable files.
9817
@end defmac
9818
 
9819
@defmac COLLECT_EXPORT_LIST
9820
If defined, @code{collect2} will scan the individual object files
9821
specified on its command line and create an export list for the linker.
9822
Define this macro for systems like AIX, where the linker discards
9823
object files that are not referenced from @code{main} and uses export
9824
lists.
9825
@end defmac
9826
 
9827
@defmac MODIFY_JNI_METHOD_CALL (@var{mdecl})
9828
Define this macro to a C expression representing a variant of the
9829
method call @var{mdecl}, if Java Native Interface (JNI) methods
9830
must be invoked differently from other methods on your target.
9831
For example, on 32-bit Microsoft Windows, JNI methods must be invoked using
9832
the @code{stdcall} calling convention and this macro is then
9833
defined as this expression:
9834
 
9835
@smallexample
9836
build_type_attribute_variant (@var{mdecl},
9837
                              build_tree_list
9838
                              (get_identifier ("stdcall"),
9839
                               NULL))
9840
@end smallexample
9841
@end defmac
9842
 
9843
@deftypefn {Target Hook} bool TARGET_CANNOT_MODIFY_JUMPS_P (void)
9844
This target hook returns @code{true} past the point in which new jump
9845
instructions could be created.  On machines that require a register for
9846
every jump such as the SHmedia ISA of SH5, this point would typically be
9847
reload, so this target hook should be defined to a function such as:
9848
 
9849
@smallexample
9850
static bool
9851
cannot_modify_jumps_past_reload_p ()
9852
@{
9853
  return (reload_completed || reload_in_progress);
9854
@}
9855
@end smallexample
9856
@end deftypefn
9857
 
9858
@deftypefn {Target Hook} int TARGET_BRANCH_TARGET_REGISTER_CLASS (void)
9859
This target hook returns a register class for which branch target register
9860
optimizations should be applied.  All registers in this class should be
9861
usable interchangeably.  After reload, registers in this class will be
9862
re-allocated and loads will be hoisted out of loops and be subjected
9863
to inter-block scheduling.
9864
@end deftypefn
9865
 
9866
@deftypefn {Target Hook} bool TARGET_BRANCH_TARGET_REGISTER_CALLEE_SAVED (bool @var{after_prologue_epilogue_gen})
9867
Branch target register optimization will by default exclude callee-saved
9868
registers
9869
that are not already live during the current function; if this target hook
9870
returns true, they will be included.  The target code must than make sure
9871
that all target registers in the class returned by
9872
@samp{TARGET_BRANCH_TARGET_REGISTER_CLASS} that might need saving are
9873
saved.  @var{after_prologue_epilogue_gen} indicates if prologues and
9874
epilogues have already been generated.  Note, even if you only return
9875
true when @var{after_prologue_epilogue_gen} is false, you still are likely
9876
to have to make special provisions in @code{INITIAL_ELIMINATION_OFFSET}
9877
to reserve space for caller-saved target registers.
9878
@end deftypefn
9879
 
9880
@defmac POWI_MAX_MULTS
9881
If defined, this macro is interpreted as a signed integer C expression
9882
that specifies the maximum number of floating point multiplications
9883
that should be emitted when expanding exponentiation by an integer
9884
constant inline.  When this value is defined, exponentiation requiring
9885
more than this number of multiplications is implemented by calling the
9886
system library's @code{pow}, @code{powf} or @code{powl} routines.
9887
The default value places no upper bound on the multiplication count.
9888
@end defmac
9889
 
9890
@deftypefn Macro void TARGET_EXTRA_INCLUDES (const char *@var{sysroot}, const char *@var{iprefix}, int @var{stdinc})
9891
This target hook should register any extra include files for the
9892
target.  The parameter @var{stdinc} indicates if normal include files
9893
are present.  The parameter @var{sysroot} is the system root directory.
9894
The parameter @var{iprefix} is the prefix for the gcc directory.
9895
@end deftypefn
9896
 
9897
@deftypefn Macro void TARGET_EXTRA_PRE_INCLUDES (const char *@var{sysroot}, const char *@var{iprefix}, int @var{stdinc})
9898
This target hook should register any extra include files for the
9899
target before any standard headers.  The parameter @var{stdinc}
9900
indicates if normal include files are present.  The parameter
9901
@var{sysroot} is the system root directory.  The parameter
9902
@var{iprefix} is the prefix for the gcc directory.
9903
@end deftypefn
9904
 
9905
@deftypefn Macro void TARGET_OPTF (char *@var{path})
9906
This target hook should register special include paths for the target.
9907
The parameter @var{path} is the include to register.  On Darwin
9908
systems, this is used for Framework includes, which have semantics
9909
that are different from @option{-I}.
9910
@end deftypefn
9911
 
9912
@deftypefn {Target Hook} bool TARGET_USE_LOCAL_THUNK_ALIAS_P (tree @var{fndecl})
9913
This target hook returns @code{true} if it is safe to use a local alias
9914
for a virtual function @var{fndecl} when constructing thunks,
9915
@code{false} otherwise.  By default, the hook returns @code{true} for all
9916
functions, if a target supports aliases (i.e.@: defines
9917
@code{ASM_OUTPUT_DEF}), @code{false} otherwise,
9918
@end deftypefn
9919
 
9920
@defmac TARGET_FORMAT_TYPES
9921
If defined, this macro is the name of a global variable containing
9922
target-specific format checking information for the @option{-Wformat}
9923
option.  The default is to have no target-specific format checks.
9924
@end defmac
9925
 
9926
@defmac TARGET_N_FORMAT_TYPES
9927
If defined, this macro is the number of entries in
9928
@code{TARGET_FORMAT_TYPES}.
9929
@end defmac
9930
 
9931
@deftypefn {Target Hook} bool TARGET_RELAXED_ORDERING
9932
If set to @code{true}, means that the target's memory model does not
9933
guarantee that loads which do not depend on one another will access
9934
main memory in the order of the instruction stream; if ordering is
9935
important, an explicit memory barrier must be used.  This is true of
9936
many recent processors which implement a policy of ``relaxed,''
9937
``weak,'' or ``release'' memory consistency, such as Alpha, PowerPC,
9938
and ia64.  The default is @code{false}.
9939
@end deftypefn
9940
 
9941
@deftypefn {Target Hook} const char *TARGET_INVALID_ARG_FOR_UNPROTOTYPED_FN (tree @var{typelist}, tree @var{funcdecl}, tree @var{val})
9942
If defined, this macro returns the diagnostic message when it is
9943
illegal to pass argument @var{val} to function @var{funcdecl}
9944
with prototype @var{typelist}.
9945
@end deftypefn
9946
 
9947
@deftypefn {Target Hook} {const char *} TARGET_INVALID_CONVERSION (tree @var{fromtype}, tree @var{totype})
9948
If defined, this macro returns the diagnostic message when it is
9949
invalid to convert from @var{fromtype} to @var{totype}, or @code{NULL}
9950
if validity should be determined by the front end.
9951
@end deftypefn
9952
 
9953
@deftypefn {Target Hook} {const char *} TARGET_INVALID_UNARY_OP (int @var{op}, tree @var{type})
9954
If defined, this macro returns the diagnostic message when it is
9955
invalid to apply operation @var{op} (where unary plus is denoted by
9956
@code{CONVERT_EXPR}) to an operand of type @var{type}, or @code{NULL}
9957
if validity should be determined by the front end.
9958
@end deftypefn
9959
 
9960
@deftypefn {Target Hook} {const char *} TARGET_INVALID_BINARY_OP (int @var{op}, tree @var{type1}, tree @var{type2})
9961
If defined, this macro returns the diagnostic message when it is
9962
invalid to apply operation @var{op} to operands of types @var{type1}
9963
and @var{type2}, or @code{NULL} if validity should be determined by
9964
the front end.
9965
@end deftypefn
9966
 
9967
@defmac TARGET_USE_JCR_SECTION
9968
This macro determines whether to use the JCR section to register Java
9969
classes. By default, TARGET_USE_JCR_SECTION is defined to 1 if both
9970
SUPPORTS_WEAK and TARGET_HAVE_NAMED_SECTIONS are true, else 0.
9971
@end defmac
9972
 
9973
@defmac OBJC_JBLEN
9974
This macro determines the size of the objective C jump buffer for the
9975
NeXT runtime. By default, OBJC_JBLEN is defined to an innocuous value.
9976
@end defmac

powered by: WebSVN 2.1.0

© copyright 1999-2024 OpenCores.org, equivalent to Oliscience, all rights reserved. OpenCores®, registered trademark.